https://orcid.org/0000-0001-5184-8587
Correction
18 Jan 2024: Sapkota M, Pereira L, Wang Y, Zhang L, Topcu Y, et al. (2024) Correction: Structural variation underlies functional diversity at methyl salicylate loci in tomato. PLOS Genetics 20(1): e1011125. https://doi.org/10.1371/journal.pgen.1011125 View correction
Figures
Abstract
Methyl salicylate is an important inter- and intra-plant signaling molecule, but is deemed undesirable by humans when it accumulates to high levels in ripe fruits. Balancing the tradeoff between consumer satisfaction and overall plant health is challenging as the mechanisms regulating volatile levels have not yet been fully elucidated. In this study, we investigated the accumulation of methyl salicylate in ripe fruits of tomatoes that belong to the red-fruited clade. We determine the genetic diversity and the interaction of four known loci controlling methyl salicylate levels in ripe fruits. In addition to Non-Smoky Glucosyl Transferase 1 (NSGT1), we uncovered extensive genome structural variation (SV) at the Methylesterase (MES) locus. This locus contains four tandemly duplicated Methylesterase genes and genome sequence investigations at the locus identified nine distinct haplotypes. Based on gene expression and results from biparental crosses, functional and non-functional haplotypes for MES were identified. The combination of the non-functional MES haplotype 2 and the non-functional NSGT1 haplotype IV or V in a GWAS panel showed high methyl salicylate levels in ripe fruits, particularly in accessions from Ecuador, demonstrating a strong interaction between these two loci and suggesting an ecological advantage. The genetic variation at the other two known loci, Salicylic Acid Methyl Transferase 1 (SAMT1) and tomato UDP Glycosyl Transferase 5 (SlUGT5), did not explain volatile variation in the red-fruited tomato germplasm, suggesting a minor role in methyl salicylate production in red-fruited tomato. Lastly, we found that most heirloom and modern tomato accessions carried a functional MES and a non-functional NSGT1 haplotype, ensuring acceptable levels of methyl salicylate in fruits. Yet, future selection of the functional NSGT1 allele could potentially improve flavor in the modern germplasm.
Author summary
Tomato fruit release a wide variety of volatiles and these volatiles are responsible for its unique aroma. One of the important volatiles in tomato flavor is methyl salicylate. Methyl salicylate is important for plant defense, but tomato taste panels routinely associate this volatile with low liking as it features the characteristic odor of wintergreen. Several enzymes involved in the biosynthesis of methyl salicylate have been identified in tomato but the genetic diversity at the gene region has not yet been fully explored. In this study, we investigate the genetic variations at the four known gene regions controlling fruit methyl salicylate levels using a diverse collection ranging from red-fruited wild tomatoes to modern and heirloom varieties. We identified extensive genomic structural variations at two of the four loci. The genetic interaction between two loci determines the volatile levels. We also showed dramatic selection pressures ensued at these loci during evolution and domestication of the tomato. Our study greatly improves the understanding of the genetic basis of methyl salicylate production in tomato fruits and gives breeders far better tools to create agrosystem-specific adapted commercial germplasm.
Citation: Sapkota M, Pereira L, Wang Y, Zhang L, Topcu Y, Tieman D, et al. (2023) Structural variation underlies functional diversity at methyl salicylate loci in tomato. PLoS Genet 19(5): e1010751. https://doi.org/10.1371/journal.pgen.1010751
Editor: Nathan M. Springer, University of Minnesota, UNITED STATES
Received: December 16, 2022; Accepted: April 19, 2023; Published: May 4, 2023
Copyright: © 2023 Sapkota et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All the raw sequencing data used in this study are publicly available in NCBI (https://www.ncbi.nlm.nih.gov/; SRA: SRP150040, SRP045767, SRP094624, and PRJNA353161).
Funding: This work was funded by NSF IOS 2151032 (EvdK and DT) and NSF IOS 1732253 (EvdK). M.S. is the recipient of the John Ingle Innovation in Plant Breeding Award and Roger Boerma Plant Breeding Excellence Scholarship Award. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Plants synthesize and emit many volatile organic compounds that play key roles in reproduction, communication, and stress responses. The volatiles emitted from flowers allow plants to attract and guide pollinators to ensure fertilization and reproduction [1,2], whereas aroma from fruits allure frugivores to assist in seed dispersal [3–5] and fend off pre- and post-harvest diseases [5–7]. Volatiles play important roles in intra- and inter-plant communications as these compounds facilitate the signaling about the danger to distant parts of same or to neighboring plants [8,9]. Furthermore, volatiles can act as direct defenses, either by repelling the danger [10,11], or intoxicating herbivores and pathogens [12,13]. Volatiles may also serve as signals to attract herbivores’ natural enemies such as parasitoids and predators [14–18] and even predatory birds [19].
The complex tomato fruit volatile aroma is produced by derivatives from six major biochemical pathways: (i) fatty acid or lipid-derived; (ii) branched chain amino acid (BCAA)-derived such as leucine (Leu) and isoleucine (Ile); (iii) phenylalanine (Phe)-derived including phenylpropanoids and phenolics; (iv) terpenoid-derived; (v) carotenoid-derived; and the (vi) acetate ester volatiles [20–22]. Tomato fruit release a wide variety of volatiles, with emission rates peaking at ripening [23]. These volatiles comprise only 10−7 to 10−4 of the fresh fruit weight and are responsible for its unique aroma [24]. In general, well-liked tomatoes feature a complex volatile aroma that is high in certain volatiles and low in others, and properly balanced in yet other volatiles [20,25].
Red-fruited tomato is edible and evolved in the tomato clade of the Solanum genus in the Solanaceae family. The red-fruited S. pimpinellifolium (SP) last shared a common ancestor with the green-fruited species approximately two million years ago [26,27]. Within the red-fruited clade, SP evolved into wild S. lycopersicum cerasiforme (SLC) that then became domesticated into S. lycopersicum lycopersicum (SLL) around seven thousand years ago [28,29]. The huge variation in volatile production in tomato from SP to SLL indicates that its domestication placed strong selection pressures on this trait [25]. One of the important volatiles in tomato flavor is methyl salicylate which is produced in the Phe-derived pathway. Tomato taste panels routinely associate this volatile with low liking as it features the characteristic odor of wintergreen. However, some varieties with moderate to high levels of methyl salicylate rate well for overall liking. This may be the result of concentrations of or interactions with other volatiles resulting in overall good flavor [25,30,31].
Several enzymes involved in the biosynthesis of methyl salicylate and its glucoside conjugation have been identified in tomato (Fig 1) [32–35]. Methyl salicylate is synthesized from salicylic acid, which is produced from the isochorismate and the Phe-derived pathways [36]. The process of methylation of salicylic acid into methyl salicylate is catalyzed by the enzyme SALICYLIC ACID METHYLTRANSFERASE 1 (SlSAMT1; Solyc09g091550) [32]. Another group of enzymes, METHYLESTERASE 1–4 (SlMES1-4: Solyc02g065240-80), demethylates methyl salicylate to salicylic acid to maintain homeostasis [35]. The methyl salicylate can also be glycosylated into glycoconjugate compounds for transport and storage in the vacuole [37]. The two major enzyme families involved in the formation and degradation of glycosylated methyl salicylate are: glycosyltransferases and glycosidases. Glycosyltransferases (eg. SlUGT5; Solyc01g095620) catalyze the formation of diglycoside conjugates [33]. These diglycoside conjugates are hydrolyzed to methyl salicylate by glycosidases. One glycosyl transferase, Non-Smoky Glycosyltransferase 1 (NSGT1; Solyc09g089585), encodes a fruit ripening-induced enzyme that adds a glucose moiety to the methyl salicylate diglycoside conjugate resulting in a triglycoside. This triglycoside conjugation is irreversible and prevents the emission of the volatile upon tissue disruption and ripening [21,34]. Of the enzymes in the methyl salicylate biosynthesis and conjugation pathway, loss of function in either SlMES1-4 or NSGT1 results in high methyl salicylate in fruits [34,35].
