| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Neurobiology and Behavior, State University of New York, Stony Brook, New York 11794-5230; 2 Howard Hughes Medical Institute, State University of New York, Stony Brook, New York 11794-5230
Submitted 14 June 2003; accepted in final form 8 August 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
; Bremmer et al. 2001). A limitation of this approach is that the activation cannot be attributed to a single cell type. Other strategies involve the use of electrode arrays, but even when using electrode arrays to record activity, individual cells being monitored are not usually identified (Warren et al. 2001). An alternative approach is to use calcium signals to report neuronal activity because electrical activity of neurons leads to increases in calcium levels via calcium influx through voltage-dependent calcium channels. We have been applying this approach to monitor activity patterns of individual neurons in larval zebrafish (Fetcho and O'Malley 1995; O'Malley et al. 1996). In our past work, neurons were retrogradely labeled with calcium indicators by manually injecting the indicators into the muscle or the spinal cord. The transparency of larval zebrafish allowed imaging of neurons in the spinal cord or brain in the intact animal by using confocal microscopy. The main problem with dye injections, however, is that neural circuitry may be disrupted as a result of damage from the injection. Hence, injections are most appropriate for studying neurons with long axonal projections that can be backfilled without severe disruption of the circuits in the region of the CNS being studied. Even then, only a subset of the cells of any one type might be labeled, thus limiting the populations that can be sampled in one animal.
The limitations of invasive techniques can potentially be overcome by generating transgenic fish expressing a calcium-sensitive derivative of GFP (cameleon) in neurons (Miyawaki et al. 1997, 1999). Cameleon is a hybrid protein in which cyan fluorescent protein (CFP) and yellow fluorescent protein (YFP) are linked by calmodulin and an M13 calmodulin-binding domain. On an increase in calcium concentration, calmodulin binds calcium and interacts with M13. The resulting conformational change of the protein increases the efficiency of fluorescence resonance energy transfer from CFP to YFP. When excited by a wavelength appropriate for CFP excitation, an increase in calcium concentration causes an increase in the YFP/CFP fluorescence intensity ratio. Thus cameleon serves as a ratiometric calcium indicator.
Cameleon, originally shown to function as a fluorescent calcium sensor in cultured cells, was used in its first in vivo application to detect calcium transients in pharyngeal muscle in Caenorhabditis elegans (Kerr et al. 2000). This report also showed calcium transients in neurons after direct electrical stimulation of the cells. In the past year, several genetically encoded indicators, including cameleon, have been used to monitor neuronal activity in flies (Fiala et al. 2002; Liu et al. 2003; Reiff et al. 2002; Wang et al. 2003; Yu et al. 2003). No one has yet demonstrated whether cameleon (or any other genetically encoded indicator) can detect calcium transients in neurons that occur in association with normal behavior of vertebrates. Here, we show that cameleon can detect calcium transients in motoneurons and a type of interneuron in an intact zebrafish during an escape behavior. In addition, we report the production of a stable transgenic line of zebrafish with cameleon expression in all neurons, demonstrating that the calcium buffering associated with its expression does not preclude the production of viable vertebrate lines. Our work shows the feasibility of using genetically encoded calcium indicators to monitor activity of neurons in behaving vertebrates.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
The yellow cameleon 2.1 (YC2.1) cDNA was kindly provided by R.Y. Tsien, University of California San Diego. The YC2.1 construct was found to contain a single nucleotide deletion in the coding sequence resulting in disruption of the stop codon. We inserted synthetic oligonucleotides to restore the stop codon such that the amino acid sequence at the C-terminus of the protein was DELYNKN instead of DELYK. To express the YC2.1 cDNA in Rohon Beard (RB) cells, we subcloned the promoter/enhancer derived from the Islet-1 gene (Higashijima et al. 2000), the modified YC2.1, and the SV40 poly(A) signal (Invitrogen) into pBluescript-SK (Stratagene). For expression in motoneurons and interneurons by stochastic labeling, a similar construct was generated by substituting the Islet-1 sequences with the goldfish alpha-tubulin promoter kindly provided by D. Goldman, University of Michigan.
For generating stable transgenic lines expressing cameleon in all neurons, a slightly modified version of the zebrafish HuC promoter was used (Park et al. 2000). The DNA fragment we used contains the promoter (from –2771), the first exon, the first intron, and a short portion of the second exon. The initiation ATG codon for the endogenous HuC is located near the end of the first exon. This ATG was destroyed by linking two different PCR products. The first PCR-fragment started from –2771 and ended just before the ATG. The second PCR-fragment started just after the ATG and ended near the 5' end of the second exon. To amplify the first PCR-fragment (3.2 kb), plasmid DNA that contains the HuC promoter sequence (a gift from A. Chitnis), was used as a template. To amplify the second PCR fragment (5.6 kb), zebrafish genomic DNA was used as a template. Because the initiation codon was removed, the final 9-kb DNA fragment no longer contained a translation initiation site. This DNA fragment can be used as a promoter cassette to drive expression of other genes that have their own translation start sites. The final HuC-cameleon construct was generated by subcloning the HuC promoter, the modified YC2.1 and the SV40 poly(A) signal into pBluescript-SK.
Fish sample preparation
All experiments were performed in accordance with National Institutes of Health guidelines under protocols approved by the Stony Brook Institutional Animal Care and Use Committee.
DNA injection into single-cell zebrafish (Danio rerio) embryos was carried out as previously described (Higashijima et al. 1997). Experiments were performed on fish after 2–3 days of development at 28.5°C. By this stage, the fish show a consistent escape behavior following a sudden head tap. All experiments were done at room temperature. In transient expression experiments, DNA-injected fish were anesthetized with 0.02% 3-aminobenzioic acid ethyl ester (MS222) and screened briefly for fluorescent neurons by using a x20 Fluar 0.75 NA objective (Zeiss) on an inverted microscope equipped with standard epifluorescence. For generation of stable lines, DNA-injected fish were raised to adulthood, and their progeny were screened for the presence of fluorescence (Higashijima et al. 1997). Approximately 25 fish were screened until we found one that produced fluorescent embryos. Those embryos (HuC-cameleon) were raised to adulthood and crossed with each other to get homozygous embryos. Three-day-old homozygous embryos were used for experiments on RB neurons in the stable line.
