A cross–nearest neighbor/Monte Carlo algorithm for single molecule localization microscopy defines interactions between p53, Mdm2, and MEG3

Abstract

The functions of long noncoding (lnc)RNAs such as MEG3 are defined by their interactions with other RNAs and proteins. These interactions, in turn, are shaped by their subcellular localization and temporal context. Therefore, it is important to be able to analyze the relationships of lncRNAs while preserving cellular architecture. The ability of MEG3 to suppress cell proliferation led to its recognition as a tumor suppressor. MEG3 has been proposed to activate p53 by disrupting the interaction of p53 with Mdm2. To test this mechanism in the native cellular context, we developed a new cross–nearest neighbor/Monte Carlo algorithm to quantify the association of the localizations of p53, Mdm2, and MEG3 in U2OS cells using two-color direct stochastic optical reconstruction microscopy (dSTORM), a single-molecule localization microscopy (SMLM) technique. We characterized the methodology using simulations and validated it using the rapamycin-induced binding of FKBP12 and mTOR. We applied the algorithm to examine the association between p53 and MEG3, and p53 and Mdm2. In contrast to previous models, our data support a model in which MEG3 modulates p53 independently of the interaction with Mdm2. This new tool in the SMLM toolbox enables the interrogation of sparse, unclustered molecular interactions using standard fluorescent probes while preserving their cellular context (e.g. subcellular compartmentalization, supramolecular organization, and intercellular heterogeneity).