- Research article
- Open Access
The pivotal role of aristaless in development and evolution of diverse antennal morphologies in moths and butterflies
BMC Evolutionary Biology volume 18, Article number: 8 (2018)
Antennae are multi-segmented appendages and main odor-sensing organs in insects. In Lepidoptera (moths and butterflies), antennal morphologies have diversified according to their ecological requirements. While diurnal butterflies have simple, rod-shaped antennae, nocturnal moths have antennae with protrusions or lateral branches on each antennal segment for high-sensitive pheromone detection. A previous study on the Bombyx mori (silk moth) antenna, forming two lateral branches per segment, during metamorphosis has revealed the dramatic change in expression of antennal patterning genes to segmentally reiterated, branch-associated pattern and abundant proliferation of cells contributing almost all the dorsal half of the lateral branch. Thus, localized cell proliferation possibly controlled by the branch-associated expression of antennal patterning genes is implicated in lateral branch formation. Yet, actual gene function in lateral branch formation in Bombyx mori and evolutionary mechanism of various antennal morphologies in Lepidoptera remain elusive.
We investigated the function of several genes and signaling specifically in lateral branch formation in Bombyx mori by the electroporation-mediated incorporation of siRNAs or morpholino oligomers. Knock down of aristaless, a homeobox gene expressed specifically in the region of abundant cell proliferation within each antennal segment, during metamorphosis resulted in missing or substantial shortening of lateral branches, indicating its importance for lateral branch formation. aristaless expression during metamorphosis was lost by knock down of Distal-less and WNT signaling but derepressed by knock down of Notch signaling, suggesting the strict determination of the aristaless expression domain within each antennal segment by the combinatorial action of them. In addition, analyses of pupal aristaless expression in antennae with various morphologies of several lepidopteran species revealed that the aristaless expression pattern has a striking correlation with antennal shapes, whereas the segmentally reiterated expression pattern was observed irrespective of antennal morphologies.
Our results presented here indicate the significance of aristaless function in lateral branch formation in B. mori and imply that the diversification in the aristaless expression pattern within each antennal segment during metamorphosis is one of the significant determinants of antennal morphologies. According to these findings, we propose a mechanism underlying development and evolution of lepidopteran antennae with various morphologies.
Olfaction is one of the most essential senses for animal survival. Accordingly, morphology and function of olfactory organs have been targets of natural selection in a wide range of animals. Terrestrial mammals with olfactory acuity, such as carnivores (e.g., cats and dogs) and ungulates (e.g., cattle and deer), have deeply convoluted olfactory epithelia [1, 2]. Nocturnal and cave-dwelling animals tend to have more sensitive olfactory organs than their diurnal or non-cave-dwelling relatives [3, 4]. In insects, the main olfactory organ is a pair of antennae on the head, which carry many odor-sensing organs on their surfaces. Antennae are segmented multiple times along the proximodistal (PD) axis and have a simple rod-shaped structure in many insect species. In Lepidoptera (moths and butterflies), however, those using pheromones for long-distance communication, such as nocturnal species, often have antennae with protrusions on or around the ventral side of each antennal segment. In extreme cases, such as Bombyx mori (silk moth), protrusions form remarkably long comb-like structures, which we term “lateral branches” (Figs. 1 and 2d). Since odor-sensing organs in such moths are densely packed and form the olfactory epithelium on the ventral side of each antennal segment, antennae with protrusions or elongated lateral branch structures can accommodate a wide surface area of the olfactory epithelium. In addition, protrusion structures aligned like a comb extend the time during which the pheromone or odorant plume passes through the antenna [5, 6]. These physical properties of protrusions or lateral branches in moths enhance the efficiency of olfactory reception and are believed to be ecologically important. In contrast, diurnal species of Lepidoptera, such as almost all butterflies, have simple antennae without protrusions or lateral branches (e.g. Fig. 1, Papilio xuthus; Additional file 1: Figure S1). Antennae with protrusions or lateral branch structures are thought to have originated from the filiform antenna of the primitive lepidopteran insects, such as micropterigid moths (e.g. Fig. 1, Micropterigidae gen.) and to have been acquired independently several times during the evolution of Lepidoptera (Fig. 1, Additional file 1: Figure S1) [7, 8].
Antennal development has been studied using branchless antennae of non-lepidopteran insects, including the fruit fly, Drosophila melanogaster. It has been suggested that antennal development along the PD axis is largely conserved: several transcription factor genes are expressed in specific regions along the PD axis and regulate the growth and differentiation of each region. Antennae are largely subdivided into proximal, medial, and distal parts by the combinatorial expression of homothorax (hth)/extradenticle (exd) and/or Distal-less (Dll) in each region [9,10,11,12,13,14,15]. The distal part, expressing only Dll, is further subdivided into several segments, and other transcription factor genes are expressed in the specific segment(s) among Dll-expressing segments. For example, aristaless (al) is specifically expressed in the most distal segment [15,16,17]. In Drosophila, according to studies of antennal development [9,10,11, 18] and the serial homology of antennae to legs [19,20,21,22,23,24,25], Wingless (Wg; a member of the WNT family), Decapentaplegic (Dpp; a member of the TGF-β family), and ligands of the epidermal growth factor receptor (EGFR) have been suggested to act as morphogens regulating the region-specific expression of these transcription factor genes along the PD axis in the antenna (for a review of leg development, see [26,27,28,29]). Wg and Dpp are expressed in the ventral and dorsal regions, respectively, as continuous stripes along the PD axis [30, 31]. EGFR ligands are thought to be produced from the distal tip as in the leg [32, 33]. Along with these PD patterning mechanisms, Notch signaling is also repeatedly activated along the PD axis and regulates antennal segmentation . Although the understanding of antennal development along the PD axis has been advanced, the mechanism by which protrusions or lateral branches are formed in each antennal segment and how they have evolved remain largely unknown.
