The skeletal ontogeny of Astatotilapia burtoni – a direct-developing model system for the evolution and development of the teleost body plan
The experimental approach to the evolution and development of the vertebrate skeleton has to a large extent relied on “direct-developing” amniote model organisms, such as the mouse and the chicken. These organisms can however only be partially informative where it concerns secondarily lost features or anatomical novelties not present in their lineages. The widely used anamniotes Xenopus and zebrafish are “indirect-developing” organisms that proceed through an extended time as free-living larvae, before adopting many aspects of their adult morphology, complicating experiments at these stages, and increasing the risk for lethal pleiotropic effects using genetic strategies.
Here, we provide a detailed description of the development of the osteology of the African mouthbrooding cichlid Astatotilapia burtoni, primarily focusing on the trunk (spinal column, ribs and epicentrals) and the appendicular skeleton (pectoral, pelvic, dorsal, anal, caudal fins and scales), and to a lesser extent on the cranium. We show that this species has an extremely “direct” mode of development, attains an adult body plan within 2 weeks after fertilization while living off its yolk supply only, and does not pass through a prolonged larval period.
As husbandry of this species is easy, generation time is short, and the species is amenable to genetic targeting strategies through microinjection, we suggest that the use of this direct-developing cichlid will provide a valuable model system for the study of the vertebrate body plan, particularly where it concerns the evolution and development of fish or teleost specific traits. Based on our results we comment on the development of the homocercal caudal fin, on shared ontogenetic patterns between pectoral and pelvic girdles, and on the evolution of fin spines as novelty in acanthomorph fishes. We discuss the differences between “direct” and “indirect” developing actinopterygians using a comparison between zebrafish and A. burtoni development.
KeywordsDirect-development Indirect-development Astatotilapia burtoni Embryonic development Axial skeleton Appendicular skeleton Evo-devo Fin Teleost
Days post fertilization
Hours post fertilization
Teleost fishes are undoubtedly amongst the most diverse of vertebrates [1, 2, 3, 4]. It therefore may seem paradoxical that the contribution of the fish model systems to the understanding of the development and evolution of the natural diversity of vertebrate body plans is relatively minor. This observation is made while acknowledging the enormous contribution of the model fish species zebrafish and medaka to the understanding of the processes of early development, such as germ layer induction, gastrulation and the subsequent formation of the anterior-posterior body axis (e.g. [5, 6, 7, 8, 9]). However, anatomical structures that develop later during development have remained surprisingly underexposed compared to for instance mammals. Where we have a very good understanding of the genetic networks underlying the differentiation and patterning of the axial and appendicular skeleton in mice, and the potential consequences of the evolution of these networks for changes in the body plan (e.g. [10, 11, 12, 13, 14, 15, 16, 17]), there is less experimental evidence regarding the orthologous networks from fish. In many aspects this goes beyond a mere confirmation of the conservation of deeply homologous gene networks, but concerns essential questions regarding ancestral vertebrate characters that have disappeared from tetrapods, or novelties that arose in the fish lineages. Open questions in this field for instance relate to the transcription factor code that determines the positioning of median fins, the differentiation of dermal versus endochondral ossifying fin elements (including the secondary loss of the dermal part in tetrapods), and the further diversification of the fins and scale skeleton in teleosts.
One restraining factor appears related to the choice of fish model systems. Embryologists frequently make the distinction between “direct” and “indirect” developing organisms [18, 19]. The term “direct-development” is used for organisms whose embryos develop directly with adult features (i.e. resembling miniature adults). Examples of such species are amniotes, such as the mouse and the chicken. “Indirect-developing” organisms pass through a prolonged free-living and feeding larval stage that only partially resembles the adult [4, 18, 19]. The typical example of indirect-development is as occurs in frog species that know a tadpole stage as part of their life history [20, 21]. This type of development is characteristic for many anamniotes, and indirect-development occurs in many gradations, and also may differ for different characters within a given species [4, 18, 19]. In ray-finned fish (actinopterygians), the occurrence of indirect or direct modes of development is strongly related to the life history strategy of the species, and the presence of a distinct larval stage (also see the discussion section) [18, 19]. Direct-developing species usually have large yolk rich eggs, and the embryo completes its development while living off the maternal yolk supply, transforming directly into a feeding juvenile (i.e. a “definitive phenotype”) [18, 19]. Indirect-developing species usually have sparser yolk supply, and the onset of feeding precedes the completion of the body plan. In these species, development is only completed during a free feeding larval stage before transformation into a juvenile fish occurs [18, 19].
