Atomic force microscopy (AFM) is a versatile tool that, among many applications, allows researchers to image single molecules, all the way down to the atomic level. This technique produces a topographic map of molecular structure. However, imaging details of the internal structure of a large molecule such as a protein complex is typically not possible with AFM.

In recent work, Ozgur Sahin of Columbia University and his postdoc Duckhoe Kim developed a method that can be applied to peer inside a protein complex by targeting specific chemical groups and tugging on them with an AFM cantilever. They created short, single-stranded DNA probes that are functionalized with a binding moiety on one end. These DNA probes are designed to hybridize with complementary DNA probes tethered to an AFM cantilever. Sahin and Kim designed this cantilever-tethered probe sequence to contain two different regions that hybridize to two different complementary DNA molecules used as the binding probes. By doing this, they could use the different AFM force-time waveforms generated as a result of hybridization to one complementary sequence or another to distinguish distinct binding interactions.

Sahin and Kim initially experimented with the cantilever-tethered probe sequence by testing its interaction with two different complementary DNA probe molecules attached to a surface. “When the AFM detects an interaction, the computer highlights the corresponding pixel in the image with a specific color,” explains Sahin. “We noticed that the highlighted pixels were clustered into regions less than 2 nanometers and more typically less than 1 nanometer.” A careful statistical analysis of the data revealed that their AFM setup had subnanometer resolution. Because they achieved such high resolution, the researchers were encouraged to apply the approach to locate specific chemical moieties within a protein complex.

Chemically specific imaging using complementary DNA labels and AFM. Figure reproduced from Kim and Sahin, Nature Publishing Group.

They put the method to the test with the well-studied streptavidin-biotin complex. They made two biotinylated DNA probes with distinct sequences and designed a cantilever-tethered DNA probe with two regions complementary to the two biotinylated probes. By moving the AFM cantilever over streptavidin complexes on a surface, they could detect locations where biotinylated DNA probes were present. After using some straightforward geometry to calculate distances between biotin-binding sites in a single streptavidin complex, they found that the measured locations of the sites were within 2 angstroms of their locations in the known crystal structure, indicating good agreement.

Sahin notes that because many proteins and complexes are difficult to crystallize, the method could have complementary value to traditional structure determination tools. He believes that the method will be readily adaptable to problems such as studying interactions between DNA or RNA and proteins. With further development, the approach could also be applied to make measurements of distances between amino acids, which in principle could be used to determine protein structure. “I anticipate that chemically linking DNA labels to specific amino acids would be the primary challenge in doing this,” says Sahin.

However, Sahin is more excited about the possibilities of this single-molecule approach in studying protein motions, something that is particularly difficult using crystallography. “The fact that our method works in physiologically relevant conditions makes it an attractive option to probe dynamics of molecules,” he says.