Neuronal Nicotinic Receptors in the Locust Locusta migratoria

We have identified five cDNA clones that encode nicotinic acetylcholine receptor (nAChR) subunits expressed in the nervous system of the locust Locusta migratoria. Four of the subunits are ligand-binding α subunits, and the other is a structural β subunit. The existence of at least one more nAChR gene, probably encoding a β subunit, is indicated. Based on Northern analysis and in situ hybridization, the five subunit genes are expressed. locα1, locα3, andlocβ1 are the most abundant subunits and are expressed in similar areas of the head ganglia and retina of the adult locust. Because Locα3 binds α-bungarotoxin with high affinity, it may form a homomeric nAChR subtype such as the mammalian α7 nAChR. Locα1 and Locβ1 may then form the predominant heteromeric nAChR in the locust brain. locα4 is mainly expressed in optic lobe ganglionic cells and locα2 in peripherally located somata of mushroom body neurons. locα3 mRNA was additionally detected in cells interspersed in the somatogastric epithelium of the locust embryo, suggesting that this isoform may also be involved in functions other than neuronal excitability. Transcription of all nAChR subunit genes begins approximately 3 days before hatching and continues throughout adult life. Electrophysiological recordings from head ganglionic neurons also indicate the existence of more than one functionally distinct nAChR subtype. Our results suggest the existence of several nAChR subtypes, at least some of them heteromeric, in this insect species.

In insects, neuromuscular transmission is mediated by glutamate, whereas acetylcholine is the principal neurotransmitter in the nervous system (1). A large body of evidence suggests the existence of both muscarinic and nicotinic acetylcholine (nAChR) 1 receptors in the insect brain, with nAChR-coding RNAs having been identified in several species, including the fruit fly Drosophila (2), the locust Schistocerca (3), the tobacco hornworm Manduca (4), and the peach-tomato aphid Myzus (5). Considerable pharmacological diversity of nicotinic receptors is indicated by the existence of ␣BTX-sensitive and -insensitive receptors (6,7) and by the rather wide variation of responses to nicotinic and non-nicotinic drugs of insect neurons and membrane preparations (8,9). In particular, the nAChR of Locusta migratoria was suggested to have mixed nicotinic and muscarinic pharmacology (10), which could correlate with the greater evolutionary age of orthoperians as compared with dipterians.
Vertebrate neuronal nicotinic receptors are quite diverse (11), with to date eight ␣ subunits and three ␤ subunits cloned in the rat. Of these, the ␣7, ␣8, and ␣9 subunits have the unique ability to form functional homomeric receptors (12)(13)(14). Various combinations of the other ␣ and ␤ subunits also give rise to functional receptors, as is exemplified by combinations of ␣4 and ␤2 subunits and of ␣3 and ␤4 subunits expressed in hippocampal neurons (15). The stoichiometries of heteromeric neuronal nAChRs are not yet established.
The homo-oligomeric receptors appear to be the evolutionarily oldest (16), which has led to the suggestion that the invertebrate neuronal nAChR from L. migratoria, given its broad pharmacology, may be an ␣7-like homo-oligomeric receptor (17,18). In Drosophila, five different putative nAChR subunits have been identified, three of which contain the two adjacent cysteines that are characteristic of ligand-binding ␣ subunits (2).
The subunit compositions and stoichiometries of insect nicotinic receptors are still unknown. This is in part due to the fact that expression in Xenopus oocytes of insect nAChR subunit RNA and cDNA has generally proven to be difficult (19). For the same reason, it has not yet been possible to determine the electrophysiological and pharmacological properties of single subtypes of insect nicotinic receptors.
In the present study, we show that in the locust L. migratoria at least six different genes exist that encode nAChR subunits. Four of these genes encode for ␣ subunits. Although we were unable to demonstrate heterologous functional expression in Xenopus oocytes of single subunits, or combinations of subunits, in situ hybridization studies show that the identified subunits are expressed in vivo and probably form functional receptors with different quarternary structures. Because insect nAChRs represent important targets for insecticides (20,21), the structural and functional characterization of such receptors may be useful in the context of rational drug design.

EXPERIMENTAL PROCEDURES
Materials-Locust eggs were supplied by Futtertierversand Hintze, Berlin (Germany). Dr. August Dorn (Institute of Zoology, University * This work was supported by the Stiftung fü r Innovation Rheinland/ Pfalz, the Fonds der Chemischen Industrie, and BAYER AG, Leverkusen. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AJ000390 -000393.
Mainz) provided live adult locusts, and Dr. Heinz Breer (Institute of Zoology, Stuttgart-Hohenheim; Germany) provided muscle and ganglia tissue from adult locusts.
