Microbial rhodopsins are photosensitive ion channels or pumps which have been widely used as optogenetic tools, where visible light illumination triggers neural excitation or depression. In 2011, Dr. Adam E. Cohen’s group at Harvard University found that certain proton-pumping rhodopsins could be “run in reverse” – meaning that fluctuations in the cellular membrane potential could affect light absorption of these seven-span transmembrane proteins¹. From a structural point of view, this electrochromic effect is mediated by the voltage-dependent acid-base equilibrium of a Schiff base in the retinal chromophore binding pocket. The altered proton electrochemical potential changes the absorption and emission spectra of the rhodopsin, thus creating a genetically-encoded voltage indicator (GEVI).

While rhodopsin-based GEVIs have high voltage sensitivity, a major drawback is their low fluorescence quantum yield. In spite of many efforts to improve the brightness, often through site-directed mutagenesis and screening², the quantum yield of rhodopsin-based GEVIs are still approximately two orders of magnitude weaker than most synthetic fluorescent dyes. One solution to this problem is to use Förster resonance energy transfer (FRET). Voltage-induced changes in the rhodopsin absorption spectrum could modulate the quenching efficiency of an appended fluorescent reporter, in a mechanism called electrochromic FRET (eFRET). For example, rhodopsin-fluorescent protein (FP) fusions have been created to achieve voltage sensing^{3, 4}. While these eFRET GEVIs have significantly improved brightness, their voltage sensitivity has been quite limited due to the low energy transfer efficiency between the bulky FP and the retinal chromophore. A back-of-the-envelope calculation suggests that the optimal FRET efficiency is ~67% (assuming in the shot noise-limited regime), whereas most FP-based eFRET GEVIs have their FRET efficiencies below 20%⁵.

To improve the FRET efficiency and hence voltage sensitivity, we turned to synthetic fluorophores as the reporter, which are substantially smaller than FPs. The challenge is: how can we achieve site-specific modification of rhodopsin protein scaffold with synthetic fluorophores? Among various bioorthogonal protein labeling techniques, we chose PRIME (PRobe-Incorporation Mediated by Enzyme) strategy to construct the hybrid voltage indicator⁶. In PRIME, a bioorthogonal functional group is enzymatically conjugated to a 13-amino acid peptide (LAP peptide). Compared to FP-fusion, this strategy effectively reduces the tag size from ~230 amino acid residues down to 13 residues, which greatly shortens the distance between retinal and the fluorescent reporter.

We started bioorthogonal rhodopsin engineering with screening LAP insertion sites in the extracellular region of the transmembrane protein⁵. Whereas most rhodopsins had defective membrane trafficking when LAP was inserted, one candidate, Acetabularia acetabulum rhodopsin II (Ace2), showed good tolerance to LAP insertion. We then compared the labeling efficiency of LAP inserted at various locations in Ace2 and eventually identified the first extracellular loop as the most ideal site, in terms of both high expression level and good membrane trafficking. This indicator was named as Flare1, for fluorophore ligation-assisted rhodopsin eFRET voltage indicator. Flare1 was applied to optically map the gap junction-mediated electrical conduction in HEK293T over hundreds of micrometers, and it could detect the simulated AP with a sensitivity of -28%⁵.

However, Flare1 has several problems that prevent its application to neuronal cell culture: 1) reagents used in copper-assisted alkyne-azide cycloaddition (CuAAC) interferes with neuronal electrophysiology; 2) Flare1-Cy5 has very low sensitivity (ΔF/F₀ = -5.4% per 100 mV). To solve these problems, we assessed several bioorthogonal reactions for higher reaction kinetics and lower neural toxicity, including strain-promoted alkyne-azide cycloaddition (SPAAC) and inverse-electron-demand Diels-Alder (IEDDA) reactions. Our efforts identified IEDDA as the most suitable method in terms of better biocompatibility, shorter labeling time and higher signal-to-background ratio. More specifically, an engineered bacterial lipoic acid ligase (LplA)⁶ was used to site-specifically conjugate a trans-cyclooctene (TCO) moiety to the LAP peptide, which then reacts with a tetrazine-conjugated fluorophore to achieve covalent linkage between the fluorophore and the rhodopsin.

We further engineered the protein scaffold to improve its expression level in neurons by screening Ace2 Asp81 mutants (the proton acceptor) and optimizing the linker between LAP and Ace2. These efforts have led to the construction of hybrid voltage indicators (HVI) that exhibited significantly higher voltage sensitivity than most FP-based eFRET GEVIs, owing to short distance and increased spectral overlap between the fluorophore donor and rhodopsin’s retinal quencher.

We created a palette of HVI indicators by conjugating different colored fluorophore-tetrazine conjugates. Among these, the most red-shifted HVI-Cy5 could faithfully report neuronal spikes in the far-red emission range. Taking advantage of its high SNR and red-shifted spectrum, we demonstrated HVI-Cy5 could be paired with optogenetic actuators and green/red-emitting fluorescent indicators, allowing multiplexed imaging and all-optical electrophysiology in cultured neurons. This expanded toolbox could facilitate high-throughput screening of agonists/antagonists of neuronal ion channels in the future. 

To summarize, HVIs have simultaneously offered high brightness/photostability and genetic targetability, thus combining advantages of both chemistry and biology. The hybrid strategy employed by HVI also highlights the importance of bioorthogonal protein engineering.

To read more about this work: https://doi.org/10.1038/s41557-021-00641-1.

References

1. Kralj, J.M., Hochbaum, D.R., Douglass, A.D. & Cohen, A.E. Electrical Spiking in Escherichia coli Probed with a Fluorescent Voltage-Indicating Protein. Science 333, 345-348 (2011).
2. Hochbaum, D.R. et al. All-optical electrophysiology in mammalian neurons using engineered microbial rhodopsins. Nat Methods 11, 825-833 (2014).
3. Zou, P. et al. Bright and fast multicoloured voltage reporters via electrochromic FRET. Nat Commun 5, 4625 (2014).
4. Gong, Y.Y. et al. High-speed recording of neural spikes in awake mice and flies with a fluorescent voltage sensor. Science 350, 1361-1366 (2015).
5. Xu, Y. et al. Hybrid Indicators for Fast and Sensitive Voltage Imaging. Angew Chem Int Ed Engl 57, 3949-3953 (2018).
6. Liu, D.S. et al. Diels-Alder cycloaddition for fluorophore targeting to specific proteins inside living cells. J Am Chem Soc 134, 792-795 (2012).