In principle, optical DNA-sequencing protocols have the advantage of reading long strands of DNA in real time and at high speeds. In practice, however, reading long DNA strands is a challenge with current methods, which require high concentrations and suffer from short-chain loading bias. To overcome these limitations, a research team led by Meni Wanunu at Northeastern University in Boston has now developed an efficient voltage-controlled DNA-loading technology that enables single molecule, real time (SMRT) sequencing of long DNA strands at ultralow concentrations.

In SMRT sequencing, the replication of DNA by a single DNA polymerase is optically recorded using fluorescently labeled dNTP analogues. The DNA replication occurs in a zero-mode waveguide (ZMW), the base of a small cavity in an opaque film which allows molecules to freely flow in and out; but, unlike a standard waveguide, which guides light, the ZMW's diameter is too small for light to enter. Since the film and ZMWs are both opaque, the background fluorescence from the dNTP analogues in the surrounding medium is blocked from reaching the detectors. However, since dNTP analogues are integrated into the DNA at the bottom of the waveguide, the fluorescence emission can still exit the waveguide at that location and be recorded.

Artistic rendering of DNA polymerase replicating DNA in a nanopore zero-mode waveguide. Credit: Image: Ella Maru Studio.

Until now, it has been a challenge to pack long strands of DNA into the zeptoliter-sized (a trillion times smaller than a nanoliter) waveguide cavity. For example, a 10 kilobase-pair (kb) DNA strand has an effective diameter, as it flows and bends in liquid, of over five times that of the ZMW. In previous work, the team introduced the nanopore ZMW (NZMW), in which the nanopore enabled the application of a voltage along the length of the waveguide cavity. The strong electric field pulls electrically charged DNA molecules into the NZMW, “allowing the DNA to cross over a huge entropic barrier,” as described by the paper's first author, Joseph Larkin. This field “is actually compressing the DNA,” he notes. The early NZMW technology, which demonstrated electrophoretic DNA packing, relied on a highly fluorescent silicon nitride substrate, and thus it was incompatible with SMRT sequencing.

“We had to replace the silicon nitride with another material,” recalls Wanunu. The team fabricated the waveguides on a 20-nm-thick glass layer deposited on a silicon nitride membrane. Then the silicon nitride was selectively etched, which left a glass substrate and reduced the photoluminescence background by 40-fold. With a nearly background-free substrate, researchers studied the loading and sequencing of DNA in the NZMW.

Not only did the team find a nearly uniform loading probability for DNA strands from 1–48.5-kb long, but it also found a loading efficiency five to six orders of magnitude greater than that possible by diffusion alone. By functionalizing the waveguide base with biotin, streptavidin on the DNA polymerase bound the biotin and held the DNA in the waveguide for sequencing. With the biotin functionalization, DNA strands remained in the waveguide indefinitely; whereas without the surface chemistry, molecules escaped when the voltage was turned off.

The team concluded by sequencing a 3-pM sample of a 20-kb DNA strand with their redesigned NZMW. They demonstrated a 67% single-read accuracy with typical read lengths of 1.6 kb. For future work, “we are thinking about scalability,” explained Wanunu. “Currently each pore is made by hand,” Larkin said. By integrating the waveguides onto a porous substrate, the team hopes to fabricate large arrays of NZMWs for high-throughput SMRT DNA sequencing.