The gray boxes represent the two pathways leading to salicylic acid. The arrows represent the direction of the reaction. The enzyme catalyzing the reaction is listed next to the arrow. The genes names are: SAMT1: Salicylic Acid Methyltransferase 1; MES1-4: Methylesterase 1–4; UGT5: UDP-glycosyltransferase 5; NSGT1: Non-smoky Glycosyltransferase 1 [32–35]
NSGT1 is known for the glycosylation of guaiacol, a smoky volatile that, similarly to methyl salicylate, negatively affects consumer liking. Based on genome structural variations (SV), NSGT1 is represented by five haplotypes that are grouped into functional and non-functional categories based on the presence of a functional enzyme [38]. Similarly, MES also exhibits SVs [35], but the diversity at this locus has not yet been fully explored in tomato. The genetic diversity at the other two known genes, SlSAMT1 and SlUGT5, has also not been explored in red-fruited tomato. The genetic characterization of these loci can provide evolutionary insights and lead to the identification of haplotypes that might be beneficial for crop improvement.
In this study, we analyzed a unique and genetically well-characterized tomato collection of fully wild SP, semi-domesticated SLC and ancestral SLL landraces from South and Central America (Varitome Collection) as well as a subset of cultivated accessions that includes heirloom and modern types (SLL_CUL). Collectively, they are referred to as red-fruited tomato in this study. We investigated the genetic diversity at the known genes controlling methyl salicylate levels in tomato fruits and uncovered extensive genome SV underlying the trait in the red-fruited tomato clade. We also show that different selection pressures ensued at these loci during domestication.
Results
Distribution of methyl salicylate levels in ripe, red-fruited tomato fruits
Quantification of fruit methyl salicylate in 166 accessions from the Varitome Collection and 143 cultivated types showed that the levels ranged from 0.003 nanogram per gram fruit weight per hour (ng/gfw/hr) to 29.360 ng/gfw/hr, representing a 9000-fold difference (Fig 2A). Most accessions showed less than 1 ng/gfw/hr of methyl salicylate in the fruits. Significantly higher levels were found in SP from Northern Ecuador and SLC from Ecuador, whereas ancestral landraces, SLL, featured low levels of the volatile. Most of the SLL_CUL produced low levels of methyl salicylate with a few exceptions. We observed variation in other sub-populations, but overall, the fruits emitted low levels of the volatile (Fig 2B).
(A) Histogram representing the distribution of methyl salicylate in ripe tomato fruits from the Varitome Collection and heirlooms. X-axis represents the methyl salicylate levels in fruits whereas, y-axis denotes the count. (B) Methyl salicylate levels in tomato accessions grouped by sub population. SP_SECU: SP from Southern Ecuador, SP_PER: SP from Peru, SP_NECU: SP from Norther Ecuador, SLC_ADM: SLC admixture, SLC_ECU: SLC from Ecuador, SLC_PER: SLC from Peru, SLC_SM: SLC from San Martin, SLC_CA: SLC from Central America, SLC_MEX: SLC from Mexico, SLL: SLL in Varitome Collection and SLL_CUL: modern and heirloom accessions. The letters in the boxplots indicate the significant differences among different sub population evaluated by Duncan’s test (α < 0.05). The x-axis represents different sub population whereas, y-axis denotes the methyl salicylate levels in fruits.
SNP, INDEL and SV-based GWAS of methyl salicylate levels in ripe tomato fruits
To investigate the genetic bases of the variation in methyl salicylate levels, we performed a genome wide association study (GWAS) using the whole genome variant information of the Varitome Collection. We identified several GWAS loci associated with fruit methyl salicylate levels (Fig 3 and S1 Table) using Single Nucleotide Polymorphisms (SNPs), Insertion-Deletions (INDELs) and Structural Variants (SVs). The identified GWAS loci included MES on chromosome 2 (chr 02), NSGT1 on chr 09, and SAMT1 also on chr 09. For MES, the locus was identified by all three variant types, namely a 4.7 kb deletion at position SL.40ch02:34,464,833, a SNP at SL4.0ch02:34,895,803, and an INDEL at SL4.0ch02:34,943,905 (Fig 4A–4C). The MES locus contains four Methylesterase genes (Solyc02g065240-80) and the 4.7 kb deletion partially spans the third exon of Solyc02g065260 (SlMES3), whereas the SNP and INDEL mapped 400 kb and 467 kb downstream of the MES locus, respectively. The accessions carrying the 4.7 kb deletion produced significantly higher amounts of methyl salicylate compared to the accessions without the deletion (Fig 4A) and the variant explained 28% of total methyl salicylate variation in the Varitome Collection. Similar to the SV at MES, accessions having alternate alleles of the associated SNP and INDEL produced significantly higher levels of methyl salicylate in tomato fruits (Fig 4B and 4C). For NSGT1, the locus was identified by a SNP at SL4.0ch09:65,404,388 (4.3 kb downstream of NSGT1) and an INDEL at SL4.0ch09: 65,401,256 (7.4 kb downstream of NSGT1), explaining 40% of the variation in the population. These overlapping GWAS loci carried three alleles: the reference allele, the alternate (SNP or INDEL) allele, or the 15 kb deletion. Accessions with the deletion at the locus showed significantly higher levels of methyl salicylate compared to accessions without the deletion (Fig 4D and 4E). For SAMT1, the locus was detected by a SNP at SL4.0ch09: 67,197,164 which was located 293 kb downstream of the gene. The accessions carrying the reference allele (GG) produced lower levels of methyl salicylate compared with the alternative allele (GA) (Fig 4F). Similar to SAMT1, the closest significant variant for UGT5 on chr 01 was an INDEL (SL4.0ch01: 85,548,458) that mapped to approximately 6.5 Mb downstream the gene. For this variant, all accessions with reference allele produced low methyl salicylate levels in fruits, whereas lines with a nucleotide insertion showed significantly higher levels of the volatile (Kruskal-Wallis test p-value: 6.4e-06) (Fig 4G).
(A) GWAS based on SNPs. (B) GWAS based on INDELs. (C) GWAS based on SVs. The x-axis represents different chromosomes and y-axis denotes the -log10(p-value) of the variants.
Boxplots representing the distribution of methyl salicylate levels among different alleles of the (A) SV, (B) SNP, and (C) INDEL identified at MES locus from GWAS. Boxplots representing the distribution of methyl salicylate levels among different alleles of the (D) INDEL and (E) SNP identified at NSGT1 locus from GWAS. (F) Boxplots representing the distribution of methyl salicylate levels among different alleles of SNP identified at SAMT1 locus from GWAS (G) Boxplots representing the distribution of methyl salicylate levels among different alleles of the INDEL identified at UGT5 locus from GWAS. (H) The table provides the information of significant variants identified from GWAS along with the known gene loci. The reference alleles are colored red. The p-value is obtained from Kruskal-Wallis test. The percentage of variation explained (PVE) is calculated using respective variants in the ANOVA model. The letters in the boxplots indicate the significant differences among different alleles evaluated by Duncan’s test (α < 0.05).
Genome landscape of MES, SAMT1 and UGT5 loci in the red-fruited tomatoes
Previously, five haplotypes of NSGT1 were identified based on the SVs at the locus of which haplotype IV and V were associated with higher methyl salicylate levels [38]. To gain more insight about the haplotype diversity of the other genes in this pathway, we investigated the genomic landscape at MES, UGT5 and SAMT1 loci with the focus on SVs. For MES, we compared the genomic sequences of the three highest methyl salicylate accumulating accessions (BGV006852: 28.388 ng/gfw/hr, BGV006931: 17.316 ng/gfw/hr and BGV006148: 14.472 ng/gfw hr) and the three lowest accumulating accessions (BGV007151: 0.0094 ng/gfw hr, BGV008098: 0.0097 ng/gfw/hr and BGV007867: 0.0099 ng/gfw/hr). The high methyl salicylate producing accessions carried several deletions at MES and multiple mismatches when compared to the Heinz1706 reference genome. On the other hand, the genomic landscape of the low methyl salicylate producing accessions corresponded more closely to the Heinz1706 reference genome (S1 Fig). Further comparisons based on SVs (S2 Table) showed additional haplotypes in red-fruited tomato to a total of nine (Fig 5 and S3 Table). We next investigated each of the predicted genes at MES to ascertain functionality with respect to the potential to encode a full-length protein (S2 Fig). The results showed the predicted functional genes at the locus ranged from zero (haplotype 9) to four (haplotype 7). MES1 was likely functional in every haplotype except 9. MES3 was truncated in several haplotypes, missing 97 amino acids from its C terminal end except in haplotype 1. It is known that the truncated allele of MES3 retains at least partial enzyme activity and is therefore considered functional [35]. For MES2 and MES4, several haplotypes appeared to carry a functional allele for one or both genes. MES4 was either full length or was truncated resulting from a 48 bp deletion in the first exon (S2 Table).
