— June 2017 —
I’ve done a good bit of research on CRISPR, though I am certainly no expert.
I thought I would lay out what I know here, for anyone who might want a place to start on the topic, or a reference.
Hope you enjoy, I’ll do my best to link to sources.
Table of Contents
- The Enzymatic Mechanism
- Applications and Limitations
- 2017 Patent Dispute
- Ethical Concerns
- With Great Power…
CRISPR (short for Clustered Regularly Interspaced Short Palindromic Repeats) is the name given to small, patterned segments of the genome of bacteria and archaea (collectively, prokaryotes) that, together with CRISPR-associated (“Cas”) proteins and RNA “tags”, allow for the generation of highly-customizable, highly-efficient, targeted proteins that splice DNA/RNA.
This system was discovered by accident in the late 1980s, and was later demonstrated to function as a quasi-“immune system” for prokaryotic organisms:
- When attacked by a virus or plasmid (nefarious snippet of foreign DNA), short segments of the viral DNA are stored between sets of recognizable palindromic nucleotide markers in the organism’s genome;
- Specialized enzymes transcribe RNA-segments from the stored viral DNA and use them as tags to recognize and cleave matching segments of invader DNA when the next attack occurs.
These snippets of viral DNA and the associated Cas family of proteins act as a dynamic, inheritable defense mechanism against phages and plasmids.
Since their initial discovery, the Cas9 protein system and associated RNAs have been found to act as a simple, cheap and highly efficient sequence-specific DNA cleavage system in all organisms in which it has been tested to date. Setting aside its significance as a biological system of study, as an engineering tool, the CRISPR-Cas9 system has emerged as the most promising, powerful and potentially harrowing genetic editing/engineering paradigm ever discovered.
There are a few things I suggest mulling over as we slog through the details of this system and their discovery for a while…
This story is remarkable for two main reasons:
(1) prokaryotes have an immune system: the simplest organisms we know of — organisms whose population is higher in our bodies than that of our own cells — have a robust and dynamic nucleotide-based defense mechanism against viruses and plasmids that we have discovered only in in the past 30 years; In fact, there wasn’t even a guess as to the function of this region of their genome until 1995;
(2) the defense mechanism we’ve discovered is evidence of Lamarckian (read: non-Darwinian) evolution*: this is an adaptive mechanism which a prokaryote uses to edit its own genome directly, which of course allows the immunity to be inheritable, but also results in a change to the organism’s genome and its expression within it’s own life cycle, and passes that trait on; this is no different (on a prokaryotic scale) from the classic case of a giraffe growing it’s neck out to reach fruit on a tree. These organisms are acquiring traits in their lifetimes that they did not have when they were born, and passing them off to progeny.
If there has to be only one reason why a lay person should care about this story, it’s the latter. Prokaryotic organisms existed long before humans, have participated in and shaped our evolution as a species, and continue to be important to our understanding of our lives and our well-being — and our model for the way in which they evolve was, until very recently, at least partially (if not totally) wrong.
There are certainly more pressing and dramatic reasons why this story is important, especially with regard to the near future, but they are mostly ethical, practical, and/or legal — aspects which are probably best left to someone else. I do mention the 2017 Patent Dispute and Ethical Concerns briefly toward the end.
We’re going to get in to the weeds here a little bit, but my goal is to keep this as readable as possible, without sacrificing accuracy.
(This story is peppered with the names of scientists and their collaborators. I wavered on whether to include them as footnotes, or to include them inline, but I decided that this is their story, so they shouldn’t be left out of it, or relegated to fine print).
Here’s the story of how we discovered the CRISPR-Cas system, and how it works:
In 1987, Yoshizumi Ishino et al. conducted a gene-sequence and gene-product analysis experiment to investigate a particular region of the E. coli genome, named the iap gene locus, which was thought to code for an enzyme (called a “gene product”) involved in the maturation/processing of another important set of proteins. In the course of analyzing the sequence and suggesting a likely gene product, they found an unusual nucleotide structure in the 3’ flanking region of the locus:
Five highly homologous sequences of 29 nucleotides were arranged as direct repeats with 32 nucleotides as spacing…. So far, no sequence homologous to these has been found elsewhere in procaryotes, and the biological significance of these sequences is not known.
