Microbes and bacteriophages: an arms race

Scientists have found out why it is difficult for bacteriophages to fight the immune system of bacteria

Alexander Markov, "Elements"

The never-ending arms race between bacteriophage parasites and their bacterial victims underlies the rapid evolution of both.

1

The system of inherited acquired immunity CRISPR-Cas, widespread in prokaryotes, reliably protects its owners from viruses. However, viruses, constantly mutating, are able to quickly (sometimes in less than a day) overcome the immunity of any individual victim. As it turned out, the effectiveness of the CRISPR system is ensured by the fact that thanks to its work, different bacteria learn to recognize the virus from different parts of its genome. As a result, the ways of protecting bacteria from this virus become so diverse that no point mutations help viruses effectively adapt to the collective defense of victims. This, in turn, contributes to the evolution of special viral genes that suppress the work of the CRISPR system as a whole, and bacteria respond to this with the evolution of new variants of the CRISPR system – but such changes require more time.

The operation of the CRISPR-Cas system is based on the fact that a small fragment cut from foreign (for example, viral) DNA that has penetrated into a bacterial cell is inserted into a special section (CRISPR locus) of the bacterial genome. Each CRISPR locus contains many such inserts ("spacers"), which are pieces of the hereditary material of parasites (viruses, plasmids, mobile elements). On the basis of spacers, RNA molecules complementary to a section of the parasite genome are synthesized. These RNAs in combination with Cas proteins are then used to identify and neutralize foreign DNA with the same sequence of nucleotides. Thus, if viral DNA once penetrated into a cell, but the cell managed to survive and embedded a piece of the viral genome into its chromosome, then subsequent attempts by the same viruses to infect this cell or its descendants will be stopped quickly and effectively.

However, viruses do not tend to remain "the same" for a long time. Due to random mutations and selection, they are able to bypass the immune defense of victims. In order for this spacer to lose its effectiveness, it is enough that the fragment of the viral genome complementary to it at least slightly changed. Therefore, viruses successfully and sometimes very quickly overcome the acquired immunity of bacteria due to point mutations. On the other hand, CRISPR systems are very widespread in prokaryotes and, apparently, provide their owners with reliable protection. What allows these systems to compete on an equal footing with rapidly evolving parasites?

Geneticists and microbiologists from the UK, France and the USA have suggested that an important contribution to the effectiveness of CRISPR is made by the fact that in response to the same viral infection, even genetically identical bacterial cells insert different spacers into their genome corresponding to different parts of the virus genome. As a result, the population of victims quickly acquires genetic diversity, which greatly complicates the evolutionary task facing viruses. By acquiring a point mutation that protects against a single spacer, viruses will be able to infect only a small part of the victim population. Fortunately for bacteria, the bacteriophage cannot determine in advance which spacers a given cell has: this will become clear only when it injects its DNA into it, and then it will be too late to change its mind. Therefore, most of the phages in the polymorphic population of victims are doomed to death, even if the phages now and then have point mutations that protect against this or that spacer. To protect itself from many different spacers at once, the phage needs to simultaneously acquire a whole complex of necessary point mutations, which is extremely unlikely, because mutations are random.

These assumptions were tested on Pseudomonas aeruginosa (Pseudomonas aeruginosa) bacteria and DMS3vir phages. To begin with, the authors made sure that the CRISPR system really reliably protects bacteria from this type of phages. In populations of "wild" bacteria, the phages introduced by the researchers completely died out in just 5 days. In bacterial cultures with the CRISPR system disabled, the viruses felt more at ease: for 30 days, while the experiment was going on, the viruses did not die out in any of the lines. In the first case, all bacteria, as expected, have developed an inherited acquired immunity based on CRISPR. Bacteria with the CRISPR system disabled have also developed some protection, but on a different basis: mutations have spread among them, changing the surface protein receptor that the phage clings to in order to infect the cell. Apparently, the second method turned out to be less effective and did not allow populations to completely get rid of the parasite, as did populations with a working CRISPR system.

The authors emphasize that the complete extinction of viruses in wild-type populations is an unexpected result, since it is known that viruses, in principle, are able to bypass CRISPR immunity with the help of point mutations. Maybe it's all about the diversity of spacers formed in bacteria inside each of the experimental populations? The fact that spacers really turn out to be different has been shown in previous experiments. Now it was necessary to prove that this diversity contributes to the effectiveness of collective immune defense. To do this, it is necessary to compare the resistance to viruses in populations with different levels of diversity of spacers.

To obtain such populations, the authors took their experimental lines that defeated viruses with CRISPR and isolated 48 individual clones from them (that is, they took 48 individual cells and received numerous offspring from each). Scientists expected (and these expectations were later confirmed) that spacers in all or almost all clones would be different. Then, from these clones, populations of five types were composed, differing in the level of genetic diversity: populations from one clone (monoculture) and mixed populations composed of 6, 12, 24 and 48 clones.

These populations were then exposed to viruses. This time, scientists were primarily interested in the evolution of viruses, namely their ability to overcome the immune defenses of victims. Therefore, the observations lasted only 3 days – enough time for noticeable evolutionary changes in viruses, but not enough for the experimental bacterial populations to have time, creating new spacers, to align in terms of genetic diversity.

