INVITED REVIEW
Sugar Metabolism and Fruit Development in the Tomato
2017 Volume 86 Issue 4 Pages 417-425
Details
2017 Volume 86 Issue 4 Pages 417-425
Sugars are strongly related to fruit yield and quality, playing a critical role in fruit set, growth, ripening, and composition. The tomato (Solanum lycopersicum) is not only an important horticultural crop, but also a useful experimental model plant that can be used to further our understanding of fruit physiology. Therefore, in this review, we consider sugar metabolism and fruit development in the tomato. We begin by discussing how the sugar content of tomato fruit has been successfully increased in a tomato introgression line containing a chromosome segment from a wild relative, and how this has furthered our understanding of the mechanism controlling sugar content. We then outline current knowledge around how sugar sensing and signaling, proton pumps, and auxin affect sugar accumulation and fruit set. The prevention of fruit abscission by auxin, which is transported by PIN auxin efflux carriers and vacuolar proton phyrophosphatase (V-PPase), may retain sucrose transport to the fruit to inhibit programed cell death (PCD) and ensure successful fruit set. There is believed to be a trade-off between fruit sugar content and yield. However, fruit size and yield do not appear to decrease in the tomato introgression line IL8-3 and sucrose-induced repression of translation (SIRT)-engineered tomatoes, which contain higher fruit sugar contents. Future research needs to investigate the factors involved in sugar sensing and signaling, in addition to the sugar metabolic enzymes that have previously been studied for horticultural applications.
In fruit, sugars provide sweetness, which is the most important determinant of fruit quality. The relationship between sugar content, measured as the soluble solids content, and fruit yield has been evaluated in a number of horticultural studies (e.g., Zhang et al., 2015). Sugars are also important in the generation of turgor pressure to promote fruit cell expansion, and as signal molecules to control fruit development and metabolism, although we often tend to forget these additional functions. Thus, sugars are closely related to fruit yield and quality, playing a critical role in fruit set, growth, ripening, and composition.
Because sugar is of such great importance to both horticulture and plant physiology, a large number of studies have investigated sugar transport, metabolism, and accumulation in fruit. Arabidopsis is often used as a model plant for understanding general plant physiology, the accumulated knowledge from which allows other plant species to be analyzed more easily. Therefore, a model plant should also be used to investigate fruit physiology to facilitate research on other fruit species. The tomato (Solanum lycopersicum) is a useful model for plants with fleshy fruits due to its relatively small genome size and short generation time, and the availability of a routine transformation technique (Giovannoni, 2004). Furthermore, the whole genome sequence has already been reported for the tomato (The Tomato Genome Consortium, 2012), and genetic resources and databases are available for this species, such as the National BioResource Project Tomato (http://tomato.nbrp.jp/indexEn.html; Shikata et al., 2016) and the Sol Genomics Network (https://solgenomics.net).
The tomato is also an important horticultural crop, having the largest production volume (t) globally (FAOSTAT) and the highest economic value in Japan (Ministry of Agriculture, Forestry and Fisheries) of all fruits and vegetables. Therefore, in this review, we focus on the tomato to examine the current state of knowledge regarding the relationship between sugar content and fruit development. We begin by describing how tomato fruit with increased sugar content have been produced using introgression lines containing a chromosome segment from a wild relative and what these can tell us about the mechanism controlling sugar content. We then discuss how sugar sensing and signaling, proton pumps, and auxin affect sugar accumulation and fruit set. It should be noted that the tomato and several other fruit crops produce sucrose as a primary photosynthate, whereas rosaceous fruit trees such as those of apple, pear, and peach synthesize and translocate sorbitol. Sorbitol metabolism is not covered in this review, but further information can be found in Kanayama (1998, 2009). In addition, the relationship between sugars and stress responses, including the increase in fruit sugar content during cultivation under stress conditions, is not considered in this review, further details around which can be found in articles cited in Kanayama and Kochetov (2015).
