Isolation of biologically active peptides from the venom of Japanese carpenter bee, Xylocopa appendiculata

Background Mass spectrometry-guided venom peptide profiling is a powerful tool to explore novel substances from venomous animals in a highly sensitive manner. In this study, this peptide profiling approach is successfully applied to explore the venom peptides of a Japanese solitary carpenter bee, Xylocopa appendiculata (Hymenoptera: Apoidea: Apidae: Anthophila: Xylocopinae: Xylocopini). Although interesting biological effects of the crude venom of carpenter bees have been reported, the structure and biological function of the venom peptides have not been elucidated yet. Methods The venom peptide profiling of the crude venom of X. appendiculata was performed by matrix-assisted laser desorption/ionization-time of flight mass spectroscopy. The venom was purified by a reverse-phase HPLC. The purified peptides were subjected to the Edman degradation, MS/MS analysis, and/or molecular cloning methods for peptide sequencing. Biological and functional characterization was performed by circular dichroism analysis, liposome leakage assay, and antimicrobial, histamine releasing and hemolytic activity tests. Results Three novel peptides with m/z 16508, 1939.3, and 1900.3 were isolated from the venom of X. appendiculata. The peptide with m/z 16508 was characterized as a secretory phospholipase A2 (PLA2) homolog in which the characteristic cysteine residues as well as the active site residues found in bee PLA2s are highly conserved. Two novel peptides with m/z 1939.3 and m/z 1900.3 were named as Xac-1 and Xac-2, respectively. These peptides are found to be amphiphilic and displayed antimicrobial and hemolytic activities. The potency was almost the same as that of mastoparan isolated from the wasp venom. Conclusion We found three novel biologically active peptides in the venom of X. appendiculata and analyzed their molecular functions, and compared their sequential homology to discuss their molecular diversity. Highly sensitive mass analysis plays an important role in this study. Electronic supplementary material The online version of this article (doi:10.1186/s40409-017-0119-6) contains supplementary material, which is available to authorized users.

Mass spectrometry-guided venom peptide profiling has become an indispensable tool for rapid, accurate, and highly sensitive screening of the novel venom substances [9][10][11][12]. It contribute to accelerate the elucidation of the venom substances in the molecular structure level. Recently, we have successfully applied mass spectrometryguided venom peptide profiling to explore novel venom substances of social and solitary wasps [13,14]. In conjunction with our continuous research program on the isolation and biological study of the Hymenoptera venom substances, we were interested in the venom of the Japanese carpenter solitary bee Xylocopa appendiculata (Hymenoptera: Apoidea: Apidae: Anthophila: Xylocopinae: Xylocopini). We considered that the target venom is a challenging sample because of the following reasons: the crude venom of carpenter bees showed significant biological effects such as lethal activities in a small bird and mice [15]; the sting of Xylocopa virginica and Xylocopa vioracea seems to be as painful in humans as are honeybee stings [15]; although the significant biological effects of the crude venom has been suggested, the biologically active peptides of the carpenter bee venom including that of X. appendiculata has not been isolated yet; it is difficult to collect X. appendiculata because of their solitary lives; and only a small amount of the venom substances is available due to the fact that the venom sac of X. appendiculata is smaller than those of honeybees and vespid wasps (Fig. 1).

Sample preparation
Fifteen female bees of X. appendiculata were collected in Osaka and Sakai, Japan. The venom sacs were dissected and homogenized with water (50 μL). The venom extract was applied to MALDI-TOF MS and HPLC analyses.

