The Genesis of Gene Editing: From Mendel to CRISPR-Cas9

The limitations of ZFNs and TALENs, however, were not lost on the broader scientific community. The high cost of protein engineering, low editing efficiency in many hard-to-transfect cell types, and steep learning curve for new labs meant that precise genome editing remained a niche tool, accessible only to well-funded specialized labs. The revolution that would democratize gene editing was already hiding in plain sight, in the genomes of the smallest life forms on Earth: bacteria.

The story of CRISPR-Cas9 begins not with a deliberate search for gene editing tools, but with a curious observation of repetitive DNA sequences in bacterial genomes. In 1987, Japanese researcher Yoshizumi Ishino and his team at Osaka University were studying the *iap* gene in *Escherichia coli* when they stumbled upon a strange pattern: five identical 29-base-pair repeats separated by 32-base-pair unique spacer sequences, all clustered together in the bacterial genome. At the time, the function of these repeats was completely unknown, and the paper describing them was largely overlooked for nearly two decades. Over the next 15 years, similar repetitive sequences were discovered in more than 40% of all sequenced bacterial genomes and nearly 90% of all sequenced archaea, leading researchers to coin the acronym CRISPR in 2002: *Clustered Regularly Interspaced Short Palindromic Repeats*. The “palindromic” descriptor refers to the fact that each repeat sequence reads the same forward and backward on complementary DNA strands, a structural feature that would later prove critical to the system’s function.

For years, CRISPR sequences were a biological mystery. The breakthrough came in 2005, when three independent research groups—led by Philippe Horvath at the Danish biotechnology company Novozymes, Rodolphe Barrangou at North Carolina State University, and Eugene Koonin at the National Institutes of Health (NIH)—hypothesized that CRISPR sequences were part of a bacterial adaptive immune system. Their reasoning was straightforward: the unique spacer sequences between CRISPR repeats matched sequences from bacteriophages (viruses that infect bacteria) and plasmids, suggesting bacteria were storing fragments of invading genetic material as a “memory” of past infections. The hypothesis was validated two years later, in a landmark experiment published in *Science* by Horvath and Barrangou’s team. They worked with *Streptococcus thermophilus*, a bacterium used in yogurt production that is frequently infected by phage viruses. When they compared phage survival rates between *S. thermophilus* strains with a CRISPR spacer matching the phage genome and strains without the spacer, the results were unambiguous: 92% of strains with the matching spacer survived phage infection, compared to just 8% of strains without the spacer. When they deleted the spacer sequence from the CRISPR locus, the previously resistant bacteria lost all immunity to the phage. This experiment proved that CRISPR sequences were not just repetitive junk DNA, but a functional immune memory system.

The next critical step was identifying the molecular machinery that carried out CRISPR-mediated immunity. Researchers quickly discovered that CRISPR loci are transcribed into a long precursor RNA that is cut into small “CRISPR RNAs” (crRNAs), each matching one of the stored phage spacer sequences. These crRNAs work in tandem with “CRISPR-associated” (Cas) proteins, a family of nucleases that can cut DNA. Early studies identified multiple Cas proteins with different functions, but the key breakthrough for gene editing came in 2011, when Emmanuelle Charpentier, then at Umeå University in Sweden, published work showing that a single Cas protein, Cas9 from the pathogenic bacterium *Streptococcus pyogenes*, was sufficient to cut invading phage DNA when guided by a crRNA and a second small RNA called the trans-activating crRNA (tracrRNA). Charpentier’s work showed that Cas9 acted as a molecular scissors that could be programmed to cut any DNA sequence complementary to the crRNA, as long as the target sequence was adjacent to a short 3-base-pair “protospacer adjacent motif” (PAM) with the sequence 5'-NGG-3' (where “N” can be any nucleotide). The PAM requirement is a critical safety feature: it ensures Cas9 only cuts invading phage DNA (which always has a PAM adjacent to the protospacer) and not the bacterial CRISPR locus itself (which lacks a PAM next to the stored spacer sequences).

Charpentier’s discovery caught the attention of Jennifer Doudna, a biochemist at the University of California, Berkeley, who had spent years studying RNA structure and function. The two met at a 2011 microbiology conference in Puerto Rico, and quickly decided to collaborate to test whether the CRISPR-Cas9 system could be repurposed as a programmable genome editing tool. Their team, which included postdoctoral researcher Martin Jinek, set out to simplify the system by fusing the crRNA and tracrRNA into a single “guide RNA” (gRNA) that could direct Cas9 to any target sequence. In a landmark paper published in *Science* on June 28, 2012, Doudna and Charpentier’s team demonstrated that the simplified CRISPR-Cas9 system could be programmed to cut any arbitrary DNA sequence in a test tube, simply by designing a 20-nucleotide gRNA sequence complementary to the target. The paper’s implications were immediately obvious to the gene editing community: unlike ZFNs and TALENs, which required custom protein engineering for every new target site, CRISPR-Cas9 only required the synthesis of a cheap, 20-base RNA molecule. The barrier to entry for genome editing had just dropped from thousands of dollars and weeks of work to less than $10 and 24 hours.

While Doudna and Charpentier’s 2012 paper proved CRISPR-Cas9 could cut DNA in vitro, the critical question of whether it could edit genomes in living cells was answered just months later by two independent teams. The first, led by Feng Zhang at the Broad Institute of MIT and Harvard, published a paper in *Science* on January 3, 2013, demonstrating that CRISPR-Cas9 could edit genes in both mouse and human cells. Zhang, a neuroscientist who had previously used ZFNs and TALENs to study brain development, had attended the same 2011 conference where Doudna and Charpentier presented their early CRISPR work, and immediately recognized the technology’s potential. His team tested two delivery methods for CRISPR-Cas9: transfection of plasmids encoding Cas9 and the gRNA, and direct delivery of a pre-assembled ribonucleoprotein (RNP) complex of purified Cas9 protein and synthetic gRNA. The RNP method proved far more efficient, with editing rates of 10-50% in human embryonic stem cells, compared to 1-5% for plasmid delivery. Critically, Zhang’s team also demonstrated that CRISPR-Cas9 could be used to introduce precise edits via homology-directed repair (HDR) by co-delivering a donor DNA template, opening the door to gene correction rather than just gene knockout.

The second independent validation of CRISPR-Cas9 in mammalian cells came from George Church’s lab at Harvard Medical School, which published a paper in *Cell* on the same day as Zhang’s *Science* paper. Church’s team used CRISPR-Cas9 to edit the *CCR5* gene in human cells, introducing a 32-base-pair deletion that mimics a naturally occurring mutation conferring resistance to HIV infection. Edited cells showed no expression of the CCR5 protein on their surface, and were completely resistant to HIV infection in vitro. This experiment was the first demonstration that CRISPR-Cas9 could be used to edit human genes for therapeutic purposes, and it immediately sparked a wave of research into CRISPR-based gene therapies for genetic diseases, infectious diseases, and cancer.

To understand why CRISPR-Cas9 represented such a dramatic leap over first-generation editing tools, it is useful to compare the technical requirements for designing a new editing experiment. For ZFNs, each 3