INFORMATION ABOUT PROJECT,
SUPPORTED BY RUSSIAN SCIENCE FOUNDATION

The information is prepared on the basis of data from the information-analytical system RSF, informative part is represented in the author's edition. All rights belong to the authors, the use or reprinting of materials is permitted only with the prior consent of the authors.

 

COMMON PART


Project Number16-15-10323

Project titleDeveloping novel type of genetically-encoded calcium sensors for combined ex vivo visualization of calcium and genomic activities in nerve system.

Project LeadSubach Fedor

AffiliationMoscow Institute of Physics and Technology,

Implementation period2016 - 2018

Research area 05 - FUNDAMENTAL RESEARCH IN MEDICINE, 05-106 - Neurobiology

Keywordsgenetically-encoded fluorescent protein, circulaly-permutated fluorescent protein, calcium sensor, redox-sensor, fluorescence microscopy, neuronal network, neuronal activity, brain, cfos immediate early gene, plasticity and memory


 

PROJECT CONTENT


Annotation
This project is directed to development of the original approach for ex vivo visualization of calcium activity in neuronal system by creation of a new type of genetically encoded calcium sensors, namely slow calcium sensors. This approach will be applied to address the fundamental yet unresolved task of neurobiology, i.e. combined visualization of neuronal calcium and genomic activities caused by the processes of neuronal plasticity in the brain. It is known that formation of long-term rearrangements in neural cells in normal and abnormal brain is accompanied by changes in both calcium activity and genes expression. However, problem on mapping of these changes on the level of individual neurons of the brain is still unresolved. To address this fundamental scientific problem we suggest developing approach that for the first time will allow us to perform ex vivo combined visualization of calcium and genomic activities of neurons in the models of different types of physiological stimulations and activations of the brain during associative learning. Today solution of this problem is hindered because of limitations for approaches connected with simultaneous visualization of calcium and genomic activities in the same neurons of the brain. Genomic response of neurons to the learning occurs slowly, i.e. in minutes or in some cases during tens of minutes. Available calcium sensors have fast kinetics of dissociation from calcium ions, i.e. during 1-3 seconds calcium dissociates completely. Therefore, efficient fixation of calcium activity of neurons ex vivo even immediately after experimental exposure is not possible. The only available calcium sensor-integrator CaMPARI allows performing such ex vivo visualization of calcium activity but it requires continuous in vivo illumination of neurons by toxic UV light of 405 nm. This limitation restricts its application by models of small non-vertebrates, transparent nerve system of embryos, or olfactory bulb and non-deep layers (2/3 layers) of neocortex of mammals with transcranial optical windows attached to the skull. To avoid these limitations we suggest original approach for ex vivo visualization of calcium activity in the brain of vertebrates that is based on new type of slow calcium sensors to be developed. The main feature of new generation of calcium sensors will be slow kinetics of dissociation of calcium ions and after episode of calcium activity sensor will stay in fluorescent calcium-bound state at least for 5-20 minutes. This time will be sufficient for its fixation that is necessary for followed ex vivo visualization of cellular allocation of calcium activity in nerve system. We suggest several clue advantages of the proposed technique for ex vivo registration of neuronal calcium activity with slow calcium sensors over existing approaches with fast calcium sensors: 1) Possibility to visualize calcium activity in any brain structure, including deep one such as amygdala as well as in several structures simultaneously, e.g. in cortex, hippocampus and amygdala; 2) The method is suitable for all types of learning and behavioral experiments on freely moving animal and without any restrictions that in the case of CaMPARI sensor are connected with illumination of studied local brain region with UV light; 3) Since there is no need to install an cranial optical windows or implant optical fiber probes, so the labor required for these surgeries is reduces; 4) There are no effects from the brain damage observed during toxic UV light irradiation and visualization of deep structures of the brain; 5) The possibility of combining developed method with other indicators and molecular cellular markers using IHC staining. Combined visualization of neuronal calcium signaling and genomic activation in the behavioral learning experiments on mammals are complex project. Realization of this versatile project will be possible as a result of combined expertize of molecular biological group for the development of genetically-encoded fluorescent sensors and proteins from one hand and neurobiological group for neurobiological studies of cellular mechanisms of associative memory and experience-dependent gene expression from another hand. In terms of research for this project: 1) Method for ex vivo visualization of calcium activity in the mice brain will be developed in several steps. 1.1) Using directed molecular evolution approach in bacteria we expect to develop and optimize novel class of genetically-encoded GFP-like fluorescent calcium indicators having slow calcium kinetics of dissociation using approach for directed molecular evolution in bacterial system. 1.2) Protocol for ex vivo fixation of slow sensors will be developed 1.3) We anticipate to characterize and select optimized variants of slow calcium sensors in mammalian cells and neuronal culture. 1.3) In vivo and ex vivo visualization of calcium activity with developed sensors will be demonstrated in different structures of the mice brain at different sensory stimuli. 2) For the first time we expect to demonstrate combined ex vivo visualization of genomic and calcium neuronal activities at cellular resolution in one of the brain structure of mice to study mechanism of formation and retrieval of associative memory. Combined genomic and calcium neuronal activities will be visualized using ex vivo confocal fluorescence microscopy in the mice behavioral model of classical fear conditioning learning with presentation of sound stimuli. Calcium activity will be registered ex vivo according to fluorescence of developed slow calcium sensors and genomic activity – via EGFP expression as a result of c-fos early gene activation invoked by neuronal activity in transgenic reporter mice line. In summary, developed method for combined ex vivo visualization of calcium and genomic neuronal activities opens up possibility for cellular mapping of dual signal, i.e. calcium-genomic activity in the brain of model animals at very different behavioral impacts, sensory, physiological and pharmacological stimulations as well as plastic rearrangements in normal brain and at brain pathologies. Developed method for visualization of doubled calcium-genomic neuronal activity will match different applications utilized in neurobiological laboratories around the world.

