Regulation of Parkinson's disease-associated genes by Pumilio proteins and microRNAs in SH-SY5Y neuronal cells

Abstract

Parkinson's disease is the second most common age-related, neurodegenerative disease. A small collection of genes has been linked to Parkinson's disease including LRRK2, SAT1, and SNCA, the latter of which encodes the protein alpha-synuclein that aggregates in Lewy bodies as a hallmark of the disease. Overexpression of even wild-type versions of these genes can lead to pathogenesis, yet the regulatory mechanisms that control protein production of the genes are not fully understood. Pumilio proteins belong to the highly conserved PUF family of eukaryotic RNA-binding proteins that post-transcriptionally regulate gene expression through binding conserved motifs in the 3' untranslated region (UTR) of mRNA targets known as PUF Recognition Elements (PREs). The 3'UTRs of LRRK2, SNCA and SAT1 each contain multiple putative PREs. Knockdown (KD) of the two human Pumilio homologs (Pumilio 1 and Pumilio 2) in a neurodegenerative model cell line, SH-SY5Y, resulted in increased SNCA and LRRK2 mRNA, as well as alpha-synuclein levels, suggesting these genes are normally repressed by the Pumilio proteins. Some studies have indicated a relationship between Pumilio and microRNA activities on the same target, especially when their binding sites are close together. LRRK2, SNCA, and SAT1 each contain several putative microRNA-binding sites within the 3'UTR, some of which reside near PREs. Small RNA-seq and microRNA qPCR assays were performed in both wild type and Pumilio KD SH-SY5Y cells to analyze global and differential microRNA expression. One thousand four hundred and four microRNAs were detected across wild type and Pumilio KD cells. Twenty-one microRNAs were differentially expressed between treatments, six of which were previously established to be altered in Parkinson's disease patient samples or research models. Expression of ten miRs predicted to target LRRK2 and SNCA was verified by RT-qPCR. Collectively, our results demonstrate that Pumilios and microRNAs play a multi-faceted role in regulating Parkinson's disease-associated genes.

Fig 1. PRE and microRNA binding sites in the 3’UTRs of LRRK2, SNCA, and SAT1.

The black lines are a linear representation of the 3’UTRs of SAT1, SNCA, and LRRK2 transcripts. The number 1 indicates the nt immediately following the stop codon (5’ end of the 3’UTR). 3’UTR lengths for LRRK2 and SAT1 were based on transcript data from NCBI and 3’UTR length for SNCA was based on NCBI data as well as empirical data from expressed sequence tags [22]. Motifs with the sequence UGUA(A/U/C)AUA are considered canonical PREs and their locations on the 3’UTRs are represented by black boxes. Motifs that do not match the canonical sequence but contain a UGUA with downstream AU-rich region are considered non-canonical and are represented by gray boxes. The relative locations of predicted miR-binding sites are represented by red arrows. MiRs confirmed to be expressed in SH-SY5Y cells using sRNA-seq are bolded. MiRs with an asterisk (*) were determined to be differentially expressed based on DEseq2 analysis. MiR family conservation is based on TargetScanHuman 8.0 categorizations [71]. Broadly conserved miR families, those conserved among vertebrates, are represented by red text. All other levels of conservation are represented by black text. 8mer sites are highlighted in blue and 7mer-m8 sites are highlighted in green. 3’UTRs and predicted sites are not drawn to scale.

Fig 2. LRRK2 and SNCA RNA and alpha-synuclein protein expression increase upon knockdown of human PUMs or PRE mutation.

PUM1- and PUM2-targeting siRNAs were co-transfected into SH-SY5Y cells using electroporation to induce KD. Non-targeting siRNAs were also transfected into cells from the same population to serve as a negative control and establish relative WT levels of expression. RNA expression upon (A) PUM1 and PUM2 KD, (B) PUM1 only KD, and (C) PUM2 only KD was measured through RT-qPCR. All KDs were performed in technical triplicate and a minimum of biological triplicate. Each bar represents an analyzed mRNA. E2F1 and ISCU served as negative and positive controls, respectively, for PUM regulation. RT-qPCR data was normalized to Beta Tubulin (TUBB) expression. Fold change of normalized mRNA levels in KD condition versus WT was calculated using 2^-ΔΔCT, with values Log2 transformed and plotted on a linear y-axis. (D) Protein levels were assessed through Western blotting following double PUM1 and PUM2 KD. Shown are representative band patterns for each target from a single biological replicate. Proteins probed on the same blot are outlined in rectangles. The molecular weight markers (kDa) on the right represent the locations of protein ladder bands from each blot. Raw blot images can be viewed in full in S1 Fig. Probing for PUM2 resulted in the appearance of a non-specific band (ns) in addition to the expected bands noted by the black arrows. (E) Protein bands were quantified using BioRad Image Lab software and normalized to Beta Tubulin. E2F1 served as a negative control. ISCU and PCNA were used as positive controls for PUM regulation. The two isoforms of ISCU were quantified separately. Proteins were tested in a minimum of biological triplicate. (F) Relative normalized luciferase activity of reporters containing WT (PREwt) or mutated (PREmut) SNCA 3’UTRs. The location and sequence of the PRE and mutated nts on the SNCA 3’UTR is illustrated above the graph. All expression results were statistically analyzed using a paired, 2-tailed T-test. * = p-value ≤0.05, ** = p-value ≤0.01, *** = p-value ≤0.001, and **** = p-value ≤0.0001.

Fig 3. MicroRNA differential expression and pathway enrichment of predicted MicroRNA targets.

(A) Venn diagram of the number of unique microRNAs determined to be expressed in either WT PUM background (purple) or KD PUM background (yellow), or co-expressed in both treatments (pink; overlap) based on raw sequence reads. (B) Venn diagram defining the number of microRNAs in WT PUM only (purple), PUM KD only (yellow), or both treatments (pink; overlap) that were determined to be suitable for differential expression analysis. (C) Heatmap of DEMs hierarchically clustered. Changes in log2(ratio) expression are represented as a color change from red to blue or vice versa. (D) Enrichment of GO terms among the predicted targets of DEMs. (E) Analysis of KEGG pathway enrichment for the top 20 pathways associated with targets of DEMs within the areas of Biological Process (BP), Cellular Component (CC), and Molecular Function (MF).

Fig 4. Relative expression of microRNAs and microRNA host genes in SH-SY5Y cells in wild type and PUM knockdown states.

(A) Expression of host genes of miRs found in Table 1 that contain canonical PRE motifs in their 3’UTRs was measured using RT-qPCR. (B) MicroRNAs determined to be expressed through sRNA-seq were quantified through RT-qPCR. The control bar represents the expression in WT cells. The remaining bars represent the relative expression of the indicated genes or microRNAs in PUM1/2 KD cells. RT-qPCR data was normalized to TUBB for RNA expression and SNORD38B for miR expression. Statistical significance was determined using paired, two-tailed Student’s T-tests. * = p-value ≤0.05, ** = p-value ≤0.01, *** = p-value ≤0.001, and **** = p-value ≤0.0001.