A hemolymph major anionic peptide, HemaP, motivates feeding behavior in the sweetpotato hornworm, Agrius convolvuli


S. Nagata, Department of Applied Biological Chemistry, Graduate School of Agricultural and Life Sciences, University of Tokyo, 1-1-1 Yayoi, Bunkyo, Tokyo 113-8657, Japan
Fax: +81 3 5841 8022
Tel: +81 3 5841 5135
E-mail: anagashi@mail.ecc.u-tokyo.ac.jp


We recently identified a novel feeding-modulating peptide, hemolymph major anionic peptide (HemaP), designated Bommo-HemaP (B-HemaP), from hemolymph of the silkworm Bombyx mori. B-HemaP has a unique biological activity in modulating the regular frequency of feeding motivation, which is accompanied by increased foraging behaviors. To confirm the conservation of the HemaP-regulated feeding mechanism in lepidopteran species, we purified and sequenced two candidate peptides from the hemolymph of larvae of the sweet potato hornworm Agrius convolvuli. Unlike B. mori, A. convolvuli had two forms of HemaP, which were designated Agrco-HemaP-1 (A-HemaP-1) and Agrco-HemaP-2 (A-HemaP-2). The amino acid sequence of A-HemaP-2 was identical with that of A-HemaP-1, except for O-glycosylation on the fifth amino acid, threonine, within the N-terminal region. The amino acid sequence of A-HemaP-1/A-HemaP-2 had only 32% identity with B-HemaP. Structural analysis revealed that the carbohydrate moiety of A-HemaP-2 was an α-GalNAc residue. Injection of A-HemaP-1, A-HemaP-2 and recombinant A-HemaP-1 (rA-HemaP-1) individually caused a significant increase in foraging behaviors in A. convolvuli larvae, and no significant differences were observed among these three A-HemaPs. The CD spectra of these three A-HemaPs were quite similar, and all had α-helix-rich secondary structures. Although A-HemaP-1 and B-HemaP did not exhibit cross-reactivity at any injection doses examined, HemaP might be a conserved molecule among lepidopteran species that can modulate feeding motivation through the fluctuation of peptide levels in hemolymph.


2-aminobenzoic acid


hemolymph major anionic peptide from Agrius convolvuli




hemolymph major anionic peptide from Bombyx mori


Griffonia simplicifolia lectin I


glutathione S-transferase


hemolymph major anionic peptide


Lens culinaris agglutinin


Phaseolus vulgaris agglutinin


recombinant hemolymph major anionic peptide from Agrius convolvuli


Sophora japonica agglutinin


trifluoroacetic acid


Phytophagous insects, such as locusts [1,2] and caterpillars [3–5], show regular feeding behaviors. Similarly, we recently demonstrated that larvae of the silkworm Bombyx mori, a monophagous and phytophagous lepidopteran species, exhibit a regular feeding cycle with a frequency of approximately 2 h that is independent of circadian rhythms [5].

Various endogenous and environmental factors have been shown to function as key regulators in the initiation and termination of feeding behavior in insects [6,7]. In particular, the level of nutrients in hemolymph has been proposed to be an endogenous factor that affects feeding frequency. A theoretical biology study on nutritional complementation in the migratory locust Locusta migratoria simulated that feeding behaviors are induced by the depression of nutrient concentration in the hemolymph [8]. Therefore, feeding behaviors of insect larvae are thought to be regulated by hemolymph nutritional states, which eventually activate locomotion through an as yet unidentified neuronal regulatory system [9,10]. Physiological investigations have also indicated that the regulatory mechanisms of feeding behavior in insects are modulated by factors in hemolymph [11]. Therefore, we assumed that nutrient fluctuation in hemolymph can be detected by some factors that monitor the probability of feeding initiation [2]. In addition, long-term starvation increases foraging behaviors, especially wandering [5]. On the basis of these findings, we hypothesized that factors modulating feeding motivation contribute to the initiation and termination of feeding behaviors by controlling locomotor neurons. In several insects, food consumption is inhibited by endogenous peptidyl factors and neuropeptides [12–15]. However, little is known about the mechanisms of regulation of feeding motivation in insects at the molecular level.

Recently, we identified a novel peptidyl factor, hemolymph major anionic peptide (HemaP), from the hemolymph of B. mori larvae that is capable of increasing feeding motivation and is accompanied by increased locomotor activity [16]. Feeding-motivated B. mori larvae show specific behaviors, such as head-swaying, nibbling, walking, and ingesting, and eventually increase the time that they spend foraging [16].

B. mori HemaP, which is designated Bommo-HemaP (B-HemaP), is composed of 62 amino acids. Expression analyses have demonstrated that B-HemaP is produced by the fat body and secreted into the larval hemolymph. The mode of action of B-HemaP in vivo is distinct from that of other hormones that have regulatory roles. In fact, the levels of B-HemaP in the hemolymph are synchronized with the regularly occurring feeding period and fluctuate within a specific range. The secreted B-HemaP accumulates in the hemolymph during after a meal. After a certain threshold level of B-HemaP has been reached, larvae start to show foraging behaviors, and subsequently eat meals. The increased B-HemaP levels then return to the basal level.

