Unveiling conserved HIV-1 open reading frames encoding T cell antigens using ribosome profiling

Bertrand, Lisa; Nelde, Annika; Ramirez, Bertha Cecilia; Hatin, Isabelle; Arbes, Hugo; François, Pauline; Demais, Stéphane; Labaronne, Emmanuel; Decimo, Didier; Guiguettaz, Laura; Grégoire, Sylvie; Bet, Anne; Beauclair, Guillaume; Gross, Antoine; Ziegler, Maja C.; Pereira, Mathias; Jeger-Madiot, Raphaël; Verdier, Yann; Vinh, Joelle; Cardinaud, Sylvain; Graff-Dubois, Stéphanie; Esclatine, Audrey; Gouttefangeas, Cécile; Altfeld, Marcus; Hocqueloux, Laurent; Samri, Assia; Autran, Brigitte; Lambotte, Olivier; Rammensee, Hans-Georg; Ricci, Emiliano P.; Walz, Juliane; Namy, Olivier; Moris, Arnaud

doi:10.1038/s41467-025-56773-2

Download PDF

Article
Open access
Published: 18 February 2025

Unveiling conserved HIV-1 open reading frames encoding T cell antigens using ribosome profiling

Nature Communications volume 16, Article number: 1707 (2025) Cite this article

6425 Accesses
6 Citations
11 Altmetric
Metrics details

Subjects

An Author Correction to this article was published on 06 March 2025

This article has been updated

Abstract

The development of ribosomal profiling (Riboseq) revealed the immense coding capacity of human and viral genomes. Here, we used Riboseq to delineate the translatome of HIV-1 in infected CD4⁺ T cells. In addition to canonical viral protein coding sequences (CDSs), we identify 98 alternative open reading frames (ARFs), corresponding to small Open Reading Frames (sORFs) that are distributed across the HIV genome including the UTR regions. Using a database of HIV genomes, we observe that most ARF amino-acid sequences are likely conserved among clade B and C of HIV-1, with 8 ARF-encoded amino-acid sequences being more conserved than the overlapping CDSs. Using T cell-based assays and mass spectrometry-based immunopeptidomics, we demonstrate that ARFs encode viral polypeptides. In the blood of people living with HIV, ARF-derived peptides elicit potent poly-functional T cell responses mediated by both CD4⁺ and CD8⁺ T cells. Our discovery expands the list of conserved viral polypeptides that are targets for vaccination strategies and might reveal the existence of viral microproteins or pseudogenes.

The translational landscape of HIV-1 infected cells reveals key gene regulatory principles

Article Open access 15 January 2025

Therapeutic vaccination following early antiretroviral therapy elicits highly functional T cell responses against conserved HIV-1 regions

Article Open access 11 October 2023

Phenotypic characterization of single CD4+ T cells harboring genetically intact and inducible HIV genomes

Article Open access 27 February 2023

Introduction

Open reading frames (ORFs) of the human genome have been long annotated using restricted criteria, including the presence of canonical start (AUG) and stop codons, and the potential to encode protein longer than 100 amino-acids (aa)¹. Recent advances in detection methods challenged this definition revealing that thousands of small ORFs (sORFs) encode polypeptides or microproteins shorter than 100 aa². These sORFs are widely spread along the human genome, some overlap classical ORF, on a different frame (so-called alternative reading frames (ARFs)), but the majority is found within 5’ untranslated regions (UTR) of known genes³. Although some microproteins encoded by sORFs have been shown to play fundamental roles, for instance, in DNA repair, mitochondrial functions, RNA regulation, etc⁴., the biological functions of most remain enigmatic. Nonetheless, the thorough characterization of sORFs greatly diversified the translation landscape in healthy and malignant cells expanding, in particular, the sources of cancer antigens.

To date the characterization of sORFs mostly relies on ribosomal profiling (Riboseq) that allows unbiased assessment of actively translated mRNA sequences. However, only a limited number of studies correlated Riboseq data with the generation of peptides⁵. Combining Riboseq and mass spectrometry (MS)-based immunopeptidomics several studies revealed that a fraction of peptides presented by major histocompatibility complex (MHC) molecules (known as ligandome) is derived from sORFs encoded polypeptides that are specific or overrepresented in tumor cells^6,7,8,9. Other studies demonstrated that the tumor immune microenvironment is likely to favor the expression of human endogenous retroviruses (hERVs) that are normally silent in healthy tissues and of out-of-frame polypeptides, both constituting antigen sources for T cell immunity^10,11. Remarkably, these findings and others¹² highlighted that noncanonical translation events of sORF located in 5′UTR or out-of-frame are likely favored in the context of stress.

Viral infections represent a major stress to the cell whose translation machinery is mobilized for the translation of viral RNAs. Viruses acquired specific features to control viral mRNA translation such as sequences favouring frameshifting or internal entry sites of ribosomes¹³. Thereafter, not surprisingly during the last decades, peptides derived from alternative translation events were described in cells infected with influenza virus^14,15,16, murine leukemia virus (MLV)¹⁷, and human immunodeficiency virus-1 (here referred to as HIV)^18,19,20.

The expression of these noncanonical viral peptides often referred to as cryptic epitopes (CE), was mainly studied in the context of peptide presentation by MHC class I (MHC-I) molecules to cytotoxic CD8⁺ T cells (CTLs). To date, in the context of viral infections, only a limited number of studies have combined Riboseq and immunopeptidomic approaches to characterize the landscape and origin of viral peptides presented by MHC-I molecules. These studies confirmed that MHC-I molecules present peptides from unannotated ORFs within ARFs of the human cytomegalovirus (HCMV)²¹ and of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)²².

In the context of HIV infection, the existence of peptides derived from ARFs (ARFP) was, so far, revealed using T cell-based assays^{20,23,24,25,26}. CTLs specific to peptides derived from ARFs were indeed detected during the acute and chronic phases of infection^18,19,20,24. ARFP-specific T cells seem preferentially abundant in people living with HIV (PLWH) with favorable clinical outcomes expressing protective human leukocyte antigen (HLA) alleles²³. Further underlining the contribution of ARFP-specific CTLs in the control of viral replication, several labs have shown that HIV adapts to and escapes from ARFP-specific T cell responses by introducing mutations within ARF sequences^18,19,20,24. Remarkably, in the macaque model of simian immunodeficiency virus (SIV) infection, a single mutation within an ARF-derived epitope was strongly associated with viral rebound²⁷. Bioinformatics approaches analysing the association between HLA polymorphisms and HIV sequence variations (HLA footprint) revealed that the virus might produce peptides from ARFs buried within sense (5’ to 3’) or antisense (3’ to 5’) frames of the viral mRNA^18,19.

In the present study, we used Riboseq to delineate the translatome of HIV in infected CD4⁺ T cells. In addition to canonical viral protein-coding sequences (CDSs), we identify 98 ARFs, corresponding to sORFs that are distributed across the HIV genome including the 5′ UTR region. Using a database of complete and unique HIV genomes, we show that most ARF aa sequences are likely conserved among HIV-1 clade B and C strains, with eight ARF aa sequences being more conserved than the overlapping aa CDS sequences. Using T cell-based assays and mass spectrometry-based immunopeptidomics, we demonstrate that 43 ARFs encode viral polypeptides. In peripheral blood mononuclear cells (PBMCs) of PLWH, ARF-derived peptides elicit potent and polyfunctional T cell responses mediated by both CD4⁺ and CD8⁺ T cells. Our findings broaden the spectrum of HIV immunogenic antigens that might help in the design of efficacious vaccines and might reveal the existence of viral microproteins or pseudogenes.

Results

HIV ribosome profiling reveals 98 alternative reading frames

To date, the existence of ARFs in the HIV genome has been highlighted using indirect approaches such as T cell-based assays and HLA footprint analyses. We thus intended to provide an unbiased assessment of actively translated viral mRNA sequences in HIV-infected cells. To this end, we infected the SupT1 CD4⁺ T cell line with HIV_NL4-3, in three biological replicates, and analysed viral mRNA translation using Riboseq (Supplementary Fig. 1a). Twenty-four hours post-infection, the viability of the cells and the infection rates were analysed using a viability dye and a combination of anti-HIV-Gag and anti-CD4 antibodies. Combining Gag and CD4 staining allowed the direct detection of cells expressing HIV-Gag protein and the indirect assessment of cells expressing viral proteins that downmodulate CD4 expression such as HIV-Nef, -Vpu, or -Env²⁸. Among the three biological replicates, the viability of the cells ranged from 44 to 71%, and the rate of infected Gag⁺CD4^- and Gag⁺CD4⁺ cells from 54 to 73% (Supplementary Fig. 1b). The cells were harvested, and cellular and viral polysomes extracted. The ribosome-protected fragments (RPFs) were then isolated and sequenced (Supplementary Fig. 1a). Data were analysed using the RiboDoc tool²⁹ and inhouse bioinformatics pipelines. The length of isolated RPFs ranged between 25 and 35 nucleotides (nts). We analysed the periodicity of the reads by aligning RPFs to annotated start codons of human CDSs. The reads, in particular the 28- and 29-mers, uniformly aligned along the human CDSs with a periodicity of three nucleotides corresponding to codon triplets, thus strongly suggesting that the RPFs correspond to bona fide translation events (Supplementary Fig. 1c). The data from the three biological replicates were then combined and viral RPFs aligned to the HIV_NL4-3 genome (Fig. 1a and Supplementary Fig. 2a, b). The ribosome footprints aligned only on the forward strand throughout the viral genome. Note that in Fig. 1, each reading frame is highlighted by a specific color: brown, red, and blue for Frame 0 (F0), 1 (F1), and 2 (F2), respectively (Fig. 1a). For clarity, the RPF occupancy for the different reading frames is also presented in Supplementary Fig. 2b in three separate graphs. The number and density of RPFs were particularly high in the UTR regions of HIV RNAs (Fig. 1a). As expected, we observed patches of RPFs in the frame and covering the CDSs of known viral proteins such as HIV-Gag and -Env (Fig. 1a, in brown and red, respectively). The decrease in density and number of RPFs covering the CDS of HIV-Pol also highlighted the specific feature of the HIV-GagPol polyprotein expression, which is controlled by a -1 ribosomal frameshifting during translation elongation³⁰. Based on the number of RPFs and the length of the CDSs, we calculated the relative expression of each CDS (Fig. 1b). The relative expression of HIV-GagPol polyprotein was about 6% of HIV-Gag, which is consistent with the literature³⁰. Thereafter, our ribosome profiling of HIV translation identifies all known HIV CDSs and confirms specific features of the regulation of HIV translation such as ribosome frameshifting.

**Fig. 1: The translational landscape of HIV-1: identification of ARFs.**

Strikingly, we also observed patches of RPFs overlapping known CDSs but aligning on other reading frames (e.g., patches of RPFs in frame-2 (red) overlapping HIV-Gag CDS or RPFs in frame-2 (blue) overlapping HIV-Vif CDS) suggesting translation of ARFs in the HIV genome (Fig. 1a and Supplementary Fig. 2). In our study having no information on potential start codons, we defined an ARF as a sequence between two stop codons with minimal size of ten codons (shorter sequences being unlikely to be processed to produce MHC ligands). In order to discriminate between background whole RNA sequencing and bona fide translation events, we then determined the number of reads in a frame within each potential ARF, which allowed calculating a mean of RPF/codon for each ARF. We observed that 75% of all ARFs had less than 8 RPF/codon and decided to use this value as the minimum threshold to define the bona fide translation of ARFs. Based on these criteria, we identified 98 HIV ARF sequences distributed across the viral genome (Fig. 1a bottom panel purple lines, 1c, d and Supplementary Table 1). Examples of potential ARFs that failed or succeeded in passing our selection criteria are presented in Supplementary Fig. 2c. Note that ARFs were named based on the CDS that they overlap plus a specific number (e.g., HIV-GAG15). In Fig. 1d, four examples of selected ARFs overlapping HIV-Gag, -Vif, -Env, and -Nef CDSs are highlighted, named ARF-GAG15, ARF-VIF3, ARF-ENV4, and ARF-NEF7, respectively. Note that we also analysed the presence of these 98 ARFs in each dataset of the biological replicates: 86 ARFs were common to at least two replicates, and 12 were found in only one of the replicates. Nonetheless, since these 12 ARFs exhibited a very high coverage, we decided to maintain them in our study. Interestingly, 16 out of 98 ARFs were previously described in the literature using indirect approaches^18,19,25,26 (Supplementary Table 1). We also identified nine ARFs in the 5’ UTR region of the HIV genome, commonly referred to as uORF, which is consistent with previous studies describing translation events in the 5’ UTR regions of different viral mRNA^31,32,33,34. Note that we observed a large number of RPFs on the 5’ UTR region upstream of the major splice site (SD1) (Fig. 1 and Supplementary Figs. 2, 3).