Transient expression leads to variable labeling from fish to fish and cell to cell. Many labeled cells in a fish were too dim for imaging with a laser power that did not produce significant bleaching (see Optical recordings for details). We selected cells that were clearly visible on the CFP channel at maximal gain with laser intensities that did not produce significant bleaching. There were occasional very bright cells, which we did not study. The wide variation in transient expression is not a problem in the stable line where the intensity of the cells was in the range of those we studied in the transient expression experiments.
The fish were embedded laterally in either 1.2% agar or 1.5% low-melting-point agarose on a cover glass in a petri dish and rinsed with 10% Hank's solution to allow recovery from the anesthetic. In some preparations, fish were paralyzed by injecting 1 mg/ml of bungarotoxin (Sigma) into the ventral region of the body before embedding.
For electrically stimulating RB sensory neurons, an extracellular metal stimulating electrode was brought into contact with the skin near where the RB peripheral processes would be passing. Current was applied by a pulse stimulator (Model 2100, A-M systems, Carlsborg, WA). Because electrical stimulation was applied to the far side (topside) of the fish in relation to the objective lens, we recorded from cells located on that side.
Escapes were elicited by an abrupt touch produced by the displacement of a small glass probe attached to a piezoelectric crystal (O'Malley et al. 1996). These stimuli were applied to the head region on the far side in relation to the objective lens. Because motoneurons and circumferential descending (CiD) interneurons located on the side opposite to the tap become active during the escape (Fetcho and O'Malley 1995; Ritter et al. 2001), we imaged cells on that side.
Optical recordings
Confocal images were obtained using a Zeiss Axiovert 100M equipped with a LSM510 laser-scanning confocal imaging system. The argon 458-nm laser line was used to excite YC2.1. A Q515LP dichroic splitter (Chroma) was inserted in the emission light path to enable the simultaneous acquisition of YFP and CFP images. Emission filters of D535/30 and D485/40 (Chroma) were used for YFP and CFP, respectively. A x63 water-immersion objective (NA 0.9, Zeiss Achroplan) was used for all recordings because this objective has the long working distance needed to image deeper into the animal. Because the objective was designed for no coverslip, its use with a coverslip resulted in chromatic aberration, leading to the collection of YFP and CFP images from slightly different focal planes. Consequently, we moved the pinhole position of the YFP channel along the z axis and adjusted it to the point where the focal planes of both channels matched. Under these conditions, we found that the ratio change due to focal plane change was minimal.
Most recordings were made with 128 x 128 pixels at the fastest scan speed on the machine (
200 ms/frame). The scan area was zoomed such that 40–60 µm corresponded to the 128 pixels. In most cases, we set the focal plane near the brightest plane through a cell before each trial. We used the 15 mW laser at 75% power and then attenuated the laser light to 
5% (typically 2–5%) because greater intensities led to bleaching. The confocal aperture was set fully open to maximize light collection. We attempted to set the photomultiplier gain in both YFP and CFP channels such that mean pixel intensity of the cell area was in the middle of the 8-bit range. The CFP intensity is much dimmer than the YFP, so for dimmer cells it was not possible in some cases to use the middle of the 8-bit range even at the highest gain on the CFP channel with 5% laser intensity. Consequently, the ratios we report are not always identical from cell to cell because of differences in the gains on the two detectors used to measure the CFP and YFP. This is not a consequence of any difference in the relative amounts of CFP and YFP in the cells, which is fixed because both are part of each cameleon molecule. Amplifier offset in both channels was set such that pixel value of the background area was close to zero.
Each trial consisted of 25- to 40-s image streams (50–80 frames in the case of 2-Hz imaging). In each trial, a rapid decrease of fluorescence intensity in both YFP and CFP channels occurred in the initial frames, especially when a trial was carried out after a long recovery (darkness) time (see the initial portion of the traces in Fig. 1D). After this decay phase, the intensity of both channels became stable. Therefore electrical stimulation or tapping was applied well after images became stable, and images of this quick-bleaching phase were typically excluded from data analysis. The source of the quick bleaching is unknown. It is likely to be some property of the indicator, rather than bleaching of tissue autofluorescence, because it did not occur when we imaged areas off of the labeled neurons. The fluorescence of the imaged region recovered after a period of no illumination. This might be a consequence of diffusion of new dye into the imaged region. After functional imaging, the morphology of the cells was studied by acquiring Z stacks of confocal images using the 488-nm argon laser line, a higher laser intensity, and a reduced pinhole to obtain better images of morphological details.
|
The mean intensity of CFP and YFP fluorescence and the ratio in moving cells were analyzed by using a specialized program written in Labview (National Instruments, Austin, TX). The principle of the program is that the pixels to be analyzed in each frame are determined by identifying a fixed number of the brightest pixels in each CFP image. The number of pixels that were to be examined was determined by thresholding the cell image in the first frame of the CFP channel. By using a fixed number of the brightest CFP pixels in each frame, the cell was automatically tracked. Then the mean fluorescence intensity of the selected brightest CFP pixels and the mean intensity of the corresponding pixels in the YFP image were calculated. The YFP/CFP ratio in each frame was calculated from these mean values. Each pixel has an 8-bit integer value. As a result, there were sometimes many pixels with the same intensity. To restrict the selected pixels on each frame to the same number when there were too many pixels of equally bright intensity in a particular frame, the program randomly selected pixels from them until the fixed total number was reached. When other cells (or processes) were included in the images, the region of the image analyzed could be selected by drawing a region of interest on the image. The same selection scheme was used to make ratio images (Fig. 1B). In this case, the ratio calculation was made on a pixel-by-pixel basis and the ratio value was pseudocolored.
