Rewriting Biology

C&I Issue 7, 2020

Read time: 8 mins

By editing or even writing the DNA of microbes, synthetic biologists promise to accelerate the sustainable manufacture of agrichemicals, drug molecules and other valuable compounds, reports Anthony King

Baker’s yeast has been an essential ingredient in winemaking, brewing and bread making since ancient times. In the modern era, Saccharomyces cerevisiae is routinely grown in the lab to study cell division, ageing and cell death. But its genome is also a circuit board for synthetic biologists to tweak and rewire.  

In 2019, Jay Keasling at the University of California, Berkeley, US, rewired baker’s yeast to generate the main cannabinoid compounds from marijuana. Keasling has long sought to modify yeast and bacteria as drug factories, eliminating wasteful or expensive processes. His lab recently inserted more than a dozen genes into yeast, including copies of marijuana plant genes, to make cannabinoids such as mind-altering tetrahydrocannabinol (THC) and non-psychoactive cannabidiol (CBD) (Nature, 2020, 567, 123).

‘It takes a lot of energy – about 3% of California’s electricity – and water to grow cannabis. When you do it in fermentation tanks, the amount of energy is very low,’ says Keasling. The setup also boosts production of rarer cannabinoids, which are uneconomic to produce in plants.

Bacterial factories can also be used to produce potential new drugs. A violet-coloured pigment produced by bacteria, violacein, can kill the malaria parasite in a lab dish and in mice. ‘In its natural environment, the bacteria use the compound to kill competing bacteria,’ says Mark Wilkinson at Imperial College London, UK. Commercial violacein requires costly purification from bacteria that require specialist facilities because they can cause deadly infections. ‘We were able to start from a culture of bacteria and by day four have enough of the compound to test on the parasites,’ says Wilkinson.  

The Imperial group, led by Paul Freemont, modified E. coli to produce 28 different versions of the violacein compound, then tested them on the malaria parasite. Around half were more potent than the parent molecule, with one derivative (7-chloroviolacein) almost 20% more potent (Antimic. Agents and Chemoth., 2020, doi: 10.1128/AAC.02129-19).

Rather than modify an existing genome, Beat Christen at ETH Zurich in Switzerland is interested in authoring DNA from scratch. ‘We can harness the power of biology to develop sustainable solutions for pressing global problems,’ he says.

He is starting with a freshwater bacterium, Caulobacter crescentus, which has been studied for decades and has 4000 genes that are well understood. ‘A bacterial genome is made up of several million DNA letters, arranged in a huge circular molecule,’ says Christen, whose lab developed computational algorithms to simplify the chemical synthesis of the Caulobacter genome. 

Christen and co-workers rewrote the Caulobacter genome and cut the number of genes to 799. ‘We changed the sequence of the entire genome, streamlining it and making it really easy for chemists to synthesise,’ he says. When read three at a time, genetic sequences spell out what is called a codon. Each codon equals an amino acid, but there is redundancy. For instance, any codon starting with GG (GGU, GGC, GGA, GGG) makes glycine. ‘There are millions of ways you can encode a protein into a DNA sequence,’ explains Christen. ‘You have the flexibility to encode the same protein in many DNA sequences.’ This allowed him to leave out DNA sequences that would cause problematic knots or loops, from a synthesis point of view. The design process, however, wrecked the function of around 100 genes, which he is now working to fix.

In the past two decades, it has become possible to sequence all sorts of organisms. ‘We can now think about writing genomes,’ says Christen. ‘We don’t need to copy genes. We can actually write something new.’ You may need 20, 200 or even 500 genes to come up with a desired trait, explains Christen. He envisages constructing a complex trait from a microorganism and writing it into an artificial organism. The advantage of growing synthetic organisms, he adds, is that you can really control and embed particular desired functions.

The notion of synthetic organisms is a scary prospect for some. Christen says they can be designed to require certain nutrients, unlike wild microbes, which will keep them tied to a lab. His lab recently reported a metabolic blueprint for the engineering of novel organisms that would improve nitrogen-fixing capabilities (Mol. Syst. Biol., doi: 10.15252/msb.20199419). More than 125 megatons of nitrogen are fixed each year into ammonia-based fertilisers using the Haber-Bosch process, consuming enormous quantities of fossil fuels.  

