• Open Access

Genetic diversity of Tunisian figs (Ficus carica L.) as revealed by nuclear microsatellites


M. Trifi, Laboratoire de Génétique moléculaire, Immunologie et Biotechnologie, Faculté des Sciences de Tunis, Campus universitaire, 2092 El Manar, Tunis, Tunisia. E-mail: mokhtar.t@fst.rnu.tn


The present study portrays the achievement of the genetic polymorphism surveying and the establishment of an ecotypes identification key on the basis of simple sequence repeats data. Seventy-two Tunisian fig ecotypes in situ and ex situ conserved were analyzed using six microsatellite loci. A total of 58 alleles and 124 genotypes were revealed and permitted to evidence high degree of genetic diversity mainly explained at the intra group level. Cluster analysis based on genetic distances proved that a typical continuous genetic diversity characterizes the local germplasm. In addition, the microsatellite multilocus genotyping has permitted to unambiguously distinguish 70 well-defined ecotypes (resolving power of 97.22%). Data are discussed in relation with the reliability of the used markers to check the conformity of the plant material and to rationally manage the conservation of this crop.

Ficus carica L. (2n=26) is one of the 750 species of the genus Ficus (Berg 1989; Falistocco and Antonielli 2002). It is the oldest fruit crop known as a gynodioecious and insect pollinated species (Beck and Lord 1988; Kislev et al. 2006). Its utilization consists of ecotypes called common figs (unisexual female trees) and caprifigs (bisexual with functional male trees) occurring in similar frequencies in wild populations (Valdeyron and Lloyd 1979). This fruit crop is wide spread in the Mediterranean basin countries since it is well adapted either to different soils or climates (Mars 2003). In Tunisia, the fig germplasm consists of numerous landraces mainly selected by farmers for their fruit qualities and maintained in orchards where fruit types are unequally represented. Hodgson (1931) reported the prevalence of the “Smyrna” (crossbreeding) ecotypes in southern Tunisia, while the “common” (parthenocarpic) and “Smyrna” ones are equally spread in the northern Tunisia. Thus, a wide phenotypic diversity characterizes the large number of ecotypes distinguishable by taste; color and flavor of fruits (Rhouma 1996; Mars et al. 2007). In addition, fig trees represent the principal component of several agro-ecosystems in the southern areas such as at the Jessours region (Matmata, Beni Khédache and Douiret) and constitute the second fruit crop in the Tunisian oases (Mars 1995, 2003). In spite of its ancient cultivation, fig groves remain sporadic and consisted of a few plantations where superior fruity cultivars have been recently established in the central and the northern areas cultivars (Mars et al. 2007). It should be stressed that since several decades, the local germplasm is seriously threatened by severe genetic erosion due to biotic and abiotic stresses such as intensive urbanization and the fig mosaic disease. Moreover, problems of synonymy are still occurring since cultivars’ appellation is mainly based on the fruit parameters. As a consequence, the precise number of varieties is unknown. Therefore, establishment of strategies aiming at the conservation and the evaluation of the local germplasm has become imperative. Therefore, recent investigations (prospecting and collection of cuttings) have been made and permitted to identify about 100 ecotypes that are ex situ conserved in different collections at the “Institut Supérieur Agronomique, (ISA)” of Chott Meriem, the “Institut des Régions Arides (IRA)” of Medenine, the “Centre de Recherches Phoénicicoles, (CRPh)” of Degache, the “Commissariat Régional au Développement Agricole, (CRDA)” of Gafsa and the “Groupement Interprofessionnel des Fruits, (GIF)” of Sbikha (Lahbib 1984; Mars et al. 1994; Rhouma 1996; Mars et al. 2007). However, the conservation management of this relatively high number of existing ecotypes with different synonyms and variable commercial values requires the development of reliable methods aiming at their fingerprinting. For this purpose, studies have reported the use of pomologic, morphometric, horticultural traits and isozyme markers to examine the genetic diversity within and between ecotypes housed in these collections (Valdeyron and Crossa-Raynaud 1950; Ben Salah et al. 1995; Hedfi et al. 2003; Chatti et al. 2004a; Salhi-Hannachi et al. 2003). However, a lack of polymorphism levels is occurring since these parameters are highly influenced by the environmental conditions and/or the plant development stage. To overcome this inconvenience, the use of molecular markers such as random amplified polymorphic DNA (RAPDs), inter simple sequence repeats (ISSRs) and random amplified microsatellite polymorphism (RAMPO) have been developed in Tunisian figs (Chatti et al. 2004b, 2007; Salhi-Hannachi et al. 2004). These studies have proved that a high genetic diversity characterizes the Tunisian fig germplasm. Nevertheless, these studies are less rewarding by regards of the some limitation of the used markers and/or the relatively little number of ecotypes involved. Therefore, we become interested in the use of microsatellites or simple sequence repeats (SSRs) known for their reliability in plant population genetic diversity assessment. Such markers are of advantages over the mentioned markers since microsatellites: (1) are very common in the eukaryotes’ genomes (Wang et al. 1994; Smulders et al. 1997), (2) are of co-dominant and of mendelian inhertitance, (3) they permit to detect high polymorphisms at the infra specific level (Testolin et al. 2000) and, (4) are efficient to unambiguously distinguish genotypes (Zehdi et al. 2004). Moreover, we have successfully reported the use of SSRs to survey the genetic diversity in sixteen Tunisian fig ecotypes (Saddoud et al. 2005). Therefore, taking into account the advantages of the microsatellites, we have enlarged the number of Tunisian ecotypes in situ and ex situ conserved in order to have an overview of the genetic diversity structuration in the local germplasm with the use of SSRs.

