Logo of jexbotLink to Publisher's site
J Exp Bot. 2009 Jan; 60(1): 237–248.
Published online 2008 Nov 25. doi:  10.1093/jxb/ern281
PMCID: PMC3071768

Identification of growth processes involved in QTLs for tomato fruit size and composition


Many quantitative trait loci (QTLs) for quality traits have been located on the tomato genetic map, but introgression of favourable wild alleles into large fruited species is hampered by co-localizations of QTLs with antagonist effects. The aim of this study was to assess the growth processes controlled by the main QTLs for fruit size and composition. Four nearly isogenic lines (NILs) derived from an intraspecific cross between a tasty cherry tomato (Cervil) and a normal-tasting large fruit tomato (Levovil) were studied. The lines carried one (L2, L4, and L9) or five (Lx) introgressions from Cervil on chromosomes 1, 2, 4, and 9. QTLs for fruit size could be mainly associated with cell division processes in L2 and L9, whereas cell expansion was rather homogeneous among the genotypes, except Cervil for which the low expansion rate was attributed to low cell plasticity. The link between endoreduplication and fruit size remained unclear, as cell or fruit sizes were positively correlated with the cell DNA content, but not with the endoreduplication factor. QTLs for fruit composition reflected differences in water accumulation rather than in sugar accumulation, except in L9 for which the up-regulation of sucrose unloading and hexose transport and/or starch synthesis was suggested. This may explain the increased amount of carbon allocated to cell structures in L9, which could be related to a QTL for fruit texture. In Lx, these effects were attenuated, except on fruit size and cell division. Finally, the region on top of chromosome 9 may control size and composition attributes in tomato, by a combination of QTL effects on cell division, cell wall synthesis, and carbon import and metabolism.

Keywords: Cell division and expansion, endoreduplication, fruit quality, near isogenic line, osmotic regulation, quantitative trait locus, Solanum lycopersicum, starch, sugar and acid contents


Advances in genetics and genomics in the last 10 years have boosted knowledge of the molecular control of fruit quality. At the same time, they have pointed out the complexity of the control of single fruit quality traits such as fruit size, shape, firmness, or sugar and acid content (Fernie et al., 2006; Causse et al., 2007). For tomato in particular, a great number of quantitative trait loci (QTLs) for these quality traits has been fine-mapped thanks to the identification of numerous molecular markers (Saliba-Colombani et al., 2001). In only a few cases have the genes underlying the main QTL effect been identified, such as, for instance, fw2.2 which is involved in fruit size regulation (Cong et al., 2002), or Lin5 controlling soluble solid content in fruit (Fridman et al., 2000). Favourable alleles could be introgressed into different tomato recipient lines by marker-assisted selection (Lecomte et al., 2004), providing the so-called near isogenic lines, which carry single or several regions of interest in a homogenous genetic background. Thus, the advent of new molecular techniques provided a powerful tool to optimize the use of genetic diversity, by bringing together, in one genotype, alleles that maximize a given trait (Paterson et al., 1988; Bai and Lindhout, 2007). In spite of these advances, breeding programmes for fruit organoleptic quality are still facing difficulties, partly due to the fact that most quality traits are polygenic and strongly influenced by environment. In tomato, ∼30 QTLs affecting fruit size and mass (Grandillo et al., 1999; Cong et al., 2002), and 23 QTLs controlling the content in soluble solids (Fridman et al., 2000) have been mapped.

One of the major objectives in tomato breeding regarding quality traits is to introduce QTLs for high soluble solid content observed in small- or green-fruited cultivars or in wild species into large fruited cultivated species. Progress towards this objective is slow, first because of the co-localizations of QTLs which create some antagonist effects, and the presence of several QTLs with low or less than additive effects (Causse et al., 2007), and secondly because of interactions between QTLs and the environment or genetic background (Chaïb et al., 2007). Dissecting complex traits into elementary physiological processes may help identify the genetic control of quality traits and also help with the search for candidate genes. It may be especially useful in NILs differing from the parental line by only small genome regions; seeking the physiological processes involved in phenotypic variations may provide direct information on the gene functions or regulatory steps potentially introgressed, as it is expected that only a few processes drive the variation in one particular trait. The capacity of ecophysiological approaches to analyse genetic variability has been illustrated in a few studies. For instance, cell division was shown to be involved in a QTL for tomato fruit size (Bertin et al., 2003), and ecophysiological models describing organ function have been successfully used to study the co-locations of QTLs for model parameters and for organ traits (Quilot et al., 2005; Tardieu et al., 2005). These studies elucidated some processes actually involved in genetic variations.

The aim of this study was to identify the ecophysiological processes involved in genetic variations underlying changes in tomato fruit size and composition. The main processes putatively involved in the regulation of these two traits have been quantified on four QTL-NILs derived from an intraspecific cross between a particularly good-tasting cherry tomato line (Cervil) and a large-fruited line with a common taste (Levovil). These lines were nearly isogenic to their recipient line (as defined in Van Berloo et al., 2001), from which they differed by the presence of one or several QTLs for quality traits introgressed from Cervil (Causse et al., 2002). These traits were related to fruit size, soluble content, sourness and sweetness, and texture. Fruit size is the result of cell division and expansion processes. In tomato, cell division is restricted to ∼10–20 d after anthesis (Bertin et al., 2007) and, in a given environment, the final fruit weight is highly correlated to the number of pericarp cells (Bünger-Kibler and Bangerth, 1983; Bertin, 2005). Cell growth and expansion depend on cell wall plasticity and accumulation of water and carbon compounds. Admittedly, cell expansion is driven by the turgor pressure induced by osmotic pressure, which is mostly determined by soluble sugars, acids, and potassium salts in tomato fruit (Ho et al., 1987). Thus, sugar and acid contents not only determine fruit sweetness and sourness, respectively, but they also play an active role in fruit osmotic pressure and, thus, in fruit size regulation. During cell expansion, endoreduplication of DNA has been shown to be involved in growth regulation in many species (Sugimoto-Shirasu and Roberts, 2003; Bertin et al., 2007). The final fruit composition, usually expressed on a fruit fresh weight basis, results from carbon import and dilution by the accumulated water (Ho, 1996). The fruit dry matter content varies >2-fold among genotypes, and consists of one-half soluble sugars, and one-quarter organic acids (Davies and Hobson, 1981). Temporal variations in composition during fruit development may indicate different causes for variations in final composition. For instance, during the first half period of fruit development, part of the sugars are stored as starch, an osmotically inactive reserve of carbon, whose size is related to the final sugar content (Ho, 1986).