Expected results
1) Using directed molecular evolution approach in bacteria we expect to develop and optimize new type of genetically encoded GFP-like fluorescent calcium indicators having slow dissociation kinetics from calcium ions and allowing fixation of calcium-bound and calcium-free states to provide bases for ex vivo imaging. 2) We anticipate to characterize and select optimized variants of calcium sensors in mammalian cells and neuronal culture. 3) We expect to characterize dynamics of developed slow calcium sensors in vivo in mice brain cortex using 2P-microscopy and miniature microscope nVista HD. 4) For the first time we plan to demonstrate ex vivo visualization of calcium activity on brain slices for different structures of mice brain. 5) For the fist time we anticipate to realize combined ex vivo visualization of genomic and calcium activities of neuronal populations during associative memory formation in the brain of mice; these data will be collected for the first time with cellular resolution for the mice behavioral model of classical fear conditioning learning. Project addresses one of fundamental problems of neurobiology – simultaneous visualization of calcium and genomic activities in the same cells of nerve system. This goal is anticipated to be achieved by developing new type of GFP-based genetically-encoded calcium sensors having slow dissociation kinetics from Ca2+ ions and allowing fixation of neuronal calcium activity for ex vivo visualization. Addressing the problem suggested in this project opens up possibility for cellular mapping of dual signal, i.e. calcium-genomic activity in the brain of model animals at very different behavioral impacts, sensory, physiological and pharmacological stimulations as well as plastic rearrangements in normal brain and at brain pathologies. Developed method for visualization of doubled calcium-genomic neuronal activity will match different applications utilized in neurobiological laboratories around the world.


 

