Ecogeographic variability and genetic diversity associated with seed albumins, globulins and prolamins patterns in Vicia taxa from Algeria

Genetic variability was studied in 78 populations of locally collected Vicia L. taxa for seed albumins, globulins and prolamins patterns by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) along with an ecogeographic characterization of sites investigated. 131, 119 and 98 bands were respectively used for albumin, globulin and prolamin cluster analysis. Dendrograms based on the Jaccard index and the UPGMA method were generated and the degree of genetic diversity between and within taxa was evaluated. Five clusters were generated from albumins, six from globulins and four from prolamins patterns. The results reflect the great diversity of storage proteins and a high correlation was obtained between the three studied fractions. Several accessions present specific bands which could be used as a discriminatory marker both on intra and interspecific levels. No clear relationships were seen between the groups according to their geographical origin. Data obtained from ecogeographic investigation can be used for future collecting missions.


Background
The genus Vicia belongs to the Legumes, family Leguminosae which is considered one of the largest families of flowering plants and represents tremendous morphological, ecological and genetic diversity. Vicia L. comprises about 210 species widely distributed along Europe, Asia and the American regions (Hanelt and Mettin 1989). In Algeria, there are 26 species belonging to three series (Quézel and Santa 1962). The genus Vicia has the capacity to fix atmospheric nitrogen (Nemecek et al. 2008). Vetch seeds contain more than 20% crude protein and relatively high amount of lysine, leucine, arginine, phenylalanine and tyrosine (Darre et al. 1998). Maxted (1993) pointed out that there had been 20 major classifications of the group since Linnaeus. Kupicha (1976) has subdivided the genus into two subgenera (Vicilla, Vicia) which have been further subdivided into 17 and 5 sections, respectively. The subgenus Vicia sensu Maxted (1993) contains 9 sections including sections Vicia, Hypechusa and Narbonensis. Section Cracca sensu Kupicha (1976) belongs to subgenus Vicilla. Morphological approach is rather difficult to estimate the all genetic diversity in the genus (Haider and El-Shanshoury 2000). Seed proteins are physiologically stable and easy to manipulate (Ladizinsky and Hymowitz 1979). Considerable insight has been drained as to their structure and synthesis during seed development and to their role as storage proteins (Higgins 1984). Electrophoretic analysis of seed storage proteins was used in testing genetic associations in Vicia at generic, specific and intraspecific levels, along with morphological characterization (Ladizinsky and Hymowitz 1979;Mirali et al. 2007;Hameed et al. 2009;Emre et al. 2010). The use of gel electrophoresis of seed protein in phylogeny is supported by the fact that mature seeds possess the same protein components unchanged with age Open Access *Correspondence: s.bechkri@gmail.com Laboratoire de Génétique Biochimie et Biotechnologies Végétales, Faculté des sciences de la nature et de la vie, Université Frères MENTOURI, 25000 Constantine, Algeria or environmental stress, and thus provide valid evidence for genetic relatedness (Crawford 1990). Potokina et al. (2003) and Mirali et al. (2007) suggested that comparison of electrophoregrams of seed proteins is useful to assess relationships among Vicia taxa.
The objective of the present study was to investigate intra and interspecific variations in 11 taxa belonging to sections Vicia, Hypechusa, Narbonensis and Cracca by SDS-PAGE of seed albumins, globulins and prolamins to test the technique for vetches identification and to clarify the genetic diversity among Vicia taxa collected from different regions of the country along with an ecogeographic characterization of sites investigated as no studies have previously been reported on electrophoretic separation of the storage proteins of the given 11 Vicia taxa from Algeria.

Plant material and taxa identification
Object of the study were 78 accessions representing 4 taxa of Sect. Vicia, 2 taxa of Sect. Hypechusa, 1 taxon of Sect. Narbonensis and 4 taxa of Sect. Cracca. Pods were randomly collected from various bioclimatic conditions of Algeria (Fig. 1). The dry seeds were stored into separate sealed paper bags at room temperature until their utilization. Informations of the investigated accessions are given in Table 1. Taxonomic identification of accessions was verified by the morphology of plants grown from seeds in a greenhouse of the laboratory of genetics, biochemistry and plants biotechnologies of Faculty of Biology in Constantine University (eastern Algeria). Taxa identification was undertaken using the key of Quézel and Santa (1962).

