Ribosome biogenesis mediates the translational increase of non-optimal codon transcripts during IFN stimulation

14 The interferon response is a signaling pathway unique to vertebrates that links the innate and 15 adaptive immune responses. Interferons signal through a cascade of factors including the JAK-16 STAT pathway to induce the transcription of hundreds of interferon-stimulated genes (ISGs). 17 Although the main interferon signal transduction pathways and ISGs have been elucidated, 18 translational regulation of ISG transcripts is not fully understood. Prior work demonstrated that 19 ribosomal protein RPL28 negatively regulates a subset of ISGs; however, we find that this effect 20 may be due to a reduction in overall ribosome availability. Multi-omics analysis of RNA-seq and 21 LC-MS/MS data reveal proteins, including several ISGs, that are translationally up-regulated in 22 IFN-β-stimulated cells depleted of ribosome biogeneis factor BOP1. Analysis of codon usage 23 demonstrates a significant reduction in codon optimality for proteins that are translationally up-24 regulated during BOP1 knockdown and IFN-β stimulation. Using reporter constructs, we 25 demonstrate that codon non-optimal reporters are translated more than codon-optimized reporters 26 in BOP1-depleted IFN-β cells. We propose that ribosome biogenesis regulates translational fine-27 tuning of integral protein production to ensure optimal interferon responses. 28 29

ribosome biogenesis, translational regulation, codon optimality, interferon response 30 31 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 3

32 The interferon (IFN) response plays a key role in innate immunity, as it orchestrates 33 signalling cascades that induce a rapid antiviral response to protect against pathogens (Le Bon 34 and Tough, 2002; Wang et al., 2017). Upon detection of pathogen-associated molecular patterns 35 (PAMPs) by pattern recognition receptors (PRRs), signaling phosphorylation cascades are 36 activated resulting in the induction of key transcription factors, IRF3, to induce IFN production 37 and secretion. IFNs, in a paracrine and autocrine manner, bind to their appropriate receptor to 38 induce a second signalling cascade, resulting in the activation of transcription factors, STAT, to 39 activate transcription of interferon-stimulated genes (ISGs). Proteins produced from the ISGs act 40 on cellular machinery, leading to an anti-pathogenic state in the cell, with the ultimate goal of 41 pathogen clearance. 42 As a crucial component of immunity, the IFN response must be finely tuned to prevent 43 diseases that can arise from an over- or under-activating immune system (Melki and Frémond, 44 2020). For example, over-activation of IFN can lead to a range of autoinflammatory diseases, 45 known as interferonopathies, which can induce varying effects within the central nervous system 46 and the skin (Melki and Frémond, 2020). The IFN response is regulated by many negative 47 feedback mechanisms. For instance, type I IFN receptor subunits are rapidly down-regulated 48 upon the binding of a type I IFN to its receptor in order to prevent the same cell from re-49 initiating additional signaling (Constantinescu et al., 1995). 50 Recently, ribosomal proteins (RPs) have been implicated in translational controls of IFN 51 signaling (Kerr et al., 2020). Reducing ribosome biogenesis has been shown to reduce mRNA 52 levels of several immune-related transcripts including HMGB2, which produces a cGAS DNA-53 sensing protein, and IFNB1, the key cytokine in the IFN-β pathway and impacts virus infection 54 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 4 by increasing HCMV replication (Bianco and Mohr, 2019). Conversely, ribosome biogenesis 55 factors SBDS and SPATA5 were identified in a CRISPR loss-of-function screen, as they were 56 required for flavivirus-encoded protein production (Ohlson et al., 2023). Ohlson and colleagues 57 propose that a threshold ribosome abundance is required for viral protein expression for 58 productive infection (Ohlson et al., 2023). Previously, we showed that during IFN-β stimulation 59 in Hela cells, RPL28 was significantly increased in both the free ribosome and polysome 60 fractions of purified ribosomes, suggesting a potential role for this RP in the IFN response (Kerr 61 et al., 2020). RPL28 depletion resulted in pre-mature and enhanced production of ISGs in IFN-β-62 treated cells (Kerr et al., 2020). The link between ribosome biogenesis and translational 63 regulation remains ambiguous, and the link to the host immune response remains to be 64 investigated. In this study, we examine the relationship between ribosome biogenesis, 65 translational controls and ISG expression. Our results point to codon usage as a parameter in 66 regulating translation of mRNAs in the IFN response, thus providing an additional novel layer of 67 control in IFN signaling. 68 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 5

69 Cell culture 70 A549 cells were cultured at 37°C in Dulbecco’s Modified Eagle’s Medium (DMEM, Gibco) 71 supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin. 72 Trypan blue exclusion cell viability assay 73 Adherent A549 cells were aspirated, washed with phosphate buffered saline (PBS), and 74 incubated at 37°C for 1 to 3 min with trypsin EDTA (Gibco) until cells began to lift, then mixed 75 with DMEM to homogenize the cells. An equal volume of trypan blue (0.4%, Invitrogen) was 76 added to 15 μL of the well-mixed cell suspension and gently mixed to prevent the formation of 77 bubbles. A 10 μL sample was loaded onto a slide and inserted into an automated cell counter 78 (Countess II, Thermo Fisher). Cell counts were performed using an optimized “A549” method 79 which has been preset with cell size and shape settings specific to the cell line, while adjusting 80 live and dead cell parameters with each slide as needed to ensure representative cell counts for 81 all samples. Two-sample unpaired t-tests assuming equal variances were performed in Microsoft 82 Excel on the cell viability obtained from three biological replicates. 83 siRNA knockdown 84 A total of 3.0 x 105 A549 cells were seeded into 6-well plates and incubated at 37°C for 24 h. 85 Knockdowns were performed by adding 5 μL Dharmafect 1 (Dharmacon) to 195 μL Opti-MEM 86 (Thermo Fisher) in one tube and 2.5 μL 20 μM target SMARTpool siRNA (Dharmacon) to 197.5 87 μL Opti-MEM in a second tube, then incubating at room temperature for 5 min. The contents of 88 tubes A and B were combined and incubated at room temperature for 20 min. Media was 89 aspirated from cells and replaced with this transfection mixture and 1.6 mL of antibiotic-free 90 DMEM supplemented with 10% fetal bovine serum, allowing for incubation at 37°C for 48 h. 91 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 6 Trifluoroethanol cell pellet lysis 92 Cells were washed with PBS, scraped for collection, and pelleted in a microcentrifuge at 93 maximum speed for 15 min at 4°C. Any remaining PBS was removed by micropipette and the 94 pellet was flash frozen in liquid nitrogen. Cells were resuspended by vortexing (VWR) with 150 95 μL of 50% trifluoroethanol (TFE) followed by sonication (Fisher Scientific) for 5 min in an ice 96 bath. Samples were reduced and alkylated for 10 min at 95°C in a solution of 100mM Tris pH 97 8.5, 10 mM tris(2-carboxyethyl)phosphine (TCEP, Bioshop), and 40 mM chloroacetamide 98 (CAA, Acros Organics). Samples were diluted to at most 10% TFE with 50 mM ammonium 99 bicarbonate. Lys-C/trypsin (Promega) was added at a 1:50 ratio of protease:protein and incubated 100 overnight at 37°C, followed by additional lys-C/trypsin at a 1:200 ratio incubated for 6 h at 101 37°C. Following TFE cell pellet lysis and digestion, peptide samples were desalted and purified 102 on C-18 STAGE tips as described.