Introduction

Androconia are specialized wing scales exclusively present in adult males of certain butterfly species. In some butterfly families, including Pieridae, Lycaenidae, Danaidae, and Nymphalidae, there is a wide variety of forms and distribution patterns of androconia (Kristensen and Simonsen 2003). The morphological features of androconia are often useful as taxonomic criteria for several species (e.g., Hall and Harvey 2002; Ômura et al. 2015). Because their forms are distinct from ordinary wing scales, androconia are believed to be involved in the storage and emanation of scent substances. Indeed, in many butterfly species, various volatile substances have been extracted from male wings containing androconia, and several of them serve as sex pheromones in close-range courtship displays (Vane-Wright and Boppré 1993). In Pieridae, adult males of the genus Pieris (Pierinae) have characteristic alar androconia (Warren 1961), and Pieris rapae (L.), P. napi (L.), and P. brassicae (L.) adult males emit male-specific scents, including sex pheromones (Andersson et al. 2003, 2007; Yildizhan et al. 2009).

Despite lacking characteristic external organs such as androconia, several butterfly species emit male-specific scents from their wings. North American sulfur butterflies Colias philodice Godart and C. eurytheme Boisduval (Pieridae, Coliadinae) possess characteristic secretory cells, instead of androconia, behind the ordinary wing scales in the costal region on the dorsal surface of the hind wings (Rutowski 1980; Sielberglied and Taylor 1978). Such intermembranous cells are specific to males and regarded as the secretory (scent-producing) organs based on their histological features (Rutowski 1980). Chemical analyses and behavioral investigations have confirmed that hexyl myristate, palmitate, and stearate serve as aphrodisiac sex pheromones in C. philodice, while 13-methylheptacosane, heptacosane, and nonacosane serve that role in C. eurytheme (Grula et al. 1980; Sappington and Taylor 1990a, b). Despite their low volatility, these compounds elicited clear electroantennographic (EAG) responses, indicating that these sulfur butterflies receive them as olfactory cues for mating (Grula et al. 1980).

The genus Colias has more than 80 species and is a large group of sulfur butterflies inhabiting North and South America, Eurasia, and Africa (Savela 2014). Since Colias butterflies often have a sympatric distribution with closely related species, this genus has been useful for studying the mechanisms of sexual selection and species recognition. Ultraviolet patterns on the wings are thought to be essential for precopulatory isolation and have been intensively studied for many Colias species (Brunton 1998; Rutowski et al. 2007; Sielberglied and Taylor 1973). On the other hand, the involvement of chemical signals in mating selection has been investigated in only 2 species, C. philodice and C. eurytheme. It remains unclear whether the 3 chemical traits found in previous studies, i.e., sexual dimorphism in epicuticular composition, species specificity of male-specific components, and the presence of secretory cells in the hind wing membrane, are widely conserved in other Colias species. Many comparative studies are needed to better understand the significance of chemical signals in the precopulatory isolation of Colias butterflies. The present study was conducted on the Japanese subspecies of the eastern pale clouded yellow Colias erate poliographus Motschulsky to examine (1) the presence of male-specific components, (2) quantitative changes in the components with aging, and (3) the distribution of the components and possible secretory cells of male-specific components.

Materials and methods

Insects

Adults of C. erate were obtained from our laboratory stock cultures originated from females collected in the field at Higashihiroshima (34.23ºN, 132.42ºE, Hiroshima Prefecture). Gravid females were placed in plastic containers (35 × 50 × 25 cm) to allow laying of eggs on fresh leaves of Trifolium repens (white clover) under incandescent lamps. Larvae were reared on potted plants of white clover at 25 °C under a photoregime of 16L:8D. Pupae were placed in rearing cages (30 × 30 × 50 cm) until adult emergence. Within 24 h of eclosion, adults were sexed and kept individually in plastic containers (9 cm height, 12 cm inner diameter). They were gently held between pure papers with fingers, their proboscis made to extend using forceps, and were fed with 15 % aqueous sucrose once daily. Several males and females were placed together in mesh cages (40 × 60 × 40 cm) to allow mating.

Extraction

Adult butterflies were frozen to death at −20 °C and subjected to 3-min extractions with purified dichloromethane at room temperature. We conducted the following 5 extractions. (1) At 3 days after eclosion, 13 males and 13 females were individually extracted with 2 ml of the solvent (whole-individual extracts). (2) At 1, 3, and 7 days, 5 adults of each sex were extracted together using 10 ml of the solvent (extracts for age-dependent variation). (3) Six 3-day-old males were divided into 3 parts (forewings, hind wings, and body), and each divided part of each male was individually extracted with 0.6 ml of the solvent (divided-individual extracts). (4) Forewings and hind wings of 3-day-old males were divided into 8 regions (see Fig. 1), and each region of 5 males was individually extracted with 0.6 ml of the solvent (divided-wing extracts). For each wing region, a total of 2 samples were prepared from 10 males. (5) The wing regions of F3 and H1 (see Fig. 1) were dissected from 3 males that were 3 days old. Using cotton wool washed with dichloromethane, all scales were separated from the wing membrane. The membranes and scales of each wing region were individually extracted with 0.6 ml of the solvent (extracts of wing membranes and scales).