In a second set of Riboseq experiments, using lactimidomycin (LTM), a drug that stops the initiation step of translation and thus leads to the accumulation of ribosomes at start sites, we then define the start codons of each ARF. To further enrich ribosomes at start codons, LTM was combined with puromycin (PMY) to dissociate elongating ribosomes from mRNAs³⁵. As previously, SupT1 CD4⁺ T cells were infected with HIV_NL4-3, in three biological replicates, cell viability and infection rates were analysed as in Supplementary Fig. 1b. Cells were treated with LTM and PMY prior harvest, RPF isolation and sequencing. The RPFs of cellular mRNAs were analyzed for periodicity and metagene. Cellular RPFs were highly enriched at start codons, thus strongly suggesting that the combined treatment of LTM and PMY allowed the accumulation of ribosomes at start codons in treated cells (Supplementary Fig. 4). Viral RPFs were then aligned to the viral genome, and RPFs in frame with the newly identified ARFs analysed. We obtained heterogenous results with 71 ARFs where multiple RPF occupancy could be observed and 27 ARFs where it was not possible to distinguish them from the background. Taking into account the three replicates, we observed that the mean intensity of RPFs for all ARFs was 62 reads, and we decided to use this value as the minimum threshold. We also excluded RPFs that accumulated 15 nucleotides downstream of stop codons. Based on these restrictive criteria, we listed the potential initiation sites that are shared by the three biological replicates (Supplementary Table 2). For 26 ARFs, a single start position was identified, while for the others, one or two patches of RPFs were observed (Supplementary Table 2). Note that in the same ARF previously defined from stop-to-stop, translation might start at different initiating sites. Interestingly, among all potential initiation codons, 18 % were closely related to the classical ATG methionine start codon (near-cognate).

Using Riboseq, we, therefore, reveal the existence of 98 ARFs distributed across the HIV genome that are actively translated in infected CD4⁺ T cells. We confirm and extend previous work showing that other regions than annotated HIV CDSs are indeed translated.

Identified ARFs are conserved among HIV clade B and C isolates

Assuming that these ARF-encoded peptides might be a reservoir of genetic novelty and a source of T cell antigens, we analysed their aa sequence conservation among clade B and C HIV isolates. Using the Los Alamos HIV Sequence Database, we created in-house databases of clade B and C HIV sequences containing complete HIV genome without aberrant mutations within known CDSs and isolated from different individuals. We obtained 1609 and 411 clade B and C sequences, respectively. Using UGENE software, we then aligned the ARFs that we identified in the HIV_NL4-3 genome to the databases (Fig. 2). Owing to the diversity of potential translation initiation sites for most ARFs, we decided to perform this analysis using the ARF aa sequence from stop-to-stop. The ARF-encoded peptide aa sequence conservation ranged from 48% to 98% with a median of 88% among the clade B (Fig. 2A) and from 42% to 98% with a median of 89% among the clade C (Fig. 2B). We did not notice a particular trend in aa sequence conservation when comparing the reading frames in which the ARFs are located in relation to the overlapping CDS (Supplementary Fig. 5a). In contrast, statistically significant differences in aa sequence conservation could be observed depending on the overlapping CDS (Supplementary Fig. 5b). The most significant difference in aa sequence conservation was found between ARFs overlapping pol and env genomic regions, which is consistent with the fact that env CDS is the most variable region. There was no difference in ARF-encoded peptide aa sequence conservation between clades B and C regardless of the reading frame from which the ARFs are translated (Supplementary Fig. 5c). Finally, we did not observe a correlation between ARF-encoded peptide aa sequence conservation and ARF relative expression (Supplementary Fig. 5d).

**Fig. 2: Amino-acid conservation of ARF-encoded peptide sequences among HIV-1 clade B and C strains.**

We then investigated whether it is the conservation of the CDS that imposes the conservation of the ARFs. To this end, we analysed the percentage of aa sequence conservation of the overlapping CDS and of the third frame (i.e., corresponding neither to the CDSs nor to the ARFs sequences). To simplify the analysis, ARFs that overlap two ORFs were excluded. We observed significantly higher conservation of CDS regions compared to the ARF and third frame, with a median of 93, 88, and 88% aa conservation, respectively (Fig. 2C and Supplementary Fig. 5e). Interestingly, the most variable ARF among clades B and C, ARF-ENV3, overlaps the highly variable V1-loop of HIV-Env glycoprotein³⁶ (Fig. 2). These observations further highlight that ORFs encoding the known viral proteins required for viral replication are under positive selection. In addition, there is no statistical difference in aa sequence conservation between the ARFs and the third frames (median of 88% aa conservation for both, Supplementary Fig. 5e), which might imply that there is a low or no selection pressure on ARFs. However, it is particularly interesting to note that eight ARFs sequences are more conserved than their overlapping ORF: ARF-GAG9, -VIF5, -VPU1, -ENV4, -ENV9, -ENV20, -ENV21, and -NEF3 (Fig. 2C, ARF names highlighted with black dots).

Overall, our HIV sequence analysis strongly suggests that the 98 noncanonical transcripts identified by ribosome profiling are likely conserved among HIV clade B and C clinical isolates. Although, most ARFs are probably not under selective evolutionary pressure, eight ARFs are more conserved than their overlapping CDSs.

To determine whether all these ARFs readily encode viral polypeptides, we then used two complementary approaches, monitoring immune responses to ARF-derived polypeptides using PBMCs of PLWH and using biochemical immunopeptodomics approaches to isolate directly, in infected cells, ARF-derived peptides.

42 ARFs encode viral polypeptides eliciting ARF-specific T cell responses

We first selected a set of peptides to be synthetized and then tested their capacity to elicit T cell responses using PBMCs of PLWH, using a cultured IFNγ-ELISPOT assay (Supplementary Fig. 6a). ARF-derived peptides (ARFP) were selected based on their aa sequence conservation among HIV isolates and their predicted capacity to bind HLA molecules exhibiting high prevalence in the general population (Supplementary Table 3). For ARFs with long potential aa sequences, several peptides were synthetized and tested (Supplementary Table 3). The peptide length varied in order to include as much as possible multiple potential epitopes within the same aa sequences. We then monitored T cell responses to these ARFP, using blood samples from individuals under antiretroviral treatment (ART) or individuals who naturally control viral replication without treatment, so-called elite controllers (EC) (Supplementary Table 4). In total, 96 different ARFP were synthesized and distributed in nine different peptide pools (Supplementary Table 3). As a control, we included peptides derived from canonical HIV proteins (HIV) and from non-HIV common viral peptides (CEF). The HIV-canonical peptide pool was composed of previously characterized T cell epitopes³⁷ (Supplementary Table 5).

PBMCs from PLWH were seeded in 24-well plates and loaded with the different peptide pools. Seven days later, a fraction of the cells was loaded with the peptide pool used for the initial culture, and T cell reactivity was monitored using an IFNγ-ELISPOT assay (Supplementary Fig. 6a, b). Although with different magnitudes, the PBMCs from all donors reacted in a specific manner to the pool of peptides derived from canonical HIV proteins (HIV+) compared to the negative control (HIV-) (i.e cells from the same T-cell cultures but loaded with DMSO-containing medium during the ELISPOT assay) (Supplementary Fig. 6b). Out of the seven samples from PLWH, six reacted to at least one ARF-derived peptide pool (Supplementary Fig. 6b). Cells from donor EC-1 reacted to eight ARFP pools. In contrast, we did not identify ARFP-specific response for donor ART-3, due to the high IFNγ secretion observed in negative controls (wells without peptide re-stimulation (−)). The response to the pool of classical HIV peptides was also particularly low in this donor (HIV+) (Supplementary Fig. 6b). Overall, several ARFP pools (including Pool-1, -2, -3, -4, and -6) induced very strong and specific IFNγ secretions in PBMCs from multiple HIV-positive donors (Supplementary Fig. 6b). The cells were then allowed to expand further by renewing the cytokine cocktail and tested on day 12, after the initial culture, for their capacity to react to individual peptides of the pools tested positive at day 7 (Fig. 3 and Supplementary Fig. 7). Note that despite the high background of IFNγ secretion observed at day 7, we tested the capacity of the cells from donor ART-3 to react to individual peptides of the pools 4, 6, and 8 because they showed a tendency to induce specific responses at day 7 (see Supplementary Fig. 6, ART-3). In addition, T cell responses where the spots appeared visually particularly large on day 7, even if not exceeding the positivity threshold in terms of spot numbers, were also tested on day 12. The reactivity of each PLWH to the individual ARFP is exhaustively presented in Supplementary Fig. 7. ARFP-induced IFNγ⁺-T cell responses are summarized in Fig. 3 and pictures of the corresponding wells of the IFNγ-ELISPOT plates, are shown in Supplementary Fig. 7.

**Fig. 3: Detection of T cell responses specific to ARF-derived peptides (ARFP) in PBMCs of PLWH.**

At day 12, with the exception of EC-2, all samples from PLWH exhibited very strong or saturated responses to classical HIV peptides. Remarkably, the PBMCs of all donors reacted to at least one individual ARFP (Fig. 3). Among the donors, the median of ARFP recognized was 6. However, responses were very heterogeneous with donor EC-1 reacting to 33 ARFP and EC-2 only to a single one (Fig. 3a). A total of 60 ARFP-specific T cell responses were identified across PLWH samples (Fig. 3a). Fourteen ARFP elicited T cell responses in samples from two different individuals (Fig. 3b). Altogether, the results using samples from PLWH revealed 46 unique ARF-derived peptide-specific T cell responses. The magnitude of T cell responses was heterogeneous between and among samples from PLWH, with an intensity ranging from 13 to above 1250 (saturated) SFU/10⁶ PBMCs (Fig. 3c) with a median of 311 SFU/10⁶ PBMCs, but highly significant over background (Supplementary Fig. 8a). For instance, despite exhibiting a broad ARFP-specific T cell response, the magnitude of most T cell responses from donor EC-1 were low as compared with those from other donors (Fig. 3 and Supplementary Fig. 8b). Finally, we did not observe a significant difference in the magnitude of T cell responses between ART and EC individuals (Supplementary Fig. 8b). Remarkably, ARFP-specific-T cell responses target peptides encoded by ARFs distributed all along the HIV genome sequence from the 5’ UTR to nef CDS (Fig. 3d). Although the magnitude of uORF-specific T cell responses seemed particularly high (Supplementary Fig. 8c), ARPF-induced T cell responses were not significantly dominated by T cells recognizing ARFP from one specific HIV genomic region, neither UTR nor CDS (Supplementary Fig. 8d).

Using the exact same protocol, the entire set of ARF-derived peptide pools was also tested, at day 7, on PBMCs from non-infected donors (n = 3), and on day 12, peptide deconvolution was performed on pools stimulating IFNγ-responses (Supplementary Fig. 9a, b). Cross-reactivity to one peptide pool was observed for EFS 1619 and EFS 5174 samples (Supplementary Fig. 9a) but the peptide cross-reactivity was not confirmed when testing individual peptides of the pool. Donor EFS 0685 exhibited a high background on all tested peptide pools. Peptide deconvolution showed a cross-reactivity to three peptides that were then excluded from the study (Supplementary Fig. 9b).

Overall, we demonstrate here that T-cell responses from PLWH target 46 ARF-derived peptides encoded by 42 different ARFs (Fig. 3d). Of note, for most ARFP that elicited T cell responses, the corresponding aa sequence could be found at least partially in the aa sequence encoded from the translation initiation sites, identified in our LTM + PMY Riboseq experiments (see Supplementary Table 2 in red). We thus reveal that at least 42 ARFs identified in our ribosome profiling encode viral polypeptides and elicit T cell responses in vivo in the course of natural infection.

