If you’ve been spending any time around virologists (or, really, anywhere) over these last few pandemic-stricken years, you may be sick of hearing about a particular viral protein called spike. For the uninitiated, spike is a cell-surface glycoprotein encoded by the SARS-CoV-2 virus which has gained infamy for its crucial molecular role: spike is the protein which SARS-CoV-2 uses to bind to and enter your cells, it’s a major target recognized by your immune system to fight the virus, and it’s also a hotspot for immune-escaping mutations (in fact, you probably first met spike while reading about a novel and alarming SARS-CoV-2 variant of concern). While spike is undoubtedly an important molecular feature of SARS-CoV-2, it’s not the only tool in the virus’ toolbox—the SARS-CoV-2 genome encodes at least ten other genes (known as Open Reading Frames, or ORFs) which are all under varying degrees of selective pressure as the virus evolves and spreads. A new study from the Bedford lab in the Vaccine and Infectious Disease Division at Fred Hutch, led by MD-PhD student Cassia Wagner and recently published in Nature Communications, tells a story about an enigmatic viral gene which you probably haven’t heard of, called ORF8.
While ORF8 lives in the shadow of its more well-known siblings in the SARS-CoV-2 genome, this didn’t stop researchers worldwide from noticing an interesting trend early in the pandemic: as SARS-CoV-2 evolved in its new human hosts, so did ORF8. And unlike spike—which predominantly accumulates missense mutations altering specific amino acid residues in the protein—ORF8 appeared to additionally accumulate large deletions and nonsense mutations (which introduce premature stop codons) which altogether cause functional loss of this viral protein. This apparent selection against ORF8 is so pervasive that as of late 2023, it’s been estimated that over 90% of SARS-CoV-2 in circulation lacks a functional ORF8! So, what gives? Is losing ORF8 function somehow beneficial for SARS-CoV-2 pathogenesis or transmission? As it turns out, this evidence alone is not enough to conclusively support this notion—it could be, for example, that ORF8 mutations co-occur with (or ‘piggyback’ on) other fitness-enhancing viral mutations, explaining their rise in frequency over time. In their recent paper, Wagner and colleagues capitalized on the extensive SARS-CoV-2 surveillance and sequencing efforts in Washington State and used phylogenetic approaches to weigh in on these competing hypotheses.
Querying SARS-CoV-2 sequencing data collected in Washington from the beginning of the pandemic to March 2023, Wagner and team tabulated the number and type of mutations in every SARS-CoV-2 gene. Consistent with previous observations, this analysis revealed an outsized number of premature stop codons (nonsense mutations) in ORF8—in fact, ORF8 had more nonsense mutations than any other SARS-CoV-2 gene in their dataset. To put these mutations in an evolutionary context, the team mapped them onto a phylogenetic tree constructed from the SARS-CoV-2 genome sequences; here, ORF8 nonsense mutations tended to appear in ‘clusters’ of closely related viral variants (see figure below). These clusters varied in size—notably, some viral clades descendent from the well-known Alpha and Omicron variants showed near-universal ORF8 knockout—but they were significantly larger than clusters containing knockouts of any other viral gene. Importantly, the team found that these observations hold true even when considering intrahost SARS-CoV-2 evolution (that is, the competition between different viral variants during an infection of a single host). Altogether, these results are consistent with altered selection of ORF8 during SARS-CoV-2 evolution.
So, ORF8 shows evidence of selection distinct from other genes during SARS-CoV-2 evolution—but is loss of ORF8 actually favored? Put another way, is loss of ORF8 actually under positive selection during viral evolution? To get at this, Wagner and colleagues used publicly-available phylogenetic data to calculate a metric called dN/dS for each of the SARS-CoV-2 genes. In a nutshell, dN/dS compares the rate of nonsynonymous substitutions (mutations which affect the amino acid sequence) to the rate of synonymous substitutions (mutations that don’t change the amino acid sequence but have an equal probability of occurring as nonsynonymous mutations) in a gene under evolution. A nonsynonymous substitution rate which is lower than the synonymous substitution rate (that is, dN/dS < 1) is taken to indicate negative selection, while the converse (dN/dS > 1) indicates positive selection. Focusing on nonsense mutations, Wagner and colleagues found that viral genes like spike and nucleocapsid (which are essential to viral function and therefore predicted to strongly select against nonsense mutations) indeed showed dN/dS values much lower than 1. It was quite a different story for ORF8, however, which showed a dN/dS value of just over 1. To substantiate these results, the team employed several different statistical modeling methods which led them to conclude that variant clusters with nonsense ORF8 mutations grow roughly five times faster than clusters with synonymous ORF8 mutations (and that this wasn’t the case for other viral genes).
Overall, these results strongly suggest that ORF8 loss is under positive selection during SARS-CoV-2 evolution in humans. Why might this information be important to know? For one, SARS-CoV-2 mutations often affect the pathogenicity of the virus; in fact, Wagner and colleagues conclude their study by presenting clinical evidence that ORF8 loss lessens the severity of COVID-19 in patients. Therefore, understanding the mutational landscape of this virus and which mutations are driving selection (as opposed to ‘piggybacking’ on other mutations) has implications for where we focus our research efforts and perhaps even how we treat different viral variants. This study also highlights that there is more to SARS-CoV-2 evolution than spike—while our mechanistic understanding of how ORF8 loss might benefit the virus is still limited (partially because scientists are still unsure what exactly ORF8 does in the first place), it provides a foundation upon which to build more in-depth research efforts and serves as a template for how we might study other cases of similar selection in viruses of public health concern.
The spotlighted work was funded by the Achievement Rewards for College Scientists, the Howard Hughes Medical Institute, the National Institutes of Health, the National Science Foundation, and the Centers for Disease Control and Prevention.
Wagner, C., Kistler, K. E., Perchetti, G. A., Baker, N., Frisbie, L. A., Torres, L. M., Aragona, F., Yun, C., Figgins, M., Greninger, A. L., Cox, A., Oltean, H. N., Roychoudhury, P., & Bedford, T. 2024. Positive selection underlies repeated knockout of ORF8 in SARS-CoV-2 evolution. Nature Communications. 15(1), 3207.
Science Spotlight writer David Sokolov is a graduate student in the Sullivan Lab at Fred Hutch. He studies how cancer cells modify their metabolism to facilitate rapid proliferation and accomodate tumor-driving mitochondrial defects. He's originally from the east coast and has bachelors' and masters' degrees from West Virginia University. Outside of the lab, you'll find him enjoying the outdoors, playing music, or raising composting worms in his front yard.