Pan-viral ORFs discovery using massively parallel ribosome profiling
Science
12 Jun 2025
Vol 388, Issue 6752
pp. 1218-1224
Editor’s summary
The functional annotation of virus genomes is tricky, not least because it can involve culturing viruses, some of which are dangerous to humans. This task is also challenging for computational approaches because it can be difficult to identify virus open reading frames. Weingarten-Gabbay et al. devised a mass ribosome profiling method that avoids the hazards of virus culture by using fragments of viral genomes inserted into plasmids that are then transfected into cells. This strategy detects peptides that regulate virus protein expression in ways that modify the course of infection and mediate potent host immune responses, can survey noncanonical translation in many diverse viruses simultaneously, and offers a tool for vaccine development. —Caroline Ash
Abstract
Defining viral proteomes is crucial to understanding viral life cycles and immune recognition but the landscape of translated regions remains unknown for most viruses. We have developed massively parallel ribosome profiling (MPRP) to determine open reading frames (ORFs) across tens of thousands of designed oligonucleotides. MPRP identified 4208 unannotated ORFs in 679 human-associated viral genomes. We found viral peptides originating from detected noncanonical ORFs presented on class-I human leukocyte antigen in infected cells and hundreds of upstream ORFs that likely modulate translation initiation of viral proteins. The discovery of viral ORFs across a wide range of viral families—including highly pathogenic viruses—expands the repertoire of vaccine targets and reveals potential cis-regulatory sequences.
Despite advances in sequencing viral genomes, functional annotations of these genomes have lagged behind. Beyond the annotated canonical open reading frames (ORFs), viral genomes encode noncanonical ORFs that do not fulfill the classical definition of ORFs—i.e., they do not start with ATG nucleotides (nt) and/or are shorter than 100 amino acids (aa) in length. These noncanonical ORFs and the resulting microproteins modulate viral infection (1–3), contribute to the immune response to viruses (4–8), and regulate gene expression (9–11). However, the deviation from classical ORF features complicates their detection by computational approaches and relies mostly on experimental measurements.
The development of ribosome profiling (also termed Ribo-seq) has transformed our ability to detect translated regions across genomes (12). Ribosome profiling utilizes deep sequencing of ribosome-bound mRNA fragments to determine ribosome occupancy, indicating translated regions at single nucleotide resolution. It has uncovered many noncanonical ORFs in mammalian cells, yeast, bacteria, and viruses, including upstream ORFs (uORFs) and upstream-overlapping ORFs (uoORFs) in 5′UTRs, short ORFs in noncoding RNAs, and overlapping internal ORFs in annotated coding sequence (iORFs) (13, 14).
The landscape of translated regions, however, is still unknown for most viruses. Ribosome profiling has only been used to profile a handful of viruses as a result of the following challenges: Each virus demands a specific culturing system, some viruses cannot be cultured in the lab, and highly pathogenic viruses require high-containment facilities. Moreover, most viruses are genetically diverse, necessitating a method that can evaluate multiple variants in parallel.
We have leveraged an oligo synthesis library (15–17) and combined it with ribosome profiling to perform pan-viral ORF discovery. We measured the translation of 20,170 synthetic sequences from 679 viral genomes in two human cell lines, under stress conditions associated with viral infection, and when expressed by cap-dependent or internal ribosome entry site (IRES)-dependent translation. We estimated ORF discovery using the annotated coding sequences (CDSs) and previously reported noncanonical ORFs. We compared ribosome footprints in the synthetic library and natural infection with four viruses. We then examined the function of the detected ORFs in two processes: class-I human leukocyte antigen (HLA-1) presentation and uORF-mediated translation regulation.
Results
Massively parallel ribosome profiling to identify ORFs
We developed massively parallel ribosome profiling (MPRP) to measure translated regions across hundreds of human viruses (Fig. 1A). We used oligonucleotide library synthesis technology to encapsulate thousands of viral sequences in a single pooled experiment. Each oligo contained 200 nt of viral sequence, flanked by constant primers. We cloned the library into an overexpression plasmid downstream of a CMV promoter and upstream of a Woodchuck hepatitis virus post-transcriptional regulatory element used to further enhance expression. To monitor the translation of ORFs in the designed oligos, we excluded ATG start codons on the plasmid. We transfected the pooled library plasmid into HEK293T or A549 cells. We then performed a modified protocol of ribosome profiling (18) after treating cells with either cycloheximide (CHX) to inhibit elongating ribosomes or lactimidomycin (LTM) to inhibit initiating ribosomes. We mapped deep sequencing reads representing ribosome footprints to the synthetic library and identified ORFs using probabilistic inference of codon activities by an EM algorithm (PRICE) a computational method for the detection of ORFs in ribosome profiling experiments (19).
Fig. 1. Design of oligonucleotide synthetic library and MPRP measurements.
(A) Illustration of the MPRP. (i) Synthetic library amplification using constant primers. (ii) Cloning of library into the overexpression vector. (iii) Transient transfection of plasmid pool into HEK293T or A549 cells for 24 hours. (iv) Treatment of cells with either LTM or CHX and performance of ribosome profiling protocol. (v) Mapping of deep sequencing reads to the synthetic library. (vi) Inferring of translated ORFs using PRICE (19). (B) Design of the tested synthetic oligonucleotides: (i) ORFs detected previously by ribosome profiling in infected cells with either the intact start codon or a GCC mutation. (ii) Tiling oligos encompassing complete viral transcripts. (iii) Oligos spanning the 5′UTR and the first 140 nt of annotated viral CDSs. For the region containing the CDS, two oligos were designed: the WT sequence and a start-codon–mutated oligo. (C) Comparison of the number of ribosome footprints mapped to 15,000 oligos in two biological replicates in HEK293T cells. R = 0.92, Pearson correlation. (D) Comparison of the number of ribosome footprints mapped to 15,000 oligos in HEK293T and A549 cells. R = 0.89, Pearson correlation. Oligos with shared sequences with the adenoviral E1A/B genes and the SV40 large T-antigen endogenously expressed by HEK293T cells are shown in green, blue, and pink. (E) Comparison of the number of ribosome footprints mapped to 1163 identical oligos in two synthetic libraries in HEK293T cells. R = 0.79, Pearson correlation.