A previous study on the highly branched antenna of B. mori, which has two lateral branches per segment (Figs. 1 and 2d), has revealed patterns of gene expression, cell proliferation and cell death in lateral branch formation during metamorphosis . Upon the onset of metamorphosis, many antennal patterning genes, including those involved in morphogen signaling, dramatically change their expression pattern from the conserved one described above to a segmentally reiterated, lateral branch-associated one. In the pupal antenna, each of these genes is expressed in a specific region within one segment, and the same expression pattern is repeated segmentally. Their expression patterns within each segment prefigure lateral branches. These expression patterns are not observed in other insects studied so far and a distinctive feature of the branch-associated antenna of B. mori. Importantly, just before the elongation of lateral branches, Dll expression disappears in the future olfactory epithelium, while it is strongly upregulated in the region surrounding the future olfactory epithelium and weakly in the remaining part (Fig. 2a) . Furthermore, al is expressed in the two small regions just dorsal to both sides of the future olfactory epithelium (Fig. 2b) . The al expression domains are included in the strong Dll expression domain (compare Fig. 2a and b). Each lateral branch then elongates during subsequent pupal development so that the region around the intersection between the al expression domain and the olfactory epithelium becomes the distal tip. Consequently, cells derived from the overlap between Dll and al expression contribute most of the dorsal side of each lateral branch, while the ventral side is covered by the olfactory epithelium (Fig. 2e) . During the elongation of lateral branches, cells in the al expression domain abundantly proliferate, whereas cell death occurs almost ubiquitously. These observations predict that localized cell proliferation regulated by the branch-associated expression of antennal patterning genes is fundamental to lateral branch formation in the B. mori antenna. Furthermore, genes involved in morphogen signaling, such as wingless (wg) and rhomboid (rho; encoding an EGFR ligand activator), also show segmentally reiterated, branch-associated expression during metamorphosis (Fig. 2c) , implying their involvement in the regulation of antennal patterning gene expression within each antennal segment.
Here, we describe the results of functional analysis on Dll and al during lateral branch formation in the B. mori antenna. We also investigated the involvement of WNT, EGFR and Notch signaling in lateral branch formation. Our results show that the branch-associated expression of al and Dll is essential for lateral branch formation and that al expression is positively regulated by Dll and WNT signaling but restricted by Notch signaling. In addition, analyses of al expression in pupal antennae of several lepidopteran species, which are different in the extent of protrusions, revealed that the al expression pattern shows a striking correlation with antennal shapes, whereas the segmentally reiterated expression pattern was observed irrespective of antennal morphologies. Together with the results from functional analysis in B. mori, this observation implies that the variation in the al expression pattern within each antennal segment during metamorphosis is one of the significant determinants of antennal morphologies. According to these findings, we propose a mechanism underlying development and evolution of lepidopteran antennae with various morphologies.
Requirement for Dll function in lateral branch formation
To investigate Dll function specifically in lateral branch formation during B. mori antennal development, the electroporation-mediated RNA interference (RNAi)  was conducted. siRNAs against Dll (Dll-siRNAs) were injected and incorporated to developing antennal cells by the electroporation (see Methods). To specifically inhibit the segmentally reiterated, branch-associated expression during the pupal stage and minimize the defect in the basic PD development, we applied Dll-siRNAs 1 day before metamorphosis. All antennae subjected to Dll-RNAi (Dll-RNAi antennae), but not those subjected to RNAi against Enhanced Green Fluorescent Protein (EGFP) as a negative control, showed regions with missing or greatly reduced lateral branches (Fig. 3a, b). RNAi is expected to be induced only in cells incorporating enough siRNAs and thus, antennae subjected to RNAi possibly contain both RNAi induced and non-induced cells. It appeared, therefore, that lateral branches were missing or greatly reduced in the regions where Dll expression was depleted. In contrast to the substantial defects in lateral branches, segmentation appeared to be unaffected and segment sizes seemed relatively normal even in the region exhibiting extensive defects in lateral branch formation (Fig. 3C), although the overall length of antennae is somewhat reduced. This indicates that in our experimental condition, there is only a little, if any, influence on Dll function for the basic PD development and the defects in lateral branch formation were mainly resulted from the depletion of Dll expression in the pupal stage. Thus, Dll function during metamorphosis appears to be essential specifically for lateral branch formation. Interestingly, antibody staining using the anti-Al antibody revealed that the segmental expression of al was lost in many regions in Dll-RNAi antennae at the pupal stage (Fig. 3d). Therefore, Dll appears to positively regulate al expression within each segment during metamorphosis (see Fig. 4e).
Significance of al function in lateral branch formation
Intriguingly, small and fine protrusions were frequently seen in the regions showing extensive defects in lateral branch formation in adult Dll-RNAi antennae (Fig. 3c, c’). Small protrusions, possibly corresponding to those observed in the adult antennae, were also found in pupal Dll-RNAi antennae (Fig. 3d, d’). The anti-Al antibody staining showed that they were always associated with the remaining al expression (Fig. 3d, d’). Together with the previous finding that the al expression domain within each antennal segment is the site of abundant cell proliferation and contributes to most of the dorsal side of the lateral branch , this tight association of al expression with the epithelial protrusion may suggest that the localized expression of al gives epithelial cells protrusive activity to elongate the lateral branch. If this is true, it is expected that the elimination of al activity would lead to the loss of the lateral branch. To test this, we knocked down al function specifically during metamorphosis with the electroporation of antisense morpholino oligomers against al (al-MO; see Methods). The anti-Al antibody staining signals were missing in several regions of pupal antennae subjected to al-MO (Fig. 3g, g’), confirming that al expression was removed in several regions by this treatment. In adult al-MO antennae, missing or greatly reduced lateral branches were observed (Fig. 3e, f, Additional file 2: Figure S2). These data indicate that the localized al expression within each antennal segment during metamorphosis regulates lateral branch formation, possibly by activating proliferation of al-expressing cells.