Electrophysiological recordings
Zebrafish larvae, 3–4 days old, were anesthetized with 0.02% Tricaine-S (Western Chemical) in an extracellular recording solution that contained (in mM) 134 NaCl, 2.9 KCl, 1.2 MgCl2, 2.1 CaCl2, 10 HEPES buffer, and 10 glucose, adjusted to pH 7.8 with NaOH (Drapeau et al. 1999; Legendre and Korn 1994). The preparations were paralyzed with 0.01 mM D-tubocurarine (Sigma) added to the recording solution, which was bubbled with ambient air and superfused continuously at 26°C.
Larvae were pinned to a silicone elastomer (Sylgard)-lined glass-bottom petri dish with short pieces (2 mm) of fine tungsten wire (0.001 in) pushed through the notochord— one pin placed near the swim bladder and another near the anus. The skin between the two pins was removed with a pair of fine forceps. Collagenase (0.1%, Sigma) in recording solution was applied to the preparation for 3–5 min to enzymatically prepare the muscle fibers for removal. The collagenase solution was washed off, and a large bore (15 µm diam) glass microelectrode attached to the patch electrode holder was used to aspirate away muscle fibers overlying a small section (2–3 segments) of the spinal cord.
Standard whole cell recording techniques were used to record the activity of RB neurons in vivo at 26°C. Patch electrodes (15 M
) were pulled on a Flaming/Brown micropipette puller (P-97, Sutter Instruments) from borosilicate glass (1.5 mm OD, 0.86 mm ID, Warner Instrument) and filled with patch solution (in mM) 125 K gluconate, 2 MgCl2, 10 HEPES buffer, 10 EGTA, and 4 Mg ATP, adjusted to pH 7.2 with KOH. We did not correct for junction potential as we were only concerned with the firing of the cell in response to stimuli. Neurons were labeled with 0.1% Sulforhodamine B (Sigma) added to the patch solution, and fluorescent images were acquired with a CCD camera (C-72-CCD, Dage MTI, Michigan City, IN), a frame grabber (LG3, Scion, Frederick, MD), and imaging software (Scion National Institutes of Health Image, Scion) for morphological identification.
For whole cell recordings, the preparation was observed on an upright microscope (BX51WI, Olympus) fitted with infrared DIC optics using a water-immersion objective (x40, 0.80 NA, Olympus). Positive pressure (50 mmHg) was applied to the patch electrode as it approached the exposed surface of the spinal cord. The tip of the electrode was carefully lowered until it broke into the cord. RB neurons were targeted for recording based on their size, shape, and position in the cord. Release of positive pressure formed a gigaohm seal once the tip of the patch electrode was directly apposed to the sensory neuron, and suction pulses were applied to break the seal for whole cell voltage recordings. Pulses of current (10–700 pA, 500-ms duration) were injected into the recorded neuron to determine its firing threshold and characteristics (single or multiple spikes).
To replicate the stimulation of RB neurons in the imaging experiments, an extracellular metal stimulating electrode was positioned on dorsal muscle one to two segments caudal to the exposed spinal cord. The stimulating electrode was placed ipsilateral to recorded RB neurons because these cells innervate ipsilateral skin. Current pulses (0.2-ms duration at 1 Hz) were applied by a stimulus isolation unit (DS2, Digitimer). The current amplitude was slowly increased to find firing threshold in the recorded sensory neuron. Once threshold was determined, either a single stimulus (0.2-ms duration) or trains of stimuli (2, 5, 15; 0.2 ms duration at 50 Hz) were applied and the sensory neuron activity recorded.
Whole cell voltage was recorded with a MultiClamp 700A (Axon Instruments), filtered at 30 kHz, and digitized at 63 Hz with a DigiData series 1322A (Axon Instruments) controlled by pClamp 8.2 software (Axon Instruments). Off-line analysis was done with both Matlab (Mathworks) and Clampfit 8.2 (Axon Instruments).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
, 1999) (results not shown), verifying the utility of 458-nm excitation for imaging cameleon. Ratio changes in single RB neurons on electrical stimulation
To test the utility of cameleon in fish neurons, we first asked whether cameleon could be used to detect calcium transients in a single RB neuron, the primary sensory neurons located in the spinal cord in larval fish. RB neurons were selected because they are relatively big (9 µm diam; large arrow in Fig. 1A) and easily activated in a controlled manner by electrical stimulation of the skin, as their highly branched peripheral processes extend into the skin (small arrow in Fig. 1A). Moreover, the stimulus parameters could be varied over a wide range to determine how the signals from cameleon change with increasing activation of the cell. Cameleon was expressed in RB neurons using a fragment of the promoter/enhancer of the Islet-1 gene (Higashijima et al. 2000), which has been shown previously to be functional in zebrafish neurons. Among several types of cameleon, each having a different affinity for calcium, we used a high-affinity version, YC2.1 (Miyawaki et al. 1999), to maximize the signal. The experiments were carried out by transient expression where transgenes are stochastically retained by a small number of cells at later stages. The DNA construct was injected into one cell stage embryos, and the injected embryos were analyzed at 2–3 days of development.
Experiments were first carried out in paralyzed fish because the paralysis eliminated potential artifacts resulting from large movements produced by the fish attempting to escape after the stimulus. On brief electrical stimulation of the skin, an increase in YFP and decrease in CFP was detected with a clear increase in YFP/CFP ratio, as shown in Fig. 1, B–D (4 cells, each from a different fish were studied in total). Images were collected at several different time intervals ranging from 30 Hz (video rate) to 2 Hz with a fixed scan speed. For faster imaging, the scanning area was reduced, whereas for slower imaging, blank time (nonscanning time) was inserted between frames. When the neurons were stimulated with repetitive stimuli (5–15 brief stimuli), a 12–18% increase in ratio was observed. Examples are illustrated in pseudocolor images in Fig. 1B and in plots of the ratio over time for trials at lower (2 Hz in Fig. 1C) and higher (10 Hz in Fig. 1D) sampling rates. Although responses were evident at low and high sampling rates, at high rates, bleaching occurred. This is evident in the 10-Hz imaging in Fig. 1D, where the baselines on the YFP and CFP channels after the trial are lower than just before the response. This is much reduced in the 2-Hz imaging in Fig. 1C.
