Polycomb group proteins confer robustness to aposematic coloration in the milkweed bug, Oncopeltus fasciatus

Abstract

Aposematic coloration offers an opportunity to explore the molecular mechanisms underlying canalization. In this study, the role of epigenetic regulation underlying robustness was explored in the aposematic coloration of the milkweed bug, Oncopeltus fasciatus. Polycomb (Pc) and Enhancer of zeste (E(z)), which encode components of the Polycomb repressive complex 1 (PRC1) and PRC2, respectively, and jing, which encodes a component of the PRC2.2 subcomplex, were knocked down in the fourth instar of O. fasciatus. Knockdown of these genes led to alterations in scutellar morphology and melanization. In particular, when Pc was knocked down, the adults developed a highly melanized abdomen, head and forewings at all temperatures examined. In contrast, the E(z) and jing knockdown led to increased plasticity of the dorsal forewing melanization across different temperatures. Moreover, jing knockdown adults exhibited increased plasticity in the dorsal melanization of the head and the thorax. These observations demonstrate that histone modifiers may play a key role during the process of canalization to confer robustness in the aposematic coloration.

1. Background

Many organisms have warning coloration to advertise their unpalatability and/or toxicity [1]. These warning colours, called aposematic coloration, are characterized by conspicuous coloration and contrasts in colours, which are more easily recognized and remembered by predators [1]. The evolution of aposematic coloration has long been debated because individuals with novel aposematic coloration among a cryptic population would be at a higher risk of potential predation before predators developed an association between coloration and distastefulness [2–4]. Intermediate steps involving facultative displays of aposematism have been proposed to overcome this issue. In some vertebrates, hidden aposematic coloration, where aposematic coloration is confined to specific locations on an otherwise cryptic body, has been suggested to serve as intermediate steps in the evolution of aposematism [5]. In other cases, phenotypically plastic aposematism [6–9] may allow aposematism to be expressed in a condition-dependent manner, such as only when the density of the prey is high [9]. In the latter cases, selection should favour the eventual evolution of robust coloration from previously plastic traits to facilitate learning and recognition by potential predators [9]. Many species in fact exhibit robust genetically determined aposematic coloration [10,11].

The evolution of robust phenotypes from environmentally induced, phenotypically plastic traits is a process known as genetic assimilation [12–14]. Phenotypically plastic traits readily change their phenotypes in response to the environment. Over time, if a specific trait confers higher fitness, then the trait will become genetically stabilized by a process known as canalization so that it becomes robust to environmental and genetic perturbations [12]. Given the selective advantage of being robust, aposematic coloration in many species is likely to be canalized and therefore offers an opportunity to study the molecular underpinnings of canalization.

Recent studies have suggested that chromatin structure can influence phenotypic plasticity and robustness of traits [15]. Chromatin is the molecular structure consisting of DNA wrapped around histone octamers. Histones can be altered by chemical modifications, such as the addition or removal of acetyl, methyl and/or ubiquitin groups to histone tails. Such chemical changes cause the chromatin structure to switch between the transcriptionally active euchromatin and the transcriptionally silent heterochromatin. Histone modifiers catalyse these changes and have been implicated in the environmental sensitivity of gene expression [16]. Polycomb group (PcG) proteins are epigenetic regulators that repress gene expression through histone methylation and ubiquitination. Their temperature responsiveness is demonstrated in Drosophila melanogaster, where PcG-regulated genes have typically been shown to be more actively transcribed at lower temperatures than those at higher temperatures [17–19] (see also Voigt & Froschauer [20] for exceptions). Because PcGs and other chromatin regulators can impact temperature-dependent gene expression changes, loci encoding these regulators as well as their targets have been shown to be under strong selective pressures and undergo adaptive evolution [21–24]. PcGs have also been implicated in conferring phenotypic plasticity [25] or phenotypic robustness [21] of various traits including melanization. Thus, PcG genes are candidates for canalization of aposematic coloration.