The most frequently used fish model system, zebrafish (Danio rerio), follows a typical pattern of indirect-development, with many adult features, such as the dermal skeleton and large parts of the trunk endoskeleton, only developing after several weeks of larval life [22, 23, 24, 25]. The development of medaka (Oryzias latipes) is somewhat more direct, with a developed caudal fin present at hatching, but the other median fins appear only after several weeks in swimming and feeding larvae . Such indirect-development brings a number of complications to the study of the development of adult characters. For genetic strategies, it is relevant that most developmental genes are involved in several biological processes (pleiotropy). Only very few zebrafish mutants identified in screens for embryonic characters  do survive until adulthood [7, 8]. For instance, one developmental bottleneck that has been reported in mutant zebrafish seems related to impaired swim bladder development [6, 26, 27]. This defect becomes lethal around the start of feeding stages, and prevents these fish from developing adult traits . Forward genetic screens focusing on adult phenotypes will per definition circumvent this problem for the mutated genes identified [7, 8]. However, this strategy will not detect genes whose early loss of function is not compatible with the formation of a coherent organism that is able to perform tasks related to feeding, locomotion and cognition. As important transcription factors and signal transduction pathways are used repeatedly during development [29, 30], focusing on viable adult mutant phenotypes will bias towards genes primarily involved in restricted terminal differentiation events, omitting numerous genes that contribute fundamentally to the establishment of the Bauplan. A further complication related to indirect-development is that growth amongst free feeding larvae varies strongly [22, 25, 31], making it difficult to accurately predict development and to obtain synchronized series of material. Therefore, a direct-developing fish model system, i.e. a species that shows a rapid development from egg to juvenile, without relying on an extended free-swimming larval period, would be a welcome addition to the field. This is particularly relevant where it concerns the understanding of adult characters of the fish body plan that may otherwise remain partially inaccessible. Here, we describe the development of the skeleton of the Haplochromine mouthbrooding cichlid Astatotilapia burtoni, with particular focus on the post-cranial parts. Mouthbrooding species are known to generally have a more direct development than free spawning species , and we show that this species develops an adult body plan and skeleton within 2 weeks after fertilization, without the presence of a specialized larval period.
A. burtoni exhibits strong sexual dimorphism, with colorful dominant males and plain females (Fig. 1a). Males are characterized by the “egg spots” on the anal fin, which play an important function in the mating ritual of Haplochromine cichlids, functioning as egg dummies [33, 37, 38]. During mating, the female deposits her eggs and takes them up in her mouth, after which the male fertilizes them while she approaches his egg spots [32, 39]. Mating usually takes place in the early morning hours, takes 5–15 min and its occurrence is frequently only noticed subsequently through the presence of mouthbrooding females, which can be recognized by the angular protrusion of their lower jaw, forming a pronounced “chin” (Fig. 1d). Embryos can be collected by gently opening the female’s mouth inside a container filled with aquarium water, and letting them drop out by gravity. Females of which eggs were removed typically spawn again within 2 weeks.
General overview of A. burtoni embryonic development
Eggs and early embryos of A. burtoni typically have an oblong shape and tend to rest on their side (which is different from zebrafish or Xenopus eggs, which usually have the animal pole oriented upwards). The blastodisc is located on the animal pole of the yolk (blastula shown Fig. 2a, stage 5). In contrast to zebrafish or medaka, the yolk is not transparent but opaque yellow, which at early stages (fertilization day to 1 dpf) can make observations on the embryonic Anlagen difficult. On the day of fertilization, the embryos pass through blastula stages and enter gastrulation. Gastrulation is completed by the second day, and the anterior-posterior axis can be distinguished by the formed neural tube (Fig. 2b, c, stage 10). Anteriorly, a thickening of the neural tube indicates the forming head (Fig. 2c). At these early stages, the chorionic membrane surrounds the embryos, and dechorionation is difficult without puncturing the yolk. Damaged embryos will usually not survive and will dissociate rapidly. If dechorionated embryos are required at these stages (for instance for in situ hybridization), it is usually more practical to dechorionate them after fixation, or alternatively, to dechorionate and immediately fix the embryos.
At 2 dpf, embryos (Fig. 2d, e, stage 14) have a clearly formed head with unpigmented eyes, otic vesicles, and the isthmus of the midbrain-hindbrain boundary is apparent. At 3 dpf, head development has progressed, and eyes have developed light pigmentation. (Fig. 2f, g, stage 15). A beating heart with relatively little blood circulation can be observed, as well as pectoral fin buds (Fig. 2g). At 4 dpf (Fig. 2h, i, stage 17), pigmentation in the eyes has increased, and the head has lifted upwards from the yolk. The heart is well developed with strong blood circulation, and extensive vasculature is running across the yolk. At 4 dpf, embryos hatch from their chorionic membrane, which can often be induced by gentle swirling of the petri dishes or pipetting up and down of the embryos.
