Notch and Delta Control the Switch and Formation of Camouflage Patterns in Caterpillars.

In most Papilio species, a younger larva mimics bird droppings but changes its pattern to match host plant colors in its final instar. This change is determined by juvenile hormone (JH) during the JH-sensitive period (JHSP) early in the fourth instar. Recently, we found that homeobox genes control the pre-pattern formation specifically during JHSP, but the molecular mechanisms underlying final patterning and pigmentation at molt are unknown. By knockdown of Delta and Notch in Papilio xuthus larvae, we here showed that these genes define the edge and pigmentation area in final patterns, during and even after JHSP, suggesting that they bridge the JHSP and molt. Knockdown of Delta in Papilio machaon led to similar phenotypic changes, and knockdown of Notch caused pigmentation loss in twin spots of the silkworm Multilunar (L) mutant. Our findings suggest the importance of the Notch signaling pathway in caterpillars' adaptive evolution of color pattern formation.

Introduction

Camouflage, mimicry, and aposematism, observed among many insects, are successful adaptive strategies to protect animals from predators (Bates, 1862; Caro et al., 2016; Cott, 1957; Poulton, 1890; Wallace, 1889). In most swallowtail butterflies in the genus Papilio, younger caterpillars show body coloration patterns that mimic bird droppings, which change during their final instar stage (Gaitonde et al., 2018; Igarashi, 1979). In the Asian swallowtail butterfly, Papilio xuthus, the younger larva (first–fourth instar stages) mimics bird droppings, with a black and white color pattern (mimetic pattern, Figure 1A) (Futahashi and Fujiwara, 2008; Prudic et al., 2007). However, the fifth instar larva switches to a cryptic greenish body, with a pair of eyespots and some dark green stripes, including a V-shaped marking across the abdominal 4 (A4) to 5 (A5) segments, which may help the caterpillar conceal itself in the host plant (Cryptic pattern, Figure 1A). Interestingly, the Papilio larva, such as Papilio machaon (the old-world swallowtail butterfly), which feeds on herbaceous or narrow-leaved plants (mainly umbellifers), shows a bold stripe in each segment in its final instar (Figure 1B). This pattern is totally different from that of other Papilio species, such as P. xuthus, which feeds on broad-leaved plants, such as citrus leaves (Gaitonde et al., 2018; Igarashi, 1979; Prudic et al., 2007). This phenomenon indicates that the color pattern of the final instar larvae has evolved to conceal them in the host plants (Caro et al., 2016), although the molecular mechanisms underlying these types of cryptic pattern formations are largely unknown.

Figure 1

Color Pattern of Lepidopteran Caterpillars and a Schematic Diagram Showing Events during the Pattern Formation

(A) Dorsal view of fourth and fifth instar stage larvae of P. xuthus. Left panel, dorsal view of fourth instar larva showing body color pattern mimicking bird droppings. An enlarged side view with dashed red lines shows the presumptive V-shaped marking. A4 and A5: the fourth and fifth abdominal segments, respectively. Middle panel: a dorsal view of fifth stage larva with cryptic green body pattern. The enlarged side view shows the V-shaped marking in the black dashed box. The red arrow indicates the border line location, and a separate border line is displayed on the side of the arrow. Right panel: the color pattern of the fifth stage larva under UV light. The enlarged side view shows the fluorescent white line on the anterior side of the V-shaped marking. The yellow arrow indicates location of the white line, and a separate fluorescent border line is displayed on the side of the arrow. The dashed boxes indicate the enlarged regions of A4 and A5 under UV light, respectively. Scale bars: 5 mm.

(B) Fourth and fifth instar larvae of P. machaon and final instar larva of B. mori (L mutant).

(C) Schematic diagram of critical events and the underlying hormonal regulation. During the first 20 h of the fourth instar (JH sensitive period, JHSP), JH titer decline triggers the prepatterning process, including expression of abd-A and Abd-B. During the last 20 h of the fourth instar, a 20E-induced gene network regulates pigmentation and cuticle formation. The information on hormone changes was adapted and modified from Futahashi and Fujiwara (2008), and Futahashi et al. (2010).

