1) Bacterially derived Protein G/A as a secondary probe for DNA-PAINT imaging
We looked for small, affordable, and universal probes for labeling the primary antibodies to be used as binders against the proteins of interest. In this regard, we utilized commonly known antibody binders derived from bacteria called Protein G and Protein A. These proteins are used for purifying the antibodies from serum samples due to their high specificity. We hypothesized that Protein G/A can be used as secondary binders against primary antibodies. We developed protocols to conjugate Protein G/A with a short single-stranded DNA. These DNA conjugated Protein G/A could bind to primary antibodies with high nanomolar affinity, allowing us to replace secondary antibodies as labeling probes. Since Protein G/A are much smaller compared to secondary antibodies, they introduce much shorter linkage errors. To demonstrate the applicability of Protein G/A as a secondary binder, we conjugated Protein G/A with single-stranded DNA which we used as a docking strand in DNA-PAINT imaging (Figure 1). For labeling the cellular targets, we first incubated the chemically cross-linked cells with primary antibodies and then with DNA conjugated protein G/A. By using DNA-PAINT imaging, we demonstrated that Protein G/A improves the localization precision. We imaged both primary-secondary antibodies and primary antibody-Protein G/A complex. We found that on average Protein G/A labeling delivers better performance with reduced linkage error of around 10 nm compared to primary-secondary antibodies. We also tested the Protein G/A as a binder against surface receptor protein EGFR and found that it localizes with better precision than primary-secondary antibodies alone.
2) DNA-Origami platform for quantitative evaluation of labeling probes:
Super-resolution imaging requires high affinity probes against targets of interest that introduce minimal artifacts with high specificity and labeling efficiency. In order to quantify the target species, the stoichiometry of labeling probes on target species needs to be delineated. Several different labeling probes have been introduced or are in the development that have high potential to improve the quality of super-resolution imaging. However, we currently lack a toolbox to quantify the labeling probes. In this project, we developed a DNA origami-based platform decorated with antigens at specified positions enabling us to quantify the binding stoichiometry and achievable resolution of labeling probes (Figure 2). We immobilized the origami nanostructures with antigens on PEG-passivated glass surface using biotin-neutravidin-biotin interactions. We then tested the binding efficiency of antibodies and nanobody against the respective antigens. We found that commonly used primary-secondary antibodies result in compromised labeling efficiencies compared to high affinity nanobody. In addition, our results indicated that antibody-based labeling introduces large clusters for single antigens which could be misled as interactions of multiple antigens. On the other hand, we found that high-affinity nanobody labeling results in a predictable cluster size and stoichiometry. We conclude that labeling probes need to be examined for their labeling stoichiometry and achievable resolution or induced cluster size before implementing them for quantitative super-resolution imaging. Our technology will be universally applicable to test the imaging probes by placing the antigens at specified positions on DNA nanostructures to use in quantitative super-resolution imaging.
3) Peptide-PAINT imaging using transient peptide-peptide interactions:
In our search for an ideal labeling probe that introduces minimal artifacts while delivering the highest labeling efficiency, we wondered about using the transient binding interactions of short peptide coils. Several labs have been working in this direction to develop nanomolar affinity coiled coil probes for fluorescence microscopy. In their efforts, they have tagged a protein of interest with a short docking peptide coil. A complementary peptide coil tagged with a fluorescent protein, called imager coil, was then expressed in the cells. Because of high affinity, the docking coil binds to the imager coil, forming a sandwich of less than 5 nm in size, hence providing a shortest labeling probe. We adopted one of the coiled-coil interactions to obtain kinetics similar to that of DNA-PAINT. We named this technique as Peptide-PAINT. We found that a 18-20 amino acids length coil interacts with its partner coil with binding kinetics twice faster than DNA-PAINT, enabling us to image as much faster with Peptide-PAINT. Peptide-PAINT can be performed in physiological ionic concentrations, hence opening opportunities for live-cell Peptide-PAINT imaging.
4) Secondary-probe based unlimited multiplexed DNA-PAINT imaging:
In this approach, we developed a multiplexed imaging approach that will allow us to image unlimited numbers of proteins of interest with DNA-PAINT imaging. To demonstrate the concept, we created 42 different DNA origami nanostructures each carrying unique 20-nt single-stranded DNA extensions. We then hybridize with a complementary DNA strand on the origami structures one at a time. The hybridized DNA strand carries a docking strand at one end to be able to image that particular origami species using DNA-PAINT. After finishing the imaging of that particular origami species, we remove that DNA strand using toe-hold displacement. We then continue to image the next origami species by hybridizing the corresponding complementary strand. By going through the cycles of hybridizing, imaging, and removing the secondary strand, we showed that we could image 42 different origami species. We note that 42 is not the upper limit, the technology could be extended to many more targets. Currently, we are working to demonstrate the technology in the cellular world.
