How to Hack the Genome

TECHNOLOGY FEATURE HOW TO HACK THE GENOME After tackling the genomes of bacteria and yeast, synthetic biologists are setting their sights on rewriting those of more complex organisms, including humans. POWER AND SYRED/SCIENCE PHOTO LIBRARY PHOTO AND SYRED/SCIENCE POWER

Human chromosomes under a scanning electron microscope.

BY JEFFREY M. PERKEL blue light. “When you add it all up, it’s quite a In a 2015 letter2 to the journal Trends in sophisticated project,” Voigt says. Biotechnology, Carr asked: “Is there a syn- nder typical laboratory conditions, And it’s not the only one. Synthetic biol- thetic biology equivalent of the sound barrier, strain JF1 of the bacterium Escherichia ogy is awash with projects of similar or or of the speed of light?” The question was coli looks like any other — a spatter even greater complexity. Improvements rhetorical because immutable limits clearly Uof yellow-tinged colonies on an amber agar in techniques for synthesizing and editing exist — the growth rate, for example, cannot plate. But bathe the colonies in wavelengths DNA have brought reduced costs and enor- be infinitely fast. But what constitutes a sound of red, green or blue light and their cells con- mous precision, helping biologists to build barrier in synthetic biology is evolving, he says. vert chemicals in the growth medium into from scratch or re-engineer the genomes of Designs that were on the cusp of feasibility a pigments in a pattern that matches that of the microorganisms such as E. coli and brewer’s few years ago are now practical. coloured light to which they were exposed, yeast (Saccharomyces cerevisiae). Synthetic- Researchers who once struggled to produce a yielding a muted and blurred image that is biology researchers are now having seri- few kilobases of synthetic DNA are now build- reminiscent of a 1970s Polaroid. ous discussions about re-engineering the ing whole genomes on the scale of megabases. Christopher Voigt, whose lab at the Massa- genomes of more-complex organisms, In March 2016, sequencing and synthetic biol- chusetts Institute of Technology in Cambridge including humans, although substantial ogy pioneer Craig Venter and his colleagues created the intricate genetic circuit that drives hurdles stand in the way. For instance, the announced that they had pruned and rewrit- this transformation, reported in May 2017 manipulation of large pieces of DNA pre- ten the genome of the bacterium Mycoplasma that his team had used the system to recreate sents technical challenges, and despite the mycoides from about 1 megabase to 531 kilo- a multicoloured geometric illustration of falling cost of DNA synthesis, the cost is still bases to create a ‘minimal’ genome3 — the lizards by Dutch artist M. C. Escher1. That prohibitive when billions of bases must be smallest set of genes that is required for life. exercise was just for fun, he says, and a way to rewritten. In August 2016, researchers led by George demonstrate the state-of-the-art in synthetic “Results over the past two years have cer- Church, a geneticist at Harvard Medical School biology. But it was not easy: the circuit con- tainly increased my optimism that we may be in Boston, Massachusetts, and Nili Ostrov, a tained 18 genes and 32 regulatory elements, able to do some really profound engineering in postdoctoral researcher in his lab, reported spread over 4 small circular molecules of DNA animals,” says Peter Carr, a synthetic biologist that they had produced a bacterium, dubbed known as plasmids, and 46,198 base pairs of at the MIT Lincoln Laboratory in Lexington, ‘rE.coli-57’, in which seven codons — the tri- DNA. It responds separately to red, green and Massachusetts. plets of nucleotides that encode particular