Fruit sugar content and fruit size are important determinants of yield and quality, and are controlled by quantitative trait loci (QTLs). The large fruit size of commercial tomato varieties has resulted from increases in cell division and the number of locules, which are controlled by several QTLs. The genes underlying the QTLs controlling fruit size have already been cloned, including the negative regulator of cell division FW2.2 (Frary et al., 2000), the YABBY-like transcription factor FASCIATED (Cong et al., 2008), and the WUSCHEL that encodes a homeodomain protein that regulates stem cell fate, LOCULE NUMBER (Muños et al., 2011). However, the QTLs that control the sugar content of fruit remain unconfirmed.
In the tomato, fruit sugar content is known to be controlled by Brix9-2-5, which encodes the functional amino acid polymorphism of cell wall invertase (LIN5). This plays a key role in fruit sink strength, affecting sugar content through the alteration of kinetics (Fridman et al., 2000, 2004). Plant cells contain two types of acid invertases–cell wall invertase and vacuolar invertase. Cell wall invertase facilitates sugar uptake into fruit cells via the hexose transporter by hydrolyzing sucrose into hexose in the apoplast (Gear et al., 2000), while vacuolar invertase is known to determine the sugar composition (sucrose/hexose ratio) via a sucrolytic reaction in the vacuole in some fruits (Itai et al., 2015). However, the genes and physiological mechanisms that underly the promotion of sink strength are not yet well understood for the other QTLs that control fruit sugar content.
Some wild relatives of the tomato have high fruit sugar content, and so this genetic variation may be useful for investigating and developing QTLs that control fruit sugar content. Since there is large genetic variation between domesticated tomatoes and their wild relatives (e.g., two times higher sugar content in one wild relative; data not shown), cross breeding can be used to develop tomato cultivars with high sugar content. There are, however, some issues with such cross breeding, as the genomes of wild relatives contain some unfavorable traits and there are many interacting QTLs that determine fruit sugar content, resulting in complicated segregation. Consequently, to address these issues, a set of tomato introgression lines (ILs) has been produced by crossing Solanum lycopersicum with its wild relative S. pennellii (Eshed and Zamir, 1995), each of which contains only a single short chromosome segment from S. pennellii in the background of the S. lycopersicum genome. Thus, unfavorable traits from the wild relative have mostly been removed and the high sugar content trait on the chromosome segment can be simply analyzed as a single genetic factor.
Some of these ILs produce fruit containing a higher sugar content than their parental cultivar M82 (Baxter et al., 2005b). In fact, the sugar content after pyramiding three independent ILs was more than 50% higher than that of M82 (Gur and Zamir, 2004). The line IL8-3 contains one of the major QTLs for sugar content and it has been shown that the increased sugar content in its ripe fruit results from an increased hexose content, likely caused by high activity of ADP-glucose pyrophosphorylase (AGPase), which is a key enzyme in starch biosynthesis in plants (Ballicora et al., 2004), and the enhanced accumulation of starch during the middle stage of fruit development (Ikeda et al., 2013). Other studies have also shown that starch accumulation in young fruit affects the sugar content of ripe fruit based on salt stress treatment experiments (Yin et al., 2010) and the introgression of the AGPase gene from another wild tomato relative, S. habrochaites (Petreikov et al., 2006; Schaffer et al., 2000). In contrast, both the hexose and sucrose contents are increased in the line IL9-2-5, in which Brix9-2-5 has been studied in detail, as described above (Baxter et al., 2005a), while the high sugar content in another wild relative, S. chmielewskii, results from enhanced sucrose accumulation due to low vacuolar invertase activity in the fruit (Yelle et al., 1991). Therefore, it is likely that the mechanism that leads to the increase in sugar content in IL8-3 fruit is different from those occurring in IL9-2-5 and S. chmielewskii. In contrast, the hexose and starch contents were also higher in IL4-4, IL7-3 (Baxter et al., 2005b), and the introgression line containing the S. habrochaites AGPase gene, as described above, and so these lines may have the same mechanism as IL8-3.