Electrophoresis
Non-reducing SDS-PAGE was undertaken according to Laemmli (1970). Bromophenol blue was added to the extraction buffer to follow proteins movement in the gel. 15, 8 and 70 μl of respectively albumins, globulins and prolamins supernatants were placed on biphasic polyacrylamide gels (12%). 10 μl of a protein molecular weight marker (BIO-RAD Precision Plus Protein Standards) containing ten proteins (10,15,20,25,37,50,75,100,150 and 250 kDa) was used as standard. Tris-glycine (pH 8.3) was used as electrode buffer. Runs were carried out at a voltage of 60 V and 500 mA overnight. Gels were stained by Coomassie Brilliant Blue R, then images were scanned using ImageScannerIII.

Ecogeographic parameters of investigated sites
The five ecological factors of Mediterranean climate (annual rainfall, average of the maximum temperature of the hottest month, average of a minimum temperature of the coldest month, Emberger coefficient and altitude) were used to characterize sampling stations. A global positioning systems (GPS GARMIN eTrex ® model 30) was used to collect coordinates of sites investigated. Data recorded to ONM (National Office of Meteorology, Algeria) were used to characterize the climate of sites investigated (Table 2). Data recorded to CLIMATE-DATA.ORG (http://fr.climate-data.org/) were used for five stations (Mila, Ain Temouchent, Tipaza, El Tarf and Blida).

Climatic data correction
Correction of precipitations and temperatures data based on extrapolations for different altitudinal points were undertaken (Table 3), according to the works of Seltzer (1946) as explained by Bechkri and Khelifi (2016).

Calculation of the bioclimatic coefficient of Emberger (1955) and definition of the bioclimate
The pluviothermic Emberger quotient (Q2) is determined by three major climate factors. Stewart's formula (1969) was used in the present study. Details of calculations were reported by Bechkri and Khelifi (2016).

Data analysis
The mobility and the frontal report of each band were calculated. The size marker standard curve was traced. The graphical equation and the coefficient of determination Alt altitude, P annual rainfall, M and m are the average maximum temperature of the hottest month and the average of the minimum of the coldest month, respectively a Data from "http://climate-data.org"

Reference station Latitude Longitude Alt. (m) P (mm) m (°C) M (°C)
Jijel (  allowed the calculation of the molecular weight of each band. In this method, "absence" contributed equally to "presence" in the calculation of dissimilarity. Present bands were scored 1 and absent bands were scored 0. For each fraction, a binary matrix was constructed. A dendrogram was produced by the UPGMA based on Jaccard index (J) between protein patterns. Analyses were carried out using XLSTAT (Pearson edition, version 2014.5.03). For ecogeographic parameters, Euclidean distances (Romesburg 1990) were used in the estimation of the genetic resemblance. Matrix including the five ecological parameters of each accession was used to elaborate a dendrogram using UPGMA. Analyses were carried out with STATISTICA (version 6.1 program). The possible correlation between albumins, globulins and prolamins patterns, was evaluated by a Mantel test (Mantel 1967) based on Pearson's correlation (XLSTAT Pearson edition, version 2014.5.03). The same test was used to test geographical matrix with seed albumins, globulins and prolamins matrices.

Seed proteins variability
The three fractions electrophoregrams are represented by some accessions illustrated in Fig. 2. Figure 3 presents dendrograms generated using UPGMA and Jaccard's index.

