Click here  for article in "Word" format.

These  article has been originally published as Reichard, M., Polacik, M. & Sedlácek, O. (2009) Distribution, colour polymorphism and habitat use of the African killifish, Nothobranchius furzeri, the vertebrate with the shortest  lifespan. Journal of Fish Biology, 74, 198-212.

Note: When cited or referred to, please use the reference to original paper.

 

 

 

 

Distribution, colour polymorphism and habitat useof the African killifish, Nothobranchius furzeri, the vertebrate with the shortest lifespan

 

M. Reichard1,2, M. Polacik1 and O. Sedlácek3

 

1Institute of Vertebrate Biology, Academy of Sciences of the Czech Republic, Kvetná 8, 603 65 Brno, Czech Republic; 2Department of Biology, University of Leicester, University Road, Leicester LE1 7RH, UK; 3Deparment of Ecology, Faculty of Science, Charles University, Vinicná 4, 128 44 Prague, Czech Republic

 

 

 

Author to whom correspondence should be addressed:

Martin Reichard[1], Institute of Vertebrate Biology, Academy of Sciences of the Czech Republic, Kvetná 8, 603 65 Brno, Czech Republic

 

 

Running headline: Nothobranchius furzeri distribution


Intensive collection in southern Mozambique across and outside the potential range of Nothobranchius furzeri, the species with the shortest recorded lifespan among vertebrates used as a model in ageing research, revealed that, contrary to previous data, it is a widespread species. It occurs in small freshwater pools south of the Save River and north of the Incomati River, including basins of the Limpopo, Changane, Chefu, Mazimechopes and Vaneteze rivers. During collection in February 2008 (the second part of the rainy season), populations were strongly female-biased (mean proportion 28 % of males across 19 populations) and there was a spatial pattern in female bias among metapopulations. Populations varied in the proportion of male colour morphs. Fourteen populations were composed exclusively of the red male phenotype, 3 populations of the yellow male phenotype and 12 populations were mixed. Overall, the red phenotype was more common, but there was strong geographical variation in morph proportion, with yellow males more abundant at the periphery and red male dominance in the centre of the range of N. furzeri in the Limpopo basin. Nothobranchius furzeri was sympatric with N. orthonotus (35 % of investigated pools) and N. rachovii (27 % of sites). Analysis of habitat use of N. furzeri is presented; N. furzeri was associated with pools containing a soft muddy substrate and turbid water.

 

Key words: adult sex ratio; Cyprinodotiformes; fish-habitat association; geographic range; male colour polymorphism; species coexistence

 

Introduction

The genus Nothobranchius (Cyprinodontiformes, Nothobranchiidae) is a group of

small (range 3-15 cm, median 5 cm), short-lived (3-12 months) fish that inhabit isolated pools throughout the savannah region of eastern and central Africa. Their distribution ranges from southern Sudan to KwaZulu Natal in South Africa (Wildekamp, 2004). The genus includes about 50 currently described species, separated into 5 well-defined clades (Huber, 2000; Wildekamp, 2004). The life history of Nothobranchius is adapted to annual desiccation of their habitat. Fish hatch after the start of rainy season, grow rapidly and become sexually mature within a few weeks. After reaching sexual maturity, they reproduce daily and females lay 5-50 eggs each day (Haas, 1976a). Eggs are spawned into a substratum and remain there after pool desiccation. While the habitat is dry, embryos survive in a developmental diapause until the next rainy season (Wourms, 1972; Wildekamp, 2004). All Nothobranchius fishes are extremely sexually dimorphic and dichromatic; males are robust and colourful and females are pale yellow or brown. The bright male colouration is sexually selected (Haas, 1976b) and species specific (Huber, 2000). Many species occur in several colour forms that are either sympatric or allopatric (Huber, 2000; Wildekamp, 2004). Several Nothobranchius species may co-occur in sympatry in the same pool (Huber, 2000).

The Nothobranchius furzeri Jubb has the shortest recorded lifespan amongst all vertebrates (Valdesalici & Cellerino, 2003; but see Depczynski & Bellwood, 2005). In nature, its lifespan is limited by habitat existence, but its survival is similarly short in captivity, with a sharp increase in mortality at the age of 6 weeks and a maximum post-hatch lifespan of less than 12 weeks (Valdesalici & Cellerino, 2003). Given that mortality in N. furzeri is intrinsic and ageing fish show severe tissue degradation (e.g. accumulation of lipofuscin in the liver) and a sharp decrease in cognitive functions and locomotor activity (Valenzano et al., 2006a), N. furzeri has been established as a model species in ageing research (Genade et al., 2005; Valenzano et al., 2006b). It has been used for pharmacological studies of lifespan extension (Valenzano & Cellerino 2006) and offers potential for investigating the genetic mechanisms controlling ageing. While the laboratory line shows a lifespan of less than three months (Valdesalici & Cellerino, 2003), there is supposedly large variation in lifespan and life-history traits among different N. furzeri populations (Terzibasi et al., in review).

