Making changes to genes with CRISPR has the potential to cure diseases and feed the world, if we can learn to control it. Now it looks like viruses hold the solution

THERE is a technology that could tackle some of life’s most pressing problems, from disease to malnutrition. It could fix medical conditions such as cystic fibrosis and sickle cell anaemia simply by changing a bit of genetic code. It could eliminate malaria by making male mosquitoes infertile. It could wipe out pests that destroy crops. And it could modify other organisms to increase their usefulness, helping to create foods that are tastier and more nutritious.

This is the promise of CRISPR, a biochemical tool at the forefront of a gene-editing revolution. Produced naturally by bacteria, CRISPR has gained rock-star status among scientists in the decade since its extraordinary potential was first recognised, and it is already starting to live up to the promise. But behind all the excitement lurk some dark questions. What if the editing goes wrong? What if it has undesired effects? What if we can’t stop it? Without a means to keep CRISPR on target and halt it in its tracks when needed, gene editing could have disastrous consequences – both for human health and for the planet.

What we need is an off-switch, one that can be used at will. Researchers around the world have spent years trying to find one, largely by investigating various biochemical solutions. However, it turns out that the answer may be right under our noses. In an evolutionary face-off between CRISPR-producing bacteria and the viruses that infect them, nature has already designed anti-CRISPR. The challenge now is to harness this evolved off-switch to our own ends and usher in the golden age that gene editing promises.

Viruses, such as the one that causes covid-19, don’t just pose a threat to humans – they attack all living organisms, including bacteria. In the ancient bacteria-virus rivalry, CRISPR is one of the weapons bacteria have evolved to combat bacteriophages, the name given to viruses that infect them (see “Evolutionary arms race”).

CRISPR forms part of many bacterial genomes. It is made up of repeating DNA sequences interspersed with fragments of genetic code left behind by phages from past viral attacks . When a phage invades again, the bacterium makes RNA copies of these CRISPR regions. These bits of genetic material then hook up with a particular protein, an enzyme called Cas. They latch on to matching sequences in the invading virus’s genome, and the accompanying Cas protein snips the viral DNA strand, destroying the phage. In effect, CRISPR works as a sort of genetic memory of past viral attacks that confers immunity against future ones.

Elegant editing

Given the system’s simplicity and elegance, it is perhaps unsurprising that researchers eventually spotted CRISPR’s potential as a gene-editing tool. The discovery won a Nobel prize in 2020 for biochemist Jennifer Doudna at the University of California, Berkeley, and Emmanuelle Charpentier, now director of the Max Planck Unit for the Science of Pathogens in Germany. In research published in 2012, they presented a CRISPR system that contained genetic sequences of their choice, rather than ones from phages, along with a Cas enzyme called Cas9. With this tool, biologists can home in on a specific DNA sequence and make a cut at a precise location. This allows them to disable a target gene or excise a faulty one and replace it with a working version.

CRISPR-Cas9 has since been used successfully many times to genetically edit cells in the lab. But for it to be an effective medical therapy, it must be delivered directly to cells in the human body either physically, such as by injection, or with a vector, usually an engineered virus that encodes the desired genes. In 2020, a team in the US achieved this for the first time, injecting CRISPR into the eyes of someone with an inherited form of blindness caused by a single mutation. Precisely targeting other parts of the body is harder, however. The issue is how to get CRISPR only to the cells of interest, while also ensuring that enough editing takes place in them to see the changes you want. With vectors “there is no ‘magic bullet’ – it’s a bit of a shotgun approach”, says molecular biologist Erik Sontheimer at the University of Massachusetts Medical School. This is where concern begins to creep in.

Many scientists worry about the consequences if gene editing is left unchecked. “Expression of Cas9 in the wrong place, or for too long, is going to be very dangerous,” says microbiologist Alan Davidson at the University of Toronto, Canada. The main problem is that CRISPR can zero in on sequences that are similar to, but not an exact match for, its target, and Cas9 is then able to cut such sequences. As a result, there is a risk that, while being used to treat a genetic disease, CRISPR-Cas9 could cause harmful changes elsewhere in a person’s genome – so-called off-target edits.

Beyond medical uses, the consequences of uncontrolled gene editing are equally concerning. In an application called a gene drive, CRISPR-Cas9 can be used to boost the prevalence of certain genes in a population by editing them to increase their chances of being passed on to the next generation. A gene drive could be used to great effect: to eradicate a vector-borne disease such as malaria, for example, by promoting a gene that makes male mosquitoes infertile or one that prevents females from biting. But there is a danger that such genetically edited organisms might run amok in the environment with unintended consequences.

Take back control

We clearly need a way to more closely control CRISPR-Cas9. That is where anti-CRISPR comes in. CRISPR has been found lurking in the genomes of half of all sequenced bacteria. However, some phages have evolved their own system to fight back.

Their defence consists of small proteins called anti-CRISPRs (Acrs) encoded in their genome. When a phage infects a bacterium, it injects its genetic material and then hijacks the host’s genetic machinery to make copies of its own genes. The Acr genes are among the first to be expressed, which means that anti-CRISPR can get straight to work to block the bacteria’s CRISPR response. It uses a variety of mechanisms, including attaching directly to the Cas enzyme and preventing CRISPR-Cas from binding to DNA.

Anti-CRISPR was discovered by accident in 2012, just as the CRISPR gene-editing revolution was taking off. Working in Davidson’s lab, microbiologist Joseph Bondy-Denomy was surprised to find that phages infecting a pneumonia-causing bacterium weren’t being destroyed by the microbe’s CRISPR system. Looking more closely, he discovered that the virus had genes capable of inactivating the bacteria’s defence. At first, Bondy-Denomy didn’t realise the magnitude of his discovery. Back then, no one was thinking about the problem of keeping CRISPR-Cas9 under control, let alone ways to do so. Nevertheless, he continued studying Acrs, finding them in a range of other phages. For a while, he and his colleagues had the field to themselves.

