Post-Translational Modifications Modulate the HLA-DR3 Restricted Epitope Landscape of Sjögren’s Associated Autoantigens Authors: Danmeng Li, Alexandria Voigt, Cuong Q. Nguyen Published: 2026 ABSTRACT Background and Objectives: Sjögren’s disease (SjD) is a chronic autoimmune disorder in which the immune system attacks the glands that produce tears and saliva, leading to symptoms such as dry eyes and dry mouth. If left untreated, SjD can also cause inflammation and damage to other parts of the body, including the skin, lungs, kidneys, and nervous system, and increase the risk of developing lymphoma. The human leukocyte antigen (HLA) class II molecule HLA-DR3 is strongly associated with SjD. Materials and Methods: To investigate how post-translational modifications (PTMs) influence the presentation of SjD-associated autoantigens by HLA-DR3, we employed a computational framework to determine the binding of PTM-mimic peptides to HLA-DR3. We further supported the in-silico results with in-vitro experiments. Results: Our analysis revealed that PTM-mimic substitutions at canonical anchor positions rarely improved predicted binding affinity using the Stabilized Matrix Method, with most modifications resulting in reduced affinity. However, a comprehensive analysis of full-length SjD-associated autoantigen sequences (Ro60, Ro52, La) identified discrete regions with high densities of PTM-eligible anchor sites, specifically, the Ro60 HEAT solenoid, Ro52 RING/B-box/PRY-SPRY modules, and the La motif-RRM1 region, suggesting that PTMs may alter epitope presentation in a sequence-dependent manner. Experimental validation of selected PTM-mimic peptides showed enhanced T cell responses, which were associated with increased binding affinity to HLA-DR3. Structural modeling of a representative complex revealed that PTM-mimic peptides adopt a slightly shifted backbone orientation and altered side-chain positioning, leading to a larger peptide–DR3 interaction interface. Conclusions: These findings provide new insights into the role of PTMs in shaping the immunogenicity of SjD-associated autoantigens and highlight the potential for PTM-mimic peptides to modulate T cell responses in SjD. 1. Introduction Sjögren’s disease (SjD) is a chronic autoimmune disease characterized by lymphocytic infiltration of the salivary and lacrimal glands, resulting in xerostomia and xerophthalmia, with additional systemic manifestations in a substantial proportion of patients. It is one of the most common connective tissue diseases, affecting approximately 0.5–1% of adults and showing a strong female predominance [1]. Serological detection of autoantibodies against Ro/SSA (Ro60 and Ro52/TRIM21) and La/SSB remains central to diagnosis and is incorporated in the 2016 ACR-EULAR classification criteria [2]. Despite their diagnostic utility, these ribonucleoproteins are ubiquitous intracellular molecules, and the processes that make them immunogenic in SjD remain poorly understood. Post-translational modifications (PTMs) have emerged as important contributors to immune recognition in several autoimmune diseases. In rheumatoid arthritis, the citrullination of joint proteins generates neoepitopes that stimulate the production of anti-citrullinated protein antibodies [3]. Additionally, carbamylation produces homocitrulline-containing antigens associated with severe phenotypes, even in patients lacking anti-citrullinated protein antibodies (ACPAs) [4]. In systemic lupus erythematosus (SLE), the apoptosis-associated phosphorylation of histones creates immunogenic epitopes [5], and oxidative modifications, such as 4-hydroxynonenal-modified Ro60, accelerate autoimmunity in vivo [6]. Recent reviews have highlighted that citrullination, acetylation, oxidation, and phosphorylation can reshape self-proteins in ways that promote their recognition by autoreactive lymphocytes [7]. Together, these findings suggest that PTMs can modify the biochemical and immunological properties of self-antigens, thereby contributing to the breakdown of immune tolerance. In SjD, several observations suggest a similar role for PTMs in influencing the antigenicity of Ro and La proteins. Ro52 undergoes phosphorylation that affects its E3 ubiquitin ligase activity and may expose regions of the molecule to immune recognition [8]. Oxidized Ro60 promotes accelerated autoimmune responses in experimental systems [9]. More recently, studies in salivary gland epithelial cells (SGECs) have identified dysregulated phosphorylation pathways, including increased ETS1 and phospho-ETS1 levels [10] as well as elevated lipid-reactive oxygen species (ROS)-associated STAT4 phosphorylation driven by reduced GPX4 expression [11]. These findings support the idea that Ro60, Ro52, and La are exposed to PTM-rich environments within disease-relevant