It was Victoria Gray’s first time in London and, despite a sleepless plane ride across the Atlantic Ocean, she wasn’t about to skip sightseeing. While crossing Trafalgar Square, Gray paused briefly to reflect on her experience. “I would never have been able to walk this long before,” she told a NPR reporter. “I feel like I got a second chance.”

Four years earlier, in 2019, Gray had become the first patient with sickle cell anemia — a genetic disorder that causes red blood cells to become sticky and rigid — to receive an experimental treatment using CRISPR genome editing. The treatment, now known as Casgevy, became the first CRISPR-based therapy to gain FDA approval, in 2023. Gray, in London to discuss the significance of her recovery at the Third International Summit on Human Genome Editing, described Casgevy as “a new beginning for people with sickle cell disease.”

Despite its association with genome editing,¹ CRISPR didn’t start out as a tool for fighting genetic disease. Instead, for billions of years, bacteria have used CRISPR systems to defend against invasion by viruses known as bacteriophages. Certain CRISPR components can add short DNA sequences from the genomes of defeated viruses into the bacterium’s own genome, creating a type of protective “memory.” These sequences, known as protospacers, can be found between short, repetitive DNA motifs — an observation that gave CRISPR, which stands for “clustered regularly interspaced short palindromic repeats,” its name.

Collectively, these repeat-protospacer regions are known as CRISPR arrays. The core of the CRISPR immune response is a guide RNA (gRNA) that binds to a CRISPR-associated (Cas) protein. Taken together, these components form what is known as a ribonucleoprotein (RNP) complex. If the same virus invades the cell a second time, the gRNA’s spacer sequence will bind to the matching viral DNA sequence, then be cut by the Cas protein. This allows bacteria to remember past viral infections and fight them off without mistakenly cutting their own genomes.

Since the early 2010s, biologists have been exploiting CRISPR’s ability to cut a diverse set of specific DNA sequences by altering gRNA sequences. Notably, this has led to the development of new medicines to treat genetic diseases — Casgevy was the first of these to gain FDA approval and is used to treat two blood disorders, called sickle cell disease and beta thalassemia. Dozens more clinical trials, based upon similar gene-editing technologies, are now underway. This article provides a summary of the major CRISPR systems,² including the naturally occurring CRISPR-Cas9, -Cas12, and -Cas13 systems, as well as base editors, prime editors, and the recently uncovered bridge RNA system.

Click here to download a large poster showing all of these genome editing systems.

{{signup}}

Cas9

Discovery

Cas9 became the first CRISPR effector (the protein that cuts or modifies DNA) engineered for genome editing in 2012, when Jennifer Doudna’s group at UC Berkeley and Emmanuelle Charpentier’s group at Umeå University showed that Cas9 could be guided to virtually any DNA sequence, not just natural targets from invading viruses, by tinkering with its gRNA sequence.³ Given the system’s bacterial origin, however, it was not guaranteed to work well in other organisms. Within a year, though, Feng Zhang’s lab at MIT showed that Cas9 worked in human cells, a result that other groups quickly confirmed in a flurry of papers.

The best-known Cas9 protein comes from a bacterium, Streptococcus pyogenes, that causes pharyngitis in young children. The protein is so good at cutting DNA that it still ranks among the most efficient Cas9 proteins and is almost certainly the most widely used genome editor today. Researchers have since discovered variants of the protein with additional useful qualities.

For example, NmeCas9 (discovered in a pathogenic microbe, Neisseria meningitidis) makes off-target DNA cuts less often than the Cas9 from S. pyogenes, making it more specific. GeoCas9 (taken from a thermophilic bacterium, Geobacillus stearothermophilus), on the other hand, is able to function and cut DNA at much higher temperatures. Researchers have also discovered a slew of small Cas9s — some less than half the size of the S. pyogenes protein — whose compactness makes them far easier to package into viral vectors and deliver into the human body.