A few years later, Nakata et al. were able to identity many instances of the same 29 base pair repeat with 32/33 base pair spacer sequences downstream of the iap gene, and elsewhere in the E. coli genome, in addition to the genomes of two other bacterial organisms (in the Shigella and Salmonella genera). In 1995, Mojica et al. identified the same structure in two species of Archaea, and suggested that it is involved in replicon (prokaryotic chromosome) partitioning — essentially positing the structure acts like “borders” between gene loci. Five years later, Mojica (with different collaborators) published a MicroCorrespondence in the Journal of Molecular Microbiology highlighting the significance of this structure, suggesting it may be essential to the genome of all prokaryotes.
The first published usage of the mnemonic “CRISPR” came from a paper by Jansen et al. in 2002, which also identified four CRISPR-associated (cas) genes located right next to CRISPR loci (many organisms have multiple separate CRISPR gene regions), and suggested that the two gene loci have a functional relationship. In addition, the CRISPR-cas loci were found to be largely conserved within a species but not conserved between species. The cas genes showed motifs consistent with proteins involved in DNA metabolism (processing) and gene expression (turning on/off transcription).
Bolotin et al., in 2005, discovered that the spacers between CRISPRs were of extrachromosomal origin (taken from another genome) due to homology with phage gene sequences. They suggested (presciently):
[T]he spacer elements are the traces of past invasions by extrachromosomal elements… [and] provide the cell immunity against phage infection, and more generally foreign DNA expression, by coding an antisense RNA.
Pourcel et al. echoed this suggestion in a paper published in the same volume that experimentally recorded the addition of a new spacer to the CRISPR locus, and confirmed that the new spacer was taken from prophage (attacker) DNA.
The following year, Makarova et al., working out of the National Institutes of Health, published an exhaustive sequence analysis of the cas gene locus, allowing for the classification of 6 distinct superfamilies of genes therein (cas1-6). In addition, they made sequence-similarity arguments hypothesizing the activity of cas-derived proteins, and gave the first rough sketches of the hypothetical workings of the CRISPR-Cas system (which they dub “CASS”) as a defense mechanism against invading phages, and the enzymatic system for incorporation of new CRISPR units into the locus. They also demonstrated that new inserts to the locus showed almost no similarity with gene loci from closely related strains, which suggested rapid (non-evolutionary-time) turnover of the genes. Seeing the potential of the system, they remarked,
“it seems that, once the psiRNA [“prokaryotic silencing RNA”] mechanism described here is investigated experimentally, it could be exploited to silence any gene in organisms that encode CASS.”
The first comprehensive, direct experimental demonstration that the CRISPR-cas locus is responsible for inheritable adaptive phage immunity came in 2007 (just ten years ago), in a short, landmark paper by Barrangou et al., published in Science. This paper definitively showed that CRISPR spacers are responsible for targeted phage resistance (both inherited and acquired), and that cas-derived proteins are the enzymatic machinery behind phage DNA neutralization. (N.B. in the early months of 2017, this paper has been cited nearly two thousand times).
It is important to note that at this time the enzymatic mechanism responsible for phage resistance was still largely a “black box”. Research up to this point was able to demonstrate that the CRISPR-cas system acts as a nucleic acid-based phage-resistance/“immunity” system, but the mechanics of the system were still largely unknown — the real power of the system (the Cas enzymatic machinery, and its applicability in other organisms) was yet to be discovered…
The Enzymatic Mechanism
The Early Research
In the following years, the nucleic acid-/enzyme-level mechanism behind CRISPR-Cas-mediated phage immunity was pieced together.