In insects, PcGs act through three primary protein complexes: the Polycomb repressive complex 1 (PRC1), PRC2 and the pleiohomeotic-repressive complex. The canonical PRC1 includes dRING, Polyhomeotic, Posterior sex combs and Polycomb [26,27]. PRC1 is associated with monoubiquitination of histone H2A at lysine 119 [28]. PRC2 includes Enhancer of zeste (E(z)), Embryonic ectoderm development, Suppressor of zeste and Chromatin assembly factor 1, p55 subunit (Caf1−55) [29]. PRC2 can associate with two additional proteins, Jing and Jarid2, which modify the activity of PRC2 [30]. PRC2 is involved in mono-, di- and trimethylation of lysine 27 on histone H3 [29].

The milkweed bug, Oncopeltus fasciatus, is notable for its aposematic orange and black coloration that protects against predation [31,32]. Praying mantids, for example, co-occur with O. fasciatus and can recognize the luminance contrasts of adult O. fasciatus and avoid them [31,32]. Adult O. fasciatus have blackheads with a V-shaped orange pattern, black pronotum with an orange border, a black scutellum, forewings that comprise an orange proximal leathery section with a black band and a black distal membranous portion. The black patterns are owing to the deposition of melanin [33,34] and are unique to the adult stage. Depending on the section of the body, the melanic coloration of O. fasciatus exhibits variation in plasticity. The dorsal adult wing melanization is robust and is only minimally affected by rearing temperature [35], possibly to ensure better recognition by potential predators. In contrast, the ventral adult abdominal melanization of O. fasciatus exhibits extensive phenotypic plasticity in response to rearing temperatures [35,36]. This variability in plasticity of melanic patterns offers us an opportunity to probe potential mechanisms underlying canalization.

In this study, the role of histone modification on robustness and phenotypic plasticity was examined in O. fasciatus. As PcGs have been implicated in the temperature-dependent plasticity of abdominal melanization in D. melanogaster [21,37], we hypothesized that PcGs may also play a role in plasticity and robustness of aposematic coloration of O. fasciatus at different rearing temperatures. Prior studies in hemimetabolous insects—insects that undergo incomplete metamorphosis, such as O. fasciatus—have found that PcGs regulate patterning and specification of segmental identity during the embryonic stage [38]; whether or not PcGs might play a role in hemimetabolous adult pigmentation has yet to be explored. Therefore, we examined the role of several genes encoding members of PcGs: Pc and E(z), which encode components of PRC1 and PRC2, respectively, and jing, which encodes a zinc-finger protein associated with the PRC2 subcomplex PRC2.2 [30,39]. E(z) is a histone methyltransferase, which silences gene expression through methylation of histone H3 on lysine 27 (H3K27) [29,40,41]. In D. melanogaster, E(z) has been shown to modulate temperature-sensitive target genes [18]. Jing, a homologue of the mammalian AEBP2, appears to aid in optimal PRC2 function by stabilizing the PRC2 complex [42]. Jing has been shown to play several roles during development including the differentiation of the central nervous system midline, trachea and wing veins, and the proximodistal axis establishment and segmental development of legs in D. melanogaster [39,43]. Importantly, Jing has also been shown to modulate abdominal pigmentation in D. melanogaster [43].

Using RNA interference (RNAi), we knocked down the expression of Pc, E(z) and jing, and examined the impact on the robustness of aposematic coloration in O. fasciatus. These genes encode proteins in different polycomb repressive complexes, and our findings indicate that they impact plasticity in distinct ways. We found that E(z) and jing knockdown led to increased plasticity and a loss of dorsal melanization robustness across a temperature gradient. In contrast, Pc knockdown caused increased melanization across all temperatures.

2. Methods

(a). Animals

O. fasciatus were raised at 26.5°C on organic sunflower seeds and water in plastic containers. For the temperature experiments, O. fasciatus were raised separately at 20, 26.5 and 33°C under a 16 h L : 8 h D photoperiod.