The initial rise of the cameleon response was reasonably slow with a maximum sometimes reached after hundreds of milliseconds. The time course of decay of the calcium transients was much slower with the return to baseline typically taking 10 s with somewhat longer decays after application of more stimuli. The slow time course allowed reliable detection of the response of a cell even at the low sampling rates. We used the slow, 2-Hz imaging as our standard experimental interval because we could detect responses with minimal bleaching in each trial and therefore collect many trials.
We varied the number of stimuli to determine the ability of cameleon to resolve the strength of neuronal activity. In particular, we wanted to ask whether cameleon could detect calcium transients associated with one or very few action potentials. As shown in Fig. 2A, we were able to detect responses to a single, brief shock (0.2-ms duration, 23 µA), which produced increases in ratio of 3–6%. Similar results were obtained in two additional cells (data not shown). The current strength that was needed to evoke a detectable ratio change with a single shock was different from preparation to preparation, presumably because it depended on the relative distance between the stimulation electrode and the RB peripheral processes running nearby. The weakest stimulation condition we achieved was 8 µA with a 0.3-ms duration. When repetitive stimuli were applied, an increase in ratio was detected (Fig. 2A) that was most evident in the decay phase of the response. It was less obvious in the peak of the responses, which might have been missed as a result of the slow imaging we used to avoid bleaching. The observations with repetitive stimuli are consistent with the idea that, with the increase in stimuli, more spikes were induced, resulting in higher calcium concentrations.
|
The interpretation of the electrical stimulation experiment just described depends on how the RB cells respond to the stimulus. Prior work indicated that similar brief electrical stimuli produce only single action potentials in Xenopus RB neurons (Clarke et al. 1984), which in addition only fire single spikes on current injection. To examine the response properties of RB cells in zebrafish, we performed whole cell patch recordings from nine RB cells in six fish. On DC injection, all nine cells fired only single action potentials, even with large depolarizations (Fig. 2B). In six of the cells, we were able to test their response to electrical stimuli applied to the body. We found that application of a single brief (0.2 ms) electrical stimulus, as we applied in our cameleon imaging experiments, led to only a single spike (Fig. 2, C and D). Trains of stimuli led to a series of spikes, with one spike per stimulus (Fig. 2D). The electrophysiology thus verifies that the RB cells fire single spikes in response to our stimuli and further supports our conclusion that cameleon could detect the calcium concentration change associated with one or a few action potentials in RB cells.
For cameleon to be a useful tool, it is important that it can record responses in many trials. This was the case. Although fluorescence intensity slowly decreased with repeated trials at 2 Hz, the loss was minor. This allowed us to carry out over 50 trials on a single cell, with each trial consisting of 40 s of imaging. Importantly, almost the same size ratiometric signal was detected in both the first and the last trials in a series. YFP tended to show more of a decrease in fluorescence intensity within a single trial than CFP. This was emphasized when very bright excitation was used or when an increased number of exposures were collected during faster imaging, such as in video rate imaging. This uneven bleaching caused a slight downward trend of the ratio. In our usual experimental conditions, however, the effect was minimal, and there was no need to modify the raw data to visualize the ratio change.
After studying paralyzed fish, we performed the same experiments in unparalyzed fish held in place with agar. In this preparation, electrical stimulation of the skin often elicited an escape behavior, causing displacement of the cell. After the escape, the agar returned the cell close to its initial position (Fig. 3A; escape occurred at the marked frame, *). To reliably analyze the data, we wrote software that automatically tracked the moving cell and calculated the mean fluorescence intensity in both CFP and YFP channels (see METHODS). As shown in Fig. 3, B and C, the RB neurons in unparalyzed fish showed a clear ratio rise in response to electrical stimulation even in the presence of the movement from the escape. In the example of Fig. 3C, there is a gap in the trace at the * when the cell briefly moved out of the field of view. After the *, both YFP and CFP were lower than before the escape because the cell, although back in view, had not returned to the original, brightest optical section through the cell. Both YFP and CFP then rose as the cell gradually moved back into the original optimal focal plane. Nonetheless, the ratio showed a clear increase consistent with a calcium increase in response to the stimulus. Similar results were obtained in repeated trials in the same cell (Fig. 3, B and C) or in other cells (5 cells, each from a different fish were studied).
|
Even though the fish moved in response to the stimulus when not paralyzed, the time course of ratio change in these trials was very similar to that in the trials done in paralyzed fish, strongly suggesting that the observed increases in ratio were results of calcium influx and not movement artifacts. However, there was still a concern that movement artifacts might, to some extent, have contributed to the change in ratio. This led us to examine possible effects of movement on ratio value. First, we systematically investigated the effect of focal plane changes by moving the stage along the z axis. This revealed that the ratio changed minimally if the decrease of fluorescence intensity of a cell from the brightest plane was within 10% (roughly a focal plane change of within ±2 µm). Displacement of cells in our trials was, in most cases, within this range, except for the very frame when the escape was occurring. Next, we investigated the effect of cell movement alone by applying electrical stimulation to a place far away from the imaged RB cell or by applying an abrupt tap to the head with a glass probe to elicit escape movements without activation of the imaged RB cell. In both cases, little change in ratio was observed. Figure 3D shows an example of the head tap experiment. Taken together, these results indicate that movement artifacts had little to no effect on the ratio.