A database search suggested the presence of three HemaPs from lepidopteran species, including Samia cynthia ricini, Spodoptera frugiperda, and Galleria mellonella, which share 19–37% identity with B-HemaP [16,17]. In the present study, we purified two forms of HemaP from the sweet potato hornworm, Agrius convolvuli [18], which belongs to the family Sphingidae and includes the tobacco hornworm Manduca sexta, in order to confirm the conservation of HemaP in other insects. We also examined the biological activities of A. convolvuli HemaPs that are related to feeding behaviors, in order to determine the conservation of the HemaP-modulating feeding system among HemaP-possessing species. Finally, we discuss the structural characteristics of HemaP and its related peptides.


Identification of two forms of HemaP from A. convolvuli hemolymph

Our previous study found that B-HemaP is a major peptide in the hemolymph of the silkworm, B. mori [16]. In order to identify HemaP in other lepidopteran species, we first analyzed the peptide fraction of larval hemolymph of the sweet potato hornworm, A. convolvuli, by MALDI-TOF MS (Fig. 1A). Two predominant ion peaks at m/z 6692 and m/z 6894 close to that of B-HemaP (m/z 6885) were observed, and we therefore purified these two peptides by three steps of reversed-phase HPLC (RP-HPLC) (Fig. 1B–D). As these two peptides were found to be A. convolvuli HemaPs, as described later, we designated the peptides with m/z 6692 and m/z 6894 as Agrico-HemaP-1 (A-HemaP-1) and Agrico-HemaP-2 (A-HemaP-2), respectively. A comparison of the peak areas of these two peptides in the RP-HPLC chromatogram showed that the concentration of A-HemaP-1 in A. convolvuli larval hemolymph was approximately three-fold higher than that of A-HemaP-2. The N-terminal sequence analyses of A-HemaP-1 and A-HemaP-2 identified 30 and 25 amino acids, respectively (Fig. S1A,B). These two peptides were separately digested with trypsin, V8 protease, or thermolysin, and the resulting peptide fragments were separated by RP-HPLC (Fig. S1C,D). Internal sequence analyses of the separated peptide fragments and the N-terminal sequences revealed that A-HemaP-1 and A-HemaP-2 had identical amino acid sequences, composed of 60 amino acids, except that the fifth amino acid from the N-terminus of A-HemaP-2 could not be identified by sequence analyses (Figs 1E and S1). The amino acid sequence of A-HemaP-1/A-HemaP-2 showed 32% identity with that of B-HemaP. The amino acid sequence alignment generated by clustalw revealed that the sequence identities among A-HemaP-1/A-HemaP-2 and other lepidopteran HemaPs were 27–41%. In contrast, the amino acid sequence of M. sexta HemaP showed high identity (79%) with that of A-HemaP-1/A-HemaP-2 (Fig. 1E).

Figure 1.

 Purification and identification of A-HemaP. (A) MALDI-TOF MS spectrum of the peptide fraction from A. convolvuli larval hemolymph. (B) The profile of the second purification step by RP-HPLC from A. convolvuli larval hemolymph. The arrows indicate peaks containing A-HemaP-1 and A-HemaP-2. The dashed line represents the acetonitrile concentration. (C) MALDI-TOF MS spectrum of the purified peptide with an ion peak at m/z 6692. Isolated A-HemaP-1 showed two ion peaks at m/z 3346 and m/z 6692, which correspond to double-charged and single-charged A-HemaP-1, respectively. (D) MALDI-TOF MS spectrum of the purified material with an ion peak at m/z 6894. An ion peak at m/z 3447 indicates a double-charged ion peak. (E) Alignment of the amino acid sequences of HemaPs from A. convolvuli, M. sexta (GR919872), B. mori (CK505940), S. cynthia ricini (DC871482), Sp. frugiperda (FP366628), and G. mellonella (P85216). Amino acids that are identical or similar to those of A-HemaP-1/A-HemaP-2 are represented by black and gray background, respectively. A-HemaP-1 and A-HemaP-2 have identical amino acid sequences, except for the fifth amino acid, which differs by a GalNAc modification (indicated with an asterisk).

Structure determination of the glycosylated fifth amino acid of A-HemaP-2

To determine the fifth amino acid of A-HemaP-2, we prepared tryptic peptides that included this amino acid from A-HemaP-2 and A-HemaP-1. These peptides, with protonated molecular masses of 1950 Da (p1950) and 1746 Da (p1746), respectively, were confirmed as the N-terminal 15-residue fragments by MALDI-TOF MS and N-terminal sequencing (data not shown). In the ESI ion-trap MS/MS analysis of p1950, when an ion peak at m/z 975.0, which is corresponding to a double-charged ion peak derived from a parent ion peak at m/z 1950, was selected as the precursor ion, the fragmentation spectra provided an amino acid sequence of p1950 from Asp1 to Gln15 (Fig. 2A) that was identical to that of p1746. In an MS/MS analysis of p1950, no signal from post-translational modification was observed. The data revealed an ion peak at m/z 1746.5 and its double-charged ion peak at m/z 873.7, suggesting that the post-translationally modified moiety was incidentally removed from p1950 (Fig. 2A). Amino acid sequencing by MS/MS analysis of the ion peak at m/z 873.7 derived from the MS/MS analysis of p1950 showed a sequence that was identical to the partial amino acid sequence of p1746 (Thr5 to Gln15) (data not shown). The finding that the difference in the molecular mass between p1950 and p1746 was 204 Da and that the fifth amino acid of A-HemaP-2 was not detected by Edman degradation indicated that the threonine was post-translationally modified, most likely with an O-linked N-acetylhexosamine.

