SARS-CoV-2 competes with host mRNAs for efficient translation by maintaining the mutations favorable for translation initiation

SARS-CoV-2 continuously optimizes its codon usage to adapt to human lung environment

Viruses need to utilize the resources from host cells to reproduce themselves. RNA translation rate, which is largely determined by codon usage, is the rate-limiting step across the life cycle of viruses. Adapting to the codon usage of hosts would help virus better proliferate. We retrieved the time-course mutation profile of millions of world-wide SARS-CoV-2 sequences. For synonymous mutations, we defined whether a mutation elevate or reduce the relative synonymous codon usage (RSCU). We found that if a synonymous mutation in SARS-CoV-2 increases the RSCU (calculated from human lungs), denoted as delta RSCU > 0, then this mutation is positively selected because the allele frequency (AF) of this mutation increases with time, and vice versa. The results suggest that in SARS-CoV-2, the synonymous mutations that increase codon optimality are beneficial to the virus and are favored by natural selection. For the first time, we used the dynamics of allele frequency to demonstrate that SARS-CoV-2 is continuously optimizing its codon usage to adapt to human lungs. Nevertheless, adaptation to other human tissues cannot be excluded. These results warn us that under this global pandemic, synonymous mutations in SARS-CoV-2 should not be automatically ignored since they indeed change the fitness of the virus.

Analysis of 3.5 million SARS-CoV-2 sequences reveals unique mutational trends with consistent nucleotide and codon frequencies

BACKGROUND

Since the onset of the SARS-CoV-2 pandemic, bioinformatic analyses have been performed to understand the nucleotide and synonymous codon usage features and mutational patterns of the virus. However, comparatively few have attempted to perform such analyses on a considerably large cohort of viral genomes while organizing the plethora of available sequence data for a month-by-month analysis to observe changes over time. Here, we aimed to perform sequence composition and mutation analysis of SARS-CoV-2, separating sequences by gene, clade, and timepoints, and contrast the mutational profile of SARS-CoV-2 to other comparable RNA viruses.

METHODS

Using a cleaned, filtered, and pre-aligned dataset of over 3.5 million sequences downloaded from the GISAID database, we computed nucleotide and codon usage statistics, including calculation of relative synonymous codon usage values. We then calculated codon adaptation index (CAI) changes and a nonsynonymous/synonymous mutation ratio (dN/dS) over time for our dataset. Finally, we compiled information on the types of mutations occurring for SARS-CoV-2 and other comparable RNA viruses, and generated heatmaps showing codon and nucleotide composition at high entropy positions along the Spike sequence.

RESULTS

We show that nucleotide and codon usage metrics remain relatively consistent over the 32-month span, though there are significant differences between clades within each gene at various timepoints. CAI and dN/dS values vary substantially between different timepoints and different genes, with Spike gene on average showing both the highest CAI and dN/dS values. Mutational analysis showed that SARS-CoV-2 Spike has a higher proportion of nonsynonymous mutations than analogous genes in other RNA viruses, with nonsynonymous mutations outnumbering synonymous ones by up to 20:1. However, at several specific positions, synonymous mutations were overwhelmingly predominant.

CONCLUSIONS

Our multifaceted analysis covering both the composition and mutation signature of SARS-CoV-2 gives valuable insight into the nucleotide frequency and codon usage heterogeneity of SARS-CoV-2 over time, and its unique mutational profile compared to other RNA viruses.

Evidence Supporting That C-to-U RNA Editing Is the Major Force That Drives SARS-CoV-2 Evolution

Mutations of DNA organisms are introduced by replication errors. However, SARS-CoV-2, as an RNA virus, is additionally subjected to rampant RNA editing by hosts. Both resources contributed to SARS-CoV-2 mutation and evolution, but the relative prevalence of the two origins is unknown. We performed comparative genomic analyses at intra-species (world-wide SARS-CoV-2 strains) and inter-species (SARS-CoV-2 and RaTG13 divergence) levels. We made prior predictions of the proportion of each mutation type (nucleotide substitution) under different scenarios and compared the observed versus the expected. C-to-T alteration, representing C-to-U editing, is far more abundant that all other mutation types. Derived allele frequency (DAF) as well as novel mutation rate of C-to-T are the highest in SARS-CoV-2 population, and C-T substitution dominates the divergence sites between SARS-CoV-2 and RaTG13. This is compelling evidence suggesting that C-to-U RNA editing is the major source of SARS-CoV-2 mutation. While replication errors serve as a baseline of novel mutation rate, the C-to-U editing has elevated the mutation rate for orders of magnitudes and accelerates the evolution of the virus.