We tested the quality of ribosome footprints by examining host-mapped reads, showing the expected read length distribution, footprint enrichment in the CDS, and trinucleotide periodicity (fig. S1). To verify that our experimental system captures the translation of an ORF embedded within a synthetic oligo, we performed ribosome profiling on cells transiently transfected with a full-length or truncated green fluorescent protein (fig. S2).
We designed a “pilot” library to estimate viral ORF discovery using MPRP. We tiled 30 mRNAs annotated in the genomic datasets of human cytomegalovirus (HCMV) and herpes simplex virus type 1 (HSV-1) (fig. S3A). We observed the expected low occupancy in the 5′UTR and enrichment of ribosomes at the start codon (fig. S3B). However, we also observed high occupancy in oligos along the CDSs that do not contain the CDS start codon. In the context of the full transcript, these alternative initiation sites are preceded by multiple start codons, making it unlikely that the ribosome will initiate at these positions. We reasoned that a tiling approach may result in false-positive ORF discovery and that MPRP has higher accuracy at the beginning of annotated CDSs and 5′UTRs. Thus, we decided to focus on this region for the design of the pan-viral library, which encompasses the majority of ORF initiation sites (13).
We designed a pan-viral library of 15,000 oligos to screen for novel ORFs in the 5′UTRs and the beginning of the CDS of 3976 genes in 679 viral genomes. For each gene, we designed three oligos: a wild-type (WT) oligo containing 60 nt of the 5′UTR and 140 nt of the CDS, a mutated oligo in which the annotated start codon was mutated to GCC nts, and a further-upstream oligo in the 5′UTR region, starting −260 nt relative to the CDS start codon (Fig. 1B).
To gauge the reproducibility of the MPRP measurements, we compared ribosome occupancy on the synthetic oligos in different experiments. Measurements of the pilot and the pan-viral libraries were consistent between biological replicates in HEK293T cells (R = 0.81 and R = 0.92, respectively; fig. S4 and Fig. 1C). We found high agreement between ribosome occupancies on the pan-viral library in HEK293T and A549 cells (R = 0.89, Fig. 1D). We also designed a set of 1163 oligos with identical sequences in both the pilot and pan-viral libraries and found good agreement between oligos that were independently synthesized, cloned, and measured (R = 0.79, Fig. 1E).
To imitate the cellular environment in infected cells, we repeated MPRP when inducing stress. We treated cells with poly(I:C) to model viruses that form double-strand RNA and Thapsigargin to induce endoplasmic reticulum stress. We confirmed the expected phosphorylation of translation regulators (PKR and eIF2alpha, fig. S5A) and innate immune sensors (IRF3 and STAT1, fig. S5B). Notably, we observed phosphorylation of PKR, eIF2alpha, IRF3, and STAT1 in cells expressing a plasmid without treatment, suggesting that transient transfection by itself induces a similar stress response. Accordingly, we found a significant agreement between MPRP measurements (fig. S5C) in nontreated and poly(I:C)-treated A549 and HEK293T cells (R = 0.773 and R = 0.767, respectively, fig. S5D and Fig. 5E). These results indicate that sequences tested in MPRP are exposed to a cellular environment associated with viral infection. MPRP measurements on the pan-viral library uncovered 5381 ORFs including 4208 noncanonical ORFs (table S1) for further study.
Estimating ORF discovery using annotated viral coding sequences
To estimate the detection of ORFs, we examined the distribution of ribosome footprints across oligos representing the first 140 nt of annotated viral CDSs (Fig. 1B). We performed metagene analysis for all the CDSs from each family by computing the average ribosome footprints in each position relative to the annotated start codon. In most of the 21 viral families tested, we found the expected enrichment of ribosome footprints in the CDS and in the correct reading frame, indicating a robust identification of annotated CDSs in MPRP across different viruses (Fig. 2A and fig. S6).
Fig 2. Annotated CDSs measurements in MPRP and infected cells.
(A) Metagene analysis for four viral families with CHX and LTM inhibitors showing the average ribosome footprints in each position. See figs S6 and S7 for 21 viral families. (B) Comparison of the average number of ribosome footprints between oligos containing the WT start codon (upper graph) with those in which the annotated start codon was mutated to GCC (lower graph). Shown are the average ribosome footprints in each position across 3777 oligos. (C) ORF discovery using PRICE. Number of ORFs detected in each position (ORF start position). (D) Mirror plot showing the number of ribosome footprints mapped to the first 200 nt of the Nucleocapsid transcript in VSV-infected cells and MPRP. (E) Similar to (D) for IAV NEP and PB1 transcripts.
We next examined ribosome footprints in the presence of an LTM inhibitor, which inhibits 80S complexes at the initiation site. We found a clear enrichment in the annotated start codons for 19 of the 21 viral families tested (Fig. 2A and fig. S7). Using reverse genetics, we found a significant 3.7-fold reduction in the number of footprints on 3777 oligos in which we mutated the annotated start codon to GCC (P < 10−288, Fig. 2B and fig. S8).
We used the annotated CDSs to assess the performance of our computational pipeline for inferring ORFs in the MPRP experiment. When running PRICE, we did not indicate which viral oligos contain annotated CDSs, to evaluate their discovery in an unbiased fashion. PRICE successfully captured 1136 of 3976 viral CDSs initiating at the annotated start codon (28%, accounting for 31% of the total ORFs detected by PRICE, Fig. 2C), with a high correlation between the number of reads in two biological replicates (R = 0.93, fig. S9). The observance of 28% of the annotated CDSs may represent inherent limitations of the MPRP, assaying the translation of viral sequences outside the context of viral infection (see Discussion). Nevertheless, we observe specific enrichment at the annotated start codon (15.3 times higher than the next most abundant position).
Since many viruses rely on IRESs for protein translation, we tested whether the mode of ribosome recruitment affects ORF discovery. We cloned the pan-viral library into a bicistronic plasmid downstream of the EMCV or Polio IRES and repeated MPRP in HEK293T cells (figs. S10A and S11 and tables S2 and S3). We observed the expected enrichment of ribosomes on annotated CDSs and a reduction in start codon mutated oligos when expressing the library from both IRESs (fig. S10B). We found a significant overlap between PRICE-detected CDSs in the cap-dependent and IRES-dependent MPRPs (hypergeometric P-value < 10–20 and P-value < 10−8 for EMCV and Polio IRES, respectively) and agreement between the number of ribosome footprints on shared ORFs (R = 0.75 and R = 0.83 for EMCV and Polio IRES, respectively, fig. S10, C and D). Finally, we found a similar pattern of ribosome footprints in the cap- and IRES-dependent MPRPs on CDSs of viruses that rely on IRESs for translation initiation from the Flaviviridae and Picornaviridae families (fig. S10, E and F). These results suggest that at least for some proteins, ORF discovery was independent of the mechanism by which the ribosome was recruited to the transcript.