tissues. However, existing studies largely focus on individual residues or specific signaling pathways and do not define how PTMs distributed along the full length of these antigens might influence the repertoire of peptides available for antigen presentation by human leukocyte antigens (HLAs) or specifically SjD-associated or risk HLAs. Therefore, a broader analytical process is needed to examine the impact of PTMs on Ro60, Ro52, and La. The key question is whether these modifications reshape the landscape of peptides capable of enhancing antigen presentation by HLAs, specifically SjD-risk alleles such as HLA-DR3. To address this gap, we undertook a comprehensive, sequence-wide in silico analysis comparing predicted HLA-DR3 binding between native and PTM-modified forms, complemented by experimental validation of selected epitopes. This integrated approach provides the foundation for determining whether PTM-driven changes have the potential to influence antigen processing and T-cell response in the autoimmune process of SjD. 2. Materials and Methods 2.1. Selection of SjD-Associated Autoantigenic Peptides Full-length human Ro60, Ro52, and La sequences (Ro60: UniProt P10155, 525 amino acids (aa); Ro52: UniProt P19474, 475 aa; La: UniProt P05455, 408 aa) were scanned using a sliding-window approach that generated overlapping 15-mer peptides at a step size of one residue. This approach follows established computational epitope scanning strategies applied to autoimmune antigens [12]. The overlapping peptide dataset provided a sequence-wide and unbiased survey of potential HLA-DR3-restricted epitopes of Ro52, Ro60, and La antigens. We also utilized a second peptide dataset using the Stabilized Matrix Method (SMM), which had been previously performed [13]. The SMM-peptide dataset consists of 15-mer peptides computationally predicted to bind HLA-DR3, derived from Ro60, Ro52, and La. 2.2. Binding Core Prediction and Anchor Position Assignment For each peptide in overlapping- and SMM-peptide datasets, the predicted 9-mer binding core and the corresponding anchor positions were obtained using the NetMHCIIpan 4.1 binding core predictor [14] (Technical University of Denmark, Lyngby, Denmark) restricted to HLA-DR3. The predicted core was treated as the reference register for subsequent substitution analysis. Anchor residues at positions P1, P4, P6, and P9 identified in this manner were recorded to determine PTM-eligible substitution sites in this study. 2.3. Post-Translational Modification–Mimic Substitutions PTM-mimic residues were introduced only at canonical HLA-DR3 anchor positions and only when a substitution had an established biochemical precedent for modeling the charge or polarity associated with the corresponding modification. Deamidation was approximated by replacing glutamine with glutamate, which removes the amide group while preserving the side-chain geometry [15]. Citrullination was modeled by substituting arginine with glutamine, a widely used approach to neutralize the guanidinium group and mimic the loss of positive charge [16,17]. Phosphorylation was modeled by replacing serine or threonine with aspartate to introduce a negative charge, a widely used phospho-mimetic strategy [18,19]. Acetylation of lysine was mimicked by substituting lysine with glutamine to eliminate the ε-amino charge while preserving side-chain size [20,21]. Methionine oxidation was approximated by replacing methionine with glutamine, which increases polarity relative to the thioether side chain. This has been used as an oxidation-like mimic in peptide-MHC studies [15]. Aromatic hydroxylation was represented by substituting phenylalanine with tyrosine, which introduces a hydroxyl group while maintaining aromaticity [22]. These six substitution classes comprised the PTM-mimic variants evaluated in this study. In the SMM-peptide dataset, for each eligible anchor position, a PTM-mimic 15-mer peptide was generated by replacing only the anchor residue while maintaining all other native positions. Each PTM-modified peptide was paired with its corresponding native peptide of identical length and predicted register. 2.4. HLA-DR3 Binding Prediction All native and PTM-mimic peptides were evaluated using the IEDB NetMHCIIpan 4.1 BA predictor [14] (Technical University of Denmark, Lyngby, Denmark) restricted to HLA-DRB1*03:01 (DR3). Predictions were recorded as IC50 (nM) and percentile rank. The log2(Native/PTM IC50) values were calculated to quantify both the magnitude and direction of PTM-associated affinity changes. This transformation places increases and decreases in predicted binding on a common fold-change scale, allowing for direct comparison of affinity shifts across peptides with different baseline IC50 values. The use of log2-transformed fold-change values follows standard practice for comparing peptide-MHC affinity shifts [23]. 