Mechanism

Before Cas9 can cut DNA, it must grab onto its gRNA molecule in order to form an active RNP complex. In wild microbes, Cas9 cuts DNA using two separate RNA molecules, called the CRISPR RNA (crRNA) and trans-activating RNA (tracrRNA). (One of the great innovations reported in the Doudna and Charpentier labs’ 2012 paper was showing that these two RNA segments could be combined into a single guide RNA, a simplification making the gene-editing tool much easier to work with.) After the crRNA binds to the Cas9 protein, it binds the tracrRNA to create an intermediate, which is cleaved by the host cell housekeeping enzyme RNase III. This mature crRNA contains components from both of the original RNA molecules and guides the CRISPR-Cas9 complex to its target sequence.

The RNP complex scans DNA for a short sequence, typically two to four nucleotides long, called a protospacer-adjacent motif (PAM).⁴ Once it finds this PAM, the complex checks the nearby DNA against its spacer region in the gRNA. If the two sequences match, Cas9 uses two catalytic domains to cut both DNA strands, creating a double-strand break. The PAM-binding requirement reduces the amount of time that the complex spends searching for its target, allowing the enzyme to quickly find and cut a desired sequence. After binding to a DNA sequence, the “target strand” containing the PAM is cleaved by the HNH protein domain, whereas the “non-target strand” is cut by a second domain in the Cas9 protein, called RuvC.

After the DNA is cut, a cell repairs the damage in one of two ways. Scientists can sneakily co-opt these natural repair processes to insert or modify DNA within the organism’s genome. In one repair pathway, non-homologous end joining (NHEJ), the cut is pasted directly back together by cellular machinery. Since the nucleotides adjacent to the cut may be damaged by DNA cleavage, bases are often removed to facilitate efficient ligation. By cutting DNA with CRISPR, researchers can therefore scramble the genetic code in a way that disables the targeted gene.

Another pathway, homology-directed repair (HDR), can occur if the cell has another piece of DNA with a similar sequence. In nature, this allows organisms to precisely repair damaged genes to match other intact copies. By cutting DNA and also providing a template for repair, however, biologists can use HDR to insert specific sequences at the cut site. Both of these repair processes occur probabilistically, meaning that scientists cannot always control which of the two outcomes happens. However, researchers have discovered various ways to bias cells towards higher rates of HDR, either by suppressing NHEJ or by designing more efficient donor templates.

Off-target activity

As remarkable as it is that CRISPR-Cas9 can cut DNA programmably, it’s not a perfect system. In fact, one of the largest safety concerns in gene-editing is “off-target activity,” which occurs when cuts are made at unintended sequences that don’t precisely match the gRNA spacer. This happens because the complex can sometimes cut DNA even after binding a target region that has one or more mismatches with the spacer; the complex registers the sequence as “close enough” due to the physics of the reaction. Such off-target cuts are obviously a big concern when developing human therapies, because they can cause mutations that lead to cancer or other diseases.

Researchers have engineered more specific enzymes — possessing drastically reduced off-target effects — by altering either the structure of the Cas9 protein itself or its gRNA. For example, a classic study from Keith Joung’s lab at Harvard Medical School showed that shortening the gRNA spacer by 2-3 nucleotides heavily reduced off-targets as it destabilized, and thus decreased, mismatched binding. There is usually a trade-off in such engineering feats, though, as more specific Cas9 variants also tend to be slower or exhibit lower on-target editing efficiencies.

Tools

Because CRISPR-Cas9 can be “programmed” to target nearly any desired DNA sequence, researchers have repurposed the tool in all sorts of creative ways to do other types of gene modifications. One approach, for example, involves fusing catalytically deactivated versions of Cas9 (meaning the enzyme can still bind to target DNA, but cannot cut it) to other proteins with useful properties.

For example, one can fuse a deactivated Cas9 protein to so-called “transcriptional regulators” to either activate or repress genes without directly editing the genome at all. Researchers have also created a broad spectrum of engineered Cas9 proteins with improved functionality, especially increased precision and specificity. Similarly, they have identified Cas9 variants that recognize shorter PAM sequences in order to increase the enzyme’s genomic targeting space.

Therapeutics

Cas9 was the basis of Casgevy, the first-ever CRISPR medicine approved by the U.S. F.D.A. in late 2023 to treat the genetic blood disorders sickle cell anemia and beta thalassemia. With Casgevy, patients’ bone marrow cells are removed, edited in the lab, then reintroduced to produce healthier blood cells.

