Abstract
Calcium imaging with protein-based indicators is widely used to follow neural activity in intact nervous systems. The popular GCaMP indicators are based on the calcium-binding protein calmodulin and the RS20 peptide. These sensors report neural activity at timescales much slower than electrical signaling, limited by their biophysical properties and trade-offs between sensitivity and speed. We used large-scale screening and structure-guided mutagenesis to develop and optimize several fast and sensitive GCaMP-type indicators. The resulting ‘jGCaMP8’ sensors, based on calmodulin and a fragment of endothelial nitric oxide synthase, have ultra-fast kinetics (rise times, 2 ms) and still feature the highest sensitivity for neural activity reported for any protein-based sensor. jGCaMP8 sensors will allow tracking of larger populations of neurons on timescales relevant to neural computation.
Introduction
Measurement of Ca2+-dependent fluorescence using genetically encoded calcium indicators (GECIs) is one of the most widely used methods for tracking neural activity in defined neurons and neural networks 1. Recent advances have been driven in a virtuous cycle of new methods for in vivo microscopy 2–4 and engineered GECIs with higher response and sensitivity. In particular, the green fluorescent protein- (GFP-) based GCaMP sensors 5–8 have been iteratively engineered to enhance the signal-to-noise ratio (SNR) for detection of Ca2+ entering neurons through voltage-gated channels. The widely used GCaMP6 8 and jGCaMP7 6 sensors enable detection of single action potentials under favorable conditions. They can be used to monitor the activity of large groups of neurons using two-photon microscopy and wide-field fluorescence imaging 2. They have also been used to measure activity-induced calcium changes in small subcellular compartments such as dendritic spines 8–11 and axons 12, 13.
Electrical signals propagate through neural circuits over timescales of milliseconds. Determining how the activity of one set of neurons influences another and ultimately animal behavior requires tracking activity on concomitant time scales. Action potentials produce essentially delta function-like calcium currents, yielding large, rapid (<1 ms) increases in cytoplasmic calcium 14. Free calcium ions typically bind fluorescent calcium indicators very rapidly. For example, millisecond timescale detection of action potentials has been demonstrated with synthetic fluorescence calcium indicators in vivo 15–17. However, the kinetics of GECI fluorescence change is limited by sensor biophysics; for instance, in response to single action potentials in pyramidal neurons GCaMPs have fluorescent rise-times (50%) on the order of 100 ms 6–8, 18–20. Consequently, GCaMPs are often used to map relatively static representations of neural information, rather than tracking the rich dynamics in neural circuits 21, 22.
Previous attempts to improve GCaMP kinetics have been only partially successful. Among the GCaMP6 and jGCaMP7 indicators, the “f” (fast) variants were optimized for kinetics. They have risetimes of ∼50 ms, but with reduced sensitivity compared to their slower siblings (“s”, sensitive variants). Generally, attempts to improve ΔF/F0 are associated with a slowing of kinetics 6, 8, 19, 20. The mechanisms underlying this trade-off are not simply due to binding affinity for Ca2+ and in general are not well-understood. For example, the kinetics are sensitive to mutation of the RS20-CaM interface, far from the Ca2+-binding EF hands 8, 19, 20. In addition to these point mutants, the RS20 peptide has been swapped for that from CaM- dependent kinase kinase CaMKK-α/β (ckkap peptide) in the “XCaMP” sensors 23 – as well as the red GECIs R-CaMP2 24 (actually an RGECO variant) and K-GECO1 25 – with mixed effects on affinity, kinetics, and ΔF/F0.
Here we present bright, sensitive GCaMP sensors with dramatically improved kinetics. jGCaMP8 sensors include: jGCaMP8s (fast rise, slow decay, sensitive), jGCaMP8f (fast rise, fast decay), jGCaMP8m (medium decay). All jGCaMP8 sensors have nearly 10x faster fluorescence rise-times than previous GCaMPs and can track individual spikes in neurons with spike rates of ∼50 Hz. jGCaMP8 sensors are also more linear than previous GCaMPs, allowing robust deconvolution for spike extraction. The jGCaMP8 sensors were tested in vivo in mice, flies and fish, and were found to provide better performance across all metrics relevant to imaging neural populations in vivo.
Results
Sensor design and optimization
Various calmodulin-binding peptides (Supp. Table 1) were chosen from the Protein Data Bank and cloned into GCaMP6s in place of the RS20 peptide. Based on fast kinetics, saturating ΔF/F0, apparent Kd, Hill coefficient, and apparent brightness, we prioritized variants based on peptides from endothelial nitric oxide synthase (PDB 1NIW; peptide “ENOSP”) and death-associated protein kinase 1 (1YR5; peptide “DAPKP”) for optimization (Methods). The two linkers 26 were systematically mutated, and sensors were screened for high signal change and retained fast kinetics. ∼35 promising sensors were then tested in response to action potentials (APs) elicited in cultured neurons in 96-well plates (Methods). Action potentials produce essentially instantaneous increases in calcium 14 and are therefore ideal to screen for GECIs with fast kinetics 27. Fluorescence changes were extracted from multiple single neurons per well. Sensors were evaluated according to several properties (Supp. Table 2): sensitivity (response to 1 AP), dynamic range (response to a high-frequency train of 160 APs), kinetics (rise and decay times), and baseline brightness. Sensors based on DAPKP showed fast decay time and good sensitivity compared to jGCaMP7f – but with slow rise times (Supp. Table 2). Sensors with ENOSP had similar sensitivity and significantly faster rise and decay times than jGCaMP7f.
We prioritized ENOSP-based sensors for further optimization. ENOSP variant jGCaMP8.410.80 (linker 1 Leu-Lys-Ile) showed 1.8-fold faster half-rise time and 4.4-fold faster half-decay time than jGCaMP7f, with similar resting brightness and dynamic range, but 35% lower 1-AP response. We solved the crystal structure of jGCaMP8.410.80 (Fig. 1A, Supp. Fig. 1A, Supp. Table 3). The structure is similar to previous GCaMPs; the major differences are at the 3-way interface between cpGFP, CaM, and the new ENOSP peptide; the slight twist of the ENOSP peptide relative to the RS20 peptide; and at the first helix of EF-hand 1 (Supp. Fig. 1A). In jGCaMP8.410.80, Ile 32 (occupying the same space as GCaMP5G-Glu60) packs closely (both hydrophobic and C-H/π interactions) with Tyr352 (Tyr380 in GCaMP5G), allowing the tyrosine to penetrate more deeply into the cpGFP-CaM interface, and forming a water-mediated hydrogen bond network with the chromophore (Supp. Fig. 1B). This tyrosine was the core improvement in GCaMP5 26 – increasing brightness and ΔF/F0 – the ENOSP peptide appears to further facilitate this interaction. Guided by the structure, we targeted interface sites (Supp. Fig. 1C) for site-saturation mutagenesis and tested the variants in cultured neurons for higher sensitivity and retained fast kinetics in detecting spikes. Several single mutations improved properties (Supp. Table 2), particularly residues near the ENOSP C-terminus and the cpGFP-CaM interface. Beneficial point mutations were combined in subsequent rounds of screening.28, 29
Mutagenesis and screening in neurons covered 776 total sensor variants, of which 683 (88%) produced detectable responses to 1 AP (Supp. Fig. 2, Supp. Table 2). Kinetics were improved relative to the previous fast sensor jGCaMP7f. Specifically, compared to jGCaMP7f, the half-rise time (trise,1/2) was significantly shorter in 48% of screened variants, the time-to-peak fluorescence (tpeak) was significantly shorter in 47%, and the half-decay time (tdecay,1/2) was significantly shorter in 40%. Sensitivity (1-AP ΔF/F0) was higher than jGCaMP7f in 19%, and only 2% of variants had increased saturation response (160-AP ΔF/F0). Together, the mutagenesis produced a large set of variants with significant improvement in kinetics and sensitivity (Supp. Table 2).
jGCaMP8 characterization
Three high-performing “jGCaMP8” variants were selected for additional characterization (Fig. 1B-D, Supp. Fig. 3). jGCaMP8f (“fast”) exhibited 1-AP trise,1/2 of 7.0±0.7 ms, and 1-AP tpeak of 24.9±6.0 ms, more than 3- and 5-fold shorter than jGCaMP7f, respectively. We note that the rise-time measurements in cultured neurons were limited by the frame rate of the camera (200 Hz) and thus constitute an overestimate. jGCaMP8s (“sensitive”) exhibited 1-AP ΔF/F0 of 1.1±0.2, and 1-AP signal-to-noise ratio (SNR) of 41.3±10.4, approximately twice that of the most sensitive GECI to date, jGCaMP7s. jGCaMP8m (“medium”) is a useful compromise between sensitivity and kinetics: it exhibits 1-AP ΔF/F0 and 1-AP SNR comparable to jGCaMP7s, and kinetics comparable to jGCaMP8f, with the exception of a slower half-decay time (tdecay,1/2, 134±14 vs. 92±22 ms; Fig. 1C). The fast kinetics and high sensitivity of the jGCaMP8 indicators allowed resolution of electrically evoked spikes at frequencies of up to 40 Hz (Fig. 1D). When stimulated with short bursts consisting of 3 and 10 APs, the jGCaMP8 sensors retained fast kinetics and high sensitivity. Overall, the jGCaMP8 series exhibited significant, multi-fold improvements across several parameters over previous GECIs.