Marraffini and Sontheimer published work demonstrating that Cas-protein interference works by targeting DNA directly (in contrast with RNA interference, or RNAi, mechanisms that had been previously observed in eukaryotic organisms) by showing that short sequences of RNA transcribed from the CRISPR locus (dubbed “crRNA”) were responsible for the targeting of particular phage DNA segments.
They were also able to show in a later paper that elements of the spacer region flanking the processed crRNA spacer region are responsible for self versus non-self recognition. Work by Deveau et al. uncovered a small, conserved sequence of nucleotides flanking the proto-spacer (the name given to the segment of phage DNA complementary to the CRISPR spacer), later termed the protospacer-adjacent motif (PAM), and determined that, in addition to allowing for self versus non-self discrimination as shown previously, it also is necessary for recognition by the Cas enzymatic complex (confirming suggestion put forward of Horvath et al. from the same year).
In a paper published in Cell in 2009, Hale et al. would provide the first experimental isolation and detailed study of a Cas effector complex in P. furriosis. They were the first to show that the effector complex is comprised of the spacer-derived RNA (psiRNA) and an attached, conserved 5’ spacer-derived “psi-tag”, together with a complex of Cas proteins; they also demonstrated that this particular Cas protein complex splices invader RNA strands complementary to a specific site, 14 base-pairs from the 3’ end of the psi-RNA. In a 2010 paper in Nature, Garneau et al. would demonstrate that this mechanism works on both bacteriophage and plasmid DNA, and that the mechanism cleaves double-stranded DNA as well. Taken together, these results suggested that the CRISPR-Cas system could be used as a targeted cleavage mechanism for essentially any nucleic acid sequence. In 2011, Deltcheva et al. discovered another CRISPR-associated RNA product (trans-activating CRISPR RNA, or “tracrRNA”, for short) which is complementary to the repeat region, and is important for the maturation of the crRNA before it can guide Cas proteins to the target invader sequence.
Sapranauskas et al., in 2011, were the first to demonstrate that the CRISPR3/Cas gene locus from S. thermophilus could be transferred to E. coli (in a different phylogenetic kingdom), and would provide protection against plasmid and phages. They also confirmed the result that the proto-spacer adjacent motif (PAM) is necessary for immunity, and demonstrated that Cas9 is the only protein necessary for CRISPR-mediated cleavage (in this particular CRISPR-Cas system).
Faced with the growing interest in the CRISPR-Cas system, several teams of researchers began publishing work intended to classify different variations of the system, and standardize nomenclature. In 2011, Makarova et al. examined the CRISPR/Cas system from an evolutionary point of view; this work is responsible for the delineation of the three main classes of CRISPR systems (types I, II, and III) and subtypes therein, which differ both in the way new phage DNA spacers are incorporated into the locus, and the way in which the crRNA is processed by the various Cas proteins before guiding the effector complex to the target; they also presented the taxonomic distribution of the three main systems across Archaea and Bacteria. As an aside, Dr. Kira S. Makarova appears to be one of the community’s foremost expert on the evolution of the CRISPR-cas locus.
The first team of researchers to demonstrate that the CRIPSR-Cas could be used as a broadly-applicable efficient, targeted genome editing tool was Jinek et el., in 2012. This paper signaled a major shift in focus, with respect to the interest in the CRISPR-Cas system from prokaryotic curiosity to powerful tool. However, before the methodology of Jinek et al. can be appreciated, it will be necessary to provide a rough sketch of the workings of the system — from spacer acquisition to phage-DNA-intervention.
The Gory Details
So… we really get in to the weeds here. If you don’t care that much about the specifics of how the CRISPR mechanism works from start to finish, skip to the next section
When a CRISPR-endowed prokaryote is challenged by a virus or plasmid, and foreign DNA/RNA is released into the cytoplasm, Cas1 and Cas2 (the only proteins conserved across all three types of CRISPR-Cas system) form a complex which binds to the invading strand, cleaves off a protospacer, and incorporates the new spacer in to the CRISPR locus. The protospacer is not randomly chosen by the Cas1-Cas2 complex: a 2-5 nucleotide protospacer adjacent motif (PAM) serves as a recognition signal for the initiation of protospacer cleavage; in type II CRISPR, one nucleotide from the PAM is inserted as the last nucleotide of the new repeat during incorporation.
