Klann et al. EvoDevo (2021) 12:8 https://doi.org/10.1186/s13227-021-00178-x EvoDevo

REVIEW Open Access Variation on a theme: pigmentation variants and mutants of anemonefsh Marleen Klann1, Manon Mercader1, Lilian Carlu1, Kina Hayashi1,2,3, James Davis Reimer3,4 and Vincent Laudet1,5*

Abstract Pigmentation patterning systems are of great interest to understand how changes in developmental mechanisms can lead to a wide variety of patterns. These patterns are often conspicuous, but their origins remain elusive for many marine fsh species. Dismantling a biological system allows a better understanding of the required components and the deciphering of how such complex systems are established and function. Valuable information can be obtained from detailed analyses and comparisons of pigmentation patterns of mutants and/or variants from normal pat‑ terns. Anemonefshes have been popular marine fsh in aquaculture for many years, which has led to the isolation of several mutant lines, and in particular color alterations, that have become very popular in the pet trade. Additionally, scattered information about naturally occurring aberrant anemonefsh is available on various websites and image platforms. In this review, the available information on anemonefsh color pattern alterations has been gathered and compiled in order to characterize and compare diferent mutations. With the global picture of anemonefsh mutants and variants emerging from this, such as presence or absence of certain phenotypes, information on the patterning system itself can be gained. Keywords: Pigmentation, Anemonefsh, Variation, Mutants

Introduction erythrophores that are distinguished by color and con- Body coloration, or pigmentation, plays an essential part tain carotenoids and/or pteridines; (3) silver/iridescent in every animal’s survival strategy. Pigmentation is not iridophores which contain guanine crystals, and (4) white only important for predator avoidance or protection leucophores which bear uric acid. against UV radiation, but also for reproductive success Typically, chromatophore precursor/progenitor cells and more generally social interactions [10, 21, 33]. In fsh, are established during embryonic development and set skin pigmentation depends on the distribution of pig- aside to later on contribute larval/adult chromatophores. ment cells, also called chromatophores, which are deriva- In zebrafsh (Danio rerio), currently the major model spe- tives of the neural crest, and are typically classifed based cies for pigment pattern development in teleost fsh, two on the colored pigments or types of crystals they bear as distinct populations of precursors are formed, one for well as their ultrastructure [15]. Te major teleost chro- xanthophores and one for melano- and iridophores [5]. matophores are (1) brown blackish melanophores that In medaka (Oryzias latipes), the xanthophore precursor contain melanin; (2) yellow/orange/red xanthophores/ is still multipotent and can produce either xanthophores or closely related leucophores [40]. In zebrafsh, chroma- tophores must be formed continuously throughout their *Correspondence: [email protected] life to maintain their characteristic striped pattern [25, 1 Marine Eco‑Evo‑Devo Unit, Okinawa Institute of Science 45]. and Technology, 1919‑1 Tancha, Onna‑son, Okinawa 904‑0495, Japan Full list of author information is available at the end of the article

© The Author(s) 2021. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat​ iveco​ ​ mmons.org/​ licen​ ses/​ by/4.​ 0/​ . The Creative Commons Public Domain Dedication waiver (http://creat​ iveco​ mmons.​ org/​ publi​ cdoma​ in/​ ​ zero/1.0/​ ) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Klann et al. EvoDevo (2021) 12:8 Page 2 of 16

Patterns in general can be formed by two main princi- were observed and captured in the wild. Other mutants ples: (1) cell autonomously by (genetic) pre-positioning are the result of aquaculture crosses, either made on pur- cues or (2) by cell–cell communication [46]. Te mech- pose or by accident. In this review, we aim to describe anisms for stripe formation (horizontal main axis) in the main anemonefsh mutants available in addition to zebrafsh are relatively well understood. All three chro- natural variants, as they could be signifcant in provid- matophore cell types (melanophores, xanthophores and ing clues on the underlying genetic mechanisms of pig- iridophores, no leucophores are present in zebrafsh) ment pattern formation in these fsh. Here, the term interact and communicate to establish the adult color “mutant” therefore refers to an established line from an pattern of dark stripes and bright interstripes. Some evi- aquaculture company, with the strain name written in dence suggests that this mode of patterning follows the parentheses. For these lines it is known that the observed Turing model, according to which the number of stripes pigmentation phenotype is based on a fxed genetic trait, increases with the growth of the fsh [41]. Interactions or in other words, a mutation. Breeding schemes and out- between zebrafsh chromatophores include long-range as comes detailed and provided on aquaculture companies’ well as short-range efects [14, 46] such as: (1) iridophores websites can be used to infer heritability, although this attract melanophores in high numbers and induce their will need to be genetically confrmed in the future. Te aggregation into prospective stripe areas (positive long- term “variant” is used for wild individuals that display a range); (2) iridophores exclude melanophores from inter- coloration alteration, as it generally remains unknown stripes (negative short-range), (3) xanthophores exclude if the phenotype is based on a fxed, inheritable muta- melanophores from interstripes (negative short-range), tion. However, often these wild variants are based on a and (4) positive and negative interactions between irido- mutation, since wild-caught aberrant individuals have phores and xanthophores being required to confne the served on several occasions as founders in the establish- shape of interstripes (short range). Similarly, the addition ment of new strains in aquaculture (as mentioned above). of bars (vertical main axis) in the cichlid Copadichromis Although there are some polymorphic traits in several azureus requires cell–cell communication, with xantho- species of anemonefsh (numbers of bars, degree of mela- phores and iridophores following the pattern of melano- nism, bar regression with age), these will not be the focus phore distribution [18]. of the present paper as these are complex situations that Extensive zebrafsh and medaka mutant collections are often related to environmentally driven phenotypic have facilitated research on many aspects of pigmenta- plasticity. Studies of several of these polymorphic traits tion, such as fate specifcation, survival, proliferation, are currently underway in our laboratory. and diferentiation [23, 24, 46]. In both taxa, research- ers have mainly focused on embryonic phenotypes, and Color pattern formation in anemonefshes there is an extensive overlap of identifed parallel phe- Anemonefshes (subfamily Amphiprioninae) are a mono- notypes. Te strategy of using mutants identifed during phyletic group within the Pomacentridae (damselfsh) large genetic screenings and then subsequently identify- family, comprising 30 species in two genera: Amphiprion ing the genes underlying their unusual phenotypes has (29 species) and Premnas (monospecifc) [54]. Tey are therefore been very fruitful. However, such a strategy well known for their ability to establish mutualistic sym- has been limited to established model organisms that are biotic relationships with sea anemones and their brilliant well set up in terms of mutagenesis, genetic screening, color patterns have been proposed to be either a social and experimental genetics [26, 47]. Although research on display allowing species recognition [58] or an apose- both zebrafsh and medaka have efectively led to great matic trait warning of harmful stings from their host [36]. scientifc breakthroughs in our understanding of pig- Adult anemonefshes display a relatively simple and ment pattern formation, they represent only a small frac- robust color pattern comprising zero to three vertical tion of living biodiversity, and therefore new Evo-Devo white bars usually with a black outline on an orange- model species have been developed in the past few years. to-red body (Fig. 1A). Interspecifc variations manifest Among these newer models, anemonefshes are color- mainly in the number of white bars, as well as the color ful coral reef fsh that ofer great promise to enlarge and of the body and the fns [31, 52, 58]. Tis characteristic deepen our knowledge of color pattern formation [54– color pattern has been classifed into four diferent cat- 56]. Anemonefsh are relatively new marine model organ- egories according to the appearances of adults: (1) spe- isms and therefore scientifc mutant screenings have not cies without bars (Fig. 1B); (2) species with one bar on yet been performed, but pigmentation mutants are avail- the head (Fig. 1C); (3) species with two bars, one on the able thanks to aquaculture companies that aim to sell head and one on the body (Fig. 1D), and (4) species with novel “fancy” fsh for the aquarium market. Some of these three bars, one on the head, one on the body and one on anemonefsh mutants are derived from rare variants that the caudal peduncle/tail (Fig. 1E) [58]. Moreover, a few Klann et al. EvoDevo (2021) 12:8 Page 3 of 16

