key: cord-102898-eyyd7ent
authors: Rizvi, Vaseef A.; Sarkar, Maharnob; Roy, Rahul
title: Translation regulation of Japanese encephalitis virus revealed by ribosome profiling
date: 2020-07-17
journal: bioRxiv
DOI: 10.1101/2020.07.16.206920
sha: 
doc_id: 102898
cord_uid: eyyd7ent

Japanese encephalitis virus (JEV), a neurotropic flavivirus, is the leading cause of viral encephalitis in endemic regions of Asia. Although the mechanisms modulating JEV virulence and neuroinvasiveness are poorly understood, several acquired mutations in the live attenuated vaccine strain (SA14-14-2) point towards translation regulation as a key strategy. Using ribosome profiling, we identify multiple mechanisms including frameshifting, tRNA dysregulation and alternate translation initiation sites that regulate viral protein synthesis. A significant fraction (~ 40%) of ribosomes undergo frameshifting on NS1 coding sequence leading to early termination, translation of NS1′ protein and modulation of viral protein stoichiometry. Separately, a tRNA subset (glutamate, serine, leucine and histidine) was found to be associated in high levels with the ribosomes upon JEV infection. We also report a previously uncharacterised translational initiation event from an upstream UUG initiation codon in JEV 5′ UTR. A silent mutation at this start site in the vaccine strain has been shown to abrogate neuroinvasiveness suggesting the potential role of translation from this region. Together, our study sheds light on distinct mechanisms that modulate JEV translation with likely consequences for viral pathogenesis.

downstream polyprotein, 2) Significant modulation in levels of a distinct subset of ribosome-bound tRNAs 48 that cannot be explained by virtue of codon usage and 3) Translation from an upstream ORF (uORF) using 49 a non-canonical initiation codon in the 5 UTR region of JEV. These events signify several strategies of 50 translational regulation during viral polyprotein synthesis along with features which could aid the virus in 51 neuroinvasion. Overall, our findings display the potential of translation governing factors in +ssRNA viruses' 52 pathogenesis by evaluating their molecular underpinnings. At 18h post infection (pi), cells were treated with either cycloheximide (Sigma-Aldrich, 100µg/ml, 5min) 63 or harringtonine (LKT laboratories, 2µg/ml, 2min) followed by cycloheximide treatment for 5min. Cells 64 were rinsed with ice-cold PBS containing 100µg/ml cycloheximide. Dishes were submerged in a liquid 65 nitrogen reservoir for 10s followed by scraping over dry ice in polysome lysis buffer (20mM Tris-HCl pH 66 7.5, 150mM NaCl, 5mM MgCl 2 , 1mM DTT, 1% (v/v) Triton X-100, 100µg/ml cycloheximide and 25U/ml 67 TURBO DNase I (Life Technologies)). Cells were collected and triturated with a 26G needle 20 times and 68 clarified by centrifugation (13000g, 20min, 4 • C). Supernatant was collected and treated with 100U/µl of 69 RNase I (Ambion) for 1h at room temperature with gentle mixing followed by inactivation with 40U of 70 SUPERase-In RNase inhibitor (Ambion). Cell extracts were passed through Sephacryl S400 spin columns 71 (GE) after pre-equilibration with polysome lysis buffer and ribosome-bound mRNA eluates were collected by 72 centrifugation at 600g, 1min, 4 • C [16] . 73 

Library preparation 74 RNA was extracted from eluates and total lysate using TRIzol reagent (ThermoScientific). For RNA-Seq, 75 total RNA was fragmented using 10x fragmentation reagent (Ambion) according to manufacturer's protocol. 76 Libraries were prepared according to original protocols of Ingolia and colleagues [17] with minor modifications.

Briefly, RNA was resolved over 15% polyacrylamide TBE-urea gel using electrophoresis and a broad range of 78 fragments (25 -70 nts) were purified from the gel to capture both ribosome bound mRNA and tRNA segments.

