PSF distortion and mislocalization by dielectric nanoparticles in single-molecule microscopy

Colloidal particles are nano- to micro-sized particles (NP) commonly dispersed in a liquid medium, occurring in various shapes and materials. These NPs can be functionalized with a multitude of bio-molecules such as DNA, antibodies, and peptides for various applications. The surface functionalization plays an integral role in determining the application and performance of the NP, emphasizing the importance of characterizing and understanding it. Standardized methods for characterizing the surface functionalization rely on ensemble averages. These methods are fast and cost-effective but lack the possibility of characterizing particle-to-particle differences. Recent advances in single-molecule microscopy have enabled the visualization and quantification of chemical groups on colloidal particles but still do not allow for the precise localization of molecules on particles in three dimensions.
A particle of a different refractive index than the surrounding aqueous solution interacts with the emitted light from a molecule. The presence of the NP results in a distorted image on the camera and a subsequent mislocalized emitter. Where mislocalizations are well characterized for gold nanoparticles because of the amplified single-molecule emission due to plasmon resonances, they have never been investigated for dielectric particles. Although the effect is less pronounced, the distortion must be considered.
Current methods for mapping functional sites on dielectric NPs involve the introduction of cylindrical lenses to complement the 2D projection with a third dimension. However, Huijben et al. have recently described the point-spread function deformation in an analytical model and utilized it in fitting experimental images, allowing for better localization of molecules on nanospheres.
In the past three years, I have taken up the challenge of visualizing distortions caused by nanoparticles and correctly analyzing distorted PSFs to overcome the limitations in the current analysis of dielectric micro- and nanospheres.
The first step in this project was to show that mislocalization exists when imaging and analyzing dielectric colloids. Both simulations and experimental results indicate that the presence of the NP changes the size of the PSF. It appears that emitters located at the equator of the bead exhibit the smallest PSF width when using a 200 nm polystyrene sphere with DNA-PAINT. Fitting the data with a symmetrical 2D Gaussian therefore, leads to the wrong interpretation of having a high localization precision but in facts is a combined effect of defocusing and PSF distortion.
I visualized the PSF distortions by increasing the bead size to 1 μm and fitted the data with the analytical PSF model. I have demonstrated that the analytical PSF model concurred with real experimental PSFs imaged from long-lasting emitters (quantum dots) on polystyrene spheres in multiple focal planes.
Finally, I applied that method to DNA PAINT data and utilized DNA PAINT statistics to correct for artifacts that are expected from focal plane, bead radius or bead position variations.
The methods I developed allow for 3D localization of functional sites on dielectric particles from a single plane image. This creates a valuable feedback system for the optimization of conjugating molecules to colloids.