species of anemonefsh show pronounced polymorphism of the prospective bar and push back the melanophores (regarding the numbers of bars as well as background present in this area so that they create a black edge [58]. color), and four species exhibit a long dorsal stripe However, since TAE 684-treated larvae possess melano- (Fig. 1F) [58]. Tis diversifcation can be traced back to phores (intermingled with xanthophores in the orange- successive evolutionary losses of bars from posterior to colored areas and forming a normal pattern on the fns) anterior. Bar formation in various anemonefsh species general development and survival is evidently not nega- is clearly developmentally constrained, and no species tively afected by the lack of iridophores. exhibit a single bar on the body or tail only, and there Even though iridophores often provide silver/bluish- is always either a head bar only, or head and body bars, greenish refective color, in anemonefshes they mainly respectively. Moreover, there is an anterior to posterior contribute white. Tis is due to the diferent arrange- sequence of bar acquisition during larval/juvenile devel- ment of the guanine crystals present. When crystals are opment, but some anemonefsh species subsequently lose aggregated in platelets that are precisely organized, they the posterior-most bar(s) during adult color maturation give of an iridescent color, while they appear whitish [58]. We have recently shown that the adult pigmentation when the crystal platelets are less organized as incom- pattern of A. ocellaris and A. percula are, as in zebrafsh ing light is scattered in various directions [15, 59]. In A. [35], under the control of thyroid hormones, but with the ocellaris the majority of iridophores contained plate- distinct diference that in both anemonefshes the main lets that are arranged parallel to one another in discrete target of these hormones seems to be the iridophores stacks organized concentrically around the nucleus [55, responsible for white bar formation [57]. All this suggests 56]. In zebrafsh it has been recently shown that there are that there is an underlying robust patterning system that diferent subtypes of iridophores that are part of stripes couples bar formation with antero-posterior cues, but the and interstripes [17]. Our own recent observations in nature of this patterning system is currently unknown. anemonefsh also suggest that distinct types of irido- As for other fsh, the larval pigmentation pattern of phores may exist, although their respective distributions anemonefshes (comprising one or two horizontal stripes and functions in the fnal pattern are not yet clear (M. of single stellate melanophores) is very diferent from Miyake, M.K and V.L, unpublished observation). the adult pattern. Because the adult bar pattern does not change or adjust with size but follows an anterior–poste- Aquaculture of anemonefshes rior sequence [58], it is highly unlikely that it develops in Anemonefshes comprise a large proportion of the orna- a Turing-like fashion. For A. ocellaris it has been shown mental fsh trade, and can be wild-caught or captive- that three chromatophore subtypes contribute to the bred. Many species reproduce regularly in captivity and mature body coloration: (1) orange skin color results from survival rates are relatively high when rearing and feed- xanthophores with interspersed round melanophores; (2) ing are optimized [44, 53]. A. ocellaris is among the fve white skin comprises iridophores and interspersed stel- most-imported ornamental fsh species in the US [49]. lar melanophores, and (3) black skin is made of a dense As with most domesticated animals, selective breeding number of round melanophores [55, 56, 58]. Iridophores eforts have been used to establish diferent strains for are not only responsible for providing the white color but ornamental purposes. Several aquaculture companies seem to be necessary for the correct positioning of mel- have focused on anemonefsh breeding and have pro- anophores as well. Treatment of A. ocellaris juveniles duced various color mutations/alterations (see below). with TAE 684 (an inhibitor of tyrosine kinase recep- Due to the higher survival rate of anemonefsh larvae in tors alk and ltk, known to be expressed in iridophores) aquaculture compared to in the wild, there is a higher results in a dose dependent loss of the bars—white color chance of mutant fsh being successfully raised, if the and black outline. Tis indicates that melanophores mutation is not lethal. depend on iridophore signaling for their correct assem- Many variations are referred to using the same name by bly and migration to form the adjacent black border. Tis most companies, which makes it easier for the customer hypothesis is supported by the observation that during (and researcher) to navigate. However, some mutations white bar formation iridophores appear at the location have diferent names, such as “Gladiator/DaVinci” or

(See fgure on next page.) Fig. 1 Phylogenetic relationships and categories of white bar appearance in adult anemonefshes. A Phylogenetic tree of 27 anemonefsh species (adapted from the published work of Litsios et al. [31] and [52]. Three species could not be included into the tree because they are either rare with little genetic information available (A. fuscocaudatus) or hybrid species (A. leucokranos and A. thiellei). B No white bars; A. ephippium. C One white bar on the head; A. nigripes. D Two white bars, one on head and one on the body; A. bicinctus. E Three white bars, one on the head, one on the body and one on the tail; A. ocellaris. F Dorsal white stripe; A. sandaracinos Klann et al. EvoDevo (2021) 12:8 Page 4 of 16

A A. allardi B

A. omanensis

A. bicinctus

A. latifasciatus

A. chagosensis no white markings A. nigripes C A. chrysogaster

A. akindynos

A. mccullochi

A. polymnus one white bar on the head A. sebae

A. sandaracinos D

A. pacificus

A. akallopisos

A. perideraion

A. chrysopterus two white bars, one on the head and one on the body A. ephippium E A. melanopus

A. rubocinctus

A. barberi

A. frenatus three white bars, one on the head, one on the body A. clarkii and on one the tail A. tricinctus F

A. latezonatus

A. ocellaris

A. percula

P. biaculeatus one dorsal stripe Klann et al. EvoDevo (2021) 12:8 Page 5 of 16

A. ocellaris “Black/Darwin” *

“Misbar” “Gladiator/DaVinci” “Snowflake” “Wide Bar Gladiator” “Caramel/Mocha” “Black Misbar” “Zombie/ Tangerine”

“Extreme Misbar/ “Naked” “Wyoming White” “Frostbite” “Mocha Gladiator” “Black Ice “Black Snowflake” “Midnight” “Domino/ Nearly Naked” Snowflake” Extreme Misbar” *

“Black Frostbite” “Black Storm” “Midnight Lightning”

“Snow Storm” Fig. 2 Common A. ocellaris strains and their origins. The information that has been used to create Fig. 2 is publicly available on the homepages of larger aquaculture companies in the US, such as ORA, S&R, and Proaquatix. The asterisks indicate a (spontaneous) mutation of the variants: a mutated “Mocha Gladiator” was used to establish the “Black Storm” retail line, while a mutated “Black” was used to create the “Zombie” retail line

“Platinum/Maine Blizzard”, and thus care must be taken. In aquaculture, food plays an important role in fsh col- While a few mutations are restricted to specifc compa- oration [64]. Te skin color of captive-bred A. ocellaris nies and are often hard to fnd (such as “Xcalibour” or juveniles can be enhanced when providing diets that con- “Galaxy”), many are widely available within the trade. tain natural sources of carotenoids [43]. It is for this rea- Continuing ongoing breeding eforts by several com- son that many aquaculture companies add carotenoids as panies will surely lead to many more retail lines in the supplements to fsh food. Astaxanthin, β-carotene, and future. To date more than 40 color pattern variations/ canthaxanthine are often used as supplements, but in A. mutations are available, most prominently of A. ocella- ocellaris it has been shown that each carotenoid type pre- ris, P. biaculeatus, A. percula and A. clarkii (Additional sent in the skin varies with regard with the dietary sup- fle 1: Table S1). Te scheme in Fig. 2 displays common plements [65]. Apart from coloration, it has also been A. ocellaris strains and their origins. Te information shown that diet and fsh density (and other ecological that has been used to create Fig. 2 is publicly available on parameters) can afect white bar formation in A. percula the homepages of larger aquaculture companies in the and A. ocellaris, leading to the so-called “Misbar” pheno- United States, such as ORA (https://​www.​orafa​rm.​com/​ types [2, 9]. Terefore, it is not always easy to determine produ​cts/​fsh/​clown​fsh/), S&R (https://​www.​seaan​dreef.​ if a given color variant has a genetic or environmental com/​marine-​ornam​ental-​fsh/​clown​fsh/), and Proa- basis, and this is particularly relevant for fsh taken from quatix (https://​www.​proaq​uatix.​com/​clown​fsh/). Some the wild. strains, like “Snow Storm”, are the result of crossings running for several generations and are usually associ- A sort of material and methods: characterization ated with a higher retail price and often lower numbers of color mutants/abnormalities in anemonefshes of available individuals. It is noteworthy that many color Tis section describes and categorizes known mutants or mutations can be combined with morphological muta- abnormalities in coloration of diferent anemonefsh spe- tions. Tis is for example the case of “Longfn” that has cies found in the wild or in the aquarium trade. Te main been introduced in many diferent color mutants, such as source for wild anemonefshes with abnormal coloration “Domino” (“Longfn Domino”), “Black Storm” (“Longfn have been companies that export wild specimens, such as Black Storm”), or “Black Ice” (“Longfn Black Ice”). RVS Fishworld or others, which are often featured on the Klann et al. EvoDevo (2021) 12:8 Page 6 of 16

