Assessment of genetic diversity of selected cowpea landraces from Nigeria based on simple sequence repeat markers

Understanding the genetic diversity of cowpea (Vigna unguiculata L. Walp.) landraces is useful for effective characterization and ex-situ conservation of germplasm. The analysis of genetic diversity of eighteen cowpea landraces collected from five agro-ecological zones in Nigeria was reported in this study. Five individuals per landrace were genotyped with six polymorphic microsatellite markers. Three to 5 alleles with a mean of 3.833 were detected. Mean Polymorphic information content (PIC) and observed heterozygosity of the markers were 0.5721 and 0.2433, respectively. Analysis of Molecular Variance (AMOVA) showed that variation due to agroecological zone constituted 24%, while variations among and within landraces as well as within individuals constituted 25%, 17% and 33%, respectively. Landraces collected from the humid rainforest zone showed high within landrace diversity and were not significantly different (P ≥ 0.001) from other landraces collected from the same zone. Landraces from the savannah zones showed low within landrace diversity and homozygous across all loci. Consequently, among landrace diversity was higher in the savannah zone with landraces collected from guinea savannah been the most diverse, followed by landraces from the derived savannah and Sudan savannah. Mantel test showed positive and significant correlation (r= 0.377, p= 0.01) between genetic and geographical distance of landrace collections. The findings are important for up-to-date characterizations of cowpea germplasm in Nigeria for improved breeding programs.


Introduction
Cowpea (Vigna unguicuilata L.Walp.) (2n = 2x = 22) belongs to the family Fabaceae. It is an annual grain legume predominantly grown in the northern parts of Nigeria and widely consumed across the country especially for its high protein (23-29 %) content (Boukar et al., 2016). The seed is rich in lysine, tryptophan, folic acids and vitamins (Nielson et al., 1993). The young leaves and immature pods are consumed in some parts of the country while stems, leaves and vines serve as animal feeds. Africa accounts for about 95 % of the total world production and Nigeria is the largest producer of cowpea in the world accounting for 61 % production in Africa and 58 % worldwide (IITA, 2015). Nigerian cowpea landraces show high variability in seed shapes, sizes, colours, texture, pigmentation and growth habits. Farmers predominantly grow the brown-seeded and white-seeded types which are most preferred by consumers (Mishili et al., 2009).
Generally, landraces are the most diverse populations in cultivated plants (Frankel et al., 1995). They are diverse mixtures of different genotypes and thus show variability within and between accessions. They represent valuable resources that can be explored for the introgression of new genes in varietal improvement (Hedge and Mishra, 2009;Xu et al., 2010). Landraces are generally defined as a population of cultivated plants with a historical background, identity and are adapted to local environment but without any breeding improvement (Camacho-Villa et al., 2005). They are also called 'farmer-developed accessions' or 'traditional varieties'. Resource-poor farmers grow these unimproved landraces despite the availability of improved cultivars (Bellon and Hellin, 2011;Kamara et al., 2012). In pulses like pigeonpea and chickpea, traditional landraces or selections from them are released directly as varieties (Hegde and Mishra, 2009).
Since early 1970s, numerous cowpea landraces have been sampled from different parts of Nigeria and preserved at the seed bank of the Genetic Resources Centre of the International Institute of Tropical Agriculture (IITA) and the National Centre for Genetic Resources and Biotechnology (NACGRAB), Ibadan, Nigeria. Socio-economic changes and drought however, led to a dramatic reduction of cowpea landraces cultivation recently (African Centre for Biodiversity, 2015) and probably to the disappearance of local populations. Data on diversity assessment of these farmer-developed accessions are scarce. To stem this loss of genetic variation, conservation and reconnaissance of existing biodiversity are fundamental. Characterization of most cowpea accessions in these gene banks are based mainly on morphological data which are fraught with environmental variations (Nkongolo, 2003;Ibrahima et al., 2013). Molecular markers, particularly Simple Sequence Repeats (SSR), are playing an increasingly important role in assessment, characterization and conservation of plant genetic resources. The sequences are abundant and randomly distributed throughout the genome, highly polymorphic, inherited codominantly, and are not influenced by environmental variations and have shown great potential for various genetic studies (Tautz, 1989;Ibrahima et al., 2014;Jingade et al., 2014). They have been used to assess the diversity of various cowpea germplasm from different countries (Li et al., 2001;Ogunkanmi et al., 2008;Asare et al., 2010;Gupta and Gopalakrishma, 2010;Ogunkanmi et al., 2014;Adesoye et al., 2016;Wamalwa et. al., 2016).
Genetic diversity between cowpea landraces has been reported, however, few studies have analyzed the diversity within landraces that are held ex situ (Singh et al., 1991;Gomez et al., 2004). Most studies would bulk DNA from several individuals per landrace (Toklu et al., 2009;Adetiloye et al., 2013) thus masking the inherent variation within each landrace. Molecular assessment of diversity within and between cowpea landraces will help to ascertain variations that exist within them and classify them into distinct groups for enhanced breeding programme. Therefore, the objectives of this study were to assess the genetic diversity within and among cowpea accessions earlier collected in various parts of Nigeria and conserved in IITA and NACGRAB gene banks.