Cas9 is also being used in a number of ongoing clinical trials for everything from HIV to protein folding disease and lymphoma. Cas9 also featured in the first in vivo CRISPR clinical trials, which edit patients’ non-germline cells directly inside the body. These include Intellia Therapeutics’ trial for the heart disease transthyretin amyloid cardiomyopathy, and Editas Medicine and Allergan’s partnered trial to treat LCA10, a form of blindness.

Related effectors

In recent years, researchers have identified evolutionary precursors of Cas9, the ancient proteins from which Cas9 likely evolved. Cas9 is believed to have stemmed from a protein called IscB, which in turn evolved from an even older enzyme called IsrB. Both IscB and IsrB are RNA-guided enzymes — just like Cas9 — but are associated with transposons, or small DNA elements that “cut and paste” themselves into genomes. Although the full roles of IscB and IsrB are not known, they are thought to help these transposons spread more easily.

Cas12

Discovery

Cas12 was initially characterized in 2015 by Feng Zhang’s lab at the Broad Institute⁵ and was quickly adopted as a complementary genome editor to Cas9. The Cas12 variant in the Zhang lab’s paper — now known as Cas12a — has a DNA cleavage mechanism quite similar to Cas9, but subsequent work has revealed that the broader Cas12 family is the most biochemically diverse set of CRISPR proteins. Since 2015, researchers have discovered many Cas12 subfamilies — or groups with common evolutionary origins — including the ancestral Cas12f, the virus-encoded CasLambda and CasΦ, and the RNA-targeting Cas12a2 and Cas12g.

Mechanism

Cas12 forms an operational RNP complex just by association with its crRNA alone — no tracrRNA required. It searches DNA for its PAM, then checks the nearby DNA against its crRNA. Upon recognizing a target, Cas12 cleaves both DNA strands at different positions using a single RuvC domain, leaving behind “sticky ends,” or overhangs. These cuts can then be repaired using either NHEJ or HDR. Because the cuts are staggered, though, HDR naturally occurs at higher rates compared to the “blunt,” double-strand breaks made with Cas9. This allows for more precise insertions and, generally, higher specificity.

Off-target activity

While Cas12 can have off-target effects, it is generally considered more specific than Cas9 because it is less tolerant of mismatches in its target DNA sequence. When Cas12 encounters a mismatch between its spacer and its DNA target, it’s able to detach itself more easily than Cas9, giving Cas12 enhanced backtracking ability. Additionally, when Cas12 binds its DNA target, the RuvC domain unleashes a large amount of collateral cleavage, cutting ssDNA in its vicinity regardless of sequence. If Cas12’s on-target cleavage is a pair of molecular scissors, its off-target activity is more like a paper shredder.

Tools

Cas12 engineering efforts have primarily focused on improving its genome editing activity while maintaining its high specificity, which typically involves introducing mutations to stabilize target DNA binding. Researchers have also further increased Cas12’s specificity and altered it to target new PAM sequences. Perhaps the best-known application of the system, however, is not a genome editing tool at all: Cas12 has been used in diagnostic tools that detect specific nucleic acids. When developing a SARS-CoV-2 diagnostic, for example, researchers used special, single-stranded DNA molecules that emit fluorescent light when cleaved. Upon binding to even a small amount of target dsDNA, Cas12 begins cutting the ssDNA to emit fluorescent light. The system’s high level of collateral cleavage acts as a signal amplifier, allowing highly specific nucleic acid detection.

Therapeutics

There are no FDA-approved therapies built upon Cas12, but the enzyme is currently being used in multiple clinical trials. For instance, there are ongoing trials for Cas12-based medicines aimed at both diseases currently treated by Casgevy. Editas Medicine’s sickle cell anemia treatment was the first Cas12 medicine to begin clinical trials, while Shanghai-based Vitalgen Biopharma recently began a phase II trial using Cas12b to treat beta thalassemia.