2.5. Dataset Integration The native and PTM pairs were analyzed within the sliding-window dataset derived from full-length Ro60, Ro52, and La. The SMM-peptide dataset was mapped to its corresponding positions within the dataset and annotated accordingly. All unique native and PTM pairs were included in the computational analysis pipeline. Structural modeling and in vitro validation were performed on selected peptide pairs. 2.6. Data Visualization Data visualizations for computational analyses were generated using Python 3.10 [24]. Dataset handling was performed using Pandas, version 3.0.1 [25], and plots were created with Matplotlib v3.7 [26]. The same axis scaling, smoothing parameters, and color mapping were applied across Ro60, Ro52, and La to ensure comparability among figures. 2.7. Validation of the Selected PMT-Mimics Using NOD.DR3 Splenocytes Four pairs of native and PTM 15-mer peptides were selected and synthesized (Alan Scientific, Gaithersburg, MD, USA) at >90% purity (HPLC-confirmed) and reconstituted at 1 mg/mL in sterile buffer. Screening of the PTM peptides was conducted using splenocytes from NOD-DR3 mice carrying the human DR3 HLA haplotype. Briefly, NOD.DR3 spleens were mashed through a 70 μm strainer, and red blood cells were lysed from the single cell suspension using lysis buffer (0.802% NH4Cl, 0.084% NaHCO3, and 0.037% EDTA) for 13.5 min. Each peptide was incubated at 100 ug/mL with 2 × 105 splenocytes (37 °C, 5% CO2, 48 h). Where indicated, plates were coated (4 °C, overnight) with 50 µL of 10 µg/mL anti-CD3 (BD Pharmingen, Franklin Lakes, NJ, USA). After washing the plate, 5 µg/mL anti-CD28 (BD Pharmingen, Franklin Lakes, NJ, USA) was added to the wells prior to the addition of other well contents. The supernatant was then collected to measure the secreted IL-2 (BioLegend, San Diego, CA, USA) according to the manufacturer’s instructions. The plate was run on Tecan NanoQuant, Männedorf, Switzerland, and the results were analyzed with Python 3. 2.8. Structural Modeling and Interface Analysis Structural comparison of native and PTM-mimic peptides was performed for four representative Ro60-derived peptide pairs. Predicted HLA-DRB1*03:01-peptide complexes were generated using TCRmodel, Rockville, MD, USA [27], which provides template-based modeling of class II peptide–MHC interactions. Models for native and PTM-mimic peptides were obtained separately and aligned in PyMOL, New York, NY, USA (The PyMOL Molecular Graphics System, Version 2.5 Schrödinger, LLC.), using the DR3 α/β heterodimer as the structural reference. Peptide–HLA interface properties were quantified using PDBsum, Cambridge, UK (RRID:SCR_006511) [28]. PDBsum-generated metrics were extracted directly from the server output without manual modification. Structural figures were prepared in PyMOL. 2.9. Determining the Specificity of the PTM-Mimics Using SjD-Associated T Cell Receptor (TCR) Transduced T Cells To generate SjD-associated TCR-transduced cells, lentivirus expressing the TCR from SjD pathogenic T cells was produced as described previously [29]. To isolate antigen-presenting cells (APCs), splenocytes of NOD.DR3 were isolated as described previously and stained with 3 µL PE anti-CD3, followed by incubation with anti-PE microbeads (Miltenyi Biotec, Dresden, Germany). Labeled cells were run on a MACS MS column to isolate APCs. Isolated APCs were treated with Mitomycin C (20 min, 37 °C), then plated on a 24-well plate at a density of 2.5 × 105 cells per well. Peptides were added at a concentration of 100 µg/mL. As a negative control for peptide specificity, an unrelated peptide, ribonucleotide reductase subunit 1 from human alphaherpesvirus 1, residues 709–721 (NVTWTLFDRDTSM), was used. Plates were incubated (5% CO2, 37 °C, 48 h), supplemented with additional media, and then incubated for an additional 24 h. Flow cytometry was performed using a Cytek Aurora (Cytek Biosciences, Fremont, CA, USA) to detect TCR expression via green fluorescent (GFP). Results were analyzed with FlowJo v10 and graphed on Python 3.10. 2.10. Statistical Analysis All quantitative data are presented as mean ± SEM. Paired comparisons between native and PTM-modified peptides were analyzed using a two-tailed paired t-test. Statistical analyses were performed using GraphPad Prism version 10.0.0 for Mac OS X (GraphPad Software, Boston, MA, USA). Statistical significance was defined as p < 0.05. 