In the present investigation, processes driving fruit growth and composition have been quantified during fruit development on two parental lines and four QTL-NILs grown in similar conditions. The results outlined the main processes involved in size and composition differences between cherry and large-fruit genotypes (Cervil and Levovil), and allowed the effects of QTLs for quality among the QTL-NILs to be analysed. Genes putatively involved in these QTLs were discussed further.

Materials and methods

Plant material

The initial QTL analysis was performed on a population of recombinant inbred lines (RILs) developed from an intraspecific cross, between Cervil, a cherry tomato [Solanum lycopersicum var. cerasiforme (Dun.) Gray] with 7 g fruits, good taste, and high aroma intensity, and Levovil (a S. lycopersicum Mill. line, hereafter called Lev) with 125 g fruits and normal taste (Causse et al., 2002). Based on the QTL map, five regions (located on chromosomes 1, 2, 4, and 9, respectively) were introgressed in the Lev genetic background (Lecomte et al., 2004). Two QTLs for fruit fresh weight were detected on chromosome 2. QTLs for fruit composition (soluble solid, dry matter, and sugar contents, and titratable acidity) were present on each of the five regions, as well as a QTL for fruit texture (firmness, elasticity, mealiness, and meltiness) (Fig. 1). The introgressed lines were produced as described in Chaib et al. (2006). They were obtained after three backcrosses on the recipient line followed by three selfing generations (Lecomte et al., 2004), and they carried homozygous alleles from Cervil at the five regions. One introgressed line cumulated these five regions (Lx), and four isogenic lines carried a single introgressed region on chromosome 2 (L2), 4 (L4), and 9 (L9).

Fig. 1.
Molecular map showing regions of interest on chromosomes 1, 2, 4, and 9, carrying the QTL for organoleptic quality, based on an intraspecific RIL population derived from a cross between a cherry tomato line and a large fruit line. Distances in Kosambi ...

Culture conditions

Plants were grown under standard glasshouse conditions in Avignon, France (43°55′N; 4°52′E). Planting took place in February at a density of 2 plants m−2, and the day–night temperature set-point was 20–18 °C. Inflorescences were pollinated by bumblebees.

The different genotypes were planted in four blocks of six plants randomly distributed over the greenhouse compartment. Trusses were pruned in order to homogenize truss size along the stem within each genotype. The maximum number of flowers left on each inflorescence was 12 for Cervil and Lx, nine for L2, eight for L9, and six for L4 and Lev. These inflorescence sizes corresponded to the average source–sink regulation by the plant itself when inflorescences are left unpruned (except Cervil which can set a larger number of fruits). The plant growth analysis performed at the end of the experiment (Table 1) suggested that the distribution of dry matter was homogenous among the genotypes, except in Cervil and to a lesser extent in Lx, for which the source–sink ratio was probaby higher than in the other lines, despite the low leaf area. In these two genotypes, the surplus of dry matter was accumulated in leaves (higher specific leaf weight).

Table 1.
Plant growth analysis performed at the end of the experiment

Observations and fruit sampling

For each genotype, fruit diameter was non-invasively measured twice a week with a calliper square on one basal and one tip fruit of the first five trusses from eight plants. Fruit sampling took place from April to May, on the first seven trusses of different plants. Flower buds and fruit ages (from anthesis to red ripe stage) were related to the time of flower anthesis, which was recorded individually. To avoid unbalanced source–sink relationships, one to two fruits per truss were sampled on a given plant, and then the plant was excluded from further sampling. Fruits were weighed and sized, cut into two halves when large enough, and then distributed for the following measurements. Fruit dry matter content was assessed after 5 d in a ventilated oven at 80 °C. Pericarp cell number was measured on basal and tip fruits to account for the gradient of cell number within the inflorescence which increases with fruit size. The pericarp was weighed, and cell suspensions were obtained after maceration in a pectinase solution, as described in Bertin et al. (2002). A minimum of eight samples per fruit were counted. The mean volume of an average cell was estimated by dividing the pericarp volume by the number of cells. The ploidy level of cells from the pericarp was measured by flow cytometry, as described in Bertin et al. (2007). Histograms were analysed with the WinMDI software (version 2.8) to determine the relative number of nuclei containing different amounts of DNA expressed as C values (from 2C to 512C). Three measurements were made in each fruit, when allowed by its size, and the average value was considered. The mean endoreduplication factor (EF) was calculated as proposed by Coockson et al. (2006):

An external file that holds a picture, illustration, etc.
Object name is jexbotern281fx1_ht.jpg

where p is the number of peaks of different DNA content (maximum=8) in the sample, Ei is the number of endocycles performed by nuclei in peak i (E1=0, E2=1, E3=2….E8=7), pi is the number of nuclei in the peak of value Ei, and ptot is the total number of nuclei in the sample. EF indicates the number of endocycles an average cell of the tissue has undergone.

The ploidy index (Pi) is the C level of an average cell of the tissue:

An external file that holds a picture, illustration, etc.
Object name is jexbotern281fx2_ht.jpg

where Ci is the C value of nuclei in peak i (C1=2, C2=4, C3=8….C8=256).