ARF-specific T cell responses are mediated by CD4⁺ and CD8⁺ T cells with a polyfunctional profile

We then asked whether CD8⁺ and/or CD4⁺ T cells mediate ARFP-specific T cell responses and studied the quality of these responses. The quality of HIV-specific T cell activation, defined as the capacity to produce multiple antiviral cytokines/chemokines, rather than the magnitude of T cell responses has been linked to disease outcome³⁸. Therefore, we characterized the capacity of ARFP-specific T cells to produce MIP-1β, IFNγ, IL2, CD107a, and TNFα using Intracellular Cytokine Staining (ICS) (Fig. 4). PBMCs from two ART and four EC individuals (ART-2 was not tested in ICS due to a lack of sample availability) were seeded in 24-well plates and loaded with the individual ARFPs that induced T cell responses in the previous cultured IFNγ-ELISPOT assay (one ARFP per well). In total, we tested 47 ARF-derived peptides, 45 peptides which induced IFNγ-responses in the cultured IFNγ-ELISPOT assay plus two ARFP (one new ARFP (A05) and one (C12) inducing a slight but not significant T cell response in the first ELIPSOT) (Supplementary Table 6). In addition to the HIV+ peptide pool (HIV+), we included immunodominant peptides derived from classical HIV proteins with known HLA binding capacity³⁷. This pool of immunodominant peptides was adapted to the HLA allotype of each individual (Supplementary Table 7). At day 12, peptides inducing an IFNγ secretion were then used to monitor T cell activation using ICS combined with CD4, CD8, and CD3 extracellular staining (Fig. 4).

**Fig. 4: ARF-derived peptide (ARFP)-specific T cell responses are mediated by CD4⁺ and CD8⁺ T cells exhibiting polyfunctional cytokine profiles.**

As illustrated with the ARF-derived peptide G02 and the HIV-Env-derived peptide env01, peptide loading led to high production levels of MIP-1β, IFNγ, IL2, CD107a, and TNFα by CD4⁺ T cells as compared to the negative controls (NS_ARF and NS_ORF, respectively) (Fig. 4a). Overall, 27 peptides induced the secretion of at least one cytokine (Fig. 4b). Most T cell responses, 89%, were mediated by CD4⁺ T cells (Fig. 4b, c). The ARF-derived peptides A05, G02, and G05 induced both CD8⁺ and CD4⁺ T cells (Fig. 4a right panel and 4b).

CDS-specific CD4⁺ T cell responses were characterized by a significantly higher proportion of cells producing individual cytokines as compared with ARFP-specific CD4⁺ T cell responses (Supplementary Fig. 10a) which was not the case for CD8⁺ T cell responses (Supplementary Fig. 10b). However, although the frequency of responses against ARFP was lower than against CDS-derived peptides, ARFP stimulated polyfunctional responses in both CD4⁺ and CD8⁺ T cells (Fig. 4d, e). Since tri-functional T cell responses are often considered polyfunctional and might be associated with protection from disease progression³⁹, we focused our analysis on activated T cells with a polyfunctional profile with at least 3 functional responses. Clustering the data, we observed that 80% of ARFP-specific CD4⁺ T cells secreted at least three cytokines simultaneously compared to 93% for canonical HIV-specific CD4⁺ T cells (Fig. 4d). Ninety-one percent of CD8⁺ T cell responses directed against canonical HIV peptides and 100% of ARFP-specific CD8⁺ T cell responses harbored a tri-functional phenotype, respectively (Fig. 4e). Profiles and frequencies of cytokines responses induced by each ARF- and CDS-derived peptide for the six PLWH are presented as heatmaps for CD4⁺ and CD8⁺ T cells (Supplementary Fig. 10c, d). It was interesting to note that upon G02 peptide stimulation, the proportion of activated T cells secreting five cytokines simultaneously was higher than for some canonical HIV or even CEF peptide stimulations (see donor EC-3 donor, Supplementary Fig. 10c). In ART and EC individuals, we did not observe a significant difference in the capacity of ARF- and CDS-encoded peptides to induce polyfunctional cytokine productions (at least three functions) among neither CD4⁺ nor CD8⁺ T cells (Fig. 4f, g). There is also no significant influence of the clinical status on the polyfunctional profile of ARF- and CDS-specific T cell responses (Fig. 4f).

Finally, since the cytokine combination secreted by T cells might influence disease outcome⁴⁰, we dissected ARF- and CDS-specific T cell cytokine secretion patterns. To this end, we analyzed the 32 possible cytokine combinations (out of the five cytokines detected: MIP-1β, IL2, INFγ, CD107a, and TNFα) secreted by CD4⁺ and CD8⁺ T cells upon ARF- or CDS-derived peptide stimulation. Although the frequencies of CD4⁺ T cells targeting ARFP were lower than that of CDS-derived peptides, activated CD4⁺ T cells exhibited similar cytokine secretion patterns whether targeting ARF- or CDS-derived peptides (Supplementary Fig. 10e). The majority of polyfunctional ARFP-specific CD4⁺ T cells produced TNFα, CD107a, IFNγ, and MIP-1β simultaneously. Among ARF- and CDS-specific CD4⁺ T cells, the most prevalent monofunctional categories were TNFα⁺ and Mip-1β⁺ cells, respectively (Supplementary Fig. 10e). ARF- and CDS-specific CD8⁺ T cell responses displayed similar pentafunctional profiles (Supplementary Fig. 10e).

We then sought to analyze ex vivo ARF- and CDS-specific T cell responses in PBMCs of PLWH without in vitro expansion. To this end, cells from EC-1 and EC-3 were loaded with the individual peptides that induced a potent T cell activation in cultured IFNγ-ELISPOT and/or ICS (Supplementary Fig. 11). Remarkably for EC-1, four out of the six ARFP tested in ICS induced significant CD4⁺ T cell responses and one peptide stimulated CD8⁺ T cells (Supplementary Fig. 11a, left panel). For donor EC-3, the four ARFP induced significant CD4⁺ T cell responses (Supplementary Fig. 11b, left panel). Overall, the frequency of ex vivo ARF- and CDS- specific T cells was very low, limiting the analysis of the polyfunctional profile of the cells (Supplementary Fig. 11a, b, pie charts on the right panels). Nonetheless, bivalent and pentafunctional ARFP-specific ex vivo T cell responses could be detected for single ARF-derived peptides (Supplementary Fig. 11).

Taken together, our multiparametric study showed that ARF-derived peptides are predominantly recognized by CD4⁺ T cells from ART and EC donors and, to a lesser extent, by CD8⁺ T cells. Following in vitro T cell expansion, the percentage of ARFP-specific T cell activation was lower than CDS-specific T cell responses. Nonetheless, some ARF-derived peptides induced a higher magnitude of T cell activation than CDS-derived peptides. ARF-specific T cells were readily detected ex vivo in the PBMCs of PLWH. ARF-derived peptide stimulation induced polyfunctional T cell activations.

Identification of a naturally presented HLA-A*02:01-restricted ARF-derived peptide on infected primary CD4⁺ T cells

MHC ligands derived from sORFs that are specific or overrepresented in tumor cells were recently identified combining Riboseq and immunopeptidomic approaches^6,7,8. In the context of viral infection, two studies also highlighted that HLA class I molecules can present peptides from ARFs of HCMV and SARS-CoV-2^21,22. We thus intended to identify ARF-derived peptides presented by HLA class I and II molecules on the surface of HIV-infected cells (Supplementary Fig. 12a). We performed de novo experiments and also reanalyzed LC-MS/MS data from previously published immunopeptidomic experiments that used HIV-infected cells. To identify HIV-derived HLA-restricted peptides from all potential ARFs, the reviewed human proteome supplemented with the six-frame translated genome of HIV_NL4-3, taking into account all potential peptides ≥8 aa encompassed between two stop codons, was used as a reference. For our experiments, we used the CD4⁺ T cell lines SupT1 (used in the Riboseq experiment) and C8166, which express HLA class I and both HLA class I and II molecules, respectively. In order to bypass the downmodulation of HLA molecule expression mediated by HIV-Nef, we used a nef-deficient HIV_NL4-3 isolate (HIV_NL4-3∆Nef)⁴¹. Cells were readily infected, with more than 50 and 20% of HIV-gag⁺ SupT1 and C8166 cells, respectively (Supplementary Fig. 12b). For both cell lines, although we obtained a large number of HLA class I ligands (5673 for SupT1 and 7794 for C8166) and we detected known HIV-derived HLA class I ligands, we could not identify any ARF-derived HLA-restricted peptides. For C8166 cells, we additionally identified 4834 HLA class II-presented peptides from which 13 were derived from the HIV-Gag protein (Supplementary Fig. 12c). To our knowledge, these HIV-derived peptides are rare examples of natural HLA class II ligands identified in infected cells. Nonetheless, we did not detect any ARF-derived peptide. We thus reanalyse the immunopeptidomics data from our previously published work using HIV_NL4-3-infected primary CD4⁺ T cells⁴² and we identified seven HIV-derived peptides (Fig. 5a), one of which, ILINGQYSL, is derived from an ARF overlapping the pol gene (Supplementary Fig. 12e). Spectral validation using an isotopic labeled synthetic peptide confirmed the identification (Supplementary Fig. 12d). The ILINGQYSL peptide is annotated as a potent binder of HLA-A*02:01. Note that this ARF did not pass through the selection criteria of our initial Riboseq profiling because of its low coverage per codon (2.5 RPF/codon). However, it was readily identified in our Riboseq performed with LTM + PMY. LTM + PMY treatment led to the accumulation of RFPs at the leucine position 3222, suggesting that it might be the start codon of the ARF from which the ILINGQYSL peptide is derived (Supplementary Table 2, bottom line). This leucine is highly conserved in strains from clade B and poorly conserved for clade C (85 versus 15% conservation) suggesting that other translation initiation events might occur in clade C isolates. Nonetheless, this ARF, is highly conserved (above 90% of aa conservation) among both HIV clades.

**Fig. 5: MS-based identification of ARF-derived HLA-presented peptides and polyfunctional T cells responses.**

We then tested the capacity of the ILINGQYSL peptide (YSL-I) and an elongated form of YSL-I, LHSFGWVMNSILING (YSL-II), which has a high predictive score for binding to HLA-DR molecules (e.g., HLA- DRß1*03:01, *04:01, *07:01 according to the SYFPEITHI server), to activate T cell responses in PBMCs of PLWH. Using our cultured IFNγ-ELISPOT assay, we observed in PBMCs of EC-1 and EC-3 donors, intermediate to very strong T cell responses targeting the YSL-I peptide (Fig. 5b). The intracellular cytokine assay revealed that the YSL-I peptide is recognized by CD4⁺ T cells with a polyfunctional profile equivalent to responses specific to immunodominant CDS-derived peptides (Fig. 5c, pie charts). Responses to YLS-I peptide were particularly strong in CD4⁺ T cells from the EC-3 donor (up to 30% of CD4⁺ T cells) and were also detected, but to a weaker extent, in the cells from donors EC-1, EC-2, and EC-4 (Fig. 5d). Two donors also exhibited weak specific CD4⁺ T cell responses to the YLS-II peptide (Fig. 5e). Overall, in our cultured ELISPOT assay, 5 donors reacted to either YSL-I or YSL-II peptides (Fig. 5f). In addition, YSL-I-specific T cell responses with polyfunctional profiles were readily detected during ex vivo peptide stimulations in PBMCs of PLWH (Supplementary Fig. 11), clearly demonstrating that the YSL-ARF readily encodes viral polypeptides.

Therefore, using two complementary and independent approaches, detection of ARF-derived peptide-specific T cells in the PBMCs of PLWH and direct isolation of one HLA-bound ARF-derived peptide using mass spectrometry-based immunopeptidomics, we readily demonstrate that the HIV ARFs that we identified by ribosome profiling encode viral polypeptides capable of inducing broad T cell responses.

Discussion

We provide here the characterization of the HIV translatome in CD4⁺ T cells. We revealed the existence of at least 98 ARFs distributed across the HIV genome, including the 5′UTR, that are actively translated in infected CD4⁺ T cells. Most ARFs are likely conserved among HIV clade B and C clinical isolates. We demonstrated that at least 43 ARFs can be translated into viral polypeptides, since ARF-derived polypeptides were targeted by CD4⁺ and CD8⁺ T cells from PLWH. Remarkably, ARF-derived peptides induced a polyfunctional T cell memory response that is reminiscent of the ones targeting CDS-derived immunodominant epitopes. Finally, we identified a conserved ARF-derived epitope, ILINGQYSL, naturally presented on HIV-infected primary CD4⁺ T cells by the HLA-A*02:01 molecule. ILINGQYSL is translated from a highly conserved ARF and induces ex vivo a polyfunctional T cell activation. Using two complementary approaches, we demonstrated that the ARFs identified are actively translated into viral polypeptides with the capacity to induce potent T cell immunity.

