Differential and inefficient splicing of a broadly expressed Drosophila erect wing transcript results in tissue-specific enrichment of the vital EWG protein isoform

Abstract

In this report, we document an unusual mode of tissue-enriched gene expression that is primarily mediated by alternative and inefficient splicing. We have analyzed posttranscriptional regulation of the Drosophila erect wing gene, which provides a vital neuronal function and is essential for the formation of certain muscles. Its predominant protein product, the 116-kDa EWG protein, a putative transcriptional regulator, can provide all known erect wing-associated functions. Moreover, consistent with its function, the 116-kDa protein is highly enriched in neurons and is also observed transiently in migrating myoblasts. In contrast to the protein distribution, we observed that erect wing transcripts are present in comparable levels in neuron-enriched heads and neuron-poor bodies of adult Drosophila. Our analyses shows that erect wing transcript consists of 10 exons and is alternatively spliced and that a subset of introns are inefficiently spliced. We also show that the 116-kDa EWG protein-encoding splice isoform is head enriched. In contrast, bodies have lower levels of transcripts that can encode the 116-kDa protein and greater amounts of unprocessed erect wing RNA. Thus, the enrichment of the 116-kDa protein in heads is ensured by tissue-specific alternative and inefficient splicing and not by transcriptional regulation. Furthermore, this regulation is biologically important, as an increased level of the 116-kDa protein outside the nervous system is lethal.

FIG. 1

Schematic representation of ewg exons, cDNAs, and PCR primers. (A) Genomic map of the ewg locus and two cDNAs, MPA-1 and SC3, that share the SC3 ORF shown as filled boxes (adapted from reference 9). A map of all characterized ewg exons (A to J) and the nomenclature of alternative spliced introns is shown below the SC3 cDNA. Alternative splicing occurs only in introns 3 and 6. The new exons E and I are present within introns 3c and 6, respectively. (B) The primers are named according to the intron excision events that they were used to assess; for example, 2F and 2R amplify transcripts that span intron 2. Primers designated In were used to amplify intron-containing transcripts; all In primers with the exception of In1R and In3cF are within introns. Note that introns 1 and 6 are not drawn to scale. (C) Sequence of the 74-bp ewg exon. The underlined nucleotide T is a silent nucleotide polymorphism, as a C in this position was found by genomic sequencing. (D) Sequence of the body-enriched 38-bp exon I.

FIG. 2

Characterization and comparison of ewg splicing in wild-type adult tissues. All RT-PCR assays were carried out with DNase I-treated total RNA isolated from 2-day-old heads (H) or bodies (B). The italicized letters below each pair of lanes represent the specific splice events as outlined in Fig. 1B, e.g., primers 3aF and 3aR for intron 3a splicing. These data are summarized in Table 2, which also lists the lengths of PCR products. rp49 transcripts were used as a control. Molecular size markers (GIBCO-BRL) are shown in lanes 11 and 16. Note that splicing of introns 3a, 3c, and 6 in heads is mostly in the mode of the SC3 cDNA.

FIG. 3

Abundance of ewg transcripts is independent of polyadenylation and is equal in heads (H) and bodies (B). (A) RT-PCR using random-primed or gene-specific-primed cDNAs. The RT reaction shown in lanes 1 to 4 was primed with a primer in exon H, an exon common to all ewg transcripts. The RT reaction shown in lanes 6 to 9 was primed with random hexamers. The italicized letters below each pair of lanes represent the specific splice events assayed as outlined in Fig. 1B. The uppermost bands in lanes 2, 4, 7, and 9 show ewg transcripts containing intron 3c. The 408-bp band in lanes 2 and 7 contains exon E. Molecular size markers (GIBCO-BRL) are shown in lane 5. (B and C) Cycle titration of PCRs using primers 4F and 5R to amplify parts of ewg transcripts common to all ewg transcripts. cDNAs were synthesized with a gene-specific primer in exon H. Aliquots were removed from the PCR beginning at cycle 18 and continuing until cycle 28. Quantitation of bands reveals that PCR is in the linear range (data not shown). Note that the intensities of bands in both heads and bodies are similar at all cycles.