Fig. 1
figure 1

The 8 wing regions (F1–F4 and H1–H4) of Colias erate males for divided-wing extracts

Full size image

These crude extracts were subsequently filtered, concentrated to 100 μl under a nitrogen stream at 10 °C, and stored at −20 °C until the chemical analyses.

Chemical analyses

Gas chromatography–mass spectrometry (GC–MS) analyses were performed using a Shimadzu QP5000 mass spectrometer coupled with a Shimadzu GC-17A gas chromatograph. The samples were analyzed using an Agilent J&W DB-1 capillary column (0.25 mm ID × 15 m, 0.25 μm film thickness) by using a splitless injector at 280 °C, with an oven temperature program of 50 °C (1 min isothermal) and an increase of 10 °C/min to 280 °C. Compounds were identified by comparing the retention times and mass spectra with those of authentic samples. Quantitative estimates of individual compounds were obtained by comparing the peak intensities with that of pentacosane.

Authentic samples

All chemicals, including authentic samples, were commercial products from Tokyo Chemical Industry (TCI). Hexyl myristate, palmitate, and stearate were synthesized by esterification of 1-hexanol (purity 98 %) and the corresponding carboxylic acid (>97 %) using sulfuric acid as catalyst (Shimizu et al. 2012). 13-Methylheptacosane was synthesized by Wittig reaction of n-tetradecyltriphenylphosphonium bromide (97 %) and 2-tetradecanone (97 %), and catalytic reduction of the resulting methylalkenes with Pd/C and hydrogen (Guedót et al. 2009). Octadecanal was synthesized by PCC oxidation of 1-octadecanol (98 %) (Ginzel et al. 2006). These synthetic products were purified by flash chromatography using silica gel and identified by GC–MS.

Microscopy

Microscopy was conducted for forewings and hind wings dissected from 3-day-old males and females. The forms and distribution patterns of wing scales were examined using an Olympus SZX-7 stereomicroscope. For scale-removed wings, the location of possible secretory organs was investigated using a Keyence BZ-9000 fluorescent microscope under bright conditions and digital Z-stacking mode. Scanning electron microscopy (SEM) was performed using a Jeol JSM-5400 scanning electron microscope. The preparations for SEM were mounted on a stage with adhesive tape and sputter-coated with gold. To describe the location of sex-specific microstructures on wings, the wing veins and cells were labelled as in the Comstock-Needham system (Miller 1969).

Results

Sexual differences in chemical composition of whole-individual extracts

The whole-individual extracts of males and females displayed remarkably different patterns of total ion chromatograms (Fig. 2). The extracts contained few highly volatile components but were abundant in less volatile components, in which we identified 19 compounds: 10 linear alkanes, 3 hexyl esters, 2 higher fatty acids, 1 branched alkane, 1 aliphatic aldehyde, 1 acyclic triterpene, and 1 cholesterol (Table 1). The male and female extracts had 4 major components with average amounts of more than 1 μg per individual. Hexyl palmitate was the predominant component in males, but it was present at less than 200 ng in females. In contrast, tricosane was the only major component in females. Pentacosane, heptacosane, and nonacosane were the major components shared by both sexes. With respect to sex differences in the quantity of each compound, 8 linear alkanes and octadecanoic acid were significantly more abundant in females than in males (p < 0.05 or 0.01, Welch’s t-test), while 3 hexyl esters and octadecanal were significantly more abundant in males than in females (p < 0.001, Welch’s t-test).

Fig. 2
figure 2

Typical total ion chromatograms obtained from the whole-individual extracts of Colias erate males (upper) and females (lower). Chromatograms were run on an Agilent J&W DB-1 capillary column (0.25 mm ID × 15 m), programmed from 50 °C (1 min isothermal) to 280 °C at 10 °C/min. Peak numbers correspond to compound numbers in Table 1

Full size image
Table 1 Chemical composition of the crude extract of 3-day-old Colias erate adults
Full size table

Age-dependent variation in amounts of designated compounds

To evaluate age-dependent quantitative changes, the average amounts of 9 designated compounds were calculated for either sex at days 1, 3, and 7. These compounds were 4 male-specific compounds (octadecanal and 3 hexyl esters), 4 linear alkanes as the major components (tricosane, pentacosane, heptacosane, and nonacosane), and 13-methylheptacosane. Among them, 3 hexyl esters, 13-methylheptacosane, heptacosane, and nonacosane were the sex pheromones of 2 North American Colias males (Grula et al. 1980; Sappington and Taylor 1990a, b). In both sexes, the amounts of the 4 linear alkanes increased after eclosion and reached a plateau at day 3 (Table 2). In contrast, the amounts of octadecanal, 3 hexyl esters, and 13-methylheptacosane in males increased until day 7. Although females contained small amounts of hexyl palmitate and 13-methylheptacosane at day 3, these quantities increased until day 7.