We then compared the pattern of ribosome footprints in MPRP and native viral infection. We performed ribosome profiling in cells infected with vesicular stomatitis virus [(VSV) Indiana strain], influenza A virus [(IAV) Puerto Rico 8 H1N1 strain], and Hepatitis C Virus [(HCV) genotype 2] (fig. S12). We found a high congruence between the location of footprints originating from MPRP and infected cells in CDSs from VSV (Fig. 2D and fig. S13), IAV (Fig. 2E and fig. S14), and HCV (fig. S15).
Noncanonical ORFs discovery
We estimated the discovery of noncanonical ORFs. As part of the pilot library, we included oligos comprising 60 nt upstream and 140 nt downstream of the initiation codons of 716 ORFs identified in HCMV-infected cells by ribosome profiling (20), termed here “Ribo-seq ORFs” (Fig. 1B). For each ORF, we also designed a mutated oligo in which we replaced the reported start codon with GCC.
We found clear enrichment of ribosome footprints along the reported Ribo-seq ORFs with trinucleotide periodicity indicating translation in the correct reading frame. Notably, MPRP detected ribosome footprints on both canonical and noncanonical Ribo-seq ORFs, including ORFs with a non-AUG start codon and short ORFs in the length of 20 aa or less (Fig. 3, A to C). Mutating the start codon to GCC of 284 Ribo-seq ORFs resulted in a substantial reduction of ribosome footprints compared with WT oligos (Fig. 3D). Our findings confirmed the non-AUG start codons reported by Stern-Ginossar et al., (20) and demonstrate how MPRP can be used to functionally characterize the initiation site of noncanonical ORFs. PRICE ORF prediction detected 152 of the 716 Ribo-seq ORFs (21%, accounting for 25% of the total ORFs detected by PRICE) with 11.7-fold enriched initiation at the reported start codon compared to the next abundant position (Fig. 3E).
Fig 3. Noncanonical ORF measurements in MPRP and infected cells.
(A to C) Metagene analysis of oligos containing the sequence of ORFs that were identified by ribosome profiling of HCMV-infected cells (20). The average of ribosome footprints in each position. (D) Comparison of the average number of ribosome footprints between oligos containing the WT start codon (upper graph) to those in which the reported start codon was mutated to GCC (lower graph). Shown are the average ribosome footprints across 284 oligos, containing Ribo-seq ORFs with lengths of 7 to 45 aa. (E) ORF discovery using PRICE. Number of ORFs detected in each position (ORF start position). (F) Mirror plot showing the number of ribosome footprints in HCMV-infected cells and MPRP. The purple box highlights uORF2, which encodes a ribosome arrest peptide (21) (G) Mirror plot showing the number of ribosome footprints in IAV-infected cells and MPRP. The purple box highlights an internal overlapping iORF in the +1 reading frame. (H) Percentages of ribosome footprints mapped to 0, +1, and −1 reading frames in the region encoding the internal ORF (upper panel) and outside this region (lower panel). (I) Ribosome footprints on the M1 coding sequences of six additional IAV strains from the MPRP experiment.
We found similar ribosome footprints on PRICE-predicted uORFs and iORFs in MPRP and cells infected with HCMV, VSV, and IAV (Fig. 3, F to G, and fig. S16). Consistent with HCMV-infected cells, MPRP detected a prominent peak at the end of an uORF in the UL4 5′UTR. This uORF was shown to inhibit the translation of the main CDS by causing ribosome stalling (21) (Fig. 3F), indicating that MPRP can provide information on functional stalling sites. Ribosome profiling of cells infected with the H1N1 strain of IAV confirms an iORF in the +1 frame of M1 detected by MPRP (Fig. 3, G to H). Notably, the design of the pan-viral library allowed us to uncover this iORF in six additional influenza strains, including the H5N1 bird flu strain responsible for the current cattle outbreak (Fig. 3I).
Expanding the repertoire of viral antigens
Noncanonical ORFs contribute to the pool of peptides presented on the HLA-I complex to cytotoxic T cells (5, 6, 7, 22–28). Notably, peptides from noncanonical ORFs can be enriched on the HLA-I complex in comparison to canonical peptides, with some eliciting stronger T cell responses (7, 26). Thus, noncanonical ORFs discovery can provide new targets for vaccines and insights on the interaction between viruses and the immune system.
To assess the contribution of noncanonical ORFs detected by MPRP to HLA-I presentation, we reanalyzed two immunopeptidome datasets from cells infected with either HCMV (19) or vaccinia virus (VACV) (29) (fig. S17). We found five distinct peptides from four noncanonical ORFs in HCMV: an uORF in the 5′UTR of UL4 (VLSAKKLS, and VLSAKKLSSL), an uORF in the 5′UTR of UL148 (FAKSKTIGL), an uoORF in the 5′UTR of UL135 (YPAPRPQAI), and an N-terminal extended isoform of the UL36 protein (VMDDLRDTL) (Fig. 4A). In VACV-infected cells, we found two distinct HLA-I peptides from an ORF located upstream to I7L (ILFFHVLLY) and from an iORF overlapping the coding region of L3L (HRNKIINAEK) (Fig. 4B). Of the seven detected peptides, six were predicted to be good binders by HLAthena (30) (MSi rank ≤ 2) to at least one of the expressed HLA-I alleles, and all the noncanonical ORFs were supported by a peptide with a good prediction score (table S4).
Fig 4. HLA-I peptides derived from noncanonical ORFs in HCMV and VACV.
(A) HLA-I peptides detected in four noncanonical ORFs of HCMV identified by MPRP. (B) HLA-I peptides originating from two noncanonical ORFs in VACV. (C) Comparison of HLA-I presentation from annotated and noncanonical ORFs in HCMV. For each ORF, we present the number of total HLA-I peptides detected in HCMV immunopeptidome (not unique) normed by the ORF length. P < 10−3, Wilcoxon rank-sum test.