Blue inverted triangles represent deletions, dark blue boxes indicate likely functional genes, light blue boxes indicate non-functional genes whereas the light red boxes represent truncated genes. Next to each haplotype are the accessions belonging to the haplotype.
We next constructed a phylogenetic tree using the genome sequence of MES including 10 kb that flank the four genes at the locus (Fig 6). Haplotype 1 was found only in a few accessions that did not cluster with any other haplotype (Fig 6 and S4 Table). Haplotype 3 was also distinct and differed from haplotype 1 by a transposon insertion at the C terminal end of SlMES3 resulting in a truncated version of the gene and a remaining protein segment that was annotated as Solyc02g065270 (S3 Fig). Haplotypes 2 and 4 were also more ancestral, both showing a deletion in the promoter of SlMES1 compared to other haplotypes. Haplotype 2 carries the 48 bp deletion in the first exon of SlMES4, truncating the protein to 119 amino acids instead of 279 amino acids, whereas haplotype 4 has an intergenic deletion between SlMES3 and SlMES4. Haplotypes 2 and 4 were clustered together implying a close evolutionary relationship in addition to a similar SV structure. The most common haplotypes were 6 and 7, found in 122 and 120 accessions, respectively. Haplotype 6 included the reference genome Heinz1706 and M82 and haplotype 7 included Moneymaker. Compared to haplotype 5, haplotypes 6 and 7 carried transposon insertions in one of the introns of SlMES2 and in the intergenic region between SlMES2 and SlMES3. Haplotype 7 was the only haplotype for which all four genes were predicted functional, whereas haplotype 6 carried a premature stop codon in SlMES2, likely rendering this gene non-functional (S4 Table). Therefore, haplotype 6 appeared to be derived from haplotype 7. Haplotype 8 clustered in a separate subclade and lacked SlMES3 as well as a non-functional SlMES4 due to a frameshift (at SL4.0ch02:34,471,277) and three missense (at SL4.0ch02:34,471,287, SL4.0ch02:34,471,546, and SL4.0ch02:34,475,508) mutations in the latter gene. The last haplotype, haplotype 9 consisted of several large deletions at the locus resulting in truncation, internal deletions, and elimination of all four MES genes, and was found only in 13 SLL_CUL accessions (e.g., Rio Grande in S4 Fig). Clearly, the huge variation in haplotype diversity even among the fully wild accessions, suggested that selection predating domestication shaped the genome structure at this locus.
Different colors of accessions and branches represent different MES haplotypes. The outer concentric circle represents the grouping of accessions (SP, SLC and SLL). The numbers on the branches represent the bootstrap values. S. pennellii is used as outgroup.
For SAMT, the genome sequence showed three tandemly duplicated Methyltransferase genes (Solyc09g091530-50). The three genes were found in all accessions and the locus did not show extensive SV except for a single 255 bp deletion (SL4.0ch09: 66,899,792–66,900,047) in the promoter of SlSAMT1 (Solyc09g091550) that was found in 27 accessions. The SAMT1 (Solyc09g091550) is known to regulate methyl salicylate levels in tomato fruit [32]. Since we did not identify any other large structural variants, we proceeded with the haplotype analysis using SNPs and small INDELs at the 21.3 kb locus (S5 Fig and S5 Table). A total of 530 variants were identified and were comprised of 399 SNPs and 131 INDELs. Among them, 83% of the variants were found in regulatory regions. Of the coding region variants, the known SAMT1 gene Solyc09g091550 carried two synonymous and seven missense mutations compared to the Heinz 1706 reference genome for which one or more were found in 34 accessions. This means that the majority of red-fruited accessions in this study encoded the same SAMT1 protein sequence as Heinz1706. For another SAMT gene, Solyc09g091530, ten synonymous, nine missense, two frameshift, and two nonsense mutations were found compared with the Heinz1706 reference genome. The frameshift and nonsense mutations resulted in null alleles that were found in seven SP, three SLC and one SLL. Lastly, we found six synonymous, seven missense and one frameshift mutation in Solyc09g091540. The frameshift mutation was found in only three accessions (2 SLC and 1 SLL) and these also carried the predicted null mutations in Solyc09g091530. The haplotype analysis for SAMT1 using the 530 variants at the 21.3 kb locus showed four alleles (S5A Fig and S3 and S5 Tables). Haplotype I was found in 19 SP, 13 SLC and three SLL_CUL and had the 255 bp deletion in the upstream regulatory region of SlSAMT1 as well as six of the 11 accessions with null mutations in Solyc09g091530. Haplotype II was found in eight SP and two SLC, including two SP with a premature stop codon in Solyc09g091530. Haplotype III appeared in 106 SLC, 18 SLL and 37 SLL_CUL accessions, including three accessions (two SLC and one SLL) with the frameshift mutation in Solyc09g091530 and Solyc09g091540. Haplotype IV was identical to the Heinz1706 reference genome and found in the majority of SLL_CUL. Haplotype I of SlSAMT was the most ancestral haplotype based on the phylogenetic tree constructed from the consensus genomic sequence of the locus and included most of the SP accessions (S6 Fig). Other haplotypes, II and III were found to cluster separately under distinct clades in the phylogenetic tree. The phylogenetic analysis did not include haplotype IV since it was found only in SLL_CUL.
At the UGT5 locus, the only SV was a 403 bp deletion (SL4.0ch01:79,099,508) found in 50 accessions and located 8 kb upstream of the gene. We identified 132 SNPs and 29 small INDELs at the 5.7 kb UGT5 locus compared to the Heinz1706 reference genome. Most of the variants (150) were found in the regulatory regions and only 11 variants (six missense, four synonymous and one frameshift mutation) were found in the gene. The frameshift mutation was a likely null allele and was found in 41 accessions. Haplotype analysis using the variants in the 5.7 kb region resulted into four distinct haplotypes (S7A Fig and S3 and S6 Tables). Haplotype I was found in nine SP accessions and carried one missense and one synonymous mutation in the SlUGT5 compared to Heinz reference genome. Haplotype II was found in 18 SP, 23 SLC and two SLL_CUL. All the accessions in this haplotype carried two missense mutations and a frameshift mutation in the gene compared to Heinz1706. Furthermore, the accessions in haplotype I and II also carried the 403 bp deletion. Haplotype III included the Heinz1706 reference genome and was found in 31 SLC, 11 SLL, and 59 SLL_CUL. Haplotype IV and was the most common haplotype and was found in all accessions except SP. Accessions belonging to haplotype IV were similar to Heinz1706 within the gene region whereas Heinz1706 upstream regulatory region was more similar to haplotype III accessions. Phylogenetic analysis showed haplotype II as the most ancestral haplotype (S8 Fig). Haplotype I was clustered separately in the tree followed by a clade with both haplotypes III and IV.
Association of the haplotypes at MES, NSGT1, SAMT1 and UGT5 loci with methyl salicylate levels
We evaluated the association of the haplotypes at these the loci with methyl salicylate levels, starting with MES and NSGT1. Among the nine MES haplotypes, haplotype 2 produced higher levels of methyl salicylate than the other haplotypes suggesting a loss-of-function allele (Fig 7A). Haplotype 4 and haplotype 9 also produced relatively high levels of methyl salicylate (Haplotype 4: PI129033: 16.61 ng/gfw/hr and haplotype 9: TS-209: 29.36 ng/gfw/hr), however these haplotypes represented only a few accessions (n = 5 in haplotype 4 and n = 13 in haplotype 9). Most accessions with haplotypes 1, 3, 5, 6, 7, or 8 produced low levels of methyl salicylate, with exceptions, suggesting the presence of functional alleles that converted the volatile into salicylic acid. The variation at NSGT1 is represented by five haplotypes [38]. These haplotypes are classified in two groups based on expected functionality: the functional NSGT1 haplotypes I, II and III and the non-functional NSGT1 haplotypes IV and V. The non-functional NSGT1 haplotypes produced significantly higher levels of methyl salicylate compared to the functional NSGT1 haplotype (Fig 7B). The combined effect of the NSGT1 and MES haplotypes showed a strong interaction between the loci. MES haplotype 2 and a non-functional NSGT1 haplotype IV or V resulted in the highest methyl salicylate levels (Fig 7C). Previous results from a biparental cross also supported the notion that MES and NSGT1 strongly interact with respect to the methyl salicylate levels [35]. Other haplotype combinations carried too few accessions to reliably ascertain association with volatile levels, although a trend between haplotype 4 and NSGT1 V suggested that this MES haplotype was non-functional.
