Research Article |
Corresponding author: Martin H. Villet ( martin.villet@gmail.com ) Academic editor: Ben Price
© 2020 Chantal L. Taylor, Nigel P. Barker, Helen M. Barber-James, Martin H. Villet, Lyndall L. Pereira-da-Conceicoa.
This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Citation:
Taylor CL, Barker NP, Barber-James HM, Villet MH, Pereira-da-Conceicoa LL (2020) Habitat requirements affect genetic variation in three species of mayfly (Ephemeroptera, Baetidae) from South Africa. ZooKeys 936: 1-24. https://doi.org/10.3897/zookeys.936.38587
|
This study investigates genetic diversity in three species of Ephemeroptera, one eurytopic and therefore widespread (Afroptilum sudafricanum) and two stenotopic and thus endemic (Demoreptus natalensis and Demoreptus capensis) species, all of which co-occur in the southern Great Escarpment, South Africa. Mitochondrial DNA was analysed to compare the genetic diversity between the habitat generalist and the two habitat specialists. Afroptilum sudafricanum showed no indication of population genetic structure due to geographic location, while both Demoreptus species revealed clear genetic differentiation between geographic localities and catchments, evident from phylogenetic analyses and high FST values from AMOVA. In addition, the phylogenetic analyses indicate some deeper haplotype divergences within A. sudafricanum and Demoreptus that merit taxonomic attention. These results give important insight into evolutionary processes occurring through habitat specialisation and population isolation. Further research and sampling across a wider geographic setting that includes both major mountain blocks of the Escarpment and lowland non-Escarpment sites will allow for refined understanding of biodiversity and associated habitat preferences, and illuminate comparative inferences into gene flow and cryptic speciation.
cytochrome oxidase 1, genetic diversity, habitat specialisation, haplotype, phylogeography, mayfly
Greater genetic diversity within a lineage is regarded as increasing its resilience to environmental change (
Aquatic insects have a winged adult stage that is generally considered to have relatively strong dispersal ability (
Genetic variation between populations is related to the ability of their members to disperse, and a high degree of genetic structure has been observed among populations of some South African winged aquatic (
The aim of this study is to use three model species of mayfly to test the hypothesis that habitat-restricted taxa have greater phylogeographical structure than habitat-generalist species. Afroptilum sudafricanum Lestage is a common, widespread African species, occurring in a range of ecological conditions, including different flow regimes and a wide altitude range (
The southern Great Escarpment forms an 800-km-long stretch of mountain complexes extending from the Nuweveldberge in the west to the Eastern Cape Drakensberg in the east. Ancient erosional features divide the mountains into five main blocks that range in altitude from 1 600–3 000 m a.s.l., making the area interesting for study of dispersal-limited groups.
Nymphs of A. sudafricanum, D. capensis, and D. natalensis were collected from 21 rivers in the Eastern Cape Great Escarpment, relating to 12 study areas within the Escarpment and non-Escarpment sites (Table
Collecting localities (Site and river name) and non-zero sample sizes for each species from each site. The GenBank sequence accession numbers for each sample are listed in Suppl. material
Locality | Longitude/Latitude | A. sudafricanum | D. capensis | D. natalensis | Demoreptus sp. |
---|---|---|---|---|---|
Escarpment sites | |||||
Eastern Cape Drakensberg | |||||
Barkley East 1: Diepspruit | -30.751, 27.546 | 3 | 1 | ||
Barkley East 2: Diepspruit | -30.757, 27.552 | 3 | |||
Barkley East 3: Diepspruit | -30.718, 27.54 | 3 | 1 | ||
Barkley Pass 1: Marais Hoek | -31.215, 27.686 | 3 | |||
Barkley Pass 4: Ben Wyvie | -31.173, 27.971 | 3 | 3 | ||
Barkley Pass 5: Lymore Lodge | -31.172, 27.854 | 2 | |||
Rhodes 1: Hawkshead | -30.676, 27.884 | 3 | 2 | ||
Rhodes 2: Tiffindell | -30.674, 27.904 | 3 | 1 | ||