MALDI-TOF MS and MS/MS analysis
Matrix-assisted laser desorption/ionization-time of flight mass spectroscopy (MALDI-TOF MS) and tandem mass spectrometry (MS/MS) analysis was performed by Ultra-Flex speed (Bruker Daltonics, Germany). MS and MS/MS analyses (Additional files 1 and 2) were performed in the linear positive ion mode and reflector positive mode, respectively. α-Cyano-4-hydroxycinnamic acid (CHCA), trifluoroacetic acid (TFA) and all other reagents were purchased from Wako Pure Chemical Industries, Ltd. (Osaka, Japan) or Nacalai Tesque (Kyoto, Japan). The matrix solution was prepared as follows. CHCA was dissolved in 3:7 acetonitrile/H 2 O (0.1% TFA) to obtain saturated matrix solution. The crude venom or peptide sample were mixed with the matrix solution on a plate, dried for 5 min at ambient temperature, and provided for the mass analysis. Molecular mass and peptide sequencing were analyzed using the FlexAnalysis 3.4 software and BioTools 3.2 (Bruker Daltonics, Germany). The monoisotopic molecular mass was estimated in a range of m/z 1000~5000 for short peptides or m/z 5000~20,000 for PLA 2 .

Peptide sequence analysis and synthesis
The purified venom peptides were sequenced by automated Edman degradation using ABI model 477A (Applied Biosystems, USA). Peptides were synthesized by Fmoc chemistry using a Shimadzu PSSM-8 automated peptide synthesizer (Shimadzu, Japan), and purified by reverse-phase HPLC. The identity and purity of the peptides were confirmed by MALDI-TOF MS. The synthetic Xac-1 and Xac-2 were employed for circular dichroism (CD) analysis, liposome leakage assay, antimicrobial and hemolytic activity tests.

Liposome leakage experiments
Liposomes were prepared from lecithin from egg yolk (phosphatidylcholine approx. 70%; Nacalai Tesque). The lecithin (28 mg) was dissolved in chloroform (5 mL). The solution was concentrated in vacuo and maintained under the reduced pressure for 10 h to remove the solvent. The dried lecithin was hydrated in 4 mL of 70 mM calcein (Sigma-Aldrich) in aqueous NaOH (pH 7.5). After sonication for 10 min, the vesicles were passed through a column of Sephadex TM G-50 (GE Healthcare) in H 2 O to remove free calcein. The first 5 mL of eluent was collected as calcein-encapsulated vesicles. Water (0.8 mL) was added to the liposome suspension (0.2 mL) in a cuvette. After 10 min, 0.5-20 μL of 10 mM solution of mastoparan (Peptide Institute, Inc., Japan) or Xac-1 was added to the cuvette. Fluorescence intensity of calcein was measured by Hitachi P-4500 fluorometer (excitation wavelength of 460 nm and emission wavelength of 530 nm). A 1% (v/v) solution of Triton X-100 was used as a positive control to obtain maximum fluorescent value at 100% leakage of calcein.
PCR products were cloned into plasmids using pGEM-T Easy Vector System (Promega, USA). Plasmids were purified with Wizard Plus SV Minipreps DNA Purification System (Promega) and sequenced on an ABI PRISM 310 Genetic Analyzer (Life Technologies) or 3130 Genetic Analyzer (Life Technologies) with BigDye Terminator v3.1 Cycle Sequence kit (Life Technologies). To obtain complete sequences of PLA 2 cDNA, 3′-and 5′-RACEs (rapid amplification of cDNA ends) were performed using a SMART RACE cDNA Amplification kit (Clontech, USA) according to the supplier's instructions. F3 (5′-CGG CGC CGT AAG GTT CAC GTA CTT C) and R1 (5′-GCT GAA GGA GAC CGA CGC CTG TTG T-3′) primers were used for 3′-and 5′ RACEs, respectively. The obtained PCR products were also cloned into a vector, and sequenced as described above.

