The Greek alphabet has 24 letters. A little less than half was spent on the names for the variants of the coronavirus, but it is already clear that this is not the limit. It will hardly be possible to stop the evolution of SARS-CoV-2, and, probably, other alphabets will be used one day. It remains for us to adapt to this - and learn to predict how dangerous this or that newcomer will turn out to be. We will tell you how such a forecast can be built and what you should pay attention to in order to take measures in time - or, conversely, do not panic once again when you hear about the appearance of another fit or sigma on the horizon.
The delta coronavirus was designated Variant of Concern on May 10, 2021. Then he was not yet called "delta", and WHO experts argued about whether he was really more infectious than his predecessors. But the third wave of the pandemic - which, as we now understand, was largely caused by the delta - by that time had been going on for almost three months.
- November 11, 2020 - the genome of the new variant appears for the first time in the GISAID database
- February 5, 2021 - epidemiologists notice the same option for the second time
- March 10, 2021 - Covid outbreak begins in India
- March 24, 2021 - Indian government reports new variant of coronavirus detected in the country
- March 30, 2021 - the sequence of the new version appears on github. It is reported that it is spreading in clusters and has already reached Europe, the USA, Australia and Canada, and in India it is responsible for about 20 percent of covid cases.
- April 2021 - “delta” appears for the first time in the base of the Russian consortium Corgi
- May 10, 2021 - Delta is recognized by WHO as an option of concern
- May 20, 2021 - the head of Rospotrebnadzor announces the appearance of the "delta" option in Russia
It is now pointless to argue about whether a new wave of a pandemic could have been avoided - in the world, or at least in India or Russia taken separately. But it would be good to think about how to prepare for the next option, in order to catch it still on the way and, if possible, avoid casualties.
Coronavirus Alphabet for August 2021
Variants of concern:
ɑ (B.1.1.7)- discovered in the UK in September 2020
β (B.1.351)- discovered in South Africa in September 2020
ɣ (P.1)- discovered in Brazil in December 2020
Δ (B.1.617.2)- discovered in India in December 2020
Variants of interest:
η (B.1.525) - discovered in Nigeria in December 2020
κ (B.1.617.1) - discovered in India in December 2020
ƛ (C.37) - discovered in Peru in December 2020
ι (B.1.526) - discovered in the US in December 2020
Options that remain under observation (for further monitoring):
ε (B.1.427 / B.1.429) - discovered in the US in September 2020
ζ (P.2) - discovered in Brazil in January 2021
θ (P.3)- discovered in the Philippines in January 2021
Fyodor Kondrashov, an evolutionary biologist at the Austrian Institute of Science and Technology, thinks sequencing could help.
“In order to understand which option is dangerous and which is not dangerous,” he advises, “it is very useful to get as much data as possible. Now we say: how bad - 10 percent of the population had some variant, but 50 percent did. And imagine that we would see that some variant appears in 0.01 percent of the infected, and now in 0.1 percent. This is a completely different situation."
This means that in every region of every country it is necessary to equip a genetic laboratory - to put a boy with a sequencer to keep track of all the varieties of coronavirus that the population suffers from. And he shouted "wolves!" If he noticed a suspicious movement in the bushes - before the animals began to rush at people.
This plan, however, does not seem realistic even to its author. Kondrashov admits that such a system will require such monetary investments that will hardly ever pay off.But let's imagine that it did happen - at least in a single country. In each of its cities, the authorities installed a sequencer that reads viral genomes without a break for lunch and sleep. It remains only to understand: what kind of movement is suspicious? What properties of the new variant should we look at to worry about before the outbreak starts?
Where is the mutation?
The genome of SARS-CoV-2 contains almost 30 thousand nucleotides. But most of the mutations that are heard today crowd in a region of a couple of hundred nucleotides - this is the RBD (receptor binding domain) of the S-protein: the place where the virus sticks to the ACE2 molecule on the cell surface.
Genome of the coronavirus SARS-CoV-2
This does not mean, of course, that the rest of its genome is not mutated by the coronavirus. Each new variant carries dozens of changes, most of them affecting other regions of the S-protein or other genes altogether. And some of these changes could seriously affect the biology of the virus. So, in the alpha variant (we talked about it in the text "We've got a new one"), a mutation was found that increases the production of the viral protein orf9. It is responsible for suppressing the host's immunity - which means that such a mutation could allow the virus to survive longer in the body.
Meanwhile, in other varieties of coronavirus, geneticists noticed a mutation in the orf1ab gene, which, on the contrary, inhibited the reproduction of the virus. It interfered with the work of the Nsp1 protein, which prevents the host cell from producing any proteins other than viral ones.
