Pharmacological Intervention of PIEZO1 for Butterfly Eyespot Color Patterns in Junonia orithya

Abstract

Background: PIEZO channels are mechanoreceptors expressed in various cells. Their contributions to animal development are not entirely clear. According to the physical distortion hypothesis, developmental organizers for butterfly wing eyespots receive and release mechanical signals in pupal wing tissues during development, initiating a calcium signaling cascade and gene expression changes. Objectives: We tested the possible involvement of PIEZO1 in butterfly wing color pattern formation, according to the physical distortion hypothesis. Methods: We performed a pharmacological intervention of PIEZO1, focusing on the eyespots of Junonia orithya. Chemical modulators of PIEZO1 and the actin cytoskeleton were injected into pupae immediately after pupation during the critical period of color pattern determination, and the eyespot color patterns of the emerging adult wings were analyzed. We also tested dimethyl sulfoxide (DMSO) because it was used as a solvent. Results: DMSO significantly enlarged most eyespots examined. In contrast, the specific PIEZO1 activator Jedi2 induced significant reduction in the dorsal hindwing eyespots. Another specific PIEZO1 activator, Yoda1, also induced similar changes, although less clearly. The mechanosensitive channel blocker GsMTx4 produced compromised eyespots in an individual, although statistical support for modification was weak. The actin polymerization activator phalloidin induced blue foci in the ventral forewing eyespots. PIEZO expression in the pupal wings was demonstrated by RT-PCR. Conclusions: These results suggest that eyespot organizers in butterfly wings may employ a PIEZO-mediated mechanotransduction pathway to regulate eyespot color patterns, supporting the physical distortion hypothesis. These results highlight the importance of PIEZO in developmental organizers in animals.

1. Introduction

All cells, regardless of their biological status, are subjected to various types of mechanical stress, including membrane tension, cellular crowding pressure, and shear stress. To sense such mechanical energy at the molecular and cellular levels, organisms are equipped with various types of mechanoreceptors [1,2,3,4,5,6]. Among them, PIEZO channels are versatile mechanoreceptors that are expressed not only in excitable cells but also in various nonexcitable cells in mammals [1,2,3,4,5,6]. Since their discovery [7], PIEZO channels have been studied intensively to reveal their involvement in various physiological processes [1,2,3,4,5,6]. PIEZO channels are large nonselective cation channels with more than 2500 amino acid residues [1,2,3,4,5,6,7,8,9,10,11,12,13,14]. PIEZO channels cooperate with cytoskeletal actin fibers [14,15,16,17] and the extracellular matrix [17,18,19], and various types of mechanical stress can activate them [20]. Toward pharmacological intervention of PIEZO channel activities in humans and other animals, the PIEZO1 activators Yoda1 [21,22,23,24] and Jedi1/2 [24] and the inhibitor GsMTx4 [25,26] have been discovered [27].

Notably, PIEZO channels play various roles not only in animal homeostasis and somatosensory responses but also in cell fate determination and other processes during development [28,29,30,31,32,33,34]. However, their roles in animal development remain rudimentary. Moreover, PIEZO channels have been studied intensively in mammalian systems, but their roles in nonmammalian animals such as insects are not well known, although PIEZO is evolutionarily conserved among vertebrates and invertebrates including insects [35,36]. We are of the opinion that the butterfly wing system provides an excellent opportunity to highlight the possible role of PIEZO channels in insect development because it has been speculated that color pattern determination in butterfly wings may involve mechanical distortion of the wing epithelial (epidermal) sheet, which is called the physical distortion hypothesis [37].

Some butterfly species, especially those in the family Nymphalidae, have distinctive eyespots on their wing surfaces. Adult wings are produced from the wing tissues during the pupal stage, and the prospective wing color patterns are determined during the critical period within several hours after pupation. Pupal wing tissues are composed of two opposing cellular sheets between which hemolymph circulates [38,39,40]. In nymphalid butterfly pupae, the positions of prospective eyespot foci are identifiable as pupal cuticle spots [41]. The focal cells at the center of the prospective eyespot are known to function as a developmental organizer, which dictates the developmental fates of surrounding immature scale cells in terms of scale colors to be expressed, on the basis of two kinds of experimental evidence. First, the transplantation of focal cells into a background area of pupal wing tissue induces an ectopic eyespot at the transplanted site [42,43,44,45]. Second, physical damage to focal cells reduces or eliminates eyespots [42,46,47]. The focal cells have been visualized in vivo as epithelial cells that are not morphologically very different from the surrounding cells [48]. Classically, putative morphogen molecules from focal cells are believed to form molecular gradients by diffusion (or by reaction-diffusion) to specify positional information for immature scale cells, producing a circular eyespot pattern [43,46]. However, this gradient model (and other modified models) is not satisfactory for explaining highly diverse butterfly wing color patterns [49]. More critically, molecular gradients must be established within a critical period of color pattern determination, i.e., within several hours immediately after pupation, over hundreds or thousands of cells [49]. This very long-range signaling must be achieved irrespective of the wing tissue size of various butterfly species within a relatively short period of time. This scaling problem in butterfly wings [50] needs a logical explanation.