After hatching (Fig. 3a, stage 18), embryos remain oriented on their sides until 7 dpf when they will right themselves and start swimming actively. Although many aspects of the body plan are apparent at this stage, we prefer to refer to them as “embryos” instead of “larvae”, according to the definitions used by Ballon 1999  (and see discussion), since they do not actively feed yet. Between day 7 and day 13 (stage 20 – stage 26), embryonic development is completed, and the body wall has closed over the gut (Fig. 3b).
After 2 weeks of development, the juvenile fish have attained the adult body plan and start feeding. This is also approximately the time when juveniles would first venture out of the mother’s mouth during natural incubation . In comparison to indirect-developing species, such as zebrafish  and medaka , development in A. burtoni stands out by its rapid development of the adult morphology, without the interlude of a prolonged, free-living and feeding larval stage.
Osteology and axial meristics of the A. burtoni post-cranial body plan
The axial skeleton of A. burtoni consists of 27 (N = 2) or 28 (N = 7) vertebrae including the urostyle, the terminal vertebral element that forms part of the caudal fin urophore complex. The axial skeleton can be divided into pre-caudal, caudal and ural regions. The pre-caudal region consists of two anterior ribless vertebrae (N = 9), followed by 11 (N = 2) or 10 (N = 7) rib-bearing vertebrae. In addition to the ribs, additional rib-like dermal bones are present. These elements have been commonly referred to as “dorsal ribs”, and their homology to ribs in non-teleosts has been intensely disputed (see Britz and Barsch ). The current view is that these elements are teleost specific intermuscular bones that form in the myoseptum between epaxial and hypaxial musculatures, and have no homology to the ribs present in basal actinopterygians, sarcopterygian fishes or tetrapods [53, 54]. Here, we will refer to these structures as epicentrals following Patterson and Johnson 1995 . The epicentrals are present from the 1st till the 9th (N = 1), 10th (N = 7) or 11th (N = 1) vertebra, and project laterally from the centra at a more dorsal position than the true ribs. Neural arches, which envelope the spinal cord, are present on all vertebrae except the urostyle. Haemal arches, which envelope and protect vasculature, define the start of the caudal region, and their occurrence is always mutually exclusive with the presence of ribs (N = 9). The vertebral column ends in the urophore complex, here defined following Monod, 1968 , as the vertebral elements that directly support the caudal fin rays, which consists in A. burtoni of two vertebrae plus the urostyle (see the paragraph on the caudal fin below).
The dorsal and anal fins are supported proximally by endochondral radials that are lodged internally between the neural and haemal spines (which are formed by the distal extensions of the neural and haemal arches). The radials articulate through the distal radials with the dermal fin rays, which form the external support of the fins. We prefer not to use the common ichthyological term “pterygiophore” here, which can refer to radials as well as rays in the anal and dorsal fins. Therefore, this indication is not informative for the embryonic origin of these elements, nor for the homology relationships between the supports of the median and paired fins, for the latter of which the expression pterygiophore is in general not used.
In cichlids, the dorsal and anal fins consist of two different classes of fin rays, namely spines anteriorly and soft-rays posteriorly (Fig. 4 a, b). Spines differ from soft-rays by stronger ossification, lack of segmentation, and ending in a sharp point instead of bifurcating [55, 56]. Fin spines have evolved as a defense mechanism against gape-limited predators , and are considered a hallmark innovation for acanthomorph fish [55, 58], the relatively young but very successful radiation of teleosts to which cichlids belong (although it needs to be noted that fin spines have evolved multiple times independently in lineages such as carps, catfish and sturgeons ). The dorsal fin of A. burtoni has 13 (N = 2) or 14 (N = 7) spines, and 9 (N = 3) or 10 (N = 6) soft-rays. In the anal fin there are 3 (N = 9) spines and 9 (N = 7) or 8 (N = 2) soft-rays. The osteology of pectoral, pelvic and caudal fin complexes are discussed in more detail in their respective sections (Figs 7, 9, 10 respectively).
Development of the axial skeleton - vertebrae, arches and ribs
The vertebral column of vertebrates consists of two features: i) vertebral centers, which form around an embryonal notochord and form the central axis of the trunk skeleton, and ii) the haemal and neural arches, which attach to the centra ventrally and dorsally, and envelope the axial blood vessels and the spinal cord . Although in the adult skeleton, centra and arches are seamlessly fused into a single morphological unit, these structures have distinct developmental and evolutionary trajectories [51, 59]. The accessory elements to the spinal column in the pre-caudal region considered here are the ribs and the epicentrals [53, 54].