Our previous study on P. xuthus showed that decreased titer of juvenile hormone (JH) during the early fourth instar stage (0–20 h after the third larval ecdysis) induced the larval coloration switch from the bird-dropping pattern to the host plant pattern (Figure 1C). This phase in early fourth instar stage is, thus, called the JH-sensitive period (JHSP). In contrast, that primitive group of these Papilio species (subgenus Heraclides) retains the bird-dropping phenotype throughout its larval life (Gaitonde et al., 2018; Igarashi, 1979; Prudic et al., 2007). This indicates that the JH-dependent pattern switch appeared later during the Papilio evolution. The pigmentation in the fifth instar larva occurs during the fourth molting stage (Futahashi and Fujiwara, 2007), about 3 days after the JHSP, according to the gene network orchestrated by 20-hydroxyecdysone (20E, known as the molting hormone) (Futahashi et al., 2010; Futahashi and Fujiwara, 2007) (Figure 1C). Thus, there is a 3-day gap between the two processes, i.e., the JH-dependent fate decision for the color switch and the actual execution of coloration by 20E, whereas the detailed molecular process during this time remains unclear.

To explore the molecular mechanisms underlying the JH-dependent color switch in P. xuthus, we recently carried out a transcriptome-wide screening for genes whose expression varies during the JHSP among different color pattern regions (Jin et al., 2019). Among the 20 candidates identified as the JH-regulated genes, we found that three homeobox genes, clawless, abdominal-A (abd-A), and Abdominal-B (Abd-B) are responsible for the final instar color patterning (Jin et al., 2019). Importantly, the final instar color patterning during the fourth molting period was inhibited by knockdown of these genes during the JHSP, but not afterward. This indicates that three homeobox genes function only during the JHSP; thus, we named these genes “prepatterning genes” (Figure 1C). However, the tightly regulated timing of these genes' functionality does not fully explain the developmental gap between JH-dependent fate decision and 20E-regulated pattern pigmentation, which motivated us to search for another type of gene bridging these two processes.

Among the 20 JH-regulated genes described earlier, Delta, encoding a specific ligand in the Notch signaling pathway, is of particular interest to us (Artavanis-Tsakonas et al., 1999; Fehon et al., 1990; Liao and Oates, 2017). The Delta ligand and the Notch signaling pathway are known to be involved in various types of cell-cell communication and cell patterning modulation during development (Artavanis-Tsakonas et al., 1999; Micchelli et al., 1997; Yoon and Gaiano, 2005). Therefore, we set out to investigate the functional role of Delta and the receptor gene Notch in the cryptic color formation, especially as bridging factors between the JH- and 20E-dependent processes. In this study, functional analyses, using electroporation-mediated RNA interference (RNAi) (Ando and Fujiwara, 2013), showed that Delta and Notch, but not the other ligand gene Serrate (Fleming et al., 1990), are involved in the formation of marking patterns in the fifth instar larval body in P. xuthus. Remarkably, knockdown of Delta and Notch partially disrupted the final instar larval patterning, even after the JHSP, indicating that these genes regulate patterning after the prepatterning genes complete their roles, which contributes to progression from this stage into the fourth molting period.

To further determine whether the Notch signaling pathway's role in pattern formation during the larval stage is conserved among Lepidoptera, we studied their function in P machaon, which has a different type of final instar larval pattern, and the silkworm Bombyx mori mutant Multilunar (L), which shows twin spot markings in each segment (Figure 1B). Our results from P. xuthus and the other two species revealed an important function of the Notch signaling pathway in camouflaging pattern formation, which has evolved and diverged among the lepidopteran caterpillars.

Results

Characteristics of V-Shaped Marking in the Fifth Instar Larval Stage of P. xuthus

In this study, we focused on the formation of the V-shaped marking, which could be seen on the dorsal side of the final instar larva of P. xuthus (Figure 1A). The marking lies across the fourth and fifth abdominal segments (A4 and A5) as a dark green stripe with two left-to-right symmetrical triangle patterns. Each half of the V-shaped marking is surrounded by two diagonal edges, one on the anterior and the other on the posterior side (Figure 1A). The anterior edge of the V-shaped marking consists of two lines, a white line adjacent to the green region and a black line next to the white line (Figure 1A). We named the set of white and black lines the “border line” of the V-shaped marking (Figure 1A, middle panel). Although the thin border line was observed consistently at the anterior edge, the black and white lines (in reverse orientation to anterior edge) appeared intermittently at the posterior edge. Interestingly, we noticed that the white line of the border absorbs UV light and emits florescent light (Figure 1A, right panel), which was, thus, used for phenotypic observation in RNAi analyses. In addition, the intersegmental color marking lines within the A1, A6, and A8 segments consist of white and black lines on the posterior side in each segment. Similar UV fluorescence was seen in the white lines of the dorsal side in the A1, A6, and A8 segments and inside the eyespot region of the T3 segment (Figure 1A, right panel).