Figure 2.

 Identification of the fifth amino acid and structure determination of the carbohydrate moiety of A-HemaP-2. (A) ESI ion-trap MS/MS spectrum from the peptide at [M + 2H]2+ of 975.0 (a parent ion peak at m/z 1950 was used as a precursor ion). An ion peak at m/z 873.7 corresponding to a double-charged ion of an ion peak at m/z 1746.5 (indicated by asterisks) was used as a precursor ion for MS/MS/MS analyses. The fragment ion peaks were assigned as y-type and b-type ion series. (B) Identification of the structure of the carbohydrate moiety by RP-HPLC. The 2-ABZ-labeled derivative of the carbohydrate moiety released from A-HemaP-2 was subjected to RP-HPLC (upper profile). Retention times of 2-ABZ derivatives were compared with those of 2-ABZ-labeled authentic carbohydrates (lower profile). The dashed line represents the acetonitrile concentration. (C) Identification of the linkage type of the carbohydrate moiety to the fifth threonine with lectin blotting. The peptides (500, 250, 100 and 10 ng of A-HemaP-2 and 200 ng of A-HemaP-1) were blotted onto a membrane. The specific recognition of lectins is as follows: GSL I, α-linkages of GalNAc and galactose residues; SJA, β-linkages of GalNAc and galactose residues; LCA, an α-linkage of a mannose residue; PHA-E, N-linked biantennary sugar chains.

We next tried to determine the chemical structure of the carbohydrate moiety attached to A-HemaP-2. The carbohydrate moiety released by acid hydrolysis was derivatized with 2-aminobenzoic acid (2-ABZ) and analyzed by RP-HPLC [19]. The retention time of the 2-ABZ-labeled carbohydrate moiety derived from A-HemaP-2 in the RP-HPLC analysis was consistent with that of 2-ABZ-labeled GalNAc (Fig. 2B), indicating that the attached carbohydrate moiety on A-HemaP-2 was GalNAc.

The linkage of GalNAc to the fifth threonine was determined by lectin blot analyses for specific linkages of carbohydrate chains with Griffonia simplicifolia lectin I (GSL I), Sophora japonica agglutinin (SJA), Lens culinaris agglutinin (LCA), and Phaseolus vulgaris agglutinin (PHA-E). Of the lectins examined, only GSL I, which is specific for α-linkages of GalNAc and galactose recognized A-HemaP-2. The other lectins did not recognize A-HemaP-2 (Fig. 2C), indicating that the GalNAc residue was linked to the fifth threonine via an α-linkage.

Cloning of A-HemaP-1/A-HemaP-2 cDNA

A cDNA encoding A-HemaP from the fat body of a fifth-instar larva was cloned by degenerate PCR and subsequent 5′-RACE and 3′-RACE. The results showed that A-HemaP-1/A-HemaP-2 cDNA included an ORF of 246 bp that encoded 82 amino acids, with a putative 22-residue signal peptide at its N-terminus (Fig. 3A). Tissue distribution analysis of A-HemaP-1/A-HemaP-2 by RT-PCR showed that A-hemap was expressed in the fat body, ovary, and testis (Fig. 3B), which is consistent with the expression profile of B-hemap [16].

Figure 3.

 cDNA cloning and tissue distribution of A-HemaP-1/A-HemaP-2 mRNA. (A) The nucleotide sequence of A-HemaP-1/A-HemaP-2 cDNA from A. convolvuli and its deduced amino acid sequences. The stop codon is indicated by an asterisk. A putative signal peptide sequence is underlined. A putative polyadenylation signal in the 3′-noncoding region is underlined with a dashed line. (B) Tissue distribution of A-hemap in A. convolvuli larvae as dtermined with RT-PCR. A. convolvuli ribosomal protein S18 was used as an experimental and transcriptional control for an endogenous transcription. FG, foregut; MG, midgut; HG, hindgut; MT, malpighian tubules; FB, fat body; CNS, central nervous system; Br, brain; Ov, ovary; Ts, testis.

Biological activities of A-HemaP-1 and A-HemaP-2 in larvae

To investigate whether A-HemaP-1 is able to increase the time spent foraging, as previously observed for B-HemaP [16], we injected seven doses of A-HemaP-1 (0.1–50.0 μg) into A. convolvuli larvae. Larvae injected with 0.5–10.0 μg of A-HemaP-1 had increased foraging behavior as compared with vehicle-injected control larvae (Fig. 4A,B). In contrast, larvae injected with higher doses of A-HemaP-1 (20.0 and 50.0 μg) did not exhibit an increase in foraging behavior (Fig. 4A). The most effective dose was 1.5 μg, and the dose-dependency was quite similar to that observed in B. mori [16]. These results suggested that larval feeding motivation reached a maximal level with an optimal dose of A-HemaP-1 in A. convolvuli larvae. We next assessed the activities of A-HemaP-1, A-HemaP-2 and recombinant A-HemaP (rA-HemaP-1) (Fig. S2) by injecting 1.5 μg of each peptide. Larvae injected with each of the three A-HemaPs increased the time spent foraging as compared with vehicle-injected larvae. No significant differences (P > 0.95) were observed among the three peptides (Fig. 4C). To examine whether the attachment of the carbohydrate moiety may affect biological activity, we compared the activity of A-HemaP-2 with that of A-HemaP-1 at seven doses (0.5–5.0 μg) (Fig. S3). At all doses, no significant difference (P > 0.13) was observed between A-HemaP-1 and A-HemaP-2, indicating that the carbohydrate moiety does not affect the biological activity. These results suggest that the regulatory mechanisms of feeding behavior controlled by HemaP are conserved among lepidopteran species.

