As a general perception, consumers do not like the taste of modern tomato cultivars and claim that many heirloom varieties have better taste and aroma quality (Tieman et al., 2012). This is because, the nutritional levels of fruits and vegetables decrease as a result of intensive breeding studies (Klee & Tieman, 2013). From this point of view, it is necessary to investigate the sources of variation for the traits of interest (Acosta-Quezada et al., 2015) because today's modern varieties can be developed in terms of taste and aroma (Ruiz et. al., 2005). On the other hand, the improvement and commercialization of landraces may empower small scale farmers and enable them to generate more income (Agong et al., 2001). For this purpose, many studies were carried out previously to understand and document the composition of the tomato fruit and its variation (Mata et al., 2000; Schauer et al., 2005; Ruiz et al., 2006; Marconi et al., 2007; Suarez et al., 2008; Pareira et al., 2013; Acosta-Quezada et al,. 2015). The study discussed below is the first to examine the diversity for organic acids of local tomato landraces spread to the Aegean-Mediterranean side of Anatolia.
Mean values of fruit weight (g), length (mm) and fruit diameter (mm) are provided in Table 2. According to these results, great variation was determined among 19 landraces. For instance, more than 8-fold difference was found between the weight of the smallest and largest tomato fruit. If we describe the fruit size as a combination of three traits such as weight, diameter and length, it is easy to say that the differences between the fruit size also prove the genetic diversity among the landraces. The same landraces were previously studied by Kaya (2012) and great variability was determined among the landraces. The size of the fruit may be related to whether the plant is genetically inclined to bloom or not (Cavicchi &Silvetti, 1976; Grandillo et. al., 1999). On the other hand, this situation can also be affected by growing conditions such as fertilizing. Since the growing conditions in these studies were kept as constant as possible, it would not be wrong to say that the reason for the diversity in fruit sizes is genetic factors.
Titrable acidity (TA) screened in tomato fruits among landraces showed a wide diversity. The landrace that contain the highest TA (TR63233) may be suggested as a good source for breeding programs in the future to increase TA in fruits. Contrary, the landrace which gave the lowest TA (TR72508) may also be a candidate to improve the fruit quality, especially for tree tomatoes (Boyes & Strübi, 1997). But, TA ratio alone is not a sufficient indicator for studies to be carried out in the future to improve the taste. Researchers generally focused on SSC, TA and pH and tried to increase these traits (Hewitt & Garvey, 1987; Triano & St Calair, 1995). Moreover scientist tried to explain the relationships between the traits such as acidity, SSC, sugars, to determine the components of the tomato taste (Baldwin et. al., 1991; Malunda et. al., 1999; Agong et. al. 2001; Marconi et. al., 2007; Acosta- Klee & Tiaman, 2013; Quezada et. al., 2015). The tomato taste is a complex of several components and it can be said that, sweetness is directly affected by SSC and reducing sugars, while sourness is affected by the amount of soluble solid content, pH and TA (Stevens et al. 1977, 1979; Baldwin et al. 1998). Galiana-Balaguer et. al (2006), cited that it is important to know which genes control the traits while planning a breeding study in tomatoes. Genes who control the traits such as TA and SSC are polygenic and this creates some difficulties in breeding programs (Saliba Colombani et al., 2001; Fulton et al., 2002). Like TA, SSC is another important trait for taste and quality of tomatoes. The range of variation of SSC in 19 tomato landraces was between 4.77–7.07%. Our study is in agreement with previous research (Mata et. al., 2000; Agong et. al., 2001; Galiana-Balaguer et. al., 2006; Kaya, 2012).
The organic acid content of the tomato landraces were analyzed by ANOVA in Table 3 but it is not sufficient in determining which landraces are different from others. Therefore, One-way ANOM tests were carried out with data which were found statistically significant with the ANOVA test in order to determine the upper and lower decision lines for the landraces. Our results regarding the predominant organic acid being citric acid are in agreement with previous studies as expected (Thorne & Effiuvwevwere, 1998; Marconi et al., 2007; Acosta-Quezada at al., 2015). Furthermore, the wide variation among landraces was determined in terms of citric acid. The One way ANOM graph of the citric acid content of the landraces are given Fig. 3. Results show that significant differences and wide variation were detected among landraces (p ≤ 0.05) in terms of CA. The highest CA content were determined in the Ege 8 (105.73 mg.g− 1), Ege 6 (78.12 mg.g− 1), Ege 3 (74.03 mg.g− 1), TR49646 (64.67 mg.g− 1) and Ege 5 (64.91 mg.g− 1) landraces. These 5 landraces were determined over the range of upper decision line according to one way normal ANOM test. On the contrary, 3 landraces (TR66062, TR62707 and TR69155) gave the lowest CA contents. According to the results obtained, the five highest landraces which gave the highest citric acid concentration may be suggested as a source for a breeding program. Organic acid concentrations in tomato vary according to maturity and the cultivar (Baldwin et al., 1991). Tomato fruits had been harvested at red maturity phase for each accession in our trials, so that it may be claimed that the variation between the accessions depends largely on the genetic factor. Also, Choi et al. (2014), have attributed phenotypic differences in tomatoes largely to genetic factors. These variations in the composition profile of citric acid can be used for the introgression of favorable traits from landraces into the genetic background of the cultivated species. It is claimed that, the tomato taste intensity is perceived higher when the amount of sugar and organic acids are higher (Stevens et al., 1977; Bucheli et al. 1999; Galiana-Balaguer et al., 2006). But, the citric acid concentration is not the only indicator for a breeding program. The taste and flavor of tomato is the sum of sugars, acids, and many other different volatile chemicals. These chemicals could affect the consumer preferences (Klee & Tieman, 2013). For instance, the ratio malic to citric acid should be lower than 0.5. At higher levels the taste turns sour because the malic acid has been reported aproximately14% more sour than citric acid (Petro-Turza, 1987; Yılmaz, 2001). Therefore, dozens of factors affecting taste and aroma of tomatoes should be taken into account and breeding programs should be carefully planned.