We compared the jGCaMP8 sensors to the XCaMP series (green XCaMP variants XCaMP-G, XCaMP-Gf, and XCaMP-Gf0 23, side-by-side in cultured neurons. The 1-AP ΔF/F0 was significantly higher for all jGCaMP8 sensors; the 1-AP SNR was significantly higher for jGCaMP8m and jGCaMP8s, 1-AP trise,1/2 was significantly shorter for all jGCaMP8 sensors, 1- AP tpeak was significantly shorter for jGCaMP8f and jGCaMP8m, and tdecay,1/2 was significantly shorter for jGCaMP8f, when evaluated against all XCaMP sensors (Supp. Fig. 3, Supp. Table 4). The baseline fluorescence of the jGCaMP8 series was similar to jGCaMP7f, and significantly higher than the XCaMP sensors (Supp. Fig. 4). Photobleaching was also similar between jGCaMP7f and the jGCaMP8 sensors (Supp. Fig. 5).
GECIs with linear (i.e., Hill coefficient ∼1) fluorescence responses to AP trains provide a larger effective dynamic range for quantifying spike rates and facilitate applications such as counting spikes within trains. We tested GCaMP sensors with bursts (83 Hz) containing different numbers (1-40) of action potentials. Given their higher sensitivity, 8m and 8s showed saturation behavior for smaller numbers of spikes compared to jGCaMP7. However, they behaved nearly linearly up to 10 spikes (Supp. Fig. 6). Finally, fluorescence recovery after photobleaching (FRAP) revealed that the jGCaMP8 variants showed similar diffusion in neurons compared to previous GECIs 27 (Supp. Fig. 7A-C) and independent of calcium (Supp. Fig. 7D), suggesting that they do not have altered cellular interactions.
Imaging in Drosophila adult visual system and larva
GCaMP responses to visual stimulation were compared in Drosophila laminar monopolar L2 neurons (Fig. 2A), part of the OFF-motion visual system 28–30. Imaging was performed where L2 dendrites connect to columns in medulla layer 2. These non-spiking neurons depolarize during light decrease and hyperpolarize during increase. Fluorescence responses to visual stimulation were measured in multiple single neurons in individual animals (Fig. 2B, Supp. Fig. 8A). XCaMP-Gf was too dim to image (Supp. Fig. 8B-C) and was excluded from further study. At light-dark and dark-light transitions, all jGCaMP8 variants showed significantly faster rise, and 8m showed faster decay, respectively, than 7f (Fig. 2C,D; half-rise times: 7f, 128±11 ms; 8f, 76±8; 8m, 58±6; 8s, 80±8 ms; decay times: 7f, 277±29 ms; 8m, 137±21). 8m and 8f also showed markedly larger ΔF/F0 than 7f following light-on (Fig. 2B-C). All three jGCaMP8 indicators revealed a negative off-response after light-off (i.e., hyperpolarization below baseline), whereas 7f was too slow (Fig. 2C). Flies were next subjected to light on-off stimulation at frequencies from 0.5-30 Hz. 8m and 8f showed much higher spectral density than 8s across all frequencies, and higher than 7f above 2 Hz (Supp. Fig. 8D). Visual stimuli of progressively shorter lengths were shown, from 25 ms dark flash down to 4 ms. 8m and 8f showed higher ΔF/F0 at all stimulus lengths (Supp. Fig. 8E-top). Quantification of stimulus detection by the discriminability index d’ 31 showed that 8m and 8f provide markedly superior stimulus detection above noise than 7f and 8s across all stimulus frequencies (Supp. Fig. 8E-bottom). The jGCaMP8 variants were somewhat dimmer than 7f because of lower expression (Supp. Fig. 8B-C, Supp. Fig. 9) but were sufficiently bright to provide high-SNR imaging.
Next, we imaged the GECIs at presynaptic boutons of the larval neuromuscular junction in response to precise electrical stimulation of motor axons (Supp. Fig. 10). jGCaMP8 variants showed large responses, with faster rise and decay times than 7f (Supp. Fig. 10B, D, E). The jGCaMP8 series detect individual stimuli much better than 7f at low frequencies and easily resolve spikes in 20 Hz trains (Supp. Fig. 10H), whereas previous GECIs cannot.
Imaging in zebrafish optic tectum
We also imaged 8f, 7f, and 6f in zebrafish as histone H2B fusions (which improves cell segmentation 32; Methods). Transgenic Tg(elavl3:H2B-8f), Tg(elavl3:H2B-7f) and Tg(elavl3:H2B-6f) larval zebrafish were shown flashing visual stimuli and GCaMP transients were imaged in the optic tectum. 8f showed markedly faster rise (Supp. Fig. 11E) and decay (Supp. Fig. 11F) times than its predecessors.
Imaging neural populations in mouse primary visual cortex
We next tested the jGCaMP8 sensors in L2/3 pyramidal neurons of mouse primary visual cortex (V1). We made a craniotomy over V1 and infected neurons with adeno-associated virus (AAV2/1-hSynapsin-1) 33(Methods) encoding jGCaMP8 variants, 7f 6, or XCaMP-Gf 23. After three weeks of expression, mice were lightly anesthetized and mounted under a custom two-photon microscope. Full-field, high-contrast drifting gratings were presented in each of eight directions to the contralateral eye (Fig. 3A). Two-photon imaging (30 Hz) was performed of L2/3 somata and neuropil (Methods).
With the jGaMP8 indicators, visual stimulus-evoked fluorescence transients were observed in many cells (Fig. 3B,C; three representative cells shown for 8s and 8f) and were stable across trials (Supp. Fig. 12). All sensors produced transients with rapid rise and decay (Fig. 3B-E). Nearly identical responses were measured after long-term expression of jGCaMP8 (five additional weeks; Supp. Fig. 13). XCaMP-Gf was much dimmer (10x) than the others (Supp. Fig. 14A-B) with few responsive cells, precluding sensitivity analysis. As protein expression levels were similar across indicators (Supp. Fig. 14C-D), XCaMP-Gf is deficient in maturation or brightness in vivo and was not studied further.
The contrast changes in visual stimuli were tracked faithfully by sensor responses (Fig. 3B,C). Consistent with in vitro characterization, jGCaMP8f showed significantly shorter half-decay time (median[1st-3rd quartiles] = 84[32-153] ms) than 7f (110[41-223] ms, P < 0.05) and comparable to 8m (84[32-165] ms) and XCaMP-Gf (91[48-155] ms) (Fig. 3E). On the other hand, 8s decay was significantly slower than the other indicators.
We quantified indicator sensitivity both as the proportion of labeled neurons responsive 6, 8 to visual stimuli (Fig. 3F) and as the cumulative distribution of peak ΔF/F0 across cells (Fig. 3G). Significantly more responsive cells were seen for 8s and 8m than for 8f and 7f (Fig. 3F; P < 0.001). Furthermore, 8s was dramatically right-shifted relative to the other indicators (Fig. 3G), reflective of its high sensitivity and saturating ΔF/F0. SNR of visually evoked fluorescence transients was significantly higher for 8s than for other sensors, followed by 8m and 7f, then by 8f (Fig. 3G).
Orientation tuning was similar for all sensors, except that 8m and 8s revealed a larger proportion of neurons with low orientation selectivity (Supp. Fig. 15). A plausible explanation for this is that the high-sensitivity indicators detect activity of GABAergic interneurons that is missed by the other sensors. (Interneurons yield smaller fluorescence responses 8, and have much less sharp tuning than excitatory neurons 34.) This hypothesis is supported by experiments with simultaneous imaging and electrophysiology (see below).
Simultaneous imaging and electrophysiology
To quantify GECI responses to neural activity, we combined two-photon imaging (122 Hz) and loose-seal, cell-attached electrophysiological recordings (Fig. 4A). We compared fluorescence changes and spiking across sensors (n=40 cells from 8 mice, 8f; n=47 cells from 7 mice, 8m; n=49 cells from 7 mice, 8s; n=23 cells from 5 mice, 7f; Supp. Fig. 16, Supp. Fig. 17, Supp. Table 5). Fluorescent signals for cell body regions of interest (ROIs) were corrected for neuropil signal 6, 8 (Supp. Fig. 18). All jGCaMP8 variants produced large fluorescence transients even in response to single action potentials (APs) (Fig. 4B-D).