Preparation of a Locust-specific Genomic nAChR Probe-A L. migratoria genomic library was constructed using the lambdoid phage NM1149 in combination with the hfl mutant Escherichia coli strain POP13b, which suppresses the lytic cycle of non-recombinant phages. Chromosomal DNA was prepared from muscle tissue of adult locusts as described (22). DNA was restricted with EcoRI, fractionated on a 0.7% agarose gel, and fragments from 3.8 to 5.4 kilobase pairs length were isolated. Samples of 100 ng of fractionated DNA were ligated with 1 g of EcoRI-restricted NM1149-DNA and packaged. Packaging efficiency was 8 ϫ 10 5 lytic plaques per 100 ng of Locusta DNA.
The E. coli strain POP13b was infected with 1.5 ϫ 10 6 recombinant plaques, and the genomic library was screened using a [␣-32 P]dATP randomly labeled 650-bp fragment from the Drosophila ARD cDNA (23). Filters were hybridized overnight and washed with low stringency (2ϫ SSC, 55°C; 20ϫ SSC, saline sodium citrate: 0.3 M sodium citrate, 3 M NaCl). After 48 h exposure two positive signals were identified. These phages were plated with lower density and hybridized. Two recombinants were isolated and characterized. A 177-bp AluI fragment of one of these recombinants coded for the full second transmembrane domain (TM2), the following short loop, and part of the third transmembrane domain (TM3) of a nAChR. This fragment was used as screening probe in further experiments.
Cloning of the First 200 Amino Acids of the Loc␣1 cDNA Clone-RNA was prepared from d9 embryos, and single strand cDNA was prepared from 5 g of total RNA using a mixture of oligo(dT) 12-18 and random primers. Reaction was carried out using 400 units of Superscript Reverse Transcriptase (Life Technologies, Inc.) following the provider's protocol.
Construction of Full-length cDNAs-The BamHI/EcoRI fragment of the rat ␣3 nAChR cDNA (24) and the locust ␣2, ␣3, and ␤ cDNAs were cloned into pBluescript® KSϩ (Stratagene). A StuI site was introduced into the 5Ј-end of these clones by PCR without changing the Arg/Pro amino acid sequence at this site. PCR was performed with 1 ng of each clone under the same conditions as described above. Amplification of the locust clones was carried out using a 3Ј-primer, designed to the pBluescript® KSϩ sequence (5Ј-AACAGCTATGACCATG-3Ј), and the following specific 5Ј-primers containing a silent mutation (in order to introduce a StuI site; see in bold letters) (loc␣2, 5Ј-ACCGCCTCATCAG-GCCTGTCACCAACAACTCCGA-3Ј; loc␣3, 5Ј-ACCGCCTCATCAGGC-CTGTCGGCAACAACTCGGA-3Ј; loc␤, 5Ј-ACAAGCTCATCAGGCCT-GTGCAGAACATGACGCA-3Ј). Amplification of the rat ␣3 nAChR fragment was carried out with 2 ng of cloned DNA using the following primers: 5Ј-GTAAAACGACGGCCAGT-3Ј, 5Ј-CCTCCGAATTCTC- CGGA-GATGATCTCGTTGTAAT-3Ј). PCR products were cloned into pBluescript® KSϩ and sequenced. The StuI site was then used to fuse the rat ␣3 nAChR signal sequence to each of the locust cDNA clones.
In Situ Hybridization-In situ hybridization was performed with digoxigenin-labeled RNA probes. RNA probes were randomly labeled with digoxigenin according to the protocol provided by the manufacturer (Boehringer Mannheim, Germany). In order to obtain isoformspecific probes, sequence regions with lowest homology were selected, i.e. for loc␣1 and loc␣2 clones from the "cytoplasmic loop" between TM3 and TM4, for loc␤ from parts of the 5Ј-translated region, for loc␣3 from the 3Ј-untranslated region, and for loc␣4 from the 5Ј-untranslated region. The isoform specificity of the RNA probes was confirmed by Southern analysis.
Whole embryos and head ganglia with attached optic lobes and retina from adult locusts were dissected out of the eggshell and individual adult insect, respectively, placed in embedding medium (Tissue Tec, Miles), and shock-frozen with dry ice. They were kept at Ϫ70°C until use. 6 -9-m frozen sections were obtained at Ϫ20°C using a Slee (Mainz, Germany) cryostat.