And yet, virologists are primarily interested in the spike protein. “It's just that it largely determines the properties of the virus from the point of view of an external observer,” explains Georgy Bazykin, an evolutionary biologist at Skoltech, “that is, the immune system and those cells into which it tries to penetrate. This is both the virus's first line of attack and the first line of defense."
It is at the spike that the immune system "looks" when it selects antibodies to it. Natural selection seems to be looking there too: at least, mutations in it more often than others coincide in different variants of the virus. Therefore, we will have to look at the spike too - in order to understand whether the "old" antibodies will lose their effectiveness against the new variant.
This, however, is not surprising. Exactly the same thing happens with other viruses - be it HIV or influenza - virologists are primarily interested in surface proteins. And this simplifies our task: despite the fact that sequencers usually read the entire viral genome, we know in advance where to expect trouble.
She is real?
Boys who shout "wolves!" Are notoriously wrong. Sequencers are also mistaken. “There are some places in the coronavirus genome,” Bazykin explains, “where some sequencing technologies stumble. It seems to you that there is cytosine, although in fact there is uracil there."
This means that the mutation that the sequencer "saw" may not actually exist. But if such a mistake is relatively easy to recognize and correct, then it is much more difficult to separate the own mutation of a new variant from the many small changes that the virus has accumulated during its lifetime inside the host.
Human cells are protected from viruses by proteins of the APOBEC family. They edit viral RNA, randomly converting cytosines into uracils, hoping to break a gene. And since every coronavirus sample we put into the sequencer comes from a specific patient, we never know for sure how to interpret the detected mutations. This may be an important property of a new version, common to all its copies, or it may be the result of the work of APOBEC. But since these proteins are the same for all people, they will invariably make similar edits. “It is very difficult to distinguish functional parallelism from parallelism associated with editing,” complains Bazykin, “and if you see that some mutation has happened on your evolutionary tree of the coronavirus a thousand times independently, this, unfortunately, does not necessarily mean that this mutation is useful for the virus."
Therefore, just a boy with a sequencer is not enough to compose a mutational portrait of a new version.If we want to be sure that changes in its genes are really important and supported by selection, we will have to prove this in vitro: collect a mutant spike protein and check whether it binds better to its target or worse - to antibodies.
So, presumably, spike proteins look in different variants of coronavirus
One could, perhaps, entrust this task to a computer - and limit ourselves to a model of the altered protein. But both interlocutors N + 1 agree that so far we can do it pretty badly. The boy with the computer model has no faith yet, so we cannot do without the boy with the test tube.
Do we know her?
Some mutations do not need modeling - those that were included in the previous variants of the coronavirus and therefore have already been studied far and wide. Genetics discussed some of them so zealously that they even gave them human names - to make it easier to pronounce. So the replacement D614G (here D and G are the codes for the initial and final amino acids, and 614 is their position in the protein), which makes the spike protein sticky to the ACE2 receptor, was named Doug. And E484K, which makes the virus less visible to antibodies, became Erik.
Together with the name, mutations gain popularity in the scientific community, and at the same time the status of potentially dangerous. That is why, for example, the delta plus option (AY.1), which was first noticed in India at the end of April, caused great concern - it even forced, for example, the UK to limit air traffic with Portugal. In addition to the substitutions typical for the usual delta, the AY.1 variant acquired the Karen mutation (Karen, K417N). We have already met her in the genome of the South African version of beta and we know that this meeting does not bode well: it is believed that it is "Karen" that allows beta to escape from binding to antibodies.
Mutations that have earned a name
Doug (Doug, D614G):Available in all existing variants. Increases binding to ACE2.
Eric (Eric, E484K):Available in beta, gamma, eta, theta, iota, and zeta options. Escapes antibodies.
Karen (Karen, K417N): Available in beta version of the coronavirus. Reduces binding to ACE2. Escapes antibodies.
Kent (Kent, K417T): Available in a gamma version. Reduces binding to ACE2. Escapes antibodies.
Leif (L18F): Available in some beta versions. Escapes antibodies.
Nelly (Nelly, N501Y): Alpha, beta, scale, theta have. Increases binding to ACE2.
Pooh (P681H): Alpha and theta options. Increases infectivity (helps to penetrate the cell bypassing ACE2).
Sean (S477N): Found in some varieties from Australia and New York. Escapes antibodies.