Our experiments allowed us to resolve fluorescence transients with much higher effective temporal resolution than the nominal 122 Hz frame rate. Specifically, fields of view were chosen so that the patched neuron occupied <20% of the frame’s scan lines (Supp. Fig. 19). Since neurons were scanned at random phases with respect to AP fluorescence transients, neurons of interest were sampled with an effective temporal resolution of 500-600 Hz. All three jGCaMP8 variants showed rise (0 - 50%) times <5 ms, approximately 5x faster than 7f (Fig. 4C-E). Peak responses for all jGCaMP8 indicators were also larger than for 7f (Fig. 4D-E). To study spike-time estimation, we first binned action potential doublets with respect to their inter-spike interval length. The jGCaMP8 indicators resolved individual action potentials from doublets at spike rates of up to 100 Hz (Fig. 4F). We subsequently grouped spike bursts based on the number of APs (from 1 to 5) in a 20 ms integration window. All sensors show monotonic increases in fluorescence response with AP count, with the jGCaMP8 sensors responding more linearly than jGCaMP7f (Fig. 4G). This greater linearity is consistent with neuronal culture results and lower Hill coefficients in purified protein (Supp. Table 6).
The synapsin-1 promoter yields expression in many neurons, including both pyramidal cells and fast-spiking (FS, presumably parvalbumin expressing) interneurons 35 – which are interspersed in our imaged ROIs. Out of our recorded neurons, we identified the subset of FS spiking interneurons by their high spike rates and short spike durations (Supp. Fig. 20A)36. All three jGCaMP8 sensors produced robust responses (Supp. Fig. 20B; ∼3% ΔF/F0 on average, with responses up to 5%) to single APs in FS interneurons, much larger than GCaMP6s (∼1% ΔF/F0)8.
Taken together, in mouse cortex in vivo, the jGCaMP8 sensors show excellent single-spike detection and spike train deconvolution, spike time estimation, good expression, strong performance in fast-spiking interneurons, and no evidence of adverse effects of long-term expression.
We also tested the jGCaMP8 variants alongside 6f and 7f in mouse cerebellar Purkinje cell dendritic arbors with the Pcp2-Cre mouse line (Supp. Fig. 21A-B). All jGCaMP8 variants showed faster rise-time than 6f and 7f (Supp. Fig. 21C-D), and decay-time was markedly faster for 8m and 8f (Supp. Fig. 21E).
Spike train modeling with jGCaMP8
Calcium-evoked fluorescence is an indirect measure of neural activity 8, 37. A large body of work has been devoted to estimating spike trains from calcium imaging data. The performance of these deconvolution algorithms is largely limited by linearity, sensitivity and kinetics of the calcium-dependent sensors 18, 38–40. We tested the effects of the faster kinetics, superior linearity, and high SNR of the jGCaMP8 indicators on state-of-the-art spike-fitting models 37 (Methods), using our simultaneous imaging and electrophysiology data (Fig. 4; Fig. 5A). We compared the variance explained across linear and non-linear (“sigmoid”) models and measured to what extent nonlinearities are required to fit fluorescence dynamics for different indicators (Fig. 5B).
Linear models performed much better for jGCaMP8 than GCaMP6 or jGCaMP7f in fitting fluorescence traces (8m and 8s had the best variance explained, at 85.0 ± 1.8% and 78.3 ± 2.8%, respectively; mean ± s.e.m.; Supp. Table 7), reflecting their improved linearity (Fig. 5B-C; Supp. Table 7), SNR, and kinetics (Supp. Fig. 22A-G; Supp. Table 8). Model estimates of rise- and decay-time constants are consistent with direct measurement (Supp. Fig. 22C, F). Moreover, the model showed that the jGCaMP8 indicators maintain linearity over a wide range of activity regimes, whereas jGCaMP7f showed both sub-linearity at low and supra-linearity at high spike rates (Fig. 5C; Supp. Fig. 22H). Consistent with their linearity and high sensitivity, 8m and 8s outperformed other indicators in recovering spike timing in widely used deconvolution algorithms 38 (Supp. Fig. 22I). Altogether, the linear model showed excellent performance at fitting activity profiles generated from the three jGCaMP8 indicators.
Discussion
jGCaMP8s is the most sensitive GECI available, and jGCaMP8f is the fastest. jGCaMP8m shows intermediate sensitivity and speed. The rapid kinetics lead to high SNR for small transients and excellent deconvolution of spikes in trains.
The jGCaMP8 indicators feature a different calmodulin (CaM)-binding peptide (from endothelial nitric oxide synthase) than previous GECIs (mostly the RS20 peptide from myosin light-chain kinase). The XCaMP sensors as based on the ckkap peptide from CaM-dependent kinase kinase CaMKK-α/β. In our hands, the XCaMP sensors were quite dim in all preparations – even with high-level expression – precluding side-by-side comparison. 43–45
In addition to the fast kinetics and large fluorescence response, the jGCaMP8 indicators are more linear with respect to [Ca2+] and spike number. This linearity provides more interpretable data. The new sensors will provide better estimates of the number and timing of transients, both for action potentials as demonstrated here, and for synaptic inputs – which will provide a clearer picture of the size and relative timing of inputs into spines and dendrites in single neurons and populations.
Previous generations of “fast GCaMPs” 19, 20 suffered from low signal-to-noise, as ΔF/F0 was compromised in the pursuit of kinetics. The combination of large-scale peptide grafting and massive screening of linker and interface mutants was able to overcome the intrinsic barriers to simultaneously maintaining fast kinetics and high SNR. The jGCaMP8 indicators showed similarly good performance after expression for several months as previous GCaMPs. Long-term, high-level GCaMP expression can result in “cytomorbid” cells with altered trafficking 7 and even epileptiform activity in mice 41. The sensor GCaMP-X 42 incorporates a second calmodulin-binding peptide as a “protection motif,” which decreases calcium channel disruption and nuclear accumulation. A similar strategy could be employed with the new sensors, although it is possible that incorporation of this second peptide would disrupt their fast kinetics and high SNR.
The jGCaMP8 indicators have improved kinetics across neuron culture, flies, fish, and mice. Furthermore, 8m and 8f also performed remarkably well in fast-spiking (likely parvalbumin) cortical interneurons, with large SNR and rapid kinetics, despite the small free calcium increases per action potential in these neurons 44. The jGCaMP8 indicators will be the reagents of choice for most calcium imaging applications.
Methods
All surgical and experimental procedures were conducted in accordance with protocols approved by the Institutional Animal Care and Use Committee and Institutional Biosafety Committees of Howard Hughes Medical Institute (HHMI) Janelia Research Campus, and of the corresponding committees at the other institutions.
Sensor design
We surveyed the Protein Data Bank for unique structures of calmodulin (CaM) in complex with a single peptide. Twenty-nine peptides were sufficiently different from the RS20 peptide sequence used in previous GCaMPs to warrant testing (Supp. Table 1). The structures of these complexes were superimposed on the GCaMP2 structure (PDB 3EK4) in PyMOL, and amino acids were added or removed to bring all peptides to a length estimated to work well in the GCaMP topology. Synthetic DNA encoding each of the 29 peptides replaced the RS20 peptide in the bacterial expression vector pRSET-A-GCaMP6s. Of the initial sensors, 20/29 sensed calcium. All 20 had lower saturating ΔF/F0 than GCaMP6s, all but three had weaker Ca2+ affinity (apparent Kd) than GCaMP6s, all but one had lower cooperativity (Hill coefficient, n), and many were dimmer (Supp. Table 1). Several sensor variants showed much faster Ca2+ decay kinetics, as determined by stopped-flow fluorescence on purified protein (Supp. Table 1). Based on fast kinetics, saturating ΔF/F0, apparent Kd, Hill coefficient, and apparent brightness, we prioritized variants based on the peptides from endothelial nitric oxide synthase (PDB 1NIW; peptide “ENOSP”) and death-associated protein kinase 1 (1YR5; peptide “DAPKP”) for optimization (Supp. Table 1).
Sensor optimization
These two sensor scaffolds were optimized in protein purified from Escherichia coli expression. Libraries were constructed to mutate the linker connecting the peptide to cpGFP (linker 1) 26 and screened for high signal change and retained fast kinetics. The linker connecting cpGFP and CaM (linker 2) was similarly mutated on top of variants from the optimization of linker 1. Out of 4000 ENOSP-based variants and 1600 DAPKP-based variants, 23 and 10 respectively had fast kinetics and high saturating ΔF/F0 in purified protein (data not shown).
Guided by the structure of jGCaMP8.410.80, we targeted 16 interface positions for site-saturation mutagenesis: 7 in ENOSP, 4 on cpGFP, and 5 on CaM (Supp. Fig. 1B). Sensor variants were tested in cultured neurons for higher sensitivity in detecting neural activity while maintaining fast kinetics. Several single mutations improved properties (Supp. Table 2), particularly residues near the ENOSP C-terminus and the cpGFP-CaM interface. Beneficial point mutations were combined in subsequent rounds of screening. Ten additional CaM positions (Supp. Fig. 1B) surrounding ENOSP were next subjected to site-saturation mutagenesis. Finally, mutations (Supp. Fig. 1B) from the FGCaMP sensor (developed using CaM and RS20-like peptide sequences from the fungus Aspergillus niger and the yeast Komagataella pastoris) 43, 44 were introduced to improve biorthogonality and/or kinetics.
Sensor screen and characterization in solution
Cloning, expression, and purification of sensor variants in Escherichia coli, calcium titrations, pH titrations, kinetic assay, and photophysical analysis were performed essentially as described before 26.