internet site reefbuilders.com. Additionally, the photo- Class I: imbalance of chromatophore subtypes graph-sharing platform “fickr” (https://​www.​fickr.​com/​ A very prominent example is provided by geographi- explo​re) was searched for images that were tagged as cally restricted wild melanistic populations, where the “clownfsh”. Similarly, the “iNaturalist” (https://​www.​inatu​ orange color is replaced by black color, resulting in ralist.​org/) website was searched for “Amphiprion” and black anemonefshes with white bars as seen in (1) com- “Premnas”, respectively. Detailed information (species, pletely black A. ocellaris found around Darwin, North location, and category of alteration) and a link is given Australia, (2) completely black A. clarkii from Ogasa- for all wild individuals that were recovered (Additional wara Island (Japan) and the Philippines (Fig. 3A), (3) fle 2: Table S2) as well as for many retail lines (Additional completely black A. frenatus from southern Japan and fle 1: Table S1). Due to copyright restrictions, it is not the Philippines (Fig. 3B), and (4) partially black A. per- possible to illustrate some of the natural variants, and in cula found around the Solomon Islands. All retail lines these cases schematic drawings are provided instead (e.g., for “Black/Darwin” A. ocellaris (Fig. 4A) are ofspring Fig. 3G and P, Fig. 4R–T). from wild-caught individuals, while the “Onyx” A. per- It must be noted that species identifcation from images cula retail lines come from selective breeding eforts. can be erroneous and should therefore be regarded with Even though all these individuals do not appear to lack caution, particularly for species that are difcult to dis- a specifc chromatophore subtype, the number of mel- tinguish in the wild (for example A. frenatus and A. mel- anophores is highly increased, and dominates the color anopus) or for species that are closely related and can be of nearly the entire fsh, apart from the white bars. hybridized in captivity. For example, A. ocellaris and A. Many wild individuals with reduced numbers of iri- percula are allopatric sister species that can be hybrid- dophores and often decreased black edging are known, ized in captivity (“Snow Onyx” or “Black Photon”) and it leading to the so-called “Misbar” phenotype. Te can be difcult to distinguish between them based only degree of mis-barring is highly variable, typically only on pictures. afecting one of the bars, predominantly the body bar. Since images of adult individuals were primarily As mentioned above, this phenotype can be linked to retrieved from the websites, the classifcation of pigmen- environmental factors, as well as to mutations. Several tation pattern abnormalities presented here is based on anemonefsh species show this rather common pheno- the adult phenotype only. Tis is in contrast with clas- type and are widely distributed: A. ocellaris (Fig. 3C, sifcations established in zebrafsh or medaka, which are D), P. biaculeatus (Fig. 3E) and A. polymnus (Fig. 3F) typically based on embryonic and larval pigmentation from the Philippines, A. bicinctus from the Red Sea, patterns [23, 24]. Within the text rare natural variants A. chrysopterus from the Marshall Islands, A. melano- and commercial aquaculture lines are presented together pus from the Coral Sea, and A. clarkii and A. percula because the phenotypes are often similar to each other. from the Solomon Islands. Both wild A. melanopus and However, for clarity, two separate fgures are presented, A. clarkii individuals have been employed to establish one detailing wild variants (Fig. 3), and the other com- the respective retail lines “Naked Cinnamon” (Fig. 4B) mercial lines (Fig. 4). and “Deluxe Clarkii” (Fig. 4C). For A. ocellaris, retail From the analyses of all these scattered data, the frst lines come from selective breeding of a few founder conclusion is readily apparent; that two broad categories individuals with severely reduced numbers of irido- of pigmentation phenotypes can be recognized in adult phores. “Naked” A. ocellaris (Fig. 4D) display a solid anemonefshes, namely (1) imbalance of chromatophore orange body with no white bars and no black outline, subtypes, and (2) irregular patterning mechanisms. We and are believed to be homozygous for a mutation of an will therefore describe and discuss these two categories unknown gene. “Misbar” and “Extreme Misbar” A. ocel- successively. laris (Fig. 4E) retain some white color in the position

(See fgure on next page.) Fig. 3 Color variants observed or taken from the wild. A Black A. clarkii, B Black A. frenatus, C Misbar A. ocellaris, D Misbar A. ocellaris, E Misbar P. biaculeatus, F Misbar A. polymnus, G “Golden Clownfsh”, H “Picasso”-typed A. ocellaris, I “Picasso”-typed A. ocellaris, J “Picasso”-typed A. ocellaris, K “Picasso”-typed A. frenatus, L “Picasso”-typed A. polymnus, M “Snowfake”-typed A. perideraion, N “Snowfake”-typed A. akindynos, O “Snowfake”-typed A. frenatus, P “Lightning” P. biaculeatus, Q “Xcalibour”-typed A. sandaracinos, R A. perideraion with extra elements, S A. ocellaris with extra elements, T A. polymnus with extra elements, U P. biaculeatus with extra elements, V “Wide Bar”-typed A. ocellaris, W “Wide Bar”-typed A. clarkii, X A. ocellaris with missing head bar. Photo credit: A–F, H–L, O, and S–X were kindly provided by RVS Fishworld; G and P are schematic drawing made by MK; M and R were taken by KH; N was taken by MM, and Q was kindly provided by Ishigakijima Kamome Diving Klann et al. EvoDevo (2021) 12:8 Page 7 of 16

A black A.c. B black A.f. C A.o. “Misbar”

D A.o. “Misbar” E P.b. “Misbar” F A.pol. “Misbar”

G A.o. “Golden Clownfish” H A.o. “Picasso”-type I A.o. “Picasso”-type

J A.o. “Picasso”-type KLA.f. “Picasso”-type A.pol. “Picasso”-type

M A.peri. “Snowflake”-type NOA.f. “Snowflake”-type P P.b. “Lightning”

A.a. “Snowflake”-type QRA.s. “Xcalibour”-type S A.o. extra elements T A.pol. extra elements

A.peri. extra elements extra elements “Wide Bar”-type “Wide Bar”-type missing head band UVP.b. A.o. W A.c. X A.o. Klann et al. EvoDevo (2021) 12:8 Page 8 of 16

of the head, body and/or tail bar and are supposed to polymnus (Fig. 3L). Te A. percula retail line is called represent the heterozygous state of the same mutation. “Picasso” (Fig. 4J), and more and more species have been Tese assumed genetic states need to be confrmed bred showing “Picasso”-type mutations, such as A. ocel- scientifcally, and for now have been inferred from laris “Gladiator/DaVinci” (Fig. 4K), A. clarkii “Galaxy”, A. breeding outcomes provided by ORA. For example, if bicinctus “Spotcinctus” (Fig. 4I), A. sebae, and P. biacu- “Misbar”x”Misbar” are crossed most ofspring will be leatus “Goldfake” (Fig. 4L). For A. percula “Picasso”, A. “Misbar”, some will be “Naked” and occasionally there ocellaris “Gladiator/DaVinci”, and P. biaculeatus “Gold- will be wild-type fsh (ORA, personal communication). fake” it is known that heterozygous mutants exhibit If the body color is black instead of orange, the retail the described phenotype, while homozygous mutants lines are called “Midnight” (Fig. 4F) and “Domino” display a nearly complete expansion of the white color (Fig. 4G), respectively. Te “Midnight” strain arose as on the body and are called “Platinum/Maine Blizzard” a separate mutation and is not the result of a specifc (Fig. 4M), “Wyoming White” (Fig. 4N), and “Gold Nug- crossing scheme involving “Naked”. get” (Fig. 4O), respectively. As discussed above, these Tere is only one known wild-caught individual, from assumed genetic states need to be confrmed scientif- Indonesia (most likely A. ocellaris), that lacks both mel- cally, and have been inferred from breeding outcomes anophores and iridophores—known as the “Golden provided by ORA. For example, if “Picasso”x”Picasso” are Clownfsh” (Fig. 3G). Te absence of all melanophores, crossed, ofspring will comprise “Picasso”, “Platinum” and including the black outline of the white bars, the fns and wild-type fsh. Similarly, if “Gladiator”x”Gladiator” are the orange-colored areas is striking. Tis contrasts with crossed the ofspring will include “Gladiator”, “Wyoming individuals that lack iridophores only, because these indi- White” and wild-type fsh. Even though viability is low, all viduals exhibit normal black coloration of the fns and ofspring of “Gold Nugget”x”Gold Nugget” will be “Gold orange-colored areas (see “Naked” A. ocellaris Fig. 4D). It Nugget”. is unclear if a retail line has been completely established A second mutation is characterized by irregular white yet, but Similan Farms Clownfsh has a strain labeled bars with exaggerated and jagged edges. In the aquarium “Super Yellow Percula” (showing very limited black col- trade, this phenotype is known as “Snowfake” and is oration in fns) that is reminiscent of this phenotype. only known in A. ocellaris (Fig. 4P). Tere are few wild A special case is presented by the hypomelanic A. individuals known with similar phenotypes, but the ocellaris that are called either “Zombie” (Fig. 4H) or altered area is comparatively small as seen in A. peride- “Tangerine” in the aquarium trade. To the best of our raion (Fig. 3M), A. akindynos (Fig. 3N), and A. frenatus knowledge, this phenotype has not been discovered in (Fig. 3O). the wild yet. Although melanophores are most likely pre- A third mutation, only known from P. biaculeatus, sent in this color morphotype, no or very little melanin results in a network-like disruption of all three white bars, (black color pigment) is produced. that often also spreads over the body, and is called “Light- ning”. Tis mutation has been found mainly in Papua Class II: irregular patterning mechanisms New Guinea, and wild founder individuals (Fig. 3P) have Variations in bar shape occur regularly, are highly vari- been used for all retail lines available (Fig. 4Q). able between individuals, and vary greatly in their extent A special case is presented by a phenotype found only – from a small bulge to extensive shape deformations. In in A. sandaracinos, which is characterized by a cross- general, there seem to be several diferent genetic muta- like mark (a horizontal branching of the dorsal stripe) tions underlying these patterning abnormalities. just posterior of the eyes. A. sandaracinos usually has Te most commonly observed mutation results in no bars, and only a single dorsal stripe. However, in bars that are interconnecting, with an irregular but this morphotype it appears as if a head bar has started smooth outline, and often also include extra white mark- to form in the dorsal-most region only. Contrary to ings in between bars. Wild examples include A. ocella- all other bar shape alterations, this phenotype is very ris (Fig. 3H–J), A. percula, A. frenatus (Fig. 3K), and A. reproducible between individuals in the wild (Fig. 3Q),