(A) MES haplotypes, (B) NSGT1 haplotypes, and (C) across different MES and NSGT1 haplotypes. The letters in the boxplots indicate the significant differences among different haplotypes evaluated by Duncan’s test (α < 0.05).
The four SAMT1 haplotypes were not statistically different in methyl salicylate accumulation (S5B Fig). Even so, SAMT1 haplotype II seemed to produce higher levels of methyl salicylate but was represented by too few accessions to significantly associate volatile levels with the haplotype. The 255 bp deletion showed no association with the volatile levels and thus was unlikely causative (S5C Fig). Interestingly, the 11 accessions carrying a null mutation in Solyc09g091530 were found to produce low levels of methyl salicylate contrary to expectations (S5D Fig). Similar to SAMT1, there was no significant difference among the UGT5 haplotypes (S7B Fig). UGT5 haplotype I showed a trend of higher methyl salicylate levels but was found in only nine accessions which is too few for a meaningful analysis. We also did not find a significant association of the 403 bp deletion with the methyl salicylate levels in fruits (S7C Fig). Overall, we could not show association of a causative variant at SAMT1 and UGT5 with methyl salicylate levels. This may be due in part to the large effect of MES and NSGT1 on volatile levels, and the relatively low number of accessions that carried the same haplotype at the two major loci to allow a thorough characterization of the two minor loci.
Expression of SlMES1-4 in MES locus
All haplotypes at the MES locus, except haplotype 9, carried putative functional copies of one to several MES genes. Especially SlMES1, in all accessions except haplotype 9, appeared to be full length and functional (Fig 5). Since the function of SlMES1 is known [35], we sought to investigate the putative role of other members at the locus in methyl salicylate production in ripe tomato fruit. We investigated their expression by Real-Time Quantitative Reverse Transcription Polymerase Chain Reaction (qRT-PCR) (Fig 8). We selected 37 Varitome accessions that represented eight MES haplotypes (Haplotype 1–8). In accessions belonging to haplotype 1 and 3, SlMES3 was well expressed in ripe tomato fruit, whereas the expression of the other MES genes was low or undetectable. Accessions belonging to these haplotypes produced very low amounts of methyl salicylate, ranging from 0.03 to 1.03 ng/gfw/hr (Fig 7A), implying a functional allele. In haplotype 2, 4, and 8, the expression of all four MES genes was low to non-detectable. Especially haplotypes 2 and 4 were associated with high accumulation of methyl salicylate, likely due to the lack of expression of any of the MES genes. The two accessions representing haplotype 5 showed slightly higher expression of SlMES1, SlMES3 and SlMES4, compared to haplotype 2, 4, and 8. In general for haplotype 6 and 7, SlMES3 was highly expressed followed by SlMES1 and last SlMES4. The expression of SlMES2 was undetectable in nearly all accessions suggesting that this gene was not important for methyl salicylate accumulation in the fruit. As expected, the expression patterns of MES were negatively correlated with the fruit methyl salicylate levels. Haplotypes producing high level of methyl salicylate: haplotype 2 and 4, showed low to non-detectable expression of MES genes. On the other hand, haplotypes that produced low levels of methyl salicylate, haplotypes 1, 3, 6 and 7, SlMES1 and/or SlMES3 were highly expressed in ripe fruits. A summary of MES haplotypes, gene structure, and expression is shown in Table 1.
The bar diagram represents the mean expression normalized by ACT4 expression with ± standard deviation. The number below the accessions represents the MES haplotype.
“+” represents the gene is highly expressed and “-” represents either low or no gene expression. *For haplotype 9, expression levels were predicted based on genome structure of the haplotype.
Evolution of the MES and NSGT1 loci during tomato domestication and selection
SP and SLC accessions that were collected in Ecuador showed higher levels of methyl salicylate whereas SLL and most of SLL_CUL featured low levels of the volatile (Fig 2B). This suggested selection by the terrestrial ecosystem and/or domestication on volatile production, presumably through the MES and NSGT1 loci. To investigate how these two loci were selected during the evolution of tomato, we evaluated the haplotype distribution in the Varitome Collection and cultivated accessions (Fig 9). Most of the SLL_CUL, the majority of SLL, and many SLC carried the MES haplotype 6 and 7, suggesting strong selection for these alleles during domestication (Fig 9A) as these haplotypes were found to produce low levels of methyl salicylate (Fig 7A). The distribution map also showed that haplotypes 6 and 7 spread across several regions from Ecuador, Peru, to Colombia and Mexico. Not surprisingly, the most common allele in Ecuador was MES haplotype 2, associated with high levels of methyl salicylate. For NSGT1, we found the nonfunctional haplotypes IV and V to be the most common haplotypes in SLL_CUL. The proportion of accessions belonging to NSGT1 haplotype IV and V was found to increase from SP (0%) to SLC (41.32%), SLL (44.44%) and SLL_CUL (81.81%) (Fig 9B). Based on the distribution map, haplotype I, II, III had spread across several regions in South and Central America whereas haplotype V was found only in Ecuador and northern Peru. Together, these data suggested that functional NSGT1 (Haplotype I, II, III) were under selection in SP but not in cultivated tomato, whereas the opposite was true for MES. For the two loci combined, it appeared that a strong selection for higher levels of methyl salicylate were particularly important to certain SP and SLC accessions from Ecuador and Peru (Fig 9C), with a non-functional haplotype for both MES and NSGT1 (Fig 9A and 9B). The high levels of methyl salicylate might have contributed to the survival in the specific ecosystem that is common to that region.
Proportion bar stack plot and distribution map of accessions in Varitome Collection (A) MES haplotypes (B) NSGT1 haplotypes. (C) Distribution map of methyl salicylate levels in the accessions. The base layer of the map was generated using R package “rnaturalearth”, which uses the public domain map dataset “Natural Earth” (https://www.naturalearthdata.com/).
Discussion
Methyl salicylate is an important plant volatile that functions as a signaling molecule within and between plants. Its role in defense responses in leaves is relatively well understood. It serves as an herbivore and pathogen-induced volatile that attracts parasitoids and predators of the herbivores. Plants emit methyl salicylate when attacked by herbivores, sending cues to parasitoids and other predators [11,39–41]. Furthermore, methyl salicylate acts as a signaling molecule to activate the genes that are important in regulating the plants’ response to biotic and abiotic stress [35,42–44]. Methyl salicylate also functions as an inter-plant signal that activates disease resistance and the expression of defense-related genes in surrounding plants [39,43]. Even though its role in defense responses is well established, the function of methyl salicylate in ripe fruit is not well understood. We show that the levels of this volatile vary dramatically in red-fruited tomato fruits and are particularly high in accessions originating from Ecuador. This suggests that methyl salicylate in ripe fruits may be beneficial to plants and that certain levels of this volatile in tomato are acceptable to consumers. It is likely that high levels of methyl salicylate in ripe fruits could serve as either an attractant to frugivores or deterrent to insect pests that often function as vectors of pathogenic microbes and viruses. No studies have been conducted on the effect of methyl salicylate levels on fruit health, but external applications and pretreatment of this volatile have been reported to reduce post-harvest losses in tomatoes. Methyl salicylate application reduces chilling injury [45] alleviates post-harvest fungal decay [46,47], and suppresses loss of several desirable aromatic compounds [48]. Hence, even though the volatile is associated with low consumer liking, methyl salicylate could be important in reducing the impact of biotic and abiotic stressors on the plant and plant products such as the fruit.