Figure 4.

 Biological activity of A-HemaP-1 determined with A. convolvuli larvae. (A) Effect of A-HemaP-1 on the time spent in foraging of sample-injected A. convolvuli larvae (n = 4–5) as compared with vehicle-injected larvae (n = 5). Injection doses of A-HemaP-1 were 0.1, 0.5, 1.5, 5.0, 10.0, 20.0 and 50.0 μg per larva. Foraging behaviors were observed for 120 min. **P < 0.00005 and *P < 0.05; one-way ANOVA followed by a post hoc Tukey test. Values represent the mean + standard deviation. (B) Representative behavioral patterns of A-HemaP-1-injected (upper panel) and vehicle-injected (lower panel) A. convolvuli larvae. The arrows indicate the timing of sample injection. The black boxes indicate ingesting behavior, and white boxes indicate foraging behaviors, including head-swaying, nibbling, and walking. (C) Relative activities of A-HemaP-1 (n = 5), A-HemaP-2 (n = 5), and rA-HemaP-1 (n = 4), based on the time spent in foraging as compared with vehicle-injected larvae (n = 5). *P < 0.001; one-way ANOVA followed by a post hoc Dunnett test. Values represent mean + standard deviation.

CD spectral analyses of A-HemaP-1 and A-HemaP-2

CD spectral analyses of B-HemaP have revealed that it has a typical α-helix-rich secondary structure [16]. As A-HemaP-1/A-HemaP-2 and other HemaPs, including B-HemaP, have a seven-residue repeating pattern that contains hydrophobic amino acids at the first and fourth positions (Fig. S4), these peptides may assume similar α-helical secondary structures. Therefore, we obtained CD spectra to investigate the secondary structure of three A-HemaPs. The CD spectra of the isolated A-HemaP-1, A-HemaP-2, rA-HemaP-1 and rB-HemaP showed similar patterns, with two minimum values at 208 and 222 nm in NaCl/Pi, which are typical features of peptides containing an α-helix-rich structure (Fig. 5). In addition, the similar pattern of the CD spectra of A-HemaP-1 and A-HemaP-2 indicated that the carbohydrate moiety of A-HemaP-2 does not affect the secondary structure formation of peptide backbones (Fig. 5).

Figure 5.

 Secondary structure analyses of A-HemaP and B-HemaP. CD spectra of A-HemaP-1 (solid line), A-HemaP-2 (dashed line), rA-HemaP-1 (solid gray line) and B-HemaP (dashed gray line) were measured in NaCl/Pi at room temperature.

Cross-reactivity between A-HemaP-1 and B-HemaP

As native A-HemaP-1, native A-HemaP-2 and rA-HemaP-1 exhibited comparable biological activity (Fig. 4) and formed an α-helix-rich secondary structure (Fig. 5), similarly to B-HemaP, we next analyzed cross-reactivity between A-HemaP-1 and B-HemaP. We injected B-HemaP into A. convolvuli larvae and A-HemaP-1 into B. mori larvae, and then observed the foraging behaviors. We found that larvae injected with HemaP from the different species did not exhibit increased foraging behavior at any dose (Fig. 6), indicating that there is no cross-reactivity between B-HemaP and A-HemaP-1.

Figure 6.

 Cross-reactivity between HemaPs from A. convolvuli and B. mori. (A) Biological activities of A-HemaP-1 and B-HemaP injected into B. mori larvae were determined by comparison with vehicle-injected larvae (n = 5). The injection doses of A-HemaP-1 were 0.1, 0.5, 1.5, 5.0, 10.0, 20.0 and 50.0 μg per larva, and that of B-HemaP was 1.5 μg per larva as a positive control. (B) Biological activities of B-HemaP and A-HemaP-1 injected into A. convolvuli larvae were determined by comparison with vehicle-injected larvae (n = 5–6). The injection doses of B-HemaP were 0.1, 0.5, 1.5, 5.0, 10.0, 20.0 and 50.0 μg per larva, and that of A-HemaP-1 was 1.5 μg per larva as a positive control. Foraging behaviors were observed for 120 min. ***P < 0.0005; one-way ANOVA followed by a post hoc Tukey test. Values represent mean + standard deviation.