Many animal color patterns are likely probabilistically determined and have been explained by reaction-diffusion models. In contrast, butterfly imago color patterns are deterministically determined, unlike other animal color patterns, but are similar to other anatomical entities. Neither as a simple or modified gradient model nor as a typical reaction-diffusion model, the induction model (as opposed to “deduction” models and for developmental “induction”) has been proposed to explain the butterfly color patterns on the basis of observational data but not on the basis of theoretical assumptions [49]. In the induction model, the signal behaviors are described as rolling balls on a surface with uniformly decelerated motion [51], suggesting mechanical signal transduction for color pattern determination. Observational studies of cuticle spots and their associated wing tissues underneath have revealed that the tissues are physically distorted at the focal site [41,48]. Transmission electron microscopy (TEM) images have shown that there is no clear physical space suitable for long-distance morphogen diffusion and gradients in butterfly pupal wing tissues in a conventional sense [40].

Importantly, butterfly wing color patterns can be modified by applying temperature shock (TS) [42,52], and the same TS-type modifications can be induced by injecting sodium tungstate [53] or other seemingly unrelated chemicals [54,55] into pupae. Notably, an additional TS-type modification inducer, fluorescent brightener 28 (FB28), was discovered to function in chitin binding and the inhibition of cuticle formation [56]. The discovery of FB28 as a TS-type modification inducer strongly argues for the functional importance of the extracellular matrix (ECM), including the pupal cuticle, in the process of adult wing color pattern determination [56,57]. In fact, the pupal cuticle appears to mediate morphogenic signal propagation for wing color patterns in butterflies [57]. The covering materials over the wing tissues (i.e., cuticle layers) are evident for the dorsal forewings and may be less evident for the ventral forewings and the hindwings. However, recent TEM images of the pupal wing tissues revealed that there are cuticle layers on the surfaces of the dorsal and ventral forewings and hindwings [40]. Together with the fact that the hindwings also have “cuticle spots” [48], both the forewings and hindwings are likely to employ the same developmental mechanisms for the color pattern formation.

On the basis of these observational and experimental results, the physical distortion hypothesis has been proposed [37], which concretizes the induction model. The physical distortion hypothesis states that physical distortion of the epithelial sheet by mechanical force functions as a morphogenic signal for color pattern determination in butterfly wings [37]. If so, mechanosensitive receptors may be expressed in pupal wing tissues, and they may respond to physical stress. Consistent with this idea, spontaneous calcium waves are released from the prospective eyespot focus [58], and physical damage induces ectopic eyespots after intensive calcium signals at damage sites [58,59].

In this study, we tested the possible involvement of PIEZO1 in the color pattern determination of butterfly eyespots. We used the blue pansy butterfly Junonia orithya for pharmacological intervention of PIEZO1 (Figure 1). This butterfly has multiple eyespots suitable for quantitative and qualitative evaluations of color pattern changes on its wings (Figure 1). We injected two PIEZO1 activators, Jedi2 [24] and Yoda1 [21,22,23,24], and one inhibitor, GsMTx4 [25,26], into the abdomen of pupae. We also injected an actin polymerization activator, phalloidin [60,61], based on the PIEZO-actin relationship [14,15,16,17]. Because these activators and inhibitors were dissolved in dimethyl sulfoxide (DMSO) for injection (excluding GsMTx4 in water), the potential effects of DMSO on color patterns were also evaluated.

2. Materials and Methods

2.1. Butterfly Rearing

Female butterflies of J. orithya (Linnaeus, 1758) were collected on the Nishihara Campus of the University of the Ryukyus in Okinawa-jima Island, Okinawa, Japan. The collected female butterflies were confined in a glass tank (300 mm × 300 mm × 300 mm) together with the host plants to obtain eggs. Hatched larvae were reared in a plastic container (35 mm× 125 mm × 125 mm) with leaves of Plantago asiatica at approximately 27 °C under L16:D8 light conditions.

2.2. Chemical Treatments

Chemical injection was performed at the right or left side of the pupal abdomen not to damage the wing tissues physically within 5 h after pupation via an Ito microsyringe MS-05 (Fuji, Shizuoka, Japan). The following chemicals were dissolved in DMSO (FUJIFILM Wako Chemicals, Osaka, Japan) and injected: Jedi2 (1.0 mg/mL; Tocris Bioscience, Bristol, UK), Yoda1 (7.1 mg/mL; Tocris Bioscience), and phalloidin (0.050 mg/mL; FUJIFILM Wako Chemicals). DMSO alone was also injected similarly. GsMTx4 (1.0 mg/mL; Tocris Bioscience) was dissolved in ultrapure water and was similarly injected. The injection volume was 2.0 μL per pupa. The highest concentrations at the maximal solubility were first used for injections. When this was not suitable due to high mortality, the concentrations were lowered to obtain surviving individuals. The total number of treated individuals, the number of individuals with successful eclosion (males and females), and the percentages of successful eclosion among the treated individuals (eclosion rate or ER) are shown in

Table 1. Number of individuals used for the injections in this study *.

Butterfly Species	Chemical	Trial No.	Total Number of Treated Individuals (Both Sexes)	Number of Individuals with Successful Eclosion (Both Sexes)	Percentage of Successful Eclosion (Eclosion Rate, ER)	Number of Females with Successful Eclosion
J. orithya	Jedi2	#1	30	15	50%	8
	Yoda1	#1	32	11	34%	5
	GsMTx4	#1	11	9	82%	8
		#2	9	9	100%	7
	Phalloidin	#1	25	15	60%	7
	DMSO	#1	33	20	61%	13
		#2	55	17	31%	8

* Numbers of individuals with different siblings were recorded as entries of different trial numbers in this table.