In this study, the RSET tag (His6 tag-Xpress epitope-enterokinase cleavage site), which had been carried over from the pRSET-A cloning vector in earlier work, was removed from all sensors: constructs simply encode a hexa-histidine affinity tag: Met-His6 tag-peptide-linker 1- cpGFP-linker 2-CaM. For the screen of linkers replacing RS20 (previously mistakenly referred to as “M13”), libraries of sensors in the pRSET-A bacterial expression vector were generated using primers containing degenerate codons (NNS) with Q5 site-directed mutagenesis (New England BioLabs) and transformed into T7 Express competent cells (New England BioLabs). A sequence encoding six repeats of the Gly-Gly-Ser tripeptide was designed as a highly flexible, presumably non-CaM-binding negative control. We expressed the new variants, as well as the presumptive Gly-Gly-Ser negative control and GCaMP6s as a positive control, in Escherichia coli T7 Express. Single colonies were picked and grown in 800 µL ZYM-5052 autoinduction medium containing 100 µg/mL ampicillin in 96 deep-well blocks for 48 hours at 30°C. Cells were collected by centrifugation, frozen, thawed, and lysed. Clarified lysate was used to estimate the dynamic range by measuring fluorescence in the presence of 1 mM Ca2+ or 1 mM EGTA.
For protein purification, T7 Express cells containing sensors were grown at 30°C for 48 hours in ZYM-5052 autoinduction medium with 100 µg/mL ampicillin. Collected cells were lysed in 1/50 volume of B-PER (Thermo Fisher) with 1 mg/mL lysozyme and 20 U/mL Pierce Universal Nuclease (Thermo Fisher) and subsequently centrifuged. Supernatants were applied to HisPur Cobalt Resin (Thermo Fisher). The resin was washed with 20 column volumes of 20 mM Tris, pH 8.0, 300 mM NaCl, 1 mM imidazole, followed by 10 column volumes of 20 mM Tris, pH 8.0, 500 mM NaCl, 5 mM imidazole. Proteins were eluted into 20 mM Tris, pH 8.0, 100 mM NaCl, 100 mM imidazole.
For calcium titrations, sensors were diluted 1:100 in duplicate into 30 mM MOPS, pH 7.2, 100 mM KCl containing either 10 mM CaEGTA (39 µM free calcium) or 10 mM EGTA (0 µM free calcium). As before, these two solutions were mixed in different amounts to give 11 different free calcium concentrations. GCaMP fluorescence (485 nm excitation, 5 nm bandpass; 510 nm emission, 5 nm bandpass) was measured in a Tecan Safire2 plate reader (Tecan). The data was fit with a sigmoidal function using KaleidaGraph (Synergy Software) to extract the apparent Kd for Ca2+, the Hill coefficient, and dynamic range.
koff was determined at room temperature using a stopped-flow device coupled to a fluorimeter (Applied Photophysics). Each sensor variant in 1 µM Ca2+ in 30 mM MOPS, pH 7.2, 100 mM KCl was rapidly mixed with 10 mM EGTA in 30 mM MOPS, pH 7.2, 100 mM KCl. Fluorescence decay data was fit with a single or double exponential decay function.
For pH titrations, purified proteins were diluted into pH buffers containing 50 mM citrate, 50 mM Tris, 50 mM glycine, 100 mM NaCl and either 2 mM CaCl2 or 2 mM EGTA, which were pre-adjusted to 24 different pH values between 4.5 and 10.5 with NaOH. A sigmoidal function was used to fit fluorescence versus pH, and the pKa value was determined from the midpoint.
Sequence & structural analysis of variants
Linker1 encodes Leu-Glu in GCaMP6s (and indeed, in all previous RS20-based GCaMP sensors – this linker was extensively mutated in the GCaMP5 screen 26, but the best variant, GCaMP5G, retained Leu-Glu); we first mutated Leu-Glu to fully degenerate 2-amino acid (aa) sequences and screened for variants with both high signal change and retained fast kinetics. Following selection of the best 2-aa linkers, these variants were expanded to libraries of 3-aa linkers by addition of fully degenerate residues. After optimization of Linker 1, Linker 2 was mutated from Leu-Pro, to which it had been selected in GCaMP5G 26, the parent of GCaMP6 and GCaMP7. Linker 2 mutagenesis was similar to that for Linker 1, but alternative Linker 2 sequences either slowed kinetics or decreased ΔF/F0, and Linker 2 was thus retained as Leu-Pro.
In addition to 8f/m/s, several other variants may be of interest, including 455, 543, 640, 707, and 712 (Supp. Table 2). All promising variants contain, in addition to the Leu-Lys-Ile linker 1, additional mutations to the ENOSP peptide: Asn19Thr and Ser24Ile appear in every variant except 712, Ser26Arg appears in every variant but jGCaMP8s (with Ser26Met), jGCaMP8m has Ala25Gly, and 712 has Met28Ser. Every variant contains the Gln88Glu mutation at the CaM-GFP interface. Further mutations include Phe286Tyr (8s, 8m, and 707); Glu288Gln (707); Gln315Leu (8f), Gln315His (8s, 707), Gln315Lys (455); Met346Gln (543); and Met419Ser (640). Of these, Phe286Tyr comes from the FGCaMP sensor; all others are unique to this work. Importantly, GCaMP6s data from both purified protein and cultured neurons are essentially identical between this work (lacking the RSET tag) and previous work (with it) (data not shown) – implying that the RSET tag does not noticeably modulate GCaMP function in protein and neuronal culture and that observed jGCaMP8 improvements stem from the peptide substitution and other mutations.
Photophysical measurements
All the measurements were performed in 39 μM free Calcium(+Ca) buffer (30 mM MOPS, 10 mM CaEGTA in 100 mM KCl, pH 7.2) or 0 μM free Calcium(-Ca) buffer (30 mM MOPS, 10 mM EGTA in 100 mM KCl, pH 7.2). Absorbance measurements were performed using UV-Vis spectrometer (Cary 100, Agilent technologies), and fluorescence excitation-emission spectra were measured using a fluorimeter (Cary Eclipse, Varian Inc.). ΔF/F0 between proteins in +Ca and -Ca buffer was calculated from the fluorescence emission spectra. Quantum yield measurements were performed using an integrating sphere spectrometer (Quantaurus, Hamamatsu) for proteins in +Ca buffer. Extinction coefficients were determined via the alkali denaturation method, using extinction coefficient of denatured GFP as a reference (ε = 44000 M-1cm-1 at 447 nm).
Two-photon spectroscopy
The two-photon excitation spectra were performed as previously described 26. Protein solutions of 1-5 μM concentration in +Ca or -Ca buffer were prepared and measured using an inverted microscope (IX81, Olympus) equipped with a 60X, 1.2NA water immersion objective (Olympus). Two-photon excitation was obtained using an 80 MHz Ti-Sapphire laser (Chameleon Ultra II, Coherent) with sufficient power from 710 nm to 1080 nm. Fluorescence collected by the objective was passed through a short-pass filter (720SP, Semrock) and a bandpass filter (550BP200, Semrock) and detected by a fiber-coupled Avalanche Photodiode (APD) (SPCM_AQRH-14, Perkin Elmer). The obtained two-photon excitation spectra were normalized to 1 μM concentration and subsequently used to obtain the action cross-section spectra (AXS) with fluorescein as a reference (Average AXS from 45, 46).
Fluorescence correlation spectroscopy (FCS) was used to obtain the 2P molecular brightness of the protein molecule. The peak molecular brightness was defined by the rate of fluorescence obtained per total number of emitting molecules. 50-100 nM protein solutions were prepared in +Ca buffer and excited with 930 nm wavelength at various power ranging from 2-30 mW for 200 sec. Emission fluorescence was collected by an APD and fed to an autocorrelator (Flex03LQ, Correlator.com). The obtained autocorrelation curve was fit to a diffusion model 47 to determine the number of molecules <N> present in the focal volume. The 2-photon molecular brightness (ε) at each laser power was calculated as the average rate of fluorescence <F> per emitting molecule <N>, defined as ε = <F>/<N> in kilocounts per second per molecule (kcpsm). As a function of laser power, the molecular brightness initially increases with increasing laser power, then levels off and decreases due to photobleaching or saturation of the protein chromophore in the excitation volume. The maximum or peak brightness achieved, <εmax>, represents a proxy for the photostability of a fluorophore.
Screening in neuronal cell culture
GCaMP variants were cloned into an hSyn1-GCaMP-NLS-mCherry-WPRE expression vector, and XCaMP variants (XCaMP-G, XCaMP-Gf, XCaMP-Gf0) were cloned into an AAV-hSyn1-XCaMP-NES vector. We used the nuclear export sequence (NES) for the XCaMP sensors as this was how they were characterized in the original publication23. As this excludes the XCaMP sensors from the nucleus, where Ca2+ signals are slower 48, whereas the variants developed here were not explicitly excluded (although GCaMPs without an explicit NES are nevertheless fairly nuclear-excluded), this will make the XCaMPs appear faster than they really are compared to the GCaMP indicators.
