A million of insect species have been identified so far, displaying a staggering variety of adult morphologies. To elucidate mechanism how such insect morphologies are developed at a molecular level, we investigated developmental process of the two-spotted cricket, Gryllus bimaculatus, as a typical hemimetabola, and compared with that of Drosophila as an extensively studied typical holometabola. We analyzed expression patterns of homeobox genes of engrailed (en) and Distal-less (Dll) during development. In early embryos, en is expressed in the posterior compartments of body segments and developing appendages, while Dll is expressed in the distal region corresponding to the telopodite of developing appendages. Interestingly, these expression patterns are very similar to those observed in Drosophila imaginal discs. In the case of Dll, we found that its expression pattern, which is similar to each other in various appendages at early stages, changes in progress with elongation and segmentation, depending on the type of appendages. Late expression patterns of Dll are classified into three types: Dll expression in the entire region of the antenna, in a distal region of the cercus, and in distal and middle regions of the leg, maxillary and labial palpus, indicating that Dll expression patterns are closely related to segmentation patterns of the appendages. Furthermore, since Dll is intensely expressed in both sides of the femur-tibia articulation of the leg, we considered that Dll is involved in positioning of articulation during the late appendage development. Hence, our results indicated that although common molecules are involved in development of insect appendages, the variety of the morphologies depends on pattern and timing of their expressions.
Insects are vast in number and about a million of species have been identified so far. Their enormous variations in body forms bring ability of adaptation to various environments on the earth. Especially, insects bear several types of appendages such as legs, antenna, mandible, cercus, etc., and their shapes and functions are highly specialized. Recently, mechanisms on development of various appendages of Drosophila have been investigated from a molecular biological point of view (Bryant, 1993; Campbell et al., 1993; North and French, 1994; Bonini and Choi, 1995) and various molecules involved in development of appendages have been identified. For instance, it is known that a segment polarity gene of engrailed (en) is an essential gene for determination of the posterior compartment of Drosophila body segments (Karr et al., 1989; Patel, 1994a, b) and the imaginal discs (Kornberg et al., 1985). On the other hand, a homeobox gene, Distal-less (Dll), is known to one of the essential genes for development of the Drosophila appendage. Dll is concentrically expressed in appendage imaginal discs (Diaz-Benjumea et al., 1994). Since a Drosophila mutant lacking Dll activity lost the distal parts of appendages (Cohen and Jürgens, 1989; Cohen et al., 1989), Dll is considered to be involved in the proximodistal pattern formation in the appendage.
Although we obtained huge information on developmental mechanism of Drosophila appendages, we have still not known about the mechanisms of other insect appendages, for example, butterfly wings, cricket legs, cockroach legs, etc. One of the questions for other insects than Drosophila is whether the information obtained on Drosophila is applicable to other insects. Recently, Panganiban et al. (1994) demonstrated that during early development of larval appendages in Precis coenia, Dll is expressed in the appendages, as observed in Drosophila. Furthermore, Panganiban et al. (1995) found that Dll is also expressed in the appendages of two crustaceans, Artemia franciscana and Mysidopsis bahia, myriapods, Ethmostigmus rubripes, and chelicerates, Argiope argentata. On the other hand, en expression has been reported in the grasshopper, Schistocerca americana (Patel et al., 1989a, b), in the beetle, Tribolium castaneum (Patel, 1994a), and its expression pattern closely resembles that in Drosophila. These results indicated that common molecules such as en and Dll are likely to function in appendage formation of other insects similarly as found in Drosophila. Thus, the expressions of en and Dll in developing insect appendages may be used as positional markers: en is a posterior compartment marker (Patel et al., 1989a) and Dll is a distal marker (Popadić et al., 1996). Furthermore, comparison of the expression patterns of these genes among different appendage types will give us further understanding on structures of insect appendages.