. Among the adult individuals with successful eclosion, the numbers of female individuals with successful eclosion are also shown in Table 1. We first examined the relationship between the DMSO-treated group and the untreated group using two sibling groups and obtained correction factors for each element and subelement to make the quantitative comparisons between the untreated group and the experimental group possible.

We focused on females because females have larger and clearer eyespots than males do in this sexually dimorphic species, with the exception of the phalloidin-induced phenotypic analysis of scales. Because there are color pattern variations among individuals of this species, untreated female individuals were saved for comparison with treated individuals from the same sibling group. Individuals treated with Yoda1 and GsMTx4 (first trial) were siblings and were compared to the same untreated group. The wings of all the treated and untreated individuals used in this study are presented in Supplementary Figures S1–S5. These specimens of J. orithya were stored in our laboratory at the University of the Ryukyus, Okinawa, Japan.

2.3. Image Analyses

Eyespot color pattern modifications induced by these chemicals were quantitatively evaluated based on image analyses (Figure 2). Anterior and posterior dorsal forewing eyespots (Eyespots A and B in Figure 1), anterior and posterior ventral forewing eyespots (Eyespots C and D in Figure 1), and anterior and posterior dorsal hindwing eyespots (Eyespots E and F in Figure 1) were subjected to area measurements using a Keyence Digital Microscope VHX-7000 (Osaka, Japan). Owing to the ambiguity of the eyespot boundary in the forewing, we measured the area of the eyespot core disk on the dorsal and ventral sides of the forewing (Figure 2). The area to be measured (region of interest, ROI) was objectively recognized using a threshold automatically specified by the built-in software. However, we visually confirmed the soundness of each ROI. The core disk area value was divided by the squared value of the width of the compartment in which the eyespot core disk was present to standardize the wing size variation among individuals. To obtain the width of the compartment, a line on the central structural scales perpendicular to the wing veins was drawn, and its length was measured (Figure 2). We also measured the area of the blue focus (Figure 2). In the hindwing eyespots, the entire eyespot area, the core disk area including the blue focus, and the width of the compartment were measured as above. The area values of an individual were corrected by the corresponding correction factors for the DMSO effect when applicable. The mean value of the untreated group was set to one, and the other values were adjusted accordingly.

With these area values obtained above, we performed bisided unpaired t tests after the F test for equal variance using JSTAT 16.1 (Yokohama, Japan) and Microsoft Excel (Microsoft Office 365). One GsMTx4-treated individual in the first trial of the GsMTx4 injection was omitted from the quantification of the dorsal hindwing eyespots because their eyespots were too deformed to accurately quantify their sizes. As a matter of convention, we considered p < 0.05 statistically significant for each comparison. However, considering that 16 traits were individually examined in one treatment, statistical significance at the level of p < 0.05 may be obtained by chance. This argument is reasonable, considering that two sibling results were not completely the same (see two trials of DMSO and GsMTx4 in the Section 3). In this sense, we consider that lower p-values (especially at the level of p < 0.001) were more statistically and physiologically important than higher p-values when interpreting the present statistical results.

2.4. RT-PCR and DNA Sequence Analyses

To examine the expression of PIEZO1 or its related PIEZO-type mechanosensitive ion channel at the mRNA level, we first stored pupae within 24 h after pupation at –80 °C. The pupal wing tissues were dissected out using forceps and scissors and subjected to RNA extraction using NucleoSpin RNA Plus (Takara Bio, Kusatsu, Shiga, Japan) in accordance with the manufacturer’s protocol. We then performed RT-PCR (reverse transcriptase-polymerase chain reaction) using PrimeScript II High-Fidelity RT-PCR Kit (Takara Bio). RT reaction conditions were as follows: 30 °C (10 min), 42 °C (30 min), 95 °C (5 min), and 4 °C. For PCR reactions (both the first and nested PCR reactions), we used TaKaRa Ex Premier DNA Polymerase (Takara Bio) under the following conditions: 94 °C (1 min) and then 35 cycles of 98 °C (10 s), 55 °C (15 s), and 68 °C (30 s).

Primer sequences were designed in reference to PIEZO mRNA sequences predicted by genomic DNA sequences reported from the painted lady butterfly Vanessa cardui (GenBank Accession Number: XM_047105143.1) and its Hawaiian relative Vanessa tameamea (GenBank Accession Number: XM_064220187) because the genera Vanessa and Junonia are reasonably close to each other phylogenetically in Lepidoptera [62,63,64]. The PCR primers were synthesized by and obtained from Eurofins Genomics (Tokyo, Japan). The primers designed from V. cardui for the first PCR were: 5′-CATTTTTGCCAGTCTACCAGATGTC-3′ and 5′-CAGTTGTTCATCGGAATCGCTGC-3′. The PCR product with this primer set was expected to be 640 bp from V. cardui. The primers designed from V. cardui for the nested PCR were: 5′-GTCTACCAGATGTCGGCACAATCGC-3′ and 5′-CATCGGAATCGCTGCGAGGTAATC-3′. The PCR product with this nested primer set was expected to be 621 bp from V. cardui. The primers designed from V. tameamea for the first PCR were: 5′-CACAATCGCATAATATAGACGTCTTCAC-3′ and 5′-CATTGAAAGTGTACATGACGAGCATATC-3′. The PCR product with this primer set was expected to be 642 bp from V. tameamea. The primers designed from V. tameamea for the nested PCR were 5′-GACGTCTTCACGGAGCAGGACTACAC-3′ and 5′-CGAGCATATCACAGTTGTTCATCGGAATC-3′. The PCR product with this nested primer set was expected to be 607 bp from V. tameamea.

