I’ve long been a believer in taking lessons from nature, not the laboratory, to understand biology, ecology, and human disease. That’s particularly true of natural infections.

When I first learned that Covid-19 was caused by a coronavirus, I hit the books and went back to its natural history. The most recent episodes, SARS and MERS, came and went after being successfully contained. But as I looked deeper into coronaviruses in general, I realized their tendency wasn’t to come and go. Quite the contrary, if you look at the human coronaviruses that cause a third of the seasonal colds we catch like clockwork, the same strains come back again and again—suggesting that coronaviruses, like influenza, might be capable of slipping past immunity acquired from previous infections and returning year after year.

The possibility that SARS-CoV-2, like these cold-causing coronaviruses, would come back to haunt us concerned me deeply, especially when it didn’t figure into early discussions about how to control the Covid-19 pandemic. Some even went so far as to argue that the only way out was through—allowing nature to take its course, even if it meant losing millions of lives in the process, in hopes that population immunity would develop and safeguard those who survived. Those who advanced this argument failed to recognize that such an outcome goes against the very nature of coronaviruses. If decades of research tell us that it is typical for one strain to come back, why would this one be any different?

At first my thoughts, spurred by the discovery that the virus had the ability to alter the immune system, turned to the notion that the immunity we develop against SARS-CoV-2 might be short-lived—rendering the population immunity argument moot. Quite a bit of evidence suggests this, including the observation that the natural immune response to critical illness from Covid-19 is more robust than in asymptomatic and mild cases.

While insufficient time has passed between now and even the very first Covid-19 infections for us to know with certainty whether immunity truly persists, another fundamental assumption underpinned hopes for population immunity that begged scrutiny—that coronaviruses were intrinsically more stable than other RNA viruses just because they had an error proofing mechanism. The textbooks I consulted did indeed suggest that a low rate of mutation resulted in little variation between strains.

The first hint, however, that this might not be the case was the D614G spike mutation that, when it emerged early on in the pandemic, increased the transmissibility of the virus by about tenfold—meaning transmission could happen with fewer particles and in less time. This observation, made later that spring, wasn’t taken seriously by many, who perhaps preferred alternative explanations. Now we know that D614G was a clear warning for what is occurring now—the widespread emergence of variants built on the same framework, but passed from one host to another with even greater ease, leading to transmission rates at least 20 times higher than those that came before.

More researchers have begun sequencing viral samples from Covid-19 patients to identify new and emerging variants, and where they look, they find, first in Britain, then South Africa, and now everywhere from Brazil, Japan, and Germany to Ohio, Illinois, and California. They’re also taking a closer look at how coronaviruses come to vary in the first place—and the new findings are absolutely startling. Certain mutations, not just in the spike protein but across the entire viral genome, could assist the virus in achieving immune escape, allowing it to become more transmissible, virulent, or both.

If SARS-CoV-2 were to gain the ability to mutate its way past our immune defenses, it wouldn’t be the first human coronavirus to do so. In December, a team of Seattle-based researchers published a preprint study (currently undergoing peer review) that set out to determine whether the antibodies of patients who were infected with the human coronavirus 229E one year could protect others from strains that emerged several years later. The antibodies were prepared and administered in the form of convalescent sera, an experimental treatment derived from the blood plasma of recovered patients. Sera for an “early” strain of 229E was tested for its potency against a “future” or later strain, and vice versa.

Seasonal variation in the fraction of positive CoV tests in Stockholm, Sweden. Test results between 2010 and 2019.
Seasonal variation in the fraction of positive CoV tests in Stockholm, Sweden. Test results between … [+] POTENTIAL IMPACT OF SEASONAL FORCING ON A SARS-COV-2 PANDEMIC DOI: HTTPS://DOI.ORG/10.4414/SMW.2020.20224 PUBLICATION DATE: 16.03.2020

Though 229E isn’t nearly as hot a topic as SARS-CoV-2 these days, it is one of several coronaviruses that have been causing us colds for more than half a century. According to the study, sera from 2020 proved adept at neutralizing earlier coronaviruses—a good sign and likely testament to the strength of natural immunity. But sera from 1984, though plenty protective against a variant of 229E from that year, was ten times less effective against a variant from 1992. Similarly, sera collected in 1990 could neutralize the 1984 virus, but nothing more recent, while 1995 sera neutralized all viruses from the preceding decade, then nothing thereafter. Antibodies developed against “early” strains, these results show, faltered when confronted with “future” strains—suggesting that part of the problem lies not with our ability to develop immunity to coronaviruses, but their ability to adapt to us.

