|Home | About | Journals | Submit | Contact Us | Français|
A desire to better understand the role of voltagegated sodium channels (NaVs) in signal conduction and their dysregulation in specific disease states motivates the development of high precision tools for their study. Nature has evolved a collection of small molecule agents, including the shellfish poison (+)-saxitoxin, that bind to the extracellular pore of select NaV isoforms. As described in this report, de novo chemical synthesis has enabled the preparation of fluorescently labeled derivatives of (+)-saxitoxin, STX-Cy5, and STX-DCDHF, which display reversible binding to NaVs in live cells. Electrophysiology and confocal fluorescence microscopy studies confirm that these STX-based dyes function as potent and selective NaV labels. The utility of these probes is underscored in single-molecule and super-resolution imaging experiments, which reveal NaV distributions well beyond the optical diffraction limit in subcellular features such as neuritic spines and filopodia.
Voltage-gated sodium channels (NaVs) serve as an obligatory component of the molecular machinery that controls initiation and propagation of action potentials in electrically excitable cells (Caldwell and Levinson, 2004; Catterall, 1992). These integral membrane complexes are expressed as a ~260 kDa pore forming α-subunit with one or two ancillary 33–36 kDa β-proteins. Neuronal activity is shaped in large part by intrinsic differences in the ten known mammalian NaV α-subtypes and four β-subtypes, their expression levels, tissue distribution, and cellular patterning (Catterall et al., 2005). Not surprisingly, dysregulation of NaV function is associated with profound pathologies such as neurodegenerative disease, epilepsy, and chronic pain (Catterall et al., 2008). A desire to elucidate the underlying biochemical mechanisms that influence NaV expression and distribution and to understand how such changes alter signal conduction in neuronal cells motivates the development of molecular tools to probe dynamic processes associated with this protein class.
Distribution of NaVs to discrete cellular locales within a neuron is achieved through a combination of targeted intracellular transport, membrane trafficking, and endocytosis (Leterrier et al., 2010; Catterall, 2010; Dib-Hajj and Waxman, 2010; Cusdin et al., 2008). Nav clustering at the axon initial segment and nodes of Ranvier plays a key role in action potential generation and saltatory conduction in myelinated nerve cells (Leterrier et al., 2011). Dynamic regulatory pathways also function to tune specific electrophysiological properties of the cell in response to firing activity, neuronal development, myelination, and nerve injury. Recent studies present compelling evidence that functional NaVs assemble in sub-diffraction-sized dendritic structures, intimating a multiplicity of functions for these proteins beyond axonal depolarization (Remy et al., 2010). Such data is beginning to reveal how compartmentalization of NaVs in cellular substructures can play critical roles in fine-tuning synaptic transmission. Deciphering the mechanisms that impart spatiotemporal control of NaV distribution is thus tightly entwined with questions of how these channel proteins orchestrate specific neurological processes (Duclohier, 2009). To this end, new technologies for visualizing functional NaVs in living cells at the resolution of individual proteins are desired.
Our present understanding of NaV function at the systems level is shaped by electrophysiology recordings of ionic currents in living cells and immunohistochemical studies on fixed samples (Lorincz and Nusser, 2010; Trimmer and Rhodes, 2004). Although GFP labeling of individual NaV isoforms presents a viable tool for NaV visualization (Zimmer et al., 2002a, 2002b; Hallaq et al., 2006), attachment of a fluorescent protein to the NaV α-subunit has been shown to greatly diminish plasma membrane expression levels and to significantly alter ion conduction properties (Maue, 2007; Baker et al., 2007; Schofield et al., 2008; Ataka and Pieribone, 2002). The latter problem similarly disadvantages use of peptide toxin-based imaging tags to track NaV dynamics, as does the complex and/or poorly defined binding kinetics of such probes (Massensini et al., 2002; Beaurepaire et al., 2004; Nakada et al., 2003).
Nature has provided a collection of topologically unique guanidinium-derived poisons that bind reversibly and with nanomolar affinity to the outer mouth of the NaV pore (Figure 1A) (Llewellyn, 2009; Tikhonov and Zhorov, 2011; Scheib et al., 2006). Owing to their exquisite selectivity and favorable on/off kinetics, these molecules are recognized as indispensable tools for electrophysiology. For our purposes, such toxins serve as blueprints for the design of NaV-selective imaging agents (Strichartz et al., 1995). Herein, we describe the synthesis and characterization of fluorescent small molecules patterned after the paralytic shellfish poison, (+)-saxitoxin (STX) (Andresen and Du Bois, 2009). These reagents enable real-time imaging of native NaVs in live cells at the single-molecule level with subdiffraction localization (Lord et al., 2010). The unprecedented spatial and temporal information provided by super-resolution and single particle tracking analysis of these STX-based fluorescent dyes underscores their utility as tools for NaV study.