(Rappsilber, Ishihama and Mann, 2003) 103 LC-MS/MS (RPL28 knockdown) 104 Purified peptides were analyzed using a quadrupole time of flight mass spectrometer (Impact II; 105 Bruker Daltonics) on-line, coupled to an Easy nano LC 1200 HPLC (Thermo Fisher) using 106 nanoBooster with methanol and a Captive spray nanospray ionization source (Bruker Daltonics). 107 Buffer A was composed of 0.1% formic acid and 2% acetonitrile in H2O and buffer B was 108 composed of 0.1% formic acid and 80% acetonitrile in H2O. Samples were resuspended in buffer 109 A and 100 ng were loaded onto the instrument running a standard 90-minute gradient. The 110 Impact II was set to data-dependent acquisition. 111 Protein identification and quantification (RPL28 knockdown) 112 A database search was performed using MaxQuant version 1.6.17 using label-free quantitation 113 (LFQ).(Cox and Mann, 2008) Search parameters included the variable modifications methionine 114 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 7 oxidation and N-terminal acetylation, and the fixed modification carbamidomethylation of 115 cysteine residues. A 1% false discovery rate was applied at the protein and peptide level. Mass 116 spectra were searched against a FASTA database for the Homo sapiens (H. sapiens) proteome 117 from UniProt dated 2020. 118 Proteomic statistical and bioinformatic analyses (RPL28 knockdown) 119 Bioinformatic analysis was performed in R using LFQ intensities for quantitation, and 120 differential expression analysis was performed using limma. 121 LC-MS/MS (BOP1 knockdown) 122 Purified peptides were analyzed using a quadrupole time of flight mass spectrometer (timsTOF 123 Pro; Bruker Daltonics) operated in DIA-PASEF mode, coupled to an NanoElute UHPLC system 124 (Bruker Daltonics) with Aurora Series Gen2 (CSI) analytical column (25 cm x 75 μM 1.6 μM 125 FSC C18, with Gen2 nanoZero and CSI fitting; Ion Opticks, Parkville, Victoria, Australia). 126 Buffer A was composed of 0.1% formic acid and 0.5% acetonitrile in H2O and buffer B was 127 composed of 0.1% formic acid in acetonitrile. Samples were resuspended in buffer A and loaded 128 onto the instrument running a standard 30-minute gradient. 129 Protein identification and quantification (BOP1 knockdown) 130 A database search was performed using DIA-NN version 1.8.1 using label-free quantitation 131 (LFQ) (Cox and Mann, 2008). Mass spectra were searched against a FASTA database for the H. 132 sapiens proteome from UniProt. Search parameters included the fixed modifications N-terminal 133 methionine excision and carbamidomethylation of cysteine residues, with no variable 134 modifications. A 1% false discovery rate was applied at the protein and peptide level. 135 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 8 Proteomic statistical and bioinformatic analyses (BOP1 knockdown) 136 Bioinformatic analysis was performed in RStudio using LFQ intensities for quantitation. Proteins 137 were filtered to a minimum quantification of 25% across all samples. Filtered data was fitted to a 138 linear model by limma (Ritchie et al., 2015). Gene ontology (GO) over-representation analyses 139 were performed with the Panther web application (www.panterdb.org) using Fisher’s exact test 140 and multiple testing correction of the p-value with the Benjamini-Hochberg procedure (FDR) 141 (Mi et al., 2019; Thomas et al., 2022). 142 RNA extraction 143 RNA samples were harvested with 500 μL TRI-zol (Invitrogen) and total cellular RNA was 144 extracted according to the Direct-zol RNA Miniprep Kit, including the optional DNase I 145 treatment (Zymo Research, R2052). 146 RNA-seq 147 RNA samples were sequenced by the Sequencing Facility at the School of Biomedical 148 Engineering (University of British Columbia) as follows. Sample quality control was performed 149 using the Agilent 2100 Bioanalyzer or the Agilent 4200 Tapestation. Samples were prepared 150 following the standard protocol for the Illumina Stranded mRNA prep (Illumina). Sequencing 151 was performed on the Illumina NextSeq2000 with Paired End 59 bp x 59 bp reads. Sequencing 152 data was de-multiplexed using Illumina’s BCL Convert. De-multiplexed read sequences were 153 then aligned to the Homo sapiens (hg38 no Alts, with decoys) reference sequence using 154 DRAGEN RNA app on Basespace Sequence Hub (Goyal et al., 2017). 155 Transcriptomic statistical and bioinformatic analyses 156 Differential expression analysis was performed with DESeq2 in RStudio (Love, Huber and 157 Anders, 2014). 158 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 9 Codon usage bioinformatic analysis 159 APPRIS annotations and transcript sequences were downloaded from ensembl 160 (www.ensembl.org) (release 112) (Harrison et al., 2024). Codon adaptation index (CAI) for each 161 sequence was calculated using a locally installed version of the Perl script 162 (CAIcal_ECAI_v1.4pl) (Puigbò, Bravo and Garcia-Vallve, 2008). 163 Codon optimality constructs 164 The plasmids containing the extreme optimal and extreme non-optimal sequences were 165 synthesized by Twist Biosciences into the pTwist CMV vector with insertion by HindIII and 166 NheI. The extreme optimal and extreme non-optimal sequences were originally described and 167 validated by Wu and colleagues; however, we included an ATG at the beginning of the extreme 168 non-optimal region to initiate translation (Wu et al., 2019). The use of capital and lowercase 169 letters separates the discrete components of the reporter, as outlined in each header. 170 T7-5’Bglo-ExOpt-GSGP2A-Nluc-3’Bglo 171 AAGCTTTAATACGACTCACTATAAGGAGacatttgcttctgacacaactgtgttcactagcaacctcaaacagacac172 catggccggcgataacggacccgaagatcgtgacaacggcgacgatggaggttatgctggaaagggagtcggaggcccaaaccctgg173 aaacggcaccttccctggggggttctacggttattatggagccaagggggatttcgacatcgtcgctttcgggtactatggccgtcctatcgg174 acctgggatcattcagaacttcgatgctgcttacgccgctgctatgccaattgagaaggaagatcccgctccatatattttccaggggggtaa175 cgaaaagaacggaaccgctatcgtcggcgatgcaggaatggaaaaggatgactatggggaggaggtcgatcccgacccaatcatggata176 tgaacggtgagaccggggcatacaaggctgccgacgccggtacccgttatggtgaaatggaacccgctgccgaagatttcgccgacgac177 caggagccaccagcctatgtcttcatcattaaggacatgcagggtccctattatgcagccaacttcggggaggacggtttcgaaggagctaa178 ggatttcggcatgaccaacaccggcggtggtaaccgtgagatgaaggggtatgagttcgaacaggccgaagacggggaaaagcgtgaa179 gaggaggagcctggcgacattaagtacatgggctatggtaacgccaaggctgccggaggccagattgagatggcaatgggcggtgcag180 ggggatccggagctactaacttcagcctgctgaagcaggctggagacgtggaggagaaccctggacctAGCGGATCCAGGC181 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 10 CTATGGTGTTCACCTTAGAGGACTTCGTGGGGGACTGGAGGCAGACAGCAGGCTACA182 ACCTGGACCAAGTGCTGGAGCAGGGAGGCGTAAGTTCACTATTCCAGAATCTGGGTG183 TCAGCGTCACACCCA TCCAAAGGATTGTATTGAGCGGAGAAAATGGCCTGAAGATTG184 ACATCCATGTCATCATTCCATATGAGGGGCTTTCCGGTGACCAGATGGGCCAGATTGA185 GAAGATCTTCAAGGTGGTCTACCCTGTGGATGACCACCACTTTAAAGTTATTCTCCAT186 TATGGGACGCTGGTGATTGATGGAGTCACCCCCAACATGATCGACTACTTTGGCAGAC187 CTTATGAAGGCATCGCTGTGTTTGATGGGAAGAAAATAACTGTTACTGGTACCCTGTG188 GAACGGGAATAAAATAATAGATGAGCGCCTCATCAACCCAGATGGTTCTTTGCTCTTC189 AGAGTGACTATCAATGGAGTTACAGGCTGGCGGCTTTGCGAGCGAATCCTGGCCTGA190 gctcgctttcttgctgtccaatttctattaaaggttcctttgttccctaagtccaactactaaactgggggatattatgaagggccttgagcatctg191 gattctgcctaataaaaaacatttattttcattgcaa 192 T7-5’Bglo-ExNonOptwstartATG-GSGP2A-Nluc-3’Bglo 193 AAGCTTTAATACGACTCACTATAAGGAGacatttgcttctgacacaactgtgttcactagcaacctcaaacagacac194 catgcgaagtcaatcactgtcgacgtctctatcgctaaggcgcgtacgaagtttgaaacggagggcgcttctcccgctcagacgcaaaaca195 gtatgcttttgtcatatattgaggcatccgacgagagtggttctttgcctcacaactgcgagatttcgatgtagcatatctgttgcgataaatagta196 attgcgtatctttacatctgcaccataattcgtgggtttggttaaggcatagaagctcacacctgtgtcggccgagccatgtacaatctttacttc197 