Related effectors

Similar to Cas9, Cas12 has a smaller transposon-encoded ancestor — an RNA-guided nuclease called TnpB. This protein is also capable of programmable genome editing, albeit at much lower levels than Cas12a. TnpB is predicted to assist its transposon in moving around its host genome and is found in an even broader range of transposons and bacteria than IscB. A specific lineage of TnpB also gave rise to Fanzor, the first genome editor found in a eukaryotic organism. (Fanzors have been identified in everything from clams and algae to butterflies.)

Cas13

Discovery

Cas13 binds to and cuts RNA instead of DNA. It was first characterized in 2016 by Feng Zhang’s lab.⁶ Since then, three more principal Cas13 subfamilies have been discovered, as well as multiple potential evolutionary intermediates.

Mechanism

Like Cas12, Cas13 requires only a crRNA to form an active RNP complex. After complex formation, it searches for a motif similar to a PAM, called a protospacer flanking sequence (PFS). PFSs are even shorter than PAMs, typically consisting of only a single nucleotide.⁷ After binding to its PFS, Cas13 identifies its target site and cleaves the single-stranded RNA molecule using two protein domains, called HEPNs, in tandem.

Off-target effects

Researchers have published conflicting findings on Cas13’s levels of off-target activity, and it remains one of the great enigmas of the CRISPR field. Like Cas12, Cas13 displays collateral cleavage, arbitrarily cutting RNA molecules in its vicinity after binding its target sequence. However, these cuts can’t easily be differentiated from off-target cuts at sites where Cas13 binds a mismatched spacer since both of the protein’s modes of activity target ssRNA. For this reason, Cas13 off-target activity is usually reported as the sum of the two. Despite known collateral activity in test tubes and bacteria, Cas13 initially showed few to no off-targets in human cells. Since high levels of nonspecific RNA-cutting would be toxic to living cells, Cas13 had skirted a key barrier to therapeutic development. However, later work found that Cas13 did have off-target activity and cause toxicity in mammalian cells, creating an inconsistency yet to be resolved.

Tools

Having seen the utility of Cas9- and Cas12-based tools for editing genes and regulating their expression, researchers began developing tools based on Cas13 to accomplish similar tasks targeting RNA. Feng Zhang’s lab created an RNA base editing system by fusing Cas13 to the ADAR2 deaminase. This enzyme changes the identity of RNA bases by chemically removing amino groups, allowing biologists to make specific changes to RNA transcripts, rather than simply cutting them. The Cas13 is first “guided” to its target RNA sequence, and then the ADAR2 deaminase modifies a specific nucleotide to create a new sequence.

Researchers in the Cress and Doudna labs also created a tool, similar to Cas9-based transcriptional repressors, using a catalytically deactivated Cas13. To do this, they mutated the amino acids used by Cas13 to cut RNA, allowing it to simply bind its targeted transcript, blocking the RNA strand from being translated into protein. Cas13 has also been a crucial component of nucleic acid detection systems, which have been further developed for rapid detection of viruses, including SARS-CoV-2.

Therapeutics

The first Cas13-based clinical trial was approved in 2024. It will test a treatment developed by the Shanghai-based company HuidaGene Therapeutics for age-related macular degeneration, which can cause blindness. The disease is caused by the accumulation of vascular endothelial growth factor (VEGF), a protein involved in the formation of blood vessels. The treatment uses Cas13 to destroy the VEGF messenger RNA and prevent harmful buildups of its encoded protein.

Related effectors

Cas13 likely evolved from a non-CRISPR bacterial immune system called Abi. Abi functions quite differently from CRISPR and cannot be reprogrammed for genome editing.

{{signup}}

Base editors

Creation

Base editors are engineered CRISPR effectors that can change individual nucleotides without creating double-strand breaks. Base editors are built by fusing a Cas9 protein to a deaminase enzyme, which chemically removes atoms from DNA bases to change their identities. Importantly, the Cas9 proteins used to make base editors are not “normal.” They are, instead, “nickases” made by mutating Cas9’s DNA-cutting amino acids to partially impair their function. In 2016, David Liu’s lab at Harvard University developed a cytidine base editor (CBE), capable of converting C nucleotides into T, and G nucleotides into A, by fusing a Cas9 nickase to a cytidine deaminase enzyme.