In vertebrates, somites not only contribute to the postcranial skeleton, but also to the formation of the posterior skull. In amniotes, the occipital somites are the anterior-most five somites, which form part of the skull base [60, 61, 62], while in anamniotes (axolotl and zebrafish), only the first three somites contribute to the occipital arch [52, 63, 64]. We used immuno-staining with the muscle sarcomere specific antibody MF20  in combination with Alcian blue staining to visualize the correspondence between the basioccipital and the somites in A. burtoni. The first basioccipital condensations, which form around 100 hpf, are visible in somite 1, and by 110 hpf extend up to the somite 2–3 border, while the neural arch condensations of the first vertebra develop inside somite 4 (Fig. 5b). This pattern is consistent with the anamniote pattern, whereby the first three somites contribute to the occiput. In actinopterygian fishes, the ancestral condition, as present in basal non-teleosts (e.g. Amia calva and gar (Lepisosteus sp.)) , is the fusion of one or more vertebral elements with the posterior skull. As has been previously reported for other teleost species, we do not observe any contribution of pre-formed vertebral elements to the skull. Therefore, the occipital-trunk transition in A. burtoni conforms to the generalized teleostean condition as proposed by Patterson and Johnson . Namely, it consists of three myotomal segments associating with the skull, which as in other anamniotes, derives from the first three occipital somites, followed by two ribless vertebrae, and the first rib is present on the third vertebrae.
Development of the dorsal and anal fins
Not only do dorsal and anal fins develop virtually synchronously, also the anterior spiny part and the posterior soft-rayed part develop largely simultaneously. Only the middle soft-ray domain is slightly ahead of the more posterior soft-ray and the anterior spine domains. Fin spines are considered an evolutionary invention associated with the rise of acanthomorph (spiny-rayed) fishes , and it has been argued that the spiny and soft-ray domains of the dorsal fin comprise distinct developmental and evolutionary modules . Their developmental sequence in A. burtoni strongly suggests that they are produced under influence of the same developmental signals and that they acquire their different identity only later through subsequent patterning as typical for serially homologous elements [69, 70].
Development of the pectoral fins and girdles
During the earliest stages of fin development (84–102 hpf, Fig. 7c and stage I and II, Fig. 7d), the pectoral fin and girdle are visible as two separate cartilaginous Anlagen for the scapula-coracoid and the fin radials, respectively. The cartilaginous Anlage for the radials is generally referred to as endoskeletal disk. In A. burtoni this structure is very transient (84–102 hpf, Fig. 7c and stage I, II, Fig. 7d), in contrast to its prolonged persistence in indirect developing teleosts, such as zebrafish, where it remains present during larval stages and does not segment into the radials until the transition into adults  (and see discussion). Although the cleithrum is a dermally ossifying bone , we find that it stains weakly using Alcian blue before its ossification, and can be seen lying rostral from the other girdle and fin elements (indicated in 84–138 hpf, Fig. 7c). In the Alcian blue stained series, the post-cleithrum is difficult to detect but is visible through its differential contrast, and can be seen behind the endoskeletal disk and radials (indicated in 126 and 192 hpf, Fig. 7c).
Sequential segmentation of the endoskeletal disk gives rise to the mesopterygial radials. The first segmentation, which partially separates the domain of the two anterior radials from the domain of the two posterior radials, occurs around 126 hpf (Fig. 7c and stage III, Fig. 7d). This is followed by the next segmentation anteriorly and posteriorly at around 156 hpf (Fig. 7c and stage IV, Fig. 7d). Initially, the radials stay joined at their distal edge, and segmentation progresses in anterior to posterior direction reaching full segmentation at 240 hpf (Fig. 7c and stage VI, Fig. 7d). The propterygium becomes visible at the anterior edge of the partially segmented endoskeletal disk at 156 hpf (Fig. 7c and stage IV, Fig. 7d), although it is not clear whether its formation is due to a de novo cartilage condensation, or whether it segments off the endoskeletal disk. The first distal radials become visible around 174 hpf, form in an anterior to posterior sequence, and reach the position of the most posterior proximal radial at 240 hpf (and subsequently their development continues more posteriorly). As is the case for the propterygium, we have not been able to determine whether the distal radials form through de novo condensations or by splitting from the fringe of the preformed cartilaginous elements of the radials.
The most distal elements of the fins, the dermal fin rays, do not form via cartilage precursors, yet stain lightly using Alcian blue (as is the case for the cleithrum and the dorsal and anal fin rays). In teleosts it is now well established that the fin rays, like the radials, are derived from mesoderm [76, 77], and they form from mesenchyme within an epidermal sheet, the fin fold  (indicated in 102 to 138 hpf, Fig. 7c and stage I, II, Fig. 7d). The first fin rays become visible around 156 hpf (Fig. 7c and stage IV, Fig. 7d), and appear in anterior to posterior sequence (as reported in ref. ) until stage 240 hpf (Fig. 7c and stage VI, Fig. 7d), when they have formed along the full anterior-posterior margin of the fin.
