The Drosophila gene CheB42a is a novel modifier of Deg/ENaC channel function

Abstract

Degenerin/epithelial Na(+) channels (DEG/ENaC) represent a diverse family of voltage-insensitive cation channels whose functions include Na(+) transport across epithelia, mechanosensation, nociception, salt sensing, modification of neurotransmission, and detecting the neurotransmitter FMRFamide. We previously showed that the Drosophila melanogaster Deg/ENaC gene lounge lizard (llz) is co-transcribed in an operon-like locus with another gene of unknown function, CheB42a. Because operons often encode proteins in the same biochemical or physiological pathway, we hypothesized that CHEB42A and LLZ might function together. Consistent with this hypothesis, we found both genes expressed in cells previously implicated in sensory functions during male courtship. Furthermore, when coexpressed, LLZ coprecipitated with CHEB42A, suggesting that the two proteins form a complex. Although LLZ expressed either alone or with CHEB42A did not generate ion channel currents, CHEB42A increased current amplitude of another DEG/ENaC protein whose ligand (protons) is known, acid-sensing ion channel 1a (ASIC1a). We also found that CHEB42A was cleaved to generate a secreted protein, suggesting that CHEB42A may play an important role in the extracellular space. These data suggest that CHEB42A is a modulatory subunit for sensory-related Deg/ENaC signaling. These results are consistent with operon-like transcription of CheB42a and llz and explain the similar contributions of these genes to courtship behavior.

Figure 1. CheB42a encodes a protein with sequence similarity to aryl sulfotransferase 1A.

Top line shows predicted aminoa acid sequence of Drosophila melanogaster CheB42a, and second and third lines show predicted amino acid sequences of Ornithorhynchus anatinus (platypus) and Homo sapiens (human) aryl sulfotransferase. Green, conserved residues; yellow, conserved residues in two of the aligned proteins; blue, similar residues. Accession numbers for the aligned proteins are: CheB42a, NM_206043.2; O_anatinus, NP_001121091; H_sapiens, NP_003157.1.

Figure 2. The CheB42a/llz locus is preferentially expressed in non-neuronal sensory structures.

A–B. Driving expression of GFP with the CheB42a/llz promoter revealed expression in cells in the front legs of males. Co-expression of GFP and DsRed containing a nuclear localization signal revealed expression in relatively large cells with a single labeled nucleus. Morphology of the CheB42a/llz-expressing cells suggested the cells might enwrap another cell type; white dotted line represents boundaries of a potentially enwrapped cell with neuronal morphology (see Movies S1 and S2 for 3-D reconstruction of typical CheB42a/llz cells). C. Pan-neuronal elav promoter driving GFP expression in the same male leg segment as shown in panel A. Typical neuronal sensory morphologies such as sensory cilia extending to the sensory hairs are apparent.

Figure 3. CheB42a and llz are expressed in the same chemosensory-related cells.

A. The expression levels of both CheB42a and llz were reduced in homozygous Poxn mutants relative to heterozygote controls. Total RNA was extracted from male appendages and was analyzed by quantitative RT-PCR. Data are mRNA in Poxn relative to wild-type control. Expression levels were normalized to the housekeeping gene rp49. y-axis represents arbitrary mRNA fold-difference units, with control expression levels designated as 1 unit. B. Genetic ablationCheB42a/llz-expressing cells reduced expression of both transcripts. The CheB42a promoter-GAL4 line was crossed to UAS-rpr, which induces cell-death . Analysis as in panel A.

Figure 4. Genetic ablation of CheB42a/llz expressing cells does not affect male courtship behavior.

A. Expression of the proapoptotic genes hid and rpr in CheB42a/llz-expressing cells using the UAS-GAL4 system did not affect male courtship latency (time from female introduction until the male shows any courtship related behaviors). Parental lines crossed to the reciprocal wild-type genetic background were used as controls. Although there were significant differences between the two parental controls, neither control line was statistically different from the focal cross (ANOVA). N  =  (Che/llz-GAL4 x yw, 33; Che/llz-GAL4 x UAS-rpr, hid, 52; w¹¹¹⁸ x UAS-rpr, hid, 27). B. Genetic ablation of Che/llz-expressing cells resulted in a mild effect on male courtship index (the proportion of time a male spends courting once courting started). Genotypes and sample sizes as in A. *, P<0.05.