For negative controls of RT-PCR, an RNA sample was treated without the addition of reverse transcriptase (but with the addition of ultrapure water instead), but other steps were identical to experimental samples. For a positive control of RT-PCR, an RNA sample and corresponding primers provided in the PrimeScript II High Fidelity RT-PCR Kit were used according to manufacturer’s protocol. The expected size of the positive control RT-PCR product was 462 bp.

The PCR products, together with a DNA size marker, FastGene 1 kb DNA Ladder (NIPPON Genetics, Tokyo, Japan), were subjected to agarose gel electrophoresis (1.0% agarose in TAE, 100 V). The DNA was stained with RedSafe Nucleic Acid Staining Solution (iNtRON Biotechnology, Seongnam-si, Gyeonggi-do, Republic of Korea) and was observed using a UV transilluminator. The original gel image was presented in Supplementary Figure S6. The positive PCR products were purified with NucleoSpin Gel and PCR Clean-up (Takara Bio) and sequenced in both directions by direct Sanger dideoxy sequencing at Eurofins Genomics. The sequences were examined with BLAST (Basic Local Alignment Search Tool; https://blast.ncbi.nlm.nih.gov/Blast.cgi; accessed on 1 July 2025) and confirmed to be a novel PIEZO gene from J. orithya. The longest cDNA (mRNA) sequences obtained in the present study were then deposited in GenBank (GenBank Accession Numbers: PV866951 and PV866952).

3. Results

3.1. DMSO

We first injected DMSO because it was used as a solvent for Jedi2, Yoda1, and phalloidin in this study. In qualitative visual inspections, the DMSO-treated individuals (n = 13 in the first trial and n = 8 in the second trial) appeared to show normal color patterns, but surprisingly, these individuals might show relatively large eyespots in comparison to the untreated individuals both in the first trial (Figure 3a,b) and in the second trial (Figure 3c,d).

Quantitative evaluations revealed that this possible eyespot enlargement was indeed the case. In the first sibling group, the DMSO-treated individuals (n = 13) showed statistically significant enlargement in all eyespots (or their subelements) examined (Eyespots A–F) in comparison to the untreated individuals (n = 16) (Figure 4a–f). Similarly, in the second sibling group, the DMSO-treated individuals (n = 8) presented statistically significant enlargement in all six eyespots examined (Eyespots A-F) compared with the untreated individuals (n = 7) (Figure 5a–f). However, there appeared to be response variations among different eyespots and among different sibling groups. We obtained correction factors for each element (eyespot) and subelement (outer black ring, orange ring, and core disk) in each trial, and the correction factors from these two trials were averaged to obtain the final correction factors (Appendix A 

Table A1. Correction factors for the DMSO solvent.

Area of Interest	Sibling #1	Sibling #2	Average
Eyespot A (whole)	0.682	0.597	0.639
Outer black ring of Eyespot A	0.672	0.590	0.631
Blue focus of Eyespot A	0.739	0.633	0.686
Eyespot B (whole)	0.687	0.766	0.726
Outer black ring of Eyespot B	0.699	0.786	0.742
Blue focus of Eyespot B	0.609	0.672	0.640
Eyespot C (whole)	0.597	0.743	0.670
Eyespot D (whole)	0.816	0.778	0.797
Eyespot E (whole)	0.933	0.650	0.791
Outer black ring of Eyespot E	1.008	0.711	0.860
Orange ring of Eyespot E	1.014	0.620	0.817
Core disk of Eyespot E	0.725	0.612	0.669
Eyespot F (whole)	0.833	0.728	0.780
Outer black ring of Eyespot F	0.865	0.672	0.769
Orange ring of Eyespot F	0.919	0.757	0.838
Core disk of Eyespot F	0.660	0.752	0.706

), which were used to eliminate the DMSO effect from the results of Jedi2, Yoda1, and phalloidin in the following sections.

3.2. Jedi2

We here injected the hydrophilic PIEZO1 activator Jedi2. In qualitative visual inspections of the Jedi2-treated individuals (n = 8), the overall dorsal color patterns did not differ from those of the untreated individuals (n = 15) (Figure 6a). However, a potentially noteworthy feature seemed to be the enlargement of the blue focus area within the core disk (Figure 6b). On the ventral side, again, the overall color patterns did not change (Figure 6c), but the outer black ring may be fortified in Eyespots C and D (Figure 6d).