(b). mRNA isolation and cDNA synthesis

Whole bodies of O. fasciatus were placed in TRIzol and frozen until they were processed. The RNA was isolated using standard chloroform extraction. RNA was treated with DNase (Promega), and cDNA was generated from 1 μg of the RNA using the RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific) following the manufacturer’s instructions.

(c). Double-stranded RNA synthesis and injection

Fragments of Pc, E(z) and jing were amplified using the primers listed in electronic supplementary material, table S1. The polymerase chain reaction (PCR) products were then inserted into a TOPO TA vector (Thermo Fisher Scientific). Following the transformation of Escherichia coli cells with this plasmid, the cells were grown, and the plasmid was purified using a Miniprep kit (Qiagen). After sequencing to verify the correct insertion, the plasmids were linearized using the restriction enzymes, SpeI or Not1. Single-stranded RNA (ssRNA) was generated using T3 and T7 MEGAscript kits (Thermo Fisher Scientific). Equal amounts of the ssRNA were combined to generate a 2 µg/µl solution. The ssRNA was then annealed to form double-stranded (dsRNA) as described by Hughes & Kaufman [44]. The dsRNA and the ssRNAs were run on a gel to verify proper annealing.

(d). Double-stranded RNA injections

Fourth-instar nymphs were injected with 1 μg (0.5 μl) of Pc, E(z), jing and ampicillin resistance (amp^r ) dsRNA using a syringe and a pulled borosilicate needle. The dsRNA-treated O. fasciatus were then reared at 20, 26.5 or 33°C. Whole bodies of adult O. fasciatus were kept frozen at −20°C. The ventral abdomen of each bug was fixed in 3.7% formaldehyde and mounted in a 70% glycerol : 30% PBS solution or an 80% glycerol : 20% water solution and then imaged.

(e). Analysis of body size and melanization

To examine the effects of knockdowns, dsRNA-injected animals were reared at 26.5°C, and the legs and wings of adults were imaged. For melanization plasticity studies, the total area and melanic pigmented areas of abdominal segments A3 to A5 were measured at each temperature. The proportion of melanization was standardized by dividing the area of melanization by the area of the abdominal segment. The entire area of the forewing and the areas with melanic pigmentation were also measured for each temperature. The area of melanization was normalized by dividing the area of melanic pigmentation by the area of the entire wing. The amount of the orange area on the head was estimated by measuring the linear distance between the ocelli and the portion of this line that is orange. The amount of melanization in the scutellum was analysed by measuring the triangular area visible externally and the amount of melanization. All measurements were analysed using ImageJ (https://imagej.nih.gov/nih-image/). Raw data are available on Dryad [45].

(f). Knockdown verification

To verify knockdown of Pc, E(z) and jing, amp^r, Pc, E(z) and jing dsRNA was injected into two fourth-instar nymphs each and collected as fifth-instar nymphs 3 days after the moult. cDNA of amp^r , Pc, E(z) and jing knockdown animals was synthesized from RNA as described above. Semi-quantitative PCR was performed with ribosomal protein subunit3 (RPS3) serving as a loading control. For Pc, E(z) and jing, the PCR was run for 30, 35 or 40 cycles. For the RPS3 primers, the PCR was run for 20, 25 or 30 cycles. Semi-quantitative PCR verified the knockdown of Pc, E(z) and jing (electronic supplementary material, figure S1).