Tomato metabolome analyses have been performed using the cultivated varieties, wild relatives, and transgenic plants, and comparative analyses using tomato ILs have successfully detected QTLs involved in fruit metabolism (Overy et al., 2005; Schauer et al., 2006). Comparative transcriptome analyses have also been performed using tomato ILs to investigate which genes are involved in secondary metabolite accumulation (Di Matteo et al., 2013; Lee et al., 2012). An integrated omics approach, which includes transcriptome and metabolome analysis, has been shown to be useful for investigating the regulation of fruit ripening and secondary metabolite accumulation in tomato fruit (Lee et al., 2012). Most omics studies in the tomato have focused on only a single stage of fruit development, particularly the ripe stage. However, omics analysis needs to be conducted throughout fruit development due to the potential interactions between the different stages; for example, the fruit sugar content at the ripe stage may be affected by the fruit starch content at an early developmental stage (Dinar and Stevens, 1981; Ikeda et al., 2013; Petreikov et al., 2006). Therefore, Ikeda et al. (2016) used an integrated omics approach in IL8-3 and its parental cultivar M82 from 10 days after flowering to ripening. Interestingly, they detected markedly different metabolite patterns not only at the ripe stage, when the sugar content was higher in IL8-3 than M82, but also at 20 days after flowering, whereas the metabolite patterns were similar between the two genotypes at other stages of fruit development (10 and 30 days after flowering). However, the transcript patterns did not exhibit any clear difference between the two genotypes at any stage of fruit development. A previous report that considered environmental effects on tomato fruit metabolism showed that enzyme activities are more stable than metabolite levels, which are sensitive to environmental changes (Biais et al., 2014). Therefore, the dynamic changes in metabolites may be caused by a limited number of factors associated with transcript levels or enzyme activities.
The cell wall invertase LIN6 and the sucrose synthase TOMSSF have been shown to be important in the dynamic regulation of carbohydrate metabolism in IL8-3 fruit (Ikeda et al., 2016), and exhibit higher levels of their transcripts in IL8-3 than its parental cultivar M82 during the early stage of fruit development. LIN6 and TOMSSF are major members of the cell wall invertase and sucrose synthase families in fruit, which are known to be key enzymes involved in fruit sink strength (D’Aoust et al., 1999; Fridman et al., 2000). Since these enzymes are not located in the IL8-3 region of the tomato chromosome, their transcription is likely enhanced by transcription factors that are located on a chromosome segment that was narrowed through map-based cloning (Ikeda et al., 2016; Fig. 1).
Proposed model for the increased sugar and amino acid contents of IL8-3 tomato fruit. IL8-3 and its parental cultivar M82 exhibit quite different metabolite patterns at 20 days after flowering and the ripe stage. A putative transcription factor in the restricted chromosome region in IL8-3 promotes gene expression of the cell wall invertase LIN6 and the sucrose synthase TOMSSF, which play a key role in sink metabolism, resulting in the accumulation of starch in fruit at 20 days after flowering. Starch accumulation during development results in an increased sugar content at the ripe stage. Furthermore, sugar promotes the expression of the glutamate synthase gene (SlGOGAT), which plays a role in the first step of amino acid production, leading to an increase in the contents of glutamate (Glu) and other amino acids at the ripe stage.
The amino acid content of tomato fruit is increased alongside sugar content in plants grown under salt stress conditions (Zushi and Matsuzoe, 2011) and in transgenic plants in which sucrose-induced repression of translation (SIRT) is genetically modified, as described later in this article (Sagor et al., 2016). The amino acid content is also higher in the ripe fruit of IL8-3 as a result of the high sugar content enhancing the expression of the glutamate synthase gene SlGOGAT, which plays a key role in amino acid synthesis (Ikeda et al., 2016). The higher content of amino acids, especially glutamate, in IL8-3 fruit is significant, as amino acids are known to be determinants of “umami”, which is one of the five basic tastes. The finding that the middle stage of fruit development (20 days after flowering) is an important determinant of the high sugar content at the ripe stage in IL8-3 is useful information for tomato cultivation.
The genetic manipulation of sugar sensors and signaling, as well as sugar-metabolizing enzymes, could be useful for enhancing sugar accumulation and regulating sugar composition, although much less progress has been made in this area compared with the genetic manipulation of sugar-metabolizing enzymes. Although we do not yet have a comprehensive understanding of sugar sensors and signaling in plant cells, some sensor and signaling systems have been identified (Li and Sheen, 2016). One group of these systems triggers signaling via direct sugar-binding (e.g., hexokinase, HXK), while the other indirectly modulates signaling proteins via sugar-derived bioenergetics molecules and metabolites (e.g., sucrose-non-fermentation-related protein kinase 1, SnRK1).