In human cells, recent advances have revealed that sORFs are widely spread throughout the human genome, some overlapping different frames of classical ORFs or locating within the 5′ UTR of known genes³. Remarkably, these sORFs encode microproteins or polypeptides smaller than 100 aa^2,4. In fact, it has been estimated that 85% of the translation products originate from non-annotated regions of the human genome and mostly from out-of-frame sequences of CDS^9,43. We identified 98 ARFs within the HIV-1 genome, ranging from the 5′ UTR region to the nef gene. Nine ARFs are located in the 5′ UTR region. Several studies have recently demonstrated that polypeptides can be encoded in the 5′ UTR regions of viral genomes^{8,9,32,33,34,44,45}. These ARFs located at the 5′ UTR are probably translation regulatory elements of downstream CDS as observed in the human genome⁴⁶.

In our dataset, in addition to the 5′ UTR, most ARFs were located within gag, pol, and env genes, which correlate with the size of the respective CDS. Previous studies, using HLA footprint or predictive ORF approaches, also found ARFs overlapping gag, pol, and nef sequences^{18,19,20,25,26}. Only 16 HIV ARFs identified in our ribosomal profiling were in common with ARFs already described in the literature^19,25,26. This relatively low number might be due to an overestimation of the number of ARFs by the various prediction methods or the lack of thorough annotations of some ARFs in the previous studies. It can be also related to the rigorous criteria that we applied in our study for the selection of ARFs. This is illustrated by the ARF overlapping gag CDS and encoding the Q9VF epitope that we previously characterized in HIV-infected cells^20,24 whose coverage (5.9 RPF/codon) here is just below our selection criteria. However, we have previously shown that HIV_NL4-3-infected cells are not recognized by Q9VF-specific CTL due to a proteasomal cleavage of the Q9VF epitope^20,24. Remarkably, we also did not identify RPFs aligning in a frame on the minus strand of HIV. In particular, we did not identify the CDS encoding the antisense protein of HIV-1 (ASP) which overlaps the env gene, although several lines of evidence suggest that it is expressed in vivo^47,48. The lack of ASP CDS detection might be due to specific regulation of translation from the 3’ UTR that could be downmodulated in lymphoid cells such as CD4⁺ T cells used in our study⁴⁹. Nonetheless, we report here the identification of ARFs translated from the vif, vpr, vpu, tat, and env sequences as well as ARF overlapping tat and rev exons.

Combining ribosome profiling with LTM and PMY, we intended to define the translation initiation sites of the newly described ARFs. We obtained heterogenous results: for 26 ARFs a single start site was identified while for the others one or two initiation sites were delineated by patches of RPFs (Supplementary Table 2). Interestingly, among all potential initiation codons, 18% were closely related to the classical ATG methionine start codon. However, these results strongly suggest that a variety of mechanisms might be responsible for ARF expression. Indeed, a large part of eukaryotic and viral CDSs or ARFs initiate translation by noncanonical mechanisms such as leaky scanning, stop codon readthrough, ribosome shunting, re-initiation, IRES, and frameshifting¹³. Due to the diversity of potential translation initiation sites for most ARFs, we defined the ARF using flanking stop codons which also probably has an influence on the threshold of ARF RFP/codon and thus ARF identification.

Based on these criteria, the length of the identified ARFs ranged from 10 to 101 aa and can therefore be considered as sORF⁵⁰. Using an in-house HIV database of clade B and C strains, we propose that the ARF-encoded aa sequences might be conserved among both clades. Overall, we observed that the overlapping CDSs are significantly more conserved than the ARFs and the third frame, suggesting that ARFs might be under a selective pressure weaker than classical CDSs. Remarkably, the ARF, encoding the ILINGQYSL peptide, is highly conserved (above 90% of aa conservation) among HIV clade B and C, which is probably explained by the fact that it overlaps the sequence encoding the catalytic site of the HIV reverse transcriptase (RT). This sequence of the RT is crucial for HIV replication and is among the most conserved regions of the virus⁵¹. Nonetheless, 8 ARFs appeared to be more conserved than their overlapping CDS sequences. We cannot exclude, thus far, that these 8 ARFs might encode polypeptides with functional viral proprieties.

uORF in the 5′UTR of HIV could influence Gag expression. This hypothesis has been addressed in the accompanying manuscript submitted together with Emiliano Ricci. Indeed, using mutagenesis and reporter systems, we show that extensive uORF translation from HIV-1 transcripts conditions the expression of Gag CDS. Other ARFs in the HIV genome might also regulate the expression of the main HIV CDSs. They might also correspond to irrelevant by-standard translation events imposed by the tertiary structure and sequences of HIV mRNAs. To this regard, we observed a strong accumulation of reads, in the three frames, 30 nucleotides upstream of SD1, which is known to fold into a stable stem-loop structure (Supplementary Fig. 3). Interestingly, this region is also translated from various frames since we could detect T cell responses targeting peptides derived from two different ARFs that overlay this patch of RPF (Fig. 3 and Supplementary Fig. 3). Overall, we believe that the strong patch of reads in this region might be due to several factors including, the presence of SD1 in all viral transcripts, the formation of stable hairpin loops (containing SD1 and the RNA dimerization motif (DIS) upstream of SD1) but also to translation of ARFs.

As mentioned, ARFs might also encode polypeptides with biological functions. Indeed, it is becoming evident that short polypeptides (below 100 aa), that were largely ignored so far, can exert various functions in the development of viral infection. New identified ARFs might as well be seen as a genomic reservoir of unselected sequences allowing the emergence of de novo genes. Indeed, pervasive translation of short peptides derived from presumed non-coding regions might expose these ARF-encoded polypeptides to selection, allowing, from time to time, the passage of ARFs into the world of established, regulated, and selected products⁵², as observed for ASP⁵³.

In infected cells, MHC-I epitopes originate from viral proteins but also from truncated or misfolded viral polypeptides, also called DRiPs (defective ribosomal products)⁵⁴. DRiPs are labile products degraded shortly after translation, allowing rapid loading of MHC-I molecules and thus CTL recognition within minutes after viral infections⁵⁵. Remarkably, DRiPs were initially identified in HIV-infected cells⁵⁶. As discussed above, it is likely that ARF-produced viral polypeptides belong to DRiPs. In the present work, we readily demonstrate that at least 43 ARFs, identified by Riboseq, produce viral polypeptides which can induce T cell responses in PLWH. Some of these ARF-derived peptides are targeted by CTLs. Highlighting the importance of ARFP-specific T cells in the course of HIV infection, others and we previously demonstrated that ARFP-specific CTL recognize infected cells^18,19,20 and exert a selection pressure on the virus in vivo^18,19,24.

To identify naturally presented ARF-derived HLA ligands in HIV-infected cells, we analyzed the immunopeptidome of two infected CD4⁺ T cell lines. Unfortunately, we could not identify any ARF-derived HLA-presented peptide. These might be due to several factors including the sensitivity of shotgun mass spectrometric discovery approaches. Indeed, even in the context of the immense technical improvements in the last decades, it remains a challenge to identify low-abundance peptides with high turnover rates. In addition, the immunopeptidome is a highly dynamic, rich, and complex assembly of peptides, which is shaped by several factors, including accessibility to the antigen processing and presentation machinery e.g., specificities of proteasomes and cellular protases or of individual HLA allotypes⁵⁷. Nonetheless, our analysis of the immunopeptidomics data, published in our previous work⁵⁸, coming from infected primary CD4⁺ T cells allowed us to identify one ARF-derived peptide, the HLA-A*02:01-restricted peptide ILINGQYSL, thus providing a direct demonstration that ARF-derived peptides are naturally presented by HLA molecules. Probably due to the low number of HIV-infected samples that we used in our study, we did not detect CTL responses targeting this peptide. However, CD4⁺ T cell responses against this peptide or an elongated version of it was observed in five PLWH. Therefore, our results suggest that alternative translation events may give rise to epitopes recognized independently or concomitantly by CD4⁺ and CD8⁺ T cells and presented on different HLA class I and HLA class II alleles. Overall, we observed that the majority of T cell responses targeting ARF-derived peptides were mediated by CD4⁺ T cells.

As a matter of fact, HIV-specific CD4⁺ T cells play an important role in HIV infection. The breadth and specificity of HIV-specific CD4⁺ T cell responses are associated with improved viral control and low viremia during acute and chronic infection, respectively^59,60. In particular, the generation of Gag-specific CD4⁺ T cells and their maintenance correlates with a better disease outcome and viral control, respectively^59,60. There is also evidence of an association between certain HLA-DRß1 alleles and Gag-specific CD4⁺ T cell activity and delayed disease progression⁶¹. HIV-specific CD4⁺ T cells can also exert direct antiviral functions by eliminating infected cells and/or inhibiting viral replication⁶². Remarkably, others and we have shown that CD4⁺ T cells can recognize peptides derived from newly-synthetized, so-called endogenous antigens in virus-infected cells, including HIV-infected cells^63,64 or in tumor cells⁶⁵. In fact, depending on the subcellular localization, the trafficking, and the nature of the antigen itself, different pathways are involved in the degradation and presentation of endogenous antigens by MHC-II molecules. This includes components of the MHC-I processing pathway, such as proteasome and TAP^66,67, autophagy⁶⁸ and receptors regulating vesicular trafficking⁶⁹. In the field of cancer, it was recently reported that CD4⁺ T cells also recognize peptides derived from unannotated CDSs and neoantigens^70,71,72,73.

To get some insights on the quality of ARFP-specific T cell responses, we analyzed the polyfunctional profiles and the pattern of cytokine secretions of these CTL and CD4⁺ T cell responses in PBMCs of PLWH. We show here that CD4⁺ and CD8⁺ T cell responses against ARFP exhibit a polyfunctional profile with 80% of cells secreting at least three cytokines simultaneously. The pattern of secretion does not differ between previously described T cell responses targeting immunodominant epitopes and ARFP-specific T cell responses, suggesting that the latter are as potent as classical T cell responses in controlling viral replication.

Although using a small cohort of PLWH with suppressed viral replication, we were intrigued by the large heterogeneity of ARF-specific T cell responses. Several parameters might influence the magnitude and the breadth of HIV-specific T cell responses, including the individual HLA allotype and T cell repertoire, the senescence of T cells, the antiretroviral treatment, and the size of the active viral reservoir, defined as cells carrying a transcriptionally active viral genome⁷⁴. Remarkably, the later has been recently linked to the magnitude and functions of HIV-specific CD4⁺ and CD8⁺ T cells^75,76. Interestingly, this active reservoir is mainly composed of defective viral genomes⁷⁵, which might favor the expression of alternative reading frames and, thus, the activation of ARF-derived peptide-specific T cells. In the future, it will be of interest to study ARF-specific T cell responses with regards to the active reservoir and the emergence of new dominant T cell clonotypes after prolonged ART treatment⁷⁷.

In conclusion, using ribosome profiling, we defined here the translatome of HIV in infected CD4⁺ T cells. We demonstrated that the HIV genome harbors ARFs that can be translated into viral polypeptides used by the immune system to initiate potent T cell responses. Both in cancer and in HIV-infected cells, ARF-derived polypeptides represent promising targets to promote the control of tumor growth or viral replication. Understanding how the translation of noncanonical genomic regions is regulated and its involvement in various cellular or viral processes will certainly help in fighting cancer and infections.

Methods

Virus and infection

VSV-G pseudotyped HIV_{NL4-3 XCS} and HIV_NL4-3∆Nef were produced by transfection of 293T cells using a CAPHOS kit (Sigma-Aldrich) as in ref. ⁶³. Viruses were pseudotyped to favor viral entry and to achieve a high infection rate within a short time. SupT1 CD4⁺ T cells were cultured with RPMI GlutaMax 1640 (Gibco) complemented with 10% FBS (Dutscher) and 1% penicillin/streptomycin. About 1 × 10⁹ SupT1 CD4⁺ T cells were infected with 4055 ng of VSV-G-HIV_{NL4-3 XCS} HIV-Gag-p24 for 3 h in IMDM plus 10 mM HEPES (Gibco/Thermo Fisher Scientific) supplemented with 2 µg/ml of DEAE-dextran (Sigma). Three biological replicates were performed corresponding to 300 million cells infected at day 0. Twenty hours post-infection a fraction of the cells (2 × 10⁵) was harvested, stained with CD4-Vio770 antibody (Miltenyi) and a Live/Dead (LVD) fixable violet dye (Invitrogen), fixed with 4% paraformaldehyde (ChemCruz), and permeabilized (PBS1X, BSA 0,5%, saponin 0,05%). Cells were then stained with an HIV-1 Gag-p24 specific antibody (KC57-PE, Beckman Coulter). Sample acquisition and analysis were performed using a BD LSRFortessa and FlowJo v10.8 Software, respectively (both from BD Life Sciences). Mock-infected cells were used as a negative control. Twenty-four hours post-infection, cycloheximide (CHX) was then added to the remaining cell cultures (100 µg/ml, 10 min, 37 °C, Sigma). Cells were washed in cold PBS containing CHX and lysed at 4 °C (10 mM Tris-HCl pH 7.5, 100 mM KCl, 10 mM Mg2+ acetate, 1% Triton X-100 and 2 mM DTT). Glass beads were then added to the cell supernatant and incubated on an orbital shaker (5 min, 4 °C). Beads were removed by centrifugation and the supernatant was quickly frozen in liquid nitrogen and stored at −80 °C. To identify the start codons, prior harvest, cells were treated with Lactimidomycin (LTM; 5 μM, 30 min, 37 °C) together with Puromycin (PMY, 25 μM) for the last 20 min. Cells were lysed in abovementioned lysis buffer containing LTM and PMY.