3. Results

(a). Effects of PcG knockdown on adult morphology

Fourth-instar nymphs were injected with Pc, E(z) and jing dsRNA to determine the effects of PcG gene knockdown on adult phenotypes. At 26.5°C, the Pc knockdown adults had increased melanization in the head, the pronotum, the thorax, the wings and the abdomen compared to the amp^r dsRNA-injected adults. The heads of the Pc knockdowns were missing the orange V-shape pattern that was present in the amp^r dsRNA-injected adults (figures 1 and 2a ). The forewings of O. fasciatus develop as hemelytra with a sclerotized proximal section, which is orange with a black band, and a membranous distal section, which is black. In the Pc knockdown nymphs, the melanization of the proximal band was expanded primarily along the veins in the forewing, and only a small portion remained orange (figures 1 and 3a ). In addition, the forewings of Pc knockdown adults had a significantly reduced wing length-to-width ratio compared to the amp^r dsRNA-injected adults, indicating that the wings were broader than those of the amp^r dsRNA-injected adults (electronic supplementary material, figure S2E); the total wing area was not significantly different from that of amp^r dsRNA-injected adults (electronic supplementary material, figure S2A). The thorax was mostly black with greater expansion of the melanized areas relative to the amp^r dsRNA-injected controls (figures 1a and 4a,b ). In addition, the morphology of the scutellum was altered such that the scutellum had duller posterior tip compared to the amp^r dsRNA-injected adults (figure 4). In both female and male Pc knockdowns, the second abdominal segment had black pigmentation, whereas the amp^r dsRNA-injected control lacked melanic pigmentation in this segment (electronic supplementary material, figures S3 and S4). The fifth abdominal segment in the Pc knockdown adults also had larger areas of pigmentation than the amp^r dsRNA-injected controls. The black bristles on the lateral sides of the abdomen were expanded posteriorly in most of the segments (electronic supplementary material, figures S3 and S4, inset). In contrast, in the amp^r dsRNA-injected control animals, only the anterior portion of each segment had the black bristles on the lateral side of the abdomen (electronic supplementary material, figures S3 and S4, inset).

Figure 1.

Effects of Pc, E(z) and jing knockdown on adult phenotypes at 20, 26.5 and 33°C. Dorsal (left) and ventral (right) whole-body views of the Pc knockdown, E(z) knockdown, jing knockdown and amp^r dsRNA-injected control of female O. fasciatus at various temperatures.

Figure 2.

Pigmentation plasticity of the head. (a) The heads of amp^r , Pc, E(z) and jing dsRNA-injected adults reared at different temperatures. (b) Measurement of head melanization. (c) Reaction norms of head melanization. (d) Normalized amounts of head melanization. The amount of head melanization for each treatment was divided by the average amount of melanization for the respective knockdowns at 26.5°C. Results of the one-way ANOVA with Tukey’s honestly significant difference (HSD) test conducted for values at 20 and 33°C are represented by the letters where distinct letters indicate statistically significant differences.

Figure 3.

The E(z) and jing knockdowns had increased plasticity of melanization in the forewings compared to the amp^r dsRNA-injected adults. (a) Forewings of amp^r, Pc, E(z) and jing dsRNA-injected O. fasciatus at different temperatures. The arrowhead indicates the melanin in the proximal band blending with the melanized membranous portion of the forewing. (b) Diagram showing the area measured to determine the amount of melanization of the wings. The proportion was calculated by measuring the total melanized area and dividing it by the total area of the forewing. (c) E(z) and jing knockdown wings had increased plasticity in the forewings as a function of temperature. (d) Normalized amount of forewing melanization. The amount of melanization for each treatment was divided by the average amount of melanization for the respective knockdowns at 26.5°C. Results of the one-way ANOVA with Tukey HSD conducted for values at 20 and 33°C are represented by the letters where distinct letters indicate statistically significant differences.

Figure 4.

The effects of amp^r , Pc, E(z) and jing dsRNA injections on the thoracic segments. (a,b) The dorsal (a) and ventral (b) views of the animals at different temperatures. (c) The measurements taken to determine the amount of melanization on the scutellum. (d) Quantified amount of melanization for amp^r and jing dsRNA-injected adults. (e) Amount of melanization normalized to the amount of melanization at 26.5°C. Asterisks indicate statistically significant differences (Student’s t‐test, p < 0.0001).