gctgtaaaacttctaggtcgaggctcctgcaaacgtccacgttatcatgctgtaattcacgatcgagactatgttcccatcttgtgtccaaatgct198 taacactacaacgcaggctaataatactattgcaacttgtgactttgactagctcagtggttcggcacttgaattggacagttcaccaaactacg199 agtcatacatggctcactaataaatggacgttgcactctacaccgagctttactcgctcacgaccgacggcgttatccctttttacgagttccat200 acttgtatggtcttcccacctgtgcctcagactgctacgacaatttacacactcccgaagcgtgctacaccggctcaggatatgcgcgctgtc201 acggtggcgcgtaaaaatacgacaaagcagtccgttgagttgtagcacattagtaactagtggatccggagctactaacttcagcctgctga202 agcaggctggagacgtggaggagaaccctggacctAGCGGATCCAGGCCTATGGTGTTCACCTTAGAGG203 ACTTCGTGGGGGACTGGAGGCAGACAGCAGGCTACAACCTGGACCAAGTGCTGGAG204 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 11 CAGGGAGGCGTAAGTTCACTATTCCAGAATCTGGGTGTCAGCGTCACACCCA TCCAA205 AGGATTGTATTGAGCGGAGAAAATGGCCTGAAGATTGACATCCATGTCATCATTCCAT206 ATGAGGGGCTTTCCGGTGACCAGATGGGCCAGATTGAGAAGATCTTCAAGGTGGTCT207 ACCCTGTGGATGACCACCACTTTAAAGTTATTCTCCATTATGGGACGCTGGTGATTGAT208 GGAGTCACCCCCAACATGATCGACTACTTTGGCAGACCTTATGAAGGCATCGCTGTGT209 TTGATGGGAAGAAAATAACTGTTACTGGTACCCTGTGGAACGGGAATAAAATAATAGA210 TGAGCGCCTCATCAACCCAGATGGTTCTTTGCTCTTCAGAGTGACTATCAATGGAGTT211 ACAGGCTGGCGGCTTTGCGAGCGAATCCTGGCCTGAgctcgctttcttgctgtccaatttctattaaaggttc212 ctttgttccctaagtccaactactaaactgggggatattatgaagggccttgagcatctggattctgcctaataaaaaacatttattttcattgcaa 213 T7-5’Bglo-Rluc-3’Bglo 214 AAGCTTTAATACGACTCACTATAAGGAGacatttgcttctgacacaactgtgttcactagcaacctcaaacagacac215 cATGACTTCGAAAGTTTATGATCCAGAACAAAGGAAACGGATGATAACTGGTCCGCA216 GTGGTGGGCCAGATGTAAACAAATGAATGTTCTTGATTCATTTATTAATTATTATGATTC217 AGAAAAACATGCAGAAAATGCTGTTATTTTTTTACATGGTAACGCGGCCTCTTCTTATT218 TATGGCGACATGTTGTGCCACATATTGAGCCAGTAGCGCGGTGTATTATACCAGACCTT219 ATTGGTATGGGCAAATCAGGCAAATCTGGTAATGGTTCTTATAGGTTACTTGATCATTA220 CAAATATCTTACTGCATGGTTTGAACTTCTTAATTTACCAAAGAAGATCATTTTTGTCG221 GCCATGATTGGGGTGCTTGTTTGGCATTTCATTATAGCTATGAGCATCAAGATAAGATC222 AAAGCAATAGTTCACGCTGAAAGTGTAGTAGATGTGATTGAATCATGGGATGAATGGC223 CTGATATTGAAGAAGATATTGCGTTGATCAAATCTGAAGAAGGAGAAAAAATGGTTTT224 GGAGAATAACTTCTTCGTGGAAACCATGTTGCCATCAAAAATCATGAGAAAGTTAGA225 ACCAGAAGAATTTGCAGCATATCTTGAACCATTCAAAGAGAAAGGTGAAGTTCGTCG226 TCCAACATTATCATGGCCTCGTGAAATCCCGTTAGTAAAAGGTGGTAAACCTGACGTT227 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 12 GTACAAATTGTTAGGAATTATAATGCTTATCTACGTGCAAGTGATGATTTACCAAAAAT228 GTTTATTGAATCGGACCCAGGATTCTTTTCCAATGCTATTGTTGAAGGTGCCAAGAAG229 TTTCCTAATACTGAATTTGTCAAAGTAAAAGGTCTTCATTTTTCGCAAGAAGATGCAC230 CTGATGAAATGGGAAAATATATCAAATCGTTCGTTGAGCGAGTTCTCAAAAATGAACA231 AGGGCgctcgctttcttgctgtccaatttctattaaaggttcctttgttccctaagtccaactactaaactgggggatattatgaagggccttg232 agcatctggattctgcctaataaaaaacatttattttcattgcaa 233 RNA transfection 234 Cells were 60-80% confluent prior to RNA transfection. To create the transfection solution, 5 μL 235 Lipofectamine 3000 (Thermo Fisher) was added to 195 μL Opti-MEM (Thermo Fisher) in one 236 tube and 1 μg of capped and tailed RNA was added with Opti-MEM up to 200 μL in a second 237 tube, then incubated at room temperature for 5 min. The contents of tubes A and B were 238 combined and incubated at room temperature for 15 min. Media was aspirated from cells and 239 replaced with this transfection mixture and 1.6 mL of antibiotic-free DMEM supplemented with 240 10% fetal bovine serum. 241 Luciferase assay 242 Cells were washed with PBS and lysed in 500 μL 1X Passive Lysis Buffer for 15 min with gentle 243 rocking (Promega). 25 μL of lysate was combined with 25 μL of the respective luciferase reagent 244 (Promega; E2920, N1110) and immediately read on Spark microplate reader (Tecan). 245 Sucrose gradient fractionation 246 Approximately 2x106 A549 cells incubated in PBS with 0.1 mg/mL cycloheximide for 10 min at 247 37°C then lysed in mammalian lysis buffer (0.3 M NaCl, 15 mM Tris-Cl pH 7.5, 15 mM MgCl2, 248 1% Triton X-100, 0.1 mg/mL cycloheximide, 1 mg/mL heparin). Samples were centrifuged at 249 13000 rpm for 10 min at 4°C. The supernatant was then gently added to a 10-50% sucrose 250 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 13 gradient. Gradients were spun at 35000 rpm at 4°C for 2 h 20 min in an ultracentrifuge 251 (Beckman Coulter). Sucrose gradients were fractionated (Brandel) and the following settings 252 were used for detection: sensitivity 0.5, chart speed 60, peak separator off, noise filter 1.5, 253 manual, 0.375 mL/min. The surface area of the 60S peak was calculated in Fiji.(Schindelin et al., 254 2012) The total surface area was calculated by cutting out and weighing the trace. 255 256

257 Ribosomal protein RPL28 depletion results in an increase in ISG proteins 258 We previously showed that depletion of RPL28 by siRNA transfection in IFN-β-treated 259 Hela cells resulted in an increase in ISG expression (Wu et al., 2015; Kerr et al., 2020). We 260 investigated whether this effect extends to other cells. We chose lung epithelial A549 cells as a 261 model as the innate immune signaling pathway is relatively intact in these cells (Wu et al., 2015). 262 Towards this, RPL28 was depleted by transfecting siRNAs in A549 cells for 48 h, followed by 263 treating cells with IFN-β. Cells were harvested and lysed cells at 8, 12, and 24 h post-transfection 264 and cell lysates were analyzed by LC-MS/MS. As controls, we also analyzed protein expression 265 by LC-MS/MS of transfected cells with siRNAs targeting RPS26, RPS28, or an siRNA control 266 targeting firefly luciferase. Both RPS26 and RPS28 have previously been implicated in the 267 immune response (Cui et al., 2014; Gripp et al., 2014; Ferretti et al., 2017; Wei et al., 2019). 268 We first investigated whether depletion of RPL28 had effects on cell viability by using 269 trypan blue cell staining at 24 and 48 h post-transfection. Cell viability was similar between 270 RPL28 siRNA and control FLuc-siRNA transfected cells at 24 h post-transfection (83.3% 271 compared to 82.3%, p-value = 0.29) and at 48 h post-transfection (74.3% compared to 73.3%, p-272 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 14 value = 0.37) (Fig. 1A). Thus, RPL28 knockdown did not have any obvious effects on cell 273 viability in A549 cells. 274 We next used LFQ to examine whole-proteome changes in cells transfected with RPL28 275 siRNAs vs control-transfected cells. RPL28 was reduced by 37% in the knockdown condition as 276 compared to control cells, thus validating the knockdown approach (p-value = 0.0079). To 277 determine whether RPL28 depletion affects ribosome abundance as has been previously 278 suggested (Robledo et al., 2008), differential expression analyses were performed across the 279 three RP knockdown conditions (FLuc control vs RPL28, RPS26, RPS28) over time. V olcano 280 plots depicted the targeted RPs across time as a control and across the RP knockdowns (Fig. 1C-281 F). Each individual RP knockdown showed a depletion of its target as expected (RPL28 37%; 282 RPS26 29%; RPS28 18%), but other RPs were also reduced in each RP knockdown condition 283 (Fig. 1G). Specifically, in RPL28 knockdown, 96% of the ribosomal proteins from both 284 ribosomal subunits were downregulated (73/76 RPs) and 36% were significantly downregulated 285 (27/76 RPs). In the RPS26 knockdown, 95.9% of ribosomal proteins were reduced (70/76 RPs), 286 with 63% being significantly reduced (48/76 RPs). Furthermore, 99% of ribosomal proteins were 287 reduced during RPS28 knockdown, with 68% being significantly reduced (52/76 RPs). We also 288 found that several key ribosome biogenesis proteins were down-regulated. In contrast, the 289 majority of ISGs were up-regulated upon RP knockdown, including several ISGs that were 290 previously reported (Fig. 1G) (Kerr et al., 2020). These results demonstrated that depleting 291 RPL28, RPS26 or RPS28 led to a broad reduction of ribosomal proteins from both subunits, thus 292 suggesting that overall ribosome abundance is reduced. 