Figure 5. CHEB42A and LLZ can form a protein complex.

A. Schematic of transfected proteins. B. Co-expression of tagged CHEB42A-HA and LLZ-GFP proteins. Immunoprecipitation with anti-HA antibody co-precipitated LLZ. “mix” indicates an experiment in which equal amounts of protein from the singly-transfected cells were mixed prior to immunoprecipitation as a control for non-specific interactions. Expression of llz produced two protein bands, the expected lower molecular mass band, plus a band of higher molecular mass. We do not know the identity of the more slowly migrating band; it might represent an LLZ-containing multiunit complex that is resistant to SDS denaturation or might represent post-translational modifications of LLZ . We also attempted to detect HA-CHEB42A after immunoprecipitating LLZ-GFP, but were not successful. C. CHEB42A did not affect LLZ surface expression. COS7 cells were transfected with llz and either GFP or CheB42a. Surface expression level was estimated with biotinylation of surface proteins followed by neutravidin precipitation. To estimate total protein, protein input was directly blotted with anti-GFP antibody (right panel). We did not observe the larger LLZ' band in the surface expression study. We speculate that differences in sample processing protocols between the IP and surface expression studies could affect solubility and detection of larger protein complexes.

Figure 6. CHEB42A increases current from the mammalian DEG/ENaC protein ASIC1a.

A–C. Effect of CHEB42A on H⁺-gated ASIC1a currents. Xenopus oocytes were injected with ASIC1a, GFP or CheB42a cDNAs alone or in combinations indicated. Bathing solution pH was reduced to 5 during time indicated. A. example of H⁺-gated currents from the ASIC1a channel. B. CHEB42A increases ASIC1a H⁺-gated currents. Asterisk indicates P<0.01 (two-tail t-test). Number of oocytes studied: 9 GFP, 9 CheB42a and GFP, 20 ASIC1a and GFP, and 27 ASIC1a and CheB42a. Data are mean ± SEM. C. CHEB42A did not affect the pH dependence of ASIC1a current. Closed symbols indicate ASIC1a and CHEB42A (n = 15) and open symbols indicate ASICla + GFP (n = 12). D. CHEB42A did not affect ASIC1a surface expression. Experiment was done as in Fig. 5B.

Figure 7. CHEB42A produces a secreted peptide.

A. Schematic of the construct used to study CHEB42A. CheB42a cDNA was tagged with GFP at the 5′ end and 3xHA tags at the 3′ end and was cloned into the mammalian expression vector pcDNA3.1 (Invitrogen). TM, transmembrane peptide as predicted by the SignalP 3.0 Server (http://www.cbs.dtu.dk/services/SignalP/). predicted transmembrane domain. Predicted signal peptide cleavage site. T, N, and C represent the predicted full length and cleaved products. B. CHEB42A is a secreted protein. Cells were transfected with tagged cDNA (CheB42a) or pcDNA3.1 as a control. Proteins from cell lysates or cell media were immunoprecipitated (IP) with either anti-GFP (G) or anti-HA (H) antibodies, and then blotted for either tag. T, N, and C indicate the predicted peptides from Fig. 7A. C. Effects of CheB42a-conditioned media on ASIC1a currents. ASIC1a was expressed in Xenopus oocytes as in Fig. 6. Four ASIC1a-expressing oocytes were tested for pH-dependent currents with and without conditioned media from CheB42a-expressing cells. The same oocytes (N = 4) were stimulated with unconditioned medium (pH 5.5), followed by stimulation with CheB42a-conditioned medium. There was no obvious effect of the conditioned media on pH-activation of ASIC1a channels. Currents were normalized in the conditioned state relative to the control media. All oocytes showed significant ASIC1a-dependent currents in response to pH 5.5, which does not elicit any currents in non-injected oocytes.