Development of the pelvic fins and girdles
Contrary to the pectoral fins, the pelvic fins play a relatively subordinate role in fish locomotion, and have been frequently lost or become highly modified during evolution [80, 81, 82]. In the acanthomorph (spiny-rayed) fish, such as cichlids, the position of the pelvic girdle is located anteriorly below the pectoral girdle (in contrast to its ancestral position near the anus) [55, 83, 84] (see also Fig. 4a). During ontogeny, however, the pelvic appendages still arise at a conserved position around the anus, and migrate anteriorly to their final location after protrusion of the fin buds . During development, pelvic fins appear at a much later stage than pectoral fins. In cichlids, the two pelvic fins and girdles are located side by side on the ventral midline. In contrast to the pectoral fins, the pelvic fin complex includes only a single girdle element (often referred to as basipterygium [73, 83]), which directly articulates with the fin rays, and, unlike in zebrafish , there are no free radial elements in A. burtoni. The most anterior fin ray has the identity of a spine, and is unsegmented and unbranched (Fig. 9a).
In adult teleosts, the pectoral and pelvic appendages and girdles are very different in their appearance. Also in more basal fish species, there exist significant differences between pectoral and pelvic appendages, contributing to the discussion considering their serial homology . During the development of the pelvic and pectoral fins and girdles in A. burtoni, we noticed a hitherto, to our knowledge, undescribed pattern of similarity between the development of pelvic and pectoral appendages, whereby both first appear as a forked chondrogenic condensations corresponding to the girdle elements, followed by the formation of the endochondral disk in between the two distal girdle extensions. Perhaps as a result of the staging series used, this pattern has not been described in other teleosts, such as zebrafish  and sea bream . Although this pattern may reflect a homologous as well as an analogous developmental trajectory, it nevertheless provides a striking resemblance between these appendages, and could represent a basic ontogenetic trajectory underlying the formation of both pectoral and pelvic girdles and appendages, worthy of further studies.
Development of the caudal fin
The first elements to appear in A. burtoni, at approximately 100–120 hpf, are the chondrogenic condensations of the hypurals and parahypural, starting with hypural I (developmental sequence shown in Fig. 10c). The haemal and neural arches on pre-ural vertebrae 2 and 3 develop slightly later, around 132 hpf, in phase with the development of the haemal and neural arches of more anterior vertebrae. On the dorsal side of the urophore complex, this leaves a distinct gap at the position of the future epurals and uroneural. Also, during this developmental window (100–132 hpf), the notochord flexes strongly dorsally. Only at 156-180 hpf, the epurals and the terminal hypural V become visible as cartilaginous elements. The uroneural develops as a membrane bone without a cartilaginous intermediate and ossifies during subsequent stages of development (visible at 260 hpf – exact time point of ossification not determined). This pattern of development is highly similar to that described for other teleosts such as sea bream  and zebrafish [24, 48]. Interestingly, there appears to be a clear heterochronic shift in development of the derived hypurals/parahypural and epurals, and the flanking unmodified haemal and neural arch of the urophore complex. The unmodified haemal and neural arch form in concert with the other neural and haemal arches of the posterior axial skeleton, while most of the ventrally located hypurals/parahypural form earlier, and the dorsally located epurals form later. Therefore, the teleost urophore complex shows a temporal modularization between ventral and dorsal elements. We would like to suggest that this dorso-ventral heterochrony, with a precocious expansion of the ventral skeletal elements in combination with a delayed formation of dorsal elements, in fact might facilitate the dorsal flexion of the vertebral column, which is so characteristic of the homocercal caudal fin. Also, the formation of the uroneural, as a late ossifying membrane bone instead of as an endochondrally ossifying bone with a cartilage precursor (which is the default state for neural arches), appears compatible with this idea, as it could contribute to the process of dorsal flexion.
Development of scales in A. burtoni
We assayed the developmental appearance of scales using Alizarin red staining and in situ hybridization using the ossification marker collagenX. The first scale Anlagen become visible on day 7 along the posterior midline (7 dpf, Fig. 11b). By day 8, this line of squamation has extended anteriorly, and two new rows of scales have appeared, one ventrally and one dorsally of the initial row (8 dpf, Fig. 11b). By day 9, the three rows have continued expanding anteriorly, while at the same time again two more rows of scales have appeared ventrally and dorsally (9 dpf, Fig. 11b). This pattern of scale appearance resembles that described for other fish, such as zebrafish  and other teleosts . Ossification, or at least bone mineralization, as assayed by Alizarin red staining, occurs slightly later during development, and also proceeds in a posterior to anterior sequence, whereby at 13 dpf the scales along the posterior midline are mineralized, while more anteriorly mineralization is still absent (13 dpf, Fig. 11c).
