Functional study of regulatory elements in CDK5R1 3’-UTR: evidence on post-transcriptional gene expression regulation

Background CDK5R1 encodes for p35, a neuron-specific activator of cyclin-dependent kinase 5 (CDK5), whose activity plays a central role in neuronal migration during central nervous system development. Cdk5r KO mice have severe cortical lamination defects and suffer from adult mortality and seizures. The active CDK5-p35 complex is involved in several processes required for central nervous system development and functioning, as axonal regeneration, cellular differentiation, neuronal apoptosis, learning and memory processes, synaptic transmission and membrane trafficking during the outgrowth of neuronal processes. Moreover, increased CDK5 activation by p25, a proteolytic fragment containing the C-terminal portion of p35, has been implicated in the pathogenesis of several neurodegenerative disorders, such as Alzheimer’s disease, Parkinson’s disease and amyotrophic lateral sclerosis. CDK5R1 has been proposed as a candidate gene for mental retardation susceptibility in NF1 microdeletion syndrome. CDK5R1 spans for 4.17 kb on chr 17q11.2, and its coding region (1021 bp) consists of a single exon. In addition, the CDK5R1 gene displays a large 3’-untranslated region (3’-UTR). The remarkable size of CDK5R1 3’-UTR suggests a role in post-transcriptional regulation of CDK5R1 expression. Untranslated regions (UTRs) are known to play crucial roles in the post-transcriptional regulation of gene expression, including modulation of the transport of mRNA out of the nucleus, and of the translation efficiency, subcellular localization and stability. The importance of 3’-UTRs in regulating gene expression is underlined by the finding that mutations which alter the 3’-UTR can lead to serious pathology. Nucleotide patterns or motifs located in 3' UTRs can interact with specific RNA-binding proteins. The biological activity of regulatory motifs at the RNA level relies on a combination of primary and secondary structure. Interactions between sequence elements located in the 3’-UTRs and specific microRNAs have also been shown to play key regulatory roles. Results The bioinformatic study shows a high conservation degree in mammals and predicts several AU-Rich Elements (AREs) and a GY-box element. The GY−box (GTCTTCC) motif, described in many 3'−UTRs of genes involved in Notch signalling in Drosophila is likely to be involved in the formation of RNA duplexes with complementary sequences at the 5' ends of some Drosophila microRNAs in vivo. AU-rich sequences, function as potent destabilizing elements that cause rapid decay of the respective transcript; these elements are composed of a variable number of copies of the AUUUA pentamer or UUAUUUAUU nonamer. Among the predicted AREs in CDK5R1, the nt 2659-2671 ARE shows complete identity to the consensus sequence for Class I AREs, according to the ARED 3.0 definition, is highly conserved in mammals and zebrafish and is predicted to be accessibile to the binding of trans-acting factors. The effect of the 3’-UTR on gene expression was studied with the Dual-Luciferase reporter assay. The insertion of CDK5R1 3’-UTR into luciferase 3’-UTR caused a decreased luciferase activity and mRNA level in four transfected cell lines (SK-N-BE, SH-SY5Y, HEK-293 and MCF-7). The dissection of 3’-UTR into 6 fragments (C1-6), each containing at least one predicted regulatory element, allowed us to investigate the potential role of each region. All the chimeric constructs showed, in most of the studied cell lines, a general decrease of luciferase activity. In most cases these effects are likely to involve transcript stability rather than translational repression mechanisms, since reduced reporter activity levels corresponded to reduced mRNA levels. A region (C2), leading to a very strong mRNA destabilization, showed a significantly low half-life, indicating an accelerated mRNA degradation. This fragment was dissected into smaller regions and a 65 bp (named C2.11) sequence has been identified to be responsable of the decreased gene expression of the all C2 fragment, in which none regulatory elements were predicted. The existence of a stable structural motif (forming a hairpin) was predicted by both RnaProfile and SFold programs in both the analysed entire 3'-UTR and C2.11 transcripts, and it was speculated that it may have a regulatory role. We show here that the hairpin structure within the 3'-UTR influences the expression of the luciferase reporter gene by means of generation of mutants. Lowering of luciferase levels for the construct with an intact hairpin structure in contrast with almost unchanged levels for the four mutated/deleted structures confirm the importance of this sequence and suggest that its disruption may directly affect rapid mRNA degradation. Since complementary mutations restoring the hairpin structure did not restore luciferase activity, we suggest that sequence within a structure is essential for the ability of the C2.11 fragment to reduce luciferase activity. The generation of a construct with the deletion of the canonical GY-box motif revealed the inactivity of this element in all the cell lines used for the transfection experiments. The 3’ end of the transcript (C6 fragment), containing the class I ARE, specifically displays a stabilizing effect in neuroblastoma cell lines. The deletion of the canonical nt 2659-2671 ARE in the C6 fragment reduced luciferase activity mRNA levels in all the analyzed cell lines, including SK-N-BE and SH-SY5Y, strongly suggesting a stabilizing role of the canonical element in neuroblastoma-derived cells through the binding of neuronal-specific stabilizing factors. We also observed the interaction of the stabilizing neuronal RNA-binding proteins nELAV with the CDK5R1 transcript in SH-SY5Y cells by immunoprecipitation and UV cross-linking experiments allowed us to observe that C1, C2 and C6 subregions show affinity for nELAV proteins in vitro. microRNA (miRNA) target site prediction with PicTar software found several target sites along the entire CDK5R1 3’-UTR without clustering. Between the 20 miRNAs predicted to bind CDK5R1, miR-15a, miR-103 and miR-107 present a high number of target sites with a free energy <-20 kcal/mol which point to these three miRNAs as the most probable to be functional. Nine pre-miRNAs have been shown to be expressed in the cell lines of interest. A preliminary quantitative analysis points to an inverse correlation of expression between miR-107 and the proteic product of CDK5R1, p35, that correlates with the negative role of miRNA on expression of proteins encoded by their target transcripts. Conclusions Our findings have shown the presence of several regulatory elements in CDK5R1 3’-UTR, and for a few of them we found a destabilizing or stabilizing function. The 3’-UTR seems to contain some regulatory elements implicated in rapid mRNA turnover which, as a consequence, maintain the steady-state transcript at low levels, and others which have a cell-specific stabilizing effect on the transcript that may contribute to rapidly increase the expression of CDK5R1 during specific biological processes. Our findings also support the hypothesis that CDK5R1 gene expression is post-transcriptionally controlled in neurons by ELAV-mediated mechanisms. This is the first evidence of the involvement of 3’-UTR in the modulation of CDK5R1 expression. The fine tuning of CDK5R1 expression by 3’-UTR may have a role in central nervous system development and functioning, with potential implications in neurodegenerative and cognitive disorders. The large 3’-UTR of CDK5R1 is expected to contain further cis-regulatory elements and interact with further trans-acting molecules, creating the possibility of complex gene expression modulation.