Engineering of Genetically Encoded Bright Near-Infrared Fluorescent Voltage Indicator

Abstract

Genetically encoded voltage indicators (GEVIs) allow for the cell-type-specific real-time imaging of neuronal membrane potential dynamics, which is essential to understanding neuronal information processing at both cellular and circuit levels. Among GEVIs, near-infrared-shifted GEVIs offer faster kinetics, better tissue penetration, and compatibility with optogenetic tools, enabling all-optical electrophysiology in complex biological contexts. In our previous work, we employed the directed molecular evolution of microbial rhodopsin Archaerhodopsin-3 (Arch-3) in mammalian cells to develop a voltage sensor called Archon1. Archon1 demonstrated excellent membrane localization, signal-to-noise ratio (SNR), sensitivity, kinetics, and photostability, and full compatibility with optogenetic tools. However, Archon1 suffers from low brightness and requires high illumination intensities, which leads to tissue heating and phototoxicity during prolonged imaging. In this study, we aim to improve the brightness of this voltage sensor. We performed random mutation on a bright Archon derivative and identified a novel variant, monArch, which exhibits satisfactory voltage sensitivity (4~5% ΔF/F_AP) and a 9-fold increase in basal brightness compared with Archon1. However, it is hindered by suboptimal membrane localization and compromised voltage sensitivity. These challenges underscore the need for continued optimization to achieve an optimal balance of brightness, stability, and functionality in rhodopsin-based voltage sensors.

Keywords: brightness; directed molecular evolution; genetically encoded voltage indicator; near infrared; phototoxicity; rhodopsin.

Figure 1

Multiparameter-directed evolution of voltage sensor Archon3 in mammalian cells. (a) Screening workflow for multiparameter optimization of genetically encoded voltage sensors based on Archon3 in HEK293FT cells. CaPhos, calcium phosphate transfection. (b) Expression of the selected 7 variants in cultured hippocampal neurons. (c) Representative fluorescence optical traces showing spontaneous activity of neurons expressing different voltage sensor variants. (d) Brightness, ΔF/F, and SNR for the selected 7 variants. Data are presented as means ± S.D. n = 1~15 neurons from 1~3 cultures. (e) Photobleaching curves of variants #3, #4, and #7 under continuous illumination (n = 9, 8, and 15 neurons from 1~3 cultures, respectively). λem = 664 long pass at 1.5 W/mm². (f) Structural model of variant #3 (monArch) as predicted by AlphaFold3. The chromophore is colored in cyan. The mutation sites are highlighted in red (compared with Archon1). The two mutation residues Gly225 and Cys99, which are within 4 Å from the chromophore, are highlighted in blue.

Figure 2

Characterization of improved Archon3-based voltage sensors in primary cultured hippocampal neurons. (a) Representative fluorescence images of Archon3 and monArch visualized via EGFP (excitation at 475/34BP from an LED and emission at 527/50 nm). Scale bar represents 50 µm. (b) Normalized fluorescent intensity of voltage sensors Archon3 and monArch. The intensity of the voltage sensor was normalized to the green GFP signal. Excitation 637 nm laser light at 1.5 W/mm² and emission at 664 nm long pass. n = 6 individual neurons for Archon3 and n = 4 individual neurons for monArch. Data are represented as means ± S.D. Two-tailed Student’s t-test. (c,d) Single-trial optical recordings of (c) Archon3 and (d) monArch fluorescence responses (signal-to-noise ratio, SNR) during evoked action potential via whole-cell patch clamp.

Figure 3

Characterization of monArch in acute mouse brain slice. (a) Fluorescence image of cortical neuron expressing monArch (left) and single-trial optical voltage traces (right) showing the SNR from cell on the left. (b) The SNR per action potential across two recordings from two neurons in one brain slice. Excitation 637 nm laser light at 1.5 W/mm² and emission at 664 nm long pass. n = 2 neurons from one slice. The horizontal line indicates the mean value.

Figure 4

Characterization of soma-localized voltage sensor in cultured hippocampal neurons. (a) Representative images showing the expression of voltage sensors (red) and EGFP (green) in cultured hippocampal neurons. (b,c) Normalized fluorescent intensity of voltage sensors (b) monArch versus soma-monArch and (c) soma-monArch versus somArchon. The fluorescent intensity of the voltage sensors was normalized to the green GFP signal. Two-tailed Student’s t-test was used. (d) Photobleaching curves for somArchon and soma-monArch under continuous illumination (n = 4 and 7 neurons from same neuronal culture, respectively). Imaging acquisition rate: 852 Hz. somArchon and soma-monArch retain 93% and 89% of fluorescence after 60 s of continuous illumination, respectively. (e) Representative fluorescence optical traces of somArchon and soma-monArch. (f) Population data of ΔF/F for somArchon and Soma-monArch. n = 4 neurons from one culture for somArchon; n = 19 from 2 cultures for soma-monArch. Two-tailed Student’s t-test.