Ribosome profiling

Cell lysates were gently thawed on ice in the presence of an EDTA-free complete protease inhibitor cocktail (Roche) and 200 U of RNase Murine Inhibitor (NEB) were added. The absorbance of the crude extract obtained was measured at 260 nm. The extracts were digested using 5U/UA_260nm RNase I (Ambion) for 1 h at 25 °C. The digestion was stopped by adding a SUPErasin RNase inhibitor (500U, Invitrogen). Monosomes were then loaded on a 24% sucrose cushion and ultracentrifugated at 543,000×g on a TLA110 rotor at +4 °C. The concentrated monosomes were resuspended in 600 µl of lysis buffer. RNA were extracted by acid phenol at 65 °C, chloroform, and precipitated by ethanol with 0.3 M sodium acetate pH 5.2. Resuspended RNAs were loaded on 17% polyacrylamide (19:1) gel with 7 M urea and run in 1xTAE buffer for 6 h at 100 V. RNA fragments corresponding to 28–34nt were retrieved from gel and precipitated in ethanol with 0.3 M sodium acetate pH 5.2 in presence of 100 mg glycogen. rRNA were depleted using the Ribo-Zero Gold rRNA Removal kit (Illumina). The supernatants containing the ribosome footprints were recovered, and RNA were precipitated in ethanol in the presence of glycogen overnight at −20 °C. The RNA concentration was measured by Quant-iT microRNA assay kit (Invitrogen), and the RNA integrity and quality was verified using Bioanalyzer small RNA Analysis kit (Agilent). cDNA library from 100 ng RNA was prepared by the High-throughput sequencing facility of I2BC, using the NebNext Small RNA Sample Prep kit with 3′ sRNA Adapter (Illumina) according to the manufacturer’s protocol with 12 cycles of PCR amplification in the last step followed by DNA purification with AMPpure XP beads cleanup. Library molarity was analyzed using a Bioanalyzer DNA Analysis kit (Agilent) and an equimolar pool of libraries was sequenced with NexTSeq 500/550 High output kit v2 (75 cycles) (Illumina) with 10% PhiX.

Bioinformatical analysis

The ribosome profiling analysis was made using the RiboDoc tool (v0.9.0)²⁹. The different main steps with corresponding programs, versions and command lines used in its analysis are described below. The reference genomes are Homo_sapiens.GRCh38.104 for humans (from the Ensembl database⁷⁸) and HIV-1-pNL4-3 XCS for HIV.

The sequencing adapters were trimmed by cutadapt v4.3⁷⁹ and the lengths of the RPFs was filter to keep reads from 25 to 35 nucleotides long as there expected length is around 30 nucleotides with the following parameters:

cutadapt -e 0.125 --max-n = 1 -m 25 -M 35 -a ${adapter_sequence} -o ${output.fastq} ${input.fastq}

The removal of the rRNA reads was made by an alignment on the rRNA sequences by bowtie2 v2.5.1⁸⁰:

bowtie2 -x ${index.rRNA} -U ${input.fastq} --un-gz ${output.fastq}

The alignment on the genome was made with both hisat2 v2.2.1⁸¹ and bowtie2 v2.5.1:

hisat2 -x ${hisat2_index.genome} --no-softclip -U ${input.fastq} --un-gz ${output.fastq} -S ${output.sam_hisat2}

bowtie2 -x ${bowtie2_index.genome} --end-to-end -U ${output.fastq} -S ${output.sam_bowtie2}

The selection of the reads uniquely mapped on the genome was made with samtools v1.14⁸²:

samtools view -F 3844 -q 1 -h -o ${output.bam}

The counting of the reads corresponding to each transcript was done by htseqcount v2.0.2⁸³:

htseq-count -f bam -t CDS -i Parent --additional-attr Name -m intersection-strict --nonunique fraction ${input.bam} ${input.gff} > ${output.txt}

The qualitative analysis was made on a transcriptome made from the genome with a selection of transcripts annotated as having a 5′UTR region only (for the human genome). This analysis and the determination of the P-site offset for each read length of every sample was made by the riboWaltz v1.2.0 package⁸⁴.

To study the reading frame of the ribosome-protected fragments (RPF), each read was represented by the coordinate corresponding to the first base of the associated ribosome’s P-site. To determine where the P-site is, a P-site offset has to be defined for every read length. This step was done with the riboWaltz program⁸⁴ which looks for the first base of each read beginning of the signal.

Data sets and multiple alignments

Sequences were downloaded from the Los Alamos HIV Sequence Database (HIV-1 clades B and C) (https://www.hiv.lanl.gov/content/index). Sequences were filtered based on the presence of only one sequence per patient and excluding sequences carrying premature stop codons within CDS. Very similar sequences and incomplete sequences (more than 10 gaps/unknowns) were additionally discarded. We obtained 1609 and 412 sequences of HIV-1 clade B and clade C, respectively. The aa sequence conservation of the overlapping CDS and third frame were analyzed using the Unipro UGENE toolkit. Multiple alignments were performed using ClustalO. Statistical analysis was performed with R (https://www.R-project.org/, version 4.2.2) using Rstudio IDE (2023.03.01.554). Comparison of population distributions were performed using Kruskal–Wallis test, Wilcoxon, and Dunn tests. Plots were generated using the ggplot2 package (v3.4.2). For each aa sequence conservation analysis of ARFs within HIV-1 clades B and C, the number of sequences used ranged from 168 to 1609 and from 47 to 411 sequences, respectively. Taking into consideration that in the UTR regions, the sequences may contain microdeletion and/or truncation, another pipeline was used for the ARF overlapping the UTRs. For each nucleotide sequence, the six phases were translated. Then, ARF aa sequences were aligned recursively to the six translated sequences. The best alignments were kept, and alignments too far (±150 nt) from the putative position were eliminated. Multiple alignments were performed using Clustal Omega.

Immunopeptidome

Isolation of HLA ligands

About 5 × 10⁸ C8166 cells were infected with HIV_NL4-3∆Nef at an MOI of 0.05 and snap frozen 22 h post-infection. HLA class I and HLA class II molecules were isolated using standard immunoaffinity purification⁸⁵, using the pan-HLA class I-specific W6/32, the pan-HLA class II-specific Tü-39, and the HLA-DR-specific L243 monoclonal antibodies (produced in-house) covalently linked to CNBr-activated Sepharose (Sigma-Aldrich). Cells were lysed in lysis buffer (CHAPS (Panreac AppliChem), complete protease inhibitor cocktail tablet (Roche) in PBS) for 1 h on a shaker at 4 °C, sonicated, and centrifuged (45 min, 4000×g) and incubated again for 1 h. Lysates were cleared by sterile filtration (5 µm filter unit (Merck Millipore)) and cyclically passed through a column-based setup overnight at 4 °C. Columns were washed with PBS (30 min) and ddH₂O (1 h). Peptides were eluted by 0.2% trifluoroacetic acid (TFA), isolated by ultrafiltration (Amicon filter units (Merck Millipore)), lyophilized, and desalted using ZipTip pipette tips with C18 resin (Merck).

Mass spectrometric data acquisition

For the mass spectrometric analysis peptides were loaded on a 75 µm × 2 cm PepMap nanotrap column (Thermo Fisher) at a flow rate of 4 µl/min for 10 min⁸⁵. Subsequent separation was performed by nanoflow high-performance liquid chromatography (RSLCnano, Thermo Fisher) using a 50 µm × 25 cm PepMap rapid separation column (Thermo Fisher, particle size of 2 µm) and a linear gradient ranging from 2.4 to 32.0% acetonitrile at a flow rate of 0.3 µl/min over the course of 90 min. Eluting peptides were analysed in technical replicates in an online-coupled Orbitrap Fusion Lumos mass spectrometer (Thermo Fisher) equipped with a nanoelectron spray ion source using a data-dependent acquisition mode employing a top-speed collisional-induced dissociation (CID, normalized collision energy 35%, HLA class I peptides) or higher-energy collisional dissociation (HCD, normalized collision energy 30%, HLA class II peptides) fragmentation method. MS1 and MS2 spectra were detected in the Orbitrap with a resolution of 120,000 and 30,000, respectively. The maximum injection time was set to 50 and 150 ms for MS1 and MS2, respectively. The dynamic exclusion was set to 7 and 10 s for HLA class I and HLA class II, respectively. The mass range for HLA class I peptide analysis was set to 400–650 m/z with charge states 2+ and 3+ selected for fragmentation. For HLA class II peptide analysis mass range was limited to 400 1000 m/z with charge states 2+ to 5+ selected for fragmentation.

Data processing

For data processing, the SEQUEST HT search engine (University of Washington⁸⁶) was used to search the human proteome as comprised in the Swiss-Prot database (20,394 reviewed protein sequences, December 4 2020) supplemented with the six-frame translated genome of HIV_NL4-3, taking into account all potential peptides ≥8 aa encompassed between two stop codons without enzymatic restriction. Precursor mass tolerance was set to 5 ppm, and fragment mass tolerance to 0.02 Da. Oxidized methionine was allowed as a dynamic modification. The peptide spectrum matches (PSM) false discovery rate (FDR) was estimated using the Percolator algorithm⁸⁷ and limited to 5% for HLA class I and 1% for HLA class II. Peptide lengths were limited to 8–12 aa for HLA class I and to 8–25 aa for HLA class II. Protein inference was disabled, allowing for multiple protein annotations of peptides. HLA class I annotation was performed using NetMHCpan-4.1⁸⁸ and SYFPEITHI⁸⁹ annotating peptides with percentile rank below 2% and ≥60% of the maximal score, respectively.

Spectrum validation

Spectrum validation of the experimentally eluted peptide was performed by computing the similarity of the spectra with the corresponding isotope-labeled synthetic peptides measured in a complex matrix. The spectral correlation was calculated between the MS/MS spectra of the eluted and the synthetic peptide⁹⁰.

Participants and samples

Elite controllers (EC, n = 4) were recruited from the CO21 CODEX cohort implemented by the ANRS | MIE (Agence Nationale de recherches sur le SIDA, les hépatites virales|Maladies infectieuses émergentes). PBMCs were cryopreserved at enrollment. ECs were defined as people living with HIV (PLWH) maintaining viral loads (VL) under 400 copies of HIV-RNA/mL without treatment for more than 5 years. Efficiently treated PLWH (ART) (n = 3) were recruited at AP-HP Bicêtre Hospital (Kremlin Bicêtre, France). They are treated according to the standard of care for at least 1 year (mean of 10 years) and have undetectable viral loads using standard assays. PBMCs from anonymous HIV-negative blood donors (n = 3) were purchased from EFS (Établissement Français du sang) under the agreement number 15/EFS/022. A detailed description of the donors is provided in Supplementary Table 4, including the interquartile range for age (at the time of the study), the sex, the number of years of infection and of treatment, the drug regimen, the mean CD4 T cell counts, RNA loads and the HLA haplotypes. HIV-RNA loads were measured on-site with different real-time PCR-based assays; depending on the date of enrollment in the cohort and the assay routinely used on each site, the VL detection limit varied from 50 to 10 copies/mL.

Ethic statement

All the participants provided their written informed consent to participate in the study. The CO21 CODEX cohort and this sub-study were funded and sponsored by ANRS | MIE and approved by the Ile de France VII Ethics Committee. The study was conducted according to the principles expressed in the Declaration of Helsinki.

