Research ArticleLARGE-SCALE BIOLOGY ARTICLE

Transcriptional Regulation of Fruit Ripening by Tomato FRUITFULL Homologs and Associated MADS Box Proteins

Masaki Fujisawa, Yoko Shima, Hiroyuki Nakagawa, Mamiko Kitagawa, Junji Kimbara, Toshitsugu Nakano, Takafumi Kasumi, Yasuhiro Ito

Published January 2014. DOI: https://doi.org/10.1105/tpc.113.119453

Abstract

The tomato (Solanum lycopersicum) MADS box FRUITFULL homologs FUL1 and FUL2 act as key ripening regulators and interact with the master regulator MADS box protein RIPENING INHIBITOR (RIN). Here, we report the large-scale identification of direct targets of FUL1 and FUL2 by transcriptome analysis of FUL1/FUL2 suppressed fruits and chromatin immunoprecipitation coupled with microarray analysis (ChIP-chip) targeting tomato gene promoters. The ChIP-chip and transcriptome analysis identified FUL1/FUL2 target genes that contain at least one genomic region bound by FUL1 or FUL2 (regions that occur mainly in their promoters) and exhibit FUL1/FUL2-dependent expression during ripening. These analyses identified 860 direct FUL1 targets and 878 direct FUL2 targets; this set of genes includes both direct targets of RIN and nontargets of RIN. Functional classification of the FUL1/FUL2 targets revealed that these FUL homologs function in many biological processes via the regulation of ripening-related gene expression, both in cooperation with and independent of RIN. Our in vitro assay showed that the FUL homologs, RIN, and tomato AGAMOUS-LIKE1 form DNA binding complexes, suggesting that tetramer complexes of these MADS box proteins are mainly responsible for the regulation of ripening.

INTRODUCTION

During ripening, fruits undergo biochemical and physiological changes in flesh texture, pigmentation, sugar contents, aroma, and nutritional qualities. Understanding the mechanisms that regulate fruit ripening can help efforts to improve fruit quality. The genetically programmed system that regulates ripening of fleshy fruits includes phytohormone signaling pathways and transcription factor networks, acting both cooperatively and independently (Giovannoni, 2007; Gapper et al., 2013). For example, ethylene promotes ripening in climacteric fruits such as tomato (Solanum lycopersicum), banana (Musa spp), and apple (Malus domestica).

In the transcription factor networks that affect fruit ripening, MADS box family transcription factors serve as central regulators of ripening (Smaczniak et al., 2012a; Gapper et al., 2013). MADS box transcription factors, which contain a conserved DNA binding domain (MADS domain), are widely present in eukaryotes. In plants, MADS box proteins (type II MIKCC-type) play an essential role in almost every developmental process, including fruit development and ripening (Smaczniak et al., 2012a). The tomato MADS box gene RIPENING INHIBITOR (RIN), a homolog of Arabidopsis thaliana SEPALLATA (SEP), is one of the earliest acting key factors that regulate ripening, including both ethylene-dependent and -independent processes (Vrebalov et al., 2002). The rin mutants exhibit a severe ripening-defective phenotype in which the fruits fail to soften, do not accumulate lycopene, and do not show a climacteric rise of respiration and ethylene production (Tigchelaar et al., 1978). Other climacteric fruit species, and even the nonclimacteric fruit strawberry (Fragaria × ananassa), have SEP homologs involved in fruit ripening (Elitzur et al., 2010; Seymour et al., 2011; Ireland et al., 2013; Schaffer et al., 2013).

In addition to RIN, the tomato MADS box genes AGAMOUS-LIKE1 (TAGL1), FRUITFULL1 (FUL1; formerly named TDR4), and FUL2 (formerly MBP7) act as fruit ripening regulators (Itkin et al., 2009; Vrebalov et al., 2009; Giménez et al., 2010; Pan et al., 2010; Bemer et al., 2012; Shima et al., 2014). TAGL1 and FUL1/FUL2 are homologs of Arabidopsis SHATTERPROOF (SHP) and FUL, respectively, and their suppression results in ripening-defective phenotypes partly similar to the phenotype of rin mutant fruits. Also, a bilberry (Vaccinium myrtillus) FUL homolog, Vm-TDR4 (Jaakola et al., 2010), and a strawberry SHP homolog, Fa-SHP (Daminato et al., 2013), are associated with fruit ripening. The identification of ripening-related MADS box proteins opens opportunities for the study of ripening regulation. However, it remains less clear how these MADS box factors coordinately regulate fruit ripening, compared with regulation of flower or dry-type fruit (e.g., Arabidopsis) development, where SEP, SHP, and FUL have well-characterized genetic and molecular functions (Gu et al., 1998; Liljegren et al., 2000; Pelaz et al., 2000; Honma and Goto, 2001; Kaufmann et al., 2009; Robles et al., 2012; Smaczniak et al., 2012b).

To elucidate the transcriptional regulation mechanism of fruit ripening, several studies have identified direct targets of the master regulator RIN; these targets include the ripening-regulating genes FUL1, NON-RIPENING (NOR; NAC family transcription factor), and COLORLESS NON-RIPENING (CNR; SQUAMOSA promoter binding protein-like family transcription factor) (Fujisawa et al., 2011, 2012; Martel et al., 2011; Qin et al., 2012). Large-scale analyses identified additional direct RIN target genes (Fujisawa et al., 2013; Zhong et al., 2013), revealing that RIN directly regulates a broad range of biological processes and metabolic pathways, including ripening-related ethylene production and lycopene accumulation during ripening. Also, the tomato FUL homologs, which interact with RIN to form DNA binding protein complexes, play a crucial role in fruit ripening (Leseberg et al., 2008; Martel et al., 2011; Bemer et al., 2012; Shima et al., 2013). The FUL homologs act upstream of the ethylene signaling pathway during ripening, as shown by a phenotypic analysis of FUL1/FUL2-suppressed tomato fruits (Shima et al., 2014).

In this study, to understand the biological roles of the FUL homologs, we identified direct FUL1 and FUL2 target genes by transcriptome analysis of FUL1/FUL2-suppressed fruits and chromatin immunoprecipitation coupled with microarray analysis (ChIP-chip). The combined ChIP-chip and transcriptome analyses successfully identified direct target genes of FUL1 and FUL2 on a large scale. The comparison of FUL1, FUL2, and RIN target genes suggests that the FUL homologs regulate transcription in both RIN-associated and RIN-independent manners. We also used an in vitro assay to examine the interactions of RIN, FUL1, FUL2, and TAGL1. Our results suggest that these ripening-related MADS box proteins form higher order complexes that are responsible for the main part of ripening regulation.

RESULTS

Transcriptome Profiling of FUL1/FUL2-Suppressed Fruits by RNA Sequencing

To identify downstream targets of FUL1 and FUL2, we first examined the alterations in transcription caused by knockdown of FUL1 and FUL2. We previously used RNA interference (RNAi) in transgenic tomato plants to cosuppress FUL1 and FUL2 (Shima et al., 2014). In this study, we used fruits of a strong FUL1/FUL2-suppressed transgenic line, TF18, which exhibits a ripening-defective phenotype with reduced ethylene production and lycopene accumulation (Shima et al., 2014), for transcriptome analysis.

To obtain a transcriptome profile of FUL1/FUL2-suppressed fruits, we conducted massively parallel RNA sequencing (RNA-Seq) analysis on TF18 fruits harvested at 35 d after pollination (DAP), the same age as the preripening (mature green) stage of the wild-type fruits, which we term “G age.” We also examined fruits at 45 DAP, the same age as the ripening (pink coloring) stage of the wild-type fruits, which we term “P age.” For transcriptome comparison, we also analyzed fruits of wild type (Ailsa Craig cultivar [AC]) and a rin mutant. We validated the RNA-Seq data (Supplemental Table 1 and Supplemental Data Set 1) by quantitative RT-PCR (qRT-PCR) for nine known ripening-related genes, which showed good correlation between the two assays (Supplemental Figure 1). The RNA-Seq data showed that the upregulation of many well-characterized ripening-induced genes during wild-type ripening (represented by the lane AC_P/AC_G in Figure 1A) was suppressed in the P age TF18 fruits (TF18_P/AC_P in Figure 1A). The suppression was stronger than in P age rin mutant fruits (rin_P/AC_P in Figure 1A). Multidimensional scaling (MDS) analysis of the RNA-Seq data showed a similar tendency, placing the TF18 samples (at G and P ages) on the map far away from the AC and rin samples, which were placed near each other (Figure 1B). These results indicate that the FUL1/FUL2 suppression affects gene expression (both at the G and P ages) more strongly than the rin mutation.

Figure 1.
Figure 1.

Transcriptome Profiling of the FUL1/FUL2-Suppressed Tomato Fruits.