Considering that 1C is equivalent to 950 Mb, i.e. 950 pg DNA (Arumuganathan and Earle, 1991), the amount of DNA per cell (pg cell−1) was calculated as:

An external file that holds a picture, illustration, etc.
Object name is jexbotern281fx3_ht.jpg

Sugar and acid composition was determined on whole fruits, immediately frozen in liquid nitrogen after sampling, ground, and stored at –20 °C. Soluble sugars (sucrose, fructose, and glucose) and organic acids (malic and citric acids) were extracted and measured by HPLC analysis, and starch content was determined enzymatically (Gomez et al., 2002, 2003).

Fruit osmotic pressures due to soluble sugars or organic acids were calculated according to (Nobel, 1974):

An external file that holds a picture, illustration, etc.
Object name is jexbotern281fx4_ht.jpg

where R is the gas constant (83.1 cm3 bar K−1 mol−1), T is temperature (293 K), Wf and Wd are the fruit fresh and dry weights (g), WMi is the molecular weight of compound i, and [Gluc], [Fruc], and [Sucr] are the dry matter contents (g g−1 dm) in glucose, fructose, and sucrose, respectively.

The total fruit osmotic pressure was probably underestimated as the contribution of potassium salts was not considered. It may be higher than the contribution of hexose (Ho et al., 1987), despite some controversies (Mitchell et al., 1991).

The structural dry matter was estimated as the difference between total dry matter and soluble dry matter (here estimated as the sum of soluble sugars, starch, and acids). The amount of structural dry matter per unit cell surface was calculated as an indicator of cell wall thickness. Soluble compounds and dry matter were measured on the whole fruit, whereas cell number only concerned the pericarp. However, the pericarp accounts for ∼75% of the fruit weight, and it may be assumed that there was no genetic variability of the percentage structural matter in the locular tissue.

Dynamics of fruit growth, cell division, and expansion were plotted versus the number of days after anthesis (DAA) in order to outline differences in the length of these processes. Then, an index of development was calculated as the ratio between fruit age and the average time needed to ripen, in order to compare the temporal dynamics of different compound accumulation and DNA endoreduplication, on a comparable physiological time scale.

At the end of the experiment, three plants per genotype were sampled. Plant leaf area was measured with an area meter (LI-COR model 3000 Area Meter, Lincoln, NE, USA), and total leaf, stem, and fruit dry matter was measured after drying at 80 °C in a ventilated oven for 6 d. The average specific leaf weight (SLW) was measured on a sample of five leaves per plant harvested at different heights along the stem.

Statistical analysis

Fruit growth, cell number, cell volume, and ploidy curves were fitted to three-parameter sigmoid functions. A logistic function An external file that holds a picture, illustration, etc.
Object name is jexbotern281fx5_ht.jpgwas chosen for the cell number and ploidy index, whereas a Gompertz function An external file that holds a picture, illustration, etc.
Object name is jexbotern281fx6_ht.jpg was preferred for fruit growth and cell size, because it better accounted for the slow increase of these variables at the start of the observed period. Parameter estimation was carried out using the least squares method. Differences between two (or more) treatments were tested by comparing the sum of the residual sums of squares from individual fittings (ΣSSi), with the residual sum of squares for the common fitting (SSc), considering that the statistic

An external file that holds a picture, illustration, etc.
Object name is jexbotern281fx7_ht.jpg

follows Fisher's law with (n–1)×k and (Ndata–k) degrees of freedom. Ndata is the total number of data, n is the number of individual regressions, and k is the number of fitted parameters for each regression.

Means and variances of the relative rates of division (cell cell−1 d−1) were calculated using the bootstrap method (‘resample’ from original sample with replacement). A linear increase of the cell number was assumed between two successive close sampling dates. A Tukey test was then performed to compare the six genotypes at the different dates.

Analysis of variance (ANOVA) was performed to analyse genotype effects for traits measured at maturity. When significant effects (P <0.05) were detected, Tukey or Bonferroni tests were performed for pairwise comparisons of mean responses. If data normality and variance equality tests failed, then ANOVAs on ranks were performed, and pairwise comparisons of mean responses were analysed by Dunnet's test. When not deduced from theoretical calculations, data were presented as averages ±95% confidence intervals (CIs). Sigmastat Jandel Scientific software (V2.03, SPSS Inc., Chicago, IL, USA) was used to perform the statistical analysis.


QTLs for fruit size in relation to development, cell division, and cell expansion

Differences in final fruit size were due to different lengths of development, and to different growth rates (Fig. 2). The time from anthesis to maturity differed by a maximum of 13 d between the parental lines (45 d and 58 d in Cervil and Lev, respectively). All the QTL-NILs showed a reduced rate of growth compared with Lev, and they differed in their growth timing: the fruit growth rate of L4 and L9 peaked at ∼25–30 DAA, similarly to Lev, whereas L2 and Lx reached their maximum growth rate at ∼20 DAA, as Cervil did. Their final fresh weights ranged from 6 g in Cervil to 148 g in Lev. Differences among genotypes were significant, except between Lev and L4.

Fig. 2.
(A) Fruit fresh weight increase during fruit ageing. Curves were fitted with three-parameter Gompertz functions (R2 > 0.92). For each genotype (Lev, solid bold line; L4, solid thin line; L9, large dotted bold line; L2, large dotted ...

Dynamics of the pericarp cell population are shown in Fig. 3. The parental lines presented both the maximum and minimum number of pericarp cells (Fig. 3A). Lev cells divided until ∼25 DAA and reached a final cell number of 9.6×106. In contrast, cell division ceased at ∼12 DAA in Cervil fruits, and the final cell number was ∼1.2×106. Among the QTL-NILs, the ranks were those previously observed for fruit growth, with differences occurring in the length and rate of division. In L4 (8.7×106 cells) and L9 (6.4×106 cells) fruits, the length of the cell division period was similar to that in Lev fruits, but the division rate was lower in L9. In L2 and Lx, the short division periods and lowered division rates were both responsible for low pericarp cell number (4.2×106 in L2 and 2.2×106 in Lx). At maturity, cell numbers were significantly different among genotypes. A statistical analysis of the relative division rate (cell cell−1 d−1) of each of the six genotypes (not shown) showed that higher division rates in large-fruited genotypes were not due to a higher relative rate of cell division, but rather to the higher number of cells compared with that in small fruit-size genotypes.