Our efforts to develop a preparative route to (+)-STX have furnished an efficient and flexible synthesis of the natural toxin and structural variants, the latter of which are inaccessible by other means. This work has demonstrated that the STX receptor site can accommodate changes to the N21 position of STX (Figure 1A) with minimal loss in affinity between ligand and protein (Andresen and Du Bois, 2009). The availability of STX-NH3+ through chemical synthesis offers a convenient starting material from which to append functional groups through mild and selective conjugation strategies (Figure 1B; Figure S1 available online). For our purposes, Cy5 and DCDHF (Lord et al., 2009) fluorophores serve as bright, photostable, long wavelength emitting probes that, when appropriately derivatized, are soluble and chemically stable in aqueous cell media (Lee et al., 2008). Both Cy5 and DCDHF exhibit high total numbers of emitted photons and large fluorescence quantum yields, and possess absorbance and emission wavelengths that avoid cellular autofluorescence. The use of N-hydroxysuccinimide (NHS) ester derivatives Cy5-NHS (Kvach et al., 2008) and DCDHF-NHS (Wang, 2007) enables “post-synthetic” coupling reactions with STX-NH3+ to furnish STX-fluorophore conjugates STX-Cy5 and STX-DCDHF (STX-Cy5, λabs/em = 655/665 nm; STX-DCDHF, λabs/em = 580/640 nm in H2O; Figure S2A). Purification by reverse-phase HPLC affords the corresponding trifluoroacetate salts of these compounds as water soluble blue and violet powders, respectively, which can be stored for several months at −20°C when shielded from light.
In order to assess the potency and binding kinetics of STX-based fluorescent probes, we have conducted whole-cell voltageclamp electrophysiological studies with rat pheochromocytoma 12 (PC12) cells. This cell type finds extensive use as a model for neuronal differentiation (Vaudry et al., 2002). Exposure of PC12 cells to nerve growth factor (NGF) over a period of days induces expression of STX-sensitive isoforms NaV1.7 (Toledo-Aral et al., 1997) and NaV1.2 (Mandel et al., 1988), and engenders characteristics that include neurite extension, suspended proliferation, and gated ion current (Greene and Tischler, 1976).
Prior efforts to measure the affinity of STX toward endogenous NaVs in NGF-treated PC12 cells have employed 3H-STX to determine an apparent Kd in the low nanomolar range (3.0 ± 0.5 nM) (Rudy et al., 1982). Consistent with the nanomolar Kd for STX, whole-cell voltage-clamp recordings obtained after 4–7 day treatment with NGF give an IC50 for STX of 2.1 ± 0.2 nM (Figure 1C). Significantly, IC50 values of 40 ± 3 and 106 ± 23 nM determined for STX-Cy5 and STX-DCDHF, respectively, are within 1.5 orders of magnitude of the parent toxin. These results are commensurate with an STX binding model that orients N21 toward extracellular space (Andresen and Du Bois, 2009). As with the parent toxin, block by both STX conjugates is fully reversible, and peak Na+ ion current is restored upon perfusion of the patched cell with buffer solution (Figure S3).
We have evaluated the specificity of STX probe binding to NaVs in NGF-differentiated PC12 cells through NaV antibody colocalization experiments. Confocal images were obtained on cells exposed to phosphate buffered saline solutions of 10 nM STX-Cy5 or 15 nM STX-DCDHF. Overlap analysis following cell fixation, membrane permeabilization, and sequential staining with a pan-NaV antibody/Alexa488-conjugated secondary antibody and STX-Cy5 or STX-DCDHF (Figures 2G–2L) yields Manders coefficients of 0.94 ± 0.02 and 0.93 ± 0.08, respectively. The excellent correspondence of these dyes with Alexa488-conjugated antibody staining offers compelling support that modified STXs selectively target NaV proteins (see also Supplemental Experimental Procedures).
Preblocking experiments with saturating concentrations of STX (10 µM) give additional evidence for NaV labeling by STX-Cy5 and STX-DCDHF. Cells treated with STX prior to application of either dye conjugate show low levels of fluorescence at the extracellular membrane, as determined by confocal imaging (Figures 2A, 2B and S4A). These results are consistent with STX occupancy of the NaV outer pore precluding association of the fluorescent STX probes.