During incorporation, the first repeat adjacent to the leader (promoter) sequence is copied, and the new spacer is inserted between the copies. It is known that Cas1 is responsible for the enzymatic cleavage of the invader sequence, and that the Cas1-Cas2 complex is uniquely responsible for recognition of the CRISPR locus leader sequence and new spacer acquisition, however, there are many aspects to the adaptation phase still unknown. It is also important to note that the acquisition phase is not identical across all CRISPR types, though the general model of insertion is generally quite similar.
crRNA biogenesis and Processing
When signaled (e.g. by up-regulation through increased cytosolic concentration of cAMP), the CRISPR-cas locus is transcribed, creating precursorRNA (pre-crRNA) transcript containing strands complementary to the entire repeat-spacer sequence. In Type I and III systems, processing (cleavage) of the pre-crRNA is carried out by Cas6: the pre-crRNA transcript is cut in the repeat sequences, leaving an 8-nucleotide 5’-handle, and a 3’-handle that varies in size, and sometimes folds back onto itself to form a hairpin (stem-loop) structure.
Type II systems contain a portion of the repeat sequence that is transcribed to form what is known as trans-activating RNA (tracrRNA). Tracr-RNA contains a sequence that is complementary to a portion of the repeat sequence of the pre-crRNA transcript; upon base-pairing with this region, the tracrRNA-pre-crRNA duplex is bound by Cas9, which recruits Ribonuclease III (a ubiquitous double-stranded RNA processing enzyme), which cleaves the pre-crRNA strand, leaving a mature crRNA strand complexed with Cas9.
Effector Complex Formation
In Type I and III CRISPR systems, the mature crRNA stays bound to Cas6, which recruits the other Cas proteins to form a CRISPR-associated Complex for Antiviral Defense, dubbed “Cascade”. In Type I systems, the Cascade complex is composed of one Cse1, two Cse2, one Cas5, six Cas7 and one Cas6 proteins.
The overall complex has a seahorse-shaped architecture, with the six Cas7 proteins forming a helix-shaped backbone, which binds, stabilizes and displays the crRNA. The Type III Cascade complex, though composed of different proteins, has a similar structure to the Type I complex. Indeed, the backbone of the complex is composed of multiple copies of a protein that binds the crRNA and has been shown to be structurally homologous to Cas7. However, they are not completely homologous. In fact, in addition to differences between the Types, the Type I and III systems themselves are divided into subtypes based on differences in the structure and function of the respective effector complexes: Types I-A, I-C, I-E, and I-F, as well as Types III-A and III-B.
In Type II systems, the effector complex consists only of the Cas9 protein, crRNA guide, and tracrRNA. The Cas9 protein has been shown to consist of an α-helical lobe, which coordinates the interaction of the crRNA-target interaction, and a nuclease lobe, which recognizes the PAM and cleaves the target DNA. Cas9 stabilizes the crRNA and exposes it for base-pairing with the complementary invader strand, but the crRNA also acts as an agonist for Cas9: upon binding of the crRNA, the two lobes undergo a conformational change that allows for DNA binding and cleavage.
Target Recognition and Interference
In all three CRISPR systems, recognition of the invader DNA/RNA proceeds in roughly the same manner:
- the invading nucleotide sequence is non-specifically bound and scanned;
- if the PAM sequence is recognized, the target DNA is unwound, forming an R-loop;
- there is an initial base-pairing event at a 7-12 nucleotide sequence known as the “seed” region, followed by extended pairing with the rest of the protospacer;
- once the entire protospacer is bound, the effector complex undergoes a conformational change that allows for cleavage of the invader sequence.
It is in this final step that the three types of systems differ.