(See fgure on next page.) Fig. 4 Color mutants available from aquaculture companies. A “Darwin Black” A. ocellaris, B “Naked Cinnamon” A. melanopus, C “Deluxe Clarkii” A. clarkii, D “Naked” A. ocellaris, E “Extreme Misbar” A. ocellaris, F “Midnight” A. ocellaris, G “Domino” A. ocellaris, H “Zombie” A. ocellaris, I “Spotcinctus” A. bicinctus, J “Picasso” A. percula, K “Gladiator” A. ocellaris, L “Goldfake” P. biaculeatus, M “Platinum” A. percula, N “Wyoming White” A. ocellaris, O “Gold Nugget” P. biaculeatus, P “Snowfake” A. ocellaris, Q “Lightning” P. biaculeatus, R “Xcalibour” A. sandaracinos, S “Morse Code” P. biaculeatus, T “Wide Bar Gladiator” A. ocellaris. Apart from the schematic drawings in R–T all images were provided by ORA Klann et al. EvoDevo (2021) 12:8 Page 9 of 16

A B C

A.o. “Black/Darwin ” A.m. “Naked” A.c. “Deluxe”

D E F

A.o. “Naked” homozygote A.o. “Extreme Misbar” heterozygote A.o. “Midnight” G H I

A.o. “Domino” A.o. “Zombie/Tangerine” A.b. “Spotcinctus”

heterozygote heterozygote heterozygote J K L

A.p. “Picasso” A.o. “Gladiator/DaVinci” P.b. “Goldflake” M homozygote N homozygote O homozygote

A.p. “Platinum/Maine Blizzard” A.o. “Wyoming White” P.b. “Gold Nugget”

P Q R

A.o. “Snowflake” P.b. “Lightning” A.s. “Xcalibour”

ST

P.b. “Morse Code” A.o. “Wide Bar Gladiator” Klann et al. EvoDevo (2021) 12:8 Page 10 of 16

as well in the aquarium trade where they are called Disruptions in any of these have the potential to alter pig- “Xcalibour” (Fig. 4R). mentation patterns. Occasionally, bars are interrupted with dot-like pat- terns, or there are white dots displayed in areas where Chromatophore specifcation there usually is no bar, as found in A. perideraion Defects during specifcation are likely present if chroma- (Fig. 3R), A. frenatus, A. ocellaris (Fig. 3S), A. polymnus tophore numbers are strongly reduced, but the remaining (Fig. 3T), and A. percula. Several wild-caught P. biaculea- cells appear normal and in the correct position. tus individuals from Papua New Guinea exhibited extra Mutations related to “Naked” phenotypes (highly dots and even dashes between the bars or emerging from reduced white barring, Fig. 4B–G) are most likely caused bars (Fig. 3U), and have been used to establish the retail by genes responsible for iridophore development, in par- line “Morse Code” (Fig. 4S). ticular specifcation. As mentioned above, the reduction Another type of altered patterning is individuals show- of iridophores will indirectly reduce the number of mel- ing the normal arrangement and shape of bars, but with anophores that form the edge of the white bars, but mel- an increased width. In the aquarium trade these are anophore distribution remains normal in fn and body known as “Wide Bar Gladiator” in A. ocellaris (Fig. 4T). coloration. In zebrafsh, several genes have been linked A few wild A. ocellaris (Fig. 3V) individuals are known, as with iridophore specifcation and they can serve as an well as a local population of A. clarkii in the Philippines entry point to investigate “Naked”-typed mutants. For (Fig. 3W). It should be mentioned that the normal color example, Foxd3 expression is required for neural crest pattern of A. latezonatus displays an exceptionally wide derivatives to develop into iridophores [29]. Leucocyte middle bar. tyrosine kinase (ltk) controls iridophore establishment, proliferation, and survival, and is associated with at least Discussion: anemonefsh pigmentation mutants two mutants, “shady” (lack of iridophores) and “moon- in a wider context stone” (ectopic iridophores) [13]. Endothelin receptor Te scattered information regarding wild and commer- signaling is crucial for larval iridophores highlighted by cial pigmentation mutants of anemonefshes has been the importance of genes encoding Edn3b (Endothelin 3b) gathered in this work, and compiled and categorized into and Ednr3b (Endothelin receptor b1a) [28]. two classes, encompassing either chromatophore imbal- Apart from iridophores, other chromatophore subtypes ances (class I) or patterning irregularities (class II). might be also afected, as demonstrated in zebrafsh (e.g.: What global picture emerges from these data? Inter- the “nacre” mutant does not possess melanophores) and estingly, although several classes of mutant phenotypes guppies (e.g.: the “blue” mutant lacks xanthophores). can be identifed, there are several potential phenotypes However, in anemonefsh specifcation defects of either that do not seem to exist, predominantly the complete melanophores or xanthophores have not been discovered absence of chromatophore subtypes (lack of xantho- yet, neither in the wild (apart for the “Golden Clownfsh”, phores or lack of all chromatophores). In order to learn Fig. 3G), nor in aquaculture. Even though some retail something about pigmentation pattern formation, bio- mutant strains are nearly or completely white (A. ocel- logical processes that could lead to the observed mutant laris “Wyoming White”, Fig. 4N, A. percula “Platinum/ phenotypes must be taken into consideration and investi- Maine Blizzard”, Fig. 4M) or completely black (A. ocel- gated in detail. Even though much more data, especially laris “Midnight”, Fig. 4F) as adults, larvae and juveniles experimental data, are needed to perform these analyses, of these lines display an orange body coloration that is we will try, in the following discussion, to link selected slowly replaced by a diferent color during maturation anemonefsh mutant phenotypes to potential processes (own personal observation, correspondence with aqua- afected. culture companies). Te face and in particular the region around the mouth is usually the last to change color as What can we learn about color pattern formation shown in the “Midnight” mutant (Fig. 4F). Te “Golden from pigmentation mutants? Clownfsh” (Fig. 3G) is the only known individual that Classifcation of pigmentation phenotypes in zebrafsh appears to lack both iridophores and melanophores, leav- and medaka has allowed researchers to clarify which ing only xanthophores to color the entire body uniform biological process(es) might be afected by the underly- yellow. ing mutation. As pigmentation heavily relies on chro- All three chromatophore subtypes are afected when matophores and their spatial arrangement, the neural genes required for normal development of neural crest crest development is of particular interest. Key develop- derivatives are mutated, such as mutations of Sox10 mental processes are (1) specifcation, (2) proliferation, (SRY-related HMG-box 10). Te corresponding zebrafsh (3) diferentiation, (4) survival, and (5) patterning [23]. mutant is called “colorless” [23] and the same gene is also Klann et al. EvoDevo (2021) 12:8 Page 11 of 16

