Genetic Diversity and Differentiation of Eleven Medicago Species from Campania Region Revealed by Nuclear and Chloroplast Microsatellites Markers

The species belonging to the genus Medicago are considered a very important genetic resource at global level both for planet’s food security and for sustainable rangelands management. The checklist of the Italian flora (2021) includes a total number of 40 Medicago species for Italy, and 27 for Campania region, with a number of doubtful records or related to species no more found in the wild. In this study, 10 Medicago species native to Campania region, and one archaeophyte (M. sativa), identified by means of morphological diagnostic characters, were analyzed in a blind test to assay the efficacy of nine microsatellite markers (five cp-SSRs and four n-SSRs). A total number of 33 individuals from 6 locations were sampled and genotyped. All markers were polymorphic, 40 alleles were obtained with n-SSRs ranging from 8–12 alleles per locus with an average of 10 alleles per marker, PIC values ranged from 0.672 to 0.847, and the most polymorphic SSR was MTIC 564. The cp-SSRs markers were highly polymorphic too; PIC values ranged from 0.644 to 0.891 with an average of 0.776, the most polymorphic cp-SSR was CCMP10. 56 alleles were obtained with cp-SSRs ranging from 7 to 17 alleles per locus with an average of 11. AMOVA analysis with n-SSR markers highlighted a great level of genetic differentiation among the 11 species, with a statistically significant fixation index (FST). UPGMA clustering and Bayesian-based population structure analysis assigned these 11 species to two main clusters, but the distribution of species within clusters was not the same for the two analyses. In conclusion, our results demonstrated that the combination of the used SSRs well distinguished the 11 Medicago species. Moreover, our results demonstrated that the use of a limited number of SSRs might be considered for further genetic studies on other Medicago species.


Introduction
Fabaceae, the third-largest Angiosperm family with 751 genera and 19,400 species [1], is a highly diversified plant family and features many economically important crops, ranging from food crops to fodder species. It provides about one-third of protein for human consumption and a wide range of raw materials for industries [2]. Legumes are able to improve soil fertility by fixing atmospheric nitrogen through symbiotic bacteria (Rhizobium) and have an important role in the global nitrogen cycle [3][4][5][6].
Among the six legumes subfamilies, Papilionoideae is the richest for number of species and the most studied one, embodying 476 genera and 13,860 species [2,[7][8][9][10][11]. In this subfamily, the genus Medicago is arguably one of the most common in Mediterranean and

Plant Material
During field surveys and laboratory observations, 11 Medicago species were identified based on their morphological characters of vegetative and reproductive organs. We considered a large number of characters, such as, e.g., petiole length, leaflet margin, stipule length, stipule margin, stipule shape, inflorescence length, number of flowers per inflorescence, pod length, pod width, etc. These 11 Medicago species included one perennial, one biannual, or short-lived perennial, and nine annual species. Aiming to analyze 3 individuals for each species (Table 1), 33 specimens were collected along secondary or dust road verges in the Province of Salerno (Campania region, Southern Italy, latitude 40.07-40.77 • , longitude 14.78-15.55 • ). All collection sites were GPS georeferenced and plotted on a geographic map (Table 1, Figure 1). Table 1. Scientific names, sections, and subsections of the 11 Medicago species collected in Campania, with geographical coordinates of the collection sites (WGS84) [4]. The status in Campania follows the checklist of the Italian flora (2021).

DNA Extraction and SSR Analysis
For each Medicago species, leaf samples were harvested from three individuals representative of the population thriving in the collection site, and then stored in silica-gel before DNA extraction and purification. Leaf samples were randomly numbered to perform a blind test. Afterwards, leaves were powdered in liquid nitrogen and total DNA was extracted using a cetyl-trimethyl-ammonium bromide (CTAB) buffer, following the protocol of Doyle and Doyle (1987), as modified by Harbor, Doyle and Tai [43][44][45]. DNA quantity and quality were assessed using 1.0% agarose gel electrophoresis. Genetic analysis was carried out by means of: (a) four nuclear microsatellites (n-SSR) from M. truncatula selected on the basis of their position on the genetic linkage map [46] (MTIC 503, MTIC 559, MTIC 563, MTIC 564) (Table S1); (b) five chloroplast Simple Sequence Repeat (cp-SSR) markers comprising CCMP2, CCMP4, CCMP6, CCMP7, and CCMP10, designed for Nicotiana tabacum L. [47,48] (Table S2). ule length, stipule margin, stipule shape, inflorescence length, number of flowers per inflorescence, pod length, pod width, etc.. These 11 Medicago species included one perennial, one biannual, or short-lived perennial, and nine annual species. Aiming to analyze 3 individuals for each species (Table 1), 33 specimens were collected along secondary or dust road verges in the Province of Salerno (Campania region, Southern Italy, latitude 40.07-40.77°, longitude 14.78-15.55°). All collection sites were GPS georeferenced and plotted on a geographic map (Table 1, Figure 1).