Insects belonging to Pterygota is subdivided into two groups of hemimetabola and holometabola. Since Drosophila belongs to the latter, it is interesting to compare expression patterns of various genes found in Drosophila with those in hemimetabolan insects. In this study, we focus on the orthoptera as a typical hemimetabola. Their embryos are categorized in the short germ type, while the Drosophila embryo is in long germ type. Furthermore, the appendages of orthopterans are initially formed as buds, without forming imaginal discs as observed in Drosophila. In the orthopteran insects, we chose a two-spotted cricket, Gryllus bimaculatus, to compare the morphological and embryological differences at the molecular level with the fly, because their rearing at a large scale is easy and many eggs are obtained at a time. Since the cricket exhibits fundamental morphologies of appendages like polysegmented antenna, mouth parts of biting type, locomotive leg, and cercus as a sense organ, which are widely found in insects, we can compare developmental processes among those appendages.
Here, focusing on the development of appendages, we first described embryogenesis of G. bimaculatus by dividing developmental processes into several stages. Then, the expression patterns of en and Dll in developing appendages were observed immunohistochemically and compared with those of Drosophila imaginal discs. We found that both en and Dll expression patterns in antenna are similar to those in legs. In addition, we analyzed expression patterns of en and Dll in the various appendages. We found that Dll expression patterns were different among appendage types, although en expresses similarly in each posterior compartment. Thus, Dll is required for elaboration of the distal part of the appendage, and is involved in determination of the appendage type.
MATERIALS AND METHODS
Two-spotted cricket, Gryllus bimaculatus, was used in this study. To obtain clear results for whole-mount immunostaining, we used a mutant with white eyes (autosomal recessive; gwhite) (Fig. 1A), kindly provided by I. Nakatani of Yamagata University. During late embryogenesis and just after molt, their bodies become more transparent than in usual.
Rearing of crickets
All nymphs and adults of G. bimaculatus were reared at 27 ± 1 °C with humidity of 70 ± 2% under a 10L: 14D photoperiod, and fed on crashed dog-food and artificial rabbit-food. Under these conditions, nymphs molt every week, and the adults emerge after the 8th molt. Generation time of the cricket is about two months. The 1st–8th instar nymphs were maintained in plastic cases of 10 × 10 × 12 cm, and the adults were in cases with 15 × 15 × 20 cm. Each case contains two cotton wool-plugged plastic tubes for supplying water, and several crumpled papers as a shelter. In the case rearing the adults, a pile of moist and folded tissue papers (about 1 cm in height) was used for females to lay eggs during the dark period. Tissue papers for laid eggs were exchanged with new ones every day, and freshly laid eggs were collected with forceps. The eggs were placed separately on moist tissue papers and incubated under the same conditions as employed for rearing adult crickets (27°C and 70% humidity). Nymphs hatched out from the eggs on 13th day of incubation were transferred to the rearing cases.
Observation of embryogenesis
To observe position and form of cricket embryos within the egg at each stage, eggs were soaked in 30% bleach for 5 min and removed their chorion. Dechorionated eggs were washed thoroughly in phosphate buffered saline (pH 7.4) and then observed in the saline under dark-field illumination, using stereo-microscope. For observation of the embryo morphology, embryos with yolk masses were taken out from the eggs, and then the yolk masses were removed carefully with tungsten needles. The embryos were mounted on microscope slides and observed their detail features under a bright-field microscopy with the Nomarski optics.
To examine the localization of engrailed (en) and Distal-less (Dll) proteins in developing cricket embryos, we used a monoclonal antibody (anti-en protein), namely MAb 4D9, purchased from the Developmental Studies Hybridoma Bank (Baltimore, MD, USA) (Patel et al., 1989b), and a polyclonal antibody (anti-Dll protein), kindly provided by Dr. Sean Carroll (University of Wisconsin) (Panganiban et al., 1995).