The quantitative analysis revealed no statistically significant differences in the dorsal and ventral forewing eyespots (Eyespots A–D) and their subelements between the Jedi2-treated group (n = 8) and the untreated group (n = 15), although the blue focus area in the dorsal forewing posterior eyespot (Eyespot B) presented a relatively low p-value for enlargement (p = 0.085) (Figure 7a–d). In contrast, two of the dorsal hindwing eyespots (Eyespots E and F) and their subelements showed highly statistically significant reduction in size (Figure 7e,f). Importantly, in both eyespots, the orange ring and core disk exhibited highly significant size reductions with very low p-values. However, the outer black ring did not seem to respond to the Jedi2-treatment in both Eyespots E and F with relatively high p-values (Figure 7e,f).

3.3. Yoda1 and GsMTx4

We next injected a hydrophobic PIEZO1 activator, Yoda1, and a PIEZO1 inhibitor, GsMTx4, and examined color pattern modifications (Figure 8a–c). In qualitative visual inspections, we did not notice any clear trends among the treated individuals in these treatments (n = 5 for Yoda1; n = 8 in the first trial and n = 7 in the second trial for GsMTx4). However, a GsMTx4-treated individual presented compromised eyespots (Figure 8c). In this individual, the core disk of the dorsal hindwing anterior eyespot (Eyespot E) was minimized, and the orange ring was enlarged so that the subsequent eyespot area quantification was not possible (Figure 8c).

The quantitative data revealed that the dorsal and ventral eyespots (Eyespots A and C) in the Yoda1-treated group (n = 5) showed significant reduction in size in comparison to the untreated group (n = 11) (Figure 9a–f). Notably, the ventral forewing posterior eyespot (Eyespot D) in the Yoda1-treated group significantly increased in comparison to the untreated group (n = 11) (p = 0.048) (Figure 9d), indicating sensitivity variation among eyespots. Additionally, the core disk of the dorsal hindwing eyespots (Eyespots E and F) in the Yoda1-treated group (n = 5) significantly reduced in size (p = 0.032 for Eyespot E and p = 0.049 for Eyespot F) in comparison to the untreated group (n = 11) (Figure 9f). These results are reminiscent of those of Jedi2. However, all the potential effects of Yoda1 appeared to be weaker than those of Jedi2; indeed, all the significant p-values for Yoda1 were found in the range of 0.01 < p < 0.05.

In contrast to the eyespots in the Jedi2-treated and Yoda1-treated groups, the eyespots in the GsMTx4-treated group (n = 8 for Eyespots A-D; n = 7 for Eyespots E and F) did not significantly differ from those in the untreated group (n = 11) in all the quantitative comparisons in the first trial (Figure 10a–f). In these statistical analyses, one individual shown in Figure 8c was excluded from the quantification of the dorsal hindwing eyespots due to the difficulty in quantification.

In the second trial using a different sibling group (Figure 11a–f), the core disk (p = 0.013) and the outer black ring (p = 0.011) in the dorsal forewing posterior eyespot (Eyespot B) in the GsMTx4-treated group (n = 7) showed a significant reduction in size in comparison to the untreated group (n = 13) (Figure 11b). Additionally, the orange ring of the dorsal hindwing anterior eyespot (Eyespot E) also showed a significant size reduction (p = 0.034) (Figure 11e). The other eyespots did not show significant change in size (Figure 11a,c,d,f). However, we obtained p-values only in the range of 0.01 < p < 0.05 for GsMTx4. Together with the first trial, the overall effects of GsMTx4 on eyespot size appeared to be minimal, if any.

3.4. Phalloidin

Because PIEZO1 activation is known to contribute to actin polymerization, we reasoned that an actin polymerization activator, phalloidin, may induce color pattern modifications similar to those of Jedi2. In qualitative visual inspections of the phalloidin-treated individuals (n = 7), no color pattern changes were clearly observed on the dorsal wings (Figure 12a), but the blue focus area appeared to be enlarged on the dorsal forewing eyespots (Eyespots A and B) (n = 7) in comparison with the untreated group (n = 9) (Figure 12b). On the ventral side, no overall color pattern changes in the phalloidin-treated individuals (n = 7) were clear in comparison to those in the untreated group (n = 9) (Figure 12c). However, notably, a distinct blue focus area emerged at the center of the core disk of the ventral forewing posterior eyespot (Eyespot D) (n = 7) (Figure 12d–f), which was not present in the untreated individuals (n = 9). Indeed, this blue focus is not usually present in this species. Moreover, similar blue focal scales emerged in the miniature eyespot associated anteriorly with the ventral forewing anterior eyespot (Eyespot C) in the phalloidin-treated individuals (n = 7) (Figure 12d,g), which was not present in the untreated individuals (n = 9).

The quantitative analysis revealed that the core disk (p = 0.036) and the outer black ring (p = 0.032) in the dorsal forewing anterior eyespot (Eyespot A) were significantly smaller in the phalloidin-treated group (n = 7) than in the untreated group (n = 9) (Figure 13a), although the p-values were in the range of 0.01 < p < 0.05. In this and other eyespots, the blue focus area did not significantly change in size (Figure 13a). The dorsal forewing posterior eyespot (Eyespot B) did not differ between the phalloidin-treated group (n = 7) and the untreated group (n = 9) (Figure 13b), but the ventral forewing anterior eyespot (Eyespot C) was significantly smaller (p = 0.037) in the phalloidin-treated group (n = 7) than in the untreated group (n = 9) (Figure 13c). On the other hand, no dorsal hindwing eyespots (Eyespots E and F) and their subelements in the phalloidin-treated group (n = 7) significantly differed from those in the untreated group (n = 9) (Figure 13e,f).

