Researchers have created a new gene-editing tool called Retron Library Recombineering (RLR) that can generate up to millions of mutations simultaneously, and mutant bacterial cells ‘barcodes’ so that the entire pool can be screened at the same time. It can be used in contexts where CRISPR is toxic or not feasible, and results in better edit rates.
Although the CRISPR-Cas9 gene editing system has become the model for innovation in synthetic biology, it has major limitations. CRISPR-Cas9 can be programmed to find and cut specific pieces of DNA, but editing the DNA to create the desired mutations requires tricking the cell to use a new piece of DNA to repair the break. This bait and switch can be complicated to orchestrate, and can even be toxic to cells, as Cas9 often also cuts unintentional and untargeted sites.
Rather, alternative gene-editing techniques called recombinations perform this bait and switch by introducing another piece of DNA while a cell replicates its genome, effectively creating genetic mutations without breaking down the DNA. These methods are simple enough that they can be used in many cells at once to create complex pools of mutations for researchers to study. Determining what the effects of these mutations are, however, requires that each mutant be isolated, sequenced, and characterized: a time-consuming and inconvenient task.
Researchers at the Wyss Institute for Biologically Inspired Engineering at Harvard University and Harvard Medical School (HMS) have created a new gene-editing tool called Retron Library Recombineering (RLR) that makes this task easier. RLR generates up to millions of mutations simultaneously, and “barcode” mutant cells so that the entire pool can be screened at the same time, making it easy to generate and analyze massive amounts of data. The achievement, which has been accomplished in bacterial cells, is described in a recent article in PNAS.
“RLR allowed us to do something that is impossible to do with CRISPR: we randomly cut out a bacterial genome, transformed these genetic fragments into single-stranded DNA in situ, and used them to screen millions of sequences simultaneously,” said said co-first author Max Schubert, Ph.D., post-doctoral fellow in the lab of George Church, Ph.D., faculty member at Wyss Core. “RLR is a simpler and more flexible gene editing tool that can be used for highly multiplexed experiments, which eliminates the toxicity often seen with CRISPR and improves researchers’ ability to explore genome-level mutations.”
Let’s go back: from the riddle to the engineering tool
Retrons are segments of bacterial DNA that undergo reverse transcription to produce single stranded DNA (ssDNA) fragments. The existence of Retrons has been known for decades, but the function of the ssDNA they produce baffled scientists from the 1980s through June 2020, when a team finally figured out that Retron’s ssDNA detects whether a virus has infected the cell, part of bacterial immunity. system.
While retons were originally thought to be simply a mysterious bacterial oddity, researchers have become more interested in them in recent years because they, like CRISPR, could be used for the precise and flexible editing of genes in bacteria, yeasts and even human cells.
“For a long time, CRISPR was just seen as a strange thing that bacteria did, and figuring out how to harness it for genome engineering changed the world. Retrons are another bacterial innovation that could also provide important breakthroughs, ”said Schubert. His interest in retons was piqued several years ago due to their ability to produce ssDNA in bacteria – an interesting feature to use in a gene editing process called oligonucleotide recombination.
Recombination-based gene editing techniques require the integration of ssDNA containing a desired mutation into the DNA of an organism, which can be done in two ways. Double-stranded DNA can be physically cut (with CRISPR-Cas9, for example) to induce the cell to incorporate the mutant sequence into its genome during the repair process, or the mutant DNA strand and a single hybridization protein. strand (SSAP) can be introduced into a cell that replicates so that the SSAP incorporates the mutant strand into the DNA of the daughter cells.
“We thought that the retons should give us the ability to produce ss DNA in the cells we want to modify rather than trying to force them into the cell from the outside, and without damaging the native DNA,” which were two very compelling qualities, ”said co-first author Daniel Goodman, Ph.D., a former research fellow at the Wyss Institute who is now a Jane Coffin Childs postdoctoral fellow at UCSF.
Another appeal of retrons is that their sequences themselves can serve as “bar codes” that identify individuals within a pool of bacteria that have received each retron sequence, allowing considerably faster pooled screens of mutant strains. created with precision.