Genetic Diversity Analysis
Gene diversity and allele number (A) for each locus were obtained using Molecular kinships ver. 3.0. software, PIC values were calculated using the following formula for molecular markers [49]: where PIC i is the polymorphism information content of the i allele and f i is the frequency of amplification (presence of fragment) of the i allele in the analyzed individuals. In order to estimate the level of genetic diversity present in the species from n-SSR data, banding profiles generated by each marker were scored on the basis of the size (bp) of amplified fragments. Deviation from Hardy-Weinberg Equilibrium (HWE) was tested at both species and locus levels and inbreeding coefficients (F) were calculated using GENEPOP software (ver. 4.7.5, Montpellier, France). The following genetic diversity indices were calculated using GenAlEx software (ver. 6.503) [50]: (i) The total number of alleles (Na); (ii) The effective number of alleles (Ne); (iii) Observed heterozygosity (H o ); (iv) Expected heterozygosity (H E ); (v) Shannon's information index (I); (vi) Gene flow (Nm) was calculated using the following formula [50]: In order to assess the variance among and within species and to estimate genetic differentiation among Medicago species, analysis of molecular variance (AMOVA) was performed using 1000 permutations of the F ST value; principal coordinates analysis (PCoA) and relationships between genetic and linear geographic distances (isolation by distances, IBD), were examined using a Mantel test [51] as implemented in GenAIEx, with 1000 permutations.
The alleles banding profiles were transformed into a binary matrix of presence (1)/absence (0) of each allele and genetic relationships were visualized using cluster analysis and the R package 'pvclust' [52], based on Euclidean distance, since this method has been proved to be the most appropriate for recognizing the genetic structure extractable from the analyzed dataset.

Genetic Structure Analysis
The genetic structure was investigated with a Bayesian approach with Structure software (ver. 2.3.4) [53], through 100,000 Monte Carlo Markov Chain (MCMC) iterations, following 25,000 burn-in length for each run. Eleven independent simulations and eight replicates were conducted for each K-value to estimate group assignments.
The analyses were conducted combining two different models (admixture/no-admixture) and two options of allele frequencies among species (correlated/independent), and the other parameters were set to default values as suggested by Pritchard et al. [54]. Structure Harvester b (v0.6.94) [55] was used to select the optimal model relying on maximum likelihood and (∆K) values. ∆K based on the order rate of change of L(K) between successive K values, was used to identify the correct number of K [56].

Cp-SSRs and Statistical Analysis
where n is the number of alleles and p i the frequency of the ith allele in species. Haplotypes Diversity was calculated in the same manner with n and p i referring to haplotypes. An unweighted pair group method with arithmetic mean (UPGMA) clustering analysis was run with NTSYS pc 2.02j software [57], using a clustering algorithm based on the Jaccard similarity index [58]. The Reliability of SSR allele clustering was assessed by bootstrapping, with 1000 permutations.

