Plant material and morphological description

Seeds of P. biglobosa were obtained from 16 locations across seven states in Nigeria (Figure 1). Thirteen of the sixteen accessions studied were collected from locations in the Derived Savanna agroecological zone (AEZ); two were collected from the Guinea Savanna and one from Humid Forest (Figure 1,Table 1). The sampling of P. biglobosa accessions was not conducted systematically across all AEZs because the species is reported to be widely adapted and naturally distributed throughout the varied savannas of Sub-Saharan Africa (Lompo et al., 2016). The seed morphological characteristics of each collected accession, including seed shape, seed coat color and seed coat texture were observed. One hundred (100) seed weight in grams of each accession was also recorded.

Seed preparation and total protein extraction

Dried seeds were milled into flour and defatted with n-hexane at a flour:hexane ratio of 1:10 (w/v) prior to protein extraction (Rosa, Gueguen, Paredes-Lopez, & Viroben, 1992).

Seed total protein extraction was carried out by suspending defatted flour samples (100 mg) in 2.0 mL of pre-chilled Tris buffer 50mM Tris-HCl, pH 7.6, 1mM DDT, 150 mM NaCl and 1mM EDTA for 2h at room temperature according to the method of Zarkadas et al. (2007). The resulting homogenates were centrifuged at 10,000 g for 20 min and supernatants were stored at -20ºC until used.

One dimensional SDS-PAGE

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed according to the procedure of Laemmli (1970). Protein extracts were diluted in a sample buffer that contained 60 mM Tris-HCl (pH 6.8), 2% w/v SDS, 3.33% v/v β-mercaptoethanol, 10% glycerol and 0.05% of bromophenol blue. The samples were heated at 98°C for 5 min before loading onto a vertical slab gel (Mini-PROTEAN II Electrophoresis Cell) using a 4% (w/v) stacking and 12% gradient separating (acrylamide/bis acrylamide) gel containing 0.1% SDS. The separating polypeptide bands were calibrated using a protein molecular weight marker which ranged from 11 to 130 kDa. Electrophoresis was done at 125 V for about 2 h. At the end of the run, the gels were stained with Coomassie brilliant blue R-250 (SigmaAldrich) in methanol/water/acetic acid (40:50:10) and destained in a solution containing methanol/water/acetic acid (50:40:10).

Genomic DNA extraction, RAPD amplification and electrophoresis

For the genetic analysis, ten viable seeds of each of the sixteen accessions of P. biglobosa collected from the different locations were grown in a greenhouse. Prior to sowing, the seeds were immersed in concentrated sulphuric acid (H₂SO₄) for 15 minutes to break seed dormancy, rinsed thoroughly in running water and air dried at room temperature. Total genomic DNA was extracted from young leaves of P. biglobosa accessions. About 0.2g of a composite of 5 samples per accession was ground in extraction buffer following the cetyltrimethyl ammonium bromide (CTAB) method of Dellaporta, Wood, and Hicks (1983).

The DNA quality was evaluated using 1% agarose gel electrophoresis with known concentrations of undigested lambda DNA (Sigma, St Louis, MO, USA). Quantification of DNA was done using a spectrophotometer (Beckman Coulter DU530) at 260 nm. Extracts were diluted to obtain DNA concentrations of 25ng/µl. To generate DNA profiles, 40 decamer oligonucleotide DNA primers were initially screened for polymorphisms and only 16 which were polymorphic were used in the PCR reactions (Table 2).

To ensure reproducibility of RAPD markers, RAPD-PCR amplification was performed in triplicate for each primer and accession. The PCR reaction contained the following reagents with the final concentrations: 1x PCR buffer, dNTPs (0.2 mM), MgCl₂ (2.0 mM), primer (25 pmol), template DNA (25ng/µl), Taq polymerase (1.25 U) and nuclease-free water to 50µl. Amplification was carried out in a 2400 Perkin Elmer Gene Amp PCR thermo-cycler as follows: pre-denaturation for 3 min at 94ºC; then 40 cycles each consisting of a denaturation step for 1 min at 94ºC; an annealing step for 1 min at 36ºC; an extension step for 90 sec at 72ºC. Amplification was terminated by a final extension of 7 min at 72ºC. PCR products were electrophoresed on a 1.5% (w/v) agarose gel in 1X Tris/Borate/EDTA (TBE) buffer at 100 volts for 2h and visualized under UV light after staining with ethidium bromide.