Calcium transients in motoneurons and interneurons during escape behavior
Having established the reliability of the technique, we asked a critical question: is cameleon able to detect calcium transients associated with neuronal activity during behavior. The escape behavior was selected because it is easily elicited by applying a sudden tap to the head (O'Malley et al. 1996). To examine the ability of cameleon to detect neuronal activity during this behavior, we focused our study on those neurons that we knew to be activated in escapes, including primary motoneurons and CiD interneurons. Imaging with the more conventional indicator, calcium green dextran, showed a rise of calcium concentration in these cells during escapes (Fetcho and O'Malley 1995; Ritter et al. 2001). To achieve cameleon expression in these neurons, we used the goldfish alpha-tubulin promoter (Hieber et al. 1998). This promoter can drive gene expression in nearly every cell type in zebrafish spinal cord. Due to the nature of the transient expression system, only a limited number of neurons are labeled in DNA-injected fish. The identity of the cell analyzed was confirmed by obtaining three-dimensional morphological images of the cell (e.g., see Fig. 4, A and B) after physiological imaging.
|
The primary motoneurons extensively innervate body wall muscles (Fig. 4, A and B) and their activity is directly responsible for rapid muscle contraction (Westerfield et al. 1986). As shown in Fig. 4C, a primary motoneuron located on the contralateral side to a tap showed a rise in ratio in conjunction with an escape. We performed this experiment in five cells, each from a different fish. All of them showed an increase in ratio on escapes. When multiple trials were performed on the same cell, we could see the ratio increase in most cases, with the ratio increasing by between 4 and 7%. In some trials, we were unable to detect a ratio change. It is likely in these cases that the fish did not perform an escape behavior, although we could not completely rule out the possibility that noise may have masked a subtle increase in ratio. When multiple taps were applied in rapid succession, a bigger ratio change was usually observed, with the ratio increase approaching 
10% (Fig. 4D).
The CiD interneuron sends its axon ipsilaterally and caudally (Fig. 5A). Its activity during an escape is presumed to assist in the activation motoneurons through synaptic contacts with them (Ritter et al. 2001). Imaging of CiD neurons expressing cameleon led to results similar to those for primary motoneurons. Three CiD interneurons all showed an increase in ratio in escapes with the signal size comparable to that in primary motoneurons. Figure 5B shows an example of a CiD cell response after a tap on the contralateral side. As we observed in primary motoneurons, multiple taps evoked bigger increases in ratio in CiD cells. Figure 5C shows an example in which the first tap was followed by the second tap after an interval of a few seconds. The second tap elicited another escape, which resulted in a further increase in ratio.
|
Calcium transients in RB neurons in stable transgenic zebrafish
In the experiments described in the preceding text, we used transient expression of the DNA construct. This is primarily because the transient system was convenient as an initial approach for testing the utility of cameleon. A genetic approach, however, is of most utility if one can generate stable lines that can be propagated and crossed into other lines to examine neuronal structure and function. One potential concern in producing stable lines of animals expressing cameleon is that expression of a calcium indicator in a large number of neurons might have negative effects on the animal's viability because of chelating of calcium by cameleon.
To address this concern, we generated stable transgenic fish expressing cameleon under the control of the pan-neuronal HuC promoter, which normally directs expression of an RNA-binding protein in all neurons (Kim et al. 1996). The transgenic line we produced expressed cameleon in all neurons (Fig. 6A), similar to transgenic fish expressing GFP under the same promoter (Park et al. 2000). The transgenic fish were viable and fertile and did not show any obvious defects.
|
The expression level of cameleon in this transgenic fish was similar to that in the cells imaged in the experiments in the previous sections. The expression levels in the RB cells in the stable line are, however, much more homogeneous than expression levels in the transient expression experiments. Measurements of the intensity of a series of RB cells in the stable line showed that the SD of the intensity averaged only 19% of the mean for three sets of six cells from different regions along the body.
We asked whether cameleon could detect calcium transients in RB neurons in the stable transgenic line. Because all neurons expressed the indicator, it was generally difficult to identify cell types in this fish. However, RB neurons could be reliably imaged (Fig. 6B, 
) for three reasons. First, RB neurons are located at the dorsal edge of the spinal cord. Second, they have a characteristic oval shape with a larger area than other cells. Third, the fluorescent intensity of RB cells was stronger than other cells at the time we imaged them (3 days). This probably reflects the early differentiation of this cell type, which allows a longer time for accumulation of cameleon. Along with the identifiability of the cell type, the cells were especially suitable for imaging in the transgenic fish because in lateral view the cord is thin at the top, making background fluorescence from other cells minimal even with a large confocal aperture.
We re-performed our initial experiments in the stable line, looking at the responses in RB neurons on skin stimulation in paralyzed fish. When the skin was stimulated, a clear ratio change was observed as shown in Fig. 6C (stimulation parameters in this experiment: 10 stimuli, 15 µA, 0.2-ms duration at 50 Hz). The amplitude of the ratio change was comparable to the one obtained in transient expression experiments with a similar stimulation strength. Four cells from two animals were studied, and all of the cells examined showed clear responses.
Taken together, the results indicate that stable transgenic fish expressing cameleon in all neurons are viable and that the indicator is functional in the transgenic line. This documents the feasibility of a genetic approach to calcium imaging in a vertebrate.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
Our imaging of RB cells shows that we can detect responses with cameleon to stimuli that produce single action potentials based on our patch recording experiments. This ability to detect responses from one or very few spikes is critical for using cameleon to monitor neuronal activity in vivo. The ability to identify cells firing very few spikes means that cameleon can serve as a useful tool for identifying active cells in vivo. Based on the electrical stimulation experiments, the indicator can provide at least a rough idea of the level of activity of a cell based on the magnitude of the increase in fluorescence of that cell from trial to trial. Nevertheless, issues of indicator saturation and response variability preclude inferring the exact number of spikes from the magnitude of the calcium change. In addition, although our data indicate that we can detect even a single action potential in RB cells as well as activity during behavior in motoneurons and CiD interneurons, this does not necessarily mean that that will be possible for all neurons. The ability to detect responses may, for example, depend on where in the cell the action potential is initiated and the extent of excitability of the cell soma, both of which will influence the somatic calcium influx.
