ENGINEERING GCaMP AFFINITY AND KINETICS FOR IMPROVED TRACKING OF NEURONAL ACTIVITY

Abstract

Fluorescent calcium indicator proteins (FCIPs) are powerful tools for monitoring neural activity. However, they still have significant performance limitations compared with synthetic indicators based on the small-molecule chelator BAPTA. Because of high cooperativity originating from a calmodulin-based recombinant calcium sensor, a given GECI is only sensitive to a small part of a neuron's likely calcium concentration range, which can span a range of 0.1-10 µM. GECIs also have up to 100-fold slower reponse kinetics than BAPTA-based indicators. Overcoming limitations in range and kinetics is a key step toward monitoring spike times and firing rates in cell-type-specific brain circuits.

We are engaged in structure-based design to vary the affinity and accelerate the response kinetics of a widely used GECI, GCaMP3. We have designed more than 50 novel variants by targeted mutation of GCaMP3's calmodulin (CaM) domain and its intraprobe peptide partner, RS20. In our cuvet characterizations of purified protein, we have attained a nearly 40-fold (0.16-6 µM) range of KD without impairing per-molecule brightness. In stopped-flow biochemical measurements, off-responses to sharp decreases in calcium are more than 10 times faster than any other published GECI. Most of the gap in off-response speed between G-CaMP3 and BAPTA-based indicators could be closed without perturbing KD. In Drosophila antennal nerve axons, sensory stimulation-evoked fluorescence responses were significantly enhanced in speed and amplitude in two novel GECIs.

With our biophysical measurements, we discovered that the N-lobe of the bilobular CaM domain is required for the high-fluorescence state and the C-lobe contributes to high affinity Ca2+ binding. To account for our observations, we propose a molecular dynamics model of GCaMP3 with two kinetic pathways leading to a high-fluorescence state. First, small amounts of Ca2+ activate a slow "C-like" pathway through sequential binding to the C-lobe followed by an allosterically induced increase in calcium affinity and binding to the N-lobe. Second, large amounts of Ca2+ can activate a faster, "N-like" pathway by direct calcium binding to the N-lobe. These findings not only enrich the existing understandings of FCIPs, but also provide a framework for the future design of faster and more functionally flexible Ca2+ sensors.