Fig. 3.
Dynamics of (A) pericarp cell number and (B) volume of one mean cell during fruit ageing. Each point is the mean of 8–10 fruits, and vertical bars show 95% CIs when larger than symbols. Basal and tip fruits were pooled which explains data scattering, ...

In contrast to cell division, cell expansion of one average-sized pericarp cell varied only slightly among the four NILs. Final cell size was similar in L4 and Lev, and somewhat lower in L9, L2, and Lx (Fig. 3B). Cells expanded over the whole period of fruit development, and reached an average volume of ∼10.6 nl. Conversely, Cervil cells stopped growing at ∼20–25 DAA, and their final volume was >3 times smaller (3 nl).

Variations in fruit fresh weight were mainly correlated to variations in cell number, both among and within genotypes (Fig. 3C). No significant relationship existed between final fruit weight and final cell size among the QTL-NILs (not shown).

Is endoreduplication involved in differences in size of contrasted NILs?

The EF [Eqn (1)] is the number of endocycles an average cell has undergone. To account for differences in developmental periods among lines, an index of development was calculated as the ratio between fruit age and the average time needed to ripen. EF increased early during fruit development, to final average values of 3.0 endocycles in Lx, 3.1 in Cervil, 3.3 in Lev and L4, 3.4 in L2, and 3.5 in L9 (Fig. 4A), without significant differences among genotypes.

Fig. 4.
(A) Endoreduplication factor measured during fruit development. Each point is the mean of 7–15 fruits, and vertical bars are 95% CIs. (B) Dynamics of cell DNA content estimated from the C level of an average cell measured during fruit development ...

The mean DNA amount accumulated in cells increased during fruit development, with slight differences among lines. Cervil, Lx, and L2 had the lowest values, L4 and L9 the highest values, whereas Lev was intermediate (Fig. 4B). Fruit fresh weight and mean ploidy index [Eqn (2); not shown] or final cell DNA content and final cell size (Fig. 4C) were positively correlated [r=0.94 (P <0.01) and r=0.86 (P=0.03), respectively].

Does osmotic regulation contribute to the QTLs for fruit size?

The temporal dynamic of fruit osmotic pressure due to both sugars and acids is shown in Fig. 5. The two parental lines largely differed in their osmotic regulation. In Cervil fruits, the osmotic pressure increased from the first third of development until maturity, to ∼0.66 MPa, whereas it was rather stable in Lev fruits at ∼0.3 MPa. In L9, the fruit osmotic pressure was higher than in Lev, and it linearly increased over the whole period of development to 0.5 MPa at maturity. The dynamics in Lx, L2, and L4 were close to that observed in Lev, despite small significant differences between Lx and Lev at maturity (Fig. 5). When considered separately, the fruit osmotic pressure due to organic acids was less than half the fruit osmotic pressure due to sugars, except during the very early stage (development index of 0.1) when it was slightly higher (not shown). No global correlation between fruit size and osmotic pressure was evidenced among genotypes.

Fig. 5.
Dynamics of the fruit osmotic pressure (calculated according to Nobel, 1974) due to soluble sugar and organic acid accumulation in each genotype (Lev, solid bold line; L4, solid thin line; L9, large dotted bold line; L2, large dotted thin line; Lx, small ...

QTLs for fruit composition in relation to the pattern of carbon accumulation and partitioning

Fruit dry matter content varied >2-fold between the parental lines (from 4.4% to 10.4%), and Cervil fruits had the highest values. As for the fresh weight, Lev and L4 fruits were similar (4.6%), whereas all other QTL-NILs contained more dry matter than Lev (5.3, 6.0, and 6.6% in L2, Lx, and L9, respectively), and they were significantly different from one another.

The total amount of carbon accumulated as sugars (g soluble sugar+starch per 100 g dry matter) was higher in L9 and in Cervil fruits at some periods of development (Fig. 6A), but differences among genotypes at maturity were not significant. However, the patterns of starch and soluble sugar accumulation were quite different among genotypes. Lev fruits accumulated half the starch and twice the soluble sugars of that in Cervil fruits, and the starch pool decreased much later in Cervil fruits, which contained a low amount of starch at maturity (Fig. 6B). Patterns of starch and soluble sugar contents were similar in L2, L4, and Lev (Fig. 6B, C). In L9 and Lx, these patterns were intermediate between the parental lines, but L9 was marked by a high soluble sugar content, close to that of Lev fruits. Dynamics of individual sugars (mainly glucose and fructose) were similar to that of the sum of sugars, and the ratio between fructose and glucose was not significantly different among the lines at maturity (not shown). Excluding L9, there was a compensative relationship between soluble sugar content and starch content during fruit development.

Fig. 6.
Dynamics of (A) total carbohydrate, (B) starch, (C) soluble sugars, (D) citric acid, and (E) malic acid contents on a dry matter basis, during fruit development of each genotype (Lev, solid bold line; L4, solid thin line; L9, large dotted bold line; L2, ...

Finally, Cervil fruits contained the highest soluble sugar content on a fresh weight basis (3.3% against 1.7% in Lev, 2.0% in L2 and L4, and 2.2% in Lx), but the lowest content on a dry weight basis. In contrast, L9 fruits contained among the highest soluble sugar contents on both a fresh (2.9%) and a dry weight basis (42%), which in addition to the low water content may explain the higher osmotic pressure in this line.