The following year, Liu’s lab created an adenine base editor (ABE) using a different deaminase, called TadA, thus allowing the opposite conversions. While the original base editors were created using modified Cas9 proteins, new versions have also been developed using Cas12, TnpB, and IscB. Cas12 base editors allow targeting of PAMs inaccessible using Cas9 base editors, while TnpB and IscB base editors are smaller and therefore advantageous for therapeutic delivery purposes.

Mechanism

The Cas9 nickase binds its PAM and target DNA sequence in the same way as conventional Cas9. But before DNA cleavage can occur, the deaminase chemically modifies its target base by catalyzing the removal of an amino group. This changes the DNA sequence by converting the original base into a form that will be recognized as a different nucleotide by cellular transcription and replication enzymes. After this, the Cas9 nicks the DNA on the strand opposite the modified base, creating a break in just one strand of the DNA and triggering a special type of DNA repair known as “mismatch repair.” In the case of a successful edit, the deaminated base is maintained while the base opposite is cut out and replaced with a new, matching nucleotide.

Off-target activity

Because they are fusion proteins, base editors have multiple types of off-target activity caused by the Cas9 and deaminase components of the system, respectively. There are two types of Cas9-based off-target activity: modification of extra bases adjacent to the target base and base editing at target sites containing mismatches with the spacer sequence. For the latter type, one study from David Liu’s group found a maximum of 15 percent off-target editing, which was reduced to 5 percent by using an engineered (high-fidelity) Cas9 as the base for the nickase. Deaminase-dependent off-target activity involves deamination of random DNA and RNA bases throughout the genome.

Therapeutics

Verve Therapeutics began the first clinical trial using base editors in 2022, with the goal of treating genetic contributors to heart disease. Currently in phase 1b, their trial uses base editors to correct genetic variants of the PCSK9 gene that increase harmful cholesterol levels. Beam Therapeutics also has an ongoing phase 1/2 trial to treat sickle cell anemia by editing one of the same genes targeted by Casgevy. GenAssist Therapeutics dosed their first patient in a base editing clinical trial for Duchenne Muscular Dystrophy, or DMD, in September 2024.

In February 2025, base editors were used in the first-ever personalized CRISPR medicine, in which a five month-old infant was treated for a severe disease called carbamoyl-phosphate synthetase 1 deficiency. The disorder is rare but frequently leads to seizures, comas, and death. Researchers across several institutes worked together to develop the custom base editing treatment, which was dosed after being approved by the FDA.

Prime editors

Creation

Prime editors consist of a Cas9 nickase fused to a reverse transcriptase. Originally developed by David Liu’s group in 2019, prime editors are designed to create larger and more diverse types of genome edits than any tool described thus far. Prime editors are able to “search and replace” within DNA strands, substituting longer, custom-designed DNA sequences into a defined target site via the use of a special gRNA called a prime editing guide RNA (pegRNA). In the last five years, researchers have engineered prime editors capable of cutting out or replacing chunks of DNA up to 800 base pairs in length, such as by using two pegRNAs at once.

Mechanism

Prime editing begins, like so many other CRISPR genome editors, with a Cas9 protein searching for its PAM and target DNA. After the prime editor binds to DNA, the Cas9 nickase slices one strand of the DNA, which is then bound by the pegRNA. The pegRNA contains a template region, which is transcribed from RNA to DNA by the reverse transcriptase, starting at the end of the existing DNA sequence. For the edit to be kept, this new “flap” must bind to the other DNA strand, and the original sequence must be cleaved off. Finally, the DNA must undergo mismatch repair, meaning the cell uses the new prime edited DNA as its repair template. Once the process is completed, the DNA strand bears the reverse transcribed sequence in place of the original target.

Off-target activity

Prime editing can cause both off-target deletions due to aberrant nickase activity (meaning the protein cuts in the wrong place in the genome) and insertions at mismatched target sites. However, the overall frequency of prime editing off-targets is thought to be very low because three distinct DNA hybridization steps must take place. A 2023 study found that individual off-target insertions caused by prime editing happen at frequencies of less than 1 percent each. Researchers have been able to lower this even further by engineering more precise Cas9 nickase variants.