In Type I systems, once the seed region has base paired to the crRNA guide, the invader DNA forms an R-loop, which triggers a conformational change in the Cascade complex and encourages the recruitment of Cas3, which is an ATP-dependent helicase and a metal-dependent nuclease. Upon binding of Cas3, the DNA strand is positioned by the Cascade complex for degradation.
In Type II systems, once the seed region has been paired, Cas9 (which has already adopted a nuclease-ready conformation, thanks to binding with the crRNA) uses its own intrinsic nuclease activity to cleave the invader DNA sequence. Cas9 has an HNH-like nuclease domain that cleaves the DNA strand base paired with the crRNA, and a RuvC-like nuclease domain that cleaves the other strand. In contrast with the Type I system, Type II DNA degradation is not ATP-dependent.
The mechanism by which Type III systems target invader polynucleotides is still largely unknown, however there are a few details known about the process to date. For one, Type III-A systems appear to discriminate self from non-self DNA in a manner that is PAM-independent. Type III-B systems, remarkably, are unique among known CRISPR systems in that they target RNA instead of DNA (though it has recently been purported that the differences in activity between Type III-A and -B systems has been resolved, and they are essentially structurally/functionally homologous). Beyond this, there is much still to be discovered with regard to Type III systems.
The Protospacer Adjacent Motif
Though it has been mentioned already, it is important to note the significance of the protospacer adjacent motif (PAM). A paper by Sternberg et al. in 2014 showed that for the Type II CRISPR system both binding and cutting by Cas9 require PAM recognition, and that off-target binding affinity scales with the density of PAM motifs. More remarkably, sequences that match the complementary guide-RNA, but lack the PAM are avoided by Cas9. They also showed that base-pairing with the target DNA starts at the PAM, and that PAM interactions trigger Cas9 nuclease activity. In summary, not only does the PAM allow for self versus non-self discrimination (the key component of any successful immune system), in the Type II CRISPR-Cas9 system it is arguably the single most important factor in target recognition and binding, DNA/RNA heteroduplex formation and nuclease activity.
Applications and Limitations
With knowledge of the details of the enzymatic machinery behind CRISPR-Cas interference, the landmark work of Jinek et al. can be better understood.
To begin with, it should be noted that Type II CRISPR systems have a variety of strengths compared to Type I and III systems, re: their use as enzymatic tools. For one, the Type II system is simpler, and has been studied in greater detail. More pragmatically though, Type II systems utilize only Cas9 (instead of an enormous multi-protein effector complex) which has intrinsic helicase and nuclease activity (in contrast to Type I and III systems which require the recruitment of other Cas proteins). Through the use of various assays, Jinek et. al were able to prove that tracrRNA is required for for target DNA recognition, and that PAM recognition is an important factor in initial DNA duplex unwinding, crRNA base-pairing, and R-loop formation.
In addition, and most remarkably, they were able to show that Cas9 can be programmed using a single engineered chimeric RNA strand that performs the role of both tracrRNA and crRNA. These chimeric-RNA-guided nucleases were shown to be efficient and reliable, as well as incredibly cheap to manufacture. Also, in contrast to existing gene editing tools such as TALENs or Zinc-Finger Nucleases, no modifications need to be made to the enzymatic protein itself — all that is required is the manufacture of a single RNA strand, and the presence of the (Guanine-Guanine) PAM in the target strand.
The CRISPR Craze was about to begin.
In subsequent years, various labs were able to use the Type II CRISPR/Cas9 system to demonstrate site-specific chromosomal DNA cleavage and gene activation in the genomes of various plant, insect, and mammal species, including the human genome.
Multiple labs were quickly able to use CRISPR-Cas systems to efficiently alter the genome of plants, such as wheat and rice. Kabin Xie and Yinong Yang were able to show that the PAM is present at a frequency of about one PAM for every 10 base-pairs in the rice genome, suggesting that more than 90% of rice genes are available for targeting by CRISPR-Cas9.
In 2014, Shan et al. were able to use the CRISPR-Cas9 system to demonstrate that specific genes in rice protoplasts could be targeted and mutated within 1-2 weeks, and mutated rice plants could be successfully grown in 13-17 weeks.
