Animals

Honeybees (Apis mellifera L.), maintained at the University of Tokyo, were used for all experiments.

SDS-PAGE and determination of N-terminal sequences

Honeybees were anesthetized on ice and the antennae were excised from the honeybee with fine scissors. Crude extracts were prepared by homogenization in PBS [10 mM phosphate buffered saline, pH 7.2, containing 130 mM NaCl, 5 mM KCl and protease inhibitor cocktail Complete (Roche, Basel, Switzerland)] using a pellet mixer, followed by centrifugation at 700 × g for 10 min at 4°C. The clear supernatant was used for electrophoretic analysis. Hemolymph was collected from each insect by severing the forelegs, injecting PBS into the thorax, and then drawing the expressed clear hemolymph into a micropipette. Proteins corresponding to 0.2,0.2, and 0.1 antennae of workers, queens, and drones, respectively, and 3 [.mu]l of hemolymph were subjected to 15% SDS-PAGE (Laemmli, 1970), followed by staining with Coomassie Blue-R. Electrophoretically-separated proteins were electroblotted onto PVDF membrane Immobilon-PSQ (Millipore, Bedford, MA) and bands of interest were excised and subjected to microsequencing using a Shimadzu (Kyoto, Japan) PPSQ-21 protein sequencer.

5′ and 3′-rapid amplification of the cDNA end (RACE)

Total RNA (2 μg), extracted from the antennae of worker or drone bees, was treated with 1 U of RNase-free DNase I (Invitrogen). First- and second-strands were synthesized with the Marathon-Ready cDNA Kit (Clontech, Palo Alto, CA) and then the Marathon cDNA adaptor (Clontech) was ligated. Polymerase chain reaction (PCR) was performed with Advantage2 Polymerase Mix (Clontech) using adaptor primers 1 and 2 included in the Clontech kit and specific primers for the D-AP1 and Amwat genes (for 5′-RACE, +409 to +429 of Amwat cDNA, for 3′-RACE, +1171 to +1200 and +1330 to +1359 of D-AP1 cDNA, and +689 to +716 of Amwat cDNA were used). The amplified DNA fragments were subcloned and sequenced. All the PCR artifacts were excluded by comparing the nucleotide sequences of at least three independent clones.

Enzyme assay

Antennae were homogenized in 50 μl of PBS (0.1 M phosphate buffered saline, pH 6.8), containing 0.3 M sucrose. The homogenates were centrifuged at 10000 × g for 5 min and the supernatant saved. The pellet was resuspended in 50 μl PBS containing 0.3 M sucrose, homogenized, and centrifuged as described above. The second supernatant was combined with the first and stored at –80°C. Carboxylesterase was assayed by the method of Sheehan et al. (1979) using β-naphthyl acetate as the substrate. One unit was defined as the amount required to hydrolyze 1 nmol of substrate in 30 min at 27°C under the assay conditions used. Protein concentrations of each homogenate were estimated using the BCA protein assay reagent kit (Pierce, Rockford, IL), with bovine serum albumin as the standard protein.

Northern blot analysis and the differential display method

Total RNAs of various tissues from workers, queens, and drones were extracted as described above. Total RNA (20 μg) was subjected to 1.2% formaldehyde-agarose gel electrophoresis, and the RNA was then transferred to a GeneScreen Plus Hybridization Transfer Membrane (New England Nuclear, Boston, MA). The filter was hybridized with [³²P]-labeled D-AP1, W-AP1 and Amwat cDNA probes corresponding to +1171 to +1884, +22 to +398, and +314 to +783 of their cDNA, respectively, for 16 hr at 42°C, followed by two washings with 2XSSC (1XSSC=150 mM NaCl, 15 mM sodium citrate) containing 0.1% SDS for 5 min each, and then with 0.1XSSC containing 0.1% SDS at 50°C for 15 min each. The filter was then subjected to autoradiography.