TECHNOLOGY CELL ENGINEERING

amino acids — had been stripped out and the simplest bacterial genomes, is huge,” Venter replaced with synonymous alternatives4 in says. Success came instead from a top-down a process known as genetic recoding. And in approach that whittled away the genome to March 2017, a team led by Pamela Silver, a bio- arrive at a core set of 473 genes. But about one- chemist at the Wyss Institute for Biologically third of those have no known function. “I found TEMPLE JASMINE Inspired Engineering at Harvard University that to be just kind of a mind-blowing result,” in Boston, Massachusetts, described its initial Silver says. attempts to recode the genome of strain LT2 of And there are further challenges. Sc2.0 and the bacterium Salmonella typhimurium5, replac- other genome-rewriting projects have tended ing around 200 kilobases of genomic DNA and to steer clear of the regulatory regions of genes, eliminating a specific leucine codon in the hope but in more-complex organisms such as eukary- of preventing the transfer of genes between otes, these are often located far from the genes pathogenic microbes. that they influence, and might not yet be fully Most dramatically, in March 2017, an inter- mapped. Researchers may therefore not know national consortium led by Jef Boeke at the New which segments to rewrite, and which to leave York University Langone Medical Center and alone. It is also unclear how such large-scale Joel Bader of the Johns Hopkins University in Yeast genetically engineered to produce pigments genomic changes might affect chromatin archi- Baltimore, Maryland, reported the end-to-end from bacteria, other fungi and plants. tecture and therefore gene expression. rewrite of 5 of the 16 chromosomes of On a practical level, chromosome-sized mol- S. cerevisiae6 — a milestone in an international LEU2, which enables growth when leucine is ecules of DNA cannot be easily manipulated project called the Synthetic Yeast Genome missing. The megachunks are then slotted into without being broken, and there is no efficient Project (Sc2.0). Sc2.0 aims to optimize and an existing chromosome through homologous way to deliver them into most eukaryotic cells. synthesize the complete genome of S. cerevisiae recombination, a natural process in which one Even if scientists can deliver the DNA, they for both industrial stretch of DNA is replaced with another, rewrit- might not be able to integrate it into the genome and pure research “It’s going to ing the DNA from one end to the other. As each because most such cells are unable to perform applications. For get easier and subsequent segment is integrated, it replaces the homologous recombination as readily as yeast, instance, says Boeke, easier with time marker gene of the previous segment, swapping and their slower growth drags out each experi- by removing all DNA to build large the yeast cell’s nutritional requirements between mental step. sequences that do genomes.” uracil and leucine. Following quantitative PCR There also is the cost of synthetic DNA to not encode proteins analysis to ensure that each megachunk has consider. Silver’s team received funding from (introns), the team can assess the biological been fully incorporated, the resulting yeast the US Defense Advanced Research Pro- roles of the cellular machinery required to strain is tested for its ability to form colonies jects Agency for her work in S. typhimurium, handle those genetic elements. under relatively stringent conditions; its slowed which allowed it to negotiate a favourable price The artificial yeast chromosomes designed growth or death in comparison to the wild type for DNA synthesis. But at a per-base price for Sc2.0 have been streamlined and stabilized indicates a problem in need of repair. “The idea of US$0.10, she says, it will cost more than by deleting repeated sequences and introns, by is: integrate the megachunk, test the fitness”, and $1 million to complete her project; the human moving sequences that encode crucial pieces repeat, Mitchell explains. genome, by comparison, would cost hundreds of the protein translational machinery to a of times more. dedicated chromosome, and by eliminating WRITER’S BLOCK Yet Church says it is just a matter of time the codon TAG, which signals a stop in transla- A single megachunk can be integrated and before technology catches up with ambition. tion, and replacing it with an alternative stop tested in about two weeks, Mitchell says, assum- “My guess is, it’s going to get easier and easier codon, TAA, to facilitate protein engineering. A ing that there are no problems. Fitness testing with time to build large genomes.” customized software package called BioStudio and ‘debugging’ (error correction) “take longer enabled the team to manage the genetic than the actual build at this point”, says Boeke. PRECISION REWRITE bookkeeping required to complete such a Each chromosome completed so far pre- Genome rewrites so far have largely stuck to massive task. sented only a handful of notable ‘bugs’, he says. nature’s recipe. But ultimately, biologists hope The logistics of Sc2.0 were substantial, says Some stemmed from errors in genome anno- to impart new functions. Patrick Yizhi Cai at the University of Edinburgh, tation, whereas others were caused by codon Several projects, including the rE.coli and UK, who is the project’s international coordina- replacements that, for instance, alter the sec- S. typhimurium studies, are focusing on genetic tor. Yet the actual process of editing the yeast ondary structure of RNA. recoding, in which codons removed from the chromosomes was fairly routine, requiring just For the most part, the yeast rolled with the genome are freed for other uses. Jason Chin, a a few plates of yeast in the incubator, says Leslie punches. Yet the glitches that did crop up hint synthetic biologist at the MRC Laboratory of Mitchell, a postdoc in Boeke’s lab who led the at the challenges faced by a larger-scale engi- Molecular Biology in Cambridge, UK, has done subgroup that synthesized chromosome VI. neering project called Genome Project-write extensive work to manipulate the genetic code. The Sc2.0 team uses a strategy called SwAP-In (GP-write), which aims to rewrite the genomes He says that such recoding can advance protein to rewrite chromosomes piece by piece (see of more-complex eukaryotes. Besides the obvi- engineering, not to mention the design, testing ‘Rebuilding a chromosome’). Researchers first ous problem of scale — at 3 billion base pairs, and synthesis of new chemical polymers built assemble short single-stranded molecules of the human genome is two orders of magnitude from monomers other than standard amino DNA known as oligonucleotides into building larger than that of S. cerevisiae — the genomes acids. Other possible applications include bio- blocks of about 750 bases, and then into ‘chunks’ of more-complex organisms tend to be less containment (preventing release of an organism of 10 kilobases or fewer that, in turn, are com- well annotated. When Venter’s team first tried outside the lab) and genetic isolation (protecting bined into ‘megachunks’ of 30–60 kilobases. to build a minimal genome in M. mycoides, organisms from viral infection). Each megachunk contains one of two marker it applied a rational design, using published In 2005, while working as a postdoc with genes that enable the selection of yeast carrying genetic data to compile a list of essential Church, Farren Isaacs, now a bioengineer at these genes — in this case, URA3, which ena- genes — an approach that didn’t work. “Our Yale University in New Haven, Connecticut, bles yeast to grow in the absence of uracil, and lack of basic biological knowledge, even with began to pursue the idea of recoding the E. coli