Current knowledge of the distribution and natural history of N. furzeri is deficient. The species was described from Gona Re Zhou Game Reserve in Zimbabwe, where fish were collected in Sazale Pan in March 1968, December 1968 and January 1969 (Jubb, 1971). Sazale Pan is located 25 km from the border with Mozambique and is drain into the upper Guluene river, a tributary of the Chefu river flowing to Mozambique, an important tributary of the Chingovo basin. The Chingovo basin never reaches the ocean as it disappears in a series of shallow swamps and inland lakes on a large flat plain at an altitude of approximately 80 metres above sea level (masl). However, the topography of the area suggests that this river system may have been connected to the Limpopo basin in humid periods of the Pleistocene in its downstream part via the Changane river (Jubb,1971; Skelton, 2001).

Since its description, N. furzeri has never been collected in the Gona Re Zhou Reserve again (conservation status currently elevated to National Park) and, until recently, only the captive strain collected in the late 1960’s was maintained by hobbyists. Gona Re Zhou, originally designed as an extension of Kruger National Park of the South Africa in 1960’s to protect migrating elephants, became a centre of guerrilla activists and subjected to poaching soon after the discovery of N. furzeri and has not been accessible for sampling thereafter. In 1999, three new populations of N. furzeri were discovered along the north bank of the Limpopo river in southern Mozambique (Wood, 2000). Interestingly, the new populations differed markedly in colouration and were considered new, as yet undescribed species (Wood, 2000). While the type population of N. furzeri consists exclusively of males with a yellow stripe at the margin of the caudal fin (Jubb, 1971), all three populations (adjacent to each other) collected in the Limpopo basin were composed exclusively of males with red caudal fins. Only when the F1 captive generation produced a mixture of red and yellow phenotypes the fish were recognized as a new phenotypic morph of N. furzeri. Between 2004 and 2007, when N. furzeri was established as a model species for ageing research (Valdesalici & Cellerino, 2003) and potential field sites in Mozambique become accessible to sampling, a total of six collection trips was completed. During those collections, several other populations were discovered, including populations with a mixture of red and yellow phenotypes. The sampling has largely been conducted with a sole aim exporting wild fish and establishing captive populations either in hobby (by B. Watters in 2004, H. Hengstler in 2005, V. Gomes in 2006) or in research laboratories (by A. Cellerino and collaborators in 2004, 2007 and by S. Schories & M. Schartl in 2006) (Nothobranchius Maintenance Group, 2008; B. Watters, personal communication). Unfortunately, no collection was systematic and no precise data on distribution, demographic parameters or habitat preference are available.

Data on the distribution, ecology and demography of N. furzeri are essential for their full appreciation in studies investigating the evolutionary origin and consequences of ageing (Genade et al., 2005). In the present study, the results of a systematic survey of 28 N. furzeri populations are presented. Specifically, (1) the range extent of N. furzeri was examined along 10 transects within the potential range of N. furzeri in Mozambique, (2) sex ratio and male colour polymorphism were investigated in most sampled populations and (3) habitat preference and habitat segregation between N. furzeri and other Nothobranchius spp. (sympatric with N. furzeri) were studied.

 

Materials and Methods

The study area in Southern Mozambique (from S 25° 50’ to S 19° 17’, i.e. between the rivers Incomati and Pungue) was divided into 10 transects with respect to particular river basins. Transects followed roads accessible to a 4WD vehicle (Fig.1). The area included basins of the Incomati (comprising basins of the Mazimechopes and Vaneteze rivers), Limpopo (including the Changane basin), Save, Gorongosa, Buzi and Pungue rivers that flow into the Indian ocean and the of the endorheic Chefu basin that desiccates in a large flat pan east of the Banhine National Park (Fig. 1). Several small coastal basins were also investigated. Along Transect 1 (T1, road between Chibuto and Chicualacuala (a town at the border with Zimbabwe), along the Limpopo typically 5-20 km from the left bank of the river) 23 habitats were sampled. Along Transect 2 (T2, between the main road along the Limpopo river and the Nuaneteze river channel), no suitable pool was found. All 22 sites sampled along Transect 3 (T3, from Mapai towards Massagena) were within the Chefu basin (not connected to the Limpopo basin). Along Transect 4 (T4, the Changane basin), 13 sites were investigated along the road between Chirrunduo and Chigubo. Seven sites were sampled along Transect 5 (T5, the Incomati basin including the Vaneteze and the Mazimechopes basins, southwest of the Limpopo). At the lower Limpopo (Transect T6, downstream from the start of T4 where the Changane river meets the Limpopo, i.e. <100 km from the ocean, measured as the straight distance), 10 sites were investigated. Finally, along Transect 7 (T7, minor coastal basins) 17 sites were sampled along the main N1 road between Maputo and the Save river of which 4 sites were considered within the range of N. furzeri retrospectively, and used in further habitat analysis (Fig. 1). At Transects 8 (T8, the Save and Gorongosa basins), T9 (the Buzi basin) and T10 (the Pungue basin) 9, 5, and 8 sites were sampled, respectively. A summary of sampling sites is presented in Supplementary table I.