REPORTS


Annotation of the results obtained in 2018
1) For the development and optimization of slow and fast calcium sensors, improved versions of CAMPARI slow sensors were selected using directed molecular evolution. For this, 9 rounds of random mutagenesis and directed mutagenesis with randomized amino acid residues at the M13 peptide positions 4, 7, 8, 9 and 10 were performed. As a result, the sCAMPARI-3s9 # 102 mutant was obtained with a 47-fold fluorescent contrast and 1.7-fold reduced affinity to calcium ions compared with the sCAMPARI version obtained in 2017 (Kd = 78 nM). However, the value of the dissociation constant Kd = 78 nM was still below the optimal value of 200 nM. 1.2.1) In the case of the pRm3-31 and pRm3-39 calcium-maturation dependent sensors, the ability to visualize the spontaneous activity of neuronal cultures was characterized. To this end, we transduced neuronal cultures with AAV virus particles (rAAV) encoding pRm3-P2A-EGFP, or control R-GECO1-P2A-EGFP red indicators under the control of the CAG promoter, and incubated cultures in the presence or absence of a mixture of AP- 5 and CNQX neuronal activity inhibitors. The pRm3 calcium-maturation dependent indicator matured in cultures under conditions of spontaneous activity as effectively as under conditions that block spontaneous activity. Thus, the pRm3 calcium-maturation dependent indicator is not able to detect the spontaneous activity of the neuronal culture. 1.2.2) An attempt was made to increase the Kd of the pRm3-31 calcium-maturation dependent sensor. To this end, the F11V mutation in the M13 peptide was introduced into the pRm3-31 indicator in order to return the amino acid residue of R-GECO1, which has a reduced affinity to calcium ions Kd = 811 nM. However, such a replacement almost did not lead to a change in Kd (Kd = 43 nM). 1.2.3) Characterization of the dynamics of the developed calcium-maturation dependent red sensor in the neurons of the cerebral cortex during and after presentation of the stimulus to a free-moving mouse using a two-photon microscope. To characterize the dynamics of the calcium-maturation dependent red indicator, a model of the environment enrichment in the home cage with the help of beads threads was used. In this case, using a two-photon microscope, calcium responses of 2/3 layers neurons of the barrel fields of the somatosensory cortex were recorded. When analyzing the fluorescence dynamics during stimulation of the vibrissae, no significant increase or decrease in the fluorescence of the calcium-dependent maturation sensor was shown. Thus, the dynamics of the developed calcium-maturation dependent sensor in the cells of the barrel fields during long-term stimulation of the vibrissae was characterized. 2) By analyzing the crystal structure of the FGCaMP calcium indicator, which has as its sensory part calmodulin from fungi, other than calmodulin from animals, and by performing several rounds of directed and random mutagenesis, a version of FGCaMP with improved contrast in neurons was obtained. Based on the analysis of the FGCaMP crystal structure, we tried to improve the contrast and affinity of the FGCaMP indicator to Ca2 + ions by optimizing the Ca2 + binding components of the Aspergillus fungi, which are part of FGCaMP, by using directed mutagenesis. Our strategy was that mutations were introduced into the calmodulin of FGCaMP in the positions coordinating Ca2 + ions according to its crystal structure, as well as into the already known positions that improve the properties of other genetically encoded calcium indicators. As a result of the screening, we chose the FGCaMP/N60D/D78Y/T79R/N97D/S101D mutant, called FGCaMP3, which had the highest affinity for Ca2 + ions (Kd1 183 nM and 208 nM for forms with an absorption peak at 402 and 493 nm, respectively), and it also had a reduced contrast of the low affinity component (Kd2 contrast) at high concentrations of Ca2 + ions. Since the affinity of other calcium indicators to Ca2 + ions depends on the sequence of the M13-like peptide (Helassa et al. 2015), we created two bacterial gene libraries for the FGCaMP3 mutant with randomized sequences in positions 3, 4, 5 (TLH) or 8, 9 , 10 (IDT) in the M13-like peptide. As a result of screening libraries with varying amino acid residues in the M13 peptide, the FGCaMP3/T3L/H5K mutant (FGCaMP4; Kd1 93 nM at 493 nm excitation) was chosen, characterized by high affinity to Ca2 + ions for the Kd1 component and low contrast for the Kd2 component with low affinity to Ca2 + ions. Then there were 6 rounds of random mutagenesis of this mutant. In the process of screening, the clones were titrated by Ca2 + ions in the presence of 1 mM MgCl2, in order to bring the screening conditions closer to the conditions in the neuronal cytosol, and the half-time of association with Ca2 + ions was controlled for the mutants. After 6 rounds of screening, the best 4 mutants were tested in neuronal cultures. The most optimal based on the contrast and kinetics of association with Ca2 + ions in vitro (Kd 204 nM in the presence of Mg2 + ions, the rapid association rate to Ca2 + ions, monophasic dependence of the fluorescence intensity on the concentration of Ca2 + ions upon excitation at 493 nm) as well as half-times of fluorescence increase/drop and ΔF/F0 (62 ± 22%) in the neurons turned out to be the FGCaMP4 mutant 17-21 (as compared with the ΔR/R0 ratiometric signal ratio for FGCaMP 31%). Thus, by analyzing the crystal structure of the FGCaMP indicator, we were able to improve the response rate of the FGCaMP indicator in the culture of mouse dissociated neurons. 