Bin Shen et al. were the first to publish results confirming that CRISPR-Cas9 could be used to induce site-specific cleavage of loci within the mouse genome. In the same year, Li Dali et al. would publish work confirming these results, and performing the same protocol with rats. This work would be followed up by a study in 2014 by Wang et al. who demonstrated the use of multiplex signal RNA to simultaneously disrupt five mouse embryonic stem cell genes with high efficiency – this allows for the one-step generation of mice with tailored mutations to multiple specific genes. This break through has the potential greatly accelerate the pace of drug discovery and clinical testing.
CRISPR-Cas9 could not only revolutionize existing gene-targeting animal models, but also allow for the creation of new ones. In 2014, Dongshang Yang et al. were able to use CRISPR-Cas9 to create gene-targeting-transgenic (GTT) rabbits for the first time, generating 4 novel lines of knock out rabbits. Similar success was seen in the targeting of the genomes of various organisms, from silkworms, to frogs and zebrafish.
Multiple labs were also able to demonstrate the use of Cas9 to affect gene expression in human cells. Cong. et al. demonstrated the use of both the standard Cas9 system and a modified Cas9 “nicking” enzyme that only cleaves one strand to facilitate sequence-homology-directed genomic insertions and repairs in mammalian cells. They were also able to demonstrate the use of a single long strand composed of smaller connected RNA strands to target multiple separate gene loci in series (“multiplex” genome engineering) — essentially demonstrating the ability to create algorithmic gene-editing “code” built into in a single RNA strand. Mali et al. would echo these results, using the multiplex CRISPR-Cas9 system to target genes in three types of human cells, including pluripotent stem cells. Though these initial results were exciting, much caution was exercised in the use of this technology to alter the human genome, limiting testing to laboratory cell lines and stem cells. However, to the surprise of many in the field, Liang et al., working out of Sun Yat-sen University in Guangzhou, China published research in 2015 demonstrating they had edited the genome of human embryos for the first time.
It was soon discovered that the system is not only useful for its DNA-slicing ability, but also for its highly–specific targeting functionality alone. Qi et al. made a significant breakthrough in 2013 with the creation of the CRISPRi (short for CRISPR interference) system which utilizes dCas9, a mutated form of Cas9 that lacks all endonuclease activity. They used this neutered form of Cas9, along with guide-RNAs, to trigger the activation and repression of gene transcription in human cells. This method is fundamentally different than previous methods, which relied on the generation of sequence-specific DNA binding proteins, and opened the door for the fast, inexpensive, highly-specific, reversible gene activation and repression tools with few off-target effects. They also showed that this system could be used to regulate multiple genes at once — the small size of the targeting region of the guide-RNA (20 base-pairs) allows for a multitude of targets on the scale of (potentially) the entire human genome. This work was followed up by Pirez-Pinera et al. who used the same system to reliably and reversibly activate human genes. Gilbert et al. and Maeder et al. demonstrated the use of a similar system, with effector-domains ferried by the dCas9 to particular gene loci, to recruit proteins and cause transcriptional activation and repression in human cells.
This lift-off phase was not without signs of potential danger. Fu et al. published a paper in Nature Biotechnology demonstrating that ribonucleotide guided nucleases (RGNs), CRISPR-Cas9 among them, can cut off target sites that differ from the complementary strand to the guide by up to five nucleotides. Fu would publish another result the following year (with a different team of collaborators) showing that the use of truncated guide-RNAs (with less than twenty nucleotides the in the target-complementary segment) can reduce off-target mutagenesis by 5000-fold or more without sacrificing accuracy. However, off-target effects are still a serious concern, and (along with difficulties in the delivery of the RNA/Cas-complex to the nucleus of target cells) constitute one of the most significant roadblocks for the CRISPR-Cas9 system as a potential drug model.
The potential complications were certainly not a deterrent for applications for intellectual property over the use of the CRISPR-Cas9 system.