The primary rat culture procedure was performed as described6. Briefly, neonatal rat pups (Charles River Laboratory) were euthanized, and neocortices were dissociated and processed to form a cell pellet. Cells were resuspended and transfected by combining 5×105 viable cells with 400 ng plasmid DNA and nucleofection solution in a 25 µL electroporation cuvette (Lonza). Electroporation of GCaMP mutants was performed according to the manufacturer’s protocol.
Neurons were plated onto poly-D-lysine (PDL) coated, 96-well, glass-bottom plates (MatTek) at ∼1×105 cells per well in 100 µL of a 1:2 mixture of NbActiv4 (BrainBits) and plating medium (28 mM glucose, 2.4 mM NaHCO3, 100 µg/mL transferrin, 25 µg/mL insulin, 2 mM L-glutamine, 100 U/mL penicillin, 10 µg/mL streptomycin, 10% FBS in MEM). Typically, each plate included GCaMP6s (8 wells), GCaMP6f (8 wells), and jGCaMP7f (8 wells). Other wells were electroporated with mutated variants (4 wells per variant), for a total of 80 wells (the first and last columns in the plate were not used). Plates were left in the incubator at 37°C and 5% CO2.
On DIV 14-19, neurons underwent field stimulation and imaging6, 49. Fluorescence time-lapse images (200 Hz; total of 7 seconds) were collected on an Olympus IX81 microscope using a 10x, 0.4 NA objective (UPlanSApo, Olympus) and an ET-GFP filter cube (Chroma #49002). A 470 nm LED (Cairn Research) was used for excitation (intensity at the image plane, 0.34 mW/ mm2). Images were collected using an EMCCD camera (Ixon Ultra DU897, Andor) with 4×4 binning, corresponding to a 0.8 mm x 0.8 mm FOV. Reference images (100 ms exposure) were used to perform segmentation. Red illumination for variants co-expressing mCherry was performed with a 590 nm LED (Cairn Research) through an ET-mCherry filter cube (Chroma #49008) with an intensity of 0.03 mW/mm2. Trains of 1, 3, 40, and 160 field stimuli were delivered with a custom stimulation electrode. For sensor linearity measurements, 1, 2, 3, 5, 10, and 40 field stimuli were delivered. For frequency response experiments, two-photon imaging was performed (objective: 16x, 0.8 numerical aperture (Nikon), wavelength 940 nm, laser power: 56 mW, acquisition rate: 155 Hz; filter set: 525/50 nm (for signal collection) and a 565 nm dichroic mirror).
The responses of individual variants were analyzed as described 6, 8. The Ilastik toolkit 50 was used to segment cell bodies in the reference images. Wells with fewer than five detected neurons, and wells with poor neuronal proliferation, were discarded. Plates with more than four discarded control (GCaMP6s) wells were discarded and re-screened. The ΔF/F0, SNR, and kinetics (half-rise, half-decay, time-to-peak) metrics were computed for each cell. Median values from each well are reported to quantify performance. Each observation was normalized to the median GCaMP6s value from the same experimental batch. Baseline brightness for constructs co-expressing mCherry (in a Binder notebook and Supp. Table 2) was calculated by dividing the GFP cellular fluorescence in the beginning of the 3-AP stimulation epoch by the mCherry cellular fluorescence (for a ratiometric measurement). For comparison with XCaMP variants (in Supplementary Figure 4), no mCherry normalization was performed, but all baseline brightness values were still normalized to GCaMP6s in the same transfection week. To determine significant differences in observations between constructs, a two-tailed Mann-Whitney U-test was performed between constructs and controls (GCaMP6s or jGCaMP7f). A median ΔF/F0 trace was computed across all detected cell bodies in a well for each stimulus. Photobleaching was corrected in the 1-AP recordings by fitting a double exponential to the beginning and end segments of the fluorescence trace.
Finally, variants were filtered according to four criteria to remove noisy, non-responding clones. 1) Variants with half-rise time > 4x slower than GCaMP6s, as these represented poor fits or noise. 2) Similarly, variants with time-to-peak >3x that of 6s. 3) Variants with half-rise time < 0.1x that of 6s, as these represented sensors with poor ΔF/F0. 4) Similarly, variants with half-decay time <0.01x that of 6s. This 683 (out of 776 tested) jGCaMP8 variants, along with the controls, that showed detectable response to 1 AP in cultured neurons (Supp. Fig. 2). All the parameters measured in our screen can be examined as an interactive scatterplot in a Binder notebook.
Baseline fluorescence measurements
In a separate round of measurements from those measuring ΔF/F0, SNR, and kinetics, the baseline fluorescence of jGCaMP8 series was compared to jGCaMP7f and the XCaMP series. Due to significant week-to-week variability in baseline fluorescence, all constructs for this experiment were transfected side-by-side (2 consecutive transfection weeks, five 96-well plates). To minimize possible plate-to-plate variability within each transfected batch, the baseline fluorescence of each construct was normalized to in-plate GCaMP6s.
Fluorescence recovery after photobleaching
FRAP experiments were carried out on a Nikon Ti-E inverted microscope outfitted with a Yokogowa CSU-X1 spinning disk and an Andor DU-897 EMCCD camera. Fluorescence excitation was carried out using a solid-state laser line at 488nm, and emission was collected with 100x 1.49NA objective (Nikon Instruments) through a standard GFP filter set. Photobleaching was performed using a Bruker Mini-Scanner by focusing a 405 nm laser to a single, diffraction-limited spot for 100 ms. Cultured neurons plated in 35 mm glass-bottom dishes (MatTek) were immersed in regular imaging buffer with the addition of synaptic blockers (same as used for neuronal culture field stimulation) and 1 µM TTX to block AP generation. In a subset of experiments, the buffer was supplemented with 5 µM ionomycin. Bleaching spots were chosen to be on the soma of the neuron but distant from the nucleus. A spot was photobleached 10 times (0.1 Hz) as the cell was concurrently imaged at 25 or 50 frames per second.
For analysis, pixels within a 1.5 µm radius around the bleach spot were averaged in each frame. The resulting fluorescence trace was normalized to the mean fluorescence of an identically sized spot on the opposite side of the soma, outside the nucleus. The trace was then split into 10 epochs (each corresponding to a bleaching event) and the fluorescence fi(t) of each epoch was normalized by dividing by the fluorescence value immediately preceding the bleaching pulse (fi(tpre)) as follows: The resistant fraction was calculated as follows: where is the final fluorescence value at the end of epoch i, and the final term in the equation is the averaged fluorescence loss of all epochs after the first. This term is subtracted to account for the overall fluorescence loss with each bleaching pulse.
Crystal structure determination
All GCaMP samples for crystallization were kept in 20 mM Tris, 150 mM NaCl, pH 8.0, 2 mM CaCl2. All crystallization trials were carried out at 22°C with the hanging-drop vapor diffusion method. Commercial sparse-matrix screening solutions (Hampton Research) were used in initial screens. 1 µL of protein solution was mixed with 1 µL of reservoir solution and equilibrated against 300 µL of reservoir solution. Diffraction data were collected at beamline 8.2.1 at the Berkeley Center for Structural Biology and processed with XDS 51. The phase was determined by molecular replacement using MOLREP , and the structure of GCaMP2 (PDB 3EK4) 52, 53 without the RS20 peptide as the starting model. Refinement was performed using REFMAC 54, followed by manual remodeling with Coot 55. Details of the crystallographic analysis and statistics are presented in Supp. Table 3.
Adult Drosophila L2 Assay
GECIs were tested by crossing males carrying the variant to a w+ ; 53G02-Gal4AD (in attP40); 29G11-Gal4DBD (in attP2) females 56. Heterozygous flies were used in our experiments. Flies were raised at 21°C on standard cornmeal molasses media.
Females 3-5 days after eclosure were anesthetized on ice. After transferring to a thermoelectric plate (4°C), legs were removed, and then facing down, the head was glued into a custom-made pyramid using UV-cured glue. The proboscis was pressed in and fixed using UV- cured glue. After adding saline (103 mM NaCl, 3 mM KCl, 1 mM NaH2PO4, 5 mM TES, 26 mM NaHCO3, 4 mM MgCl2, 2.5 mM CaCl2, 10 mM trehalose and 10 mM glucose, pH 7.4, 270–275 mOsm) to the posterior side of the head, the cuticle was cut away above the right side, creating a window above the target neurons. Tracheae and fat were removed, and muscles M1 and M6 were cut to minimize head movement.
Two-photon imaging took place under a 40x 0.8NA water-immersion objective (Olympus) on a laser-scanning microscope (BrukerNano, Middleton, WI) with GaAsP photomultiplier tubes (PMTs). Laser power at 920 nm was kept constant at 8 mW using a Pockels cell. No bleaching was evident at this laser intensity. The emission dichroic was 580 nm and emission filters 511/20-25 nm. Images were 32×128 pixels with a frame rate at 372 Hz.
A MATLAB script drove the visual stimulation via a digital micromirror device (DMD, LightCrafter) at 0.125 Hz onto a screen covering the visual field in front of the right eye. A blue LED (Thorlabs M470L3) emitting through a 474/23-25 nm bandpass filter (to keep blue light from contaminating the green imaging channel) provided illumination.