MES, along with NSGT1 is a major contributor to fruit methyl salicylate levels in the tomato germplasm. Moreover, the interaction between MES and NSGT1 explained most of the variation in fruit volatile levels in a biparental mapping population [35], supporting the findings from this study. The combination of MES haplotype 2 and a non-functional NSGT1 haplotype IV or V resulted in the highest methyl salicylate levels indicating that the interaction of these two loci leads to higher levels of this volatile. Other genes in this pathway, SlSAMT1 and SlUGT5, appeared as minor players in this germplasm. The high or low methyl salicylate level haplogroups of SAMT1 and UGT5 could be explained by their genotype at MES and NSGT1 loci. SAMT1 haplotype II seemed to produce high levels of methyl salicylate. All accessions with SAMT1 haplotype II carried a non-functional MES haplotype (haplotype 2 or 4) which might explain the high volatile levels even though the production of methyl salicylate is dependent on a functional SAMT1. In addition, the 11 accessions carrying a null mutation in Solyc09g091530, whose function in methyl salicylate production is unknown, carried the functional NSGT1 haplotype which could explain the low volatile levels in the fruit. It is likely that there are other fruit-specific glycosyltransferases and methyltransferases with activity on methyl salicylate and salicylic acid respectively, downplaying individual effect of the cloned SlSAMT1 and SlUGT5 [32,33,49]. Additionally, too few accessions remained after considering the effect of the major two loci (MES and NSGT1) to conclude which was the most likely active or inactive allele at the minor loci.
The MES and NSGT1 loci have been shaped predominantly by SVs. It is known that SVs greatly affect gene expression and protein functions resulting in variations in phenotype [38,50]. MES carries several deletions and insertions, spanning the four Methyltransferase genes either partially or completely deleting genes from the locus. These SVs were likely responsible in part for the difference in MES gene expression observed among the haplotypes (Table 1). SlMES1 appeared to be the only functional gene in all haplotypes, except for haplotype 9. However, the expression of this gene was very low or undetected in haplotypes carrying the 363 bp deletion in the promoter of SlMES1 but was higher for haplotype 8 which did not have the deletion. SlMES3 was the highest expressed gene at the MES locus especially in haplotypes 1, 3, 5, 6 and 7. We did not find a correlation between gene expression and transposon insertion or truncation of SlMES3. Instead, we identified a SNP (SL4.0ch02: 34,463,357; C to T) in the first exon of SlMES3 found only in accessions in haplotypes 2 and 4. This SNP is predicted to cause an amino acid change from leucine to phenylalanine at position 83, which is one of the salicylic acid binding sites of the protein [51] and possibly consequential. This finding suggests that SlMES3 is likely functional in addition to SlMES1. For SlMES4, expression levels were associated with the truncation of the protein. Haplotypes 1, 2, 3, 4, and 8 carrying a truncated SlMES4 had low or undetectable expression, whereas the haplotypes 5, 6, and 7 carried full length SlMES4 that showed high expression.
Several events in selection and domestication of cultivated tomatoes have affected the fruit methyl salicylate levels and MES controlling the production of this volatile. Recent evolutionary models proposed by Razifard et al. [28] and Blanca et al. [29] can be compared to the MES haplotype distribution. SP from Peru, considered as the most diverse and ancestral, includes accessions with several MES haplotypes (Fig 10). During its spread north, the tomato plants would have encountered biotic and abiotic conditions associated with higher latitudes, such as changes in temperature, water availability, or disease, which likely shaped the natural selection for MES haplotypes. SP from southern Ecuador carried the functional MES haplotype (7), while SP from northern Ecuador carried the non-functional haplotypes (2 and 4). Non-functional MES haplotype and high levels of methyl salicylate may have contributed to their survival in adverse environments. It is also possible that birds or animal dispersers, having preference for minty flavor, contributed to selection and migration, but little is known about this. However, selections imposed by the new environment necessitates adaptation that might have favored the selection of a specific MES haplotype over the other. According to Blanca et al. [29], SP spread into Mesoamerica where it evolved into SLC and the ancestral SLC migrated back to South America where it hybridized with native SP populations. Following this model, the most functional MES haplotype 7, with potentially four functional genes, was selected as Peruvian and Ecuadorian SP evolved into SLC in Mesoamerica. MES haplotype 6 evolved from haplotype 7, after a premature stop-codon mutation in MES2. Upon migration back to South America, admixture with Ecuadorian SP resulted in SLC from Ecuador carrying the most diverse MES haplotypes. SLC Peru evolved from admixture with Peruvian SP resulting in at least four MES haplotypes, and together with Ecuadorian SLC spread northwards to evolve into ancestral SLL carrying only haplotype 6 and 7. On the other hand, Razifard et al. [28] proposed that SP from northern Ecuador evolved into SLC in South America carrying several domestication alleles which then migrated to Mesoamerica where it evolved into SLL. Based on this model, the functional MES haplotype 7 and its derivative haplotype 6 are found in Peruvian and Ecuadorian SP and were selected early as tomato migrated northwards to evolve into SLL. The evolution in MES locus seems to agree more with the model proposed by Razifard et al. [28] as the linear gene flow model from Peruvian and Ecuadorian SP to Mexican SLL explained the decrease in diversity at the locus. Regardless of the model, large genetic diversity at the MES locus was maintained in South America whereas migration northwards into Mesoamerican SLC and eventually SLL resulted in the selection of only two MES haplotypes 6 and 7. Thus, the evolution at the MES locus suggests a complex migratory and evolutionary path of tomato.
Linear illustration of evolutionary model proposed by A. Blanca et al. [29] and B. Razifard et al. [28]. The numbers in blue represent the MES haplotypes. The gray solid arrow represents the direction of evolution. The abbreviations are PER: Peru, ECU: Ecuador, and MA: Mesoamerica. SLC from PER and ECU includes some accessions which are admixture of SP and SLC.
The genetic variation for the known genes associated with methyl salicylate biosynthesis and conjugation in the Varitome Collection was large, especially at the MES locus. Moreover, even within each distinct haplotype, we observed wide phenotypic variation (Fig 7), suggesting that additional genetic factors control the biosynthesis process. GWAS using the Varitome Collection identified several additional QTL associated with fruit methyl salicylate levels in addition to known genes (Fig 3 and S2 Table). These QTL could harbor novel genes, either affecting the homeostasis of methyl salicylate including additional Glycosyltransferases and Glucosidases. The Varitome Collection thus, is a valuable tool to discover additional genes associated with volatile production through a genetic mapping approach. Identification of the novel genes would allow us to gain a deeper understanding of the regulation of methyl salicylate in the fruits.
The results presented herein contribute to understanding the genetics of the biosynthesis and conjugation of methyl salicylate in ripe tomato fruits. The beneficial haplotypes at the MES and NSGT1 loci could be introgressed into cultivated germplasm to obtain optimal flavor depending on the agroecosystem. Introgression of beneficial haplotypes of NSGT1 has the added advantage of decreasing levels of another undesirable volatile, guaiacol, thereby improving flavor. Thus, the genomic information from this study can be used to develop molecular tools to perform marker-assisted selection for the desirable haplotype at each locus. Intermediate levels of methyl salicylate in modern and heirloom cultivars suggest that certain amounts of this volatile are acceptable or were inadvertently selected because of a strong breeding focus on yield and disease resistance. If the latter, breeding that is focused on the optimal allele selection that favors flavor is now an attainable goal. Thus, the genetic diversity knowledge can be used to fine-tune the methyl salicylate levels to arrive at an acceptable amount of methyl salicylate for overall plant health and resistance as well as consumer satisfaction to balance the desired taste with other post-harvest traits such as shelf life.
Methods
Plant materials and phenotyping
A panel of 166 tomato accessions, the Varitome Collection, consisting of 27 wild SP, 121 semi domesticated SLC and 18 ancestral landrace SLL, as well as 143 heirloom and modern varieties (SLL_CUL) were grown and phenotyped for volatile accumulation as previously described [25,28]. Based on collection site and whole genome information, the species that comprise these accessions were further divided into three SP subpopulations (SP_SECU from Southern Ecuador, SP_NECU from Norther Ecuador and SP_PER from Peru); six SLC subpopulations (SLC_ECU from Ecuador, SLC_PER from Peru, SLC_SM from San Martin region of Peru, SLC_CA from Central, Northern South and Southern North America, SLC_MEX from Mexico, and SLC_ADM which are admixture SLC accessions); and two SLL subpopulations namely SLL comprised of ancestral landraces and SLL_CUL comprised of heirloom and modern accessions. The red ripe fruits were harvested, cut, and loaded in glass tubes. Air was passed over samples for an hour and the volatiles were collected on a SuperQ Resin column. Volatiles were then eluted off the column with methylene chloride and ran on GC-MS for analysis as previously described [25]. The volatile measurements were average of volatile levels from at least two and up to five harvests.