Temporal and Spatial Expression Patterns of Delta and Notch during the JHSP

We previously identified Delta as a differentially expressed gene during the JHSP by RNA sequencing analysis (Jin et al., 2019). To understand Delta's role in larval coloration, we first analyzed its temporal and spatial expression patterns in the V-shaped marking and its surrounding regions during the JHSP (Figure 2, left panel). We prepared mRNA from the V-shaped marking (denoted as “V”), the eyespot region on the T3 segment (E), the dorsal region on A3 (D), the green region across A3-A4 (G), and the middle region across A3-A5 (M), at 0, 6, and 12 h after the third molt. Using quantitative polymerase chain reaction (qPCR), we compared Delta's expression patterns (Figure 2, right panel, Delta). Among all five regions, we found that the Delta expression level was significantly higher in the G region at 0 and 6 h (ANOVA, p value = 0.0010 and 0.0085, respectively), whereas its expression decreased during the JHSP (Figure 2). Delta's expression was relatively lower in the other four regions, although its expression levels in E and D regions were slightly higher than those in the M and V regions.

Figure 2

Temporal and Spatial Expression Patterns of Delta and Notch

Sampling regions for mRNA are shown on the left panel, from anterior to posterior of P. xuthus larva: the eyespot region (E) on T3 segment, the dorsal region (D) on A3 segment, the green region (G) on A3-A4 segments, the middle region (M) on A3-A4 segments, and the V-shaped marking (V) on A4-A5 segments. Relative expression levels are shown on the right panel and plotted as means + SD. Epidermis samples were prepared at 0, 6, and 12 h in the JHSP. Number of samples: n = 6 in E; n = 3 in D, M, and V; n = 9 (Delta) and 6 (Notch) in W. For statistical analysis, a one-way ANOVA, followed by a Tukey HSD test, was applied to analyze differential gene expression at the same time point among different sampling regions. For the Delta gene, the p value of 0 h group was 0.0010 and of 6 h group was 0.0085. For the Notch gene, the p value of 0 h group was 0.0325. ∗p < 0.05; ∗∗p < 0.01.

Different expression levels of Delta ligand activate the Notch receptor and downstream signaling pathway in various regions, which, in turn, induces Notch expression (Artavanis-Tsakonas et al., 1999; de Celis and Bray, 1997). To further ascertain whether expression of other factors in the Notch signaling pathway changes during formation of the V-shaped making, we analyzed genes encoding Notch and another ligand, Serrate (Fleming et al., 1990). Notch gene expression was detected in all five regions, with relatively constant levels in G and D regions, whereas their expression fluctuated during the JHSP in the remaining three regions (Figure 2, right panel, Notch). The Serrate gene expression pattern appeared similar to Delta's, showing higher expression in the G region than other regions (Figure S1). Further to know whether expressions of Notch pathway genes are affected by the JH titer, we have re-analyzed expression profiles of Notch, Delta, and Serrate based on the RNA-seq data obtained in our former report (Jin et al., 2019). After a JH analog (fenoxycarb) was applied on the fourth instar larvae just after the third molt, we prepared RNA samples from each region (E, G, and V) at 0, 6, and 12 h and performed RNA-seq analysis. Figure S2 shows the region-specific repression of three genes by the JH application: repression of Delta in the G and V regions; repression of Serrate in the G region; and repression of Notch in the E and V regions.