Cryosections were fixed with paraformaldehyde (4%) for 10 min, washed 4 times for 5 min with PBS supplemented with 0.1% Tween 20 (PBS-T), and incubated in hybridization buffer (50% formamide, 5ϫ SSC, 50 g/ml tRNA, 50 g/ml heparin, 0.1% Tween 20) for 1 h at 50°C. Digoxigenin-labeled RNA probes were added, and the cryosections were incubated overnight. After five washes for 20 min in 2ϫ SSC, 50% formamide, and treatment with RNase A (25 mg/ml) and RNase T1 in 2ϫ SSC, 8 further washes were performed in which the SSC buffer was diluted out with PBS-T in a stepwise fashion. After incubation for 1 h with PBS, supplemented with 2 mg/ml bovine serum albumin and 0.1% Triton X-100, the sections were incubated for 30 min with an antidigoxigenin antibody conjugated with alkaline phosphatase (Boehringer Mannheim, Germany). After two washes with 100 mM Tris-HCl, 150 mM NaCl, pH 7.5, the sections were incubated with 100 mM Tris-HCl, 100 mM NaCl, 50 mM MgCl 2 , pH 9.5, supplemented with the dye reagents nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phos-phate. After washing the sections for 5 min in PBS, they were covered in PBS, 50% glycerol and were analyzed under the microscope.
Southern and Northern Blot Analysis-10 g of chromosomal DNA prepared from muscle tissue of adult locusts was completely restricted with EcoRI, BamHI, and HindIII, fractionated, and blotted onto a nylon membrane. The genomic blot was hybridized with the randomly primed [␣-32 P]dATP-labeled AluI fragment and was washed under moderate stringency (65°C; 1ϫ SSC).
RNA was prepared from embryos d4 -d9 after egg laying. 1 mg of RNA was used to prepare poly(A) ϩ RNA. Every 5 g of poly(A) ϩ RNA of all six developmental stages was submitted to electrophoresis and blotted onto a nylon membrane. Blots were hybridized with three random primed [␣-32 P]dATP-labeled probes: a 420-bp HinfI/HinfI fragment of the loc␣2 cDNA, a 230-bp EcoRI/SphI fragment of the loc␤1 cDNA, and a 370-bp EcoRI/SphI fragment of the loc␣1 cDNA. Blots were washed with 1ϫ SSC, 65°C.
Fragments of nAChR ␣ Subunits Obtained by Expression in E. coli and Binding Studies with ␣-Bungarotoxin-cDNA clones were obtained from the following sources: Drosophila ␣1 from Marc Ballivet, Geneva, Switzerland (25); Drosophila ␣2 from Eckart Gundelfinger, Magdeburg, Germany (26); and Torpedo ␣ from Toni Claudio, New Haven, CT. cDNA fragments of the N-terminal extracellular regions were expressed either as maltose-binding protein fusions using pMAL-c2 vector from New England Biolabs (Schwalbach, Germany) or as glutathione S-transferase fusions using pGEX-4T-1 vector (Amersham Pharmacia Biotech). Fusion proteins were prepared from isopropyl-1-thio-␤-D-galactopyranoside-induced E. coli transformants (E. coli strain DH5␣, Life Technologies, Inc.) of flask submersed cultures. E. coli pellets were sonicated in lysis buffer, and after centrifugation, the supernatant was diluted and applied to affinity chromatography. Eluted nAChR fusion proteins were dialyzed, and the purity was determined by capillary electrophoresis and SDS-PAGE. Based on these methods, the purity of fusions proteins was generally better than 85%. In selected examples (Torpedo ␣ fragment), the fusion protein was proteolytically cleaved (factor Xa), and the nAChR fragment was isolated, and terminal pep- tide sequencing and mass spectrometry were performed. These data confirmed the primary structure of the nAChR fragments. Fusion proteins were separated on a, SDS-polyacrylamide gel, which was then blotted onto a nitrocellulose membrane. The membrane was incubated with 10 Ϫ9 M 125 I-␣BTX at 4°C, 30 min, and washing was performed 5 ϫ for 30 min with 1ϫ PBS. 125 I-␣BTX was obtained from Amersham Pharmacia Biotech (Braunschweig, Germany).

Preparation of Locust Neurons and Electrophysiological
Recordings-Head (supra-esophageal) ganglia and optic lobes from individual adult L. migratoria were dissected out and placed into dissociation solution (Sigma-Aldrich, Deisenhofen, Germany). Dispase (2 mg/ml, Life Technologies, Inc., Eggenstein, Germany) was added and incubated for 5 min at 37°C. The material was then centrifuged, and the pellet was resuspended in culture buffer and was dissociated by gentle aspiration with a fire-polished Pasteur pipette (27). Cells were plated onto glass coverslips that were pre-coated with concanavalin A (400 g/ml, Sigma) and laminin (4 g/ml, Sigma). The cultures were kept at room temperature and used for electrophysiological measurements on the following 2 days.
For electrophysiological recordings, the whole cell patch clamp technique was used. Microelectrodes were pulled from borosilicate glass capillaries (Hilgenberg, Malsfeld, Germany). The resistance of the firepolished pipettes was 4 -7 megohms, using the internal and external solutions described below. All experiments were performed at room temperature (22-25°C).