Materials and Methods
Planting Materials and description of collection Eighteen cowpea accessions collected from eighteen states across five agroecological zones in Nigeria and curated in the genetic resource centers of IITA and NACGRAB, Ibadan, Nigeria were used for the experiment (Table 1). At least two states from each agroecological zone were represented in the sample. The place and coordinates of collection as well as the growth habits of the accessions were obtained from the genetic resource centers.

DNA Extraction
Two seeds per landrace were sown in five pots filled with sterilized soil. The pots were arranged in complete randomized design in a screen house. Two weeks after planting, fresh leaves were collected from five plants per landrace for DNA extraction. The leaves were collected in properly labeled nylons on dried ice pack and total genomic DNA was extracted using the modified Dellaporta et al. (1983) protocol.
Extracted DNA was washed in 300 µl of 70 % ethanol centrifuged at 3500 rpm for 10 minutes twice. DNA pellets were dried at room temperature for 1 hour, dissolved in 110 µl TE and 1 µl of RnaseA. It was gently mixed and incubated at 37 o C for 30 minutes. Quantification was done using spectrophotometer ND-1000 (Thermo Scientific, Wilmington, DE, USA). DNA was stored at -20 o C for subsequent uses. The research was carried out at the Bioscience laboratory of the International Institute for Tropical Agriculture (IITA), Ibadan, Nigeria.  (Liu and Muse 2005). Gene diversity and heterozygosity within each landrace were also calculated in Powermarker. Analysis of molecular variance (AMOVA) was calculated in GenAlex 6.5 (Peakall and Smouse 2012) based on the linear statistical model described by Weir and Cockerham (1984): Pijkz= P+Az+Bkz+Cjkz+Dijkz; Where, Pijkz indexes the i th allele in the j th individuals within k th landraces in z th agroecological zone. A represents the effects of the agroecological zone, B is the effects of landrace and C is the effects of individuals within landraces while D is the effects of alleles within each individual. Wright's F statistics for the AMOVA was based on 999 pair wise population permutation (Wright 1951;Weir and Cockerham, 1984). Nei's genetic distance matrices among landraces (Nei 1979) were used to construct dendrograms based on Unweighted Pair Group Method with Arithmetic means (UPGMA). Dendrograms were drawn and visualized in Mega 6 (Tamura et al., 2013).

Marker information
The Markers detected 3 to 5 alleles with mean allele number of 3.8333 (Table 3)  Population structure Analysis of molecular variance (AMOVA) partitioned the overall variation of the cowpea landraces into four components: Variation due to agroecological zones, variation among and within landraces and variation within individuals (Table 4). Variation due to agroecological zone of landrace collections constituted 24 %, variation among the landraces accounted for 25 %, while variations among and within individuals accounted for 17 % and 33 %, respectively. The genetic variations in the four levels were significant (p˂0.001).

Genetic diversity measure within landraces
To assess genetic diversity within each landrace; allele number, gene diversity and heterozygosity of five individual plants within each landrace were estimated ( Genetic diversity among landraces based on cluster analysis and pairwise genetic distance Dendrogram (Figure 1) based on UPGMA formed three major distinct clusters of the eighteen landraces. Cluster I comprises of Tvu-3890 (Borno) and Tvu-818 (Yobe) landraces, both of the Sahel savannah.     Table 6: Fst pairwise matrix of eighteen cowpea landraces selected from Nigeria Ns= Nonsignificant; *= P ≥ 0.05; ** = P ≥ 0.01; *** =P ≥ 0.001; 1, 2…18= Serial number of landraces as defined in table1