Table 2 Age-dependent increment of 9 compounds from Colias erate adults
Full size table

Location of male-specific compounds in the insect body

The relative abundance of each of the 19 compounds in the forewings, hind wings, and body of males was compared using divided-individual extracts (Table 3). Most compounds were distributed in all parts, with each of the 3 parts having at least 13 % of the total amount of compound. However, octadecanal, 3 hexyl esters, and 13-methylheptacosane had a characteristic distribution pattern; these compounds were less abundant in bodies (less than 8 %), but were concentrated in the wings, especially the forewings. This suggests that the possible secretory and storage organs of these compounds are localized on male wings.

Table 3 Distribution of the compounds of Colias erate 3-day-old males
Full size table

Using divided-wing extracts, 8 wing regions were compared for their relative abundance of hexyl palmitate, the predominant compound in males (Table 4). This compound was 3 times more abundant in the forewings (F1–F4) than the hind wings (H1–H4). Whereas the abundance in forewings was approximately constant among the 4 regions, that in the hind wings was biased toward the H1 and H2 regions.

Table 4 Distribution of hexyl palmitate on 8 wing regions of Colias erate 3-day-old males
Full size table

Microscopy of possible secretory organs

Stereomicroscopic observation of the wing surface confirmed that both sexes were identical in the forms and distribution pattern of wing scales, indicating that C. erate males lack characteristic structures such as androconia and sex brands. However, digital microscopy using the Z-stacking mode for the scale-removed wings revealed that males, but not females, had many rows of stain-like spots on the surface of the wing membrane (Figs. 3a, b, 4a, b). These spots were more conspicuous in the forewings than hind wings, and were attributed to sac-like intermembranous cells lying under the ordinary scale sockets, which had associated integumental swelling structures (Figs. 3c–e, 4c–e). These cells were exclusively distributed from basal to discal areas on the ventral surface of forewing cells 1A, 2A, and Cu2 and on the dorsal surface of hind wing cell Sc+R1 (Figs. 3f, 4f).

Fig. 3
figure 3

Distribution and morphology of scale sockets on the ventral surface of Colias erate forewings. a–d Male (a, c) and female (b, d) socket arrangement from basal to discal areas in the forewing cells 1A, 2A, and Cu2 in the Comstock-Needham system. Sac-like intermembranous cells (black arrows) are located at the bottom of several sockets of males (c). e Scanning electron micrograph of the socket together with integumental swelling (white arrow). f Distribution of intermembranous cells in the male forewing (white area)

Full size image
Fig. 4
figure 4

Distribution and morphology of scale sockets on the dorsal surface of Colias erate hind wings. a–d Male (a, c) and female (b, d) socket arrangement from basal to discal areas in the hind wing cell Sc+R1 in the Comstock-Needham system. Sac-like intermembranous cells (black arrows) are located at the bottom of several sockets of males (c). e Scanning electron micrograph of the socket together with integumental swelling (white arrow). f Distribution of intermembranous cells in the male hind wing (white area)

Full size image

Distribution of male-specific compounds between wing membranes and scales

The wing membranes and scales in the F3 and H1 wing regions, including intermembranous cells, were compared for their relative abundance of 9 designated compounds (Table 5). Among them, 4 linear alkanes were more abundant in the wing membranes, while octadecanal, 3 hexyl esters, and 13-methylheptacosane were more abundant in the wing scales.

Table 5 Distribution of 9 compounds on wing membrane and scales of Colias erate 3-day-old males
Full size table

Discussion

C. erate adults had a distinct sexual dimorphism in the chemical composition of epicuticular substances. Hexyl palmitate was the predominant component in males, and its quantity was approximately 40-fold higher than that in females. Hexyl myristate and stearate were detected in small amounts in males, but were absent in females. With regard to the overall chemical composition and the predominance of hexyl esters, C. erate males were identical to C. philodice males, which utilize 3 hexyl esters as aphrodisiac pheromones toward conspecific females (Grula et al. 1980). Although phylogenetic relationships within the genus Colias remain unclear, C. erate is in a distant clade from C. philodice and C. eurytheme (Brunton 1998; Pollock et al. 1998; Wheat and Watt 2008; Wheat et al. 2005). Since C. erate and C. philodice are biogeographically and phylogenetically distant, the male chemical profile characterized by these hexyl esters seems to be widely preserved within the genus Colias.