Repression of Delta Causes Ectopic V-shaped Marking in the Green Region

To further determine the Notch signaling pathway's function in color patterning of P. xuthus larva, we used a modified electroporation-mediated RNAi method (Ando and Fujiwara, 2013), which enabled us to knock down genes of interest in desired regions at specific developmental stages. Based on Delta's spatial expression, we reasoned that its imbalanced spatial expression may be involved in the V-shaped marking formation. Therefore, we introduced the Delta siRNA, through electroporation, during the late third instar stage (before head capsule slippage, HCS), and designed two types of experiments to test our hypothesis (Figure 3). In the type I experiment, we knocked down Delta in the green region (A3 to A4), resulting in an area with low Delta expression surrounded by areas with high Delta expression. In the type II experiment, we knocked down Delta in a wider area that covers both the green and future V-shaped marking (A4–A5) regions to observe the effect on the border line formation.

Figure 3

Effect of Delta knockdown

Delta knockdown was performed during the late period of third instar of P. xuthus, and the results were observed at the fourth and fifth instar stages.

(A) No phenotypic changes were observed in fourth instar larva after Delta knockdown. si-Delta, siRNA-treated side; Control, non-treated side. Scale bar: 5 mm.

(B) Confirmation of effects of knockdown using qPCR. Relative expressions were plotted as means + SD. n = 5, Student's t test was performed, ∗p = 0.0313.

(C) The fifth instar larva after Delta knockdown in the green region (type I, see text). The dashed red box shows A3 and A4 segments. Scale bar: 5 mm.

(D) Enlarged phenotypic changes of type I (same individual as in C) and type II experiment (see text). The arrowhead indicates ectopic pigmentation and emergence of a newly formed border line. Scale bar: 5 mm. The dashed red circles represent the knockdown area (PBS droplet region, see text).

(E) Enlarged ectopic border in (D, Type I). Region with high Delta expression shows green coloration, whereas that with low Delta in the knockdown area shows dark green.

In all individuals treated with Delta siRNA, we found no phenotypic changes during the fourth instar stage (n = 11; Figures 3A and S3). We confirmed the siRNA knockdown results by qPCR using samples from the epidermis, dissected from both the siRNA-treated side and the untreated control side, at 12 h of the fourth instar stage. The results showed that Delta expression decreased significantly in the siRNA-treated side (Figure 3B, n = 5, Student's t test, p = 0.0313). However, most individuals (10/11, 90.9%) showed color pattern changes in the fifth instar stage (Figures 3C and S3). To ensure siRNA knockdown results are interpreted accurately, we included negative control siRNA and found no phenotypic changes in either fourth or fifth instar larvae in P. xuthus (Figures S4 and S5).

Moreover, in both type I and II experiments, ectopic black pigmentation appeared in the green region, but not in the control side, in the fifth instar stage after Delta knockdown (Figures 3D and S3). In those individuals in which Delta was knocked down in the green region (Type I), the V-shaped marking pattern, including the anterior border line (a set of white/black lines), dark green region, and posterior black line, appeared ectopically, with irregular shapes (Figures 3D and S3, type I). This result indicates that Delta regulates the entire V-shaped marking in the green region. To further analyze Delta knockdown's effects on border formation, we marked the siRNA knockdown region by tracing the positions of phosphate buffered saline (PBS) droplets on the side of the positive electrode (+side) during electroporation (represented by dotted red circles in each figure). By comparing the patterns of PBS droplets and ectopic pigmentation, we found that the black or white lines appeared at the edge of the PBS droplet area (Figure 3D, type I, red arrowhead; Figure 3E, an enlarged photo of 3D; Figure S3, type I), indicating that they are formed at the boundary between the area with Delta knockdown and normal Delta expression level. In contrast, the dark green region appeared inside the region of PBS droplets, which represents the Delta knockdown area. In some individuals, the Delta knockdown area (PBS droplet area) expanded from the green region to the original V-shaped marking area (Figures 3D and S3, Type II). We observed that the anterior border line of the V-shaped marking was broken and curved, and the black pigmentation extended from these V-shaped markings to the green region, presumably because of redistribution of Delta expression.

