Introduction

The parallel monitoring of multiple fluorescence probes has become increasingly important for the microscopy of cellular processes in many biological disciplines. E.g. the use of spectrally distinct styryl FM dyes revealed information about endocytosis kinetics [1] as well as spontaneous recycling synaptic vesicles [2]. However, interactions between fluorescent dyes beyond an obvious overlap of their fluorescent spectra might confound experimental results and have to be carefully evaluated.

In this study, we focused on the spectrally distinct fluorescent dyes FM®2-10 or FM®1-43 and CypHer™5E. Individually, these dyes are extensively used for the analysis of synaptic transmission. Their combined use might present an interesting tool to investigate the dynamics of synaptic vesicle recycling as these dyes have individual advantages and could in principle easily be combined.

FM dyes are particularly useful for the analysis of lipid membrane trafficking. Due to their about two orders of magnitude higher fluorescence in membranes compared to polar solvents such as water [3] they offer the advantage to determine the brightness of single synaptic vesicles [4–7]. In neuronal cells these dyes offer the possibility to stain different pools of recycling synaptic vesicles dependent on the electrical stimulus [8, 9]. Thus, FM dyes allow the direct measurement of exocytosis of previously stained vesicles [10], but only an indirect measurement of endocytosis and related kinetics [10, 11]. Additionally, repetitive stimulation experiments are complex due multiple rounds of staining, washing extracellular dye away and subsequent destaining [12, 13]. Several different FM dyes with different spectra and distinct membrane departitioning characteristics have been designed [3, 14]. E.g. FM®2-10 and FM®1-43, in which FM®2-10 departitions faster corresponding to better dye unloading from synaptic vesicles [3]. Differences between FM dyes were used to analyze several aspects of synaptic vesicle recycling [1, 2].

Many of these disadvantages of FM dyes have been overcome with the introduction of the pH-dependent CypHer™5E [15–17]. CypHer™5E is a variant of the dye Cy5™, that only fluoresces in acidic conditions (pKa = 6.1) [15, 16]. Due to its pH-dependency this probe reports changes of the vesicular pH-gradient during the different phases of vesicle recycling when coupled to an antibody against the luminal domain of a synaptic vesicle protein. Recently, many questions regarding synaptic vesicle recycling have been answered with this probe [6, 18–21]. Compared to homologs of the pH-dependent GFP variant pHluorin [22–24], antibodies labelled with CypHer™5E have the advantage that they can be flexibly applied to native tissue. As pHluorins have to be genetically encoded in transgenic animal species or used in overexpression systems, sensitive physiological measurements might be obscured due to the artificial expression of the protein.

The combination of the pH-dependent probe CypHer™5E coupled to a vesicular antibody with either one of the FM dyes FM®2-10 or FM®1-43 would offer various advantages with respect to synaptic vesicle recycling. As especially FM®1-43 is widely used in single vesicle analysis [4–7], a combination with CypHer™5E give the opportunity to calibrate the CypHer™5E fluorescence to the quantal nature of synaptic transmission [25]. Especially experiments using repetitive stimulation or focusing on endocytosis would benefit from such a quantal analysis of synaptic vesicle recycling, e.g. study of drug effects [12, 26]. Furthermore as FM dyes as well as CypHer™5E coupled to an antibody are able to label individual vesicle pools a combined use could solve questions regarding the intermixing, origin and reuse of different synaptic vesicle pools [2, 9, 27]. In contrast to spectrally different FM dyes like FM®1-43 and FM®5-95 [2], which allow only one measurement of exocytosis, the use of CypHer™5E offers the possibility for multiple measurements due to the permanent binding of the antibody.

Here, we analysed the potential combination of the FM dyes FM®2-10 or FM®1-43 with CypHer™5E for parallel imaging experiments. We report the concentration-dependent quenching of CypHer™5E fluorescence by the FM dyes. Furthermore, we show how this effect influences the dynamic CypHer™5E signal monitored in in vitro experiments with primary rat hippocampal neurons. While we have found that a parallel measurement with both dyes using low FM dye concentrations is generally possible, our results might warrant a caveat for experimenters conducting similar imaging protocols.

Experimental

Cell Culture

Hippocampal neuronal cultures were prepared from 1 to 3 days old Wistar rats (Charles River, USA) as described [21]. Newborn rats were sacrificed by decapitation in accordance with the guidelines of the State of Bavaria. Hippocampi were removed from the brain and transferred into ice cold Hank’s salt solution, and the dentate gyrus was cut away. After digestion with trypsin (5 mg ml−1) cells were triturated mechanically and plated in MEM medium, supplemented with 10 % fetal calf serum and 2 % B27 Supplement (all from Invitrogen, Taufkirchen). Experiments were performed between 25 and 30 days in vitro.