RNA fragments were further dephosphorylated using T4 polynucleotide kinase (NEB) for 1h at 37 • C followed 80 by heat inactivation for 10min at 75 • C. Fragments were ligated to microRNA preadenylated linker (NEB) 81 using T4 Rnl2(tr) ligase (NEB) for 2.5h at room temperature. Ligated products were size selected and 82 purified from PAGE gel followed by reverse transcription [17] using SuperScript III (ThermoScientific) and 83 eliminating RNA by NaOH hydrolysis for 20min at 98 • C. cDNA is again size selected on a denaturing PAGE, 84 gel purified and circularised using CircLigase (Epicentre) for 1h at 60 • C followed by heat inactivation for 10min 85 at 80 • C. Circularised product is subjected to two rounds of rRNA depletion using biotinylated oligos [17] 86 and streptavidin-coated magnetic beads (NEB) according to manufacturer's protocol. The rRNA-subtracted 87 circular product is subjected to a final round of PCR amplification with Illumina adaptor primers [17] 88 using Phusion polymerase (NEB). All the libraries were then quantified, and quality checked by Genotypic 89 technology services using qubit fluorimeter, real time qRT-PCR and bioanalyzer followed by sequencing on Raw sequences were quality filtered and adaptors were trimmed using FASTX-Toolkit [18] . Reads ≥ 24 nt 95 were mapped to Mus musculus rRNA (Accession numbers NR003278, NR030686, NR003279, NR003280, 96 NR046233, GU372691) and tRNA (gtRNA database) using Bowtie2 (very-sensitive-local alignment) [19] with 97 a maximum of one mismatch and sorted into separate files. Unmapped reads were aligned to JEV strain 98 P20778 (AF 080251) and Mus musculus Refseq RNA database (mm10) without any mismatches for analyzing 99 ribosomal footprint distribution. P-site offsets were determined by metagene analysis of host mRNA using 100 ribogalaxy tool [20] for corresponding footprint lengths. Reads aligning to JEV were further mapped to single 101 nucleotide by setting respective P-site offset (+12 for 28 − 30nts and +13 for 31nt). Individual fragment not accounted for RNA levels [15] . tRNA mapping was executed using sports1.0 [21] with no mismatches 112 and default parameters for cytoplasmic and mitochondrial tRNAs. As mitochondria employ an independent 113 translation system and serves as internal control for relative quantification [22] , individual cytoplasmic tRNAs 114 were first normalised to total mitochondrial tRNA levels and further quantile normalised to derive relative 115 fold changes between the samples. Sequence comparison of 5 UTR from various JEV strains was carried out 116 using kalign with default parameters [23] . Statistical and correlation analyses were performed using in-house 117 written scripts.

A dual reporter vector, pCMV-sLuc-IRES-GFP [24] , was employed to validate expression activity from at NS1 C-terminus, with the exception of NS2B possibly due to low ribosomal velocities (Fig.2B) . This frameshift region compared to frame-wise read densities immediately upstream and downstream is consistent 195 with −1 PRF near the NS1 terminus (Fig.2C) . With similar estimates reported from WNV (30 -50%), this 196 frameshifting can result in higher levels of structural proteins and will lead to deviations suggested in viral 197 polyprotein stoichiometry [4] . However, considering the involvement of vRNA in parallel lifecycle processes, 198 accurate estimates of T.E for individual viral proteins remains to be evaluated. Since a conserved RNA 199 pseudoknot structure is shown to stimulate −1 PRF on a slippery heptanucleotide motif and generate NS1 200 in JEV [3] , we also scan for frameshift-associated pausing near the frameshit site. Although no ribopausing 201 was observed at the slippery heptanucleotide sequence, we detect significant accumulation of RPFs ∼ 100 202 nt upstream to the frameshift site (3446 nt, Fig.2A inset) . This could either be a consequence of higher 203 representation of charged and proline residues near this region or limited nuclease accessibility due to closely 204 stacked ribosomes upstream to frameshift site. In addition, we also observe significant pause sites in NS3 205 (6318 nt, Asn) and membrane CDS (892 nt, Ala). However, these sites do not represent commonly associated Studies on RNA viruses have suggested adaptations in codon usage of viral genes to the host translation [30] .