Light dimming produced a stereotypical calcium increase in L2 neurons28–30. Intensity measurements were taken in medulla layer 2 (Fig. 2A; modified from 57). A target region image was chosen by testing each focal layer with 0.5 Hz full-field visual stimulation until a layer with maximum ΔF/F0 was identified. Then 2-3 columns producing a maximum response were identified within this layer. In addition to the ROI containing these L2 columns, a background ROI was selected where no fluorescence was evident. The mean background intensity was subtracted from the mean L2 ROI. Imaging then targeted this region over a protocol involving multiple tests, as such:
Image analysis was performed using custom Python scripts. In the ΔF/F0 calculation, baseline F0 included the last 20% of images taken at the end of the light period. Stimulus onset is the light- to-dark transition. Change in fluorescence ΔF is the intensity minus baseline. ΔF/F0 is ΔF divided by baseline. The final signal is processed through a Gaussian filter (σ=3).
Imaging in the Drosophila larval neuromuscular junction
We made 20XUAS-IVS-Syn21-op1-GECI-p10 in VK00005 transgenic flies 58 and crossed them with 10XUAS-IVS-myr::tdTomato in su(Hw)attP8 x R57C10-Gal4 at VK00020; R57C10-Gal4 at VK00040 double-insertion pan-neuronal driver line. Heterozygous flies were used in our experiments. Sensor cDNAs were codon-optimized for Drosophila. The NMJ assay is as in our previous study1. Briefly, female 3rd instar larvae were dissected in chilled (4°C) Schneider’s Insect Medium (Sigma) to fully expose the body wall muscles. Segment nerves were severed in proximity to the ventral nerve cord (VNC). Dissection medium was then replaced with room temperature HL-6 saline in which 2 mM CaCl2 and 7 mM of L-glutamate were added to induce tetany – freezing the muscles in place. A mercury lamp (X-CITE exacte) light source was used for excitation, and out-of-objective power was kept less than 5 mW to reduce bleaching. Type Ib boutons on muscle 13 from segment A3-A5 were imaged while the corresponding hemi-segment nerve was stimulated with square voltage pulses (4 V, 0.3 ms pulse width, 2 s duration, 1-160 Hz frequency) through a suction electrode driven by a customized stimulator. Bath temperature and pH were continuously monitored with a thermometer and pH meter, respectively, and recorded throughout the experiment. The filters for imaging were as follows: excitation: 472/30 nm; dichroic: 495 nm; emission: 520/35 nm. Images were captured with an EMCCD (Andor iXon 897) at 128.5 frames per second and acquired with Metamorph software. ROIs around boutons were manually drawn, and data were analyzed with a custom Python script.
NMJ immunofluorescence
Variants were crossed to a pan-neuronal driver line, also containing tdTomato, (pJFRC22-10XUAS-IVS-myr::tdTomato in su(Hw)attP8 ;; R57C10 at VK00020, R57C10 at VK00040). 3rd instar larvae were filleted and fixed following standard techniques 59. Primary chicken anti-GFP (Thermo Fisher A10262, 1:1000) and secondary goat anti-chicken AlexaFluor 488 plus (Thermo Fisher A32931, 1:800) were used to stain GECIs. Primary rabbit anti-RFP (Clontech 632496, 1:1000) and secondary goat anti-rabbit Cy3 (Jackson 111-165-144, 1:1000) labeled tdTomato.
MBON-γ2α’1 immunofluorescence
Variants were co-expressed with membrane-localized myr::tdTomato using the MB077B driver. Adults 3-6 days old were harvested, brains dissected, and fixed using standard techniques. GCaMP variants were directly labeled with anti-GFP (AlexaFluor 488, Molecular Probes A-21311, 1:500). Primary Rat anti-RFP (mAb 5F8 Chromotek, 1:500) and secondary goat anti-rat Cy3 (Jackson 112-165-167, 1:1000) labeled tdTomato.
Immunofluorescence quantification
ROIs were drawn on targeted regions using custom Python scripts. Within each ROI, otsu-thresholding was used to identify regions expressing myr::tdTomato. Intensity measurements were then taken for both the variant and tdTomato within these regions. The ratio is the intensity from the green channel (variant staining) divided by the intensity from the red channel (myr::tdTomato staining).
Western blot
Protein was extracted from female brains with the same genotype used in the NMJ immunostaining. Western blots were performed following standard techniques. Each variant was stained using primary rabbit anti-GFP (Millipore Sigma) and secondary goat anti-rabbit IgG conjugated to horseradish peroxidase (HRP; Thermo Fisher). Actin was stained using mouse IgM anti-α-actin (Thermo Fisher, 1:5000) and goat anti-mouse IgG and IgM-HRP (Thermo Fisher, 1:5000). Signal was formed using SuperSignal West Dura luminescence and was imaged on a BioRad Gel imager. Band intensity was measured using FIJI. Band intensity from the variant was divided by band intensity from the actin band to determine the ratio.
Light sheet imaging of larval zebrafish
Presentation of flashes during fictive swimming
Zebrafish larvae of 6-8 dpf expressing 8f, 7f and 6f as nuclearly-localized fusions to histone H2B under the pan-neuronal elavl3 promoter were reared in conditions described in 60 and prepared for fictive behavior recording and light sheet imaging according to previously detailed methods 32. A LED torch, placed diagonally from the fish, emitted short pulses of light (100 ms) at 20- second intervals (Supp. Fig. 11A). The fish experienced no structured visual stimulus besides the constant 488 nm light sheet at all other times. To image the optic tectum, where visual responses were observed, we captured five horizontal planes at an interval of 20 µm spanning its dorsoventral axis in a single volume. We imaged at 30 Hz, a rate sufficient to capture the fast kinetics of the calcium sensors.
Identification of flash-responsive neurons
Fluorescence images were motion-corrected and segmented into cell segments by a custom preprocessing pipeline (https://github.com/zqwei/fish_processing).
To find flash-responsive neurons, we first computed a trial- and neuron-averaged response triggered on the onset of the flash stimulus for each fish for trials where the fish did not swim around the presentation of the stimulus. We then fit each of those responses to a difference between two exponentials to derive two time constants We then used the result of those fits to generate a calcium kernel with which regressors for the visual stimulus and motor output were generated. We found visual-responsive cells by looking for cells with a high visual coefficient (>75th percentile) and a low motor coefficient (<25th percentile). For each fish, we then computed another trial-averaged response triggered on the onset of the flash stimulus, this time only for visual-responsive cells and for all trials. These traces were averaged across fish to generate a variant-mean.
The variant-means were again fit to the function above to generate a calcium kernel for each variant. Finally, we once again performed regression with these calcium kernels to derive our final set of visual-responsive cells for each fish. Further analyses, such as the re-computation of the variant-mean (Supp. Fig. 11C), as well as single-cell characterization of response amplitude (Supp. Fig. 11D), half-rise time (Supp. Fig. 11E) and half-decay time (Suppl. Fig. 11F) were performed on this set of cells.
Cortex: mouse surgeries
Young adult (postnatal day 50-214) male C57BL/6J (Jackson Labs) mice were anesthetized using isoflurane (2.5% for induction, 1.5% during surgery). A circular craniotomy (3 mm diameter) was made above V1 (centered 2.5 mm left and 0.5 mm anterior to the Lambda suture). Viral suspension (30 nL) was injected in 4-5 locations on a 500 µm grid, 300-400 µm deep. Constructs included: AAV2/1-hSynapsin-1-jGCaMP8 constructs (pGP-AAV-syn1-jGCaMP8f-WPRE, Addgene plasmid #162376, 4e12 GC/mL titer; pGP-AAV-syn1-jGCaMP8m-WPRE, Addgene plasmid #162375, 2.2e12 GC/mL titer; pGP-AAV-syn1-jGCaMP8s-WPRE, Addgene plasmid #162374, 2.1e12 GC/mL titer). A 3 mm diameter circular coverslip glued to a donut-shaped 3.5 mm diameter coverslip (no. 1 thickness, Warner Instruments) was cemented to the craniotomy using black dental cement (Contemporary Ortho-Jet). A custom titanium head post was cemented to the skull. An additional surgery was performed for loose-seal recordings. 18-80 days after the virus injection, the mouse was anesthetized with a mixture of ketamine-xylazine (0.1 mg ketamine & 0.008 mg xylazine per gram body weight), and we surgically removed the cranial window and performed durotomy 61. The craniotomy was filled with 10-15 µL of 1.5% agarose, then a D-shaped coverslip was secured on top to suppress brain motion and leave access to the brain on the lateral side of the craniotomy.