In addition to the eyespot color pattern modifications described above, the phalloidin-treated individuals presented curled scales (Figure 14), although this effect is not directly related to color pattern formation in butterfly wings. Scales are normally flat at least in this butterfly species, but the right and left sides of a scale were curled up. As a result, ground scales were visible. In females, some cover scales for white structural color covering the black core disk and the orange ring in the dorsal forewing posterior eyespot (Eyespot B) and/or in the dorsal hindwing anterior and/or posterior eyespots (Eyespots E and/or F) were curled in five of seven individuals examined, although the locations of the curled scales somewhat varied among individuals (Figure 14a–f). In males, some cover scales for white structural color at the center of the dorsal eyespots and many cover scales with blue structural color located in the background (non-elementary) area were curled in one of eight individuals examined (Figure 14g–l). As a result, the black ground scales underneath the blue structural cover scales were directly observable (Figure 14g–l). Curled scales were present not only on the dorsal hindwing surface but also on all four wing surfaces including the ventral forewing (Figure 14j,k) and hindwing (Figure 14l) surfaces. On the ventral side, cover scales for pigment colors located in the background area were also curved (Figure 14j–l). Curled scales appeared to have normal colors in both sexes. These results suggest that actin filaments may be used as rigid materials to mechanically flatten growing scales.

3.5. PIEZO Expression

Here we tested the PIEZO expression in the pupal wing tissues at the mRNA level by RT-PCR using the primer sets that were designed based on PIEZO sequences from the phylogenetically related butterflies, V. cardui and V. tameamea. A weak band of the expected size (approximately 600 bp) was observed in the first round of PCR, and a strong single band of the expected size was observed in the second round (nested) PCR in both the V. indica and V. tameamea primer sets (Figure 15). The entire RT-PCR procedures were independently repeated three times (n = 3) from the wing tissue of different individuals, confirming the reproducibility of the results. Sanger dideoxy sequencing of the nested PCR product with the V. tameamea primer set revealed that the nested band indeed encoded two unique sequences of PIEZO-type mechanosensory ion channels (GenBank Accession Numbers: PV866951 and PV866952), demonstrating the PIEZO channel expression in the pupal wing tissues of J. orithya.

4. Discussion

4.1. Physical Distortion Hypothesis and PIEZO Channels

In this study, we tested the physical distortion hypothesis, which stresses the importance of mechanical transduction to determine the eyespot color patterns in butterfly wings [37]. This hypothesis is based on observations that the center of prospective eyespots in pupal wing tissues is physically distorted during the critical period of color pattern determination [40,41,48]. In this hypothesis, the putative physical distortion waves propagate to open stretch-activated channels, allowing calcium ions to enter the cell, which then induces the downstream molecular transduction cascade to change gene expression patterns. In this study, we took advantage of pharmacological intervention, in which the modification-inducing properties of chemicals were tested through injections into fresh pupae [53]. Pupae are quite resistant to injections, and pharmacological effects can be visualized as color pattern changes in adult wings after eclosion. Here, we pharmacologically tested the involvement of PIEZO1 (or its related PIEZO-type channel), which is a candidate molecule for focal signal transduction, in the color pattern determination process.

To inject chemical compounds, we are almost obligated to use DMSO for solubilization. It has been reported that DMSO injections do not induce any noticeable qualitative effects on J. orithya, although not evaluated quantitatively [65]. However, in the present study, we clearly detected an eyespot-expanding effect of DMSO. Because of this unexpected result, the area data from the experimental groups (i.e., the Jedi2-, Yoda1-, and phalloidin-treated groups) were corrected using the correction factors that were obtained from the DMSO experiments throughout this study. How DMSO affects the eyespot development is not clear, but this effect of DMSO may be mediated by the chemical modifications of the cuticle to which PIEZO1 binds directly or indirectly. Alternatively, DMSO may modify the physicochemical properties of the plasma membrane in which PIEZO1 is embedded. Interestingly and importantly, DMSO is known to induce analgesia in locust mechanosensory neurons [66].

4.2. Jedi2 on Eyespots

In J. orithya, a hydrophilic PIEZO1 activator, Jedi2, reduced the hindwing eyespots (Eyespots E and F) in size. This reduction effect of Jedi2 was so strong that it resulted in very low p-values. In contrast, the forewing eyespots (Eyespots A–D) did not significantly differ in size. These negative results for the forewing eyespots may be due to difference in the timing of signaling between these wings because they are not synchronized developmentally [40]. Interestingly, blue focus formation seemed to be upregulated in a few Jedi2-treated individuals in J. orithya, which was evident in the dorsal forewing posterior eyespot (Eyespot B), although no size difference in the blue focus area was statistically significant. Additionally, the anterior miniature eyespots associated with the dorsal forewing eyespots were probably enhanced. Together, we conclude that Jedi2 reduced the dorsal hindwing eyespots in size and enhanced the blue focus area of the forewing eyespots in the wings of J. orithya, suggesting a regulatory role of PIEZO1 in eyespot formation.