Development of the cranium in A. burtoni
The focus of our description of the developmental osteology of A. burtoni has been on the post-cranial skeleton, particularly since a very detailed description of the cranial ontogeny is already available for the closely related Nile tilapia . For completeness, we include here a brief description of the development of the chondrified element of the head, and the subsequent ossification of the jaw and skull bones, as we have observed in A. burtoni.
Direct- versus indirect-development; A. burtoni versus zebrafish
There has been considerable discussion over the definition of indirect- versus direct-development in fishes and the criteria for the delimitation of embryonal, larval and adult life-stages. Part of this is due to the different prioritization of phenomena such as hatching and the onset of feeding by different authors, which has lead to incompatible views e.g. [19, 92]. Here, we follow the synthesis provided in Ballon 1999 , and consider the embryo to larva transition as the point where an embryo becomes dependent on exogenous feeding, but only so if at this point the formation of the body plan has not been completed (i.e. the embryo has not yet become a “definitive phenotype” ). In many teleost species, the timing of the embryo to larva transition is related to the amount of yolk provided in the egg. Sparse yolk provision is related to an early transition to the larval stage, and these larvae generally possess many embryonal characters, and still lack essential aspects of the body plan [18, 19]. Embryos from species with large yolk-rich eggs, such as Haplochromine cichlids, become only dependent on feeding at a relatively later stage, when nearly all aspects of the body plan have developed (i.e. a “definitive phenotype”), making them in effect transform directly from a non-feeding embryo into a feeding juvenile fish, without the interlude of a larval stage .
For the distinction between direct- versus indirect-development, we refer to “indirect-development” in species where the embryo transforms in a larva, and completes important aspects of the formation of its body plan during feeding stages. We refer to “direct-development” in cases where the embryo transforms directly into a juvenile, i.e. a “definitive phenotype” that possesses most aspects of the adult body plan, and lacks embryonic characters, such as the endoskeletal disk and the median fin folds. Therefore, indirect- versus direct-development is not related to the absolute speed of development, or to the relative sequence of appearance of developmental landmarks. Rather, the difference between these two developmental modes lies in the presence or absence of a larval period necessary for the completion of the body plan. We can illustrate this using the development of the pectoral fins in zebrafish (after the description by Grandel & Schulte-Merker 1998 ) and A. burtoni (this article) (Fig. 14). Both species progress through a very similar developmental sequence, namely formation of the endoskeletal disk (stage II), appearance of a median cleft in the disk (stage III), appearance of an additional dorsal and ventral cleft together with the first appearance anteriorly of distal radials and fin rays (stage IV), followed by a complete separation of the four fin radials as well as posterior generation of the full complement of distal radials and fin rays (stage VI) (stages given after Fig. 7). In A. burtoni, this sequence is completed within less than 1 week of development, between day three, with the first appearance of the endoskeletal disk, and day ten, when the fin has reached its definitive phenotype (Fig. 14). During this period (and in general before and shortly after onset of feeding), growth and development in A. burtoni are very homogenous and predictable (as also indicated by the low divergence in standard length, typical for direct-development (Fig. 14)).
In the indirect-developing zebrafish, the endoskeletal disk is also formed around day three post-fertilization, but has not yet started its further differentiation at the embryo to larva transition at 5 dpf, when feeding starts. After this point, the growth of the larvae slows down and becomes strongly heterogeneous (Fig. 14, standard length plot after Parichi et al. 2009 ). Hence, milestones in the development of the pectoral fins can only be reliably indicated with a reference to standard length . From the embryo to larva transition on, the further differentiation of the endoskeletal disk is not only slower, with the fastest individual completing the formation of the pectoral fins only at about 3 weeks to 1 month , it also is heterogeneous to such an extent that individuals from the same clutch can be weeks apart in reaching the same developmental landmark. Such relatively slow and heterogeneous development can of course immensely complicate experimental approaches. In this sense, A. burtoni can function as a complementary direct-developing model system to circumvent the disadvantages associated with indirect-development.
These observations on direct- and indirect-development of course leave the issue of metamorphosis in A. burtoni unaddressed. Considering the universal presence of a metamorphic step in vertebrates and teleosts [4, 93, 94], it is likely that some form of cryptic metamorphosis is present as has been suggested for direct developing species , something which would be worth studying by investigating the underlying thyroid hormone pathways in A. burtoni.