Currents were measured using an L/M-EPC-7 patch clamp amplifier (List, Darmstadt, Germany). The holding potential was Ϫ70 mV. Current records were low pass Bessel filtered at 315 Hz and digitized at 1-kHz sample rate. Data storage and analysis were performed with the pClamp version 6.03 software package (Axon Instruments, Foster City, CA). Test substances were applied to the cells using the U-tube reversed flow technique (28) with applications of 1-2 s duration at intervals of 1 min. Acetylcholine chloride, cytisine, and coniine were obtained from Sigma. Nicotine bitartrate was obtained from RBI (Natick, MA). Drugs were stored frozen as stock solutions (10 or 100 mM in water) and thawed and diluted on the day of the experiment.

Electrophysiological Studies Suggest Several Subtypes of
Neuronal nAChR in the Locust L. migratoria-The dissociated neurons obtained from head (supra-esophageal) ganglia and optic lobes of L. migratoria and cultured for 1-2 days on glass coverslips were large, round cells between 30 and 120 m in diameter, frequently having short protrusions. These cells invariably responded to test applications of 10 M acetylcholine (ACh) with a fast inward current of between 100 and 2000 pA at Ϫ70 mV clamp potential. Dose-response curves obtained with ACh on head ganglion cells (Fig. 1) yielded an EC 50 value of 17.9 Ϯ 3.4 M (n ϭ 7), whereas an EC 50 of 19.1 Ϯ 8.9 (n ϭ 4) was measured for optic lobe cells (not shown). In both cases, other nicotinic agonists such as (Ϫ)-nicotine or cytisine elicited responses that, even at saturating concentrations, remained well below the maximal currents induced by ACh.
The time course of the response to ACh was quite variable from one cell to another. In addition to the rapidly desensitizing component present in most cells, some cells also exhibited a non-desensitizing response to ACh, suggesting the existence of (at least) two nAChR subtypes. Further evidence was provided by dose-response curves for nicotine on cells in which one or the other type of ACh response predominated (Fig. 2). In a cell with mostly rapidly desensitizing ACh-induced currents, nicotine was a weak partial agonist with EC 50 Ͼ20 M (29). By contrast, another cell with non-desensitizing ACh currents responded sensitively to nicotine, with an EC 50 of Ͻ1 M. Since most cells contained both subtypes of nAChR in varying proportions, the effective EC 50 values for nicotine were quite variable (e.g. 9.4 Ϯ 11.4 M, n ϭ 5 in optic lobe cells). Similar results were obtained with cytisine, whereas coniine acted as agonist exclusively at the non-desensitizing nAChR subtype. Taken together, these data confirm the existence of functional nAChR ion channels in locust neurons and clearly suggest that there exist at least two pharmacologically distinct subtypes of nAChR in the cells studied.
The L. migratoria Genome Contains Several Genes Encoding for nAChR Subunits-A genomic library was prepared from muscle tissue of adult L. migratoria nAChR and was screened with a 32 P randomly labeled 650-bp long fragment from Drosophila ARD2 cDNA (23) encoding the transmembrane regions TM1-TM3. We obtained among other clones a 4.1-kilobase pair long genomic fragment that displayed 69.2% homology to the Drosophila ARD2 sequence, 81.6% to the Drosophila ALS sequence, and 75.9% to the chick ␣4 sequence. An AluI fragment (177 bp) of the genomic clone, containing mostly coding sequences (140 bp) and encoding the full second transmembrane The amino acid residues are numbered according to the Loc␣2 sequence (Fig. 4). M1-M4 refer to putative transmembrane domains. domain (TM2), the following short loop, and part of the third transmembrane domain (TM3), was prepared, randomly labeled with [ 32 P]dATP, and used as probe in a genomic blot of purified DNA from muscle tissue of adult locusts that was cleaved with the restriction endonucleases EcoRI, HindIII, and BamHI. As shown in Fig. 3, at least six bands were detected in each lane, suggesting that this number represents the approximate number of nAChR-coding genes that exist in L. migratoria.
