A general problem in applying an experimental embryological approach for lizards is that few tables of developmental stages are available. Whereas tables exist for viviparous lacertid (Dufaure and Hubert,1961) and iguanid (Lemus,1967), only several developmental stages have been described for oviparous lizards (e.g., Muthukkaruppan et al.,1970; Blanc,1974; Lemus et al.,1981; Sanger et al.,2008), which are thought to be useful for developmental analysis. Many lizards, even oviparous lizards, lay eggs with well-developed embryos (Shine,1983), and dissection or laparotomy is necessary if an early-stage embryo is required, although neither method is convenient. Blanc (1974) presented a table of developmental stages for the jewel chameleon Chamaeleo lateralis. Although the chameleon lays eggs with early stage (corresponding to stage 8 of Dufaure and Hubert,1961) embryos, they cannot be bred easily in captivity, and the chameleon therefore does not seem to be suitable for “egg-consuming” experimental analysis. The parthenogenetic teiid (Cuellar,1971) lays eggs with early embryos second only to Chamaeleo among lizards studied in Shine (1983; corresponding to stage 22 of Dufaure and Hubert,1961), but they breed seasonally and the number of eggs expected from each female in a season is less than 10 (Crews,1987). Recently, a table of developmental stages of Anolis was published (Sanger et al.,2008). This work, combined with the Anolis Genome Sequencing Project (Losos et al.,2005), provides a strong tool for biological studies of lizards. Anoles have leathery-flexible egg-shells, but gekkonid lizards are known to have eggs with hard shells. There has been simple description of the embryos for three gekkonid lizards (Werner,1971), but no tables of development have been published so far. We have established a staging table for the Madagascar ground gecko Paroedura pictus, which breed all year round in captivity like Anolis and lays eggs with hard shells (Kratochvıl et al.,2006) like birds.
We describe 27 developmental stages for the Madagascar ground gecko P. pictus from oviposition to hatching. Each egg was 10 mm in diameter and 13 mm in length on average. The shells were hard but thin, allowing us to determine the position of the embryo without any special equipment after 3 days of incubation. The vitelline membrane was already associated with the shell membrane at oviposition, so that the embryo was immobile, unlike that in birds, and no albumen was found. There was little difference in morphology of the embryos just after oviposition when the eggshell was still flexible. The average period of incubation to hatching was 60 days (28°C, 50–60% humidity), but one hatched after 55 days and another after 80 days. Under the breeding and incubating conditions used (Fig. 1), no obvious sex ratio bias was found among adults, probably due to their genetic, rather than temperature-dependent, sex determination (Blumberg et al.,2002).
A series of stages of whole embryos is shown in Figure 2, and the limb buds and limbs are shown in Figure 4. Figures 3 and 4 show the gene expression pattern of Fgf8, within the eye region (Lovicu and Overbeek,1998), ectoderm and endoderm layers of the developing pharyngeal arches (Abu-Issa et al.,2002), anterior neural ridge, metencephalon/mesencephalon boundary, tail bud, apical ectodermal ridge (AER) of limb buds (Ohuchi et al.,1994; Vogel et al.,1996), and also in the otic vesicle. The chondrogenesis of limb elements at earlier stages is shown by the expression of the cartilage condensation marker Aggrecan1 (Fig. 5), which is a proteoglycan component of the cartilage matrix (Chen et al.,1995), and skeletal elements were stained with Alcian blue/Alizarin red at later stages (Fig. 6).
0 dpo (Days Postoviposition)
Egg: The egg had a soft and sticky shell just after oviposition, and it became hard after 1 hr. Somites: 11–16; embryos with 9 somites were rarely observed. Brain: Anterior neuropore is closing (Fig. 3). Fgf8 expression is present in the metencephalon/mesencephalon boundary and the anterior neural ridge (Fig. 3). Branchial arches: 1–3 (Fig. 3). Optic vesicle: Visible. Otic vesicle: Clearly formed. Amnion: Completely covers embryo.
Egg: Clear liquid separation from yolk by gravity is visible from the outside. Somites: 17–22. Brain: Anterior neuropore closed (Fig. 3). Branchial arches: 3–4 (Fig. 3). Optic vesicle: Clearly formed.
