A new split calcium sensing/reporting system based on split firefly luciferase linked to M13 and calmodulin was designed that is able to report the increase of the cytosolic calcium ions induced by mechanoreceptor stimulation by emitted light.

Motivation

While calcium influx could be detected by exogenous fluorescent dyes such as the FuraRed and Fluo-4, we needed a genetically encoded calcium sensor that would couple a change in the calcium concentration to a biologically relevant output, such as the luciferase reporter or reconstitution of a split protease the initiation of the signaling pathway. The ideal calcium sensor should be inactive at intracellular concentration of calcium and have a high response to calcium concentrations above physiological levels and should be detected by a quick and easy readout. For our intended application, the calcium sensor should also have the potential to act as the reconstitution mechanism for split proteins so that a conformational rearrangement in the presence of calcium would bring the two split protein fragments together and reconstitute the protein’s activity.

A wide pallet of genetically encoded calcium sensors are used for mapping intracellular calcium concentration Whitaker2010, including calmodulin, troponin C and aequorin Wilms2014. These reporters are based on different mechanisms of detection. From this abundant collection we chose the calmodulin (CaM)-based calcium sensors (in particular CaMeleons), since their mechanism is based on a large conformational change, allowing reconstitution of split proteins Whitaker2010.

We replaced the FRET pair of CaMeleon2.12 by a split firefly luciferase (fLuc) as it provides a distinct signal even in small amounts and has a remarkable signal-to-noise ratio. The new luciferase based calcium sensor was named fLuc2.12. The fLuc2.12 was tested on HEK293 cells, but we found that the sensor was active already in resting cells (1). We hypothesized the activation was a consequence of a close proximity of calmodulin and M13 in the fusion even in the absence of calcium binding. In order to resolve this problem we set out to test a similar sensor based on two separate molecules. Two-molecule-based CaM sensors have not been widely studied, but lower leakage in comparison to a single molecule sensor has been reported by Miyawaki et al. Miyawaki1997.

Based on the inspection of the 3D structure of the CaM-M13 complex (PDB code: 2BBM), we fused the N-terminal fragment of the split firefly luciferase to the N-terminus of M13 (nLuc:M13) and the C-terminal fragment of the split firefly luciferase to the C-terminus of calmodulin (CaM:cLuc). The split calcium sensor is represented in 2A. When transfected into HEK293T cells the sensor was expressed in the cytosol (2B).

The split luciferase was tested on live cells (3) but the ratio of the outputs from stimulated and non-stimulated cells remained low, because activation of the sensor was still detectable at the cytoplasmic concentration of calcium. In order to decrease the activation of the sensor at resting levels of calcium we introduced two mutations E31Q, E104Q in EF hand motifs, reported previously to decrease the affinity of CaM to calcium Evans2009. We prepared the single mutant E104Q and the double mutant E31Q, E104Q split calcium sensors and tested the systems on HEK293T cells (4).

Split calcium reporter with a single mutation (E104Q) introduced into calmodulin was proved to work best, whereas the sensor with two mutations (E31Q, E104Q) generated a low signal. The sensor with E104Q mutation had the highest ratio of the stimulated vs. resting cells, therefore representing the best with calcium sensor. We additionally tested whether the ratio between parts of the calcium sensor affects the fold change between signal of non-stimulated and stimulated cells (5). The response of the reported depended on the ratio, favoring an excess of the nLuc:M13, with the highest ratio close to 10.

Those reporters were later used to detect response of cells to activation of mechanoreceptors based on the ultrasound and mechanical stress, where they enables clean difference between stimulated and unstimulated cells, already few minutes after the stimulation and could be used for real time monitoring.