Isolation and Sequencing of Five cDNAs Coding for nAChR Subunits-After initial information was obtained by Northern analysis (see later for details) for the developmental stages at which nAChR subunit mRNA is expressed in the locust, we prepared a cDNA library from late embryonic stage. By using the 32 P-labeled genomic AluI fragment described above, we screened the cDNA library at relatively high stringency. We obtained several cDNA clones which coded for four ␣ subunits and one ␤ subunit. The cDNA clones loc␣2, loc␣3, and loc␤1 encoded for mature subunits, whereas clone loc␣1 missed the nucleotides encoding for the first approximately 200 amino acids of the mature protein. All four cDNA clones missed the 5Ј-terminal sequences encoding the signal sequences. Another cDNA clone isolated encoded the full N-terminal sequence of an ␣ subunit (loc␣4) but extended only to the end of transmembrane domain 2. The complete sequence of loc␣1 was obtained by PCR. Attempts to also obtain by PCR the signal sequences of the four full-length cDNA clones were not successful. The mature proteins of L. migratoria Loc␣1-␣3 subunits consist of 559, 515, and 540 amino acids, respectively, with predicted molecular masses of 61.5, 56.7, and 59.4 kDa. The Loc␤ subunit consists of 497 amino acids with a molecular mass of 54.7 kDa. The nucleotide sequences of the four full-length clones have been reported to the EMBO nucleotide sequence Data Bank (accession numbers AJ000390 -000393). Since none of the five cDNAs was identical in sequence to the exon sequence of the genomic AluI fragment used as screening probe (homologies are 79 -89% for the ␣ clones, 64% for the ␤ clone), there must exist at least one more gene in L. migratoria that encodes nAChR subunits. Fig. 4 (upper panel), the hydropathy plot of the Loc␣2 isoform is representatively shown. As is typical of all isoforms of nAChR identified, it displays the pattern of four putative transmembrane domains (TM1-4) that is common for the superfamily of ligand-gated ion channels. The sequences coding for the ␣-isoforms also contain in their N-terminal extracellular domain the four conserved cysteines (Fig. 4, lower  panel), including the two vicinal ones just in front of the first transmembrane domain, and two putative N-glycosylation sites. The cytoplasmic domain between TM3 and TM4 is quite variable in size (199, 147, 185, and 156 amino acids in Loc␣1-3 and Loc␤1, respectively) and contains several putative phosphorylation sites. In the Loc␤1 sequence, this region contains a putative phosphorylation site for cAMP-dependent kinase, as was also found in non-␣ subunits from Torpedo (30) and Drosophila (23).

Properties of Sequences and Homology to Other Insect nAChRs-In
Sequence homologies of the four full-length clones of the L. , and ␤1 as studied by Northern blot analysis. 5 g of poly(A) ϩ RNA of each embryonic day (d4 -9) were submitted to electrophoresis and blotted onto a nylon membrane. Hybridization was performed with randomly 32 P-labeled cDNA fragments of clones loc␣1, loc␣2, and loc␤1. A, 420-bp long HinfI/HinfI fragment from loc␣2; B, 370-bp long EcoRI/SphI fragment from loc␣1; C, 230-bp long EcoRI/SphI fragment from loc␤1. Size markers were mouse rRNA (28 S and 18 S). All three blots show beginning subunit mRNA expression at day 6 after egg laying, with maximal expression on d8.
migratoria nAChR subunits to those of other species are reported in Fig. 5, top and bottom. Attempts to Express the L. migratoria nAChR cDNAs and mRNA in Xenopus Oocytes-Functional expression in Xenopus oocytes of insect nAChR following injection of mRNA into the cells has been reported for the cloned ␣ subunit from the locust Schistocerca gregaria (3). Functional channels were formed that were gated by micromolar concentrations of nicotine and that were blocked by ␣BTX, BTX, strychnine, and bicuculline. These results suggest that in Schistocerca there exists a homomeric nAChR that has similar physiological properties as the nicotinic responses recorded from insect neurons. In contrast, Drosophila nAChR ␣ subunits so far could only be co-expressed in Xenopus oocytes with the chick neuronal ␤2 subunit (31).
We have undertaken various attempts, so far in vain, to express the cloned L. migratoria nAChR subunits in Xenopus oocytes. To achieve expression, we introduced by silent mutagenesis a StuI restriction site at the 5Ј-end of each clone, and by using this site, we attached rat ␣3 signal sequence to the Locusta cDNA clones. We then injected into Xenopus oocytes single subunit cDNA and cDNA mixtures of different subunits or in vitro transcribed mRNA, neither treatment led to any significant channel activity. We also were unable to reproduce for any of the loc␣ clones the successful co-expression in Xenopus oocytes of Drosophila ␣ subunit with the vertebrate ␤2 clone (31). The inability of the cloned Locusta subunits to form functional channels in Xenopus oocytes may be due to inappropriate assembly of insect receptors in this ectopic expression system (2,19), or to missing or inappropriate post-translational modifications (32,33), or to additional as yet unidentified subunits that are required for assembly and/or channel function. That functional ion channels are formed from L. migratoria nAChR subunits was shown by Hanke and colleagues (18,34) who reported electrophysiological recordings from affinity purified and reconstituted nAChR protein.