Immunostaining was performed as described previously for grasshopper embryos by Patel et al. (1989b). All embryos at stages 5–12 were stained as whole-mount preparation. To detect en proteins, the MAb 4D9 supernatant was used at 1:3 dilution and the secondary antibody, peroxidase-conjugated goat anti-mouse IgG (Jackson Immunoresearch Lab), was used at a dilution of 1:100. On the other hand, the Dll antibody was used at a dilution of 1:200 and the secondary antibody, peroxidase-conjugated goat anti-rabbit IgG (Jackson Immunoresearch Lab), was used at a dilution of 1:400. To observe the expression domain precisely, the stained embryos were dehydrated in ethanol, placed in xylene, then embedded in paraffin and serially sectioned at 5 μm thickness. Both stained whole embryos and sections were mounted on microscope slides and then observed under a Nomarski differential-interference microscope.
Definition of developmental stages in Gryllus bimaculatus
Although outline of the embryogenesis of G. bimaculatus has already reported by Miyamoto and Shimozawa (1983), detailed embryonic structures have not been described. Here, general aspects of morphologies of the developing embryos and metathoracic limbs are briefly illustrated in Fig. 2. Most of the eggs are hatched on the 13th day ± 12 hr under incubation conditions of 27 ± 1°C and 70 ± 2% humidity. We divided total developmental span into 16 stages based on morphological features of developing embryos and appendages, as summarized in Table 1. The correspondence between stages and morphologies are shown in Fig. 2.
Embryogenesis in Gryllus bimaculatus
Expression patterns of engrailed in body segments are conserved
We used a monoclonal antibody 4D9 to observe expression of engrailed (en) in G. bimaculatus embryos at stages 5–12. The en expression patterns are shown in Figs. 3 and 4. At stage 5, the en expression patterns were segmentally repeated stripes in the gnathal and thoracic segments (Fig. 3A). In addition, a stripe of en also appears in a region becoming the first abdominal segment (Fig. 3A). With the caudal extension of the embryo, the number of abdominal stripes increases and finally sixteen stripes were observed in the fully segmented embryo (Fig. 3C). These stripes are localized in the posterior compartment of every segment. No stripe of en was detected in the most caudal segment, i.e., the eleventh abdominal segment (Fig. 4D), whereas the clear en stripe was found in the posterior compartment of the tenth abdominal segment (Figs. 3C, D and 4D). In progress of development, the gnathal and thoracic stripes become wide. In the thoracic terga, en stripes are much wider than those in abdomens at stage 12 (Fig. 3F). The intensity of expression in the segments becomes maximum at stage 11 (Fig. 3D). From this stage, several en-positive cells are regularly arranged at the mid-ventral region of every body segment (Fig. 3H) and in the head (Fig. 3G). Since the en expression during neurogenesis has been reported in many organisms including arthropods, annelids, and chordates (Patel et al., 1989b), the en-positive cells may become a subset of neuroblasts, ganglion mother cells and neurons in the central nervous system. The distribution of en-positive cells during neurogenesis in G. bimaculatus closely resembles that in the grasshopper, Schistocerca americana (Patel et al., 1989b; Condron et al., 1994).
Expression patterns of engrailed in the posterior compartments of appendages
The en gene is also expressed in developing appendages such as the antenna (Fig. 3F), mandible, maxilla, labium, and thoracic legs (Fig. 4A-C), but not in the developing labrum. During elongation of these appendages, en expression is limited to the posterior compartment of each appendage. At stage 7 when each appendage bud begins to elongate laterally, the segmental en stripe extends in the posterior portion of each appendage (Figs. 3B and 4A). In thoracic leg buds, the expression in the proximal region is more intense than that in the distal region (Fig. 4A, B). At stage 11, en is also expressed in the femoral apodeme in each leg (Fig. 4C). Although no stripe of en is detected in the eleventh abdominal segment of embryo as above mentioned, the en expression is observed in developing cercus which is the appendage of the eleventh abdominal segment (Figs. 3C-F and 4D-F). With elongation of the cercus, the expression extends proximally and connects to the en stripe of the tenth abdominal segment (Fig. 4E). In this stage, the orientation of the cercus changes toward the posterior direction along the anteroposterior axis of the body, so that en appears expressed in the dorsal region of the cercus (Fig. 4E). A transverse section of the immunostained cercus reveals that en-positive cells are localized in the posterior one third (Fig. 4F). The similar expression pattern was also observed in transverse sections of other immunostained developing appendages (data not shown).