3) A version of the iYTnC2 sensor with a reverse, positive reaction to calcium ions, called YTnC, was obtained. The genetically encoded calcium indicator NTnC has an improved design due to its smaller size, GFP-like N- and C-termini, and two-fold reduced number of calcium-binding sites compared to the widely used GCaMP family of indicators. However, NTnC has a reverse phenotype, moderate calcium response and low temporal resolution. By replacing the mNeonGreen fluorescent part in NTnC with EYFP, we engineered an NTnC-like indicator, referred to as YTnC, that had a positive and substantially improved calcium response and faster kinetics. YTnC had a 3-fold higher calcium response and 13.6-fold lower brightness than NTnC in vitro. According to stopped-flow experiments performed in vitro, YTnC had 4-fold faster calcium-dissociation kinetics than NTnC. In HeLa cells, YTnC exhibited a 3.3-fold lower brightness and 4.9-fold increased response to calcium transients than NTnC. The spontaneous activity of neuronal cultures induced a 3.6-fold larger ΔF/F response of YTnC than previously shown for NTnC. On patched neurons, YTnC had a 2.6-fold lower ΔF/F than GCaMP6s. YTnC successfully visualized calcium transients in neurons in the cortex of anesthetized mice and the hippocampus of awake mice using single- and two-photon microscopy. Thus, in vivo two-photon visualization of neuronal activity in the mouse visual cortex showed that YTnC exhibited spontaneous activity in the mice visual cortex with similar sensitivity as GCaMP6s, both in the neuronal bodies and in the spines. However, the assessment of visual signal-specific responses of YTnC and GCaMP6s sensors revealed a 1.9-fold decrease (p <0.0001) of the average response ΔF/F 0.62 ± 0.23 for YTnC compared to the value of ΔF/F 1.17 ± 0.53 for GCaMP6s. To assess the applicability of the YTnC indicator for observation of in vivo neural activity with an NVista HD mini-microscope (mini-scanner) mounted on the head of an animal, we visualized the calcium activity of CA1 neurons in the hippocampus of freely moving mice during their examination of the new environment. As a result, we identified a neuron that is specifically activated on the track between 225 and 315. Thus, we have demonstrated that YTnC can be successfully used to visualize neuronal dynamics during the study of animals in a new environment and to identify the place cells in a freely moving mouse using the NVista HD mini-scanner. In addition, YTnC outperformed GCaMP6s in the mitochondria and endoplasmic reticulum of cultured HeLa and neuronal cells. As a result of this research, an article was published in 2018 in the Scientific Reports journal: doi: 10.1038 / s41598-018-33613-6 4) The kinetics of fast NTnC and insNCaMP-like sensors has been improved. 4.1) A rational and several rounds of random mutagenesis of the NTnC2 sensor were carried out. The NTnC2 sensor versions obtained in 2017 had a limited contrast in neurons due to their slow calcium binding kinetics. To improve the rate of interaction of NTnC2 with calcium, we performed its mutagenesis at position 222, corresponding to the residue 148 in EGFP, which is located near the hydroxyl group of the chromophore tyrosine ring, stabilized the cis configuration of the chromophore by introducing the serine residue into position 222 and volume residue of phenylalanine into position 239, replaced EEE residues with GGG or QQQ in the troponin part of the NTnC2 sensor, extended both linkers 1 and 2 by two randomized residues, and added one residue before linker 1 and removed two residues after linker 2 with randomization of three residues in linker 1 and one residue in linker 2, by analogy with the fast YTnC indicator. However, none of these approaches has led to a significant acceleration of the kinetics of interaction with calcium ions. Thus, using rational and random mutagenesis, we failed to improve the kinetics of the interaction of the NTnC2 sensor with calcium ions. 4.2) A rational and several rounds of random mutagenesis of the insNCaMP sensor were carried out in order to accelerate its association with calcium. In 2017, the insNCaMP sensor version, consisting of the mNeonGreen fluorescent protein and calmodulin/M13 peptide fusion, was obtained, which had a limited contrast in neurons due to the slow association with calcium ions. To increase the rate of association of insNCaMP with calcium ions, rational mutagenesis of residue 326, corresponding to position 148 of the EGFP protein, on His, Thr or Asn residues, several rounds of random mutagenesis, and extension of the linker between calmodulin and M13 peptide by 2 and 4 amino acid residues was carried out. As a result of the selection of the best variants with fast calcium binding kinetics, the final insNCaMP4, insNCaMP7, insNCaMP9, and insNCaMP10versions were developed, having different affinity for calcium ions from 105 to 373 nM, high 21-34-fold fluorescence contrast to calcium ions and fast kinetics of association-dissociation with calcium ions. In HeLa Kyoto cells, the insNCaMP9 and insNCaMP10 with Kd 198 and 373 nM respectively have the most optimal affinity to calcium ions for use in the cytosol and exhibit a dynamic range and brightness similar to that of the GCaMP6s sensor. We compared the characteristics of the obtained indicators during the visualization of the calcium neuronal activity in the CA1 field of the hippocampus of freely moving mice in the course of their examination of the new environment using the in vivo NVista HD miniscope. All four variants had a fast kinetics of interaction with Ca2 +, similar or faster compared to the control indicator GCaMP6s. It turned out that the insNCaMP7 indicator is the most optimal for visualizing neuronal activity using the NVista minicope and demonstrates an 1.4 fold greater signal-to-noise ratio than the GCaMP6s indicator. By directional mutagenesis of calcium-binding residues, it was possible to obtain an insNCaMPer47 indicator, which has a reduced affinity for calcium ions suitable for detecting calcium transients in the endoplasmic reticulum. 5) Ex vivo visualization of calcium activity in brain slices: Cells from the barrel fields of the somatosensory cortex were infected by transfection with viral particles carrying the genes of the calcium-maturation dependent pRm3 # 39 sensor or the slow sNCaMP calcium sensor. In all mice infected with the red calcium-maturation dependent pRm3 # 39-P2A-EGFP sensor it was found that the percentage of colocalization of the calcium-maturation dependent sensor with GFP was similar in stimulated and control mice. Thus, in repeated experiments using the developed calcium-maturation dependent sensor, no specific increase in calcium activity in the cells of the barrel fields was recorded after prolonged stimulation of the vibrissae. For mice infected with the slow sNCaMP sensor, as a result of analyzing the cell number (from several brain slices), after a brief (within five minutes) stimulation of the vibrissae, about 42% of the cells increase the sensor fluorescence (as opposed to 18% in control animals). Thus, according to preliminary data (based on two mice), using the sNCaMP indicator and ex vivo imaging of brain slices changes in calcium activity in the somatosensory cortex cells can be detected in response to stimulation of the vibrissae, but this result requires reliable confirmation using a larger number of mice.