Sampling was conducted between 8 and 22 February 2008 at the end of the rainy season. Pools encountered along the transect were sampled using a dip net with a triangular metal frame (45 by 45 cm, mesh size 5 mm) on a long (1.5 m) wooden pole. Typically, 15 – 40 hauls were performed at each site to allow semi-quantitative estimates of Nothobranchius abundance. Additional hauls were performed if more fish were needed for a specific analysis (e.g. colour polymorphism, habitat use study). At some habitats, a beach seine (length 2.7 m, depth 0.7 m, mesh size 4 mm) was used in addition to the dip net. Fewer dip net hauls were taken if the habitat was too small to accommodate 15 hauls. The chosen mesh size was sufficient to capture all Nothobranchius fish present at sampling sites. At each site, GPS coordinates, altitude (Garmin eMaps GPS handset), water temperature (°C), conductivity (µS.cm-1) (Hanna Combo), maximum and typical water depth (to the nearest 5 cm), vegetation (submergent, emergent littoral, Nymphaea; as absent, rare or abundant), water turbidity (clear blackwater; transparent: bottom visible; turbid: visibility 1-10 cm; very turbid: visibility <1 cm), substratum (sand, hard clay bottom, mud: a person standing on bottom sinks approximately 3-15 cm, soft mud: a person standing on substrate sinks >15 cm), area (estimated to the nearest 10 m) and time of day were recorded. The distance to the nearest stream was measured to the nearest 0.1 km by superimposing GPS coordinates on fine-resolution satellite images using Google Earth. Water temperature was not considered in statistical analyses because it fluctuated widely during the day. Waterproof dataloggers exposed at two representative pools of the study area at a depth of 10-15 cm below the surface (the middle of the water column) for 28 and 96 hours, respectively, showed that water temperature reached minimum of 20.5 °C before sunrise and maximum of 35.2 °C at 14:00 hours. All fish encountered were identified to species (Cyprinidontiformes) or generic level (other orders) according to Skelton (2001) and Wildekamp (2004). A study of within-pool habitat use of N. furzeri was performed at site 55. A total of 20 dip net points were sampled in each of the three habitat types present (Nymphaea vegetation, submergent littoral vegetation, open water) and N. furzeri presence was compared using GLM with binomial error, using water depth (measured at each sampling point) as a covariate.

For all analyses, a matrix of species presence and environmental variables was reduced by deletion of sites outside the range of N. furzeri (i.e. sites in basins where N. furzeri has not been recorded), yielding a total of 79 sites in all analyses. A one-way ANOVA was used to compare sex and colour morph ratios among parts of the N. furzeri range (defined as transects). Sex and colour morph ratios were expressed as proportions of males and red morph respectively. For these analyses, only populations with at least 9 fish (sex ratio) or 9 males (colour polymorphism) were included. Habitat characteristics at landscape scale (across individual pools) were visualized by detrended Canonical Correspondence Analysis (dCCA) using CANOCO for Windows 4.5 (ter Braak & Šmilauer, 1997). The dCCA is an ordination technique that relates species presence to variation in environmental variables. For dCCA, the species matrix was coded for 4 columns; N. furzeri, Nothobranchius ortonothus (Peters), N. rachovii Ahl and “no Nothobranchius”. The dCCA is useful for providing insights into relationships between species and habitat since it reduces a multidimensional space of habitat variables and bivariate plots allow a graphical illustration of species occurrence in relation to habitat variables. In brief, each habitat variable is illustrated by a vector; the length of each vector is proportional to the importance of the habitat variable in explaining the variability in the species matrix. The minimum adequate model was constructed by stepwise model simplification from a maximal model containing all explanatory habitat variables and quadratic terms of every continuous variable (Crawley, 2007) by Akaike Information Criterion (AIC) using GLM ANCOVA (a mixture of continuous and categorical variables) with a binary response (presence/absence) variable and log-link function in R 2.0.1 (R Core Development Team, 2006). Histograms were constructed for individual habitat variables to visualize their relationship to N. furzeri presence.

 

Results

Distribution

The presence of N. furzeri was recorded at 29 of 124 sampling sites investigated, of which 79 were retrospectively classified as within the range of N. furzeri (south of the River Save, excluding the Save and coastal plains of the Indian Ocean). A map of the N. furzeri range with sampling sites investigated is presented in Fig. 1. The findings presented here confirm previous records from the lower Limpopo, the Chefu and the Mazimechopes basins and show that pools with N. furzeri populations are not as sparse as believed previously (earlier collections identified only eight pools with N. furzeri populations). Further, two populations of N. furzeri in the Vaneteze basin (part of the Incomati basin) were recorded, which extend N. furzeri distribution in a south-eastern direction.

The recorded range of N. furzeri is 282 km across its longest axis which increases to 333 km if the type locality is included. The distance between the furthest upstream and furthest downstream populations along the Limpopo is 125 km, with an altitudinal range from 45 to 144 masl. The furthest site from the left bank of the Limpopo was 117 km at the Chefu basin and 52 km from the right bank of the Limpopo in the Vaneteze basin.

 

Sex ratio and colour polymorphism

Females dominated most populations with a mean proportion of 72 % across 19 populations where at least 25 N. furzeri were collected. Of those 19 populations, only one had a male-biased sex ratio (53 males, 34 females), while site 124 had the most female biased sex ratio (9 males, 105 females) (Table I). There was a less female-biased sex ratio in populations along T3 than along T4 and T5 (ANOVA, F3,13 = 4.7, P = 0.020; Tukey HSD tests for T3 vs T4: P = 0.041 and for T3 vs T5: P =  0.031, all other pair-wise comparisons P > 0.22). Sex ratios were 0.41 ± 0.05 in T3 populations, 0.32 ± 0.04 in T1, 0.21 ± 0.04 in T4 and 0.17 ± 0.06 in T5.