Regulation of al expression within each antennal segment during metamorphosis by WNT signaling
The significance of the localized al expression within each antennal segment during metamorphosis for lateral branch formation led us to investigate the regulatory mechanism of al expression. Since wg and rho are also expressed in a segmentally reiterated pattern during metamorphosis , the involvement of WNT and EGFR signalings in the regulation of al expression within each antennal segment during metamorphosis were investigated by the electroporation-mediated RNAi. WNT signaling was attenuated by siRNAs against armadillo (arm), encoding a transducer of the canonical WNT signaling [37, 38], and EGFR signaling by siRNAs against egfr, a gene encoding EGFR itself . Immunostaining of antennal primordia subjected to arm-RNAi with anti-Al and anti-Arm antibodies showed that al expression was lost in the arm-depleted regions (Fig. 4a-a”). In contrast, no significant change in al expression was observed in the regions where EGFR signaling seemed attenuated (judged by the lack of the anit-diphospholylated MAPK staining) in pupal antennae subjected to egfr-RNAi (Additional file 3: Figure S3). These results indicate that WNT signaling positively regulates segmental al expression during metamorphosis, whereas EGFR signaling is dispensable (Fig. 4e).
Involvement of Notch signaling in determining al expression domain within each antennal segment
We next investigated the function of Notch signaling in the regulation of al expression during metamorphosis. Notch signaling has been known to regulate segmentation of antennae through segmentally repeated Notch activation in several insect species [15, 34, 40]. Accordingly, the involvement of Notch signaling in the regulation of segmentally reiterated al expression was expected. Notch signaling was attenuated by RNAi using the electroporation or lipofection to introduce siRNA against Notch (Additional file 4; Figure S4, see Methods). Adult antennae subjected to the electroporation-mediated Notch-RNAi showed a severe defect, in which most segments were drastically fused and the overall length along the PD axis was extensively reduced (Fig. 4c). This indicates that the involvement of Notch signaling in antennal segmentation appears to be conserved also in B. mori. In contrast to the severe case, the lipofection-mediated RNAi resulted in mildly affected antennae that showed branch fusion in several regions but without considerable reduction in the overall length (Fig. 4d). This suggests that neighboring lateral branches are segregated by Notch signaling. Antibody staining of such Notch-RNAi antennal primordia showed derepression of al in the Notch-depleted regions. al was ectopically expressed in the regions between its normal expression domains (Fig. 4b-b”, open arrowheads). These results are consistent with the above idea that al activates proliferation of al-expressing cells, and the fusion of neighboring branches in Notch-RNAi antennae can be explained as follows. Cells between branches ectopically express al and would abundantly proliferate in conjunction with flanking normal branch cells, resulting in the fusion of branches. In addition to the cells between branches, the ectopic al expression was also detected in the regions dorsal to its normal expression domains (Fig. 4b-b”, arrowheads). These results indicate that Notch signaling is required to repress al expression in the regions between and dorsal to its normal expression domains (Fig. 4e). Therefore, dorsal and lateral extent of the al expression domains within each antennal segment appears to be determined by this Notch-dependent repression. On the other hand, in the regions ventral to the normal al expression domains, which include both of the olfactory epithelia and regions between them, no ectopic al expression was observed (Fig. 4b-b”, asterisk), suggesting that the lack of al expression in these regions is independent of Notch signaling.
Correlation between al expression pattern and antennal morphology in lepidopteran species other than B. mori
The significance of al function in lateral branch formation (see above) and localization of al expression to the region of abundant cell proliferation  in B. mori led us to examine al expression in pupal antennae of other lepidopteran species with various antennal morphologies in order to investigate the relationship between al expression patterns and antennal morphologies. We used three lepidopteran species, Agrius convolvuli (hawk moth), Mamestra brassicae (cabbage moth), and Papilio xuthus (swallowtail butterfly), all belonging to the same Lepidoptera subclade, Obtectomera, as B. mori (Fig. 1). The olfactory epithelium is formed on the ventral half of the antenna in all three species as in the B. mori antenna (Fig. 5c, f, i, n). P. xuthus has a simple antenna without branch structures or protrusions (Fig. 5c). In A. convolvuli, neither branch structures nor protrusions are observed in females (Fig. 5f). In males, the olfactory epithelium extensively protrudes ventrally, although no apparent branch structure is observed (Fig. 5i). M. brassicae has no branch structure but shows a slight protrusion around the ventral midline (Fig. 5n).
In all species examined, the anti-Al antibody staining of pupal antennae revealed that al is expressed in a segmentally reiterated manner as in the case of B. mori (Fig. 5a, d, g, j, l). Thus, segmental reiteration of al expression alone does not predict the antennal morphologies. Interestingly, however, there was a difference in al expression patterns within each antennal segment. In the P. xuthus antenna, which has no branch structure or protrusion, al was expressed broadly within each antennal segment (Fig. 5a, b). In the antenna of female A. convolvuli, which lacks branch structures or protrusions, al expression within each antennal segment was also detected broadly in the ventral region (Fig. 5d, e). In contrast, al expression in the pupal antenna of male A. convolvuli, whose adult antenna shows the extensive ventral protrusion of the olfactory epithelium, was restricted to the region around the ventral midline (Fig. 5g, h). In the M. brassicae antenna, showing modest ventral protrusion, al expression displayed a very interesting pattern. At the early pupal stage, al expression was observed in a broad region spanning almost the entire half of each segment (Fig. 5j, k). At the later stage, however, al expression was changed to be localized to the region around the ventral midline (Fig. 5l, m).
