The work to-date has focused on delving into the natural features of CRISPR arrays and their accompanying CRISPR biology as well as laying a foundation to apply CRISPR arrays to interrogating biological redundancy.
Within the natural features of CRISPR arrays, we discovered that the region upstream of CRISPR arrays contributes to the production of the encoded guide RNAs. This region, called the leader, was associated with other aspects of CRISPR biology but never guide RNA production. We showed for some CRISPR-Cas systems that this region interacted with the front end of the CRISPR array, promoting subsequent processing steps. As a result, the guide RNA targeting the invader most recently encountered by the cells is prioritized for defense, ensuring that the systems are primed against invaders that might reappear or could still be lurking in the environment.
Exploring other aspects of CRISPR biology beyond the CRISPR arrays has also proven fruitful. For instance, we discovered a set of novel CRISPR nucleases dubbed Cas12a2 that look for RNA targets and, upon finding their target, begin degrading virtually any nucleic acid they encounter. This activity extends to double-stranded DNA, the information storage material of cells and many invaders alike. We also discovered two clades of nucleases most closely related to Cas12a2, with one (Cas12a3) exhibiting RNA-triggered cleavage of tRNA tails. These nucleases represent the first examples in the CRISPR family in which the target-dependent enzymatic activity of the nuclease is directed away from the target to enact the immune response.
In a separate example from CRISPR biology, we exploited the discovery that the tracrRNA, a processing factor necessary to go from Cas9 CRISPR arrays to guide RNAs, could convert cellular RNAs into guide RNAs for use by Cas9. After engineering this process, we were able to achieve a technological first: recording selected cellular transcripts in single cells. This technology allows us to peer into a cell’s past while tying it to its present state. We also applied the concept to tracrRNA-dependent Cas12 nucleases that, upon target DNA recognition, collaterally cleave single-stranded DNA. This approach allowed us to harness these DNA-targeting nucleases for direct RNA detection, relying on collateral cleavage for signal amplification.
Laying the foundation for CRISPR array design, we developed a tool to predict targeting activity based on the guide RNA sequence. While many such tools exist, few have focused on using CRISPR to silence genes in bacteria. We applied machine learning with published datasets to devise an algorithm for predicting “good” guide sequences and “bad” guide sequences. We also explored how to make CRISPR arrays used by Cas9 more compact, finding that arrays can be shortened. In some cases, shortening the array even improved performance. We also used the arrays in other contexts, such as the first sRNA screens in bacteria using the gut microbe Bacteroides thetaiotaomicron as a model.
Finally, we advanced a simple system for characterizing CRISPR biology and technologies: cell-free transcription-translation (TXTL). Using. TXTL, we established new approaches for characterizing CRISPR-Cas systems involving multiple components. We also found that re-optimizing TXTL preparation allowed us to begin using linear DNA. This step makes it easier to go from designed DNA sequence to experimental testing, accelerating our ability to perform experiments.