Expression patterns of Distal-less depend on the type of appendages
We used a polyclonal antibody to observe expression of Distal-less (Dll) in G. bimaculatus embryos at stages 5∼12. Stained embryos are shown in Figs. 5 and 6. Dll expression was specifically observed in developing appendages. The initial expression was found in the metathoracic appendage bud at stage 5 (Fig. 5A). By stage 8, Dll is expressed in the primordia of labrum, antenna, maxilla, labium, thoracic legs, and preuropodium (Fig. 5B). In addition, Dll expression was also observed in the apical region of the cercus at stage 9 (Fig. 5C). As shown in Fig. 6A, initial expressions in the maxillary and labial buds are limited to the distal half regions. These regions become the telopodites defined as the distal shafts of the arthropod limbs (Snodgrass, 1935). The similar expression pattern was also observed in the thoracic appendage buds. However, no Dll expression region was found only in mandibular buds (Fig. 6A). These expression patterns are consistent with those reported for Drosophila melanogaster, Precis coenia (Panganiban et al., 1994), and Thermobia domestica (Popadić et al., 1996).
With further elongation and segmentation of the appendages, the Dll expression becomes intense and then the expression pattern changes, depending on the appendage type (Fig. 5C, D). In thoracic legs, intense expressions were observed in the distal and middle regions (Fig. 6C). We found that the distal region becomes tarsus and a part of tibia in future, whereas the middle region corresponds to the entire region of femur. The similar expression pattern was also observed in the labial and maxillary palpus (Fig. 6B, C). Dll is also expressed in two endites of the maxilla, the galea and lacinia (Fig. 6B). Interestingly, the expression is restricted to the distal region of each endite, as observed in early gnathal limb buds.
In other types of appendages such as the labrum, antenna, and cercus, no change in Dll expression patterns was observed with elongation of the appendages (Fig. 5D), although the expression patterns in the appendages are more or less different from each other. In elongating antenna, Dll is expressed in most of the distal region corresponding to future pedicel and flagellum of the antenna, but not in the basal narrow region probably becoming the scape segment (Fig. 6D). The entire region of the growing labrum and the apical region of the cercus also consist of Dll-positive cells (Figs. 5D and 6E).
Expression patterns of en and Dll are conserved in developing antenna and legs of Drosophila and G. bimaculatus
In the insect appendage formation, two types of developmental process are known: One is seen in holometabolan insects such as Drosophila. The appendages are formed as imaginal discs in the larval body. The other type is known in hemimetabolan insects: The appendages are initially formed as buds in the early embryo. In G. bimaculatus, the type of the appendage formation belongs to the latter, in which imaginal discs are not formed during appendage development (Fig. 2B). We compared the expression patterns of en and Dll in the developing antenna and legs of G. bimaculatus with those in the corresponding imaginal discs of Drosophila as shown in Fig. 7. In the Drosophila embryo, Dll expression patterns are different between antenna and leg discs, whereas the en expression in the posterior compartment is similar between them. The differences between the antenna and leg buds were also observed in G. bimaculatus (Fig. 7). Therefore, the expression patterns of en and Dll in the developing antenna and legs closely resemble those in the Drosophila imaginal discs. The similarities of en and Dll expression patterns between the two organisms imply that common molecules may be involved in development of insect antenna and legs and that their functions may be common in various insects in spite of different mode of appendage development. Since the en and Dll expression patterns in the crustaceans have been reported to resemble closely those in insects (Patel et al., 1989a; Panganiban et al., 1995), these genes may function commonly in arthropod appendage formation.