(A) Expression of representative ripening-related genes in the wild type (AC), the rin mutant, and the FUL1/FUL2-suppressed line (TF18) fruits. The samples harvested at 35 and 45 DAP are represented by G and P, respectively (see text). Sample pairs compared are shown above the heat map. ACO1 and ACO6, genes for ACC oxidase; ACS2 and ACS4, genes for ACC synthase; AP2a, an APETALA2 homolog gene; Cel2, a gene for endo-β-1,4-glucanase; CRTISO, a gene for carotenoid isomerase; E4, a gene for Met sulfoxide reductase; E8 and E8-6, genes for ACC oxidase homologs; EXP1, a gene for expansin; GGPS2, a gene for geranylgeranyl diphosphate synthase; HB-1, a gene for HD-zip homeobox protein; ISPE, a gene for 4-diphosphocytidyl-2-C-methyl-d-erythritol kinase; LoxC, a gene for lipoxygenase; MAN4, a gene for mannan endo-1,4-β-mannosidase; NP24, a gene for osmotin-like protein; NR, NEVER RIPE; PSY1, a gene for phytoene synthase; PG2A, a gene for polygalacturonase; SGR1, a gene for stay-green protein; TBG4, a gene for β-galactosidase; TIV1, a gene for acid β-fructofuranosidase (invertase); XYL1, a gene for β-d-xylosidase; Z-ISO, a gene for ζ-carotene isomerase.

(B) MDS analysis of RNA-Seq data from the fruits of AC, rin mutants, and the FUL1/FUL2-suppressed line (TF18). Each spot represents an individual sample, with three biological replicates, indicated by the name of each line, and followed by the age harvested (G or P).

Detection and Functional Classification of FUL1/FUL2-Regulated Genes

To identify FUL1/FUL2-regulated genes and clarify the role of the FUL homologs during ripening, we identified differentially expressed genes (DEGs) between samples using the RNA-Seq data. To do this, we applied the cutoff value of fold change (FC) ratio >2 or <0.5 with false discovery rate (FDR) ≤0.05 for genes that are substantially expressed in either or both samples (average counts per million reads ≥ 5). Comparison of the P age AC and TF18 fruits identified 5953 DEGs, including 2856 upregulated and 3097 downregulated by FUL1/FUL2 suppression (DEGs for FUL; Supplemental Data Set 2A). Similarly, comparison of the P and G age AC fruits identified 5978 DEGs, including 3003 upregulated and 2975 downregulated with ripening (DEGs for ripening; Supplemental Data Set 2B). Comparison of the P age AC and rin fruits identified 3821 DEGs, including 2399 upregulated and 1422 downregulated by the rin mutation (DEGs for rin; Supplemental Data Set 2C). The set of DEGs for FUL included 4809 (80%) of the DEGs for ripening, but the set of DEGs for rin included markedly fewer DEGs for ripening (1927 genes; 32%) (Supplemental Figure 2).

To provide an overview of the roles of the FUL homologs, we classified the DEGs for FUL based on the functions of their Arabidopsis homologs, as predicted by the Munich Information Center for Protein Sequences (MIPS) functional catalog database (Supplemental Data Set 2A). We found categories and subcategories (e.g., protein synthesis and differentiation of organ, tissue, and cell type) in which the frequency of genes in the up- and downregulated subsets of DEGs for FUL is significantly overrepresented (P < 0.001 by Fisher’s exact test) compared with the whole genome (Supplemental Data Set 3). Both the up- and downregulated subsets of DEGs for FUL had more overrepresented categories than the up- and downregulated subsets of DEGs for rin, respectively (Supplemental Data Set 3). Typical ripening-associated categories such as ethylene production (metabolism of Met) and carotenoid metabolism (isoprenoid metabolism and tetraterpene metabolism) were overrepresented in the downregulated subset of DEGs for rin but not in that of the DEGs for FUL, although many genes in these categories were included in the subset of DEGs for FUL (Supplemental Data Sets 2C and 3). The results suggest that the FUL homologs regulate expression of more diverse genes than RIN, which mainly regulates genes involved in ripening.

Detection of FUL1- and FUL2-Bound Genomic Regions by ChIP-chip

To detect the genomic regions bound by FUL1 and FUL2 in tomato promoters during ripening, we performed a ChIP-chip analysis. We used ChIP to collect DNA fragments bound to FUL1 and FUL2 from ripening tomato fruits at the pink coloring stage using antibodies to FUL1 and FUL2. The ChIPed and input (no ChIP) DNAs were hybridized to the NimbleGen microarrays, which contain probes designed from 2-kb promoter regions of all predicted tomato genes (ITAG2) and were previously used to identify direct RIN targets (Fujisawa et al., 2013). Hybridization signals were analyzed to assign each probe a log2-scale FC value of the ChIPed DNA relative to the input DNA. Consecutive probes with high average FC from three biological replicates at statistically significant levels (FDR ≤ 0.05; for more details, see Fujisawa et al., 2013) defined ChIP-chip peak regions (Supplemental Table 2). This analysis identified 2307 FUL1-bound regions and 2457 FUL2-bound regions (Figure 2; Supplemental Data Sets 4A and 4B). We validated the microarray data by quantitative ChIP-PCR (qChIP-PCR) for the mapped regions with the 10 highest scores each for FUL1 and FUL2. The qChIP-PCR confirmed >4-fold enrichment of all the regions analyzed, although we found no significant correlation between the ChIP-chip scores and the qChIP-PCR enrichment levels (Supplemental Figure 3), possibly due to the difference in dynamic range between the two assays.

Figure 2.
Figure 2.

Distribution of FUL1- and FUL2-Bound Regions on the Tomato Chromosomes.

Genomic positions of binding regions of FUL1 and FUL2 detected by ChIP-chip on the 12 chromosomes are indicated by bars in magenta and violet, respectively, with the log2 scale peak score. For comparison, the positions of RIN-bound regions identified in this study are indicated in red. Positions of the promoters (+, forward strand; −, complementary strand) where ChIP-chip probes were designed are indicated by blue bars.

Overlap and Motif Conservation of DNA Binding Regions of FUL1, FUL2, and RIN

We previously reported that in several ripening-related gene promoters, the binding regions of FUL1 and FUL2 overlap with each other and with RIN-bound regions (Shima et al., 2013). To examine this in more loci, we analyzed the overlaps of FUL1- and FUL2-bound regions and RIN-bound regions (1217 regions) that we identified by reanalysis of our previously examined ChIP-chip data (Fujisawa et al., 2013) with the same algorithm we used to identify the genomic regions bound by the FUL homologs. We defined an overlapping region as one factor’s bound region that, at least in part, is detected as the binding region of another factor(s). As summarized in Supplemental Table 3, 1145 (50%) of FUL1-bound and 1134 (46%) of FUL2-bound regions overlapped, at least in part, with those of RIN. Furthermore, 1810 (78%) of FUL1-bound regions and 1803 (73%) of FUL2-bound regions overlapped with each other. By contrast, 435 (19%) of FUL1-bound, 604 (25%) of FUL2-bound, and 39 (3.2%) of RIN-bound regions did not overlap with the other transcription factor bound regions. To verify that FUL1, FUL2, and RIN independently bind these nonoverlapping regions, we compared these regions with the regions detected by each of the three replicate ChIP-chip assays for the other factors. The results showed that 58 to 68% for FUL1, 55 to 70% for FUL2, and 36% for RIN of their nonoverlapping regions were not found in the other factors' bound regions in any replications, confirming that at least these regions are most likely unique to each factor (Supplemental Table 4). By contrast, considering the variation in ChIP-chip between replicates (Supplemental Table 2), the remaining regions might include undetected common regions (Supplemental Table 4).

To clarify DNA binding specificities of FUL1, FUL2, and RIN, we compared the nonoverlapping regions of each transcription factor based on the frequency of four types of CArG motif sequences: PAL type [CC(A/T)6GG], N10 type [CTA(A/T)4TAG] (West et al., 1998), and typical RIN binding motifs [C(C/T)(A/T)6(A/G)G (Ito et al., 2008) and C(A/T)8G (Fujisawa et al., 2011)] in the bound regions. The frequency of PAL and N10 motifs was 0.05 to 0.23 motif/region and 0 to 0.11 motif/region, respectively, whereas the frequency of typical RIN binding motifs was 0.72 to 1.59 motif/region (Supplemental Table 5). The difference in the frequencies between PAL/N10 and typical RIN binding motifs mostly depended on the stringency of the motif sequences. Intriguingly, the frequency of PAL sequences in the FUL2-unique regions (0.23 motif/region) was relatively higher than those in the RIN- and FUL1-unique regions (0.05 to 0.14 motif/region) compared with the other types of CArG-box motifs (Supplemental Table 5). We also searched the nonoverlapping regions for conserved sequence motifs using MEME software (Bailey et al., 2006), but we did not detect any motifs.

