Periodic Reporting for period 4 - MULTIVIsion (Multiphoton Voltage Imaging)
Reporting period: 2024-06-01 to 2025-03-31
Understanding the brain is one of the great scientific challenges of our time. This pursuit fundamentally depends on advances in physical sciences and engineering to provide novel tools and methods to perturb, record and interpret brain activity.
Information in the brain is encoded in changes in the voltage across the membrane of brain cells. Voltage imaging with genetically encoded voltage indicators (GEVIs) is a revolutionary method that allows faithful recording of the fast electrical dynamics of many genetically targeted cells in parallel. This provides an unprecedented means to record how patterns of change in this membrane voltage, called action potentials, manifest in subcellular compartments, cells, and networks across the brain, which is the only way to arrive at a fundamental understanding of brain functions like learning and memory, and of neurogenerative diseases. This is not only a means to unravel the secrets of the brain, but also a way to acquire information about the function of brain cells that could help us understand and develop treatments for diseases related to malfunctioning of brain cells.
For this promise to be fulfilled, we need voltage imaging in 3-dimensioal tissue, so we can look at the functionality of netowrks of neurons in their native environments.
The overall objective of this project are to develop a GEVI and an excitation-and imaging protocol that would allow voltage imaging deep in tissue, so that we can start looking at the detailed functionality of nerve cells in their native environments. we will then do a proof-of-principle experiment in which a model animal learns a movement task: we will try to record the change in functionality in the relevant neuronal networks.
In this project we developed a screening setup for GEVIs; developed novel nonlinear excitfation and readout options for GEVIs, and successfully investigated voltage dynamics in developing zebrafish and mouse cerebellum. aartfrom this, our screenings provided sufficient data for 2 AI based screening methods (one in which an LLM was trained on sequence and functional data of mutants to create beneficial mutations, and one where alphfold3 was used to create a database of structures of untested opsins, which were parametrized together with functional data, and clustered to find untested opsin with a high chance for voltage sensitivity. All in all we achieved the goals of this project and went beyond it in terms of adding a model system and already having a database of a large number of potential superior sensors to mine.
Information in the brain is encoded in changes in the voltage across the membrane of brain cells. Voltage imaging with genetically encoded voltage indicators (GEVIs) is a revolutionary method that allows faithful recording of the fast electrical dynamics of many genetically targeted cells in parallel. This provides an unprecedented means to record how patterns of change in this membrane voltage, called action potentials, manifest in subcellular compartments, cells, and networks across the brain, which is the only way to arrive at a fundamental understanding of brain functions like learning and memory, and of neurogenerative diseases. This is not only a means to unravel the secrets of the brain, but also a way to acquire information about the function of brain cells that could help us understand and develop treatments for diseases related to malfunctioning of brain cells.
For this promise to be fulfilled, we need voltage imaging in 3-dimensioal tissue, so we can look at the functionality of netowrks of neurons in their native environments.
The overall objective of this project are to develop a GEVI and an excitation-and imaging protocol that would allow voltage imaging deep in tissue, so that we can start looking at the detailed functionality of nerve cells in their native environments. we will then do a proof-of-principle experiment in which a model animal learns a movement task: we will try to record the change in functionality in the relevant neuronal networks.
In this project we developed a screening setup for GEVIs; developed novel nonlinear excitfation and readout options for GEVIs, and successfully investigated voltage dynamics in developing zebrafish and mouse cerebellum. aartfrom this, our screenings provided sufficient data for 2 AI based screening methods (one in which an LLM was trained on sequence and functional data of mutants to create beneficial mutations, and one where alphfold3 was used to create a database of structures of untested opsins, which were parametrized together with functional data, and clustered to find untested opsin with a high chance for voltage sensitivity. All in all we achieved the goals of this project and went beyond it in terms of adding a model system and already having a database of a large number of potential superior sensors to mine.
In the context of this project we developed multiple optical setups to help us achieve the stated goals:
- We developed an optical setup that helps us evolve GEVIs by imaging a large number of cells, which each express a mutated version of a GEVI. We can then select the cells that show a good response to our stimulus, and sequence the plasmid in it, to find the mutations that provided better functionality. A version of this setup was also used to identify aggressive cancer cells (L. You et al; Nature Biomedical Engineering, 6, 667–675 (2022) ). This setup is part of a licencing deal being set up with a commercial drug screening party.
- We developed a setup for nonlinear screening and in vivo imaging of GEVIs, including all the hardware and software needed for the evolution and nonlinear optogenetic experiments (X. Meng et al; Journal of Optics, 24, 054004 (2022)
- We created a setup for pump-probe experiments on GEVIS in cellular environments, to understand their response to nonlinear excitation under different voltages. We found that we could influence the photocycle of the GEVIs and with it their response to nonlinear excitation by making a hybrid version of the GEVI coupled to a plasmonic antenna. this influening of the photocycle makes the GEVI respond with kinetics that we otherwise do not see using nonlinear excitation and makes it possible to use them for nonlinear excitation under circumstances where they otherwise would not (Locarno et al, Adv. Materials, revision, 2025)
- Covid delayed the progress on some aspects of the project. To catch up, we created a separate in vivo imaging setup, which aims for more superficial voltage imaging in vivo, which can already start working on the proof-of-principle learning task while the other parts of the project develop further. We have taken this part of the project in two directions: we use pulsed excitation of GEVIs to perform FLIM imaging in zebrafish, invetigating their development (Wu, Silva et al, under review; Wu et al in prep); and we used voltage imaging in the cerebellum to investigate motor memory formation (Silva et al, in prep). we find correlations in complex spike occurance between different parts of the olivocerebellar loop using this voltage imaging technique.