293 BOP1 knockdown results in an increase in ISG proteins 294 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 15 Next, we sought to identify whether ribosome biogenesis was the broader mechanism 295 underlying the enhanced ISG expression in IFN-β-treated cells. Towards this, we depleted a key 296 60S subunit biogenesis factor, BOP1, through siRNA transfection. BOP1 is part of the PeBoW 297 complex that is required for 60S maturation and processing of 28S and 5.8S rRNA (Strezoska, 298 Pestov and Lau, 2000; Hölzel et al., 2005). A549 cells transfected with BOP1 or control FLuc 299 siRNAs for 48 h were subsequently treated with or without IFN-β. RNA samples were collected 300 at 8 and 12 h post-stimulation for analysis by RNA-seq while protein samples were collected at 301 8, 12, and 24 h post-stimulation for analysis by LC-MS/MS. Cells transfected with BOP1 302 siRNAs showed that BOP1 was depleted by 73% at the RNA level and 28% at the protein level 303 (Fig. 2B,D). Furthermore, sucrose gradient fractionation of BOP1-depleted cells showed an 304 average reduction of 19% in the 60S subunit (Supplemental Figure 1). Finally, cells transfected 305 with BOP1 siRNAs did not adversely affect cell viability at 24 and 48 h under these conditions 306 (Fig 2A). In summary, these results demonstrated BOP1 depletion in A549 cells that reduced 60S 307 subunit levels. 308 We next analyzed the protein changes that occur between the control and BOP1 309 knockdown conditions in IFN-β-stimulated A549 cells. At 12 h post-IFN, a subset of ISGs 310 (35/56 ISGs) was increased in BOP1-depleted cells compared to siFluc-transfected cells (Fig. 311 2C). We also identified several non-ISG proteins that were upregulated in BOP1-depleted cells 312 including SARAF, GNB4, and ADCY9 (3296/6521). A one-sided Fisher’s exact test for 313 enrichment of these subset of ISGs compared to non-ISGs amongst up-regulated proteins 314 resulted in a p-value = 0.04918, demonstrating that these ISGs in particular are over-represented 315 among the up-regulated proteins during BOP1 knockdown. These results showed that ribosome 316 biogenesis results in an increase in proteins levels in a subset of ISGs under IFN treatment. 317 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 16 A subset of proteins are translationally up-regulated during BOP1 knockdown 318 We next investigated whether the increase in ISG expression in IFN-β-treated BOP1-319 depleted cells is due to changes at the transcriptome level. Specifically, we focused on the subset 320 of ISGs that were upregulated at the protein level (Fig 2C). RNA-seq analysis revealed minimal 321 changes of the subset of ISG mRNAs, thus ruling out that transcription or mRNA decay effects 322 on differential ISG protein expression (Fig. 3A). 323 The transcriptome and proteome data suggested that the effects could be at the level of 324 translation. We compiled both datasets from RNA-seq and LC-MS/MS and analyzed the fold-325 changes of protein and RNA to infer translational effects. First, we benchmarked this approach 326 with the differential expression of A549 cells with and without 12 h of IFN-β stimulation (Fig. 327 3B) as well as in A549 cells transfected with BOP1 vs control siRNA (Fig. 3C). This first 328 benchmark alignment revealed that several proteins and genes were up-regulated, many of which 329 were ISGs, supporting that IFN stimulation results in an increase in the IFN pathway. A second 330 benchmark was performed with the differential expression of A549 cells with BOP1 knockdown 331 or a control. The second benchmark dataset demonstrated relatively normal distribution across all 332 data points, while the profile of BOP1 also validated that while both the transcript and protein 333 levels are reduced due to the siRNA knockdown (RNA 72.9% reduction; protein 28.3% 334 reduction), there is a greater effect at the transcript level. 335 Next, we performed differential expression analysis of BOP1 vs control siRNA 336 transfected cells at 12 h of IFN-β stimulation (Fig. 3D). Most ISGs showed an increase in 337 differential expression at the protein level compared to the RNA level (41/54). Two ISGs, ISG15 338 and MX1, were previously identified by Kerr et al as part of a subset of ISGs that are 339 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 17 upregulated during RPL28 knockdown (Kerr et al., 2020). These results strongly suggested that 340 these ISGs are translationally upregulated in BOP1-depleted IFN- β-stimulated cells. 341 To determine whether other proteins were up-regulated at the protein level compared to 342 the RNA level, GO term enrichment analysis was performed on the set of proteins that were up-343 regulated (Fig. 3E). GO terms involved in nucleic acid metabolism and RNA metabolism were 344 enriched, while those involved in lipid metabolism were reduced. Subsequent analyses examine 345 these datasets more closely. 346 Proteins that are translationally up-regulated have lower codon optimality 347 Our results suggested that a subset of mRNAs is translationally upregulated in BOP1-348 depleted cells. One possibility is that there may be differences in codon usage of the ISG ORF 349 that could impact translation. Codon usage of an mRNA is assessed by calculating the codon 350 adaptation index (CAI), which quantifies how closely a gene’s codon usage matches the 351 synonymous codon frequency within a reference set (Puigbò, Bravo and Garcia-Vallve, 2008). 352 Using the tool CAIcal, the CAI was calculated for the aligned RNA-seq and LC-MS/MS datasets 353 (Puigbò, Bravo and Garcia-Vallve, 2008). Using protein and mRNA datasets of IFN-beta cells 354 (12 h) transfected with BOP1 or control siRNAs at 12 hours, the aligned datasets were then 355 grouped by the CAI highlighted with distinct colours for all proteins (Fig. 4A) and for ISGs (Fig. 356 4C). The CAI values were also visualized in violin plots categorized by their differential 357 expression at the protein or the RNA level (Fig. 4B,D). Amongst all proteins, those that were 358 translationally up-regulated had a significantly lower codon optimality (0.778 vs 0.782, p-value 359 = 0.0003576). A similar trend was observed amongst ISGs, though not statistically significant 360 (0.772 vs 0.796, p-value = 0.1688). In summary, these results showed that proteins that exhibited 361 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 18 greater differential expression at the protein level compared to the transcript level have reduced 362 codon optimality as assessed by CAI. 363 BOP1 knockdown results in an increase of translation of non-optimal transcripts 364 To determine whether codon optimality could dictate translation of an mRNA in BOP1 365 depleted cells, we monitored translation of extreme optimal (opt) and extreme non-optimal (non-366 opt) luciferase reporters adapted from Wu et al (Wu et al., 2019) (Fig. 5A). The optimal or non-367 optimal constructs consisted of a codon-optimized or -non-optimized region adapted from Wu et 368 al, which was separated from a nanoluciferase reporter by a P2A sequence (Wu et al., 2019). The 369 co-transfection construct contained only a Renilla luciferase reporter. All constructs were flanked 370 by the 5’ and 3’ β-globin UTRs. In these experiments, BOP1 or control siRNA treated A549 cells 371 stimulated with mock or IFN-β (2 h) were co-transfected with 5’capped and 3’poly-A tailed 372 reporter mRNA (Fig. 5B). Cells were also co-transfected with an RLuc mRNA to control for 373 transfection efficiency. Cells were harvested and luciferase activity was monitored. Relative 374 NLuc activity was normalized to the RLuc co-transfection control to account for differences in 375 global translation and transfection As expected, untreated cells transfected with the opt-Nluc 376 reporter mRNA resulted in significantly more NLuc activity as compared to the non-opt-Luc 377 reporter, demonstrating that the codon optimality can impact translation and stability of the 378 mRNA (Fig. 5C). The mean fold change across each condition demonstrate a minimal increase in 379 the translation of the optimal reporter under the control knockdown due to the IFN-β stimulation, 380 and a decrease in the translation of the non-optimal reporter under the same conditions (Fig. 5D). 