The 20 ISSR primers used in our study generated 144 bands, which 100 % showed to be polymorphic (Table 3). The maximum and the minimum number of amplified bands ranged from 3 (UBC 818) to 10 (UBC 825 and 848), with an average of seven bands per primer. The expected heterozygosity (He) showed variation from 0.03 (ISSR2) to 0.19 (UBC 825), with an average value of 0.11. Similarly, Shannon Index (I) ranged between 0.04 (ISSR2) and 0.25 (825), with an average value of 0.16 (Tabel 3; Fig. 2).
Table 3. Inter-simple sequence repeat (ISSR) primers used in moringa with their respective total fragments, polymorphic fragments, polymorphism percentages, Shannon Index (I) and expected heterozygosity (He).
Primer
|
Total of bands
|
Polymorphic bands
|
Polymorphism (%)
|
He
|
I
|
UBC 807
|
5
|
5
|
100
|
0.09
|
0.13
|
UBC 809
|
4
|
4
|
100
|
0.15
|
0.21
|
UBC 813
|
4
|
4
|
100
|
0.12
|
0.18
|
UBC 811
|
6
|
6
|
100
|
0.07
|
0.10
|
UBC 816
|
8
|
8
|
100
|
0.13
|
0.19
|
UBC 818
|
3
|
3
|
100
|
0.10
|
0.15
|
UBC 823
|
8
|
8
|
100
|
0.11
|
0.15
|
UBC 825
|
10
|
10
|
100
|
0.17
|
0.25
|
UBC 826
|
9
|
9
|
100
|
0.10
|
0.15
|
UBC 827
|
9
|
9
|
100
|
0.07
|
0.11
|
UBC 845
|
8
|
8
|
100
|
0.09
|
0.13
|
UBC 848
|
10
|
10
|
100
|
0.15
|
0.23
|
UBC 855
|
9
|
9
|
100
|
0.11
|
0.18
|
UBC 856
|
7
|
7
|
100
|
0.14
|
0.21
|
UBC 860
|
9
|
9
|
100
|
0.15
|
0.22
|
UBC 864
|
7
|
7
|
100
|
0.05
|
0.08
|
ISSR 1
|
8
|
8
|
100
|
0.15
|
0.23
|
ISSR 2
|
4
|
4
|
100
|
0.03
|
0.04
|
ISSR 4
|
6
|
6
|
100
|
0.11
|
0.16
|
ISSR 6
|
7
|
7
|
100
|
0.12
|
0.18
|
|
7
|
7
|
100
|
0.11
|
0.16
|
Previous studies have reported ISSR primers as an effective molecular marker to characterize genetic diversity in moringa (Hassan et al. 2020; Hassanein 2018; Rajalakshmi, Rajalakshmi and Parida 2019; Saini et al. 2013). In our study, all observed bands were polymorphic. Similar findings were reported by Hassanein (2018), in which 10 ISSR primers amplified 65 bands and 90.8 % were polymorphic for both species Moringa oleifera and Moringa peregrina. Saini et al. (2013) used six ISSR primers to detect the genetic variability among eight Indian cultivars of Moringa oleifera, which was considered the most adequate marker as compared to RAPD and cytochrome P450-based markers, and the results showed a rate of 48.57% of polymorphisms. Similarly, the genetic variability of 97 accessions of Moringa oleifera from India was assessed using 15 ISSR markers and amplified a total of 100 bands, showing 59.6% of polymorphism (Rajalakshmi et al. 2019).
The Shannon’s Index infers genetic diversity, ranging from 0 to 1 in which more diversity is found in values close to 1 (Perry and McIntosh 1991). In our study, the average of Shannon’s Index was 0.16, which indicates a low genetic diversity (Silva et al. 2015). Another indicator of genetic diversity is the expected heterozygosis (He) which varied from 0.03 to 0.15 (mean of 0.11), indicating low variability in the 177 genotypes of moringa. Natural populations normally have a He values higher than zero due to the incorporation of new alleles by crossing (Silva et al. 2014). Similarly, Chaves-Bedoya et al. (2017) also found low genetic diversity in 45 accessions of moringa from Colombia, in which values of He ranged from 0.13 to 0.29. The authors reported that the genotypes from Colombia were probably from a single population or a few populations.
