Generation of Monoclonal Antibodies Detecting Specific Epitopes in Locust Antennae

Volatile semiochemicals detected by the highly sensitive chemosensory receptor cells in the antennae play an important role in regulating the behaviour of insects; they provide information on the location of food, hosts, mates or oviposition sites and may initiate specific physiological and behavioural transformations. In the desert locust, olfaction plays a central role in the detection of food sources (Greenwood and Chapman, 1984) and is thought to be involved in maturation and social interactions (Amerasinghe, 1978; Gillet, 1983). Approaches to controlling locust plagues by interfering with this important sensory system have been considered for quite a while, particularly as these methods are supposed to be environmentally safe, target-specific and do not lead to the resistance seen with pesticides (Ferenz, 1990). An essential prerequisite for developing sound strategies for manipulating the chemosensory system is a more detailed understanding of the primary mechanisms that enable locusts to perceive airborne chemical cues. Unfortunately, investigations on insect olfaction have been restricted to a few species, mainly to aspects of pheromone detection in moths (Kaissling, 1986). The antennae of the locust (Locusta migratoria) consist of about 20 segments equipped with several thousand sensilla (Boeckh, 1967), which can be classified into trichoid, basiconic and coeloconic types (Altner et al. 1981). Electrophysiological recordings have demonstrated that most of these sensillum types have an olfactory function (Boeckh, 1967; Kafka, 1970). However, at present there is very little information on the molecular machinery which enables the receptor cells in locust antennae to detect odorous molecules and to transduce the chemical stimulus into neuronal signals.

One way of investigating the specificity of olfactory receptor cells is by using specific, site-directed agents. As such ligands are not available per se, hybridoma technology can be employed to generate monoclonal antibodies (mAbs) against molecules that are selectively expressed in one particular tissue but not in others, following a differential screening paradigm (Köhler and Milstein, 1975). The availability of specific mAbs allows the application of analytical approaches disclosing the cellular localization, the molecular identity and ultimately even the function of an antigenic protein. Attempts to raise mAbs against olfactory receptor cells from vertebrates and insects have recently been made in several laboratories (Hempstead and Morgan, 1985; Hishinuma et al. 1988; Anholt et al. 1990; Strotmann and Breer, 1991).

Here we report the generation of two monoclonal antibodies which react with antennal cells of Locusta migratoria.

Adult locusts (Locusta migratoria) and embryos were obtained from the Insektarium Dr Frieshammer, Jaderberg. Day E1 was the day of oviposition, hatching was on day E13 at 28°C. Balb/c mice were supplied by the Zentralinstitut für Versuchstierzucht (Hannover).

Cell culture media were from GIBCO (RPMI 1640) and Serva (Schneider’s Drosophila medium and basal medium Eagle [BME]). 5+4 medium was made from five parts of Schneider’s Drosophila medium and four parts of BME. Calf sera and cell culture additives were purchased from GIBCO, as were complete and incomplete Freund’s adjuvants. Polyethylene glycol (PEG) 500 was supplied by Roth. Culture plates and enzyme-linked immunosorbent assay (ELISA) plates were obtained from Nunc. Nitrocellulose sheets (0.45 μm) were obtained from Schleicher and Schuell. Tetramethylrhodamine isothiocyanate (TRITC)-conjugated goat anti-mouse Ig was obtained from Sigma. Fluorescein isothiocyanate (FITC)-conjugated sheep anti-mouse Ig came from Amersham. Alkaline-phosphatase-conjugated goat anti-mouse Ig was from Promega, horseradish peroxidase (HRP)-conjugated goat anti-mouse was from BioRad. All other reagents were obtained in the purest form commercially available.

Animals were anaesthetized on ice and muscle tissue, head and thoracic ganglia and antennae were dissected and collected at 4°C. Crude membranes were prepared as described by Breer et al. (1985). Tissues were homogenized in 50mmol l⁻¹ Tris/HCl, pH6.8, 150mmol l⁻¹ NaCl, 5mmol l⁻¹ EDTA, 3mmol l⁻¹ EGTA, 0.1mmol l⁻¹ phenylmethylsulphonyl fluoride (PMSF) and centrifuged for 10min at 1000 g. The supernatant (S1) was centrifuged at 27000 g for 10min. The pellet (P2) was resuspended in 10mmol l⁻¹ Tris/HCl, pH7.4, 150mmol l⁻¹ NaCl and stored at −70°C. Protein concentration was determined according to Bradford (1976).

