QUICK SEARCH:   [advanced] Author: Keyword(s):
Home     Help     Feedback     Subscriptions     Archive     Search     Table of Contents

First published online 10 November 2004
doi: 10.1242/dev.01526

This Article Summary Full Text Full Text (PDF) Supplementary Material Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Quigley, I. K. Articles by Parichy, D. M. Search for Related Content PubMed PubMed Citation Articles by Quigley, I. K. Articles by Parichy, D. M.

Pigment pattern evolution by differential deployment of neural crest and post-embryonic melanophore lineages in Danio fishes

Section of Integrative Biology, Section of Molecular, Cell and Developmental Biology, Institute for Cellular and Molecular Biology, University of Texas at Austin, 1 University Station C0930, Austin, TX 78712, USA

View larger version (89K): [in a new window]   Fig. 1. Diverse pigment patterns of Danio fishes and their relatives. D. rerio, D. nigrofasciatus and D. kyathit all exhibit well-defined horizontal stripes of melanophores, although D. kyathit adults develop breaks and irregularities in the stripes both anteriorly (shown) and posteriorly. D. kerri exhibits a few broad and relatively diffuse melanophore stripes. D. albolineatus adults lack horizontal stripes and melanophores are evenly distributed, although larvae exhibit a transient pattern of stripes posteriorly. D. choprae have vertical bars of melanophores as adults and transient horizontal stripes at earlier stages. Devario (formerly Danio) shanensis adults have vertical bars of melanophores anteriorly, with horizontal stripes posteriorly. Tanichthys albonubes exhibit a narrow horizontal melanophore stripe with a broader pattern of evenly dispersed melanophores. Yellow coloration in fish derives from neural crest-derived xanthophores, whereas red pigment in some species derives from neural crest-derived erythrophores. Adult fish are 25-30 mm standard length.

View larger version (98K): [in a new window]   Fig. 2. Similar stripes with different melanophore numbers and colors in D. rerio and D. nigrofasciatus juveniles. (A,B) D. rerio; (C,D) D. nigrofasciatus. (A) The ventral primary melanophore stripe of D. rerio consists of numerous gray-black metamorphic melanophores (arrow). Melanophores persisting from embryonic stages at the site of the early larval ventral melanophore stripe (arrowhead) are browner than metamorphic melanophores. (B) In the D. rerio dorsal primary melanophore stripe, a few melanophores at the ventral edge are brown in color (arrowhead), where a few melanophores are derived from the early larval stripe pattern. (A',B') Schematics of fish shown in A and B, showing black and brown melanophores. No attempt is made to precisely delineate individual melanophore boundaries. (C) Unlike in D. rerio, the ventral primary melanophore stripe of D. nigrofasciatus includes numerous brown melanophores (arrowhead), in addition to black melanophores (arrow). (D) Detail showing D. nigrofasciatus brown melanophores (arrowhead) and black melanophores in the ventral primary melanophore stripe. (C',D') Schematics of fish shown in C and D.

