The Long and Winding Road: On

Jan 26, 2024

Accessibility to DNA writing technology is growing thanks to the commercialization of these techniques through services ranging from benchtop DNA printers to custom synthesis

Credit: CIPhotos/Stock/Getty Images Plus

The state of DNA synthesis is a bit like humanity’s journey to the moon—just because something has been achieved before doesn’t mean we should cease figuring out how to do it better.”

That is how Harold P. de Vladar, CEO and founder of long-synthetic DNA service provider Ribbon Biolabs, sees the vast opportunity in the DNA synthesis space. de Vladar says that the DNA synthesis equivalent to the Apollo 11 moon landing was Craig Venter’s landmark work on assembling the 5386 bp bacteriophage ΦX174 from short single strands of synthetic, commercially available DNA known as oligonucleotides.1 Venter and his team at the J. Craig Venter Institute built the ΦX174 genome using an adaptation of polymerase chain reaction (PCR) called polymerase cycle assembly2 in 2002–2003—just after the sequencing of the human genome in the early 2000s.

This problem still existed until a few years ago. “When I was still a researcher, I confronted the problem of making long DNA by wanting to synthesize a library of small phage genomes, around 4000 bp, and that was an impossible project. I wanted like 250 sequences, and this was unachievable,” de Vladar said.

About 10 years ago, de Vladar saw that a company called Twist Bioscience was starting to appear in the news. He came to the realization that people were still interested in synthesizing long DNA. “Superficially, synthesizing DNA seemed solved, but when you start looking into the details, how the cost per base per has dropped for sequencing and so on, you realize that this was far from solved” (Box 1).

The cost of sequencing has plummeted for the past 10–15 years at a pace that has outstripped Moore’s law. The cost of a human genome sequence decreased from an estimated $1 million in 2007, to $1000 in 2014, and today it is nearing $100, with the Illumina NovaSeq X series and Ultima Genomics UG100 claiming costs of $200 and $100 per genome, respectively. But the affordability of custom DNA synthesis has not kept pace and the cost remains relatively high. For the past several years, there has been a general trend toward gene synthesis costs falling to $0.01 per bp, as opposed to less than the $1–2 per Gigabase offered by the latest next-generation sequencing tools.7

In 2018, de Vladar launched Ribbon Biolabs in Vienna, Austria. The DNA synthesis company can reliably construct 1-kb fragments in a day, which can be assembled into larger molecules of 10–20 kb. To do so, Ribbon algorithmically processes a sequence into smaller pieces that get sorted into a decision tree to identify the best combination of 10- and 12-mer oligos (from their biobank of nearly 80,000 oligos) for automated assembly. The true turnaround time, including sequencing-based verification and delivery, for the small company (with just a few dozen employees), is ∼3 days. The goal for de Vladar and Ribbon Biolabs is to scale the gene synthesis from the library approach to 1000 kb per day.

By improving enzymes and assembly standards, the accuracy and number of DNA molecules that can be combined in a single step have improved. In the past decade, a plethora of powerful new DNA assembly methods—including Gibson Assembly3—have been developed. These assembly methods have been applied in the construction of a minimal bacterial genome4 and synthetic yeast chromosomes.5

Assembly approaches to DNA synthesis have several drawbacks: they can be compromised by impurities of synthetic oligonucleotides, the dependence of the method on sequence confirmation from an individual clone, and the reliance on high-fidelity proofreading PCR enzymes, which must be used to copy constructed genes to prevent mutations during amplification.

Ribbon Biolabs’ tiered approach is limited by the physical size of the synthesized molecules. “DNA that’s 20–40 kb is a really complex molecule,” said de Vladar. “The reactions are not that efficient anymore. But that’s not even the worst of the problem: longer DNA molecules break due to sheer force. The assembly itself isn’t the issue; it’s the manipulation of the DNA molecules. It’s a problem we’re also trying to solve, and we have some ideas in R&D. But it’s not going to just print 30 kb at the press of a button.”

