PASSPORT-seq: A Novel High-Throughput Bioassay to Functionally Test Polymorphisms in Micro-RNA Target Sites.

Workflow of the PASSPORT-seq bioassay. (A) 100 Reference and variant miRNA binding regions each with the same 15–20 bp flanking sequence was synthesized as an oligonucleotide pool. (B) Using the flanking universal sequences, the oligonucleotide pool was amplified and made double stranded by PCR. pIS-0 plasmid was linearized by restriction enzymes. (C) The double stranded oligonucleotides were inserted into the linear plasmid using the NEBuilderTM gene assembly system. (D) Chemically competent bacteria were transformed with the plasmid pool containing the test miRNA binding regions. Transformed bacteria were plated on four plates. (E) All colonies from the plates were harvested, combined and scaled up in liquid culture. Plasmids were isolated from the liquid culture. (F) Three cell lines were transfected with the plasmid pool and incubated for 48 h after which cDNA was prepared from total RNA. (G) miRNA binding regions were amplified using universal primers that were uniquely barcoded for replicates within cell lines and for the input plasmid pool. (H) The barcoded PCR products were combined to form the sequencing pool.

Validation of the PASSPORT-seq assay. (A) Correlation between the percent-change of variant alleles compared to respective reference alleles observed in the experimental and validation PASSPORT-seq runs. Each run contained five biological replicates tested in three cell lines. The graph includes combined average data from all three cell lines. (B) Functional mirSNPs identified by the PASSPORT-seq assay. For each SNP, the observed percent change in the expression of the variant allele compared to the respective reference allele in predicted miRNA binding site was calculated Statistically significant changes after correction for multiple testing using Benjamini and Hochberg algorithm are indicated by colored boxes. Blue boxes indicate a reduction in the variant allele expression and Orange boxes indicate increased expression. Results from the experimental and validation runs are shown. Additionally, results of the analysis with merged data from experimental and validation runs are also represented.

Functional SNPs in the cell lines. Venn diagram (Micallef and Rodgers, 2014) showing the number of unique and overlapping functional SNPs that were identified by the PASSPORT-seq assay in the three cell lines. Total number of functional SNPs identified in each cell line is indicated in parenthesis.

Comparison of the results from the PASSPORT-seq assay with those from the luciferase reporter assay. Correlation between results of the PASSPORT-seq assay (P-seq) and the luciferase reporter assay (Luc); in (A) HEK293, (B) HepG2 cells, and (C) HeLa cells. The results of the two assays are represented as the percent-change observed in the variant allele compared to its respective reference allele. Tables show the magnitude of change observed in each sample using the two different assays: < 5% = 0, ≥ 5 to < 10% = –/+, ≥ 10 to < 15% = --/++, ≥ 15% = --/+++. ‘–’ indicate decreased expression and ‘+’ indicate increased expression. The rs numbers of the SNPs can be identified based on the table.

Functional mirSNPs in pharmacogenes identified by the PASSPORT-seq assay. For each SNP, the observed percent change in the expression of the variant allele compared to the respective reference allele in predicted miRNA binding site was calculated. Statistically significant changes after correction for multiple testing using Benjamini and Hochberg algorithm are indicated by colored boxes. Blue boxes indicate a reduction in the variant allele expression and Orange boxes indicate increased expression.

Cell line-specificity of functional mirSNPs in pharmacogenes. Venn diagram (Bardou et al., 2014) showing the number of unique and overlapping functional mirSNPs in the four cell lines tested.