CRISPR-Cas and Viral Immune Memory
Spacer Acquisition, PAM Recognition, and R-Loop Conformational Cleavage
Yoshizumi Ishino sequenced the alkaline phosphatase isozyme gene (iap) in Escherichia coli in 1987 and noticed five unusual 29-base-pair palindromic DNA repeats separated by non-repeating 32-base-pair spacers. At the time, no known genetic element shared that periodic architecture. The biological function of those clustered repeats remained an unsolved puzzle for nearly two decades.
After identifying similar repeat-spacer arrays across twenty bacterial and archaeal genomes from salt marshes to volcanic vents, Spanish microbiologist Francisco Mojica ran sequence alignments on the spacers in 2003. He reported in 2005 that the spacer sequences matched fragments of bacteriophages and conjugative plasmids. Microbes carrying a matching spacer were immune to that specific virus.
Rodolphe Barrangou and Philippe Horvath proved CRISPR adaptive immunity experimentally in 2007 while studying industrial Streptococcus thermophilus yogurt cultures at Danisco. Challenging bacteria with lytic bacteriophages selected for survivors that had spliced brand-new 30-base-pair viral DNA spacers into the leader end of their CRISPR array. Deleting those acquired spacers instantly abolished phage resistance.
During viral infection, an integrase complex of four Cas1 and two Cas2 subunits captures a snippet of invading phage DNA. Acting as a rigid protein ruler spanning 33 base pairs from end to end, Cas1-Cas2 splices the viral spacer into the first repeat at the leader end. That insertion order creates a chronological tape recorder of past infections.
If a Cas nuclease cuts any 20-nucleotide DNA sequence matching its guide RNA, why doesn't it slice the bacterium's own chromosomal CRISPR array where the spacer template sits? The enzyme ignores any DNA target unless it sits right next to a short Protospacer Adjacent Motif (PAM, such as $5'\text{-NGG-}3'$ for S. pyogenes Cas9). Host CRISPR repeats lack that PAM sequence.
Eric Greene and Jennifer Doudna used single-molecule DNA curtains in 2014 to watch how Streptococcus pyogenes Cas9 searches billions of base pairs for a 20-nucleotide target without burning ATP. Rather than unzipping every sequence, Cas9 bounces by three-dimensional collisions between $5'\text{-NGG-}3'$ PAM sites. It melts only the adjacent 8-base seed region and rejects mismatches in milliseconds.
Emmanuelle Charpentier and Jennifer Doudna—alongside independent biochemical proof by Virginijus Šikšnys—demonstrated in 2012 that the natural two-part RNA guide of S. pyogenes Cas9 (crRNA paired with tracrRNA) can be fused into a single chimeric guide RNA (sgRNA). Changing the first 20 letters of that sgRNA directs the 1,368-amino-acid enzyme to cleave any chosen DNA locus.
Cryo-EM and X-ray structures by Martin Jinek, Fuguo Jiang, and Eva Nogales revealed how Cas9 avoids cutting off-target sites that match only half the guide. The HNH nuclease domain starts $35\text{ \AA}$ away from its target strand. Only after the guide RNA zips across all 20 target bases to form an R-loop does the HNH domain swing 180 degrees to cut.
Although biotechnology relies heavily on single-protein Class 2 effectors like Cas9 (Type II), Cas12a (Type V), and RNA-cleaving Cas13 (Type VI), roughly 90 percent of CRISPR systems in nature belong to multi-subunit Class 1. In Type I systems, a 400-kilodalton Cascade complex flags the viral DNA. It then recruits Cas3, an ATP-powered helicase-nuclease that shreds kilobases of phage DNA.
When Cas9 makes a blunt double-strand break in a human chromosome, cellular DNA repair pathways complete the edit. Error-prone Non-Homologous End Joining (NHEJ) glues the broken ends back together with small insertions or deletions that knock out the gene. Supplying a donor DNA template during S or G2 phase lets Homology-Directed Repair write in a precise sequence replacement.
To rewrite mutated DNA letters without making double-strand breaks, David Liu engineered Base Editors in 2016 and Prime Editors in 2019. Base editors fuse a Cas9 nickase to a cytidine or evolved deoxyadenosine deaminase to convert $\text{C}\cdot\text{G}$ to $\text{T}\cdot\text{A}$ or $\text{A}\cdot\text{T}$ to $\text{G}\cdot\text{C}$. Prime editors use a tethered reverse transcriptase to copy a pegRNA extension directly into a nicked strand.
In late 2023, regulators in the United Kingdom and the United States approved the first CRISPR-Cas9 medicine, exagamglogene autotemcel (Casgevy). Doctors harvest a patient's own blood stem cells and use Cas9 to disrupt the erythroid enhancer of the BCL11A repressor gene. Infusing the edited cells back into the patient reactivates fetal hemoglobin to cure sickle cell disease and $\beta$-thalassemia.