Fluorescent Labeling

For the exemplary FM®2-10 experiment (Fig. 1c, e) recycled synaptic vesicles were labeled with 100 μM FM®2-10 (Invitrogen, Karlsruhe). To stain the total recycling pool nerve terminals were loaded with 1,200 pulses at 40 Hz [28]. Dye was allowed to remain on the cells for 60 s after the end of the stimulation to permit complete endocytosis, and was subsequently removed by a 7 min wash-out.

Fig. 1
figure 1

Dye structure and corresponding spectra of the styryl dye FM®2-10 and the pH-dependent Cy5™ variant CypHer™5E. a Dye structure of FM®2-10 [14]. b Excitation (dashed) and emission spectra of the dye FM®2-10 for two different pH values. Typical excitation and emission filters are depicted as grey rectangles which were chosen to avoid any bleed-through of the two fluorophors in dual colour experiments. c Exemplary fluorescence images of hippocampal neurons stained with FM®2-10 following the wash-out of the external dye before and after electrical stimulation with 200 pulses at 10 Hz. d Schematic of membrane staining with FM dyes. After adding FM dye to the extra cellular medium, the dye incorporates preferentially into the membrane of hippocampal neurons (2). Dye incorporated into lipid membranes exhibits a higher fluorescence compared to dye in solution. The membrane of synaptic vesicles, which fuses with the plasma membrane upon an electrical stimulus, is stained with the dye (3). After compensatory endocytosis stained synaptic vesicles are inside the synapse (4). After washing the extracellular dye away, only fluorescent synaptic vesicles remained (5). Upon an electrical stimulus these vesicles again fuse with the membrane and lose their fluorescent dye due to the concentration gradient (6), which enables the measurement of exocytosis. An exemplary fluorescence profile of the synapses shown in (c) upon a stimulation with 200 pulses at 10 Hz is shown in (e). f Dye structure of CypHer™5E [16]. (g) Excitation (dashed) and emission spectra of the dye CypHer™5E for two different pH values. Typical excitation and emission filters are depicted. h Exemplary fluorescence images of hippocampal neurons labelled with αSyt1-CypHer™5E before and after electrical stimulation with 200 pulses at 10 Hz and after recovery. i Illustration of αSyt1-CypHer™5E labelling and its fluorescence properties. Antibodies labelled with CypHer™5E against the synaptic vesicle protein synaptotagmin1 are inside the vesicle (1). CypHer™5E is fluorescent in the acidic lumen of the synaptic vesicle. Upon an electrical stimulus vesicles fuse with the plasma membrane and the intravesicular protons are lost. CypHer™5E is less fluorescent (2). After compensatory endocytosis, (3) synaptic vesicles are reacidified and the fluorescence of CypHer™5E increases (4). Thus, αSyt1-CypHer™5E enables the measurement of exocytosis and endocytosis/reacidification of synaptic vesicles. An exemplary fluorescence profile of the synapses shown in (h) upon a stimulation with 200 pulses at 10 Hz is shown in (j)

Full size image

For the αSyt1-CypHer™5E experiments, cultured hippocampal neurons were incubated for 20 min with 0.6 μg of CypHer™5E labeled anti-Synaptotagmin1 antibody (αSyt1-CypHer™5E; Synaptic Systems, Goettingen) in extracellular medium containing 40 mM K+ to stain the total synaptic vesicle recycling pool.

Imaging

Experiments were conducted at room temperature on a Nikon TI-Eclipse inverted microscope equipped with a 60×, 1.2 NA water immersion objective and a Perfect Focus System™. Fluorescent dyes were excited by a Nikon Intensilight C-HGFI through excitation filters centered at 482 nm and 640 nm using dichroic longpass mirrors (cut-off wavelength 500 nm and 660 nm), respectively. The emitted light passed emission band-pass filters ranging from 500 nm–550 nm to 660 nm–700 nm (Semrock, Rochester) and was projected onto a cooled EM-CCD camera (iXonEM DU-885, Andor, Belfast). The filter sets were chosen to avoid any spectral overlap or bleed-through of the respective fluorophors. Cover slips were placed into a perfusion chamber (volume = 500 μl) containing extracellular medium consisting of (in mM): 144 NaCl, 2.5 KCl, 2.5 CaCl2, 2.5 MgCl2, 10 Glucose, 10 Hepes, pH = 7.4.

In FM experiments images were recorded with 200 ms exposure time and in αSyt1-CypHer™5E experiments with 900 ms exposure time at 0.5 Hz frame rate each.