One way ANOM graph of malic acid content of the landraces are given Fig. 4. Results show that significant differences and wide variation were detected among landraces (p ≤ 0.05) in terms of malic acid. The highest malic acid contents were determined in TR49646 (8.23 mg.g− 1) and Ege 8 (8.03 mg.g− 1) landraces. Also these two landraces were over range according to one way normal ANOM test which the upper limit had been determined as 1514.5 mg L− 1 in extraction solution. The result shows that, these two landraces have significantly higher malic acid contents from the other landraces. The average malic acid content of the landraces was 6.45 mg.g− 1. On the other hand, 3 landraces, which are TR63233, Ege 6 and TR61785, gave the lowest malic acid contents as 4.88, 5.11, 5.29 mg.g− 1, respectively. These landraces are below the lower decision line of 1067.10 mg L− 1 calculated by the one way ANOM test (p ≤ 0.05) in extraction solution. The MA content of the landraces is in agreement with other research (Suarez et. al., 2008; Mata et al., 2000; Galiana-Balaguer et al. 2006; A.P. Breksa III et al. 2015). The varieties labeled as TR49646 and Ege 8 can be recommended as a source of high malic acid content. One of the interesting results of the ANOM test is that, Ege 8 and TR49646 are above the upper decision line both for CA and MA. Fulton et al. (2002), reported positive correlations between malic acid and citric acid, which provides yet another proof of the relationship between CA and MA.
One way ANOM graph of oxalic acid content of the landraces are given Fig. 5. Results show that we have found narrow diversity among landraces in terms of oxalic acid content but a statistical significance was (p ≤ 0.05) detected. The highest oxalic acid content were determined in Ege 6 (188.96 mg L− 1). This result, however, was not over the upper decision line so it can be said that this landrace is not different from the other landacres except for landrace TR63233. The landrace TR63233 was determined under the lower decision line, which means that only this landrace was different from others. Suarez et al. (2008), determined the oxalic acid contents in 5 tomato cultivars between 25-37.5 mg L− 1, but Mata et al. (2000), determined a wide range for OA content of tomato cultivars and accessions between 270 mg kg− 1 and 2580 mg kg− 1 (with another unit expression as 0.27–2.58 mg.g− 1). Our results which are between 0.54 and 0.94 mg.g− 1 are in agreement with those studies. The importance of OA in tomato composition should not be overlooked because it is important from a human nutrition point of view. Generally, it is desired that the amount of oxalic acid in food is low. OA diminishes the bio-availability of calcium in the alimentary canal (Guil et al., 1996). For instance, 1 g of Calcium precipiates with 2.5 g of oxalic acid. So the bioavailability of calcium is affected by oxalic acid. If the relationship between calcium and oxalic acid is over 2.25, it is considered as insufficient source of calcium for food (Mitjavila, 1990). Therefore from this point of view, TR63233 can be recommended as a source of low OA content.
The multidimensional scaling analysis showed a wide diversity among landraces, and also distinguished those that are more or less similar or dissimilar. It can be said that the MDS analysis is a reliable tool in explaining and assigning similarities among landraces (Fig. 6). 11 landraces, circled as A, showed similarities or narrow diversity among each other and 5 landraces, circled as B also showed similarities or narrow diversity between each other. But the landraces grouped together as A and B showed a wide variation. One of the interesting results is that landraces named as Ege 8, TR72508 and TR66062 differed from all other landraces and showed a great diversity from the other 16 landraces. When data is examined carefully, it can be seen that TR66062 has the lowest CA content, Ege 8 has the second highest TA and highest CA content, and TR72500 has the third highest OA and second lowest SSC content. Thus, the places and group of landraces on the MDS analysis diagram can be explained. Overall, results indicate that, wide diversity among landraces exists for traits studied. The data explained above may give ideas and prospects for selection or use of these landraces as a germplasm source in the future.