Journal of Chemical Ecology

, 37:575

Synergistic or Antagonistic Modulation of Oviposition Response of Two Swallowtail Butterflies, Papilio maackii and P. protenor, to Phellodendron amurense by Its Constitutive Prenylated Flavonoid, Phellamurin


    • Department of Biofunctional Science and Technology, Graduate School of Biosphere ScienceHiroshima University
  • Hisashi Ômura
    • Department of Biofunctional Science and Technology, Graduate School of Biosphere ScienceHiroshima University
  • Mamoru Chachin
    • Department of Biofunctional Science and Technology, Graduate School of Biosphere ScienceHiroshima University
  • Seiji Kawano
    • Department of Biofunctional Science and Technology, Graduate School of Biosphere ScienceHiroshima University
  • Takashi A. Inoue
    • Japanese National Institute of Agrobiological Sciences

DOI: 10.1007/s10886-011-9965-9

Cite this article as:
Honda, K., Ômura, H., Chachin, M. et al. J Chem Ecol (2011) 37: 575. doi:10.1007/s10886-011-9965-9


Papilio maackii females prefer a rutaceous plant, Phellodendron amurense, for oviposition, whereas another semi-sympatric Rutaceae feeder, Papilio protenor, never exploits this plant as a host in nature. However, the larvae of both species perform well on this plant in the laboratory. Phellamurin, a flavonoid present in the organic fraction from P. amurense inhibits egg laying by P. protenor. We examined whether phellamurin is involved in the differential acceptance of P. amurense by the two butterflies. The ovipositing females of P. maackii readily accepted P. amurense and a methanolic extract of the foliage, while P. protenor rejected them entirely. However, the aqueous fraction derived from the extract elicited significant oviposition responses of similar levels from the two species. Phellamurin did not induce oviposition behavior in P. protenor females. In contrast, P. maackii was stimulated to oviposit by phellamurin at concentrations exceeding 0.2%. The response was dose-dependent and reached ca. 70% at 2% phellamurin, which is approximately equivalent to its natural abundance in young leaves of P. amurense. Since the aqueous fraction was very stimulatory to both species, the combined effect of phellamurin and the aqueous fraction on oviposition was tested. The addition of phellamurin to the aqueous fraction enhanced the ovipositional activity of P. maackii, but dramatically suppressed the oviposition response of P. protenor even at 0.1% concentration. These results, taken together with those obtained from electrophysiological recordings with foretarsal chemosensilla, indicate that phellamurin acts as an oviposition stimulant for P. maackii, and as a potent deterrent for P. protenor. The results suggest that host range expansion or host shifts may be made by ovipositing females that overcome phytochemical barriers.

Key Words

Host selectionHost shiftRutaceaeDihydroflavonol glucosideOviposition stimulantOviposition deterrentLepidopteraPapilionidae


Plant secondary metabolites can be exploited by phytophagous insects during host selection. In general, some compounds serve as stimulants/attractants for feeding or oviposition, while others act as deterrents/repellents (e.g., Schoonhoven et al., 2005; Honda et al., 2010). Although the host range or host-preference hierarchy of an insect is regionally variable, the majority of herbivorous insects use only a limited range of plant species, mostly of a single or few families, or only one plant species at the extreme. Such host specificity may be due to a species-specific capacity for metabolizing allelochemicals (Berenbaum, 1991; Feeny, 1991) and/or limitations on processing information in the central nervous system (Bernays, 2001). Extensive work on chemical mediators in oviposition by lepidopterans has demonstrated clearly that chemical attributes of plants serve as the crucial cues that permit females to discriminate and recognize their hosts (Renwick and Chew, 1994; Honda, 1995, 2005; Honda et al., 2010).

In butterflies, assessment of potential hosts is achieved mainly by drumming on the surface of foliage with foretarsi endowed with chemotactile sensilla that can perceive plant chemicals (Inoue, 2006). Acceptance or rejection of a given plant by ovipositing females is governed not only by the presence or absence of stimulatory substances, but also by other co-occurring constituents that exert inhibitory activities, hence, by a reciprocal balance of positive and negative sensory inputs received upon contact with the plant (Renwick and Chew, 1994; Honda, 1995).