478 | NATURE | VOL 547 | 27 JULY ©20172017 Mac millan Publishers Li mited, part of Spri nger Nature. All ri ghts reserved. ©2017 Mac millan Publishers Li mited, part of Spri nger Nature. All ri ghts reserved.

CELL ENGINEERING TECHNOLOGY

genome, focusing his energy on replacing the stop codon TAG throughout. REBUILDING A CHROMOSOME Because E. coli contains just 321 TAG codons, To rewrite yeast chromosome XII, the Synthetic Yeast Genome Project (Sc2.0) used a two-step process. Isaacs was able to accomplish this task by modi- Synthetic megachunks Partial synthetic chromosome XII strains fying an existing genome rather than synthesiz- Segments of synthetic DNA (coloured blocks) Through a series of hierarchical mating ing one from scratch7. Using a strategy called were inserted into the yeast genome to steps, the segments in each partial strain create six separate strains, each containing were brought together to produce a single, MAGE (multiplexed automated genome engi- one-sixth of rewritten chromosome XII. fully synthetic chromosome XII. neering) that enables multiple DNA sequences

HTTP://DOI.ORG/B9SS (2017)/AAAS to be edited at once, Isaacs and his colleagues first divided the E. coli genome into 32 segments SCIENCE and altered the TAG codons in each to the synonymous codon TAA. Next, they joined the ET AL. ET 32 modified segments into a single molecule by exploiting a natural process of genetic exchange between bacteria. To complete the recoding process, the team deleted a gene that encodes a protein known as RF1, which recognizes the