Large-Scale Identification of Direct FUL1 and FUL2 Target Genes

Using the genomic positions of the 2307 FUL1-bound and 2457 FUL2-bound regions, we found 2454 and 2701 genes as potential direct FUL1 and FUL2 target genes, respectively. These genes have one or more bound regions in the 2-kb promoter region or in other gene regions (exons, introns, or the 1-kb region downstream of the translation termination site) overlapping with the promoter region of a neighboring gene. To identify direct FUL1 and FUL2 target genes whose expression is actually regulated by FUL1 and FUL2 during ripening, we compared the sets of potential targets with the set of FUL1/FUL2-regulated genes (DEGs for FUL) identified by the RNA-Seq analysis described above. We defined the potential targets included in the DEG set as direct FUL1 or FUL2 targets. These analyses identified 860 direct targets of FUL1 (567 positively regulated and 293 negatively regulated; Figure 3A; Supplemental Data Set 5A) and 878 direct targets of FUL2 (555 positively regulated and 323 negatively regulated; Figure 3B; Supplemental Data Set 5B). The sets of direct FUL1 and FUL2 targets shared 697 genes in common (81% for FUL1 and 79% for FUL2; areas E and G in Figure 4). Using the RNA-Seq data for rin mutant fruits and the reanalyzed ChIP-chip data for RIN (described above), we also identified 262 direct RIN target genes, including 162 genes that were not found in the previous study (Fujisawa et al., 2013) (Supplemental Data Set 5C). Of the direct FUL1 and FUL2 targets, 215 (25%; areas D and G in Figure 4) and 210 (24%; areas F and G in Figure 4), respectively, were also direct RIN targets. Of the direct RIN targets, 217 (83%) were also direct targets of FUL1, FUL2, or both. All the sets of direct FUL1, FUL2, and RIN targets shared 208 genes (area G in Figure 4). Of the FUL1 and FUL2 target genes, 645 (75%; areas B and E in Figure 4) and 668 (76%; areas C and E in Figure 4), respectively, were identified as not belonging to the set of direct RIN target genes. Finally, we identified unique targets of each transcription factor and found 156 (18%; area B in Figure 4) genes for FUL1, 179 (20%; area C in Figure 4) genes for FUL2, and 45 (17%; area A in Figure 4) genes for RIN (Supplemental Data Sets 5A to 5C).

Figure 3.
Figure 3.

Identification of Direct Targets of FUL1 and FUL2 during Ripening.

Venn diagrams of potential direct target genes of FUL1 (A) and FUL2 (B) and DEGs positively and negatively regulated by FUL1 or FUL2 during ripening. DEGs positively and negatively regulated by the FUL homologs were significantly (FDR ≤ 0.05) downregulated (FC < 0.5) or upregulated (FC > 2) in the FUL1/FUL2-suppressed fruits harvested at 45 DAP compared with the same aged wild-type ripening fruits (Shima et al., 2014).

Figure 4.
Figure 4.

A Venn Diagram of Direct Targets of RIN, FUL1, and FUL2.

Of the DEGs for rin or FUL1/FUL2, genes that contain one or more binding regions in the promoters or other gene regions were identified as direct targets for the respective transcription factors (see text). The set of direct RIN targets includes the genes identified by reanalysis in this study.

To test the significance of the overlap between potential direct target genes based on ChIP-chip and DEGs based on RNA-Seq, we estimated the number of overlapping genes that we would expect to occur by chance, using the formula: Eg = Po × De/Ge, where Eg is the expected number of genes overlapping by chance, Po is the number of potential targets, De is the number of DEGs, and Ge is the number of all genes in the genome. For FUL1, we identified significantly more direct target genes (860) than expected by chance (Eg = 408, P < 0.001 by χ2 test). Similarly, for FUL2 and RIN, we also identified more direct target genes (878 and 262) than we expected by chance (Eg = 449, P < 0.001 for FUL2; Eg = 152, P < 0.001 for RIN).

Detection of Known Ripening-Related Genes among Direct FUL1 and FUL2 Targets

The combined ChIP-chip and RNA-Seq analysis showed that the FUL1 and FUL2 direct targets include known ripening-related direct RIN target genes, such as genes for 1-aminocyclopropane-1-carboxylic acid (ACC) synthase (ACS2 and ACS4), for an ACC oxidase homolog (E8), for APETALA2a, NOR, and RIN (Figure 5; Supplemental Data Sets 5A and 5B). The bound regions of FUL1, FUL2, and RIN overlapped with each other in the promoters of many ripening-related target genes (Figure 5). The binding of FUL1, FUL2, and RIN to the same regions strongly suggests that they regulate these target genes as a complex. However, this simple model did not apply to TAGL1, which has overlapping FUL1-, FUL2-, and RIN-bound regions in the promoter and was included in the sets of positive direct FUL1 and FUL2 targets, but not in the set of direct RIN targets (Figure 5; Supplemental Data Sets 5A to 5C) because the rin mutation showed little effect on TAGL1 expression.

Figure 5.
Figure 5.

Examples of FUL1- and FUL2-Bound Regions Detected by ChIP-chip in the Promoters of Ripening-Related Genes.

For comparison, RIN-bound regions identified in this study are also indicated. Genomic position and log2-scale peak score of each bound region is indicated above the 2-kb gene promoters (horizontal bars). Boxed arrows with gene symbols indicate the orientation of genes. Thin vertical lines indicate the positions of CArG-boxes in the promoter. Abbreviations are given in Figure 1.

By contrast, a few ripening-associated genes had bound regions for only one of the transcription factors. For example, the promoters of genes for HD-zip homeobox protein 1 and for geranylgeranyl diphosphate synthase contain only a FUL1-bound region; genes for 4-diphosphocytidyl-2-C-methyl-d-erythritol kinase and for 2-C-methyl-d-erythritol 2,4-cyclodiphosphate synthase contain only a RIN-bound region (Figure 5; Supplemental Data Sets 5A to 5C). We did not find any ripening-associated genes that contain only a FUL2-bound region. A gene for phytoene synthase (PSY1), a direct target of RIN (Martel et al., 2011; Zhong et al., 2013) and FUL1 (Shima et al., 2013), contains regions bound by FUL1 and RIN, but not FUL2, in the promoter. Binding regions both of FUL1 and FUL2, but not RIN, were detected in the promoters of FUL1, a gene for acid-β-fructofuranosidase (invertase; TIV1) and the genes for carotene β-hydroxylase (BCH) and 9-cis-epoxycarotenoid dioxygenase (NCED), which act in carotene conversion and in cleavage of 9-cis xanthophylls, respectively (Figure 5; Supplemental Data Sets 5A to 5C; as described below). These results indicated that these genes are regulated by FUL homologs in a RIN-independent manner.

By contrast, direct RIN targets, such as genes for polygalacturonase (PG2A), for Met sulfoxide reductase (E4) and CNR, were not included in the sets of direct FUL1 and FUL2 targets because they showed either insufficiently significant changes in expression or no detected bound region, as described previously (Fujisawa et al., 2013).

Functional Classification of Direct FUL1 and FUL2 Targets

To elucidate biological processes and pathways that FUL1 and FUL2 directly regulate during ripening, we used the MIPS classification profiles to classify the functions of direct FUL1 and FUL2 target genes. FUL1 and FUL2 targets are implicated in a wide range of biological processes, including metabolic and signaling pathways essential for fruit ripening (Supplemental Data Sets 6A and 6B). Supplemental Table 6 shows overrepresented categories with a significantly higher frequency of genes (P < 0.001 by Fisher’s exact test) in the set of direct targets than in the whole genome. Of the overrepresented categories in the sets of positive direct FUL1 and FUL2 targets, categories related to ethylene synthesis such as “metabolism of methionine” and related to responses to various external stimuli were also enriched in the set of positive direct RIN targets (Supplemental Table 6 and Supplemental Data Set 6C). Similarly, a category related to lycopene accumulation such as “tetraterpenes (carotenoids) metabolism” was also enriched in the sets of positive direct FUL1, FUL2, and RIN targets. However, we found a difference in the distribution of their targets in the carotenoid pathway and its upstream pathways. We detected direct FUL1 targets in the pathway downstream from GGDP synthesis and direct FUL2 targets in the pathway downstream of phytoene, but direct RIN targets were limited to the pathway upstream of carotenes except for the gene encoding zeaxanthin epoxidase (ZEP) (Figure 6). Together, the results suggest that FUL1 and FUL2 regulate genes involved in ethylene production or response to external stimuli in collaboration with RIN, but they regulate carotenoid metabolism partly independently of RIN.

Figure 6.
Figure 6.

Direct Target Genes of RIN, FUL1, and FUL2 in Carotenoid Synthesis and Metabolism.

Only enzymes encoded by the direct target genes are shown, with gene identifiers and tags for RIN-, FUL1-, and FUL2-targeting profiles. Gene identifiers (Solyc numbers) in red and blue indicate positively and negatively regulated genes during ripening, respectively. ABA, abscisic acid; BCH, carotene β-hydroxylase; CRTISO, carotene isomerase; CrtL-b, β-lycopene cyclase; GGDP, geranylgeranyl diphosphate; GGPS, geranylgeranyl diphosphate synthase; ISPE, 4-diphosphocytidyl-2-C-methyl-d-erythritol kinase; ISPF, 2-C-methyl-d-erythritol 2,4-cyclodiphosphate synthase; MEP, methylerythritol phosphate; NCED, 9-cis-epoxycarotenoid dioxygenase; PSY, phytoene synthase; ZEP, zeaxanthin epoxidase.

