Genetic Characterization of Tunisian Lime Genotypes Using Pomological Traits

Accessions were collected in 2013–2014 and re-collected in 2014–2015 (Table 1) throughout the Cap Bon Nord region east of Tunisia, where citrus cultivation is most widespread. Acquisitions were carried out among a wide range of stakeholders and with the presence of local governmental agencies both in farms and ex situ collections. During the collecting missions, we visited old orchards where limes were cultivated for many decades. Farmers and technical staffs from regional authorities confirmed the names of genotypes. Identification of species was performed with the help of Blondel (1978) classification.

The FCA was used to classify the three species under consideration based on the size of the cultivation area and on the number of households, as described by Sthapit et al. (2012).

The fully ripe fruits were taken from the four directions of the tree and from the interior and exterior layers of the canopy at the rate of 30 fruits per tree. These fruits were divided into 3 batches of 10 fruits to analyze the quantitative traits (Table 2) and the juice parameters (Table 3). For pomological characterization, analysis has been performed separately for each growing season. Sixteen quantitative traits, including seven parameters dealing with juice description, were measured (Table 4) and correlations among those traits were calculated (Table 5). Nineteen qualitative characters (Table 6) were chosen based on Citrus descriptors (IPGRI 1999). Fresh juice was obtained using a citrus press (Santos Classic N°11, Lyon, France). Subsequently, the juice was filtered through a 1-mm mesh sieve; weighed and volume was measured in a burette. Density was estimated in a sample of 100 ml of juice. Total soluble solids (TSS) content was determined by direct readings on a hand-held refractometer (Toledo, 30 PX) calibrated before use with distilled water. The pH was measured using a pH meter (Toledo, S22) previously calibrated. The titratable acidity (TA) of the juice was evaluated by the determination of citric acid by titration with a NaOH solution (0.1 N). The determination of vitamin C was carried out by titration with iodine solution. Data were obtained in triplicate.

For quantitative traits, all analyses were performed using SAS software (version 6.07, 1990). Descriptive statistics were performed and presented as minimum, maximum, mean standard deviation, and coefficient of variation (CV). One-way analysis of variance (ANOVA) was used, and data are presented as mean ± standard deviation (SD). Pearson index was calculated for quantitative traits. Principal component analysis (PCA) was carried out to examine the distribution of genotypes in the first plan of PCA for quantitative parameters. For qualitative data, frequency distributions were computed.

The numbers of phenotypic classes for qualitative parameters that differed for each trait were used to estimate the Shannon–Wiener diversity index (H’). It was used to characterize the phenotypic frequencies of the traits and was defined as H=∑i=1npi ln piH = \sum\nolimits_{i = 1}^n {pi\,\ln \,pi} where n is the number of phenotypic classes for a character and pi is the proportion of the total number of entries in the ith class. Each value of H was standardized by conversion to a relative phenotypic diversity index (H’) by division by H_max = ln (n) in order to express the values of H’(H/H_max) in the range of 0–1 (indicating the absence of diversity and maximum of diversity, respectively). The diversity index was classified as high (H’ ≥ 0.6), intermediate (0.40 ≤ H’ ≤ 0.60), or low (0.1 ≤ H’ ≤ 0.40), as described by Eticha et al. (2005) and Mengistu et al. (2015).

The participatory FCA was used, while regional agricultural authorities and the farmers were interviewed. It allowed categorizing C. limetta as a threatened species. Although many householders cultivated this species, it was propagated in small area. C. latifolia and C. limettioides were classified as rare species because they were cultivated in small area and by few householders (Fig. 1). Thus, special attention must be paid to these species in order to conserve them and encourage their dissemination. The most cultivated species was C. limetta in contrast to C. latifolia, which is the least cultivated (Table 1). Accessions from different species showed a wide range of variability for all the pomological traits studied. According to a recent research, Curk et al. (2016) have elucidated the origins of limes and lemons based on cytoplasmic and nuclear markers. The survey highlighted that all limes and lemons descend from Citrus medica as the direct male parent in combination with Citrus aurantium for C. limetta and a hybrid (Citrus maxima × Citrus reticulata) for C. limettioides. Among triploid limes, C. latifolia accessions Persian lime types result from the fertilization of a haploid ovule of Citrus limon by a diploid gamete of Citrus aurantifolia. As limes and lemons were vegetatively propagated by apomixes and horticultural practices, the intra-subgroup phenotypic diversity results from asexual variations (Curk et al. 2016). Two classifications of limes have been reported: lime of Pearse known as C. aurantifolia hybrid by Swingle and Reece (1967) and C. latifolia Tan. by Tanaka (1977). Instead, sweet lime of Tunisia is classified as C. limon (L.) Burm. by Swingle and Reece (1967) and C. limetta Risso by Tanaka (1977). The results can be useful for both selection of cultivars and breeding programs aiming the improvement of fruit quality. Snoussi et al. (2012) revealed that both sexual and asexual reproductions of limes cultivated in Tunisia contributed to their genetic diversity.