381 Conversely, under the BOP1 knockdown condition, both the optimal and non-optimal reporters 382 demonstrate increases in translation upon IFN-β stimulation. In order to assess the impact of the 383 knockdown, we analyzed the mean fold change in the BOP1 knockdown compared to the control 384 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 19 condition. In the unstimulated condition, there is little difference in the effect of the knockdown 385 between the optimal and non-optimal constructs. In the IFN-β-stimulated condition however, the 386 non-optimal construct exhibits a greater increase in translation due to the knockdown compared 387 to the control (Fig. 5E). These results suggest that the BOP1 knockdown results in significantly 388 greater translation of the non-optimal construct during IFN stimulation. Furthermore, to assess 389 the impact of the IFN stimulation directly, we analyzed the mean fold change in the IFN-β-390 stimulated compared to the unstimulated condition. The effect of the IFN stimulation does not 391 significantly impact the translation of the optimal construct between the control and BOP1 392 knockdown conditions. For the non-optimal construct, IFN stimulation reduced the translation in 393 the control condition; however, IFN stimulation significantly increased the translation in the 394 BOP1 knockdown condition (Fig. 5F). This work suggests that there is a specific and 395 pronounced effect in which translation of non-optimal codons is increased during reductions in 396 ribosome biogenesis, and is exacerbated during IFN-β stimulation. 397 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 20

398 The IFN response is a critical component of innate immunity and pathogen defense and 399 as such requires dynamic gene and protein expression to ensure efficient defense without 400 excessive inflammation and damage (Guillemin et al., 2022). Although this pathway is 401 fundamentally initiated through changes in gene expression, multiple layers of post-402 transcriptional regulation exist to fine-tune the IFN response and include RNA binding proteins, 403 mRNA stability, chemical modifications, and translational control (Burnett et al., 1998; Wu et 404 al., 2019; McFadden et al., 2021; Rothamel et al., 2021; Guillemin et al., 2022). Many of these 405 aspects of fine-tuning remain unclear and prompt further investigation. 406 Ribosome biogenesis is well-understood to play a key role in enabling effective 407 translation; generally, when ribosome biogenesis is perturbed, this results in broad translational 408 defects (Strezoska, Pestov and Lau, 2000; Kumar, 2021). Surprising roles for ribosome 409 biogenesis have recently been identified, such as reductions in ribosome biogenesis increasing 410 HCMV replication, which suggests the potential for nuance in the role of ribosome biogenesis in 411 translational regulation, particularly in the immune response (Bianco and Mohr, 2019). 412 How does ribosome biogenesis impact the translation of the IFN response? The 413 transcriptional landscape of the IFN response has been well-characterized across multiple 414 species, including the identification of conserved ISGs (Mostafavi et al., 2016; Shaw et al., 415 2017). In contrast, there is less data supporting the changes in proteins that occur during IFN 416 stimulation. Recent work investigating the proteome under IFN stimulation identified a 417 translational lag of up to 24 h in the synthesis of ISG proteins, and suggest a potential role for 418 ISGs that are produced earlier (Kerr et al., 2020). Further, a subset of ISGs was identified that 419 were up-regulated upon RPL28 depletion, suggesting a role of the ribosome in negatively 420 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 21 regulating the IFN response (Kerr et al., 2020). We sought to understand whether this effect may 421 be more broadly connected to translation through ribosome biogenesis. 422 Depletion of BOP1, a key ribosome biogenesis factor, (Strezoska, Pestov and Lau, 2000) 423 resulted in a significant up-regulation of ISGs after 12 h of IFN-β stimulation as compared to the 424 knockdown control. This finding suggests that perhaps the observed effects on ISGs are a result 425 of ribosome availability rather than composition. These effects are consistent to those observed 426 during RP depletion with IFN-β stimulation (Kerr et al., 2020). This ISG up-regulation was 427 demonstrated to be specifically occurring at the protein level, as no such effect was observed at 428 the gene level. Although mRNA levels do not always result in equivalent protein levels, this 429 finding pinpoints that the regulation is occurring post-transcriptionally (Schwanhäusser et al., 430 2011). The alignment of these data confirm that ISGs are translationally up-regulated during 431 BOP1 knockdown and IFN stimulation, and furthermore provide a boon of information 432 regarding levels of transcriptional and translational control that may be useful in understanding 433 how ribosome biogenesis and the IFN response lead to changes in said control. We find that there 434 is a small but significant difference in proteins that are translationally up-regulated having lower 435 codon optimality. 436 GO term analysis on proteins that exhibited a greater increase in differential as compared 437 to their transcripts identified several metabolic processes. The increase in these ranging 438 metabolic processes suggest heightened cellular activity. Moreover, an enrichment in 439 ribonucleoprotein complex biogenesis hints at increases in translation, which could relate to the 440 translational shift observed during BOP1 knockdown. There was a decrease in GO terms 441 associated with lipid metabolism, which could be a result of prioritization of metabolic processes 442 during stress such as IFN stimulation. Although IFN stimulation does result in an increase in 443 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 22 fatty acid oxidation and oxidative phosphorylation, these results were found after 24 h of IFN-α 444 (Wu et al., 2016). Temporal changes during IFN stimulation are critical in understanding the 445 dynamic nature of the IFN response and its impacts. This has been established as the changes in 446 lipid metabolism shift over the course of IFN stimulation, where 6 h of IFN-γ stimulation 447 resulted in reductions to triacylglycerol and lipid droplet levels, but 24 h of IFN-γ stimulation 448 had the opposite effect (Truong et al., 2020). Inhibiting fatty acid synthesis has been shown to 449 increase production of type I IFN, which also correlates with these findings (Kanno et al., 2021). 