Egg: Blood island is visible on one end of the egg from the outside. Somites: 23–29. Branchial arches: 4 (Fig. 3). Limb buds: Forelimb buds are start to swell. Face: Olfactory pit is formed (Fig. 3).
Somites: 30–35. Branchial arches: 5 (Fig. 3). Maxillary processes are visible (Fig. 3). Limb buds: Distinct swelling. The forelimb is wider than the hind limb (Fig. 4). Faint signal of Fgf8 expression is found in the forelimb AER (Fig. 3).
Somites: 36–40; extend to posterior of the hind limb bud. Eyes: Faintly pigmented at the rim. Limb buds: Forelimb is larger than hind limb (Fig. 4). The forelimb AER is readily detected by the typical expression pattern of Fgf8 (Fig. 3). Faint signal of Fgf8 expression is found in the hind limb AER (Fig. 3).
Egg: The position of the embryo can be detected from the outside by pigmentation of the eyes. Somites: 41–46. Eyes: Dorsal half of optic vesicle is pigmented. Limbs: Forelimb bud is still larger than hind limb bud. The forelimb bud begins to become asymmetric along the rostral–caudal axis, and it inclines posteriorly (Fig. 4). The hind limb AER is clearly shown by the expression of Fgf8 (Fig. 3).
Somites: Difficult to count accurately (more than 50). Limbs: Width of hind limb bud equals that of forelimb bud. Forelimbs are directed posteriorly. Eyes: Gray pigmented except on the underside.
Limbs: Width of hind limb bud slightly greater than forelimb bud. Forelimb buds of equal length and width (Fig. 4). Eyes: Fully pigmented but still grayish. Genital buds: Distinctively bulging at the roots of the hind limbs.
Eggs: The vitelline artery can be seen through the eggshell. Limbs: Forelimb length is slightly greater than its width. Eyes: Darkly pigmented. Face: Maxillary processes extend to the middle of the eye.
Eggs: Blood island covers almost the entire sphere of the egg. Limbs: Forelimb is distinctively longer than wide.
Limbs: Length of hind limb buds equal to width (Fig. 4).
Limbs: Proximal parts are narrowed and distal parts are flattened. Chondrogenesis in the stylopod and zeugopod is detected by Aggrecan1 staining. The knees and elbows are clearly formed. Hind limb bud equal in length to forelimb bud at this stage (Figs. 4, 5).
Limbs: Only one digit (digit IV) is visible in the autopod, and slight condensation of the digit III is shown as weak Aggrecan1 expression (Fig. 5). It shows posterior to anterior Aggrecan1 expression for cartilage condensation.
Limbs: Cartilage condensation of the forelimb digit II, III, IV, V, and the hind limb digit III, VI, V are seen, and slight condensation of the forelimb digit 1 and the hindlimb digit II are shown as weak Aggrecan1 expression (Fig. 5). Face: Mandibular processes extend to the anterior end of the eye.
Limbs: All digits clearly show expression of Aggrecan1, and slight condensation of the forelimb digit 1 is shown as weak Aggrecan1 expression (Fig. 5). Skin: Scale anlages are being formed and observed as skin roughness by PFA fixation and methanol dehydration.
Limbs: All digits can be recognized without staining (Fig. 4). Aggrecan1 expression clearly shows all five digits with joints in both the forelimb and hind limb (Fig. 5).
Limbs: Webs extend to the tip of the digits. The hind limb bud is significantly larger than the forelimb bud at this stage (Fig. 4). Face: Mouth is closed (Fig. 2). Skin: Scale anlages are well developed and observed as skin roughness on live embryos.
Limbs: Each digit is separated, but webs remain between the proximal parts of each digit (Fig. 4).
Limbs: Slight webbing can be seen between digits. Anterior–posterior asymmetry of the autopod is obvious in the hind limb (Fig. 4).
Limbs: Claws are seen on all digits (Fig. 4). All of the webs between digits are gone. Skeletal morphogenesis is almost completed.