Female 6-week-old Balb/c mice received intraperitoneal injections of 100 μg of antennal protein emulsified in complete Freund’s adjuvant and two similar injections with incomplete Freund’s adjuvant at 2-week intervals. Three days after the last immunization, the spleen was removed and the spleen cells were fused to P3-X63 Ag8.653 myeloma cells (Kearney et al. 1979) by a standard fusion protocol (Galfre and Milstein, 1981) using PEG 500 as the fusion reagent. Cells were plated in microtitre wells in hypoxanthine, aminopterin and thymidine (HAT) selection medium with 10% foetal calf serum at 37°C in a humidified atmosphere containing 5% CO₂. Antibody-producing clones were identified by ELISA and subcloned twice by limiting dilutions.

ELISA plates were coated overnight at 4°C with 100ng protein per well in 50mmol l⁻¹ NaHCO₃, pH9.5. Unbound material was washed out with 10mmol l⁻¹ Tris/HCl, pH7.6, with 150mmol l⁻¹ NaCl (TBS). After preincubation with TBS containing 1% bovine serum albumin (BSA), culture supernatants were incubated in the plates for 1h, followed by HRP-conjugated goat anti-mouse Ig (1:7500 in TBS). Each step was followed by three TBS washes. Antibody binding was visualized with 0.005% 3,5,3’,5’-tetramethylbenzidine and 0.003% H₂O₂ in 100mmol l⁻¹ sodium acetate/citrate buffer; substrate turnover was stopped after 10min with 2mol l⁻¹ H₂SO₄; absorbance was measured at 450nm with a Dynatech microplate reader MR700.

Samples of membrane preparations were diluted 1:1 in sample buffer (125mmol l⁻¹ Tris/HCl, pH6.8, 20% glycerol, 10% 2-mercaptoethanol, 2% SDS) to a final protein concentration of 1 μg μl⁻¹. Samples were heated in a boiling water bath for 3min and loaded onto 12.5% gels using the Laemmli buffer system (Laemmli, 1970). Proteins were transferred onto nitrocellulose according to Towbin et al. (1979). Non-specific binding sites were blocked by incubation of blots with 1% BSA in 10mmol l⁻¹ Tris/HCl, pH8.0, 150mmol l⁻¹ NaCl, 0.05% Tween 20 (TBST). An incubation with culture supernatants for 1h was followed by three TBST washes of 15min. Blots were subsequently incubated with alkaline-phosphatase-conjugated goat anti-mouse Ig (1:7500) and washed as described above. Antibody binding was visualized with 0.015% nitro-blue tetrazolium and 0.007% 5-bromo-4-chloro-3-indolylphosphate in 100mmol l⁻¹ Tris/HCl, pH9.5, 100mmol l⁻¹ NaCl, 5mmol l⁻¹ MgCl₂. Substrate turnover was stopped by addition of 20mmol l⁻¹ Tris/HCl, pH8.0, 5mmol l⁻¹ EDTA.

The membrane pellet from adult locust antennae was resuspended in 30mmoll⁻¹ Tris, pH8.0, 100mmol l⁻¹ NaCl, 2mmol l⁻¹ EDTA, 2mmol l⁻¹ EGTA, 0.1mmol l⁻¹ PMSF and 0.5% CHAPS and proteins were solubilized for 1h at 4°C. After a centrifugation at 27000 g for 15min, CHAPS in the supernatant was reduced to 0.05% using PD-10 columns (Pharmacia). Gel filtration was performed using a Pharmacia FPLC and Sephacryl S200 as column material with a constant flow rate of 0.5mlmin⁻¹. Absorbance at 280 nm was monitored with a single-path UV-1 monitor; 2ml fractions were collected and dotted onto nitrocellulose. Immunoreactive fractions were visualized as described for Western blots.

ELISA plates were coated with 100ng protein per cup as described above. After three washes with TBS/gelatine, the cups were incubated with 0.01% trypsin in 10mmol l⁻¹ Tris/HCl, pH7.4, 150mmol l⁻¹ NaCl for 30min at 37°C. Subsequently, the plates were washed three times with TBS/gelatine followed by the ELISA protocol described above. All solutions used subsequently were supplemented with 0.01% soybean trypsin inhibitor. Control experiments were performed with trypsin preincubated with 0.1 % trypsin inhibitor.

Alternatively, samples were incubated with 0.001% proteinase K in 20mmol l⁻¹ Tris/HCl, pH7.8, 0.1% CaCl₂ for 30min at 37°C. After three washes in TBS/gelatine, the ELISA protocol described above was followed.