View larger version (87K): [in a new window]   Fig. 3. Pigment pattern metamorphosis differs between D. rerio (A-E') and D. nigrofasciatus (F-J'). Panels shown are of selected days from a complete image series for individual, representative larvae. In A and F, the sites of early larval melanophore stripes are indicated at the dorsal and ventral margins of the myotomes (horizontal arrowheads), and at the horizontal myoseptum (squares). In D. rerio, pigment pattern metamorphosis begins with the differentiation of pioneer metamorphic melanophores over the ventral myotomes (A, arrow), with additional metamorphic melanophores (B, arrow) appearing both dorsally and ventrally over a period of several days (B-D). Adult primary stripes become evident as dispersed melanophores migrate to sites of stripe formation and additional metamorphic melanophores differentiate within the stripes (D,E). A few early larval melanophores move from the horizontal myoseptum to join the adult dorsal primary melanophore stripe (C-E, arrowheads). Near the end of pigment pattern metamorphosis the larvae have developed an adult dorsal primary melanophore stripe and an adult ventral primary melanophore stripe (1D, 1V, respectively, in panel E). The adult ventral primary melanophore stripe develops just ventral to the level of the aorta (a, in panel E), about halfway between the horizontal myoseptum and the ventral margin of the myotomes (E'). In D. nigrofasciatus, pigment pattern metamorphosis begins with early larval melanophores becoming displaced from the larval stripes (F, melanophores 1-6). Whereas some metamorphic melanophores differentiate de novo (G,H, arrows), these are markedly fewer than in D. rerio. As metamorphosis proceeds, melanophores initially present in the ventral early larval stripe (H, arrowheads) become increasingly distant from the ventral margin of the myotomes. By late stages of pigment pattern metamorphosis, a complete adult ventral primary melanophore stripe has formed (J), and both dorsal and ventral stripes contain numerous early larval melanophores. The D. nigrofasciatus ventral primary melanophore stripe develops further ventrally relative to the level of the aorta (a, J), and closer to the ventral margin of the myotome (J'), compared to D. rerio. Inset (J) shows brownish cast of an adult stripe melanophore (6) that originated in the early larval stripe. (E',J') Schematics of fish shown in E and J, showing melanophores associated with the adult primary melanophore stripes, and residual melanophores from the early larval stripes dorsally and ventrally, as determined by following individual melanophores from early larval stages throughout the image series (i.e. by analyzing cell lineage rather than by examination of final melanophore colors). For consistency with Fig. 2, melanophores that originated in the early larval pattern are shown in brown, and melanophores that differentiated during metamorphosis are shown in black. Dorsal metamorphic melanophores that will cover the dorsum and dorsal scales are omitted for clarity. Double arrowheads in B and F indicate deep, internal melanophores that are ventral to the notochord, or dorsal to the neural tube, respectively, and that do not contribute to pigment patterns beneath the skin. Standard lengths of larvae (mm): A, 6.7; B, 7.4; C, 8.6; D, 10.3; E, 11.5; F, 6.7; G, 7.1; H, 8.6; I, 9.4; J, 10.3.

View larger version (36K): [in a new window]   Fig. 4. Migration of melanophores during D. nigrofasciatus pigment pattern metamorphosis. Early larval melanophores (1, 2, 4, 5) and newly differentiating metamorphic melanophores (3) change positions as stripes form. Some changes in position are likely to reflect passive movements due to growth (e.g. increasing dorsal-ventral separation of melanophores 4 and 5), whereas others can be explained only by active rearrangements (e.g. relative dorsal-ventral positions of melanophores 1 and 2). Only selected days from the complete image series are shown. To maintain the same region of interest, images in this figure and in Figs 6, 9 and 12 are rescaled across days; exact sizes and stages are available on request.

View larger version (22K): [in a new window]   Fig. 5. Phylogenetic relationships of danios inferred from 12S and 16S rDNA sequences. Shown is a maximum-likelihood tree (branch lengths proportional to estimated divergence). Support values are percentages from Bayesian analysis followed by nonparametric bootstrapping. Taxa in red and green were chosen for analyses of pigment pattern metamorphosis based on phylogenetic position and embryo availability. The analysis supports the division of the danios into two genera, Danio and Devario, based on morphological criteria (Fang, 2003), and is in general agreement with previous molecular analyses of fewer taxa (Zardoya et al., 1996; Parichy and Johnson, 2001). GenBank Accession numbers for 12S and 16S sequences are (top to bottom): AY707450, AY707456; U21372, U21381; AF3226h58, AF322663; AY707446, AY707452; AF322663; AY707449, AY707455; AY707447, AY707453; AF322656, AF322661; U21376, U21384; AF322659, AF322664; U21377, U21377; U21375, U21370; AY707448, AY707454; U21553, U21554; AF322660, AF322665; U21378, U21386; AY707445, AY707451. Published sequences for D. aff. tweediei and D. pulcher were excluded owing to their limited length and the resulting loss of phylogenetic resolution.

