Genetic Restoration of Black Rhino In South Africa: Conservation Implications

Globally, wildlife populations are becoming increasingly small and isolated. Both processes contribute to an elevated risk of extinction, notably due to genetic factors related to inbreeding depression and a loss of adaptive potential. Wildlife translocation is a valuable conservation tool to reintroduce species to previously occupied areas, or augment existing populations with genetically divergent animals, thereby improving the viability of endangered populations. However, understanding the genetic implications of mixing gene pools is key to avoid the risk of outbreeding depression, and to maximise translocation effectiveness. In this study we used mitochondrial and microsatellite DNA collected from 110 black rhinos (Diceros bicornis minor) in Kruger National Park, South Africa, to determine levels of genetic diversity, inbreeding and relatedness. We compared this diversity with the two source populations (KwaZulu-Natal, South Africa and Zambezi River, Zimbabwe) using data from previously published studies, and assessed changes in the relative contribution of source lineages since their reintroduction in the 1970s. Our results show that Kruger’s black rhinos are genetically more diverse than those from KwaZulu-Natal, with levels closer to those from the Zambezi Valley. Furthermore, our ndings indicate a relative increase in the Zimbabwean lineage since reintroduction, suggesting a possible selective advantage. From a conservation perspective, our results demonstrate the benets of mixing multiple source populations to restore gene ow, improve genetic diversity and thereby help protect small, isolated populations from extinction.


Introduction
Compared to their historical counterparts, many wildlife populations remaining today are small and isolated with limited gene ow. These populations may exhibit an increased population differentiation and reduced genetic diversity as a function of genetic drift and/or inbreeding (Lacy 1987;Primack 2002). Negative genetic impacts are intensi ed in fenced populations, as mechanisms that evolved to reduce inbreeding, such as dispersal, are inhibited. Thus, a key challenge in wildlife conservation is to prevent or minimise the loss of genetic diversity to enable the long-term persistence of threatened species (Weeks et Table S1).

iii. Relatedness and inbreeding
Average pairwise relatedness (r) for the population was calculated using the package 'related' (Pew et al. 2015) in R version 3.6.1 (R Core Team 2019). The estimator of relatedness chosen for this analysis was based on a simulation analysis comparing different estimators using the 'compareestimators' function. Given similar performance, the Wang (2002) estimator (having the highest correlation between observed and expected values) was chosen for further analysis. Pairwise relatedness was calculated using the 'coancestry' function. Wright's inbreeding coe cient (F IS ) was calculated using GENETIX 4.05.2 (Belkhir et al. 2004). Con dence intervals for inbreeding coe cient values for each locus and over all loci in each population were obtained by bootstrapping 1000 times.

Source and founder population comparison
The combined mtDNA sequence dataset was used to compare the nucleotide (π) and haplotype (h) diversity between source (Zambezi River and KwaZulu-Natal) and founder (Kruger) populations. The genetic structure among source and founder haplotypes in the combined dataset was visualised by constructing a median-joining network (Bandelt et al. 1999) in PopART (Leigh and Bryant 2015). Finally, the relative maternal lineage contributions of the source populations were determined and compared to the initial ratio of Zambezi River and KwaZulu-Natal founder females obtained from historic records (SANParks, unpublished data).

Quality control
No evidence of genotyping errors or allelic dropout was found. Signatures of null alleles were detected at seven loci, namely SR74, IR12, SRS262, 7C, BlRh1B, DB44 and DB66; loci with null allele frequencies greater than 0.08 (SR74, IR12, SRS262, 7C and BlRh1B) were removed from subsequent analysis. ZF1  Table S3). Loci IR12 and SR74 were removed as they appeared to be sexlinked. Locus 12F (originally isolated from white rhino) and locus BlRh37D were removed from further analysis. Loci 7B and 32A, both originally isolated from white rhino, were monomorphic in this study and were also removed. Finally, individuals with more than 30% missing data (n=1) were removed from further analysis.

Diversity, relatedness and inbreeding
The nal mtDNA dataset contained 103 sequences of 469 bp in length. Four haplotypes were identi ed. These haplotypes were characterised by ve polymorphic sites, all containing transition nucleotide substitutions (G↔A and/or C↔T). Haplotype diversity (h) and nucleotide diversity (π) were 0.48 (± 0.05 SD) and 0.29 (± 0.20 SD), respectively. The nal microsatellite data set comprised 109 animals. All 13 microsatellite loci retained for analyses were polymorphic, with two to 14 alleles each ( The combined Kruger, Zambezi River and KwaZulu-Natal mtDNA control region dataset comprised 296 sequences of 363 bp after alignment. A total of seven mtDNA haplotypes were found across the three populations, containing seven polymorphic sites ( Table 2). The nucleotide and haplotype diversities calculated in the Kruger population was approximately mid-way between those of the two source populations ( Table 2). The relationship between haplotypes can be seen in Fig. 2. The Kruger population shared two haplotypes (H1 and H5) with the Zambezi River population; these two haplotypes together represented 28.15% of the Kruger black rhino in this study. The extant KwaZulu-Natal population is represented by only a single haplotype (H2), and this haplotype was shared with 67.96% of the Kruger population. Haplotype 3 (H3) was unique to Kruger, two mutational steps from both H2 and H5 (Fig. 2). Haplotypes 4, 6 & 7 were unique to the Zambezi River population. The current haplotype distribution among the three populations is illustrated in Fig. 3. Table 2 (Fig. 4).