The Loc␣3 nAChR Isoform Binds ␣-Bungarotoxin-Binding of the snake neurotoxin ␣BTX to L. migratoria nAChR isoforms was studied by Western blotting. For this purpose, fusion proteins between either maltose-binding protein or glutathione S-transferase and fragments of nAChR ␣ subunits were expressed in E. coli, and the purified fusion proteins were applied to Western blotting using as probe 125 I-labeled ␣BTX. As is representatively shown in Fig. 6, the Loc␣3 fusions (containing aa 126 -229 and aa 12-227, respectively) and (as controls) the Torpedo ␣ fusion (aa 1-246) and Drosophila ␣1 (ALS) fusion (aa 83-223) bound ␣BTX, whereas the Loc␣2 fusion (aa 1-222) and the Drosophila ␣2 fusion (aa 3-226) did not bind 125 I-␣BTX. To exclude the possibility of nonspecific binding of ␣BTX to the fusion partner, nAChR fragments were released by proteolytic cleavage from fusion proteins and were then tested in Western blots for toxin binding. These experiments (not shown) confirmed the above findings. In other Western blotting experiments (not shown), binding of ␣BTX to Loc␣3 (aa 12-290), Torpedo ␣1 (aa 143-201), rat ␣1 (aa 1-210), and rat ␣7 (aa 80 -213) was demonstrated. Selective binding of ␣BTX by Loc␣3, as compared with Loc␣2, cannot be explained by differences in sequence in the region around the two adjacent cysteines of ␣ subunits but rather appears to be due to attachment points for the toxin that are located in other sequence region(s). That the binding site for ␣BTX is discontinuously distributed within the N-terminal region of the ␣ subunit has previously been reported for nAChRs from other species (34).
From its ␣BTX binding properties, Loc␣3 is a candidate for a homo-oligomeric nAChR, such as the ␣7 subtype of mammals (11,35). As reported above though, we have been unable so far to functionally express this or other isoforms in Xenopus oocytes, as was achieved for other homo-oligomeric nAChR (13,36). We therefore do not know whether the Loc␣3 isoform indeed forms a functional channel with the typical properties of homo-oligomeric nAChRs, e.g. fast desensitization and sensitivity to epibatidine and choline (37).
Western blotting with the same fusions proteins was also performed with the antibody WF6 which competes with ACh and competitive agonists and antagonists (including ␣BTX) for binding to the Torpedo nAChR (34,38,39). Selective binding of WF6 was observed only to the Loc␣3 fusion proteins and (as controls) to those from Torpedo ␣ and D␣1 (data not shown). These observations agree with previous findings in that the attachment point patterns within the binding sites for ␣BTX and WF6 seem to be overlapping, albeit distinct (40).
Temporal and Spatial Expression of L. migratoria nAChR mRNA in the Developing and Adult Insect-Initial information on developmental stage-specific expression of L. migratoria mRNA was obtained by Northern analysis (Fig. 7). We employed randomly labeled 32 P-cDNA probes that were selected on the basis of minimal sequence homology between each other and that did not cross-react with the other cDNA clones, as was tested by Southern analysis (not shown). As is representatively shown for loc␣1, loc␣2, and loc␤1 in Fig. 7, the nAChR subunit RNAs begin to be expressed in late embryonic development, i.e. around d7 after egg laying and 3 days before hatching of these hemi-metabolic insects. Expression increased and reached a maximal level approximately 1 day before hatching. Northern blots probed with the ␤ cDNA probe always showed two transcripts (6.1 and 4.1 kilobase pairs), suggesting cross-hybridization with a second (as yet unidentified) ␤ subunit mRNA.
Temporal and spatial expression of L. migratoria nAChR mRNA was studied in further detail by in situ hybridization with subunit-specific RNA probes of frozen sections of 6 -9 m thickness. In the scheme of Fig. 8A, the areas are indicated in which nAChR mRNA was detected in the adult locust; these are the two-paired head ganglia, called mushroom bodies, the paired optic lobes, and the retina. The mushroom bodies are composed of large neurons, the cell bodies of which are located peripherally, whereas the neurites form the central neuropil. The cell bodies of the optic lobe neurons are clustered in a structure that is close to the retina (see also Fig. 9A).
As shown in Fig. 8, B-F, mRNA encoding loc␣1-3 and loc␤1, but not loc␣4, was detected in the mushroom bodies of the adult locust. The strongest expression was observed for loc␣1 mRNA, in the cytosol of neurons whose cell bodies are located in the center of the mushroom bodies. In roughly the same area, loc␣3 and loc␤1 mRNA was detected. loc␣3 mRNA was expressed to a much lower extent than loc␣1 or loc␤1. loc␣4 mRNA was not detected in mushroom bodies. loc␣2 mRNA was more abundant in peripherally located cell somata. Specificity of the hybridizations is indicated by the cellular staining pattern which was confined to the cytosol and did not include the cell nuclei (see inset of Fig. 8C). As a general observation, the neurons containing Locusta nAChR mRNA were intermingled with unlabeled nerve cells, suggesting that not all neurons of the locust express nicotinic receptors. The similar expression patterns of loc␣1, loc␣3, and loc␤1 suggest that these nAChR subunits may form a single (consisting of all three subunits) or two separate (a homo-oligomeric and a hetero-oligomeric one) receptor subtypes.