3. Results 3.1. Overview of the HLA-DR3-Specific PMT Epitope Analysis Workflow The first HLA class II associations with SjD were identified at the DR3 and DR2 loci in Caucasian populations, accounting for up to 90% of the MHC association. The HLA-DR3 allele, particularly the HLA-DRB1*0301-DQB1*0201 haplotype, has been consistently linked to SjD in various populations [30,31,32]. Therefore, we focused on HLA-DR3 to identify the PMT associated with it. To first define how post-translational modifications influence the landscape of HLA-DR3-presented epitopes derived from canonical SjD-associated autoantigens, we implemented the two-stage computational framework summarized in Figure 1. The analysis began with a reference set of 26 Ro60-, Ro52-, and La-derived peptides, previously identified using the SMM approach as high-affinity binders to HLA-DR3 [13]. This SMM-peptide dataset served as a defined baseline for evaluating PTM-dependent response. In addition, we used full-length overlapping libraries for Ro60, Ro52, and La, covering 524, 460, and 394 amino acid antigens, respectively, to map modifiable anchor sites and PTM-sensitive regions within each antigen. The overlapping libraries were generated using the sliding-window approach, evaluating every position across Ro60, Ro52, and La rather than restricting the analysis to annotated domains or a previously reported SMM-peptide dataset. In both datasets, peptides were screened for modifiable amino acids at canonical HLA-DR3 anchor positions (P1, P4, P6, and P9), expanded into PTM-mimic variants, and evaluated by comparing predicted binding affinities between native and modified sequences. Using both the reported epitope set and the full sliding-window scan yielded a unified, high-resolution map of PTM-sensitive positions. 3.2. Predictive Binding Affinity of PTM-Mimic Substitution in the SMM-Peptide Dataset The SMM-peptide dataset comprises 26 previously reported HLA-DR3-restricted epitopes derived from the Ro60, Ro52, and La autoantigens. All peptides contained at least one modifiable anchor residue and yielded a total of 56 PTM-mimic variants suitable for analysis. The native peptides showed a broad distribution of predicted affinities, with most values falling within a range typical for class II ligands of intermediate strength (Figure 2; full data shown in Supplementary Table S1). Substitution at PTM-eligible anchor residues rarely improved predicted binding. Instead, the majority of modified peptides exhibited higher IC50 values than their corresponding native sequences. In total, 48 of 56 modified peptides (85.7%) showed negative log2(Native/PTM IC50) ratios, indicating reduced predicted affinity following substitution. The extent of these changes depended on both the identity of the residue and its anchor position. Mimics of arginine citrullination (R → Q) at P6 and P9 consistently resulted in weaker binding. Lysine acetylation mimics (K → Q) showed a similar pattern. Substitutions modeling phosphorylation (S/T → D) produced more modest effects, and oxidative mimics (F → Y, M → Q) typically resulted in small decreases that did not alter qualitative affinity class. Only a small subset of peptides showed any improvement, and these gains were limited in magnitude. Overall, PTM-mimic substitutions within the SMM-peptide epitopes produced limited changes to HLA-DR3 binding. Given the narrow dynamic range observed across this set, we next asked whether regions outside the annotated epitopes might harbor positions with stronger PTM sensitivity. This rationale motivated a comprehensive, sequence-wide analysis across the full Ro60, Ro52, and La proteins. 3.3. PTM-Mimic Substitution of the Full-Length Ro60, Ro52, and La Autoantigens After evaluating PTM-mimic substitution shifts within previously reported antigenic peptides, we applied the same anchor-based framework to full-length sequences to assess the frequency and distribution of PTM-eligible sites across the entire protein sequence. Extending the analysis to full-length sequences allowed us to systematically quantify the incidence of modifiable anchor sites of Ro60, Ro52, and La. Using the screening sliding-window technique to generate the overlapping libraries, we identified 459 PTM-eligible peptides of Ro60, 413 PTM of Ro52, and 347 of La, defined as peptides containing at least one residue that could be substituted by a PTM-mimic at a canonical HLA-DR3 anchor position. We next summarized how these potential modifications were distributed across the four DR3 binding-motif positions (P1, P4, P6, and P9). To quantify the distribution of modifiable residues at these canonical anchor sites, we enumerated all possible substitutions at P1, P4, P6, and P9 across each overlapping library. The resulting positional counts are summarized in Figure 3. Across all three antigens, substitutions were not evenly distributed among anchor positions. P6 and P9 consistently exhibited the highest numbers of potential modifications, whereas P1 contributed the fewest. In Ro60, P6 and P9 accounted for 254 and 241 substitutions, respectively, compared with 169 at P4 and 139 at P1. Ro52 displayed a similar pattern, with 214 substitutions at P6 and 227 at P9 and lower counts at P4 (157) and P1 (119). La also showed elevated substitution counts at P6 (215) and P9 (167), relative to P4 (162) and P1 (111). These consistent patterns across Ro60, Ro52, and La establish the positional distribution of modifiable anchor residues prior to PTM-mimic expansion, indicating that potential PTM-permissive substitutions occur more frequently at P6 and P9 than at P1 or P4 across all three autoantigens. 