Interaction of FUL1, FUL2, RIN, and TAGL1

FUL1 and FUL2 can form heterodimers with RIN (Leseberg et al., 2008; Bemer et al., 2012; Shima et al., 2013). This is consistent with our ChIP-chip results showing that RIN and FUL homologs share binding regions on the promoters of many genes, indicating that they regulate common targets as a complex. In addition, a yeast two-hybrid screen detected interaction of RIN with a MADS box ripening regulator, TAGL1 (Martel et al., 2011). FUL1/FUL2 suppression and TAGL1 suppression resulted in similar phenotypes, including defective ripening with fruit softening (firmness was reduced similar to the untransformed control) (Itkin et al., 2009; Vrebalov et al., 2009; Giménez et al., 2010; Bemer et al., 2012; Shima et al., 2014), suggesting a functional interaction between the FUL homologs and TAGL1. In this study, to clarify complex formation of ripening-related MADS box proteins in more detail, we examined the interaction of RIN and FUL homologs with TAGL1 by in vitro gel retardation assay using CArG-box-containing DNA fragments as probes.

The gel retardation assay showed that TAGL1 was able to form DNA binding complexes with RIN and FUL2 (Figure 7A). By contrast, the formation of a DNA binding complex of FUL1 with TAGL1 was not confirmed (Figure 7A). However, if RIN was present in the synthesis mixture with TAGL1 and FUL1, the gel retardation assay detected two retarded bands (Figure 7B, lane 3): a lower unknown band in addition to the upper band at the same size as RIN-TAGL1 (Figure 7B, lane 2). When the amount of FUL1-expressing vector was decreased in the in vitro protein synthesis reaction, the signal intensity of the lower band decreased, but the intensity of the upper band increased (Figure 7C). Taking these observations together, we consider that the upper and lower bands represent DNA binding of RIN-TAGL1 and FUL1-RIN-TAGL1 complexes, respectively.

Figure 7.
Figure 7.

Gel Retardation Assays for DNA Binding Activity of the MADS Box Proteins Regulating Fruit Ripening.

(A) DNA binding complex formation between RIN, FUL1, FUL2, and TAGL1. For probe DNAs, two types of CArG motif sequences (N10 and PAL; West et al., 1998) were used.

(B) Higher order DNA binding complex formation by RIN, FUL1, FUL2, and TAGL1.

(C) Verification of the participation of FUL1 in complex formation with RIN and TAGL1. For each protein synthesis reaction, the FUL1 expression vector was added in decreasing amounts (for the lane 2 experiment, two-thirds of the vector in lane 1, for lane 3, one-third, and for lane 4, no vector), while the plasmids for RIN and TAGL1 were added equally in each experiment. In the experiments in (B) and (C), N10 was used as the probe.

FUL2, like FUL1, formed a complex with RIN and TAGL1, namely, FUL2-RIN-TAGL1 (Figure 7B, lane 5). In addition, unlike FUL1, the assay detected DNA binding of FUL2 homodimer and FUL2-TAGL1 heterodimer with a difference in DNA sequence preference: FUL2 homodimer could bind to the N10-type CArG-box but not to the PAL type, whereas FUL2-TAGL1 could bind to both types (Figure 7A). Moreover, FUL2-TAGL1 predominantly formed in the presence of FUL2 and TAGL1 rather than FUL2 homodimer, implying that the FUL2-TAGL1 heterodimer is the main form of the functional complex when they act in a RIN-independent manner.

FUL Homologs Affect Flavonoid Accumulation in Peel

A previous study on fleshy fruits found the association of a bilberry FUL homolog with flavonoid biosynthesis (Jaakola et al., 2010). To examine the association of tomato FUL homologs with flavonoid biosynthesis, we measured the accumulation of naringenin chalcone, a major flavonoid in tomato (Muir et al., 2001), in the peel of the fruits of three FUL1/FUL2-suppressed lines (TF2, TF18, and TF68) (Shima et al., 2014). We found that at P age, the fruits of all three lines had markedly less naringenin chalcone than the wild-type fruits, although the rin mutant fruits accumulated naringenin chalcone at similar levels to the wild-type fruits (Supplemental Figure 4A). Furthermore, to examine the effect of FUL1/FUL2 suppression on gene expression, we compared RNA-Seq data among the fruits of the wild type, TF18, and rin mutants, for genes involved in the flavonoid pathway (Supplemental Figure 4B). The analysis showed that FUL1/FUL2 suppression caused changes in expression of genes related to the flavonoid pathway compared with the wild type and rin mutants (Supplemental Figure 4C). Both of the positive sets of direct FUL1 and FUL2 targets included Solyc06g082540, which encodes cinnamate 4-hydroxylase, which is required for naringenin chalcone synthesis. By contrast, both of the negative sets of direct FUL1 and FUL2 targets included Solyc05g010320, which encodes a chalcone-flavonone isomerase that metabolizes naringenin chalcone (Supplemental Data Sets 5A to 5C). These two genes were not included in the set of direct RIN targets. These results suggest that the FUL homologs affect flavonoid synthesis via a RIN-independent regulatory mechanism.

DISCUSSION

Several types of transcription factors regulate tomato fruit ripening, and the MADS box factors RIN and TAGL1 act upstream of ethylene-dependent and -independent pathways. Recent studies cosuppressing FUL1 and FUL2 revealed that the FUL homologs also act as key regulators of many ripening processes, including lycopene accumulation (Bemer et al., 2012; Shima et al., 2014) and ripening-associated ethylene production (Shima et al., 2014). In this study, we identified hundreds of direct FUL1 and FUL2 target genes expressed during ripening and confirmed the formation of higher order complexes with FUL1, FUL2, RIN, and TAGL1, revealing a mode of regulation by these MADS box factors.

FUL Homologs Have More Diverse Roles in Ripening Than RIN

In this study, using RNA-Seq, we demonstrated the effect of FUL1/FUL2 suppression on gene expression in ripening tomato fruits, drastically expanding our knowledge of the role of these FUL homologs. The RNA-Seq analysis detected 5953 DEGs in the FUL1/FUL2 suppressed fruits (Shima et al., 2014), many more than a previous FUL RNAi analysis, which identified 895 genes (Bemer et al., 2012). This difference may be due to our RNAi construct, which may have suppressed FUL1 and FUL2 expression more effectively than those in the previous report (Shima et al., 2014) or the difference in cultivar (AC versus Micro-Tom) or analytical design between the two studies (RNA-Seq analysis versus microarray). The strong suppression of FUL1/FUL2 may reduce the expression of their target genes enough to decrease ripening-related ethylene production. Surprisingly, the analysis indicated that the FUL homologs affect a larger number (>3000) of positively or negatively regulated genes that change during ripening (i.e., DEGs for ripening) than RIN did (Supplemental Figure 2). This and the results of MIPS classification suggest that the FUL homologs regulate a wider range of biological processes during ripening than RIN.

FUL1 and FUL2 Have Redundant and Independent Functions and Have Functions in Common with and Distinct from RIN

Our combined ChIP-chip and RNA-Seq analysis revealed that the sets of direct FUL1 and FUL2 targets include positively and negatively regulated genes, indicating the function of the FUL homologs as activators and repressors, similar to RIN (Fujisawa et al., 2013). The analysis also revealed that FUL1 and FUL2 have 697 direct target genes in common, which represent the majority (79 to 81%) of direct FUL1 and FUL2 targets. The subset of common FUL1 and FUL2 targets included many key ripening-related genes, supporting the idea of redundant functions of FUL1 and FUL2 in the regulation of ripening, as previously described (Bemer et al., 2012; Shima et al., 2014). The FUL1 transcript was more abundant in wild type ripening than the FUL2 transcript (P age AC fruits; Supplemental Table 7); therefore, FUL1 may be more responsible for the redundant FUL activity during ripening than FUL2. The redundant FUL function may positively regulate the expression of TAGL1, a SHP homolog. This seems opposite to Arabidopsis carpel development, where FUL negatively regulates SHP (Ferrándiz et al., 2000). The existence of opposite regulatory pathways in tomato and Arabidopsis might reflect the distinct evolution of fleshy type fruits from dry type fruits.

FUL1 and FUL2 also have nonoverlapping target genes and bound regions, suggesting that each FUL homolog has additional independent functions; for example, “transcriptional control” was enriched only in the FUL1 target set (Supplemental Table 6). Although it is unclear to what extent these unique functions contribute to ripening, these results help clarify functional divergence between FUL1 and FUL2.