Peptides

ARF-derived peptides were chosen based on the aa conservation among the clade B isolates and the predictions to bind frequent HLA alleles. The binding affinity of ARF-derived peptides was evaluated using the NetMHCPan-4.1 and NetMHCIIPan-4.1⁸⁸. For the longest ARF sequences, some overlapping peptides were also designed. The synthetic peptides were then randomly distributed in 9 different pools to obtain 11 or less peptides per pool to limit peptide competition for HLA binding (Supplementary Table 3). HIV-classical epitopes were chosen based on the literature³⁷ (Supplementary Table 5). ARF-derived peptides and HIV-classical peptides were synthetized by Vivitide company and in-house at the Department of Immunology, University of Tübingen, Germany. The in-house production was performed with the peptide synthesizer Liberty Blue (CEM) using the 9 fluorenylmethyl-oxycarbonyl/tert-butyl strategy⁹¹. Non-HIV common peptides (from HCMV, EBV, and influenza virus) were obtained from Mabtech (PepPool: CEF (CD8), human, 3616-1).

12-day in vitro T cell amplification prior to ELISPOT assay

PBMCs were thawed and rested 2–3 h in IMDM (Gibco/Thermo Fisher Scientific) containing 5% human AB serum (SAB, Institut Jacques Boy), supplemented with recombinant human IL2 (rhIL-2, 10 U/ml, Miltenyi) and Dnase I (1 U/ml, New England Biolabs). Cells were washed, and 5–9 × 10⁶ PBMCs were then seeded/well in a 24-well plate in IMDM supplemented with 10% SAB, Pen/Strep, nonessential aa, and sodium pyruvate (all from Gibco/Thermo Fisher Scientific). Nevirapine (NVP, 1.2 µM, HIV reagent program) was added to inhibit potential viral replication and poly I:C (2 µg/ml, InvivoGen) to facilitate the presentation of long peptides by antigen-presenting cells⁹². PBMCs were loaded with peptide pools (each peptide at 10 µg/ml), except for the CEF pool (Mabtech) used at 5 µg/ml, and cultured overnight at 37 °C, 5% CO₂. HIV-classical peptides and/or CEF peptide pools were used as positive control for the expansion of HIV-specific and non-HIV common anti-virus-specific T cells, respectively. On days 1 and 3, T cell media were then complemented with rhIL-2 (10 U/ml) and rhIL-7 (20 ng/ml, Miltenyi) with NVP (1.2 µM), respectively, and maintained throughout the culture. On days 3, 5, 7, and 9, when the cell layer was >70% confluent, cells were split into two wells before the addition of rhIL-2, rhIL-7, and NVP. On day 7, cells were harvested, counted, and a fraction was submitted to IFN-γ ELISPOT assay using the ARF-derived peptide pools and positive controls: HIV-classical peptide and/or CEF peptide pools. The remaining cells were maintained in culture and submitted, on day 12, to IFN-γ ELISPOT assay using individual ARF-derived peptides and the controls.

Enzyme-linked immunospot assay—cultured IFN-γ ELISPOT

About 1–3 × 10⁵ cells/well were seeded in ELISPOT plates (MSIPN4550, Millipore) in IMDM supplemented with 10% SAB, Pen/Strep, nonessential aa, sodium pyruvate 10%, and HEPES 10 mM (Gibco/Thermo Fisher Scientific), loaded with either 50, 10, or 2 µg/ml of ARF-derived peptides, HIV-classical peptides and CEF peptides, respectively and incubated at 37 °C for 16 h. ELISPOT plates were pre-coated with anti-IFNγ primary antibody, and after cell incubation, IFNγ was revealed using an anti-IFNγ secondary antibody conjugated with biotin (both from Mabtech) as described²⁴. The ELISPOT analysis was performed in technical triplicates or duplicates. Spots were counted with the AID ELISPOT Reader according to standard protocols. Responses were considered positive when IFNγ production was superior to 20 spots/10⁶ PBMCs and at least twofold higher than background (cells loaded with the DMSO-containing medium used to solubilize the peptides).

Intracellular cytokine staining (ICS) assay

Cells were either cultured as described for the 12-day in vitro T cell amplification or treated directly after thawing (for the ex vivo assays). On day 12, cells were harvested, washed, and counted. About 2–10 × 10⁵ PBMCs were then seeded/well of a 96-well U-bottom plate in IMDM supplemented with 10% SAB, Pen/Strep, nonessential aa, sodium pyruvate 10%, and HEPES 10 mM. PBMCs were immediately loaded with individual ARF-derived peptides (50 µg/ml), individual HIV- classical peptides (10 µg/ml), or CEF peptide pool (2 ug/ml). Additionally, the CD107a-BV786 antibody (BD Biosciences) was included in each condition for degranulation detection. For the ex vivo assessment of T cell activation, 2–4 × 10⁶ PBMCs were seeded per well of a 96-well U-bottom plate using the same IMDM-supplemented medium. The same peptide loading and CD107a-BV786 antibody addition steps were performed. After 1 h at 37 °C, brefeldin A was added (5 µg/ml), and the incubation was carried out for an additional 5 h. PBMCs were washed in PBS and incubated with LVD and human Fc Block (BD Biosciences). Cells were then stained for 30 min on ice with a mix of cell surface antibodies: CD3-PE, CD8-SuperBright600 (Invitrogen), and CD4-AF750 (Beckman Coulter). Next, cells were washed, fixed with PFA 4%, and permeabilized (PBS + BSA 0.5%, saponin 0,05%). Cytokine production were detected by intracellular staining using IFN-γ-PerCP-Cy-5.5, MIP-1β-FITC, TNFα-PE-Cy7, and IL2-APC antibodies (all from BD Biosciences).

Flow cytometry and data analysis

Sample acquisition and analysis were performed using Cytoflex with CytExpert software (Beckman Coulter). The compensation matrix was calculated automatically by the CytExpert software after measuring negative and single-color controls. Sample analysis was performed using FlowJo v10.8 Software (BD Life Sciences). After the elimination of doublets, lymphocytes were gated on the side scatter area versus the forward scatter area pseudo-color dot plot, and dead cells were removed according to LVD staining. CD4⁺CD3⁺ events were gated versus individual gates made for the detection of MIP-1β, IFNγ, IL2, CD107a, and TNFα secretion. All gates were set based on the unstimulated control (cells loaded with the DMSO-containing medium used to solubilize the peptides). Individual cytokine secretions were then combined together using Boolean gating to create a full array of possible combinations of response patterns from the CD4⁺CD3⁺-cell subsets. The same procedure was used to CD8⁺CD3⁺ events. Positive responses were reported after background subtraction. Samples were considered positive when the % of activated cells was at least two times higher than negative controls.

Statistical analysis

Statistical significances (p values) were calculated using Prism Software (GraphPad). Statistical tests used for each individual figure are indicated in the figure legends.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The mass spectrometry immunopeptidomics data have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the PRIDE (Perez-Riverol et al, 2019) partner repository with the dataset identifier: Project accession: PXD043984. The Riboseq data have been deposited to the public functional genomics data repository GEO: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE239818 Source data are provided with this paper.

Code availability

The code package for this study is available and fully described in the above section “Bioinformatical analysis”.