Next, we evaluated the contribution of the noncanonical ORFs to HLA-I presentation. Appending MPRP-detected ORFs to the annotated HCMV ORFs increased the number of mapped HLA-I peptides by 7.4% (from 964 to 1035 peptides). Moreover, the noncanonical ORFs produced more HLA-I peptides, compared with most of the annotated ORFs (Wilcoxon rank-sum P-value <10−3, Fig. 4C).
Exposing uORFs in viral 5′UTRs
uORFs can modulate the expression of viral proteins by attenuating translation initiation at the main CDS (9–11, 31). Examining the distribution of ribosome footprints in 2418 viral genes, we identified two main clusters: a group of genes in which most of the footprints were observed in the 5′UTR with low occupancy in the CDS region (5′UTR cluster), and a group of genes in which most of the footprints were detected in the CDS with low occupancy in the 5′UTR (CDS cluster) (Fig 5, A and B). The trinucleotide periodicity observed in uORFs detected in the 5′UTR region indicates that they are actively translated by ribosomes (Fig 5C). Moreover, we found a strong signal of initiating ribosomes at the start codons of these uORFs in the presence of an LTM inhibitor.
Fig 5. Ribosome densities on uORFs and CDSs in response to eIF2alpha phosphorylation.
(A) Heatmap showing ribosome footprint densities across 2418 viral oligos. Each line represents a single viral gene and each column represents the position relative to the annotated start codon. Genes in the upper cluster (purple) had more footprints in the 5′UTR region than the CDS region and genes in the lower cluster (blue) had more footprints in the CDS than the 5′UTR. (B) Example of two individual genes from each cluster and the distribution of ribosome footprints observed in each position. (C) Metagene analysis showing the average ribosome footprints in each position along uORFs detected by PRICE, relative to the uORF start position. Shown for CHX (left) and LTM (right) inhibitors. (D) Western blot analysis of lysates from HEK293T cells treated with 40 uM sodium-arsenite for 30 and 60 min. Phosphorylated eIF2alpha was detected with a monoclonal phospho S51 antibody (upper panel). ATF4 protein was detected using a polyclonal antibody (lower panel). In both membranes, Vinculin was used as a loading control. (E) Repeat of the MPRP experiment in HEK293T cells that were treated with 40 uM sodium arsenite and in nontreated cells. Shown are heatmaps of ribosome densities across viral oligos and clusters similarly to the analysis in (A).
We hypothesized that genes in the 5′UTR cluster have uORFs that attenuate translation from the main CDS. uORFs translation response to cellular stress, including viral infection, through the phosphorylation of eIF2alpha (fig. S18). Upon eIF2alpha phosphorylation, preinitiation complexes (PICs) are more likely to scan through the uORF start codon and initiate translation at the main CDS. To test if the detected uORFs are regulated by cellular stress, we treated cells with sodium arsenite, a potent inducer of eIF2alpha phosphorylation (32). We confirmed the increase in eIF2alpha phosphorylation and the uORF-regulated ATF4 protein (Fig. 5D).
We performed MPRP in HEK293T cells treated with sodium-arsenite. We found a relative decrease in the fraction of viral genes in which ribosomes were “held” at the 5′UTR (37% to 19% in untreated and treated cells, respectively, Fig. 5E) and an increase in the fraction of genes containing ribosome footprints in the CDS (63 to 81% in untreated and treated cells, respectively). This result indicates that in response to eIF2alpha phosphorylation, ribosomes were more likely to bypass the uORFs in the 5′UTR and initiate translation at the main CDS, as expected in the case of inhibitory uORFs.
Discussion
We present a method to comprehensively screen sequences of many viruses for translated regions in a single pooled experiment. Using MPRP, we uncovered thousands of potential ORFs and provided high resolution of ribosome footprints across the 5′UTR region and the beginning of the CDS in thousands of viral genes.
Although synthetic systems can never fully recapitulate naturally infected cells, MPRP bridges the gap between current computational annotations, which rely on outdated assumptions of ORF characteristics, and traditional ribosome profiling. This is critical given that the landscape of translated regions has not yet been determined for the majority of viruses. MPRP has the capacity to identify translated ORFs in a broad spectrum of viruses and provides a tool for investigating highly pathogenic viruses. Using 200-nt-long viral fragments omits the requirement for scarce high-containment facilities. Additionally, we demonstrated how MPRP can provide rapid insights during an outbreak by exposing a noncanonical iORF in the H5N1 bird flu genome. Within a few weeks, MPRP can detect ORFs in a newly discovered virus, independently of its culturing conditions.
We exposed additional sources of viral antigens that could contribute to HLA-I presentation and T cell recognition. While T cell assays almost exclusively assess responses against canonical proteins, some T cell epitopes from noncanonical ORFs induce more potent T cell responses compared with canonical epitopes (7). Thus, the incorporation of noncanonical ORFs into T cell assays has the potential to enhance their sensitivity and facilitate the identification of vaccine targets.
We also discovered viral uORFs that likely affect gene expression regulation. We found numerous potential uORFs that exhibited hallmarks of translation and eIF2alpha phosphorylation responsiveness. These uORFs might function in the temporal regulation of viral proteins as suggested for HHV-6 and HCMV (33). Moreover, MPRP can provide insights into the potential mechanism by which uORFs exert their function, such as ribosome stalling (34).
It is important to acknowledge the limitations inherent in this study. The viral sequences examined here were assessed independently of the broader genome context and were evaluated in noninfected cells. The 200-nt synthetic oligo might exclude cis-regulatory elements such as IRESs (35), VPg proteins (36), uORFs in long 5′UTRs (37), ribosome shunting (38), microRNA binding sites (39), and pseudoknots (40). MPRP does not capture the distinctive biology occurring within cells infected by each of the many viruses studied here and therefore lacks host and/or viral proteins that regulate translation, such as SARS-CoV-2 nsp1 (41). Hence, MPRP cannot accurately detect every ORF in each virus.
Nevertheless, we have substantive evidence supporting the identification of genuine ORFs. This evidence includes elongating and initiating ribosomes on annotated and reported noncanonical ORFs, the reproducibility of measurements across different cell types, the notable decrease in ribosome footprints upon mutating AUG and non-AUG start codons, the high similarity of ribosome footprints pattern in MPRP and native virus infection, the corroborative support from HLA-I peptides identified through mass spectrometry, and the responsiveness of uORFs to stress conditions.