HXK, which is a well-known sugar sensor, plays a role in inhibiting the expression of photosynthesis-related genes (Jang and Sheen, 1994). It has dual independent functions in hexose phosphorylation and glucose sensing. Interestingly, HXK forms a complex with vacuolar H+-ATPase B1 (VHA-B1) and the 19S regulatory particle of proteasome subunit (RPT5B) to modulate the transcription of specific genes by binding their promoter region in the nucleus (Cho et al., 2006). However, the sensor and signaling systems of the other metabolic sugars, sucrose and fructose, remain relatively unknown.
A unique sucrose signaling mechanism has been reported that involves upstream open reading frames (uORF) (Wiese et al., 2004). A typical eukaryotic mRNA contains one translation start site to produce a single functional protein; however, alternative translation start sites are also recognized to form alternative open reading frames containing uORFs, which are located in the 5′-untranslated region (Kochetov, 2008). Although the roles of these uORFs are not fully understood, they may affect translation. One such uORF-mediated regulatory mechanism of translation is sucrose-induced repression of translation (SIRT), in which the translation of the normal ORF of a basic region leucine zipper (bZIP) transcription factor from Arabidopsis, AtbZIP11, is repressed by sucrose (Wiese et al., 2004).
An ortholog of AtbZIP11 in tobacco plants also exhibits uORF-mediated SIRT, and the introduction of the bZIP gene without uORF increases the sucrose concentration by releasing SIRT (Thalor et al., 2012); however, these transgenic tobacco plants also exhibited growth retardation. Therefore, a tomato homolog of the bZIP gene that lacked the uORF was expressed in tomato fruit using a ripe fruit-specific E8 promoter, leading to the fruit sugar concentration being successfully increased without growth retardation (Sagor et al., 2016). Interestingly, SIRT is sucrose-specific, and so is not sensitive to other sugars such as glucose and fructose. To the best of our knowledge, this is the first report on the sugar sensing-related engineering of fruit sugar metabolism, and the reported similar yield between the engineered and control plants suggests that this type of metabolic engineering is a promising strategy for fruit production in the future. Furthermore, in this case, the E8 promoter was used, which is ripening specific and activated only at the late stage of fruit development; therefore, the use of other fruit-specific promoters that function from earlier stages of fruit development, such as 2A11 (Amemiya et al., 2005), may result in even greater sugar accumulation in the fruit. It is likely that bZIP indirectly activates sugar accumulation through up-regulation of the amino acid synthesis-related gene, which is the transcriptional target of bZIP (Sagor et al., 2016). Since amino acids increase alongside sugars in IL8-3 fruit and in fruit grown under stress conditions to produce sweet tomatoes, as described above, the interaction between sugar and amino acid metabolism is of particular interest and the positive regulation of the metabolism of each may be useful in fruit production. Although sucrose sensors have not been identified to date, sucrose transporters, membrane-associated sucrose synthase, and other proteins that are similar to these will be investigated as potential sucrose sensors in the future (Li and Sheen, 2016).
Since fructose is sweeter than the other major sugars, sucrose and glucose, metabolic engineering to increase fructose concentration has been attempted at two different points in the sugar metabolic pathway. The first of these targeted fructokinase (FRK), which is essential for fructose metabolism following the production of fructose and hexose from sucrose that has been unloaded into fruit cells. FRK is known to be the major enzyme involved in phosphorylating fructose to fructose-6-phosphate, which is utilized in starch biosynthesis, and plants contain several FRK genes with different expression patterns and physiological roles (Granot et al., 2014; Kanayama et al., 1997; Mukherjee et al., 2015). Thus, FRK gene expression has been suppressed in an attempt to increase the fructose concentration in tomato fruit (Odanaka et al., 2002). The second approach introduced the sorbitol cycle into sucrose-translocating plants, which utilize sucrose to translocate photosynthates from the leaves to sink tissues. The conversion between glucose and fructose is usually mediated by phosphoglucoisomerase following the phosphorylation of both hexoses to hexose-6-phosphate. Therefore, based on the idea that fructose accumulates in rosaceous fruits via sorbitol dehydrogenase forming fructose from sorbitol, which is the translocatable sugar in these plants (Kanayama et al., 2005; Suzuki et al., 2001), the sorbitol dehydrogenase gene was introduced into sucrose-translocating plants along with the sorbitol-6-phosphate dehydrogenase gene, which plays a role in sorbitol synthesis from glucose-6-phosphate (Deguchi et al., 2006). However, the metabolic engineering at both regulation points resulted in an increase in sucrose concentration rather than fructose, suggesting that fructose sensing and homeostasis exist in plant cells. This also indicates that the engineering of fructose metabolism could be useful for increasing sucrose concentration. Fructose 1,6-phosphatase and FRK-like protein have been proposed as candidate fructose signaling molecules in plant cells (Cho and Yoo, 2011; Gilkerson et al., 2012). Thus, releasing this regulation may lead to an increase in fructose concentration because the genetic modification of sucrose sensing (SIRT) successfully increases sugar concentration in fruit, as described above. Developing sweet fruit by engineering the sugar composition to produce a higher ratio of “sweeter” fructose to glucose without affecting the total sugar concentration seems to be a promising strategy because the modification of sugar composition may be easier than increasing the total sugar content.