The differential display between workers and queens was performed essentially as described previously (Takeuchi et al., 2002) using total RNA extracted from queen and worker antennae and a Fluorescence Differential Display kit (TaKaRa, Tokyo, Japan) with 60 primer combinations. Bands of interest were excised from the gel and the DNA was reamplified, subcloned into a pGEM-3zf(+) vector (Promega, Tokyo, Japan), and transfected into Escherichia coli JM109. The sequences of both strands were determined.

Identification of drone-preferential antennal proteins

To identify honeybee antennal proteins expressed in a sex-preferential manner, we first compared the protein profiles of worker and drone antennae using SDS-PAGE. In a parallel experiment, the hemolymph of each insect was also subjected to SDS-PAGE to examine whether the detected bands were derived from the hemolymph contained in the antennae. Two bands of 60 kDa and 58 kDa were detected preferentially in drone antennae (Fig. 1) and the N-terminal sequence for the 60-kDa protein was determined to be LEDAPRVKTPLGAIKGYYKI. The N-terminal sequence of the 58-kDa protein was not determined, suggesting that its N-terminus was blocked. Because the 60-kDa protein, termed D-AP1 (Drone-preferential Antennal Protein-1) was not detected in the hemolymph (Fig. 1), this protein was considered to be involved in the antennal functions characteristic of drones.

Fig. 1

Detection of proteins differentially expressed in drone and worker antennae. Antennal (left panel) and hemolymph proteins (right panel) from worker (W) and drone honeybees (D) were analyzed by SDS-PAGE. D-AP1 (60 kDa) and another worker-selective 58-kDa antennal protein are indicated by an arrow and an arrowhead, respectively. Molecular weights are given in kDa on the right of the panels.

cDNA cloning of D-AP1 and its identification as a carboxylesterase

A database search revealed that the N-terminal sequence of D-AP1 was identical to a part of the deduced amino acid sequence of the EST348 clone from honeybee male antennae (DDBJ/EMBL/GenBank accession number BE844603). The deduced amino acid sequence of EST348 had 33% sequence identity with a part of Drosophila melanogaster carboxylesterase (Esterase 6 precursor, accession number M33780), suggesting that D-AP1 was a honeybee homologue of carboxylesterase. To isolate the full-length cDNA, we identified an Expressed Sequence Tag (EST) clone, EST349 (accession number BE844604), which was homologous to an insect carboxylesterase (32% identity to a part of the Esterase FE4 precursor from the Peach-potato aphid Myzus persicae) (Field et al., 1993). Reverse transcription-polymerase chain reaction (RT-PCR) analysis revealed that the EST348 and 349 clones were continuous (data not shown), and the RT-PCR product was then subcloned. Next, 3′-RACE was performed to isolate the full-length cDNA. Finally, a 2084-bp D-AP1 cDNA, which contained an open reading frame encoding 567 amino acid residues, was obtained (Fig. 2A). The polyadenylation signal was 66 bp downstream of the predicted stop codon. Comparison with the N-terminal sequence indicated that the 18-residues upstream of the N-terminus comprised a signal sequence, suggesting that D-AP1 is a secretory protein. The calculated molecular mass for the mature protein (61.8 kDa) was in good agreement with the molar mass measured using SDS-PAGE (60 kDa).

Fig. 2

Sequence analysis of D-AP1. (A) Nucleotide and deduced amino acid sequences of D-AP1 cDNA. Deduced amino acid residues coded by the cDNA are shown under the nucleotide sequences in capital letters. The termination codons are indicated by asterisks. The possible polyadenylation signal is double underlined. The determined N-terminal sequence is in bold and underlined. Signal sequence is in italics. Sequences corresponding to the EST348 and EST349 are boxed. (B) Predicted amino acid sequence of D-AP1 and alignment with carboxylesterases from three other species. MpFE, esterase FE4 of M. persicae; DmE6, esterase 6 precursor of D. melanogaster; and HsLE, liver carboxylesterase precursor of human. Gaps were introduced to obtain maximum matching. Identical residues among all five sequences are indicated with asterisks below the sequences, and conserved amino acids are in bold. The predicted amino acid residues for the catalytic triad are indicated by arrowheads and those for the oxyanion hole are indicated by crosses.