The study, we must remember, is of the seasonal coronavirus 229E, not SARS-CoV-2. But just like SARS-CoV-2, the spike protein is how 229E binds to our cellular receptors, making the two ripe for comparison. When taking a closer look at the mutations that allowed 229E to evolve, the researchers did focus on the spike, but also parts of the N-terminal domain and receptor binding domain more broadly, finding significant variability in all three. Though we’ve known for some time that SARS-CoV-2 and coronaviruses in general don’t mutate as quickly as other RNA viruses like HIV, structure-wise they appear to be flexible enough to contort into uniquely advantageous shapes—excellent for their survivability, but not so much for ours.

Beyond the actual fact of variation is the extent of variants that are possible—in other words, not just how many times a coronavirus like SARS-CoV-2 can change, but how much. In analyzing the spikes of five different variants of 229E that emerged from 1984 to 2016, the researchers found that across the entire amino acid sequence of each genome, they differed from one another by up to four percent. But in each spike’s receptor binding domain, the difference was vast—between the spike from 1984 and that from 2016, about 17 percent. This is a harbinger of the astonishing degree of variation of which the surface protein is capable. Far from thinking of these viruses as stable, we must consider them capable of extraordinary flexibility that could occur not just in the spike, but other proteins relevant to transmission and pathogenesis.

 We already know, thanks to genomic databases like GISAID, that every single amino acid in SARS-CoV-2 is capable of mutating. We also know that several mutations—from small deletions to small insertions of other genes to small rearrangements of existing components—can occur simultaneously in a single variant, and that these mutations may preserve and even enhance the spike protein. At this rate, the SARS-CoV-2 we encounter a year from now, not to mention a decade down the road, could look very different from the virus we met over a year ago. Since we can’t anticipate how different, or in what way, the best we can do is accept the possibility and give the virus less opportunities to experiment by reducing the overall disease burden.

In many respects, SARS-CoV-2 is certainly a different beast than 229E—the global death toll of two million says as much. But the leap in lethality seen in SARS-CoV-2 and not other human coronaviruses is all the more reason to revisit what we’ve already learned about the entire family. The virus that causes Covid-19 isn’t a radical departure from its predecessors. It is a familiar foe that will dependably resort to the signature moves that helped it get ahead before, including the ability to change just enough to evade the immune system.

We now have solid but not definitive evidence, published recently as a preprint, that people immunized with the Pfizer-BioNTech vaccine will be protected from new variants circulating in the UK and other countries. The data shows that antibodies in the blood of patients who have been vaccinated are capable of inactivating stomatitis virus carrying the B.1.1.7 lineage spike protein mutations. The question of whether other strains will meet the same fate, as well as whether other Covid-19 vaccines will be as successful, remains open.

Might we need to continuously modify our vaccines for Covid-19 like we do for the flu? Flu vaccines are adjusted each year to account for the major strains that arise, and a single flu vaccine often contains at least four different variants of influenza virus. Given what appears to be the rapid evolution of SARS-CoV-2 in disparate geographic regions, it seems possible, even likely, that future generations of Covid-19 vaccines will follow a similar pattern. In my next piece for this series I will go into further detail about the similarities between influenza and coronaviruses.

Random variation is an essential component of all living things. It drives diversity, and it is why there are so many different species. Viruses are no exception. Most viruses are experts at changing genomes to adapt to their environment. We now have evidence that the virus that causes Covid, SARS-CoV-2, not only changes, but changes in ways that are significant. This is the sixth part of a series of articles on how the virus changes and what that means for humanity. Read the rest: part onepart twopart threepart fourpart five, and part six.