While host-plant affiliations for papilionid butterflies include many plant families, 80% of butterflies in the genus Papilio feed on rutaceous plants. A wide array of phytochemical stimulants and deterrents responsible for oviposition has been reported for Papilio butterflies that feed on Rutaceae, Apiaceae, or Lauraceae (Honda, 1995; Honda et al., 2010). Major compounds identified include amino acid derivatives, sugar-related acids, alkaloids, flavonoids, and hydroxycinnamic acid derivatives. Among the stimulants for Papilio butterflies, some are common stimulants for other species, and others are structurally related to one another, thereby strongly suggesting the involvement of phytochemicals in host specialization on one hand, and expansion of host range on the other (Feeny, 1991). Many investigations have been conducted on assumed host shifts of papilionids, presumably originating from Aristolochiaceae, via Lauraceae, Magnoliaceae, Annonaceae, and Rutaceae, eventually leading to Apiaceae and Asteraceae (e.g., Aubert et al., 1999; Scriber et al., 2007, 2008).

In this study, we focused attention on two Rutaceae specialists, Papilio maackii and P. protenor, that have quite different host ranges in order to clarify the phytochemical and electrophysiological bases for the differential acceptance of Phellodendron amurense by ovipositing females. The two butterfly species range from temperate to boreal regions of Japan, with broad overlapping habitats on the main island (Honshu). Papilio maackii is rather montane, and prefers cool climates, while P. protenor is common in warm districts but relatively rare in montane or northern districts. Therefore, their occurrence is somewhat allopatric in most temperate regions. Papilio maackii utilizes P. amurense as a primary host plant in almost all habitats but P. prontenor females completely reject this species although the larvae perform well on this plant in the laboratory (Honda and Hayashi, 1995a). By contrast, Papilio protenor exclusively utilizes Citrus plants throughout its habitats.

Phellamurin [3,5,7,4′-tetrahydroxy-8-(3-methylbut-2-enyl)flavanone-7-O-ß-glucoside], which is present in P. amurense, strongly deters oviposition in a Kanagawa population of P. protenor demetrius (Honda and Hayashi, 1995b). Although many studies have investigated oviposition stimulants for Papilio butterflies, less attention has been paid to ecologically significant oviposition deterrents. It is important to understand how inhibitory phytochemicals may have played a role in establishing present-day host use and host-range evolution. The aim of this study was to examine how P. maackii females respond to phellamurin in their ovipositional behavior, and how the compound stimulates the foretarsal chemosensory receptors in both P. maackii and P. protenor. We also evaluated the oviposition response to P. amurense and phellamurin in a population of P. protenor demetrius from Hiroshima for comparison to previous work on a population from Kanagawa.

Methods and Materials


Larvae of Papilio maackii and P. protenor demetrius were reared at 25°C under a photoregime of 16:8 h L:D on potted plants of Phellodendron amurense and Citrus spp., respectively. These were laboratory stock cultures that originated from females collected in Hiroshima Prefecture. Adults of both species used for the behavioral bioassays were 3- to 10-d-old gravid females, which had been hand-paired or allowed to copulate in an outdoor cage (7 × 10 m, 3.5 m high). Females were hand-fed with 15% aqueous sucrose solution once daily throughout the experiments.

Extraction and Fractionation of Plant Materials

Young leaves of P. amurense (600 g) were extracted with methanol (4 l) at room temperature for 1 mo. The methanolic extract was concentrated in vacuo below 50°C, and an aliquot of the whole concentrate, after being dispersed in water, was successively partitioned with chloroform and isobutanol to give 3 fractions (1 aqueous and 2 organic). The chloroform-soluble fraction (Fr. 1) and isobutanol-soluble fraction (Fr. 2) were evaporated to dryness below 50°C, and each residue was re-dissolved in chloroform and in methanol, respectively. The water-soluble fraction (Fr. 3) was lyophilized and dissolved in 50% aqueous methanol. All fractions were stored below 0°C until use. Phellamurin was isolated from Fr. 2 (Honda and Hayashi, 1995b).