Statistical analysis
All statistical analysis were conducted using R [52] unless mentioned otherwise. The figures were generated using “ggplot2” package [53] in R. The multiple mean comparison was performed using the “duncan.test” function from “agricolae” package in R [54].
Variant calling and genome-wide association study
Raw ILLUMINA read files were downloaded from NCBI (https://www.ncbi.nlm.nih.gov/; SRA: SRP150040, SRP045767, SRP094624, and PRJNA353161). We augmented the original GWAS by Razifard et al. [28] by aligning the reads to the most recent tomato genome SL4.0 build as opposed to SL2.50. We also interrogated the three main variant types, namely SNPs, INDELs and SVs for the GWAS compared to SNPs only [35]. Raw read evaluation, filtering, alignment, and variant calling were performed as previously described [55]. The phenotype deviating from normality (p value from Shapiro test <0.01) was normalized using quantile normalization in R (52). Associations between genotype and phenotype were calculated using the FarmCPU model in GAPIT (version 3) [56] following the established protocols [35]. Association analysis was conducted using 21,893,681 SNPs, 2,735,297 INDELs and 27,477 SVs. The association results were plotted using the “ggplot2” package in R [53].
Haplotype analyses
To identify the haplotypes at the known genes regulating methyl salicylate levels, we first determined the presence of SVs at or near each by comparing the high-quality genome sequences of 28 Varitome accessions (ftp://ftp.solgenomics.net/genomes/tomato100/March_02_2020_sv_landscape/variants/). The genomes of these accessions were sequenced using Oxford Nanopore Technology and Illumina short read sequencing and, therefore, were well suited for the identification of SV at each of the loci. For the loci with evidence of SV, MES and NSGT1, SVs were identified in Integrative Genome Viewer (IGV) [57] on the pattern of short reads in the tomato accessions, and haplotypes were defined largely based on SVs. The size of the MES locus ranged from 15 kb to 35 kb, whereas the size of the NSGT1 locus was as defined previously [38]. For SlSAMT1, the locus comprised three tandem duplicated genes without large SVs. To define the haplotypes at the SAMT1 and UGT5 loci, SNPs and small INDELs within the coding region, 3 kb upstream and 1 kb downstream of the termination of transcription were extracted using VCFtools [58]. The method to generate the haplotypes based on SNPs and INDELs was described previously [55]. Variants (SNPs and INDELs) were annotated using SnpEff [59] using a locally built database of SL4.0 tomato reference genome. The FGENESH [60] tool was used to verify and predict gene models. The haplotype distribution map was created using R package “rnaturalearth” (https://CRAN.R-project.org/package=rnaturalearth) and “ggspatial” (https://CRAN.R-project.org/package=ggspatial). The package “rnaturalearth” uses the public domain map dataset “Natural Earth” (https://www.naturalearthdata.com/) to generate the base layer of the map.
Phylogenetic tree construction of the Methylesterases at MES locus
A phylogenetic tree was built using Neighbor-Joining method with the genomic sequences of genes at the MES locus and 10 kb flanking the locus. The consensus nucleotide sequences were generated using IGV and, exported in Geneious Prime v2021.2.2 (https://www.geneious.com/). The sequences were then aligned using the Clustal alignment option. The tree was constructed using Neighbor-Joining method and Tamura-Nei genetic distance model using S. pennellii (sequence obtained from SGN; https://solgenomics.net/) as an outgroup. Bootstrap support for the tree was obtained using 100 bootstrap replicates. The tree was then plotted using “ggtree” package in R [61].
qRT-PCR
qRT-PCR was performed following a protocol that was used previously [35]. Total RNA was extracted from red ripe fruits of 37 accessions and used for the qRT-PCR, including some that were collected before [35]. ACTIN4 (Solyc04g011500) was used as the control to standardize relative expression levels. The relative expression was calculated using the formula of ΔCq method using a reference gene: R = 2Cq(reference)-Cq(target), where Cq is quantification cycle, reference is ACT4, and targets are MES1, MES2, MES3, or MES4. The expression analysis figure was generated using “ggplot2” package in R [53]. The sequences of primers used in the study are provided in S7 Table.
Data access
All the raw sequencing data used in this study are publicly available in NCBI (https://www.ncbi.nlm.nih.gov/; SRA: SRP150040, SRP045767, SRP094624, and PRJNA353161).
Supporting information
S1 Fig. Genomic alignment of Varitome accessions to SL4.0 genome version at MES locus.
(A) Alignment of three highest methyl salicylate producing accessions against the SL4.0 genome build at the MES locus. (B) Alignment of three least methyl salicylate producing accessions against the SL4.0 genome build at the MES locus. Yellow boxes represent gene models.
https://doi.org/10.1371/journal.pgen.1010751.s001
(PDF)
S2 Fig. ClustwalW alignment of predicted protein sequence of SlMES1-4 of all MES haplotypes.
Alignments of (A) SlMES1 (B) SlMES2 (C) SlMES3 and (D) SlMES4
https://doi.org/10.1371/journal.pgen.1010751.s002
(PDF)
S3 Fig. Comparison of MES haplotype 2 and 5 and representation of a transposon insertions between SlMES2 and SlMES3; and within SlMES3.
The gene model below is a screenshot from SL4.0 genome version of the exact region which is the insertion showing a transposon insertion between SlMES2 and SlMES3.
https://doi.org/10.1371/journal.pgen.1010751.s003
(PDF)
S4 Fig. Alignment of Rio Grande against the SL4.0 genome build at the MES locus.
Alignment for SlMES1, SlMES2, SlMES3 and SlMES4. Yellow boxes represent gene models.
https://doi.org/10.1371/journal.pgen.1010751.s004
(PDF)
S5 Fig. Haplotype analysis of SAMT1 locus.
(A) Heatmap representing the genotypes of accessions (rows) for the polymorphisms identified (columns). Reference genotypes are represented in blue, alternate in red, heterozygous in yellow and missing data in white. The black rectangular box represents the position of the genes in the locus. (B) Distribution of methyl salicylate in red fruits in different accessions among different SAMT1 haplotypes. (C) Distribution of methyl salicylate in red fruits in different accessions with and without deletion in the SAMT1 locus. (D) Distribution of methyl salicylate between accessions with and without null mutations of Solyc09g091530.
https://doi.org/10.1371/journal.pgen.1010751.s005
(PDF)
S6 Fig. Phylogeny tree constructed using consensus DNA sequence of SAMT1 locus and 10 kb flanking region of SAMT1 locus from Varitome collection.
Different colors of accessions and branches represent different SAMT1 haplotypes. The outer two concentric circles represent the grouping of accessions (SP, SLC and SLL) and corresponding methyl salicylate levels (MeSA) in the accessions respectively. The numbers on the branches represent the bootstrap values.
https://doi.org/10.1371/journal.pgen.1010751.s006
(PDF)
S7 Fig. Haplotype analysis of UGT5 locus.
(A) Heatmap representing the genotypes of accessions (rows) for the polymorphisms identified (columns). Reference genotypes are represented in blue, alternate in red, heterozygous in yellow and missing data in white. The black rectangular box represents the position of the gene in the locus. (B) Distribution of methyl salicylate in red fruits in different accessions among different UGT5 haplotypes. (C) Distribution of methyl salicylate in red fruits in different accessions with and without deletion in the promoter of SlUGT5.
https://doi.org/10.1371/journal.pgen.1010751.s007
(PDF)
S8 Fig. Phylogeny tree constructed using consensus DNA sequence of UGT5 locus and 10 kb flanking region of UGT5 locus from Varitome collection.