In total, our study yields thousands of candidates for unexplored ORFs across hundreds of human viruses, which can enhance our understanding of viral biology and contribute to vaccine development.
Acknowledgments
We thank T. Ouspenskaia, A. Nachshon, S. Reilly, J. Xue, A. Lin, H. Metsky, A. Lercher, and O. Mizrahi for many valuable discussions. We thank N. McGlincy and N. Ingolia for sharing their detailed ribosome profiling protocol with the broad community of researchers (18), which contributed to the development of MPRP.
Funding:
This study was supported in part by grants from the National Institute of Health (NIH) (U19AI110818 to P.C.S., P01CA206978 to S.A.C., and R01CA057973 to C.M.R.), the United States Department of Agriculture (58-3022-2-031 to P.C.S.), National Cancer Institute (NCI) Clinical Proteomic Tumor Analysis Consortium program (U24CA270823 and U01CA271402 to S.A.C.), and the Dr. Miriam and Sheldon G. Adelson Medical Research Foundation (to S.A.C.). S.W.-G. is the recipient of a Human Frontier Science Program fellowship (LT-000396/2018), EMBO nonstipendiary long-term fellowship (ALTF 883-2017), the Gruss-Lipper postdoctoral fellowship, the Zuckerman STEM Leadership Program fellowship, and the Rothschild Postdoctoral Fellowship.
Author contributions:
S.W.-G. conceptualized the study, designed the synthetic libraries, designed the experiments, performed experiments, analyzed the data, and wrote the manuscript. M.R.B., A.C.S., Y.Y., C.A.F., N.L.W., and C.K.B. performed the experiments. S.K., E.K.V., D.L., K.R.C., S.A.C., and J.G.A. analyzed the mass spectrometry data of HLA-I immunopeptidomes. L.E.H. and C.M.R. contributed to data interpretation. P.C.S. supervised the work and wrote the manuscript.
Competing interests:
S.W.-G., M.R.B., A.C.S., and P.C.S. are named as co-inventors on International Patent Application PCT/US2024/048478, claiming priority to US Provisional Application 63/540,279 related to this work filed by The Broad Institute that covers massively parallel methods and techniques for evaluating open reading frames in genomes, particularly viral genomes. These methods and data unveil new proteins, immune targets, and cis-regulatory elements that can serve in the design of vaccines and commercial overexpression platforms. M.R.B. is an employee of BioNTech SE. S.K. is an employee of Genentech. S.A.C. is a member of the scientific advisory boards of Kymera, PTM BioLabs, Seer, and PrognomIQ. J.G.A. is a paid consultant for Enara Bio and Moderna. P.C.S. is a cofounder of and consultant to Sherlock Biosciences and Delve Biosciences, is on the Board of Directors and a shareholder of Danaher Corporation and Polaris Genomics, and holds equity in all the companies. All other authors declare no competing interests.
Data and materials availability:
The raw sequencing data generated in this study have been submitted to the Gene Expression Omnibus (GEO; http://www.ncbi.nlm.nih.gov/geo/) under accession number GSE272406.
License information:
Copyright © 2025 the authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original US government works. This article is subject to HHMI’s Open Access to Publications policy. HHMI lab heads have previously granted a nonexclusive CC BY 4.0 license to the public and a sublicensable license to HHMI in their research articles. Pursuant to those licenses, the Author Accepted Manuscript (AAM) of this article can be made freely available under a CC BY 4.0 license immediately upon publication. http://www.science.org/content/page/science-licenses-journal-article-reuse
Supplementary Materials
The PDF file includes:
Other Supplementary Material for this manuscript includes the following:
Data S1 to S4
- Download
- 1.75 MB
MDAR Reproducibility Checklist
- Download
- 162.24 KB
References and Notes
1
V. Lulla, A. M. Dinan, M. Hosmillo, Y. Chaudhry, L. Sherry, N. Irigoyen, K. M. Nayak, N. J. Stonehouse, M. Zilbauer, I. Goodfellow, A. E. Firth, An upstream protein-coding region in enteroviruses modulates virus infection in gut epithelial cells. Nat. Microbiol. 4, 280–292 (2019).
2
P. J. Ogden, E. D. Kelsic, S. Sinai, G. M. Church, Comprehensive AAV capsid fitness landscape reveals a viral gene and enables machine-guided design. Science 366, 1139–1143 (2019).
3
V. Lulla, A. E. Firth, A hidden gene in astroviruses encodes a viroporin. Nat. Commun. 11, 4070 (2020).
4
A. Bansal, J. Carlson, J. Yan, O. T. Akinsiku, M. Schaefer, S. Sabbaj, A. Bet, D. N. Levy, S. Heath, J. Tang, R. A. Kaslow, B. D. Walker, T. Ndung’u, P. J. Goulder, D. Heckerman, E. Hunter, P. A. Goepfert, CD8 T cell response and evolutionary pressure to HIV-1 cryptic epitopes derived from antisense transcription. J. Exp. Med. 207, 51–59 (2010).
5
N. T. Ingolia, G. A. Brar, N. Stern-Ginossar, M. S. Harris, G. J. S. Talhouarne, S. E. Jackson, M. R. Wills, J. S. Weissman, Ribosome profiling reveals pervasive translation outside of annotated protein-coding genes. Cell Rep. 8, 1365–1379 (2014).
6
N. Yang, J. S. Gibbs, H. D. Hickman, G. V. Reynoso, A. K. Ghosh, J. R. Bennink, J. W. Yewdell, Defining Viral Defective Ribosomal Products: Standard and Alternative Translation Initiation Events Generate a Common Peptide from Influenza A Virus M2 and M1 mRNAs. J. Immunol. 196, 3608–3617 (2016).