A total of 14 sites with pure red populations, 3 sites with pure yellow populations (though they contained only 2, 3, and 9 males captured) and 12 mixed populations were recorded. Pure red populations occurred at Transect 1 (Limpopo basin) and Transect 4 (Changane basin). Mixed populations were found at high altitude sites in Transect 1 (2 sites at 129 and 146 masl), along Transect 3 (5 sites between 124 and 128 masl), at 3 sites in Transect 4 (between 24 and 34 masl) and 2 sites in Transect 5 (30 and 56 masl). Exclusively yellow males were found on the right bank of the Limpopo river at two sites in the Vaneteze basin (Transect 5, with 3 and 9 males captured at 88 to 96 masl) and one site in the Mazimechopes basin (Transect 5, only 2 males collected, 48 masl). In mixed populations, red males typically (but not always) dominated (Table I). There was a significant difference in the proportion of red males among transects (ANOVA, F3,15 = 10.3, P = 0.001), with significantly fewer red males in populations along T5 (0.05 ± 0.15) than along T1 (0.89 ± 0.09), T3 (0.63 ± 0.09) and T4 (0.92 ± 0.09) (Tukey HSD tests for T5 vs all other transects P < 0.05).

 

Coexistence with other fish species

Up to three Nothobranchius species inhabited a single temporary pool. Within the N. furzeri range (79 sites), N. furzeri was recorded with N. orthonotus at 10 sites (35 % from 29 sites where N. furzeri was found), with N. orthonotus and N. rachovii at 5 sites (17 %) and with N. rachovii exclusively at 3 sites (10 %). At 16 sites, N. furzeri occurred without any other Nothobranchius species (55 %).

Other fishes recorded sympatric with N. furzeri were small cyprinids, Barbus sp. (3 sites), lungfish, Polypterus annectens brieni Poll (4 sites) and catfish, Clarias gariepinus (Burchell) (2 sites). Nothobranchius furzeri was never recorded sympatric with tilapias (Tilapia s. l. juveniles including Oreochromis mossambicus (Peters) and Tilapia rendalli (Boulenger) for which determination was confirmed on adult individuals) that occurred at 11 sampled sites within the N. furzeri range.

Habitat segregation among Nothobranchius spp. is presented in Fig. 2. The first two axes in the CCA accounted for 92.1 % of variation in species-habitat data (55.2 % for the first axis) (test for all canonical axes, F = 1.56, P = 0.018). For the first three axes, eigenvalues were 0.338, 0.073 and 0.024. The highest interspecific segregation was across the gradients of water turbidity, bottom composition and water conductivity (Fig. 2).

 

Habitat use

The dCCA revealed that N. furzeri inhabited sites with soft substratum and high turbidity (Fig. 2). The minimal adequate model (GLM with binomial error) retained conductivity, substratum, altitude, distance from the nearest river and surface area as significant factors affecting the presence of N. furzeri populations (Table II). Nothobranchius furzeri was found at altitudes between 18 and 140 masl, at sites with conductivity from 50 to 625 µS.cm-1, soft or very soft substrata and surface area typically between 50 and 625 m2 (6 - 70,000 m2 including outliers). The relationship between N. furzeri presence and distance from the nearest river was complex (Fig. 3).

Nothobranchius furzeri often inhabited simple shallow pools without any vegetation. Within pools with vegetation, N. furzeri typically occurred in Nymphaea in close proximity to open water or at the interface between vegetation and open water. At site 55 where quantitative sampling was performed, N. furzeri preferred Nymphaea vegetation (present at 12 out of 20 points) over littoral vegetation (3 of 20) and open water (0 of 20) and the difference in the use of Nymphaea vegetation compared to other two habitats was significant (GLM with binomial error, P = 0.019 for habitat type). Water depth was not a significant covariate (P = 0.372). Water temperature was not significant predictor of N. furzeri presence and was higher in flooded littoral vegetation than in open water and in Nymphaea vegetation (ANOVA, F2,16 = 107.5, P < 0.001). Water temperature in littoral vegetation was 30.8 ± 0.01 °C, 29.5 ± 0.01 °C in Nymphaea vegetation and 29.4 ± 0.01 °C in open water.

 

Discussion

The distribution of N. furzeri extends over the Incomati, Limpopo and Chefu basins in a relatively small part of southern Mozambique. Additionally, at least one N. furzeri population occurs in the Zimbabwean part of the Chefu basin (Jubb, 1971). The type locality in Sazale pan is situated at 422 masl, while all other known populations inhabit pools between 18 to 140 masl. It is likely that more N. furzeri populations occur in the upper Chefu basin, between Sazale pan and the collection sites along Transect 3 at 132 and 128 masl (sites 33 and 40), respectively. Unfortunately, the area in Zimbabwe and near the border between Zimbabwe and Mozambique is inaccessible to research for security reasons and there is little prospect of obtaining permits to sample that part of the N. furzeri range. There is a report of a collection of N. furzeri from the upper Chefu basin at the Mozambican-Zimbabwean border, though no details on sampling are available (B. Watters, personal communication). The present study included sampling at high altitude area of the Limpopo basin (up to 363 masl), but no freshwater pools were encountered higher than 256 masl and no N. furzeri populations were found higher than 140 masl in the Limpopo basin. Therefore, it is possible that high altitude populations occur solely in the upper Chefu basin. To confirm this speculation, a larger area in the upper Limpopo basin needs to be investigated, but the savannah further from the roads is inaccessible for logistical reasons (presence of land mines). Ultimately, the internal structure of the N. furzeri range can be determined using genetic markers, but current knowledge and the assumption that Nothobranchius are weak dispersers upstream (Wildekamp, 2004) lend support to a putative hypothesis that the upper Chefu basin contains the source populations of N. furzeri and that N. furzeri are dispersed by occasional extensive floods resulting from cyclones forming off Mozambique in the Indian Ocean. Such rare events may connect isolated pools and allow downstream migration and colonization. How N. furzeri colonized the high altitude sites and its spreading downstream and into adjacent basins is the subject of ongoing phylogeographic investigations that include N. furzeri and its sister, yet undescribed, species from the lower Save basin.