View larger version (128K): [in a new window]   Fig. 6. Primary role for metamorphic melanophores in adult pigment pattern formation across species. Shown are repeated images of the same region of the ventral flanks in representative individuals of D. nigrofasciatus, D. rerio, D. kyathit, D. kerri, D. albolineatus, D. choprae, and T. albonubes (compare with Fig. 1). Only selected images are shown from the complete series for each individual. Row 1, shortly after the onset of pigment pattern metamorphosis in each species. Row 8, terminal stages of pigment pattern metamorphosis when the adult pigment patterns have formed; row 8', schematics showing melanophores present at early larval stages (brown) and melanophores that differentiated during metamorphosis (black), as revealed by tracing individual melanophores throughout pigment pattern metamorphosis. Squares indicate the horizontal myoseptum; horizontal arrowheads indicate the ventral aspect of the myotome. In D. nigrofasciatus, numerous early larval melanophores relocate (arrowheads) from the early larval stripe along the ventral aspect of the myotome (horizontal arrowhead, row 1) to the adult ventral primary melanophore stripe on the flank (row 8,8'). In D. rerio and D. kyathit, early larval melanophores typically do not contribute to the compact stripes of the adult. In D. kerri, a more diffuse stripe pattern arises compared with in D. nigrofasciatus, D. rerio and D. kyathit; although a few early larval melanophores leave their initial positions (arrowheads, row 5), they typically do not enter into the adult stripes. In D. albolineatus, rare early larval melanophores leave the larval stripes (arrowhead, row 6) but do not contribute substantially to the uniformly dispersed anterior melanophores or weak melanophore stripes posteriorly. In D. choprae, a few early larval melanophores leave the larval stripes (arrowheads, row 7) but do not join the horizontal adult stripes that form during metamorphosis, or the vertical barring pattern that develops at later stages; the same early larval melanophore behaviors are seen in the vertically striped D. shanensis (I.K.Q. and D.M.P., unpublished). Finally, in T. albonubes, a few early larval melanophores (arrowheads, row 6) leave the larval stripes but do not move far onto the flank where diffuse horizontal adult stripes develop in the adult. In all panels, larvae were imaged at a 30° angle to better visualize the ventral-lateral margin of the flank and the early larval melanophores, and images are rescaled to show the same region of the flank. Slight differences in starting pigment patterns (row 1) principally reflect inter-individual variation and minor differences in developmental stage. nigrof, D. nigrofasciatus; alb, D. albolineatus; T. alb, T. albonubes. Number of larvae examined: D. nigrofasciatus, 10; D. rerio, 5; D. kyathit, 5; D. kerri, 2; D. albolineatus, 4; D. choprae, 2; T. tanichthys, 4. Overall contributions of embryonic neural crest-derived melanophores and metamorphic melanophores are similar in other regions of developing adult pigment patterns (data not shown).

View larger version (34K): [in a new window]   Fig. 7. Different fates of early larval melanophores, and variation in adult melanophore origins across species. All values are means±s.e.m. (A) Total numbers of early larval melanophores differ somewhat across species (F6,25=4.44, P<0.005); black letters within bars indicate post-hoc Tukey comparisons of means and bars sharing the same letter do not differ significantly, thus only D. nigrofasciatus and D. kyathit differ significantly from one another. Total early larval melanophores for each species comprise melanophores that remain in the early larval pigment pattern during metamorphosis (light gray), and melanophores that leave the early larval stripes during metamorphosis and localize further laterally over the flank (brown). Different proportions of early larval melanophores leave the adult stripes in the different species (arcsine transformed proportions, F6,25=41.88, P<0.0001). However, post-hoc means comparisons of numbers and proportions indicate that D. nigrofasciatus alone differs significantly from other species (brown letters within bars). (B) Pigment patterns after metamorphosis differ markedly in total melanophore numbers across species (F6,25=18.93, P<0.0001). In all species, a majority of melanophores in the adult pigment pattern are metamorphic melanophores. Numbers of early larval neural crest-derived melanophores in the adult pattern are the same as in A. In adult pigment patterns, the proportions of early larval melanophores to metamorphic melanophores differ significantly among species (arcsine transformed proportions, F6,25=54.56, P<0.0001), yet only D. nigrofasciatus differs significantly from other species in post-hoc means comparisons. nigrof, D. nigrofasciatus; kyath, D. kyathit; alb, D. albolineatus; T. alb, T. albonubes.