A. burtoni as a model system for evolutionary-developmental biology
A wide range of fish species (predominantly teleosts) is now being developed as model resources [7, 98]. Many of these fishes are selected for their phylogenetic position, rather than for their suitability to the wet-lab. Cichlid fishes have been in the spotlight for several decades because of their spectacular radiation, and their rapid evolution and speciation. In addition, A. burtoni has already acquired a place as a model system in neuro-behavioral and physiological research [99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119]. Although this species has many traits that would make it a useful developmental system, it does not seem to have caught the eye of embryologists yet. Husbandry of A. burtoni is straightforward, it has a relatively high quality genome sequence available , and the possibility for microinjection into the fertilized egg makes it tractable to genetic strategies, such as transgenesis  and gene editing . In our laboratory, we obtain high efficiency F0 mutation using CRISPR/Cas9, and we reach generation times of between 3 and 5 month.
We show here that the complete osteology of A. burtoni develops within a time span of less than 2 weeks, an aspect in which it differs strongly from established model fish model systems, such as zebrafish and medaka. This rapid development from egg to free swimming juvenile fish offers scope for different experimental approaches than in these other teleost. Together with its experimental tractability, this will make it a suitable system for studies on the fish body plan. Its further development may prove a great help in addressing the evo-devo of the truly spectacular radiation of teleost fishes.
For adult skeletal preparations nine fish were fixed in 4% buffered PFA, bleached using H2O2 in 1% KOH solution and transferred to 100% EtOH. For skeletal stains, samples were incubated for several days on a shaking platform at room temperature in 0.004% Alcian blue in 70% EtOH/30% glacial acetic acid. Subsequently they were washed overnight in 70% EtOH/30% glacial acetic acid to reduce background staining. Afterwards, fish were stained in 0.04% Alizarin red in 1%KOH followed by further clearing using enzymatic digestion with trypsin in saturated borax solution. Embryos and juvenile fish were treated similarly except that these were not digested using trypsin. For each stage reported at least 6 embryos for the same time point (+/− 6 h) were examined.
In situ hybridization
In situ hybridization was performed according to Woltering et al. 2009, 2014 [10, 121] with the difference that proteinase K treatment used was 8 min for 5 dpf, 10 min for 7 dpf and 12 min for 9 dpf. The collagenX probe (genbank locus: XM_005931031.2) was amplified from cDNA and cloned into the pGEMT vector (Promega) using primers FW: GTGAAAAAGGAAATGGAGGTGC, RV: TCACGTTGAAGCGATGAGGAAC.
MF20 immuno staining
Immuno-histological stainings for muscle sarcomeres were carried out using the MF20 antibody  obtained from the Developmental Studies Hybridoma Bank, Iowa. Briefly, embryos were fixed in MeOH at − 20 °C overnight and stained with Alcian blue as described above. Embryos were rehydrated to PBS, incubated in PBS 0.01% Triton X1000 for 2 h, washed 5 times for 5 min in PBS, blocked for minimally 2 h in PBS 10% heat inactivated lamb serum and incubated overnight at 4 °C with a dilution of 1/200 MF20 primary antibody in PBS supplemented with 10% heat inactivated lamb serum and 0.01% Tween 20. The next day embryos were washed a minimum of 5 times for 1 h at room temperature with PBS supplemented with 0.01% Tween-20 and subsequently incubated with secondary antibody (Goat-anti mouse., DyLight 488, Thermo Fisher Scientific #35502, diluted 1:200) overnight at 4 °C. The next day embryos were briefly washed in PBS 0.01% Tween-20 and imaged using standard fluorescence microscopy.
Animal husbandry and embryo culture
A. burtoni were kept in mating groups of 10–25 individuals at ~ 25 °C in the animal facility of the University of Konstanz. After collection, embryos were cultured in 100 ml plastic petri dishes in a 28 °C incubator in tapwater supplemented with 0,01μg/ml Methylene blue and penicillin/streptomycin (Sigma P4333) diluted 1:1000. To mimic the constant movement embryos would experience in a females buccal cavity, embryos were kept on a slow moving orbital shaker until 6 dph as the circulation at these stages seems to increase survival. After day 7 when embryos start to swim actively they were removed from the shaker. Particularly during pre-hatching stages it is essential to clean dishes from dead and morbid embryos as these can easily lead to infections of the whole clutch. Standard lengths (from the base of the caudal fin rays to the operculum) were measured on 10 embryos/juveniles a day from different clutches.
The authors would like to thank Claudius Kratochwil, Sandra Nappe, Dominique Leo, Yipeng Liang, Maggie Sefton, Jan Gerwin, Rebekka Höch, Alison Kickuth and other member of the Meyer laboratory for useful discussion, sharing of reagents and assistance in the lab.
This study received support from the University of Konstanz (JMW, AM), the Young Scholar Fund of the University of Konstanz (JMW), the Deutsche Forschugsgemeinschaft (AM & RFS), and the European Research Council (AM – grant number 293700 - GenAdap). None of the funding bodies had any participation in the collection, the interpretation or the analysis of the data or influence on the writing of the manuscript.