Notch Signaling Defines Border Lines in the Fifth Instar Larval Pattern

Because Delta is the specific ligand for Notch (Artavanis-Tsakonas et al., 1999), we next examined whether the Notch gene is involved in larval color pattern formation. The results from Delta RNAi knockdown suggest that the border line formation is related to the Notch signaling pathway activation; thus, we knocked down the expression of Notch via RNAi near the future V-shaped marking site (Figures 4 and S6) during the late third instar stage and examined the potential phenotypic change during the fourth and fifth instar stages. Similar to Delta knockdown, Notch knockdown in the V-shaped marking caused no color pattern change during the fourth instar stage (n = 10; Figures 4A and S6). Instead, it led to a dramatic alteration in color pattern in the fifth instar stage (n = 8; Figures 4C and S6). We confirmed that the Notch expression was highly repressed by siRNA using qPCR (Figure 4B, n = 3, p = 0.0544).

Figure 4

Effect of Notch Knockdown

Notch knockdown was performed during the late period of the third instar of P. xuthus, and the results were observed at the fourth and fifth instar stages.

(A) No phenotypic changes were seen in fourth instar larva after Notch knockdown. Scale bar: 5 mm.

(B) Confirmation of knockdown effects using qPCR. Relative expressions were plotted as means + SD. n = 3, Student's t test was performed, p = 0.0544. (C) The fifth instar larva after Notch knockdown around the V-shaped marking (type II, see text). The dashed red box shows A3 and A4 segments. Scale bar: 5 mm.

(D) Enlarged view of type I and type II (same individual as in c) experiments after Notch knockdown. The dashed red circles represent the knockdown area (PBS droplet region, see text). Ectopic and expanded pigmentation were observed in type I, but the border line disappeared in type II.

(E) Effect of Notch knockdown around the T3 to A2 segments.

When Notch was knocked down in the original V-shaped marking region (Type II), both the anterior and posterior border lines of the V-shape marking were broken and/or dislocated at the fifth instar stage in most individuals (Figures 4C and S6, type II). Moreover, we observed that the dark green pigmentation expanded into the green region from A5 to A3 and A4 segments (Figure 4D, type II). In addition, as expected, Notch knockdown in the green region (type I) led to an ectopic V-shaped marking, including the border lines and inner dark green pigmentation, a phenotypic change similar to that following Delta knockdown (Figures 4D and S6, type I).

In addition to the V-shaped marking, we have observed several border lines, including a set of white/black lines, at the anterior edge of a lateral blue band in A1 and the anterior edge of the lateral dark green band in A6 (Figure 1A). To determine Notch's function in the border line formation in the whole larval body, we knocked down Notch in the border line between A1 and A2 of two larvae (Figure 4E). At the fifth instar stage, the A1-A2 border lines were broken, and the blue pigmentation between the two borders expanded into the T3 and A1 in both larvae (Figures 4E and S6). Interestingly the black lines around the eyespot region (T3) were also disrupted (Figure 4E). These results suggested that Notch defines the border line formation and that dislocation of Delta and Notch expressions alter the position and shape of V-shaped marking, A1-A2 border lines, and eyespot marking.

We also examined whether Serrate is involved in the larval color pattern formation by Serrate knockdown. We observed no phenotypic change in either the fourth or fifth instar larvae (Figure S7). These results indicate that among genes in the Notch signaling pathway, Delta and Notch, but not Serrate, regulate color pattern formation in the last instar larva of P. xuthus.

Function of Delta and Notch after JHSP

Our previous study revealed that the three prepatterning genes' function is restricted to the JHSP because knockdown of these genes after the JHSP had no effects on color pattern formation (Jin et al., 2019). To further understand how Delta and Notch contribute to color patterning, we performed RNAi knockdown of Delta and Notch after the JHSP during the fourth instar stage. We injected siRNAs for Delta or Notch at least 24 h after the third molt (at least 4 h beyond the JHSP) and confirmed gene expression knockdown using qPCR. Following both Delta and Notch knockdown, all the fifth instar larvae showed broken and diminished border lines around the V-shaped marking in the siRNA-treated side (Figures S8 and S9, bright field panel). However, we observed no effect on the pigmentation pattern of the V-shaped marking. To examine the phenotypic change of the border lines following RNAi treatment more closely, we placed individuals under ultraviolet (UV) light and recorded fluorescent signal emitted by the white lines (Figures S8 and S9, UV light panel). All individuals treated with Delta and Notch RNAi showed a partial depletion of the fluorescent signal in the border lines (Figures S8 and S9, UV light panel, pink arrowheads), consistent with the above observation. These results indicate that Delta and Notch modulate border line formation even after the JHSP, which is different from the previously reported prepatterning gene dynamics.

