Event-based Single Molecule Localization Microscopy (eventSMLM) for High Spatio-Temporal Super-resolution Imaging

Abstract

Photon emission by single molecules is a random event with a well-defined distribution. This calls for event-based detection in single-molecule localization microscopy. The detector has the advantage of providing a temporal change in photons and emission characteristics within a single blinking period (typically, [~] 30 ms) of a single molecule. This information can be used to better localize single molecules within a user-defined collection time (shorter than average blinking time) of the event detector. The events collected over every short interval of time / collection time ([~] 3 ms) give rise to several independent temporal photon distributions (tPSFs) of a single molecule. The experiment showed that single molecules intermittently emit photons. So, capturing events over a shorter period / collection time than the entire blinking period gives rise to several realizations of the temporal PSFs (tPSFs) of a single molecule. Specifically, this translates to a sparse collection of active pixels per frame on the detector chip (image plane). Ideally, multiple realizations of single-molecule tPSF give several position estimates of the single-molecules, leading to multiple tPSF centroids. Fitting these centroid points by a circle provides an approximate position (circle center) and geometric localization precision (determined by the FWHM of the Gaussian) of a single molecule. Since the single-molecule estimate (position and localization precision) is directly driven by the data (photon detection events on the detector pixels) and the recorded tPSF, the estimated value is purely experimental rather than theoretical (Thomsons formula). Moreover, the temporal nature of the event camera and tPSF substantially reduces noise and background in a low-noise environment. The method is tested on three different test samples (1) Scattered Cy3 dye molecules on a coverslip, (2) Mitochondrial network in a cell, and (3) Dendra2HA transfected live NIH3T3 cells (Influenza-A model). A super-resolution map is constructed and analyzed based on the detection of events (temporal change in the number of photons). Experimental results on transfected NIH3T3 cells show a localization precision of [~] 10 nm, which is [~] 6 fold better than standard SMLM. Moreover, imaging HA clustering in a cellular environment reveals a spatio-temporal PArticle Resolution (PAR) (2.3lp x{tau} ) of 14.11 par where 1 par = 10-11 meter.second. However, brighter probes (such as Cy3) are capable of [~] 3.16 par. Cluster analysis of HA molecules shows > 81% colocalization with standard SMLM, indicating the consistency of the proposed eventSMLM technique. The single-molecule imaging on live cells reveals temporal dynamics (migration, association, and dissociation) of HA clusters for the first time over 60 minutes. With the availability of event-based detection and high temporal resolution, we envision the emergence of a new kind of microscopy that is capable of high spatio-temporal particle resolution in the sub-10 par regime.