This coupling is modulated by tRNA concentrations of specific codons and regulates ribosome elongation 211 dynamics as well as co-translational folding of proteins [28, 31] . In order to understand the differential (Fig.3A) . As tRNAs contain modified bases which generate truncated cDNAs during reverse transcription step 216 of library preparation [32, 33] , we perform stringent alignments to quantify ribosome-protected tRNAs [21] 217 (see methods). The tRNA levels agree well between CHX and HAR-CHX treated infected lysates (R 2 = 0.97, 218 Fig.3B ) suggesting that our results are independent of procedure or protocols used for ribosome associated 219 tRNAs [25, 33] . A subset of tRNAs also represented at higher abundance in ribosome associated fraction upon 220 JEV infection (Figs.3B and 3C) . These include glutamic acid (UUC), serine (CGA, GCU, UGA), histidine 221 (GUG) and leucine (CAG) tRNAs (Fig.3C) . Such high levels of tRN A U U C Glu could signify a strong enrichment 222 of GAA codon which is reported to be associated with ribopause sites in mammalian translatomes [29] . In tRNA profiles generated using targeted tRNA library preparations [25, 33] . Since mice infected with JEV 227 and WNV show upregulation of certain aminoacyl tRNA synthetases [34] , it is possible that translation start codon (Fig.4C) . A recent study on riboprofile of ZIKV also showed ribosomal initiation scanning from 244 out-of-frame non-canonical start codons present in the 5 UTR -CUG (uORF1) and UUG (uORF2) [13] .

However, unlike ZIKV riboprofile where initiating ribosomes tend to stall more at uORF start codons 246 compared to AUG, riboprofile of JEV (CHX) in N2a cells shows that the RPF density is ∼ 3.2x higher at the 247 canonical start codon compared to the uORF start site (Figs.4A and 2B ). Translation initiation from both JEV 248 start codons was confirmed by luciferase-based translation reporter constructs (Fig.4B) . Consistent with our 249 ribosome profile findings, polyprotein ORF (ppORF) expresses 2.7 − 4x more efficiently than uORF (Fig.4C) , 250 10/19 as expected from poor initiation context of UUG [29] . Interestingly, JEV infection appears to stimulate 251 expression from UUG start site by almost 67% suggesting viral or virus-induced host trans-regulatory factors 252 promoting uORF translation (Fig.4D) . Upon sequence comparison of 5 UTR across JEV strains, the alternate 253 start codon exhibits 100% conservation except for the attenuated strain-JEV SA14-14-2 harbouring U39A 254 mutation and generating a previously overlooked stop codon (Fig.4E) . Implications of this disrupted alternate 255 start site in the vaccine strain remains to be evaluated. are distinct from those reported in mammalian systems [29] . We also identify a subset of ribosome associated 286 tRNAs whose levels are modulated globally upon JEV infection (Fig.3) . However, such tRNA abundances 287 remain to be critically examined as recent studies indicate interference of cycloheximide in quantifying bound 288 tRNA fractions due to ribosomal conformational locks [31] . Nevertheless, this discriminating tRNA subset 289 during JEV infection could provide a possible antiviral intervention strategy. For example, a recent study 290 demonstrated impairment in WNV infectivity upon depletion of schlafen-11 which prevents WNV-induced 291 changes in a tRNA subset translating 11.8% of viral polyprotein [37] . Also, schlafen-11 was shown to bind 292 tRNAs essential for HIV protein synthesis during later stages of infection in a codon usage dependent 293 manner [38] . It is additionally possible that these tRNA might be involved in translation-independent 294 processes like viral replication (eg. retro-and bromoviruses [39] ) or novel functions with their site-specific 295 interactions on viral genome (eg. tRN A CT C Glu and tRN A CCC Gly with ZIKV [40] ). Although studies have indicated 296 adaptation in viral codon usage to host organism or tissue [41, 42] , our findings suggest its re-evaluation 297 in the context of the anticodon counterpart (tRNA) dynamics that will arise due to virus-induced cellular 298 12/19 rearrangements. Manipulating virus codon usage to attenuate the virus has recently shown promising 299 advancements in vaccination of mice models [43] [44] [45] [46] [47] . Combined with tRNA perturbation, efficacy of these 300 putative attenuated vaccine targets can be further improved.