FIG. 4

Alternative splicing of ewg transcripts is restricted to introns 3 and 6, alternative splice events are independent of each other, and splicing of introns 1, 3, and 6 is inefficient. Comparison was done between head (H) and body (B) poly(A)⁺ RNAs. (A) PCR products spanning introns 3 and 6 reveal all combinations of alternatively spliced introns. The italicized letters below each pair of lanes show the amplified section of ewg transcripts. Table 3 provides a complete listing of PCR products and summarizes the data. The forward primer was 3aF or 3cF, the return primer was RV, 6R, or 38R. Molecular size markers (GIBCO-BRL) are shown in lanes 5 and 11. In lane 8, a lambda hindIII digest was used as a marker. Note that heads and bodies show differences in the ewg transcript population and abundance due to the body-enriched usage of exons E and I, while increased inclusion of exon D occurs in heads. (B and C) Primer pairs spanning several introns reveal no additional alternative splice events. The italicized letters below each pair of lanes show the amplified section of ewg transcripts. Table 3 provides a complete listing of PCR products and summarizes the data. Molecular size markers (GIBCO-BRL) are shown in lanes 9 and 18 of panel B and lanes 1 and 8 of panel C. In lane 15 of panel C, a lambda hindIII digest was used as a marker. (D) Introns 1, 3a, 3c, and 6 are retained in polyadenylated ewg transcripts. A listing of PCR products and summary of the data are shown in Table 2. Note that differences between heads and bodies are mainly detected in the retention of introns 1 and 6 but not 3. Molecular size markers (GIBCO-BRL) are shown in lanes 5 and 13.

FIG. 5

ORFs of ewg splice isoforms deduced by RT-PCR analysis. All deduced ewg RNA isoforms are outlined from data presented in Table 2. The size of each ORF is given in amino acids (aa). Transcripts including exon E and I are not shown since the ORF terminates in exon F. Also, ORFs resulting from intron-retaining transcripts are not shown.

FIG. 6

Characterization of anti-EWG antibody and immunoblot analysis of EWG proteins in Drosophila embryos and heads. (A) Anti-EWG antibody recognizes epitopes all along the EWG protein. Different parts of EWG protein were expressed in E. coli and tagged with either six histidines (6His) or MBP at the N terminus. As controls either bacterial extracts expressing only MBP or bacterial extracts from uninduced E. coli (lanes −) were loaded. Note that the part encoded by exons B and C has a calculated molecular mass of 43 kDa but runs at ∼70 kDa. Molecular mass markers are shown on the left in kilodaltons. (B) Anti-EWG antibody recognizes one major and several minor proteins in wild-type embryos (+; lane 1), which are absent in ewg^l1 embryos (lane 2). Proteins marked with arrows are degradation products of the 116-kDa EWG protein, since they are also present in ewg^l1; EWG^NS4 embryos (NS4; lane 3). EWG^NS4 is a rescue construct where a cDNA for the 116-kDa EWG protein is fused to a elav promoter fragment restricting expression to the nervous system. Molecular mass markers are shown on the left in kilodaltons. (C) Basal expression of the 116-kDa EWG protein under a heat shock promoter is comparable to EWG levels in the wild type. Head extracts from wild-type females (Tb [+/+; +/+; +/Tb]; lane 1) are compared to head extracts from females with either one copy (HS1 [ewg^l1/ewg^l1; +/+; EWG^HS1/+] [lane 2] and HS7, Tb [ewg^l1/ewg^l1; EWG^HS7/+; +/Tb] [lane 3]) or two copies (HS1, HS7 [ewg^l1/ewg^l1; EWG^HS7/+; EWG^HS1/Tb] [lane 4]). Note that the amount of EWG protein is about half of the wild-type amount with only one copy. Molecular mass markers are shown on the left in kilodaltons. (D) The 40-kDa proteins recognized by the anti-EWG antibody in head extracts are not EWG isoforms. Head extracts from the wild type (+; lane 1) are compared to head extracts from an ewg^l1; EWG^NS4 mutant (NS4; lane 2). Molecular mass markers are shown on the left in kilodaltons.