The magnitude of ratio change obtained in our work appears to be bigger than those obtained in a C. elegans study in which neurons were subjected to direct electrical stimulation (Kerr et al. 2000). The observed ratio change in the RB experiments was 3–6% by a single shock and 12–18% by multiple (5–15) shocks. In contrast, a 4–5% ratio change from multiple shocks (100 pulses of 0.1-ms duration at 200 Hz with 300-µA intensity) was reported in the C. elegans study. Although the exact reason for this difference is unclear, it may be that the size of calcium transients in zebrafish neurons is larger than those in C. elegans neurons. The apparent larger signal size and larger soma size in zebrafish (at least relative to C. elegans; 5–10 µm diam for zebrafish neurons vs. 2 µm in C. elegans neurons) are helpful for performing experiments in this animal.
Cameleon was not especially good for resolving temporal dynamics. The peak detected with cameleon was sometimes hundreds of milliseconds after the beginning of the ratio increase. The basis for the slow rise is unknown. It may be related to the kinetics of the conformation change necessary for cameleon function. The time course of the cameleon decay was much slower than the rise, taking sometimes 10 s. A long decay time course is also observed in zebrafish spinal motoneurons, interneurons, and hindbrain neurons imaged with calcium green dextran (Fetcho and O'Malley 1995; Fetcho et al. 1998; Gahtan et al. 2002; O'Malley et al. 1996). The cameleon responses were slower than the fastest observed with calcium green, but in both cases, the decays occur over many seconds. We did not explore the relationship between cameleon expression levels and the time course of decay, but it is likely based on previous work that the time course of decay varies with the level of expression with high expression levels leading to a slower decay (Miyawaki et al. 1999). Because our main aim is to use cameleon as a tool to identify active cells in vivo, a slow time course is advantageous. It allows for slower image collection, which reduces bleaching. This in turn allows for collection of the many trials essential for studies of changes in the patterns of activity during the many behaviors produced by larval fish (Budick and O'Malley 2000; Ritter et al. 2001).
Although our aim was to use cameleon as an indirect measure of neuronal activity, ratiometric indicators are often used to measure absolute calcium levels. Relating the fluorescence changes we observed with cameleon to changes in absolute calcium levels is not easily accomplished in vivo. Calibration of calcium indicators in the intact fish is very difficult, as it is not possible to saturate the indicator with calcium or go to zero calcium without killing the fish and rendering it opaque. Previous rough estimates of calcium changes in zebrafish neurons measured with calcium green dextran suggest that calcium increases during escape behaviors by 200–300 nM (Fetcho and O'Malley 1995). These estimates are provided only to give some sense of the likely magnitude of the changes in calcium in response to our stimuli. Given our aim to use the indicator as an indirect measure of neuronal activity, the exact magnitude of the calcium changes is not as important as our ability to detect whether a cell has responded or not.
Movement is a potential problem for detecting neuronal activity during behavior. This problem is greatly reduced by the ratiometric nature of cameleon. We carefully investigated possible effects of focal plane change on ratio by systematically moving the stage along the z axis. The ratio changed little with focal plane in our experimental conditions. Consistent with this idea, we verified that the shift in position of a cell during an escape movement resulted in little change in ratio in RB neurons when the stimuli that elicited the escape movements were not ones that would activate the RB cell. Simultaneous acquisition of the CFP and YFP images is, of course, necessary to achieve this. The ratiometric measurements effectively compensated for gradual shifts in the focal plane of the cells after the completion of large escape movements. The ratiometric approach will become even more important in future work imaging larger numbers of cells, each of which might not be in the optimal, brightest focal plane.
Signal-to-noise level is also an important consideration for determining the utility of cameleon. In our experiments, in the worst case of single spike detection, the YFP/CFP ratio signal was twice the amplitude of the noise. In the largest responses, the ratio signal was five times the noise. The signal/noise level was not a serious problem in detecting the ratio change in RB neurons in paralyzed fish, but it could be a potential problem for detecting a small ratio change in moving cells in which the peak of the ratio change may not be reliably analyzed due to too much movement. For example, in the case of escape behavior, it is usually difficult to analyze the data in the frame in which the escape is occurring. Additionally, the signal/noise level increases when smaller cells are analyzed, presumably because the noise is dominated by photon noise and the smaller cross section of a small cell leads to fewer collected photons and a lower signal/noise. RB neurons are among the biggest zebrafish neurons with a nearly 10 µm diam, while there are many smaller neurons of 5–7 µm diam. We focused on big cells, but improving the signal/noise level will be important in future work to examine smaller neurons. It could, in principle, be increased by increasing excitation, but we were working at the maximum level we could without substantial bleaching of both CFP and YFP. CCD-based imaging might also enhance signal to noise; this approach is simplified by new image splitting devices that allow dual imaging with one CCD camera (Optical Insights, Santa Fe, NM). Protein expression level also affects fluorescence intensity and thus noise level. The production of transgenic lines with even higher levels of expression might help to increase signal/noise. Finally, some of the new genetically encoded single-wavelength indicators produce larger signals than cameleon and have better signal to noise (Nagai et al. 2001; Nakai et al. 2001). They are not, however, ideal for our preparation where ratiometric indicators help to reduce the problems associated with the movement of the fish.
We have used confocal microscopy in our study. The major advantage of using this instrument is its ability to acquire optical sections, eliminating background fluorescence along the z axis during recordings. Optical sectioning also enables the collection of high-quality morphological data, which were critical for unambiguous identification of cell types in our experiments. Although our studies depended on installation of the appropriate dichroic mirror and emission filters, new technology for spectral separation, now standard on the latest models of confocal microscopes (e.g., Zeiss and Leica), should facilitate the general use of genetically encoded indicators for studies of network physiology and behavior.
