ENO regulates tomato fruit size through the floral meristem development network

Significance Fruit-size increase is one of the major changes associated with tomato domestication, and it currently represents an important objective for breeding. Regulatory mutations at the LOCULE NUMBER and FASCIATED loci, the orthologues of the Arabidopsis WUSCHEL and CLAVATA3, have mainly contributed to enlarging fruit size by altering meristem activity. Here, we identify ENO as a tomato fruit regulator, which may function by regulating WUSCHEL gene expression to restrict stem-cell proliferation in a flower-specific manner. Our findings also show that a mutation in the ENO promoter was selected during domestication to establish the background for enhancing fruit size in cultivated tomatoes, denoting that transcriptional changes in key regulators have significant effects on agronomic traits.

translation of ENO protein were performed using the TNT® Quick Coupled Transcription/Translation System (Promega, Madison, WI, USA) according to the supplier's instructions. The binding activity of ENO to specific DNA sequence which included theoretical ERF binding site (GCCGTC) on the SlWUS promoter was assayed using a LightShift™ Chemiluminescent EMSA kit (Thermo Scientific, Wilmington, DE, USA). Briefly, ENO protein was incubated in a binding buffer (10 mM Tris-HCl (pH 7.5), 50 mM KCl, 1 mM dithiothreitol, 2.5% glycerol, 5mM MgCl2, 50ng/µg Poly(dI-dC), 0.05% NP-40) for 20 minutes at room temperature in the presence or absence of unlabeled (double-stranded) competitor probe. The biotin-labeled dsDNA probe was then added and the incubation continued for 20 min. dsDNA biotinylated probe was generated by amplification with SlWUS_EMSA-F and SlWUS_EMSA-R biotinylated primers (SI Appendix, Table S6). Protein-DNA complexes were resolved on 6% native polyacrylamide gels in 0.5X TBE buffer, and the biotin-labeled probes were detected according to the instructions provided with the LightShift™ Chemiluminescent EMSA kit.
Sequencing of ENO in wild and cultivated tomato accessions. Genomic DNA from accessions indicated in Dataset S3 was extracted using the DNAzol® Reagent kit (Invitrogen Life Technologies, San Diego, CA, USA). A 1.6 kb genomic region harboring the full-length ENO coding sequence was amplified by PCR in overlapping fragments of approximately 650 bp using primers listed in SI Appendix, Table S6. PCR products were sequenced by Sanger technology using the BigDye® Terminator v3.1 chemistry and the Applied Biosystems TM 3500 Series Genetic Analyzer (Applied Biosystems, Foster City, CA, USA). Sequence analysis and alignments were performed using Geneious software (25).
Allele-specific PCR for ENO InDel mutation. Allele-specific ENO transcript levels were determined by TaqMan probe using Droplet Digital PCR (ddPCR) assay. cDNA of F1 hybrids heterozygous for the InDel mutation (haplotype-1 x haplotype-9) was used for ddPCR following the method described in Kamitaki et al. (26). Briefly, 250 ng/µL of cDNA sample solution was combined with 11 µL of 2X ddPCR Supermix for Probes, 1 µL of target TaqMan probe (5 µM) and 1 μl each of forward and reverse primers (18 µM) in a total volume of 20 µL. After droplet generation using a QX200 TM automated droplet generator (Bio-Rad, Hercules, CA, USA), PCR was performed in a T100 Thermal Cycler (Bio-Rad, Hercules, CA, USA) using the following thermal cycling conditions: 1 cycle of 10 minutes at 95°C; 40 cycles of 30 seconds at 94°C followed by 1 minute at 60°C; 1 cycle of 10 minutes at 98°C. Samples were subsequently measured using a QX200 droplet reader (Bio-Rad, Hercules, CA, USA) with QuantaSoft Software. The ddPCR analysis was carried out using three biological and two technical replicates. Primer and probe sequences used are shown in SI Appendix, Table S6.
Classification of re-sequenced tomato accessions in phylogenetic groups. Whole genome short read data for 601 tomato accessions (27) were downloaded from NCBI-SRA and aligned to the S. lycopersicum reference genome version 2.50 (28) using Bowtie2 (4) with default parameters. Reads mapping to multiple locations were removed using SAMtools (29) (parameter -q 5), duplicated reads were removed using picard-tools MarkDuplicates (http://broadinstitute.github.io/picard) and InDels were realigned using GATK RealignerTargetCreator and IndelRealigner v3.8 (5). In order to classify the re-sequenced accessions in phylogenetic groups, we compared them with 1008 accessions classified in (30) using 8700 genome-wide SNPs genotyped with the SolCAP Infinium Chip. For this purpose, we genotyped the 601 re-sequenced accessions at the SolCAP Infinium array positions (as indicated in the ITAG2.4_solCAP.gff3 file available at ftp.solgenomics.net) using GATK UnifiedGenotyper (5) with default parameters. After merging the two datasets, only variants that were bi-allelic and whose alleles agreed in both datasets were kept. The resulting 5856 variants were filtered to remove loci in linkage disequilibrium using the LD pruning option of PLINK with parameters --mind 0.1 --geno 0.1 --indep 50 5 2.8 (31). A phylogenetic tree was estimated from the final matrix (1536 variants in 1609 accessions) using the ape package in R and the neighbor-joining method including S. pennellii LA0716 as a root (32). The resulting tree was plotted using the ggtree package in R (33). Tomato accessions in the tree were classified manually having into account the previously described classification (30) and their position in the tree (Dataset S4 and SI Appendix, Fig. S5).