The patterns of citric and malic acid contents were different in the two parental lines (Fig. 6C, D). In Cervil fruits, the contents of both acids remained low in the first period of development, then increased in the second half of development and decreased again during maturation. In Lev, the citric acid content decreased in the first half of development and then increased later, in contrast to the pattern of malic acid accumulation. Lev fruits were richer in malic acid and poorer in citric acid than Cervil fruits. In the four QTL-NILs, the dynamics of acid accumulation were close to those observed in Lev. At maturity, the citric acid content was significantly increased in L2 and L9 compared with Lev, whereas the malic content was significantly decreased in all NILs, especially in Lx and L9.

Allocation of dry matter between structure and soluble storage may be involved in QTLs for quality

The percentage dry matter allocated to the structural compounds rapidly stabilized during development to between 50% and 55% of total fruit dry matter. Slightly lower proportions were observed for Cervil and L9 fruits in the first half of development, hypothetically because of the high transient storage as starch (Fig. 6B), Indeed, at maturity, whereas the pool of starch has been exhausted, differences among genotypes were no longer significant (Fig. 7A). The absolute amount of structural matter per unit cell area was calculated as an indicator of cell wall thickness, which probably plays a role in fruit texture. Compared with Lev, it was 1.5 times higher in Cervil (Fig. 7B) due to the low number of very small cells, and more than double in L9 fruit, due to their high fruit dry weight (9.03 g fruit−1 versus ∼6.8 g fruit−1 in Lev or L4) combined with slightly fewer and smaller cells. L2 had the lowest value, due to the low fruit dry weight (2.7 g fruit−1) and lowest proportion of structural dry matter.

Figure 7.
(A) Dynamics of the percentage of structural dry matter estimated during fruit development. Each point is the mean of 10–15 fruits, and vertical bars are 95% CIs. (B) Estimation of structural dry weight per cell surface at maturity for the different ...


In genetics, QTL analysis allows breeders to locate genetic factors associated with variations of traits of interest, and to transfer favourable QTL alleles in recipient lines through marker-assisted selection. Seldom have ecophysiological analyses of these QTLs been performed, though they may emphasize the processes affected by these QTLs (Bertin et al., 2003; Quilot et al., 2005), and help bridge the gap between QTLs and genes (Quarrie et al., 2006). Among the processes investigated in the present study, some of them were specifically modified depending on genotypes, and could be related to QTL effects.

QTLs for fruit size mainly reflected QTLs for cell division on chromosomes 2 and 9

In many species and organs, QTLs for yield or size have often been associated with division processes (Bertin et al., 2003; Quarrie et al., 2006). Differences in fruit size between wild and domesticated tomato were also attributed to differences in cell division (Kortstee et al., 2007). Although organ size cannot be simply considered as the sum of cell size and number (Tsukaya, 2008), in the present study, the genetic variation in fruit fresh weight, as well as the environmental variability within each genotype (from 5 g to 180 g), were tightly correlated to differences in cell number (Fig. 3), without compensatory effects between number and size of cells. Two types of controls could be discerned: the first observed in L2 mainly shortened the period of cell division (60% fewer cells than in Lev), whereas the second observed in L9 mainly lowered the rate of cell division (20% fewer cells than in Lev). In Lx, which contained 80% fewer cells than Lev, both effects probably co-occurred. These results are consistent with the presence of Fw2.2 in the region introgressed in L2, a QTL key for tomato fruit size, which affects the timing of gene regulation during the early phase of mitotic activity (Frary et al., 2000; Cong et al., 2002). In addition, the results suggest that a gene(s) regulating cell division may be present on top of chromosome 9. Interestingly, the data suggested that differences in the cell division rate at the fruit level mainly resulted from different sizes of the cell population, probably originated from genetic variations in mitotic activity in the shoot apical meristem, which can affect the floral meristem size (Gyllaspy et al., 1993).

Small cells in Cervil fruits probably resulted from low cell plasticity

In contrast to division, cell expansion was hardly different in Lev and in the QTL-NILs. The lower cell size in small fruits was mostly related to the shorter period of growth, except in Cervil. Interestingly, the reduction of the length of the division period was exactly reflected in the duration of the whole period of development, supporting the idea that the first may determine the second, and might underlie fruit size (Cong et al., 2002). However, in Cervil fruits, the low cell expansion could not be ascribed only to the short period of development. Other factors, such as cell plasticity or factors driving the influx of water, were probably involved. Irreversible cell expansion is powered by the turgor pressure generated by osmotic pressure within the cell (Lockart, 1965), and the inflow of water is driven by the gradient of water potential between fruit and surrounding tissues (Lee, 1989). High osmotic pressure (low osmotic potential) in fleshy fruit has been assumed to allow the development of low water potential, and thus the maintenance of a high gradient of water potential between the plant and the fruit. However, in this study, the absence of a positive correlation between osmotic pressure and cell/fruit size, and the high structural dry weight per cell surface estimated in Cervil fruits, would promote the hypothesis of low cell plasticity.

The link between endoreduplication and tomato fruit size remains unclear

Endoreduplication might be an important process involved in QTLs for fruit size (Sugimoto-Shirazu and Roberts, 2003). Among pea genotypes (Lemontey et al., 2000) or contrasted tomato lines (Cheniclet et al., 2005), positive correlations between endoreduplication and seed or pericarp cell size have been reported. In this study, despite a 25-fold increase in fruit fresh weight, the EF changed by less than one endocycle among genotypes, and was unrelated to cell or fruit size. However, the mean cell DNA content varied by 1.5-fold, and it was positively correlated to the final cell size or to the fruit fresh weight. It remains difficult to draw conclusions about the relevance of these relationships, as they may result from different lengths of development. However, Cervil fruits contained very small cells and low DNA content compared with Lx, whose developmental period was only a few days longer.

