Technical Comments

Response to Comment on “Local reorganization of xanthophores fine-tunes and colors the striped pattern of zebrafish”

See allHide authors and affiliations

Science  17 Apr 2015:
Vol. 348, Issue 6232, pp. 297
DOI: 10.1126/science.aaa2804


Watanabe and Kondo question our conclusion that the current Turing-type model of color patterning in zebrafish requires modification. In addition to xanthophores and melanophores, iridophores are essential for stripe formation in the body, although not in the fins. A model of predictive value should accommodate the in vivo dynamics and interactions of all three chromatophore types in body stripe formation.

It has been postulated that Turing-type mechanisms underlie color pattern formation in fish, and they can be generalized to explain vertebrate coloration (13). However, experimental analyses in several organisms (4, 5), including zebrafish (610), suggest these models to be oversimplifications with limited predictive value.

In zebrafish, three chromatophore types—xanthophores, iridophores, and melanophores—are required to form the striped pattern in the body, whereas only melanophores and xanthophores are required in the fins (6, 7, 11, 12). The importance of melanophore-xanthophore interactions over short and long ranges was identified in our and Parichy's laboratories (11, 12). Recently, iridophore-dependent interactions for the body patterning have been identified (6, 7). In the absence of iridophores (Fig. 1B), the shiny appearance of stripes is lost, the number of melanophores is much reduced, and the repetitive stripe formation does not occur beyond the first two stripes. In the absence of either melanophores or xanthophores, residual stripes are formed between iridophores and the remaining cell type (6). Iridophores and xanthophores support melanophores over a long range but repel at short range. Iridophores attract xanthophores and directly influence their survival, differentiation, and recruitment to the light stripes (Fig. 2) (6, 7). Although iridophores and xanthophores have qualitatively similar interactions with melanophores, only iridophores are capable of long-range dispersal in the skin, which is important in initiating new light stripes and for the formation of a repetitive pattern (6, 8). Xanthophores and melanophores only display short-scale movements during normal development (8, 9). This has recently been supported by an independent study indicating that iridophores regulate melanophore stripe position and width (13). However, the current Turing-model does not consider iridophores.

Fig. 1 Zebrafish pigmentation mutants and pattern variation in other Danio species.

(A) Wild-type and (B to F) mutant Danio rerio (zebrafish). (G) Danio margaritatus with striking difference in body and fin patterning. (H) Danio albolineatus. (I) F1 hybrid of zebrafish and Danio albolineatus: striped on the body, not on the fins.

Fig. 2 Scheme of interactions between chromatophores types.

Black circles, melanophores; I, iridophores; yellow circles, xanthophores; red curved arrows, long-range interactions; black arrows, short-range interactions. [Reproduced from (6)]

In the body, dark and light stripes are generated sequentially; formation of light stripes always precedes the formation of dark stripes (8). The first light-stripe region is formed following a morphological prepattern provided by the horizontal myoseptum, not by iridophores; choker mutants (Fig. 1C) lack the horizontal myoseptum and make meandering stripes (6). However, the stripes in choker are of normal width and composition, suggesting that further stripe formation is a self-organizing process, requiring the interaction between all three chromatophore types (6).

The role of iridophores is not to orient the stripe pattern; the residual stripes in mutants lacking iridophores are oriented normally. Iridophores emerge along the horizontal myoseptum, and, after forming the first light stripe, disperse along the dorsoventral axis as a loose net of cells. Subsequently, they undergo patterned aggregation to form new light stripes (8). This spatial transition in organization prefigures the formation of dark stripes by melanophores appearing in the skin in situ; in mutants lacking iridophores, no stripes are added to a basic pattern. In double mutants lacking melanophores and xanthophores, iridophores cover the whole skin as a dense sheet (Fig. 1D), indicating that the transition in iridophore organization depends on their interactions with melanophores and xanthophores (68). In the connexin mutants leopard/Cx41.8 and luchs/Cx39.4 (Fig. 1, E and F), the transition of iridophores from the dense to the loose state is compromised and iridophores disperse in a dense state, leading to frequent interruptions of the dark stripe and formation of melanophore spots (14). We do not present a different model in which iridophores provide a prepattern. Rather, we describe iridophores as an intrinsic part of the pattern itself; this quality must be part of a valid model.

The body, anal fin, and tailfin display a contiguous stripe pattern in zebrafish (Fig. 1A). Watanabe and Kondo have used this to argue for a core Turing-type network driven by melanophore-xanthophore interactions leading to stripe formation, because iridophores are not participating in the striping of the fins (15). But iridophores are essential in the body (Fig. 1B), which means that stripe formation in the fins share some but not all properties of the mechanism working in the trunk. The body and fin stripes, which appear similar, have in fact distinctly different underlying pigment cell arrangements (16). In many Danio species, closely related to zebrafish, patterning in all the fins is different from that of the body, and in some species several fins display a common motif, distinct from the body pattern (Fig. 1, G and H). Notably, only the zebrafish pattern suggests periodicity along the dorsoventral axis (10). Danio albolineatus does not display a stripe pattern (Fig. 1H). Whereas F1 hybrids between zebrafish and D. albolineatus are striped on the body, the fins remain unstriped (Fig. 1I). Several zebrafish mutants differently affect the stripe pattern on the body and on the fins, suggesting differences in the underlying mechanism (Fig. 1, B, C, and F) (6, 7, 10). luchs/Cx39.4 displays striped fins but a spotted pattern on the body like leopard/Cx41.8. (Fig. 1F) (14). Thus, the seemingly contiguous pattern in the body and in the fins appears to be a derived feature and cannot be taken as an argument in favor of a core mechanism underlying both body and fin patterning.

The Turing model successfully incorporates previously known interactions but excludes iridophore-dependent interactions, suggesting that it may be of limited predictive value for the patterning of the body and, more generally, for patterning in other fish species. In the light of cellular-genetic data, the model does not reflect the real conditions. Kondo’s studies have consistently put forward melanophore-xanthophore cell sorting from an initial uniform distribution as a major underlying cellular mechanism (3). Watanabe and Kondo argue that this is not a crucial feature of their Turing-type model, which can still explain all observations in retrospect. They claim that iridophores can readily be incorporated in the existing models. This is what we have suggested should be done.

It is important to build a theoretical framework (Turing or not) that takes the experimental observations into account and that is capable of generating testable hypotheses to improve the model and adapt it to explain color patterning not just in zebrafish but also in other vertebrates.

References and Notes

  1. Acknowledgments: This work was supported by the Max Planck Society.
View Abstract

Navigate This Article