The potential effect of Jedi2 on the blue focus detected in J. orithya was similar to heat-shock-induced modifications in Junonia almana [67]. This eyespot enlargement phenotype is called the “reversed type” [67] in relation to eyespot-reduced TS-type modifications [53]. That is, the molecular mechanisms of the TS-type and reversed type modifications are essentially the same but differ only in direction. Because tungstate, FB28, and other related modification inducers likely work through changes in cellular binding to the inner surface of the pupal cuticle [56], PIEZO1 may be expressed in immature scale cells throughout the pupal wing tissue and bind directly or indirectly to the cuticle.

How does Jedi2 reduce the eyespot size? Because Jedi2 is a PIEZO1 activator, one may think that it should enlarge eyespots. On the contrary, chemical activation of PIEZO1 may induce a refractory state to mechanical signals, in which case eyespots would be reduced in size because PIEZO1 in epithelial cells cannot be activated well by the conventional levels of mechanical signals in the presence of Jedi2. Premature eyespot organizers may use PIEZO1 to establish their functional identity and then to release a mechanical signal, in which case the levels of mechanical signals would be decreased in the presence of Jedi2. More studies are needed to resolve this issue.

4.3. Yoda1 and GsMTx4 on Eyespots

We were unable to detect the effect of Yoda1, a hydrophobic activator of PIEZO1, as clearly as Jedi2, although statistically significant differences in size were obtained in the ventral forewing eyespots (Eyespots C and D) and in the dorsal hindwing eyespots (Eyespots E and F). Although the sensitivity variation among eyespots to Yoda1 cannot be explained well, these changes may be considered essentially similar to those of Jedi2 but with lower intensity. We speculate that the ability of hydrophobic Yoda1 to reach PIEZO1 was much lower than that of hydrophilic Jedi2 in this butterfly system under the experimental conditions. Yoda1 may have to reach the intracellular portion of PIEZO1 to activate it [21,22,23,24].

We did not observe any statistically significant changes in the GsMTx4-treated individuals in the first trial. However, we obtained one striking GsMTx4-treated individual whose dorsal hindwing eyespot core disk was very small but its orange ring was expanded (Eyespots E and F). In the second trial, we observed a statistically significant eyespot reduction in two eyespots (Eyespots B and E) in the GsMTx4-treated group, but their significance may be obtained by chance because of relatively high p-values.

Among the three chemicals used in this study, Yoda1 was used at the highest concentration (7.1 mg/mL), but its efficacy was low. In contrast, Jedi2 was used at a relatively low concentration (1.0 mg/mL), but its efficacy was high. Obviously, different levels of color pattern changes among these chemicals cannot be attributed to concentration differences. Rather, the color pattern changes were likely attributable to the unique chemical activities of Jedi2, Yoda1, and GsMTx4. Indeed, Jedi2 and Yoda1 are considered highly specific for PIEZO1 at least in mammals [21,22,23,24], although GsMTx4 seems to be less specific [25]. We do not know how specific they are in butterflies. The binding sites for Yoda1 and GsMTx4 may not be conserved in butterflies, but the binding site for Jedi2 may be well conserved.

In any case, an eyespot size decrease in response to the PIEZO1 activator Jedi2 argues for the view that PIEZO1 is a functional regulator of eyespot size in butterfly wings. Our RT-PCR results further support this view. We demonstrated that the pupal wing tissue expresses a PIEZO-type mechanosensory ion channel by RT-PCR. We have not identified specific cells expressing PIEZO1 in the pupal wing tissue, but wing epithelial cells are likely expressing a PIEZO-type channel, because these cells are by far the majority of the cells in this tissue used for RT-PCR.

4.4. Phalloidin on Eyespots

Given that PIEZO1 may be activated in the process of eyespot color pattern formation, PIEZO1 may then activate actin cytoskeleton rearrangement directly or through calcium influx [14,15,16,17]. Alternatively, the actin cytoskeleton may directly receive mechanical force [68]. Either way, we tested this possibility by injecting a well-known actin polymerization activator, phalloidin. We observed color pattern changes in the forewing eyespots but not in the hindwing eyespots of the phalloidin-treated individuals. This result was different from that of Jedi2, in which only the hindwing eyespots clearly decreased in size. The statistical significance of these eyespots was within the range of 0.01 < p < 0.05. Importantly, phalloidin produced a novel blue focus at the center of the core disk of the ventral forewing posterior eyespot (Eyespot D), which is not normally present in this species. This result was similar to that of Jedi2. Taken together, these findings suggest that PIEZO1-dependent or independent calcium signals and the rearrangement of intracellular actin fibers are involved in color pattern formation in butterfly wings. However, because of the inconsistency between the results of Jedi2 and phalloidin, their relationship remains ambiguous.

We also obtained curled cover scales in the phalloidin-treated individuals. This result may not be very surprising, considering that actin and chitin filaments are components of butterfly wing scales [69,70]. It is to be noted that cover scales with metallic white and blue colors generate such colors through scale microstructures and that cover scales with nonmetallic colors generate such colors through pigments. Both types of scales were sensitive to phalloidin.

4.5. PIEZO1 Activation

Notably, PIEZO1 activation depends not only on mechanical distortion of the epithelial sheet but also on the stiffness of the extracellular matrix to which cells are attached [16,17,18,19,71]. Considering that proper extracellular materials are required to propagate morphogenic signals for eyespots in butterfly wings [56,57], together with the present results, we speculate that PIEZO1 binds to the ECM (pupal cuticle) and receives planar forces from the epithelial sheet to play a major role in determining eyespot color patterns via calcium signals and the actin cytoskeleton. Because the pupal cuticle is an important mediator of morphogenic signals, the cuticle itself may be considered a signaling machinery for color pattern determination in butterfly wings [40,56,57].