These observations clearly demonstrate that al expression patterns correlate well to antennal morphologies. Together with the functional analysis in B. mori, the striking correlation between al expression patterns within each antennal segment and antennal morphologies suggests the importance of the al expression pattern within each antennal segment in determining the extent of the association with branches or protrusions (see Discussion).
As pheromone-receptive organs, insect antennae have been diversified in their morphologies according to habitat environment, especially in Lepidoptera, from simple rod-like structures to multi-branched morphologies (Fig. 1, Additional file 1: Figure S1). Previous study on B. mori has revealed the unique pattern of gene expression and localized cell proliferation during lateral branch formation . Yet, how lateral branches are formed in B. mori and various morphologies of antennae have been evolved are largely unknown. Functional analyses in B. mori described here strongly demonstrate that al expression within each antennal segment during metamorphosis is important for lateral branch formation. In addition, examination of al expression in other lepidopteran species with various antennal morphologies suggests possible importance of al expression pattern within each antennal segment on determining antennal morphologies.
Regulatory mechanism to determine the al expression domain within a segment in the B. mori antenna
Our results indicate that lateral branch formation in the B. mori antenna is regulated by al possibly through its function in activating proliferation of al-expressing cells (Fig. 3e-g’). Therefore, determination of the al expression domain appears to be important for lateral branch formation. Dll-RNAi (Fig. 3d, d’) and arm-RNAi (4a–a”) experiments indicate that the segmental al expression during metamorphosis are positively regulated by Dll and WNT signaling (Fig. 4e). The Dll expression domain larger than the al-expressing region (Fig. 2a, b)  and the segmentally reiterated wg expression (Fig. 2c)  may suggest that al expression is activated in cells both expressing Dll and receiving an appropriate level of Wg. Since arm is a common downstream effector of several WNT family members, however, contribution of WNT other than Wg cannot be excluded. The Notch-RNAi experiment (Fig. 4b-b”) clearly shows that al expression is repressed by Notch signaling in the region dorsal to the al expression domains and the region between neighboring al expression domains (Fig. 4e). In the region ventral to the al expression domains, however, al expression was not derepressed even when Notch was depleted (Fig. 4b-b”). Given the requirement of Dll in antennal al expression within each antennal segment (Fig. 3d), the absence of al derepression from the olfactory epithelium may be due, at least in part, to the lack of Dll expression there (Fig. 2a) . The absence of al derepression in the rest of the ventral region, in which Dll is expressed, can be explained if strong levels of Wg signaling repress al expression. The existence of other factors repressing al expression in the ventral region cannot be ruled out. Nonetheless, multiple regulatory inputs determine the strict expression domain of al within each segment in the B. mori antenna.
Possible involvement of al in the formation of antennae with various morphologies
Assuming that al function suggested in the B. mori antenna is conserved among other species investigated in this study, the interspecific difference in antennal morphologies can be explained as follows. In male A. convolvuli, the restricted al expression around the ventral midline may induce cell proliferation and lead to the extensive ventral protrusion of the olfactory epithelium (Fig. 6, Agrius convolvuli male). In contrast, in P. xuthus and female A. convolvuli, cell proliferation would be possibly induced by the broad al expression in the region occupying almost the entire ventral half of antennal segments and this may lead to the uniform growth of the whole ventral olfactory epithelium so that the epithelial protrusion does not occur (Fig. 6, Agrius convolvuli female and Papilio xuthus). In M. brassicae, while cell proliferation possibly induced in the initial broad domain of al expression may not form an epithelial protrusion as in the cases of P. xuthus and female A. convolvuli, a later restriction of al expression to the region around the ventral midline may alter the pattern of cell proliferation, leading to slight protrusion of the ventral epithelium (Fig. 6, Mamestra brassicae). Another interpretation is also possible: the broad al expression in P. xuthus and female A. convolvuli may have little or no activity in regulating cell proliferation, whereas changes in al expression to the restricted region may be associated with the acquisition of function in controlling cell proliferation in B. mori and male A. convolvuli. In either case, a variation in the al expression pattern within a segment may strongly influence antennal morphology. Testing of these interpretations awaits further study in these species.
Evolutionary perspectives of antennal morphologies
Considering that the branchless or non-protruding antennae of P. xuthus and female A. convolvuli seem to retain more ancestral characteristics of antennal morphology than the highly branched or extensively protruding antenna of B. mori or male A. convolvuli (Figs 2d and 5i), it seems reasonable to hypothesize that the broad expression of al is an ancestral state in the lineage leading to Obtectomera including these species (Fig. 6). Deviating from this ancestral situation, a subsequent restriction of al expression to two small regions within a segment may have led to the acquisition of two lateral branches on each segment in the lineage leading to B. mori, while its male-specific restriction to the region around the ventral midline in each segment may have resulted in the extensive ventral protrusion in males in the A. convolvuli lineage. Thus, changes in al expression within each antennal segment may have been important for morphological evolution of antennae.
