The 1950s are widely recognized as one of the most important decades in biology. This is, in large part, due to the discovery of the structure of DNA and the launch of the field of molecular biology. At roughly the same time, less flashy discoveries were made that were, arguably, just as important. Several polymerases were identified. In 1956, Arthur Kornberg discovered DNA polymerase. In 1959, Severo Ochoa discovered RNA polymerase. And also in 1959, Frederick James Bollum discovered terminal deoxynucleotidyl transferase (TdT).
Although far less famous than its Nobel-winning counterparts, TdT is the only polymerase that performs de novo synthesis of single-strand DNA. “If you give it a bunch of As, Gs, Cs, and Ts, [the enzyme] will randomly stitch those together to make a really long piece of DNA,” notes William Efcavitch, PhD, co-founder and CSO of Molecular Assemblies.
Accordingly, the idea of using a TdT-based process to write DNA has occurred to several scientists since the enzyme’s isolation, an event that predates the idea of using chemical means to synthesize DNA. Indeed, the potential for TdT-based DNA writing has been recognized since the 1960s.
Nonetheless, TdT’s potential has been hard to realize. “It’s a really hard problem,” says Dan Lin-Arlow, PhD, founder and CEO of Ansa Biotechnologies. He explains that the process of making ultra-high-fidelity oligos is very technically demanding. Tiny fractions of a percent can make the difference between a high-quality oligo and one that is useless.
Traditionally, DNA writing has been performed through a chemical process called phosphoramidite synthesis. But the TdT-based alternative is looking increasingly attractive, even though phosphoramidite synthesis sets a high bar. Lin-Arlow, for one, sees opportunity. “We think [phosphoramidite synthesis] has reached a plateau,” he declares, “and we are already way beyond that in our technology.”
Jazzed by a successful DNA build
Leslie Mitchell, PhD, founder and CEO of Neochromosome (Neo) was “so jazzed last week” because the company completed its first DNA build with enzymatically synthesized DNA. The company was born out of the Synthetic Yeast Genome Project (Sc2.0), a collaborative effort to create the world’s first designed synthetic eukaryotic genome. Acquired by Open Trons Labworks in 2021, Neo has a particular focus on building DNA with complex regions, delivering it into cells, and functionally characterizing it.
Mitchell tells GEN that at first, using the enzymatic route was a sort of dream. Then she saw this route become practical enough for Neo to complete projects and turn them into value. (These projects include finding new ways to generate optimized antibodies.) So, Mitchell continues, Neo has become an early adopter of enzymatic DNA synthesis as a platform technology that will allow cells to be engineered more extensively, expanding the future of biotech. She notes that the enzymatic route may overcome limitations that currently exist in DNA writing despite broad access to a range of synthetic DNA and amazing commercialization of those capabilities.
Natural DNA, Mitchell explains, has a lot of complexity to it. For example, it can contain homopolymer runs, repeats, and other features that are hard to replicate in synthetic DNA. It is important, she notes, that newer and different approaches are coming online, above and beyond chemical synthesis and some of the existing technologies to assemble DNA.
Enzymatically synthesizing DNA may help overcome some of the challenges of sequence constraints. Moreover, Neo and other companies that are working to overcome challenges associated with commercially inaccessible DNA are very interested in using alternative starting materials. Doing so could underpin more ambitious DNA builds, expanding sequence space and, therefore, discovery space.
Possibilities beyond DNA writing
Efcavitch could not agree more with Lin-Arlow’s point that the phosphoramidite method has reached a plateau. He adds that, in the past few decades, many applications have arisen that require longer, and higher quality, DNA. An enzymatic approach, he predicts, will facilitate these applications. Moreover, it will allow researchers to imagine applications that are currently beyond the existing chemistry.
Why? Because enzymatic approaches allow for DNA that is significantly longer than 150 nucleotides. And it’s greener and safer: you don’t need the flammable and toxic reagents that come with chemical DNA synthesis.
At the core of Molecular Assemblies’ technology, like Ansa’s, is TdT. But Molecular Assemblies is not working with the same TdT that they started with; the company needed to engineer the protein to get the performance it needs. In collaboration with Codexis, Molecular Assemblies has gone through roughly 1.3 million variants and have modified the enzyme by about 25% to ensure that it works faster and tolerates higher temperatures (to denature hairpins). The company has developed a process in which the TdT adds the nucleotide to the growing DNA strand and a blocking group is removed with a phosphatase.
Right now, Molecular Assemblies plans to provide a DNA writing service similar to those provided by companies that rely on phosphoramidite synthesis. The company is also setting its sights on RNA. The plan is to take advantage of a proprietary enzyme that incorporates RNA monomers.
Although Efcavitch won’t identify the enzyme, he discloses that it is a template-dependent polymerase that “can be tricked” into being a template independent polymerase. Efcavitch adds that a desktop instrument that could write DNA or RNA is not out of the realm of possibility.
When asked if Molecular Assemblies intends to become a DNA writing company or an RNA writing company, Efcavitch answers, “We want to be a writing company.”
Molecular Assemblies has also set its sights set on DNA data storage. But the company realizes that it must first prove itself commercially in DNA writing. Like Ansa, Molecular Assemblies has yet to launch a product. But Molecular Assemblies plans to start a “key customer program” in December.
A loop, an opening, an opportunity
While still an undergraduate, Lin-Arlow attended a class at MIT led by Eric S. Lander, PhD. It was that class, Lin-Arlow recalls, that inspired him to apply his computational interests and background to genomics. In addition to running the class, Lander introduced Lin-Arlow to Vamsi K. Mootha, MD, who was then a postdoctoral researcher in Lander’s laboratory. Lin-Arlow would work for Mootha for the next five years where, he says, he got “the tools bug.” After several more years working on the computer, Lin-Arlow wanted to “pick up a pipette and see how things are done.” A short time after joining the laboratory led by Jay Keesling, PhD, at the University of California, Berkeley, to pursue a doctorate, Lin-Arlow realized that getting his hands wet meant “lots and lots of cloning.”
