Creating a new gene in a single day could soon be possible, thanks to a new technique that mimics the way the body copies its own DNA. Though the technology needs to clear a few more hurdles, it could one day let researchers speedily rewrite microbe genes, enabling them to synthesize new medicines and fuels on the fly.
“It’s the future,” says George Church, a geneticist at Harvard University who has pioneered numerous technologies to read and write DNA for synthetic biology. “This is going to be enormous.”
Researchers have been able to make DNA since the 1970s. The traditional approach takes DNA nucleotides—the chemical letters A, G, C, and T—and adds them, one by one, to a growing chain called an oligonucleotide, or oligo. But the process, which uses a series of toxic organic reagents, is typically slow and error-prone, limiting oligos to about 200 letters—a tiny fraction of the thousands of letters that make up most genes.
Our cells make DNA differently. A variety of enzymes called polymerases read a single strand of DNA and then synthesize a complementary strand that binds to it. That has prompted dreams of re-engineering polymerases to write new DNA.
Over the decades, most researchers have settled on one particular polymerase, called terminal deoxynucleotidyl transferase (TdT), because, unlike other polymerases, it can attach new nucleotides to an oligo strand without following a DNA template strand. Natural TdT does this to write millions of new variations of genes for antibodies, which the immune system can then select from to target invaders. But the natural enzyme adds new DNA letters randomly, rather than controlling the precise sequence of letters as researchers want to do.
Scientists have tried for years to make TdT add one nucleotide at a time and stop, before repeating the process with a different nucleotide, says Sebastian Palluk, a Ph.D. student who worked on the project in chemist Jay Keasling’s lab at the Lawrence Berkeley National Laboratory in California. They started by adding chemical groups to DNA’s four bases that act as “stop” signals. So, when TdT adds a modified A to an oligo of any length, it would be prevented from adding the next base. The oligo is then fished out, washed, and treated with another compound to cut off the blocking group, readying it for the next extension.
But TdT doesn’t work well with these modified nucleotides. “TdT is very picky,” Palluk says. One such system, for example, required about an hour to add each modified base, far too slow to be practical.
Palluk says he also tried to make this approach work. “I went down that route for 2 years,” he says. But he got to talking with Daniel Arlow, a fellow Ph.D. student in Keasling’s lab who was also trying to use enzymes to synthesize DNA. Eventually, the pair settled on a novel approach. They start with four separate pools for the four separate bases, each one with copies of TdT tethered to either A, G, C, or T. To grow their oligo, they add a base from one of the pools. After TdT adds the base to the end of the oligo, it remains tethered, blocking any additional copies of the enzyme from reacting with the oligo and extending it further. The now oligos are then fished out, and the tethers are snipped off. The free TdT is washed away, and the oligo is ready for the next base to be added.
Ultimately, the approach should be cheap, Keasling says, because TdT is easy to manufacture in bacteria and yeast. It’s also fast. Most new nucleotides attach to the growing oligo in 10 to 20 seconds, Palluk, Arlow, Keasling, and their colleagues report today in Nature Biotechnology. For now, the tether snipping step still takes a minute. So synthesizing a whole gene will still likely take the better part of a day.
Church says the new approach is not quite ready to dethrone conventional DNA synthesis. So far, the group has made oligos only 10 bases long. And there are still a few writing problems, as the approach was only 98% accurate at writing DNA in the desired sequence, below the 99% accuracy of the traditional approach. “It’s cool for a first demonstration,” Palluk says. “But it’s not quite there yet.”
In order to write oligos up to 1000 bases long, the approach will likely need to be 99.9% accurate. If it gets there, Church says it could help revolutionize not just synthetic biology’s efforts to write and test new genes, but also enable efforts to write massive libraries of data in DNA to create a compact archive the firehoses of information coming from giant science projects such as astronomy surveys, which could then be fished out and read out later.
*Correction, 18 June, 3:30 p.m.: An earlier version of this story misstated Jay Keasling's affiliation. While he is a professor at the University of California, Berkeley, this research was conducted in his lab at the Lawrence Berkeley National Laboratory.