A database search revealed that the D-AP1 protein shared 38%, 33%, and 31% residue identities with carboxylesterases from M. persicae (Esterase FE4 precursor), D. melanogaster (Esterase 6 precursor), and human Homo sapiens liver carboxylesterase precursor, respectively (Fig. 2B) (Oakeshott, 1987; Munger, 1991; Field et al., 1993). In addition, all the amino acid residues required to assemble three critical motifs in the active site of carboxylesterase (Oakeshott, 1987; Cygler, 1993; Newcomb, 1997) were conserved in the D-AP1 sequence (Fig. 2B). Ser209, Glu341, and His464 of D-AP1 were assumed to form the catalytic triad. Furthermore, the Gly-X-Ser-X-Gly motif (X indicates any amino acid residue), which contains the active site Ser, was also conserved in D-AP1 (Gly207-Leu208-Ser209-Ala210-Gly211). The third conserved feature of the active site of the carboxylesterase is the oxyanion hole, which comprises three small residues. This was also predicted by the presence of Gly130, Gly131, and Ala210 in D-AP1. These results strongly suggest that D-AP1 belongs to the carboxylesterase family.

Expression analysis of D-AP1

We examined the expression of the D-AP1 gene by Northern blot hybridization using total RNA from antennae of drones and workers. A discrete 2.0-kb band was detected in the lane for drone antennae indicating that the D-AP1 gene is expressed preferentially in the drone antennae (Fig. 3A). We also examined the tissue-specificity of the expression of the D-AP1 gene in the drone and worker. The D-AP1 gene was strongly expressed in the antennae and weakly in the head, while there was no significant signal detected in the legs, thorax, or abdomen of the drone (Fig. 3B). Because the head contains the antennae, this finding indicated that the D-AP1 gene was expressed almost exclusively in the antennae of the drones. In contrast, there was a weak expression in the antennae and almost no significant expression was detected in any other body parts of the workers (Fig. 3B).

Fig. 3

Analysis of D-AP1 gene expression. Expression of the D-AP1 gene in the worker and drone antennae (A) and in various body parts of the drone (B, left panels) and worker (B, right panels) was analyzed by Northern blot hybridization. Arrows indicate the positions of the D-AP1 transcripts (top panels) and 18S rRNA bands stained with ethidium bromide (bottom panels). 18S rRNA was analyzed to indicate that approximately equal amounts of RNA were loaded in each blot. RNA size markers are given in kb.

Carboxylesterases have broad substrate specificities and are usually detected by the hydrolysis of α- or β-naphthyl acetate (Morton and Singh, 1985). To examine whether carboxylesterase activity is enhanced in drone antennae, the enzymatic activities in drone and worker antennae were compared. The specific activity in drone antennae was approximately 2.3-fold higher than in worker antennae (9.5±1.2 and 4.2±0.5 U /μg protein, respectively), consistent with the drone-preferential expression of D-AP1. The difference in specific activity, however, was much smaller than the difference in the amount of the D-AP1 transcript. Database search revealed that there was also a clone (EST324, accession number BE844579) for the other carboxyesterase isoform having 38% sequence similarity with D-AP1. We performed semi-quantitative RT-PCR and found that the EST324 mRNA was expressed almost equally between worker and drone antennae (data not shown). Thus, the discrepancy between the antennal enzymatic activity and DAP1 expression could be partly explained by the presence of multiple carboxylesterases in the honeybee antennae.