To the contrary…
2017 Patent Dispute
In 2012 Jennifer Doudna, of the University of California, Berkeley filed for a patent on the intellectual property rights to the CRISPR-Cas9 system, showing that it can be used to modify specific loci of the bacterial genome. Dr. Doudna, and her frequent collaborator Emmanuelle Charpentier, are giants in the field, and have been a part of many major discoveries (including the seminal Jinek et al. work referenced above, upon which the patent was based, and the CRISPRi system, just to name a few). Shortly after, Feng Zhang, of the Broad Institute of MIT and Harvard, was granted a patent for the use of the same CRISPR system to edit the genome of eukaryotes.
Recognizing the possibly astronomical financial incentive, the University of Berkeley sued to have the Broad Institute’s patent thrown out, arguing that Zhang was awarded a patent over the same technology Doudna and her collaborators had invented, merely used in a different context.
On February 15th, 2017, the United States Patent and Trademark Office ruled in favor of Zhang, claiming the Broad Institute’s legal representation had successfully convinced government that the innovation was specific to the eukaryotic environment, and not merely an obvious re-application of Doudna’s invention. However, although Zhang’s patent claim over the use in eukaryotes was upheld, Doudna still retains her patent over the use of the CRISPR-Cas9 system “not restricted to any environment.” Interviewed about the decision, Doudna commented,
They will have a patent on green tennis balls. We will get a patent on all tennis balls.
Whether or not the University of California will appeal the ruling is yet to be seen. Considering the precedent set by the upholding of Zhang’s patent, it will be interesting to see how future patent battles play out, if other labs attempt to patent equally specific implementations of CRISPR-Cas technology. Given the multitude of Types and Subtypes of CRISPR I, II and III systems, the potential for new innovations specific to various systems/organisms seems high.
It is tempting to underestimate the potential of the CRISPR-Cas system. In many ways, it is just provides improvement of gene-editing capabilities that had existed already for decades, albeit allowing them to be more quickly, cost-effectively and accurately implemented. However, a few recently published studies should give the community and the broader population pause. In particular, the gene targeting and stability improvements enabled by the CRISPR-Cas9 system would allow for the first plausible implementation of a working synthetic gene drive.
Gene drives are genes that encourage, or force, their own expression. Simple examples include targeted nucleases which cut the complementary strand of a homologous chromosome during replication, and force the cell to reproduce a copy of the only remaining intact gene, or endonucleases that copy themselves into chromosomes that lack the genes that code for them, but there are many more types. Recall, however, that the CRISPR-Cas system is a nucleic-acid-based system: it is generated by the transcription of CRISPR repeats and spacers and cas genes (DNA) and translation to the Cas proteins that coordinate with them and act on them — indeed, were it not for the protospacer adjacent motif, the CRISPR-Cas machinery would cleave itself out of existence.
With this is mind, it is not hard to imagine a simple CRISPR-Cas9-based gene drive:
- simply insert into the chromosome a DNA sequence that will code for a chimera of tracrRNA along with the complementary strand to the gene(s) you would like to remove/alter;
- adjacent to this sequence (or not, for prokaryotes), attach a copy of the gene that codes for Cas9,
et voilà. Once this locus is copied and translated, the tracrRNA-crRNA strand will complex with the Cas9 protein and slice the complementary strand(s). In contrast to previous gene-deletion methods (in which subsequent reproduction with a wild-type organism would give a 50% chance of the offspring inheriting the spliced-out gene), even if the organism obtains a copy of the deleted gene from a future reproductive partner, the gene for Cas9 and guide RNA would remain waiting to splice it back out.
Now, there are many complications with this simple model. As sketched out above, this gene drive would be vulnerable to drive resistance through random mutations and repairs that would alter the specificity/affinity of the guide-RNA, natural sequence polymorphisms, or even the evolution of a method to combat the nuclease. But CRISPR-Cas9 makes it more plausible than ever that humans will be able to make permanent changes to entire populations of organisms. Indeed, a 2016 paper by Hammond et al. demonstrated a CRISPR-Cas9 gene drive system that rendered sterile female mosquitos which act as vectors for malaria, with transmission rates to progeny of 92-99% (depending on the gene locus targeted). Promising as it may be, this study alone suggests the need for extreme caution (and perhaps governmental regulation) as these tools are refined.