Both the E(z) and jing knockdown adults had a scutellum with a more pointed tip relative to that of amp^r dsRNA-injected adults (figure 4). In the E(z) knockdown adults reared at 26.5°C, the melanization of the head, the thorax and the abdomen was similar to that observed in the amp^r dsRNA-injected controls (figures 1a , 2a and 4a,b ; electronic supplementary material S3 and S4). On the wing, however, the proximal melanic band was expanded along the veins although the degree of expansion was much weaker than that of the Pc dsRNA-injected animals (figure 3a ). The jing knockdown adults also had limited effects on the adult melanization at 26.5°C. The melanization of the head, the thorax and the abdomen was similar to that observed in the amp^r dsRNA-injected controls (figures 1a , 2a and 4a,b ; electronic supplementary material S3 and S4). On the forewing, the posterior portion of the proximal melanic band was expanded distally such that the black coloration merged with the black membranous portion of the wing (figure 3a , arrowhead). Taken together, at 26.5°C, Pc knockdown caused a major increase in the melanization of the entire body, while E(z) and jing knockdown caused a minor expansion of melanization on the wings.

We also examined the morphology of the legs as knockdowns of Pc and E(z) in a holometabolous insect have been shown to cause partial homeotic transformations of thoracic segment-specific leg identities and tarsus-to-tibia transformations [46]. We measured the lengths of each leg segment and determined the third thoracic (T3) leg segment-to-first thoracic (T1) leg segment ratios and the T3-to-second thoracic (T2) leg segment ratios (electronic supplementary material, figure S5). These ratios did not differ significantly between the knockdown and amp^r dsRNA-injected adults (electronic supplementary material, figure S5C), indicating that at least when injected into fourth-instar nymphs, the PcG dsRNAs do not cause homeotic transformations of leg identities. In addition, the tibia–tarsus ratios did not differ between the knockdown adults and the amp^r dsRNA-injected adults, indicating that the tarsus did not acquire a tibia-like morphology (electronic supplementary material, figure S5C). These results demonstrate that the knockdowns of PcG genes had minimal impacts on the leg morphology. The legs can therefore be used as a proxy for body size. Although E(z) and Pc knockdown adults had slightly increased leg segment lengths, the alterations were minor and inconsistent across different segments (electronic supplementary material, figure S5B). Thus, no major changes in body size were noted across the different knockdown animals.

(b). jing knockdown increases phenotypic plasticity of head melanization

We next explored the roles of Pc, E(z) and jing in regulating the temperature-dependent plasticity and robustness of melanic pigmentation. In the amp^r dsRNA-injected adults, the melanization of the dorsal side decreased at higher temperatures (figures 1–4). The heads of the Pc knockdown adults were black at all temperatures (figures 1 and 2). On the dorsal side, the heads of the E(z) and jing knockdown adults did not have the orange V-shape pattern and were mostly black at 20°C (figure 2a ). At higher temperatures, the orange area expanded to be similar to that seen in amp^r dsRNA-injected animals. To see whether the degree of plasticity was increased, the 20 and 33°C measurements were normalized to the amount of melanization observed at 26.5°C. A two-way ANOVA revealed that there was a statistically significant interaction between the effects of the dsRNA injection and temperature (F_6,164 = 4.4121, p = 0.0004) (table 1). Since Pc knockdown heads were completely black at all temperatures, we focused on the reaction norms for amp^r , E(z) and jing dsRNA-injected adults (figure 2c,d ). A comparison of head melanization across the dsRNA treatments showed that the melanization of the head of the jing knockdown had the greatest increase in melanization at 20°C relative to those reared at 26.5°C. The increase in the melanization of E(z) knockdown head was statistically indistinguishable from that of the amp^r dsRNA-injected control heads and the jing knockdown heads. At 33°C, no significant differences were observed in the amount of reduction of melanization across the amp^r , jing and E(z) dsRNA-injected animals (figure 2d ). These results demonstrate that jing RNAi caused the heads to have increased plasticity. In contrast, because the head of Pc knockdown adults were all black at all temperatures, the plasticity was completely removed.

Table 1.