The present study portrays the evidencing of SSRs in a set of 72 local ecotypes originated from different Tunisian areas and the achievement of their fingerprinting. In addition, data have permitted not only to determine the between fig landraces and geographic groups’ relationships but also to establish a cultivars’ identification key. Data are discussed in relation to the true-to-typeness control of Tunisian fig cuttings.

Material and Methods

Plant material

A total of 72 ecotypes listed in Table 1 were used in this study. These are in situ preserved in orchards located in the north of Tunisia (Testour, Utique and Raf Raf) and Kerkennah islands or ex situ conserved the CRPh Degache, the CRDA Gafsa, the ISA Chott Mariem and the INRA Medenine collections (Fig. 1). Plant material consisted of young leaves randomly sampled from adult trees and frozen in nitrogen liquid until their use. Depending on their origin, landraces are clustered in three groups namely North, Centre and South.

Table 1.  Origin and names of the 72 Tunisian fig ecotypes used to assess genetic diversity based on SSRs (D: Douiret; B: Béja; Bk: Béni Khedache; G: Gafsa; K: Kerkennah; R: Raf Raf; S: Sahel; Ts: Testour; Tz: Tozeur; U: Utique; Caprifigs are labeled with an asterisk).
North“U; R; Ts; B”Bither Abiadh (T); Zidi (T); Djebbi (T); Wahchi (B); Zergui (B); Khortoumi (B). Dhokkar* (U); Mestiri* (R); Marsaoui (U); Zidi (U); Kerkeni (U); Goutti (R); Chetoui (R); Harragui (R); Zidi (R); Bither Souri (R); Bither Arbi (R); Marsaoui (R)
Centre “S; K”Dchiche Assal (S); Kahli A (S); Zidi (S); Soltani 1 (S); Soltani 2 (S); Kahli 1 (S); Kahli 2 (S); Hemri 1 (S); Hemri 2 (S); Bither abiadh 1 (S); Bither abiadh 2 (S); Bither abiadh 3 (S); Bidhi 1 (S); Bidhi 2 (S); Baghali (S); Zidi (S); Besbessi (S); Goutti (S); Chetoui (S); Ghabri (S); Khadhri (S); Abiadh (K); Temri (K); Bither (K); Bouang(K) ; Baghli (K). Dhokkar* (K); Jrani* (S); Assafri* (S)
South “D+Bk+Z”“Tz+G”Assal boudchiche (G); Bither abiadh (G); Sawoudi (G); Mlouki (G); Gaa Zir (G); Khadouri (G); Soltani (G); Mokh bagri (Tz); Bouslames (Tz); Khalt (Tz); Tounsi (Tz); Hamri (Tz); Khzami (Tz); Grichy (Tz); Zidi (Tz); Bither (Tz); Tayouri Asfar (D); Sawoudi (Bk); Makhbech (Z); Hammouri (Bk); Zaghoubi (Bk); Wedlani (Bk). Dhokkar* (G); Dhokkar* (Z); Dhokkar*(Tz)
Figure 1.

Distribution of the five different regions on which fig ecotypes were originated.

DNA isolation

DNA was extracted from 1 g of plant material according to Dellaporta et al. (1983) procedure yielding about 100 to 150 μg. DNAs were spectrophotometrically quantified and their integrity was performed after analytic electrophoresis (Sambrook et al. 1989).

Primers and PCR assays

A total of six microsatellite primers previously isolated by Khadari et al. (2001) and identified as MFC2, MFC3, MFC5, MFC6, MFC7 and MFC8 were used (Table 2).

Table 2.  Summary of SSRs as evidenced in the Tunisian fig gerplasm.

Total58 124

PCR amplifications were performed as reported in Saddoud et al. (2005). The amplified banding patterns were firstly checked on 2% agarose gels visualized with ethidium bromide staining under UV light. SSRs were then resolved on non-denaturing polyacrylamide gels (10%) and revealed by ethidium bromide staining according to Sambrook et al. (1989).