Differences in dry matter content and allocation between structural and storage compounds may explain QTLs for fruit sugar content and texture

In many genetic studies, the chemical composition of fruits is analysed on a fresh weight basis, as this is the most relevant to assess the perception of taste by consumers (Causse et al., 2003). In the present study, the higher sugar content on a fresh weight basis in small fruit size genotypes could be attributed to their high dry matter content (low water content), rather than to the increase of sugar accumulation in the dry matter. In accordance, QTLs for sugar or soluble solid contents (Fig. 1) co-localized with QTLs for dry matter content on the bottom of chromosome 2 and the top of chromosome 9. In contrast, the QTLs for sugar content on a dry matter basis were unexplained, and even in contradiction to the highest sugar content observed in L9 fruits.

In L9 and Cervil fruits, slightly lower proportions of fruit dry matter were allocated to structures in the first half of development, hypothetically because of the allocation of carbon for starch synthesis, whose turnover reduced the differences among genotypes at maturity. The amount of carbon allocated to cell structures may be involved in fruit texture (Bourne, 2002), a complex quality trait which implies fruit firmness, mealiness, and juiciness. QTLs for texture have been detected on chromosomes 4 and 9 (Causse et al., 2002; Chaib et al., 2006), but these QTLs depend on the environment (Chaïb et al., 2007). These authors reported a higher firmness, and a lower elasticity and mealiness of L9 and Lx fruits compared with Lev, though differences were not always significant. If one assumes that increasing cell wall thickness probably contributes to increase fruit firmness and to decrease fruit elasticity, the present analysis agreed with the results reported by Chaib et al. (2007) for L9. These results are supported by a recent study which outlined different chemical composition and structure of cell walls in L9 and Lx fruits (Quemener et al., 2007). In contrast, high juiciness and low mealiness of L4 fruits (Chaib et al., 2007) would be related to large cells, high water content, and a low amount of structural dry matter per cell surface.

Genes involved in the control of starch synthesis are present on top of chromosome 9

In agreement with the literature (Ho, 1986) which suggests that the size of the starch pool serves as a reservoir for the synthesis of soluble sugars, the peak of starch content and the final sugar content in percentage fresh weight were positively correlated. In contrast to Cervil or L2, L9 fruits also accumulated large amounts of soluble sugars during the phase of starch accumulation, resulting in the highest content of sugars in the dry matter. Starch accumulation may be enhanced by an increase of the main enzyme activity [ADP glucose pyrophosphorylase (ADPGppase)] or of the level of synthesis of the substrate (glucose 6P) (Robinson et al., 1988; Nguyen-Quoc and Foyer, 2001); it is also triggered by the amount of sucrose unloaded in the fruit (D'Aoust et al., 1999; N'Tchobo et al., 1999), which is higher in cherry tomatoes than in large-fruited cultivars (Islam and Khan, 2001). The presence of the apoplastic cell wall invertase gene lin5 (Fridman and Zamir, 2003; Fridman et al., 2004) and of a hexose transporter (Causse et al., 2004), together with the absence of a gene regulating the ADPGppase on chromosome 9 (Schaffer et al., 2000), supports the hypothesis that high starch accumulation in L9 was stimulated by the higher sucrose unloading, and hexose transport within the cells (Ruan et al., 1995, 1997; Ho, 1996; Schaffer et al., 1999), as observed in an other study (Li et al., 2002). However, the fact that high starch, but not high soluble contents, were recovered in Lx suggested the involvement of several levels of regulation.

In cherry tomatoes, the gluconeogenesis, which leads to the refixation of CO2 inside the fruit by photosynthesis or by dark fixation as malate, was shown to promote starch synthesis and dry matter accumulation (Ho, 1996). The hypothesis of such a regulation in Cervil and in L9 fruits may be put forward, as they contained less malate and more starch than the other genotypes.


By analysing the main ecophysiological processes involved in fruit growth and composition, it was possible to outline major differences among QTL-NILs involved in the genetic control of final fruit size and composition. These differences are in accordance with molecular or genetic analysis already performed on these tomato lines, but they also pointed to some new hypotheses, in particular for L9. This line presented singular patterns of growth and accumulation of dry matter and soluble solids, and it differed from the parental lines by accumulating different QTL effects, implying cell division, cell wall synthesis, and sugar and acid metabolism. The top of chromosome 9, in contrast to chromosome 2, seems an interesting region to accumulate fruit size and composition attributes, as sugar content was increased without substantial loss of fruit size, which represents a real challenge for breeders.


We thank Dr Diane Beckles for her valuable comments on the manuscript. This work was partly supported by the FP6 European Eu-Sol Project PL016214.