Synaptic boutons were stimulated by electric field stimulation (platinum electrodes, 10 mm spacing, 1 ms pulses of 50 mA and alternating polarity) with 200 pulses at 10 Hz; 10 μM 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX, Tocris Bioscience,Bistrol) and 50 μM D-amino-5-phosphonovaleric acid (D,L-AP5, Tocris Bioscience, Bistrol) were added to prevent recurrent activity.

Data Analysis

All image analysis was performed using custom-written routines in MATLAB (The MathWorks, Natick, MA). Recorded image stacks were used to automatically define peak regions of interest (ROI) of synaptic bouton size [29], where a stimulation-evoked fluorescence decrease occurred in difference images. The average of each ROI was calculated for each image to obtain fluorescence intensity profiles for each bouton. Graphs were generated using MATLAB (The MathWorks Inc., Natick), Microsoft® Excel 2010 (Microsoft® Corporation, Redmond) or Origin® 8.5.1 (OriginLab Corporation, Northampton, Massachusetts).

Fluorescence Spectrophotometry

All fluorometric measurements were performed on a Cary Eclipse fluorescence spectrophotometer (Agilent Technologies Deutschland GmbH, Boeblingen).

All solutions were measured in quartz cuvettes (sample volume = 400 μl) at an excitation wavelength of 473 nm or 640 nm with excitation and emission slits set at 5 nm. FM®2-10, FM®1-43, CypHer™5E and αSyt1-CypHer™5E were solved in respective concentrations in extracellular medium with pH = 7.4 or pH = 5.3.

For the determination of FM dye excitation and emission spectra, samples, including calibration samples, were mixed with 20 % 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS) [30, 31], which is a zwitterionic detergent used to increase the absolute fluorescence of FM dyes (Fluka™, Buchs) [31].

Results and Discussion

Properties of FM Dyes and αSyt1-CypHer™5E

Styryl dyes like FM®2-10 and FM®1-43 are widely used to analyse synaptic vesicle recycling [1, 2, 6, 10, 21]. The dye molecules (Fig. 1a) consist of a lipophilic tail region, which causes the dye to partition into lipids such as cellular membranes, a fluorescent middle region with two aromatic rings and a positively charged head, which prevents the dye from flipping across the membrane [3]. The fluorescence of these dyes shows a marked increase of the fluorophore core that lies in the middle of the FM dye structure, when they are inserted into lipid membranes [14]. Thus almost the entire signal is emitted by dye molecules in membranes in contrast to dye dissolved in polar solutions such as water [3, 31]. In our measurements FM dyes exhibited a 100-fold increase in fluorescence when CHAPS was added to the extracellular medium. Above a critical concentration this soluble amphiphile form stable aggregates called micelles [32]. FM dye molecules partition into these CHAPS micelles [31]. Hereby the lower dielectric constant of the lipid solvent (compared with non CHAPS) is responsible for the increase in fluorescence [31, 33]. The critical micelle concentration (CMC) of CHAPS is 6.50 mM [34] and is slightly reduced in our system due to the used extracellular medium [35]. Numerous similar dyes with different properties such as shifted fluorescence spectra or modified tails have been generated [3, 14]. The spectra of FM dyes are pH independent (Fig. 1b and Fig. S1B), which is crucial for the analysis of synaptic vesicle recycling as the intravesicular lumen is acidic with a pH of about 5 compared to neutral pH of the cytoplasm and extracellular space. This is due to the activity of the vacuolar H+ ATPase and is required for neurotransmitter accumulation [24, 36, 37]. Specific staining and destaining protocols of the dyes have been used in the past two decades to visualize many aspects of synaptic vesicles recycling [1, 2, 6, 7, 38]. An exemplary fluorescence image of hippocampal neurons stained with FM®2-10 is shown in Fig. 1c and the corresponding typical staining scheme with exocytosis kinetics is shown in Fig. 1d, e.

Recently, the pH-dependent Cy5™ variant CypHer™5E (Fig. 1f) has been introduced in the analysis of synaptic vesicle recycling [6, 21]. CypHer™5E fluoresces in the acidic milieu of synaptic vesicles (pH = 5.3), but not if it is exposed to the extracellular pH (pH = 7.4; Fig. 1g). In accordance with previous studies we found a reduction of the CypHer™5E fluorescence at pH = 7.4 to 30.19 % compared to pH = 5.3 [16, 18]. Coupled to a monoclonal antibody directed against the intravesicular domain of synaptotagmin1 (αSyt1-CypHer™5E), it enables the selective labelling of functional terminals by antibody uptake and visualization of synaptic exo- and endocytosis cycles according to the stimulation-dependent change in fluorescence (Fig. 1h–j).