3.4. Representation of PTM-Mimic Classes of Ro60, Ro52, and La Autoantigens To characterize the biochemical diversity of potential modifications, all PTM-eligible peptides were converted into mimic variants representing six predefined PTM classes: deamidation, phospho-mimic, acetylation-like, arginine-modification mimic, oxidation-like, and aromatic hydroxylation. This resulted in 803 PTM-mimic peptides for Ro60, 717 for Ro52, and 654 for La. As illustrated in Figure 4A, the representation of PTM classes varied considerably among the three antigens. Ro60 produced disproportionately large numbers of phospho-mimic (191) and acetylation-like (177) peptides, followed by arginine-modification mimics (124), aromatic hydroxylation (107), oxidation-like (104) peptides, and deamidation (100) peptides. Ro52 was enriched for deamidation-type (168) and phospho-mimic (143) variants, followed by arginine-modification mimics (133), whereas oxidation-like peptides were comparatively rare (40). In La, acetylation-like mimics predominated (212 variants), with smaller but still substantial contributions from deamidation (141) and phospho-mimic (124) variants. Oxidation-like (26) and arginine-modification mimics (63) occurred at much lower frequencies. These differences underscore that each antigen provides a unique substrate for PTM-driven substitution, shaped by its intrinsic residue composition and the distribution of anchor-accessible positions. The diversity of PTM classes represented also suggests that multiple biochemical pathways—including phosphorylation, deamidation, oxidation-like changes, and lysine modifications—could alter epitope presentation across these antigens. We next examined how these PTM-eligible anchor sites were distributed across each antigen’s linear sequence. Mapping the number of modifiable anchor residues in each overlapping 15-mer peptide revealed discrete high-density regions along all three antigens (Figure 4B). In Ro60, prominent peaks were centered approximately around residues 70–150, 355–390, 425–460, and 480–510. Ro52 exhibited multiple enriched intervals located near residues 50–90, 160–200, 250–285, and 420–455. La displayed three broader high-density regions, roughly spanning residues 70–150, 265–300, and 325–400. Windows containing three or more PTM-eligible anchors were largely confined to these intervals, whereas the remaining sequence showed lower and more uniform densities. These profiles indicate that PTM-eligible anchor sites cluster within defined sequence ranges rather than being evenly dispersed across each antigen (Supplementary Table S2). 3.5. Domain-Mapped PTM-Eligible Epitopes of Ro60, Ro52, and La Autoantigens To relate PTM-dependent changes in predicted HLA-DR3 binding to the structural organization of each antigen, we mapped log2(Native/PTM IC50) values onto annotated protein domains. The three antigens displayed distinct profiles, and each contained well-defined regions where PTM-mimic substitutions were associated with lower predicted IC50 values. For instance, as shown in Figure 5A, the strongest decreases in the predicted IC50 of Ro60 were located within HEAT ring 2 and HEAT ring 3. These regions contained many windows with consistently negative log2(Native/PTM IC50) values across multiple PTM classes. The HEAT ring 1 and the distal C-terminal portion showed weaker and less consistent responses. The Y-RNA-binding interface also included several PTM-responsive windows, although the magnitude of these shifts was smaller. Ro52 displayed two major PTM-sensitive regions. One was located within the RING and B-box domains, and the other spanned the PRY-SPRY domain. Both regions contained groups of windows with lower predicted IC50 values after PTM substitution. The coiled-coil segment between these domains showed only scattered responsive positions and produced smaller shifts (Figure 5B). Lastly, La showed a different distribution. Strong PTM-associated effects occurred within the La motif, RRM1, and parts of the central linker. These regions formed several clearly defined intervals, with lower predicted IC50 values. Additional responsive positions were present in RRM2 and near the C terminus, but these changes were smaller and less continuous (Figure 5C). In summary, in the three antigens, positive log2(Native/PTM IC50) values were rare and appeared only at isolated positions. These domain-mapped profiles indicate that PTM-eligible epitopes are concentrated in specific structural regions rather than being evenly distributed along the antigen sequences. 