Assessing the genetic diversity of 48 groundnut ( Arachis hypogaea L . ) genotypes in the Guinea savanna agroecology of Ghana , using microsatellite-based markers

1 Department of Crop Sciences, Tshwane University of Technology, Private Bag X680, Pretoria 0001, South Africa. 2 International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), Hyderabad 502324, India. 3 International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), BP 30, Bamako Mali. 4 Chemistry Department, Tshwane University of Technology, Private Bag X680, Pretoria 0001, South Africa.


INTRODUCTION
Groundnut (Arachis hypogaea L.) is one of the most important oilseed crops in the world.Currently, China, India and Nigeria account for largest groundnut production in the world, with Ghana ranking 12 th in land area under groundnut cultivation (FAO, 2014).Groundnut is therefore the most important grain legume in Ghana (MoFA-SRID, 2011) and is grown in all agro-ecologies of the country with the three northern regions accounting for 70% of production (Tsigbey et al., 2003).The grain, fodder, processed oil and cake serve as a major source of cash income for smallholder famers.Groundnut grain contains 20 to 30% protein and 40 to 55% oil (Asibuo et al., 2008) and is therefore important for nutritional security.As a nodulated legume, groundnut can contribute substantial amounts of symbiotic N to cropping systems, which ranges from 43 to 171 kg N ha -1 in Zambia, Ghana and South Africa (Dakora et al., 1987;Nyemba and Dakora, 2010;Mokgehle et al., 2014).In Ghana, the haulms serve as high-protein fodder for livestock during the dry season (Martey et al., 2015).Despite the importance of groundnut to food and nutritional security and its contribution to soil fertility, increased production is constrained by a variety of factors.The yields obtained on farmers' fields are less than 1.0 t ha -1 due largely to biotic and abiotic stress (Naab et al., 2005;2009).For example, foliar diseases such as early and late leaf spot caused by Cercospora arachidicola S. Hori and Cercosporidium personatum Berk.and Curt., respectively, are known to reduce groundnut yields on farmers field (Naab et al., 2009).Groundnut rosette virus and insect pest damage can also reduce groundnut yield (Padi, 2008).Abiotic factors such as low phosphorus and calcium in soils, as well as erratic rainfall, can cause poor yields of groundnut in Ghana (Abubakari et al., 2012;Rademacher-Schulz et al., 2014).The use of improved genotypes is a cost-effective and environmentally-safe approach to increasing yields in fields of resource-poor farmers (Holbrook and Stalker, 2003).Therefore, developing groundnut genotypes with tolerance to both biotic and abiotic stress has the potential to achieve higher yields on farmers' fields (Holbrook and Stalker, 2003).Traditionally, morphological characters have been used to describe traits such as seed distinctiveness, uniformity and stability of genotypes (Holbrook and Stalker, 2003).However, this method is sometimes influenced by the environment and is labour intensive.In recent times, molecular markers (especially DNA-based markers) have been employed as an alternative to the use of morphological traits.Even then, using a combination of the two (that is, morphological characters and DNA markers) has been found to offer a more comprehensive characterisation of plant genotypes (Holbrook and Stalker, 2003).Many DNA-based molecular markers have been used to characterise groundnut.These include restriction fragment length polymorphism (RFLP), amplified fragment length polymerphism (AFLP), sequence characterised amplified regions (SCARs), random amplified polymorphic DNA (RAPDs), and simple sequence repeats (SSR) or microsatellites (Kochert et al. 1991;Garcia et al. 1996;He and Prakash 1997;Hopkins et al., 1999;Cuc et al., 2008;Carvalho et al., 2010).More recently, single nucleotide polymorphism (SNP) has also been applied to the characterization of groundnut (Barkley et al., 2011;Nagy et al., 2012).However, SSR markers appear to have wider application because of their presence in genomes of all living organisms, their high level of allelic variation, their codominant way of inheritance and their potential for automated analysis (Rakoczy-Trojanowska and Bolibok, 2004).SSR markers have thus remained the common routine tool used in the breeding and genetic analysis of groundnut (Pandey et al., 2012a).In this study, 48 groundnut genotypes, comprising advanced breeding lines and farmer varieties that exhibited varying levels of resistance to drought, foliar diseases and aflatoxin contamination, were screened with 53 SSR markers in order to assess (i) the genetic diversity among the groundnut genotypes, and (ii) their potential as parental material in future groundnut breeding programs.