Effect of FM Dyes on CypHer™5E Fluorescence–Spectrophotometry

First, we measured the influence of different concentrations of FM®2-10 on the emission spectrum of αSyt1-CypHer™5E. We used typical concentrations of FM®2-10, as they are used in live-cell experiments ranging from a relatively low concentration of 25 μM to the highest of 400 μM [1, 3, 27, 39]. With increasing concentrations of FM®2-10 the emission intensity of αSyt1-CypHer™5E decreased (Fig. 2a; Pearson’s r = −0.90, p < 0.01). Additionally, we analyzed the effect of 200 μM FM®2-10 on the emission maximum of pure CypHer™5E and measured a reduction of 49.46 % ± 0.79 % in accordance with the reduction of 46.36 % for αSyt1-CypHer™5E (see Fig. 2a, c). We also analyzed the αSyt1-CypHer™5E emission spectra as a function of typical FM®1-43 concentrations ranging from 2.5 μM to 100 μM (Fig. 2b; Pearson’s r = −0.97, p < 0.01) [1, 2, 6, 30, 31]. At the highest concentrations of FM®2-10 and FM®1-43, emission spectra with CHAPS [30, 31] are depicted, which lead to a small shift in the emission spectrum, but not to a further reduction indicating that FM dyes do not directly quench αSyt1-CypHer™5E (Fig. 2a, b). We used a concentration of 325.27 mM CHAPS which is high above the CMC and thus not only micelles are formed also at such high concentrations surfactants could form worm-like structures and bilayers, which encapsulate extracellular medium.

Fig. 2
figure 2

Fluorescence characteristics of the probe CypHer™5E as a function of FM dye concentration. Emission spectra of CypHer™5E at pH = 5.3 for different concentrations of the styryl dyes FM®2-10 (a) and FM®1-43 (b). At the highest FM dye concentrations CHAPS was added (molar ratios, CHAPS/FM®2-10 = 813.25 and CHAPS/FM®1-43 = 3,253). c Maximum emission intensities normalized to controls as a function of FM®2-10 (black squares) and FM®1-43 (red circles) concentration

Full size image

With increasing concentrations of FM dyes the emission intensity of αSyt1-CypHer™5E decreased (Fig. 2c). This reduction is not a bleaching effect of the fluorescent dye CypHer™5E, as multiple measurements revealed no change in amplitude of the emission spectra (Fig. S2; Pearson’s r = −0.25, p = 0.44). Furthermore a dimming of αSyt1-CypHer™5E fluorescence by interception of the excitation and emission light was excluded. Therefore the absorptions at the excitation (640 nm) and emission (664 nm) wavelengths of CypHer™5E were measured in the presence of the lowest and highest concentration of the respective FM dyes [40]. For the absorption path length of 5 mm the intensities after passing through the FM dyes containing solution were in the range of 93.39 % to 99.34 %, indicating a minimal effect in our system.

Effect of FM Dyes on CypHer™5E Fluorescence–Fluorescence Microscopy

To analyse the effects of FM dyes on αSyt1-CypHer™5E measurements under more physiological conditions, we performed live-cell experiments. Primary rat hippocampal neurons were labelled with αSyt1-CypHer™5E in a 40 mM K+ solution to ensure the fluorescent labelling of the entire vesicle recycling pool. The cells were then stimulated at least two consecutive times with 200 pulses at 10 Hz (Fig. 3a). The first stimulation showed the expected fluorescence time course of cells labelled with αSyt1-CypHer™5E (Fig. 3b, c) [18, 21]. The second stimulation was in the presence of FM®2-10. Depending on the concentration of the FM dye, the baseline fluorescence, which is the constant fluorescence level before the onset of the electrical stimulation, of the previously detected spots was lowered. As the protein Synaptotagmin1, similar like many other vesicular proteins, is located on the surface of the synapse [41, 42], the CypHer™5E dye is accessible for the FM dyes in solution. Thus the baseline fluorescence is quenched. Additionally, compared to control experiments the fluorescence recovery after the end of the stimulation was significantly reduced (Fig. 3d; Kruskal–Wallis test: χ2 3,15 = 8.19, p < 0.05) as FM®2-10 enters synaptic vesicles and quenches the intravesicular CypHer™5E (Fig. 3a). This quenching would probably take place between the FM dye in lipid phase of the vesicle and the CypHer™5E dye inside the synaptic vesicle (Fig. 1d, i). After a thorough washout of the external FM dye, an optional third stimulation revealed a recovery of the αSyt1-CypHer™5E fluorescence signal. The amount of quenching in the live-cell experiments exhibits a similar characteristic as the data obtained from spectrophotometry (Fig. 3e; Pearson’s r = 0.97, p < 0.05). Furthermore, the effect of FM®1-43 was tested. Increasing concentrations of the dye reduced the fluorescence signal of αSyt1-CypHer™5E and was again comparable to the spectrophotometer data (Fig. 3f; Pearson’s r = 0.97, p < 0.05).