Bioassay for Oviposition Response

Behavioral bioassays were conducted by a method similar to that reported previously (Honda, 1990), employing a green heart-shaped plastic plate (2–5 cm2) as a leaf surrogate. Before testing, females were screened daily to assess their responsiveness, and only those that exhibited positive responses to the foliage of P. amurense (for P. maackii) or Citrus natsudaidai (for P. protenor) and negative responses to water alone (control) were chosen. In each trial (sample presentation), the response of an individual was scored as 100% for actual egg laying or an equivalent behavior (trying to bring the ovipositor in contact with the underside of the leaf without egg deposition), 50% for half-curling the abdomen with continuous drumming (this behavior took place infrequently), and 0% for drumming only, with no attempt to oviposit (Nakayama et al., 2002). Trials were replicated at least three times for each individual, and the responses of an individual to a given sample were averaged. For all trials, merely alighting on ovipositional substrates without drumming was not counted. Oviposition response to each sample was eventually represented as the mean percentage of responses recorded from at least 10 females.

Concentrations of test solutions are all expressed as % w/v. The amount of dry test materials on the artificial leaf was estimated. For example, if a 0.5% solution of test sample was applied to the leaf, the amount of material was estimated at 50 μg/cm2. Test samples prepared from the initial extract, the three partitioned fractions (Frs. 1 to 3), and phellamurin were tested at several concentrations for their stimulatory activity in ovipositioning. For test samples of a binary mixture, Fr.3 (0.5%) was used as a base component, since it exhibited potent oviposition-stimulatory activity (more than 90% response) for both species.

Electrophysiological Recordings

The 5th tarsomere of the female foretarsi of both species possesses bristles with a large number of long (major) and short sensilla trichodea (Ryan, 2002) that are situated between 2 lateral and a midline row of spines. The foretarsi also bear a small number of similar lateral sensilla. The responses of the foretarsal chemosensilla to phellamurin, Fr. 3, and their mixtures were examined electrophysiologically. To record responses from the sensilla, tip recording was performed using an ER-1 extracellular amplifier (Cygnus Technology, Inc., USA) and a PowerLab 4/25 converter (ADInstruments Pty Ltd., Australia). Data sampling was conducted at the rate of 10 k/s (low pass filter: 3 kHz; high pass filter: 100 Hz; and sampling gain: 100). Forelegs were amputated at the base of the tibia and fixed on the stage with double-faced adhesive tape. A sharpened tungsten needle (indifferent electrode) was inserted into the first tarsomere. Under a binocular microscope, a glass capillary tube (recording electrode) filled with a test solution was brought into contact with the tip of a medial long trichoid sensillum of the 5th tarsomere. Successive stimulations of a sensillum were of 1 s duration and were separated by an interval of at least 3 min. One to 5 sensilla of the same type were randomly selected from the same preparation. A total of 18 sensilla from 6 females of P. maackii and 19 sensilla from 7 females of P. protenor were subjected to the measurement. Spike trains were analyzed with Spike Histogram software (ADInstruments Pty Ltd., Australia). The frequencies (F) and average intensities (I) of spikes with amplitudes higher than 10 μV were quantified, regardless of their spike types, using 5 recordings for each stimulus. Test materials (stimuli) were dissolved in 6–12% aqueous ethanol containing 10 mM NaCl as an electrolyte.


Oviposition Responses to the Foliage, Extract, and Fractions of Phellondendron amurense

Ovipositing P. maackii females responded strongly to the foliage, methanol extract, and Fr. 3, and moderately to Fr. 2., while Fr. 1 (a mixture of less polar substances) was weakly active (Fig. 1). It was noteworthy that P. maackii females responded positively to Fr. 2, as this isobutanol-soluble fraction has been reported to be almost inactive by itself for the Rutaceae-specialized Papilio species thus far investigated (Honda and Hayashi, 1995a; Nakayama et al., 2002; Murakami et al., 2003; Chachin et al., 2007).
Fig. 1

Oviposition responses (mean±SE) of the two species (Papilio maackii and Papilio protenor) to the foliage of Phellodendron amurense, the methanol extract of the foliage, and fractions of the extract prepared as described in the text. The extract and fractions were tested at 0.5%

In marked contrast, P. protenor females responded positively only to Fr. 3, and totally rejected the other fractions (Fig. 1). The response of P. protenor females of the Hiroshima population to each sample was virtually identical to that of the Kanagawa population reported previously (Honda and Hayashi, 1995a, b).