As a final control to assess the specificity of NaV labeling by STX-Cy5 and STX-DCDHF, we have performed both confocal and wide-field single-molecule imaging experiments on Chinese hamster ovary (CHO) cells transfected with a plasmid encoding for the α -subunit of rat NaV1.4 (rNaV1.4). Previous work in our group has established the ability of modified STX derivatives to effect channel block with this construct, which shows 70% sequence homology to NaV1.2 and 71% sequence homology to NaV1.7 expressed in PC12; all three rat isoforms share a similar p-loop sequence that defines the STX binding site (Figure S5C). Both ensemble and single-molecule images of native CHO cells in the presence of STX-Cy5 or STX-DCDHF show no fluorescence emission from the cytoplasmic membrane or from within the cell. By contrast, membrane localization of the dyes is readily evident in CHO cells transfected with rNaV1.4 (Figures S5A and S5B; see also Supplemental Experimental Procedures). These data provide strong corroborating support for the selective association of STX-dye conjugates with NaVs.
The cellular distribution of NaVs in live cells has been explored indirectly through voltage-clamp measurements. These data show increased sodium ion current density in excised neurites relative to the cell body of intact cells (O’Lague et al., 1985). Corresponding immunocytology experiments to mark the location of NaVs in NGF-differentiated PC12 cells appear in a single report (Toledo-Aral et al., 1997). Such images are obtained with a specific antibody raised against NGF-induced NaV1.7 and reveal upregulated channel populations in neurite growth cones and in the membrane of the cell soma. In accord with these findings, confocal images obtained with STX-Cy5 and STX-DCDHF against NGF-treated PC12 cells highlight specific labeling of NaVs at neuritic processes and in the somatic membrane (Figures 2G–2L). Regions of enhanced fluorescence intensity are focused at growth cones and in the periphery of the cell body, data that concurs with recent reports describing clustering of NaVs in unmyelinated neurons (England et al., 1990; Henry et al., 2006, Zeng and Tang, 2009).
In confocal experiments performed with sub-IC50 concentrations of STX-Cy5 and STX-DCDHF (10 and 15 nM, respectively), membrane-associated STX-fluorophores are clearly visible above the background emission of the unbound probe. This unique brightness enhancement is especially pronounced with STX-DCDHF (Lord et al., 2009) (Figure S2B). Z-stack cross sections acquired at multiple time points over 15 min to 6 hr reveal that fluorescence is restricted to the cell surface, a finding that demonstrates STX-Cy5 and STX-DCDHF neither passively diffuse across the cell membrane nor activate internalization pathways (Figure S4D). When cells are incubated in media containing STX-fluorophore, imaging experiments can be performed in an incubator-equipped microscope over several days with no observable deleterious effect on cell morphology or viability. Consistent with a reversible ligand-protein binding interaction, the fluorescence signal from either probe is effectively washed-off the cell by replacement of the medium (Figures 2C, 2D, and S4B). The time course for fluorescence intensity decay in these experiments is within the same order of magnitude as the off-rate kinetic measurements from electrophysiology recordings (Figures S3 and S6).
Transcriptional profiling and electrophysiology data suggest that a measurable level of NaVs exists in the plasma membrane of undifferentiated PC12 cells (Mandel et al., 1988). To query the effect of NGF treatment on NaV expression, we have compared the integrated fluorescence intensity of STX-Cy5 and STX-DCDHF in identical populations of PC12 cells cultured with and without NGF (4–7 days). Undifferentiated cells exposed to 10 nM STX-Cy5 or 15 nM STX-DCDHF show uniform fluorescence emission at the outer membrane with average intensities of ~20% relative to NGF-treated cells (Figures 2E, 2F, and S4C). These findings are in agreement with NaV occurrence in undifferentiated cells and reflect the substantial upregulation in protein expression caused by NGF (Ginty et al., 1992; Toledo-Aral et al., 1995; Hilborn et al., 1998). The ability to detect small basal populations of NaVs in live cells could help further our understanding of how posttranslational protein modification is used to regulate neuronal cellular plasticity (Cantrell and Catterall, 2001).
In principle, single-molecule tracking analysis can differentiate among physical processes associated with NaV behavior, including membrane diffusion, cytoskeleton-associated transport, and endocytotic turnover (Matsuoka et al., 2009; Kusumi et al., 2010). Such experiments can also distinguish between populations of NaVs that reside permanently in a distinct state or represent dynamic transitions of a single population—information unavailable through ensemble imaging (Vrljic et al., 2009). We have employed STX-Cy5 and STX-DCDHF in single particle tracking experiments and instantaneous velocity measurements to characterize the motion of individual NaVs in NGF-differentiated PC12 cells. The availability of fluorescent markers that preserve the motility of NaVs is paramount for this type of study, favoring the use of small molecule dyes over fluorescent fusion protein constructs. We have capitalized on the longer residence binding time of STX-DCDHF relative to STX-Cy5 in the plasma membrane of PC12 cells to examine single-molecule NaV tracks on the somatic surface (vide infra).