Two-way ANOVA results.

		df	F	p‐value
head	gene	3	0.6214	0.6021
	temperature	2	67.695	<0.0001
	gene × temperature	6	4.4121	0.0004
wing	gene	3	7.7871	<0.0001
	temperature	2	65.4683	<.0001
	gene × temperature	6	7.4752	<.0001
scutellum	gene	1	4.4451	0.0386
	temperature	2	33.2042	<.0001
	gene × temperature	2	18.4331	<.0001
A3 female	gene	3	6.2111	0.0007
	temperature	2	142.7434	<0.0001
	gene × temperature	6	9.3711	<0.0001
A4 female	gene	3	2.0669	0.1098
	temperature	2	73.01	<0.0001
	gene × temperature	6	1.7372	0.1207
A5 female	gene	3	6.3473	0.0006
	temperature	2	16.4447	<0.0001
	gene × temperature	6	7.2443	<0.0001
A3 male	gene	3	1.5204	0.2159
	temperature	2	117.3304	<0.0001
	gene × temperature	6	6.087	<0.0001
A4 male	gene	3	0.9735	0.4097
	temperature	2	95.0001	<0.0001
	gene × temperature	6	4.9816	0.0002
A5 male	gene	3	4.8241	0.0039
	temperature	2	25.5526	<0.0001
	gene × temperature	6	5.2587	0.0001

(c). jing and E(z) knockdown increases phenotypic plasticity of wing melanization

We next examined the effects of these gene knockdowns on the melanization of the forewings. The forewings of normal adults have a black proximal band and a black distal membranous portion. The amount of melanization in the proximal band varied depending on the gene that was knocked down and the temperature. When normalized to the melanization at 26.5°C, there was a statistically significant interaction between the effects of the dsRNA injection and temperature (F_6,187 = 7.4752, p < 0.0001) (table 1). Pc knockdown led to increased melanization of the wings across all three temperatures and limited plasticity (figure 3a ). In contrast, plasticity was observed in amp^r , E(z) and jing dsRNA-injected animals. We therefore focused on the amount of melanization in the entire wing for amp^r , E(z) and jing dsRNA-injected animals and found that at 20°C, increased melanization was observed in the E(z) and jing knockdown adults (figure 3a,c,d ). At 26.5°C, the wings from E(z) knockdown adults had reduced melanization and jing knockdown adults had even greater reduction in melanization (figure 3a,c,d ). At 33°C, jing knockdown wings had the least melanization (figure 3a,c,d ). When normalized to the 26.5°C melanization, the E(z) and jing knockdown wings had the greatest increase in melanization at 20°C relative to the wings at 26.5°C (figure 3d ). Overall, forewings from the jing knockdown adults exhibited the greatest amount of plasticity although E(z) knockdown adults also exhibited higher plasticity than amp^r dsRNA-injected adults. Although the Pc knockdown animals exhibited altered melanization in the hindwings, the hindwings—which are normally hidden behind the forewings—overall did not exhibit obvious changes in melanization across temperatures for any of the knockdown treatments (electronic supplementary material, figure S6).

(d). jing RNAi increases phenotypic plasticity of thoracic melanization

jing knockdown adults exhibited thoracic melanization plasticity, whereas E(z) and Pc knockdown did not appear to increase in plasticity (figure 4). Pc knockdown adults had increased melanization in their thoraces across all temperatures and were mostly black (figure 4). Because only jing knockdown adults showed notable changes in plasticity relative to the amp^r dsRNA-injected adults, we focused on jing knockdown adults. To quantify the plasticity in the jing knockdown thorax, the scutellum was isolated and the amount of melanization was quantified. Compared to the amp^r dsRNA-injected adults, the jing knockdown adult had decreased melanization at 33°C (figure 4a,b,d,e ). When normalized to the melanization at 26.5°C, jing exhibited a significantly higher amount of melanization at 20°C and significantly reduced melanization at 33°C (figure 4e ). The interaction between the effects of the dsRNA injection and temperature was statistically significant (F_2,70 = 18.4331, p < 0.0001) (table 1). Thus, the removal of jing led to an increased plasticity in the scutellum. The ventral thorax of jing knockdown adults was also observed to be mostly devoid of melanization at 33°C, indicating that jing normally maintains robustness of melanization in the thorax.