Identification of worker-preferential antennal proteins

We next identified honeybee antennal proteins expressed in a caste-preferential manner and compared protein profiles of extracts of antennae from workers and queens using SDS-PAGE. A worker-preferential 10-kDa protein and a queen-preferential 8-kDa protein were detected (Fig. 4). We determined the N-terminal sequence of the 10-kDa protein, which we termed W-AP1 (Worker preferential Antennal Protein-1), as EELYSDKYDYVNIDEILAND. The N-terminal sequence of the 8-kDa protein could not be determined because it required a very large sample. A database search revealed that the N-terminal sequence of W-AP1 was identical to a part of the deduced amino acid sequence of a 546-bp EST clone, EST117, from honeybee female antennae (accession number BE844373). EST117 contained an open reading frame encoding 116 amino acid residues, and a polyadenylation signal 129 bp downstream of the stop codon. Two other EST clones, EST118 and EST142 (accession numbers, BE844374 and BE844397, respectively), were also partial cDNAs for WAP1, as both overlap with EST117. The assembled consensus cDNA encoded a 116-amino acid protein (Fig. 5A). Comparison with the N-terminal sequence of mature W-AP1 indicated that the 19-residues upstream of the N-terminus was a signal sequence, suggesting that W-AP1 is a secretory protein. The calculated molar mass for the mature protein, 11.5 kDa, was in agreement with the molar mass measured using SDS-PAGE.

Fig. 4

Detection of proteins differentially expressed in worker and queen antennae. Antennal (left panel) and hemolymph proteins (right panel) from worker (W) and queen honeybees (Q) were analyzed by SDS-PAGE. W-AP1 (10 kDa) and a protein detected preferentially in queen antennae (8 kDa) are indicated by an arrow and an arrowhead, respectively. Molecular weights are given in kDa on the right side of the panels.

Fig. 5

Sequence analysis of W-AP1. (A) Nucleotide and deduced amino acid sequences of W-AP1 cDNA. The nucleotide-coding region is shown in capital letters. The termination codons are indicated by asterisks. A potential polyadenylation signal is double underlined. The N-terminal sequence is indicated in bold and underlined. Signal sequence is in italics. (B) Predicted amino acid sequence of W-AP1 and alignment with other chemosensory proteins. Asp3c from A. mellifera; SgCSP, CSP-sg4 from S. gregaria; Pap10, p10 from P. americana; DmOSD, OSD from D. melanogaster, and CcCLP, CLP-1 from Cactoblastis cactorum. The four conserved cysteines are marked by solid triangles above the sequences. Identical residues among all five sequences are indicated with asterisks below the sequences, and conserved amino acids are in bold. Alignment was optimized by introducing several gaps (–).

A database search revealed that the deduced amino acid sequence of the mature protein shared 43%, 40%, 39%, 38%, and 31% residue identities with chemosensory proteins (CSP) from Schistocerca gregaria (CSP-sg4), Cactoblastis cactorum (CLP-1), D. melanogaster (OS-D), Periplaneta americana (p10), and Apis mellifera (ASP3c) (accession numbers, AF070964, U95046, U05244, AF030340, and AF481963, respectively) (Maleszka and Stange, 1997; Kitabayashi et al., 1998; Angeli et al., 1999; Briand et al., 2002) (Fig. 5B). Furthermore, W-AP1 possessed all four cysteine residues that are conserved among the CSP family. These results strongly suggest that W-AP1 belongs to the CSP family.

Identification of Amwat, the worker-enriched antennal transcript

To further identify gene(s) differentially expressed between castes, we employed the differential display method using total RNA extracted from queen and worker antennae. By screening approximately 1000 bands, several candidate bands were detected that were differentially expressed in queen and worker antennae. Among them, we identified a gene termed Amwat (Apis mellifera worker enriched antennal transcript) that was expressed preferentially in the worker antennae. Northern blot analysis revealed a strong Amwat mRNA signal of about 0.8 kb in size in worker antennae, whereas there was a weak signal in queen antennae (Fig. 6A). We also compared the expression between workers and drones, and the results indicated a preferential expression in the worker antennae (Fig. 6B). These findings suggest a worker-preferential function of the Amwat protein in the antennae.