Single 50 ms epifluorescence imaging frames of differentiated PC12 cells obtained in the presence of STX-DCDHF clearly show individual NaVs on the somatic membrane. Throughout this region, small clusters of NaVs are also observed (Figure 3B). Illustrative tracks of single molecules plotted versus time reveal heterogeneous trajectories that are characteristic of confined diffusion, Brownian motion, and directed transport (Qian et al., 1991) (Figure 3D). Although exhaustive analysis of the modes of motion of NaV is beyond the scope of this report, for the purposes of a preliminary characterization, ~4,000 single NaVs in the somatic membrane were tracked using 12.8 nM STX-DCDHF at low imaging intensity to reduce photobleaching. As revealed by these data, labeled NaVs in the membrane of the soma exhibit an average visibility time of 250 ±7 ms and display multiple trajectories before disappearing. Selection of 370 trajectories with 10 steps or more allowed us to compute short-time diffusion coefficients from mean square displacements. The measured D distribution was analyzed under the assumption that all recorded values arise from underlying random Brownian motion characterized by a single diffusion coefficient, Do, which yields a well-known probability distribution (see also Supplemental Experimental Procedures). The expected distribution is designated by the smooth red curve (Figure 3C), with the average diffusion coefficient, Davg = 0.036 ± 0.027 µm2/s (N = 370) taken as Do. Departures from the smooth curve indicate heterogeneous motion, as observed (Figure 3C). The measured diffusion coefficients are comparable to reported values for NaVs restricted to specific membrane regions in neurons (Leterrier et al., 2010; Nakada et al., 2003; Brachet et al., 2010).
Analysis of the trajectories of single NaVs in neuritic projections of differentiated PC12 cells at 100 ms temporal resolution is made possible by capitalizing on the prolonged visibility time of STX-DCDHF at the cell membrane (Figure 3E). Tracking instantaneous velocity of five individually resolved NaVs undergoing directed motion affords lower and upper speed limits of 680 and 970 ± 10 nm/s (59 mm/day and 85 mm/day), respectively, along the length of the neurite. These values are consistent with reported rates of intracellular axonal transport of certain neuronal proteins and vesicles (Brown, 2003).
The ability to spatially resolve single molecules in the dynamic environment of a living cell permits real-time tracking of the distribution of molecules over a small region of interest. In the cellular membrane, where the motion of labeled proteins is restricted to two dimensions, changing fluorescence patterns can be used to trace morphological fluctuations in these structures. Here, rapid photobleaching of NaV bound STX-Cy5 affords a visibility time of 100 ± 10 ms (Figures 4I and Figure S8A; see also Supplemental Experimental Procedures). This short visibility time is well suited for rapid stochastic sampling of single NaV locations within filopodia of differentiated PC12 cells (Figures 4C–4G). Locations of individual NaVs can be summed into a single image by maxplotting the data, a method in which the highest intensity value of an individual pixel over a fixed time interval is taken to be the value of that pixel (Figures 4C–4G and S7A). In this manner, NaV spatial distribution can be used to outline growth cone motility of cellular filopodia at a temporal resolution of 10 s. Analysis over a 24 min period reveals dynamic growth and retraction of filopodia at a maximum extension/retraction rate of 360 nm/s with an average value of 80 ± 90 nm/s (Figures 4H and Figure S7C). Notably, the observed rates are commensurate with previously determined traction rates measured in neurons (Mallavarapu and Mitchison, 1999; Lu et al., 1997; Sheetz et al., 1992; Kress et al., 2007; Costantino et al., 2008; Laishram et al., 2009). These rapid, noncontinuous motions are reminiscent of the sliding of actin filaments (Mallavarapu and Mitchison, 1999) and suggest that analogous cytoskeleton associated mechanisms may play a significant role in the trafficking of filopodial NaVs.