Availability of data and materials
A. burtoni stocks and or embryos are available upon reasonable request.
JMW concept and design of the study, performed experiments and wrote the manuscript; MH performed experiments and helped with writing the manuscript; RFS performed experiments and commented on the manuscript; VN performed experiments; AM contributed to the design of the study and the writing of the manuscript. All authors read and approved the manuscript.
Experiments were performed in accordance with the rules of the animal research facility of the University of Konstanz, Germany and have been granted permission by the Tierschutzbeauftragten (Az. #T15/05TFA).
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 1.Nelson JS. Fishes of the world. 4th ed. Hoboken, New Jersey: John Wiley & Sons; 2006.Google Scholar
- 2.Helfman GS, Collette BB, Facey DE. The diversity of fishes. Oxford: Blackwell Science; 1997.Google Scholar
- 20.Carl TF, Klymkowsky MW, Richardson MK, Hanken J. Direct-developing frogs: a novel system for studying the development and evolution of the vertebrate limb. Dev Biol. 1998;198(1):200.Google Scholar
- 30.Basson MA. Signaling in cell differentiation and morphogenesis. Cold Spring Harb Perspect Biol. 2012;4(6):1–21.Google Scholar
- 32.Fernald RD, Hirata NR. Ontogeny of social-behavior and body coloration in the African cichlid fish Haplochromis-Burtoni. Z Tierpsychol. 1979;50(2):180–7.Google Scholar
- 36.Astatotilapia burtoni. The IUCN Red List of Threatened Species 2006: e.T60462A12359157. [ https://doi.org/10.2305/IUCN.UK.2006.RLTS.T60462A12359157.en. ].
- 45.Keivany R. Pectoral girdle bones of eurypterygian fishes. Int J Aquat Biol. 2014;2(5):253–74.Google Scholar
- 46.Lauder GV. Function of the caudal fin during locomotion in fishes: kinematics, flow visualization, and evolutionary patterns. Am Zool. 2000;40(1):101–22.Google Scholar
- 50.Monod T. Le complexe urophore des poissons téléostéens. Dakar: Institut Fondamental d'Afrique Noire; 1968.Google Scholar
- 53.Patterson C, Johnson GD. The Intermuscular bones and ligaments of teleosten fishes, vol. 559. Washington, D.C.: Smithsonian institution pres; 1995.Google Scholar
- 55.Helfman GS, Colette BB, Douglas EF: The diversity of fishes. In: The diversity of fishes 10 edn. Oxford: Blackwell publishing; 1997. 32, 34, 35, 238, 239.Google Scholar
- 56.Moyle PB, Cech JJ: Fishes: an introduction to ichthyology. In: Fishes: an introduction to ichthyology. Englewood Cliffs: Prentice-Hall; 1982. 19, 288.Google Scholar
- 57.Price SA, Friedman ST, Wainwright PC. How predation shaped fish: the impact of fin spines on body form evolution across teleosts. Proc Biol Sci. 2015;282(1819):1–9.Google Scholar
- 66.Metscher B, Ahlberg P: Origin of the teleost tail: phylogenetic frameworks for developmental studies. In: Major events in early vertebrate evolution. Edited by P.E. A. London: CRC press; 2001. pp. 333–49.Google Scholar
- 71.Gegenbauer C. Elements of comparative anatomy. London: MacMillan and co.; 1878.Google Scholar
- 72.Mabee PM. Developmental data and phylogenetic systematics: evolution of the vertebrate limb. Am Zool. 2000;40(5):789–800.Google Scholar
- 78.Thorogood P. The development of th teleost fin and implications for our understanding of tetrapod limb evolution. In: Hinchcliff J, Hurle J, Summerbell D, editors. Developmental Patterning of the Vertebrate Limb. New York: Plenum Press; 1991.Google Scholar
- 86.Lauder GV. Flexible fins and fin rays as key transformations in ray-finned fish. In: Dial KP, Shubin N, Brainerd EL, editors. Great transformations in vertebrate evolution. Chicago: University of Chicago press; 2015. p. 31–45.Google Scholar
- 87.Gosline WA. The Perciform Caudal Skeleton. Copeia. 1961;(3):265. https://www.jstor.org/journal/copeia?decade=1960.
- 90.Sire JY, Arnulf I. The development of squamation in 4 Teleostean fishes with a survey of the literature. Jpn J Ichthyol. 1990;37(2):133–43.Google Scholar
- 92.Kovac V, Copp GH. Prelude: looking at early development in fishes. Environ Biol Fish. 1999;56(1–2):7–14.Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.