The opposite was observed when evaluating genetic materials from India where more genetic diversity is found since it is the center of origin (Muluvi et al. 1999). A high level of genetic diversity was found evaluating seven advanced breeding lines from different locations in India (Kumar et al. 2017). Wu et al. (2010) also observed higher genetic diversity for genotypes from India and Myanmar (He 0.36 to 0.76) using microsatellite markers. Ganesan et al. (2014) used 19 primers SSR to assess the genetic diversity among 300 genotypes of 12 populations from northern (Himachal Pradesh) and southern (Tamil Nadu) India. They also found a high genetic diversity in the Indian collection and reinforce the idea that moringa was originated in northern India and was progressively establish in southern part where it became more diverse. In another study using genotypes of natural populations from India, Malawi and Kenya showed that Nei’s average values ranged from 0.040 (Kenya) and 0.122 (Indian population), and the highest level of genetic diversity was found within Indian populations (Muluvi et al. 1999). In our study, the highest coefficient of Nei's Genetic Distance was found between two accessions from the Florida exchange, M4 and M 18 (0.310). On the other hand, the closest pair of accession is M23 and M24 (0.048), both from Brazil. Nei’s genetic distance matrix obtained with pairwise accessions is shown in Table 4.
The molecular variance was analyzed by AMOVA (Table 5). The results revealed 56% variance within accessions and 44% variance among accessions. Results reported by Ganesan et al. (2014) and Rajalakshmi et al. (2019) showed 86% and 95% of the variation was found within the population, respectively. Since moringa is a cross-pollinated plant is expected that most variation is found within the population (Leone et al. 2015). Moreover, different factors may affect the genetic variability in moringa populations, such as aleatory mating patterns, genetic drifts caused by changes in allelic frequency, spontaneous mutation, and migration events of alleles within the population (Lakshmidevamma et al. 2021).
Table 5. Analysis of molecular variance (AMOVA) showing the genetic variation within and among accessions from Moringa Genebank of Embrapa Tabuleiros Costeiros based on 20 ISSR primers. Df = degree of freedom, SS= sum of squares, MS =mean of squares, Est. var. = estimate of variance, % = percentage of total variation.
Source of variation
|
|
df
|
SS
|
MS
|
Est. Var.
|
%
|
GST
|
Among Accessions
|
|
24
|
1597.69
|
66.57
|
7.99
|
44 %
|
0.440 ***
|
Within Accessions
|
|
152
|
1549.39
|
10.19
|
10.19
|
56 %
|
|
Total
|
|
176
|
3146.98
|
-
|
18.92
|
100 %
|
|
***p<0.001
The relationship of all 177 genotypes was revealed by genetic distance of Rogers (Fig. 3). Moringa genotypes were categorized into two groups. One cluster was composed of M1, M2, M3, M4, M5, M6, M7, M8, M9, M10, M11, M12, and M13, in which genotypes M4.1, M4.2, M4.6, and M4.7 are highlighted for being in a different subgroup from the others. The second cluster is composed of M14, M15, M16, M17, M18, M19, M20, M21, M22, M23, M24, and M25. Inside of that, three sub-clusters were identified, which include: most of the genotypes from M16 in one sub-cluster; M14 and M15 in the second sub-cluster; and the others in the third sub-cluster. In general, the cluster analysis did not correlate with the geographic origin of the accessions.
Similarly, the Bayesian analysis provided by the Structure software grouped moringa accesses in two broad clusters (K=2) (Fig. 4). Group I (red) included all genotypes from the Florida exchange (M1, M2, M3, M4, M5, M6, M7, M8, M9, M10, M11, M12, M13, whereas accession M11, M12, and M13 were categorized as admixture (values of membership higher than 0.2). On contrary, all accessions from Brazil (M19, M20, M21, M22, M23, M25) were ranked in Group II (green), except for M17 and M18 which were from the Florida exchange. The accessions M14, M15, and M16 were also categorized as an admixture for having membership values lower than 0.8.