7
S. Weingarten-Gabbay, S. Klaeger, S. Sarkizova, L. R. Pearlman, D.-Y. Chen, K. M. E. Gallagher, M. R. Bauer, H. B. Taylor, W. A. Dunn, C. Tarr, J. Sidney, S. Rachimi, H. L. Conway, K. Katsis, Y. Wang, D. Leistritz-Edwards, M. R. Durkin, C. H. Tomkins-Tinch, Y. Finkel, A. Nachshon, M. Gentili, K. D. Rivera, I. P. Carulli, V. A. Chea, A. Chandrashekar, C. C. Bozkus, M. Carrington, N. Bhardwaj, D. H. Barouch, A. Sette, M. V. Maus, C. M. Rice, K. R. Clauser, D. B. Keskin, D. C. Pregibon, N. Hacohen, S. A. Carr, J. G. Abelin, M. Saeed, P. C. Sabeti, K. Lavin-Parsons, B. Parry, B. Lilley, C. Lodenstein, B. McKaig, N. Charland, H. Khanna, J. Margolin, A. Gonye, I. Gushterova, T. Lasalle, N. Sharma, B. C. Russo, M. Rojas-Lopez, M. Sade-Feldman, K. Manakongtreecheep, J. Tantivit, M. Fisher Thomas; MGH COVID-19 Collection & Processing Team, Profiling SARS-CoV-2 HLA-I peptidome reveals T cell epitopes from out-of-frame ORFs. Cell 184, 3962–3980.e17 (2021).
8
S. W.-G. D.-Y. Chen, S. Sarkizova, H. B. Taylor, M. Gentili, L. R. Pearlman, M. R. Bauer, C. M. Rice, K. R. Clauser, N. Hacohen, S. A. Carr, J. G. Abelin, M. Saeed, P. C. Sabeti, The HLA-II immunopeptidome of SARS-CoV-2. Cell Rep. 43, 113596 (2024).
9
A. Chen, Y. F. Kao, C. M. Brown, Translation of the first upstream ORF in the hepatitis B virus pregenomic RNA modulates translation at the core and polymerase initiation codons. Nucleic Acids Res. 33, 1169–1181 (2005).
10
P. S. Gould, A. J. Easton, Coupled translation of the respiratory syncytial virus M2 open reading frames requires upstream sequences. J. Biol. Chem. 280, 21972–21980 (2005).
11
R. S. Shabman, T. Hoenen, A. Groseth, O. Jabado, J. M. Binning, G. K. Amarasinghe, H. Feldmann, C. F. Basler, An upstream open reading frame modulates ebola virus polymerase translation and virus replication. PLOS Pathog. 9, e1003147 (2013).
12
N. T. Ingolia, S. Ghaemmaghami, J. R. S. Newman, J. S. Weissman, Genome-wide analysis in vivo of translation with nucleotide resolution using ribosome profiling. Science 324, 218–223 (2009).
13
N. T. Ingolia, L. F. Lareau, J. S. Weissman, Ribosome profiling of mouse embryonic stem cells reveals the complexity and dynamics of mammalian proteomes. Cell 147, 789–802 (2011).
14
N. T. Ingolia, J. A. Hussmann, J. S. Weissman, Ribosome Profiling: Global Views of Translation. Cold Spring Harb. Perspect. Biol. 11, a032698 (2019).
15
S. Weingarten-Gabbay, S. Elias-Kirma, R. Nir, A. A. Gritsenko, N. Stern-Ginossar, Z. Yakhini, A. Weinberger, E. Segal, Comparative genetics. Systematic discovery of cap-independent translation sequences in human and viral genomes. Science 351, aad4939 (2016).
16
S. Weingarten-Gabbay, R. Nir, S. Lubliner, E. Sharon, Y. Kalma, A. Weinberger, E. Segal, Systematic interrogation of human promoters. Genome Res. 29, 171–183 (2019).
17
J. J. Seo, S.-J. Jung, J. Yang, D.-E. Choi, V. N. Kim, Functional viromic screens uncover regulatory RNA elements. Cell 186, 3291–3306.e21 (2023).
18
N. J. McGlincy, N. T. Ingolia, Transcriptome-wide measurement of translation by ribosome profiling. Methods 126, 112–129 (2017).
19
F. Erhard, A. Halenius, C. Zimmermann, A. L’Hernault, D. J. Kowalewski, M. P. Weekes, S. Stevanovic, R. Zimmer, L. Dölken, Improved Ribo-seq enables identification of cryptic translation events. Nat. Methods 15, 363–366 (2018).
20
N. Stern-Ginossar, B. Weisburd, A. Michalski, V. T. K. Le, M. Y. Hein, S.-X. Huang, M. Ma, B. Shen, S.-B. Qian, H. Hengel, M. Mann, N. T. Ingolia, J. S. Weissman, Decoding human cytomegalovirus. Science 338, 1088–1093 (2012).
21
K. Ito, S. Chiba, Arrest peptides: Cis-acting modulators of translation. Annu. Rev. Biochem. 82, 171–202 (2013).
22
J. Chen, A.-D. Brunner, J. Z. Cogan, J. K. Nuñez, A. P. Fields, B. Adamson, D. N. Itzhak, J. Y. Li, M. Mann, M. D. Leonetti, J. S. Weissman, Pervasive functional translation of noncanonical human open reading frames. Science 367, 1140–1146 (2020).
23
H. D. Hickman, J. W. Mays, J. Gibbs, I. Kosik, J. G. Magadán, K. Takeda, S. Das, G. V. Reynoso, B. F. Ngudiankama, J. Wei, J. P. Shannon, D. McManus, J. W. Yewdell, Influenza A Virus Negative Strand RNA Is Translated for CD8+ T Cell Immunosurveillance. J. Immunol. 201, 1222–1228 (2018).
24
N. J. Maness, A. D. Walsh, S. M. Piaskowski, J. Furlott, H. L. Kolar, A. T. Bean, N. A. Wilson, D. I. Watkins, CD8+ T cell recognition of cryptic epitopes is a ubiquitous feature of AIDS virus infection. J. Virol. 84, 11569–11574 (2010).
25
T. Ouspenskaia, T. Law, K. R. Clauser, S. Klaeger, S. Sarkizova, F. Aguet, B. Li, E. Christian, B. A. Knisbacher, P. M. Le, C. R. Hartigan, H. Keshishian, A. Apffel, G. Oliveira, W. Zhang, S. Chen, Y. T. Chow, Z. Ji, I. Jungreis, S. A. Shukla, S. Justesen, P. Bachireddy, M. Kellis, G. Getz, N. Hacohen, D. B. Keskin, S. A. Carr, C. J. Wu, A. Regev, Unannotated proteins expand the MHC-I-restricted immunopeptidome in cancer. Nat. Biotechnol. 40, 209–217 (2022).