Data analysis

Data on 100-seed weight of all accessions were subjected to ANOVA and means were separated using Duncan’s multiple range tests (p≤0.05). Distribution frequencies of the qualitative traits were calculated using Microsoft Excel 2010. The protein bands and RAPD amplified DNA fragments were scored qualitatively, whereby (0) stands for the absence and (1) stands for the presence of bands in the profile of each accession. All values were pooled together to generate a binary data matrix, which was analyzed using the Numerical Taxonomy and Multivariate Analysis System (Rohlf, 2002). Genetic relationship among accessions was evaluated based on Jaccard’s similarity index (Jaccard, 1908). Dendrograms were generated by using the Unweighted Pair Group Method with Arithmetic Mean (UPGMA). Analysis of genetic diversity (GD) was calculated according to the method of Nei (1987). Polymorphic information content (PIC) for each primer was determined from allele frequencies (Nei & Li, 1979).

Variation in seed morphology

Improvement of the quality of indigenous fruit trees across Sub-Saharan Africa involves assessment of fruit/seed traits; this has shown significant variation both among and within provenances and has enhanced selection in breeding programmes (Ræbild et al., 2012). In this study, significant differences (p≤0.05) were observed in the 100-seed weight of P. biglobosa collected from 16 different locations (Table 3), which ranged from 17.93g (NH/2016/P09) to 26.11g (NH/2016/P05). The highest 100-seed weight was recorded for accession NH/2016/P05 collected from Agyaragu, Nassarawa State; this was significantly (p≤0.05) higher than that of all other accessions. The lowest 100-seed weight was recorded for NH/2016/P09 collected from Kwatan-Sule, Benue State. The mean 100-seed weight (21.46g) of all accessions of P. biglobosa collected from the Derived Savanna zone was higher than the mean 100-seed weight collected from the Guinea Savanna and Humid Forest (Figure 2a). This variation in seed weight can be attributed to a wide range of factors such as environment (soil, climate), anthropogenic activities and phenotypic plasticity, thus promoting the presence of ecotypes.

(Leakey, 2017) reported that fruit and/or kernel size was greater in more humid sites for Adansonia digitata and Vitellaria paradoxa; but greater in drier sites for Balanite aegyptiaca. Also, results from tests of B. aegyptiaca and Prosopis Africana indicated that provenances from drier sites had significantly better aboveground growth than provenances from more humid sites (Leakey, 2017). It is recommended that fruit trees germplasm should be collected in drier sites for future plantings in parklands, especially as this germplasm appears to be better adapted to dry conditions (Leakey, 2017).

The seeds of P. biglobosa accessions collected from the different locations in this study had distinct morphology and exhibited variation for seed shape, seed coat colour and seed coat texture (Table 3). Seeds of the sixteen P. biglobosa accessions were predominantly round oval in shape, black in colour with smooth seed coat and were mostly obtained from the Derived Savanna zone (Figure 2B, C, D). Seeds collected from the two sites in Guinea Savanna differed in shape and texture as seeds from one site were flat oval, black and smooth (NH/2016/P03) (Figure 3B); while seeds from the other site were round oval, black and rough (NH/2016/P12) (Figure 3C). Seeds collected from the same agroecological zone (Derived Savanna) exhibited distinct variation in colour and texture, NH/2016/P15 seeds were flat oval, brown and smooth (Figure 3E); while seeds of NH/2016/P04 were flat oval, black and wrinkled.

The morphological variation may be due to gene flow across regions because of seed dispersal since P. biglobosa is an open pollinated tree crop having main pollinators such as bat, honeybees, rodents and humans (Lassen, 2016; Lassen, Ræbild, Hansen, Brødsgaard, & Eriksen, 2012; Lompo et al., 2017). This variation offers great potential for future selection and domestication.

Lack of a clear association between genome size and morphological, geographical, or ecological patterns of differentiation has been reported for two species of Artemisia (Asteraceae) (Dobes et al., 2019; Xaxars et al., 2016).