(e). Segment-specific effects of jing, E(z) and Pc RNAi on ventral abdominal pigmentation plasticity

The abdomens of O. fasciatus are sexually dimorphic. Therefore, the males and females were analysed separately. In the amp^r dsRNA-injected females, at 26.5°C, the females typically have one large band in the abdominal segment A3 and two spots in A4 (electronic supplementary material, figure S3). At 20°C, these melanic patterns expanded, and two additional spots appeared in both A2 and A5. At 33°C, two spots developed in A3 and A4 (electronic supplementary material, figure S3). In the males, at 26.5°C, bands developed in A3 and A4 (electronic supplementary material, figure S4). At 20°C, the bands expanded, and two additional spots appeared in A2 and A5 (electronic supplementary material, figure S4). At 33°C, two spots developed in A3 and A4 (electronic supplementary material, figure S4). In addition, on the lateral sides of each of the abdominal segment, small black spots developed on the anterior portion along the lateral margins of each segment (insets in electronic supplementary material, figures S3 and S4).

In both the male and female abdominal segments of Pc knockdowns, A2 through A5 developed large bands at 20°C unlike other knockdown animals (electronic supplementary material, figures S3 and S4). At both 26.5 and 33°C, two melanic spots appeared on A2 through A4 of the female. In the males reared at 26.5°C, two bands developed on A3 and A4 (electronic supplementary material, figure S4). At 33°C, two spots developed on A3 and A4 (electronic supplementary material, figure S4), and occasionally also on A2 (not shown). For both females and males, the melanic marks on the anterior portion of the lateral margins were expanded posteriorly (insets in electronic supplementary material, figures S3 and S4).

In the E(z) knockdown adults, the plasticity in abdominal melanization was similar to that of amp^r dsRNA-injected adults (electronic supplementary material, figures S3, S4, S7B and S8). Measurements of melanized area normalized to the measurements at 26.5°C demonstrated that the change in melanization was similar to that seen in amp^r dsRNA-injected adults except in the A5 where females exhibited greater plasticity and males exhibited less plasticity compared to the amp^r dsRNA-injected adults (electronic supplementary material, figure S8). In the jing knockdown adults, increased plasticity was seen in A3 for both sexes and in A4 for the males (electronic supplementary material, figure S8). Overall, no consistent change in plasticity was observed in the ventral abdominal melanization.

4. Discussion

Canalization is an important mechanism by which organisms evolve developmental stability and robustness. For traits like aposematic coloration, evolution of robust phenotypic development amid environmental fluctuations may contribute to increased survival and fitness. In this study, we explored the potential role of epigenetic regulation on adult melanin plasticity and robustness in O. fasciatus.

(a). Pc is involved with patterning in the abdomen and wings

The knockdown of Pc led to significantly more melanization in the second and fifth abdominal segments, the lateral edge of each abdominal segment, the head and the forewings. The pigmentation of O. fasciatus in the body overall is regulated by the melanin pathway, in which dopamine is converted to black melanin [34]. Our head and the abdomen of the Pc knockdown nymphs resemble those seen when ebony and black are knocked down [33]. Ebony and black both play a role in N-β-alanyldopamine (NBAD) synthesis, and knockdown of ebony or black leads to an increased amount of dopamine and melanization [33]. Thus, Pc may be involved in reducing melanin synthesis and may promote the production of NBAD instead. In the ebony and black knockdowns, the orange areas of the wings were darker such that the entire wing became black. In contrast, when Pc was knocked down, the shape of the melanic pattern was altered. This indicates that Pc may be involved in regulating the establishment of the pattern in addition to reducing melanin synthesis. Alternatively, because the veins supply melanin [34] and Pc knockdown leads to an expansion of the melanization along the veins, Pc may impact how much melanin precursor is transported through the veins.