Limbs: Fingertip morphologies are slightly different from those in the previous stage, especially in the hind limb where tip of digit IV is hollowed at the root of the claw, although it is almost identical to that in the previous stage in the forelimb (Fig. 4). Ossification in the limb skeleton is starting in the middle of each long bone. Phalangeal formula is 2-3-4-5-3 for both forelimb and hind limb (Fig. 6). Skin: Pigmentation has started, but the skin is still almost completely translucent.
Skin: Pigmentation is in progress, but the color pattern is still faint. Scales can be detected at high magnification.
Skin: Color pattern is distinct.
Skin: Translucent and mucosa-like surface texture. Limbs: Central parts of long bones are well ossified (Fig. 6).
Skin: Scales are visible. Genital organs: Still protruded externally.
Eggs: A few eggs hatch at this time. Skin: Scales are clearly visible. Genital organs: Being held inside the body.
Eggs: Hatching; just before hatching, color pattern is seen through the eggshell. Skin: Dry and whitish, ready to shed. Yolk sac: Some individuals still have the yolk sac, but this is completely absorbed in 1 day.
Most lizards produce a soft, leathery eggshell, but gekkonid geckoes, including Paroedura, form a hard eggshell like most Archosauromorpha (Packard et al.,1982). This feature makes it easy to treat and observe the embryos in ovo (Fig. 1B). Sanger et al. (2008) presented a table of developmental stages for Anolis, and a database of the Anolis Genome Sequencing Project (Losos et al.,2005) is available online. Anole eggs, however, are not easy to treat after oviposition because they have soft leathery shells. Eggs with hard shells can be windowed (Fig. 1B), micro-operated, and subsequently incubated, and P. pictus can be bred year-round in captivity like Anolis. Moreover, we were able to clone several gene cDNA fragments of P. pictus easily by referring to the Anolis Genome Database. Therefore P. pictus would be a useful model animal for experimental studies on lizard development.
Mammals and birds have mainly been used as representatives of amniote animals in experimental embryology. They are markedly different from each other, so it is difficult to investigate the evolutionary process from amphibian to amniote by studies of these two end groups alone. It would be valuable to have a range of reptiles available for experimental embryology as additional amniote models. Some preliminary work has been done on turtles (Burke,1989; Gilbert et al.,2001; Tokita and Kuratani,2001; Nagashima et al.,2007) and crocodiles (Ferguson,1981,1987; Tissir et al.,2003), but squamates form the largest and most diverse of living reptilian groups. The position of Testudines (turtles, tortoises) remains uncertain, but Lepidosauromorpha (squamates and related groups) are thought to have diverged from the Archosauromorpha (birds, dinosaurs, crocodiles, and related groups) before 250 Ma (Evans,2003; Kumazawa,2007). Among squamates, snakes and slow worms (Anguis) have been used for the study of limblessness (Cohn and Tickle,1999; Raynaud et al.,1995; Gomez et al.,2008), Hemiergis lizards have been used to study digit loss (Shapiro et al.,2003), and anoles have been introduced as representative squamates for developmental studies on morphological diversity among Anolis species (Sanger et al.,2008). However, given the flexibility of its shell, the potential of Anolis for experimental evo-devo studies is limited. In contrast, gecko eggs are useful for in ovo observation and manipulation of the late embryo because they have hard rigid shells, which are easily windowed and re-sealed (Fig. 1B). Moreover, Anolis and Paroedura are representatives of distinct major squamate clades, Iguania and Gekkota, respectively, and it is clearly advantageous to have squamate diversity represented as fully as possible.
Comparison with other amniote embryos shows some characteristics of the gecko embryo. P. pictus embryonic stage at oviposition corresponds to Lacerta stage 22–24 of Dufaure and Hubert (1961), mouse stage 13–14 of Theiler (1989), chick stage 11–13 of Hamburger and Hamilton (1951), and is slightly earlier than the late pre–limb-bud stage of Anolis (Sanger et al.,2008). This is one of the earliest stages at oviposition among squamates investigated, third only to that of the chameleon (Blanc,1974) and the whiptail lizard (Cuellar,1971) among species described in Shine (1983). While 16 somites were formed a day in the chick embryo, only 5–6 somites were formed a day in the gecko embryo. It takes 60 days from oviposition to hatching for the gecko, and it takes 19 days from Hamburger and Hamilton stage (HH) 12 to hatching for the chick. Thus, the gecko embryo incubated at 28°C developed approximately three times slower than the chick embryo incubated at 37°C and twice as slow as Anolis incubated at 27°C (Sanger et al.,2008). The average incubation periods for the agamid lizard, iguanid lizard, skinks and geckoes at 28°C have been shown to be around two months (Muthukkaruppan et al.,1970; Lemus et al.,1981; Ji et al.,2002; Thompson and Russell,1999; Viets et al.,1993; Werner,1971). Taken together with results of other studies (e.g., Hubert,1985), these observations suggest that anole embryos develop characteristically faster than do embryos of other lizards even after consideration of their stage differences at oviposition. The developmental speed of the gecko is thought to be a common speed among lizards.