Fig. 1

Table 4 describes the minimum, maximum, mean, standard deviation, and CV for each variable studied. The weight of fruit varied from 41.3 g (Lime16) to 157.9 g (LimePal 2) per fruit with a mean of 77.3 g per fruit. High variability between cultivars was observed for fruit weight, which is confirmed by the relatively high CV (32%). The number of seeds per fruit varied from 1.2 (Lime3) to 5.4 (Lime8) with a mean of 3.3 and a CV of 35%. The weight of juice was highly variable (CV = 42%), ranging from 6.5 (Lime19) to 69.9 (Limpal2) with a mean of 29.9. The volume of juice varied from 50 ml (Lime18) to 152 ml (Lime8), with a mean of 80 ml (CV = 32%). These four parameters were the most discriminant between cultivars based on CVs. The fruit rind thickness (CV = 21%) range from 1 mm (Lime6 and Lime14) to 2 mm (LimPal 2 and Lime3). Diameter of fruit axis (CV = 19%) range from 26.19 (Lime4) to 48.9 mg·dm⁻³ (LimPal2 and Lime8).

In order to estimate correlation between quantitative parameters based on the data measured on Tunisian limes, Pearson's correlation coefficients were estimated (Table 5). A significant correlation among several quantitative parameters was observed. Weight of fruit was significantly and positively correlated with fruit diameter (r = 0.81; p ≤ 0.01), diameter of fruit axis (r = 0.75; p ≤ 0.05), weight and volume of juice (r = 0.97; p ≤ 0.01; r = 0.62; p ≤ 0.01, respectively), fruit length (r = 0.61; p ≤ 0.05), width of fruit skin (r = 0.61; p ≤ 0.05), and width of epicarp at equatorial plane (r = 0.60; p ≤ 0.05). Fruit diameter was significantly and positively correlated with the diameter of fruit axis (r = 0.68; p ≤ 0.01) and weight of juice (r = 0.84; p ≤ 0.01). Highly significant and positively correlations were observed between fruit skin width and width of pericarp at equatorial plane (r = 0.77; p ≤ 0.01). Number of segment was significantly and positively correlated with TSS (r = 0.63; p ≤ 0.01); diameter of fruit axis was significantly and positively correlated with weight of juice (r = 0.75; p ≤ 0.01).

Seed color was the most discriminant trait with four different phenotypes (Table 6). Similarly, fruit skin color, fruit surface texture, adherence of albedo to pulp, adherence of segment walls, vesicle length and thickness, fruit axis, and cotyledon color were also more discriminant compared with all the other parameters. Uniformity of pulp color and cross-section shape of axis were monomorphic for all the accessions studied (Table 6). H’ ranged from 0 for both cross-section shape and pulp color uniformity to 0.98 for both thickness of segment walls and vesicle thickness (Table 6) with a mean value of 0.61. Moreover, other parameters showed high values of H’: fruit skin (epicarp) color (H’ = 0.96), shape of fruit base (H’ = 0.93), seed color (H’ = 0.84), pulp (flesh) color (H’ = 0.82), and fruit surface texture (H’ = 0.8). According to the Shannon diversity index, we assume that shape of fruit apex, fruit skin epicarp color, fruit surface texture, thickness of segment walls, pulp (flesh color), vesicle thickness, seed shape, and seed color were the most discriminant qualitative parameters. Pulp (flesh color), fruit skin epicarp color, and fruit surface texture are definitely used as selection criteria throughout the supply and consumption chain. It is well known that genetics, environment, and cultural practices interact to define the eventual main fruit traits (weight, diameter, length, and width of skin).

PCA was performed based on fruit and juice quantitative parameters. The results showed that 60.3% of the total variability is accounted for the first three principal components (PCs). The first two PCs account 35.5% and 15.5% of the total variability (Fig. 2 and Table 7). The PC1 positively correlated with weight and diameter of fruit, diameter of axis, and weight of juice. The PC2 positively correlated with fruit rind (mesocarp) thickness, number of segments, and vitamin C content. The projection of lime cultivars in the plan 1–2 of the PCA allows the discrimination of the species C. limettioides and Lime8 from the other genotype (Fig. 2). Regarding C. latifolia and C. limetta, we did not observe any significant discrimination. Accessions from these cultivars were grouped together. Regarding TSS and content of vitamin C, both LimPal1 and Lim-Pal2 exhibited the highest values. Moreover, PCA has also distinguished C. limettioides (LimPal 1 and LimPal 2) and Lime 8 from all the other cultivars. Lime8 genotype, which belongs to C. limetta species, is characterized by large fruit. For this reason, it has been clustered with genotypes LimPal1 and LimPal2, which belongs to C. limettioides species. This species can be selected for breeders in order to improve fruit size (weight and diameter of fruit) and yield of juice. On the basis of the same descriptors of Citrus (IPGRI 1999), similar findings were recorded, referring to Saddoud Debbabi et al. (2013) for the main parameters correlated with the first axis of PCA (weight, diameter, and length of fruit). Phenotypic characterization have shown their efficiency for many crops, such as carrot (Mezghani et al. 2014) and wheat (Mengistu et al. 2015), and for many fruit trees such as fig (Saddoud et al. 2008; Gaaliche et al. 2012), olive (Hannachi et al. 2008), apricot (Ruiz & Egea 2008), apple (Mratinić & Fotirić Akšić 2011), and cornelian cherry (Moradi et al. 2019). This evaluation is necessary to achieve the developmental program and genetic improvement of the lime species. The outcomes of this study will be very useful for Tunisian Gene Bank and for a good identification and documentation of Citrus genetic resources. Although pomological characterization is low cost method and has many advantages, it remains limited in the number of characters and is limited in use. The characterization could be improved through the involvement of molecular markers, for example Simple Sequence Repeat (SSR) that allows the study of molecular diversity and the establishment of fingerprints of the cultivars studied.

Fig. 2