(b). E(z) and jing buffer against temperature fluctuation

Although Pc repressed overall melanization across all temperatures, it did not appear to regulate phenotypic plasticity as knockdown of Pc led to consistently increased melanization across all temperatures. In contrast, the dorsal melanization of E(z) and jing knockdown adults exhibited increased sensitivity to temperature. E(z) and jing knockdowns exhibited increased temperature-dependent phenotypic plasticity in the wings compared to the Pc knockdown and the control animals (figure 3). jing knockdown, in particular, increased the amount of plasticity exhibited in various tissues, including the head (figure 2) and the thorax (figure 4). jing has also previously been shown to be involved in regulating abdominal pigmentation of D. melanogaster [43]. These observations suggest that E(z) and jing play a role in maintaining robustness of melanization. Previous research has also shown that epigenetic regulators, such as histone deacetylases and PcG proteins, play a role in the nutrition-sensitive plasticity of the mandibles of male beetles Gnatocerus cornutus [47]. Thus, PcGs may play important roles in the process of canalization and the evolution of plasticity.

PcG-regulated genes have been shown to evolve through temperature-dependent selection. Studies of different D. melanogaster populations have demonstrated that the temperature sensitivity of PcG target genes has diverged between tropical and temperate regions with reduced plasticity found in populations from temperate regions [20,48]. In addition, variants of the PcG genes themselves have been shown to be under selection in northern temperate populations of D. melanogaster, leading to the evolution of thermal plasticity [21]. Thus, PcG genes and their targets can evolve to shape phenotypic plasticity and robustness.

At this point, we do not know which genes are regulated by PcGs in O. fasciatus. Because PcGs generally repress gene expression, we propose that PRC2 may repress melanization genes at lower temperatures in the head and the wings. When E(z) or jing are knocked down, gene expression is de-repressed at lower temperatures, leading to increased melanization of the head and wings. In D. melanogaster, abdominal pigment plasticity has been shown to be regulated by tan, a gene encoding N-β-alanyldopamine hydrolase, which catalyses the hydrolysis of NBAD to dopamine [49,50]. The promoter of tan is highly acetylated at low temperatures, leading to darker pigmentation [49]. Whether or not tan is regulated by PcGs in O. fasciatus remains to be seen.

(c). Histone modification as a key contributor to tissue-specific canalization of traits

In this study, we demonstrated the importance of E(z) and Jing in maintaining robustness of the dorsal pigmentation. The orange and black pigmentation of Oncopeltus serves as warning coloration to deter potential predators. We propose that robustness of the dorsal pigmentation is essential for predators’ ability to recognize the aposematic colour patterns and that PcG genes and/or their targets may have played an important role in canalizing the phenotypes across various temperatures. In contrast, the ventral abdominal patterns, which likely are not under similar selective pressures by predators, were highly variable and not consistently impacted when components of PRC2 were knocked down. Similarly, the hindwings did not exhibit any detectable changes in melanization. Thus, epigenetic regulators may evolve in a tissue-specific manner to confer robustness and plasticity to specific traits. We propose that the evolution of robustness in aposematic coloration and possibly more broadly in other traits involve evolution of histone modifiers. Therefore, studies on robustness would benefit from consideration of histone modification.

Ethics

This work did not require ethical approval from a human subject or animal welfare committee.

Data accessibility

Data are available in the Dryad Digital Repository [51].

Supplementary material is available online [52].

Declaration of AI use

We have not used AI-assisted technologies in creating this article.

Authors’ contributions

M.T.: conceptualization, formal analysis, investigation, writing—original draft, writing—review and editing; L.P.: funding acquisition, investigation, methodology, writing—review and editing; E.C.: formal analysis, investigation, methodology, writing—review and editing; M.H.: investigation, methodology, writing—original draft; L.T.: investigation, methodology, writing review and editing; Y.S.: conceptualization, funding acquisition, investigation, methodology, project administration, supervision, writing—original draft, writing—review and editing.

All authors gave final approval for publication and agreed to be held accountable for the work performed therein.

Conflict of interest declaration

We declare we have no competing interests.

Funding

This work was supported by the National Science Foundation grant IOS-2002354, the Dorothy and Charles Jenkins Distinguished Chair in Science funds to Y.S., and funds provided by Wellesley College.