Discussion
The Kruger black rhino population offered a unique opportunity to evaluate the outcome of mixing the two remnant south-central black rhino source populations. Using mtDNA and microsatellite markers, we found that mixing gene pools substantially enhanced both the mitochondrial and nuclear diversity of the founded Kruger population relative to the source D. b. minor population in South Africa. Together with the low levels of relatedness and no evidence of non-random mating, our results con rm that the Kruger black rhino population is a diverse, outbred, panmictic population. This study provides a baseline for informing black rhino metapopulation management strategies and indicates that Kruger black rhinos would be ideal candidates for translocation and reintroduction efforts aimed at improving diversity in other D. b. minor black rhino populations.
Maintaining adequate levels of genetic diversity is essential for ensuring both the short-term health and long-term survival of isolated populations of endangered species. Small population numbers, genetic drift and/or inbreeding may all contribute to a substantial loss in genetic diversity of such populations, and consequently may negatively impact their viability. For example, the small, reintroduced population of black rhinos in Addo Elephant National Park, South Africa has comparatively low genetic diversity and high relatedness relative to their source populations, resulting in low tness, manifesting as low population growth rate and reduced male survival ( We also found two Zambezi River haplotypes (H1 and H5) within the Kruger population; the remaining Kruger haplotype (H3) was reported in a single captive Zimbabwean black rhino (Fernando et al. 2006). Thus at least two (or three, if including H3) of the six known Zimbabwean haplotypes (33-50%) have been retained in the Kruger black rhino population. It is also possible that with more extensive sampling, additional Zimbabwean haplotypes would be detected. Future studies that directly compare the nuclear contributions of the two source populations within the current Kruger population would also provide further insight into this genetic admixture.
While mixing individuals from different source populations may increase genetic diversity and reduce the likelihood of inbreeding depression, it may increase the risk of outbreeding depression (Edmands 2007). For example, if source populations are under unique environmental pressures (e.g., different climates or habitats), local adaptations may arise, especially in long-isolated populations. Thus, outbreeding with genetically diverse individuals is counterproductive if the hybrid offspring face lowered tness due to the loss of locally adapted genetic variants (Edmands 1999). A classic case of outbreeding depression occurred when two subspecies populations of Alpine ibex (Capra ibex) were translocated from the Sinai Peninsula and Turkey into the European Alps. Unfortunately, the introduced Ibex bred earlier in the season than their European counterparts, resulting in hybrid offspring born in midwinter, reducing survival and ultimately leading to the hybridised herd's extinction (Templeton 1986). However, this is example is an exceptional case, and outbreeding depression is seldom seen in practice (Frankham et al. 2011;Ralls et al. 2018).
Outbreeding depression from mixing KwaZulu-Natal and Zambezi River black rhinos is unlikely when considering primary risk factors, such as chromosomal differences, lack of gene ow for more than 500 years, and substantial environmental differences between populations (Frankham et al. 2011). The Zambezi River and KwaZulu-Natal black rhino populations were historically connected (Kotzé et al. 2014) and a healthy population of translocated KwaZulu-Natal black rhino in Malilangwe, Zimbabwe suggests that the different environment between populations is unlikely to contribute to outbreeding depression. Furthermore, the increase in Zimbabwean lineage proportion seen in the extant Kruger population (relative to the ratio of founder females) contradicts any potential loss of local adaptation; if anything, selection over the generations may have favoured the more diverse Zambezi River black rhino. Further research, however, is required to test whether a selective advantage or stochastic events are responsible for the lineage proportion increase seen in this study.
In conclusion, this study indicates that the admixture of black rhinos from different gene pools substantially enhanced both the nuclear and mitochondrial diversity of the founded Kruger population relative to the source D. b. minor population in South Africa. In the absence of threat alleviation, metapopulation management strategies (such as population supplementation through translocation) aimed at increasing the range and securing the genetic health of black rhino are critical. The improved genetic diversity found in the Kruger population is encouraging for the long-term survival of this subspecies as a managed metapopulation within South Africa, possibly improving its adaptive potential to respond to environmental change. Given the encouraging levels of diversity observed, this also makes the Kruger black rhino population an ideal source candidate for founding new populations or improving the genetic variation (and thus reducing extinction risk) for genetically depauperate D. b. minor populations in South Africa.