Different colors of accessions and branches represent different UGT5 haplotypes. The outer two concentric circles represent the grouping of accessions (SP, SLC and SLL) and corresponding methyl salicylate levels (MeSA) in the accessions respectively. The numbers on the branches represent the bootstrap values.
https://doi.org/10.1371/journal.pgen.1010751.s008
(PDF)
S1 Table. List of significant loci associated with methyl salicylate levels and their respective variants from SNP, INDEL and SV GWAS, their position, p-value, LOD values and cloned genes.
https://doi.org/10.1371/journal.pgen.1010751.s009
(XLSX)
S2 Table. List of SVs identified in MES locus and used for defining MES haplotypes.
https://doi.org/10.1371/journal.pgen.1010751.s010
(XLSX)
S3 Table. Methyl salicylate levels (ng/gfw/hr) in Varitome collection and heirlooms and their respective MES and NSGT1 haplotype.
https://doi.org/10.1371/journal.pgen.1010751.s011
(XLSX)
S4 Table. Annotation of all the variants (SNPs and INDELs) in the MES locus in Varitome collection.
The MES haplotype information is provided in the column besides the accession names. The annotation information is provided along each variant. 0|0 represents reference allele, 1|1 represents alternate allele, and 1|0 represents heterozygous allele. The yellow highlighted variants are the variants discussed in the text.
https://doi.org/10.1371/journal.pgen.1010751.s012
(XLSX)
S5 Table. Annotation of all the variants (SNPs and INDELs) in the SAMT1 locus in Varitome collection.
The SAMT1 haplotype information is provided in the column besides the accession names. The annotation information is provided along each variant. 0|0 represents reference allele, 1|1 represents alternate allele, and 1|0 represents heterozygous allele. The yellow highlighted variants are the variants discussed in the text.
https://doi.org/10.1371/journal.pgen.1010751.s013
(XLSX)
S6 Table. Annotation of all the variants (SNPs and INDELs) in the UGT5 locus in Varitome collection.
The UGT5 haplotype information is provided in the column besides the accession names. The annotation information is provided along each variant. 0|0 represents reference allele, 1|1 represents alternate allele, and 1|0 represents heterozygous allele. The yellow highlighted variants are the variants discussed in the text.
https://doi.org/10.1371/journal.pgen.1010751.s014
(XLSX)
S7 Table. qRT PCR primers for the expression analysis of SlMES1-4 and ACT4.
https://doi.org/10.1371/journal.pgen.1010751.s015
(XLSX)
S8 Table. Raw data to generate Figures 4, 6, 8, and 9.
The data are presented in different worksheets.
https://doi.org/10.1371/journal.pgen.1010751.s016
(XLSX)
Acknowledgments
We thank the Georgia Advanced Computing Resource Center, UGA, for providing the computation resources.
References
- 1. Raguso RA. Wake up and smell the roses: the ecology and evolution of floral scent. Annual review of ecology, evolution, and systematics. 2008:549–69.
- 2. Schiestl FP. Ecology and evolution of floral volatile-mediated information transfer in plants. New Phytologist. 2015;206(2):571–7. pmid:25605223
- 3. Pichersky E, Gershenzon J. The formation and function of plant volatiles: perfumes for pollinator attraction and defense. Current opinion in plant biology. 2002;5(3):237–43. pmid:11960742
- 4. Borges RM, Ranganathan Y, Krishnan A, Ghara M, Pramanik G. When should fig fruit produce volatiles? Pattern in a ripening process. Acta Oecologica. 2011;37(6):611–8.
- 5. Rodríguez A, Alquézar B, Peña L. Fruit aromas in mature fleshy fruits as signals of readiness for predation and seed dispersal. New Phytologist. 2013;197(1):36–48. pmid:23127167
- 6. Fallik E, Archbold DD, Hamilton-Kemp TR, Clements AM, Collins RW, Barth MM. (E)-2-Hexenal can stimulate Botrytis cinerea growth in vitro and on strawberries in vivo during storage. Journal of the American Society for Horticultural Science. 1998;123(5):875–81.
- 7. Rodríguez A, San Andrés V, Cervera M, Redondo A, Alquézar B, Shimada T, et al. Terpene down-regulation in orange reveals the role of fruit aromas in mediating interactions with insect herbivores and pathogens. Plant Physiology. 2011;156(2):793–802. pmid:21525333
- 8. Karban R, Shiojiri K, Huntzinger M, McCall AC. Damage-induced resistance in sagebrush: volatiles are key to intra-and interplant communication. Ecology. 2006;87(4):922–30. pmid:16676536
- 9. Qualley A, Dudareva N. Plant volatiles. eLS. 2010.
- 10. De Moraes CM, Mescher MC, Tumlinson JH. Caterpillar-induced nocturnal plant volatiles repel conspecific females. Nature. 2001;410(6828):577–80. pmid:11279494
- 11. Kessler A, Baldwin IT. Defensive function of herbivore-induced plant volatile emissions in nature. Science. 2001;291(5511):2141–4. pmid:11251117
- 12. Andersen RA, Hamilton-Kemp TR, Hildebrand DF, McCracken CT Jr, Collins RW, Fleming PD. Structure-antifungal activity relationships among volatile C6 and C9 aliphatic aldehydes, ketones, and alcohols. Journal of Agricultural and Food Chemistry. 1994;42(7):1563–8.
- 13. Vancanneyt G, Sanz C, Farmaki T, Paneque M, Ortego F, Castañera P, et al. Hydroperoxide lyase depletion in transgenic potato plants leads to an increase in aphid performance. Proceedings of the national academy of sciences. 2001;98(14):8139–44. pmid:11416166
- 14. Aartsma Y, Bianchi FJ, van der Werf W, Poelman EH, Dicke M. Herbivore-induced plant volatiles and tritrophic interactions across spatial scales. New Phytologist. 2017;216(4):1054–63. pmid:28195346
- 15. Bouwmeester H, Schuurink RC, Bleeker PM, Schiestl F. The role of volatiles in plant communication. The Plant Journal. 2019;100(5):892–907. pmid:31410886
- 16. Gasmi L, Martínez-Solís M, Frattini A, Ye M, Collado MC, Turlings TC, et al. Can herbivore-induced volatiles protect plants by increasing the herbivores’ susceptibility to natural pathogens? Applied and environmental microbiology. 2019;85(1):e01468–18. pmid:30366995
- 17. Couty A, Kaiser L, And DH, Pham-Delegue MH. The attractiveness of different odour sources from the fruit–host complex on Leptopilina boulardi, a larval parasitoid of frugivorous Drosophila spp. Physiological Entomology. 1999;24(1):76–82.
- 18. Melo JWS, Lima DB, Pallini A, Oliveira JEM, Gondim MG. Olfactory response of predatory mites to vegetative and reproductive parts of coconut palm infested by Aceria guerreronis. Experimental and Applied Acarology. 2011;55(2):191–202. pmid:21499777
- 19. Mäntylä E, Alessio GA, Blande JD, Heijari J, Holopainen JK, Laaksonen T, et al. From plants to birds: higher avian predation rates in trees responding to insect herbivory. PLoS One. 2008;3(7):e2832. pmid:18665271
- 20. Tieman D, Bliss P, McIntyre LM, Blandon-Ubeda A, Bies D, Odabasi AZ, et al. The chemical interactions underlying tomato flavor preferences. Current Biology. 2012;22(11):1035–9. pmid:22633806
- 21. Rambla JL, Tikunov YM, Monforte AJ, Bovy AG, Granell A. The expanded tomato fruit volatile landscape. Journal of Experimental Botany. 2014;65(16):4613–23. pmid:24692651
- 22. Martina M, Tikunov Y, Portis E, Bovy AG. The genetic basis of tomato aroma. Genes. 2021;12(2):226. pmid:33557308
- 23. Dudareva N, Pichersky E, Gershenzon J. Biochemistry of plant volatiles. Plant physiology. 2004;135(4):1893–902. pmid:15326281
- 24.
Hui YH, Chen F, Nollet LM, Guiné RP, Martín-Belloso O, Mínguez-Mosquera MI, et al. Handbook of fruit and vegetable flavors: Wiley Online Library; 2010.