We were concerned that production of stable lines of fish expressing cameleon might be difficult if the buffering of calcium by the indicator compromised the health of the neurons and therefore the viability of the fish. Our demonstration that we could generate a stable line with cameleon expression in all neurons shows that this is not an insurmountable problem, although we cannot yet rule out more subtle behavioral effects of the cameleon expression. The HuC-cameleon line we made has so many labeled cells in the CNS that it is not ideal for studies of their function. It may, however, prove more useful for peripheral neurons in ganglia, olfactory epithelium, or the lateral line where the labeled cells are not embedded among many other cells as in the CNS. The viability of the HuC-cameleon line makes it likely that lines with cameleon targeted to specific subsets of neurons can be generated as well. Indeed, the major advantage of genetic indicators is the ability to target their expression to a subset of cells and to noninvasively label large populations of local interneurons that would be difficult to fill with indicator in other ways. The increasing number of promoter fragments that can direct expression to subsets of neurons, as well as the successful application of bacterial artificial chromosome recombination technology in zebrafish, offers the prospect of generating many such lines for functional studies (Jessen et al. 1998, 1999).
In the past year, genetically encoded indicators have been used to study activity in neuronal circuits of flies (Fiala et al. 2002; Liu et al. 2003; Reiff et al. 2002; Wang et al. 2003; Yu et al. 2003). These studies demonstrate the power of the approach to look at patterns of activity in invertebrate models. Our work extends this to an important vertebrate model. Zebrafish are an especially powerful model for optical studies because their transparency at larval stages allows imaging of neurons throughout the CNS of intact animals. They offer the additional advantage of many mutant lines with disruptions of behavior (Granato et al. 1996). The ability to use genetically encoded indicators in zebrafish will allow for stable lines expressing these indicators to be crossed into mutants to simultaneously study both structural and functional disruptions, just as GFP lines are now being used to study developmental deficits in mutants (Ono et al. 2001; Zeller et al. 2002).
|
DISCLOSURES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: J. Fetcho, Dept. of Neurobiology and Behavior, Life Science Bldg., SUNY, Stony Brook, NY 11794-5230 (E-mail: Jfetcho{at}notes.cc.sunysb.edu).
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
Bremmer F, Schlack A, Shah NJ, Zafiris O, Kubischik M, Hoffmann K, Zilles K, and Fink GR. Polymodal motion processing in posterior parietal and premotor cortex: a human fMRI study strongly implies equivalencies between humans and monkeys. Neuron 29: 287–296, 2001.[ISI][Medline]
Budick SA and O'Malley DM. Locomotor repertoire of the larval zebrafish: swimming, turning and prey capture. J Exp Biol 203: 2565–2579, 2000.[Abstract]
Clarke JD, Hayes BP, Hunt SP, and Roberts A. Sensory physiology, anatomy and immunohistochemistry of Rohon-Beard neurons in embryos of Xenopus laevis. J Physiol 348: 511–525, 1984.
Drapeau P, Ali DW, Buss RR, and Saint-Amant L. In vivo recording from identifiable neurons of the locomotor network in the developing zebrafish. J Neurosci Methods 88: 1–13, 1999.[ISI][Medline]
Fetcho JR, Cox KJA, and O'Malley DM. Monitoring activity in neuronal populations with single-cell resolution in a behaving vertebrate. Histochem J 30: 153–167, 1998.[ISI][Medline]
Fetcho JR and O'Malley DM. Visualization of active neural circuitry in the spinal cord of intact zebrafish. J Neurophysiol 73: 399–406, 1995.
Fiala A, Spall T, Diegelmann S, Eisermann B, Sachse S, Devaud JM, Buchner E, and Galizia CG. Genetically expressed cameleon in Drosophila melanogaster is used to visualize olfactory information in projection neurons. Curr Biol 12: 1877–1884, 2002.[ISI][Medline]
Gahtan E, Sankrithi N, Campos JB, and O'Malley DM. Evidence for a widespread brain stem escape network in larval zebrafish. J Neurophysiol 87: 608–614, 2002.
Granato M, van Eeden FJ, Schach U, Trowe T, Brand M, Furutani-Seiki M, Haffter P, Hammerschmidt M, Heisenberg CP, Jiang YJ, Kane DA, Kelsh RN, Mullins MC, Odenthal J, and Nusslein-Volhard C. Genes controlling and mediating locomotion behavior of the zebrafish embryo and larva. Development 123: 399–413, 1996.[Abstract]
Hieber V, Dai X, Foreman M, and Goldman D. Induction of alpha1-tubulin gene expression during development and regeneration of the fish central nervous system. J Neurobiol 37: 429–440, 1998.[ISI][Medline]
Higashijima S, Hotta Y, and Okamoto H. Visualization of cranial motor neurons in live transgenic zebrafish expressing green fluorescent protein under the control of the islet-1 promoter/enhancer. J Neurosci 20: 206–218, 2000.
Higashijima S, Okamoto H, Ueno N, Hotta Y, and Eguchi G. Highfrequency generation of transgenic zebrafish which reliable express GFP in whole muscles or the whole body by using promoters of zebrafish origin. Dev Biol 192: 289–299, 1997.[ISI][Medline]
Jessen JR, Meng A, McFarlane RJ, Paw BH, Zon LI, Smith GR, and Lin S. Modification of bacterial artificial chromosomes through chi-stimulated homologous recombination and its application in zebrafish transgenesis. Proc Natl Acad Sci USA 95: 5121–5126, 1998.