Fig. 6

Worker-preferential expression of Amwat in antennae. Total antennal RNA was subjected to Northern blotting using the Amwat cDNA probe. Arrows indicate the Amwat transcripts (top panels) and 18S rRNA bands stained with ethidium bromide (bottom panels). RNA size markers are given in kb. (A) Hybridization with total RNA from antennae of workers and queens. (B) Hybridization with total RNA from antennae of workers and drones.

To obtain a full-length cDNA, 5′- and 3′-RACE were performed using total RNA from worker antennae and primers based on the differential display product. As a result, an 863-bp consensus cDNA, which is in good agreement with the mRNA size estimated by Northern blot analysis, was obtained (Fig. 7A). The sequence analysis revealed an open reading frame that encodes 87 amino acid residues. The initiator methionine was followed by a highly hydrophobic sequence of 17 amino acid residues, which is typical for signal sequences of secretory protein precursors. According to the 01, 03 rule (Heijne, 1986), cleavage of the signal sequence from the precursor would occur between amino acid positions 19 (S) and 20 (Q).

Fig. 7

Sequence analysis of Amwat. (A) Nucleotide and deduced amino acid sequences of Amwat cDNA. The nucleotide-coding region is shown in capital letters. The termination codons are indicated by asterisks. The sequences corresponding to the cDNA obtained by the differential display are boxed. The predicted signal sequence is in italics. (B) Comparison of amino acid sequence of Amwat with those encoded by RH61650 of D. melanogaster (Dm) and EST328 of M. sexta (Ms). Predicted signal sequences are shown in italics. Identical residues among all sequences are indicated with asterisks below the sequences. Predicted mature peptides are underlined, and flanking putative processing sites are shown in bold.

A database search revealed that this protein sequence had homologies with the predicted amino acid sequences of the 514- and 690-bp EST clones from D. melanogaster head (accession number BI631960) and M. sexta antennae (Accession number AI187536) of unknown functions (47 and 44% identity, respectively) (Fig. 7B). The N-terminus of these deduced amino acid sequences is highly hydrophobic, like theAmwat protein, suggesting that they are signal sequences for secretory proteins. A much higher similarity was observed in the C-terminal region, which is preceded by a recognition sequence of proprotein convertase 2, RXXR (Day et al., 1998). Proteolysis in these sequences generates highly conserved small proteins (peptides) between species (61% and 52% identities between A. mellifera and D. melanogaster or M. sexta, respectively). The above results suggest that the Amwat protein and its homologues belong to a novel secretory peptide family conserved among insects.

Tissue-specific distribution of W-AP1 and Amwat transcripts

We examined the expression of W-AP1 and Amwat mRNAs in different body parts of workers. A strong single W-AP1 transcript of approximately 0.6 kb was detected in the antennae but only weakly in the legs and head, suggesting it has antennal-selective functions (Fig. 8A). In contrast, the Amwat transcript was detected strongly in the legs, on which contact chemoreceptors (gustation) reside (Gullan and Cranston, 1994), and only weakly in the antennae and thorax (Fig. 8B). Because the thorax contains legs, this finding indicates that this transcript is expressed predominantly in the legs, selectively in the antennae, and not in other body parts, suggesting that the Amwat gene products have a unique role in the general chemosensory organs.

Fig. 8

Analysis of the W-AP1 and Amwat gene expression in various body parts. Total RNAs from worker antennae, legs, head, thorax, and abdomen were subjected to Northern blot analyses using W-AP1 (A) or Amwat (B) cDNA probes. Arrows indicate positions for each transcript (top panels) and 18S rRNA bands stained with ethidium bromide (bottom panels). RNA size markers are given in kb.