View larger version (65K): [in a new window]   Fig. 8. Adult hybrid phenotypes exclude major-effect roles for genes isolated as D. rerio melanophore mutants. (A) Control (wild-type) D. rerioxD. nigrofasciatus hybrids develop adult dorsal and ventral primary melanophore stripes (large arrow) with melanophore numbers similar to those of D. rerio, but with fewer total stripes and fewer melanophores in the secondary melanophore stripes (small arrow) that develop as the fish grow (Parichy and Johnson, 2001). (B) Detail of dorsal primary melanophore stripe. (C) Tester hybrid for the tfap2a (lockjaw) mutant, a sibling to the hybrid in A. Despite the absence of melanophores in tfap2a mutant D. rerio (Knight et al., 2004), tester hybrids have as many melanophores as control hybrids, suggesting that tfap2 does not contribute substantially to the different numbers of melanophores between wild-type D. rerio and D. nigrofasciatus. Minor individual variation in secondary melanophore and stripe numbers does not segregate with tfap2 alleles (data not shown). (D) Detail of dorsal primary melanophore stripe in tester hybrid, showing a similar number of melanophores to the control in B.

View larger version (100K): [in a new window]   Fig. 9. Danio rerio mutants exclude and identify pathways for evolutionary changes in D. nigrofasciatus. Shown are selected images of representative larvae that were imaged throughout pigment pattern metamorphosis. Schematics (bottom row) illustrate the locations of early larval melanophores (brown) and metamorphic melanophores (black), as determined by the tracing of individual cells from the early larval pigment pattern into the adult pigment pattern (dorsal scale-associated melanophores are omitted for clarity). (A-D) ednrb1 mutant D. rerio develop an adult pattern of stripe and spots, superficially similar to D. nigrofasciatus (Parichy and Johnson, 2001). Nevertheless, the underlying mode of pigment pattern metamorphosis differs from D. nigrofasciatus, as few early larval melanophores contribute to the developing adult stripes. Arrow in A indicates a newly differentiated metamorphic melanophore. (E-H) puma mutant D. rerio exhibit a severe reduction in metamorphic melanophore numbers, whereas early larval melanophores (arrowheads) spread laterally over the flank, similar to D. nigrofasciatus. (I,J) picasso mutant D. rerio also have fewer metamorphic melanophores, and increased persistence of early larval melanophores (arrowheads). (M-P) Hybrids between D. rerio and D. nigrofasciatus exhibit fewer metamorphic melanophores than D. rerio, yet early larval melanophores only rarely contribute to the adult stripes (arrowheads), similar to D. rerio but unlike D. nigrofasciatus (compare with Fig. 3E,J). A few early larval melanophores at the horizontal myoseptum persist into the adult pigment pattern (as in D. rerio), but early larval melanophores along the ventral myoseptum typically do not join the developing adult ventral primary melanophore stripe (as in D. rerio, but unlike D. nigrofasciatus). (Q-T) Hybrids between puma mutant D. rerio and D. nigrofasciatus exhibit early larval melanophore behaviors similar to those seen in D. nigrofasciatus. Although a few early larval melanophores leave their initial positions in control hybrids (P'), these cells are increased in number in puma tester hybrids, particularly among melanophores in the vicinity of the anal fin (arrowheads, T). Sites of early larval melanophore stripes are indicated at the dorsal and ventral margins of the myotomes (horizontal arrowheads), and at the horizontal myoseptum (squares) in A, E, I, M and Q.