Rhodes 3: Tenahead | -30.696, 28.150 | 3 | 1 | ||
Maclear 1: Vuvu River | -30.603, 28.216 | 5 | |||
Stomberg | |||||
Stomberg 3: Lana River | -31.163, 26.602 | 3 | |||
Stomberg 4: Lemonfountain | -31.416, 26.842 | 3 | |||
Winterberg-Amatole | |||||
Elansberg 1: Elandsberg | -32.506, 26.903 | 3 | |||
Winterberg 1: Fanella falls | -32.363, 26.385 | 2 | 3 | ||
Winterberg 2: Fanella falls | -32.380, 22.967 | 3 | |||
Winterberg 3: | – | 5 | |||
Sneeuberg | |||||
Sneeuberg 1: Fish River | -32.227, 24.954 | 2 | |||
Sneeuberg 2: Melkriver | -32.243, 24.941 | 2 | 3 | 3 | |
Kamdeboorberg 1: Buffelsrivier | -32.177, 24.016 | 3 | 2 | ||
Kamdeboorberg 3: Waterkloof | -32.353, 23.890 | 2 | 2 | ||
Nuweveldberge | |||||
Nuweveldberge 1: Maijiesvlei | -32.102, 22.636 | 1 | |||
Non-Escarpment sites | |||||
Grahamstown | |||||
Grahamstown CR: Coleridge River | -33.349, 26.618 | 2 | |||
Grahamstown KP: Kap River | -33.351, 26.858 | 5 | |||
Grahamstown KR: Kowie River | -33.349, 26.560 | 5 | |||
Grahamstown PM: Palmiet River | -33.370, 26.476 | 5 | |||
KwaZulu-Natal | |||||
KwaZulu-Natal KK: Karkloof River | -29.338, 30.307 | 5 | |||
KwaZulu-Natal LR: Lions River | -29.492, 30.108 | 5 | |||
KwaZulu-Natal UM: Umgeni River | -29.477, 30.261 | 1 | |||
86 | 11 | 12 | 1 |
A related species of Baetidae, Baetis rhodani Pictet, was used as the outgroup for phylogenetic analyses, and relevant sequence data (
DNA was extracted using the Invisorb Spin Tissue Mini Kit following manufacturer’s protocol (Invitek, Berlin, Germany). Extraction was non-destructive, using internal body digestion, which ensured the preservation of the exoskeleton for future morphological analysis (housed in the Albany Museum, Makhanda, South Africa, along with additional material that is stored in the collection, listed under the GEN catalogue.)
Two mitochondrial gene regions were amplified: cytochrome c oxidase subunit I (COI) and small subunit ribosomal RNA (16S). A 528-bp section of the COI regions of D. natalensis and D. capensis was successfully amplified with the standard ‘universal’ primer pair, LCO1490 and HCO2198 (
The polymerase chain reaction (PCR) was performed in a 50 μl volume using the following thermal regime: 95 °C for 5 min, 35 cycles of 95 °C for 45 s, 50 °C for 45 s, and 72 °C for 90 s, followed by a final extension period of 72 °C for 5 min. PCR amplifications were checked for the presence of amplified PCR products by gel electrophoresis (0.5% agarose gel stained with SYBR green) and viewed with a UV-transilluminator. Successful PCR products were cleaned up using the Invisorb PCRapace® Quick purification kit (Invitek, Berlin, Germany) and cycle-sequenced in both directions using the primers used for amplification, the ABI Big Dye Sequencing kit v.3.1 (following manufacturer’s instructions (Applied Biosystems)), and a ABI Genetic Analyzer 3500 (Applied Biosystems).
Sequence trace files were assembled and edited using Sequencher 3.0 (DNA sequence analysis software, Gene Code Corporations, Ann Arbor, MI USA, http://www.genecodes.com). The sequences were then aligned in MEGA v.6 (
Each gene was tested for substitution saturation using plots of transitions and transversions against F84 distance in DAMBE v7.0.58 (
Congruence between the COI and 16S datasets was assessed using the partition homogeneity test (PHT) in PAUP* (
Bayesian Inference (BI) analyses were conducted with MrBayes v.3.1.2 (
Maximum likelihood (ML) analyses were conducted with 2 000 bootstrap replicates using the GARLI (Genetic Algorithm for Rapid Likelihood Inference) on XSEDE via the CIPRES (Cyberinfrustructure for Phylogenetic Research) Science Gateway v3.3 (
Parsimony analyses were performed in PAUP* version 4.0b10 (
Data characteristics and summary of the parsimony analysis. The number of specimens with sequence data (ntax), total number of base pairs (bp), parsimony informative (# Pi), and percent parsimony informative (% Pi) is reported. The results of the parsimony are summarised with the number of trees retained (# trees), tree length (score), Consistence Index (CI) and Retention Index (RI). The summary of the models for the Maximum Likelihood analysis (ML) selected by jModeltest.