26
M. V. Ruiz Cuevas, M.-P. Hardy, J. Hollý, É. Bonneil, C. Durette, M. Courcelles, J. Lanoix, C. Côté, L. M. Staudt, S. Lemieux, P. Thibault, C. Perreault, J. W. Yewdell, Most non-canonical proteins uniquely populate the proteome or immunopeptidome. Cell Rep. 34, 108815 (2021).
27
S. R. Starck, N. Shastri, Nowhere to hide: Unconventional translation yields cryptic peptides for immune surveillance. Immunol. Rev. 272, 8–16 (2016).
28
J. Holly, J. W. Yewdell, Game of Omes: Ribosome profiling expands the MHC-I immunopeptidome. Curr. Opin. Immunol. 83, 102342 (2023).
29
E. Lorente, A. Barriga, E. Barnea, C. Palomo, J. García-Arriaza, C. Mir, M. Esteban, A. Admon, D. López, Immunoproteomic analysis of a Chikungunya poxvirus-based vaccine reveals high HLA class II immunoprevalence. PLOS Negl. Trop. Dis. 13, e0007547 (2019).
30
S. Sarkizova, S. Klaeger, P. M. Le, L. W. Li, G. Oliveira, H. Keshishian, C. R. Hartigan, W. Zhang, D. A. Braun, K. L. Ligon, P. Bachireddy, I. K. Zervantonakis, J. M. Rosenbluth, T. Ouspenskaia, T. Law, S. Justesen, J. Stevens, W. J. Lane, T. Eisenhaure, G. Lan Zhang, K. R. Clauser, N. Hacohen, S. A. Carr, C. J. Wu, D. B. Keskin, A large peptidome dataset improves HLA class I epitope prediction across most of the human population. Nat. Biotechnol. 38, 199–209 (2020).
31
A. G. Hinnebusch, I. P. Ivanov, N. Sonenberg, Translational control by 5′-untranslated regions of eukaryotic mRNAs. Science 352, 1413–1416 (2016).
32
D. E. Andreev, P. B. F. O’Connor, C. Fahey, E. M. Kenny, I. M. Terenin, S. E. Dmitriev, P. Cormican, D. W. Morris, I. N. Shatsky, P. V. Baranov, Translation of 5′ leaders is pervasive in genes resistant to eIF2 repression. eLife 4, e03971 (2015).
33
Y. Finkel, D. Schmiedel, J. Tai-Schmiedel, A. Nachshon, R. Winkler, M. Dobesova, M. Schwartz, O. Mandelboim, N. Stern-Ginossar, Comprehensive annotations of human herpesvirus 6A and 6B genomes reveal novel and conserved genomic features. eLife 9, e50960 (2020).
34
T. E. Dever, I. P. Ivanov, M. S. Sachs, Conserved Upstream Open Reading Frame Nascent Peptides That Control Translation. Annu. Rev. Genet. 54, 237–264 (2020).
35
N. Stern-Ginossar, S. R. Thompson, M. B. Mathews, I. Mohr, Translational Control in Virus-Infected Cells. Cold Spring Harb. Perspect. Biol. 11, a033001 (2019).
36
I. Goodfellow, The genome-linked protein VPg of vertebrate viruses - a multifaceted protein. Curr. Opin. Virol. 1, 355–362 (2011).
37
A. G. Hinnebusch, Translational regulation of GCN4 and the general amino acid control of yeast. Annu. Rev. Microbiol. 59, 407–450 (2005).
38
L. A. Ryabova, M. M. Pooggin, T. Hohn, Viral strategies of translation initiation: Ribosomal shunt and reinitiation. Prog. Nucleic Acid Res. Mol. Biol. 72, 1–39 (2002).
39
P. Schult, H. Roth, R. L. Adams, C. Mas, L. Imbert, C. Orlik, A. Ruggieri, A. M. Pyle, V. Lohmann, microRNA-122 amplifies hepatitis C virus translation by shaping the structure of the internal ribosomal entry site. Nat. Commun. 9, 2613 (2018).
40
I. Brierley, S. Pennell, R. J. C. Gilbert, Viral RNA pseudoknots: Versatile motifs in gene expression and replication. Nat. Rev. Microbiol. 5, 598–610 (2007).
41
K. Schubert, E. D. Karousis, A. Jomaa, A. Scaiola, B. Echeverria, L.-A. Gurzeler, M. Leibundgut, V. Thiel, O. Mühlemann, N. Ban, SARS-CoV-2 Nsp1 binds the ribosomal mRNA channel to inhibit translation. Nat. Struct. Mol. Biol. 27, 959–966 (2020).
42
C. B. Matranga, K. G. Andersen, S. Winnicki, M. Busby, A. D. Gladden, R. Tewhey, M. Stremlau, A. Berlin, S. K. Gire, E. England, L. M. Moses, T. S. Mikkelsen, I. Odia, P. E. Ehiane, O. Folarin, A. Goba, S. H. Kahn, D. S. Grant, A. Honko, L. Hensley, C. Happi, R. F. Garry, C. M. Malboeuf, B. W. Birren, A. Gnirke, J. Z. Levin, P. C. Sabeti, Enhanced methods for unbiased deep sequencing of Lassa and Ebola RNA viruses from clinical and biological samples. Genome Biol. 15, 519 (2014).
43
X. Adiconis, D. Borges-Rivera, R. Satija, D. S. DeLuca, M. A. Busby, A. M. Berlin, A. Sivachenko, D. A. Thompson, A. Wysoker, T. Fennell, A. Gnirke, N. Pochet, A. Regev, J. Z. Levin, Comparative analysis of RNA sequencing methods for degraded or low-input samples. Nat. Methods 10, 623–629 (2013).
44
A. L. Balish, J. M. Katz, A. I. Klimov, Influenza: propagation, quantification, and storage. Curr. Protoc. Microbiol. 15, mc15g01s3 (2013).
45
U. J. Buchholz, S. Finke, K. K. Conzelmann, Generation of bovine respiratory syncytial virus (BRSV) from cDNA: BRSV NS2 is not essential for virus replication in tissue culture, and the human RSV leader region acts as a functional BRSV genome promoter. J. Virol. 73, 251–259 (1999).