  • Arumuganathan K, Earle E. Estimation of nuclear DNA content of plants by flow cytometry. Plant Molecular Biology Reports. 1991;9:208–218.
  • Bai Y, Lindhout P. Domestication and breeding of tomatoes: what have we gained and what can we gain in the future? Annals of Botany. 2007;100:1085–1094. [PMC free article] [PubMed]
  • Bertin N. Analysis of the tomato fruit growth response to temperature and plant fruit load in relation to cell division, cell expansion and DNA endoreduplication. Annals of Botany. 2005;95:439–447. [PMC free article] [PubMed]
  • Bertin N, Borel C, Brunel B, Cheniclet C, Causse M. Do genetic makeup and growth manipulation affect tomato fruit size by cell number, or cell size and DNA endoreduplication? Annals of Botany. 2003;92:415–424. [PMC free article] [PubMed]
  • Bertin N, Gautier H, Roche C. Number of cells in tomato fruit depending on fruit position and source–sink balance during plant development. Plant Growth Regulation. 2002;36:105–112.
  • Bertin N, Lecomte A, Brunel B, Fishman S, Génard M. A model describing cell polyploidization in tissues of growing fruit as related to cessation of cell proliferation. Journal of Experimental Botany. 2007;58:1003–1013. [PubMed]
  • Bünger-Kibler S, Bangerth F. Relationship between cell number, cell size and fruit size of seeded fruits of tomato (Lycopersicon esculentum Mill.), and those induced parthenocarpically by the application of plant growth regulators. Plant Growth Regulation. 1983;1:143–154.
  • Bourne MC. Food texture and viscosity: concept and measurement. 2nd edn. New York: Academic Press; 2002.
  • Causse M, Buret M, Robini K, Verschave P. Inheritance of nutritional and sensory quality traits in fresh market tomato and relation to consumer preferences. Journal of Food Science. 2003;68:2342–2350.
  • Causse M, Chaïb J, Lecomte L, Buret M, Hospital F. Both additivity and epistasis control the genetic variations for fruit quality traits in tomato. Theoretical and Applied Genetics. 2007;115:429–442. [PubMed]
  • Causse M, Duffé P, Gomez MC, Buret M, Damidaux R, Zamir D, Gur A, Chevalier C, Lemaire-Chamley M, Rothan C. A genetic map of candidate genes and QTLs involved in tomato fruit size and composition. Journal of Experimental Botany. 2004;55:1671–1685. [PubMed]
  • Causse M, Saliba-Colombani V, Lecomte L, Duffé P, Rousselle P, Buret M. QTL analysis of fruit quality in fresh market tomato: a few chromosome regions control the variation of sensory and instrumental traits. Journal of Experimental Botany. 2002;53:2089–2098. [PubMed]
  • Chaïb J, Devaux M-F, Grotte M-G, Robini K, Causse M, Lahaye M, Marty I. Physiological relationships among physical, sensory, and morphological attributes of texture in tomato fruits. Journal of Experimental Botany. 2007;58:1915–1925. [PubMed]
  • Chaïb J, Lecomte L, Buret M, Causse M. Stability over genetic backgrounds, generations and years of quantitative trait locus (QTLs) for organoleptic quality in tomato. Theoretical and Applied Genetics. 2006;112:934–944. [PubMed]
  • Cheniclet C, Rong WY, Causse M, Frangne N, Bolling L, Carde JP, Renaudin JP. Cell expansion and endoreduplication show a large genetic variability in pericarp and contribute strongly to tomato fruit growth. Plant Physiology. 2005;139:1–11. [PMC free article] [PubMed]
  • Cong B, Liu J, Tanksley SD. Natural alleles at a tomato fruit size quantitative trait locus differ by heterochronic regulatory mutations. Proceedings of the National Academy of Sciences, USA. 2002;99:13606–13611. [PMC free article] [PubMed]
  • Coockson SJ, Radziejwoski A, Granier C. Cell and leaf size plasticity in Arabidopsis: what is the role of endoreduplication? Plant, Cell and Environment. 2006;29:1273–1283. [PubMed]
  • D'Aoust MA, Yelle S, Nguyen-Quoc B. Antisense inhibition of tomato fruit sucrose synthase decreases fruit setting and the sucrose unloading capacity of young fruit. The Plant Cell. 1999;11:2407–2418. [PMC free article] [PubMed]
  • Davies JN, Hobson GE. The constituents of tomato fruit—the influence of environment, nutrition and genotype. Critical Reviews in Food Science and Nutrition. 1981;15:205–280. [PubMed]
  • Fernie AR, Tadmor Y, Zamir D. Natural genetic variation for improving crop quality. Current Opinion in Plant Biology. 2006;9:196–202. [PubMed]
  • Frary A, Nesbitt TC, Frary A, et al. fw2.2: a quantitative trait-locus key to the evolution of tomato fruit size. Science. 2000;289:85–88. [PubMed]
  • Fridman E, Pleban T, Zamir D. A recombination hotspot delimits a wild-species quantitative trait locus for tomato sugar content to 484 bp within an invertase gene. Proceedings of the National Academy of Sciences, USA. 2000;97:4718–4723. [PMC free article] [PubMed]
  • Fridman E, Zamir D. Functional divergence of a syntenic invertase gene family in tomato, potato, and Arabidopsis. Plant Physiology. 2003;131:603–609. [PMC free article] [PubMed]
  • Fridman E, Carrari F, Liu YS, Fernie AR, Zamir D. Zooming on a quantitative trait for tomato yield using interspecific introgressions. Science. 2004;305:1786–1789. [PubMed]
  • Gomez L, Rubio E, Augé M. A new procedure for extraction and measurement of soluble sugars in ligneous plants. Journal of the Science of Food and Agriculture. 2002;82:360–369.
  • Gomez L, Rubio E, Lescourret F. Critical study of a procedure for the assay of starch in ligneous plants. Journal of the Science of Food and Agriculture. 2003;83:1114–1123.
  • Grandillo S, Ku H-M, Tanksley SD. Identifying loci responsible for natural variations in fruit size and shape in tomato. Theoretical and Applied Genetics. 1999;99:978–987.
  • Gyllaspy G, Ben-David H, Gruissem W. Fruits: a developmental perspective. The Plant Cell. 1993;5:1439–1451. [PMC free article] [PubMed]
  • Ho LC. Metabolism and compartmentation of translocates in sink organs. In: Cronshaw J, Lucas WJ, Giaquinta RT, editors. Phloem transport. New York: A.R. Liss Inc.; 1986. pp. 317–324.
  • Ho LC. The mechanism of assimilate partitioning and carbohydrate compartmentation in fruit in relation to the quality and yield of tomato. Journal of Experimental Botany. 1996;47:1239–1243. [PubMed]