Seed storage protein analysis

Seed storage protein profiling has been used as an effective tool for varietal identification and determination of phylogenetic relationship in several plant species populations (Emre, 2009; Hameed, Qureshi, Nawaz, & Iqbal, 2012; Machuka, 2001).

In this study, the SDS-PAGE profile of sixteen accessions of P. biglobosa showed similar electrophoretic patterns with respect to the number of protein bands and their band intensities (Figure 4). The uniformity in the protein band profiles of all sixteen accessions suggests a low level of genetic diversity, which also corroborates earlier studies on P. biglobosa using seed protein electrophoresis (Adesoye et al., 2015).

Occurrence of a narrow genetic base from the accessions of P. biglobosa would suggest that they resulted not only from a common gene pool, but also from the outcrossing nature of the species over a long time, leading to a low level of inter-population diversity. Low variation was observed in thirteen mungbean varieties using SDS-PAGE (Hameed et al., 2012) and also in genetic diversity assessment of groundnut using SDS-PAGE (Javaid et al., 2004).

The number of clearly visible protein bands ranged from eight to nine among the sixteen accessions of P. biglobosa with molecular weights ranging from 11 to 130 kDa. Two major clusters of bands were observed between 11−17 kDa and 34−53 kDa and one minor band was observed above 130 kDa (Figure 4). This is similar to results ofAdesoye et al. (2015) on albumin and globulin fractions of P. biglobosa in which two major bands were observed at the 16 and 50 kDa regions.

It has been suggested that high molecular weight proteins facilitate the development of seed hardness (Coelho, Bellato, Santos, Ortega, & Tsai, 2007); therefore, the presence of a high molecular weight band observed above the 130 kDa in this study may account for the seed hardness of P. biglobosa. The seeds of P. biglobosa are very hard and require treatment with sulphuric acid to induce germination.

A polymorphic polypeptide with molecular weight of approximately 24 kDa was absent in NH/2016/P04 but present in all other accessions (Figure 4A) and can be used for its identification. Differences in the presence or absence of bands among accessions are indicative of differences in the genes controlling the different polypeptide subunits (Osanyinpeju & Odeigah, 1998). Also, the presence or absence of polypeptide bands has been found to be linked to the expression of certain characteristics such as wrinkled seeds and insect resistance (Odeigah & Osanyinpeju, 1996; Rao & Pernolett, 1981). The seed shape of NH/2016/P04 was observed to be wrinkled and the absence of the 24 kDa band in this accession may be responsible for this seed texture as it was the only accession expressing this characteristic. Seed storage proteins are non-enzymatic and have the sole purpose of providing proteins (nitrogen and sulphur source) required during germination and establishment of a new plant (Hameed et al., 2012). Legume seed storage proteins are mostly 7S and 11S globulins, which tend to be deficient in sulphur-containing amino acids (Tang & Sun, 2010). In mung bean, SDS-PAGE revealed that 11S globulin was composed of two bands: 40 kDa and 24 kDa, 8S vicilin was composed of 60 kDa, 48 kDa, 32 kDa and 26 kDa bands; and basic 7S globulin was composed of 28 kDa, 17 kDa and 16 kDa bands (Hameed et al., 2012; Mendoza, Adachi, Bernardo, & Utsumi, 2001).

In this study, peptides with molecular weights of 40 kDa and 24 kDa were detected that may correspond to 11S globulin, peptides with molecular weights of 48 kDa and 32 kDa were detected that may correspond to the 8S vicilin subunit, while 28 kDa and 16 kDa peptides were also detected that may correspond to the 7S globulin subunit.Barakat (2004) reported low levels of protein polymorphism among six cultivars of soybean in which five major bands were identified at the 72, 36, 32, 20 and 16 kDa regions.

In African yam bean, SDS-PAGE separated identical globulin, albumin and vicilin patterns for all 26 accessions studied (Machuka, 2001).Strelec, Has-Schon, and Vitale (2012) reported that barley varieties could be partially discriminated by albumin/globulin banding patterns using SDS-PAGE, whereas this was not possible with native PAGE which showed more or less identical protein patterns for all varieties.

Through the use of SDS-PAGE on total proteins it is possible to detect a useful band polymorphism to explore the diversity of P. biglobosa as was detected in this study, albeit at a low level.