To be sure, we cannot completely exclude the possibility that the color pattern modifications observed in this study were nonspecific side effects of the chemicals injected into pupae. However, stress responses to thapsigargin (an intracellular calcium releaser that inhibits calcium-ATPase on the endoplasmic reticulum), ionomycin (a calcium ionophore that increases cytosolic calcium ions), and geldanamycin (an Hsp90 inhibitor) that are expressed as an overall darkness of the wings in J. orithya are readily distinguishable from tungstate-induced TS-type modifications [65] and from modifications obtained in this study. The expansion of the outer black ring of the hindwing eyespots (Eyespots E and F) has been reported in response to thapsigargin, ionomycin, and geldanamycin [65], which may be similar to the results of the DMSO treatment. This similarity can be understood as a result of calcium ion upregulation in the cytosol by thapsigargin, ionomycin, and DMSO. Geldanamycin is also known to induce intracellular calcium upregulation [72,73].

The color pattern determination steps must be executed during the critical period after pupation. One study on the cold-shock response in J. orithya revealed that major color pattern changes occur until 30 h postpupation [74]. However, late treatment at 18 h (and later) postpupation resulted in less severe color pattern modifications [74]. We speculate that late treatment with Jedi2, Yoda1, and GsMTx4 likely results in less severe eyespot size changes in this butterfly species.

4.6. Sequences of Eyespot Development

Although we cannot exclude the possibility that non-PIEZO-type mechanosensitive ion channels such as TRP channels may also be involved in butterfly wing color pattern development, we speculate that PIEZO1 is responsible for receiving mechanical signals at the prospective eyespot focus, causing calcium influx into epithelial organizing cells. In butterfly pupal wing tissues, a sequence of eyespot development may be described below based on the present and other studies (Figure 16). The eyespot focal site for the eyespot organizer is first determined by mechanical waves from the previous organizers, which distort the epithelium and activate PIEZO1 (Step 1 in Figure 16). Calcium influx via PIEZO1 may induce actin polymerization, tight cellular binding to the cuticle, and the formation of a focal indentation (in the case of the dorsal hindwing). Spontaneous slow calcium waves travel long distances, probably more than one millimeter, from the organizing cells [58]. Prospective eyespots are likely equipped with calcium signaling machinery [75]. This influx via PIEZO1 may trigger calcium release from intracellular calcium stores or mitochondria [58]. The distortion hypothesis states that in response to calcium influx, eyespot focal cells may change their volume or density to release mechanical signals [37]. Epithelial cell division may be triggered by PIEZO1 activation [29]. In this way, mechanical signals are then released from the eyespot organizer to produce secondary indentations (Step 2 in Figure 16). The primary and secondary indentations correspond to the inner core disk and the outer black ring, respectively. The area in between may then be specified by unknown inhibitory signals that inactivate the activator signals (i.e., distortion signals) (Step 3 in Figure 16). After the determination of the prospective colors, scales with the prospective colors are produced (Step 4 in Figure 16).

In the process of color pattern determination, the calcium influx via PIEZO1 may upregulate the expression of several important genes involved in terminal differentiation, such as a transcription factor Distal-less [76,77,78,79], the Wnt family [80,81,82,83,84,85,86], and other genes [87,88,89,90,91,92,93,94,95,96,97,98,99,100], resulting in graded gene expression. Thus, classical chemical morphogens may then play an important role. Additionally, the Hippo signaling pathway may be a candidate pathway to be activated because this pathway regulates some genes that are known to be expressed in butterfly wings [101,102,103]. Calcium ions through PIEZO1 may also trigger mechanisms similar to cellular extrusion [104,105,106,107,108,109,110,111,112,113,114], which may result in the indentation of the focal area in the dorsal hindwing. The nematic order of actin filaments may also play a role in organizer function in butterflies, as in other systems [115,116,117,118,119,120,121].

5. Conclusions

The present study pharmacologically tested the possible involvement of the PIEZO1 mechanosensitive channel in color pattern formation in butterfly wings, according to the physical distortion hypothesis. Surprisingly, DMSO enlarged eyespots, which tended to mask the effects of chemical compounds. This unique DMSO effect should be cautiously taken into account when evaluating the effect of DMSO-dissolved compounds and should be investigated regarding its mechanism in the future. After removing the DMSO effect using correction factors, we demonstrated that a PIEZO1 activator, Jedi2, clearly reduced eyespots, and the effect of another activator, Yoda1, was similar, although not extensive. The effect of an inhibitor, GsMTx4, was minimal. An actin polymerization activator, phalloidin, also reduced eyespot areas, although not very clearly, and enhanced the blue foci. PIEZO expression was detected in the pupal wing tissues. These results suggest that PIEZO1 or a similar PIEZO channel may play an essential role in eyespot color pattern determination, supporting the physical distortion hypothesis. Overall, this study emphasizes the importance of mechanical signals and PIEZO1 in organizer function during development in insects and probably in other animals. A precise characterization of the mechanical signals should be performed in the future.