FUL1 and FUL2 may function together with and independently of RIN, as found in the carotenoid and flavonoid pathways. Of the targets of the FUL homologs in the carotenoid pathway, the genes encoding phytoene synthase 1 (PSY1), ζ-carotene isomerase, carotenoid isomerase, and β-lycopene cyclase are also direct RIN targets, but the genes involved in carotene conversion and its downstream processes, such as BCH and NCED, are probably not direct RIN targets (Figure 6). This clearly indicates that the FUL homologs regulate the overall pathway, but RIN is specialized to regulate lycopene accumulation, as previously described (Fujisawa et al., 2013). These findings show the divergent roles of RIN and the FUL homologs during ripening. We also clearly show that FUL homologs, but not RIN, regulate the level of a flavonoid, naringenin chalcone. This implies that the functions of FUL homologs in flavonoid biosynthesis are conserved between bilberry (Jaakola et al., 2010) and tomato.

Higher Order Complexes of MADS Box Proteins in Ripening Regulation

Our analysis revealed that 217 genes, including many key ripening-related genes, are direct targets of RIN and either FUL1 or FUL2, or both. Based on this and our previous analyses (Shima et al., 2013), we consider that RIN plays its principal role (i.e., fruit ripening regulation) as part of a DNA binding complex with the FUL homologs. Moreover, this study confirmed the RIN-TAGL1 heterodimer formation, consistent with previous studies using yeast two-hybrid assays (Leseberg et al., 2008; Martel et al., 2011; Bemer et al., 2012; Shima et al., 2013) and higher order complex formation by RIN, FUL1/FUL2 and TAGL1. The higher order complexes may be tetramers formed by coupling RIN-TAGL1 with RIN-FUL1 or RIN-FUL2 (Supplemental Figure 5). In the tetramer, FUL1 and FUL2 can be substituted for each other, consistent with previous reports of little or no effect of single FUL1 knockdown on fruit ripening (Jaakola et al., 2010; Bemer et al., 2012). We assume that the tetramers are mainly responsible for ripening regulation. In contrast with the common direct targets of FUL homologs and RIN, transcriptional regulation of FUL1/FUL2-independent direct RIN targets might be accomplished by RIN homodimers (Ito et al., 2008) or complexes of RIN and other factors such as TAGL1. We assume that RIN and TAGL1 form a tetramer of two RIN-TAGL1 heterodimers (Supplemental Figure 5), as observed in Arabidopsis where SEP3 and AGAMOUS (AG), homologs of RIN and TAGL1, respectively, form a tetramer (Smaczniak et al., 2012b).

The possible tetramer combinations may be analogous to the floral quartet model, a refinement of the classic ABC model for the Arabidopsis MADS domain proteins that regulate flower organ specification. In the quartet model, floral organs are specified by tetramers composed of two dimers with organ-specific combinations of four classes of MADS domain proteins (Theissen and Saedler, 2001; Smaczniak et al., 2012b). Tetramer formation by RIN, FUL1/FUL2, and TAGL1 can explain the similar ripening defects, such as inhibition of ethylene and carotenoid biosynthesis, commonly observed in the rin mutant (Tigchelaar et al., 1978; Vrebalov et al., 2002), the TAGL1 knockdown lines (Itkin et al., 2009; Vrebalov et al., 2009; Giménez et al., 2010), and the FUL1/FUL2-suppressed fruits (Bemer et al., 2012; Shima et al., 2014). Furthermore, tetramer formation can also explain the observation that RIN binding to target sites occurs in a CNR-dependent manner (Martel et al., 2011) because the reduced expression of FUL1 (TDR4) in Cnr mutant fruits (Eriksson et al., 2004) may affect tetramer formation. In MADS box complex formation, SEP proteins may function as the glue that mediates the assembly of MADS box protein complexes (Immink et al., 2009). Our results suggest that RIN, which belongs to the SEP subfamily, also conserves the assembly function to form ternary complexes (Supplemental Figure 5) similar to those assembled by Arabidopsis SEP3, such as FUL-SEP-SEEDSTICK (AG family) and SHP (AG family)-SEP-AG (Immink et al., 2009).

We also found that 75 to 76% of the direct FUL1 and FUL2 targets are not direct RIN targets, suggesting that the FUL homologs regulate the expression of these targets in a RIN-independent manner. Moreover, 18 to 20% of direct targets were RIN independent and exclusively targets of either FUL1 or FUL2. This means that each factor has a unique regulatory function. We consider that a FUL2-TAGL1 heterodimer may be a plausible component of the regulation of unique FUL2 targets. As shown by our gel retardation assay, FUL2 expands its DNA binding ability by interacting with TAGL1, which may allow FUL2 to recognize both N10 and PAL-type CArG-boxes, the latter of which are enriched in the FUL2 unique bound regions in the tomato genome (Supplemental Table 5). However, TAGL1 seems not to interact with FUL1 to form a DNA binding dimer complex and FUL1 is unable to form a DNA binding homodimer (Figure 7A) or heterodimer with FUL2 (Supplemental Figure 6). Thus, FUL1 may interact with unidentified partner(s) for the regulation of its unique direct targets (Supplemental Figure 5). Other possible candidates for FUL1 partners include an AG homolog (TAG1) and two SEP homologs (TM5 and TM29), based on their strong expression in ripening fruits (Supplemental Table 7). So far, however, the involvement of these candidates in fruit ripening remains unclear (Pnueli et al., 1994a, 1994b; Ampomah-Dwamena et al., 2002). Identification of a FUL1 partner will improve our understanding of transcriptional regulation of ripening and also refine the quartet model for fruit development.

METHODS

RNA-Seq

Total RNA samples were prepared from fruits of the wild type (AC cultivar), a FUL1/FUL2-suppressed line (TF18, generated by transformation of AC) (Shima et al., 2014), and rin mutants (a nearly isogenic line LA3754, obtained from the Tomato Genetics Resource Center [University of California, Davis]) of tomato (Solanum lycopersicum) at 35 and 45 DAP by SDS-phenol extraction and subsequent elimination of polysaccharides with 2-butoxyethanol (Schultz et al., 1994). The RNAs were further purified with the RNeasy Plus Kit (Qiagen). RNA quality was confirmed using a 2100 Bioanalyzer (Agilent). The RNA samples were analyzed with a HiSequation 2000 sequencer (Illumina) according to the manufacturer’s instructions. qRT-PCR for verification of the RNA-Seq data was performed using Thunderbird SYBR qPCR mix (Toyobo) with a 7300 Real-Time PCR system (Applied Biosystems) as previously described (Shima et al., 2014). Primer sequences for qRT-PCR are listed in Supplemental Table 8. Three biological replicates for each stage of the lines were analyzed. MDS analysis was conducted by calculating the coefficient of variation of expression between samples for the top 500 genes that best distinguish the samples. The coefficient was calculated from normalized RNA-Seq data of FUL1/FUL2-suppressed (TF18), wild-type (AC), and rin mutant fruits at G and P ages (three biological replicates for each stage) using the R Bioconductor package edgeR (http://www.bioconductor.org/packages/2.12/bioc/html/edgeR.html) (Robinson et al., 2010; Robinson and Oshlack, 2010). Distances on the plot represent the coefficient values.

Detection of DEGs

Single-end RNA-Seq reads (∼100 nucleotides) that passed through a filter for low-quality sequence were aligned using the Bowtie 2 program (Langmead and Salzberg, 2012) with protein coding sequences of the tomato predicted genes provided by the International Tomato Annotation Group version 2 (ITAG2; http://solgenomics.net/genomes/Solanum_lycopersicum/index.pl) (Tomato Genome Consortium, 2012). We used Bowtie 2 with the settings that allow 0 mismatches in a seed alignment to be adopted and that allow only uniquely mapped reads to be counted. Raw read counts for the genes were normalized and used for detection of DEGs between samples using the R Bioconductor package edgeR (Robinson et al., 2010; Robinson and Oshlack, 2010). The significance of FC ratio of DEGs was tested using P value from three independent samples estimated by exact test. P values were adjusted for multiple testing with the Benjamini and Hochberg method to control FDR (Benjamini and Hochberg, 1995). DEGs were selected using the following criteria: average counts per million ≥5 among the samples compared, and FC >2 or <0.5 with FDR ≤ 0.05. The raw read counts were also normalized using the reads per kilobase of protein coding sequence per million mapped reads method (Mortazavi et al., 2008) to express relative transcript levels within each sample.

ChIP-chip

ChIP-DNA from ripening tomato fruits (AC cultivar) was prepared using antibodies to FUL1 and FUL2 (Shima et al., 2013), and the subsequent hybridization analyses using Roche NimbleGen microarrays for tomato gene promoters were conducted as described previously (Fujisawa et al., 2013). Three independently prepared samples were used in separate hybridizations (including one dye-swap). Data analysis for detection of FUL1- and FUL2-bound regions was conducted as follows: The log2-transformed and normalized values of each probe in three replicates were averaged and subjected to peak detection (FDR ≤ 0.05) by sliding window analysis using NimbleScan software (for more detail, see Fujisawa et al., 2013). For comparison, the previously examined ChIP-chip data for RIN (Fujisawa et al., 2013) were also reanalyzed with the same algorithm. Verification of the binding by qChIP-PCR was basically conducted as previously described (Fujisawa et al., 2013) using Thunderbird SYBR qPCR mix. Primer sequences for qChIP-PCR are listed in Supplemental Table 8. Three biological replicates were analyzed.