Mantel test
A Mantel test based on Pearson's correlation was used to highlight correlations between the matrices of albumins (matrix A), globulins (matrix B) and prolamins (matrix C). The p value was calculated from the distribution of r(AB) using 10,000 permutations with the value of r(AB.C) = 0.3099. This test showed significant correlation between the three fractions studied since the calculated p-value (<0.0001) is below the significance level of alpha (0.05 = 5%). Concerning the correlation between ecogeography and seed proteins, r values were −0.0012, −0.0039 and 0.0166 respectively for albumins, globulins and prolamins. p-values are 0.8233, 0.9319 and 0.3689 respectively for the three fractions. Thus, Mantel test showed no significant correlation between ecogeography and protein patterns since the calculated p-values are below the significance level of alpha.

Discussion
The discrimination in the genus Vicia into subgenera, sections and subsections was undertaken by several studies, based on morphological and cytological analyses (Hanelt and Mettin 1989;Kupicha 1976;Leht 2009). In the present work, seed storage proteins and ecogeographic parameters were used.

Seed proteins variability
The differences among accessions were observed and all eleven taxa can be recognized by their protein profiles. Samples within each taxon showed a different number of bands with different molecular weights. Thus, intraspecific heterogeneity is obtained. A positive correlation was exhibited between seed globulin, seed albumin and seed prolamin contents (highly significant). Our results partially confirmed classification of Vicia by Kupicha (1976), Hanelt and Mettin (1989) and Leht (2009) at subgeneric and sectional levels. According to Osborne (1924), proteins are classified into albumins, globulins, prolamins and glutelins based on their solubility which is a convenient method to initiate the discrimination of the seed storage proteins from a species that has not been studied in detail (Ribeiro et al. 2004). The differences in the three fractions profiles of individual seeds was expected since Mudzana et al. (1995) and Goodrich et al. (1985) found that there was variability in the total seed storage protein profiles of individual seeds within a subspecies. This was probably due the cross fertilization nature of the genus.

Albumins patterns
Albumins cluster analysis revealed five major clusters, differing only in the relative position of some accessions in subgroups. Populations of V. monantha subsp. calcarata belonging to section Cracca of the subgenus Vicilla (sensu Kupicha) are linked to samples of V. lutea (Sect. Hypechusa) which indicates a close relationship between the two subspecies of V. lutea when it is difficult to determine distinct groups which could be individually identified as eu-lutea or as vestita. There are bands which are specific of some accessions and can be used as markers to discriminate samples at interspecific level. Discrimination at intraspecific level is also obtained by albumins patterns. The use of albumins proved to be helpful in revealing interspecific variability and intraspecific diversity in the studied taxa. Some bands are specific constant markers for each taxon and can be discriminated bay their electrophoregrams. Other bands are common of several taxa. Albumins present the highest bands number which indicates major role of albumin heterogeneity in discriminating the Vicia samples (Mustafa 2007).

Globulins patterns
Cluster analysis of globulins patterns revealed six basic groups. A low distance can be observed between samples of V. lutea subsp. vestita or by samples of V. lutea subsp. eu-utea and V. lutea subsp. vestita. V. narbonensis (sect. Narbonensis) and V. lutea subsp. vestita (sect. Hypechusa) present close globulin profiles. These results concord with those of Jaaska (1997), Jaaska and Leht (2007) and Shiran and Raina (2014), which showed the species of sections Hypechusa as sister to the clade of section Narbonensis. High distances are observed between V. narbonensis (sect. Narbonensis) and V. sativa (sect. Vicia) and between species of V. sativa or between V. lutea subsp. eu-lutea (sect. Hypechusa) and V. sativa.
Globulins are the major storage proteins present in seeds of legumes (Freitas et al. 2000) and differences both at intraspecific and interspecific levels can be obtained by globulins patterns. A good example in this case is the one of sample 43 (V. monantha subsp. calcarata) characterized by 3 specific bands (51.74, 78.27 and 115.04 kDa). The unique population of V. leucantha also has a specific band (119.96 kDa). Samples of V. monantha subsp. calcarata are a good example for intraspecific heterogeneity as shown by accessions 102 and 43, characterized by one specific band each.