Cluster analysis

A dendrogram was constructed based on protein banding patterns using Jaccard’s similarity coefficients (Table 4). The dendrogram grouped the sixteen accessions into four clusters at 0.93 similarity coefficient (Figure 5). The first cluster grouped together the largest number (thirteen) of accessions which consisted of NH/2016/P01, NH/2016/P02, NH/2016/P05, NH/2016/P16, NH/2016/P15, NH/2016/P06, NH/2016/P13, NH/2016/P12, NH/2016/P11, NH/2016/P10, NH/2016/P09, NH/2016/P08 and NH/2016/P07. The majority of the members of this cluster had similarity coefficient of 1.00 suggesting them to be genetically identical. Cluster II consisted of NH/2016/P14. Accession NH/2016/P03, which was collected in Minna, Niger state and belonging to the Guinea Savanna agroecological zone, was separated alone into Cluster III, displaying pattern of geographic origin. Cluster IV consisted of NH/2016/P04 and was the most divergent of the sixteen tested accessions. The genetic similarity coefficient was 0.85 among NH/2016/P03, NH/2016/P04 and NH/2016/P14. Clustering showed pattern of geographic origin as eleven of the thirteen accessions grouped in Cluster I were collected from the Derived Savanna agro-ecological zone.

Genetic distance values give some idea of the level of genetic variability among selected species or accessions. The variation in genetic diversity among the sixteen P. biglobosa accessions ranged from 0.85 to 0.92 depicting a low genetic diversity (Table 4). All accessions in Cluster I had a similarity coefficient of 0.92 in any pairwise comparisons with NH/2016/P03, NH/2016/P04 and NH/2016/P14. The low levels of variability observed in P. biglobosa is congruent withSina (2006) who reported that the genetic distances among P. biglobosa trees in Burkina Faso were low (between 0 and 0.240), indicating that the populations were similar enough to belong to the same genetic group. Occurrence of a narrow genetic base from the accessions of P. biglobosa suggests that they may have resulted from a common gene pool and that fixation of variants through genetic drift has probably occurred.

Amusa et al. (2014) had earlier reported that close natural populations like P. biglobosa exhibited high genetic similarity and low genetic distance attributed to a high rate of exchange of gene flow between populations. Low genetic diversity was also observed in mung bean germplasm based on electrophoresis of seed storage proteins where a dendrogram analysis grouped the tested genotypes into three clusters at 93% homology (Hameed et al., 2012). Results of this study is similar to that obtained in cowpea in which seven landraces were separated into different clusters based on their geographical origin (Alghamdi et al., 2019). The high level of similarity in the accessions of P. biglobosa suggests that the seed storage proteins are highly conserved.

RAPD polymorphism

Genetic polymorphism is an indication of evolutionary adaptation, which has a main role in the survival of species in a changing environment (Stevens, Salomon, & Sun, 2007). In this study, Random Amplified Polymorphic DNA (RAPD) amplification profiles revealed a total of 256 alleles out of which 163 were polymorphic showing 63.67% polymorphism among the studied accessions. This relatively high polymorphism suggests a significant amount of genetic diversity among the accessions. A similar result was obtained among 11 soybean cultivars in which RAPD amplification profiles revealed 61.7% polymorphism using five primers (El-Kholy, 2013).

Alleles, ranging in size from 250 to 1500 basepairs (bp) and above, were scored for estimation of genetic relationship among the sixteen P. biglobosa accessions (Figure 6). An average of 10 alleles per primer was obtained in this study ranging from a minimum of four alleles using primer OPT05 to a maximum of 15 alleles using primer OPB10 (Table 2).

Genetic diversity is one of the important indices for the evaluation of genetic diversity among crop species (Narzary, Mahar, Rana, & Ranade, 2009; Tamboli et al., 2016). The genetic diversity in this study ranged from 0.41 to 0.93, suggesting a significant amount of variation among the 16 accessions evaluated (Table 2). An earlier report on genetic diversity assessment of 23 accessions of P. biglobosa from different agro-ecological zones using RAPD indicated weak genetic diversity (expected heterozygosity, H_E= 0.05–0.18 and observed number of alleles, ONA = 1.11–1.65) (Amusa et al., 2014). The significance of the polymorphic information content (PIC) value is such that they are used to evaluate the amount of genetic diversity and are categorized as high (PIC > 0.05), medium (PIC < 0.05) and low (PIC < 0.25) (Abedian, Talebi, Golmohammdi, & Sayed-Tabatabaei, 2012). Polymorphic information content ranged from 0.39 for primer OPT05 to 0.93 for primer OPB10 (Table 2).