Genetic Diversity at Nuclear Microsatellites
The four n-SSR were found to be highly polymorphic, with allele number per locus ranging from 8 (MTIC 559) to 12 (MTIC 564), as reported in Table 2. A total of 40 alleles were amplified from the DNA of the 11 Medicago species, with a mean value of 10 alleles per SSR locus. Gene diversity (He) per locus ranged from 0.695 (MTIC 563) to 0.861 (MTIC 503) with an average of 0.799. The Polymorphism Information Content (PIC) for each primer was in the range of 0.672 to 0.847 with an average of 0.780. The results of n-SSR (Table 2) suggested a moderate level of genetic diversity of the studied Medicago species: Na values ranged from 1.2 (POL) to 2.2 (SAT and SCU), Ne values from 1.2 (POL) to 2.2 (SCU and ORB), HO and He values ranged from 0.00 (POL) to 0.33 (SAT) and from 0.11 (POL) to 0.46 (ARA), respectively. Within the 11 species, the average number of alleles revealed by the surveyed loci was 6.3; it ranged from 3 (POL) to 8 (SAT, ARA, LIT and MRX). Although the number of migrants (Nm) ranged from −0.64 (LUP) to 0.49 (RUG).
Inbreeding indices (F) deviated from zero for almost all the species, they were in the range of −0.33 (LUP) to 1.00 (POL) and showed heterozygote deficiencies: a total loss of heterozygosity in M. polymorpha and a loss of 33% of homozygosity in M. lupulina ( Table 3).
The results of the overall AMOVA (Table 4) indicated that 52% of the variation was due to differences among species, while the remaining 48% was due to the variation within groups.
The first two axes of the PCoA ( Figure 2) explained 70.65% of the total variation and clearly separated M. sativa and M. scutellata from all the other species, with an accumulated variance of 42.21% and 28.44%, respectively. Although the F ST values ranged from 0.03 to 0.333 (Table S3), indicating a moderate or great genetic differentiation, the two highest differentiations were observed between M. scutellata and M. murex, and M. scutellata and M. polymorpha, whilst the lowest differentiation was found between M. minima and M. rugosa.  The clustering, based on Euclidean distances (Figure 3), showed a partition of individuals in three main groups (1-3 in Figure 3), which were separated by an Euclidean distance value of 2.75 and 2.56, respectively. In order to read these results and for a better approximation to an unbiased p-value (AU), AU values were adopted instead of BP-values [59]. Nonetheless, approximately AU values were not significant (AU < 95) for most branches. However, individuals belonging to the 3 different Medicago sections were located in different clusters. The first two axes of the PCoA (Figure 2) explained 70.65% of the total variation and clearly separated M. sativa and M. scutellata from all the other species, with an accumulated variance of 42.21% and 28.44%, respectively. Although the FST values ranged from 0.03 to 0.333 (Table S3), indicating a moderate or great genetic differentiation, the two highest differentiations were observed between M. scutellata and M. murex, and M. scutellata and M. polymorpha, whilst the lowest differentiation was found between M. minima and M. rugosa. The clustering, based on Euclidean distances (Figure 3), showed a partition of individuals in three main groups (1-3 in Figure 3), which were separated by an Euclidean distance value of 2.75 and 2.56, respectively. In order to read these results and for a better approximation to an unbiased p-value (AU), AU values were adopted instead of BP-values [59]. Nonetheless, approximately AU values were not significant (AU < 95) for most branches. However, individuals belonging to the 3 different Medicago sections were located in different clusters. The first smaller branch of cluster 1 included individuals from Medicago and Spirocarpos subsection Rotatae; the cluster 2 included individuals belonging to Orbiculares and to two Spirocarpos subsections (Pachyspirae and Spirocarpos). All the others grouped in cluster 3 and included species belonging to Spirocarpos section and a number of subsections (e.g., Rotatae, Lupularia, etc.).
The 33 individuals of 11 species were divided into almost the same groups by the UPGMA dendrogram based on the Jaccard similarity index as by the Pvclust clustering method; although well distinguished (Figure 4), M. scutellata and M. sativa formed a sep- The first smaller branch of cluster 1 included individuals from Medicago and Spirocarpos subsection Rotatae; the cluster 2 included individuals belonging to Orbiculares and to two Spirocarpos subsections (Pachyspirae and Spirocarpos). All the others grouped in cluster 3 and included species belonging to Spirocarpos section and a number of subsections (e.g., Rotatae, Lupularia, etc.).
The 33 individuals of 11 species were divided into almost the same groups by the UPGMA dendrogram based on the Jaccard similarity index as by the Pvclust clustering method; although well distinguished (Figure 4)