required for chromatophore development in guppies Diferentiation processes [12]. To the best of our knowledge, no anemonefsh lack- Alterations of cell diferentiation processes can include ing all three chromatophore subtypes has yet been found. a variety of phenotypes, such as reduced or altered pig- Neural crest derivatives do not only specify chromato- mentation or abnormal cell morphology. phore subtypes, but also craniofacial cartilage, peripheral A frst example for a diferentiation defect is the “Zom- and enteric neurons as well as glia [4, 22]. It is possible bie” mutation (Fig. 4H), which is most likely caused by that in anemonefsh mutations that afect these addi- the lack of melanin (albinism). In albinism, melanophores tional cell types are lethal and therefore have not been are present in their normal arrangement and distribu- identifed so far. tion but they are de-pigmented, either completely lacking Of course, it is difcult to go further with analyses of melanin or displaying highly reduced melanin levels [39, the existing mutants without more detailed information. 63]. Since melanophores are present in these mutants, But from the data available so far, very interesting and regulatory genes, such as Mitfa (microphthalmia tran- focused questions emerge: Do xanthophores and melan- scription factor a) or kita (receptor tyrosine kinase a), are ophores in anemonefsh or rather their precursors have unlikely to be afected in these mutants. Genetic muta- other functions that if disrupted lead to highly increased tions in a gene involved in melanogenesis can potentially mortality? Alternatively, it may be that the specifcation result in reduced or absent melanin synthesis leading to process for these cells can take alternative routes, so even reduced pigmentation levels. Te biosynthetic pathway if one gene is knocked out, others can replace it so that during which L-tyrosine is converted to melanin through the precursor is still established. But why should this several enzyme catalyzed steps involves many genes, situation be diferent for iridophores? All these questions such as MC1R (melanocortin 1 receptor), tyr (tyrosi- are awaiting molecular and developmental characteriza- nase), tyrp1 (tyrosinase-related protein-1), and dct/tyrp2 tion of these specifcation mutants. (dopachrome tautomerase). Te most well documented example is MCR1, which contributes to melanin-related Chromatophore proliferation polymorphisms in several animal species, including the Proliferation defects can also be indicated by a strongly guppy Poecilia reticulata [61], zebrafsh [51], and vari- reduced number of chromatophores as well as normal ous vertebrates [19]. Another potential candidate is oca2 and correct positions of the remaining chromatophores. (oculocutaneous albinism II). When its function is dis- Many retail lines display a pronounced shift towards rupted, the frst step of melanin synthesis, the conversion either partially or completely white or black morpho- of L-tyrosine to L-DOPA, is impaired, and this situation types. However, this does not necessarily mean that the is responsible for the evolution of albinism in the cavefsh establishment of the xanthophore lineage is negatively Astyanax mexicanus [27]. Any of the above-mentioned afected. For example, larvae and juveniles of many of genes might be responsible for the “Zombie” mutants these lines (such as “Darwin Black”, “Black Storm”, “Snow in anemonefsh. Interestingly, the frst “Zombie” mutant Storm”, and “Wyoming White”) usually display an orange frst arose from a pair of “Black/Darwin” A. ocellaris body coloration that is slowly replaced by black color (Fig. 4A). Juvenile “Zombie” mutants display a bright yel- during maturation. Tis suggests that initial xanthophore low-orange body coloration with white bars devoid of the specifcation and proliferation processes are unafected, black edge and have red eyes. Later on, individuals turn but their subsequent fate remains to be analyzed. Possi- a darker shade of orange-reddish to reddish-brown. Tis ble scenarios include (1) melanophores (or iridophores) suggests that some melanin is produced and accumu- replacing xanthophores; (2) xanthophores changing lates, resulting in the darker shades of mature individu- chromatophore subtype identity, or (3) xanthophores als. Because the original “Zombie” arose from completely remaining present but becoming covered by other chro- black parents, melanophores cover the entire body and matophores. For example, anemonefsh skin chromato- the little melanin that is produced is evenly distributed. phores are distributed within the epidermis as well as the Another example of altered pigment cell diferentia- dermis [55, 56] and it is possible that changes in the top tion is the icy blue, iridescent hue (instead of the nor- layers cover the coloration of the deeper layers. In this mal white coloration) that is usually more prominent hypothesis, it is assumed that xanthophores remain pre- near black coloration. Tis altered pigmentation feature sent in a deeper epidermal layer, while melanophores (or is most often seen in black and white morphotypes of A. iridophores) spread in a more upper layer and therefore ocellaris mutant lines, such as “Black Frostbite” or “Black overlay the orange color. Regardless of the exact mecha- Snowfake” (see also https://​reefb​uilde​rs.​com/​2018/​03/​ nism, proliferative processes of melanophores (or irido- 05/​blue-​clown​fsh/). It is noteworthy that some species phores, respectively) are enhanced above normal levels of anemonefsh such as A. chrysopterus, A. frenatus, A. during juvenile color maturation. melanopus, and P. biaculeatus, show an increased bluish Klann et al. EvoDevo (2021) 12:8 Page 12 of 16

appearance at the black border as part of their normal Terefore, it is unclear if the underlying mechanisms are pigmentation patterns. Tis phenotype is most likely inheritable, and the molecular and cellular mechanisms linked to the arrangement and organization of guanine responsible for the anterior to posterior sequence of bar platelets and crystals within the iridophores. As men- appearance/formation remain to be analyzed. tioned above, an iridescent color can be observed when Pigmentation research on anemonefshes ofers the crystals are aggregated in platelets that are precisely unique opportunity to study boundary formation— organized, while they appear whitish when the crystal with melanophores forming a distinct border between platelets are less organized as light is scattered in various the other two chromatophore subtypes, which is much directions [15, 59]. Guanine platelet formation requires harder to investigate in zebrafsh or medaka (zebrafsh: gmps (guanine monophosphate synthetase) expression diferently colored stripes, medaka: uniform color). It [42] and a mutation in this gene could potentially alter can therefore be assumed that novel processes will be platelet formation. However, since the blue-iridescent revealed from such future analyses. Like the underly- hue of iridophores usually appears near the black bor- ing mechanisms for anterior–posterior patterning, the der, it is also possible that the presence of melanophores genetic cause(s) for mutations that result in abnormal bar is partially responsible for guanine crystals to be orien- shape are rather hard to predict. Te failure of diferent tated diferently. Interestingly, the existence of two difer- subtypes of chromatophores to communicate and inter- ent subtypes of iridophores has been shown recently in act could be responsible for altered boundary formation. zebrafsh [17] and the authors suggest that the presence Indeed, preliminary data suggests that melanophores of melanophores might afect the patterning of both iri- restrict iridophore expansion (our own observation). dophore subtypes. Terefore, if melanophore regeneration and migra- tion abilities are impaired, this could potentially lead to Pigmentation patterning altered bar shapes. As outlined above, diferences can Patterning mechanisms are likely afected if a subset of be observed between several types of bar shape muta- the pattern is absent or altered in any way (e.g.: position, tions. “Picasso”-typed mutations (Fig. 4I–L) are the most extra elements). In general, these changes are very hard widespread and have been found in at least eight difer- to predict and will most likely afect tissues other than ent species of anemonefsh (A. percula, A. ocellaris, A. the neural crest. clarkii, A. bicinctus, A. sebae, A. frenatus/A. melanopus, In anemonefshes, major patterning factors concern A. polymnus, and P. biaculeatus). Moreover, the striking the anterior–posterior succession of white bars. Both the similarities between heterozygous and homozygous phe- position as well as the sequence of white bar appearance notypic appearances (compare Fig. 4J/4 M with 4 K/4 N appears highly constrained. All species of anemonefsh with 4L/4O) in three species (A. percula, A. ocellaris, P. that possess white bars have them at the same position: biaculeatus) suggests that the underlying genetic mecha- (1) the transition of head to body; (2) the middle of the nisms are very similar, if not the same, while phenotypic body, and (3) the caudal peduncle/tail. Tis suggests that appearance remains highly variable. Comparative devel- the same morphological landmarks are used to establish opmental analyses of diferent species would be helpful the exact position of the bars. For the white body bar, it to investigate when color alterations frst appear and if has been suggested that the transition between anterior they are similar in diferent species. Furthermore, genetic spines and posterior soft rays of the dorsal fn acts as a analyses of all species showing “Picasso”-typed mutations pre-positioning cue [58]. More specifcally, an indenta- would show if the same genes are indeed afected. tion (made by longer anterior spines and shorter poste- Preliminary research on “Snowfake” A. ocellaris rior spines) is more pronounced in species that have two (Fig. 4P) suggest that a local decrease of melanophores or three white bars and less obvious in species with zero forming the black edge might enable the iridophores or one bar [58]. In addition to the exact position of the to expand into neighboring areas, leading to the jagged white bar, their developmental appearance is also highly black outline of this mutation (own observation). conserved and constrained, as mentioned above. No spe- cies exhibits a single bar on the body or tail only. So far, Ecological and evolutionary implications of color pattern only one wild-caught fsh has been found that exhibits a variations disruption in this normal anterior–posterior patterning Te ecological functions of color patterns in anemone- (Fig. 3X). Tis specimen lacked the head bar but showed fshes are still unclear, probably because experimental normal formation of the body and tail bars. As there usu- studies are scarce. However, it has been hypothesized ally is an anterior to posterior sequence of bar acquisi- that color might play a role in various functions, such tion during larval/juvenile development, this specimen is as (1) concealment and camoufage either from preda- very fascinating, but unfortunately, its fate is unknown. tors [36, 58] or from resident fsh in the case of juveniles Klann et al. EvoDevo (2021) 12:8 Page 13 of 16

