The genetic architecture of divergence between threespine stickleback species

The genetic and molecular basis of morphological evolution is poorly understood, particularly in vertebrates. Genetic studies of the differences between naturally occurring vertebrate species have been limited by the expense and difficulty of raising large numbers of animals and the absence of molecular linkage maps for all but a handful of laboratory and domesticated animals. We have developed a genome-wide linkage map for the three-spined stickleback (Gasterosteus aculeatus), an extensively studied teleost fish that has undergone rapid divergence and speciation since the melting of glaciers 15,000 years ago. Here we use this map to analyse the genetic basis of recently evolved changes in skeletal armour and feeding morphologies seen in the benthic and limnetic stickleback species from Priest Lake, British Columbia. Substantial alterations in spine length, armour plate number, and gill raker number are controlled by genetic factors that map to independent chromosome regions. Further study of these regions will help to define the number and type of genetic changes that underlie morphological diversification during vertebrate evolution.

Along the speciation continuum in sticklebacks

Speciation can be viewed as a continuum, potentially divisible into several states: (1) continuous variation within panmictic populations, (2) partially discontinuous variation with minor reproductive isolation, (3) strongly discontinuous variation with strong but reversible reproductive isolation and (4) complete and irreversible reproductive isolation. Research on sticklebacks (Gasterosteidae) reveals factors that influence progress back and forth along this continuum, as well as transitions between the states. Most populations exist in state 1, even though some of these show evidence of disruptive selection and positive assortative mating. Transitions to state 2 seem to usually involve strong divergent selection coupled with at least a bit of geographic separation, such as parapatry (e.g. lake and stream pairs and mud and lava pairs) or allopatry (e.g. different lakes). Transitions to state 3 can occur when allopatric or parapatric populations that evolved under strong divergent selection come into secondary contact (most obviously the sympatric benthic and limnetic pairs), but might also occur between populations that remained in parapatry or allopatry. Transitions to state 4 might be decoupled from these selective processes, because the known situations of complete, or nearly complete, reproductive isolation (Japan Sea and Pacific Ocean pair and the recognized gasterosteid species) are always associated with chromosomal rearrangements and environment-independent genetic incompatibilities. Research on sticklebacks has thus revealed complex and shifting interactions between selection, adaptation, mutation and geography during the course of speciation.

Large scale transgenic and cluster deletion analysis of the HoxD complex separate an ancestral regulatory module from evolutionary innovations

The ancestral role of the Hox gene family is specifying morphogenetic differences along the main body axis. In vertebrates, HoxD genes were also co-opted along with the emergence of novel structures such as limbs and genitalia. We propose that these functional recruitments relied on the appearance, or implementation, of regulatory sequences outside of the complex. Whereas transgenic human and murine HOXD clusters could function during axial patterning, in mice they were not expressed outside the trunk. Accordingly, deletion of the entire cluster abolished axial expression, whereas recently acquired regulatory controls were preserved.

Murine c-mpl: a member of the hematopoietic growth factor receptor superfamily that transduces a proliferative signal

The murine myeloproliferative leukemia virus has previously been shown to contain a fragment of the coding region of the c-mpl gene, a member of the cytokine receptor superfamily. We have isolated cDNA and genomic clones encoding murine c-mpl and localized the c-mpl gene to mouse chromosome 4. Since some members of this superfamily function by transducing a proliferative signal and since the putative ligand of mpl is unknown, we have generated a chimeric receptor to test the functional potential of mpl. The chimera consists of the extracellular domain of the human interleukin-4 receptor and the cytoplasmic domain of mpl. A mouse hematopoietic cell line transfected with this construct proliferates in response to human interleukin-4, thereby demonstrating that the cytoplasmic domain of mpl contains all elements necessary to transmit a growth stimulatory signal. In addition, we show that 25-40% of mpl mRNA found in the spleen corresponds to a novel truncated and potentially soluble isoform of mpl and that both full-length and truncated forms of mpl protein can be immunoprecipitated from lysates of transfected COS cells. Interestingly, however, although the truncated form of the receptor possesses a functional signal sequence and lacks a transmembrane domain, it is not detected in the culture media of transfected cells.