In Silico Motif Analysis

FUZZNUC program included in the EMBOSS package (Rice et al., 2000) was used to search DNA sequences of the bound regions unique to FUL1, FUL2, and RIN for CArG-box motif sequences [PAL type, CC(A/T)6GG; N10 type, CTA(A/T)4TAG (West et al., 1998); RIN binding, C(C/T)(A/T)6(A/G)G and C(A/T)8G (Ito et al., 2008; Fujisawa et al., 2011)].

Functional Gene Classification

This study used the functional gene annotation previously conducted (Fujisawa et al., 2011, 2013) for analysis of direct target genes of FUL1, FUL2, and RIN. Briefly, the tomato ITAG2 predicted gene products were tagged by Arabidopsis thaliana proteins based on a similarity search using BLASTP (Altschul et al., 1997). The ITAG2 genes were functionally annotated using the information for tagged Arabidopsis proteins in a functional catalog database provided by MIPS (http://www.helmholtz-muenchen.de/en/ibis/resourcesservices/services/funcat-the-functional-catalogue/index.html) (Ruepp et al., 2004). Significance of enrichment of genes in the subsets of direct targets relative to those in the tomato genome (the whole set of ITAG2 genes) was calculated by Fisher’s exact test for each MIPS category.

Gel Retardation Assay

The expression vectors for RIN, FUL1, and FUL2 were previously constructed (Ito et al., 2008; Shima et al., 2013) and used for cell-free protein synthesis. The expression vector for TAGL1 (pEU-TAGL1) was constructed as follows: The TAGL1 cDNA was amplified with primers specific to TAGL1 (Supplemental Table 8), digested with BamHI and NcoI, and inserted into the pEU-3a vector (Ito et al., 2002). Protein synthesis was performed by in vitro translation using the TnT SP6 quick-coupled transcription/translation system (Promega). Probe DNA was prepared by PCR using plasmids (pGEM-N10 and pGEM-PAL; Ito et al., 2008) that harbor a DNA fragment including the CArG-box sequence (N10, CTATTTATAG; PAL, CCTAATTAGG) (West et al., 1998) as templates with fluorescein isothiocyanate–conjugated primers, as described previously (Ito et al., 2008). Protein-DNA binding reactions were performed according to the method described previously (Riechmann et al., 1996). The reaction was electrophoresed in a 5% polyacrylamide/bisacrylamide (60:1) gel in 0.5× TBE (44.5 mM Tris-borate/1 mM EDTA), and fluorescein isothiocyanate–labeled probes were detected using a Typhoon 8600 (GE Healthcare Bio-Sciences).

Measurement of Naringenin Chalcone

The 45-DAP fruits of the wild type AC, rin mutant, and three FUL/FUL2 suppressed lines (TF2, TF18, and TF68) were harvested, and the peel of the fruits were used for the experiment. Three biological replicates for each stage of the lines were analyzed. Naringenin chalcone accumulation in the peel was measured by liquid chromatography–tandem mass spectrometry analysis according to the method described (Iijima et al., 2008), except for a modification of the internal standard concentration (1 μg/mL of formononetin). An Agilent 1100 system (Agilent) coupled to a 4000QTRAP (AB Sciex) was used for the analysis, and the data were analyzed using Analyst software (version 1.5.2; AB Sciex).

Accession Numbers

Sequence data from this article can be found in the GenBank/EMBL libraries or the tomato unigene database (for genes with SGN number; http://solgenomics.net/search/transcripts/unigene) under the following accession numbers: ACS2, X59139; ACS4, M88487; Actin,U60482; a gene encoding clathrin adaptor complexes medium subunit (CAC), SGN-U314153; Cel2, U13055; CNR, DQ672601; FUL1, SGN-U578128; FUL2, SGN-U581535; GRAS gene (Solyc07g052960), AK327648; PSY1, EF157835; PG2A, X14074; RIN, AF448522; rin, AF448523; and TAGL1, AY098735. The ITAG2 identifiers for tomato genes from this article are shown in Supplemental Data Sets 1, 2, 5, and 6. The RNA-Seq and ChIP-chip data are MIAME compliant and have been deposited in a MIAME-compliant database (Gene Expression Omnibus accession numbers GSE49289 and GSE49125, respectively) at the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/geo/). Supplemental Data Sets 1 to 6 are deposited in the DRYAD repository: http://dx.doi.org/10.5061/dryad.4f7n5.

Supplemental Data

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

Acknowledgments

We thank the Tomato Genetic Resource Center (University of California, Davis) for providing seeds of tomato rin mutant. This work was supported by Scientific Technique Research Promotion Program for agriculture, forestry, fisheries, and food industry of Japan to Y.I.

AUTHOR CONTRIBUTIONS

M.F., M.K., T.K., and Y.I. designed the experiments. M.F., T.N., Y.S., M.K., J.K., and Y.I. prepared samples. M.F., T.N., Y.S., and H.N. performed research. M.F. and Y.I. analyzed the data and wrote the article.

Footnotes

Glossary

ChIP-chip
chromatin immunoprecipitation coupled with microarray analysis
RNAi
RNA interference
RNA-Seq
RNA sequencing
DAP
days after pollination
qRT-PCR
quantitative RT-PCR
MDS
multidimensional scaling
AC
Ailsa Craig
DEG
differentially expressed gene
FC
fold change
FDR
false discovery rate
MIPS
Munich Information Center for Protein Sequences
qChIP-PCR
quantitative ChIP-PCR
ACC
1-aminocyclopropane-1-carboxylic acid
  • Received October 8, 2013.
  • Revised December 11, 2013.
  • Accepted December 19, 2013.
  • Published January 10, 2014.