[6], (2) species recognition and co-existence [58], and that in turn would join or establish colonies. One reason (3) individual identifcation [8]. Alteration of color pat- could be that juveniles with color alterations are less suc- tern via mutation might then afect the adaptive value of cessful in joining an existing colony precisely because the color pattern trait, and hence the ftness of the indi- they exhibit a diferent color pattern. Te investigation of vidual. Indeed, some color pattern mutations observed the ecological function(s) of color patterns would greatly in the aquarium trade are never or very rarely observed beneft from experimental studies using mutants, which in the wild, for example almost completely white mor- could provide new insights on this poorly understood photypes or individuals with no bands in species that aspect of anemonefsh ecology. For example, in meso- usually have bands. Tis suggests that mutations which cosm experiments it would be interesting to study how result in drastic color pattern alterations might have a juveniles with diferent white bar patterns are diferen- negative efect on the survival of individuals in the wild tially afected during the recruitment phase when joining and are therefore negatively selected against. However, an existing colony. Moreover, the efect of changes in bar other mutations are more widespread (like melanism, patterns on the aggressiveness of conspecifcs could also and “Picasso”-typed bar shape alterations), and individu- be studied. Tese types of analyses are currently under- als survive long enough to reach reproductive positions. way in our laboratory. Tis indicates that these mutations have no or negligible Finally, it should be noted that some polymorphic vari- negative efects on survival, as the ascension to breeding ations in bar shape are observed more frequently in the positions is a lengthy process as juveniles go up the social wild but are not regarded to be the results of mutations. hierarchy [7]. In the aquarium trade, A. ocellaris is by Tese include, for example, (1) the shape of the middle far the anemonefsh species that exhibits the most color bar in A. polymnus, (2) the head bar in A. nigripes which mutations, followed by P. biaculeatus (Additional fle1: can be joined dorsally or display a gap, and (3) the ten- Table S1). In the wild, most mutant individuals are P. bia- dency of all three bars to disappear in a ventral to dorsal culeatus or A. perideraion, and have been noted particu- fashion in older individuals of P. biaculeatus and A. ocel- larly from within the Coral Triangle area. Ecological feld laris. Te last case may be linked to senescence related studies within the Ryukyu Archipelago revealed rela- degeneration processes. tively high rates of abnormal coloration in A. perideraion Research has suggested that approximately half of all (1.8%) and A. sandaracinos (2.56%), but negligible rates coral reef fsh species have evolved in the last 5 Myr [3]. for the remaining four anemonefsh species present in While much of this massive radiation of new species is this region (A. clarkii, A. frenatus, A. ocellaris, and A. pol- thought to have been driven by biogeographic condi- ymnus) (Kina Hayashi, unpublished observation). From a tions, colors and the patterns of fsh are also thought to limited Internet search using the photo sharing platform have played important roles [48], perhaps via stabilizing fickr (https://​www.​fickr.​com/​explo​re), a total of 36,195 species boundaries [3]. Recent phylogenetic analyses of pictures that were tagged as “clownfsh” were screened Amphiprioninae with molecular clock analyses show that for images of wild mutant individuals. Only approxi- a large majority of anemonefsh species have also evolved mately half of those images displayed natural marine within the last 5 Myr, likely due to biogeography com- environments, from which 34 aberrant anemonefsh bined with ecological speciation related to host-anemone images were retrieved (0.18%). Tis fgure is an only gross associations [30, 32]. Tis Amphiprioninae phylogenetic estimate and most likely represents a minimum of wild framework combined with recent phylogenetic data on mutant anemonefshes. On the other hand, it may be that anemones [62] and the analyses of coloration and pat- wild anemonefsh variants are more noticeable and rec- terning as suggested in this review promise to open reational divers might take more photographs of unusual exciting new avenues of research that will allow a better individuals, and clearly more data from feld observations understanding of the evolution of anemonefsh pigmen- are needed to better establish the true rates of abnor- tation diversity. mal coloration of anemonefsh. As well, wild-caught specimens with aberrant color patterns are defnitively Concluding remarks on conservation issues targeted by export companies to be sold to aquaculture Te conservation of coral reefs and therefore anemone- companies or private collectors [37]. fshes is an important topic facing critical problems [34]. From these observations, it can be concluded that most It must be stated clearly that the introduction of non- variants are rare and therefore selected against, imply- native anemonefsh species or mutants that have been ing that variations do not provide any advantages. It is artifcially produced into coral reef ecosystems would be possible that even though mutant individuals may reach disastrous. Natural hybridization can play an important breeding positions, the underlying mutation is usually not role in the evolution and radiation of taxa by increasing transmitted to an increased number of aberrant juveniles, the genetic source [11, 16, 31], which is the case for the Klann et al. EvoDevo (2021) 12:8 Page 14 of 16

two naturally occurring hybrid species: A. leucokranos (A. Additional fle 1: Mutant anemonefsh retail lines. chrysopterus x A. sandaracinos) and A. thiellei (A. ocel- Additional fle 2: Wild anemonefsh variants. laris x A. sandaracinos). It seems unlikely that natural hybridization events will result in species with more com- Acknowledgements plex color patterns, because all species of anemonefsh This paper is dedicated to “Mama Lightning”, a female Premnas biaculeatus that show rather similar, “normal” color patterns. Te high- was the progenitor of the “Lightning Maroon” anemonefsh strain. She was est degree imaginable is a hybrid with three horizontal captured in Papua New Guinea over 10 years ago as a young male and died recently in captivity as a female after giving rise to a large progeny currently bars and one dorsal stripe. However, the artifcial breed- known as the “Lightning Maroon” strain. ing of hybrids is often seen as a negative when the con- The rights to use their images were kindly provided by ORA farm (Oceans, servation of wild populations is taken into regards [1]. Reefs & Aquariums, https://​www.​orafa​rm.​com/​produ​cts/​fsh/​clown​fsh/) and RSV Fishworld (http://​rvsf​shwor​ldinc.​com/​index.​aspx). We thank Yi-Hsien Su Popular hybrid strains include “Blood Orange/Mai Tai” for approaching us to write this paper. (A. ocellaris x P. biaculeatus), “Snow Onyx” (A. ocellaris “Snowfake” x A. percula “Onyx”), and “Black Photon” Authors’ contributions MK, MM, LC, KH, and JDR searched for and/or contributed pictures of wild (A. ocellaris “Black Darwin” x A. percula “Onyx”) lines. anemonefsh variants. MK and VL conceptualized and designed the study. MK Future breeding eforts, especially by aquaculture com- wrote most of the article, with contributions from MM, JDR, and VL. All authors panies, will show if higher color pattern complexities of have read and approved the fnal manuscript. retail lines can be achieved; for example, (1) combining Funding the “Lightning Maroon” phenotype of P. biaculeatus with Work from our laboratory is supported by an OIST KICKS Grant. A. ocellaris, or (2) introducing the “Snowfake” mutation Availability of data and materials of A. ocellaris into other anemonefsh species. Te intro- All data generated or analyzed during this study are included in this published duction of hybrids into a natural system can lead to the article (and its Addditional fles 1, 2). loss of parental species, in particular when one of them is relatively rare. Moreover, if two species are bred that Declarations have similar morphological appearances, sometimes the Ethics approval and consent to participate. ofspring may not clearly be recognized as they are often Not applicable. an intermediate between the parent species. If such speci- mens are subsequently sold as a pure-bred species, breed- Consent for publication. Not applicable. ing eforts to establish designer anemonefsh, as well as conservation eforts, are threatened. Efective manage- Competing interests ment of hybrids from a conservation perspective is still a The authors declare that they have no competing interests. topic under debate [1, 20, 50, 60]. Because the long-term Author details outcomes of hybridization are unpredictable, care should 1 Marine Eco‑Evo‑Devo Unit, Okinawa Institute of Science and Technology, 2 be taken when acquiring and handling hybrids or sus- 1919‑1 Tancha, Onna‑son, Okinawa 904‑0495, Japan. Present Address: Marine Eco‑Evo‑Devo Unit, Okinawa Institute of Science and Technology, 1919‑1 Tan‑ pected hybrids. cha, Onna‑son, Okinawa 904‑0495, Japan. 3 Molecular Invertebrate Systematics If these conservation issues are seriously taken, it must and Ecology Lab, Graduate School of the Engineering and Science, University 4 be emphasized that aquaculture trade anemonefsh of the Ryukyus, 1 Senbaru, Nishihara, Okinawa 903‑0213, Japan. Tropical Biosphere Research Center, University of the Ryukyus, 1 Senbaru, Nishihara, mutants ofer a great entry point to better understand the Okinawa 903‑0213, Japan. 5 Marine Research Station, Institute of Cellular genetics and development of complex and polymorphic and Organismic Biology (ICOB), Academia Sinica, 23‑10, Dah‑Uen Rd, Jiau Shi, pigmentation patterns. Anemonefsh can be raised easily I‑Lan 262, I‑Lan, Taiwan. in laboratory husbandry systems and are open to devel- Received: 14 April 2021 Accepted: 2 June 2021 opmental and genetic manipulation [38, 54] as well as to genomic analyses [54]. Terefore, identifying the genes afected in various mutants, manipulating them in labo- ratory environments, testing the efect of these mutations References in the complex social system of captive anemonefsh col- 1. Allendorf FW, Leary RF, Spruell P, Wenburg JK. The problems with hybrids: setting conservation guidelines. Trends Ecol Evol. 2001;16:613–22. onies, and studying the polymorphism and evolutionary 2. Avella MA, Olivotto I, Gioacchini G, Maradonna F, Carnevali O. The role patterns of these genes in wild populations will certainly of fatty acids enrichments in the larviculture of false percula clownfsh shed light on how and why brilliant pigmentation pat- Amphiprion ocellaris. Aquaculture. 2007;273:87–95. 3. Bellwood DR, Goatley CHR, Bellwood O. The evolution of fshes and corals terns emerged in coral reef fshes. on reefs: form, function and interdependence. Biol Rev. 2017;92:878–901. 4. Bronner ME, LeDouarin NM. Development and evolution of the neural Supplementary Information crest: an overview. Dev Biol. 2012;366:2–9. The online version contains supplementary material available at https://​doi.​ 5. Budi EH, Patterson LB, Parichy DM. Post-embryonic nerve-associated org/​10.​1186/​s13227-​021-​00178-x. precursors to adult pigment cells: genetic requirements and dynamics of morphogenesis and diferentiation. PLoS Genet. 2011;7:e1002044. Klann et al. EvoDevo (2021) 12:8 Page 15 of 16