 

Publications

1. Барыкина Н.В., Доронин Д.А., Субач О.М., Сотсков В.П., Плюснин В.В., Ивлева О.А., Груздева А.М., Куницына Т.А., Ивашкина О.И., Лазуткин А.А., Малышев А.Ю., Смирнов И.В., Варижук А.М., Позмогова Г.Е., Петкевич К.Д., Анохин К.В., Ениколопов Г., и Субач Ф.В. NTnC-like genetically encoded calcium indicator with a positive and enhanced response and fast kinetics. Scientific Reports, - (year - 2018).

2. Ланин A. A. , Чеботарев A.С., Барыкина Н.В., Субач Ф.В. , и Желтиков A.M. The whither of bacteriophytochrome-based near-infrared fluorescent proteins: Insights from two-photon absorption spectroscopy. Journal of Biophotonics, - (year - 2018).

3. Груздева А.М., Ивашкина О.И., Субач Ф.В., Анохин К.В. ОПТИЧЕСКИЙ ИМИДЖИНГ НЕЙРОНОВ МОЗГА, УЧАСТВОВАВШИХ В ПОЛУЧЕНИИ НОВОГО ОПЫТА: НОВЫЕ IN VIVO И EX VIVO ПРИМЕНЕНИЯ ГЕНЕТИЧЕСКИ КОДИРУЕМЫХ КАЛЬЦИЕВЫХ СЕНСОРОВ Конференция <<Оптогенетика и оптофармакология>> 11-15 апреля 2018 г., - (year - 2018).