Released into the wild, on purpose or by accident, mutated organisms have a non-negligible chance of significantly altering the population at large. While it is straightforward to envision protocols to curb the unwanted proliferation of organisms the size of insects and larger, the proliferation of mutated microscopic organisms will be harder to control. The potential for unintended adverse effects due to genetic changes with this scope and magnitude is unprecedented.
With Great Power…
Putting aside caution for the moment, it is thrilling to envision the potential of these systems to perform feats previously thought prohibitively difficult, or impossible.
In the words of van der Oost et al. (2014):
in terms of applications of CRISPR-associated nucleases in general, and Cas9 in particular, the sky seems to be the limit.
For instance, as the CRISPR system has been shown to be found in about 90% of archaea and 40% of bacteria. Though there are few known examples of archaea that act as pathogens or parasites, examples of bacterial pathogens are manifold.
Considering CRISPR-containing bacteria already have the Cas machinery “installed,” all that is required for targeted interference in bacterial pathogenesis is a method for getting a guide-RNA-tracrRNA chimera into the cell, and in to the nucleus (both present a significant challenge).
As a bactericide, the CRISPR system has an advantage of over previous methods in the potential impossibility for resistance if the guide-RNA is sufficiently well designed: one can imagine a guide-RNA that targets a gene that is both organism-specific and essential; any mutation in the segment of DNA being targeted that would allow for resistance through evasion of guide-RNA recognition could potentially cripple the function of an essential protein; any mechanism developed to inhibit the Cas9 protein responsible for splicing would cripple the organism’s own quasi-immune system.
In eukaryotes, CRISPRs’ potential as a therapeutic agent is currently hampered in large part by the challenge of getting both the guide-RNA chimera and the Cas9 enzyme into the cell. Accordingly, methods of guide-/tracrRNA delivery are a major current focus for cellular and molecular pharmacology. Truncated guide-RNA results show a glimmer of promise to this end: the shorter the RNA strand to be delivered, the easier it will be to design a method for delivery. However, there are areas in which CRISPR-Cas9 can be put to use more easily, for instance in cancer therapies.
Gene editing of T-cells may enable efficient targeted antibody-mediated attack of cancer cells, and the ability to shield antibodies from recognition and subsequent destruction by cancer cells — immune cells removed from patients could be quickly and accurately edited and reintroduced. In fact, a CRISPR-Cas9 based drug that would do exactly that was approved for clinical trials in the US in June of 2016. In November of 2016, a CRISPR-Cas9-edited immune cells was administered to a human patient for the first time by a team led by Dr. Lu You a Sichuan University in Chengdu as part of a clinical trial for the treatment of aggressive lung cancer.
Though some suggest the excitement about CRISPR is exaggerated, all signs seem to indicate that pharmaceutical interest in CRISPR-Cas systems as potential therapeutics is still only starting to pick up steam — once the first successful CRISPR-Cas system passes clinical trials and heads to the market, rampant expansion in applications for potential CRISPR-Cas-based therapeutics patents can certainly be expected. It is telling that even in January of 2013, when CRISPR-Cas9 research was still an emerging field, Bayer AG announced it was investing 300 million dollars into a joint project with a startup, CRISPR Therapeutics, known as Casebia Therapeutics, to work on breakthroughs in blood disorders, blindness, and congenital heart disease. As of early 2017 there are already dozens of CRISPR-focused therapeutics companies (notable among them: Intella, Caribou Biosciences, Inc., and ERS genomics), some of which have already had multimillion dollar Initial Public Offerings (IPOs).
Faced with CRISPR’s seemingly limitless potential and power, one can only hope that excitement and wonderment are shown in equal measure to caution and humility.