Translation of bi-directional transcripts enhances MHC-I peptide diversity

Abstract

Antisense transcripts play an important role in generating regulatory non-coding RNAs but whether these transcripts are also translated to generate functional peptides remains poorly understood. In this study, RNA sequencing and six-frame database generation were combined with mass spectrometry analysis of peptides isolated from polysomes to identify Nascent Pioneer Translation Products (Na-PTPs) originating from alternative reading frames of bi-directional transcripts. Two Na-PTP originating peptides derived from antisense strands stimulated CD8⁺ T cell proliferation when presented to peripheral blood mononuclear cells (PBMCs) from nine healthy donors. Importantly, an antigenic peptide derived from the reverse strand of two cDNA constructs was presented on MHC-I molecules and induced CD8⁺ T cell activation. The results demonstrate that three-frame translation of bi-directional transcripts generates antigenic peptide substrates for the immune system. This discovery holds significance for understanding the origin of self-discriminating peptide substrates for the major histocompatibility class I (MHC-I) pathway and for enhancing immune-based therapies against infected or transformed cells.

Keywords: MHC-I epitope; Pioneer Translation Products; bi-directional transcripts; bi-directional translation; reverse strand antigenic peptides.

Figure 1

Identification of Nascent Pioneer Translation Products (Na-PTPs). The bioinformatics workflow for analyzing nascent peptides derived from non-exonic sequences is illustrated. Polysomes from cells treated with cyclohexamide to freeze the nascent peptides on ribosomes were isolated and subjected to RNA sequencing (RNA seq.) and mass spectrometry (MS) analysis. QC: quality control of input reads. Alignment: alignment of reads to reference genome hg38.p14. SNP and INDEL: detection of single nucleotide polymorphism and insertions/deletions by VarScan2. Genome creation: creation of the sample-specific genome sequence with incorporated mutations. Intron extraction: extraction of the intron sequences for each sample. Translation to 6 Reading Frames: translation into six possible reading frames and creation of sample specific database. Redundancy reduction: sample specific databases were merged and redundancy reduction for the same sequences was applied. SFDB: The resulting six frame proteomic database: Measured MS data were first searched against Human taxonomy. Spectra unassigned in this search were then used for the search against the SFDB. Verification: candidate sequences were verified against the hg38.p14 database for multiple occurrences in genome and blasted against Human SwissProt database to exclude previously described peptides. See also Supplementaryay Figures S1, S2.

Figure 2

Visualization of detected Nascent Pioneer Translation Products (Na-PTPs) identified by proteomics search and manually verified by human genome reference hg38.p14 and SFDB databases. Detected peptides are shown as orange stripes. Green arrows indicate direction of open reading frames and red arrows the direction of translation of PTPs from antisense strand corresponding to genes CWF19L1, IQSEC2 and ZNF615. For ZNF615 two splice variants with different origin of exon are shown. The exact position of the peptide in the chromosome (column “Chrom”) is shown in the “Na-PTP position” column. Peptide TLLTETGAGR from the antisense strand of ZNF615 was also detected in H1299 cells. See also Table 1 and Supplementary Figures S1, S2 .

Figure 3

Verification of translated antisense transcripts. (A) Forward and reverse primers were used to estimate the expression of the sense and antisense strands of the IQSEC2 and CWF19L1 genes in Hek293 and H1299 cell lysates. (B) The relative amount of sense and antisense CWF19L1 and IQSEC2 transcripts. (C) Comparing total vs. ribosome-associated levels of sense and antisense transcripts for CWF19L1 and IQSEC2.

Figure 4

IFN-γ production by CD8⁺ T cells stimulated with Na-PTP peptides in PBMCs from nine healthy donors. (A) Sequences of the five peptides (P1 to P5). (B) Mean fold change (FC) in relative IFN-γ levels for each peptide (P1 to P5) and the positive control (HER2) across all donors. (C) Individual bar charts displaying the relative IFN-γ levels (fold change (FC)) for each donor across peptide conditions, compared to the baseline (dashed line at FC = 1). Statistical significance was assessed using a one-sample t-test, comparing the observed mean IFN-γ level for each peptide to a theoretical mean of 1. P values < 0.05 are indicated on the figure.

Figure 5

Peptide substrates derived from the antisense strand are processed and presented on MHC-I molecules. (A) The SL8 (SIINFEKL) peptide is derived from chicken ovalbumin (OVA) and presented on murine Kb MHC-I molecules. Four OVA constructs were generated in which SL8 was expressed from the sense or antisense strand. OVA corresponds to the full-length OVA whereas OVA SHORT corresponds to exons 5, 6 and 7. (B) The relative expression of the transcripts shows that the sense and antisense strands are expressed. (C) CD8+ T cell assay using OT-1 cells that specifically detect SL8 presented on Kb molecules shows that translation of both strands generates SL8 peptides. The data show an average of five independent experiments (unpaired t test, ***p-value <0.001, **p-value<0.005).