ADAPTED FROM ZHANG ADAPTED codon TAG. (A related protein, RF2, recognizes the codon TAA.) Survival of the modified E. coli 100% following removal of this otherwise essential gene showed that their recoding process had 66% worked. For many researchers, existing technolo- gies provide all the power they need to hack 33% the genome. eGenesis, located in Cambridge, Massachusetts, is using the gene-editing tool CRISPR to turn pigs into sources of transplant- altering the coding sequence of a particular (cellocad.org) to make that possible. Research- able organs. Luhan Yang, cofounder of the com- gene inadvertently weakened the promoter of ers specify genetic-circuit designs in a pro- pany, explains that the idea is to pare back the an overlapping gene, reducing the fitness of gramming language called Verilog, and Cello pig genome using CRISPR to remove sequences the strain. produces the DNA sequences that are required encoding proteins that might evoke an immune “There’s idiosyncratic information in the to make them work9. response in people. New genes encoding pro- genome,” explains Chin, and it can only be deci- Despite their seeming simplicity, the genomes teins that help to make the pig tissues compat- phered experimentally. of microorganisms demonstrate an incredible ible with humans can also be introduced. “We capacity for subtle genetic control, Voigt says. think dozens of modifications probably would CIRCUIT CITY “We’re almost taunted by what exists in nature.” suffice,” she says. Other researchers are developing genetic But through a combination of genome editing, Yet the approach is different for more- circuitry to imbue genomes with new genome synthesis and cleverly designed tools, expansive projects. Ostrov and Church’s rE.coli functionality. researchers are slowly rebalancing the scales. ■ project, for instance, removed the TAG stop Generally, these circuits — such as Voigt’s codon and two codons each for serine, arginine image-capturing strain of bacteria — are built Jeffrey M. Perkel is technology editor for and leucine from E. coli to create a 57-codon up from simpler designs that use proteins Nature. strain4. That work required 62,214 changes, called transcription factors as positive or nega- 1. Fernandez-Rodriguez, J. et al. Nature Chem. Biol. which the team made using bottom-up DNA tive input and output signals. Wilson Wong, a 13, 706–708 (2017). synthesis rather than top-down editing. With so biomedical engineer at Boston University in 2. Zakeri, B. & Carr, P. A. Trends Biotechnol. 33, 57–58 many necessary genetic modifications, Ostrov Massachusetts, builds his designs instead using (2015). 3. Hutchison, C. A. III et al. Science 351, aad6253 explains, “we might as well make the genome enzymes known as recombinases that invert or (2016). from scratch.” delete segments of DNA — a design strategy 4. Ostrov, N. et al. Science 353, 819–822 (2016). None of these genome-hacking studies called BLADE (Boolean logic and arithmetic 5. Lau, Y. H. et al. Nucleic Acids Res. 45, 6971–6980 (2017). has actually built a chromosome-sized mol- through DNA excision). 6. Richardson, S. M. et al. Science 355, 1040–1044 ecule of DNA in one continuous stretch. Most Wong says that BLADE frees researchers (2017). commercial suppliers of synthetic DNA rely from the difficulty of linking circuits to one 7. Isaacs, F. J. et al. Science 333, 348–353 (2011). 8. Weinber, B. H. et al. Nature Biotechnol. 35, 453–462 on a decades-old method of synthesis that is another, which requires the output strength of (2017). unsuitable for producing molecules longer one circuit to be matched with the anticipated 9. Nielsen, A. A. K. et al. Science 352, aac7341 (2016). than about 200 nucleotides. Church’s team, as input of the next. do most groups pursuing genome synthesis, In one demonstration, Wong and his team assembled the DNA it required hierarchically. created a Boolean logic look-up table8 — a CORRECTION It purchased pre-made segments of genes that genetic circuit, about 10 kilobases long, that is The Technology Feature ‘The architecture were 2–4 kilobases long, assembled them into capable of turning into any of 16 possible logic of structured DNA’ (Nature 546, 687– 50-kilobase-long blocks in yeast using homol- gates depending on whether 6 recombinases are 689; 2017) erroneously credited Paul ogous recombination, and transferred these present. Rothemund for designing the experiments. completed segments into E. coli. The team then Wong’s team essentially designed that cir- In fact, the work was done by his postdoc, deleted the corresponding region of the E. coli cuit using a pencil and paper. But ultimately, Ashwin Gopinath. Furthermore, the genome, and tested the resulting strain of bac- synthetic biologists hope to build their examples of shapes created by Shawn teria for fitness. designs in silico. Voigt, working with Douglas Douglas should have been protractors, According to Ostrov, the recoding process Densmore, an electrical engineer at Boston twisted ribbons and an octahedral globe. went smoothly, despite a few bugs. For example, University has developed a tool called Cello

©2017 Mac millan Publishers Li mited, part of Spri nger Nature. All ri ghts reserved. ©2017 Mac millan Publishers Li mited, part of Spri nger Nature. All ri ghts r27ese rJULYved. 2017 | VOL 547 | NATURE | 479