Aim The formation history of Africa’s current river basins remains largely unknown. In order to date changes in landscape and climate, we studied the biogeography of the African freshwater fish with the largest natural distribution. We also validated biogeographic units.

Location Continental Africa.

Taxon Clarias gariepinus sl.

Methods We investigated mitochondrial cytb sequences of 443 individuals from 97 localities, using a haplotype network and a genetic landscape analysis. We inferred a dated phylogeny using maximum likelihood and Bayesian inference approaches and reconstructed ancestral areas with S-DEC and S-DIVA models. Microsatellite genotyping complemented the mitochondrial approach in the Congo basin, where the latter revealed complex patterns.

Results Limited differentiation is found in northern and south-western Africa, and sharp genetic differentiation in the continent’s East and Centre. Populations with affinities to neighbouring basins occur at the edges of the Congo province. High diversity exists in the South of the Congo basin. The Zambezi province is partitioned into eastern, central and western sectors. In the East, specimens were related to those from the Congo. In the West, they were similar to Southern representatives. Phylogenetic inference placed the origin of C. gariepinus in the East Coast, with intraspecific diversification starting around the Great Lakes. These events occurred ca. 4.8-1.65 and 2.3-0.8 MYA, respectively.

Main conclusions Clades of C. gariepinus sl. show a clear geographic signature. The origin of C. gariepinus in the East Coast and diversification around the Great Lakes coincided with periods of increased aridity. Low genetic differentiation in northern and southern Africa may result from connectivity during recent periods of higher rainfall. In contrast to other widespread African freshwater fishes, colonisation rather than extinction seemed to mediate distribution patterns. This can be explained by a high ecological tolerance. We highlight the species’ suitability to study landscape and climate evolution at various scales.

We genotyped 280 adult C. gariepinus from eight sites at seven loci, Cga01, Cga02, Cga05, Cga09, Cga11 (Galbusera, Volckaert, Hellemans, & Ollevier, 1996), Cba 19 and Cba 20 (Yue, Kovacs, & Orban, 2003). We amplified all loci using the QIAGEN Multiplex PCR kit (QIAGEN, Venlo, the Netherlands) in two multiplex and one singleplex reaction.

Multiplex 1: Cba 19 and Cba 20; Annealing temp 58-52 °C (TD) and 52°C ; Number of cycles: 7 (TD) and 28; Primer concentrations: 0.05 µm (both)

Multiplex 2: Cga 01, Cga 05, Cga 09 and Cga 11 ; Annealing temp 54°C ; Number of cycles: 25; Primer concentrations: 0.1 µm (Cga 01), 0.05 µm (others)

Singleplex: Cga 02; Annealing temp 54°C ; Number of cycles: 25; Primer concentration: 0.2 µm

We ran multiplex PCR products with an internal size standard (500LIZ, Applied Biosystems), on an ABI 3130 automated capillary DNA sequencer (Applied Biosystems, Foster City, USA). We conducted fragment analysis in GENEMAPPER v4.0 (Applied Biosystems) and checked for the potential occurrence of null alleles and scoring errors, with MICRO-CHECKER v2.2.3 and DROPOUT v2.0. Scoring errors were detected at loci Cga09 and Cba19, which were excluded from further analyses. We ran 10-30% of the samples twice for all markers to verify reproducibility.

codes:

KIN; Kinshasa; Lower Congo; -04°18’N; 15°20’E

KIS; Kisangani; Middle Congo; 00°31’N; 25°11’E

DBL; Bukama; Upper Congo; -09°11’N; 25°51’E

KAD; Dianda; Kando, Lulua; -10°45’N; 25°50’E

LUF; Kapolowe; Upper Lufira; -11°21’N; 26°46’E

KYU; Kyubo; Middle Lufira; -09°31’N; 27°02’

KAF; Lubumbasi; Kafubu, Luapula; -11°42’N; 27°29’E

LUT; Mwaba; Lutshipuka, Luapula; -10°14’N; 28°20’E