There are two proton pumps in the vacuolar membranes of plant cells: vacuolar proton ATPase (V-ATPase) and vacuolar proton phyrophosphatase (V-PPase), which generate an electrochemical gradient to transport metabolites including sugars and organic acids through the vacuolar membrane (Maeshima, 2000). Fruit-specific suppression of the expression of the subunit gene for V-ATPase has previously been shown to affect fruit growth and sugar composition (Amemiya et al., 2005). However, it has also been proposed that V-PPase contributes to sugar and organic acid accumulation in fruit cells in grapes, pears, and peaches (Etienne et al., 2002; Suzuki et al., 2000; Venter et al., 2006). Therefore, a detailed expression analysis was performed in tomato fruit prior to demonstrating this role using a transgenic approach (Mohammed et al., 2012). In this analysis, a high expression of the V-PPase gene was unexpectedly found not during the later stage of fruit development, which is important in metabolite accumulation, but during the early cell division stage, and this high expression was induced by pollination. The suppression of V-PPase gene expression by RNAi with a promoter that was specific to the early stage of fruit development retarded fruit development, revealing the novel role of V-PPase in early fruit development. The localization of V-PPase mRNA in young seeds and their vicinities at the early stage of fruit development suggests that V-PPase does not directly affect cell division in fruit pericarp tissue, but rather promotes auxin transport from the seeds, which has previously been reported as a novel function of V-PPase (Li et al., 2005).
Fruit set is practically promoted by the addition of the synthetic auxin 4-chlorophenoxyacetic acid to tomato flowers, and is experimentally promoted by the application of auxin transport inhibitors to the pedicel and by the ovary-specific expression of the auxin synthesis gene (Carmi et al., 2003; Serrani et al., 2010), indicating that auxin is a key component involved in fruit set. PIN and AUX/LAX are known to be auxin efflux and influx carriers in plant cells and their tissue, and so their cellular localization may affect organ differentiation and growth, as described in the model regarding Arabidopsis root development (Kramer, 2004). The tomato contains 10 PIN genes and five AUX/LAX genes, which are differentially expressed (Pattison and Catalá, 2012).
Auxin concentration varies between different tomato fruit tissues (Pattison and Catalá, 2012), making it difficult to evaluate the physiological roles of auxin. However, auxin accumulation can be visualized using a reporter gene consisting of the auxin-inducible promoter with the β-glucuronidase gene (DR5::GUS). The use of this reporter gene showed that during the early stage of fruit development, auxin activity spreads to all of the fruit tissues and then unexpectedly localizes in the peduncle (Nishio et al., 2010). The expression patterns of the PIN family genes also appear to correspond with this auxin transport. In general, fruit abscission is promoted by ethylene and prevented by auxin (Murayama et al., 2015), and so the first step in fruit set is likely the prevention of fruit abscission by the accumulation of auxin in the peduncles. In the proposed model of tomato fruit set, PIN and V-PPase in the seeds and their vicinities and another family member of PIN in all of the fruit tissues transport auxin from the seeds to the peduncles via the fruit tissues, and the low expression of the PIN gene in the peduncles leads to the accumulation of auxin, which prevents fruit abscission by ethylene (Fig. 2).
