ELAV-Mediated 3′-End Processing of ewg Transcripts Is Evolutionarily Conserved Despite Sequence Degeneration of the ELAV-Binding Site

Regulation of alternative mRNA processing by ELAV (embryonic lethal abnormal visual system)/Hu proteins is mediated by binding to AU-rich elements of low complexity. Since such sequences diverge very rapidly during evolution, it has not been clear if ELAV regulation is maintained over extended phylogenetic distances. The transcription factor Erect wing (Ewg) is a major target of ELAV in Drosophila melanogaster and coordinates metabolic gene expression with regulation of synaptic plasticity. Here, we demonstrate evolutionary conservation of ELAV regulation of ewg despite massive degeneration of its binding site and of associated elements in the regulated intronic 3′-end processing site in distantly related Drosophila virilis. In this species, the RNA-binding part of ELAV protein is identical to D. melanogaster. ELAV expression as well as expression and regulation of ewg are also conserved. Using in vitro binding assays and in vivo transgene analysis, we demonstrate, however, that the ELAV-binding site of D. virilis is fully functional in regulating alternative splicing of ewg intron 6 in D. melanogaster. Known features of the ELAV-binding site, such as the requirement of multiple poly(U) motifs spread over an extended binding site of ∼150 nt and a higher affinity to the 3′ part of the binding site, are conserved. We further show that the 135-bp ELAV-binding site from D. melanogaster is sufficient for ELAV recruitment in vivo. Hence, our data suggest that ELAV/Hu protein-regulated alternative RNA processing is more conserved than anticipated from the alignment of degenerate low-complexity sequences. ALTERNATIVE mRNA processing is major mechanism to generate molecular diversity and organismal complexity from the limited number of genes present in higher eukaryotes. Through alternative splicing and polyadenylation, more than one mRNA can be generated from a single gene that differs in the encoded protein and/or alters expression or localization of the encoded protein post-transcriptionally (Matlin et al. 2005; Soller 2006; Chen and Manley 2009; Licatalosi and Darnell 2010). In humans, alternative splicing and polyadenylation occur in 92–94% and in >50% of genes, respectively, and are particularly abundant in the brain (Licatalosi and Darnell 2006; Li et al. 2007; Wang et al. 2008; Neilson and Sandberg 2010). Our understanding of the regulation of alternative mRNA processing, however, is limited. Since RNA-binding regulators are generally well conserved, but noncoding parts of pre-mRNAs that harbor regulatory sequences for its processing diverge very rapidly during evolution, it is not clear if and how evolutionary conservation is maintained at the sequence level. ELAV (embryonic lethal abnormal visual system)/Hu family proteins are prominent regulators of alternative mRNA processing in the brain and are widely used neuronal markers (Soller and White 2004; Hinman and Lou 2008). ELAV/Hu proteins are proto-type RNA-binding proteins that contain three RNA recognition motifs (RRMs). The RRMs of ELAV/Hu proteins, as in many other RRM-containing proteins, are highly conserved with identities of 52–82% between human Hu and Drosophila ELAV. Despite the high sequence conservation of ELAV/Hu family proteins, however, their number varies between different clades, suggesting either highly dynamic functions or redundancy among individual family members (Samson 2008). The founding member of this family of proteins, neuron-specific ELAV from Drosophila, has been shown to affect gene-specific alternative pre-mRNA processing of erect wing (ewg), neuroglian (nrg), and armadillo (arm) (Koushika et al. 2000). The neuronally expressed human homologs HuB-C and the ubiquitously expressed HuR were initially assigned cytoplasmic functions, but also regulate pre-mRNA processing (Antic et al. 1999; Kasashima et al. 1999; Brennan and Steitz 2001; Zhu et al. 2007, 2008). ELAV/Hu proteins preferentially bind to U-rich sequences that are abundant in introns and untranslated regions. Furthermore, ELAV/Hu proteins have been shown to multimerize, suggesting an important role of this feature in achieving target