Cortex: two-photon population imaging
Mice were kept on a warm blanket (37°C) and anesthetized using 0.5% isoflurane and sedated with chlorprothixene (20–30 µL at 0.33 mg/mL, intramuscular). Imaging was performed with a custom-built two-photon microscope with a resonant scanner. The light source was an Insight femtosecond-pulse laser (Spectra-Physics) running at 940 nm. The objective was a 16× water immersion lens with 0.8 numerical aperture (Nikon). The detection path consisted of a custom filter set (525/50 nm (functional channel), 600/60 nm (cell targeting channel) and a 565 nm dichroic mirror) ending in a pair of GaAsP photomultiplier tubes (Hamamatsu). Images were acquired using ScanImage (vidriotechnologies.com)62. Functional images (512 × 512 pixels, 215 × 215 µm2; or 512 × 128 pixels, 215 × 55 µm2) of L2/3 cells (50–250 µm under the pia mater) were collected at 30 Hz or 122 Hz. Laser power was up to 50 mW at the front aperture of the objective unless stated otherwise for the XCaMP-Gf experiments.
Cortex: loose-seal recordings
Micropipettes (3–9 MΩ) were filled with sterile saline containing 20 µM AlexaFluor 594. Somatic cell attached recordings were obtained from upper layer 2 neurons (50-200 µm depth from brain surface) visualized with the shadow patching technique 63. Spikes were recorded either in current clamp or voltage clamp mode. Signals were filtered at 20 kHz (Multiclamp 700B, Axon Instruments) and digitized at 50 kHz using Wavesurfer (wavesurfer.janelia.org/). The frame trigger pulses of ScanImage were also recorded and used offline to synchronize individual frames to electrophysiological recordings. After establishment of a low-resistance seal (15–50 MOhm), randomized visual stimulation was delivered to increase the activity of the cells in the field of view. In a small subset of recordings, we microstimulated the recorded neuron in voltage-clamp recording mode by applying DC current to increase its firing probability 64.
Cortex: visual stimulation
Visual stimuli were moving gratings generated using the Psychophysics Toolbox in MATLAB (Mathworks), presented using an LCD monitor (30 × 40 cm2), placed 25 cm in front of the center of the right eye of the mouse. Each stimulus trial consisted of a 2 s blank period (uniform gray display at mean luminance) followed by a 2 s drifting sinusoidal grating (0.05 cycles per degree, 1 Hz temporal frequency, eight randomized different directions). The stimuli were synchronized to individual image frames using frame-start pulses provided by ScanImage.
Cortex: post hoc anatomy
After the loose-seal recording sessions, mice were anesthetized with a mixture of ketamine-xylazine (0.1 mg ketamine & 0.008 mg xylazine per gram body weight) and were trans-cardially perfused with 4% PFA in 1X DPBS. The brains were extracted and post-fixed overnight in the perfusing solution. The brains were sectioned at 50 µm thickness, blocked with 2% BSA+ 0.4 Triton-100 (in PBS) for 1 h at room temperature, incubated with primary antibody (Rb-anti-GFP, 1:500, Invitrogen, #G10362) for 2 days at 4℃, and secondary antibody (AlexaFluor 594 conjugated goat anti-Rb, 1:500, Invitrogen, #A-11012) overnight at 4℃. The sections were mounted on microscope slides in Vectashield hard-set antifade mounting medium with DAPI (H-1500, Vector). Samples were imaged using a TissueFAXS 200 slide scanner (TissueGnostics, Vienna, Austria) comprising an X-Light V2 spinning disk confocal imaging system (CrestOptics, Rome, Italy) built on an Axio Imager.Z2 microscope (Carl Zeiss Microscopy, White Plains, NY) equipped with a Plan-Apochromat 20x/0.8 M27 objective lens.
Cortex: analysis of in vivo two-photon imaging
The acquired data was analyzed using MATLAB (population imaging) or Python (imaging during loose-seal recordings). In the MATLAB pipeline, for every recorded FOV, we selected ROIs covering all identifiable cell bodies using a semi-automated algorithm, and the fluorescence time course was measured by averaging all pixels within individual ROIs, after correction for neuropil contamination (r = 0.7), as described in detail in 8. We used one-way ANOVA tests (P< 0.01) for identifying cells with significant increase in their fluorescence signal during the stimulus presentation (responsive cells). We calculated ΔF/F0 = (F − F0)/F0, where F is the instantaneous fluorescence signal and F0 is the average fluorescence in the interval 0.7 s before the start of the visual stimulus. For each responsive cell we defined the preferred stimulus as the stimulus that evoked the maximal ΔF/F0 amplitude (averaging the top 25% of ΔF/F0 values during the 2 s of stimulus presentation). The half-decay time was calculated as follows: for each responsive cell we averaged its ΔF/F0 response to the preferred stimulus over five trials. We also calculated the standard deviation of the averaged baseline signal during 0.7 s before the start of the stimulus. Only cells where maximal ΔF/F0 amplitude was higher than four standard deviations above the baseline signal were included in the analysis. The time required for each trace to reach half of its peak value (baseline fluorescence subtracted) was calculated by linear interpolation. The fraction of cells detected as responsive was calculated as the number of significantly responsive cells over all the cells analyzed. The cumulative distribution of peak ΔF/ F0 responses included the maximal response amplitude from all analyzed cells, calculated as described above for each cell’s preferred stimulus. The orientation sensitivity index (OSI) was calculated as before 6, 8, by fitting the fluorescence response from individual cells to the eight drifting grating stimuli with two Gaussians, centered at the preferred response angle (Rpref ) and the opposite angle (Ropp). The OSI was calculated as , where Rorth is the orthogonal angle to the preferred angle.
The movies recorded during loose-seal recordings were motion-corrected and segmented with the python implementation of Suite2p (github.com/MouseLand/suite2p) 65. The ROI corresponding to the loose-seal recorded cell was then manually selected from the automatically segmented ROIs. For this dataset, we could calculate the neuropil contamination for most of the movies and got a distribution with a median of r_neu ∼ 0.8 (Supp. Fig. 18), so we used this value uniformly for neuropil correction. Calcium events were defined by grouping action potentials with a 20 ms inclusion window. Then we calculated ΔF/F0 = (F − F0)/F0, where F is the instantaneous fluorescence signal and F0 was defined separately for all calcium events as the mean fluorescence value of the last 200 ms before the first action potential in the group. We also calculated a global ΔF/F0 trace (ΔF/F0)global, where we used the 20th percentile of the fluorescence trace in a 60 s long running window as the F0,global. In the analyses we only included calcium events where this (ΔF/F0)global value was less than 0.5 right before the action potential, to include only events starting near baseline fluorescence values, in order to exclude non-linear summation and saturation. Traces in Fig. 4 were filtered with a Gaussian kernel (σ = 5 ms).
Cerebellum: mouse surgeries
Young adult (postnatal day 42-98) male C57BL/6J (Jackson Labs) mice were anesthetized using isoflurane (2.5% for induction, 1.5% during surgery). A circular craniotomy (3 mm diameter) above medial crus I (2 mm left and 1 mm posterior to the midline junction of the interparietal and occipital bones). Viral suspension (200 nL) was injected in 2 locations near the center point at a depth of 300-400 µm. Constructs injected included: AAV2/1-CAG-FLEx-jGCaMP8 constructs (pGP-AAV-CAG-FLEx-jGCaMP8f-WPRE, Addgene plasmid #162382; pGP-AAV-CAG-FLEx-jGCaMP8m-WPRE, Addgene plasmid #162381; pGP-AAV-CAG-FLEx-jGCaMP8s-WPRE, Addgene plasmid #162380; pGP-AAV-CAG-FLEx-jGCaMP7f-WPRE, Addgene plasmid #104496; and pGP-AAV-CAG-FLEx-jGCaMP6f-WPRE, Addgene plasmid #100835; all viruses diluted to 4e12 GC/mL titer).
Purkinje cell-specific expression was induced by co-injection of virus expressing Cre under control of a promoter fragment from the Purkinje cell protein 2 (Pcp2; a.k.a. L7) gene (AAV2/1- sL7-Cre, 5.3e10 GC/mL titer) 66. A 3 mm diameter circular coverslip glued to a donut-shaped 3.5 mm diameter coverslip (no. 1 thickness, Warner Instruments) was cemented to the craniotomy using dental cement (C&B Metabond, Parkell Inc.). A custom titanium head post was cemented to the skull.
Cerebellum: two-photon imaging
Head-restrained mice were allowed to freely locomote on a wheel. Imaging was performed with a custom-built two-photon microscope with a resonant scanner. The light source was a Mai Tai sapphire laser (Spectra Physics) running at 920 nm. The objective was a 16× CFI LWD Plan fluorite objective water immersion lens with 0.8 numerical aperture (Nikon). The detection path consisted of a bandpass filter (525/50 nm) and a 565 nm dichroic mirror directed towards a photomultiplier tube ( Hamamatsu). Images were acquired using Scan Image (vidriotechnologies.com) 62. Functional images (512 x 32 pixels, 215 x 27 µm2) of Purkinje cell dendrites (50–250 µm below the pia mater) were collected at 283 Hz. Laser power was up to 50 mW at the front aperture of the objective.