References

  1. ↵
    1. Altschul S.F.,
    2. Madden T.L.,
    3. Schäffer A.A.,
    4. Zhang J.,
    5. Zhang Z.,
    6. Miller W.,
    7. Lipman D.J.
    (1997). Gapped BLAST and PSI-BLAST: A new generation of protein database search programs. Nucleic Acids Res. 25: 3389–3402.
    OpenUrlAbstract/FREE Full Text
  2. ↵
    1. Ampomah-Dwamena C.,
    2. Morris B.A.,
    3. Sutherland P.,
    4. Veit B.,
    5. Yao J.L.
    (2002). Down-regulation of TM29, a tomato SEPALLATA homolog, causes parthenocarpic fruit development and floral reversion. Plant Physiol. 130: 605–617.
    OpenUrlAbstract/FREE Full Text
  3. ↵
    1. Bailey T.L.,
    2. Williams N.,
    3. Misleh C.,
    4. Li W.W.
    (2006). MEME: discovering and analyzing DNA and protein sequence motifs. Nucleic Acids Res. 34: W369–W373.
    OpenUrlAbstract/FREE Full Text
  4. ↵
    1. Bemer M.,
    2. Karlova R.,
    3. Ballester A.R.,
    4. Tikunov Y.M.,
    5. Bovy A.G.,
    6. Wolters-Arts M.,
    7. Rossetto Pde.B.,
    8. Angenent G.C.,
    9. de Maagd R.A.
    (2012). The tomato FRUITFULL homologs TDR4/FUL1 and MBP7/FUL2 regulate ethylene-independent aspects of fruit ripening. Plant Cell 24: 4437–4451.
    OpenUrlAbstract/FREE Full Text
  5. ↵
    1. Benjamini Y.,
    2. Hochberg Y.
    (1995). Controlling the false discovery rate - A practical and powerful approach to multiple testing. J. R. Stat. Soc. Series B Stat. Methodol. 57: 289–300.
    OpenUrl
  6. ↵
    1. Daminato M.,
    2. Guzzo F.,
    3. Casadoro G.
    (2013). A SHATTERPROOF-like gene controls ripening in non-climacteric strawberries, and auxin and abscisic acid antagonistically affect its expression. J. Exp. Bot. 64: 3775–3786.
    OpenUrlAbstract/FREE Full Text
  7. ↵
    1. Elitzur T.,
    2. Vrebalov J.,
    3. Giovannoni J.J.,
    4. Goldschmidt E.E.,
    5. Friedman H.
    (2010). The regulation of MADS-box gene expression during ripening of banana and their regulatory interaction with ethylene. J. Exp. Bot. 61: 1523–1535.
    OpenUrlAbstract/FREE Full Text
  8. ↵
    1. Eriksson E.M.,
    2. Bovy A.,
    3. Manning K.,
    4. Harrison L.,
    5. Andrews J.,
    6. De Silva J.,
    7. Tucker G.A.,
    8. Seymour G.B.
    (2004). Effect of the Colorless non-ripening mutation on cell wall biochemistry and gene expression during tomato fruit development and ripening. Plant Physiol. 136: 4184–4197.
    OpenUrlAbstract/FREE Full Text
  9. ↵
    1. Ferrándiz C.,
    2. Liljegren S.J.,
    3. Yanofsky M.F.
    (2000). Negative regulation of the SHATTERPROOF genes by FRUITFULL during Arabidopsis fruit development. Science 289: 436–438.
    OpenUrlAbstract/FREE Full Text
  10. ↵
    1. Fujisawa M.,
    2. Nakano T.,
    3. Ito Y.
    (2011). Identification of potential target genes for the tomato fruit-ripening regulator RIN by chromatin immunoprecipitation. BMC Plant Biol. 11: 26.
    OpenUrlCrossRefPubMed
  11. ↵
    1. Fujisawa M.,
    2. Nakano T.,
    3. Shima Y.,
    4. Ito Y.
    (2013). A large-scale identification of direct targets of the tomato MADS box transcription factor RIPENING INHIBITOR reveals the regulation of fruit ripening. Plant Cell 25: 371–386.
    OpenUrlAbstract/FREE Full Text
  12. ↵
    1. Fujisawa M.,
    2. Shima Y.,
    3. Higuchi N.,
    4. Nakano T.,
    5. Koyama Y.,
    6. Kasumi T.,
    7. Ito Y.
    (2012). Direct targets of the tomato-ripening regulator RIN identified by transcriptome and chromatin immunoprecipitation analyses. Planta 235: 1107–1122.
    OpenUrlCrossRefPubMed
  13. ↵
    1. Gapper N.E.,
    2. McQuinn R.P.,
    3. Giovannoni J.J.
    (2013). Molecular and genetic regulation of fruit ripening. Plant Mol. Biol. 82: 575–591.
    OpenUrlCrossRefPubMed
  14. ↵
    1. Giménez E.,
    2. Pineda B.,
    3. Capel J.,
    4. Antón M.T.,
    5. Atarés A.,
    6. Pérez-Martín F.,
    7. García-Sogo B.,
    8. Angosto T.,
    9. Moreno V.,
    10. Lozano R.
    (2010). Functional analysis of the Arlequin mutant corroborates the essential role of the Arlequin/TAGL1 gene during reproductive development of tomato. PLoS ONE 5: e14427.
    OpenUrlCrossRefPubMed
  15. ↵
    1. Giovannoni J.J.
    (2007). Fruit ripening mutants yield insights into ripening control. Curr. Opin. Plant Biol. 10: 283–289.
    OpenUrlCrossRefPubMed
  16. ↵
    1. Gu Q.,
    2. Ferrándiz C.,
    3. Yanofsky M.F.,
    4. Martienssen R.
    (1998). The FRUITFULL MADS-box gene mediates cell differentiation during Arabidopsis fruit development. Development 125: 1509–1517.
    OpenUrlAbstract
  17. ↵
    1. Honma T.,
    2. Goto K.
    (2001). Complexes of MADS-box proteins are sufficient to convert leaves into floral organs. Nature 409: 525–529.
    OpenUrlCrossRefPubMed
  18. ↵
    1. Iijima Y.,
    2. et al
    . (2008). Metabolite annotations based on the integration of mass spectral information. Plant J. 54: 949–962.
    OpenUrlCrossRefPubMed
  19. ↵
    1. Immink R.G.,
    2. Tonaco I.A.,
    3. de Folter S.,
    4. Shchennikova A.,
    5. van Dijk A.D.,
    6. Busscher-Lange J.,
    7. Borst J.W.,
    8. Angenent G.C.
    (2009). SEPALLATA3: The ‘glue’ for MADS box transcription factor complex formation. Genome Biol. 10: R24.
    OpenUrlCrossRefPubMed
  20. ↵
    1. Ireland H.S.,
    2. Yao J.L.,
    3. Tomes S.,
    4. Sutherland P.W.,
    5. Nieuwenhuizen N.,
    6. Gunaseelan K.,
    7. Winz R.A.,
    8. David K.M.,
    9. Schaffer R.J.
    (2013). Apple SEPALLATA1/2-like genes control fruit flesh development and ripening. Plant J. 73: 1044–1056.
    OpenUrlCrossRefPubMed
  21. ↵
    1. Itkin M.,
    2. Seybold H.,
    3. Breitel D.,
    4. Rogachev I.,
    5. Meir S.,
    6. Aharoni A.
    (2009). TOMATO AGAMOUS-LIKE 1 is a component of the fruit ripening regulatory network. Plant J. 60: 1081–1095.
    OpenUrlCrossRefPubMed
  22. ↵
    1. Ito Y.,
    2. Kitagawa M.,
    3. Ihashi N.,
    4. Yabe K.,
    5. Kimbara J.,
    6. Yasuda J.,
    7. Ito H.,
    8. Inakuma T.,
    9. Hiroi S.,
    10. Kasumi T.
    (2008). DNA-binding specificity, transcriptional activation potential, and the rin mutation effect for the tomato fruit-ripening regulator RIN. Plant J. 55: 212–223.
    OpenUrlCrossRefPubMed
  23. ↵
    1. Ito Y.,
    2. Ozawa A.,
    3. Sawasaki T.,
    4. Endo Y.,
    5. Ochi K.,
    6. Tozawa Y.
    (2002). OsRALyase1, a putative F-box protein identified in rice, Oryza sativa, with enzyme activity identical to that of wheat RALyase. Biosci. Biotechnol. Biochem. 66: 2727–2731.
    OpenUrlCrossRefPubMed
  24. ↵
    1. Jaakola L.,
    2. Poole M.,
    3. Jones M.O.,
    4. Kämäräinen-Karppinen T.,
    5. Koskimäki J.J.,
    6. Hohtola A.,
    7. Häggman H.,
    8. Fraser P.D.,
    9. Manning K.,
    10. King G.J.,
    11. Thomson H.,
    12. Seymour G.B.
    (2010). A SQUAMOSA MADS box gene involved in the regulation of anthocyanin accumulation in bilberry fruits. Plant Physiol. 153: 1619–1629.
    OpenUrlAbstract/FREE Full Text
  25. ↵
    1. Kaufmann K.,
    2. Muiño J.M.,
    3. Jauregui R.,
    4. Airoldi C.A.,
    5. Smaczniak C.,
    6. Krajewski P.,
    7. Angenent G.C.
    (2009). Target genes of the MADS transcription factor SEPALLATA3: Integration of developmental and hormonal pathways in the Arabidopsis flower. PLoS Biol. 7: e1000090.
    OpenUrlPubMed
  26. ↵
    1. Langmead B.,
    2. Salzberg S.L.
    (2012). Fast gapped-read alignment with Bowtie 2. Nat. Methods 9: 357–359.
    OpenUrlCrossRefPubMed
  27. ↵
    1. Leseberg C.H.,
    2. Eissler C.L.,
    3. Wang X.,
    4. Johns M.A.,
    5. Duvall M.R.,
    6. Mao L.
    (2008). Interaction study of MADS-domain proteins in tomato. J. Exp. Bot. 59: 2253–2265.
    OpenUrlAbstract/FREE Full Text
  28. ↵
    1. Liljegren S.J.,
    2. Ditta G.S.,
    3. Eshed Y.,
    4. Savidge B.,
    5. Bowman J.L.,
    6. Yanofsky M.F.
    (2000). SHATTERPROOF MADS-box genes control seed dispersal in Arabidopsis. Nature 404: 766–770.
    OpenUrlCrossRefPubMed
  29. ↵
    1. Martel C.,
    2. Vrebalov J.,
    3. Tafelmeyer P.,
    4. Giovannoni J.J.