The two colour morphs that are expressed in male N. furzeri have largely overlapping ranges. In the type population in the upper Chefu basin, only yellow males were found (Jubbs, 1971). In the lower Chefu basin, all populations were composed of a mixture of yellow and red males, while only 7 out of 13 populations in the Limpopo basin contained any yellow males. The yellow morph also dominated in the Incomati basin in the south-western part of the N. furzeri range, with exclusively yellow males found in the Vaneteze basin and yellow-dominated populations in the Mazimechopes basin. In contrast, the red morph dominated in most populations in the Limpopo basin that includes the centre and the largest part of the N. furzeri range. Colour morphs are often associated with different life-histories and behavioural tactics and their coexistence may be enabled by fluctuation in environmental and population factors (Sinervo et al., 2000). There are some preliminary data suggesting that N. furzeri populations from more humid areas have a longer lifespan and later commencement of physiological changes linked to senescence (Terzibasi et al., review), but nothing is currently known about the association between differences in colouration and life history.

The sex ratio was almost exclusively female-biased, but the degree of female bias varied among transects. In fish, there is a great variability in sex determination and genetic and environmental determination of sex has evolved multiple times (Mank et al., 2006). In a closely related species, Nothobranchius guentheri (Pfeffer 1893), sex is determined by a complex combination of sex chromosomes (Ewulonu et al., 1985) and it is likely that sex determination in N. furzeri is also genetic. No sex related differences in mortality rate or biases in the adult sex ratio were reported in laboratory studies of life expectancy (Valdesalici & Cellerino, 2003; Genade et al., 2005; Valenzano et al., 2006a, 2006b; Terzibasi et al., in review) suggesting increased male extrinsic mortality in natural populations. Male colouration in Nothobranchius spp. is sexually selected (Haas, 1976b) and showy sexual displays may be associated with increased mortality risk (Hunt et al., 2004). Another plausible explanation is that high male mortality stems from severe male-male disputes over occupancy of superior positions in spawning arenas that are located at particular places within the habitat (Haas, 1976a) and where females likely prefer to lay their eggs. Male-male competition is intense in Nothobranchius and often involves serious injuries to subordinate males (Huber, 2000).

The habitat of all Nothobranchius is characterized by vertisol soils on alluvial deposits (Wildekamp, 2004) and Nothobranchius populations do not occur in pools developed on laterite soils, because only vertisols provide suitable soil structure for survival of dormant eggs during the dry season (Wildekamp, 2004). The present study revealed that other habitat factors are also significantly associated with the presence of N. furzeri populations. The best predictors of N. furzeri presence were a soft muddy substratum and very turbid water (Figs 2 and 3). The water turbidity was likely associated with disturbance by domestic cattle. The cattle, their hoofprints and dung were often encountered at the N. furzeri sites, but no quantitative data were collected to substantiate this observation. Nothobranchius furzeri co-occurred with two other Nothobranchius species at several sites. Multivariate analysis of habitat associations revealed that N. orthonotus is intermediate in its habitat requirements between N. rachovii, a coastal plain species (Wildekamp, 2004), and N. furzeri. Polypterus annectens brieni and C. gariepinus sometimes co-occur with N. furzeri and are their potential predators. In contrast, N. furzeri was not recorded in sympatry with tilapias, which are otherwise abundant in many pools in the study area. This is likely to be a consequence of the different habitat requirements of tilapias, such as connection to permanent habitats.

It is believed that the present data on the distribution and ecology of N. furzeri will contribute to the understanding of the association between environmental conditions, life expectancy and the evolution of rapid onset of senescence. The latest advance of research on N. furzeri ageing revealed that a laboratory line derived from a population from a high altitude/low precipitation site (Transect 3 in the present study) showed a shorter lifespan and more rapid onset of cognitive decline than a laboratory line derived from a population inhabiting a low altitude/high precipitation site (low altitude site at Transect 1 in the present study) (Terzibasi et al., in review). Ageing in those laboratory strains also differed from an established laboratory strain originated in Gona Re Zhou (Valdesalici & Cellerino, 2003; Terzibasi et al., 2007). While such results make N. furzeri an ideal model species for studies on genetic control of life-history traits (Genade et al., 2005; Terzibasi et al., 2007), data presented here show that a larger scale study using a series of natural populations along clines of altitude and environmental variables is possible and can bring stronger insights into the evolution of ageing. In particular, geographic variation in habitat duration is expected to be reflected in variation in developmental rates, ageing and lifespan. Preliminary data on six N. furzeri populations showed high inter-population genetic distances at a neutral genetic marker (a fragment of cox1 gene in the mitochondrial genome) (Terzibasi et al. in review). This deep genetic structuring makes the system highly amenable to research into the genetic basis of the ageing process, including the identification of critical genes (and their expression) affecting the ageing process. Experimental gerontology has been so far largely limited to mutant strains of laboratory organisms (Terzibasi et al., 2007), but recent progress in genomic procedures in natural populations of non-model species should enable synthesis of ecological, evolutionary and molecular approaches (Ellegren & Sheldon, 2008).  Unlike the genes linked to the short lifespan in laboratory organisms, which are often coupled with serious developmental deficiencies (Terzibasi et al., 2007), the genes responsible for rapid ageing in N. furzeri have been selected in the context of life history evolution under natural conditions and are expected to result from adaptive trade-offs between costs and benefits associated with their functions. Hence, it is hoped that the present study constitutes an important step to our understanding of ageing in natural populations.

