By Leo Barolo
Genome-wide CRISPR screening is becoming more and more accessible to a wide range of research topics. With new resources on Campus facilitating the use of the technology, researchers from different fields are using it in their research for the first time.
As new researchers consider genome-wide CRISPR screening, experts in the field share what is exciting about the technology and who should consider using it. Dustin Rubinstein, director of the UW Genome Editing and Animal Models (GEAM) facility and part of the group that first published the use of CRISPR/Cas9 in Drosophila, Gaelen Hess, a screening pioneer working on mammalian cells, and Jason Peters, a bacteria researcher pushing the boundaries of the technology, share their insights on how this tool is changing how we do research.
What is Genome-wide CRISPR screening?
From drug discovery to understanding cancer pathways, many research projects start with the same first step: identify which genes are responsible for a specific phenotype of interest. This is done through screening, a process of assessing a large number of gene perturbations (such as mutations or knockdowns) to identify which of them are responsible for the outcome or phenotype of interest.
CRISPR systems were originally identified as part of bacterial immune systems and have now been exploited to alter an organism’s genome. Systems like CRISPR cutting are based on a bacterial enzyme that is directed to mutagenize a specific region of the genome based on targeting via a guide RNA (gRNA). Other systems, like CRISPR interference (CRISPRi) and CRISPR activation (CRISPRa), don’t edit genes directly, instead controlling gene expression. Through CRISPR, researchers are able to delete, repress, up-regulate, or otherwise alter a genomic locus in different ways.
In the early days, whole-genome screening was done through arrayed libraries of mutant cells – several stacks of microtiter plates with each well containing a cell line with a different gene deleted or mutated. All the plates were then exposed to the experimental treatment, and cells could be phenotyped individually. This type of screening can assay thousands of mutant cells in parallel. However, it is a high-maintenance and costly process that is now being outclassed by pooled screening, a technology offered to UW researchers by GEAM.
Pooled screening compiles all mutant cells into a pooled library, generated using a complex pool of guide RNAs that direct the CRISPR enzyme to target different genes to knock out in each cell. To identify new pathways, researchers investigate the representation of guide RNAs in the starting pool, then expose the mutant pool to their drug or selection of interest, and finally compare the resulting pool to the original.
By doing so, cells in which a critical gene was knocked out will die and are depleted in the final pool. Thus, comparing quantitative sequence reads of guide RNAs or their corresponding DNA barcode (used to identify cells) before and after selection reveals what gene mutations were selected.
Genome-wide screening is particularly useful when much is unknown about a system and broad questions are being asked.
Screens were traditionally conducted in parallelized assays, with thousands of mutants in microtiter plates assayed individually. But more recent methods focus on pooled selections, in which all mutants are grown in a single pool and read out by deep sequencing.
Peters mentions that pooled screening allowed him to find unexpected information. “We found that even well-characterized antibiotics can have effects outside of their known target in bacterial cells—the unbiased nature of our pooled CRISPRi screens have revealed unexpected insights.”
“The strength of a screen is that you are giving a chance for anything to be seen” – Dr. Hess.
Hess (highlighted recently in G&T), Peters, and Rubenstein highlighted several key considerations – and the biggest challenges – on getting started with CRISPR screening.
Designing a gRNA library and getting it into cells.
CRISPR enzymes are directed to genomic loci through guide RNAs (gRNA). Thus, the first challenge is designing and generating a gRNA library that minimizes off-target effects. GEAM facilitates researchers’ connections – and access to discounted rates – with companies like Millipore-Sigma, which offer gRNA library design and oligo synthesis. These libraries are then cloned into either a plasmid or lentivirus backbone and typically delivered to Cas-engineered cell lines.
To facilitate library transfer into diverse bacteria, the Peters lab developed the bacterial Mobile-CRISPRi system. This system is agnostic to the bacterial species and integrates both Cas enzyme and gRNAs into the genome. Once integrated, bacteria can be pooled and grown in the absence of plasmid selection. This offers new possibilities for screening — Peters’ group is using the system to study antibiotic function and genes important for pathogen infection in mice.
Choosing amenable cell lines
Hess points out that not all cell lines are tolerant of CRISPR machinery, and lentivirus infection efficiency varies across lines.
“The closer you get to primary cells [from a living organism, as opposed to immortalized cell lines], the more sensitive they are,” says Hess.
Rubenstein outlined several criteria to select appropriate cell lines: they should divide rapidly, be easy to handle, and be efficiently transfected. “Early first-step controls can give an insight into how well a line would work,” says Rubenstein. But another solution is to use proven cell lines and those already engineered with CRISPR enzymes.
Peters points out limitations in bacteria, since some strains have natural restriction enzymes that degrade plasmid DNA encoding the CRISPRi system. Once again, early tests for transformability can guide which bacterial species will work for CRISPR screening.
Measuring phenotypes
Most researchers use pooled screens to track phenotypes related to growth, which makes the measurement process straightforward – by comparing the starting composition of the strains in the pool to the after-selection composition, one can easily determine which strains were sensitive to the condition and dropped out of the population while resistant strains grew more.
CRISPR screens that target regulatory elements, however, may be more challenging, because the effects may be subtle. Hess explains that, as the phenotypes are weak, “the researcher should know what they are looking for and whether they can get a big enough change” due to the genetic perturbation.
Other methods besides growth rate can also be used to distinguish mutant phenotypes. “What you can do is to take a fluorescent-protein gene and make a promoter fusion to it – now your library has different levels of fluorescence, and you can do things like FACS sorting,” illustrates Peters. Alternatively, one can introduce strain-specific barcodes into the 3’ untranslated region of the reporter gene, and simply sequence the reporter RNA and quantify the barcodes in the transcript. “The limitation is that you need a very large number of reads to count the reporter RNA accurately,” warns Peters.
Sub-libraries
A major consideration in genome-wide screens is making sure the phenotypic assay is sensitive and measuring only the phenotype of interest. Hess encourages people to develop their experiments using sub-libraries – small screens targeting 10-15% of a genome. “A lot of genome-wide libraries are already subdivided into these” he explains, “If that doesn’t work, your genome-wide screening wouldn’t work either, so it saves money.” He explains this is particularly important because while sequencing is usually the greatest cost of screens, preparing the libraries can cost just as much, especially for genome-wide sequencing.
He also attests that using sub-libraries facilitates trying different conditions – “If you don’t know the best assay, you can try 2-3 ways quickly and figure out the optimal strategy”.
Talk to someone who is doing it
Peters encourages researchers considering high-throughput CRISPR screening to talk to people who are already doing it. “One of the nice things about this campus is that people are willing and happy to talk to you,” he explains, “They want to talk about their research – they are not trying to shield it from you.”
Additionally, he recommends researchers review articles about the topic and think through the whole process before getting started – “Don’t start building a library and then realize you have no way of getting it to your cells!” he cautions.
There are several opportunities for high-throughput screening at UW-Madison. GEAM (a shared facility with the Carbon Cancer Center and UW Biotechnology Center, UWBC) is now widely accessible, with cost- and time-effective methods for reliable genome-edited cell lines. “UWBC works with several different types of edits and cell lines in order to assist as many types of research as possible,” explains Rubinstein.
GEAM has the equipment to streamline the process to track outgrowth, such as automatic incubators and robotic arms that move plates from the incubator to a plate imager to track clonal outgrowth. They can verify clonality, track fluorescent markers, and process 66 plates, each with almost 100 clones. “There is a lot of interest and now there is automation behind it to fulfill it,” Rubinstein points out.
Researchers interested in using GEAM services should contact the Genome Editing Facility through genomeediting@biotech.wisc.edu