Data analysis

Genetic polymorphism in each population was evaluated by the mean number of alleles, alleles frequencies, the observed heterozygosity (Hobs) and expected heterozygosity (Hexp) (Nei 1978) using Genetix software version 4.04 (Belkhir 2001). The Wright inbreeding coefficient (Fis) was computed according to Weir and Cockerham (1984) using GENEPOP 3.1 (Rousset and Raymond 1995). A positive value of (Fis) indicates a deficit in heterozygotes in comparison with the Hardy-Weinberg equilibrium expectations. The total genetic diversity (Ht), the mean genetic diversity within population (Hs) and the genetic differenciation among popualtions (Gst) (Nei 1978) were calculated using Genetix software version 4.04 (Belkhir 2001). The non parametric analysis of molecular variance (AMOVA) procedure was used to describe the population's structure and the variability among clones within and between regions as reported by Excoffier et al. (1992) using the ARLEQUIN package v.2000 (Schneider et al. 2000).

Genetic distances among ecotypes were estimated according to Nei (1972). Dendrograms were constructed with unweighted pair group method with arithmetic averaging (UPGMA) cluster analysis using the program Neighbor of the PHYLIP ver. 3.57c software (Felsenstein 1995) and constructed using the TreeView ver. 1.5 software (Page 1996). The ecotypes’ identification key was established as described by Ould Mohamed Salem et al. (2001).

Results and discussion

Starting from total cellular DNAs used as templates in presence of the tested primers, the SSR patterns exhibited more than six alleles per locus with distinguishable homozygous and heterozygous individuals. As reported in Table 2, a total of 58 alleles ranged in number from 6 (MFC7) to 14 (MFC3, MFC8) and sized from 118 bp (MFC5) to 228 bp (MFC2) were scored. The mean number of alleles per locus is of 9.66 per locus (Table 4). The Hobs values scored in all accessions varied from 0.51 (MFC3) to 0.84 (MFC6) with an average of 0.7 suggesting that the Tunisian germplasm is characterized by a high level of polymorphism at the DNA level (Table 3). This is in agreement with the multilocus of expected heterozygicity (Hexp) that did not significantly differed among the defined groups (Table 4). In addition, for all ecotypes, the observed heterozygosity was lower than the expected values according to the Hardy-Weinberg (HW) equilibrium suggesting a deficiency of heterozygous individuals over the HW expectation.

Table 4.  Genetic diversity indices for the three defined groups.
GroupHexpHobsFisp-valueMean number of alleles/locus
All accessions0.820.700.1280.0009.66
Table 3.  Expected (Hexp) and observed (Hobs) heterozygosity in each group by locus computed using Genetix 4.04.
GroupNorthCentreSouthAll accessions
H valuesHexpHobsHexpHobsHexpHobsHexpHobs


Allele frequencies ranged from 0.6 for locus MFC5 allele 124 (North) to 0.023 for locus MFC2 allele 228, MFC3 allele 144 and MFC5 allele 118 (Centre). The mean number of alleles as well as their frequencies varied among the defined groups since 6.83 and 9 alleles were scored in the North and the Centre respectively (Table 4). Moreover, group's specific alleles have been scored. This is well exemplified in the case of alleles 124, 150, 194 (MFC3) and 182 (MFC6) that are only evidenced in the Centre group as well as alleles 141 (MFC8) and 143 (MFC3) scored only in the North group. Compared to data reported previously using these loci in this crop, the present study has permitted to evidence relatively higher number of alleles. In fact, a total of 23 alleles were scored either in French or Moroccan figs (Khadari et al. 2003). Similarly, the allele sizes scored in this work are different from those obtained by these authors. This is well exemplified in the case of the MFC6 that has been described by Khadari et al. (2001) with a size ranged from 291 to 311 bp and exhibiting sizes varying from 168 to 188 in Tunisian figs. This result strongly supported the large diversity that characterises the local fig germplasm.

