Calcium Green FlAsH as a genetically targeted small-molecule calcium indicator

Abstract

Intracellular Ca(2+) regulates numerous proteins and cellular functions and can vary substantially over submicron and submillisecond scales, so precisely localized fast detection is desirable. We have created a approximately 1-kDa biarsenical Ca(2+) indicator, called Calcium Green FlAsH (CaGF, 1), to probe [Ca(2+)] surrounding genetically targeted proteins. CaGF attached to a tetracysteine motif becomes ten-fold more fluorescent upon binding Ca(2+), with a K(d) of approximately 100 microM, <1-ms kinetics and good Mg(2+) rejection. In HeLa cells expressing tetracysteine-tagged connexin 43, CaGF labels gap junctions and reports Ca(2+) waves after injury. Total internal reflection microscopy of tetracysteine-tagged, CaGF-labeled alpha(1C) L-type calcium channels shows fast-rising depolarization-evoked Ca(2+) transients, whose lateral nonuniformity suggests that the probability of channel opening varies greatly over micron dimensions. With moderate Ca(2+) buffering, these transients decay surprisingly slowly, probably because most of the CaGF signal comes from closed channels feeling Ca(2+) from a tiny minority of clustered open channels. With high Ca(2+) buffering, CaGF signals decay as rapidly as the calcium currents, as expected for submicron Ca(2+) domains immediately surrounding active channels. Thus CaGF can report highly localized, rapid [Ca(2+)] dynamics.

Figure 1

In vitro and intracellular titration of CaGF fluorescence. (a) Structure of CaGF bound to a tetracysteine-containing peptide or protein. (b) In vitro calibration of CaGF; fluorescence increased ~ten-fold (concentration is indicated in red above the traces). (c) A large gap junction between two HeLa cells expressing recombinant connexin 43 tagged with a tetracysteine in its C terminus and stained with CaGF. (d–g) Intra-cellular calibration of CaGF bound to cx43-GFP-C4. A field of cells with an en-face and a perpendicular gap junction (red and blue arrows, respectively) bathed in zero calcium plus 1 mM EGTA (d); the same field in 15 mM calcium (e); raw data of CaGF titration (f). The calcium concentrations are indicated by numbers: 1 = 15 mM; 2 = zero calcium plus 1 mM EGTA; 3 = 25 μM; 4 = 100 μM; 5 = 400 μM and 6 = nominal zero estimated at 3 μM. (g) Dose-response curves constructed from the data shown in f (red, K_d = 50 μM; blue, K_d = 61 μM).

Figure 2

CaGF reports calcium dynamics of L-type calcium channel activation. (a) A schematic representation of an α_1C channel tagged with two tetracysteine peptides (2TC-α_1C). (b,c) A typical bright field image of HEK293 cells (b) and a TIRF image of the same field following CaGF staining, illustrating seven numbered cells (c). Cells 1, 2 and 3 were successfully transfected with the 2TC-α_1C construct. Five ROIs are indicated on cell 1 (small magenta ROI, blue, light blue, yellow and black), and one gray ROI is shown on cell 4. (d) The membrane potential of cell 1 was clamped at −80 mV and subsequently stepped to +20 mV for 1.4 s. External Ca²⁺ was set at 15 mM and FPL 64176 was added. (e) Raw data for CaGF fluorescence is heterogeneous with micron-scale hot spots (magenta and blue) at one end of the cell. (f) A plot of F/F₀ illustrates that in this cell the signal rise is fastest and largest at the hot spots.

Figure 3

Ca²⁺ transients in response to three 20-ms depolarization pulses. (a) A schematic representation of an α_1C channel tagged with GFP and the optimized tetracysteine peptide N4 (N4-GFP-α_1C). (b) A TIRF image showing the locations of the color-coded and numbered ROIs. The black ROI is the entire cell (not shown). (c) Current trace with three short depolarization pulses. Time is measured from the beginning of the first pulse. 1 mM BAPTA was present in the pipette; the external solution contains 15 mM Ca²⁺ and 10 μM FPL 64176. (d) Three fluorescent traces aligned in time with the current trace. The imaging rate was 100 f.p.s. (e) Two additional fluorescent traces registered at ROIs placed on hot spots. (f) A second current applied to the same cell, 2 min after the trace shown in c, using the protocol of Figure 2d. (g) Fluorescent signals at the ROI shown in c in response to the 960-ms depolarization. (h) Comparison of the responses of the red ROI to the two protocols, depicted as F/F₀ (the color of the trace from d was switched to gray).

Figure 4

CaGF shows rapid kinetics. The upper panel depicts the current elicited by the same protocol described in Figure 3c, except that time was measured from the beginning of the tail current and 40 mM EGTA was present in the patch pipette. Below is the corresponding fluorescent signal, where F/F_fit is the photobleaching-corrected signal. The fluorescent response to the tail current (the fourth pulse) is the largest and has almost the same fast time course as the current, which demonstrates the fast K_off of CaGF.

Figure 5

Simulation of CaGF signals. All panels show results after 20 ms of depolarization for 7 pS calcium channels with 100 mV driving force. Channels or clusters of channels are assumed to be uniformly distributed. (a) CaGF activated by an isolated open channel. Shown is the predicted fractional activation of CaGF as a function of lateral distance from the channel (red squares, 1 mM BAPTA; black circles, 40 mM EGTA; blue triangles, 4.5 mM EGTA; gray inverted triangles, small-molecule buffers absent) and a hypothetical dye with 100-fold higher affinity (1 mM BAPTA, magenta diamonds; no buffer, cyan circles). (b) Fraction of bound CaGF (thin lines, open symbols) and depletion of buffers (thick line, filled symbols) around a cluster of 64 channels; buffers as in a. (c) Free calcium around a cluster of 64 channels. (d) Effect of cluster size. Filled symbols, response in 1-μm hot spot. Open symbols, whole-cell response.