6. Buston PM. Forcible eviction and prevention of recruitment in the clown 30. Litsios G, Pearman PB, Lanterbecq D, Tolou N, Salamin N. The radia‑ anemonefsh. Behav Ecol. 2003;14:576–82. tion of the clownfshes has two geographical replicates. J Biogeogr. 7. Buston PM. Mortality is associated with social rank in the clown anem‑ 2014;41:2140–9. onefsh (Amphiprion percula). Mar Biol. 2003;143:811–5. 31. Litsios G, Salamin N. Hybridisation and diversifcation in the adaptive 8. Buston PM. Size and growth modifcation in clownfsh. Nature. radiation of clownfshes. BMC Evol Biol. 2014;14:245. 2003;424:145–6. 32. Litsios G, Sims CA, Wüest RO, Pearman PB, Zimmermann NE, Salamin N. 9. Chambel J, Severiano V, Baptista T, Mendes S, Pedrosa R. Efect of stocking Mutualism with sea anemones triggered the adaptive radiation of clown‑ density and diferent diets on growth of Percula Clownfsh, Amphiprion fshes. BMC Evol Biol. 2012;12:212. percula (Lacepede, 1802). Springerplus. 2015;4:183. 33. Marshall NJ. Communication and camoufage with the same “bright” 10. Dale J, Dey CJ, Delhey K, Kempenaers B, Valcu M. The efects of life history colours in reef fshes. Philos Trans R Soc Lond B Biol Sci. 2000;355:1243–8. and sexual selection on male and female plumage colouration. Nature. 34. McClanahan TR. Wilderness and conservation policies needed to avoid a 2015;527:367–70. coral reef fsheries crisis. Mar Policy. 2020;119:104022. 11. DiBattista JD, Rocha LA, Hobbs J-PA, He S, Priest MA, Sinclair-Taylor TH, 35. McMenamin SK, Bain EJ, McCann AE, Patterson LB, Eom DS, Waller ZP, Bowen BW, Berumen ML. When biogeographical provinces collide: Hamill JC, Kuhlman JA, Eisen JS, Parichy DM. Thyroid hormone-depend‑ hybridization of reef fshes at the crossroads of marine biogeographical ent adult pigment cell lineage and pattern in zebrafsh. Science. 2014. provinces in the Arabian Sea. J Biogeogr. 2015;42:1601–14. https://​doi.​org/​10.​1126/​scien​ce.​12562​51. 12. Dick C, Arendt J, Reznick DN, Hayashi CY. The developmental and genetic 36. Merilaita S, Kelley JL. Scary clowns: adaptive function of anemonefsh trajectory of coloration in the guppy (Poecilia reticulata). Evol Dev. coloration. J Evol Biol. 2018;31:1558–71. 2018;20:207–18. 37. Militz TA, McCormick MI, Schoeman DS, Kinch J, Southgate PC. Frequency 13. Fadeev A, Krauss J, Singh AP, Nüsslein-Volhard C. Zebrafsh leucocyte and distribution of melanistic morphs in coexisting population of nine tyrosine kinase controls iridophore establishment, proliferation and clownfsh species in Papua New Guinea. Mar Biol. 2016;163:200. survival. Pigment Cell Melanoma Res. 2016;29:284–96. 38. Mitchell LJ, Tettamanti V, Marshall NJ, Cheney KL, Cortesi F. CRISPR/Cas9- 14. Frohnhöfer HG, Krauss J, Maischein HM, Nüsslein-Volhard C. Iridophores mediated generation of biallelic G0 anemonefsh (Amphiprion ocellaris) and their interactions with other chromatophores are required for stripe mutants. bioRxiv. 2020. https://​doi.​org/​10.​1101/​2020.​10.​07.​330746. formation in zebrafsh. Development. 2013;140:2997–3007. 39. Montoliu L, Grønskov K, Wei A-H, Martínez-García M, Fernández A, 15. Fujii R. Cytophysiology of fsh chromatophores. Int Rev Cytol. Arveiler B, Morice-Picard F, Riazuddin S, Suzuki T, Ahmed ZM, Rosenberg 1993;143:191–255. T, Li W. Increasing the complexity: new genes and new types of albinism. 16. Gainsford A, Van Herwerden L, Jones GP. Hierarchical behaviour, habitat Pigment Cell Melanoma Res. 2014;27:11–8. use and species size diferences shape evolutionary outcomes of hybridi‑ 40. Nagao Y, Suzuki T, Shimizu A, Kimura T, Seki R, Adachi T, Inoue C, Omae zation in a coral reef fsh. J Evol Biol. 2015. https://​doi.​org/​10.​1111/​jeb.​ Y, Kamei Y, Hara I, Taniguchi Y, Naruse K, Wakamatsu Y, Kelsh RN, Hibi M, 12557. Hashimoto H. Sox5 functions as a fate switch in medaka pigment cell 17. Gur D, Bain EJ, Johnson KR, Aman AJ, Pasoili HA, Flynn JD, Allen MC, development. PLoS Genet. 2014;10:e1004246. Deheyn DD, Lee JC, Lippincott-Schwartz J, Parichy DM. In situ diferentia‑ 41. Nakamasu A, Takahashi G, Kanbe A, Kondo S. Interactions between tion of iridophore crystallotypes underlies zebrafsh stripe patterning. Nat zebrafsh pigment cells responsible for the generation of Turing patterns. Commun. 2020;11:6391. PNAS. 2009;106:8429–34. 18. Hendrick LA, Carter GA, Hilbrands EH, Heubel BP, Schilling TF, Le Pabic P. 42. Ng A, Uribe RA, Yieh L, Nuckels R, Gross JM. Zebrafsh mutations in gart Bar, stripe and spot development in sand-dwelling cichlids from Lake and paics identify crucial roles for de novo purine synthesis in vertebrate Malawi. EvoDevo. 2019;10:18. pigmentation and ocular development. Development. 2009;136:2601–11. 19. Hoekstra HE. Genetics, development and evolution of adaptive pigmen‑ 43. Nhan HT, Minh TX, Liew HJ, Hien TTT, Jha R. Efects of natural dietary tation in vertebrates. Heredity. 2006;97:222–34. carotenoids on skin coloration of false Clownfsh (Amphiprion ocellaris 20. Jackiw RN, Mandil G, Hager HA. A framework to guide the conserva‑ Cuvier, 1830). Aquacult Nutr. 2019;25:662–8. tion of species hybrids based on ethical and ecological considerations. 44. Olivotto I, Chemello G, Vargas A, Randazzo B, Piccinetti CC, Carnevali O. Conserv Biol. 2015. https://​doi.​org/​10.​1111/​cobi.​12526. Marine ornamental species culture: from the past to “Finding Dory.” Gen 21. Kelley JL, Fitzpatrick JL, Merilaita S. Spots and stripes: ecology and Comp Endocrinol. 2017;245:116–21. colour pattern evolution in butterfyfshes. Proc R Soc B Biol Sci. 45. Parichy DM, Spiewak JE. Origins of adult pigmentation: diversity in pig‑ 2013;280:20122730. ment stem cell lineages and implications for pattern evolution. Pigment 22. Kelsh RN. Sorting out Sox10 functions in neural crest development. BioEs‑ Cell Melanoma Res. 2015. https://​doi.​org/​10.​1111/​pcmr.​12332. says. 2006;28:788–98. 46. Patterson LB, Parichy DM. Zebrafsh pigment pattern formation: insights 23. Kelsh RN, Brand M, Jiang YJ, Heisenberg CP, Lin S, Hafter P, Odenthal J, into the development and evolution of adult form. Ann Rev Genet. Mullins MC, van Eeden FJ, Furutani-Seiki M, Granato M, Hammerschmidt 2019;53:505–30. M, Kane DA, Warga RM, Beuchle D, Vogelsang L, Nüsslein-Volhard C. 47. Patton EE, Zon LI. The art and design of genetic screens: zebrafsh. Nat Zebrafsh pigmentation mutations and the processes of neural crest Rev Genet. 2001;2:956–66. development. Development. 1996;123:369–89. 48. Puebla O, Bermingham E, Guichard F, Whiteman E. Colour pattern as a 24. Kelsh RN, Inoue C, Momoi A, Kondoh H, Furutani-Seiki M, Ozato K, single trait driving speciation in Hypoplectrus coral reef fshes? Proc R Soc Wakamatsu Y. The Tomita collection of medaka pigmentation mutants as B Biol Sci. 2007;274:1265–71. a resource for understanding neural crest cell development. Mech Dev. 49. Rhyne AL, Tlusty MF, Schofeld PJ, Kaufman L, Morris JA Jr, Bruckner AW. 2004;121:841–59. Revealing the appetite of the marine aquarium fsh trade: the volume 25. Kelsh RN, Sosa KC, Owen JP, Yates CA. Zebrafsh adult pigment stem cells and biodiversity of fsh imported into the United States. PLoS ONE. are multipotent and form pigment cells by a progressive fate restriction 2012;7:e35808. process. BioEssays. 2017;39:1600234. 50. Richards Z, Hobbs J. Hybridisation on coral reefs and the conservation of 26. Kirchmaier S, Naruse K, Wittbrodt J, Loosli F. The genomic and genetic evolutionary novelty. Curr Zool. 2015;61:132–45. toolbox of the teleost medaka (Oryzias latipes). Genetics. 2015;199:905. 51. Richardson J, Lundegaard PR, Reynolds NL, Dorin JR, Porteous DJ, Jackson 27. Klaassen H, Wang Y, Adamski K, Rohner N, Kowalko JE. CRISPR mutagen‑ IJ, Patton EE. mc1r Pathway regulation of zebrafsh melanosome disper‑ esis confrms the role of oca2 in melanin pigmentation in Astyanax sion. Zebrafsh. 2008;5:289–95. mexicanus. Dev Biol. 2018;441:313–8. 52. Rolland J, Silvestro D, Litsios G, Faye L, Salamin N. Clownfshes evolution 28. Krauss J, Frohnhöfer HG, Walderich B, Maischein HM, Weiler C, Irion U, below and above the species level. Proc R Soc B Biol Sci. 2018. https://​doi.​ Nüsslein-Volhard C. Endothelin signalling in iridophore development and org/​10.​1098/​rspb.​2017.​1796. stripe pattern formation of zebrafsh. Biol Open. 2014;3:503–9. 53. Roux N, Logeux V, Trouillard N, Pillot R, Magré K, Salis P, Lecchini D, Bes‑ 29. Lister JA, Cooper C, Nguyen K, Modrell M, Grant K, Raible DW. Zebrafsh seau L, Laudet V, Romans P. A star is born again: methods for larval rearing Foxd3 is required for development of a subset of neural crest derivatives. of an emerging model organism, the False clownfsh Amphiprion ocellaris. Dev Biol. 2006;290:92–104. J Exp Zool Mol Dev Evol. 2021. https://​doi.​org/​10.​1002/​jez.b.​23028. Klann et al. EvoDevo (2021) 12:8 Page 16 of 16