Annotation of the results obtained in 2016
The main aim of the project is development of unique approach for ex vivo visualization of calcium dynamics in nerve system by means of creation of genetically encoded calcium indicators of novel type, i.e. slow calcium sensors and application of this new approach to fundamental unresolved problem of neurobiology – combined visualization of calcium and genomic activities within the same neurons of the brain that occur as a result of processes of neuronal plasticity. In the year of 2016 we have found and developed two types of genetically encoded green fluorescent calcium indicators using directed molecular evolution in bacteria. These two types of calcium sensors, named NCaMP and NTnC, respond to Са2+ by notable fluorescence changes but have slow and fast temporal characteristics, respectively. Slow type NCaMP sensors have “classic” design, that includes circularly permutated version of mNeonGreen fluorescent protein and sensory part that is composed of calmodulin and M13-peptide similar to those in GCaMP6s. Their properties were characterized in vitro. NCaMP sensors responded to calcium ions by increase of their fluorescence and had high contrast of 7-9-, 24-148- and 21-25-fold. They had molecular brightness that was superior over commonly used GCaMP6s indicator. Reversible NCaMP sensors had slow calcium kinetics, i.e. their half-times of dissociation-association with Са2+ varied in a range of 5-10 min and 2-5 min, respectively. Irreversible NCaMP sensors had even more slow kinetics of dissociation from Са2+, i.e. half-time of dissociation from Са2+ reached 17 min, but this process was irreversible and repeated Са2+ addition did not cause increase of its fluorescence. Hence, as a result of search and development of slow calcium indicators we have found calcium sensors that irreversibly dissociate from Са2+ ions, as well as slow reversible sensors with different characteristic times of dissociation-association from Са2+ ions. To demonstrate possibility of fixation of slow sensors we accomplished their fixation and studied time dependence of fluorescent contrast in bacteria. After fixation with paraformaldehyde slow reversible and irreversible calcium NCaMP indicators demonstrated high contrasts between calcium-bound and calcium-free states of 4-6 and 10-fold, respectively. These contrasts preserved unchanged at least during 5 days after fixation. Hence, developed slow calcium NCaMP sensors of second type allow fixation of their calcium-bound and calcium-free states in bacteria. Properties of slow reversible sensors were characterized in mammalian cells and on neuronal cultures. In mammalian cells NCaMP indicators demonstrated even cytoplasmic localization and dramatically slowed down kinetics of dissociation from Ca2+ as compared with control fast sensor such as R-GECO. We registered reaction of NCaMP indicators in response to electrical stimulation of neurons from neuronal cultures using patch-clamp. In contrast to fast sensors such as GCaMP6s and NTnC, electrical stimulation of neurons expressing slow reversible NCaMP indicators resulted in significantly slower responses. Hence, retardation of accociation/dissociation kinetics from Ca2+ ions for NCaMP sensors was observed on neuronal culture and it was similar to results from in vitro experiments. Hence, these data suggest that slow reversible sensors of NCaMP type may be applied for fixation and ex vivo registration of calcium neuronal activity as a result of stimuli representation. NTnC fast sensor of second type has a novel design that combines the benefits of FRET-based sensors of the Twitch family and circularly-permutated FP-based sensors. We have characterized the main features of this new intensiometric sensor both in vitro and in vivo. NTnC has high brightness due to the high intrinsic brightness of its parental protein, mNeonGreen. It also maintains other beneficial characteristics of the parental protein, such as pH stability and monomeric behavior. NTnC has an inverted phenotype, i.e., it reduces its fluorescence upon the binding of Ca2+ ions. Compared with other commonly used sensors, it is ~100 amino acids smaller in size and has half the number of Ca2+-binding sites. The insertion version of the mNeonGreen protein that we have developed and its use for a new design of genetically encoded calcium indicators may help to improve the brightness of other single-fluorophore-based sensors and may assist in the generation of other mNeonGreen-based sensors. We have expressed NTnC in mammalian and neuronal cells and characterized its features in vivo. NTnC could reliably report variations in Ca2+ ion levels induced by ionomycin in mammalian cells and by spontaneous activity in dissociated neuronal cultures. Specific responses of the NTnC indicator to calcium ions have been confirmed with a control mutant, NTnC/166D+ /202D+ with Asp amino acid insertions at positions 166 and 202, in which the affinity to Ca2+ ions is abolished. The availability of such a control for NTnC opens up possibilities for further research to distinguish the impact of Ca2+ from those of other factors (such as pH) that may affect the fluorescence of the indicator. Using whole-cell patch clamp recording, we have revealed similar kinetics of Ca2+ responses in neurons expressing NTnC and GCaMP6s. We have found that NTnC shows sensitivity to single APs similar to that of GCaMP6s. This sensitivity correlates with the higher affinity of NTnC to Ca2+ ions and with its linear response at low Ca2+ concentration changes, which is due to the decreased number of Ca2+-binding sites. Furthermore, our results indicate similar impacts of NTnC and GCaMP6s expression on the electrophysiological features of the MEA-plated neurons. Finally, we have explored the potential of NTnC for in vivo applications. NTnC has revealed stimulus-evoked neuronal calcium ion activity in vivo in the visual cortex of awake mice with the help of two-photon microscopy. NTnC can also successfully visualize neuronal activity in vivo in the visual cortex of freely moving mice using an nVista head-mounted miniature microscope (notably, the high brightness and inverted phenotype of NTnC facilitated the installation procedure preceding in vivo imaging). During in vivo experiments, NTnC revealed less neuronal calcium ion activity than did GCaMP6s, which might be related to the limited dF/F dynamic range of NTnC. Importantly, the enhanced baseline brightness and inverted fluorescence responses of NTnC facilitate the identification of neurons at low resting concentrations of calcium ions and, thus, may be valuable for in vivo applications where background fluorescence makes detection of such cells problematic. We expect that further exploration of NTnC-like designs, with the aim of enhancing its d F/F dynamic range, may result in sensors with performance levels similar or superior to those of sensors with conventional designs. We have successfully applied two approaches for visualization of neuronal activity in the mice brain cortex using two-photon microscopy and developed NTnC sensor and commonly used GCaMP6s. One approach included monitoring and analysis of calcium neuronal activity in somatosensory cortex in response to mechanical stimulation of vibrissa of awake mice. In the second approach we registered calcium neuronal activity in visual cortex of awake mice in response to such complex stimulus representation as grating moving in different directions. Development of these approaches allowed optimization of experimental conditions for successful monitoring of calcium activity in animals with fixed head for further characterization of dynamics of slow and fast calcium sensors. We have successfully developed approach for visualization and analysis of calcium dynamics in mice cortex of freely-moving mice in response to stimulus representation with the help of GCaMP6s and NTnC fast calcium sensors. Establishment of this method enabled optimization of conditions for successful registration of calcium activity from freely-moving mice for further characterization of dynamics of slow calcium sensors with the help of miniaturized nVista HD microscope.