Despite the employment of translation inhibitors to enrich ribosome associated fragments, both UTRs 302 display unique nuclease protected regions. While reads in 3 UTR fail to exhibit phasing, RCS2 region of 5 303 DB displays significant nuclease protection. This region, arising from duplication of CS2 region in 3 DB, is 304 observed across JEV and DENV serotypes and mutations in CS2 region have also shown to decrease translation 305 and replication in the latter [48] . Despite the absence of a polyA tail, it was shown that A-rich sequences 306 flanking these DB structures interact with polyA binding protein (PABP) and enhance translation by serving 307 as a circularisation factor [49] . We speculate that CS2 and RCS2 duplicate sequences in 3 UTR facilitate a 308 similar interaction and possibly modulate vRNA translation by circularisation and ribosome recycling. We 309 also observe a striking nuclease protection profile in 5 UTR upon comparing CHX and HAR-CHX datasets 310 which exhibit characteristics of an uORF (Fig.4A ). Translational activity of the uORF was validated using 311 reporter constructs (Fig.4B ). While the functional significance of the putative peptide remains unknown, a 312 close overlap of uORF stop codon (100 th nt) with main ORF's start codon (96 th nt, Fig.4C ) could directly 313 impact ribosome loading and usage rates-key parameters in regulating protein synthesis [50] . Although a 314 canonical structure of 5 UTR is shared amongst flaviviruses [51] , this strategy of translation initiation might 315 suggest the enhanced translational efficiency of JEV and ZIKV [13] over DENV [9] . This site exhibits high 316 conservation across JEV strains with the exception of vaccine strain (SA14-14-2) harbouring a silent mutation.

A reverse genetic approach incorporating U39A mutation in wild type JEV SA14 backbone reduced mice 318 neuroinvasiveness but not neurovirulence of the virus with no significant differences in viral titres under 319 in vitro conditions [5] . However, it remains to be tested whether this intervention is a consequence of the 320 putative short peptide encoded from the uORF or a cis acting element important for RNA transport as 321 suggested in tick borne encephalitis virus [52] . A similar strategy of uORF and polyprotein expression is 322 employed by enteroviruses where the uORF encoded protein was shown to facilitate virus growth in gut 323 epithelial cells [14] . This remarkable display of tropism using short uORFs, along with other cis acting 

Estimated global incidence of Japanese encephalitis: a systematic review

The Viruses and Their Replication. Fields Virology

NS1 of Flaviviruses in the Japanese Encephalitis Virus Serogroup Is a Product of Ribosomal Frameshifting and Plays a Role in Viral Neuroinvasiveness

Non-canonical translation in RNA viruses

Genetic Determinants of Japanese Encephalitis Virus Vaccine Strain SA14-14-2 That Govern Attenuation of Virulence in Mice

Comparison of the live-attenuated japanese encephalitis vaccine SA14-14-2 strain with its pre-attenuated virulent parent SA14 strain: Similarities and differences in vitro and in vivo

A single nucleotide mutation in NS2A of Japanese encephalitis-live vaccine virus (SA14-14-2) ablates NS1 formation and contributes to attenuation

Genomewide analysis in vivo of translation with nucleotide resolution using ribosome profiling

Dengue Virus Selectively Annexes Endoplasmic Reticulum-Associated Translation Machinery as a Strategy for Co-opting Host Cell Protein Synthesis

Cellular gene expression during hepatitis c virus replication as revealed by ribosome profiling

