Skip to main content
Premium Trial:

Request an Annual Quote

Researchers Demonstrate Multiplex Codon Editing in the Human Genome


BALTIMORE – Researchers led by a team at Harvard Medical School have taken a step toward whole-genome recoding in human cells, demonstrating that the human amber stop codon TAG can be simultaneously edited in dozens of genes in a single experiment.

In a proof-of-concept study published last month in Nature Communications, the researchers, led by Harvard Medical School geneticist George Church, converted the TAG stop codon in 33 essential genes in a human cell line via a single transfection, marking the first large-scale recoding of the human genome.

This paper "pushed the limit" of how many different loci can be targeted at once for base editing in the human genome, said Eriona Hysolli, one of the study's corresponding authors and a former postdoctoral researcher in Church's lab.

According to Hysolli, the study, which was more than half a decade in the making and was conducted in collaboration with researchers from the Chinese Academy of Sciences, was built upon the Church group's previous accomplishment that replaced all UAG stop codons with UAA across the entire Escherichia coli genome. Hysolli, now head of biological sciences at Colossal Biosciences, said the team had to overcome a multitude of technical obstacles in order to extend recoding to the human genome.

For one, she said, compared with the relatively compact bacterial genomes, mammalian genomes are generally much larger and more complex. Besides having a greater number of genes, they also tend to harbor numerous intergenic regions that separate genes far apart from each other. Additionally, because many genome editing tools are adapted to work in prokaryotes, they often cannot function efficiently or are not yet available for mammalian cells.

There are also hurdles associated with multiplex base editing to tackle several genes at once. These include making the different targeted loci in the genome accessible, delivering a larger payload with multiple guide RNAs (gRNAs) into a cell efficiently, and curtailing the off-target promiscuity of the editing tools.

To explore the feasibility of genome-wide recoding in human cells, the researchers looked to the TAG stop codon, which represents fewer targets, given it is the least commonly used human codon, and can be theoretically edited to TAA by cytosine base editors.

To cope with the scale of the human genome, the team developed a python-based software, named Genome Recoding Informatics Toolbox (GRIT), that can automate the process of part design for recoding.

Based on the human genome reference GRCh38.p13 and the Online Gene Essentiality (OGEE) Database, GRIT identified a total of 6,700 TAG codons within the HEK293T cell line that was used in the study, of which 6,648 are editable across the human haploid genome by base editors. Moreover, the software discovered 1,947 TAG codons that are located in essential genes, of which 1,937 are editable.

Because multiplex TAG to TAA editing requires multiple gRNAs and base editors protein to be simultaneously delivered into a single mammalian cell, the researchers designed and synthesized artificial DNA fragments, or so-called gBlocks, that can each carry five individual gRNA cassettes and remain stable when introduced into mammalian cells.

After optimizing a strategy for multiplexed base editing, the researchers showed they were able to successfully convert up to 33 TAG codons to TAA at once out of 47 target sites via a single transfection.

To evaluate the on- and off-target efficiencies of the recoding, the team performed whole-genome sequencing on highly modified clones and compared the results with the negative control. While the off-target effects in this study were "definitely significant," most of them were found outside of the coding regions, Hysolli said.

Calling the study "very impressive," Magomet Aushev, a genome editing expert at the Wellcome Centre for Mitochondrial Research at Newcastle University in the UK, said this paper laid a foundation for other researchers who hope to achieve multiplex gene editing.

"One of the cool things that CRISPR brought to the genome editing table was multiplexing," he said, adding that unlike previous efforts, which mostly deployed CRISPR to modify repetitive sequences in the genome, this study targeted different genes using different guide RNAs, which is "really cool."

With all UAG stop codons replaced with UAA, the Church group previously engineered E. coli to be resistant to certain bacteriophages by depleting release factor 1 (RF1), which is responsible for terminating translation for UAG and UAA.

Similarly, the Harvard researchers and their collaborators envisaged that by converting the TAG stop codon to TAA genome wide and replacing the endogenous eukaryotic release factor 1 (eRF1) with engineered eRF1 variants, they might be able to generate virus-resistant human cell lines — a goal for the Genome Project–write (GP-write) project, an initiative of the Human Genome Project jointly spearheaded by Church and a handful of other researchers.

Aushev said that while the recent publication is "a really good initial study" that tested the waters of how far scientists can go with the current technology for human genome recoding, there is still a vast knowledge gap to fill when it comes to achieving virus-resistant human cell lines via whole-genome recoding.

"Base editing is like fixing a tire, while genome engineering is like building a whole car," he said. "Maybe a technology that is meant to fix tires is not the best way to go. … Maybe there needs to be some innovation on the genome engineering side to develop some newer technologies [that are] more ambitious."

Beyond the technical challenges, scientists also do not know the possible side effects of whole-genome recoding in humans, Aushev pointed out. "It could be that maybe nothing happens, or maybe something bad happens. It's just difficult to say because this has never been done."

Mirroring Aushev's point, Hysolli also thinks there is still a long way to go before scientists can harness human genome recoding for applications such as manufacturing of cell therapies or virus-proof therapeutics production.

Meanwhile, Hysolli said there are several potential strategies that scientists can work on to help improve the efficiency and multiplex capability of the TAG conversion. These include continuing to develop new base editors to boost editing efficiency while minimizing off-target effects, improving guide RNA delivery capability, and designing new delivery mechanisms to achieve more efficient transfections.

"It's always a little bit more challenging to work on the mammalian genome side," Hysolli said. "We have just laid the foundation for how we can potentially achieve this work."