Species | Dataset | ntax | Characters | Parsimony analysis | Model | |||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
bp | #Var | # Pi | % Pi | # trees | Score | CI | RI | ML analysis | BI analysis | |||
A. sudafricanum | COI | 88 | 649 | 217 | 192 | 29.6 | 5 000 | 421 | 0.601 | 0.932 | GTR+I+G | GTR+I+G |
COI+16S | 88 | 1191 | 380 | 336 | 28.2 | 5 000 | 645 | 0.662 | 0.939 | TIM3+I+G | GTR+I+G | |
Demoreptus spp. | COI | 24 | 528 | 164 | 159 | 30.1 | 8 | 302 | 0.745 | 0.922 | TVM+G | GTR+I+G |
Molecular diversity was investigated using the COI datasets. The number of variable sites (S), number of haplotypes (Hap) and haplotype diversity (Hd), Nucleotide diversity (p) and neutrality tests (Tajima’s D and Fu’s FS) were calculated in DNAsp (
COI sequences (649 bp) were obtained from 86 individuals and 16S sequences (542 bp) obtained from 59 individuals of A. sudafricanum. Shorter (528 bp) COI sequences were obtained for 24 Demoreptus individuals (for D. natalensis, N = 12; D. capensis, N = 11; unidentified, N = 1). DNA characteristics for each gene dataset are summarised in Table
The parsimony analyses’ results are summarised in Table
Bayesian inference phylograms of A. sudafricanum for gene markers COI (left) and COI + 16S (right). Support for major nodes shown in the order Bayesian Inference / Maximum Parsimony / Maximum Likelihood (BI/ML/MP). Bars next to clades refer to distinct clades that are colour-coded according to the study areas found within that clade (see colour legend), except for the widespread grade which is designated by a solid black line. Branches bearing outgroups have been omitted to save space and their position is depicted by a dashed line.
The phylogenetic analysis of the habitat-restricted D. capensis and D. natalensis clearly indicated strong genetic structure corresponding to geographic location (Fig.
Bayesian inference phylogram of Demoreptus spp for the COI gene marker. Support for major nodes is shown in the order Bayesian Inference / Maximum Parsimony / Maximum Likelihood (BI/ML/MP). Bars next to clades refer to distinct clades that are colour-coded according to the study areas found within that clade (see colour legend). Baetis rhodani Pictet was used as the outgroup.
MJN analysis collapsed the 86 A. sudafricanum COI sequences into 60 haplotypes (Fig.
Haplotype characteristics and Neutrality tests for A. sudafricanum, D. capensis and D. natalensis.
Species | Haplotype characteristic | ||
---|---|---|---|
Number of haplotypes (Hap) | Nucleotide diversity (Pi) | Number of variable sites (S) | |
A. sudafricanum | 60 | 0.07508 | 129 |
A. sudafricanum (unresolved) | 28 | 0.01998 | 67 |
D. capensis | 8 | 0.08592 | 101 |
D. natalensis | 6 | 0.01881 | 21 |
Median-joining network of A. sudafricanum based on COI haplotypes generated in this study. The network was estimated using the median-joining algorithm in PoPArt v.1.7 with epsilon = 0. Each circle represents a different haplotype and the size of a circle correlates with the number of individuals assigned to that haplotype. Only haplotypes found in more than one sample are numbered. Colours indicate the geographic origin of sequences; black dots indicate unsampled or extinct haplotypes.
The MJN analyses for D. capensis retrieved eight haplotypes (Hd = 0.9273, S = 101), six of which were singletons and D. natalensis retrieved six haplotypes (Hd = 0.8636, S = 21) including three singletons (Fig.
Distribution of A. sudafricanum, D. natalensis and D. capensis COI haplotypes across the study area. The map shows the study areas defined in Table
Nucleotide diversities (Pi) are reported in Table
The AMOVA results for A. sudafricanum revealed that over all localities, 52.33% of the total variance was explained by variation among populations (df = 10, Va = 12.073) while 47.67% (df = 75, Vb = 10.998) was explained by variation within populations (Table
Median-joining networks of D. natalensis and D. capensis based on COI haplotypes generated in this study. The network was estimated using the median-joining algorithm in PoPArt v.1.7 with epsilon = 0. Each circle represents a different haplotype and the size of a circle correlates with number of individuals belonging to that given haplotype. Only haplotypes found in more than one sample are numbered. Colours indicate the geographic origin of sequences; black dots indicate unsampled or extinct haplotypes.
The measure of population differentiation due to genetic structure (FST) was much lower for A. sudafricanum compared to the Demoreptus species (Table
One-level AMOVA results for A. sudafricanum, D. capensis and D. natalensis showing percentage variation among and within populations and the fixation index (FST). Significant p-values (< 0.05) are set in bold.
Species/clade | % variation | FST | |
---|---|---|---|
Among | Within | ||
A. sudafricanum | 52.33 | 47.67 | 0.52327 |
A. sudafricanum unresolved group | 39.43 | 60.57 | 0.39426 |
D. capensis | 94.83 | 5.17 | 0.94827 |
D. natalensis | 95.39 | 4.61 | 0.95393 |
This study considered evidence of the phylogenetic structure of three species of Baetidae corresponding to two different habitat requirements. Results indicate that habitat-restricted Demoreptus species have greater maternal genetic structure than widespread A. sudafricanum, showing notable genetic differentiation associated with geographic localities and catchments. This is evident from the haplotype networks in a MJN analysis, FST values from an AMOVA and the phylogeographical structure indicated by phylogenetic trees.
