Fig 1.
Measuring allelic mRNA occupancy on polysomes.
The flowchart depicts the strategy used to fractionate cellular components for RNA analysis and to determine Allelic Expression Imbalance (AEI), defined as a ratio of reference allele (REF) to alternative allele (ALT). Chromatographic peaks obtained from SNaPshot [25] represent relative allelic abundance in gDNA and mRNA (as cDNA). Significant deviation from unity (after normalization to allelic ratios of gDNA or plasmid DNA) indicates unequal polysomal occupancy of allelic mRNA. RNPs represent cytosolic mRNA either polysome-free or bound to ribonucleotide particles.
Fig 2.
Allelic mRNA expression measurement in polysomes of transfected cell lines and cells with native expression of NAT1 and ABCB1.
A. Effect of OPRM1 118A>G polymorphism on cytoplasmic mRNA expression and polysome loading; an A>G ratio >1 indicates reduced levels of the minor G allele relative to the A allele. Equal amounts of OPRM1 plasmids containing 118A and G were co-transfected into CHO cells. The 118A/G allelic ratios, measured with SNaPshot, represent the mean ± s.(n = 6), *p<0.049 for monosome versus cytoplasmic and **0.003 for monosome versus polysomes, two-tailed Student’s t-test. These profiles are representative of results from 3 independent cultures. B-C. AEI measurement in polysomes of lymphoblastoid cells with native expression of NAT1 and ABCB1. LCL (1 x 107 cells) heterozygous for *4/*10 were selected for polysome preparation. N-acetyltransferase 1 (NAT1) *10 allelic mRNA increased polysomal loading (lower *4/*10 ratios, p<0.01, corrected for gDNA ratios (B). Ribosomal occupancy of ABCB1 (MDR1) 3435C>T allelic mRNA. ABCB1 3435C>T allelic mRNA ratios (C). Reduced 3435C/T ratios demonstrated a significant reduction of the 3435T allele in the cytoplasmic mRNA extract, with no significant difference in any gradient fraction. These profiles are representative of gradients done with extracts from 3 independent cultures (mean ± s.d., n = 6).
Fig 3.
RNA classes in cytosol and polysomes from three LCLs, measured by RNA-Seq.
After amplification of total cytosol RNA fraction and polysome fractions with NuGen, equal amounts of RNA were subjected to RNA-Seq, and sequence reads were aligned to annotated RNA to determine expression levels. A. Average expression level (normalized to FKPM) in cytoplasmic and polysome fractions from 3 different LCLs after exclusion of rRNA, tRNA, and mt-RNA. The pseudogenes and lncRNAs are reduced on polysome fractions. In cytosol lncRNA represent antisense-RNA (21.3%), lincRNA (4.7%) and lncRNA (74%). On polysomes the lncRNA distribution was drastically different: antisense-RNA (73%), lincRNA (14%) and lncRNA (13%). Error bars represent expression s.d. from 3 different LCL cells. B-D. Hierarchical clustering of the profiles of protein coding genes (B), pseudogenes (C) and lncRNA (D) performed with similarity indices (see Materials and Methods) for cytosol and polysome fractions. Observed difference of similarity indices between cytosol and polysome clusters is higher in lncRNA than in other RNA classes (note the difference on vertical scale).
Table 1.
Protein coding mRNAs showing significant differences of polysome to cytosol ratios, adjusted p-value < 0.05.
Table 2.
Noncoding RNAs showing significant differences of polysome to cytosol ratios, adjusted p-value < 0.05.
Table 3.
Top-scoring 20 microRNAs most strongly enriched on polysomes and most strongly polysome-depleted compared to cytosol, non-adjusted p-value < 0.05.
Fig 4.
microRNA expression in cytosol and polysomal loading in three LCLs.
A. Hierarchical clustering of microRNA expression profiles between cytosol and polysomes. microRNAs highly enriched on polysomes such as let-7a, mir-1275, and mir-145 had docking sites on mRNAs also enriched on polysomes. B. Hierarchical clustering of microRNA expression profiles measured with similarity indices (see Materials and Methods) reveal significant differences of microRNA profiles between cytosol and polysome fractions, showing reduced relative distances in comparison to cytosolic fractions.
Fig 5.
mRNA isoforms consistently different between cytoplasmic and polysome identified by RNA seq in three LCLs.
Among all isoforms detected for a gene, some isoforms significantly differed in abundance between cytosol and polysomes. Isoforms showing similar pattern in all 3 LCLs are shown. Differences consistent among all cell lines suggest the possibility of isoform specific polysome recruitment. The results represent mean ± s.d. (n = 3). The isoforms are identified in S3 Table.
Table 4.
Genes expressing RNA isoforms with differential loading onto polysomes, and displaying different allelic RNA ratios in cytosolic and polysomal extracts.
Targets with > 2 fold difference in allelic ratio between cytosol and polysome were analyzed for isoform differences. Results show both enrichment and depletion of isoforms on polysomes.
Fig 6.
Allelic RNA ratios determined in whole transcriptome and translatome (polysomal RNAs) of three LCLs.
A. Allelic RNA ratio differences measured in genes with two or more heterozygous SNPs are shown. To improve accuracy of allelic ratios measured using sequencing data, the ratios were measured at multiple SNPs of same transcript. B. Validation of mRNA allelic ratios obtained with RNA-Seq using SNaPshot assay. Both methods yielded similar allelic ratio differences between cytosol and polysomes.
Fig 7.
mRNA isoforms differentially loaded onto polysomes, and in addition displaying twofold different allelic RNA ratios (major/minor alleles) between cytosol and polysomes.
A. Isoforms underrepresented on polysomes. B. Isoforms enriched on polysomes. Allelic mRNA ratios reflect the sum of all isoforms where the marker SNP is located. The allelic mRNA ratios are shown above the respective bar.
Table 5.
Genes implicated as having genetic variants affecting polysomal loading (either directly or via RNA isoforms) were tested for associations with clinical phenotypes (all GWAS from dbGaP) and eQTLs.