View larger version (133K): [in a new window]   Fig. 10. Fewer metamorphic melanophore precursors in D. nigrofasciatus revealed by in situ hybridization for the melanoblast markers dct (A-F) and tyr (G-L). (A) During the early stages of pigment pattern metamorphosis in D. rerio (e.g. Fig. 3B), primary metamorphic melanophores (arrowhead) differentiate over the middle of each ventral myotome; these and a few unmelanized cells (arrow) stain for dct. (B) The corresponding region in D. nigrofasciatus is devoid of primary metamorphic melanophores and unmelanized dct+ cells, although the early larval melanophores located further ventrally are dct+ (arrowhead). (C) At middle metamorphic stages in D. rerio (e.g. Fig. 3C,D), unmelanized (arrow) and melanized dct+ cells are abundant over the ventral myotome in the vicinity of the ventral primary melanophore stripe. (D) In the corresponding region of D. nigrofasciatus, only melanized cells express detectable levels of dct, even after overdevelopment (data not shown). (E) At middle metamorphic stages in D. rerio, unmelanized dct+ melanoblasts (arrow) are abundant in the vicinity of the dorsal primary melanophore stripe. (F) In D. nigrofasciatus, unmelanized dct+ cells (arrow) are infrequent compared with D. rerio, although a few are present and differentiate as melanophores (arrowhead), showing melanin in addition to dct staining. (G-L) Staining for tyr expression is similar to staining for dct. Shown are similar stages and positions to the corresponding panels in A-F. Scale bars: 100 µm for A,B,G,H; 40 µm for C-F,I-L.

View larger version (106K): [in a new window]   Fig. 11. L-dopa staining for tyrosinase activity reveals fewer melanoblasts in D. nigrofasciatus compared with in D. rerio. Shown are larvae during middle stages of pigment pattern metamorphosis before (A,C) and after (B,D) incubation with L-dopa. (A,B) In D. rerio, melanoblasts are revealed by new melanin deposition (arrowheads show locations of cells before and after incubation). (C,D) In D. nigrofasciatus, few melanoblasts are revealed in general, and no new cells are observed in the region shown.

View larger version (77K): [in a new window]   Fig. 12. Non-autonomous factors underlying the differences in pigment pattern metamorphosis between D. rerio and D. nigrofasciatus, revealed by interspecific genetic mosaic analyses. Shown are selected days in the development of two representative chimeras (n=10), taken from a complete image series through pigment pattern metamorphosis. (A-D) D. nigrofasciatus cells transplanted into albino mutant D. rerio. Melanized donor melanophores differentiate at embryonic stages within the early larval melanophore stripes (arrowheads, A). Yet these donor melanophores fail to contribute to the ventral primary melanophore stripe, as for host melanophores. Subsequently, donor metamorphic melanophores differentiate over the flank and contribute to adult primary melanophore stripes located at positions that are indistinguishable from host stripes. Arrow in D marks the primary ventral melanophore stripe (a, aorta; compare with Fig. 3E). (F-I) D. nigrofasciatus cells transplanted to nacre mutant D. rerio. Despite the absence of host melanophores, donor early larval melanophores still fail to contribute to the ventral primary melanophore stripe, which forms in the normal position for D. rerio (arrows, G-I). In this individual, a secondary adult melanophore stripe comprising late-appearing metamorphic melanophores has started to form ventrally (small arrow, I). Schematics (D',I') illustrate the locations of early larval melanophores (brown) and metamorphic melanophores (black), as revealed by following individual cells throughout development. Dorsal scale-associated melanophores are omitted for clarity.

View larger version (15K): [in a new window]   Fig. 13. Model for the development of early larval and adult neural crest derivatives. In embryos, neural crest (nc) cells develop into a variety of derivatives, including early larval glial cells (ELG) and early larval melanophores (ELM). Some cells are set aside as post-embryonic stem cells or specified precursors (m) that will be recruited to differentiate during metamorphosis. During the larval-to-adult transformation, metamorphic glia (MG) (Parichy et al., 2003) and metamorphic melanophores (MM) differentiate from latent precursors, with melanophores passing through a melanoblast intermediate (MB, expressing the melanophore lineage markers dct and tyr; gray circle). Additionally some early larval melanophores can transit the metamorphic boundary and persist into the adult pigment pattern, although relatively few do so in D. rerio or in most of the other species examined in this study. In D. nigrofasciatus, there has been a reduction (right red arrow) in metamorphic melanophores owing to an early block in this lineage, possibly associated with a puma-dependent pathway; concomitantly, there has been an increase (left red arrow) in the number of early larval melanophores persisting into the adult pigment pattern.