Immunohistochemical examination was carried out on cryostat sections of antennae from adult locusts. Specimens were cut into small segments and fixed for 4h by immersion in 4% paraformaldehyde in phosphate-buffered saline (PBS), pH7.4, at 4°C. After three rinses in PBS, they were transferred for 24h to ice-cold 30% sucrose in PBS. Fixed and cryoprotected tissues were embedded in Tissue Tec (Miles Inc.) and rapidly frozen by immersion in isopentane (−80°C). Sections (10 μm) were cut on a Reichert and Jung cryostat model 2800 E, thaw-mounted on chrome-alum/gelatin-coated slides and air-dried for 30min. Sections were treated with 0.1% Triton X-100 in PBS for 2min and additional binding sites were blocked with 1% BSA in PBS for 30min.

For double-labelling experiments, mAb B14/6D2 was purified from culture supernatant by caprylic acid precipitation according to Reik et al. (1987) and conjugated to FITC according to The and Feltkamp (1970). Sections were sequentially incubated with mAb S1/5D5 followed by TRITC-conjugated goat anti-mouse Ig (1:400) and FITC-conjugated mAb B14/6D2 for 1h each. Each of the steps was followed by three PBS washes for 5min. Sections were mounted in Citifluor (Amersham) to retard fading of fluorescence during microscopy and were examined under a Zeiss epifluorescence microscope.

Antennal cells from adult and embryonic locusts were prepared by cutting off an antenna, opening it with a razor blade and carefully scraping out the cells. The cells of ten antennae were collected in 1ml of 5+4 medium and carefully dissociated with a Pasteur pipette. The preparation was filtered through a 100 μm gauze and centrifuged for 5min at 1000 g. The pellet was resuspended in 5+4 medium and the cells were plated in 3.5cm culture dishes. They were covered with 5+4 medium and kept at 27°C. For immunohistochemical examination, cells were allowed to attach, the culture supernatant was aspirated and the cells were fixed with 4% paraformaldehyde in PBS for 10min at room temperature. Visualization of antibody binding was performed as described for tissue sections.

To inhibit protein biosynthesis, the left antenna from hatching locusts was cut off and transferred to 5+4 medium supplemented with 100 μmol l⁻¹ cycloheximide for 6h; the right antenna from the same animal served as a control and was incubated in 5+4 medium without additives. After that period, immunochemical examination of dissociated cells with mAb B14/6D2 was carried out as described above.

Hybridoma cell lines were produced from the fusion of a myeloma cell line (P3-X63 Ag8.653) with spleen cells from mice immunized with membrane preparations from locust antennae. Antibodies secreted by hybridoma cells were monitored for reactivity with membrane preparations from antennae as well as from nervous and muscle tissue using ELISA assays. This differential screening approach led to the identification of antibodies that displayed distinct immunoreactivity with antennal preparations but only limited or no reactivity with membranes from ganglia and muscle tissue. The monoclonal antibody (mAb) designated S1/5D5 showed particularly high binding to antennal and ganglionic membrane proteins, whereas the mAb designated B14/6D2 specifically reacted with antennal membranes (Fig. 1). Therefore, these antibodies were chosen for further investigation. Both mAbs were found to be of the IgG₁ subtype.

In phase-contrast images from longitudinal sections of antennal segments the multicellular organization of sensory organelles, as already described for other grasshopper species, is evident (Fig. 2). Slifer et al. (1959) have found that the number of sensory neurones associated with an insect sensillum varies from a few to more than 50. Indirect immunofluorescence analyses on cryostat sections were performed to explore the topochemical localization of the antigens in locust antennae. The distribution of immunoreactivity for both antibodies in longitudinal sections of segments from locust antennae is shown in Fig. 3. By application of a double labelling technique the reactive sites for both antibodies were visualized in the same section. It is immediately obvious that the antennal nerve is heavily labelled by the S1/5D5 antibody (Fig. 3A), whereas the B14/6D2 antibody does not react with it (Fig. 3B). In the sense organs, the surface of several large cell bodies, which are probably receptor neurons, is labelled by the S1/5D5 antibody. In contrast, the immunoreactivity of the B14/6D2 antibody is distributed between these cell clusters. Control experiments were performed using non-relevant antibodies of the same subtype (IgG₁); these antibodies showed no reaction.