Oviposition Response to Phellamurin

Females of P. maackii had positive oviposition responses to phellamurin in a dose-dependent manner (Fig. 2). With 2.0% phellamurin, corresponding to the concentration of the compound found in young leaves of P. amurense (Honda and Hayashi, 1995b), the response was 72.4%. Ovipositing females of P. protenor did not respond to phellamurin.
Fig. 2

Oviposition responses (mean±SE) of the two species (Papilio maackii and Papilio protenor) to various concentrations of phellamurin . The number of individuals tested is given in parentheses or brackets

Mixtures of Phellamurin and Fr. 3

As shown in Fig. 3, the addition of phellamurin enhanced oviposition responses of P. maackii females to Fr. 3. In P. protenor, the stimulatory activity of Fr. 3 was suppressed upon addition of a very small dose (0.1%) of phellamurin. There was almost complete rejection at higher doses (0.2% or more). This tendency is similar to that reported previously for the Kanagawa population (Honda and Hayashi, 1995b), though individuals of the Hiroshima population appear to be more sensitive to phellamurin.
Fig. 3

Dose-response relationships for phellamurin mixed with Fr. 3. The oviposition behavior of the two species (Papilio maackii and Papilio protenor) expressed as mean±SE for various concentrations of phellamurin mixed with Fr. 3 (0.5%). The number of individuals tested is given in parentheses or brackets

Electrophysiological Responses

Recordings from foretarsal taste sensilla were conducted with aqueous solutions of ethanol (solvent), phellamurin, Fr. 3, and their mixtures. Sucrose was used as the reference stimulus. Figure 4 shows typical recording traces for P. maackii and P. protenor. The sensilla of both species were unresponsive to the solvent (12% ethanol) and 62.5 mM sucrose (traces 1 and 6), but exhibited unambiguous responses to stimulation with the other solutions. The sensilla of P. maackii were more intensely stimulated with phellamurin and/or Fr. 3 than were those of P. protenor, with the former displaying spikes of lower amplitude. In both species, phellamurin appeared to evoke regular spikes from 1 neuron (trace 2). The frequencies (F; mean±SD) and intensities (μV) (I; mean±SE) of spikes evoked in the sensilla of P. maackii and P. protenor were (F) 39 ± 19, (I) 210 ± 14, and (F) 29 ± 9, (I) 341 ± 15, respectively.
Fig. 4

Electrophysiological recordings from the foretarsal medial large sensilla (s. trichodea) for the two species (Papilio maackii and Papilio protenor). Examples of traces in response to aqueous solutions of (1) 12% ethanol, (2) 0.2% phellamurin +12% ethanol, (3) 0.5% Fr. 3 + 10% ethanol, (4) 0.5% Fr. 3 + 0.1% phellamurin +6% ethanol, (5) 0.5% Fr. 3 + 0.2% phellamurin +12% ethanol, and (6) 62.5 mM sucrose. Each sample contained NaCl (10 mM) as an electrolyte and all traces were from the same sensillum. Amplitude scale bar: 500 μV

In contrast, Fr. 3 , which contained less than 62.5 mM sucrose, elicited several types of spikes with different amplitudes, and its stimulation mode for P. maackii was diverse as compared to that for P. protenor (trace 3). The spike trains of both species were so complicated that we were only able to approximately evaluate their response patterns by lumping together all spike types. The mean spike frequencies and intensities of P. maackii and P. protenor were (F) 102 ± 9, (I) 312 ± 19, and (F) 26 ± 6, (I) 407 ± 20, respectively.