Furthermore, not everyone agrees that assembly is DNA synthesis. “When we talk about DNA synthesis, we have to be a little cautious because some people talk about oligonucleotide synthesis and other people talk about gene assembly,” said Sylvain Gariel, cofounder and COO of DNA Script, a company that chemically synthesizes DNA.

Since the structure of DNA was first solved 70 years ago, substantial milestones have been achieved, paving the way for the DNA synthesis industry. For the past 40 years, scientists have taken steady steps to figure out the chemistry behind the step-by-step creation of DNA, nucleotide by nucleotide. Chemical methods were developed to reliably provide short DNA chains, typically <200 nucleotides (oligonucleotides). These methods were made to work best with automatic synthesizers, which are now needed for gene engineering and sequencing.

Next, scientists devised methods to create DNA that is longer and more complex than oligonucleotides by using enzymes with or without DNA templates. For instance, the basic technology for DNA Script is template-independent enzymatic oligonucleotide synthesis. Companies have turned these chemical synthesis methods into products and services, such as benchtop DNA printers and custom synthesis, which make DNA synthesis possible for people who are not experts.

On March 9, 2023, Ansa Biotechnologies, Inc. reported that they had successfully made the world’s longest DNA oligonucleotide in a single synthesis. The 1005-base sequence codes for a key part of an adeno-associated virus vector used to develop gene therapy. It has complex features, such as strong secondary structures and a high GC content, that make it very hard to make with traditional methods that require putting together shorter oligonucleotides.

As a metabolite engineer, Gariel was limited by the acquisition of DNA constructs. He had to wait weeks or even months to obtain long stretches of DNA for tens or even hundreds of genes. At DNA Script, Gariel is making strides to commercializing a benchtop DNA synthesis device called Syntax to bypass this bottleneck (Figure 1).

“Our end goal in the market was, if I can do it on my bench stuff, I’m suddenly cutting out all the times that the service provider requires,” Gariel explained. “I’m not giving control to a service provider as long as I have an instrument and the reagents that I need to run it. I can do it on the benchtop without any specific infrastructure requirements other than a molecular biology space and an electric plug.”

The Syntax instrument is a benchtop DNA synthesizer similar in size to the HiSeq sequencer from Illumina. This synthesizer can generate 60-bp oligonucleotides in a pure form for immediate use within 6 h.

To make DNA synthesis safer and more accessible, DNA Script has addressed the danger posed by phosphoramidite chemistry (Box 2), both to the user and the environment. Gariel notes that this is why this chemistry comes with strong constraints in terms of infrastructure and dedicated labor.

Advances in solid-phase synthesis—the synthesis of chemical compounds whereby the reactant molecule is chemically bound to an insoluble material and reagents are added in the solution phase—inspired the groundbreaking development of phosphoramidite chemistry for DNA synthesis in the 1980s by Marvin H. Caruthers of the University of Colorado Boulder.8 In the 1980s, the first automated DNA synthesizer resulted from a collaboration with Caruthers, and was based on Caruthers’ work elucidating the chemistry of phosphoramidite oligonucleotide synthesis. Applied Biosystems used this method to make the first automated DNA synthesizer. This made it easier for people to get synthetic oligonucleotides. But the phosphoramidite method of making DNA has some problems, such as phosphoramidite that does not stay stable on the bench, the need to use a lot of organic solvents, and the inability to make poly-repeat sequences.

Synthesizing, processing, and purifying oligonucleotides is a labor-intensive process that is still mostly done by service providers. Therefore, synthesis capabilities have become centralized within specialty reagent manufacturers. Leading companies such as Agilent Technologies, GenScript, Integrated DNA Technologies, ThermoFisher, TriLink, Dharmacon, Twist Bioscience, and others make custom DNA (and RNA) on demand in a variety of formats. There is a range of instruments for users who wish to decrease the lead time for such services, such as Cytiva’s ÄKTA oligonucleotide synthesizers, which can be purchased and operated on a daily basis.

In the past, molecular biology used short DNA sequences such as primers for polymerase chain reaction or probes for molecular detection, amplification, and change. Now, scientists are looking for longer sequences with different parts, such as whole genomes with accuracy down to a single base, which must be put together from scratch. The phosphoramidite method cannot work with such long sequences because it is less effective at making pure DNA after about 200-bp oligonucleotide sequences.