Proposed model for the roles of sugars and auxin in fruit set and early fruit development. Auxin promotes fruit sink activity and cell division, and represses the action of ethylene and abscisic acid (ABA). PIN and V-PPase transport auxin from the seeds to the peduncles, where PIN gene expression is low, and prevent fruit abscission to retain sugar transport to the fruit. The sugars in the fruit inhibit PCD, and activate cell division and sugar metabolism to promote fruit set and early fruit development.
The prevention of fruit abscission retains sugar transport from the plant body to the fruit, inhibiting programed cell death (PCD) and promoting fruit development (Fig. 2). Sucrose that is transported to fruit tissues is usually degraded to hexose by cell wall invertase and the resultant glucose is then imported to fruit cells to inhibit PCD as a signal; however, a reduced glucose signal under stress conditions induces fruit abortion by PCD (Ruan et al., 2012). The prevention of fruit abscission by PIN and V-PPase described above may retain sucrose transport to the fruit to inhibit PCD. In addition, the activation of sugar uptake and metabolism in sink tissues by auxin and sugars has been reported in various crops (Kanayama et al., 1998; Koch et al., 1992; Lee et al., 1997; Ofosu-Anim et al., 1996, 1998; Xu et al., 1996), and may also be related to the inhibition of PCD and fruit development. In parthenocarpic fruit, sugar metabolism is also activated by auxin (Tang et al., 2015). Furthermore, the same mechanism seems to occur in pollination-dependent and -independent fruit set, with sucrose metabolism-related enzymes being involved in each, suggesting the importance of sugar metabolism in fruit set (Wang et al., 2009).
Tomato V-PPase mRNA also localizes in the fruit vascular tissues, in addition to the seeds and their vicinities (Mohammed et al., 2012). Cell wall invertase does not show localization in tomato ovaries before flowering, but interestingly does localize in the fruit vascular tissues at fruit set to transduce the glucose signal by supporting sugar unloading to the fruit, promoting cell division and the resultant fruit set (Palmer et al., 2015). Therefore, the vascular tissue-specific expression of the V-PPase gene may be related to this mechanism involving cell wall invertase because V-PPase may play a role in sugar accumulation, as described above. Alternatively, the role of V-PPase in auxin transport may support sugar unloading to the fruit through vascular tissue development because auxin is involved in vascular differentiation and development (Mattsson et al., 2003).
In this review, we considered the high sugar content trait, and the role of sugars in fruit set and the early stage of fruit development in the tomato. There is believed to be a trade-off between fruit sugar content and yield. For example, while the sugar content of tomato fruit increases as a result of the increased ratio of sucrose to hexose through the suppression of vacuolar invertase gene expression, the fruit size decreases (Klann et al., 1996); fruit size and yield generally decrease as fruit sugar content increases in tomato plants grown under salinity stress conditions (Saito et al., 2008; Sarkar et al., 2008). However, fruit size and yield do not appear to decrease in IL8-3 and SIRT-engineered tomato plants, which contain a higher fruit sugar content, as described above; and in other ILs, the fruit yield does not always decrease even if their fruits exhibit high sugar contents (Baxter et al., 2005b).
Recently, a breeding strategy to improve both the fruit sugar content and yield of tomato has been proposed using a breeding simulation with genomic information to apply genomic selection to tomato breeding (Yamamoto et al., 2016). In cases where there is no trade-off between fruit sugar content and yield, the increased fruit sugar content is likely accompanied by increased photosynthesis, decreased respiration, or increased sugar distribution to fruit. The physiological mechanism behind the improvement in both fruit sugar content and yield, and cultivation and breeding strategies that exploit this can be elucidated by analyzing the plant materials described in this review. Furthermore, a novel factor that controls the sink-source balance has been found, the modification of which allows fruit yield to be significantly increased without any associated decrease in hexose concentration in the tomato (Bermudez et al., 2014). The fraction of dry matter that is distributed to the fruit is used as an indicator of fruit yield (Higashide et al., 2015). Factors that are involved in sugar sensing, signaling, and the sink-source relationship should be further investigated, as well as the sugar metabolic enzymes that have been studied previously, to improve fruit quality and yield.