Although the spike trains of P. maackii in response to the mixtures of phellamurin and Fr. 3 (traces 4 and 5) were too complicated to analyze with the software available, the mean spike frequency and intensity [(F) 104 ± 4 and (I) 271 ± 27] recorded by stimulation with 0.5% Fr. 3 plus 0.2% phellamurin (trace 5) were similar to those observed with Fr. 3 alone. The sensory response pattern in P. protenor to the same mixture (trace 5) was less complex. The impulses evoked by stimulation with mixture [(F) 68 ± 20 and (I) 426 ± 21] may have represented an additive response because the spike count of the mixture was virtually equivalent to the sum total of spikes evoked by the respective samples.


The rejection of Phellodendron amurense by ovipositing females of Papilio protenor can be ascribed to negative effects of phellamurin in the foliage. Fraction 1 had no oviposition-inhibitory effect (Honda and Hayashi, 1995a). The mixture we tested (Fig. 3) composed of 0.5% Fr. 3 and 0.25% phellamurin, approximated the relative concentrations in the living plants., We conclude that the deterrent activity of phellamurin far outweighs the potent stimuli evoked by water-soluble substances (Fr. 3), thus resulting in persistent avoidance of the plant by females. The essential agreement on the dose-response relationship for phellamurin between Kanagawa and Hiroshima populations (located 650 km apart) indicates that the circumvention of phellamurin by P. protenor females is not a regionally-expressed but species-specific trait.

Fraction 2 (isobutanol-soluble substances) consists mainly of flavonoid glycosides, including phellamurin as a major component (Honda and Hayashi, 1995b). Thus, phellamurin is most likely responsible for the positive response to Fr. 2 by P. maackii females. Phellamurin alone exerted considerable oviposition-stimulatory activity for P. maackii, in contrast to other flavonoid stimulants that are effective only when they operate together with other structurally unrelated substances (Ohsugi et al., 1985; Nishida et al., 1987; Honda, 1990). Furthermore, marked synergism of multiple components in host recognition generally prevails in Papilio butterflies (Renwick and Chew, 1994; Honda, 1995; Honda et al., 2010). In addition to its direct action in stimulation of oviposition, phellamurin synergistically enhanced the responses of P. maackii females to Fr. 3 (Fig. 3). Thus, for P. maackii, phellamurin acts as an independent or a synergistic oviposition stimulant probably involved in host preference.

The foretarsal sensilla of the two butterfly species exhibited species-specific electrophysiological responses to Fr. 3 and phellamurin. In both species, the spikes generated by stimulation with Fr. 3 (trace 3) seem to originate from at least 2 or more different sensory receptor cells that are responsive to some as yet uncharacterized oviposition stimulants present in the fraction. Despite an enormous difference in its effect on the two butterflies’ behavior, phellamurin elicited one type of spikes from both species. In P. protenor, phellamurin (a deterrent), which is received by a specific cell distinct from those involved in the reception of stimulant(s), would transmit a behaviorally negative signal to the central nervous system. In P. maackii, however, phellamurin might stimulate one of the neurons that are responsive to components in Fr. 3.

Electrophysiological experiments have been carried out for some butterfly species in relation to oviposition behavior (e.g., Roessingh et al., 1991; Du et al., 1995; Städler et al., 1995). Investigations of 2 pierids, Pieris napi and P. rapae, revealed that one of the taste neurons responsive to glucosinolates (oviposition stimulants) is also sensitive to cardenolides (oviposition deterrents), and further have provided evidence for peripheral interactions of stimulant compounds with a neuron responding to deterrents (Du et al., 1995; Städler et al., 1995). Other instances of interference with chemoreceptor activity caused by inhibitory chemicals have been reported for other lepidopterans (e.g., Schoonhoven and Fu-Shun, 1989; van Loon, 1996; Qiu et al., 1998). None of the recordings we observed, however, found clear indications of such peripheral interaction of phellamurin with Fr. 3 sensitive cells (trace 5). Therefore, sensory events at the periphery, unlike an interference mechanism, seem likely to underlie the rejection of phellamurin-containing plants by ovipositing P. protenor females.