Using a different method, longer sequences must be put together from smaller strands in stages where mistakes are fixed. In this way, enzymatic approaches are most appealing because they can be used on a large scale, are able to target specific groups, and are good for the environment. Enzymes can mediate mismatch recognition, enabling the selective annealing of complementary strands, reduce the number of steps in each elongation cycle by eliminating the need for coupling reagents, and decrease the dependence on organic solvents. Enzymes can promote synthesis with or without DNA templates, through amplification or the synthesis of de novo sequences.

“Those chemicals are incredibly harsh—they’re volatile and carcinogenic,” said Gariel. “Typically, you don’t want to deal with them in the lab. We designed an enzymatic DNA synthesis (EDS) technology that is water-based, which becomes incredibly versatile and easy to automate.”

Matthew Hayes, chief technology officer and cofounder of Evonetix, says that most attempts to make DNA have focused on shrinking the technology. Evonetix has taken a very different approach. “We’ve asked ourselves, ‘What’s wrong with DNA synthesis?’ ‘Why is it fundamentally difficult to make long pieces of DNA, and can we come up with a solution that solves that from the ground up?’” Hayes explained.

“There are people today who can make machines that will make a small number of high-yield oligos and PCR primers using column-based synthesis. But if you want to synthesize an array of low molecular quantity but a high diversity pool, you’re pretty much restricted to service providers.” Evonetix plans to give users instruments that can synthesize a pool of oligos using semiconductor-based chips (Figure 2). “We’re also able to instruct our machine to assemble that pool into a much longer double stranded DNA template,” Hayes said.

This is not the first time that benchtop DNA synthesizers have hit the market. Emily Leproust, CEO of Twist Bioscience, remembers when people stopped buying benchtop DNA synthesizers. “I started my PhD in DNA synthesis in 1996, and we bought one of the last decentralized desktop oligo synthesizers in 1998,” Leproust told GEN Biotechnology.

“I’m from a generation that remembers when people stopped buying DNA synthesizers. The reason why they stopped is that it was just cheaper and faster to send it out. A lot of people have forgotten that we used to have the benchtop DNA synthesizer, and that market just died because the centralized approach is faster and cheaper.”

Among the several thousand Twist customers, Leproust receives a few requests each year asking if users can get the chip and DNA synthesis machine in their laboratory, to which she would oblige. “There are maybe a few niche applications where you are willing to pay a lot more to have the synthesis done onsite,” said Leproust. “But if it’s onsite, it’s going to be a lot more expensive—you will never beat the price in a centralized facility.”

Leproust says Twist has “unlimited scalability. When I had my desktop DNA synthesizer in graduate school, I could make eight oligos. Now, the desktops can do around 96. But what if you want to make 500 oligos? You have to run the instrument five times in a row. If you want 10,000 oligos, you have to run it 100 times in a row! Our range is super flexible.”

DNA synthesis companies such as Twist offer the ability to make complex and large amounts of DNA quickly at prices that are finally becoming more affordable. Twist’s silicon-based high-throughput gene synthesis platform makes high-quality gene fragments for as little as $0.07 per bp and perfect clonal gene sequences verified by Next-Generation Sequencing for as little as $0.09 per bp.

“Our technology can make millions of oligos at the same time, up to 300 bases with a quality of about one error in 2–3000 bases, and we can assemble up to 5000 bases in a construct,” she said. “If you want fragments of 1 kb in length, we can do that for $70 and ship them to you in four days, and we can make more than 1 million of these 1000-base fragments a year.”

Although Leproust thinks that Twist’s technology can push past 300-base oligos, the synthesis of the oligos is relatively slow compared with stitching them together. “If we were to make 1,000 bases all by synthesis, it would not be faster than what we shipped today in four days and you’ll lose out on speed, quality, and cost,” said Leproust. “There’s some vanity in saying, ‘I can make a 1,000-bp fragment.’ But if it’s more expensive, slower, and of lower quality, that’s not what the customer wants. We’re more into what moves the needle for the customer. I’m not into running science experiments. We are a high-throughput, high-scale, and low-cost operation with a very nice gross margin, and we have to deliver revenue growth quarterly.”