We have described several points of embryonic morphology, limb buds in particular. The external morphology of the limb bud and limb is often considered to be a guidepost of embryonic staging (see Hamburger and Hamilton,1951; Sanger et al.,2008, for example). In the gecko embryo, both limb buds start to swell and develop at 2 dpo, and this corresponds to the early limb bud of the chick limb bud at HH17/18 and mouse limb bud at TS15. The position of swelling for the limb field was at the level of somites 6 to 13 for the forelimb bud and at the level of somites 26 to 32 for the hind limb bud. Corresponding regions in the chick and mouse embryos are located at somites 15–20 (chick) and somites 8/9–13/14 (mouse) for the forelimb buds and at somites 25/26–31/32 (chick) and somites 23/24–28/29 (mouse; Burke,1995), indicating that the gecko embryo has a relatively long space for the flank region between the forelimb and hind limb buds. Outgrowth of the forelimb bud was initiated earlier than that of the hind limb bud, and the forelimb bud was always larger than the hind limb bud. This tendency toward earlier initiation of forelimb bud outgrowth is also seen in the chick and mouse limb development, whereas hind limb development precedes forelimb development in the chick embryo. In the 12 dpo gecko embryo, forelimb and hind limb buds showed morphological resemblance to those of the HH25 chick embryo and TS19 mouse embryo. They showed a paddle-like shape with slightly flattened distal parts. In both chick and mouse embryos, it takes 2.5 days for limb development to progress from HH17/18 to HH25 and to progress from TS14 to TS19. Thus, development in gecko embryos proceeds much more slowly than that in chick and mouse embryos. After 2 weeks of incubation, some differences to mouse or chick embryos were seen. Gecko autopods had extended anteriorly and posteriorly much more than those in chick and mouse embryos (Fig. 4; compare with the limbs in Hamburger and Hamilton,1951; Theiler,1989). It is possible that this extension reflects the relatively long digit I and digit V. At 14 dpo, the first visible condensation of digits is detectable on the line of distal branching and extension of the postaxial elements (ulna and fibula) in the zeugopod region. This line is termed the “primary axis,” and the first digit visible on the primary axis is classified as digit IV (see Burke and Feduccia,1997, and references therein). As in alligator and turtle limbs, we identified the first visible digit as digit IV in the gecko limb bud. Subsequently, digit V appears immediately posterior to the primary axis, and digits III, II, and I form in a posterior-to-anterior sequence.
Although there are many differences among amniote embryos (e.g., in contrast to the chick embryo; Bininda-Emonds et al.,2007, chondrogenesis of the forelimb occurred slightly earlier than that of the hind limb in the gecko embryo, see also Richardson et al.,1997), the preliminary results discussed above demonstrate the potential for comparing and evaluating the differences among distant species by combining tables of developmental stages and by appropriate analysis. For example, the scale placode of the gecko has been shown to arise by 18 dpo. When comparing limb chondrogenesis with that of the chick and mouse, gecko 18 dpo corresponds to chick HH29–HH30 and mouse TS21–22. Feather bud formation has been shown to begin at HH29–HH30 (Hamberger and Hamilton,1951; Widelitz et al.,2000), and hair placode formation has been shown to begin at TS21-22 (Theiler,1989; Millar,2002). Thus, limb chondrogenesis and integument morphogenesis proceed coordinately in these three species. There has been an accumulation of data on feather morphogenesis and hair morphogenesis obtained from experimental analysis (Wu et al.,2004; Sawyer et al.,2005), and a similar approach can now be applied to lizards as we have succeeded in operating and subsequently incubating gecko embryos (zone of polarizing activity [ZPA] transplantation experiments, data not shown). This would allow further understanding and elucidation of evolutionary processes of amniote integuments. The results of further studies on the development of Paroedura should provide valuable insights into evolutionary processes underlying amniote characters as a whole, but also the roots of variation (axial length, limb proportions, skull differences) within Squamata as a major living clade.