Genetic Structure at Nuclear Microsatellites
To assign individuals to one or more estimated groups (K), a Bayesian Markov Chain Monte Carlo approach, implemented in Structure (ver. 2.3) [53], was applied under the no-admixture model and with the assumption of independent allele frequencies between species. At K = 2, Medicago species were assigned into groups belonging to different sections. At K = 3 (

Genetic Structure at Nuclear Microsatellites
To assign individuals to one or more estimated groups (K), a Bayesian Markov Chain Monte Carlo approach, implemented in Structure (ver. 2.3) [53], was applied under the no-admixture model and with the assumption of independent allele frequencies between species. At K = 2, Medicago species were assigned into groups belonging to different sections. At K = 3 ( subgroups: M. sativa, M. scutellata, and M. rugosa were allocated to one subgroup (cluster 1, blue); M. lupulina, M. arabica, M. muricoleptis, and M. murex in cluster 2 (green), while the cluster 3 (red) was formed by M. littoralis, M. minima, and M. orbicularis. It is worthwhile to note that M. polymorpha had the highest average ancestry coefficient (inferred proportion of membership) from the cluster 1 (0.7) then from cluster 2 (0.3). Consequently, K=3 was recognized as the most appropriate value able to describe the genetic structure of the 11 Medicago species studied ( Figure 5).

Genetic Analyses by Chloroplast Microsatellites
The statistics results for the five cp-SSR markers are summarized in Table 2. The loci showed a high level of genetic diversity. In total, 56 polymorphic alleles were amplified, ranging from 7 alleles per locus in the case of CCMP6, to 17 alleles per locus in the case of CCMP10, with an average of 11. Gene diversity (He) per locus ranged from 0.694 (CCMP6) to 0.899 (CCMP10), with an average of 0.833. Meanwhile, Polymorphism Information Content (PIC) of each primer was in the range of 0.694 (CCMP6) to 0.899 (CCMP10), with an average of 0.833. These results demonstrate that the cp-SSR markers used were enough informative to justify further Medicago species genetic diversity analysis. CCMP10 was the cp-SSR showing the highest ability to distinguish among the different analyzed species.
Gene diversity (He) for the 11 species (Table 3)  The AMOVA (Table 4) showed that genetic variation was mainly within species (94%), rather than among species (6%). Although genetic differentiation among species was found from moderate to high, the PHI-PT values for haplotypes ranged from 0.083 to 0.600, the highest differentiation was observed between M. orbicularis and M. murex.
Mantel test for isolation by distance among species did not show any significant correlation between pairwise PHI-PT and geographic distance (R 2 = 0.012, p = 0.158) [60].
Unbiased cluster analysis based on cp-SSR markers was performed with the Numerical Taxonomy Multivariate Analysis System (NTSYS-PC-ver. 2.2) ( Figure 6) [3]. A dendrogram was built via UPGMA and Jaccard similarity coefficients were used to reveal the similarity among the 11 Medicago species. The results of clustering analysis with pvclust based on Euclidian distances revealed the presence of two macro-clusters (A and B). These two clusters included different sub-clusters, A1, A2, B1, B2, and B3. In particular, A1 was formed by the only M. murex, its specimens were well separated from M. sativa and M. scutellata cluster (A2). In the case of the macro cluster B, the three sub-clusters B1, B2, and B3 showed that different Medicago species grouped together, even if the three clusters were well separated. However, the identified subclusters comprised all the individuals of the same species (e.g., B1 grouped all the specimens belonging to M. rugosa, B2 all those belonging to M. littoralis, B3 the ones belonging to M. lupulina). In general, the AU value was quite well supported by data obtained, in particular in the case of the subgroups (AU comprised between 53% and 99%).

Pvclust-R-Package
The dendrogram based on Euclidian distance calculated for the most informative microsatellite (5 chloroplast and 4 nuclear), showed three major clusters (A, B, and C; Figure  7). M. murex formed an independent cluster (A − AU = 100%), M. scutellata and M. sativa, belonging to two different sections, formed a distinct cluster (B) and their relative branches showed a high AU values (95%), which explained that those subdivisions are strongly supported by the data. The third cluster (C- Figure 7), including the large part of the samples, was divided in six subclusters able to well distinguish the Medicago species: C1 grouped M. orbicularis belonging to section Orbiculares; C2 cluster comprised the specimens belonging to M. littoralis, two belonging to M. minima and one to M. polymorpha; the subcluster C3 incorporated the samples of M. rugosa and one of M. minima; C4 was identified as the cluster including M. lupulina specimens; C5 was formed by the M. murex; C6 grouped the samples belonging to M. arabica.