Recently, it has been suggested that mutations relevant to the parallel evolution of morphology among different species are not randomly distributed in the genome, but occur in specific genes . This conceptual gene, named an input-output gene, is expressed under the control of the integrated information from multiple upstream patterning genes and in turn regulates the expression of a battery of downstream effector genes involved in determining a final tissue structure . Change in the expression pattern of the input-output gene by its regulatory mutations can alter the expression pattern of a battery of the downstream effector genes without any effects in the upstream patterning information. Thus, only the specific structure can be changed by regulatory mutations in the input–output gene without any effects on other structures. Such morphological change is expected to have minimal deleterious effects and if it is adaptive, associated regulatory mutations in the input–output gene can easily spread in the population . As described here, al expression domains within each antennal segment during metamorphosis are variable with a pronounced correlation with antennal morphologies among different species in Lepidoptera. This correlation suggests that al is one of the input–output genes for antennal morphogenesis in Lepidoptera.
The segmentally reiterated expression of al observed in all of three species investigated in this study (Fig. 5a, d, g, j, l) suggests that al, and possibly other antennal patterning genes, had already been recruited to the segmentally reiterated expression pattern in an ancestor of the lineage leading to Obtectomera. Since the antennae of P. xuthus and female A. convolvuli lack branch structures or protrusions (Fig. 5c, f), however, this recruitment alone may not have led to the acquisition of branch structures or protrusions. It may have led to the uniform expansion of the olfactory epithelium and/or determination of the specific cell type(s) in a segmentally reiterated manner. Therefore, the acquisition of the segmentally reiterated expression appears to be cryptic from the viewpoint of antennal morphology but a prerequisite for further changes in the expression within each segment, which may have directly led to the formation of branch structures or protrusions. Once the segmentally reiterated al expression had been acquired, it might have been relatively easy to change its expression within a segment. This may be one of the forces facilitating the parallel evolution of antennal morphologies at least within, and possibly also outside of, Obtectomera. In light of the input-output gene concept, acquisition of the morphologically cryptic, segmentally reiterated expression in the ancestral lepidopteran species might have conferred on al the potential to be an input–output gene and thus, a target for evolutionarily relevant mutations, and allowed the parallel evolution of antennal morphologies in Lepidoptera. This stepwise mechanism may be one of general features of morphological evolution.
In conclusion, our results show that in the B. mori antenna, al is essential for lateral branch formation and its expression domain appears to be determined strictly by the combinatorial function of Dll, WNT signaling and Notch signaling. Variation in the al expression patterns within each antennal segment, but not the segmentally reiterated expression itself, appears to be one of the significant factors for determining various antennal morphologies in Lepidoptera. Further research based on these findings will provide insights for understanding antennal evolution and general features of morphological evolution.
B. mori (N4 strain) and M. brassicae were reared on artificial diet (NIHON NOSANKO, Yokohama, Japan). A. convolvuli was reared on artificial diet containing sweat potato leaf extract. P. xuthus was reared on tangerine leaves. All insects were reared at 25 °C in long day condition (L:D = 16 h: 8 h) successively in our laboratory. S. cynthia was gifted from Dr. Z. Kajiura. Male animals were used for analysis unless otherwise noted.
The onset of pupal development (pupal molt) was monitored using a USB connection type CCD web camera equipped with an automatic infrared LED illumination system (GR-CAM130N2, Groovy, Tokyo, Japan). Pupal antennal primordia dissected at the stage when their epithelia were completely retracted from pupal cuticle were considered as samples at the early developmental stage (24 h after pupation [P24h] in B. mori, P72h in M. brassicae). Those dissected 12 or 48 h later before cuticle sclerotization were considered as samples at the late developmental stage (P84h in M. brassicae, P120h in A. convolvuli and P. xuthus).
Antibody staining and in situ hybridization
Antibody staining and in situ hybridization was conducted essentially as described previously [35, 43]. The following antibodies were used: guinea pig anti-Al (1:1000) , goat anti-Dll (1:2, dF-20, Santa Cruz Biotechnology, Dallas, USA; pre-adsorbed with silkworm larval epidermal powder), rat anti-Pros (1:5, a gift from F. Matsuzaki), mouse anti-Notch (1:40, C17.9C6, Developmental Studies Hybridoma Bank, Iowa City, USA), mouse anti-Arm (1:40, N2 7A1, Developmental Studies Hybridoma Bank, Iowa City, USA), mouse anti-diphospho ERKI/II (1:250, M8159, Sigma-Aldrich, St. Louis, USA), and fluorophore (Alexa Fluor 488, 555, 647)-conjugated secondary antibodies (1:100, Thermo Fisher Scientific, Waltham, USA or Jackson ImmunoResearch, West Grove,USA). Riboprobes for in situ hybridization of wg and rho in B. mori were generated with DIG or Biotin RNA labeling kit (Roche, Basel, Switzerland) using cDNA as a template. Primers used are listed in Additional file 5: Table S1. The detailed procedures are described in Additional file 6.
The electroporation- and lipofection-mediated RNAi
siRNAs were designed using siDirect version 2.0  (Additional file 7: Table S2), and chemically synthesized and annealed (Fasmac, Atsugi, Japan). The electroporation-mediated RNAi was conducted as described previously . In brief, 0.5 μL of 300 μM siRNA solution was injected through the tip of a larval antenna 1 day before metamorphosis. Immediately after injection, droplets of PBS were placed on the antenna and the lateral side of the head. Then, platinum electrodes were inserted into the PBS, and 5 square pulses of voltage (280 ms/s., 45 V) were applied. 24 h after the treatment, antennal primordia were collected for immunohistochemistry. We found that higher voltage or more pulses resulted in malformation of an antenna without siRNA.
