Adenosine Deaminases Acting on RNA (ADARs) function to co-transcriptionally deaminate specific (or non-specific) adenosines to inosines within pre-mRNAs, using double-stranded RNAs as substrate. In both Drosophila and mammals, the best-studied ADAR functions are to catalyze specific nucleotide conversions within mRNAs encoding various ligand- or voltage-gated ion channel proteins within the adult brain. In contrast, ADARs within developing fly embryos have scarcely been studied, in part because they contain little or no editase activity, raising interesting questions as to their functional significance. Quantitative RT-PCR shows that two major developmentally regulated mRNA isoform classes are produced (full-length and truncated), which arise by alternative splicing and also alternative 3'-end formation. In situ localization of specific dADAR mRNA isoforms during embryogenesis reveals that the full-length class is found primarily within the developing germ band and central nervous system, whereas the truncated isoform is mostly located in gut endothelium. Developmental Western immunoblots show that both isoform classes are expressed into protein during embryogenesis. Both the rnp-4f 5'-UTR unspliced isoform and the full-length dADAR mRNA primarily localize in the embryonic germ band and subsequently throughout the developing central nervous system. Previous studies have shown that some rnp-4f pre-mRNAs are extensively edited by dADAR in the adult brain. Computer predictions suggest that intron-exon pairing promotes formation of an evolutionarily conserved secondary structure in the rnp-4f 5'-UTR, forming a 177-nt RNA duplex resembling an editing site complementary sequence, which is shown to be associated with splicing failure and to generate a long isoform. Taken together, these observations led us to explore the possibility that interaction between rnp-4f pre-mRNA and nuclear full-length dADAR protein may occur during embryogenesis. In dADAR null mutants, rnp-4f 5'-UTR alternative splicing is significantly diminished, suggesting a non-catalytic role for dADAR in splicing regulation. A working model is proposed which provides a possible molecular mechanism.