In the present study, we identified two HemaPs from Aconvolvuli larval hemolymph. A-HemaP-1 and A-HemaP-2 represent the second instance where feeding modulating peptides have been isolated and characterized from lepidopteran species. In contrast to that of B. mori, Aconvolvuli larval hemolymph had two HemaPs of different sizes, A-HemaP-1 and A-HemaP-2, the latter of which carried an α-GalNAc moiety at the fifth threonine. Despite the presence of the carbohydrate moiety in A-HemaP-2, this peptide exhibited a similar α-helix-rich secondary structure as A-HemaP-1. In addition, A-HemaP-1 and A-HemaP-2 caused almost comparable foraging behavior activities in Aconvolvuli larvae.

The amino acid sequence features of HemaPs

Although six sequences of HemaPs show relatively low identity (19–37%) with one another (Fig. 1E), including B-HemaP and other HemaPs, all of these peptides share the potential to form α-helix-rich secondary structures that are characterized by similar repetitive-patterned hydrophobic amino acids, leading to an amphiphilic structure (Fig. S4). Indeed, predictprotein (http://www.predictprotein.org), a secondary structure prediction algorithm, predicted that five HemaPs would contain two α-helical regions within their sequences. Moreover, the second α-helical region within the N-terminus of B-HemaP was predicted to be divided into two regions. In addition, the two α-helical regions are located at the same positions in the HemaPs, strongly suggesting that these peptides form a similar conformation. An alignment of the six HemaPs also revealed that these peptides exhibit three conserved regions: the N-terminus, an intermediate (from the 30th to the 40th amino acids) region, and the C-terminus. Although A-HemaP-1/A-HemaP-2 and B-HemaP form similar α-helix-rich secondary structures (Fig. 5), no cross-reactivity was observed between these two peptides (Fig. 6), indicating that the three conserved domains in the HemaPs may not be directly related to biological activity. An exceptionally high sequence identity was found between A-HemaP-1/A-HemaP-2 and M. sexta HemaP (79%) as compared with those between A-HemaP-1/A-HemaP-2 and other HemaPs (27–41%) (Fig. 1E), because they belong to the same family, the Sphingidae. For this reason, A-HemaP-1/A-HemaP-2 and M. sexta HemaP may exhibit cross-reactivity.

An α-GalNAc residue in A-HemaP-2

Although two ion peaks derived from A-HemaP-1 and A-HemaP-2 were detected (Fig. 1A), the carbohydrate moiety was not cleaved by MALDI-TOF MS analysis of the isolated A-HemaP-2 (Fig. 1D), indicating that the fifth threonine of A-HemaP-1 was partially O-glycosylated. Prediction with predictprotein indicated that an α-helical region located near the N-terminus of A-HemaP-1/A-HemaP-2 started from the 11th amino acid, which is close to the fifth threonine carrying an α-GalNAc, indicating that the carbohydrate moiety might affect the biological activity and α-helical structures of A-HemaP-2. However, Figs 4 and 5 show that the carbohydrate moiety of A-HemaP-2 did not influence its biological activity or secondary structure. Therefore, both A-HemaP-1 and A-HemaP-2 are biologically active compounds that can exert a regulatory role in feeding motivation, although the reason for partial glycosylation on A-HemaP remains to be elucidated.

Feeding behavior of lepidopteran larvae regulated by the fluctuation of HemaP levels in hemolymph

When Aconvolvuli larvae were injected with various doses of A-HemaP-1/A-HemaP-2, we found that the larvae injected with 0.5–5.0 μg of both peptides increased the time spent foraging (Figs 4A and S3). In contrast, the larvae injected with 20.0 and 50.0 μg of A-HemaP-1/A-HemaP-2 showed a decrease in feeding motivation. The decreased biological activity induced by higher doses of A-HemaP-1/A-HemaP-2 was also observed in B-HemaP-injected B. mori larvae [16]. These results suggest that HemaP functions by fluctuating within a defined range.

In order to satisfy their nutritional requirements, insects have to select a complementary diet according to imbalances in nutrients. A nutritionally diluted diet leads to a decrease in the nonfeeding period and an increase in food intake in L. migratoria [20] and Bmori larvae (unpublished observation). These results suggest that feeding behavior is more strongly influenced by nutritional requirements than volumetric requirements when sufficient amounts of food are present. Because the B-HemaP level in hemolymph fluctuates according to regularly occurring feeding behaviors and different states of feeding [16], we hypothesize that the elevation of HemaP levels in hemolymph is related to decreases in the levels of other nutrients in the hemolymph. To date, it is still unknown how the feeding cycle is driven in vivo and what factors are involved. HemaP may be a candidate factor that controls the feeding cycle, although HemaP has only been identified in lepidopteran insects. Although the regulatory system of feeding motivation by HemaP seems to be specific for lepidopteran species, similar regulatory mechanisms using relatively abundant peptides or proteins might be present in other groups of insects.