The results confirmed the authors' expectations. In three days, the viruses completely died out in all populations of victims made up of 24 and 48 clones, and in many populations made up of 12 clones. In the least diverse populations made up of one or six clones, viruses in most cases survived.

Every few hours during this three-day experiment, a part of the viruses was removed for detailed analysis. Phages were subjected to genome-wide sequencing, and also introduced into pure cultures of each of the 48 bacterial clones to see in which cases the viruses learned to overcome the immune defenses of the victims.

It turned out that those viruses that evolved for three days in bacterial monocultures, in most cases, fixed mutations that make the corresponding spacer ineffective. And this happened, as a rule, already in the first day. Only five of the 48 monocultures were not able to cope with viruses. As it turned out, in three cases out of these five, the bacteria formed not one, but two or more antiviral spacers.

Viruses that have evolved in bacterial populations composed of 6 and 12 clones have developed resistance to one or another spacer only in a few cases. Well, those poor devils who had to evolve in the most diverse populations of victims, made up of 24 and 48 clones, did not learn to overcome the protection of any of the original clones.

The genetic diversity of victims prevents viruses from overcoming their immune defenses. The figure shows the results of experiments on infection with viruses that co-evolved with different populations of bacteria, each of the 48 original bacterial clones. Each of the five columns corresponds to a series of experiments with one of the five levels of victim diversity (from left to right: 1 clone, 6, 12, 24, 48 clones). In each experiment, viruses were taken for analysis after 0, 16, 24, 40, 48, 64 and 72 hours after infection (6 vertical rows inside each column; the time in days after infection is signed at the bottom – d.p.i.). Each portion of the viruses obtained in this way was introduced into the pure culture of each of the 48 bacterial clones and looked at whether the viruses could multiply (red squares) or die out (green squares). Thick black horizontal lines separate the repetitions, that is, different experimental populations with a given level of diversity. The figure shows that in bacterial monocultures, viruses in most cases learned to overcome immune defenses in less than a day; this rarely happened in populations of 6 and 12 clones, and never in populations of 24 and 48 clones. A drawing from the discussed article in Nature.

The reason, of course, is not that these viruses had "correct" mutations with a reduced frequency. Mutagenesis most likely proceeded the same way in all viruses. It's just that in a monoculture, a single mutation that provides protection from this spacer immediately gives the virus a huge advantage (since it allows infecting any bacterium), and in a polymorphic culture of 48 clones, exactly the same mutation will ensure success for the mutant virus only with a probability of 1/48. Even if he is incredibly lucky and he injects his DNA into exactly such a bacterium, from whose spacer he is now protected, his descendants will again face the same problem. And it will only get worse as the number of bacterial clones that have become vulnerable decreases. However, it usually does not even come to this: selection practically does not support point mutations in viruses that protect against individual spacers, and viruses die out before they have time to overcome the protection of any of the 48 clones.

These conclusions are confirmed by the fact that the number of mutations found in the genomes of viruses that evolved in monocultures of victims turned out to be significantly higher than those of viruses that evolved in mixed cultures: in the first case, selection supported beneficial mutations in viruses, and in the second it did not. In addition, it turned out that bacteria from mixed cultures did not acquire new antiviral spacers in three days, whereas such spacers appeared in monocultures. This is also logical: viruses that are dangerous for bacteria have not appeared in mixed cultures, and viruses that break through the old protection have appeared in monocultures, and the CRISPR system has created additional spacers.

Thus, the diversity of spacers generated by the CRISPR system really makes an important contribution to its effectiveness. If the system responded to each infection by creating the same spacer in all bacteria, viruses would easily cope with such immunity. But since spacers are different every time, point mutations and selection become an insufficiently effective evolutionary strategy for viruses.

This explains the effectiveness of the CRISPR system and its widespread use in prokaryotes. It remains to understand why bacteriophages have not yet died out, since this system is so good. The answer to this question is partly already known: recently, special genes have been found in phages that suppress the CRISPR system as such (see: J. Bondy-Denomy et al., 2013. Bacteriophage genes that inactivate the CRISPR/Cas bacterial immune system). In this regard, the following question arises: if there are viral genes that completely disable CRISPR, then why haven't all viruses acquired them? And what can bacteria oppose to these genes? However, the answer to this question is also already partly known: there are many different variants of the CRISPR system, each of which is vulnerable only to some variants of the anti-CRISPR genes and is protected from others. And to contain a bunch of additional genes in your genome is an expensive pleasure for viruses, in which selection usually supports the compactification of the genome (this increases the rate of reproduction of the virus).

Apparently, antagonistic coevolution (arms race) between phages and bacteria goes on in parallel at different levels and on different time scales. The formation of new spacers and the production of point mutations by viruses that make this spacer ineffective takes only a few days or even hours. The development of new anti-CRISPR genes or new variants of the CRISPR system that are invulnerable to these genes may take thousands or millions of years (although in the world of viruses and microbes, everything that was once invented by someone can then be quickly transferred from hand to hand by horizontal transfer). The end of this race is not in sight, and it is unlikely that anyone will come out of it as the "final winner". However, knowledge of its mechanisms will allow people to at least partially take control of it in the future.

Source: van Houte et al., The diversity-generating benefits of a prokaryotic adaptive immune system // Nature. 2016.

Portal "Eternal youth" http://vechnayamolodost.ru  18.04.2016