Expression of nAChR isoforms in the retina and optic lobes is shown in Fig. 9. In the retina, loc␣1 mRNA was most abundant, followed by loc␤1 and loc␣3 mRNA. loc␣2 and loc␣4 mRNA were absent. Transcripts of the three subunits were located in primary and secondary pigment cells which surround the crystal cones. Whereas loc␣4 mRNA was absent in the retina (Fig.  9F), it was detected in the ganglionic cell bodies of the lamina ganglionaris of the optic lobes where it was located in cell bodies that are situated close to the fibrous part of the optic lobe (Fig. 9G). As concluded above from the expression patterns in the mushroom bodies, Loc␣1, Loc␣3 and Loc␤1 may form a single or two separate nAChR subtypes.
As demonstrated in Fig. 10, the expression patterns of nAChR isoforms in the d8 locust embryo are quite different from those in the adult locust. Whereas loc␤1 mRNA was abundantly expressed in the protocerebral lobe, the developing mushroom body (Fig. 10F), the ␣ subunits were not expressed in this area, to any comparable extent. This suggests that the medulla interna; o, ommatidium; pc, photoreceptor cell; ppc, primary pigment cell; spc, secondary pigment cell. Bars, 20 m. In the areas depicted, mRNA coding for the subunits ␣1, ␣3, and ␤1 was detected in primary and secondary pigment cells of single ommatidia in the retina. In contrast, ␣4 mRNA was expressed in ganglionic cells of the optic lobe. Dashed lines in G indicate the approximate borders between ganglionic cell bodies and the lamina ganglionaris. Inset, ␣4 mRNA was located in the perinuclear region of ganglionic cell bodies.  (G and H). The abbreviations used are: np, neuropil; pl, protocerebral lobe; e, ectoderm; ola, optic lobe anlage; se, somatogastric epithelium; gcb, ganglionic developing mushroom bodies of embryonic day 8 do not yet express functional nAChR and that pre-expression of a structural subunit may be required for functional assembly of heteromeric locust nAChRs. loc␤1 mRNA was also detected in the ectoderm anlage from which the eye develops (Fig. 10F).
Outside of the head area, loc␣3 mRNA was mainly detected in isolated cells interspersed in the somatogastric epithelium (Fig. 10G). Additional limited expression of loc␣3 mRNA was observed in peripheral ganglion cells distributed in the mesenchyme below the somatogastric epithelium. In contrast, loc␤1 mRNA was exclusively found in peripheral ganglionic cells (Fig. 10H).

DISCUSSION
At Least Six nAChR Subunits Are Expressed in the Locust-They form several functionally distinct subtypes. In the course of this study, we have identified in the locust L. migratoria six nAChR subunit genes. Three cDNAs encoding ␣ subunits and one cDNA encoding a ␤ subunit were obtained as full-length clones. In addition we have isolated a partial nAChR ␣ subunit cDNA and have identified in Northern blots (Fig. 3) a sixth mRNA which probably encodes another ␤ subunit. This is the largest number of nAChR subunit genes so far identified in an invertebrate species, and our results clearly contradict previous suggestions of a single homomeric nAChR in L. migratoria (17). In SDS-PAGE of Torpedo nAChR, the four subunits are better separated than expected from their differences in molecular mass (between 50.2 and 57.6 kDa) (50). In contrast, the gel pattern of the affinity purified Locusta nAChR protein suggested a single polypeptide of approximately 65 kDa (17, whereas the molecular weights calculated from the amino acid sequences vary by approximately the same amount (between 54.7 and 61.5 kDa) as the Torpedo subunits. These differences in SDS-PAGE resolution are probably due to variations in the levels of posttranslational modifications.
Our cloning data clearly suggest that several subtypes of nicotinic receptors exist in the locust. These findings are supported by electrophysiological recordings from head ganglia neurons ( Figs. 1 and 2) which indicate at least two pharmacologically distinct nAChR subtypes (9).
Based on Northern blot and in situ hybridization studies (Figs. 7-10), all six of the identified nAChR subunit genes are expressed. From the local positions and sizes of cells in the mushroom body area in which nAChR-mRNA was detected, expression is probably confined to neurons. This is in contrast to an immunohistochemical study of nAChR expression in the locust Schistocerca gregaria (41) which suggested expression also in glia cells. The general staining pattern with the antibody used in that study is consistent with the present study in that all principal neuropils were positive.