3.6. Structural Modeling of Four Representative Native and PTM-Mimic Peptide-HLA-DR3 Complexes To investigate how PTM-mimic substitutions influence peptide recognition by HLA-DR3, we modeled peptide-HLA-DR3 complexes for four selected native/PTM-mimic pairs and performed comparative interface analysis to quantify peptide-HLA-DR3 interactions. Interface residues and contact metrics, including interface area, salt bridges, hydrogen bonds, and non-bonded contacts, were quantified for each peptide-HLA-DR3 complex. Across the four peptide pairs, PTM-mimic substitution produced measurable but non-uniform changes in those metrics, with distinct and peptide-dependent effects on the DR3α and DR3β chains (Table 1). For Pair 1, the PTM substitution altered the distribution of peptide-HLA contacts between the two chains, with increased DR3 engagement accompanied by reduced DR3 interface area. Notably, changes in hydrogen bonding were not aligned with interface area, decreasing for the DR3α chain while increasing for the DR3β, indicating that contact density and bonding patterns were redistributed rather than uniformly scaled (Figure 6A). Pair 2 similarly showed a redistribution of contacts, characterized by reduced DR3α interface area and a relative shift toward DR3β contributions following PTM substitution (Figure 6B). In contrast, Pair 3 exhibited concurrent increases in interface engagement for both DR3α and DR3β, with the most pronounced gain observed in DR3β non-bonded contacts among the four pairs (Figure 6C). Pair 4 showed the largest increase in DR3β interface area, together with increased hydrogen bonding and non-bonded contacts, while DR3α interactions were maintained or modestly enhanced (Figure 6D). Despite the peptide- and chain-specific differences in the interface metrics quantified above, structural superposition analysis revealed a conserved overall binding mode across all four peptide pairs. A visibly altered backbone conformation was observed in one peptide pair, whereas the remaining three pairs showed highly similar backbone trajectories, with the predicted binding register preserved in all cases. 3.7. Experimental Validation of PTM Showed an Enhanced T Cell Response To experimentally assess whether predicted HLA-DR3 binding by PTM-mimic differences corresponds to measurable T cell responses and whether PTM-mimic substitutions capture functionally relevant features of the corresponding native PTMs, we selected four native and PTM peptide pairs for experimental evaluation as a proof-of-concept. T cell responses were assessed using two complementary assays: T-cell response is defined by IL-2 secretion (quantified by ELISA) and proliferation (tracked via GFP reporter expression). The sequences and predicted HLA-DR3 binding affinities of the four Native-PTM peptide pairs are presented in Table 2. The first three pairs were derived from the Ro60 overlapping analysis library. These peptides were predicted to exhibit substantially increased HLA-DR3 binding affinity following PTM-mimic substitution compared with their corresponding native peptides based on the lower IC50 concentrations following the PTM substitutions. The last peptide pair was selected from the previously reported SMM-peptide dataset of Ro60 and was predicted to bind HLA-DR3 with similarly high affinity in both WT and PTM-mimic forms. As presented in Figure 7A, using NOD-DR3 splenocytes, PTM-mimic peptides elicited higher IL-2 secretion than native peptides for pairs 1 and 2. For pair 1, IL-2 increased from 13.70 ± 2.25 pg/mL to 27.17 ± 3.27 pg/mL and from 0.65 ± 0.39 pg/mL to 4.12 ± 0.52 pg/mL for pair 2. In contrast, pairs 3 and 4 did not show increased IL-2 secretion in response to the native PTM peptides, with mean IL-2 decreased from 16.98 ± 1.31 to 13.37 ± 1.29 pg/mL for pair 3 and from 10.29 ± 3.01 to 6.35 ± 0.85 pg/mL for pair 4. The positive control induced robust IL-2 release, whereas blank and negative controls showed minimal activity. Furthermore, using mitocycin C-treated APCs with SjD-associated TCR-transduced Jurkat T cells expressing GFP fusion protein, we showed that PTM-mimic peptides corresponding to pairs 1 and 2 exhibited a significant increase in GFP+ events compared with native peptides. Pair 1 increased from 2281 ± 47 to 3361 ±253, and pair 2 increased from 1454 ± 184 to 3589 ± 322. For pairs 3 and 4, GFP+ events were comparable between native and PTM-mimic peptides, indicating limited PTM-dependent effects on GFP+ activation under the conditions tested (Figure 7B). The data suggest that PTM could modulate IL-2 secretion and T cell proliferation at various responses. 