Look – no cells

‘We still have a way to go in being able to create, control and re-programme cellular behaviour,’ says Michael Jewett, biological engineer and Director of the Center for Synthetic Biology at Northwestern University, Illinois, US. ‘If I look back 15 to 20 years, we were still mainly editing and writing programmes in just a few genes. It is now more commonplace to be looking at tens of genes.’

We end up with this tug-of-war between what cells want to do, guided by their own evolutionary history, and what we as engineers want.
Michael Jewett Center for Synthetic Biology, Northwestern University, Illinois, US

However, a genetically modified yeast cell does not want to make insulin for us. ‘We end up with this tug-of-war between what cells want to do, guided by their own evolutionary history, and what we as engineers want.’ His solution is cell-free systems. Instead of a cell being converted into a bio-factory, only the molecular components are put to work.  

‘Think of it like opening up the hood of a car and taking the engine out, and repurposing it to do something else,’ says Jewett. This involves ripping away the cell wall of E. coli, and treating the insides as a molecular factory, which is fed raw materials and information, to generate desirable compounds. 

One early example was a paper-based, cell-free diagnostic system for Zika virus, developed by Keith Pardee at the University of Toronto, Canada. He worked with James Collins at Massachusetts Institute of Technology in Boston, US, to take gene circuits out of cells and embed them with the freeze-dried machinery of cells into paper (Cell, 2016, 165, 1255). This set-up usually requires living cells.

Jewett also developed a riboswitch to detect fluoride in water (ACS Synth Biol, 2019, 9, 10). This turns on the expression of a gene, which makes an enzyme that allows for a yellow colour to be produced in the presence of fluoride, a groundwater contaminant. ‘It is as easy as a pregnancy test. You can then decide whether your water is safe to drink within minutes with just a hand-held device,’ Jewett explains. Mammoth Biosciences, Sherlock Biosciences and Stemloop are companies using cell-free diagnostics to identify SARS-CoV-2. 

The market for therapeutic proteins was valued at $93.14bn in 2018 and will grow to $172.87bn through 2022.
More than 125 megatons of nitrogen are fixed each year into ammonia-based fertilisers using the Haber-Bosch process, consuming enormous quantities of fossil fuels.
15 to 20
Fifteen to 20 years ago, cell-free systems relied on high-energy phosphate compounds costing hundreds of dollars/g, but now inexpensive substrates like glucose suffice.
Imperial College researchers tweaked E. coli to produce 28 different versions of violacein, an antimalarial compound. Around half were more potent than the parent compound.
Protein therapeutics

Jewett also advocates cell-free systems as a way of synthesising medicines at smaller scale and facilitating local biomanufacturing.

His group is currently making a conjugate vaccine against a bacterial infection, for less than $5, in less than an hour, on demand. They reported on this on the preprint server bioRxiv (doi:

Sutro Biopharma in the US, meanwhile, separates out cellular components required to produce proteins, from small peptides to complex proteins. It describes this as ‘effectively a liquid handling solution to protein synthesis,’ and uses the platform to manufacture next generation cancer and autoimmune therapeutics. It currently makes two cancer treatments, now in clinical trials, at a cell-free manufacturing facility in San Carlos, California. ‘One of the key advantages of these cell-free systems is speed. You don’t need to go to cell cultures, which can be slow and lead to weeks of development time per iteration,’ says Jewett.  

This can allow for rapid response to an emerging outbreak. Swiftscale Biologics was founded in 2019 on the back of research developed by Jewett at Northwestern and Matthew DeLisa at Cornell University. Its cell-free platform enzymatically modifies complex glycosylated proteins, at speed. Usually, it may take three weeks for manufacturing a batch of mammalian cells to make a therapeutic antibody and many repeats are needed to refine it and optimise production. ‘We can make medicines in 12 to 18 hours using the cell-free approach,’ says Jewett.