Phylogeographical structure of habitat generalist, A. sudafricanum retrieved six distinct, well-supported clades and one widespread grade of individuals from widespread (Escarpment and non-Escarpment) sites across the sampling area. Afroptilum sudafricanum was best represented with a haplotype network (Fig.
The habitat specialist species, D. natalensis and D. capensis are rheophilic and found on rock faces associated with waterfalls and large bedrock sections in shallow but fast-flowing sections of mountain streams. Analyses indicate restricted gene flow over distance and across catchments, a possible consequence of isolation by habitat limitations in mountainous areas. Distinct clades retrieved from phylogenetic analyses show a close association with geographic locality. Demoreptus natalensis returned clades and haplotypes exclusive to Sneeuberg and Winterberg areas; the Eastern Cape Drakensberg clade included two study areas (Barkly Pass and Rhodes areas); and Kamdebooberg was unresolved. Demoreptus capensis had a similar result, but the Rhodes area returned a separated clade with a well-supported, long branch. Suggestively, the samples of A. sudafricanum and D. capensis collected at Rhodes both occupy long branches in their respective phylogenetic analyses (Figs
Preliminary re-examinations indicate morphological differences between D. capensis from Rhodes and D. capensis from other localities, and between D. natalensis from Barkly East and D. natalensis from other localities (HMBJ, pers. obs.); these characters will be documented in a subsequent taxonomic study. Other areas in the Drakensberg range in KwaZulu-Natal and Lesotho should be sampled to investigate the range of this mitochondrial clade and whether it occurs throughout high altitude, mountainous areas. A caveat is that the Demoreptus population analyses involve limited sample sizes from few localities, which can produce misleading clustering (
Previous studies on South African species have found genetic differentiation according to catchments in both animals with limited dispersal ability (
The high support values for some geographically localised clades within A. sudafricanum and the two Demoreptus species could indicate the presence of cryptic species or local haplotype filtering and mutation due to protracted isolation (
The observed deep haplotype divergences in all three species studied and the recent population expansion in A. sudafricanum may be explained by possible Quaternary glaciation in the Drakensberg area, where small glaciers formed as low as 2100 m on south-facing slopes (
However, because they are asexually and maternally inherited, strongly divergent haplotypes that originated in relict populations may not reflect contemporary mating pattern if those isolated populations’ ranges subsequently expand to restore potential panmixis (
These results help to illuminate some of the evolutionary processes occurring in mayfly species and highlight the effect of habitat-specificity on haplotype diversity and partitioning within a species. While all three species have qualitatively similar levels of dispersal potential in terms of flight, they show differences in gene flow, suggesting that other processes, such as species-specific habitat requirements, may contribute to genetic population structure. These results have implications for the conservation of riverine organisms, the reintroduction of locally extinct taxa and the rehabilitation of disturbed environments (
In South Africa, it is legislated that catchments are used as management units (
In addition, dispersal among adjacent catchments has implications for the recovery of lotic systems following disturbance (
This study highlights the importance for future studies on community structure, biodiversity, and biomonitoring, where the taxonomic accuracy of species identification is crucial (
We thank Matthew Janks, Lizzie Gaisford, Sarah Macqueen, David Taylor and Duncan Stodart for field assistance; Riaan Strauss and Mardi Nolands (Rhodes University) for administrative assistance; Alistair Barker (Rhodes University) and Taryn Bodil (South African Institute for Aquatic Biodiversity, Makhanda) for assisting with sequencing; and Jean-Luc Gattolliat, Gavin Gouws and an anonymous reviewer for their useful inputs. All species used in this study are neither CITES-listed species nor endangered species according to regional Red Lists or South Africa’s Threaten or Protected Species (ToPS) legislation, so no special permission was necessary for sampling. Sampling was collected with the routine permissions of the relevant land owners in terms of Section 63 of the Nature and Environmental Conservation Ordinance, 1974 (Ordinance 19 of 197 4), Sections 24 and 25 of Environmental Conservation Decree, 1992 (Decree No. 9 of 1992, former Transkei) and Sections 20 and 21 of the Nature Conservation Act 1987 (Act No. 10 of 1987, former Ciskei), permit numbers CRO 14/12CR and CRO 15/12CR, and Rhodes University permit number RU5-BE-939/001-2013. Research funding was provided by the Rhodes University Council to CLT and by the National Research Foundation (NRF) of South Africa (Grant Unique Number GUN 2069059) to NPB. Any opinion, findings and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the National Research Foundation. The Eastern Cape Province Department of Sport, Recreation, Arts and Culture (DSRAC) supported the Albany Museum. The authors have declared that no competing interests exist.
List of GenBank sequence accession numbers for each sample
Data type: molecular data