In 2016, Bondy-Denomy and his colleagues found Acrs capable of disabling the Cas9 enzyme, the one used in the vast majority of gene-editing studies. By then, he had his own lab at the University of California, San Francisco. His former colleague April Pawluk discovered this same protein, AcrIIA4, simultaneously. Now CRISPR researchers did take notice. With the problem of control widely recognised, they began to pile in to anti-CRISPR research. Within months, a team including Doudna had delivered AcrIIA4 into human cells along with CRISPR-Cas9, allowing them to limit gene editing to a brief period, so minimising the problem of off-target edits. Timing when the Acr was administered provided another layer of control, allowing them to switch off Cas9 either abruptly or gradually.

There is still much to discover about Acr proteins, but their potential to regulate gene editing is clear. “It’s a very exciting time to think about this Acr strategy as one option – or even the premier option,” says Bondy-Denomy. “What it gives you, is the ability to have a genetic off-switch encoded with Cas9.”

That has significant advantages over other possible approaches. For a start, it would minimise the number of therapies patients are exposed to compared with using a separate drug to inhibit Cas9. It also means you can be sure that Cas9 and AcrIIA4 are in the same place at the same time. And it can give you more control to stop and start gene editing by allowing the activity of the two proteins to be toggled back and forth – potentially by using light. Already, one group has engineered a version of AcrIIA4 that can be turned on and off by shining light onto it, a technique known as optogenetics.

On target

Anti-CRISPR might even help solve the problem of getting gene editing to occur only in certain cells. By tinkering with Acrs, it is possible to produce CRISPR-Cas9-Acr complexes that are permanently off in non-target parts of an organism. Sontheimer’s team demonstrated this in the first successful study of anti-CRISPR in a living organism. The researchers created an Acr that was active unless it was in the presence of a snippet of RNA found only in liver cells. They then added this to CRISPR-Cas9 so that gene editing occurred only in a mouse’s liver. Such an approach could potentially be used in any organ that contains a unique RNA molecule.

“We need a way to control gene editing – that is where anti-CRISPR comes in”

Combining Acrs with CRISPR offers long-term benefits too. Viral vectors are currently the most common method for getting CRISPR-Cas9 genes into cells, but these then remain there indefinitely because they become incorporated into the cell’s DNA. As a result, there is a risk that, somewhere down the line, Cas9 will somehow become active again and make undesired edits in the genome. This can be prevented if the gene-editing sequence also contains an off-switch.

Progress with anti-CRISPR has been remarkably rapid, but huge questions remain. One is whether it is safe to administer Acrs to people. “I don’t think that they have any risks that aren’t present with using CRISPR or, in fact, any foreign proteins that you would introduce into people,” says Davidson. Nevertheless, both Acrs and CRISPR-Cas9 are of non-human origin so could generate an immune response that would inactivate them and might cause damaging inflammation. Cas9 has already been seen to generate antibodies in mice. However, Acrs are around 100 times smaller than Cas9, which means they stand less chance of being recognised by any antibodies that might be produced in response to them. Bondy-Denomy thinks that adding Acrs is unlikely to make things any worse.

The biggest challenge will be making anti-CRISPR work in practice. To control gene editing using Acrs, we need to find ways to deliver them to the right place inside the body and reliably control them once they are there. We aren’t capable of that yet. “We’re still just scratching the surface,” says Davidson. That is hardly surprising, given that this idea is just a few years old. Bondy-Denomy, for one, believes it will happen one day. But to make the life-saving potential of CRISPR a reality, anti-CRISPR needs to generate the same level of interest and creative research as its nemesis. “It’s really important to get it on everybody’s radar,” says Bondy-Denomy.

Evolutionary arms race

The defence mechanism that bacteria have evolved in response to the viruses that infect them is ingenious. Known as CRISPR, it consists of two elements: a stretch of DNA that can target viruses the bacteria has encountered and an enzyme that then chops up the invaders. This ability to target and destroy has brought CRISPR to the attention of researchers aiming to develop gene editing (see main story). However, in the natural world, CRISPR’s power seems to be waning. Some bacteria-infecting viruses have evolved a protein-based counter-attack called anti-CRISPR – and it is almost always successful.

It is a bit of an evolutionary puzzle why natural selection keeps CRISPR going in bacteria that meet resistance from anti-CRISPR. One possible explanation is that CRISPR has acquired other useful functions. For example, it seems to help some bacteria form biofilms – diverse communities of microbes that have many advantages for the survival of their inhabitants. In addition, some bacteria use CRISPR to help regulate the expression of their genes. It may also have other uses yet to be discovered.

Another reason bacteria maintain CRISPR is that it is still effective against viruses with anti-CRISPR in certain circumstances. The most important factor appears to be the relative size of the virus and host populations. The CRISPR system takes energy to run, but its big advantage is that it can respond rapidly. So, when bacteria are under attack from just a few viruses, CRISPR can eliminate them before they proliferate and activate their anti-CRISPR, saving the microbes from having to expend too much energy.

Never-ending battle

It probably took thousands of years for bacteria to evolve CRISPR. “It requires huge genetic innovation,” says Edze Westra at the University of Exeter, UK, who studies the evolutionary ecology of bacterial immunity. Yet, its evolutionary future is uncertain. All we can be sure of is that, in the arms race for survival, bacteria will continue to evolve innovative defences against viruses and viruses will evolve ways to fight back.