Plant materials
In this study, a total of 48 groundnut genotypes exhibiting varying levels of resistance to drought, foliar diseases and aflatoxin contamination were used.These included 45 advanced breeding lines supplied by the International Crop Research Institute for the Semi-Arid Tropics (ICRISAT) and three commonly-grown varieties used by farmers in Northern Ghana (Table 1).

Plant DNA extraction
Total plant genomic DNA was extracted from young newly emerged leaves of 12-day-old plants using a modified CTAB protocol (Mace et al., 2003).The quality and quantity of DNA were estimated by running the extracted DNA on a 0.8% agarose gel stained with ethidium bromide.The DNA samples were diluted to 5 ng/μl prior to use in PCR analysis.

RESULTS AND DISCUSSION
A total of 53 SSR markers with PIC values ≥ 0.5 were selected and used in this study (Table 2).
Twenty five (25) of the 53 SSR markers (47%) successfully amplified polymorphic fragments in all the 48 groundnut genotypes tested (Table 3).These 25 SSR markers amplified a total of 164 alleles (Table 3).However, the number of alleles per marker was highly variable, and ranged from two for marker GM1357 to 21 for GM1577, with an average of 6.56 alleles per marker.Major allele frequency per marker also ranged from 0.13 to 0.91, which indicated the presence of allelic variants.Marker GM1577 (0.13) recorded the lowest allele frequency, with GM2053 (0.91) and S003 (0.91) showing the highest (Table 3).Large variability was observed between and among the markers for gene diversity, which ranged from 0.172 to 0.929.Markers GM2053 and S003, which were found to have the highest major allele frequency, revealed the lowest gene diversity.In contrast, marker GM1577 which showed the lowest major allele frequency, exhibited the highest gene diversity.Although, all the markers selected for use in this study had a PIC value ≥ 0.5 Pandey et al. (2012a), the PIC values obtained here ranged from a low 0.16 for markers GM2053 and S003 to 0.92 for GM1577, yielding a PIC mean value of 0.57.The observed variation in PIC values in this study could be attributed to genotypic differences in the groundnut material used.To assess the diversity among the 48 groundnut genotypes, Jaccard's similarity coefficient was calculated using NTSYSpc v. 2.1 and a dendrogram generated based on the unweighted pair-group method with arithmetic mean (UPGMA) procedure (Figure 1).Overall, two major clusters were formed at 72% coefficient of similarity.Cluster I consisted mainly of genotypes that matured at 85 to 90 days after sowing (early maturity) with the exception of NKATIESARI, which matured at 110 days after planting (medium maturity).Cluster II showed no clear-cut demarcation as it contained both earlyand late-maturing genotypes.However, the majority of the genotypes in cluster II exhibited tolerance to foliar diseases and aflatoxin contamination.Two sub-clusters (sub-cluster IA and IB or IIA and IIB) were formed within each cluster at 75% coefficient of similarity.Sub-cluster IA contained two early-maturing genotypes and IB    Although the groundnut genotypes were grouped based on their phenotypic attributes, the clustering from DNA analysis did not match these phenotypic groupings.Genotypes from different groups were found to cluster together irrespective of their phenotypic characters, and this observation is consistent with the findings of other studies which also found low genetic diversity within cultivated groundnut (Jiang et al., 2007;Janila et al., 2013).Based on these results, there is a strong possibility that the majority of genotypes tested in this study shared a similar pedigree since few parents have historically been used in groundnut breeding programs (Nigam, 2000;Janila et al., 2013).The limited number of parents used in groundnut breeding programs stems from the fact that the hybridization of two possible diploid ancestors (Arachis duranensis and Arachis ipaensis), followed by chromosome doubling, resulted in the allotetraploid genome of cultivated groundnut (Nagy et al., 2012;Shirasawa et al., 2013;Janila et al., 2013), which introduced a crossing barrier with its wild diploid ancestors and therefore limited the sources of allelic variability needed for groundnut improvement (Nagy et al., 2012;Janila et al., 2013).