specificity in a complex cellular environment (Kasashima et al. 2002; Soller and White 2005; David et al. 2007; Toba and White 2008). The ewg gene encoding a transcriptional regulator homologous to human NRF-1 is a major target of ELAV in Drosophila as ewg transgenes can rescue post-embryonic development and viability of elav mutants (Haussmann et al. 2008). ELAV is required for splicing of the last intron 6 of ewg that results in EWG protein expression (Soller et al. 2008). In ewg intron 6, ELAV binds distal of a poly(A) site and inhibits 3′-end processing in vitro and in vivo (Soller and White 2003). At this regulated poly(A) site, ELAV-binding requires a number of short poly(U) motifs spread over an extended binding site of ∼135 nt (Soller and White 2005). ELAV forms a defined dodecameric complex in vitro, and the importance of complex formation for ewg intron 6 splicing is indicated by the requirement of multiple poly(U) motifs. These poly(U) motifs, however, can be variably positioned in the ELAV-binding site as indicated by the presence of deletions in very closely related species. Also, introduction of spacer sequences minimally affects ELAV regulation of ewg intron 6 splicing. These features make it unlikely that target-specific binding depends upon the formation of a higher-order RNA structure encompassing the ELAV-binding site (Soller and White 2005; Soller et al. 2010). The ewg gene integrates multiple signaling pathways (e.g., Notch, Wnt/wingless, TGF-β, and AP-1) in coordinating neuronal metabolism and synaptic plasticity (Haussmann et al. 2008; Haussmann and Soller 2010). Given the importance of ELAV-mediated regulation of ewg in Drosophila melanogaster, we were wondering if the mechanism of ELAV-dependent splicing of ewg intron 6 is evolutionarily conserved in the distantly related Drosophila virilis that separated ∼40–60 MYA, a phylogentic distance similar to mice and humans. Since the RNA-binding part of ELAV protein is identical in D. virilis ELAV (Yao and White 1991), we further anticipated gaining insights into the evolution of RNA-processing signals and the underlying mechanism governing gene-specific target RNA recognition by ELAV. Our analysis of ELAV and EWG expression shows evolutionary conservation in D. virilis since both ELAV and EWG proteins are restricted to neurons and ewg transcripts are broadly expressed as in D. melanogaster. Furthermore, we identified a functional poly(A) site in D. virilis at a similar position in the regulated intron 6 as in D. melanogaster and demonstrate binding of ELAV to the vicinity of this poly(A) site. By using reporter transgenes in D. melanogaster, we show that a 600-nt-long region containing the regulated poly(A) site from D. virilis provides full functionality and is regulated in an ELAV-dependent manner in neurons. Intriguingly, however, the ELAV-binding site in ewg intron 6 diverged such that it is not recognized with sequence alignment algorithms in a genomic context due to its low complexity. The ELAV-binding site in D. virilis ewg intron 6 extends over ∼150 nt, and as for the D. melanogaster sequence, multiple and spaced poly(U) motifs are required for binding and regulation of intron 6 splicing. Within this sequence the importance of the 3′ part for ELAV binding in vitro and splicing regulation in vivo is also conserved. In addition, our analysis demonstrates the flexibility of regulatory elements involved in 3′-end processing since the distance of the cleavage site relative to the poly(A) site recognition element (AAUAAA) is not conserved. Given the massive sequence degeneration of the ELAV-binding site and the redundancy of ELAV-binding motifs at a genomic scale, it has not been clear if the 135-bp sequence identified from D. melanogaster is sufficient to recruit ELAV for binding in the natural context of Drosophila neurons. Using a number of reporter transgenes, we demonstrate that the choice of the promoter has no role in the recruitment of ELAV to its binding site in ewg and that the 135-bp sequence from D. melanogaster is sufficient for ELAV recruitment in a heterologous context. Our data demonstrate evolutionary conservation of pre-mRNA processing by ELAV/Hu proteins that are mediated by degenerate low-complexity sequences.