 

Acknowledgements

The study was supported by Czech Science Foundation (206/06/P152) and Association for Studies on Animal Behaviour. All fieldwork complied with legal regulations of Mozambique (collection permit DPPM/053/7.10/08 and sample export permit 013/MP/2008 of the Mozambican Ministry of Fisheries). The authors thank R. Spence, R. J. Wootton and two anonymous referees for comments on the ms and M. Dušková for drafting river network on the Fig. 1. M. R. conceived and designed research, analyzed the data and wrote the paper. All authors collected data and commented on the research design and manuscript.

 

 

 

References

ter Braak, C. J. F. & Šmilauer, P. (1997). Canoco for Windows Version 4.02. Wageningen: Centre for Biometry.

Crawley, M. J. (2007). The R Book. New York: John Wiley.

Depczynski, M. & Bellwood, D. (2005). Shortest recorded vertebrate lifespan found in a coral reef fish. Current Biology 15, 288-289.

Ellegren, H. & Sheldon, B. C. (2008). Genetic basis of fitness differences in natural populations. Nature 452, 169-175.

Ewulonu, U. K., Haas, R. & Turner, B. J. (1985). A multiple sex chromosome system in the annual killifish, Nothobranchius guentheri. Copeia 1985, 503-508.

Genade, T., Benedetti, M., Terzibasi, E., Roncaglia, P., Valenzano, D. R., Cattaneo, R. & Cellerino, A. (2005). Annual fishes of the genus Nothobranchius as a model system for aging research. Aging Cell 4, 223-233.

Haas, R. (1976a). Behavioral biology of the annual killifish, Nothobranchius guentheri. Copeia 1976, 81-91.

Haas, R. (1976b). Sexual selection in Nothobranchius guentheri (Pisces: Cyprinidontidae). Evolution 30, 614-622.

Huber, J. H. (2000). Killi-Data 2000. Updated Checklist of Taxonomic Names, Collecting Localities and Bibliographic References of Oviparous Cyprinodont Fishes (Cyprinodontiformes). Paris: Cybium.

Hunt, J., Brooks, R. & Jennions, M. D.,Smith, M. J., Bentsen, C. L. & Bussiere, L. F. (2004). High-quality male field crickets invest heavily in sexual display but die young. Nature 432, 1024-1027.

Jubb, R. A. (1971). A new Nothobranchius (Pisces, Cyprinodontidae) from Southeastern Rhodesia. Journal of American Killifish Association 8, 12-19.

Mank, J. E., Promislow, D. E. L. & Avise, J. C. (2006). Evolution of alternative sex-determining mechanisms in teleost fishes. Biological Journal of the Linnean Society 87, 83-93.

R Development Core Team (2006). R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing.

Sinervo, B., Svensson, E. & Comendant, T. (2000). Density cycles and an offspring quantity and quality game driven by natural selection. Nature 406, 985-988.

Skelton, P. (2001). A Complete Guide to Freshwater Fishes of Southern Africa. Cape Town: Struik Publishers.

Terzibasi, E., Valenzano, D. R. & Cellerino, A. (2007). The short-lived fish Nothobranchius furzeri as a new model system for aging studies. Experimental Gerontology 42, 81-89.

Terzibasi, E., Valenzano, D. R., Benedetti, M., Roncaglia, P., Cattaneo, A., Domenici, L. & Cellerino, A. (in review). Aging phenotype of wild-derived lines of the annual fish Nothobranchius furzeri originating from habitats with differences in variables related to extrinsic mortality. PLOS One.

Valdesalici, S. & Cellerino, A. (2003). Extremely short lifespan in the annual fish Nothobranchius furzeri. Proceedings of the Royal Society London B Suppl. 279, S189-S191.

Valenzano, D. R. & Cellerino, A. (2006). Resveratrol and the pharmacology of aging - A new vertebrate model to validate an old molecule. Cell Cycle 5, 1027-1032.

Valenzano, D. R., Terzibasi, E., Cattaneo, A., Domenici, L. & Cellerino, A. (2006a). Temperature affects longevity and age-related locomotor and cognitive decay in the short-lived fish Nothobranchius furzeri. Aging Cell 5, 275-278.

Valenzano, D. R., Terzibasi, E., Genade, T., Cattaneo, A., Domenici, L. & Cellerino, A. (2006b). Resveratrol prolongs lifespan and retards the onset of age-related markers in a short-lived vertebrate. Current Biology 16, 296-300.

Wildekamp, R. H. (2004). A World of Killies: Atlas of the Oviparous Cyprinidontiform Fishes of the World. Volume IV. Elyria: AKA.