The mouse Ulnaless mutation deregulates posterior HoxD gene expression and alters appendicular patterning

The semi-dominant mouse mutation Ulnaless alters patterning of the appendicular but not the axial skeleton. Ulnaless forelimbs and hindlimbs have severe reductions of the proximal limb and less severe reductions of the distal limb. Genetic and physical mapping has failed to separate the Ulnaless locus from the HoxD gene cluster (Peichel, C. L., Abbott, C. M. and Vogt, T. F. (1996) Genetics 144, 1757–1767). The Ulnaless limb phenotypes are not recapitulated by targeted mutations in any single HoxD gene, suggesting that Ulnaless may be a gain-of-function mutation in a coding sequence or a regulatory mutation. Deregulation of 5′ HoxD gene expression is observed in Ulnaless limb buds. There is ectopic expression of Hoxd-13 and Hoxd-12 in the proximal limb and reduction of Hoxd-13, Hoxd-12 and Hoxd-11 expression in the distal limb. Skeletal reductions in the proximal limb may be a consequence of posterior prevalence, whereby proximal misexpression of Hoxd-13 and Hoxd-12 results in the transcriptional and/or functional inactivation of Hox group 11 genes. The Ulnaless digit phenotypes are attributed to a reduction in the distal expression of Hoxd-13, Hoxd-12, Hoxd-11 and Hoxa-13. In addition, Hoxd-13 expression is reduced in the genital bud, consistent with the observed alterations of the Ulnaless penian bone. No alterations of HoxD expression or skeletal phenotypes were observed in the Ulnaless primary axis. We propose that the Ulnaless mutation alters a cis-acting element that regulates HoxD expression specifically in the appendicular axes of the embryo.

Ecological selection against hybrids in natural populations of sympatric threespine sticklebacks

Experimental work has provided evidence for extrinsic post-zygotic isolation, a phenomenon unique to ecological speciation. The role that ecological components to reduced hybrid fitness play in promoting speciation and maintaining species integrity in the wild, however, is not as well understood. We addressed this problem by testing for selection against naturally occurring hybrids in two sympatric species pairs of benthic and limnetic threespine sticklebacks (Gasterosteus aculeatus). If post-zygotic isolation is a significant reproductive barrier, the relative frequency of hybrids within a population should decline significantly across the life-cycle. Such a trend in a natural population would give independent support to experimental evidence for extrinsic, rather than intrinsic, post-zygotic isolation in this system. Indeed, tracing mean individual hybridity (genetic intermediateness) across three life-history stages spanning four generations revealed just such a decline. This provides compelling evidence that extrinsic selection plays an important role in maintaining species divergence and supports a role for ecological speciation in sticklebacks.

Lateral line diversity among ecologically divergent threespine stickleback populations

The lateral line is a mechanoreceptive sensory system that allows fish to sense objects and motion in their local environment. Variation in lateral line morphology may allow fish in different habitats to differentially sense and respond to salient cues. Threespine sticklebacks (Gasterosteus aculeatus) occupy a diverse range of aquatic habitats; we therefore hypothesized that populations within the G. aculeatus species complex might show variation in the morphology of the lateral line sensory system. We sampled 16 threespine stickleback populations from marine, stream and lake (including benthic and limnetic) habitats and examined the distribution, type and number of neuromasts on different regions of the body. We found that the threespine stickleback has a reduced lateral line canal system, completely lacking canal neuromasts. Although the arrangement of lines of superficial neuromasts on the body was largely the same in all populations, the number of neuromasts within these lines varied across individuals, populations and habitats. In pairwise comparisons between threespine sticklebacks adapted to divergent habitats, we found significant differences in neuromast number. Stream residents had more neuromasts than marine sticklebacks living downstream in the same watershed. In two independent lakes, benthic sticklebacks had more trunk neuromasts than sympatric limnetic sticklebacks, providing evidence for parallel evolution of the lateral line system. Our data provide the first demonstration that the lateral line sensory system can vary significantly between individuals and among populations within a single species, and suggest that this sensory system may experience different selection regimes in alternative habitats.