The spatial expression patterns of the Locusta nAChR mRNAs are consistent with the existence of several receptor subtypes. Loc␣1, Loc␣3, and Loc␤1 are the isoforms most abundantly expressed in the head ganglia and the retina, whereas Loc␣4 is mainly expressed in optic lobe ganglionic cells. Based on its ␣BTX binding properties (Fig. 6), Loc␣3 is a candidate for a homomeric nAChR. Extending this line of arguments, Loc␣1 and Loc␤1 may form the predominant hetero-oligomeric nAChR of the locust. In the absence of functional expression studies in Xenopus oocytes or other ectopic expression systems, it cannot be excluded though that Loc␣1, Loc␣3, and Loc␤1 together form a heteromeric nAChR (42)(43)(44)(45). The early pres-ence of Loc␤1 in the embryo of the locust suggests that its expression may be a prerequisite for functional assembly of some hetero-oligomeric nAChR subtypes. loc␣2 and loc␣4 mRNA are much less abundant and have distinctly different expression patterns, suggesting that they do not form subunits of nAChR subtypes containing ␣1, ␣3, and ␤1 subunits. To further substantiate the subunit composition of locust nAChR subtypes, immunoisolation of these nAChR using subunit-specific antisera is suggested.
Comparison of the deduced amino acid sequences of the locust nAChR subunits with those of other insect species indicate distinct subfamilies of nAChR isoforms as follows: 1) Loc␣1 and MARA1 (and the ␣1 subunit of the lepidopteran Heliothis virescens which was recently cloned in our laboratory) 2 ; 2) Loc␣2 and SBD; 3) Loc␣3, ALS, and Mp␣2; 4) Loc␤1 and ARD; and 5) SAD and ␣L1. In a phylogenetic tree (Fig. 5) that was constructed on the basis of these data, each of the five subfamilies conforms to a separate evolutionary branch. Taken together, these data argue against simplified evolutionary considerations such as that because Lipidoptera and Diptera apparently are more closely related to each other than to Orthoptera, so must be their proteins and nucleic acids. Similarly, the present data do not suffice to group the nAChR isoforms according to their physiological function(s) or pharmacological profiles.
Recent studies from several laboratories (47) 3 have identified nicotinic acetylcholine receptors to be involved in cellular activities other than neurotransmission. Thus, the human ␣3 nAChR expressed in skin keratinocytes has been reported to regulate cell adhesion and motility. The expression of Loc␣3 in cells of the somatogastric epithelium may suggest similar nonneuronal activity.
Expression of nAChR Subunits in the Locust and in Ectopic Expression Systems-Transcription of all L. migratoria nAChR subunits begins at about the same time in late embryonic development (3 days before hatching) and remains throughout adulthood. In Drosophila, transcripts of the ␣ and non-␣ genes were detected beginning from mid-aged embryos, and the levels of expression were also developmentally regulated, as observed here for the locust (for a review, see Ref. 2). As already discussed above, the spatial and temporal expression patterns suggest the formation of several functionally distinct nAChR subtypes.
The availability of four full-length cDNA clones (three ␣ and one ␤ subunit) with attached signal sequence of a vertebrate neuronal nAChR should offer excellent conditions for the functional expression of the Locusta nAChR subtypes. Unfortunately, our attempts to express single subunit cDNAs or mixtures of subunit DNAs in the frog oocytes so far were not successful, even though these studies were performed in collaboration with research groups that are well experienced in this area (31,48,49). Our unsuccessful attempts reinforce previous findings that expression of insect nicotinic receptors in Xenopus oocytes is very difficult to achieve (19). This may be due to inappropriate assembly of insect receptors in this ectopic expression system (2,19), or to missing or inappropriate posttranslational modifications, or to the existence of as yet uni-dentified (structurally unique) additional subunit(s). We presently attempt to functionally express Locusta cDNAs in a mammalian cell line in which vertebrate neuronal nAChRs have been successfully expressed (46). Alternatively, should insect receptors not well express in vertebrate expression systems, the locust receptors may be transiently expressed in Drosophila S2 cells. Functional expression of insect nAChRs in cell systems that are suited for electrophysiological studies will be essential for the elucidation of the molecular basis for the peculiar pharmacology of some insect receptors (9,10,17). Based on the present study, the unusual pharmacology of the locust nAChR may be brought about by the presence of several receptor subtypes in the same cell. Elucidation of the subunit compositions of functional locust nAChRs and of their subtypespecific pharmacology will therefore remain pressing tasks. In addition, such studies might foster the development of novel insecticides.