450 We next endeavored to identify how these ISGs may be translationally enhanced during 451 ribosome biogenesis reduction. We identified a correlation between proteins that exhibit greater 452 differential expression at the protein level compared to the transcript level and lower codon 453 optimality. When testing codon optimality directly, our findings showed that non-opt reporter 454 RNAs are increasingly translated during BOP1 knockdown and IFN-β stimulation. This work 455 suggests that there is a specific and pronounced effect in which translation of non-optimal 456 codons is increased during reductions in ribosome biogenesis, and significantly so during IFN-β 457 stimulation. We propose that this effect is a form of translational regulation to fine-tune the 458 immune response during times of stress, although this effect is likely more broadly applied 459 across many critical pathways to ensure effective translation at appropriate times (Fig. 6). Under 460 normal translation conditions, even optimal transcripts will experience some ribosomal stalling 461 and possibly ribosome collisions and thereby may induce the ribosome quality control pathways, 462 though the resulting proteins will be produced at a high level. Under BOP1 knockdown, there are 463 fewer available ribosomes and therefore a concurrent reduction in stalling and collisions as well, 464 resulting in a slight increase in translation of optimal transcripts. In contrast, under normal 465 translation conditions, non-optimal transcripts can lead to greater stalling and ribosome 466 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 23 collisions, which is known to result in mRNA instability and reduced protein production.(Wu et 467 al., 2019) When BOP1 is knocked down, the reduction in ribosomes also reduces the stalling and 468 collisions occurring on non-optimal transcripts, resulting in a translational up-regulation of non-469 optimal transcripts. In other words, when ribosome biogenesis is perturbed, this may result in a 470 translational shift towards non-optimal transcripts, resulting in altered ratios of protein 471 expression. 472 Previously, Wu et al. demonstrated that the effect of codon optimality on transcript 473 stability and protein production is dependent on layers translational regulation that may be 474 occurring in cells, such as the translational shut-off that often occurs during virus infection.(Wu 475 et al., 2019) During HSV-1 infection, they found that genes enriched in optimal codons were 476 down-regulated and genes enriched in non-optimal codons were up-regulated. This work further 477 supports our findings that during translational perturbations such as BOP1 knockdown, 478 translation shifts towards non-optimal transcripts. Additional research is needed to further clarify 479 the extent to which codon optimality impacts mRNA stability as opposed to translation 480 efficiency. 481 These findings may impact the design of mRNA therapeutics, as one must consider the 482 balance of CAI for both RNA stability and codon usage. mRNA therapeutics often employ codon 483 optimization to enhance protein production by facilitating faster translation; however, if lower 484 codon optimality can also promote effective translation, especially in the context of immune 485 stresses such as the IFN response, over-optimization could be counterproductive, emphasizing 486 the importance of tailoring mRNA therapeutics for specific conditions. 487 The experiment determining the effect of IFN-β stimulation on translational differences 488 due to codon optimality was performed by performing a two-hour treatment prior to the mRNA 489 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 24 transfection; however, to more closely replicate the earlier proteomics and transcriptomics 490 experiments, IFN-β could be supplemented during the transfection as well. Further, to expand 491 upon these results and pinpoint a specific mechanism, the effect of ribosome collisions in the 492 observed results could be tested. BOP1 and ZNF598 could be knocked down, treated with IFN-β 493 then co-transfected with mRNA for the extreme optimal or extreme non-optimal NLuc reporters 494 and an RLuc reporter for analysis by luciferase assay. ZNF598 is a ubiquitin ligase which senses 495 ribosome collisions and initiates the ribosome-associated quality control pathway, although a 496 recent study identified that ribosome collision is not a strict requirement for ubiquitination by 497 ZNF598 and that stalling alone can lead to degradation (Juszkiewicz et al., 2018; Miścicka et al., 498 2024). More broadly, to assess global differences in translation and ribosome loading during 499 BOP1 knockdown, Ribo-seq could provide information on ribosome occupancy across a range of 500 transcript codon optimality. 501 These findings provide a comprehensive dataset for the investigation of the effects of 502 BOP1 knockdown during IFN-β stimulation, including establishing a transcriptome and 503 translatome, as well as an alignment of differential expression providing insights on translational 504 regulation. Notably, we observed that ISGs exhibited translational up-regulation during BOP1 505 knockdown and IFN-β stimulation. Upon analysis of the CAI for proteins that were 506 translationally up-regulated and down-regulated, we found that all translationally up-regulated 507 proteins have a small but significant reduction in CAI. ISGs that were translationally up-508 regulated shared this same trend with a greater difference in CAI, although not statistically 509 significant. Functional analysis of the impacts of codon optimality revealed that BOP1 exerts an 510 additional layer of translational control on codon optimality, which we propose is due to the 511 availability of ribosomes impacting ribosome collision and degradation. We propose that this is a 512 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 25 form of translational fine-tuning of the IFN response, as a means for the cell to produce 513 protective proteins during times of cellular stress where translation machinery may be inhibited, 514 but this requires further investigation on the functional mechanisms at play. 515 516

517 The authors thank the Foster and Jan labs for their support, insight, and guidance, and 518 particularly Jaden Chen for their assistance on the codon optimality work. This research was 519 supported by the Natural Sciences and Engineering Research Council, Genome Canada, and 520 Genome BC. 521 522 Competing interests: The authors declare there are no competing interests. 523 524 Authorship 525 BNH, EJ, and LJF conceptualized this work and designed the experiments. BNH, RS, and NL 526 performed experiments. SF performed the codon adaptation index analysis. BNH prepared the 527 first manuscript draft; EJ and LJF edited and prepared the final draft with BNH. 528 529 Data availability statement 530 The mass spectrometry data have been deposited to the ProteomeXchange Consortium via the 531 PRIDE (Perez-Riverol et al., 2022) partner repository with the dataset identifier PXD060209. 532 533 The RNA-seq data have been deposited to the Gene Expression Omnibus (Edgar, 2002) with the 534 dataset identifier GSE281848. 535 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 26 536