In case of Notch-RNAi, the lipofection-mediated RNAi was conducted as a mild RNAi treatment. Lipofection mixture of siRNAs (61.25 μM each, Additional file 7: Table S2) was prepared by mixing TransFast transfection reagent (Promega, Madison, USA) and the same volume of siRNA solution. After incubating at room temperature for 15 min, 2 μL of lipofection solution was injected through the tip of an antenna of the late 4th instar (semifinal instar) larvae.
The adult antennae were photographed after eclosion under a stereomicroscope or a compound microscope.
We also tried to knock down al function by RNAi. There are at least two al homologs in the B. mori genome (al1/BMgn006008 and al2/BMgn006007, SilkDB/KAIKObase). We designed four different siRNAs for each of al1 and al2, and used them in various combinations. However, we failed to deplete either of al1 or al2 mRNA effectively by unknown reason (data not shown). Therefore, we tried using MO to knock down al function as described below.
Morpholino oligomer (MO) treatment
The electroporation was used to incorporate MO as in the electroporation-mediated RNAi treatment, except that 0.5 μL of 1 mM MO solution was injected and 5 square pulses of voltage (280 ms/s., 41 V) were applied. After the treatment, the antennal primordia were collected for immunohistochemistry between P24h and P48h. The adult antennae were photographed after eclosion under a stereomicroscope.
To knock down al function, we designed MO against al1. Possible isoforms of al1 mRNA in B. mori N4 strain was reconstructed by Trinity RNA-seq assembler on DDBJ Read Annotation Pipeline  using public RNA-seq data derived from female embryos 72 h after egg laying (DRR015667, NCBI Sequence Read Archive). We identified two isoforms with different translation initiation sites in the first exons (Additional file 8: Figure S5, Additional file 9: Figure S6). Their exon-intron structures were identified by mapping the mRNA sequences onto B. mori genome  using Exonerate software (ver. 2.2.0, with options ‘-m est2genome --score 1200’) . Fluorescein labeled MO was designed at one of the two common splicing donor sites within the homeobox to skip the second exon encoding the polypeptide including N-terminal half of a homeodomain (Gene Tools, Philomath, USA; Additional file 8: Figure S5, Additional file 10: Table S3). Standard Control oligo with 3’ Fluorescein (Gene Tools, Philomath, USA; Additional file 10: Table S3) was used as a negative control. The electroporation with al1-MO resulted in loss of the anti-Al staining signals and inhibition of lateral branch formation (Fig. 3e-g’). Three possibilities can be considered to explain these observations: first, the anti-Al antibody could detect both Al1 and Al2 protein, and al2 may not be expressed significantly in the antenna; second, the antibody used here may only detect Al1 protein but even if al2 is expressed in the antenna, it may have no or little function in the lateral branch formation; third, there may be a cross regulation between al1 and al2 and knock down of al1 alone may result in simultaneous downregulation of al2, leading to inhibition of both al1 and al2 activities. In any case, results of the MO experiment indicate that al activity for the branch formation is depleted successfully.
Scanning electron microscopy (SEM)
SEM images were collected using scanning electron microscopes (Miniscope, Hitachi, Tokyo, Japan; JSM-5600LV, JEOL, Tokyo, Japan). Samples were attached to a pedestal with glue or nail polish without chemical fixation and analyzed according to the manufacturer’s instruction.
For the bright field images, images from several planes of focus were projected using Helicon focus (Helicon Soft, Kharkov, Ukraine) with Radius = 50 and Smoothing = 1. Brightness and contrast of the images were adjusted using Photoshop CS5.5 and CS6 (Adobe systems, San Jose, USA).
Negus VE. Introduction to the Comparative Anatomy of the Nose and Paranasal Sinuses. Ann R Coll Surg Engl. 1954;15(3):141–73.
Van Valkenburgh B, Smith TD, Craven BA. Tour of a Labyrinth: Exploring the Vertebrate Nose. Anat Rec. 2014;297(11):1975–84.
Protas M, Jeffery WR. Evolution and development in cave animals: from fish to crustaceans. WIREs Dev Biol. 2012;1(6):823–45.
Balkenius A, Rosén W, Kelber A. The relative importance of olfaction and vision in a diurnal and a nocturnal hawkmoth. J Comp Physiol A. 2005;192(4):431–7.
Vogel S. How much air passes through a silkmoth’s antenna? J Insect Physiol. 1983;29(7):597–602.
Bau J, Justus KA, Loudon C, Cardé RT. Electroantennographic resolution of pulsed pheromone plumes in two species of moths with bipectinate antennae. Chem Senses. 2005;30(9):771–80.
Kawahara AY, Breinholt JW. Phylogenomics provides strong evidence for relationships of butterflies and moths. Proc R Soc Lond B Biol Sci. 2014;281(1788):20140970.
Scoble MJ. The lepidoptera: form, function, and diversity. Reprint ed. Oxford: Oxford University Press; 1992.
Dong PD, Chu J, Panganiban G. Coexpression of the homeobox genes Distal-less and homothorax determines Drosophila antennal identity. Development. 2000;127(2):209–16.
Dong PDS, Chu J, Panganiban G. Proximodistal domain specification and interactions in developing Drosophila appendages. Development. 2001;128(12):2365–72.
Dong PDS, Dicks JS, Panganiban G. Distal-less and homothorax regulate multiple targets to pattern the Drosophila antenna. Development. 2002;129(8):1967–74.
Ronco M, Uda T, Mito T, Minelli A, Noji S, Klingler M. Antenna and all gnathal appendages are similarly transformed by homothorax knock-down in the cricket Gryllus bimaculatus. Dev Biol. 2008;313(1):80–92.
Jockusch EL, Williams TA, Nagy LM. The evolution of patterning of serially homologous appendages in insects. Dev Genes Evol. 2004;214(7):324–38.