  • Ho LC, Grange RI, Picken AJ. An analysis of the accumulation of water and dry matter in tomato fruit. Plant, Cell and Environment. 1987;10:157–162.
  • Islam S, Khan S. Seasonal fluctuations of carbohydrate accumulation and metabolism of three tomato (Lycopersicon esculentum Mill.) cultivars grown at seven sowing times. Journal of Horticultural Science and Biotechnology. 2001;76:764–770.
  • Kortstee AJ, Appeldoorn NJG, Oortwijn MEP, Visser RGF. Differences in regulation of carbohydrate metabolism during early fruit development between domesticated tomato and two wild relatives. Planta. 2007;226:929–939. [PubMed]
  • Lecomte L, Duffé P, Buret M, Servin B, Hospital F, Causse M. Marker-assisted introgression of 5 QTLs controlling fruit quality traits into three tomato lines revealed interactions between QTLs and genetic backgrounds. Theoretical and Applied Genetics. 2004;109:658–668. [PubMed]
  • Lee DR. Vasculature of the abscission zone of tomato fruit: implications for transport. Canadian Journal of Botany. 1989;67:1898–1902.
  • Lemontey C, Mousset-Déclas C, Munier-Jolain N, Boutin P. Maternal genotype influences pea seed size by controlling both mitotic activity during early embryogenesis and final endoreduplication level/cotyledon cell size in mature seed. Journal of Experimental Botany. 2000;51:167–175. [PubMed]
  • Li X, Xing J, Gianfagna TJ, Janes HW. Sucrose regulation of ADP-glucose pyrophosphorylase subunit genes transcript levels in leaves and fruits. Plant Science. 2002;162:239–244. [PubMed]
  • Lockhart JA. An analysis of irreversible plant cell elongation. Journal of Theoretical Biology. 1965;8:264–275. [PubMed]
  • Mitchell JP, Shennan C, Grattan SR. Developmental changes in tomato fruit composition in response to water deficit and salinity. Physiologia Plantarum. 1991;83:177–185.
  • Nguyen-Quoc B, Foyer CH. A role for ‘futile cycles’ involving invertase and sucrose synthase in sucrose metabolism of tomato fruit. Journal of Experimental Botany. 2001;52:881–889. [PubMed]
  • Nobel PS. Introduction to biophysical plant physiology. San Francisco: W.H. Freeman and Co; 1974.
  • N'Tchobo Y, Dali N, Nguyen-Quoc B, Foyer CH, Yelle S. Starch synthesis in tomato remains constant throughout fruit development and is dependent on sucrose supply and sucrose synthase activity. Journal of Experimental Botany. 1999;50:1457–1463.
  • Paterson AH, Lander ES, Hewitt JD, Peterson S, Lincoln SE, Tanksley SD. Resolution of quantitative traits into Mendelian factors, using a complete linkage map of restriction fragment length polymorphisms. Nature. 1988;335:721–726. [PubMed]
  • Quarrie SA, Pekic Quarrie S, Radosevic R, Rancic D, Kaminska A, Barnes JD, Leverington M, Ceoloni C, Dodig D. Dissecting a wheat QTL for yield present in a range of environments: from the QTL to candidate genes. Journal of Experimental Botany. 2006;57:2627–2637. [PubMed]
  • Quemener B, Bertrand D, Marty I, Causse M, Lahaye M. Fast data preprocessing for chromatographic fingerprints of tomato cell wall polysaccharides using chemometric methods. Journal of Chromatography A. 2007;1141:41–49. [PubMed]
  • Quilot B, Kervella J, Génard M, Lescourret F. Analysing the genetic control of peach fruit quality through an ecophysiological model combined with a QTL approach. Journal of Experimental Botany. 2005;56:3083–3092. [PubMed]
  • Robinson NL, Hewitt JD, Bennett AB. Sink metabolism in tomato fruit 1. Developmental changes in carbohydrate metabolizing enzymes. Plant Physiology. 1988;87:727–730. [PMC free article] [PubMed]
  • Ruan YL, Patrick JW. The cellular pathway of postphloem sugar transport in developing tomato fruit. Planta. 1995;196:434–444.
  • Ruan YL, Patrick JW, Brady C. Protoplast hexose carrier activity is a determinate of genotypic difference in hexose storage in tomato fruit. Plant, Cell and Environment. 1997;20:341–349.
  • Saliba-Colombani V, Causse M, Langlois D, Philouze J, Buret M. Genetic analysis of organoleptic quality in fresh market tomato. 1. Mapping QTLs for physical and chemical traits. Theoretical and Applied Genetics. 2001;102:259–272.
  • Schaffer AA, Petreikov M, Miron D, et al. Modification of carbohydrate content in developing tomato fruit. Hortscience. 1999;34:1024–1027.
  • Schaffer AA, Levin A, Oguz I, Petreikov M, Cincarevsky F, Yeselson Y, Shen S, Gilboa N, Bar M. ADPglucose pyrophosphorylase activity and starch accumulation in immature tomato fruit: the effect of a Lycopersicon hirsutum-derived introgression encoding for the large subunit. Plant Science. 2000;152:135–144.
  • Sugimoto-Shirasu K, Roberts K. ‘Big it up’: endoreduplication and cell-size control in plants. Current Opinion in Plant Biology. 2003;6:544–553. [PubMed]
  • Tardieu F, Reymond M, Muller B, Granier C, Simonneau T, Sadok W, Welcker C. Linking physiological and genetic analyses of the control of leaf growth under changing environmental conditions. Australian Journal of Agricultural Research. 2005;56:937–946.
  • Tsukaya H. Controlling size in multicellular organs: focus on the leaf. PLoS Biology. 2008;6:1373–76. [PMC free article] [PubMed]
  • Van Berloo R, Aalbers H, Werkman A, Niks RE. Resistance QTL confirmed through development of QTL-NILs for barley leaf rust resistance. Molecular Breeding. 2001;8:187–195.

Articles from Journal of Experimental Botany are provided here courtesy of Oxford University Press
PubReader format: click here to try


Save items

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • Compound
    PubChem chemical compound records that cite the current articles. These references are taken from those provided on submitted PubChem chemical substance records. Multiple substance records may contribute to the PubChem compound record.
  • MedGen
    Related information in MedGen
  • PubMed
    PubMed citations for these articles
  • Substance
    PubChem chemical substance records that cite the current articles. These references are taken from those provided on submitted PubChem chemical substance records.
  • Taxonomy
    Taxonomy records associated with the current articles through taxonomic information on related molecular database records (Nucleotide, Protein, Gene, SNP, Structure).
  • Taxonomy Tree
    Taxonomy Tree

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...