54. Roux N, Salis P, Lee S-H, Besseau L, Laudet V. Anemonefsh, a model for 61. Tezuka A, Yamamoto H, Yokoyama J, van Oosterhout C, Kawata M. The Eco-Evo-Devo. EvoDevo. 2020;11:20. MC1R gene in the guppy (Poecilia reticulata): genotypic and phenotypic 55. Salis P, Lorin T, Laudet V, Frédérich B. Magic traits in magic fsh: polymorphisms. BMC Res Notes. 2011;4:31. understanding color pattern evolution using reef fsh. Trends Genet. 62. Titus B, Benedict MC, Laroche R, Gusmão LC, Van Deusen V, Chiodo T, 2019;35:265–78. Meyer CP, Berumen ML, Bartholomew A, Yanagi K, Reimer JD, Fujii T, Daly 56. Salis P, Lorin T, Lewis V, Rey C, Marcionetti A, Escande ML, Roux N, Besseau M, Rodríguez E. Phylogenetic relationships among the clownfsh-hosting L, Salamin N, Sémon M, Parichy D, Volf JN, Laudet V. Developmental and sea anemones. Mol Phylogenet Evol. 2019;139:106526. comparative transcriptomic identifcation of iridophore contribution to 63. Tsetskhladze ZR, Canfeld VA, Ang KC, Wentzel SM, Reid KP, Berg AS, white barring in clownfsh. Pigment Cell Melanoma Res. 2019;32:391–402. Johnson SL, Kawakami K, Cheng KC. Functional assessment of human 57. Salis P, Roux N, Huang D, Marcionetti A, Mouginot P, Reynaud M, Salles coding mutations afecting skin pigmentation using zebrafsh. PLoS ONE. O, Salamin N, Pujol B, Parichy D, Planes S, Laudet V. Thyroid hormones 2012;7:e47398. regulate the formation and environmental plasticity of white bars in 64. Vissio PG, Darias MJ, Di Yorio MP, Pérez Sirkin DI, Delgadin TH. Fish skin clownfshes. PNAS. 2021. https://​doi.​org/​10.​1073/​pnas.​21016​34118. pigmentation in aquaculture: The infuence of rearing conditions and its 58. Salis P, Roux N, Soulat O, Lecchini D, Laudet V, Frédérich B. Ontogenetic neuroendocrine regulation. Gen Comp Endocrinol. 2021;301:113662. and phylogenetic simplifcation during white stripe evolution in clown‑ 65. Yasir I, Qin JG. “Efect of dietary carotenoids on skin color and pigments fshes. BMC Biol. 2018;16:90. of False Clownfsh Amphiprion ocellaris, Cuvier.” J World Aquac Soc. 59. Schartl M, Larue L, Goda M, Bosenberg MW, Hashimoto H, Kelsh RN. What 2010;41:308–18. is a vertebrate pigment cell? Pigment Cell Melanoma Res. 2016;29:8–14. 60. Taylor RS, Manseau M, Horn RL, Keobouasone S, Golding GB, Wilson PJ. The role of introgression and ecotypic parallelism in delineating intraspe‑ Publisher’s Note cifc conservation units. Mol Ecol. 2020;29:2793–809. Springer Nature remains neutral with regard to jurisdictional claims in pub‑ lished maps and institutional afliations.

Ready to submit your research ? Choose BMC and benefit from:

• fast, convenient online submission • thorough peer review by experienced researchers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations • maximum visibility for your research: over 100M website views per year

At BMC, research is always in progress.

Learn more biomedcentral.com/submissions