 

Publications

1. Барыкина Н.В., Субач О.М., Доронин Д.А., Сотсков В.П., Рощина М.А., Куницына Т.А., Малышев А.Ю., Смирнов И.В., Азиева А.М., Соколов И.С., Пяткевич К.Д., Бурцев М.С., Варижук А.М., Позмогова Г.E., Анохин К.В., Субач Ф.В. и Ениколопов Г.Н. A new design for a green calcium indicator with a smaller size and a reduced number of calcium-binding sites Scientific Reports, v. 6, article number: 34447 (year - 2016).


Annotation of the results obtained in 2017
The main aim of the project is development of unique approach for ex vivo visualization of calcium dynamics in nerve system by means of creation of genetically encoded calcium indicators (GECIs) of novel type and application of this new approach to fundamental unresolved problem of neurobiology – combined visualization of calcium and genomic activities within the same neurons of the brain that occur as a result of processes of neuronal plasticity. In the year of 2017 we continued to develop green calcium indicators of two types that react to calcium by changing their fluorescence, but with slow (sNCaMP and NTnC2) and fast (FGCaMP, iYTnC, iYTnC2, and insNCaMP) temporal characteristics, respectively. We also engineered the first versions of indicators that had calcium-dependent maturation, and versions of slowed photoconvertable CAMPARI-like sensors. Slow type sNCaMP sensors had “classic” design, that included circularly permutated version of mNeonGreen fluorescent protein and sensory part that was composed of calmodulin and M13-peptide similar to those in GCaMP6s. The best slow sNCaMP-like sensor, developed in the year of 2016, had a too high affinity for calcium ions that was not optimal for work in neurons. Using directed mutagenesis we found a version of the sNCaMP sensor, which had a reduced affinity for calcium ions, without affecting its slow kinetics. Next we slowed down the kinetics of the sNCaMP sensor with reduced affinity for calcium ions using random mutagenesis followed by screening. As a result, we found mutants with calcium association and dissociation half-times of 0.7-1.6 and 3.8-14.4 minutes, respectively. These times should be optimal for experiments with animal stimulation for 2-5 minutes to saturate the sensor with calcium ions and the subsequent perfusion and fixation of the sensor. Thus, we selected final versions of slow sNCaMP sensors, optimized for ex vivo fixation of the neuronal calcium activity after stimulation. To study the in vivo dynamics of the slow calcium sensor sNCaMP, we used the animal model with mechanical stimulation of vibrissae and two-photon microscopy. The fast and slow calcium dynamics of the sensor were evaluated in neurons of the 2/3 layer of the barren fields of the somatosensory cortex. Investigation of the fast fluorescence dynamics of the sNCaMP slow calcium indicator did not reveal a change in the fluorescence of the sensor on short time intervals. During longer two-photon imaging for an hour, sNCaMP showed a decrease in fluorescence in the cells with a F/F values of 121 ± 17% and a half-time of fluorescence decay of about 20 minutes. Thus, sNCaMP sensor demonstrated a slow dynamics in the mouse brain cells during two-photon in vivo imaging, optimal for ex vivo visualization. To create slow calcium sensors with calcium-dependent maturation, we obtained rational and random bacterial libraries based on green G-GECO1.2 and GCaMP6s and red R-GECO calcium indicators. The screening of the libraries was performed during their maturation in the presence and absence of calcium. As a result, we found mutants that matured and became fluorescent in bacteria only in the presence of calcium ions in libraries based on the red calcium indicator R-GECO. To evaluate the calcium-dependent-maturing sensors for ex vivo visualization of calcium activity in the brain of mice in response to sensory stimulation, we used an animal model of environment enrichment in a home cage. Three days after the viral infection, the experimental mice were exposed to the threads with beads uniformly distributed in the home cage, according to the protocol of Yang et al., 2009. The presence of mice in such an enriched environment was shown to cause significant activation of neurons in the somatosensory cortex. The control mice were left in the standard home cages. After five days of the exposure to the enriched environment, the animals were perfused and brain sections were analyzed. In the case of mice that received sensory stimulation, a higher percentage of red/green vs non-red/green fluorescent cells expressing calcium-dependent-maturing red sensors and green GFP was detected, compared to the control group. Hence, the ex vivo protocol for the visualization of calcium activity on brain sections using calcium-dependent-maturing sensors was successfully developed. The developed red calcium-dependent-maturing sensors allow the ex vivo identification of the cells in a somatosensory brain region that exhibited specific calcium activity in response to prolonged sensory stimulation. For the development of a slower and more contrast version of the green-to-red photoconvertable calcium sensor-integrator CAMPARI, we obtained and screened random bacterial libraries of CAMPARI. As a result, we found mutants named sCAMPARI that showed a 3.5-5-fold slower calcium dissociation kinetics and a 2.4-12 improved contrast, compared to the original CAMPARI. The slowed calcium dissociation kinetics and the increased contrast of CAMPARI should improve the efficiency and specificity of its calcium-dependent transition from the green to red during photoconversion, and enables the development of an alternative approach for ex vivo visualization of calcium activity. The next important specific aim was development of calcium indicators that have as a sensory part a calmodulin from other organisms, different from calmodulin from animals that is used in slow NCaMP sensors developed in the previous year. With this aim we created calcium indicators with different combinations of CaMs and M13-like peptides or calcineurins from Aspergillus fungi and Komagataella yeast. From these combinations, we chose the best one and using directed molecular evolution in bacteria, we developed a new genetically encoded cpFP-based green calcium indicator, FGCaMP, with a novel sensory M13-peptide/CaM pair from Aspergillus fungi. We have characterized the main features of this new ratiometric indicator both in vitro and in vivo. In vitro, FGCaMP demonstrates the highest brightness and dynamic range combination among other ratiometric indicators such as GEX-GECO, Pericam and Y-GECO. It has other beneficial characteristics, such as high photostability, monomeric behavior and fast kinetics of dissociation from Ca2+ ions. FGCaMP has a ratiometric phenotype, i.e., upon binding of Ca2+ ions the large Stokes shift green fluorescence reduces in its 402-form and increases in its 493-form. In contrast to the intensiometric Ca2+ sensors such as GCaMP6 variants and GECOs, ratiometric changes in the FGCaMP fluorescence enable quantitative measurements of Ca2+ concentrations at around 300 nM in live cells and help to visualize cells both at low and high Ca2+ concentrations Using site-directed mutagenesis, we found mutations that enhance the fluorescence response and FGCaMP indicator affinity to Ca2+ concentration changes of 0-1000 nM that occur during neuronal activity. The FGCaMP indicator and its enhanced versions FGCaMP2 and FGCaMP3 that include a novel sensory part from fungi may be further characterized in vivo and used as templates for the development of calcium indicators that have different fluorescent colors and properties. We further characterized the features of the FGCaMP indicator in the cytoplasm of mammalian cells and on neuronal cultures and demonstrated its novel sensory part provides beneficial high mobility. FGCaMP could reliably visualize variations in Ca2+ ion concentrations either induced by ionomycin in mammalian cells or by spontaneous activity in dissociated neuronal cultures. Using FRAP experiments we have found that in contrast to G-GECO1.2 and GCaMP6s standard green GECIs, FGCaMP has high mobility in the mammalian cells at low Ca2+ concentrations. This property may be advantageous in terms of cytotoxicity and fluorescence response in living cells. Using whole-cell patch clamp recording, we revealed faster kinetics of Ca2+ responses in neurons expressing FGCaMP compared with those expressing GCaMP6s. We also found that FGCaMP shows a lower sensitivity to APs