Collection of Eggs and Embryos
All of the eggs used for staging were obtained from individuals of Paroedura pictus, which were hatched and kept in our laboratory. The breeding room was maintained in a 14-hr: 10-hr light–dark cycle at 28°C with 50–60% humidity all year round. Water dishes and shelters were set on a substrate of sand (Fig. 1). The geckoes were fed crickets dusted with calcium powder 4–7 times a week. Females were allowed to oviposit naturally. Each female laid two eggs once every 7 to 10 days. Eggs used to determine normal embryonic stages were collected twice a day and incubated at 28°C until opened. Embryos were dissected out from the eggs and observed in Tyrode saline (Tyrode,1910). At least five embryos were examined for each period of incubation.
Partial fragments of P. pictus cDNAs of Aggrecan1 and Fgf8 were obtained by degenerate polymerase chain reaction (PCR) with a cDNA template from P. pictus 7 dpo (days postoviposition) embryos. Degenerate primer sequences were as follows: aggrecan1F, GAA RTT CAC CTT CCM RGA RGC; aggrecan1R, TRT ARC AGA TGG CRT CRT ANC; fgf8F, GGA AGC AGC TCC GGA TCT AYC ARY TNT A; fgf8R, TGG CCC TTG GGG TAC CKY TTC ATR AAR T. PCR conditions for Aggrecan1 were 4 min at 95°C; 40 cycles of 30 sec at 95°C, 30 sec at 60°C, and 2 min at 72°C; and 15 min at 72°C. PCR conditions for Fgf8 were 3 min at 95°C; 40 cycles of 45 sec at 94°C, 30 sec at 50°C, and 2 min at 72°C; and 5 min at 72°C. Obtained fragments were cloned by using a TOPO cloning Kit (Invitrogen), confirmed by sequencing (GenBank accession numbers: Aggrecan1, EU927304; Fgf8, EU927305), and used as templates for RNA probe synthesis as described in the digoxigenin (DIG) RNA labeling kit protocol (Roche Applied Science).
In Situ Hybridization
Embryos were fixed in 4% phosphate-buffered paraformaldehyde and dehydrated in graded methanol. Then the embryos were transferred into 5% H2O2/methanol for the purpose of inactivation of intrinsic enzymes and subsequently transferred into methanol. Whole-mount in situ hybridization was performed as described previously (Yonei et al.,1995; Yonei-Tamura et al.,1999). DIG-labeled RNA probes for Fgf8 or Aggrecan1 were used for whole-mount in situ hybridization. Embryos were treated with 1 and 40 μg/ml Proteinase K for Fgf8 and Aggrecan1, respectively.
Alcian Blue/Alizarin Red
Embryos were fixed in 10% formalin/Tyrode and dehydrated by soaking in ethanol. For pigmented embryos, 1/10 vol. H2O2 was added for bleaching. The composition of Alcian blue/Alizarin red solution was 0.005% alizarin red, 0.03% Alcian blue, 60% ethanol, and 5% acetic acid, slightly modified from that reported by Masaki and Ide (2007). Embryos were stained for 3–10 days and cleared with 1% KOH for several hours and then stored in glycerol.
We thank Susan Evans for valuable discussion and critical reading of the manuscript. We also thank Hitoshi Yokoyama for his support on applying for permission for animal experiments. The authors thank Q. P. EGG CORPORATION for Calhope (Eggshell calcium powder) sample donation. M.M. and K.T. were funded by the Ministry of Education, Culture, Sports, Science and Technology of Japan; K.T. was funded by KAKENHI (Grant-in-Aid for Scientific Research) on Priority Areas “Comparative Genomics” and the Toray Science Foundation; and M.N. was funded by the Fujiwara Natural History Foundation of Japan.