It would be useful, of course, to be able to predict the properties of mutations even before they get into the sequencer and get names. We are gradually moving in this direction thanks to Jesse Bloom's group in Seattle. Together with his colleagues, Bloom is exploring the spectrum of possibilities of the coronavirus using saturation mutagenesis: scientists make all possible substitutions in each of the amino acids of the S-protein and force the yeast cells to produce these mutant proteins. It turns out a library of yeast cultures, each of which exposes its own version of the S-protein on the cell surface, which differs from the original by one amino acid. Then, the luminous receptor ACE2- is added to each culture and it is calculated how often it is deposited on yeast cells. Based on the results of the experiment, Bloom builds red-blue tables, in which the redder the cell, the stronger the binding - and the more potentially dangerous the mutation. “If anyone wants to predict the evolution of a covid,” says Bazykin, “here are the best data on which to do this.”
Binding tables of mutant spike proteins with ACE2. Horizontal - position in the protein, vertical - amino acid substitution option
Elsewhere, Bloom and colleagues collide their modified proteins with antibodies - and draw conclusions about which substitutions are potentially more slippery.From his works, for example, it follows that “Karen” (K417N) is one of the most unpleasant substitutions in position 417, only the hypothetical “Katya” (K417I) can be worse, and “Eric” (E484K) is not as scary as his possible brother "Emil" (E484L), whom we have never met before.
This is how different amino acid substitutions help the spike protein escape from antibodies. The larger the letter, the stronger the escape. The color indicates the binding strength of the mutant protein to ACE2
What else is in the genome?
The problem is that most of these mutations are far from new. During the time that the coronavirus has been walking around the world, it has multiplied so many times in cells that each replacement should inevitably arise. But why, among the many mutations, only a few have become entrenched is still unclear. Perhaps the fact is that each of the substitutions is not particularly dangerous one by one - and they only "shoot" together when they are going to a new version.
For example, in the delta spike protein, two amino acids have changed at once.
- The E484Q mutation was suspiciously reminiscent of Erica (E484K) - which was already known to bind better to the ACE2 receptor and worse to antibodies.
- The P681R mutation was very similar to Pooh (P681H), which is responsible for the infectiousness of the virus. The changes in amino acid 681 make the S protein site more sticky to the human enzyme furin. The virus meets furin inside the cells, before going outside. Furin cuts the S protein (this happens inside the cells) and makes it more sticky to another enzyme, TMPRSS2. When the virus prepares to infect the next cell, it can meet TMPRSS2 on its surface, which cuts the S-protein again - and then the protein can adhere directly to the membrane of the new cell without the participation of ACE2.
The union of the two mutations turned out to be mutually beneficial: the first helped the second to withstand the onslaught of antibodies, and the second allowed the first to spread more widely in the population. True, it is unclear why it took them so long to settle in one genome. Maybe the fact is that coronavirus genomes do not recombine with each other very intensively and do not exchange regions - which means that the virus has to rely on chance. In other coronaviruses, however, recombination has been described. But in SARS-CoV-2, according to Bazykin, it is not easy to detect it - the genomes of different variants are quite similar to each other, and it is not always possible to distinguish recombination from point replacement.
Obviously, over time, the number of replacements in coronavirus variants will only grow. In 2020, the Dag mutation (D614G) spread so successfully that now it is not even indicated in the lists of features of one or another variant - it is found in each of them. The same can happen with the delta: its mutations can become a new norm if it is firmly entrenched in the population and itself becomes the soil for further evolution (which is already happening little by little, as can be seen in the example of the “delta plus” variant).
Location of different mutations in the beta spike protein
Therefore, Jesse Bloom's tables should be replaced by multidimensional matrices - attempts to calculate in advance how several amino acid substitutions within one protein will interact at once. There is no hope for computer modeling here, and even more so. And this is not an easy task for saturating mutagenesis. “There are a couple of dozen positions in a protein that are known to be important,” estimates Bazykin, “and each of them may contain 20 different amino acids. To check in pairs all combinations of mutations, you will have to make 19x19x190 new variants of the genotype - that is, about 70 thousand. It is still too expensive. " If the list is reduced to five positions, only 3610 measurements are needed - which, according to the scientist, sounds more realistic. “It would be a very good task,” he agrees, “if anyone has the money and opportunities. And I think it will be done."
Who is around?
And so, let's say, we examined in all details the genome of the next variant and studied what risks each of its mutations carries. Even if we understand what tricks he has up his sleeve, this is not enough to predict how he will behave in human society.