Function of the Notch Signaling Pathway in Larval Color Patterns of P. machaon and the Silkworm

To explore the Notch signaling pathway's function in regulating larval color patterns in other lepidopteran species, we carried out RNAi knockdown in P. machaon, a species closely related to P. xuthus (Figures 1B and 5A), and in a larval marking mutant L of the silkworm B. mori (Figures 1B and 5B), a species distantly related to P. xuthus.

Figure 5

Delta Knockdown in P. machaon Larva and Notch Knockdown in B. mori L Mutant Larva

(A) The dashed red circles represent the Delta knockdown area in P. machaon larva. The effects were shown in fourth instar (left) and fifth instar larva (right). siRNA knockdown was performed at the last stage of the third instar larva. The red dashed box shows the A4-A6 segments. The enlarged dorsal view, on the right edge, shows that the black pigmentation expanded ahead after knockdown of Delta on A5. The dashed red circles represent the knockdown area. Scale bar: 5 mm.

(B) The effects of Notch knockdown on fifth instar larva of the B. mori L mutant. siRNA knockdown was performed at the late fourth instar stage, and the phenotypic change was observed at the fifth stage. The red arrowhead shows that black and brown pigmentation was depleted after Notch knockdown. The right panel represents a model (right) of Notch knockdown. In the knockdown area of Notch (indicated by dashed red circle), the melanin pigmentation disappeared. Scale bar: 5 mm.

Similar to P. xuthus, P. machaon changes its coloration pattern between the fourth and fifth instar larval stages, resulting in black stripes on the green body in each segment (Igarashi, 1979). When the Delta siRNA was injected around the A5 segment at the late third instar, there was no phenotypic change at the fourth instar stage (Figures 5A and S10). In the fifth instar stage, however, the anterior edge of the black stripe in A5 (Figure 5A, enlarged dorsal view; Figure S10) and A6 (Figure S10) expanded to the anterior side. In addition, thin black stripes appeared at the anterior edge of A5 (Figure 5A). However, no phenotypic change was observed at the posterior edge of the black stripe (Figures 5A and S10). We concluded that Delta repression affected mainly the anterior part of black stripes in P. machaon, consistent with the results of Delta knockdown in P. xuthus.

We further examined the silkworm L mutant, which shows pairs of black brown twin spots on each larva body segment (Yamaguchi et al., 2013) (Figure 1B). It was reported that Wnt1 (wingless) gene is responsible for L phenotype, and region-specific upregulation of Wnt1 causes the twin spot markings (Yamaguchi et al., 2013). Unexpectedly, Notch knockdown in the twin spot region during the fourth instar stage caused loss of black pigmentation in the knockdown area (Figures 5B and S11). We also examined the functions of homologs of Delta, Serrate, and fringe genes (LeBon et al., 2014; Sato et al., 2008), which are critical factors in the Notch signaling pathway, in twin spot marking of the L mutant. The region-specific expression of these genes in B. mori (L) larva varies during the development and between regions (Figure S12). However, RNAi knockdown of Delta, Serrate, and fringe showed no effect on pigmentation or color patterning in the final instar larval stage.

In summary, these results indicate that the Notch signaling pathway regulates multiple larval color patterns among the distantly related lepidopteran species. Although the detailed mechanisms underlying hormonal regulation, color patterning, and pigmentation may differ and need to be studied further, the Notch signaling pathway serves as an evolutionarily conserved switch shaping the insect body color pattern and is used repeatedly through evolution for camouflage formation.