High-resolution analysis of coronavirus gene expression by rna sequencing and ribosome profiling

Tomer Israely, Nir Paran, Michal Schwartz, and Noam Stern-Ginossar. The coding capacity of SARS-CoV-2

The translational landscape of Zika virus during infection of mammalian and insect cells

An upstream protein-coding region in enteroviruses modulates virus infection in gut epithelial cells

Protein-directed ribosomal frameshifting temporally regulates gene expression

Isolate and sequence ribosome-protected mRNA fragments using size-exclusion chromatography

The ribosome profiling strategy for monitoring translation in vivo by deep sequencing of ribosome-protected mRNA fragments

Fast gapped-read alignment with Bowtie 2

RiboGalaxy: A browser based platform for the alignment, analysis and visualization of ribosome profiling data

SPORTS1.0: A Tool for Annotating and Profiling Non-coding RNAs Optimized for rRNA-and tRNA-derived Small RNAs

Tracking the Missing Footprints of Idle Ribosomes

Kalign -An accurate and fast multiple sequence alignment algorithm

Functional incompatibility between the generic nf-kb motif and a subtype-specific sp1iii element drives the formation of the hiv-1 subtype c viral promoter

Genome-wide Translation Profiling by Ribosome-Bound tRNA Capture

Small noncoding RNA modulates Japanese encephalitis virus replication and translation in trans

Integrated in vivo and in vitro nascent chain profiling reveals widespread translational pausing

Ribosome pausing, a dangerous necessity for co-translational events

Ribosome profiling of mouse embryonic stem cells reveals the complexity and dynamics of mammalian proteomes

Repertoires of tRNAs: The Couplers of Genomics and Proteomics

Codon optimality, bias and usage in translation and mRNA decay

Characterizing Expression and Processing of Precursor and Mature Human tRNAs by Hydro-tRNAseq and PAR-CLIP

Landscapes of Ribosome Dwell Times and Relationships with Aminoacyl-tRNA Levels in Mammals

Virus-induced transcriptional changes in the brain include the differential expression of genes associated with interferon, apoptosis, interleukin 17 receptor A, and glutamate signaling as well as flavivirus-specific upregulation of tRNA synthetases

Membranous replication factories induced by plus-strand RNA viruses

Quantitative Analysis of the Hepatitis C Virus Replication Complex

Schlafen 11 Restricts Flavivirus Replication

Codon-usage-based inhibition of HIV protein synthesis by human schlafen 11

Parallels among positive-strand RNA viruses, reverse-transcribing viruses and doublestranded RNA viruses

COMRADES determines in vivo RNA structures and interactions

The extent of codon usage bias in human RNA viruses and its evolutionary origin

Codon usage bias and the evolution of influenza A viruses. Codon Usage Biases of Influenza Virus

Modulation of Poliovirus Replicative Fitness in HeLa Cells by Deoptimization of Synonymous Codon Usage in the Capsid Region

Virus attenuation by genome-scale changes in codon pair bias

Reduction of the Rate of Poliovirus Protein Synthesis through Large-Scale Codon Deoptimization Causes Attenuation of Viral Virulence by Lowering Specific Infectivity

Live attenuated influenza virus vaccines by computer-aided rational design

Deliberate reduction of hemagglutinin and neuraminidase expression of influenza virus leads to an ultraprotective live vaccine in mice

Identification of cis-acting elements in the 3 -untranslated region of the dengue virus type 2 RNA that modulate translation and replication

Poly(A)-binding protein binds to the nonpolyadenylated 3 untranslated region of dengue virus and modulates translation efficiency

The key parameters that govern translation efficiency

Viral RNA switch mediates the dynamic control of flavivirus replicase recruitment by genome cyclization. eLife

Dendritic transport of tick-borne flavivirus rna by neuronal granules affects development of neurological disease

The authors thank Prof. P.N. Rangarajan (BC, IISc) for providing JEV P20778 virus and Neuro2a cells;