Fifteen to 20 years ago, cell-free systems relied on high-energy phosphate compounds costing hundreds of dollars/g, but now inexpensive substrates like glucose suffice. In 2019, Jewett’s lab reported a cell-free system that could generate almost 3g/L of protein – cellular systems in industry usually make between 2 and 8g/L. His group made a commodity organic compound (2,3-butanediol) at around 80g/L in E. coli fed on glucose, with around 10g/L per hour at peak productivity (Metabol. Engin. Commun., 2020, e00114). ‘One to 4g/L for a small molecule is the typical manufacturing volumetric productivity,’ says Jewett. 

Another example involved valinomycin, a natural compound made by several Streptomyces strains, which shows antimicrobial, insecticidal and antiviral effects. After optimising Streptomyces in his lab, Jewett initially achieved yields of mg/L in 2015. Recently, his lab in collaboration with Jian Li’s group at Shanghai Tech University, China, expressed the necessary enzymes in a cell-free system and boosted the yields of this natural product 5000-fold, to around 30 mg/L of valinomycin (Metabol. Engin., 2020, 60, 37).

According to a report by Research and Markets in December 2019, the market for therapeutic proteins was valued at $93.14bn in 2018 and will grow to $172.87bn through 2022. Large-scale commercial fermentation systems are here to stay, but cell-free systems could fill gaps, such as by faster or more local manufacturing of valuable pharma products.

This will include vaccines, but also products that are difficult for cells to make, such as those that are inherently toxic. In a recent example, Swiftscale is using cell-free system to focus on fast manufacturing of neutralising antibody therapeutics for Covid-19, partnering with San Francisco-based Centrivax.

Alien DNA
Famously, all life on Earth is formed by a DNA code made up of four bases, denoted by the letters G, T, C and A. But some biologists are inventing new letters. Floyd Romesberg while at the Scripps Research Institute in California expanded the genetic alphabet to six letters in 2012. Then, he created the world’s first semi-synthetic organism by putting an extra ‘alien’ base pair into E. coli in 2014. Adding two extra letters (X, Y) meant bacteria could generate new kinds of amino acids and proteins. In 2017, his lab reported it had stabilised its semisynthetic organism while it used a six-letter, three-base-pair alphabet.
In 2014, Romesberg set up a company called Synthorx, which began clinical trials in 2019 with a recombinant cytokine with anti-tumour activity; this causes fewer side effects than an existing recombinant cytokine (Proleukin). In December 2019, Sanofi struck a deal to buy Synthorx for $2.5bn.
‘The extra letters in the genetic alphabet means you can write more codons in the lexicon of transfer RNA, which means that you can code more amino acids,’ says Steven Benner, founder of the Foundation for Applied Molecular Evolution. ‘That is worth $2.5bn because it turns out that adding an unnatural amino acid to a pharmaceutical protein, converts a cytokine with all sort of bad side effects to one that didn’t.’
Benner first created ‘unnatural’ bases in the 1980s. In February 2019, he created a synthetic, eight-letter genetic language for the first time, effectively creating four new forms of DNA that could be transcribed into RNA. The aim was to test the limits of molecular information storage that combines Watson-Crick hydrogen bonding with Schrodinger’s requirement for crystal-like universality, constructing an alien genetic system from eight letters (Science, 2019, 363, 884).
By tweaking the four regular DNA letters, Benner added S, B, P and Z. ‘What we were doing here was showing that you understand at a deep molecular level how natural DNA works. You do that by making DNA of your own,’ says Benner, who refers to this kind of work as a grand challenge in design, rather than narrow hypothesis-driven research. Grand challenge research, says Benner, forces a scientist into uncharted terrain, where they are forced to address unscripted questions and solve unscripted problems. Romesberg described the study as a landmark, since it implies that there is nothing special about the four chemicals that evolved on Earth.
Benner warns against returning to the premise that biology is simple, which was pursued by the company GeneX, headquartered in Gaithersburg, Maryland, US, in the 1980s and resurrected in the 1990s after advances in genetic sequencing. ‘They said we just had to learn rules for how you edit amino acids, proteins and we are going to have interchangeable parts,’ says Benner. ‘Proteins are not interchangeable parts. They are not chips on a silicon wafer.’