Twist is working on several fronts to improve their DNA synthesis as a service. One dimension of Twist’s R&D is for the silicon chip. Leproust said that they are trying to create higher density DNA synthesis. “We can make a million oligos on the current chip, which is the size of a big cell phone or a 96-well plate,” said Leproust (Figure 3). “The new chip is 24 times smaller, about the size of a postage stamp, and we can make 256 million oligos on these. That’s an improvement in amount and cost over 6,000 × . Our roadmap for miniaturization and to lower the cost is to go from 256 million to 3 billion, then 10 billion, then 50 billion oligos per chip.”

The second dimension pertains to speed. It currently takes 4 days from ordering to receiving 1000 bp synthesized DNA on the FedEx truck. Leproust thinks Twist can probably cut that in half to 2 days. A third consideration is the length of the gene fragments. Twist’s current maximum is about 1.8 kb, but Leproust wants to stretch that to 5 kb. The fourth aspect is the mass that the customer needs. Leproust says some customers want femtomoles, whereas others might need 100 kg. Twist is on the low side of this range but is working on extending the quantities that customers want.

In an increasingly crowded field, the best way to take the lead in DNA synthesis is to make things orders of magnitude better than the state of the art. That is why Michael Chen, CEO and founder of Nuclera, decided to pivot. “We didn’t see a trajectory where enzymatic DNA synthesis would result in a 10 × or 100 × improvement in something for the customer.”

Chen and his cofounders settled on a different question: What is the bottleneck in research today? The key bottleneck for all the companies in the EDS space, Chen argues, is actually making the proteins to write DNA.

Trained as an X-ray crystallographer, “people often asked me what the bottleneck in the process was?” Chen said. “The natural assumption is that it’s the crystallography, which took me 3–6 months to learn. But what took me years was just getting good protein in the first place to carry out my PhD project! That sounds a lot like drug discovery, where researchers are spending months, maybe even years, to get target proteins. Today’s protein scientists are in the lab for weeks performing an incredibly labor-intensive process to grow, manipulate, break, and purify cells. It’s a multi-billion-dollar industry, so it’s a huge opportunity.”

Chen saw an opportunity to take Nuclera’s core digital microfluidics technology and collaborate in a strategic partnership with ePaper display creator E Ink. “The supply chain and research to make ePaper displays have a lot to do with the type of droplet automation technology that we’re working on and commercializing at Nuclera. We’ve made [the eDrop lab-on-a-chip technology] to move thousands of droplets in a programmable fashion in what we call a ‘pipette and forget’ process. The researcher pipettes directly into the cartridge and then can walk away.”

Whether using assembly or chemical oligonucleotide synthesis methods, the generation of long DNA is a reality and has a growing list of biotechnological applications beyond making proteins and synthetic life. DNA synthesis can be used to make vaccines, gene therapies, DNA data storage, DNA origami (Box 3), nanobots,6 and plants that are resistant to climate change. This has led to a huge demand for mass production of long synthetic DNA and, with it, a surge in the development and commercialization of DNA synthesis techniques.

DNA origami consists of using short oligos (“staples”) to guide the folding of a long (“scaffold”) strand and enable the assembly of complex 2D and 3D structures, including curved nanostructures. In 2012, DNA origami was used to put together complex nanostructures such as gated transmembrane channels.9,10

That same year, an approach for the design of complex 3D DNA structures was reported based on the assembly of short oligonucleotides (without a scaffolding strand) termed DNA bricks.11 The key to this approach lies in linking structural units (8-mer) to physical location, in effect creating an eight-nucleobase address for each position in a discrete 3D space. The use of unique addresses made it possible to assemble 3D structures based on predesigned sets of short DNA oligomers (“bricks”)—each DNA brick a 32-mer, designed as four contiguous eight-base domains targeting adjacent physical addresses, reminiscent of a 2 × 1 LEGO® brick.