Angelini DR, Kaufman TC. Functional analyses in the hemipteran Oncopeltus fasciatus reveal conserved and derived aspects of appendage patterning in insects. Dev Biol. 2004;271(2):306–21.
Angelini DR, Kikuchi M, Jockusch EL. Genetic patterning in the adult capitate antenna of the beetle Tribolium castaneum. Dev Biol. 2009;327(1):240–51.
Miyawaki K, Miyawaki K, Inoue Y, Mito T, Fujimoto T, Matsushima K, Shinmyo Y, Ohuchi H, Noji S. Expression patterns of aristaless in developing appendages of Gryllus bimaculatus (cricket). Mech Dev. 2002;113(2):181–4.
Campbell G, Tomlinson A. The roles of the homeobox genes aristaless and Distal-less in patterning the legs and wings of Drosophila. Development. 1998;125(22):4483–93.
Chu J, Dong PDS, Panganiban G. Limb type-specific regulation of bric a brac contributes to morphological diversity. Development. 2002;129(3):695–704.
Snodgrass RE. Principles of Insect Morphology. New York, London: McGraw-Hill, Inc.; 1935.
Boxshall GA. The evolution of arthropod limbs. Biol Rev. 2004;79(02):253–300.
Gehring W. Bildung eines vollstandigen Mittelbeines mit Sternopleura in der Antennenregion bei der Mutante Nasobemia (Ns) von Drosophila melanogaster. Arch Julius Klaus Stift Vererbungsforsch Sozialanthropol Rassenhyg. 1966;41:44–54.
Struhl G. A homoeotic mutation transforming leg to antenna in Drosophila. Nature. 1981;292(5824):635–8.
Casares F, Mann RS. Control of antennal versus leg development in Drosophila. Nature. 1998;392(6677):723–6.
Angelini DR, Kaufman TC. Insect appendages and comparative ontogenetics. Dev Biol. 2005;286(1):57–77.
Shubin N, Tabin C, Carroll S. Deep homology and the origins of evolutionary novelty. Nature. 2009;457(7231):818–23.
Kojima T. The mechanism of Drosophila leg development along the proximodistal axis. Develop Growth Differ. 2004;46(2):115–29.
Estella C, Voutev R, Mann RS. A dynamic network of morphogens and transcription factors patterns the fly leg. Curr Top Dev Biol. 2012;98:173–98.
Kojima T. Developmental mechanism of the tarsus in insect legs. Curr Opin Insect Sci. 2017;19:36–42.
Kojima T. Proximodistal Development of the Leg in Drosophila – Creation of New Developmental Fields by Gene Regulatory Interaction and Tissue Growth. Proc Arthropod Embryol Soc Jpn. 2017;48:1–11.
Diaz-Benjumea FJ, Cohen B, Cohen SM. Cell interaction between compartments establishes the proximal-distal axis of Drosophila legs. Nature. 1994;372(6502):175–9.
Lecuit T, Cohen SM. Proximal–distal axis formation in the Drosophila leg. Nature. 1997;388(6638):139–45.
Campbell G. Distalization of the Drosophila leg by graded EGF-receptor activity. Nature. 2002;418(6899):781–5.
Galindo MI, Bishop SA, Greig S, Couso JP. Leg Patterning Driven by Proximal-Distal Interactions and EGFR Signaling. Science. 2002;297(5579):256–9.
Rauskolb C, Irvine KD. Notch-mediated segmentation and growth control of the Drosophila leg. Dev Biol. 1999;210(2):339–50.
Ando T, Kojima T, Fujiwara H. Dramatic changes in patterning gene expression during metamorphosis are associated with the formation of a feather-like antenna by the silk moth, Bombyx mori. Dev Biol. 2011;357(1):53–63.
Ando T, Fujiwara H. Electroporation-mediated somatic transgenesis for rapid functional analysis in insects. Development. 2013;140(2):454–8.
Siegfried E, Wilder EL, Perrimon N. Components of wingless signalling in Drosophila. Nature. 1994;367(6458):76–80.
Clevers H, Nusse R. Wnt/β-Catenin Signaling and Disease. Cell. 2012;149(6):1192–205.
Sopko R, Perrimon N. Receptor Tyrosine Kinases in Drosophila Development. Cold Spring Harb Perspect Biol. 2013;5(6):a009050.
Mito T, Shinmyo Y, Kurita K, Nakamura T, Ohuchi H, Noji S. Ancestral functions of Delta/Notch signaling in the formation of body and leg segments in the cricket Gryllus bimaculatus. Development. 2011;138(17):3823–33.
Davidson EH, Rast JP, Oliveri P, Ransick A, Calestani C, Yuh CH, Minokawa T, Amore G, Hinman V, Arenas-Mena C, Otim O, Brown CT, Livi CB, Lee PY, Revilla R, Rust AG, Pan Zj, Schilstra MJ, Clarke PJ, Arnone MI, Rowen L, Cameron RA, McClay DR, Hood L, Bolouri H. A Genomic Regulatory Network for Development. Science. 2002;295(5560):1669–78.
Stern DL, Orgogozo V. Is Genetic Evolution Predictable? Science. 2009;323(5915):746–51.
Sato M, Kojima T, Michiue T, Saigo K. Bar homeobox genes are latitudinal prepattern genes in the developing Drosophila notum whose expression is regulated by the concerted functions of decapentaplegic and wingless. Development. 1999;126(7):1457–66.
Yasunaga K, Saigo K, Kojima T. Fate map of the distal portion of Drosophila proboscis as inferred from the expression and mutations of basic patterning genes. Mech Dev. 2006;123(12):893–906.
siDirect version 2.0. http://sidirect2.rnai.jp. Accessed 15 Jan 2018.
