Enzymatic DNA Synthesis: Going to Great Lengths

ABOVE: In pursuit of longer, higher quality DNA, scientists turn to enzymatic synthesis approaches © istock.com, Dr_Microbee

The demand for DNA is everywhere: cell and gene therapy, vaccine development, protein engineering, data storage, and biomanufacturing.¹ To drive these endeavors, scientists develop approaches for customizing and mass producing synthetic DNA on demand to bypass the limitations and low-yield extraction of naturally occurring DNA. Since the 1980s, one approach has reigned supreme in this field: chemical DNA synthesis, namely the phosphoramidite method.² From primers for driving polymerase chain reaction to guide RNA for CRISPR gene editing, 10-100 bases long, chemically synthesized, short oligonucleotides form a bedrock of life sciences research.³

There’s been this dream for decades. What if we could just print out an entire gene directly, one base at a time?

 —Daniel Lin-Arlow, Ansa Biotechnologies

As scientists set their sights on cell and gene therapies and synthetic biology applications, a growing interest in longer, gene-length oligonucleotides is pushing chemical synthesis to its limits. Over the last decade, scientists interested in synthesizing longer strands of DNA looked to the past for inspiration, where they found a one-of-a-kind molecule. Their efforts culminated in a new approach: enzymatic DNA synthesis.⁴ Although the method is still in its infancy, investors are flocking to a growing list of biotechnology companies, lured by the possibility that this technology will fuel the next big wave of scientific breakthroughs. 

Breaking a Bottleneck in Oligonucleotide Synthesis

Bioengineer Daniel Lin-Arlow, a cofounder and chief scientific officer at the enzymatic DNA synthesis company Ansa Biotechnologies, arrived at Massachusetts Institute of Technology (MIT) in 2002 for his freshman year of college. A self-described computer kid growing up, he was eager to apply his interest in programming to the life sciences. He enrolled in an entry level biology class cotaught by Eric Lander, a geneticist at MIT and one of the leaders of the Human Genome Project. Lander’s teaching style included puzzles to reason with and approach quantitatively, which resonated with Lin-Arlow. Excited, he wanted to work with Lander. Fortunately, Lander was on the lookout for interdisciplinary researchers as well. “He basically begged the computer people in the class to come help them understand all the data that they were creating,” said Lin-Arlow, who joined Lander’s team as a research assistant.

Lin-Arlow got his first taste of being a scientist as he wrote software tools that later drove discoveries about gene regulation. “I find it very satisfying to build tools for scientists to help them get their work done faster,” said Lin-Arlow. This passion continued to shape his career. 

After a stint as a computational biologist in industry, Lin-Arlow returned to academia. In 2012, he started his graduate research studies in the laboratory of Jay Keasling at the University of California, Berkeley (UC Berkeley), where synthetic biology was picking up steam.

“I was really captivated by the idea that you could program cells using DNA to do useful stuff,” said Lin-Arlow. “I really quickly discovered that it just took way too long to get the DNA that we needed for any experiments.” 

         

He needed DNA that was thousands of bases long; he could design these sequences on his computer in an afternoon, but it would take weeks to assemble the DNA. “It was crystal clear to me that if we were going to be able to fulfill the promise of engineering biology, we needed to tighten up that learning loop, and the biggest bottleneck that I could see was getting the DNA,” said Lin-Arlow. 

In 1981, the biochemist Marvin Caruthers and his colleagues at the University of Colorado, Boulder solved a similar problem when they introduced the phosphoramidite method for short oligonucleotide synthesis, the approach that still underlies the manufacturing of most synthetic DNA.² The sequential addition of a new nucleoside phosphoramidite, an analog of an A, T, C, or G nucleotide, generates an elongating oligonucleotide chain that builds atop a solid support system.³ 

“[Phosphoramidite chemistry] is beautiful and it’s really brilliant work by Marvin Caruthers and others in the field,” said Marcel Hollenstein, a biochemist at the Pasteur Institute. However, Hollenstein noted that despite finetuning over the years, the method has an upper limit of generating around 300-bases-long oligonucleotides. 

“The biggest issue is the [coupling] efficiency,” said Max Ryadnov, a chemist at the National Physical Laboratory. The repeated use of harsh chemicals across the cycle increases DNA damage while reducing its purity, so the coupling of a nucleotide analog to the previously added base is never 100 percent efficient.^1,4 With each synthesis cycle, error creeps in. As the oligonucleotide length increases, the yield decreases. This is referred to as the stepwise yield, which provides a measure of how much of the generated material is correct after a certain number of cycles. 

“If you don’t have 99.99 percent, at least, then you will lose in yield [and] you will lose purity,” said Ryadnov. For example, a coupling efficiency of 99.5 percent, which is typical of phosphoramidite synthesis, achieves a 36.7 percent yield after 200 cycles. With a seemingly small dip to a coupling efficiency of 98.5 percent, the yield plummets to 4.9 percent. 

It’s an exception in the polymerase world. 