“It’s 90% cloning and 10% testing [to determine] whether your design has the phenotype you want,” Lin-Arlow says. This procedure seems woefully inefficient to Lin-Arlow, particularly since he comes from a computing background. He complains, “You’re not going to get the next biotechnological revolution on top of all of this artisanal DNA construction.”
Lin-Arlow decided to develop a method to print out whatever DNA constructs scientists need for an experiment. He wanted to create something like Amazon Prime for DNA. “You don’t even think about it,” he explains. “It just comes the next day.”
Lin-Arlow started working on an enzymatic DNA synthesis method in early 2013. In 2015, he was introduced to Sebastian Palluk, PhD, who was then working on DNA synthesis at Technische Universität Darmstadt. (Today, Palluk is the CTO of Ansa.) The two began working together, and Palluk moved to the States. After just a month on the same side of the ocean, the two had a new idea. It was, in Lin-Arlow’s words, a “once-in-a-career ‘Aha!’ moment.”
In this moment, the scientists thought of a new way they could use TdT to incorporate nucleotides. More specifically, they realized that they didn’t need a terminator. Instead, they could use TdT-nucleotide conjugates. Each conjugate would consist of an individual copy of TdT and a single, preloaded nucleotide.
Ansa’s process starts with DNA on a solid support (like a chip). Then, the DNA is exposed to a TdT-nucleotide conjugate, whereupon the tethered nucleotide becomes covalently attached to the end of the primer. At this point, the nucleotide-extended primer is still attached to the whole enzyme, which blocks any other additions from being made. After a wash step, a second enzyme is added to cleave the linker, and the Tdt is released. Now the primer is not just one base longer, it is also ready for another base to be added (Palluk et al. Nat. Biotechnol. 2018; 36: 645–650).
Using the polymerase-nucleotide conjugate, Ansa can make “very long oligos with very high quality,” Lin-Arlow asserts. The company won’t specify how long “very long” might be, nor will it say how expensive the DNA will be. Ansa does claim, however, that it can synthesize oligos that exceed the length of other synthesized oligos “by a considerable margin.”
More details are expected from Ansa next year, which is also when the company plans to start working with early-access customers. Eventually, Ansa will have an online ordering portal where customers can order their DNA. The services Ansa plans to offer will be similar to those provided by companies that already provide custom-made DNA.
The 70-person company is focused on making complex sequences (current DNA writing companies reject a lot of sequences that are complex and difficult). Customers developing gene and cell therapies, for example, struggle to get the DNA they need. This is one area where Ansa hopes to make a difference.
The limitation is in the length. And oligos of a certain length can be made up to a certain quality. Lin-Arlow notes that if you want something longer than that—even if that something is just 500 base pairs—existing vendors have to stitch multiple oligos together. Worse, existing processes are unable to make certain sequences and can fail unpredictably, resulting in delays or outright failure.
Ansa hopes to provide a service that can give scientists any DNA sequence they want (subject to biosecurity prudence). The ability to write DNA, the company explains, is an incredible lever on the whole biotechnology industry. If you can improve DNA writing, Lin-Arlow declares, you have the opportunity to move the needle on many things.
DIY DNA writing
The company DNA Script, which is split between Paris and San Francisco, was not the first company to launch in the enzymatic DNA writing space. But it is the first to have launched a product. Unlike the other companies in this space, DNA Script does not provide a service. Rather, it sells an instrument, the Syntax, so that customers can make DNA in their own laboratories.
One of the biggest advantages of being able to write DNA in-house, notes Thomas Ybert, PhD, co-founder and CEO of DNA Script, is speed. “Of course,” he says, “you can wait for your internet like you were doing like 20 years ago. You will have the same amount of information in the end, but you will have had to wait for it. And waiting has a tremendous cost on the projects.”
The Syntax makes oligos of up to 80 nucleotides commercially, but Ybert says that working in-house, the Syntax can make oligos of up to 350 nucleotides. Ybert adds that there are no theoretical limits and that the company plans to announce soon that it can go even longer. DNA Script is also working on an enzymatic RNA synthesis technology.
One area where DNA Script sees its technology having an impact is in helping artificial intelligence (AI) develop and test new products. Specifically, DNA Script could feed AI the data it needs.
Imagine, Ybert says, designing the next protein that will make spider silk, which is difficult to produce in a stable and inexpensive way. There would be a feedback loop between the AI that is designing the protein, the DNA to make the protein, and the test of the product. In that feedback loop, the wheel can be cranked so fast that new fabrics are ready in weeks.
Another area Ybert points to is personalized medicine. One possibility is giving point-of-care facilities DNA-based biomanufacturing abilities. Ybert says the “ultimate goal of the technology” is to allow an mRNA oncology vaccine to be made for a patient and injected within a day of diagnosis.
It took a long time for the phosphoramidite technology to become a real industry, Ybert notes. But enzymatic technology is maturing more quickly. This is the case, he says, partly because the new technology has more potential and is simply more powerful.
Except for the Syntax, enzymatically made oligos cannot be ordered from a company today. It’s a race to the beginning. But the race is happening in a market already crowded by veteran companies that use more established technology.
There will certainly be room for more than one enzymatic company in this space. And some of the veteran, chemical DNA synthesis companies may befriend the enzymatic companies. In other cases, there will be stiff competition. Either way, the introduction of multiple enzymatic DNA writing options—and the impact they may have on the world of biology—is right around the corner.