    (2011). The tomato MADS-box transcription factor RIPENING INHIBITOR interacts with promoters involved in numerous ripening processes in a COLORLESS NONRIPENING-dependent manner. Plant Physiol. 157: 1568–1579.
    OpenUrlAbstract/FREE Full Text
  30. ↵
    1. Mortazavi A.,
    2. Williams B.A.,
    3. McCue K.,
    4. Schaeffer L.,
    5. Wold B.
    (2008). Mapping and quantifying mammalian transcriptomes by RNA-Seq. Nat. Methods 5: 621–628.
    OpenUrlCrossRefPubMed
  31. ↵
    1. Muir S.R.,
    2. Collins G.J.,
    3. Robinson S.,
    4. Hughes S.,
    5. Bovy A.,
    6. Ric De Vos C.H.,
    7. van Tunen A.J.,
    8. Verhoeyen M.E.
    (2001). Overexpression of petunia chalcone isomerase in tomato results in fruit containing increased levels of flavonols. Nat. Biotechnol. 19: 470–474.
    OpenUrlCrossRefPubMed
  32. ↵
    1. Pan I.L.,
    2. McQuinn R.,
    3. Giovannoni J.J.,
    4. Irish V.F.
    (2010). Functional diversification of AGAMOUS lineage genes in regulating tomato flower and fruit development. J. Exp. Bot. 61: 1795–1806.
    OpenUrlAbstract/FREE Full Text
  33. ↵
    1. Pelaz S.,
    2. Ditta G.S.,
    3. Baumann E.,
    4. Wisman E.,
    5. Yanofsky M.F.
    (2000). B and C floral organ identity functions require SEPALLATA MADS-box genes. Nature 405: 200–203.
    OpenUrlCrossRefPubMed
  34. ↵
    1. Pnueli L.,
    2. Hareven D.,
    3. Broday L.,
    4. Hurwitz C.,
    5. Lifschitz E.
    (1994a). The TM5 MADS box gene mediates organ differentiation in the three inner whorls of tomato flowers. Plant Cell 6: 175–186.
    OpenUrlAbstract/FREE Full Text
  35. ↵
    1. Pnueli L.,
    2. Hareven D.,
    3. Rounsley S.D.,
    4. Yanofsky M.F.,
    5. Lifschitz E.
    (1994b). Isolation of the tomato AGAMOUS gene TAG1 and analysis of its homeotic role in transgenic plants. Plant Cell 6: 163–173.
    OpenUrlAbstract/FREE Full Text
  36. ↵
    1. Qin G.,
    2. Wang Y.,
    3. Cao B.,
    4. Wang W.,
    5. Tian S.
    (2012). Unraveling the regulatory network of the MADS box transcription factor RIN in fruit ripening. Plant J. 70: 243–255.
    OpenUrlCrossRefPubMed
  37. ↵
    1. Rice P.,
    2. Longden I.,
    3. Bleasby A.
    (2000). EMBOSS: The European Molecular Biology Open Software Suite. Trends Genet. 16: 276–277.
    OpenUrlCrossRefPubMed
  38. ↵
    1. Riechmann J.L.,
    2. Krizek B.A.,
    3. Meyerowitz E.M.
    (1996). Dimerization specificity of Arabidopsis MADS domain homeotic proteins APETALA1, APETALA3, PISTILLATA, and AGAMOUS. Proc. Natl. Acad. Sci. USA 93: 4793–4798.
    OpenUrlAbstract/FREE Full Text
  39. ↵
    1. Robinson M.D.,
    2. McCarthy D.J.,
    3. Smyth G.K.
    (2010). edgeR: A Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26: 139–140.
    OpenUrlAbstract/FREE Full Text
  40. ↵
    1. Robinson M.D.,
    2. Oshlack A.
    (2010). A scaling normalization method for differential expression analysis of RNA-seq data. Genome Biol. 11: R25.
    OpenUrlCrossRefPubMed
  41. ↵
    1. Robles J.A.,
    2. Qureshi S.E.,
    3. Stephen S.J.,
    4. Wilson S.R.,
    5. Burden C.J.,
    6. Taylor J.M.
    (2012). Efficient experimental design and analysis strategies for the detection of differential expression using RNA-Sequencing. BMC Genomics 13: 484.
    OpenUrlCrossRefPubMed
  42. ↵
    1. Ruepp A.,
    2. Zollner A.,
    3. Maier D.,
    4. Albermann K.,
    5. Hani J.,
    6. Mokrejs M.,
    7. Tetko I.,
    8. Güldener U.,
    9. Mannhaupt G.,
    10. Münsterkötter M.,
    11. Mewes H.W.
    (2004). The FunCat, a functional annotation scheme for systematic classification of proteins from whole genomes. Nucleic Acids Res. 32: 5539–5545.
    OpenUrlAbstract/FREE Full Text
  43. ↵
    1. Schaffer R.J.,
    2. Ireland H.S.,
    3. Ross J.J.,
    4. Ling T.J.,
    5. David K.M.
    (2013). SEPALLATA1/2-suppressed mature apples have low ethylene, high auxin and reduced transcription of ripening-related genes. AoB Plants 5: pls047.
    OpenUrlAbstract/FREE Full Text
  44. ↵
    1. Schultz D.J.,
    2. Craig R.,
    3. Cox-Foster D.L.,
    4. Mumma R.O.,
    5. Medford J.I.
    (1994). RNA isolation from recalcitrant plant tissues. Plant Mol. Biol. Rep. 12: 310–316.
    OpenUrlCrossRef
  45. ↵
    1. Seymour G.B.,
    2. Ryder C.D.,
    3. Cevik V.,
    4. Hammond J.P.,
    5. Popovich A.,
    6. King G.J.,
    7. Vrebalov J.,
    8. Giovannoni J.J.,
    9. Manning K.
    (2011). A SEPALLATA gene is involved in the development and ripening of strawberry (Fragaria x ananassa Duch.) fruit, a non-climacteric tissue. J. Exp. Bot. 62: 1179–1188.
    OpenUrlAbstract/FREE Full Text
  46. ↵
    1. Shima Y.,
    2. Fujisawa M.,
    3. Kitagawa M.,
    4. Nakano T.,
    5. Kimbara J.,
    6. Nakamura N.,
    7. Shiina T.,
    8. Sugiyama J.,
    9. Nakamura T.,
    10. Kasumi T.,
    11. Ito Y.
    (2014). Tomato FRUITFULL homologs regulate fruit ripening via ethylene biosynthesis. Biosci. Biotechnol. Biochem., in press.
  47. ↵
    1. Shima Y.,
    2. Kitagawa M.,
    3. Fujisawa M.,
    4. Nakano T.,
    5. Kato H.,
    6. Kimbara J.,
    7. Kasumi T.,
    8. Ito Y.
    (2013). Tomato FRUITFULL homologues act in fruit ripening via forming MADS-box transcription factor complexes with RIN. Plant Mol. Biol. 82: 427–438.
    OpenUrlCrossRefPubMed
  48. ↵
    1. Smaczniak C.,
    2. Immink R.G.,
    3. Angenent G.C.,
    4. Kaufmann K.
    (2012a). Developmental and evolutionary diversity of plant MADS-domain factors: Insights from recent studies. Development 139: 3081–3098.
    OpenUrlAbstract/FREE Full Text
  49. ↵
    1. Smaczniak C.,
    2. et al
    . (2012b). Characterization of MADS-domain transcription factor complexes in Arabidopsis flower development. Proc. Natl. Acad. Sci. USA 109: 1560–1565.
    OpenUrlAbstract/FREE Full Text
  50. ↵
    1. Theissen G.,
    2. Saedler H.
    (2001). Plant biology. Floral quartets. Nature 409: 469–471.
    OpenUrlCrossRefPubMed
  51. ↵
    1. Tigchelaar E.C.,
    2. Mcglasson W.B.,
    3. Buescher R.W.
    (1978). Genetic-regulation of tomato fruit ripening. HortScience 13: 508–513.
    OpenUrl
  52. ↵
    1. Tomato Genome Consortium
    (2012). The tomato genome sequence provides insights into fleshy fruit evolution. Nature 485: 635–641.
    OpenUrlCrossRefPubMed
  53. ↵
    1. Vrebalov J.,
    2. Pan I.L.,
    3. Arroyo A.J.,
    4. McQuinn R.,
    5. Chung M.,
    6. Poole M.,
    7. Rose J.,
    8. Seymour G.,
    9. Grandillo S.,
    10. Giovannoni J.,
    11. Irish V.F.
    (2009). Fleshy fruit expansion and ripening are regulated by the tomato SHATTERPROOF gene TAGL1. Plant Cell 21: 3041–3062.
    OpenUrlAbstract/FREE Full Text
  54. ↵
    1. Vrebalov J.,
    2. Ruezinsky D.,
    3. Padmanabhan V.,
    4. White R.,
    5. Medrano D.,
    6. Drake R.,
    7. Schuch W.,
    8. Giovannoni J.
    (2002). A MADS-box gene necessary for fruit ripening at the tomato ripening-inhibitor (rin) locus. Science 296: 343–346.
    OpenUrlAbstract/FREE Full Text
  55. ↵
    1. West A.G.,
    2. Causier B.E.,
    3. Davies B.,
    4. Sharrocks A.D.
    (1998). DNA binding and dimerisation determinants of Antirrhinum majus MADS-box transcription factors. Nucleic Acids Res. 26: 5277–5287.
    OpenUrlAbstract/FREE Full Text
  56. ↵
    1. Zhong S.,
    2. Fei Z.,
    3. Chen Y.R.,
    4. Zheng Y.,
    5. Huang M.,
    6. Vrebalov J.,
    7. McQuinn R.,
    8. Gapper N.,
    9. Liu B.,
    10. Xiang J.,
    11. Shao Y.,
    12. Giovannoni J.J.
    (2013). Single-base resolution methylomes of tomato fruit development reveal epigenome modifications associated with ripening. Nat. Biotechnol. 31: 154–159.
    OpenUrlCrossRefPubMed
Back to top

Table of Contents

Print
Download PDF
Email Article
Citation Tools
Request Permissions
Transcriptional Regulation of Fruit Ripening by Tomato FRUITFULL Homologs and Associated MADS Box Proteins
Masaki Fujisawa, Yoko Shima, Hiroyuki Nakagawa, Mamiko Kitagawa, Junji Kimbara, Toshitsugu Nakano, Takafumi Kasumi, Yasuhiro Ito
The Plant Cell Jan 2014, 26 (1) 89-101; DOI: 10.1105/tpc.113.119453
del.icio.us logo Digg logo Reddit logo Twitter logo CiteULike logo Facebook logo Google logo Mendeley logo

Jump to section

In this issue

View this article with LENS