Jessen JR, Willett CE, and Lin S. Artificial chromosome transgenesis reveals long-distance negative regulation of rag1 in zebrafish. [letter]. Nat Genet 23: 15–16, 1999.[ISI][Medline]
Kerr R, Lev-Ram V, Baird G, Vincent P, Tsien RY, and Schafer WR. Optical imaging of calcium transients in neurons and pharyngeal muscle of C. elegans. Neuron 26: 583–594, 2000.[ISI][Medline]
Kim CH, Ueshima E, Muraoka O, Tanaka H, Yeo SY, Huh TL, and Miki N. Zebrafish elav/HuC homologue as a very early neuronal marker. Neurosci Lett 216: 109–112, 1996.[ISI][Medline]
Legendre P and Korn H. Glycinergic inhibitory synaptic currents and related receptor channels in the zebrafish brain. Eur J Neurosci 6: 1544–1557, 1994.[ISI][Medline]
Liu L, Yermolaieva O, Johnson WA, Abboud FM, and Welsh MJ. Identification and function of thermosensory neurons in Drosophila larvae. Nat Neurosci 6: 267–273, 2003.[ISI][Medline]
Miyawaki A, Griesbeck O, Heim R, and Tsien RY. Dynamic and quantitative Ca2+ measurements using improved cameleons. Proc Natl Acad Sci USA 96: 2135–2140, 1999.
Miyawaki A, Llopis J, Heim R, McCaffery JM, Adams JA, Ikura M, and Tsien RY. Fluorescent indicators for Ca2+ based on green fluorescent proteins and calmodulin [see comments]. Nature 388: 882–887, 1997.[Medline]
Nagai T, Sawano A, Park ES, and Miyawaki A. Circularly permuted green fluorescent proteins engineered to sense Ca2+. Proc Natl Acad Sci USA 98: 3197–3202, 2001.
Nakai J, Ohkura M, and Imoto K. A high signal-to-noise Ca(2+) probe composed of a single green fluorescent protein. Nat Biotechnol 19: 137–141, 2001.[ISI][Medline]
O'Malley DM, Kao Y-H, and Fetcho JR. Imaging the functional organization of zebrafish hindbrain segments during escape behaviors. Neuron 17: 1145–1155, 1996.[ISI][Medline]
Ono F, Higashijima S, Shcherbatko A, Fetcho JR, and Brehm P. Paralytic zebrafish lacking acetylcholine receptors fail to localize rapsyn clusters to the synapse. J Neurosci 21: 5439–5448, 2001.
Park HC, Kim CH, Bae YK, Yeo SY, Kim SH, Hong SK, Shin J, Yoo KW, Hibi M, Hirano T, Miki N, Chitnis AB, and Huh TL. Analysis of upstream elements in the HuC promoter leads to the establishment of transgenic zebrafish with fluorescent neurons. Dev Biol 227: 279–293, 2000.[ISI][Medline]
Reiff DF, Thiel PR, and Schuster CM. Differential regulation of active zone density during long-term strengthening of Drosophila neuromuscular junctions. J Neurosci 22: 9399–9409, 2002.
Ritter DA, Bhatt DH, and Fetcho JR. In vivo imaging of zebrafish reveals differences in the spinal networks for escape and swimming movements. J Neurosci 21: 8956–8965, 2001.
Wang JW, Wong AM, Flores J, Vosshall LB, and Axel R. Two-photon calcium imaging reveals an odor-evoked map of activity in the fly brain. Cell 112: 271–282, 2003.[ISI][Medline]
Warren DJ, Fernandez E, and Normann RA. High-resolution two-dimensional spatial mapping of cat striate cortex using a 100-microelectrode array. Neuroscience 105: 19–31, 2001.[ISI][Medline]
Westerfield M, McMurray JV, and Eisen JS. Identified motoneurons and their innervation of axial muscles in the zebrafish. J Neurosci 6: 2267–2277, 1986.[Abstract]
Yu DH, Baird GS, Tsien RY, and Davis RL. Detection of calcium transients in Drosophila mushroom body neurons with camgaroo reporters. J Neurosci 23: 64–72, 2003.
Zeller J, Schneider V, Malayaman S, Higashijima S, Okamoto H, Gui J, Lin S, and Granato M. Migration of zebrafish spinal motor nerves into the periphery requires multiple myotome-derived cues. Dev Biol 252: 241–256, 2002.[ISI][Medline]
This article has been cited by other articles:
|
|
 |
|
  T. Hendel, M. Mank, B. Schnell, O. Griesbeck, A. Borst, and D. F. Reiff Fluorescence Changes of Genetic Calcium Indicators and OGB-1 Correlated with Neural Activity and Calcium In Vivo and In Vitro J. Neurosci., July 16, 2008; 28(29): 7399 - 7411. [Abstract] [Full Text] |
in 1st frame) collected at 500-ms intervals in a trial from an unparalyzed fish in which electrical stimuli (10 pulses of 0.2-ms duration at 35 µA and 100 Hz) were applied to the skin resulting in an escape, which led to movement of the fish and the neuron. Frames read from left to right in successive rows with the 1st in each row numbered. The electrical stimuli were applied to the skin at frame 5 (*). The cell in this frame is distorted because it moves briefly out of the frame after the stimulus. The neuron is back in view by the next frame in the sequence. B: time course of YFP emission (top), CFP emission (middle), and YFP/CFP ratio (bottom) in the RB neuron shown in A. On electrical stimulation, the movement of the escape behavior led to a large decrease in fluorescence intensity in both YFP and CFP channels (at the time marked, * on the YFP plot) when the cell briefly moved out of focus. A ratio could not be calculated for this frame. The cell returns rapidly to nearly the original focal plane and an increase in ratio is seen after the escape. C: another trial performed in the same RB neuron with the same stimulation as in A and B. Fluorescence intensity of both channels decreased immediately after the escape in this trial (
in top and middle) due to a shift of focal plane. Nonetheless, the ratio reports a clear rise of calcium concentration (
in bottom). D: plots for the same RB neuron after a head tap, which leads to an escape but does not activate the RB cell. There was a decrease of fluorescence intensity in both YFP and CFP channels after the escape event (*), again due to a focal plane change. The ratio, however, was stable, indicating that the movements themselves do not significantly affect the ratio. Scale bar in frame 1 of A = 10 µm.
