Harvard Researchers Encode 169 Bytes of Text in DNA Using a Semiconductor Chip

Researchers at Harvard University have developed a semiconductor chip capable of simultaneously synthesizing 64 distinct DNA sequences, leveraging an enzymatic process in an aqueous environment and finely controlled electrical currents. The study, published in Nature Electronics, sets a new benchmark for enzymatic DNA synthesis and identifies protection chemistry, rather than electronics, as the next technical hurdle to overcome.

Synthetic DNA is a fundamental tool in modern biology and medicine; it is used in diagnostics, genomic engineering, and cancer research. The industrial production of custom DNA is currently primarily based on phosphoramidite chemistry, a well-established process capable of generating millions of sequences in parallel but dependent on toxic organic solvents and centralized facilities.

Enzymatic DNA synthesis is emerging as a more sustainable alternative, more akin to the biological mechanism by which cells replicate their genetic material and potentially suitable for more compact and accessible tools. Up until now, however, this methodology had proven incapable of approaching the productivity of phosphoramidite chemistry: the practical limit was about a dozen sequences that could be synthesized simultaneously.

The team led by Donhee Ham has pushed that limit to 64 distinct sequences, each up to 39 nucleotides long. The chip features 64 synthesis sites on its surface, each structured with two concentric ring electrodes surrounding the DNA strands anchored at the center. DNA synthesis occurs one nucleotide at a time: after each addition, a temporary blocking group must be removed before the next nucleotide can bind. This step, known as deprotection, is triggered by an acidic environment, that is, a low pH.

To spatially control this reaction, the chip's electronics inject current into the internal electrode of a chosen site, generating protons that lower the localized pH around the DNA strands and thus enable their enzymatic elongation. At the same time, the external electrode absorbs the diffusing protons, preventing the low pH area from expanding to adjacent sites. By repeating this operation site by site in each synthesis cycle, the chip simultaneously constructs 64 different DNA sequences.

The hardware platform wasn't initially designed for this purpose. The electronic chip had been developed in Ham's lab by Jeffrey Abbott, then a doctoral student, for intracellular recording of neurons at a population scale. Modifying the surface electrodes then allowed the group to extend the same electronic architecture to DNA synthesis.

"A distinctive feature of the chip was the precision current injection, which we used to permeabilize neuronal membranes for intracellular access," said Ham. "At some point, we wondered if that same current control could be redirected from cells to molecules—replacing the electrodes aimed at neurons with pairs of ring electrodes capable of localizing the pH for DNA synthesis. It worked."

In addition to shorter-term biological applications—synthetic biology, diagnostics—the team explored the implications for digital data storage on DNA, a field requiring synthesis volumes far exceeding current capacities. Using the 64 produced sequences, researchers encoded a 169-byte text message, demonstrating the feasibility of the principle on this platform.

"Data storage on DNA requires synthesis to operate at a scale far beyond current demands," stated Woo-Bin Jung, co-first author of the study. "That's why enzymatic synthesis in water can make a difference. If one could synthesize many more than 64 sequences in parallel, it might offer an ecologically sustainable path to writing DNA on a very large scale."

To assess potential improvements, the group attempted to increase the density of synthesis sites on the chip, bringing them closer together. The attempt was unsuccessful, but it produced one of the most relevant findings of the entire work. Through systematic experiments, researchers established that the issue wasn't in the chip's electronic architecture, which correctly localized pH at the selected sites, but in the deprotection chemistry employed in the study. The low pH does not directly remove the blocking group from the DNA; rather, it generates intermediate molecules that perform deprotection, and these molecules can diffuse to nearby sites, undermining the spatial confinement that the electronics had ensured for the pH.

"The chip did what we asked it to do: it localized the low pH at the selected sites," said Han Sae Jung, co-first author of the study and postdoctoral researcher at Harvard. "The limit came from the deprotection chemistry, not from silicon. This clearly indicates the next step for the field: to develop a more directly acid-driven deprotection chemistry capable of keeping pace with the chip."