Pvclust-R-Package
The dendrogram based on Euclidian distance calculated for the most informative microsatellite (5 chloroplast and 4 nuclear), showed three major clusters (A, B, and C; Figure 7). M. murex formed an independent cluster (A − AU = 100%), M. scutellata and M. sativa, belonging to two different sections, formed a distinct cluster (B) and their relative branches showed a high AU values (95%), which explained that those subdivisions are strongly supported by the data. The third cluster (C- Figure 7), including the large part of the samples, was divided in six subclusters able to well distinguish the Medicago species: C1 grouped M. orbicularis belonging to section Orbiculares; C2 cluster comprised the specimens belonging to M. littoralis, two belonging to M. minima and one to M. polymorpha; the subcluster C3 incorporated the samples of M. rugosa and one of M. minima; C4 was identified as the cluster including M. lupulina specimens; C5 was formed by the M. murex; C6 grouped the samples belonging to M. arabica.

Unweighted Pair-Group Method with Arithmetic Averages Using NTSYS-PC Software
The dendrogram derived from Jaccard coefficient (Figure 8) based on similarity matrix of the Medicago species showed three major groups (A, B, and C), and the similarity coefficient ranged from 5 to 82%. The cluster A can be further divided into three subgroups (A1, A2, and A3) having different degrees of similarity; the A1 was represented by M. littoralis (subsection Pachyspirae), A2 comprised M. orbicularis (section Orbiculares), whilst A3 included 2 of the 3 specimens of M. polymorpha. The group B embodied five species of the Spirocarpos section, with four sub-clusters; M. rugosa formed a separated subgroup B1; B2 included M. murex, whilst B3 and B4 were formed by M. muricoleptis and M. lupulina, respectively. However, one sample of M. arabica, included in cluster B4, showed roughly 35% of similarity with M. lupulina.
Finally, this dendrogram highlighted the fact that M. arabica specimens had a very high biodiversity; in fact, their Jaccard similarity index was equal to 20% between ARA2 and ARA3. Similarly, the specimens identified as M. minima showed a very low similarity index, and were wide spread among all the identified clusters.
The cluster C embodied M. scutellata and M. sativa, belonging to two different sections, (e.g., Spirocarpos and Medicago, respectively).

Unweighted Pair-Group Method with Arithmetic Averages Using NTSYS-PC Software
The dendrogram derived from Jaccard coefficient (Figure 8) based on similarity matrix of the Medicago species showed three major groups (A, B, and C), and the similarity coefficient ranged from 5 to 82%. The cluster A can be further divided into three subgroups (A1, A2, and A3) having different degrees of similarity; the A1 was represented by M. littoralis (subsection Pachyspirae), A2 comprised M. orbicularis (section Orbiculares), whilst A3 included 2 of the 3 specimens of M. polymorpha. The group B embodied five species of the Spirocarpos section, with four sub-clusters; M. rugosa formed a separated subgroup B1; B2 included M. murex, whilst B3 and B4 were formed by M. muricoleptis and M. lupulina, respectively. However, one sample of M. arabica, included in cluster B4, showed roughly 35% of similarity with M. lupulina.
Finally, this dendrogram highlighted the fact that M. arabica specimens had a very high biodiversity; in fact, their Jaccard similarity index was equal to 20% between ARA2 and ARA3. Similarly, the specimens identified as M. minima showed a very low similarity index, and were wide spread among all the identified clusters.
The cluster C embodied M. scutellata and M. sativa, belonging to two different sections, (e.g., Spirocarpos and Medicago, respectively).