A number of reports on the use of SSR markers to characterise groundnut have produced results similar to those obtained in this study.For example, Mace et al. (2006), who used 23 SSR primers to study 22 groundnut genotypes with varying levels of resistance to rust and early leaf spot, recorded 52% polymorphism with a PIC value ≥ 0.5.In a study with 31 groundnut genotypes that exhibited different levels of resistance to bacterial wilt, Jiang et al. (2007) also found that 29 of the 78 SSR primers were polymorphic, and amplified a total of 91 polymorphic loci with an average of 2.25 alleles per primer.Similarly, Tang et al. (2007) employed 34 SSR markers to determine the genetic diversity in four sets of 24 accessions from the four botanical varieties of cultivated groundnut, and found that 16 primers were polymorphic.This led to the conclusion that abundant inter-variety SSR polymorphism exists in groundnut.
A recent study which assessed the diversity of 11 groundnut genotypes using 17 SSR markers, also recorded 24% polymorphism (Shoba et al., 2010).Mondal and Badigannavar (2010) similarly used 26 SSR primers to amplify 136 bands and showed that 76.5% were polymorphic in 20 cultivated groundnut genotypes that differed in resistance to rust and late leaf spot disease.It is therefore interesting that, in this study, a total of 164 bands were amplified using 54 SSR markers on 48 groundnut genotypes, and 25 were found to be polymorphic.But more importantly, the map position and distance of some of the polymorphic markers used in this study (Table 3) have recently been identified on the consensus genetic map of the Arachis genome (Shirasawa et al., 2013).Furthermore, they have also been mapped on different linkage groups from the consensus groundnut maps created to identify QTLs associated with foliar disease resistance and drought tolerance (Table 4).Marker GM1911 in this study is, for example, linked to a drought tolerance QTL (Ravi et al., 2011;Gautami et al., 2012), while markers GM1577 and GM1991 are linked to QTLs associated with tolerance to late leaf spot disease (Sujay et al., 2012).No doubt, the identification of polymorphisms associated with these important agronomic traits has potential for advancing groundnut improvement in Ghana, as they can be used in QTL mapping, and/or marker-assisted breeding activities (for example, marker-assisted backcrossing and markerassisted recurrent selection).
Taken together, the results of this study demonstrate that SSR markers can be very effective in discerning variations among the 48 different groundnut genotypes despite their close relatedness, a finding consistent with other studies (Hopkins et al., 1999;Cuc et al., 2008;Carvalho et al., 2010).The mean PIC value of 0.57 suggests that the primers were highly polymorphic (Pandey et al., 2012b) and can be applied to different groundnut populations in breeding programs.The clustering of groundnut genotypes in this study was independent of their phenotypic attributes, and thus confirmed the low level of genetic variability in cultivated groundnut (Pandey et al., 2012a;Janila et al., 2013).The relatively low genetic diversity within the genotypes used in this study was likely due to the fact that they were crosses generated from a breeding program.The SSR markers could discern variations and differentiate between the closely related groundnut genotypes, makes this technology a powerful tool for genomic characterisation of groundnut.The relatively diverse genotypes identified in this study are potential candidates for use as parental material in future studies to advance groundnut breeding in Ghana.

Figure 1 .
Figure 1.Genetic diversity among 48 groundnut genotypes generated using the unweighted pair group method with arithmetic mean (UPGMA) procedure based on the Jaccard's similarity coefficient created with NTSYSpc v. 2.1.

Table 1 .
Biological traits of groundnut germplasm used in this study and their sources.

Table 2 .
Description of SSR markers used in this study and their sources.

Table 3 .
(Shirasawa et al., 2013)olymorphism of markers used in this study, calculated using PowerMarker v.3.25.Map position and distances were obtained from the integrated consensus map of cultivated groundnut(Shirasawa et al., 2013).