Structural similarities between HemaPs and other proteins

B-HemaP and A-HemaP-1/A-HemaP-2 are amphiphilic α-helix-rich peptides. An amphiphilic α-helix-rich structure may interact with small molecules, such as lipids, through hydrophobic interactions, as observed for apolipophorin. In insect hemolymph, there are other major proteins, such as apolipophorin-III (ApoLp-III), which form an amphiphilic α-helix-rich structure [21]. ApoLp-IIIs are component of insect lipoproteins that have been well characterized as molecules associated with lipophorin particles containing diacylglycerol, phospholipid, and hydrocarbon, and participate in lipid transfer in hemolymph [22–27]. HemaP and ApoLp-IIIs share three predominant characteristics: (a) they have a linear peptide without any cysteine residue; (b) they are enriched in α-helix structure; and (c) they are major components in the hemolymph that are secreted by the fat body. Furthermore, ApoLp-IIIs are widely conserved within insect species, with low sequence identity (< 10%) [23,28–31], and HemaPs also show low identity with each other. In addition, some insect ApoLp-IIIs are glycosylated, but the α-helical secondary structures and biological activities are not influenced by this modification [30,32,33]. Although the functional similarity between HemaPs and ApoLp-IIIs has not been elucidated, structural similarities between these peptides may help in the elucidation of the molecular mechanism of HemaP-induced feeding behavior.

The present study has demonstrated that no cross-reactivity occurrs between B-HemaP and A-HemaP-1 (Fig. 6). Therefore, a specific amino acid region of HemaP contributes to species-specific biological activity. In order to address the molecular mechanisms of regulation of feeding behavior by HemaP, we also have to assess the biochemical characterization of HemaP for species specificity. We predicted that HemaPs possess a specific region that is essential for biological activity and contributes to species-specific biological activity. Therefore, it is necessary to determine the amino acid sequence that is essential for biological activity. We also demonstrated the presence of HemaPs in lepidopteran species with biological activity related to feeding behavior. Although the new paradigm of a regulatory role for HemaP has not yet been fully elucidated [16], HemaP may be one of the biological oscillators that synchronizes with feeding rhythmicity. In our previous study, several neuropeptides were characterized as factors that influence feeding behaviors in B. mori larvae [34]. Therefore, we will need to consider the various factors that coordinately control regularly occurring feeding behavior in lepidopteran insect larvae as a total body event. The regulation of HemaP production in the fat body and metabolic regulation by HemaP are important issues for estimating the total nutritional body balance in insects.

Experimental procedures

Reagents and chemicals

The chemicals and reagents used in the present study were purchased from Nacalai-tesk (Osaka, Japan), and HPLC reagents [acetonitrile and trifluoroacetic acid (TFA)] were purchased from Kanto Kagaku (Tokyo, Japan).

Experimental animals

A. convolvuli larvae were reared on an artificial diet at 27 ± 1 °C with 70% ± 10% humidity under long-day conditions (16 h of light and 8 h of dark), as described previously [35]. B. mori larvae were reared on Silk Mate 2S (Nippon Nosan kogyo, Yokohama, Japan) under the same conditions as used for A. convolvuli.

Purification and identification of A-HemaP

Approximately 20 mL of hemolymph was collected from 25 last-instar A. convolvuli larvae, acidified with TFA at a final concentration of 0.1%, and centrifuged at 20 400 g at 4 °C for 15 min. The supernatant was applied to a Sep-Pak C18 cartridge (Waters, Milford, MA, USA), which was washed with water/0.1% TFA, and then eluted with 60% acetonitrile/0.1% TFA. The eluate was subjected to three further purification steps by RP-HPLC, with the same PEGASIL300-ODS column (4.6 × 250 mm; Senshukagaku, Tokyo, Japan), and with monitoring by absorbance at 225 nm. In the first step, the column was eluted with a linear gradient of 10–60% acetonitrile in 0.1% TFA over 25 min at a flow rate of 1 mL·min−1. In the second step, the elution was performed with a 5-min linear gradient of 10–36% acetonitrile, a 25-min linear gradient of 36–48% acetonitrile and a 5-min linear gradient of 48–60% acetonitrile in 0.1% TFA at a flow rate of 1 mL·min−1. In the third step, the elution was performed with a 5-min linear gradient of 10–36% acetonitrile, a 30-min linear gradient of 36–48% acetonitrile and a 5-min linear gradient of 48–60% acetonitrile in 0.1% TFA at a flow rate of 1 mL·min−1. The amount of peptide was weighed after lyophilization, or was estimated by UV absorbance at 225 nm by comparison with BSA as a standard. Purified A-HemaP-1 and A-HemaP-2 were separately digested with trypsin (Roche, Basel, Switzerland), V8 protease (Roche) and thermolysin (Sigma, St. Louis, MO, USA) at 37 °C for 1 h, and the reaction mixtures were then purified by RP-HPLC with a PEGASIL300-ODS column (4.6 × 250 mm; Senshukagaku), which was eluted with a linear gradient of 0–60% acetonitrile in 0.1% TFA over 40 min at a flow rate of 1 mL·min−1. The isolated A-HemaP-1/A-HemaP-2 and peptide fragments were sequenced with a Procise cLC protein sequencer (Applied Biosystems, Foster City, CA, USA).

Analyses by MALDI-TOF MS

Mass spectra were measured on a Voyager DE STR mass spectrometer (Applied Biosystems) in the positive ion mode. Samples (1 μL) were prepared by mixing with the supernatant (1 μL) of α-cyano-4-hydroxycinnamic acid saturated in 60% acetonitrile as the matrix. Angiotensin I, angiotensinogen, insulin and myoglobin were used as external calibrants.