 —Thomas Ybert, DNA Script

Ultimately, the yield determines how useful the material is. “We need to go beyond what phosphoramidite can do in terms of performance,” said Thomas Ybert, cofounder and chief executive officer at DNA Script, a biotechnology company that develops DNA printers fueled by enzymatic DNA synthesis technology.

With direct synthesis of any material close to the length needed for a gene (averages around 3,000 bases) out of the picture, scientists developed innovative approaches to stitch together many short oligonucleotides to create DNA material up to several hundred kilobases.⁵ Lin-Arlow said that oligonucleotide assembly is still error prone, failing randomly and unpredictably. Furthermore, the method struggles with certain types of sequences, including GC-rich fragments. 

“There’s been this dream for decades. What if we could just print out an entire gene directly, one base at a time?” said Lin-Arlow. 

Chemical Versus Enzymatic DNA Synthesis From synthetic biology and gene therapy to data storage and biofuel production, the demand for synthetic DNA is rising.

Chemical synthesis Phosphoramidite oligonucleotide synthesis is the gold-standard method for generating synthetic DNA. Scientists use the method to assemble modified nucleotides—phosphoramidite nucleosides—with the help of chemicals.1           (1) Chemicals remove the 5’-hydroxyl protecting group. The nucleosides are rich in exposed hydroxyl and amino groups, so scientists use protecting groups to prevent unintended reactions with compounds used in DNA synthesis. Chemicals remove the protecting groups to facilitate nucleoside coupling. (2) Chemicals displace the diisopropylamino group. (3) Chemicals cap unreacted 5’-hydroxyl groups. An oxidation step creates a more stable coupling. The oligonucleotide is ready for another round of synthesis. (4) Chemicals stabilize coupling. Once the oligonucleotide reaches the desired length, additional chemicals remove the remaining protecting groups and cleave the chain from its solid support. (5) Chemicals cleave the oligonucleotide and remove protecting groups. (6) Continued exposure to harsh chemicals damages the DNA and reduces yields.  Scientists use short oligonucleotides as primers for polymerase chain reaction and next-generation sequencing, in DNA microarrays and fluorescence in situ hybridization, and in antisense therapies.

Enzymatic synthesis Enzymatic DNA synthesis is an emerging technology that offers several advantages over chemical methods. Scientists use two main approaches to achieve template-independent oligonucleotide assembly.2           (1) Protected deoxynucleotide triphosphate (dNTP) and tethered dNTP approaches both use terminal deoxynucleotidyl transferase (TdT), a unique and specialized DNA polymerase that does not require a primer template to construct DNA. (2) An incoming protected dNTP or tethered TdT-dNTP conjugate couples to the previously added base.  (3) Mild, aqueous wash reagents. The deblocking or untethering step is followed by a wash step to remove any lingering reagents, including blocking agents, enzymes, and unbound dNTP. The oligonucleotide is ready for another round of synthesis. (4) A cleavage reagent separates the oligonucleotide from the base. Once the desired length is reached, chemicals cleave the oligonucleotide from the solid support. (5) Reactions occur in mild, aqueous conditions, which limits the use of harsh chemicals that cause DNA damage. Therefore, enzymatic approaches can generate longer strands, such as gene fragments, with a lower error rate.

Enzymatic DNA Synthesis Using Terminal Deoxynucleotidyl Transferase

In the phosphoramidite method, waves of chemicals facilitate the coupling between modified nucleosides. In nature, a complex slurry of enzymes orchestrates DNA synthesis. Scientists have previously explored using enzymes to drive de novo synthesis, but they had limited success.^6,7 However, around 60 years ago, a unique molecule with far-reaching influence emerged. 

In the 1960s, Frederick Bollum and Lucy Chang discovered and characterized a DNA polymerase isolated from calf thymus: terminal deoxynucleotidyl transferase (TdT).^8,9 Unlike other known polymerases, TdT doesn’t require a primer or template to start building DNA.¹⁰ “It’s an exception in the polymerase world,” said Ybert. Instead, it adds nucleotides at random to an existing DNA molecule. Specifically, it’s responsible for generating diversity in the T cell receptor, which allows the immune system to detect and respond to a variety of threats. 

The indiscriminate addition of nucleotides made this specialized polymerase a great candidate for enzymatic DNA synthesis, but scientists struggled to control its activity. In his 1962 paper in The Journal of Biological Chemistry, Bollum speculated that scientists could achieve a nonrandom and therefore custom DNA product if they could find a way to wrangle the enzyme to add one base at a time.⁸ Chemically synthesized oligonucleotides, which scientists could guide in a stepwise manner, crossed this threshold first and went on to dominate the field. Quietly in the background, enzymatic synthesis ambled along.⁴ Then in the 2010s, two approaches for regulating TdT-mediated synthesis emerged as viable competitors to the phosphoramidite method. 