Antimicrobial activity
According to the procedure [16,17], antimicrobial activities of Xac-1 and Xac-2 were evaluated using Escherichia coli (NBRC14237) and Micrococcus luteus (NBRC 12708) as a gram-negative bacterium, Stapylococcus aureus (NBRC12708) as a gram-positive bacterium, and the yeast Saccharomyces cerevisiae (NBCR 10217). To compare the potency, MIC values of mastoparan were evaluated. Bacteria were grown in 2 mL Trypticase soy broth, and yeasts in Sabouraud dextrose broth for 16 h with shaking at 200 rpm as a pre-culture. Subsequently, 0.1 mL pre-culture medium was inoculated into 2 mL of fresh medium. It was cultivated for 2-3 h until A 600 = 0.5. The cultivated medium was diluted with PBS solution. The diluted microbial broth (100 μL) was mixed with peptide solutions (11 μL) in 96-well plates and incubated for 3 h. After 3 h incubation, two times concentrated medium were added and 96-well plates were reincubated for 16 h. Microbial growth was measured by Spectra MAX 190 microplate reader at A 600 .

Hemolytic activity
According to the procedure described by Shigeri et al. [12], hemolytic activities of Xac-1 and Xac-2 were tested. Heparinized rat whole blood from Wistar rats (male, 6 weeks old) was washed twice in NaCl/Pi (100 mM NaCl, 7.5 mM Na 2 HPO 4 and 2.5 mM NaH 2 PO 4 ) by centrifugation at 900 g and suspended in NaCl/Pi to a concentration of 0.5% (v/v). NaCl/Pi and NaCl/Pi containing 0.2% Triton X-100 were used as controls for 0 and 100% hemolysis, respectively. Xac-1 and Xac-2, as well as mastoparan and melittin were employed as comparable standards.

Histamine releasing activity
The histamine-releasing activities of Xac-1 and Xac-2, mastoparan, and melittin were determined with rat peritoneal mast cells, as previously described [17]. The histamine-releasing activity was defined as the ration of the extracellular to the total amount of histamine. Spontaneous histamine-releasing activity was 6.9 ± 0.3%.

MALDI-TOF MS and HPLC analysis of the crude venom extract of X. appendiculata
The crude venom extract of X. appendiculata was subjected to MALDI-TOF MS analysis ( Fig. 2a and  b). The MSPP analysis in a range of m/z 1000~5000 (Fig. 2a) indicated that peptides in a range of m/z 1850~2200 are the major in the venom of X. appendiculata. Characteristic ion signal at m/z 16508 was observed in a range of m/z 5000~20000 (Fig. 2b). Having these profiles, the crude venom was subjected to HPLC purification with a C18-reversed phase column to provide eight major fractions (A to H) (Fig. 3). To prove this possibility, GFVALLKKLPLILKHLH-NH 2 was prepared and subjected to HPLC analysis to compare retention time. Retention times of the synthetic and naturally occurring peptide were identical. As a result, the peptide of fraction F was determined to be GFVALLKKLPLILKHLH-NH 2 . In a similar manner, the peptide of fraction G was identified as GFVALLKKLPLILKHLP-NH 2 (Additional file 2).
These peptide sequences were not registered in the NCBI database (http://www.ncbi.nlm.nih.gov/) and Hymenoptera Genome Database (http://hymenopteragenome.org/). Thus, we named these novel peptides Xac Edman degradation of the peptide of fraction D provided a partial sequence: IIFVG TKWCG NGNVA EGPED LGSLK E-. Sequence similarity searches showed that the partial sequence conserved a 70% identity with those of PLA 2 s isolated from the bumblebee Bombus hypocrite (Apidae: Apinae: Bombini) and the social honeybee A. mellifera [18]. We hypothesized that this peptide would be a PLA 2 homolog and attempted molecular cloning and RACE to elucidate the full nucleotide sequence encoding this peptide (Fig. 4). The resulting sequence (DDBJ/ GenBank/EMBL accession no. AB731659) was compared with those of PLA 2 s isolated from bee venom, indicating that the PLA 2 homolog conserves characteristic amino acid residues associated with the catalytic activity of PLA 2 s of honey and bumblebees [18,19]. It is speculated that the PLA 2 of X. appendiculata is a product of a post-translational modification due to the fact that the molecular mass number of fraction D ([M + H] + m/z 16508) was not identical to that of the peptide estimated by the molecular cloning.
Physicochemical properties of Xac-1 and Xac-2: helical wheel projection analysis, CD spectroscopy analysis, and liposome leakage assay The helical wheel projection of Xac-1 and Xac-2 was made by the database program (http://www.tcdb.org/ progs/?tool=pepwheel) [20]. The results suggest that these peptides possess amphiphilic helical structures in which positively charged amino acid resides, histidines and lysines are arranged on one side and hydrophilic residues on the other side (Fig. 5). To obtain an analytical proof, CD spectra of Xac-1 was measured. Xac-1 exhibited a mostly disordered conformation in aqueous solution whereas a higher αhelical content in 50% TFE solution (Fig. 6). The presence of two negative dichroic bands at 208 and 222 nm was consistent with the preferential formation of α-helix. Subsequently, we analyzed liposome leakage properties of Xac-1 (Fig. 7). Xac-1 revealed liposome degradation activity in which its potency was almost the same as that of mastoparan.