The genetic basis of divergent pigment patterns in juvenile threespine sticklebacks

Animal pigment patterns are important for a range of functions, including camouflage and communication. Repeating pigment patterns, such as stripes, bars and spots have been of particular interest to developmental and theoretical biologists, but the genetic basis of natural variation in such patterns is largely unexplored. In this study, we identify a difference in a periodic pigment pattern among juvenile threespine sticklebacks (Gasterosteus aculeatus) from different environments. Freshwater sticklebacks exhibit prominent vertical bars that visually break up the body shape, but sticklebacks from marine populations do not. We hypothesize that these distinct pigment patterns are tuned to provide crypsis in different habitats. This phenotypic difference is widespread and appears in most of the freshwater populations that we sampled. We used quantitative trait locus (QTL) mapping in freshwater-marine F2 hybrids to elucidate the genetic architecture underlying divergence in this pigmentation pattern. We identified two QTL that were significantly associated with variation in barring. Interestingly, these QTL were associated with two distinct aspects of the pigment pattern: melanophore number and overall pigment level. We compared the QTL locations with positions of known pigment candidate genes in the stickleback genome. We also identified two major QTL for juvenile body size, providing new insights into the genetic basis of juvenile growth rates in natural populations. In summary, although there is a growing literature describing simple genetic bases for adaptive coloration differences, this study emphasizes that pigment patterns can also possess a more complex genetic architecture.

Karyotype differentiation between two stickleback species (Gasterosteidae)

The stickleback family (Gasterosteidae) of fish is less than 40 million years old, yet stickleback species have diverged in both diploid chromosome number (2n) and morphology. We used comparative fluorescence in situ hybridization (FISH) on 2 stickleback species, Gasterosteus aculeatus (2n = 42) and Apeltes quadracus (2n = 46), to ascertain the types of chromosome rearrangements that differentiate these species. The A. quadracus karyotype contains more acrocentric and telocentric chromosomes than the G. aculeatus karyotype. By using bacterial artificial chromosome probes from G. aculeatus in our FISH screen, we found that 6 pericentric inversions and 2 chromosome fusions/fissions are responsible for the greater number of acrocentric and telocentric chromosomes in A. quadracus. While most populations of G. aculeatus have an XX/XY sex chromosome system, A. quadracus has a ZZ/ZW sex chromosome system, as previously reported. However, we discovered that a population of A. quadracus from Connecticut lacks heteromorphic sex chromosomes, providing evidence for unexpected sex chromosome diversity in this species.

Sensory trait variation contributes to biased dispersal of threespine stickleback in flowing water

Gene flow is widely thought to homogenize spatially separate populations, eroding effects of divergent selection. The resulting theory of 'migration-selection balance' is predicated on a common assumption that all genotypes are equally prone to dispersal. If instead certain genotypes are disproportionately likely to disperse, then migration can actually promote population divergence. For example, previous work has shown that threespine stickleback (Gasterosteus aculeatus) differ in their propensity to move up- or downstream ('rheotactic response'), which may facilitate genetic divergence between adjoining lake and stream populations of stickleback. Here, we demonstrate that intraspecific variation in a sensory system (superficial neuromast lines) contributes to this variation in swimming behaviour in stickleback. First, we show that intact neuromasts are necessary for a typical rheotactic response. Next, we showed that there is heritable variation in the number of neuromasts and that stickleback with more neuromasts are more likely to move downstream. Variation in pectoral fin shape contributes to additional variation in rheotactic response. These results illustrate how within-population quantitative variation in sensory and locomotor traits can influence dispersal behaviour, thereby biasing dispersal between habitats and favouring population divergence.