Prolamins patterns
Cluster analysis of prolamins patterns revealed four basic groups. Few studies have been reported concerning the utilization of prolamins patterns to discriminate Vicia taxa in comparison with albumins and globulins. The classification obtained using the UPGMA showed that samples belonging to the same taxon close together in the clusters. A common profile was observed for samples 1 and 4 belonging to V. lutea subsp. vestita, but this taxon also showed other patterns. It is may be due to the fact that native wild populations are composed of a mixture of genotypes which provide survival advantages in varied environmental conditions. Outcrossing could also be an explanation of diversity in the accessions studied, as indicated for several types of vetch (Hanelt and Mettin 1989;Mirali et al. 2007). The sample 10 (V. sativa subsp. obovata) showed differences in its electrophoregram compared to the other taxon members, and might be considered an ''off type'' as proposed by De la Rosa and Gonzalez (2010). Prolamins patterns are a good discriminatory marker in Vicia taxa at both intraspecific and interspecific levels especially for V. sativa samples as the subspecies belonging to this species can be characterized by specific bands as in the case of samples 6, 10, 57, 82 and 95 of V. sativa subsp. obovata or samples 48, 77 and 42 belonging to V. sativa subsp. cordata. V. leucantha, V. tenuifolia and V. narbonensis are also characterized by specific bands which could be considered as markers at interspecific level.

Interspecific and intraspecific variation V. sativa s.l. (Sect. Vicia)
Vicia sativa is the most polymorphic species of the genus Vicia and the debate about its taxonomic classification is extensive. In the present work, the all studied taxa of V. sativa s.l (section Vicia) are found in the same group on the basis of albumins homology and can be found in three clusters which indicates a close relationship between subspecies of V. sativa when it is difficult to observe separate groups which could be identified as obovata, consobrina, cordata or angustifolia. In our previous paper using plant morphology (Bechkri and Khelifi 2016), our results have demonstrated that in the V. sativa using morphological traits alone do not provide a stable grouping. A close relationship between samples can be seen. The picture generated between the phylogenetic trees may be due to the possible phylogenetic instability of these taxa as indicated by Leht (2009).

V. narbonensis (Sect. Narbonensis) and V. lutea (Sect. Hypechusa)
On the basis of albumins patterns, all populations of V. narbonensis belong to the same subcluster except for the accession 23 which is linked to accessions of V. sativa.
Globulins patterns linked all accessions of V. narbonensis together. Accessions of V. narbonensis clustered together using prolamins profiles. Albumins, globulins and prolamins patterns joined all samples of V. lutea in the same subcluster with no discrimination between subsp. eulutea and subsp. vetsita. These observations show that there is an overlap between accessions of these two subspecies. The utilization of seed storage proteins shows a close relationship between the taxa when it is difficult to distinguish groups which could be identified as eu-lutea or as vestita. Albumins and globulins profiles not link accessions of V. lutea and V. narbonensis. Prolamins patterns of the present study concord with those of Jaaska (1997) and Jaaska and Leht (2007) and Shiran and Raina (2014) which showed the species of sections Hypechusa as sister to the clade of section Narbonensis. Our data revealed that subgenus Vicia is a well-separated subgenus and agreed with the results based on morphology reported by Diklic (1972) and with results on phylogenetic relationships (Potokina et al. 1999;Leht 2009). Seed albumins, globulins and prolamins patterns showed V. lutea samples to form an homogenous group. The same findings were reported by Przybylska and Zimniak-Przybylska (1997).

V. monantha, V. tenuifolia and V. leucantha (Section Cracca)
Samples of V. monantha clustered together in two different groups on the basis of albumins patterns with no distinction between the two subspecies calcarata and cinerea. An exception is observed for two samples ( Kupicha (1976) in the sub-genus Vicilla, section Cracca. Thus, the three species attributed to the section Cracca are joined in a separate group in the present work. V. tenuifolia and V. leucantha, are grouped in one subcluster of closely related taxa that provided strong homologous variation with shared characters. As a consequence, the treatment of V. tenuifolia, V. monantha and V. leucantha in the section Cracca is supported. In the present analysis, V. leucantha, the species transferred by Ball (1968) to his section Ervum, is in the same clade with the remaining Cracca species. In spite of this, our analysis of seed proteins supports Kupicha's placement of V. leucantha in section Cracca as was also done by Davis and Plitmann (1970).