Discussion
In addition to morphological identification and study of diagnostic traits, molecular characterization can help to solve a number of taxonomic ambiguities [30]. In the present study, we assessed, for the first time, whether nuclear and/or chloroplast microsatellites may be suitable for the differentiation of 11 Medicago species (and relative sections) collected in South Italy in Campania region.
The high level of synteny between legume genomes allowed the transferability of SSR markers developed in M. truncatula [33] (more than 500 SSR) to the other Medicago species [21,[34][35][36]. Contrariwise, chloroplast genomes have lower evolutionary rate, are not recombining predominantly, maternally inherited, highly conserved across genera, and often distributed throughout non-coding regions.
Compared to morphological data, molecular markers can provide highly reliable information, in fact, they are insensitive to environmental variations and, furthermore, not subject to personal interpretation. Previous studies showed that molecular markers are suitable for revealing phylogenetic relationships among different Medicago species and also to estimate the genetic diversity. In fact, different molecular markers (IRAP, REMAP, ISSR, SSR, RAPD, and AFLP) have been used to estimate the genetic diversity and evaluate the phylogenetic relatedness in Medicago species [32,48,[61][62][63][64][65][66][67][68][69][70][71].
A noteworthy result obtained with the present study is that, on a quite limited surveyed area, a significant number of Medicago species were present, including the rare M. muricoleptis species, which was recorded in the Campania region for the first time only in 2019 [72]. The species richness in the study area is highly remarkable, regardless the level of anthropic pressure on the area and the intensive land uses (e.g., agricultural, industrial, tourist etc.). However, this apparent paradox is in line with the renewed tolerance to disturbance of many Medicago species [73].

Discussion
In addition to morphological identification and study of diagnostic traits, molecular characterization can help to solve a number of taxonomic ambiguities [30]. In the present study, we assessed, for the first time, whether nuclear and/or chloroplast microsatellites may be suitable for the differentiation of 11 Medicago species (and relative sections) collected in South Italy in Campania region.
The high level of synteny between legume genomes allowed the transferability of SSR markers developed in M. truncatula [33] (more than 500 SSR) to the other Medicago species [21,[34][35][36]. Contrariwise, chloroplast genomes have lower evolutionary rate, are not recombining predominantly, maternally inherited, highly conserved across genera, and often distributed throughout non-coding regions.
Compared to morphological data, molecular markers can provide highly reliable information, in fact, they are insensitive to environmental variations and, furthermore, not subject to personal interpretation. Previous studies showed that molecular markers are suitable for revealing phylogenetic relationships among different Medicago species and also to estimate the genetic diversity. In fact, different molecular markers (IRAP, REMAP, ISSR, SSR, RAPD, and AFLP) have been used to estimate the genetic diversity and evaluate the phylogenetic relatedness in Medicago species [32,48,[61][62][63][64][65][66][67][68][69][70][71].
A noteworthy result obtained with the present study is that, on a quite limited surveyed area, a significant number of Medicago species were present, including the rare M. muricoleptis species, which was recorded in the Campania region for the first time only in 2019 [72]. The species richness in the study area is highly remarkable, regardless the level of anthropic pressure on the area and the intensive land uses (e.g., agricultural, industrial, tourist etc.). However, this apparent paradox is in line with the renewed tolerance to disturbance of many Medicago species [73].

Intra Species Diversity
Allele number of nuclear SSR within species varied from 3 to 10, with the highest value observed for M. scutellata, M. arabica, and M. minima, and the lowest one for M. murex.
For all of the species studied, the results show a low amount of heterozygosity (H o varied between 0.00 and 0.33, and H e between 0.11 and 0.44). The highest fixation index was found in M. muricoleptis and the lowest one in M. polymorpha. The high self-pollination detected in M. polymorpha (H o = 0.00, H e = 0.11) likely contributed to lower the overall level of the observed heterozygosity. A low heterozygosity was also observed in the case of M. lupulina, confirming the data of Yan and co-workers [67], who reported that the H o within M. lupulina populations was 0.017, ranging between 0.00 and 0.04. Nevertheless, the limited heterozygosity cannot be due to outcrossing events, since it was found that the breeding system of M. lupulina varies from complete self-pollination to extensive outcrossing and, although honeybees show great interest for M. lupulina flowers, under natural conditions, neither pollinating insects nor wind are absolutely necessary for its fertilization [74,75].