Discussion

In this study, we showed that Delta or Notch knockdown disrupted the coloration patterns of the final instar larva of P. xuthus, regardless of whether the siRNA was injected before or after the JHSP (Jin et al., 2019) (Figure 6A). This manner of gene action is different from that of other prepatterning genes, such as the abd-A and Abd-b, whose functions are limited to the JHSP (Figure 6A), or that of the downstream pigmentation gene, such as Laccase 2, which mediates the melanization process during every molting period (Futahashi and Fujiwara, 2007; Futahashi et al., 2011; Jin et al., 2019). Delta was previously reported to be upregulated by reduced JH level in the green region (Jin et al., 2019), and expression of Delta and Notch was repressed by an increased JH level during the JHSP (Figure S2). Combining these JH-dependent expression reports with functional analyses, we reason that Delta and Notch activity initiate during the JHSP and continue, throughout the intermolt period (~3 days), until the molting period. Previously, we showed that knockdown of the prepatterning genes abd-A and Abd-B led to depletion or ectopic formation of the entire V-shaped marking (Jin et al., 2019) including the full set of border lines, which shares some similarity with the Delta and Notch knockdown results (Figures 3 and 4). Based on these observations, we concluded that the Notch signaling pathway transmits positional information for the final instar color pattern through the JHSP to the fourth molting. The Notch signaling pathway has been implicated previously in insect development, such as embryogenesis (Hoppe and Greenspan, 1986; Liao and Oates, 2017) and wing formation during metamorphosis (de Celis and Bray, 1997; Micchelli et al., 1997). However, our findings on Notch and Delta functions in larval color pattern formation provides new insights into how the local pattern change occurs at just one molting stage without changing whole body plan.

Figure 6

Effects of siRNA Knockdown and Genes Involved in Color Pattern Switch in Papilio Larva

(A) Three types of genes involved in the P. xuthus color pattern switch. Prepatterning genes (cll, abd-A and Abd-B) function only during the JHSP, whereas the pigmentation gene (Laccase 2) functions at every molting period (Jin et al., 2019). Delta and Notch function continuously throughout the JHSP and the intermolt of fourth instar, which is different from timing of activity of the prepatterning and pigmentation genes.

(B–D) The schematic diagrams show the phenotypic change after Delta knockdown (B) and Notch (C) in P. xuthus larva and Delta knockdown in P. machaon (D). (B) Delta knockdown results in low Delta expression area inside the dashed red circle, where the ectopic dark green coloration occurs (2). Between the green (higher Delta expression) and the dark green regions, border lines, consisting of a white and a black line, formed at both anterior and posterior sides (1). (C) Notch knockdown depletes border line (1) and expands dark green region (2), probably due to the disrupted Delta expression by Notch repression. The dashed red circle represents the Notch knockdown region. (D) Delta knockdown in P. machaon larva causes ectopic melanization and expansion of the black stripe area into the green region. The dashed red circle represents the Delta knockdown region.

It is intriguing that Notch and Delta greatly impact the border line formation even after the JHSP. This may imply that the boundary information of color patterns is defined and maintained at later stages of color pattern switch (Figures 6B–6D). Notch's continuous activation is critical for cell adhesion (Bao, 2014), proliferation (Shimura et al., 2009), and other epidermal cell functions involved in color patterning. In conclusion, Notch and Delta determine more detailed shape and color of cryptic marking of the final instar stage, after the prepatterning genes initiate the regional information of cryptic color patterns during the JHSP (Figure 6A).

One important finding from the Delta knockdown is that the border lines of marking appeared between areas with decreased Delta expression and the surrounding areas with normal expression. This indicates that different Delta expression levels in adjacent regions initiates formation of the border line or the pigmentation area and defines their subsequent positions (Figure 6B). The Notch signaling pathway is known to be involved in cell-cell communication during development (Artavanis-Tsakonas et al., 1999; Bray, 2016; de Celis et al., 1998; Heitzler and Simpson, 1991; Liao and Oates, 2017; Micchelli et al., 1997), and its lateral inhibition results in different cell fates (del Alamo et al., 2011; Liao and Oates, 2017; Sprinzak et al., 2010). In the classic model, lateral inhibition occurs in a short range of neighboring cells (Heitzler and Simpson, 1991) because it requires direct contact between the Notch ligand and its receptor, both of which are transmembrane proteins. The border line formation in the adjacent regions following Delta (Figures 3D, 3E, and 6B) and Notch (Figures 4D and 6C) knockdown, therefore, is consistent with the lateral inhibition theory.