Analyses by ESI ion-trap MS/MS

ESI MS/MS analyses were performed on an ion-trap mass spectrometer (Esquire 3000 plus; Bruker Daltonik, Bremen, Germany) in the positive ion mode, and data were acquired by scanning an m/z range from 200 to 1800. The lyophilized sample was dissolved in 0.1% formic acid (final concentration: 1 pmol·μL−1) and then subjected to MS by direct infusion. To identify the amino acid sequence of a peptide corresponding to a peak at m/z 1950, MS/MS analysis of a double-charged precursor ion at m/z 975.0 was selected to provide daughter ion peaks. The flow rate was 3 μL·min−1. The pressure of the nebulizer gas (N2) was set at 10 psi, and the dry gas was applied at a flow rate of 3.5 L·min−1 at 250 °C. MS/MS data were analyzed with biotools (Bruker Daltonics, Billerica, MA, USA).

2-ABZ labeling of the carbohydrate moiety

The carbohydrate moiety was released from A-HemaP-2 (110 ng) by hydrolysis in 15% TFA, at 100 °C for 4 h, according to a method reported previously [36]. The released carbohydrate moiety was N-acetylated with the following procedure. The reaction mixture was applied to a Sep-Pak C18 cartridge (Waters), and then eluted with 0.1% TFA. The eluate was lyophilized and then dissolved in ice-cold saturated sodium bicarbonate (42.5 μL), to which acetic anhydride (1.1 μL) was added immediately. The solution was agitated gently at room temperature for 10 min. Acetic anhydride (1.1 μL) was then added again to the reaction mixture, which was agitated for 20 min at room temperature [37]. The reaction mixture was then evaporated to dryness under reduced pressure, and lyophilized. For labeling of the carbohydrate moiety with 2-ABZ, the lyophilized material was dissolved in 0.1% TFA (50 μL), to which 2-ABZ solution (1 μL) [30 mg of 2-ABZ, 20 mg of NaBN3CN, and 1 mL of sodium acetate/sodium borate (4% CH3COONa, 2% H3BO3/methanol)] was added, and incubated at 80 °C for 45 min [19]. Authentic 2-ABZ derivatives of GlcNAc and GalNAc (Wako, Osaka, Japan) were prepared as described above. The reaction mixture was subjected to TLC on a silica gel plate (developed with 1-butanol/ethanol/H2O, 4 : 1 : 1), and 2-ABZ derivatives were extracted from the TLC plate with 90% acetonitrile. They were analyzed on an ESI-TOF mass spectrometer (JMS-T100LC; JEOL, Tokyo, Japan) by direct infusion. The resulting 2-ABZ-labeled carbohydrate from A-HemaP-2 was compared with 2-ABZ-labeled GalNAc and GlcNAc by retention time in RP-HPLC with a PEGASIL300-ODS column (4.6 × 250 mm; Senshukagaku). The elution was performed with a linear gradient of 4.5–18% acetonitrile per 50 mm acetic acid over 50 min at a flow rate of 1 mL·min−1, modified from a previous report [19]. 2-ABZ derivatives were detected by fluorescence with 360 nm for excitation and 425 nm for emission.

Lectin blot analyses

Purified A-HemaP-2 (500, 250, 100 and 10 ng) and A-HemaP-1 (200 ng) were dotted onto a nitrocellulose membrane (Hybond-+; GE-Healthcare, Little Chalfont, UK). After baking at 80 °C for 1 h, the membrane was blocked by submerging it in NaCl/Tris (137 mm sodium chloride, 2.7 mm potassium chloride, 25 mm Tris, pH 7.4) containing 0.05% Tween-20 and 5% BSA for 30 min at room temperature. After being washing three times with NaCl/Tris, the membrane was incubated with each lectin solution – GSL I for α-linkages of GalNAc and galactose residues, SJA for β-linkages of GalNAc and galactose residues, LCA for an α-linkage of mannose residues, and PHA-E for N-linked biantennary sugar chains (10 μg·mL−1) – in NaCl/Tris (Biotinylated Lectin Kit II; Vector Laboratories, Burlingame, CA, USA) for 1 h at room temperature. After being washed again, the membrane was incubated with horseradish peroxidase-conjugated avidin (1 : 3000 dilution) for 1 h at room temperature. After being washed, the membrane was incubated with ECL reagent (SuperSignal West Pico Luminol; Thermo, Waltham, MA, USA) for detection of signals with a lumino image analyzer (LAS-3000; Fujifilm, Tokyo, Japan).