Although not all solutions have been released on the market, long-range DNA synthesis is clearly possible and improving in terms of price, speed, and quality. And there is an increasing number of solutions in all customer segments that seek to shape the future of biotechnology using long synthetic DNA.

The development of DNA synthesis technologies may also be relevant in materials science and nanotechnology. Nucleic acids are a remarkably versatile source of novel materials and nanodevices. Design and evolution combined with the sub-angstrom resolution offered by nucleic acid scaffolds may enable not only the assembly of intricate structures but also confer them with properties such as tight ligand binding, catalytic activity, and different conformational states. Chemical modification of nucleic acids, and in particular wholesale replacement of the natural framework with synthetic genetic polymers (xeno nucleic acids [XNAs]) capable of evolution, has the potential to create novel structures and devices as well as greatly extend their functionalities—for instance, generating XNA devices that are stable in vivo and that can be used in diagnostic applications capitalizing on fluorescence or catalytic properties.

Ongoing work to expand the chemistry that is compatible with enzymatic synthesis and evolution will add to the expanding palette of XNAs available for the construction of XNA nanostructures and devices enabling novel functions and improved performance in challenging environments and conditions.

References

1. Smith HO, Hutchison CA, Pfannkoch C, et al. Generating a synthetic genome by whole genome assembly: φX174 bacteriophage from synthetic oligonucleotides. Proc Natl Acad Sci U S A 2003;100(26):15440–15445; doi: 10.1073/pnas.2237126100 Crossref, Medline, Google Scholar

2. Stemmer WPC, Crameri A, Ha KD, et al. Single-step assembly of a gene and entire plasmid from large numbers of oligodeoxyribonucleotides. Gene 1995;164(1):49–53; doi: 10.1016/0378-1119(95)00511-4 Crossref, Medline, Google Scholar

3. Gibson DG, Young L, Chuang R-Y, et al. Enzymatic assembly of DNA molecules up to several hundred kilobases. Nat Methods 2009;6(5):343–345; doi: 10.1038/nmeth.1318 Crossref, Medline, Google Scholar

4. Casini A, Storch M, Baldwin GS, et al. Bricks and blueprints: methods and standards for DNA assembly. Nat Rev Mol Cell Biol 2015;16(9):568–576; doi: 10.1038/nrm4014 Crossref, Medline, Google Scholar

5. Mitchell LA, Wang A, Stracquadanio G, et al. Synthesis, debugging, and effects of synthetic chromosome consolidation: synVI and beyond. Science 2017;355(6329):eaaf4831; doi: 10.1126/science.aaf4831 Crossref, Medline, Google Scholar

6. Douglas SM, Bachelet I, Church GM. A logic-gated nanorobot for targeted transport of molecular payloads. Science 2012;335(6070):831–834; doi: 10.1126/science.1214081 Crossref, Medline, Google Scholar

7. Wetterstrand KA. DNA Sequencing Costs: Data. 2022. [Last accessed: March 1, 2023]. Google Scholar

8. Beaucage SL, Caruthers MH. Deoxynucleoside phosphoramidites—A new class of key intermediates for deoxypolynucleotide synthesis. Tetrahedron Lett 1981;22(20):1859–1862; doi: 10.1016/S0040-4039(01)90461-7 Crossref, Google Scholar

9. Langecker M, Arnaut V, Martin TG, et al. Synthetic lipid membrane channels formed by designed DNA nanostructures. Science 2012;338(6109):932–936; doi: 10.1126/science.1225624 Crossref, Medline, Google Scholar

10. Burns JR, Stulz E, Howorka S. Self-assembled DNA nanopores that span lipid bilayers. Nano Lett 2013;13(6):2351–2356; doi: 10.1021/nl304147f Crossref, Medline, Google Scholar

This article was originally published in the April 2023 issue of the GEN Biotechnology journal. GEN Biotechnology, published by Mary Ann Liebert, Inc., is a marquee peer-reviewed journal publishing outstanding original research and perspectives across all facets of the biotech industry.