Distinct startle responses are associated with neuroanatomical differences in pufferfishes

Despite the key function of the Mauthner cells (M-cells) in initiating escape responses and thereby promoting survival, there are multiple examples of M-cell loss across the teleost phylogeny. Only a few studies have directly considered the behavioral consequences of naturally occurring M-cell variation across species. We chose to examine this issue in pufferfishes, as previous research suggested that there might be variability in M-cell anatomy in this group of fish. We characterized the M-cell anatomy and fast-start responses of two pufferfish species, Tetraodon nigroviridis and Diodon holocanthus. T. nigroviridis showed robust fast-starts to both tactile and acoustic startling stimuli. These fast-starts occurred with a latency typical of M-cell initiation in other fish, and retrograde labeling of spinal-projection neurons revealed that T. nigroviridis does have M-cells. By contrast, D. holocanthus only rarely exhibited fast-start-like behavior, and these responses were at a substantially longer latency and were much less extensive than those of T. nigroviridis. Using three complementary anatomical techniques we were unable to identify obvious M-cell candidates in D. holocanthus. These results provide a clear correlation between M-cell presence or absence and dramatic differences in fast-start behavior. The rich diversity within the pufferfish clade should allow future studies investigating the factors that contribute to this correlated anatomical and behavioral variation.

Reduction of sexual dimorphism in stream-resident forms of three-spined stickleback Gasterosteus aculeatus

Sexual dimorphism in geometric body shape and external morphology was compared between marine and stream-resident forms of three-spined stickleback Gasterosteus aculeatus collected from North America and Japan. Some aspects of sexual dimorphism were shared between ecotypes: males had larger heads than females with no significant effect of ecotype on the magnitude of sexual dimorphism. By contrast, a significant sex-by-ecotype interaction was found for body depth. Males tended to have deeper bodies than females in both forms, but the magnitude of sexual dimorphism was reduced in stream-resident forms. Although females were generally larger in standard length and had larger pelvic girdles, significant sexual dimorphism in these traits was not consistently found across populations or ecotypes. These results suggest that some aspects of sexual dimorphism were shared between ecotypes, while others were unique to each population. The results further suggest that ecology may influence the evolution of sexual dimorphism in some external morphological traits, such as body depth.

Genetic and physical mapping of the mouse Ulnaless locus

The mouse Ulnaless locus is a semidominant mutation which displays defects in patterning along the proximal-distal and anterior-posterior axes of all four limbs. The first Ulnaless homozygotes have been generated, and they display a similar, though slightly more severe, limb phenotype than the heterozygotes. To create a refined genetic map of the Ulnaless region using molecular markers, four backcrosses segregating Ulnaless were established. A 0.4-cM interval containing the Ulnaless locus has been defined on mouse chromosome 2, which has identified Ulnaless as a possible allele of a Hoxd cluster gene(s). With this genetic map as a framework, a physical map of the Ulnaless region has been completed. Yeast artificial chromosomes covering this region have been isolated and ordered into a 2 Mb contig. Therefore, the region that must contain the Ulnaless locus has been defined and cloned, which will be invaluable for the identification of the molecular nature of the Ulnaless mutation.

Mapping the midkine family of developmentally regulated signaling molecules

Midkine (Mdk) and heparin-binding neurotrophic factor (Hbnf)/pleiotrophin (Ptn) comprise the Midkine family of developmentally regulated signaling molecules. We have determined the chromosomal localization of these genes in the mouse by use of single-strand conformation polymorphisms (SSCPs), which facilitated the typing of Mdk and Hbnf alleles in recombinant inbred (RI) strains and interspecific backcrosses. Mapping was performed relative to other cloned genes, as well as simple sequence length polymorphisms (SSLPs) in the interspecific backcrosses. Mdk maps to mouse Chromosome (Chr) 2, linked to the Hoxd gene cluster. Hbnf maps to proximal mouse Chr 6, linked to the Cftr and Cpa genes. Comparative mapping of human MDK and HBNF employing species-specific polymerase chain reaction (PCR) primers and human monochromosomal somatic cell hybrids assigns MDK to human Chr 11 and HBNF to human Chr 7q32-qter.