BALTIMORE – With almost $26 million in recent Series B financing under its belt, synthetic biology company Molecular Assemblies is gearing up to commercialize its enzymatic DNA synthesis technology through an early-access customer program later this year.
The method, which the San Diego-based company calls Fully Enzymatic Synthesis (FES), involves a proprietary two-step process that deploys enzymes to achieve nucleotide incorporation and 3'-deprotection. The firm claims FES can yield "longer, purer pieces of synthetic DNA" than the existing chemical DNA synthesis methods, although it has yet to show specific performance data.
According to Molecular Assemblies Cofounder and CSO William Efcavitch, enzymatic DNA synthesis requires two main steps, a polymerase adding a nucleotide to the growing oligonucleotide and the removal of a reversible terminator on the 3' end of the last nucleotide to allow the synthesis step to repeat. Similar to existing enzymatic DNA synthesis technologies, Efcavitch said, FES also requires a version of terminal deoxynucleotidyl transferase, or TdT, for nucleotide incorporation. However, instead of using a chemical process to deprotect the nucleotide, FES uses a phosphatase.
Calling the enzymatic deprotection "a deliberate strategy," Efcavitch said a key differentiator between the FES technology and others is that it involves an "extremely novel" and patent-protected cleanup step.
"Unless your cycle efficiency is 100 percent, you end up with some byproducts during the synthesis," he explained, adding that purity is especially important when it comes to synthesizing oligos that are 150 nucleotides in length or longer. Since the 3' reversible terminator used in FES is only removable with an enzyme, he said, Molecular Assemblies scientists were able to devise a built-in cleanup step to get rid of those unwanted materials.
The company also boasts that FES can produce especially long DNA sequences. "We pride ourselves as [making] 'longmers,'" Molecular President and CEO Michael Kamdar said, referring to the company's ability to produce oligonucleotides of 150 nucleotides and above. However, he declined to disclose the length of the longest oligo the company has successfully synthesized to date. "The reality is, it's a good question, but it's not the question," Kamdar said, adding that "what you have to look at is all the other variables," including incorporation efficiency and cycle time.
Meanwhile, Efcavitch said the turnaround times for the enzymatic incorporation and deprotection steps "are numbers that we don't give out at this point in time."
"Customers don't care about cycle time," he added, "what they want to know is the total turnaround time to get the product in their hands." To that end, with its built-in cleanup step, FES has "shortened some of the back end" by obviating the need for post-synthesis purification, he said, although he did not reveal the total turnaround time. In addition, Molecular Assemblies did not provide further information regarding the error rate for FES.
According to Aaron Hammons, head of the life science tools business at Codexis, Molecular Assemblies' collaborator and investor, the coupling speed — or the speed of conjugating one nucleotide to the acceptor — of the TdT enzyme used in FES has been improved from three and a half hours to less than 90 seconds with engineering. A protein biotechnology company based in Redwood City, California, Codexis has invested in Molecular Assemblies' Series A and B financing, according to Hammons, and has engineered the current TdT enzyme used in the FES method. Although the proprietary TdT enzyme is "exclusive to Molecular Assemblies," Hammons made clear that Codexis holds intellectual property for the enzyme and that "there is no joint IP in the venture."
Mirroring similar opinions in the field, Hammons also believes that cycle speed and error rate are important for enzymatic DNA synthesis to commercially compete with traditional phosphoramidite-based methods. For de novo synthesis of a large nucleic acid fragment, he pointed out, "if each conjugation takes three and a half hours, you're gonna be there a while." As for the error rate, Hammons said the conversion efficiency of the TdT enzyme was "well above 99.5 percent" and can "compete easily with phosphoramidite chemistry" after "over 40 rounds of evolution."
Hammons said Codexis was not involved in designing Molecular Assemblies' phosphatase for deprotection, so he did not know specifics about that enzyme. But he said the TdT enzyme was engineered with the deprotection step in mind, adding that generally when both enzymes are involved, there is a wash step in between.
"Whether an enzyme is involved [with deprotection], you still have to provide a buffer with a particular pH," said Thomas Ybert, CEO and cofounder of DNA Script, the French synthetic biology company that, ahead of Molecular Assemblies and other companies, has already launched a commercial platform for its enzymatic DNA synthesis technology.
According to Ybert, DNA Script has also investigated different options for deprotection, including using enzymes, during R&D. "The logical advantage of using an enzyme for the deprotection is to achieve faster and better completion of the [process]," he argued. "But if you can achieve those without enzymes, it's even better."
For one, he said, cutting down the use of enzymes is "a great opportunity" to decrease the synthesis cost, since they are "among the biggest" line items for cost. "Because the protection occurs at every cycle," Ybert added, "if you have an enzyme for the deprotection, that means that you have to use this enzyme at every cycle," swelling the overall production costs.
DNA Script's own deprotection method deploys "a buffer with a particular pH condition" that "anyone could buy from the regular reagent provider," Ybert said. He added that although it is impossible for any method to achieve 100 percent success rate, for DNA Script's approach, nucleotides that remain protected are below the level of detection. In terms of deprotection speed, he said that the company's approach can currently complete full deprotection in less than 30 seconds. "You want to achieve the full cycle, including the deprotection step, as fast as possible," he stressed. "If the deprotection takes a long time, it become very difficult [from] a practical standpoint to do it."
Ybert said the current total turnaround time for DNA Script to synthesize a 200-some-base oligo takes a bit more than a full day. Overall, the company is able to achieve 99.6 percent per-cycle efficiency for the entire DNA synthesis, which he claimed exceeds the current cycle efficiency of phosphoramidite-based chemistry. When it comes to synthesis length, he said the company can make 300-nucleotide oligos at this point, with the highest internal record of around 360 nucleotides.
However, he noted that for gene building — which many synthetic biology companies, including DNA Script and Molecular Assemblies, are aspiring to accelerate — "the [DNA] length that you can achieve is part of the equation, but the equation is more complex."
"Of course, long fragments can help in some aspects of gene building," he said, "but shorter fragments can help as well." What equally matters, he explained, is the efficiency to stitch these synthesized DNA fragments together, whether it is with long or short building blocks.
With the money raised in the Series B, Molecular Assemblies plans to start an early-access program and deliver enzymatically synthesized DNA to customers, likely in the third quarter of this year, Kamdar said. The goal for the program is not only to receive feedback on the technology but also to turn the early adopters into customers, with the eventual goal to fully commercialize the technology in 2023, he said.
Although Kamdar did not disclose who the key customers are, he said they are "industry leaders in their respective fields," including CRISPR, vaccine development, and agriculture.
But unlike DNA Script, which primarily focuses on building enzymatic DNA synthesis benchtop instruments, Molecular Assemblies is "focused on a service model, just like the way you would get chemical oligos," said Kamdar.
"We're primarily after consumers who want a lot of synthetic DNA," said Efcavitch, adding that "when someone comes to you and says, 'we need hundreds of thousands of long oligonucleotides,' you're not going to do that on a benchtop synthesizer."
"You can never say never," he added. "But we'll crack the market with a service model."