The genus Citrus is one of the plant taxa exploited worldwide by Rutaceae-feeding Papilio butterflies. Phytochemical data compiled to date show that flavonoid compounds abundant in Citrus species are flavones, flavanones, and flavonols, while compounds like phellamurin, which is a C-prenyl substituted 2,3-dihydroflavonols (flavanonols), are rarely found in Citrus (Tripoli et al., 2007; Ribeiro et al., 2008; Abad-García et al., 2009). Citrus plants are not sympatric with P. amurense at most locations in Japan; Citrus plants generally lack cold hardiness, and thus grow along coastal areas in temperate to warm regions. Phellodendron amurense is distributed in lowlands in cold latitudes but usually occurs at higher elevations in temperate regions. Therefore, it seems likely that not only P. protenor but other Citrus feeders, including P. xuthus (Honda and Hayashi, 1995b), have had little chance in their evolutionary history to encounter P. amurense. It is surprising that the larvae are able to ingest phellamurin given the female aversion (Honda and Hayashi, 1995a). The genetic grounds for such differential adaptation of adults and larvae to a given plant have been studied in other Papilio butterflies (e.g., Thompson et al., 1990).

Some flavonoids are known as oviposition stimulants for some papilionid butterflies (Ohsugi et al., 1985; Nishida et al., 1987; Honda, 1995), whereas flavonoids present in non-host plants often play a significant inhibitory role in regulation of oviposition (Nishida et al., 1990). We previously proposed that the ancestors of Papilionidae might have most heavily relied in host recognition on simple phytochemicals such as amino acid derivatives, hydroxycarboxylic acids, and/or polyols that are closer to primary metabolites (Nakayama et al., 2003). By contrast, flavonoids are relatively advanced secondary metabolites (Harborne and Baxter, 1993). In this context, we presume that flavonoids serve as modulators that govern host choice by ovipositing female, and that positive responses to flavonoids may be an apomorphic trait in the Papilionidae.

According to mitochondrial DNA phylogeny of the Papilionidae (Yagi et al., 1999; Zakharov et al., 2004), Japanese swallowtails in the tribe Papilionini comprise at least 2 subclusters; the subgenus Menelaides that includes typical Citrus feeders such as Papilio protenor and P. memnon, and the subgenus Achillides that includes Papilio maackii and P. bianor, both of which utilize Phellodendron amurense as a natural host. Papilio protenor and P. maackii may have diverged from a common ancestor, eventually exploiting distinct host plants and establishing their respective niches. It is noteworthy that other Menelaides members, Papilio polytes and Papilio helenus, also can complete their immature stages on P. amurense, and its aqueous fraction exerts strong to weak oviposition-stimulatory activity to P. xuthus and P. polytes (Honda and Hayashi, 1995a; Nakayama et al., 2002). Papilio maackii prefers to live in cool and montane zones, where rutaceous plants available for larval food are restricted to P. amurense only, particularly in northern Japan. In light of the present-day species richness of Papilionidae in tropical and temperate regions (Scriber, 1995), it is hypothesized that female chemosensory adaptation of the ancestors of P. maackii to an unusual (prenylated) flavonoid might have enabled them to colonize P. amurense and to extend their habitat further to boreal regions (as far as ca. 53°N latitude). In addition, P. bianor, another member of Achillides, has been observed in the field to lay eggs on Citrus plants on rare occasions and can marginally survive on them up to adulthood (Endo and Nihira, 1990). Positive oviposition responses to ancestral host plants also have been reported for species of the Papilio glaucus group (Mercader and Scriber, 2008). These and our finding that P. protenor females are very susceptible to phellamurin also seem to lend support to a possible host shift from Citrus plants to P. amurense by the common ancestor of Menelaides and Achillides.

The present finding that the stimulus evoked by a single phytochemical is interpreted differently by related butterfly species, suggests the involvement of chemoreceptor mutation(s). This may afford an opportunity for investigating the sensory and evolutionary mechanisms of chemical mediation of host range expansion and host shifts in the Papilionidae.


This work was supported in part by a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science to K. Honda (No. 14560039).

Copyright information

© Springer Science+Business Media, LLC 2011