Diversity among Species
This study included 11 Medicago species (whit a number of globally poorly studied species) and has been the first survey of this type for Campania region and Italy, so that it is not possible to compare with similar study cases or previous national investigations. Although genetic diversity has fundamental importance for species survival [76,77], few studies are available on less common and endangered species, such as the island endemic M. citrina and others [64]. In fact, at the global level, genetic studies are mainly concentrated on M. truncatula, M. polymorpha, on the M. sativa-M. falcata complex [78], and on M. truncatula, from which widely used SSR markers were identified, as reported in Diwan et al. [62]. In the same vein, Min et al. [78] found that the mean value of information content of the SSR polymorphisms detected in diverse accession of M. truncatula, was as high as 0.71. This number is usually more than 0.70 for annual medics [62,78].
Lesins and Lesins [22] and Small and co-workers [24] considered a variety of genetic and morphological traits, such as chromosome number, presence of woody tissue, or cotyledon structure that support recognition of infrageneric taxa and the delimitation of species within the genus Medicago [13]. However, results presented by Steele et al. [13] suggest that section Lupularia, containing the two species M. lupulina and M. secundiflora, should no longer be recognized. The same authors also propose considering a reduced subsection Leptospirae, with M. lupulina, M. coronata, M. disciformis, M. minima, and M. tenoreana. Interestingly, we detected a relatedness among M. lupulina and M. minima in the PCoA plot (UPGMA tree was drawn using Maximum Parsimony method). Although they are traditionally included in two different sections, PCoA plot and UPGMA dendrogram clearly highlighted a limited genetic distance between M. sativa and M. scutellata, which is thought to be a polyploid derivative of a hybrid between a 2n = 16 species and a 2n = 14 [20]. In another global study of the Medicago genus [77], although based on isoenzyme banding pattern, among fifty Medicago species, representing eight sections (Spirocarpos, Lunatae, Buceras, Medicago, Hymerocarpos, Lupularia, Orbiculares, and Heynianae sections), the dendrograms, based on cluster analysis of isozyme data, showed relatedness among them. Otherwise, population genetics studies on Medicago species in Iran have been mainly limited to M. sativa [68]. On the basis of the data produced with this study on 11 Medicago species assayed through a selection of highly informative nuclear and chloroplast markers, it is possible to conclude that, in most cases and with few exceptions, the distinction of the sections (and subsections) of the genus Medicago is supported by the detected genetic diversity.

Effectiveness of N-SSR and Cp-SSR Markers
Importantly, n-SSR were more effective than cp-SSR markers to separate Medicago species from the genetic point of view. Our n-SSR-based clustering of Medicago species chiefly agreed with the recent taxonomic classifications [4]. In fact, the species belonging to Spirocarpos section were grouped together in the same clade, and species belonging to Orbiculares section formed a distinct group. Nonetheless, for Medicago section M. sativa specimens were grouped in the same clade together with M. scutellata, and this could be due to the small number of microsatellites markers used.
Furthermore, n-SSR markers placement of the species based on Bayesian clustering analysis (which sets individuals to groups in relation with genotype), agreed with their placement based on pvclust non-Bayesian clustering approach. An unexpected clustering of M. polymorpha specimens was observed, which is not surprising and it resulted from a low level of polymorphism within species and the limited number of DNA SSR-markers used.
At the same time, the microsatellites employed in our study (both chloroplast and nuclear) may provide a useful tool for the management of germplasm repository. In fact, such an assay would be particularly relevant and effective, with a relative low cost and time spent, and for the possibility of automation. Our results confirm, once more, the potentiality and the effectiveness of the SSRs in genetic study and, in particular, it provides a successful experience in Medicago species discrimination and suitable for further studies on this genus.

Conclusions
This study focused on the discrimination power of n-SSR and cp-SSR for 11 Medicago species collected in a very limited area, and including M. muricoleptis, a species recorded for the first time in 2019 in Campania region. Nuclear microsatellites have proven very effective for species discrimination. Otherwise, cp-SSR resulted informative as well, being related to the common phylogenetic origin of Medicago species. Therefore, this innovative approach, which combined data from both n-SSR and cp-SSR, has been proven highly informative and allowed to distinguish Medicago species in relation to their common phylogenetic origin.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/genes13010097/s1, Table S1: Molecular and genetic information on the four nuclear microsatellites loci used in the study; Table S2: Molecular and genetic information on the five chloroplast microsatellites used in the study; Table S3

Conflicts of Interest:
The authors declare no conflict of interest.