Fig. 3
figure 3

Live cell experiments of hippocampal neurons labelled with αSyt1-CypHer™5E and subsequently stained with FM®2-10 or FM®1-43. a Experimental procedure. Functional synapses of rat hippocampal neurons were labelled with αSyt1-CypHer™5E, which fluoresces due to the intravesicular acidic pH. A control stimulus of 200 pulses at 10 Hz was delivered before adding various concentrations of FM®2-10 or FM®1-43 or DMSO as control. The cells were stimulated a second time with 200 pulses at 10 Hz to analyse the influence of FM dyes on CypHer™5E intensity. Optionally, after washing away the extracelluar FM dye, neurons were stimulated with 200 pulses at 10 Hz. b Exemplary average fluorescence profiles of αSyt1-CypHer™5E without FM dye, with 100 μM FM®2-10 and after washout. c Exemplary fluorescence images of hippocampal neurons labelled with αSyt1-CypHer™5E before and after electrical stimulation with 200 pulses at 10 Hz and after recovery. Same region at the end of the protocol stained with FM®2-10. Scale bar, 10 μm. d Average fluorescence profiles of αSyt1-CypHer™5E for different concentrations of FM®2-10 (≥3 experiments with >400 boutons for each). All profiles are normalized to respective controls. Relative quenching as a function of FM®2-10 (e) and FM®1-43 (f) concentration for live cells experiments (blue and red circles) and spectrophotometer data (black squares)

Full size image

In accordance with a previous study, we observed that high concentrations of FM dyes resulted in a lower fusion probability (10 μM FM®1-43: 16.73 % ± 7.43 % compared 10 % to 17 % reduction estimated by Zhu. Y. and Stevens C.F.) [43]. However, this effect did not tamper our quenching analysis as the fluorescence profiles were normalized to their respective minimum as well as to the baseline before stimulation (Fig. 3d).

Conclusion

In this study, we report the quenching effect of the FM dyes FM®2-10 and FM®1-43 on CypHer™5E fluorescence. Besides the reduction of the emission spectrum of CypHer™5E by FM dyes in spectrophotometric measurements, we showed that this effect is also apparent in in vitro experiments with primary rat hippocampal neurons labelled with these two dyes. The spectrophotometric data revealed that this reduction in fluorescence is due to the quenching effect of FM dyes on CypHer™5E fluorescence rather than an inhibition of endocytosis as could be supposed to fluorescence microscopy data alone. The vesicle-like complexes built by the used high concentrations of the detergent CHAPS in the spectroscopy study are similar to synaptic vesicles in the in vitro experiments arguing for the comparable results obtained by these experiments.

We showed that the decrease of CypHer™5E fluorescence by FM dyes is a process of fluorescence quenching as we were able to exclude dimming by interception of excitation and emission light [33, 40]. Analysis of the Stern–Volmer relationship [44] revealed a linear relation between the concentration of the quencher (FM dyes) and the ratio of fluorescence of CypHer™5E with and without the quencher, indicating a dynamic or a static quenching mechanism (Fig. S3A, B). The reduced fluorescence recovery after the end of the stimulation indicated that the quenching process requires close proximity to the fluorophore and cannot occur when FM dyes and CypHer™5E are separated by cell and vesicle membranes [40]. Furthermore the recovery of the αSyt1-CypHer™5E fluorescence signal due to washout of FM dyes almost excluded a static quenching mechanism as possible complexes formed by FM dyes and CypHer™5E are unlikely to be reformed. Any form of resonance energy transfer and dark quenching could additionally be excluded as the emission spectra of CypHer™5E do not overlap with the absorption spectra of the investigated FM dyes. However, the increase of the light scattering background of the CypHer™5E emission spectrum (Fig. 2a) together with the more effective quenching of FM®1-43 compared to FM®2-10 (Fig. 2c) give a hint to a dynamic quenching via photoinduced electron transfer upon van der Waals contact [45, 46]. Our data suggest that a parallel measurement with both dyes using low FM dye concentrations is generally possible, even though a small amount of quenching of CypHer™5E by FM dyes have to be accepted. Special care has to be taken regarding kinetic measurements when using both dyes.