Discussion
Simple sequence repeat markers were idea to discriminate alleles within and between landraces selected from some states across five agroecological zones of Nigeria. Markers detected 1 to 5 alleles depending on the type of SSR marker. VM 31 detected 5 alleles and a PIC value of 0.6681 while VM 9 and VM 53 detected 3 alleles each with corresponding PIC values of 0.5641 and 0.4174, respectively. In earlier reports, Adetiloye et al. (2013) reported 2 to 4 alleles in some Nigerian cowpea landraces based on SSR markers while Ogunkanmi et al. (2014) reported allele number of 2 to 5 using 14 SSR markers to assay cultivated cowpea collected across Africa. The average PIC value of 0.5721 obtained in this study indicates that the markers were polymorphic. Bostein et al. (1980) stated that PIC value ˃ 0.5 is highly polymorphic and suitable to discriminate alleles of germplasm. Polymorphic information content values of 0.37 and 0.38 lower than the one obtained in this study have previously been reported in cowpea SSR markers (Ogunkanmi et al., 2008 Genetic diversity among five individuals within each cowpea landraces constituted 17 % of the total variation while variation within the 90 individuals across the loci constituted 33 % of the total variation in this study. There was no genetic diversity within NG/MR/11/11/066 landrace collected from Jos; consequently, gene diversity and heterozygosity values were zero. Most of the landraces assessed, especially those from the savannah zones had low gene diversity and consequently were homozygous across all loci. However, landraces from the humid forest zone had high gene diversity and heterozygous across the loci. Different levels of genetic diversity within cowpea landraces have previously been reported. Gomez et al. (2004) reported that two-thirds of the variation observed in Nicaraguan common bean landraces was distributed within the landraces; while Chen et al. (2017) reported a 52 % within population variation in 105 cowpea accessions collected from China, Kenya, Nigeria and Niger. However, Ali et al. (2015) reported a relative low diversity of 9 % within 231 cowpea landraces obtained from Sudan. In this study, genetic diversity within the landraces collected from the savannah agroecological zones is relatively low. This is however common in cowpea as the crop is predominantly self-pollinating in nature with moderate to high outcrossing levels (Gepts 1993;Graham and Ranalli, 1997). We refrained to adjudge any reason to the high genetic diversity observed within the landraces collected from the humid forest zone. Similar result was reported by Gomez et al. (2004) in which one out of four zones assessed in their study recorded high genetic diversity within common bean landraces. They attributed the high genetic diversity of the landraces within the zone to outcrossing at the farm level with resultant gene flow between diverse individuals of same or different landrace(s). Therefore, landraces with broad genetic bases remain reservoir of novel genes for introgression in improving desirable traits in cowpea breeding programs.
Based on the dendrogram, most landraces belonging to same agroecological zones formed same cluster except NG-MR-11-11-066 (Jos landrace) and Tvu-4320 (Gombe) of guinea savannah which clustered with landraces of the derived savannah because of their low within landrace diversity. Similarly, Tvu-663 (Kano landrace) clustered with landraces of humid forest zone because of its high within landrace diversity. Pairwise Fst values among landraces of the humid zone showed that the genetic distances among most of the landraces of this zone were not significant. The genetic distance among the landraces NG/AT/APR/09/017 (Ogun), Tvu-9304 (Delta), NG/OA/MAR/09/010 (Edo), Tvu-4053 (Imo), and Tvu-10862 (Ondo) was not significant. Similarly, the difference between NG/SA/07/0133 (Kaduna) and NG/AO/11/08/44 (Katsina) both of Sudan savannah zone was not significant while Tvu-3926 (Abuja) and Tvu-7842 (Oyo), belonging to derived savannah zone were genetically identical. This could be attributed to 'founder effect' in which landraces within the same zone have common ancestor. It could also be due to farmers' preference to adapt specific landrace(s) to an agroecology and exchange seeds because of their good performance. Nevertheless, some landraces though collected from the same agroecological zones were significantly different as found in the landraces of the guinea savannah and Sudan savannahs, respectively.
Genetic diversity among landraces and agroecological zones contributed 25 % and 24 %, respectively to the total genetic variation. reported 23 % genetic diversity among 33 cowpea genotypes assessed in their study. Genetic diversity among landraces has been attributed to spatial differentiation among sites of collections (Gomez et al., 2004;Boezkowska and Tarezyk, 2013). In this study, genetic distance was positive and significantly correlated with geographic distance of seed collection (r= 0.377, p= 0.01), indicating that the variation among landraces could have been due to geographical distance among landraces. However, landraces with short distances between collection sites are not significantly different as seen in humid forest zone and the Sahel savannah zones.

Conclusions
Simple sequence repeat marker was effective in dissecting genetic diversity within and among cowpea landraces collected from five agroecological zones in Nigeria. The use of SSR marker is important as it provides useful information complementary to phenotypic and morphological characterization of these landraces held ex situ in the gene banks. In this study, there was high genetic diversity within the cowpea landraces collected from the humid forest zone. This in turn resulted in low diversity between the landraces in this zone. Conversely, there was low genetic diversity within the landraces collected from the savannah agroecological zones with consequent high genetic diversity among the landraces; especially within guinea savannah zone. The understanding of the genetic bases of the landraces assessed in this study and the distinct grouping of the landraces will help to plan future breeding programs to maximize heterosis in desired traits in cowpea breeding.