Change history

06 March 2025
A Correction to this paper has been published: https://doi.org/10.1038/s41467-025-57621-z

References

Andrews, S. J. & Rothnagel, J. A. Emerging evidence for functional peptides encoded by short open reading frames. Nat. Rev. Mater. 15, 193–204 (2014).
Martinez, T. F. et al. Accurate annotation of human protein-coding small open reading frames. Nat. Chem. Biol. 16, 458–468 (2020).
Article ADS CAS PubMed MATH Google Scholar
Ji, Z., Song, R., Regev, A. & Struhl, K. Many lncRNAs, 5’UTRs, and pseudogenes are translated and some are likely to express functional proteins. eLife 4, e08890 (2015).
Article PubMed PubMed Central Google Scholar
Schlesinger, D. & Elsässer, S. J. Revisiting sORFs: overcoming challenges to identify and characterize functional microproteins. FEBS J. 289, 53–74 (2022).
Article CAS PubMed MATH Google Scholar
Yewdell, J. W. MHC class I immunopeptidome: past, present, and future. Mol. Cell. Proteomics 21, 100230 (2022).
Article CAS PubMed PubMed Central Google Scholar
Chong, C., Coukos, G. & Bassani-Sternberg, M. Identification of tumor antigens with immunopeptidomics. Nat. Biotechnol. 40, 175–188 (2022).
Article CAS PubMed Google Scholar
Nelde, A. et al. Upstream open reading frames regulate translation of cancer-associated transcripts and encode HLA-presented immunogenic tumor antigens. Cell. Mol. Life Sci. 79, 171 (2022).
Article CAS PubMed PubMed Central Google Scholar
Ouspenskaia, T. et al. Unannotated proteins expand the MHC-I-restricted immunopeptidome in cancer. Nat. Biotechnol. 40, 209–217 (2022).
Article CAS PubMed Google Scholar
Ruiz Cuevas, M. V. et al. Most non-canonical proteins uniquely populate the proteome or immunopeptidome. Cell Rep. 34, 108815 (2021).
Article CAS PubMed MATH Google Scholar
Bartok, O. et al. Anti-tumour immunity induces aberrant peptide presentation in melanoma. Nature 590, 332–337 (2021).
Article ADS CAS PubMed MATH Google Scholar
Smith, C. C. et al. Endogenous retroviral signatures predict immunotherapy response in clear cell renal cell carcinoma. J. Clin. Invest. 128, 4804–4820 (2018).
Article PubMed PubMed Central MATH Google Scholar
Starck, S. R. et al. Translation from the 5’ untranslated region shapes the integrated stress response. Science 351, aad3867 (2016).
Article PubMed PubMed Central MATH Google Scholar
Firth, A. E. & Brierley, I. Non-canonical translation in RNA viruses. J. Gen. Virol. 93, 1385–1409 (2012).
Article CAS PubMed PubMed Central MATH Google Scholar
Bullock, T. N. & Eisenlohr, L. C. Ribosomal scanning past the primary initiation codon as a mechanism for expression of CTL epitopes encoded in alternative reading frames. J. Exp. Med. 184, 1319–1329 (1996).
Article CAS PubMed MATH Google Scholar
Yang, N. et al. 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).
Article CAS PubMed MATH Google Scholar
Zanker, D. J. et al. Influenza A virus infection induces viral and cellular defective ribosomal products encoded by alternative reading frames. J. Immunol. 202, 3370–3380 (2019).
Article CAS PubMed MATH Google Scholar
Gaur, A. & Green, W. R. Role of a cytotoxic-T-lymphocyte epitope-defined, alternative gag open reading frame in the pathogenesis of a murine retrovirus-induced immunodeficiency syndrome. J. Virol. 79, 4308–4315 (2005).
Article CAS PubMed PubMed Central Google Scholar
Bansal, A. et al. CD8 T cell response and evolutionary pressure to HIV-1 cryptic epitopes derived from antisense transcription. J. Exp. Med. 207, 51–59 (2010).
Article CAS PubMed PubMed Central Google Scholar
Berger, C. T. et al. Viral adaptation to immune selection pressure by HLA class I–restricted CTL responses targeting epitopes in HIV frameshift sequences. J. Exp. Med. 207, 61–75 (2010).
Article ADS CAS PubMed PubMed Central Google Scholar
Cardinaud, S. et al. Identification of cryptic MHC I–restricted epitopes encoded by HIV-1 alternative reading frames. J. Exp. Med. 199, 1053–1063 (2004).
Article CAS PubMed PubMed Central MATH Google Scholar
Erhard, F. et al. Improved Ribo-seq enables identification of cryptic translation events. Nat. Methods 15, 363–366 (2018).
Article CAS PubMed PubMed Central MATH Google Scholar
Weingarten-Gabbay, S. et al. Profiling SARS-CoV-2 HLA-I peptidome reveals T cell epitopes from out-of-frame ORFs. Cell 184, 3962–3980.e17 (2021).
Article CAS PubMed PubMed Central MATH Google Scholar
Bansal, A. et al. Enhanced recognition of HIV-1 cryptic epitopes restricted by HLA class I alleles associated with a favorable clinical outcome. J. Acquir. Immune Defic. Syndr. 70, 1–8 (2015).
Article CAS PubMed PubMed Central MATH Google Scholar
Cardinaud, S. et al. CTL escape mediated by proteasomal destruction of an HIV-1 cryptic epitope. PLoS Pathog. 7, e1002049 (2011).
Article CAS PubMed PubMed Central Google Scholar
Champiat, S. et al. Influence of HAART on alternative reading frame immune responses over the course of HIV-1 infection. PLoS ONE 7, e39311 (2012).
Article ADS CAS PubMed PubMed Central Google Scholar
Garrison, K. E. et al. Transcriptional errors in human immunodeficiency virus type 1 generate targets for T-cell responses. Clin. Vaccine Immunol. 16, 1369–1371 (2009).
Article CAS PubMed PubMed Central MATH Google Scholar
Maness, N. J. et al. AIDS virus specific CD8+ T lymphocytes against an immunodominant cryptic epitope select for viral escape. J. Exp. Med. 204, 2505–2512 (2007).
Article CAS PubMed PubMed Central MATH Google Scholar
Wildum, S., Schindler, M., Münch, J. & Kirchhoff, F. Contribution of Vpu, Env, and Nef to CD4 down-modulation and resistance of human immunodeficiency virus type 1-infected T cells to superinfection. J. Virol. 80, 8047–8059 (2006).
Article CAS PubMed PubMed Central Google Scholar
François, P., Arbes, H., Demais, S., Baudin-Baillieu, A. & Namy, O. RiboDoc: a docker-based package for ribosome profiling analysis. Comput. Struct. Biotechnol. J. 19, 2851–2860 (2021).
Article PubMed PubMed Central Google Scholar
Namy, O., Moran, S. J., Stuart, D. I., Gilbert, R. J. C. & Brierley, I. A mechanical explanation of RNA pseudoknot function in programmed ribosomal frameshifting. Nature 441, 244–247 (2006).
Article ADS CAS PubMed PubMed Central Google Scholar
Bencun, M. et al. Translational profiling of B cells infected with the Epstein-Barr virus reveals 5’ leader ribosome recruitment through upstream open reading frames. Nucleic Acids Res. 46, 2802–2819 (2018).
Article CAS PubMed PubMed Central MATH Google Scholar
Finkel, Y. et al. The coding capacity of SARS-CoV-2. Nature 589, 125–130 (2021).
Article ADS CAS PubMed MATH Google Scholar
Whisnant, A. W. et al. Integrative functional genomics decodes herpes simplex virus 1. Nat. Commun. 11, 2038 (2020).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Yang, Z. et al. Deciphering poxvirus gene expression by RNA sequencing and ribosome profiling. J. Virol. 89, 6874–6886 (2015).
Article CAS PubMed PubMed Central MATH Google Scholar
Gao, X. et al. Quantitative profiling of initiating ribosomes in vivo. Nat. Methods 12, 147–153 (2015).
Article CAS PubMed MATH Google Scholar
Curlin, M. E. et al. HIV-1 envelope subregion length variation during disease progression. PLoS Pathog. 6, e1001228 (2010).
Article CAS PubMed PubMed Central Google Scholar
HIV Molecular Immunology (Theoretical Biology and Biophysics Group T-10, Los Alamos National Laboratory, 2022).
Sauce, D., Elbim, C. & Appay, V. Monitoring cellular immune markers in HIV infection: from activation to exhaustion. Curr. Opin. HIV AIDS 8, 125–131 (2013).
Article CAS PubMed Google Scholar
Larsen, M. et al. Evaluating cellular polyfunctionality with a novel polyfunctionality index. PLoS ONE 7, e42403 (2012).
Article ADS CAS PubMed PubMed Central Google Scholar
Vingert, B. et al. HIV controllers maintain a population of highly efficient Th1 effector cells in contrast to patients treated in the long term. J. Virol. 86, 10661–10674 (2012).
Article CAS PubMed PubMed Central MATH Google Scholar
Schwartz, O., Maréchal, V., Gall, S. L., Lemonnier, F. & Heard, J.-M. Endocytosis of major histocompatibility complex class I molecules is induced by the HIV–1 Nef protein. Nat. Med. 2, 338–342 (1996).
Article CAS PubMed Google Scholar
Ziegler, M. C. et al. HIV-1 induced changes in HLA-C∗03: 04-presented peptide repertoires lead to reduced engagement of inhibitory natural killer cell receptors. AIDS 34, 1713–1723 (2020).
Article CAS PubMed MATH Google Scholar
Ingolia, N. T., Hussmann, J. A. & Weissman, J. S. Ribosome profiling: global views of translation. Cold Spring Harb. Perspect. Biol. 11, a032698 (2019).
Article CAS PubMed PubMed Central Google Scholar
Zhang, J. & Crumpacker, C. HIV UTR, LTR, and epigenetic immunity. Viruses 14, 1084 (2022).
Article CAS PubMed PubMed Central MATH Google Scholar
Arias, C. et al. KSHV 2.0: a comprehensive annotation of the Kaposi’s sarcoma-associated herpesvirus genome using next-generation sequencing reveals novel genomic and functional features. PLoS Pathog. 10, e1003847 (2014).
Article PubMed PubMed Central Google Scholar
Orr, M. W., Mao, Y., Storz, G. & Qian, S.-B. Alternative ORFs and small ORFs: shedding light on the dark proteome. Nucleic Acids Res. 48, 1029–1042 (2020).
Article CAS PubMed MATH Google Scholar
Bet, A. et al. The HIV-1 antisense protein (ASP) induces CD8 T cell responses during chronic infection. Retrovirology 12, 15 (2015).
Article PubMed PubMed Central Google Scholar
Vanhée-Brossollet, C. et al. A natural antisense RNA derived from the HIV-1 env gene encodes a protein which is recognized by circulating antibodies of HIV+ individuals. Virology 206, 196–202 (1995).
Article PubMed Google Scholar
Laverdure, S. et al. HIV-1 antisense transcription is preferentially activated in primary monocyte-derived cells. J. Virol. 86, 13785–13789 (2012).
Article CAS PubMed PubMed Central MATH Google Scholar
Finkel, Y., Stern‐Ginossar, N. & Schwartz, M. Viral short ORFs and their possible functions. Proteomics 18, 1700255 (2018).
Article PubMed PubMed Central Google Scholar
Jacobo-Molina, A. et al. Crystal structure of human immunodeficiency virus type 1 reverse transcriptase complexed with double-stranded DNA at 3.0 A resolution shows bent DNA. Proc. Natl Acad. Sci. USA 90, 6320–6324 (1993).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Ruiz-Orera, J., Verdaguer-Grau, P., Villanueva-Cañas, J. L., Messeguer, X. & Albà, M. M. Translation of neutrally evolving peptides provides a basis for de novo gene evolution. Nat. Ecol. Evol. 2, 890–896 (2018).
Article PubMed Google Scholar
Cassan, E., Arigon-Chifolleau, A.-M., Mesnard, J.-M., Gross, A. & Gascuel, O. Concomitant emergence of the antisense protein gene of HIV-1 and of the pandemic. Proc. Natl Acad. Sci. USA 113, 11537–11542 (2016).
Article ADS CAS PubMed PubMed Central Google Scholar
Yewdell, J. W., Antón, L. C. & Bennink, J. R. Defective ribosomal products (DRiPs): a major source of antigenic peptides for MHC class I molecules? J. Immunol. 157, 1823–1826 (1996).
Article CAS PubMed Google Scholar
Dolan, B. P. et al. Distinct pathways generate peptides from defective ribosomal products for CD8+ T cell immunosurveillance. J. Immunol. 186, 2065–2072 (2011).
Article CAS PubMed MATH Google Scholar
Schubert, U. et al. Rapid degradation of a large fraction of newly synthesized proteins by proteasomes. Nature 404, 770–774 (2000).
Article ADS CAS PubMed MATH Google Scholar
Bassani-Sternberg, M. & Coukos, G. Mass spectrometry-based antigen discovery for cancer immunotherapy. Curr. Opin. Immunol. 41, 9–17 (2016).
Article CAS PubMed Google Scholar
Ziegler, M. C. et al. HIV-1 induced changes in HLA-C*03: 04-presented peptide repertoires lead to reduced engagement of inhibitory natural killer cell receptors. AIDS 34, 1713–1723 (2020).
Article CAS PubMed MATH Google Scholar
Ranasinghe, S. et al. HIV-specific CD4 T cell responses to different viral proteins have discordant associations with viral load and clinical outcome. J. Virol. 86, 277–283 (2012).
Article CAS PubMed PubMed Central MATH Google Scholar
Schieffer, M. et al. Induction of Gag-specific CD4 T cell responses during acute HIV infection is associated with improved viral control. J. Virol. 88, 7357–7366 (2014).
Article PubMed PubMed Central MATH Google Scholar
Ranasinghe, S. et al. Association of HLA-DRB1-restricted CD4⁺ T cell responses with HIV immune control. Nat. Med. 19, 930–933 (2013).
Article CAS PubMed PubMed Central MATH Google Scholar
Zheng, N., Fujiwara, M., Ueno, T., Oka, S. & Takiguchi, M. Strong ability of Nef-specific CD4+ cytotoxic T cells to suppress human immunodeficiency virus type 1 (HIV-1) replication in HIV-1-infected CD4+ T cells and macrophages. J. Virol. 83, 7668–7677 (2009).
Article CAS PubMed PubMed Central Google Scholar
Coulon, P.-G. et al. HIV-infected dendritic cells present endogenous MHC class II–restricted antigens to HIV-specific CD4+ T cells. J. Immunol. 197, 517–532 (2016).
Article CAS PubMed MATH Google Scholar
Veerappan Ganesan, A. P. & Eisenlohr, L. C. The elucidation of non-classical MHC class II antigen processing through the study of viral antigens. Curr. Opin. Virol. 22, 71–76 (2017).
Article CAS PubMed MATH Google Scholar
Fonteneau, J. F., Brilot, F., Münz, C. & Gannagé, M. The tumor antigen NY-ESO-1 mediates direct recognition of melanoma cells by CD4+ T cells after intercellular antigen transfer. J. Immunol. 196, 64–71 (2016).
Article CAS PubMed Google Scholar
Lich, J. D., Elliott, J. F. & Blum, J. S. Cytoplasmic processing is a prerequisite for presentation of an endogenous antigen by major histocompatibility complex class II proteins. J. Exp. Med. 191, 1513–1524 (2000).
Article CAS PubMed PubMed Central Google Scholar
Tewari, M. K., Sinnathamby, G., Rajagopal, D. & Eisenlohr, L. C. A cytosolic pathway for MHC class II–restricted antigen processing that is proteasome and TAP dependent. Nat. Immunol. 6, 287–294 (2005).
Article CAS PubMed Google Scholar
Dengjel, J. et al. Autophagy promotes MHC class II presentation of peptides from intracellular source proteins. Proc. Natl Acad. Sci. USA 102, 7922–7927 (2005).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Sarango, G. et al. The autophagy receptor TAX1BP1 (T6BP) improves antigen presentation by MHC-II molecules. EMBO Rep. 23, e55470 (2022).
Article CAS PubMed PubMed Central Google Scholar
Cachot, A. et al. Tumor-specific cytolytic CD4 T cells mediate immunity against human cancer. Sci. Adv. 7, eabe3348 (2021).
Article ADS CAS PubMed PubMed Central Google Scholar
Hu, Z. et al. Personal neoantigen vaccines induce persistent memory T cell responses and epitope spreading in patients with melanoma. Nat. Med. 27, 515–525 (2021).
Article CAS PubMed PubMed Central MATH Google Scholar
Kreiter, S. et al. Mutant MHC class II epitopes drive therapeutic immune responses to cancer. Nature 520, 692–696 (2015).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Sahin, U. et al. Personalized RNA mutanome vaccines mobilize poly-specific therapeutic immunity against cancer. Nature 547, 222–226 (2017).
Article ADS CAS PubMed Google Scholar
Baxter, A. E., O’Doherty, U. & Kaufmann, D. E. Beyond the replication-competent HIV reservoir: transcription and translation-competent reservoirs. Retrovirology 15, 18 (2018).
Article PubMed PubMed Central Google Scholar
Dubé, M. et al. Spontaneous HIV expression during suppressive ART is associated with the magnitude and function of HIV-specific CD4+ and CD8+ T cells. Cell Host Microbe 31, 1507–1522.e5 (2023).
Article PubMed PubMed Central MATH Google Scholar
Takata, H. et al. An active HIV reservoir during ART is associated with maintenance of HIV-specific CD8+ T cell magnitude and short-lived differentiation status. Cell Host Microbe 31, 1494–1506.e4 (2023).
Article CAS PubMed PubMed Central MATH Google Scholar
White, E. et al. Clonal succession after prolonged antiretroviral therapy rejuvenates CD8+ T cell responses against HIV-1. Nat. Immunol. 25, 1555–1564 (2024).
Article CAS PubMed MATH Google Scholar
Yates, A. D. et al. Ensembl 2020. Nucleic Acids Res. 48, D682–D688 (2020).
CAS PubMed Google Scholar
Martin, M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet. journal 17, 10–12 (2011).
Article Google Scholar
Langmead, B. & Salzberg, S. L. Fast gapped-read alignment with Bowtie 2. Nat. Methods 9, 357–359 (2012).
Article CAS PubMed PubMed Central MATH Google Scholar
Kim, D., Paggi, J. M., Park, C., Bennett, C. & Salzberg, S. L. Graph-based genome alignment and genotyping with HISAT2 and HISAT-genotype. Nat. Biotechnol. 37, 907–915 (2019).
Article CAS PubMed PubMed Central Google Scholar
Li, H. & Durbin, R. Fast and accurate short read alignment with Burrows–Wheeler transform. Bioinformatics 25, 1754–1760 (2009).
Article CAS PubMed PubMed Central MATH Google Scholar
Anders, S., Pyl, P. T. & Huber, W. HTSeq—a Python framework to work with high-throughput sequencing data. Bioinformatics 31, 166–169 (2015).
Article CAS PubMed MATH Google Scholar
Lauria, F. et al. riboWaltz: optimization of ribosome P-site positioning in ribosome profiling data. PLoS Comput. Biol. 14, e1006169 (2018).
Article PubMed PubMed Central MATH Google Scholar
Nelde, A., Kowalewski, D. J. & Stevanović, S. Purification and identification of naturally presented MHC class I and II ligands. Methods Mol. Biol. 1988, 123–136 (2019).
Article CAS PubMed MATH Google Scholar
Eng, J. K., McCormack, A. L. & Yates, J. R. An approach to correlate tandem mass spectral data of peptides with amino acid sequences in a protein database. J. Am. Soc. Mass Spectrom. 5, 976–989 (1994).
Article ADS CAS PubMed MATH Google Scholar
Käll, L., Canterbury, J. D., Weston, J., Noble, W. S. & MacCoss, M. J. Semi-supervised learning for peptide identification from shotgun proteomics datasets. Nat. Methods 4, 923–925 (2007).
Article PubMed Google Scholar
Reynisson, B., Alvarez, B., Paul, S., Peters, B. & Nielsen, M. NetMHCpan-4.1 and NetMHCIIpan-4.0: improved predictions of MHC antigen presentation by concurrent motif deconvolution and integration of MS MHC eluted ligand data. Nucleic Acids Res. 48, W449–W454 (2020).
Article CAS PubMed PubMed Central Google Scholar
Rammensee, H., Bachmann, J., Emmerich, N. P., Bachor, O. A. & Stevanović, S. SYFPEITHI: database for MHC ligands and peptide motifs. Immunogenetics 50, 213–219 (1999).
Article CAS PubMed Google Scholar
Toprak, U. H. et al. Conserved peptide fragmentation as a benchmarking tool for mass spectrometers and a discriminating feature for targeted proteomics. Mol. Cell. Proteomics 13, 2056–2071 (2014).
Article CAS PubMed PubMed Central MATH Google Scholar
Sturm, T. et al. Mouse urinary peptides provide a molecular basis for genotype discrimination by nasal sensory neurons. Nat. Commun. 4, 1616 (2013).
Article ADS PubMed MATH Google Scholar
Schuhmacher, J. et al. Simultaneous identification of functional antigen-specific CD8+ and CD4+ cells after in vitro expansion using elongated peptides. Cells 11, 3451 (2022).
Article CAS PubMed PubMed Central MATH Google Scholar

