|Group Leader:||Christiane Nüsslein-Volhard|
|Phone:||+49 7071 601-489|
|Fax:||+49 7071 601-384|
|Staff:||Alphabetical List | Alumni|
The zebrafish, Danio rerio, owing its name to the striking stereotypic pattern of horizontal blue and golden stripes, has emerged as the model organism for the genetic analysis of colour pattern formation in vertebrates. During the last years an ever increasing number of adult viable mutants with altered colour patterns have been collected, and novel approaches in lineage tracing in individual fish were developed, providing the unique opportunity to access the genetic and cell biological background of the complex and protracted developmental process in this species.
Most animals are coloured; colours not only protect them from harmful UV irradiation, but they also play important roles in inter- and intra-species communication. Colour patterns, often strikingly elaborate, serve as camouflage in predator-avoidance, they have roles in kin recognition and in mate selection. Colour patterns are highly variable and evolve rapidly, and as direct targets of natural and sexual selection they are of high evolutionary relevance. An understanding of the mechanisms that underlie pigmentation and colour pattern formation is an important step towards comprehending the evolution of bio-diversity.
The pigment-producing cells in vertebrates originate from the neural crest, a transient primordium of multi-potent cells located at the dorsal neuro-ectodermal ridge, from which progenitor cells migrate out into the periphery to develop a variety of structures and tissues. Neural crest evolved at the protochordate-vertebrate transition; it is a developmental innovation that allowed vertebrates to become both, large and colourful. Of highest evolutionary significance is the formation of a true head with neural crest-derived skull bones and jaw.
Using genetic techniques for long-term cell-labelling we discovered recently that the pillar cells of the gills, the respiratory organs of fish that predate the appearance of the skull, are also neural crest-derived; whereas the scales, dermal bone structures protecting the body, are not (Mongera et al, 2013, Mongera and Nüsslein-Volhard, 2014). Other neural crest-derived structures include the neurons of the peripheral nervous system, the glia, and the pigment cells.
Birds and mammals have only one pigment cell type, the melanocyte, producing the pigment melanin (although in two different types, brownish-black eumelanin and reddish pheomelanin) that is secreted into the skin or the integumentary appendages, feathers and hairs. In contrast, basal vertebrates such as fish, amphibia and reptiles develop several pigment cell types – the chromatophores - producing different colours.
Zebrafish develop two independent colour patterns: a larval pattern is formed directly from migrating cells of the neural crest. The adult pattern emerges, together with other adult structures such as fins and scales, during a period of about four weeks, called metamorphosis (Figure 1).
The striking stereotypic pattern of horizontal blue and golden stripes in zebrafish arises from a multi-layered arrangement of three different types of pigment cells below the dermis; black melanophores, yellow xanthophores and silvery/blue iridophores. Melanophores exist in the dark blue stripe in a compact extended shape, and fill the space leaving little room between the cells. Iridophores and xanthophores are present in both the light and dark stripes, albeit in different shapes: silvery-whitish dense iridophores form the light stripe, whereas in the dark stripes iridophores cover the melanophores as a net of loose cells appearing blue. Xanthophores are compact and roundish in the light stripes, whereas in the top layer of the dark stripe regions they are stellate and faint (Figure 2), (Singh et al, 2014, Mahalwar et al. 2014). The exact superposition of the different pigment cells is important for the contrast and brightness of the colour pattern.
The colour pattern of zebrafish is different in different body regions: the striped pattern is restricted to the lateral trunk and tail, the anal and the caudal fins. The pectoral, pelvic and the dorsal fins do not show a striped pattern. The dorsal and ventral regions of the fish also display different colouration with dark dorsal aspects and shiny ventral regions.
In closely related Danio species the adult colour patterns display conspicuous differences ranging from stripes with different width and orientation, over spots to no pattern at all. As in zebrafish, the patterns are different in the body regions, but unlike zebrafish, the dorsal, anal and tail fins often display a common theme, quite distinct from that of the body (Figure 3). Remarkably, the larval patterns in all these different Danio species are quite similar.
Genetic screens have identified a number of mutations leading to viable fish that display changes in the adult colour pattern. According to the mutant phenotypes, three major classes of genes may be distinguished:
1. Genes required for the formation of one chromatophore type.
They affect only one type of chromatophore directly, however other cell types are indirectly impaired. Kit signalling is required for the establishment and the survival of embryonic and early metamorphic melanophore progenitors as suggested from the analyses of sparse and sparse-like. Melanophores require the transcription factor MitfA for differentiation; they are absent in nacre/mitfA mutants. Iridophores cell-autonomously require Leukocyte tyrosine kinase for iridophore-fate specification as suggested by the analysis of shady/ltk, and they require rose/Ednrb endothelin signals for expansion of their population in the skin. Xanthophores are absent in pfeffer/Csf1ra mutants. (Figure 4)
The analysis of these mutants has shown that the establishment of the stable stripe pattern in adult fish depends on positive and negative interactions within and between all three pigment cell types, which means that pigment cells do not independently fill in an invisible pre-pattern. Strikingly, iridophores are not participating in the striping of the fins, supporting the notion that the mechanism of stripe formation differs in fins and body (Maderspacher and Nüsslein-Volhard, 2003, Frohnhöfer et al. 2013, Krauss et al. 2013).