Cerebellum: analysis of in vivo two-photon imaging
Purkinje cell dendrite movies were captured during free locomotion without applied stimulation. Movies were motion-corrected and converted to ΔF/F0 traces using the Python implementation of CaImAn 67. Individual events within the traces were identified by finding adjacent local maxima in ΔF/F0 variance that had one local maximum in the ΔF/F0 trace between them. Statistics for individual events were calculated by fitting the equation: where t0 is the start of the peak, τrise is the rise time constant shaped by α; τ1 and τ2 are the decay time constants, Fmax is the maximum amplitude of the trace above the starting point Fstart, and F1 and F2 are component amplitudes.
Spike-to-fluorescence model
We developed a phenomenological model that converts spike times to a synthetic fluorescence time series 37. This ‘spike-to-fluorescence’ (S2F) model consists of two steps. First, spikes at times {tk} are converted to a latent variable, c(t), by convolution with two double-exponential kernels: τr and τd1, τd2 are the rise time and decay times, respectively. In our model, we required τd1< τd2 (i.e., fast and slow components), with r representing the ratio of the weight for the fast component to that of the slow one. ni(t)∼N(0,σ 2) is Gaussian-distributed ‘internal’ noise. c(t) was truncated at zero if noise drove it to negative values. We tested the performances of models with various choices of rise and decay times: (1) one rise and one decay time, (2) one rise and two decay times, and (3) two rise and two decay times. Using cross-validation, we found that one rise and two decay times models fit pyramidal cells (as described above), whereas interneurons were fit well by one rise and a single decay time (as described above with r = 0). Subsequently, c(t) was converted to a synthetic fluorescence signal through a sigmoidal function: k is a non-linearity sharpness parameter, c1/2 is a half-activation parameter, and Fm is the maximum possible fluorescence change. ne(t)∼N(0,σ 2) is Gaussian-distributed external noise. For comparison, we also generated a S2F linear model with ΔF/FSynth(t) = Fmaxc(t) + F0, where Fmax is a scaling parameter (we kept the naming as max to clarify the relationship to other models); F0 is the baseline.
Variance explained
Variance explained measures the goodness-of-fit of an S2F model, as . Here we used only the period t with spike rate >0 Hz after spikes in the calculation, where the instantaneous spike rate at time t is estimated by a boxcar-rolling average over a 600 ms time window.
Statistics
Exact statistical tests used for each comparison, as well as n, are listed in the main text and figure legends. Box-whisker plots throughout the manuscript indicate the median and 25th–75th percentile range; whiskers indicate the shorter of 1.5 times the 25th–75th range or the extreme data point. For Fig. 3E, full statistics are: jGCaMP7f vs XCaMP-Gf: P = 1.0; jGCaMP7f vs jGCaMP8f: P=0.013; jGCaMP7f vs jGCaMP8m: P = 0.029; jGCaMP7f vs jGCaMP8m: P = 0.010; jGCaMP7f vs jGCaMP8s: P = 0.010; XCaMP-Gf vs jGCaMP8f: P = 1.0; XCaMP-Gf vs jGCaMP8m: P = 1.0; XCaMP-Gf vs jGCaMP8s: P = 0.0027; jGCaMP8f vs jGCaMP8m: P = 1.0; jGCaMP8f vs jGCaMP8s: P < 0.001; jGCaMP8m vs jGCaMP8s: P < 0.001. *P < 0.05; ***P < 0.001; ns, not significant. For Fig. 3F, data passed Shapiro-Wilk normality test (α = 0.05 level). For Fig. 3F, full statistics are: jGCaMP7f vs jGCaMP8f: P = 0.83; jGCaMP7f vs jGCaMP8m; P = 0.0184; jGCaMP7f vs jGCaMP8s: P < 0.001; jGCaMP8f vs jGCaMP8m: P < 0.001; jGCaMP8m vs jGCaMP8s: P = 0.23. For Fig. 3G, full statistics are: jGCaMP7f vs jGCaMP8f: P < 0.001; jGCaMP7f vs jGCaMP8m; P = 1.0; jGCaMP7f vs jGCaMP8s: P < 0.001; jGCaMP8f vs jGCaMP8m: P < 0.001; jGCaMP8m vs jGCaMP8s: P < 0.001 (Kruskal-Wallis test with Dunn’s multiple comparison test was used to compare the magnitude of response across groups).
In situ stability of the jGCaMP8 sensors
The jGCaMP8 indicators were about half as bright as 7f in Drosophila adult and larval NMJ (and had lower levels of protein expression), whereas in mouse cortex they had similar brightness and expression levels. The mechanisms underlying the decreased expression levels in Drosophila are unclear but may involve species-specific variation in the “N-end rule” 68 and ubiquitin exoligation at the N-terminus, which was shortened by truncation of the RSET affinity tag in the jGCaMP8 variants.
Reagent distribution and data availability
DNA constructs and AAV particles of jGCaMP8s, jGCaMP8m, and jGCaMP8f (pCMV, pAAV- synapsin-1, pAAV-synapsin-1-FLEX, and pAAV-CAG-FLEX) have been deposited at Addgene (#162371-162382). Sequences have been deposited in GenBank (#OK646318-OK646320). Drosophila stocks were deposited at the Bloomington Drosophila Stock Center (http://flystocks.bio.indiana.edu); Drosophila UAS and lexAOp plasmids are at Addgene (#162383-#162388). Fish lines are available on request. Most datasets generated for characterizing the new sensors are included in the published article (and its Supplementary Information files). Additional datasets are available from the corresponding authors on reasonable request.
Author contributions
Designed project: YZ, ERS, LLL. Led the project: YZ, KS, IK, JPH, LLL. Optimized peptides for grafting: YZ, ERS. Mutagenesis: YZ, GT, JPH. Fly: DB, JZ, GCT. Fish: JXL, SN, MBA, ZW. Cerebellum: GJB, SS-HW. Photophysics: YZ, RP, IK. Crystallography: YZ. Purified protein experiments: YZ. Cortex: MR, YL, KS. Simultaneous imaging & electrophysiology: MR. Spike to fluorescence: ZW, MR, KS. Cultured neuron screen design: IK, JPH, KS. Cultured neuron experiments: YL, DR, AT, CJO, RZ, JPH, IK. Histology: DR, AT, IK. Analysis methodologies: YZ, MR, YL, KS, IK, DB, JZ, ZW. Coordination of shipments & deposits: YZ, JPH. Wrote paper: YZ, MR, YL, KS, DB, LLL, with contributions from all the authors.
Competing interests
YZ, ERS, JPH, IK, and LLL are inventors of US Patent Application 63082222, “Genetically Encoded Calcium Indicators and Methods of Use,” which covers the jGCaMP8 sensors.
Supplementary Tables
Supp. Table 1. Biophysical properties of initial sensors with different calmodulin-binding peptides used in this study.
Supp. Table 2. Sensor variants run through cultured neuron assay. ΔF/F0, half-rise time, time-to-peak, half-decay time, signal-to-noise ratio, and normalized F0 given for each of 1, 3, 10, 160 AP. P-values computed by the Mann-Whitney U test for each variant, compared to 6s, given. Nucleic acid and protein sequence of each variant, as well as detailed name (if any) shown.
Supp. Table 3. Crystal structure determination of jGCaMP8.410.80.
Supp. Table 4. Comparison of sensitivity and kinetics of jGCaMP8 to XCaMP-G, -Gf, and -Gf0 sensors. Colors in each cell indicate whether the value was significantly higher for jGCaMP8 (yellow), XCaMP (blue), or not statistically different (no color), as evaluated with Dunn’s multiple comparisons test (P-values in cells).
Supp. Table 5. Data on recorded cells, including GECI, mouse, neuron # per each mouse, putative cell type (pyramidal, interneuron), recording length, recording mode, total # of recorded spikes, and median spike SNR.
Supp. Table 6. Biophysical characterization in purified protein. Apparent Kd, Hill coefficient (cooperativity), saturating ΔF/F0, off and on-rates (8f is modeled as having a 2-component decay curve), pKa in both Ca2+-free and Ca2+-saturated states, excitation, emission, and absorption wavelength maxima, and extinction coefficient of Ca2+-free and Ca2+-saturated states. Values are n = 3, mean ± std. err.
Supp. Table 7. Statistics of the degree of nonlinearity of sensors measured by the difference of variance explained by S2F sigmoid from linear model (mean ± std.dev.).
Supp. Table 8. Statistics of S2F parameter fits (mean ± std.dev.).
Supplementary Figures
Acknowledgements
This work is part of the GENIE Project at the Howard Hughes Medical Institute, Janelia Research Campus. The GENIE Project is led by JPH, WLK, and IK; the Steering Committee includes GCT, KS, ERS, and LLL. This work was supported by the Howard Hughes Medical Institute, and NIH U19 NS104648 and NIH R01 NS045193 to SS-HW and NIH F32 MH120887 to GJB. We would like to thank the Berkeley Center for Structural Biology for use of beamline 8.2.1. We are grateful to the Viral Tools, Cell and Tissue Culture, Molecular Biology, Media Prep, Vivarium/Aquarium, Anatomy and Histology, and Fly Facility Shared Resources at Janelia for technical assistance with numerous parts of the project.
Footnotes
We forgot to upload the eight Supplementary Tables with the first submission. They are now here.