Cloning of A-HemaP-1/A-HemaP-2 cDNAs

Total RNA was extracted with TRIzol Reagent, according to the manufacturer’s instructions (Invitrogen, Carlsbad, CA, USA), from the fat bodies of A. convolvuli larvae. Total RNA was treated with DNase I (TaKaRaBio, Shiga, Japan) to remove contaminated genomic DNA. Purified total RNA was reverse transcribed with SuperScript III reverse transcriptase (Invitrogen), with the primer 5′-GCCCAGTGAATTGTAATACGACTCACTATAGGGAGGCGGT(24)-3′. PCR amplification was performed with Ex Taq DNA polymerase (TaKaRaBio). The primers were designed on the basis of the identified amino acid sequence of A-HemaP, as follows: forward primer, 5′-GTNATHAAYTTYGAYGTNAT-3′; and reverse primer, 5′-TTNGGNGCYTCYTTRTTRTC-3′. The PCR product was subcloned into a pGEM-T easy vector (Promega, Madison, WI, USA), and this was followed by sequencing with Big Dye Terminator v3.1 and a DNA sequencer, ABI PRISM 310 Genetic Analyzer (Applied Biosystems). A full-length cDNA encoding A-HemaP was obtained by 5′-RACE and 3′-RACE. 3′-RACE was performed with a primer set of 5′-AGCCAAGCCGAAAAAGCCTTCGAC-3′ and SY-IIIA 5′-AAGGAGTGGTATCCAGTGTGCTGG-3′. cDNA for 5′-RACE was synthesized from purified total RNA with primers SY-CDS 5′-AAGGAGTGGTATCCAGTGTGCTGG T(31)-3′ and SY-CDIII 5′-AAGGAGTGGTATCCAGTGTG CTGGG-3′. PCR amplification was performed with forward primer SY-IIIA and reverse primer 5′-GTTGACGAATTCTTCGAAGTTC-3′. Nested PCR was performed with forward primer SY-IIIA and reverse primer 5′-GAAGTTCTTCTTAATCGTGTCTGC-3′. PCR products were subcloned into a pGEM-T easy vector (Promega), and then sequenced.

Tissue distribution of A-HemaP-1/A-HemaP-2 mRNA

Total RNA from foregut, midgut, hindgut, Malpighian tubules, fat body, central nervous system, brain, ovary and testis was extracted with TRIzol reagent (Invitrogen). RT-PCR was performed with specific primers based on the identified cDNA encoding A-HemaP-1/A-HemaP-2, with cDNAs prepared as described above: forward primer, 5′-GATGCTCCCGCAACTGAGAGC-3′; and reverse primer, 5′-TCACTTGGGGGCTTCCTTGTTG-3′. A partial fragment of cDNA encoding ribosomal protein S18 was amplified with forward primer 5′-GGGGGATAATTGCAAACCCCAAT-3′ and reverse primer 5′-AGCGATTTGTCTGGTTAATTCCGG-3′ as an experimental control.

Expression of recombinant A-HemaP-1

For construction of glutathione S-transferase (GST)-tagged rA-HemaP-1, cDNA was used to amplify the coding region of A-HemaP-1 by PCR, with the following primers: forward primer, 5′-CGCGGATCCGATGCTCCCGCAACTGAGAGC-3′, including a BamHI site (underlined) followed by a codon for methionine (bold); and reverse primer, 5′-GGCCTCGAGTCACTTGGGGGCTTCCTTGTTGTC-3′, with a XhoI site (underlined). The PCR product was ligated into the XhoI and BamHI sites of pGEX-6p-1. Then, BL21-CodonPlus (DE3), an Escherichia coli strain, was transformed with this plasmid. Recombinant A-HemaP-1 fused N-terminally with GST and a following methionine residue (GST-M-rA-HemaP-1) was collected from a soluble fraction. The recombinant protein was purified by affinity column chromatography with Glutathione Sepharose 4 Fast Flow (GE Healthcare). Affinity-purified GST-M-rA-HemaP-1 was digested with PreScission Protease (GE Healthcare) at 4 °C for 5 h. The tag-removed recombinant protein was purified with a Sep-Pak C18 cartridge (Waters). The eluate with 60% acetonitrile in 0.1% TFA was lyophilized for further preparation. Several N-terminal extra amino acid and methionines of the resulting peptide were cleaved off by incubation with 70% formic acid containing 1% BrCN at room temperature for 4.5 h. Finally, rA-HemaP-1 was purified by RP-HPLC as described for the purification of A-HemaP. The recombinant peptide was confirmed by measurement of molecular mass before CD spectral analysis and assessment of the biological activity.

CD analyses

CD spectra were measured on a Jasco-720 spectropolarimeter (JASCO Corporation, Tokyo, Japan), with a 0.1-cm-pathlength cell over the range of 200–260 nm at room temperature. The spectra were acquired every 0.5 nm with a 4-s average time per point and a 1.0-nm bandpass. The concentrations of sample peptides were 0.6 mg·mL−1. The mean residue ellipticity was calculated with the software supplied with the spectropolarimeter.

Bioassay for foraging behavior (HemaP assay)

The fifth-instar day 2 A. convolvuli larvae were reared on an artificial diet for the HemaP assay as described previously [16]. In brief, after starvation for 2.5 h, A. convolvuli larvae were fed for 30 min. The larvae were then anesthetized by submerging them in ice-cold water for 10 min. Samples dissolved in distilled water (100 μL) were injected into larvae through the fifth segment with a 27G syringe. Each of the larvae was placed in front of an artificial diet block in a dish with a diameter of 15 cm. The behavior of each larva was observed for 120 min after sample injection. All locomotor-activated foraging behaviors, including head-swaying, nibbling, walking, and ingesting, were recorded. The time spent foraging during the observation period was compared with that of vehicle-injected control larvae. Fifth-instar day 2 B. mori larvae reared on an artificial diet were used for the HemaP assay as described previously [16].


This work was supported partly by Grants-in-Aid for Scientific Research (#18780083 and 22780099) from the Ministry of Education, Science, Sports, and Culture of Japan. N. Morooka was supported by a Research Fellowship of the Japan Society for the Promotion of Science (JSPS) for Young Scientists.