Dll is expressed in the telopodite of each appendage
In the early development in G. bimaculatus, the similarities in Dll gene expression patterns were found among different appendage types except the mandible: Dll is expressed in the distal region corresponding to the telopodite of each appendage. Since the morphological analysis of the Drosophila mutant lacking the Dll activity suggested that formation of the telopodite is regulated by Dll (Cohen and Jürgens, 1989), the similarities in Dll expression patterns suggest that the initial formation of all appendages except the mandible may occur by determination of the telopodite of the appendage by Dll. In addition, en expression in the posterior compartment is conserved pattern among different appendage types. Therefore, common expression patterns of the Dll and en genes in the early appendage formation in spite of different appendage type may support a hypothesis that every appendage type was evolved from a type of limb of a myriapod-like ancestor, which was originally proposed by Snodgrass (1935). He also assumed that in the insect mandible, the telopodite region is absent and consists of only the coxopodite. It is interesting to note that expression of Dll was not observed in the cricket mandible. Furthermore, no expression of Dll gene in the mandible has been also observed in the lepidopteran insect, Precis coenia (Carroll, 1994; Panganiban et al., 1994). These facts support his assumption and suggest that mandible may evolve differently from other appendages in the insect. On the other hand, we found that expression of en was not detected in the labrum, although Dll is expressed in it. Thus, we assume that although the labrum is originally derived from an appendage, its posterior compartment may be somehow degenerated.
Late expression patterns of Dll depend on the structure of the appendage
We found that expression domains of the Dll genes become defined finely in each appendage type with the progress of morphogenesis in the appendages. For instance, since the expression domain of the Dll gene was found to correspond to the telopodite region in each appendage, we found that the boundary between coxopodite and telopodite corresponds to the articulation region of coxa-trochanter of G. bimaculatus in the leg (Fig. 6C) or to the scape-pedicel articulation region in the antenna (Figs. 6D and 8). The Dll expression patterns in the maxillary palpus (Fig. 6B) also suggest that the maxillary palpus corresponds to the telopodite, as speculated previously (Snodgrass, 1935; Machida, 1981; Bitsch, 1994). In addition, Dll expression was also found in the distal regions of the galea and lacinia, which are two endites in the maxilla, showing that they also bear the telopodite elements. Thus, we concluded that the galea and lacinia may be homologous with other appendages.
In cricket embryos, from stage 9 the Dll expression patterns begin to change with the elongation and segmentation, depending on the type of appendages (Fig. 5C, D). Final expression patterns of Dll are classified into three types. First, the expression was expanded entire region of the appendage as seen in the antenna (Fig. 6D). Second, the expression was observed in a distal region of the appendage as in the cercus (Fig. 6E). Third, as in the leg, intense expression was observed in distal and middle regions (Fig. 6C). The space between the intensely expressed regions corresponds to the position of the prospective femur-tibia articulation (Fig. 8). The maxillary and labial palpus, whose expression pattern of Dll belongs to the third type, have an articulation corresponding to the femurtibia, indicating high homology with the leg. In the maxillary palpus of G. bimaculatus, the articulation between the 2nd and 3rd segment from the distal corresponds to the femurtibia articulation in the leg. On the other hand, in the first and second types, since the space observed in the third type was not found, an articulation corresponding to the femur-tibia may be absent. Thus, we considered that the Dll expression patterns in the late appendage development are specific for the appendage type and that these expression patterns may be closely related to the segmentation patterns, especially participated in the positioning of the articulation between the femur and the tibia or between corresponding segments of the appendages.
We wish to thank Dr. Sean B. Carroll (University of Wisconsin) for providing the Dll antibody, and Dr. I. Nakatani (Yamagata University) for providing the gwhite mutant of Gryllus bimaculatus. We also thank Dr. Ryuichiro Machida (University of Tsukuba) and Dr. Takayuki Nagashima (Tokyo University of Agriculture) for their encouragement and helpful advice. This study was supported in part by Grants-in-Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture, Japan, and a grant from the Mitsubishi Foundation.