Estimates of the total genetic diversity (Ht), the mean genetic diversity within population (Hs) and the genetic differentiation among groups (Gst) are reported in Table 5. Since nearly similar values either of Hs or Ht for the six loci, we assume that the maximum of variation scored as 89.02% is maintained within groups while only 10.98% (Gst) of the variability is explained at the inter-group level. This assumption is confirmed by regards the low values of Gst scored taking into account each locus. Additionally, results of AMOVA analysis performed to study the partitioning of the molecular variance among the defined groups corroborate with this assumption. In fact, significant percentages of variance were scored among the defined groups: 6.16%, 6.32% and 5.59% respectively between North/Centre, North/South and Centre/South groups respectively (Table 6). However, 93% of this diversity is maintained at the intra group level. Similar results were reported by Salhi-Hannachi et al. (2005, 2006) using RAPDs and/or ISSRs in two Tunisian fig collections. In their study, the scored AMOVA partition indicates that more than 92% of the total genetic diversity is distributed within the collections suggesting the presence of considerable genotypic variation among fig cultivars sampled from two Tunisian collections, while only 8% of the diversity to accessions differences between regions (Salhi-Hannachi et al. 2005). The low divergence scored between the groups and the large variability detected among the cultivars could be explained by the occurrence of gene flow in the natural populations from which cultivars originated and the reproduction mode. In fact, life form and breeding systems can have significant influences on the genetic variation and its portioning (Hamrick and Godt 1996). This is in agreement with estimates of the Fis indices for the three defined groups since the scored values exhibited deviation from (HW) equilibrium among groups as well as in all ecotypes (Table 4). However, a specific test for heterozygote deficiency (U test, Rousset and Raymond 1995) indicated statistically no significant deficits of heterozygocity.

Table 5.  The total genetic diversity (Ht), the mean genetic diversity within population (Hs) and the genetic differentiation (Gst) estimates based on SSR data.
Table 6.  AMOVA for the three fig groups. p-value: significance test after 1023 random permutations.
Source of variationDFSum of squaresVariance componentsPercentage of variationp-value*



UPGMA dendrogram based on Nei's (1972) genetic distances is reported in Fig. 2. This phylogram clustered the ecotypes studied in two main clusters. The first one is composed of 13 ecotyeps and originated from the defined groups. These are identified as: Goutti, Sawoudi, Zidi, Djebbi, Hammouri, Makhbech, Kahli, Khadhri, Bither Abiadh, Mokh Bagri, Ghabri, Wahchi and Baghli. All the remaining ecotypes (59 over 72 studied) constitute the second cluster that presented three sub goups. Therefore we assume that this tree branching is made independently from the geographic origin as well as from the trees sex. This result suggests that a typically continuous genetic diversity characterizes the Tunisian fig germplasm. This is strongly supported since the caprifigs did not significantly diverge from the common fig groups and agree with the monoecious origin of Ficus that has evolved into two gynodioecious forms (Machado et al. 2001). Note worthy that similar data have been reported in Tunisian figs using other molecular markers such as RAPDs, ISSRs and RAMPOs (Chatti et al. 2004b, 2007; Salhi-Hannachi et al. 2006).

Figure 2.

UPGMA dendrogram of Tunisian fig 72 ecotypes based on Nei (1972) genetic distances estimated on SSRs.▴: North, ▪: Centre p, •: South and *: caprifigs. G: Gafsa, K: Kerkennah, Tz: Tozeur, U: Utique and Z: Zarzis.

Based on the scored multilocus genotypes, the ecotypes’ identification key was established as follows: genotypes were hierarchically ordered according to the greater number of alleles per locus. Ecotypes were then organized and those of identical fingerprints were grouped. In this case, alleles were labeled as A to F respectively for MFC2 to MFC8 and each letter is indexed with 1 to n according to the alleles’ number. The derived the ecotypes’ identification key is reported in Fig. 3. This precise diagram exhibited a resolving power of 97.22% since it has permitted to unambiguously differentiate 70 out of 72 ecotypes studied. In fact, only the Soltani2 and Kahli1 common figs are characterized by identical multilocus fingerprints. These mentioned cultivars have presented similar RAPD, ISSR and RAMPO banding profiles (Chatti et al. 2004b, 2007). At least hypothesis of homonymy could be forwarded to explain this result. The enlargement of the number of microsatellites would be of great interest to confirm our assumption. It should be stressed that the scored percentage is nearly similar to those reported in other fruit crop species such as date-palms and apricot (Zehdi et al. 2004; Sanchez-Perez et al. 2005; Krichen et al. 2006). Therefore this ecotypes’ identification key would be of a great interest in the description, the registration and the certification of plant material as well as in the rational management and Tunisian fig germplasm conservation. Moreover, since a large number of fig ecotypes are cultivated over the world and exchanges of cuttings are currently occurring, the availability of the evidenced SSRs would be strongly of great support in the true to typeness of plant material. Work is currently in progress to molecularly characterize all the known local fig ecotypes.

Figure 3.

Identification key of 72 cultivars of common fig (Ficus carica L.) representing Tunisian fig germplasm, based on multilocus genotypes.


This work was supported by grants from the Tunisian “Ministère de l'Enseignement Supérieur et de la Recherche Scientifique”. Authors would like to thank the staff of “Direction Générale de la Production Agricole”, the CRPh of Degache, the CRDA of Gafsa, CTV of Ras Djebel, Utique Ruine City for their fruitful help and cooperation.