DDBJ Read annotation pipeline. https://p.ddbj.nig.ac.jp/pipeline/. Accessed 15 Jan 2018.
Kaikobase, Public data. http://sgp.dna.affrc.go.jp/data/integretedseq.txt.gz. Accessed 15 Jan 2018.
Slater GSC, Birney E. Automated generation of heuristics for biological sequence comparison. BMC Bioinf 2005;6:31.
Regier JC, Mitter C, Zwick A, Bazinet AL, Cummings MP, Kawahara AY, Sohn JC, Zwickl DJ, Cho S, Davis DR, Baixeras J, Brown J, Parr C, Weller S, Lees DC, Mitter KT. A large-scale, higher-level, molecular phylogenetic study of the insect order Lepidoptera (moths and butterflies). PLoS One. 2013;8(3):e58568.
Sen A, Reddy GV, Rodrigues V. Combinatorial expression of Prospero, Seven-up, and Elav identifies progenitor cell types during sense-organ differentiation in the Drosophila antenna. Dev Biol. 2003;254(1):79–92.
We thank H. Kataoka for the cabbage moths, M. Shimoda and N. Ando for the hawk moth strain and their artificial diet, Z. Kajiura (National Bioresource Project, Silkworm) for the wild silkmoth strains, F. Matsuzaki for anti-Pros antibody, G. Sheng for standard control morpholino oligo, I. Oshima for the image of the micropterigid moth, J. Murray, S. Gregg, and A. Anker for the license for the use of the images of the hepialid moth, the psychid moth, and the cossid moth on Flickr website (farm4.staticflickr.com/3784/9570255774_7b57b3a269_o_d.jpg, farm8.staticflickr.com/7309/9733855147_47fd0131ee_o_d.jpg, and farm1.staticflickr.com/178/367375201_fcc0c49672_o_d.jpg), respectively.
This work was supported in part by Grant-in-Aid for Scientific Research (KAKENHI) to TK [20570198, 24570230] and to HF [22128005, 20017007, 15H05778], and Grant-in Aid for JSPS Fellows to TA [08 J08820].
Availability of data and materials
The datasets supporting the conclusions of this article are included within the article and its additional files.
Ethics approval and consent to participate
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.
Antennal morphology and phylogenetic tree of all lepidopteran families and superfamilies. The phylogenetic tree is based on the molecular phylogenetic estimation by Regier et al. . Antennal morphology of each family was quoted from Scoble . Antennal morphology of each family was categorized with the indicated color code in the upper left of the figure. The prominent bipectinate lateral branch appears to be acquired independently at least three times in the linage leading to swift moths (Hepialidae), bagworm moths (Psychidae), and the large group including Obtectomera and Cossoidea + Sessoidea + Zyaenoidea. Due to insufficient description or discrepancy between morphological classification and molecular phylogeny, several descriptions were redundantly quoted in distant families as below. (*) The same description in Papilionoidea was quoted. (**) The same description in Copromorphidae was quoted. (***) The same description in Megalopygidae was quoted. (****) The same description in Zygaenidae was quoted. (*****) The same description in Choreutidae was quoted. (TIFF 437 kb)
Negative control of MO treatment. (A) Two classes of antennal morphologies (Normal, Defective) observed in the standard control morpholino treatment. Antennae with fused or short branches were categorized as “Defective”. Scale bar, 1 mm. (B) Distribution of Normal and Defective phenotypes in the standard control morpholino and al1 MO treatments. The ratio of defective individuals in the al1 MO treatment was significantly higher than that of the standard control morpholino treatment *, p = 3.1 * 10− 3 < 0.01, Fisher’s exact test. (TIFF 605 kb)
Effect of reduced EGFR activity on al expression. Reduced EGFR activity was monitored using dpERKI/II signals. Segmentally reiterated EGFR activity was depleted in the region indicated with yellow arrowheads, whereas the native expression pattern of al (white arrowheads) was not affected, indicating that EGFR signal does not regulate induction of al expression at this stage. (TIFF 1275 kb)
Statistical analysis of mild RNAi treatment against Notch. (A) The four categories of defects in lateral branch formation. Normal, lateral branches were not fused. I, fused lateral branches at one region. II, less than 3 subsegments were fused in several regions. III, 3 or more subsegments were largely fused. (B) Effect of Notch siRNA injection was compared with Ubx siRNA and Buffer injections. Ubx was selected as the negative control gene that is not expressed in the antenna. To conduct Fisher’s exact test, categories I to III were collectively categorized as “defective”. The ratio of defective individuals in Notch siRNA injection was significantly higher compared to the other two negative control experiments (N vs. Ubx, p = 3.6 * 10− 8 < 0.05; N vs. Buffer, p = 2.7 * 10− 6 < 0.05). (TIFF 686 kb)
Primers used for preparing RNA probes. (PDF 23 kb)
Supplementary Methods. (PDF 350 kb)
Sequences of siRNA used in RNAi experiments. (PDF 271 kb)
The exon-intron structure of al1. The two al1 isoforms identified (al1A, al1B) have isoform-specific exons on the 5′ end (exon1), and share the remaining common exons (exon2-exon5). Homeobox is encoded in the region between exon2 and exon4. Orange indicates open reading frame. MO against al1 was designed at the intronic region adjacent to the 3′ end of exon2 (MO target) to skip exon2. (TIFF 60 kb)
The reconstructed exon sequences of al1 mRNA isoforms. Yellow, open reading frame; pink, homeobox. (PDF 108 kb)
Sequences of Morpholino oligomers used to knock down al1. (PDF 29 kb)