Woods, T. (2000). Trip to Mozambique, 1999. BKA Killi-News 2000, 105-107.

Wourms, J. P. (1972). The developmental biology of annual fishes. III. Pre-embryonic and embryonic diapause of variable duration in the eggs of annual fishes. Journal of Experimental Zoology 182, 389-414.

 

Electronic reference

Nothobranchius Maintenance Group (2008). Nothobranchius codes and trips. http://www.nothobranchius.org.

 

 

 

 

 

Figure captions

 

Fig. 1. Map of the southern Mozambique with river network (lines), important towns (circles) and collection transects (thick black line highlighted by shaded areas with codes T1-T7) indicated. Transects outside the range of N. furzeri (determined retrospectively) are marked by broken lines (T7-T10). Pools containing N. furzeri populations are indicated by an arrow; one arrow may represent several adjacent populations. The names of the most important rivers are presented (in italics). Note that many rivers are temporary and may not form flowing sections every rainy season.

Fig. 2. Bivariate plot of detrended Canonical Correspondence Analysis showing the position of species (triangles) and habitat variables (vectors) along the first two canonical axes. Data on altitude; area (estimated area of water surface); bottom (along the gradient of hard to soft); conductivity; distance from the nearest (temporary) stream; littoral, Nymphaea and submergent vegetation; occurrence of sand substrate; and water depth are superimposed over species presence data.

Fig. 3. The association between N. furzeri occurrence and habitat variables. Black bars denote the proportion of habitats with N. furzeri populations in a given interval of the habitat variable. The number of habitats within the intervals is indicated for each histogram. For water turbidity, the category “clear” sums blackwater (N = 2) and transparent water (N =10).

 

 

 

Table I. Sex ratio (proportion of males) and colour morph ratio (proportion of red morph) of 24 populations of N. furzeri across its range. Note that in five populations (sites 3, 4, 7, 9, 54), sex and colour morph ratios were not estimated due to small sample size, non-quantitative sampling or logistic reasons. All these populations contained red males only

 

Sex ratio

Male morphs

 

site

transect

males

females

total

sex ratio

red

yellow

Prop red

Note

1

T1

12

31

43

0.28

12

0

1.00

 

2

T1

9

22

31

0.29

9

0

1.00

 

8

T1

11

16

27

0.41

11

0

1.00

 

13

T1

11

19

30

0.37

108

3

0.97

 

23

T3

53

34

87

0.61

32

21

0.60

 

28

T3

56

100

156

0.36

48

8

0.86

 

29

T3

Not estimated

 

 

6

3

0.67

 

33

T3

61

86

147

0.41

11

50

0.18

 

34

T3

12

33

45

0.27

10

2

0.83

 

43

T1

Not estimated

 

 

6

10

0.38

 

50

T1

8

23

31

0.26

8

0

1.00

 

51

T4

6

23

29

0.21

6

0

1.00

 

53

T4

12

45

57

0.21

11

1

0.92

 

55

T4

11

95

106

0.10

10

1

0.91

 

56

T4

5

20

25

0.20

4

1

0.80

 

120

T5

10

31

41

0.24

1

9

0.10

 

121

T4

81

153

234

0.35

72

9

0.89

 

122

T4

Not estimated

 

 

40

0

1.00

 

124

T5

9

105

114

0.08

0

9

0.00

 

119

T5

2

8

10

0.20

0

2

0.00

#

123

T5

3

4

7

0.43

0

3

0.00

#

40

T3

1

3

4

0.25

0

1

0.00

#

59

T4

0

3

3

0.00

 

 

 

#

61

T4

2

0

2

1.00

2

0

1.00

#

Total

 

375

854

1229

0.28

 

 

 

 

 

# less than 9 males captured; population was not considered in statistical analysis

 

Supplementary table I. List of sampled pools within the range of N. furzeri, with their position in the transect, latitude and longitude and presence of Nothobranchius spp. indicated

Site code

 

Transect

 

Latitude (south)

Longitude (east)