4. Discussion Post-translational modifications influence antigen processing and MHC class II presentation in several autoimmune conditions. Their contributions are most clearly defined in rheumatoid arthritis, where citrullinated and carbamylated peptides show enhanced binding to DRB1*04:01 [3,4,33]. Studies with SLE samples have demonstrated similar effects with oxidized Ro60 and modified histones [5,34]. Deamidated gliadin peptides in celiac disease provide another example of PTM-dependent improvements in MHC II affinity [35,36]. These findings established that PTMs modify peptide binding and antigenicity, but their broader influence on the Ro60, Ro52, and La autoantigens of SjD has not been explored beyond isolated biochemical observations. Earlier studies focused on the phosphorylation of Ro52 [8], oxidation of Ro60 [9], or specific mimicry-related peptides [37]. None assessed how PTMs distributed across full antigen sequences might influence MHC binding, specifically the SjD-risk allele, HLA-DR3. This gap has limited the ability to evaluate whether PTM effects in SjD resemble patterns in other autoimmune diseases or whether Ro and La antigens have distinct features. The current results address this gap by characterizing how PTMs shape predicted HLA-DR3 binding across the full lengths of Ro60, Ro52, and La. Across the SMM set, most PTM-mimic substitutions (e.g., citrullination, phosphorylation, acetylation, oxidation) reduced predicted HLA-DR3 binding; only a small fraction showed improved affinity, underscoring a limited dynamic range when baseline affinity is already high. In contrast, the full-length scan revealed >2100 PTM-mimic peptides, with PTM-eligible anchor sites clustering in specific sequence motifs: Ro60 HEAT repeats, Ro52 RING/B-box/PRY-SPRY domains, and La RRM1/La-motifs. Domain-resolved maps showed that PTM-induced affinity gains were localized to these structurally permissive regions. Functional validation of four native-PTM pairs confirmed that some of the PTM-mimics could enhance T-cell responses with IL-2 secretion and proliferation. Structural modeling of a representative Ro60 peptide pair demonstrated that PTM-mimics shift side-chain orientations, enlarging the peptide–HLA interface and likely accounting for the observed differences in affinity and immunogenicity. Overall, PTMs selectively reshape the HLA-DR3 epitope landscape of Ro60, Ro52, and La, highlighting potential neoepitopes relevant to SjD pathogenesis. The sliding-window approach for overlapping peptide libraries revealed a diverse set of PTM-eligible anchor positions and more than 2100 PTM-mimic peptides of the Ro60, Ro52, and La proteins. The predicted effects did not occur uniformly. Instead, they concentrated within structurally defined regions that differed among the three antigens. Ro60 displayed the strongest predicted responses within HEAT rings 2 and 3. Ro52 showed prominent clusters in the RING region, the B-box, and the PRY-SPRY domain. The La motif, RRM1, and the central linker exhibited enriched PTM-responsive intervals. These segments are characterized by flexibility, partial solvent exposure, and known roles in RNA or substrate recognition [38]. Similar associations between PTM responsiveness and structural dynamics have been reported for modified autoantigens in SLE and rheumatoid arthritis [5,34,39]. These parallels suggest that PTMs in SjD are likewise most consequential in regions naturally accessible to modification and antigen processing. These structural properties align closely with the principles of MHC class II antigen presentation. MHC class II immunodominance requires that peptides arise from segments that are accessible, protease-sensitive, and capable of adopting registers compatible with the DR3 groove [40]. Each of the PTM-responsive regions identified above fulfills these criteria: the Ro60 HEAT-repeat solenoid undergoes small conformational shifts during RNA engagement; the PRY-SPRY domain of Ro52 contains surface loops that readily adjust; and the La motif and RRM1 include partially exposed regions that reposition during RNA binding [41,42,43]. PTMs arising in these structurally permissive environments may therefore stabilize favorable registers or enhance compatibility with pockets P1, P4, P6, or P9. This mechanism mirrors findings from other autoimmune conditions in which PTMs strengthen pocket interactions or reposition side chains to promote class II loading [44]. Structural analyses provided a framework for interpreting how PTM-mimic substitutions may influence peptide engagement within the HLA-DR3 binding groove. Across the four modeled Native-PTM peptide pairs, PDBsum interface metrics showed that