Biological activities of Xac-1 and Xac-2
Antimicrobial and hemolytic activities of Xac-1 and Xac-2 were examined. Mastoparan (14 amino acid amphiphilic peptides from wasp venom) was selected as a reference peptide because it is a representative amphiphilic peptide that shows antimicrobial and hemolytic activities due to its potent pore forming effects and mast cell degradation activities [7]. In addition, melittin isolated from the venom of A.  mellifera was used as a reference to compare hemolytic activity [3]. These results are summarized in Table 1. Xac-1 exhibited growth inhibitory effects against E. coli, S. aureus, M. luteus, and S. cerevisiae with MIC values in a range from 1.57 to 6.25 μM.
The potency is similar to that of mastoparan. Xac-2 showed almost the same or slightly less potency as Xac-1 on the antimicrobial activities using M. luteus and S. cerevisiae. Xac-1 and Xac-2 exerted hemolytic activities (37.5 and 23.5% at 100 μM), respectively. These data were compared with those of mastoparan (40.6% at 100 μM) and melittin (91.8% at 10 μM). These results indicated that the potencies of Xac-1 and Xac-2 were close to that of mastoparan, whereas these potencies were much weaker than that of melittin. Bioactive peptides isolated from ant, bee, and wasp have been shown to activate the release of histamines from rat peritoneal mast cells [17]. Both Xac-1 and Xac-2 caused a significant and dose-dependent histamine release. At a concentration of 10 μM, Xac-1 and Xac-2 displayed 58.0 and 53.0% of histaminereleasing activities, respectively. These activities comparable to mastoparan (57.6%), but less effective than melittin (84.8%).