To see if they could actually use retrons to achieve efficient recombination with retrons, Schubert and his colleagues first created circular bacterial DNA plasmids containing antibiotic resistance genes placed in retron sequences, as well. than an SSAP gene to allow integration of the retron sequence into the bacterial genome. They inserted these Retron plasmids into E. coli bacteria to see if the genes were successfully integrated into their genomes after 20 generations of cell replication. Initially, less than 0.1% E. Coli carrying the Retron recombination system incorporated the desired mutation.
To improve on this disappointing initial performance, the team made several genetic modifications to the bacteria. First, they inactivated the natural mismatch repair mechanism in cells, which corrects DNA replication errors and could therefore “fix” desired mutations before they can be passed on to the next generation. They also inactivated two bacterial genes that code for exonucleases – enzymes that destroy free-floating ssDNA. These changes dramatically increased the proportion of bacteria that incorporated the retron sequence, to over 90% of the population.
Name tags for mutants
Now that they were convinced that their ssDNA retron was incorporated into the genomes of their bacteria, the team tested whether they could use the retrons as a “shortcut” to genetic sequencing, allowing many experiments to be performed in a mixture. Because each plasmid had its own unique retron sequence which can function as a “name tag”, they felt that they should be able to sequence the much shorter retron rather than the entire bacterial genome to determine which mutation the cells are. had received.
First, the team tested whether RLR could detect known antibiotic resistance mutations in E. coli. They found that it could – backward sequences containing these mutations were present in much larger proportions in their sequencing data compared to other mutations. The team also determined that the RLR was sensitive and precise enough to measure small differences in resistance resulting from very similar mutations. Basically, collecting this data by sequencing barcodes from the entire pool of bacteria, rather than isolating and sequencing individual mutants, dramatically speeds up the process.
Then the researchers took the RLR a step further to see if it could be used on random fragmented DNA, and to find out how many retons they could use at a time. They cut out the genome of a strain of E. Coli highly resistant to another antibiotic and have used these fragments to build a library of tens of millions of genetic sequences contained in retron sequences in plasmids. “The simplicity of RLR really shone in this experiment, as it allowed us to build a much larger library than what we can currently use with CRISPR, in which we have to synthesize both a guide and a DNA sequence. donor to induce each mutation, ”says Schubert.
This library was then introduced into the RLR-optimized E coli strain for analysis. Once again, the researchers found that the retons conferring resistance to antibiotics could be easily identified by the fact that they were enriched relative to the others when sequencing the pool of bacteria.
“The ability to analyze barcode mutant libraries clustered with RLR allows millions of experiments to be performed simultaneously, allowing us to observe the effects of mutations across the genome, as well as how these mutations could interact with each other, ”said lead author George Church, who heads the synthetic biology focus at the Wyss Institute and is also a professor of genetics at HMS. “This work helps to establish a roadmap towards the use of RLR in other genetic systems, which opens up many interesting possibilities for future genetic research.”
Another feature that distinguishes RLR from CRISPR is that the proportion of bacteria that successfully integrate a desired mutation into their genome increases over time as the bacteria replicate, whereas the “one shot” method of CRISPR tends to be successful. or fail the first time. RLR could potentially be combined with CRISPR to improve its editing performance, or could be used as an alternative in the many systems in which CRISPR is toxic.
There is still work to be done on RLR to improve and standardize the edit rate, but enthusiasm is growing for this new tool. The simple and streamlined nature of RLR could make it possible to study how several mutations interact with each other and generate a large number of data points that could allow the use of machine learning to predict other effects. mutation.
“This new synthetic biology tool takes genome engineering to even higher throughput levels, which will undoubtedly lead to exciting and unexpected new innovations,” said Don Ingber, MD, Ph.D., director founder of the Wyss Institute. Ingber is also Judah Folkman Professor of Vascular Biology at HMS and Boston Children’s Hospital, and Professor of Bioengineering at Harvard John A. Paulson School of Engineering and Applied Sciences.
Other authors of the article include Timothy Wannier of HMS, Divjot Kaur of the University of Warwick, Fahim Farzadfard and Timothy Lu of the Massachusetts Institute of Technology and Seth Shipman of the Gladstone Institute of Data Science and Biotechnology.
This research was funded by the United States Department of Energy (DE-FG02-02ER63445) and the National Defense Science and Engineering Graduate Fellowship.