The Drosophila melanogaster phospholipid flippase dATP8B is required for odorant receptor function

Abstract

The olfactory systems of insects are fundamental to all aspects of their behaviour, and insect olfactory receptor neurons (ORNs) exhibit exquisite specificity and sensitivity to a wide range of environmental cues. In Drosophila melanogaster, ORN responses are determined by three different receptor families, the odorant (Or), ionotropic-like (IR) and gustatory (Gr) receptors. However, the precise mechanisms of signalling by these different receptor families are not fully understood. Here we report the unexpected finding that the type 4 P-type ATPase phospholipid transporter dATP8B, the homologue of a protein associated with intrahepatic cholestasis and hearing loss in humans, is crucial for Drosophila olfactory responses. Mutations in dATP8B severely attenuate sensitivity of odorant detection specifically in Or-expressing ORNs, but do not affect responses mediated by IR or Gr receptors. Accordingly, we find dATP8B to be expressed in ORNs and localised to the dendritic membrane of the olfactory neurons where signal transduction occurs. Localisation of Or proteins to the dendrites is unaffected in dATP8B mutants, as is dendrite morphology, suggesting instead that dATP8B is critical for Or signalling. As dATP8B is a member of the phospholipid flippase family of ATPases, which function to determine asymmetry in phospholipid composition between the outer and inner leaflets of plasma membranes, our findings suggest a requirement for phospholipid asymmetry in the signalling of a specific family of chemoreceptor proteins.

Figure 1. Mutations in dATP8B cause severe olfactory defects.

(A) Electroantennogram (EAG) and (B) electropalpogram (EPG) responses to a panel of odorants. Bars represent mean response ± SEM (n = 10), asterisks are significant differences. Responses of homozygous ll2 mutant flies (blue bars) are severely reduced when compared with controls (black bars) for all the tested odorants except CO₂ (t-test, Bonferroni, p<0.05). Heterozygote flies (grey bars) are not affected. (C) A deletion that removes dATP8B and a piggyBac element inserted in dATP8B both fail to complement the ll2 phenotype. Bars represent mean EAG responses ± SEM (n = 6–10). Trans-heterozygotes for ll2 and deficiency Df(3R)Exel8155 (blue bars) have reduced EAG response compared to controls (black bars, t-tests, Bonferroni, p<0.05), and homozygotes for the piggyBac insertion dATP8B^f05203 (red bars) and trans-heterozygotes for ll2 and dATP8B^f05203 (red/blue hatch) show similar reductions when compared to heterozygote controls (grey bars; t-tests, Bonferroni, p<0.05). (D) The mapped genomic interval for the ll2 mutant and the gene model of dATP8B (CG14741). The candidate region contained 14 annotated genes. The identified nonsense mutation in dATP8B and the insertion site of the piggyBac line dATP8B^f05203 are in the 1^st and the 10^th common coding exon respectively, affecting all the annotated transcripts. For isoforms RC, RF, RG, RH and RI the EMS mutation causes R18-X and for isoforms RB and RD the mutation causes R197-X. Coding exons are colored in orange and the 3′ and the 5′ UTR are in grey. (E) The olfactory defect in the dATP8B^f05203 line is reverted when the piggyBac insertion is precisely excised. EAG responses of homozygous dATP8B^f05203-Ex flies (red bars) to a panel of odorants were not different from controls (black bars). Bars represent mean response ± SEM (n = 5, t-test, Bonferroni). Odorants are: EA, ethyl acetate, PA, pentyl acetate, MS, methyl salicylate, OL, 1-octen-3-ol, HB, ethyl 3-hydroxybutanoate, EH, ethyl hexanoate, MH, 6-methyl-5-hepten-2-one, BZ, benzaldehyde, MP, 4-methylphenol, PO, paraffin oil (solvent blank).

Figure 2. dATP8B is required for responses of Or but not IR or Gr- expressing sensory neurons.

(A–C) Mutations in dATP8B reduce the sensitivity of Or-expressing neurons to their major ligands. Dose-response curves for neurons located in three different morphological types of sensilla are shown; (A) ab3B neurons in basiconic sensilla (Or85b) to 2-heptanone, (B) at1A neurons in trichoid sensilla (Or67d) to cis-vaccenyl acetate, (C) ac3B neurons in coeloconic sensilla (Or35a) to Z3-hexenol. In all cases sensitivity is significantly lowered for all doses in homozygous dATP8B^f05203 flies (red) compared to controls (black, t-test, Bonferroni, n = 6–9, recorded from 4–8 flies). (D–E) Mutations in dATP8B differentially alter responses of ab3A neurons (Or22a) to a major ligand ethyl hexanoate (D) and a minor ligand ethyl butanoate (E). (n = 7 sensilla, recorded from 5–6 flies). (F–I) Neurons expressing IR or Gr receptors are not affected by dATP8B mutations. Four different types of neurons are shown; (F) responses of ac2A neurons in coeloconic sensilla (IR41a and IR76b) to 1,4-diaminobutane, (G) responses of ac3A neurons in coeloconic sensilla (IR75a,b and c) to propionic acid, (H) responses of ab1C neurons in basiconic sensilla (Grs21a and 63a) to CO₂, (I) responses from single neurons in labellar taste sensilla expressing Gr genes to 100 mM sucrose (Suc) or 10 mM caffeine (Caff) (mean ± SEM). In all cases control and mutant responses are not significantly different (n = 6–10 sensilla from 3–5 flies, t-tests, Bonferroni). Controls are either wild type or heterozygous dATP8B^f05203 mutants.

Figure 3. dATP8B is expressed and required in Or-expressing olfactory receptor neurons.

(A–A′) dATP8B protein localises to the dendrites of ORNs. 14 µm thick antennal sections from wild type flies (CS-5) were stained for anti-dATP8B (magenta). Strong anti-dATP8B staining is seen in the shafts of the sensilla (the location of the outer dendrites) of basiconic (arrows) and trichoid (arrowhead) sensilla. Staining is also seen in the inner dendrites and cell bodies, however this staining is also present in dATP8B mutants. (B) The strong outer dendrite staining of anti-dATP8B is absent in dATP8B mutants (ll2/Df(3R)Exel8155), indicating it is specific for dATP8B. (C) dATP8B and Orco co-localise in the outer dendrites. 14 µm thick antennal sections from wild type flies were stained for anti-dATP8B (magenta) and Orco (green). (D) dATP8B is absent from the outer dendrites of Gr21a-expressing (ab1C) neurons. 14 µm thick antennal sections from Gr21a>mCD8:GFP flies were stained for anti-dATP8B (magenta) and anti-GFP (green). (E) In Orco-GAL4: UAS-dATP8B^RNAi flies (red bars) the EAG response is significantly reduced compared to controls (black and grey bars, t-test, Bonferroni, p<0.05) for some of the same odorants that are affected by the two dATP8B mutant alleles. Bars represent mean EAG responses ± SEM (n = 6–10). Odorants are: EA, ethyl acetate, PA, pentyl acetate, MS, methyl salicylate, OL, 1-octen-3-ol, HB, ethyl 3-hydroxybutanoate, EH, ethyl hexanoate, BZ, benzaldehyde, PO, paraffin oil (solvent blank).

Figure 4. Orco and Or22a localize normally to the dendrites in dATP8B mutants.

14 µm thick antennal sections from wild type flies (CS-5) were stained for anti-Orco or for anti-Or22a. No difference in either Orco or Or22a localisation to the outer dendrites was observed in dATP8B mutants (dATP8B^f05203) compared to control flies.