46
D. H. Schott, D. K. Cureton, S. P. Whelan, C. P. Hunter, An antiviral role for the RNA interference machinery in Caenorhabditis elegans. Proc. Natl. Acad. Sci. U.S.A. 102, 18420–18424 (2005).
47
J. A. Horwitz, M. Dorner, T. Friling, B. M. Donovan, A. Vogt, J. Loureiro, T. Oh, C. M. Rice, A. Ploss, Expression of heterologous proteins flanked by NS3-4A cleavage sites within the hepatitis C virus polyprotein. Virology 439, 23–33 (2013).
48
B. D. Lindenbach, M. J. Evans, A. J. Syder, B. Wölk, T. L. Tellinghuisen, C. C. Liu, T. Maruyama, R. O. Hynes, D. R. Burton, J. A. McKeating, C. M. Rice, Complete replication of hepatitis C virus in cell culture. Science 309, 623–626 (2005).
49
N. T. Ingolia, G. A. Brar, S. Rouskin, A. M. McGeachy, J. S. Weissman, The ribosome profiling strategy for monitoring translation in vivo by deep sequencing of ribosome-protected mRNA fragments. Nat. Protoc. 7, 1534–1550 (2012).
(0)eLetters
eLetters is a forum for ongoing peer review. eLetters are not edited, proofread, or indexed, but they are screened. eLetters should provide substantive and scholarly commentary on the article. Neither embedded figures nor equations with special characters can be submitted, and we discourage the use of figures and equations within eLetters in general. If a figure or equation is essential, please include within the text of the eLetter a link to the figure, equation, or full text with special characters at a public repository with versioning, such as Zenodo. Please read our Terms of Service before submitting an eLetter.
Log In to Submit a ResponseNo eLetters have been published for this article yet.
Information & Authors
Information
Published In
Science
Volume 388 | Issue 6752
12 June 2025
12 June 2025
Copyright
Copyright © 2025 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original U.S. Government Works.
This is an article distributed under the terms of the Science Journals Default License.
Article versions
Submission history
Received: 15 February 2024
Accepted: 10 March 2025
Published in print: 12 June 2025
Acknowledgments
We thank T. Ouspenskaia, A. Nachshon, S. Reilly, J. Xue, A. Lin, H. Metsky, A. Lercher, and O. Mizrahi for many valuable discussions. We thank N. McGlincy and N. Ingolia for sharing their detailed ribosome profiling protocol with the broad community of researchers (18), which contributed to the development of MPRP.
Funding:
This study was supported in part by grants from the National Institute of Health (NIH) (U19AI110818 to P.C.S., P01CA206978 to S.A.C., and R01CA057973 to C.M.R.), the United States Department of Agriculture (58-3022-2-031 to P.C.S.), National Cancer Institute (NCI) Clinical Proteomic Tumor Analysis Consortium program (U24CA270823 and U01CA271402 to S.A.C.), and the Dr. Miriam and Sheldon G. Adelson Medical Research Foundation (to S.A.C.). S.W.-G. is the recipient of a Human Frontier Science Program fellowship (LT-000396/2018), EMBO nonstipendiary long-term fellowship (ALTF 883-2017), the Gruss-Lipper postdoctoral fellowship, the Zuckerman STEM Leadership Program fellowship, and the Rothschild Postdoctoral Fellowship.
Author Contributions:
S.W.-G. conceptualized the study, designed the synthetic libraries, designed the experiments, performed experiments, analyzed the data, and wrote the manuscript. M.R.B., A.C.S., Y.Y., C.A.F., N.L.W., and C.K.B. performed the experiments. S.K., E.K.V., D.L., K.R.C., S.A.C., and J.G.A. analyzed the mass spectrometry data of HLA-I immunopeptidomes. L.E.H. and C.M.R. contributed to data interpretation. P.C.S. supervised the work and wrote the manuscript.
Competing Interests:
S.W.-G., M.R.B., A.C.S., and P.C.S. are named as co-inventors on International Patent Application PCT/US2024/048478, claiming priority to US Provisional Application 63/540,279 related to this work filed by The Broad Institute that covers massively parallel methods and techniques for evaluating open reading frames in genomes, particularly viral genomes. These methods and data unveil new proteins, immune targets, and cis-regulatory elements that can serve in the design of vaccines and commercial overexpression platforms. M.R.B. is an employee of BioNTech SE. S.K. is an employee of Genentech. S.A.C. is a member of the scientific advisory boards of Kymera, PTM BioLabs, Seer, and PrognomIQ. J.G.A. is a paid consultant for Enara Bio and Moderna. P.C.S. is a cofounder of and consultant to Sherlock Biosciences and Delve Biosciences, is on the Board of Directors and a shareholder of Danaher Corporation and Polaris Genomics, and holds equity in all the companies. All other authors declare no competing interests.
Data and Materials Availability:
The raw sequencing data generated in this study have been submitted to the Gene Expression Omnibus (GEO; http://www.ncbi.nlm.nih.gov/geo/) under accession number GSE272406.
License Information:
Copyright © 2025 the authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original US government works. This article is subject to HHMI’s Open Access to Publications policy. HHMI lab heads have previously granted a nonexclusive CC BY 4.0 license to the public and a sublicensable license to HHMI in their research articles. Pursuant to those licenses, the Author Accepted Manuscript (AAM) of this article can be made freely available under a CC BY 4.0 license immediately upon publication. http://www.science.org/content/page/science-licenses-journal-article-reuse
Authors
Funding Information
National Institutes of Health: P01CA206978
National Cancer Institute: U24CA270823
National Cancer Institute: U01CA271402
U.S. Department of Agriculture: 58-3022-2-031
Metrics & Citations
Metrics
Article Usage
- 0 citation in Scopus
- 0 citation in Web of Science
Altmetrics
Citations
Cite as
- Shira Weingarten-Gabbay et al.
Export citation
Select the format you want to export the citation of this publication.
View Options
View options
PDF format
Download this article as a PDF file
Download PDFLog in to view the full text
AAAS login provides access to Science for AAAS Members, and access to other journals in the Science family to users who have purchased individual subscriptions.
More options
Purchase digital access to this article
Download and print this article for your personal scholarly, research, and educational use.
Buy a single issue of Science for just $15 USD.