The Bayesian analysis provided by Structure software is a well-established tool to obtain information about population structure using bands of molecular markers (Pritchard et al. 2000). In our study, the Bayesian analysis was also a valuable grouping method to categorize moringa genotypes, being possible to identify which genotypes are considered an admixture by stablishing a threshold score (Figure 4). Even though Moringa Genebank accessions were divided into two groups and most of the accession from the University of Florida were allocated in group 1, some of them (M11, M12, M13, M14, M15, M16, M17, and M18) share similarities with accessions collected in Brazil, which belongs to group 2. Similar observations were obtained from the PCoA plot that together with structure data and cluster dendrogram can provide a more reliable outlook from the results (Figure 5).
Principal Coordinate Analysis (PCoA) with ISSR markers showed the distribution of genetic diversity among the accessions across the two axes. The percentage of variation explained was 26.01 % (axis1: 20.12% and axis2: 5.89%) (Figure 5) and agrees with both groups formed in the aforementioned cluster analysis.
A possible explanation for the lack of correlation between genetic variability versus geographic origin can be explained by the origin center of moringa. Most genetic studies with moringa are performed in Asia, more specifically in India, its origin center (Muluvi et al. 1999; Saini et al. 2013; Ganesan et al. 2014; Kumar et al. 2017; Ravi et al. 2020). In addition, studies with cultivated and natural accessions to quantify the genetic diversity of moringa across the world are considered meagre, although the conservation of genetic resources for this species by using germplasm banks is increasing (Boopathi et al. 2021).
In the literature, it is reported that moringa was introduced as an ornamental tree in United States in 1915, more specifically in southern Florida, with seeds from Cuba and Nicaragua. The Director of ECHO (Educational Concerns of Hunger Organization) became interested in moringa due to its adaptability and distributed seeds to other countries such as Haiti and Brazil. In Brazil, the author reported a seed shipment to the state of Maranhão, which resulted in 25.000 trees planted (Morton, 1991).
Despite a wide cultivation of moringa in America, there are only a few studies that attempted to assess the genetic variation of genotypes from Mexico (Avila-Treviño et al. 2017) and Colombia (Chaves-bedoya et al. 2017). Previously, the first molecular characterization of the Moringa Genebank was conducted with only 16 accessions from exchanged germplasm of University of Florida, using RAPD markers (Silva et al. 2012). It was also found low genetic diversity between the accessions which were evaluated. Since then, new accessions of moringa collected in Brazil were introduced and our study represents the first molecular evaluation of genotypes from Brazil using ISSR primers for this specie.
Shahzad et al. (2013) evaluated the genetic diversity in 131 accessions from a wild population in Pakistan and 30 accessions obtained from ECHO (Florida), which were from nine different countries: Haiti, Mexico, Belize, USA (Florida), Zimbabwe, Mozambique, Tanzania, Senegal, and India. Interestingly, even though there was a great genetic diversity in the wild collection from Pakistan, low genetic diversity was found in the accessions obtained from ECHO. Consequently, the authors suggested that those accession are probably from the same population or from a few populations since moringa was introduced from India to different countries by traders or immigrants and ECHO received seeds from theses farmers. It is reported that the fact of introduced populations may show a low genetic diversity can be explained by using related accessions when they were introduced into a new place. In addition, high selection factors can be operating to reduce even more the genetic diversity in those populations (Muluvi et al. 1999). Our findings also agree with this same pattern, that is, the collection of accessions was probably carried out in the same or a few populations, which would explain the low observed genetic diversity.
In our study, we observed a discrepancy in genetic diversity between the Moringa Genebank in Brazil and the species center of origin, indicating a low representation of moringa natural diversity. Therefore, actions to expand the bank are required to cover the species wide diversity and provide genetic material for crop improvement in future plant breeding programs to develop local varieties.