Fifteen of the sixteen polymorphic primers had PIC values greater than 0.5 indicating that RAPD can develop high-locus polymorphism, which is useful to assess genetic variability of the accessions. All primers showed an average PIC value of 0.79 and nine out of the sixteen primers exceeded the average. This relatively high polymorphism has been observed in studies of genetic diversity on soybean cultivars byChowdhury et al. (2001) and Barakat (2004).

Cluster analysis based on UPGMA revealed six distinct clusters at a genetic similarity coefficient of 0.68 on the UPGMA dendrogram (Figure 7). Cluster I consisted of accessions NH/2016/P01, NH/2016/P05, NH/2016/P12, NH/2016/P13, NH/2016/P04, NH/2016/P02, NH/2016/P09, NH/2016/P14 and NH/2016/P15. All the accessions grouped together in Cluster 1 were collected from Derived Savanna except accession NH/2016/P12 which was collected from Guinea Savanna. This suggests that gene flow between the different locations may have occurred as a result of movement of germplasm and exchange of genetic material among farmers and markets. Cluster II consisted of NH/2016/P06; Cluster III consisted of NH/2016/P07 while Cluster IV consisted of NH/2016/P03. Accessions NH/2016/P08, NH/2016/P16 and NH/2016/P11 grouping together in Cluster V were collected from Derived Savanna. Accession NH/2016/P10 collected from Ogoja, Cross River State in the Humid Forest zone grouped alone in Cluster VI. Thus, in the UPGMA analysis of the RAPD profiles, accessions from same agroecological zone grouped together, while in other cases they were placed in different clusters. For most accessions in this study there was similar clustering pattern of geographically closer accessions, indicating significant association between genetic similarity and geographical distance. This does not agree with a report of Adesoye et al. (2013) who reported that distribution of P. biglobosa genotypes among clusters did not correspond with their geographical patterns. The lowest similarity was recorded between NH/2016/P10 and NH/2016/P06, NH/2016/P07, NH/2016/P03. The highest similarity was recorded between accessions NH/2016/P12 and NH/2016/P13, followed by accessions NH/2016/P05 and both NH/2016/P12, NH/2016/P13. Accessions NH/2016/P05 and NH/2016/P13 with narrow genetic distance were collected from the same agro-ecological zone (Derived Savanna). This result corroborates findings of Hamrick and Godt (1989) in which species with small geographic ranges maintained less genetic diversity than geographically widespread species. RAPD clearly distinguished some accessions by accession specific fragments (Figure 6). Therefore, RAPD markers can successfully be used to produce variety specific fingerprints in P. biglobosa accessions and are a valuable tool for assessing genetic diversity.

Implications of findings to the conservation and sustainable use of Parkia biglobosa genetic resources

• The variation observed in qualitative traits and 100-seed weight among the sixteen accessions of P. biglobosa across the three agro-ecological zones of Nigeria indicates their potential for domestication through selection towards conservation and breeding.

• Cluster analysis of the seed protein profile identified NH/2016/P04 as genetically distinct as it was clustered alone and was the most divergent of the sixteen tested accessions. The absence of the 24kDa band in accession NH/2016/P04 distinguishes it from all the other accessions. It can be adopted as parental line for heterosis crossing and should be exploited for conservation, breeding and sustainable utilization.

• Four selected RAPD primers (OPB10, OPT07, OPB04 and OPT12) can be used to discriminate the accessions of P. biglobosa.

• The highly diversified accessions of P. biglobosa indicate potential for domestication through selection in breeding programmes.

• Most accessions from the same agroecological zone clustered together suggesting that there is a strong correlation between genetic similarity and geographic proximity.

• The homogeneity of alleles among the studied P. biglobosa accessions suggests possible loss of intraspecific genetic diversity. The weakening gene pool and diversity observed can be enhanced through more germplasm collections, particularly from the diverse agroecological zones of Nigeria for better genetic characterization using more specific markers towards conservation, breeding and sustainable utilization.