C-to-U RNA deamination is the driving force accelerating SARS-CoV-2 evolution

Understanding the molecular mechanism underlying the rampant mutation of SARS-CoV-2 would help us control the COVID-19 pandemic. The APOBEC-mediated C-to-U deamination is a major mutation type in the SARS-CoV-2 genome. However, it is unclear whether the novel mutation rate u is higher for C-to-U than for other mutation types, and what the detailed driving force is. By analyzing the time course SARS-CoV-2 global population data, we found that C-to-U has the highest novel mutation rate u among all mutation types and that this u is still increasing with time (du/dt > 0). Novel C-to-U events, rather than other mutation types, have a preference over particular genomic regions. A less local RNA structure is correlated with a high novel C-to-U mutation rate. A cascade model nicely explains the du/dt > 0 for C-to-U deamination. In SARS-CoV-2, the RNA structure serves as the molecular basis of the extremely high and continuously accelerating C-to-U deamination rate. This mechanism is the driving force of the mutation, adaptation, and evolution of SARS-CoV-2. Our findings help us understand the dynamic evolution of the virus mutation rate.

Hepatitis B virus (HBV) codon adapts well to the gene expression profile of liver cancer: an evolutionary explanation for HBV's oncogenic role

Due to the evolutionary arms race between hosts and viruses, viruses must adapt to host translation systems to rapidly synthesize viral proteins. Highly expressed genes in hosts have a codon bias related to tRNA abundance, the primary RNA translation rate determinant. We calculated the relative synonymous codon usage (RSCU) of three hepatitis viruses (HAV, HBV, and HCV), SARS-CoV-2, 30 human tissues, and hepatocellular carcinoma (HCC). After comparing RSCU between viruses and human tissues, we calculated the codon adaptation index (CAI) of viral and human genes. HBV and HCV showed the highest correlations with HCC and the normal liver, while SARS-CoV-2 had the strongest association with lungs. In addition, based on HCC RSCU, the CAI of HBV and HCV genes was the highest. HBV and HCV preferentially adapt to the tRNA pool in HCC, facilitating viral RNA translation. After an initial trigger, rapid HBV/HCV translation and replication may change normal liver cells into HCC cells. Our findings reveal a novel perspective on virus-mediated oncogenesis.

The Sponge Interaction Between Circular RNA and microRNA Serves as a Fast-Evolving Mechanism That Suppresses Non-small Cell Lung Cancer (NSCLC) in Humans

Non-small cell lung cancer (NSCLC) is one of the most lethal cancer types in the world. Currently, the molecular mechanisms and pathways underlying NSCLC oncogenesis are poorly understood. Using multiple Omics data, we systematically explored the differentially expressed circular RNAs (circRNAs) in NSCLC. We also investigated potential microRNA sponges (that absorb circRNAs) in NSCLC and downstream target genes with experimental verifications. hsa_circ_0003497 was down-regulated in NSCLC and played an inhibitory role in tumorigenesis. In contrast, miR-197-3p was up-regulated in NSCLC. hsa_circ_0003497 directly interacts with miR-197-3p and releases a target gene of miR-197-3p termed CTNND1 (a known tumor suppressor gene). Evolutionary analysis reveals fast evolution of this hsa_circ_0003497-miR-197-3p-CTNND1-NSCLC axis in mammals. This work clarified the biological functions and molecular mechanisms of how hsa_circ_0003497 suppresses NSCLC through miR-197-3p and CTNND1. We discovered molecular markers for the prognosis of NSCLC and provided potential intervention targets for its treatment.

uORF-Mediated Translational Regulation of ATF4 Serves as an Evolutionarily Conserved Mechanism Contributing to Non-Small-Cell Lung Cancer (NSCLC) and Stress Response

Diseases and environmental stresses are two distinct challenges for virtually all living organisms. In light of evolution, cellular responses to diseases and stresses might share similar molecular mechanisms, but the detailed regulation pathway is not reported yet.