Ecogeographic characterization
As the first step towards more efficient conservation is to undertake an ecogeographic study (Maxted et al. 1996), the aim of the present work was to collect ecogeographic informations from investigated stations of Vicia L. Analysis of the passport data will elucidate each taxon's geographic and ecological location. The distribution maps will be used in the planning of future collecting missions. The wide geographic ranges may explain the high degree of protein seed storage variation among accessions and should be considered in conservation programs of this Vicia taxa (El Bakatoushi and Ashour 2009). The obtained intraspecific diversity within the taxa reflects a wide geographical and ecological distribution of this species as reported by Ehrman and Maxted (1990) and Maxted (1995). Studying the species from different geographic regions and altitudes, indicates that the species may be still evolving in different pathways as reported by Ashour et al. (2005). According to Hannelt and Mettin (1989), Vicia taxa do not tolerate extreme environmental conditions. Whereas, Francis et al. (2000) report that V. sativa has good adaption to adverse environmental conditions. Cluster analysis shows that samples having differences in electropherograms and belonging to different taxa can belong to an identical bioclimate and altitudes as in the case of accessions 19 and 32 which were collected from the same locality. In parallel, there are samples with high protein homology level which are collected from stations belonging to the same bioclimate. Accessions 72 and 80 are a good example in this case. The dendrogram obtained with ecogeographic parameters did not indicate clear discrimination among accessions based on their geographical locations. The Mantel test between proteins patterns and ecogeography indicated that the correlation between proteins profiles resemblance and geographical origin is less significant. The same findings were reported by Chung et al. (2013), Potokina et al. (2003), Mirali et al. (2007) and De la Rosa and Gonzalez (2010). Considering all stations of the current paper, the studied samples of Vicia L. occur from 1 to 1222 m. Stations belong to seven different bioclimates (SH, LH, HSA, MSA, LSA, HA) and are characterized by cool, wild, warm or temperate winters. Considering all stations of the current work, V. sativa L. occurs from sea level to 880 m which is consistent with the findings of Maxted (1995). Samples of V. narbonensis were collected from sites receiving between 382.95 and 697.18 mm of precipitations and belonging to bioclimates characterized by cool or mild winter. Bennett and Maxted (1997) reported that the V. narbonensis occur over a wide range of altitudes, from sea level to 3200 m when Abd El Moneim (1992) reported that V. narbonensis adapts in areas receiving 250-300 mm annual precipitations and are characterized by low winter temperatures. Accessions of V. lutea were collected from sites belonging to four bioclimates (LH, HSA, MSA, SH) and altitudes ranged between 11 and 604 m with a minimum temperatures ranged from 2.55 to 8.64 °C. Accessions 74, 77, 91 and 98 belonging to V. monanatha occur until 1222 m and at these altitudes, they require some frost tolerance as the temperatures can drop to 0.63 °C as in the case of the accession 98. V. leucantha was collected from a site characterized by HSA bioclimate with a minimum temperature of 2.62 °C (cool winter) and an altitude of 586 m. The two samples of V. tenuifolia occurred in SH and MSA bioclimates with mild or cool winters and at altitudes of 276 and 543 m. These patterns are not necessarily a real picture of the preferred altitude of these two taxa. A larger number of accessions from each geographical location should be tested to confirm patterns.

Conclusion
The electrophoregrams obtained can be exploited as passport data for the genetic diversity of the studied taxa. Seed protein electrophoresis is a valid tool for taxa discrimination. The variability observed indicates that improvement by simple selection for these traits is possible. No significant correlation is obtained between seed proteins and ecogeography. The use of more samples from different origins is necessary to include most of the genetic determinants of these traits.