Discussion
Bees are classified into seven families including more than 16,000 described species [21]. Female bees use their venom for defense when they are exposed to dangers and predators. Bee stings are known to be painful. In contrast to the unpleasant effects of bee toxins in humans, its venom has been utilized as a remedy for centuries and recently has attracted much attention as a promising alternative and preventive medicine for the treatment of arthritis, rheumatism, pain, and cancer, etc. [8,22]. Although many biologically active peptides and enzymes have been isolated from the venom of social honeybees such as A. mellifera and a eusocial bumblebee (Megabombus pennsylvanicus), the structure elucidation of the venom substances of carpenter bees has not been well examined except for the Nakajima's study [23] on the analysis of biogenic amines in the venom of X. appendiculate. It revealed that histamine, putrescine and spermidine were detected as major biogenic amines in the venom.
Piek [15] predicted that the presence of melittin-like peptides in the venom of X. violacea by comparison with biological activities of the crude venom of X. violacea, A. mellifera, and Bombus terrestris. To the best of our knowledge, isolation of the peptide substances in the carpenter venom has not been elucidated yet.
In this study, we found two novel amphiphilic peptides, Xac-1 and Xac-2, and a new PLA 2 homolog in the venom of X. appendiculata for the first time. Our results corroborate that by Nakajima et al. [23] by clearly showing that the venom of X. appendiculata is a cocktail of biogenic amines, amphiphilic peptides, PLA 2 and the molecular constitution resembles those of honeybees and bumblebees. It is supposed that Xac-1 and Xac-2 would be a principle of the melittin-like peptide proposed by Piek [15] since biological activities of Xac-1 and Xac-2 resemble those of melittin.
Recently, the research interests for venoms has reached other families of solitary and eusocial bees (Fig. 8). These studies have unveiled the distribution of amphiphilic and biologically active peptides such as melectin from Mellecta albifrons (Apoidea: Melectini) [24], codesane from Colletes daviesanus (Colletidae) [25], osmin from Osma rufa (Megachilidae) [26], lasioglossins from Lasioglossum laticeps (Halicitidae) [27], halictines from Halictus sexcinctus (Halicitidae) [28], macropin from Macropis fulvipes (Melittidae) [29] in the bee venom. It is interesting to note that the amino acid sequences of Xac-1 and Xac-2 are similar to those of melectin and osmin isolated from the long-tonged bees, but not to those of bombolitins and melittin isolated from the social bee venom, though carpenter bees, bumblebees and honeybees are closely related. These comparable analyses indicate a possibility that Xac-1, Xac-2, melectin and osmin would be derived from a prototype amphiphilic peptide of the ancestor of solitary bees. On the other hand, melittin, bombolitins, mastoparan may have separately developed during the course of the social evolution. To prove this hypothesis, further research on isolation and biological studies on the bee venom peptides are required. PLA 2 is known to be the main enzyme component of bee venom. Previously, the presence of PLA 2 in the venom of Anthophora pauperata (Apidae) was proposed by biological and hematological studies [30]. To the best of our knowledge, structure analysis of the PLA 2 of carpenter bees has not been examined yet. We isolated the PLA 2 of X. appendiculata and found that it has a high sequence identity with PLA 2 of related species such as bumblebees and honeybees (Fig. 4) [31]. We also analyzed the molecular evolution of bee PLA 2 s using database sets ( Fig. 9) [32,33]. Interestingly, PLA 2 evolution tree did not match with the bee phylogeny that is well-established by large dataset though the characteristic amino acid residues of the PLA 2 family of honeybees and bumblebees are highly conserved in the PLA 2 of X. appendiculata. Our analysis would contribute to discuss the evolution patterns of PLA 2 s of the bee venoms.

Conclusion
We have analyzed the venom components of the solitary bee X. appendiculata and isolated novel amphiphilic peptides, Xac-1 and Xac-2, and a PLA 2  homolog. The accurate analysis and structure determination of the venom reveals that it is a cocktail of various biologically active molecules. Our study helps to understand the biological function and molecular diversity of the solitary bee venom components. Additionally, it may aid in the design of biologically active peptides based on the structures of Xac-1 and Xac-2 to develop more potent peptide analogs toward biotechnological and medical applications. Fig. 8 Amphiphilic venom peptides isolated from bees and wasps. The cladogram of bee families is based on Hedtke et al. [21]. Blue: basic amino acid residues Fig. 9 Molecular Phylogenetic analysis by Maximum Likelihood method based on PLA2 amino acid sequences (a) and phylogenetic tree proposed by Hedtke et al. [21] (b). The evolutionary history was inferred by using the Maximum Likelihood method based on the Whelan And Goldman model [32]. The tree with the highest log likelihood (-1575.2225) is shown. The percentage of trees in which the associated taxa clustered together is shown next to the branches (bootstrap value). Initial tree(s) for the heuristic search were obtained automatically by applying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using a JTT model, and then selecting the topology with superior log likelihood value. The rate variation model allowed for some sites to be evolutionarily invariable ([+I], 28.9827% sites). All positions containing gaps and missing data were eliminated. There were a total of 130 positions in the final dataset. Evolutionary analyses were conducted in MEGA7 [33]