Among 10 identified linear alkanes with carbon numbers from 21 to 31, pentacosane, heptacosane, and nonacosane were the principal components. Although these hydrocarbons commonly occurred in both sexes, 8 linear alkanes were significantly more abundant in females than in males at 3 days old. These alkanes are omnipresent cuticular hydrocarbons in many insects (Blomquist 2010). In addition to linear alkanes, 13-methylheptacosane was also present as a minor component in both sexes. This branched alkane, together with heptacosane and nonacosane, is known to act as a sex pheromone in C. eurytheme males (Sappington et al. 1990a, b). It is noteworthy that 13-methylheptacosane was present together with hexyl palmitate in C. erate males. In a previous study, 3 hexyl esters were specific to C. philodice, while 13-methylheptacosane was exclusively present in C. eurytheme (Grula et al. 1980). Since biosynthetic capabilities leading to either hexyl esters or 13-methylheptacosane coexist in C. erate, this Eurasian Colias species might preserve more ancestral traits than the 2 North American species.

Hexadecanoic and octadecanoic acids, octadecanal, squalene, and cholesterol are often identified as cuticular lipid components in many insects (Buckner 1993; Canavoso et al. 2001; Juárez and Bromquist 1993). Among them, C. erate males contained significantly larger amounts of octadecanal than females, while octadecanoic acid was significantly more abundant in females than in males. The presence of octadecanal and octadecanoic acid has not been reported in C. philodice and C. eurytheme (Grula et al. 1980; Sappington et al. 1990a, b).

Adult butterflies frequently showed age-dependent changes in their chemical composition. Especially, quantitative changes in sex pheromone components synchronize the timing of sexual maturation and have a great impact on copulation success (Nieberding et al. 2012). In the present study, the amounts of octadecanal, hexyl esters, and 13-methylheptacosane in males increased during 1–7 days after eclosion, while those of 4 linear alkanes reached a plateau at day 3. Similar age-dependent increments within 7 days after eclosion have also been reported for the sex pheromones of C. eurytheme males (Sappington et al. 1990a). Interestingly, hexyl palmitate and 13-methylheptacosane were detectable in females and increased until day 7, indicating that C. erate females also have a small amount of these synthases.

In 3-day-old males, the wings contained more octadecanal, 3 hexyl esters, and 13-methylheptacosane than the body. These compounds were especially more abundant in the forewings than in the hind wings. The abundance of hexyl palmitate was approximately equal among the 4 regions of forewings but was greater in the 2 costal regions than the 2 inner marginal regions of hind wings. These results strongly reflect in the possible location of secretory and storage organs of the male-specific compounds. Microscopic surveys revealed that C. erate males possess characteristic intermembranous cells from basal to discal areas in the forewing cells 1A, 2A, and Cu2 and the hind wing cell Sc+R1. Since these cells were absent in the same areas of female wings and morphologically identical to the secretory (scent-producing) organs of C. philodice males (Rutowski 1980), these were regarded as the sources of male-specific compounds in C. erate. In general, adult male coliadine butterflies have a patch of secretory cells, which often forms a sex brand, in the friction area of the ventral surface of the forewing or the dorsal surface of the hind wing (Vetter and Rutowski 1978). However, the distribution pattern of secretory cells differs among sulfur butterfly species; C. philodice, C. cesonia, and Nathalis iole have them on the hind wing (Vetter and Rutowski 1978; Grula et al. 1980), while Eureme lisa and E. nicippe possess them on the forewing (Rutowski 1977; Vetter and Rutowski 1978). C. erate males uniquely possess a patch of secretory cells on both the forewings and hind wings.

In the F3 and H1 regions of male wings, including secretory cells, 4 linear alkanes were more abundant in the membranes than in the scales. Octadecanal, 3 hexyl esters, and 13-methylheptacosane, in contrast, showed larger amounts in the wing scales, although these secretory cells were present in the wing membranes. The disagreement between the origin and the presence of male-specific compounds implies a possible dissemination mechanism such that these compounds originate from the intermembranous cells and are transferred to the wing scales in the F3 and H1 regions. Then, these compounds would be spread from these wing regions to other regions by the contact of wing surfaces (Rutowski 1977). C. erate males lack androconia in these wing regions, the scales of which were not differentiated from those found in other regions of male wings and in the same regions of female wings. So far, morphologically specialized scales (androconia) have not been discovered in adult male coliadine butterflies (Vetter and Rutowski 1978). These findings demonstrate that sulfur butterflies have evolved to use their ordinary wing scales as dissemination devices for male-specific compounds.

We clarified the chemical natures, quantitative changes by aging, and origins of male-specific compounds of C. erate. Further studies are needed to investigate the pheromonal activities of these compounds.