than GCaMP6s GECI. This low sensitivity correlates with the lower affinity of the FGCaMP indicator to Ca2+ ions and with its lower fluorescence responses at low Ca2+ concentration changes in the range of 0 - 1000 nM. Hence, the utilization of enhanced versions of FGCaMP such as FGCaMP2 and FGCaMP3 may address this important limitation. Finally, we have explored the potential of the FGCaMP indicator for in vivo applications. Using fluorescence microscopy FGCaMP has successfully revealed chemically-evoked neuronal Ca2+ ion activity in vivo in the nervous system of paralyzed zebrafish. During in vivo experiments, the FGCaMP indicator demonstrated a lower F/F dynamic range and revealed less neuronal calcium ion activity than GCaMP6f. Ratiometric in vivo imaging with FGCaMP allowed visualization of resting neurons before their activation. Importantly, the ratiometric fluorescence responses of FGCaMP may enable the quantitative estimation of Ca2+ ions concentrations. We believe that further exploration of the novel sensory module from Aspergillus fungi and CaMs from other fungi species may provide enhanced sensors that demonstrate superior properties yet are distinct from GECIs with conventional sensory parts from metazoans. The next specific aim was the enhancing of dynamic range of the fast NTnC calcium indicator developed in the year of 2016. To achieve this goal, three alternative approaches had been tested. The first approach consisted in increasing the contrast of the NTnC sensor itself using rational and several rounds of random mutagenesis of the original NTnC sensor. Positions for the rational mutagenesis were selected based on the mutations found during development of the NTnC sensor and the crystal structure of the NTnC indicator resolved in 2017 in collaboration with the "Protein Factory" from the Kurchatov Institute. Using directed and random mutagenesis followed by screening, we developed the positive and inverted NTnC2 sensors with contrasts in vitro of 269 and 21-fold, respectively, which were 268- and 20-times better than contrast for the original NTnC sensor. In vitro, the NTnC2 sensors had slow calcium kinetics and a brightness similar to that for the EGFP protein. To validate the functionality of several versions of the NTnC2 indicator in neurons, we compared their response during the spontaneous activity of dissociated neuronal cultures with the responses of R-GECO1 and GCaMP6f indicators. The NTnC2 indicators responded to spontaneous neuronal activity with a similar or lower response compared to R-GECO1 responses, but with significantly slower kinetics. Their slow calcium dynamics was visualized in vivo using a two-photon microscopy in the visual cortex of the mouse brain, their half-fluorescence decay times were on the order of 40 sec. Thus, on the basis of the crystallographic data for the NTnC sensor we could obtain its versions with significantly improved contrast and slow kinetics. To increase the contrast of the NTnC sensor, we applied a second alternative approach. Using directed molecular evolution in bacteria, we developed NTnC-like GECI with enhanced fluorescence contrast and kinetics by replacing the mNeonGreen fluorescent part in the NTnC indicator with EYFP one. Similarly to NTnC, the developed indicator, named iYTnC, showed an inverted response to Ca2+. We have characterized the main properties of the newly developed indicator in vitro, in cultured mammalian cells and on neuronal cultures. iYTnC has an inverted phenotype, i.e., its green fluorescence reduces upon the binding of Ca2+ ions similarly to the NTnC progenitor. Inverted phenotype makes simple visualization of the resting neurons and cells with low concentration of free calcium as well as simplifies installation of NVista like systems for in vivo imaging. Compared with other commonly used indicators, its length is ~100 amino acids shorter and it has half the number of Ca2+-binding sites similar to the NTnC indicator. In vitro, in the presence of Mg2+ ions at a concentration that resembles that in neurons, the iYTnC indicator shows fluorescence contrast which is 2-fold higher than the contrast for the NTnC indicator. iYTnC has 5-fold lower brightness and similar photostability as compared with NTnC in vitro. According to the stopped-flow experiments, iYTnC demonstrates Ca2+ dissociation kinetics 10-fold faster or similar to that for NTnC and GCaMP6f, respectively. Depending on Ca2+ concentration in the range of 300-1300 nM, Ca2+ association kinetics of iYTnC is 4.6-1.3-fold faster than that for GCaMP6f. Fast kinetics makes iYTnC perspective indicator for monitoring neuronal calcium activity in vivo. We have expressed the iYTnC indicator and characterized its response in cultured mammalian and neuronal cells. iYTnC could reliably report variations in Ca2+ ion levels induced by ionomycin in mammalian cells and according to fluorescence contrast, iYTnC outperforms NTnC indicator in 2.2-fold. iYTnC could successfully visualize the spontaneous activity of neurons in dissociated neuronal cultures similarly to NTnC GECI. To understand and overcome modest response of iYTnC in neurons we engineered its enhanced version called iYTnC2. In contrast to iYTnC, iYTnC2 demonstrated 4-fold higher F/F response in neuronal cultures as compared with iYTnC and NTnC GECIs. According to in vitro characteristics of iYTnC2 its improved contrast in neurons could be explained by the reduced Mg2+-dependence of its contrast and affinity to Ca2+ ions. Finally, we successfully utilized iYTnC2 for in vivo visualization of neuronal dynamics in hippocampal brain area of mice. We expect that further exploration of NTnC-like design, with the aim of enhancing its brightness and F/F response in neurons, may result in indicators with performance levels similar or superior to those of GECIs with conventional designs. To enhance the contrast of the NTnC indicator using another approach, we replaced the calcium-binding domain of the NTnC sensor with another one based on a fusion protein consisting of calmodulin and M13 peptide. As a result of this replacement, we found positive and inverted insNCaMP indicators with up to 70- and 20-fold enhanced contrasts as compared with the original NTnC sensor. insNCaMP indicators had in vitro calcium dissociation and association kinetics 1.6-5-fold slower and 1.5-2.5-fold faster, respectively, compared to the GCaMP6f control indicator. Thus, we could develop versions of the NTnC indicator with significantly improved contrast and fast kinetics by the replacing its calcium-binding domain. We have successfully applied two approaches for visualization of neuronal activity in the mice brain cortex using two-photon microscopy and developed NTnC sensor and commonly used GCaMP6s. One approach included monitoring and analysis of calcium neuronal activity in somatosensory cortex in response to mechanical stimulation of vibrissa of awake mice. In the second approach we registered calcium neuronal activity in visual cortex of awake mice in response to such complex stimulus representation as grating moving in different directions. Development of these approaches allowed optimization of experimental conditions for successful monitoring of calcium activity in animals with fixed head for further characterization of dynamics of slow and fast calcium sensors. We have successfully developed approach for visualization and analysis of calcium dynamics in mice cortex of freely-moving mice in response to stimulus representation with the help of GCaMP6s and NTnC fast calcium sensors. Establishment of this method enabled optimization of conditions for successful registration of calcium activity from freely-moving mice for further characterization of dynamics of slow calcium sensors with the help of miniaturized nVista HD microscope.

 

Publications

1. Доронин Д. А., Барыкина Н. В., Субач О. М., Сотсков В. П., Плюснин В.В., Ивлева О.А., Исаакова Е. А., Варижук А. М., Позмогова Г. Е., Малышев А. Ю., Смирнов И. В., Пяткевич К. Д., Анохин К. В., Ениколопов Г. Н., и Субач Ф. В. Genetically encoded calcium indicator with NTnC-like design and enhanced fluorescence contrast and kinetics BMC Biotechnology, - (year - 2017).

2. Барыкина Н.В., Субач О.М., Пяткевич К.Д., Юнг Э., Малышев А.Ю., Смирнов И.В., Богородский А.О., Борщевский В.И., Варижук А.М., Позмогова Г.Е., Бойден Э.С., Анохин К.В., Ениколопов Г.Н., Субач Ф.В. Green fluorescent genetically encoded calcium indicator based on calmodulin/M13-peptide from fungi PLOS ONE, - (year - 2017).