Download references

Acknowledgements

This work was granted by the Agence Nationale de Recherches sur le SIDA (ANRS-Maladies infectieuses émergentes, A.M. and O.N.), Sidaction for fundings (A.M. and O.N) and Agence Nationale de la Recherche (ANR-22-CE12-0041-02, O.N.). L.B., A.B., and E.L. were supported by ANRS. L.B. and M.P. were supported by Sidaction. G.B. was supported by ANR-20-IDEES-0002. We thank Remy Villette for his expertise using R, Bernard Maillere and his team for the access to the ELISPOT reader, Anne Lopes, Paul Roginski, for discussions, Frederic Suba and Clemence Richetta for access to the L3 facility of ENS-Paris-Saclay. We thank all participants of the ANRS CODEX cohort and the NIH AIDS Research and Reference Reagent Program for providing drugs. This work was supported by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation, WA 4608/1-2, J.S.W.), the Deutsche Forschungsgemeinschaft under Germany’s Excellence Strategy (EXC2180 390900677, J.S.W. and H.-G.R.), the German Cancer Consortium (DKTK, H.-G.R. and J.S.W.), the Ernst Jung Prize for Medicine (H.-G.R.), the Landesforschungspreis of Baden-Württemberg (H.-G.R.), the Wilhelm Sander Stiftung (2016.177.3, J.S.W.), the Deutsche Krebshilfe (German Cancer Aid, 70114948, J.S.W.), and the Fortüne Program of the University of Tübingen (2451-0-0, J.S.W.). The collaboration between Tübingen University and I2BC was supported by the French-German Partnership Hubert Curien Procope 2021 program, A.M. and H.-G.R.

Author information

These authors contributed equally: Annika Nelde, Bertha Cecilia Ramirez, Isabelle Hatin.

Authors and Affiliations

Université Paris-Saclay, CEA, CNRS, Institute for Integrative Biology of the Cell (I2BC), 91190, Gif-sur-Yvette, France
Lisa Bertrand, Bertha Cecilia Ramirez, Isabelle Hatin, Hugo Arbes, Pauline François, Stéphane Demais, Sylvie Grégoire, Guillaume Beauclair, Mathias Pereira, Audrey Esclatine, Olivier Namy & Arnaud Moris
Sorbonne Université, Inserm U1135, CNRS ERL 8255, Centre d’Immunologie et des Maladies Infectieuses (CIMI-Paris), 75013, Paris, France
Lisa Bertrand, Sylvie Grégoire, Anne Bet, Mathias Pereira, Raphaël Jeger-Madiot, Sylvain Cardinaud, Stéphanie Graff-Dubois, Assia Samri, Brigitte Autran & Arnaud Moris
Department of Peptide-based Immunotherapy, University and University Hospital Tübingen, 72076, Tübingen, Germany
Annika Nelde & Juliane Walz
Institute of Immunology, University of Tübingen, 72076, Tübingen, Germany
Annika Nelde, Cécile Gouttefangeas, Hans-Georg Rammensee & Juliane Walz
Cluster of Excellence iFIT (EXC2180) “Image-Guided and Functionally Instructed Tumor Therapies”, University of Tübingen, Tübingen, Germany
Annika Nelde, Cécile Gouttefangeas, Hans-Georg Rammensee & Juliane Walz
Laboratoire de Biologie et Modélisation de la Cellule, Ecole Normale Supérieure de Lyon, CNRS, UMR 5239, Inserm, U1293, Université Claude Bernard Lyon 1, 46 allée d’Italie F-69364, Lyon, France
Emmanuel Labaronne, Laura Guiguettaz & Emiliano P. Ricci
ADLIN Science, Evry-Courcouronnes, France
Emmanuel Labaronne
CIRI, Centre International de Recherche en Infectiologie, Univ Lyon, Inserm, U1111, Université Claude Bernard Lyon 1, CNRS, UMR5308, ENS de Lyon, F-69007, Lyon, France
Didier Decimo
IRIM, UMR 9004, CNRS, Université de Montpellier, Montpellier, France
Antoine Gross
Leibniz Institute of Virology, Hamburg, Germany
Maja C. Ziegler & Marcus Altfeld
ESPCI Paris, PSL University, Spectrométrie de Masse Biologique et Protéomique, CNRS UAR2051, Paris, France
Yann Verdier & Joelle Vinh
Vaccine Research Institute (VRI), INSERM-U955 (IMRB) Équipe 16, Université Paris-Est Créteil (UPEC), Créteil, France
Sylvain Cardinaud
German Cancer Consortium (DKTK) and German Cancer Research Center (DKFZ), partner site Tübingen, 72076, Tübingen, Germany
Cécile Gouttefangeas, Hans-Georg Rammensee & Juliane Walz
Centre Hospitalier Universitaire d’Orléans, Orléans, France
Laurent Hocqueloux
Université Paris Saclay, Inserm, CEA, AP-HP, UMR1184 IDMIT, Department of Internal Medicine & Clinical Immunology, Bicêtre Hospital, Le Kremlin-Bicêtre, Bicêtre, France
Olivier Lambotte
Clinical Collaboration Unit Translational Immunology, German Cancer Consortium (DKTK), Department of Internal Medicine, University Hospital Tübingen, 72076, Tübingen, Germany
Juliane Walz

Authors

Lisa Bertrand
View author publications
Search author on:PubMed Google Scholar
Annika Nelde
View author publications
Search author on:PubMed Google Scholar
Bertha Cecilia Ramirez
View author publications
Search author on:PubMed Google Scholar
Isabelle Hatin
View author publications
Search author on:PubMed Google Scholar
Hugo Arbes
View author publications
Search author on:PubMed Google Scholar
Pauline François
View author publications
Search author on:PubMed Google Scholar
Stéphane Demais
View author publications
Search author on:PubMed Google Scholar
Emmanuel Labaronne
View author publications
Search author on:PubMed Google Scholar
Didier Decimo
View author publications
Search author on:PubMed Google Scholar
Laura Guiguettaz
View author publications
Search author on:PubMed Google Scholar
Sylvie Grégoire
View author publications
Search author on:PubMed Google Scholar
Anne Bet
View author publications
Search author on:PubMed Google Scholar
Guillaume Beauclair
View author publications
Search author on:PubMed Google Scholar
Antoine Gross
View author publications
Search author on:PubMed Google Scholar
Maja C. Ziegler
View author publications
Search author on:PubMed Google Scholar
Mathias Pereira
View author publications
Search author on:PubMed Google Scholar
Raphaël Jeger-Madiot
View author publications
Search author on:PubMed Google Scholar
Yann Verdier
View author publications
Search author on:PubMed Google Scholar
Joelle Vinh
View author publications
Search author on:PubMed Google Scholar
Sylvain Cardinaud
View author publications
Search author on:PubMed Google Scholar
Stéphanie Graff-Dubois
View author publications
Search author on:PubMed Google Scholar
Audrey Esclatine
View author publications
Search author on:PubMed Google Scholar
Cécile Gouttefangeas
View author publications
Search author on:PubMed Google Scholar
Marcus Altfeld
View author publications
Search author on:PubMed Google Scholar
Laurent Hocqueloux
View author publications
Search author on:PubMed Google Scholar
Assia Samri
View author publications
Search author on:PubMed Google Scholar
Brigitte Autran
View author publications
Search author on:PubMed Google Scholar
Olivier Lambotte
View author publications
Search author on:PubMed Google Scholar
Hans-Georg Rammensee
View author publications
Search author on:PubMed Google Scholar
Emiliano P. Ricci
View author publications
Search author on:PubMed Google Scholar
Juliane Walz
View author publications
Search author on:PubMed Google Scholar
Olivier Namy
View author publications
Search author on:PubMed Google Scholar
Arnaud Moris
View author publications
Search author on:PubMed Google Scholar

Contributions

A.M. conceived and designed the project together with O.N. for the Riboseq experiments. Design and performed the experiments: L.B., A.N., B.C.R., I.H., S.G., D.D., L.G., A.B., Y.V. Analysed the data: H.A., P.F., L.B., G.B., A.N., S.D., B.C.R., I.H., S.G., Y.V., J.V, M.P., S.G.-D., J.W. O.N., and A.M. Contributed reagents/materials/analysis tools: H.A., P.F., G.B., I.H., A.G., M.C.Z., C.G., M.A., L.H., E.R., E.L., R.J.-M., S.C., A.E., A.S., B.A., O.L., H.-G.R., J.W., O.N., and A.M. Wrote the paper: L.B. and A.M. with contributions from H.A., A.N., and O.N. and all authors.

Corresponding authors

Correspondence to Olivier Namy or Arnaud Moris.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Sandra Pankow, and the other, anonymous, reviewers for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Reporting Summary

Transparent Peer Review file

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Bertrand, L., Nelde, A., Ramirez, B.C. et al. Unveiling conserved HIV-1 open reading frames encoding T cell antigens using ribosome profiling. Nat Commun 16, 1707 (2025). https://doi.org/10.1038/s41467-025-56773-2

Download citation

Received: 11 August 2023
Accepted: 29 January 2025
Published: 18 February 2025
DOI: https://doi.org/10.1038/s41467-025-56773-2