N. furzeri

N. orthonotus

N. rachovii

MZCS08 -

1

T1

24°

09.6

32°

48.1

1

1

1

MZCS08 -

2

T1

24°

03.8

32°

43.9

1

1

1

MZCS08 -

3

T1

24°

03.8

32°

43.9

1

0

1

MZCS08 -

4

T1

24°

03.8

32°

43.9

1

0

1

MZCS08 -

5

T1

24°

02.1

32°

42.1

0

0

0

MZCS08 -

6

T1

24°

02.1

32°

42.1

0

0

0

MZCS08 -

7

T1

23°

41.6

32°

36.6

1

0

0

MZCS08 -

8

T1

23°

41.6

32°

36.6

1

1

0

MZCS08 -

9

T1

23°

41.6

32°

36.6

1

1

0

MZCS08 -

10

T1

23°

37.4

32°

35.8

0

0

0

MZCS08 -

11

T1

23°

37.4

32°

35.8

0

0

0

MZCS08 -

12

T1

23°

28.4

32°

34.0

0

0

0

MZCS08 -

13

T1

23°

27.5

32°

33.8

1

0

0

MZCS08 -

14

T1

23°

08.1

32°

24.4

0

0

0

MZCS08 -

15

T1

23°

06.0

32°

21.7

0

0

0

MZCS08 -

16

T1

23°

05.1

32°

20.2

0

0

0

MZCS08 -

17

T1

23°

05.0

32°

19.8

0

0

0

MZCS08 -

18

T1

23°

00.8

32°

14.2

0

0

0

MZCS08 -

19

T1

22°

41.5

32°

02.0

0

0

0

MZCS08 -

20

T3

22°

44.7

32°

05.4

0

0

0

MZCS08 -

21

T3

22°

39.0

32°

16.3

0

0

0

MZCS08 -

22

T3

22°

32.3

32°

28.5

0

0

0

MZCS08 -

23

T3

22°

30.5

32°

33.0

1

0

0

MZCS08 -

24

T3

22°

30.5

32°

33.0

0

0

0

MZCS08 -

26

T3

22°

30.2

32°

34.2

0

0

0

MZCS08 -

27

T3

22°

28.9

32°

37.2

0

0

0

MZCS08 -

28

T3

22°

28.9

32°

37.2

1

0

0

MZCS08 -

29

T3

22°

27.0

32°

38.8

1

0

0

MZCS08 -

30

T3

22°

23.3

32°

40.1

0

0

0

MZCS08 -

31

T3

22°

23.3

32°

40.1

0

0

0

MZCS08 -

32

T3

22°

23.3

32°

40.1

0

0

0

MZCS08 -

33

T3

22°

21.8

32°

41.9

1

0

0

MZCS08 -

34

T3

22°

08.8

32°

49.5

1

1

1

MZCS08 -

35

T3

22°

10.9

32°

52.0

0

0

0

MZCS08 -

36

T3

22°

14.4

32°

54.9

0

0

0

MZCS08 -

37

T3

22°

20.7

32°

48.1

0

0

0

MZCS08 -

38

T3

22°

20.8

32°

47.8

0

0

0

MZCS08 -

39

T3

22°

21.8

32°

44.4

0

0

0

MZCS08 -

40

T3

22°

21.8

32°

43.5

1

0

0

MZCS08 -

41

T3

22°

31.8

32°

29.4

0

0

0

MZCS08 -

42

T3

22°

32.8

32°

27.9

0

0

0

MZCS08 -

43

T1

23°

18.4

32°

32.1

1

1

1

MZCS08 -

44

T1

23°

26.9

32°

33.7

0

0

0

MZCS08 -

49

T1

24°

12.9

32°

50.0

0

0

0

MZCS08 -

50

T1

24°

12.9

32°

50.0

1

0

0

MZCS08 -

51

T4

24°

23.4

32°

53.7

1

0

0

MZCS08 -

52

T4

24°

22.8

32°

55.4

0

0

0

MZCS08 -

53

T4

24°

22.2

32°

57.0

1

0

0

MZCS08 -

54

T4

24°

22.2

32°

57.0

1

0

0

MZCS08 -

55

T4

24°

21.8

32°

57.7

1

0

0

MZCS08 -

56

T4

24°

21.8

32°

57.7

1

0

0

MZCS08 -

57

T4

24°

19.2

33°

01.9

0

0

0

MZCS08 -

58

T4

24°

17.8

33°

03.3

0

0

0

MZCS08 -

59

T4

24°

17.8

33°

03.4

1

1

1

MZCS08 -

60

T4

24°

17.8

33°

03.4

0

0

0

MZCS08 -

61

T4

24°

14.4

33°

09.8

1

1

0

MZCS08 -

62

T6

24°

27.5

33°

00.9

0

0

0

MZCS08 -

63

T6

24°

32.8

33°

07.7

0

0

0

MZCS08 -

64

T6

24°

32.9

33°

07.8

0

0

0

MZCS08 -

65

T6

24°

37.0

33°

16.7

0

0

0

MZCS08 -

66

T6

24°

39.9

33°

21.2

0

0

0

MZCS08 -

67

T6

24°

40.4

33°

24.6

0

1

1

MZCS08 -

68

T6

24°

48.6

33°

30.5

0

1

1

MZCS08 -

69

T7

25°

3.7

33°

60.0

0

0

0

MZCS08 -

70

T7

25°

07.1

33°

48.9

0

0

0

MZCS08 -

71

T7

25°

46.4

33°

40.5

0

0

0

MZCS08 -

72

T7

25°

47.6

33°

40.4

0

0

0

MZCS08 -

73

T6

24°

59.7

33°

34.0

0

0

0

MZCS08 -

74

T6

25°

00.0

33°

34.9

0

0

1

MZCS08 -

115

T6

24°

59.9

33°

34.3

0

0

0

MZCS08 -

116

T5

24°

44.8

33°

07.1

0

0

0

MZCS08 -

117

T5

24°

41.0

33°

05.5

0

0

0

MZCS08 -

118

T5

24°

29.0

32°

54.3

0

0

0

MZCS08 -

119

T5

24°

25.1

32°

46.7

1

0

1

MZCS08 -

120

T5

24°

19.5

32°

43.2

1

1

0

MZCS08 -

121

T4

24°

21.5

32°

58.4

1

1

0

MZCS08 -

122

T4

24°

18.2

33°

02.8

1

0

0

MZCS08 -

123

T5

24°

38.8

32°

26.7

1

0

0

MZCS08 -

124

T5

24°

35.6

32°

24.3

1

0

0

 

 

 



[1] Email: reichard@ivb.cz, tel.  +420 543 422 522, fax. +420 543 211 346