Furthermore, in the Delta knockdown experiment, it is intriguing that the entire V-shaped pattern appeared ectopically in the green region (Figure 3D, enlarged side view), with white and black lines in its anterior edge, the whole pigmented Delta knockdown area, and black and white lines in the posterior edge (in a reverse order to the anterior edge) (Figure 6B). The order of white/black lines in the ectopic marking is consistent with the original V-shaped marking, in which the white line always appears near the green region and the black line shows in neighboring dark green region. This indicates the ectopically created pattern still follows the anterior-posterior orientation of the larval segments, indicating that the Notch signaling pathway only regulates pigmentation pattern formation after the body plan has been determined. We observed some yellow spots located within and around V-shaped marking in P. xuthus (Futahashi et al., 2012) and within bold black stripes in P. machaon (Yamaguchi et al., 2013), which did not change following Delta knockdown (Figures 1 and 3D). Therefore, the Notch signaling does not affect formation of this spot, which was regulated by Wnt1 in P. machaon (Yamaguchi et al., 2013). Moreover, the diagonal V-shaped marking exists across A3 to A5 segments, although the detailed mechanism underlying this intersegmental color pattern is unknown. One possible explanation is that the Notch-Delta pathway is modulated in a head-to-tail direction across the entire body, presumably by a morphogenic element such as Wnt (Collu et al., 2014; Sonnen et al., 2018).

The Notch expression pattern corresponded to the eyespot and midline color pattern in wings of the nymphalid butterflies (Reed and Serfas, 2004). Furthermore, in Papilio polytes, the enhancer of split gene, which is downstream to the Notch signaling pathway (Bailey and Posakony, 1995; Bray, 2006), showed a specific spatial-temporal expression patterns across wings (Iijima et al., 2019). By knockdown experiments in three species, however, the present study revealed more clearly the functional roles of the Notch signaling pathway in regulating the coloration pattern.

Moreover, we found that Delta knockdown in P. machaon disrupted the edge line formation of the bold stripe and induced black pigmentation in the siRNA-treated region (Figures 5A and 6D), similar to the results in P. xuthus. These two Papilio species are closely related (Zakharov et al., 2004), with similar changes between the younger mimetic pattern and the final instar cryptic pattern (Shirataki et al., 2010), which helps them camouflage in different types of host plants (Prudic et al., 2007). This indicates that P. machaon also has the JHSP in the early fourth instar stage and retains a similar pattern switch mechanism. The mechanism of JH-dependent pattern switch is conserved among many Papilio species and has diverged from the ancestral group, which retains the bird-dropping pattern throughout life (Prudic et al., 2007). Although the reason why the switch mechanism has diverged remains unknown, it is speculated that coloration mimicking the host plant is more effective than the bird-dropping pattern for protection from predators because the size of the last instar larva in the Papilio species usually exceeds 5 cm.

On the other hand, the knockdown of Notch, but not Delta, caused loss of the larval spot marking in the L mutant of B. mori, which is distantly related to P. xuthus (Figure 5B). This phenotypic change is opposite to the appearance of pigments in the knockdown area of Delta or Notch in P. xuthus. In addition, knockdown of neither Serrate nor fringe affected the pigmentation in the L spot (data not shown). Notch pathway involved in the larval pattern formation is conserved in Lepidoptera, although its gene network may be different among species. Because the twin spots of the L mutant are known to be caused by 20E-induced Wnt1 expression (Yamaguchi et al., 2013), it is conceivable that the Notch signaling pathway regulates Wnt1 expression in spot formation. It is known that the melanin synthesis in the larval epidermis occurs during the molting period under the control of the ecdysone cascade (Hiruma and Riddiford, 2009; Yamaguchi et al., 2013). Thus, the Notch signaling pathway involved in the pigmentation process of three species should be interacted closely with the 20E-induced gene pathway (Futahashi and Fujiwara, 2007; Futahashi et al., 2012).

Our findings shed new light on the Notch signaling pathway's function in the camouflage color formation of lepidopteran larva, thereby bridging two important development programs, the JH-dependent patterning and 20E-regulated pigmentation.

Limitations of the Study

Although we observed fluorescence emitting from the white lines in the V-shaped marking in caterpillar of P. xuthus, its biological function is unclear. Thus, a future study will be important to understand the molecular basis underlying synthesis of the fluorescent pigment in caterpillars of butterfly and moth.

Resource Availability

Lead Contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Haruhiko Fujiwara (haruh@edu.k.u-tokyo.ac.jp).

Materials Availability

This study did not generate new unique reagents.

Data and Code Availability

The published article includes all data generated or analyzed during this study.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.