Therapeutics

Prime Medicine, a spinout from Liu’s laboratory, secured approval for the first prime editing-based clinical trial in mid-2024, and planned to test a drug candidate for a rare immune disorder called chronic granulomatous disease. The disease can be caused by multiple genetic mutations, all of which impair cells’ ability to produce infection-fighting oxygen molecules and leave patients vulnerable to severe infections. The treatment sought to correct one such gene, NCF1, by using prime editing to add missing nucleotides into the genome and thus eliminate the disease. In May 2025, however, Prime Medicine announced that they would shift their priorities to other diseases instead, including cystic fibrosis.

Bridge RNA

Discovery

Like prime editing, the bridge RNA (bRNA) system allows large-scale DNA insertions in genomes. It was first characterized by Patrick Hsu’s lab at the Arc Institute in 2024. Naturally encoded by transposons, bRNAs are not CRISPR systems but are likewise guided by RNA, giving them a similar underlying logic for genome editing. The system consists of the bRNA molecule, which functions as a guide, and a recombinase enzyme, which can exchange large chunks of DNA.

Mechanism

The active bRNA RNP complex consists of two recombinases and the bRNA molecule itself. Recombinases are proteins that exchange regions of DNA found between specific sequence motifs — in the case of the bRNA system, these motifs are dictated by the RNA sequence. The bRNA has two loops, which bind both the DNA to be inserted and the specific point where the insertion will occur; one recombinase associates with each of these loops. Both the target and donor DNA strands are cleaved by a composite RuvC active site, which forms at the interface of the recombinase dimer. The cleaved DNA ends are switched, creating an X-shaped intermediate called a Holliday junction. This intermediate is resolved by cleavage of the other ends of the two DNA strands, which are also switched to complete the recombination process. The bRNA system therefore allows genome editing without having to interact with cells’ natural DNA repair processes — and that’s a big deal because, theoretically, it means scientists can make their edits more precisely.

Off-target activity

The bRNA system results in a moderate number of off-target insertions at mismatched target sites, as well as a very small number of genomic deletions and inversions. In bacterial cells, 69 percent of insertions occurred at target sites. Hsu’s lab was able to increase this to 85 percent by elongating the section of the bRNA associating with the target DNA.

Future directions

The bRNA system is still in development, and was recently adapted to work in human cells. There are currently no therapeutics or tools based on the technology. In addition to insertions, the system’s recombination mechanism allows researchers to perform specific DNA inversions and excisions. Future Bridge RNA-based therapeutics could therefore hold promise for treating diseases requiring replacement or insertion of large DNA sequences.

Related effectors

Recently, Feng Zhang’s lab discovered a new class of genome editors that’s essentially a cross between the bRNA and CRISPR-Cas. Called TIGR-Tas, the system consists of a Tas nuclease and a unique gRNA called a tigRNA. TIGR-Tas systems can be programmed for genome editing in human cells, though only at very low levels. Tas proteins all have a structural core, called a Nop domain — also found in the bRNA recombinase — which allows them to bind RNA. This Nop domain can be accompanied by either an HNH or RuvC — the same domains that Cas9 and Cas12 use to cut DNA.

Unlike CRISPR-Cas, however, TIGR-Tas systems don’t require a PAM for cleavage. Additionally, tigRNAs have not one but two spacers, each of which binds a different strand of the dsDNA target. The sequences recognized by these spacers are right next to each other and form a composite target site, which the Tas enzyme then cuts. The Zhang lab predicts that the dual spacer architecture prevents the system from cutting its own TIGR arrays in the absence of a PAM requirement.

Conclusion

In the decade-plus since it was first engineered to serve as a programmable genome editing tool, CRISPR and its relatives’ versatility, diversity, and precision have expanded tremendously. Casgevy’s approval marked CRISPR’s entry into the clinic, while the first CRISPR genome-edited food in the U.S. — mustard greens edited to remove their bitter taste — hit grocery store shelves in 2023. Companies are also using CRISPR technologies to develop blackberries with no seeds, bananas that don’t turn brown, and tomatoes with increased levels of vitamin D.