Apart from the optical work:
- We set up a number of cell lines that are helpful to us in our GEVI evolution, characterization of GEVI dynamics and chanracterization of neural dynamics (used in a.o. Q. Li et al; Optics Express, 29, 21, 34097-34108 (2021); Flamourakis et al, Adv Func. Mat 35 (5), 2409451 (2025))
- We set up a theoretical framework for the analysis and manipulation of GEVI photocycles (Meng et al, ACS Phys Chem Au 3 (4) 320 (2023);
- We evaluated different GEVI families for dynamics under nonlinear excitations, and screened them for evolved brightness and functionality. We developed a novel GEVI with improved NIR fluorescence (Ganapathy et al, J. Bio. Chem. 299 (6) 2023). our screening setup has been augmented with AI based computational screening and we are no also sitting on a database of voltage sensitive rhodopins that we are characterizing (work in progress).
- We developed an optical setup that helps us evolve GEVIs by imaging a large number of cells, which each express a mutated version of a GEVI. We can then select the cells that show a good response to our stimulus, and sequence the plasmid in it, to find the mutations that provided better functionality. A version of this setup was also used to identify aggressive cancer cells (L. You et al; Nature Biomedical Engineering, 6, 667–675 (2022) ). This setup is part of a licencing deal being set up with a commercial drug screening party.
- We developed a setup for nonlinear screening and in vivo imaging of GEVIs, including all the hardware and software needed for the evolution and nonlinear optogenetic experiments (X. Meng et al; Journal of Optics, 24, 054004 (2022)
- We created a setup for pump-probe experiments on GEVIS in cellular environments, to understand their response to nonlinear excitation under different voltages. We found that we could influence the photocycle of the GEVIs and with it their response to nonlinear excitation by making a hybrid version of the GEVI coupled to a plasmonic antenna. this influening of the photocycle makes the GEVI respond with kinetics that we otherwise do not see using nonlinear excitation and makes it possible to use them for nonlinear excitation under circumstances where they otherwise would not (Locarno et al, Adv. Materials, revision, 2025)
- Covid delayed the progress on some aspects of the project. To catch up, we created a separate in vivo imaging setup, which aims for more superficial voltage imaging in vivo, which can already start working on the proof-of-principle learning task while the other parts of the project develop further. We have taken this part of the project in two directions: we use pulsed excitation of GEVIs to perform FLIM imaging in zebrafish, invetigating their development (Wu, Silva et al, under review; Wu et al in prep); and we used voltage imaging in the cerebellum to investigate motor memory formation (Silva et al, in prep). we find correlations in complex spike occurance between different parts of the olivocerebellar loop using this voltage imaging technique.
Apart from the optical work:
- We set up a number of cell lines that are helpful to us in our GEVI evolution, characterization of GEVI dynamics and chanracterization of neural dynamics (used in a.o. Q. Li et al; Optics Express, 29, 21, 34097-34108 (2021); Flamourakis et al, Adv Func. Mat 35 (5), 2409451 (2025))
- We set up a theoretical framework for the analysis and manipulation of GEVI photocycles (Meng et al, ACS Phys Chem Au 3 (4) 320 (2023);
- We evaluated different GEVI families for dynamics under nonlinear excitations, and screened them for evolved brightness and functionality. We developed a novel GEVI with improved NIR fluorescence (Ganapathy et al, J. Bio. Chem. 299 (6) 2023). our screening setup has been augmented with AI based computational screening and we are no also sitting on a database of voltage sensitive rhodopins that we are characterizing (work in progress).
The setups we developed are beyond the state of the art and provide imaging, optogenetics and screening capabilities unique to the lab (and now our collaborators).
The results of the screening are beyond the state of the art in sensitivity, brightness and kinetics, and beyond the state of the art in terms of our AI-based protein design.
The hybrid plasmonic sensors with improved kinetics are a novel concept beyond the state of the art.
Our in situ investigation of nonlinear dynamics of GEVIs is also beyond the state of the art. These experiments turned out harder than expected in terms of the needed alignment precision (M.-P. Chien et al; Science Advances, 7, 19, eabe3216 (2021))., but we delivered interesting evaluations of GEVI dynamics and a framework for how they can be used in optimized imaging protocols.
Our in vivo results are beyond the state of the art in that we are imaging electrical dynamics in purkinje cells int eh mouse cerebellum, which has not been achieved before; beyond this our approach to detecting developmental bioelectricity using nonlinearities in voltage sensing is new too.
The results of the screening are beyond the state of the art in sensitivity, brightness and kinetics, and beyond the state of the art in terms of our AI-based protein design.
The hybrid plasmonic sensors with improved kinetics are a novel concept beyond the state of the art.
Our in situ investigation of nonlinear dynamics of GEVIs is also beyond the state of the art. These experiments turned out harder than expected in terms of the needed alignment precision (M.-P. Chien et al; Science Advances, 7, 19, eabe3216 (2021))., but we delivered interesting evaluations of GEVI dynamics and a framework for how they can be used in optimized imaging protocols.
Our in vivo results are beyond the state of the art in that we are imaging electrical dynamics in purkinje cells int eh mouse cerebellum, which has not been achieved before; beyond this our approach to detecting developmental bioelectricity using nonlinearities in voltage sensing is new too.