Since this pandemic has been with us for a long time, not a single option comes to an empty place - it immediately enters into competition with those who are already "feeding" in this population. And they begin, as Bazykin put it, "pushing their sides." "An option arises," he says, "very good from the point of view of the virus, but it is prevented from spreading by another option, which may not be so good, but circulates with it at the same time."
In addition, the previous variants, which "grazed" on the same people, have already left behind a legacy - in the form of acquired immunity. And to those who have been ill or vaccinated with the virus, a completely different approach is needed. If in a naive population it is easy enough to spread quickly, then in an immunized population those who are invisible to antibodies can get a head start - this allows people to infect people again. Therefore, it may turn out that the same mutation does not win in all situations. For example, if it decreases binding to ACE2 and at the same time is less "visible" to antibodies (as, apparently, "Karen" acts), then in a naive population it can be expected that it will lose, and in an immunized population it will win.
As a result, the fate of even the most dangerous variants of the coronavirus can turn out to be unexpected. The delta, for example, was first noticed back in November 2020, and the common ancestor of all deltas appears to have appeared a month earlier. However, for almost half a year, this option did not manage to get ahead.
Beta escapes vaccine immunity better than most others and has spread almost all over the world. But all this did not help him overtake his rivals. Without worrying about the increase in infectiousness in advance, the beta in almost all countries is now inferior to the delta. And even in Moscow, where its share increased unexpectedly in April, beta is now almost negligible.
The same fate befell the "Moscow" version of B.1.1.523 (it did not get the Greek letter, since the WHO did not consider it suspicious enough to be included in the list of "causes of concern"). Back in April, it infected more people than alpha and beta combined - but also surrendered under the pressure of the delta.
The share of different variants of coronavirus among Moscow samples
Each option is under pressure from several sides - both the coronavirus neighbors and the human immune systems. The vector of adaptation to this pressure in different populations, therefore, will differ. Therefore, no one has yet undertaken to predict which mutations selection will support and which will discard, and whether it will change its mind after some time.
Where to go?
Nevertheless, we can already say something about the future of man and his coronavirus. We have experience from previous pandemics.
“I don’t think that evolution has reached the maximum that this virus can do in such a short time,” says Fyodor Kondrashov. Bazykin agrees with him: "by the definition of natural selection," he reminds, "the options are selected that are most transmitted." Therefore, the following suspicious options can be expected to be more infectious than the delta.
This, however, does not mean that the very nature of the disease will change. “The virus,” explains Bazykin, “doesn't care how severe the symptoms it causes. It is important to him how easily it is transmitted from one person to another, and what happens to this person does not bother him. " Therefore, it is unlikely that natural selection will specifically support mutations that make the virus more lethal. Another thing is that the more infectious the virus, the more people get sick with it at the same time - and the higher the burden on the healthcare system. This means that the risk of dying in the population may grow - and not only from covid.
But humanity does not stand still.During the time that the virus has accumulated a couple of dozen amino acid substitutions, we have created several dozen coronavirus vaccines. To escape from their action, the virus had to acquire new mutations - but the vaccination campaign reduces the number of replicating viruses in the population, and it becomes more difficult to evolve.
"The effective reproductive number," notes Bazykin, "in countries that are partially immunized and where the delta is now spreading, is about the same as it was in early 2020, when the virus was young and inexperienced, and we were young and inexperienced." In a year and a half, he learned to elude us better, and we learned to catch him better. And they ran after him fast enough that both remained in the same place. But in order to get somewhere from this place, you will have to run several times faster - that is, to supply people with immunity faster than the virus increases its infectiousness.
And there is a catch here. The harder we put pressure on the virus, the stronger the incentive it has to evolve. As Fyodor Kondrashov and his colleagues recently calculated, the likelihood of new variants of coronavirus emerging from the action of vaccines, the higher the fewer unimmunized people remain in the population. And this probability reaches its peak at about the moment when the population approaches herd immunity. And if for the initial, Wuhan version, this threshold was in the region of 60 percent of the vaccinated population, then for the delta, according to the scientist, it could be higher - about 80-85 percent.
“This is the only time when epidemiological and evolutionary considerations diverge,” says Kondrashov. - From an epidemiological point of view, when we vaccinate everyone on the sly, it gets better, better and better. From an evolutionary point of view, when we vaccinate everyone on the sly, it’s already dangerous.”
And if we want to stop the pandemic, the scientist believes, we must think about it in an evolutionary context - and try to prevent the evolution of the virus as a whole. This can be done only by slowing down its spread, which means that not only vaccines will be needed, but also masks, border checks and quarantines. And then, perhaps, there will be no need to guess about the fate of the new variant by the pattern of nucleotides in its genome.