Genotyping of re-sequenced tomato accessions for calculation of the ENO promoter, lc and fas mutant allele frequencies. Alignments from 601 re-sequenced accessions obtained as described above were genotyped for mutations at ENO, LC and FAS. The deletion in the promoter of ENO was genotyped by aligning the reads from all 601 re-sequenced accessions to a tomato reference genome where the deleted nucleotides were included, and visually inspecting the region for the presence or absence of reads. For LC, GATK UnifiedGenotyper (5) with default parameters was used to obtain the genotype at positions SL2.50ch02:47188498 (T/C) and SL2.50ch02:47188504 (A/C). Accessions with T + A alleles at these positions were considered wild type while accessions with C + C alleles were considered mutant. The mutation at FAS locus was genotyped by visually inspecting the alignments at positions SL2.50ch11:54877493-54877107 and SL2.50ch11:55171147-55171482 for signals of the inversion breakpoints previously described (34). Presence of abundant reads without pair and low coverage at both breakpoints were considered as the presence of the inversion. In all cases visualization of the alignments was performed with the Integrative Genomics Viewer genome browser (35).
Twenty-one Vintage tomato accessions were found to contain the wild-allele (absence of the ENO promoter deletion) at ENO locus (Dataset S4), which could be due to a recent introgression of a wild species in the genome of cultivated tomato. To investigate this possibility, we called variants using GATK UnifiedGenotyper (5) in the region of ENO in alignments of 139 Vintage accessions. The frequency of non-reference (alternative) bi-allelic SNPs was calculated in windows of 1000 SNPs and steps of 500 for each accession using the R Environment for Statistical Computing (7). The analysis of these frequencies along chromosome 3 shows that all vintage accessions with the ENO wild-allele contain high frequencies of non-reference alleles in genomic blocks that span the region of ENO, in contrast to all accessions carrying the ENO promoter deletion allele (SI Appendix, Fig. S6). These large blocks of non-reference alleles represent introgressions from wild species acquired during breeding. Sequence alignment and phylogenetic reconstruction of the tomato and Arabidopsis ERF subfamily group VIII was performed with MEGA6 (8). Full-length proteins were aligned using CLUSTALW (9). The phylogenetic tree was generated by the Neighbor-Joining method, using Jones-Taylor-Thornton substitution model. A bootstrap analysis with 1000 replications was performed to confirm the reliability of the constructed phylogeny. Bootstrap values of < 40% are not shown. Hierarchical tree graph of enriched GO terms for significantly differentially expressed genes between wild-type and eno mutant reproductive meristems using agriGO v2.0 software (22). Enriched GO terms were determined by false discovery rate (FDR) < 0.05 with the Fisher statistical test and the Bonferroni multi-test adjustment. To obtain an overview of enriched GO terms for the genes significantly (A) up-and (B) downregulated in reproductive meristems from eno and wild-type plants, their putative Arabidopsis homologues were used to perform a GO enrichment analysis using the Cytoscape plug-in ClueGO (Version 2.5.5) (24). The dimension of the pie chart wedges is proportional to the number of GO terms included in each GO group, which are listed in Dataset S2. The 1.6 kb of DNA sequences from the nine ENO haplotypes were aligned by using CLUSTALW (9) implemented in the msa Bioconductor R package (36). The ENO coding sequence is indicated with a red line. Exclamation marks (!) indicate fully conserved residues, while asterisks (*) indicate polymorphic sites. The tree was constructed with the neighbor-joining method using 1536 genome-wide SNPs. Data was obtained from (27,30). Colors in the branches and accessions represent the phylogenetic groups defined in by (30). All re-sequenced accessions are colored in gold. Details on groups and accessions can be found in Dataset S4. Branches of the tree used in the evolutionary analysis of ENO are marked by circular lines outside the tree.     Fig. S6. Frequency of non-reference SNPs in vintage tomatoes in the region of ENO.
The frequency of non-reference (alternative) bi-allelic SNPs was calculated in 139 vintage tomato accessions in windows of 1000 SNPs and steps of 500. Accessions with the deletion in the promoter of ENO are colored in red, while accessions without the deletion colored in blue. Blocks of high frequencies of non-reference alleles are indicative of introgressions from wild species that have occurred during breeding. Table S1. Quantification of floral organ number in wild-type (cv. P73) and eno plants.