537 Bianco, C. and Mohr, I. (2019) ‘Ribosome biogenesis restricts innate immune responses to virus 538 infection and DNA’, eLife, 8, p. e49551. Available at: https://doi.org/10.7554/eLife.49551. 539 Burnett, P.E. et al. (1998) ‘RAFT1 phosphorylation of the translational regulators p70 S6 kinase 540 and 4E-BP1’, Proceedings of the National Academy of Sciences of the United States of America, 541 95(4), pp. 1432–1437. Available at: https://doi.org/10.1073/pnas.95.4.1432. 542 Constantinescu, S. et al. (1995) ‘Expression and signaling specificity of the IFNAR chain of the 543 type I interferon receptor complex’, Proceedings of the National Academy of Sciences of the 544 United States of America, 92(23). Available at: https://doi.org/10.1073/pnas.92.23.10487. 545 Cox, J. and Mann, M. (2008) ‘MaxQuant enables high peptide identification rates, individualized 546 p.p.b.-range mass accuracies and proteome-wide protein quantification’, Nature Biotechnology, 547 26(12), pp. 1367–1372. Available at: https://doi.org/10.1038/nbt.1511. 548 Cui, D. et al. (2014) ‘The ribosomal protein S26 regulates p53 activity in response to DNA 549 damage’, Oncogene, 33(17), pp. 2225–2235. Available at: https://doi.org/10.1038/onc.2013.170. 550 Edgar, R. (2002) ‘Gene Expression Omnibus: NCBI gene expression and hybridization array 551 data repository’, Nucleic Acids Research, 30(1), pp. 207–210. Available at: 552 https://doi.org/10.1093/nar/30.1.207. 553 Ferretti, M.B. et al. (2017) ‘Rps26 directs mRNA-specific translation by recognition of Kozak 554 sequence elements’, Nature Structural & Molecular Biology, 24(9), pp. 700–707. Available at: 555 https://doi.org/10.1038/nsmb.3442. 556 Goyal, A. et al. (2017) ‘Ultra-Fast Next Generation Human Genome Sequencing Data 557 Processing Using DRAGENTM Bio-IT Processor for Precision 558 Medicine’, Open Journal of Genetics, 07(01), pp. 9–19. Available at: 559 https://doi.org/10.4236/ojgen.2017.71002. 560 Gripp, K.W. et al. (2014) ‘Diamond-Blackfan anemia with mandibulofacial dystostosis is 561 heterogeneous, including the novel DBA genes TSR2 and RPS28’, American Journal of Medical 562 Genetics. Part A, 164A(9), pp. 2240–2249. Available at: https://doi.org/10.1002/ajmg.a.36633. 563 Guillemin, A. et al. (2022) ‘Shaping the Innate Immune Response Through Post-Transcriptional 564 Regulation of Gene Expression Mediated by RNA-Binding Proteins’, Frontiers in Immunology, 565 12, p. 796012. Available at: https://doi.org/10.3389/fimmu.2021.796012. 566 Harrison, P.W. et al. (2024) ‘Ensembl 2024’, Nucleic Acids Research, 52(D1), pp. D891–D899. 567 Available at: https://doi.org/10.1093/nar/gkad1049. 568 Hölzel, M. et al. (2005) ‘Mammalian WDR12 is a novel member of the Pes1–Bop1 complex and 569 is required for ribosome biogenesis and cell proliferation’, The Journal of Cell Biology, 170(3), 570 pp. 367–378. Available at: https://doi.org/10.1083/jcb.200501141. 571 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 27 Juszkiewicz, S. et al. (2018) ‘ZNF598 Is a Quality Control Sensor of Collided Ribosomes’, 572 Molecular Cell, 72(3), pp. 469-481.e7. Available at: 573 https://doi.org/10.1016/j.molcel.2018.08.037. 574 Kanno, T. et al. (2021) ‘SCD2-mediated monounsaturated fatty acid metabolism regulates 575 cGAS-STING-dependent type I IFN responses in CD4+ T cells’, Communications Biology, 4(1), 576 pp. 1–14. Available at: https://doi.org/10.1038/s42003-021-02310-y. 577 Kerr, C.H. et al. (2020) ‘Dynamic rewiring of the human interactome by interferon signaling’, 578 Genome Biology, 21(1), p. 140. Available at: https://doi.org/10.1186/s13059-020-02050-y. 579 Kumar, V . (2021) ‘Chapter 7 - Ribosomal biogenesis in eukaryotes’, in V . Kumar (ed.) Emerging 580 Concepts in Ribosome Structure, Biogenesis, and Function. Academic Press, pp. 129–150. 581 Available at: https://doi.org/10.1016/B978-0-12-816364-1.00011-1. 582 Le Bon, A. and Tough, D.F. (2002) ‘Links between innate and adaptive immunity via type I 583 interferon’, Current Opinion in Immunology, 14(4), pp. 432–436. Available at: 584 https://doi.org/10.1016/S0952-7915(02)00354-0. 585 Love, M.I., Huber, W. and Anders, S. (2014) ‘Moderated estimation of fold change and 586 dispersion for RNA-seq data with DESeq2’, Genome Biology, 15(12), p. 550. Available at: 587 https://doi.org/10.1186/s13059-014-0550-8. 588 McFadden, M.J. et al. (2021) ‘Post-transcriptional regulation of antiviral gene expression by N6-589 methyladenosine’, Cell Reports, 34(9), p. 108798. Available at: 590 https://doi.org/10.1016/j.celrep.2021.108798. 591 Melki, I. and Frémond, M.-L. (2020) ‘Type I Interferonopathies: from a Novel Concept to 592 Targeted Therapeutics’, Current Rheumatology Reports, 22(7), p. 32. Available at: 593 https://doi.org/10.1007/s11926-020-00909-4. 594 Mi, H. et al. (2019) ‘Protocol Update for large-scale genome and gene function analysis with the 595 PANTHER classification system (v.14.0)’, Nature Protocols, 14(3), pp. 703–721. Available at: 596 https://doi.org/10.1038/s41596-019-0128-8. 597 Miścicka, A. et al. (2024) ‘Ribosomal collision is not a prerequisite for ZNF598-mediated 598 ribosome ubiquitination and disassembly of ribosomal complexes by ASCC’, Nucleic Acids 599 Research, 52(8), pp. 4627–4643. Available at: https://doi.org/10.1093/nar/gkae087. 600 Mostafavi, S. et al. (2016) ‘Parsing the Interferon Transcriptional Network and Its Disease 601 Associations’, Cell, 164(3), pp. 564–578. Available at: https://doi.org/10.1016/j.cell.2015.12.032. 602 Ohlson, M.B. et al. (2023) ‘Genome-Scale CRISPR Screening Reveals Host Factors Required 603 for Ribosome Formation and Viral Replication’, mBio, 14(2), p. e0012723. Available at: 604 https://doi.org/10.1128/mbio.00127-23. 605 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 28 Perez-Riverol, Y . et al. (2022) ‘The PRIDE database resources in 2022: a hub for mass 606 spectrometry-based proteomics evidences’, Nucleic Acids Research, 50(D1), pp. D543–D552. 607 Available at: https://doi.org/10.1093/nar/gkab1038. 608 Puigbò, P., Bravo, I.G. and Garcia-Vallve, S. (2008) ‘CAIcal: A combined set of tools to assess 609 codon usage adaptation’, Biology Direct, 3(1), p. 38. Available at: https://doi.org/10.1186/1745-610 6150-3-38. 611 Rappsilber, J., Ishihama, Y . and Mann, M. (2003) ‘Stop and Go Extraction Tips for Matrix-612 Assisted Laser Desorption/Ionization, Nanoelectrospray, and LC/MS Sample Pretreatment in 613 Proteomics’, Analytical Chemistry, 75(3), pp. 663–670. Available at: 614 https://doi.org/10.1021/ac026117i. 615 Ritchie, M.E. et al. (2015) ‘limma powers differential expression analyses for RNA-sequencing 616 and microarray studies’, Nucleic Acids Research, 43(7), pp. e47–e47. Available at: 617 https://doi.org/10.1093/nar/gkv007. 618 Robledo, S. et al. (2008) ‘The role of human ribosomal proteins in the maturation of rRNA and 619 ribosome production’, RNA, 14(9), pp. 1918–1929. Available at: 620 https://doi.org/10.1261/rna.1132008. 621 Rothamel, K. et al. (2021) ‘ELA VL1 primarily couples mRNA stability with the 3’ UTRs of 622 interferon-stimulated genes’, Cell Reports, 35(8), p. 109178. Available at: 623 https://doi.org/10.1016/j.celrep.2021.109178. 624 Schindelin, J. et al. (2012) ‘Fiji: an open-source platform for biological-image analysis’, Nature 625 Methods, 9(7), pp. 676–682. Available at: https://doi.org/10.1038/nmeth.2019. 626 Schwanhäusser, B. et al. (2011) ‘Global quantification of mammalian gene expression control’, 627 Nature, 473(7347), pp. 337–342. Available at: https://doi.org/10.1038/nature10098. 628 Shaw, A.E. et al. (2017) ‘Fundamental properties of the mammalian innate immune system 629 revealed by multispecies comparison of type I interferon responses’, PLOS Biology. Edited by H. 630 Malik, 15(12), p. e2004086. Available at: https://doi.org/10.1371/journal.pbio.2004086. 631 Strezoska, Ž., Pestov, D.G. and Lau, L.F. (2000) ‘Bop1 Is a Mouse WD40 Repeat Nucleolar 632 Protein Involved in 28S and 5.8S rRNA Processing and 60S Ribosome Biogenesis’, Molecular 633 and Cellular Biology, 20(15), pp. 5516–5528. 634 Thomas, P.D. et al. (2022) ‘PANTHER: Making genome-scale phylogenetics accessible to all’, 635 Protein Science, 31(1), pp. 8–22. Available at: https://doi.org/10.1002/pro.4218. 636 Truong, N.T.T. et al. (2020) ‘Regulation of lipid metabolism in pancreatic beta cells by interferon 637 gamma: A link to anti-viral function’, Cytokine, 133, p. 155147. Available at: 638 https://doi.org/10.1016/j.cyto.2020.155147. 639 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 29 Wang, W. et al. (2017) ‘Transcriptional Regulation of Antiviral Interferon-Stimulated Genes’, 640 Trends in Microbiology, 25(7), pp. 573–584. Available at: 641 https://doi.org/10.1016/j.tim.2017.01.001. 642 Wei, J. et al. (2019) ‘Ribosomal Proteins Regulate MHC Class I Peptide Generation for 643 Immunosurveillance’, Molecular Cell, 73(6), pp. 1162-1173.e5. Available at: 644 https://doi.org/10.1016/j.molcel.2018.12.020. 