In addition to their growing role in commercial agriculture, CRISPR technologies are being directed towards global sustainability efforts. Researchers have used CRISPR to create poplar trees with reduced lignin, a key structural component that also limits fiber processing. By editing the trees with CRISPR, the scientists were able to decrease lignin content while increasing the levels of wood carbohydrates to maintain structural integrity. Fiber can be obtained more efficiently from the edited trees, resulting in both higher yields and a decreased environmental footprint. The study increased the ratio of wood carbohydrates to lignin more than 200%, paving the way for field trials of the edited trees. Scientists have similarly used genome editing to create crops requiring less water or other non-renewable resources.

Despite Casgevy’s success in treating sickle cell, it remains a costly and labor-intensive treatment. Since the therapy requires that patients’ cells be edited outside their bodies and reinserted, it costs millions of dollars per patient and requires them to undergo chemotherapy and bone marrow transplants to allow the edited cells to proliferate. Companies have begun clinical trials where cells are edited directly in patients’ bodies, but this approach is even more difficult than external editing as it requires yet lower levels of off-target activity.

This in vivo editing strategy also requires genome editors to be packaged so that they can be delivered to the correct cells in a desired tissue or organ. Traditional approaches to this problem involve the use of viral vectors, which run the risk of triggering immune responses in patients. To combat this, scientists have developed new delivery strategies using small particles made of fat molecules — called lipid nanoparticles — as well as vehicles derived from cell membrane compounds — known as enveloped delivery vehicles. For example, Intellia’s in vivo heart disease clinical trial uses lipid nanoparticles to deliver the genome editing machinery to patients’ cells.

If the genome editing toolbox of 2012 contained only a single hammer, today’s contains a miscellany of hammers, wrenches, pliers, and screwdrivers. The diverse set of enzymes available means that translational researchers can deploy the one best suited for the disease they want to treat and optimize it. In doing so, they push the tools to be more precise, safer, and easier to use.

However, this toolbox is by no means complete. Biologists continue to discover new genome editors in nature with advantageous properties and, in the case of bridge RNA, entirely new types of chemistry. It’s quite likely that yet more systems exist elsewhere in nature, awaiting our attention.

{{divider}}

Evan DeTurk studies CRISPR-Cas genome editing and protein design at UC Berkeley. He earned his A.B. in Molecular Biology from Princeton University and also writes about science fact and fiction on Substack.

Thanks to Tong Wu for reading drafts of this, as well as Kenneth Loi, Peter Yoon, and Matthew Kan for discussions of the material. Artwork by Ella Watkins-Dulaney.

Cite: DeTurk, E. “A Visual Guide to Genome Editors.” Asimov Press (2025). https://doi.org/10.62211/15yw-55jh

{{divider}}

Footnotes

1. This field goes by a few names, including “genome editing,” “gene editing,” and “genome engineering,” all of which are used basically interchangeably. (“Genome design” has also recently entered the conversation.) For consistency, “genome editing” is used throughout this article.
2. Describing all CRISPR systems in even a cursory manner would require a much longer article. The genome editors described here are meant to facilitate a working knowledge of the genome editing field as it stands today. Notably, this article only covers class 2 CRISPR-Cas systems, which have only one protein per RNP complex. Class 1 CRISPR-Cas RNPs contain multiple proteins each and have also been utilized for genome editing, though not as thoroughly as class 2 systems.
3. The duo later shared the 2020 Nobel Prize in Chemistry for their discovery.
4. The PAM is crucial to CRISPR-Cas’s functionality as an immune system because it helps the bacterium differentiate its own DNA from that of invading viruses. If the spacer were all that was required for cleavage, CRISPR-Cas systems would target their own CRISPR arrays, but the PAM requirement prevents this.
5. At the time of this initial characterization, Cas12 was known as Cpf1. In 2017, Eugene Koonin’s group at the National Center for Biotechnology Information proposed a new nomenclature for CRISPR-Cas systems. Koonin’s group has been the prime mover in standardizing CRISPR-Cas naming conventions since they released their first classification in 2011.
6. Like Cas12, Cas13 initially had a different name (C2c2) but was renamed in the Koonin group’s 2017 article.
7. It’s worth noting that Cas13 targeting requirements are more flexible than Cas9 or Cas12. While PAM recognition is required for all DNA-targeting enzymes in those groups, some Cas13s don’t require a PFS at all.