Biochemical analyses were performed to characterize the antennal antigens. Membrane proteins from adult locusts were separated by SDS–PAGE, transferred onto nitrocellulose and probed for immunoreactivity. As can be seen in Fig. 4, the monoclonal antibody S1/5D5 strongly stained a protein band with an apparent M_r of 70000; the weakly stained band at M_r 25000 is due to an unspecific reaction of the second antibody. The B14/6D2 antigen could not be detected on Western blots. Following a size-dependent separation of antennal membrane proteins by gel filtration on Sephacryl S200, the B14/6D2 immunoreactivity was detected in the void volume (Fig. 5), suggesting that the B14/6D2 epitope is located on a large molecule. The antigens were further characterized by their sensitivity to various treatments. As documented in Table 1, the S1/5D5 antigen was very sensitive to protease treatment (trypsin and proteinase K); the B14/6D2 reactivity disappeared upon treatment of tissue sections with various detergents (Deoxycholic acid, Tween 20) or fixatives (glutaraldehyde).

To circumvent the well-known problems of characterizing antennal cells protected by the tough and impermeable cuticle (Slifer et al. 1959), clusters of cell bodies were isolated from split antennal segments. The cells were mechanically dissociated and kept alive in primary cultures for several days. As can be seen in Fig. 6, in a suspension of cultured cells from adult locust antennae the surface of some cell somata is labelled by the S1/5D5 antibody; staining with the B14/6D2 antibody was detected in a different subpopulation of cells (Fig. 6B). Thus, these cytological observations confirm the results obtained in the histological approaches (Fig. 3).

In a quantitative analysis, antennal membrane preparations from different developmental stages were assayed for immunoreactivity. The data in Fig. 7A indicate that B14/6D2 immunoreactivity was very low during embryogenesis; significant immunoreactivity appeared just before hatching, whereas S1/5D5 reactivity was already detectable in the earliest stages investigated. The level of B14/6D2 reactivity increased further during maturation; in adult antennae, the binding of antibodies was about five times that seen around hatching.

The ontogenetic development of B14/6D2 immunoreactivity could also be demonstrated at the cellular level. As can be seen in Fig. 7B, 1h after hatching a faint immunofluorescence first appeared; 6h after hatching the surfaces of several distinct somata in an isolated cell cluster were heavily labelled.

To explore whether the sudden appearance of B14/6D2 immunoreactivity was due to the exposure of epitopes on preformed molecules or whether it depended on the synthesis of new proteins, the expression of immunoreactivity was monitored on antennal cells in culture. Fig. 8 demonstrates that the level of antigen in membrane preparations from antennae before hatching increases more than fivefold after incubation for 6h in 5+4 medium, while in the presence of cycloheximide, which efficiently blocks protein synthesis, no significant increase in immunoreactivity was observed. These results indicate that the developmental pattern of epitope appearance is sustained under experimental conditions and requires expression of new proteins.

Following a differential screening paradigm, two monoclonal antibodies were generated which selectively interact with the surfaces of distinct populations of antennal cells from adult grasshoppers (Locusta migratoria). One of the antibodies (S1/5D5) recognizes a 70×10³M_r constituent on the surface of putative antennal receptor cells that is also present in the axons of the antennal nerve, whereas the B14/6D2 antigen was identified on the surface of cells that are possibly supporting cells. Both antigens may be considered as markers of specific cells in locust antennae in situ as well as in vitro. mAb S1/5D5 may be used to identify isolated receptor cells. The identification of receptor neurons in vitro is an important step towards a precise evaluation of the physiological mechanisms underlying the specific recognition and transduction processes in these highly specialized cells. Electrophysiological studies on identified receptor cells (Wegener et al. 1992a,b) will help to explore the sensory modality of isolated antennal cells and contribute to elucidating whether chemosensory receptor cells from locust are fine-tuned to detect specific pheromones, as has been described for several moth species (Kaissling, 1986), or broadly tuned for recognizing general odours relevant to adult locust.

The sudden appearance of the B14/6D2 antigen around the time of hatching, the phase of development when the sensitivity of antennal cells to odorous molecules first appears, suggests that it may represent a functionally important element of the chemosensory machinery involved in odorant detection. This idea is strengthened by the observation that the appearance of this epitope requires the synthesis of proteins and is not due to a modification (e.g. glycosylation or phosphorylation) of pre-exisiting molecules. The appearance of immunoreactivity in a critical phase of differentiation apparently reflects the expression of new gene products; mAb B14/6D2 may thus be a valuable tool for following the cellular and functional differentiation of antennal cells during development. In summary, the monoclonal antibodies generated in this study provide the first markers of antennal cells from locusts. They may be invaluable tools in future studies to elucidate the cellular and molecular mechanisms in odour recognition and chemo-electrical signal transduction in insects.

This work was supported by EC programme ‘Science and Technology for Development’ and by the Land Baden-Württemberg as part of the cooperation programme between Hohenheim and Hebrew University.