In one approach, scientists attach a blocking group to the 3’ end of a nucleotide, which prevents additional bases from binding to the growing oligonucleotide chain.^1,4 Following a wash step to remove the blocking group and any lingering reagents, including TdT and nucleotides, the cycle repeats. Ybert and his colleagues at DNA Script use this approach in their benchtop DNA printers.¹¹ Companies like DNA Script have reported a greater than 99 percent coupling efficiency for oligonucleotides around 300 bases, proving that enzymatic synthesis can match the gold standard chemical method while getting a slight edge on the length.¹² 

Lin-Arlow and fellow graduate student Sebastian Palluk, cofounder and chief technology officer at Ansa Biotechnologies, took a slightly different approach to direct TdT activity. There is some evidence that the enzyme struggles to add nucleotides that carry a blocking group to a growing DNA chain, possibly due to inefficiencies in TdT incorporating the modified structures into its binding pocket.^4,13

“[Palluk and I] had this lightning strikes, once-in-a-career idea,” said Lin-Arlow. Instead of working against the enzyme, Lin-Arlow and his colleagues linked the base region of a nucleotide directly to the enzyme to create a polymerase-nucleotide conjugate.¹⁴ After coupling to the growing DNA chain, the TdT stays put, protecting the 3’ end of the oligonucleotide until reagents wash away extra conjugates and cleave the linker. 

[Palluk and I] had this lightning strikes, once-in-a-career idea

 —Daniel Lin-Arlow, Ansa Biotechnologies

When the linker disconnects, it can leave a chemical residue or a small scar on the nucleobase, meaning that from a molecular standpoint, it is not a pure DNA molecule. This complicates the direct use of the synthesized material without additional processing, so Lin-Arlow and his colleagues developed a modified linker design that limits or avoids the creation of a scar during synthesis.¹⁵

By the time Lin-Arlow and Palluk finished their doctoral research, they had successfully generated a 10-bases long oligonucleotide. “It was not enough to really change the world, by a longshot,” said Lin-Arlow. “[Palluk] always believed that we could make oligos [that are] 1000 bases long or longer, and I think he continues to move the goalposts on that.” 

Based on this proof-of-concept study, the duo founded Ansa Biotechnologies, where they have focused on maturing the platform. Last year, Ansa Biotechnologies announced that they had successfully generated the world’s longest oligonucleotide. The 1,005 base long oligonucleotide encodes a portion of an adeno-associated virus vector, the main vehicle for the delivery of gene therapy, and includes tricky sequences, such as GC-rich regions. To their excitement, the sequence had a 99.9 percent stepwise yield. 

“But can you reach 99.99? That’s a big question still,” said Ryadnov.

“In theory, polymerase coupling should be close to 100 percent. Maybe not yet, but I’m pretty convinced that we will reach that,” Hollenstein said.

Longer, Greener Synthesis of the Future

Lin-Arlow’s goal is to use enzymatic DNA synthesis to directly make genes or gene fragments. Lin-Arlow said that Ansa Biotechnologies can simultaneously synthesize 384 distinct 1,000-mer oligonucleotides, but in the future, they hope to use an array synthesizer to generate millions of oligonucleotides at a time. At that length, they could encode a full-length gene, such as a single chain antibody.

Neither approach for enzymatic DNA synthesis is perfect, but the technology is still undergoing optimization.^1,4 For example, a re-engineered TdT or a completely new enzyme with different binding pockets could improve the elongation efficiency. “The key question there is, ‘Is it the only one? Are there any others?’ Maybe they are there, but we haven’t seen them yet, which is bizarre,” said Ryadnov. 

Enzymatic DNA synthesis is also a greener alternative to chemical synthesis. “[Enzymatic DNA synthesis] decreases the reliance on organic solvents and organic materials,” said Ryadnov. As the demand for synthetic DNA rises, a switch to enzymes could alter environmental health. In a recent study, scientists analyzed the waste generated by each synthesis platform and found that enzymatic synthesis produced one liter of 80 percent aqueous waste, while chemical synthesis generated eight liters of 100 percent hazardous waste.¹²

“The future is quite secure and bright there for DNA [synthesis],” said Hollenstein. “For RNA and modified DNA, this is much less developed.” Getting high-quality, synthetic sequences is currently a bottleneck for RNA-based therapeutics; however, researchers are exploring the potential of enzymatic approaches for oligonucleotide synthesis beyond DNA.^16,17

Enzymatic oligonucleotide synthesis might not replace chemical synthesis overnight. Rydanov noted that it took decades to develop and optimize chemical synthesis, so some scientists might be reluctant to make the switch without an incentive or a clear need (such as longer DNA). Rydanov views the two approaches as complementary: “The marriage of chemical shorter oligonucleotides or longer enzymatically enabled sequences is working, but in principle everything can be done enzymatically.” 

Correction: June 19, 2024: An earlier version of the story stated that Ansa Biotechnologies can simultaneously synthesize up to four oligonucleotides that are 1,000 nucleotides in length. They can actually synthesize 384 oligonucleotides that are 1,000 nucleotides in length at a time. The text has been corrected.