We obtained the transcriptomes and translatomes from several NSCLC (non-small-cell lung cancer) patients as well as from different species under normal or stress conditions. We found that the translation level of gene ATF4 is remarkably enhanced in NSCLC due to the reduced number of ribosomes binding to its upstream open reading frames (uORFs). We also showed the evolutionary conservation of this uORF-ATF4 regulation in the stress response of other species. Molecular experiments showed that knockdown of ATF4 reduced the cell growth rate while overexpression of ATF4 enhanced cell growth, especially for the ATF4 allele with mutated uORFs. Population genetics analyses in multiple species verified that the mutations that abolish uATGs (start codon of uORFs) are highly deleterious, suggesting the functional importance of uORFs.

Our study proposes an evolutionarily conserved pattern that enhances the ATF4 translation by uORFs upon stress or disease. We generalized the concept of cellular response to diseases and stresses. These two biological processes may share similar molecular mechanisms.

Rampant C-to-U deamination accounts for the intrinsically high mutation rate in SARS-CoV-2 spike gene

The high mutation rate of SARS-CoV-2 largely complicates our control of the pandemic. In particular, it is currently unclear why the spike (S) gene has an extraordinarily high mutation rate among all SARS-CoV-2 genes. By analyzing the occurrence of fixed synonymous mutations between SARS-CoV-2 and RaTG13, and profiling the DAF (derived allele frequency) of polymorphic synonymous sites among millions of worldwide SARS-CoV-2 strains, we found that both fixed and polymorphic mutations show higher mutation rates in the S gene than other genes. The majority of mutations are C-to-T, representing the APOBEC-mediated C-to-U deamination instead of the previously proposed A-to-I deamination. Both in silico and in vivo evidence indicated that the S gene is more likely to be single-stranded compared to other SARS-CoV-2 genes, agreeing with the APOBEC preference of ssRNA. We conclude that the single-stranded property of the S gene makes it a favorable target for C-to-U deamination, leading to its excessively high mutation rate compared to other non-S genes. In conclusion, APOBEC, rather than ADAR, is the "editor-in-chief" of SARS-CoV-2 RNAs. This work helps us to understand the molecular mechanism underlying the mutation and evolution of SARS-CoV-2, and we believe it will contribute to the control of the pandemic.

Translational Control of COVID-19 and Its Therapeutic Implication

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is the causative agent of COVID-19, which has broken out worldwide for more than two years. However, due to limited treatment, new cases of infection are still rising. Therefore, there is an urgent need to understand the basic molecular biology of SARS-CoV-2 to control this virus. SARS-CoV-2 replication and spread depend on the recruitment of host ribosomes to translate viral messenger RNA (mRNA). To ensure the translation of their own mRNAs, the SARS-CoV-2 has developed multiple strategies to globally inhibit the translation of host mRNAs and block the cellular innate immune response. This review provides a comprehensive picture of recent advancements in our understanding of the molecular basis and complexity of SARS-CoV-2 protein translation. Specifically, we summarize how this viral infection inhibits host mRNA translation to better utilize translation elements for translation of its own mRNA. Finally, we discuss the potential of translational components as targets for therapeutic interventions.

Evidence for selection on SARS-CoV-2 RNA translation revealed by the evolutionary dynamics of mutations in UTRs and CDSs

RNA translation is the rate-limiting step when cells synthesize proteins. Elevating translation efficiency (TE) is intuitively beneficial. Particularly, when viruses invade host cells, how to compete with endogenous RNAs for efficient translation is a major issue to be resolved. We collected millions of worldwide SARS-CoV-2 sequences during the past year and traced the dynamics of allele frequency of every mutation. We defined adaptive and deleterious mutations according to the rise and fall of their frequencies along time. For 5ʹUTR and synonymous mutations in SARS-CoV-2, the selection on TE is evident near start codons. Adaptive mutations generally decrease GC content while deleterious mutations increase GC content. This trend fades away with increasing distance to start codons. Mutations decreasing GC content near start codons would unravel the complex RNA structure and facilitate translation initiation, which are beneficial to SARS-CoV-2, and vice versa. During this evolutionary arms race between human and virus, SARS-CoV-2 tries to improve its cis elements to compete with host RNAs for rapid translation.