2. Genes affecting the spatial arrangement of all three chromatophore types:
Several mutants have been identified in which all three pigment cell types are present, but the stripes are interrupted, wavy or display abnormal width. These phenotypes indicate that the proteins involved are mediating the interactions between the three pigment cell types. Dominant and recessive leopard/ Connexin41.8 (leo) mutations lead to a spotted pattern in which light stripe regions invade the dark stripe regions. Connexins are subunits of gap junctions, channels that mediate the chemical and electrical coupling of neighbouring cells. Recently we identified luchs (luc), a second connexin gene with a spotted phenotype. Both leo and luc are required in melanophores and xanthophores, however they are not required in iridophores. The two connexins form heteromeric gap junctions in and between xanthophores and melanophores resulting in the instruction of iridophores to undergo a transition from the compact into the loose form characteristic for the dark stripe in an as yet unknown mechanism (Irion et al. 2014). (Figure 5)
We also collected mutants affecting stripe width: obelix mutants display broader and fewer stripes in heterozygous as well as homozygous condition. The gene encodes an inwardly-rectifying potassium channel (Kir7.1), which is required in melanophores for their communication with xanthophores. Other mutants affecting stripe width, currently being studied in our laboratory, are asterix, idefix and autobahn. (Figure 6)
3. Genes affecting the cellular environment of chromatophores:
In some mutants apparently not the pigment cells themselves but their cellular environment is affected. An example is karneol (kar), encoding Ece2, an endothelin-converting enzyme (Krauss et al. 2014). We recently identified Endothelin-3B as the ligand involved in rose/Ednrb1b signalling, that is presumably target of kar/Ece2. Endothelin-3 is an important regulator of animal colour patterns and has been suggested to coordinate localised colour differences in the skin of cheetah. However, we do not yet know the sites of expression of Endothelin-3 and other ligands such as slk/KitlgA. Other examples of genes not required in pigment cells but affecting the pattern are mau, encoding Aquaporin3a, a plasma membrane water channel (A. Eskova, unpublished) and nepomuk. It will be important to explore the as yet unclear role of the tissue environment in colour pattern formation.
The striped pattern of the adult zebrafish arises during a metamorphic period that begins approx. 3 weeks post fertilisation and lasts for about one month. During this period newly formed pigment cells emerge in the skin to generate the stripes on the flanks and in the anal- and tail fins. A big challenge for many years has been the problem to trace individual cells and their progeny from the embryo until the differentiation of the final pattern in the adult. We recently pioneered the analysis of Cre-loxP induced fluorescently labelled clones in metamorphic zebrafish to monitor the dynamics of cell divisions, cell spreading and cell shape changes during the period of stripe formation.
Our studies have shown that metamorphic iridophores, as well as melanophores, develop from segmentally organized stem cells located at the dorsal root ganglia (Figure 7) whereas adult xanthophores originate from larval xanthophores, which persist throughout larval stages and begin to proliferate in the growing skin at the onset of metamorphosis (Dooley et al. 2013, Singh et al, 2014).
Iridophores take a lead in stripe formation; they migrate through the horizontal myoseptum, which serves as a morphological pre-pattern, into the skin where they form the first light stripe of dense iridophores (Figure 8).
The iridophores continue to proliferate and spread bi-directionally as loose cells in the skin towards the dorsal and ventral aspect of the juvenile fish. They form a coherent sheet of cells covering the flank of the fish, in which additional light stripes appear by patterned aggregation of iridophores into the dense form at a given distance from the first light stripe.
Melanoblasts reach the skin at the presumptive dark stripe regions via the peripheral nerves dorsal and ventral to the first light stripe. In contrast to iridophores, melanophores do not proliferate but dramatically expand in size to form a contiguous sheet of cells forming the dark stripe. Hence, the dark and light stripes are formed by completely different cellular behaviours (Singh et al. 2014).
Prior to the emergence of iridophores, which form a contiguous middle layer, xanthophores evenly cover the body of metamorphic fish as a top layer; they adjust in shape to the incoming iridophores and melanophores (Figure 9) (Mahalwar et al. 2014). Our present view is that the transition between the loose and dense forms of iridophores pre-figures the formation of the striped pattern, which is further refined by short range interactions between all three cell types. Pattern variations are caused by changes in the spatial parameters regulating this transition in shape. (Figure 10)
April Dinwiddie, Postdoctoral fellow
Anastasia Eskova, Postdoctoral fellow
Gurpreet Kaur, Postdoctoral fellow
Hans Georg Frohnhöfer, Staff Scientist
Uwe Irion, Project leader
Christiane Nüsslein-Volhard, Research Group Leader