PTM-mimic substitution was associated with peptide- and chain-specific changes in peptide-HLA contacts, reflected in non-uniform shifts in interface area, hydrogen bonding, salt bridges, and non-bonded contacts. Rather than following a single directional trend, these metrics suggested that PTM-mimic substitutions can alter the distribution and composition of interactions across the DR3α and DR3β chains in a context-dependent manner. Consistent with this view, structural superposition further indicated that these interface differences arise without a global change in binding mode. The native and PTM-mimic peptides retained the same predicted binding register and similar overall trajectories within the groove, while localized conformational differences were concentrated near the substituted anchor residue. In some cases, these local changes extended to modest backbone displacement, supporting the idea that PTM-mimic substitutions can be accommodated through subtle, peptide-specific structural adjustments that redistribute contacts along the interface. Experimental validation of four representative Native-PTM pairs largely supported the computational predictions of the PTM-dependent response of HLA-DR3 binding. For two of the four pairs tested (Pair 1, 2), PTM-mimic peptides consistently exhibited higher T cell response than their corresponding native peptides, in agreement with the predicted increases in HLA-DR3 binding affinity. This observation was supported by the GFP proliferation assay, which suggested TCR-dependent activation and distinguished between native and PTM peptides. Pair 4 had a similar HLA-DR3 binding affinity and thus no change in T cell response. However, pair 3 deviated from this overall pattern. Although the PTM-mimic peptide was predicted to exhibit enhanced HLA-DR3 binding, the experimental observation difference between native and PTM peptides was reduced. Together, these results suggest that an in silico PTM-mimic-based screening strategy could be used to identify peptides associated with PTMs. The experimental data demonstrate that some of the peptides identified by this screening could potentially exhibit measurable functional differences when tested, consistent with predicted PTM-dependent effects on HLA-DR3 binding. Several limitations should be noted. PTM-mimic substitutions approximate but cannot fully reproduce native chemical modifications. As a proof-of-concept, only four peptide pairs of the Ro60 autoantigen were experimentally validated, and the functional consequences for antigen presentation and T-cell activation were not evaluated. Further studies are needed to experimentally evaluate whether other SjD-associated autoantigens with specific translational modifications could have a similar impact on T cell function. Structural modeling was performed for a single pair and is presented illustratively rather than comprehensively. Lastly, the study relied largely on computational analysis by applying the in silico predictive modeling of PMT; additional work must be performed to validate the PMT in human patients. Therefore, in vivo PTM mapping in salivary gland tissue will be necessary to determine which modified peptides are processed and recognized during disease development as well as to determine how these modifications contribute to T cell function and the eventual clinical symptoms of SjD. 5. Conclusions Together, these observations could connect SjD to PTM-dependent changes in antigen presentation seen in other autoimmune diseases. Salivary gland epithelial cells in SjD experience oxidative stress, disrupted phosphorylation pathways, and lipid-reactive oxygen species accumulation [11]. These conditions favor multiple PTMs, and the correspondence between the PTM-responsive regions identified here and structurally accessible domains suggests that modified peptides could be generated in vivo. Such peptides may contribute to the heterogeneous CD4+ T-cell responses reported in the disease and are consistent with earlier evidence of epitope diversification in Ro and La autoreactivity [9]. Although additional functional studies are required, these findings provide conceptual support for considering PTM-modified Ro and La peptides as potential contributors to loss of tolerance in SjD. In conclusion, this work characterizes how PTMs influence the HLA-DR3-restricted epitope landscape of Ro60, Ro52, and La. The integration of sequence-wide scanning, PTM-mimic analysis, experimental testing, and structural modeling identifies discrete PTM-eligible regions of the SjD-associated autoantigens. These findings suggest that PTMs can alter the repertoire of peptides available for class II presentation in SjD, providing a foundation for future work to identify PTM-dependent neoepitopes relevant to disease pathogenesis.