- 25. Tieman D, Zhu G, Resende MF Jr, Lin T, Nguyen C, Bies D, et al. A chemical genetic roadmap to improved tomato flavor. Science. 2017;355(6323):391–4. pmid:28126817
- 26. Särkinen T, Bohs L, Olmstead RG, Knapp S. A phylogenetic framework for evolutionary study of the nightshades (Solanaceae): a dated 1000-tip tree. BMC evolutionary biology. 2013;13(1):1–15. pmid:24283922
- 27. Pease JB, Haak DC, Hahn MW, Moyle LC. Phylogenomics reveals three sources of adaptive variation during a rapid radiation. PLoS biology. 2016;14(2):e1002379. pmid:26871574
- 28. Razifard H, Ramos A, Della Valle AL, Bodary C, Goetz E, Manser EJ, et al. Genomic evidence for complex domestication history of the cultivated tomato in Latin America. Molecular biology and evolution. 2020;37(4):1118–32. pmid:31912142
- 29. Blanca J, Sanchez-Matarredona D, Ziarsolo P, Montero-Pau J, van der Knaap E, Díez MJ, et al. Haplotype analyses reveal novel insights into tomato history and domestication driven by long-distance migrations and latitudinal adaptations. Horticulture Research. 2022;9. pmid:35184177
- 30. Klee HJ, Tieman DM. The genetics of fruit flavour preferences. Nature Reviews Genetics. 2018:1. pmid:29563555
- 31. Bruhn CM, Feldman N, GARLITZ C, Harwood J, Ivans E, Marshall M, et al. Consumer perceptions of quality: apricots, cantaloupes, peaches, pears, strawberries, and tomatoes. Journal of food quality. 1991;14(3):187–95.
- 32. Tieman D, Zeigler M, Schmelz E, Taylor MG, Rushing S, Jones JB, et al. Functional analysis of a tomato salicylic acid methyl transferase and its role in synthesis of the flavor volatile methyl salicylate. The Plant Journal. 2010;62(1):113–23. pmid:20070566
- 33. Louveau T, Leitao C, Green S, Hamiaux C, Van der Rest B, Dechy-Cabaret O, et al. Predicting the substrate specificity of a glycosyltransferase implicated in the production of phenolic volatiles in tomato fruit. The FEBS journal. 2011;278(2):390–400. pmid:21166996
- 34. Tikunov YM, Molthoff J, de Vos RC, Beekwilder J, van Houwelingen A, van der Hooft JJ, et al. Non-smoky glycosyltransferase1 prevents the release of smoky aroma from tomato fruit. The Plant Cell. 2013;25(8):3067–78. pmid:23956261
- 35. Frick EM, Sapkota M, Pereira Garcia L, Wang Y, Hermanns A, Giovannoni JJ, et al. A family of methyl esterases converts methyl salicylate to salicylic acid in ripening tomato fruit. Plant Physiology. 2023. pmid:36315067
- 36. Rivas-San Vicente M, Plasencia J. Salicylic acid beyond defence: its role in plant growth and development. Journal of experimental botany. 2011;62(10):3321–38. pmid:21357767
- 37. Baldwin IT. Plant volatiles. Current Biology. 2010;20(9):R392–R7. pmid:20462477
- 38. Alonge M, Wang X, Benoit M, Soyk S, Pereira L, Zhang L, et al. Major impacts of widespread structural variation on gene expression and crop improvement in tomato. Cell. 2020;182(1):145–61. e23. pmid:32553272
- 39. Shulaev V, Silverman P, Raskin I. Airborne signalling by methyl salicylate in plant pathogen resistance. Nature. 1997;385(6618):718–21.
- 40. De Boer JG, Dicke M. The role of methyl salicylate in prey searching behavior of the predatory mite Phytoseiulus persimilis. Journal of chemical ecology. 2004;30(2):255–71. pmid:15112723
- 41. Zhu J, Park K-C. Methyl salicylate, a soybean aphid-induced plant volatile attractive to the predator Coccinella septempunctata. Journal of chemical ecology. 2005;31(8):1733–46. pmid:16222805
- 42. Hardie J, Isaacs R, Pickett JA, Wadhams LJ, Woodcock CM. Methyl salicylate and (−)-(1R, 5S)-myrtenal are plant-derived repellents for black bean aphid, Aphis fabae Scop.(Homoptera: Aphididae). Journal of chemical ecology. 1994;20(11):2847–55. pmid:24241919
- 43. Park S-W, Kaimoyo E, Kumar D, Mosher S, Klessig DF. Methyl salicylate is a critical mobile signal for plant systemic acquired resistance. Science. 2007;318(5847):113–6. pmid:17916738
- 44. Gao H, Guo M, Song J, Ma Y, Xu Z. Signals in systemic acquired resistance of plants against microbial pathogens. Molecular Biology Reports. 2021;48(4):3747–59. pmid:33893927
- 45. Ding C-K, Wang CY, Gross KC, Smith DL. Reduction of chilling injury and transcript accumulation of heat shock proteins in tomato fruit by methyl jasmonate and methyl salicylate. Plant Science. 2001;161(6):1153–9.
- 46. Zhang X, Min D, Li F, Ji N, Meng D, Li L. Synergistic effects of L-arginine and methyl salicylate on alleviating postharvest disease caused by Botrysis cinerea in tomato fruit. Journal of agricultural and food chemistry. 2017;65(24):4890–6. pmid:28535671
- 47. Min D, Li F, Zhang X, Shu P, Cui X, Dong L, et al. Effect of methyl salicylate in combination with 1-methylcyclopropene on postharvest quality and decay caused by Botrytis cinerea in tomato fruit. Journal of the Science of Food and Agriculture. 2018;98(10):3815–22. pmid:29352462
- 48. Zeng X, Wang L, Fu Y, Zuo J, Li Y, Zhao J, et al. Effects of methyl salicylate pre-treatment on the volatile profiles and key gene expressions in tomatoes stored at low temperature. Frontiers in Nutrition. 2022;9. pmid:36276839
- 49.
Kamiyoshihara Y, Tieman DM, Klee HJ. Analyses of Plant UDP-Dependent Glycosyltransferases to Identify Their Volatile Substrates Using Recombinant Proteins. Plant Signal Transduction: Springer; 2016. p. 199–207. https://doi.org/10.1007/978-1-4939-3115-6_16 pmid:26577791
- 50. Chiang C, Scott AJ, Davis JR, Tsang EK, Li X, Kim Y, et al. The impact of structural variation on human gene expression. Nature genetics. 2017;49(5):692–9. pmid:28369037
- 51. Zubieta C, Ross JR, Koscheski P, Yang Y, Pichersky E, Noel JP. Structural basis for substrate recognition in the salicylic acid carboxyl methyltransferase family. The Plant Cell. 2003;15(8):1704–16. pmid:12897246
- 52.
R Core Team. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2021. https://www.R-project.org/
- 53.
Wickham H. ggplot2. Wiley Interdisciplinary Reviews: Computational Statistics. 2011;3(2):180–5. https://ggplot2.tidyverse.org.
- 54. de Mendiburu F, de Mendiburu MF. Package ‘agricolae’. R Package, version. 2019;1(3). https://CRAN.R-project.org/package=agricolae
- 55. Pereira L, Sapkota M, Alonge M, Zheng Y, Zhang Y, Razifard H, et al. Natural genetic diversity in tomato flavor genes. Frontiers in plant science. 2021;12:914. pmid:34149747
- 56. Wang J, Zhang Z. GAPIT Version 3: boosting power and accuracy for genomic association and prediction. Genomics, proteomics & bioinformatics. 2021;19(4):629–40. pmid:34492338
- 57. Robinson JT, Thorvaldsdóttir H, Winckler W, Guttman M, Lander ES, Getz G, et al. Integrative genomics viewer. Nature biotechnology. 2011;29(1):24–6. pmid:21221095
- 58. Danecek P, Auton A, Abecasis G, Albers CA, Banks E, DePristo MA, et al. The variant call format and VCFtools. Bioinformatics. 2011;27(15):2156–8. pmid:21653522
- 59. Cingolani P, Platts A, Wang LL, Coon M, Nguyen T, Wang L, et al. A program for annotating and predicting the effects of single nucleotide polymorphisms, SnpEff: SNPs in the genome of Drosophila melanogaster strain w1118; iso-2; iso-3. Fly. 2012;6(2):80–92. pmid:22728672
- 60. Solovyev V, Kosarev P, Seledsov I, Vorobyev D. Automatic annotation of eukaryotic genes, pseudogenes and promoters. Genome biology. 2006;7(1):1–12. pmid:16925832
- 61. Yu G, Smith DK, Zhu H, Guan Y, Lam TTY. ggtree: an R package for visualization and annotation of phylogenetic trees with their covariates and other associated data. Methods in Ecology and Evolution. 2017;8(1):28–36.