645 Wu, D. et al. (2016) ‘Type 1 interferons induce changes in core metabolism that are critical for 646 immune function’, Immunity, 44(6), pp. 1325–1336. Available at: 647 https://doi.org/10.1016/j.immuni.2016.06.006. 648 Wu, Q. et al. (2019) ‘Translation affects mRNA stability in a codon-dependent manner in human 649 cells’, eLife, 8, p. e45396. Available at: https://doi.org/10.7554/eLife.45396. 650 Wu, W. et al. (2015) ‘RIG-I and TLR3 are both required for maximum interferon induction by 651 influenza virus in human lung alveolar epithelial cells’, Virology, 482, pp. 181–188. Available at: 652 https://doi.org/10.1016/j.virol.2015.03.048. 653 654 Figure Captions 655 Figure 1. Targeted ribosomal protein knockdown results in general ribosomal protein 656 depletion and ISG up-regulation during IFN stimulation. A) Percent cell viability of A549 657 cells transfected with RPL28 or control FLuc-siRNAs for the indicated times after transfection. 658 Cell viability was measured by trypan blue staining and an unpaired t-test was performed. Error 659 bars depict standard deviation. N=3. B) RPL28 protein levels as quantified by label-free 660 quantitation (LFQ) intensities during siRNA knockdown on A549 cells as determined by mass 661 spectrometry. T-test p-value = 0.0079. N=3. 662 C) Differential expression of RPs (black) over time. N=3. D) Volcano plot showing differential 663 expression of RPs (black) and the target of the siRNA knockdown, RPL28 (red), in RPL28 664 knockdown and control conditions at all time points. All other proteins are grey. N=3. E) 665 Differential expression of RPs (black) and the target of the siRNA knockdown, RPS26 (red), in 666 RPS26 knockdown and control conditions at all time points. All other proteins are grey. N=2. F) 667 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 30 Differential expression of RPs (black) and the target of the siRNA knockdown, RPS28 (red), in 668 RPS28 knockdown and control conditions at all time points. All other proteins are grey. N=3. G) 669 Differential expression of RP depletion (resulting from RPL28, RPS26, and RPS28 knockdowns) 670 on RPs (black), ribosome biogenesis proteins (blue), ISGs (red), and all other proteins (grey) on 671 IFN-β-stimulated vs unstimulated A549 cells across all time points. Nexpt=8. Ncontrol=3. 672 673 Figure 2. BOP1 knockdown results in ISG up-regulation during IFN stimulation. A) 674 Percent cell viability of A549 cells transfected with BOP1 or control FLuc-siRNAs for the 675 indicated times post-transfection. Cell viability was measured by trypan blue staining and an 676 unpaired t-test was performed. Error bars depict standard deviation. N=3. B) BOP1 counts during 677 siRNA knockdown on A549 cells as determined by RNA-seq. An unpaired t test was performed 678 (p-value = 1.054E-7). N=3. C) Differential protein expression between A549 cells transfected 679 with BOP1 or control FLuc-siRNAs with IFN-β stimulation for 12 h. ISGs are highlighted in red, 680 BOP1 is highlighted in black, and all other proteins are in grey. N=3. D) BOP1 intensity during 681 siRNA knockdown on A549 cells as determined by mass spectrometry. An unpaired t-test was 682 performed (p-value = 1.685E-5). N=3. 683 684 Figure 3. Multi-omics alignment reveals translationally up-regulated proteins. A) Transcript 685 differential expression between A549 cells transfected with siBOP1 or the control siFLuc with 686 IFN-β stimulation for 12 h. ISGs are highlighted in red and all other proteins are in grey. B) 687 Alignment of differential expression in transcripts (x-axis) and proteins (y-axis) for A549 cells 688 stimulated with IFN-β for 12 h during BOP1 knockdown compared to control FLuc knockdown. 689 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 31 ISGs are highlighted in red, all other points are in grey. C) GO term analysis for terms up-690 regulated at the protein level. Signs (+ and -) depict up- or down-regulation respectively. 691 692 Figure 4. Translationally up-regulated proteins have lower codon optimality. A) Alignment 693 of differential expression in transcripts (x-axis) and proteins (y-axis) for A549 cells stimulated 694 with IFN-β for 12 h during BOP1 knockdown compared to control FLuc knockdown. All points 695 are coloured by CAI. Grey dashed line depicts a slope of 1. B) CAI (y-axis) for differential 696 expression results by fold change group (x-axis) including all proteins in dataset. Mean CAI of 697 higher_protein = 0.778. Mean CAI of higher_RNA = 0.782. Unpaired t-test p = 0.0003576. C) 698 Alignment of differential expression in transcripts (x-axis) and proteins (y-axis) for A549 cells 699 stimulated with IFN-β for 12 h during BOP1 knockdown compared to control FLuc knockdown, 700 plotting only ISGs. All points are coloured by CAI. Grey dashed line depicts a slope of 1. D) 701 CAI (y-axis) for differential expression results by fold change group (x-axis) including all ISGs 702 in dataset. Mean higher_protein = 0.772. Mean higher_RNA = 0.796. Unpaired t-test p = 0.1688. 703 704 Figure 5. BOP1 knockdown increases translation of non-optimal transcripts under IFN 705 stimulation. A) Codon optimality constructs adapted from Wu et al. (Wu et al., 2019). Extreme 706 optimal (Opt) or extreme non-optimal (NonOpt) sequences were followed by P2A and 707 nanoluciferase (NLuc), or a control with solely renilla luciferase (RLuc). All constructs were 708 flanked by β-globin 5’ and 3’ UTRs. Created with BioRender. B) 1% RNA gel with 500 ng of 709 each construct after capping and tailing. C) Mean fold change of NLuc and RLuc reporters in 710 control knockdown condition with no IFN stimulation. P-value = 0.009. N=3. D) Mean fold 711 change of NLuc over RLuc (y-axis) across each condition (x-axis). N=3. E) Relative luciferase 712 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint 32 activity normalized for the effect of BOP1 knockdown (y-axis) for each condition (x-axis). P-713 value = 0.01. N=3. F) Relative luciferase activity normalized for the effect of IFN-β stimulation 714 (y-axis) for each condition (x-axis). P-value = 0.01. N=3. 715 716 Figure 6. A proposed model illustrating the effects of BOP1 on codon-optimality-dependent 717 translational regulation. Non-optimal transcripts and protein in red, optimal transcripts and 718 protein in green. Created with BioRender. Under control conditions, BOP1 is present and 719 ribosome biogenesis occurs. The translation of optimal transcripts is more efficient compared to 720 non-optimal transcripts. Under conditions wherein BOP1 and therefore ribosome biogenesis is 721 reduced, the ribosomal landscape shifts. Translation efficiency shifts and non-optimal transcripts 722 are more efficiently translated. 723 724 Supplemental materials 725 Supplementary figure 1. The effect of BOP1 knockdown on the 60S fraction by sucrose 726 gradient fractionation. A) Sucrose gradient fractionation profiles during control (FLuc) and 727 BOP1 knockdowns. N=2. B) Bar chart depicts the reduction in the 60S peak normalized to total 728 surface area in the BOP1 knockdown condition compared to the control. Replicate 1 showed a 729 24% reduction in 60S during BOP1 knockdown and Replicate 2 showed a 14% reduction in 60S 730 during BOP1 knockdown. N=2. 731 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint C D E F G A B (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint C D A B (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint D A B C E (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint A B C D (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint C BA E F Bases D (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 10, 2025. ; https://doi.org/10.1101/2025.09.05.673799doi: bioRxiv preprint