The small dimensions of neuritic spines and filopodia in NGF-differentiated PC12 cells render the study of protein dynamics intractable by conventional diffraction-limited confocal or epifluorescence microscopy. Super-resolution imaging of individual proteins thus has the potential to reveal fine details of NaV locations within these intricate substructures, information that may give deeper insight into integrated channel function. Pointillist super-resolution reconstruction relies on efficient turn-on/turn-off fluorescence in order to maintain sparse populations of emitters from NaVs within the region of interest. While both saxitoxins afford efficient on/off fluorescence necessary for super-resolution imaging, STX-Cy5 exhibits a shorter visibility time and is particularly well suited for such experiments. Photophysical analysis reveals that turn-off of the STX-Cy5 fluorescence signal is primarily caused by intensity-dependent photobleaching of the dye molecule (Figures S8D and S8E). By carefully selecting pump laser intensities and exposure times, background emission of the unbound probe is negligible and single NaVs blocked by the STX probes are clearly visible (signal/ background STX-Cy5 = 1.8 ± 0.4, STX-DCDHF = 4.2 ± 0.5; see also Supplemental Experimental Procedures). With optimized imaging parameters, experiments can be performed with a fixed concentration of either STX-fluorophore in the cell medium, thus allowing continuous turn-on events to occur with the binding of new dye molecules (Figures S8A–S8C). As such, these experiments can be performed in the absence of thiol reductants or photostimulation. We view this method as a target-specific variation of points accumulation for imaging in nanoscale topography (PAINT) (Sharonov and Hochstrasser, 2006), which provides a sampling of NaV locations within these morphological structures.
Harnessing the short visibility time of STX-Cy5 enables the recording of multiple single NaV positions to yield pointillist super-resolution reconstruction with a mean statistical localization precision of 25 ± 3 nm (Figures S8F and S8G). Under these conditions, an imaging time of 5 s, averaging 5,300 ± 200 localizations/s, is needed to accumulate sufficient data to create a well defined ~300 µm2 image of the neuritic structure (Figure 5A). As each single-molecule localization event is represented by a Gaussian with fixed width and amplitude, this data also reflects the relative densities of NaVs in different cellular regions. The reconstructed images show punctate populations of NaVs at nodes along the neurite shaft and at the origins of numerous projections extending from the neurite body. Neuritic spines, ~85 nm in width, are perceptible as discrete structures containing NaVs distributed both individually and in small clusters (Figures 5D, 5F, and 5J). A quasi-real-time super-resolution movie generated using a weighted moving average algorithm reveals that these neuritic outgrowths exhibit highly fluxional lateral movements, extensions, and retractions on the timescale of hundreds of milliseconds (Figure 5L; Movie S1). These rates, on par with dendritic spine motility in neurons (Bhatt et al., 2009), enable us to visualize at nanometer precision and in quasi-real-time structural changes and specific NaV distributions in motile neuritic spines.
An intricate array of regulatory networks and dynamic trafficking mechanisms control the patterning of NaV subtypes and shape the specific conductive properties of a cell. Despite the importance of NaVs in neuronal function, methods for live cell imaging of these proteins are conspicuously few. Saxitoxin offers a unique molecular template for the design of new pharmacological and imaging agents to interrogate dynamic channel function. The singular power of chemical synthesis has made possible access to fluorescently-labeled forms of STX that maintain high affinity to the channel and reversible binding kinetics. We have described the electrophysiological characterization of these toxin conjugates and have demonstrated their utility for live-cell imaging of NaVs in NGF-treated PC12 cells. Our results further highlight the potential for STX-Cy5 and STX-DCDHF to detect small populations of NaVs in undifferentiated PC12 cells. The unique properties of these probes have been underscored in single-molecule imaging and super-resolution experiments. Such data provide analysis of NaV motility in functionally distinct regions of the excitable cell membrane and quasi-real-time super-resolution visualization of changing NaV populations in neuritic spines. These studies form part of a larger program dedicated to elucidating dynamic processes associated with NaV function and their integral role in bioelectrical conduction.
Saxitoxin derivatives are expected to show similar toxicity and symptomatology to saxitoxin, tetrodotoxin, and other sodium channel blockers and should be handled with appropriate care for safety.
To a solution of STX-NH3+ (6.6 mg, 6.4 µmol) in 635 µl of a 2:1 mixture of MeCN/pH 9.5 buffer (0.1 M aqueous NaHCO3/Na2CO3) was added Cy5-NHS (5.8 mg, 9.4 µmol, 1.5 equivalent) (Figure 1B). The mixture was stirred at room temperature for 2 hr, then acidified with 3.0 ml of 0.10% aqueous CF3CO2H. After concentrating this solution under reduced pressure, the solid material was purified by reverse phase HPLC (Altima C18, 10 µm, 10 × 250 mm column, eluting with a gradient flow over 30 min of 20:80 MeCN/ H2O with 0.1% CF3CO2H → 50:50 MeCN/H20 with 0.1% CF3CO2H, 254 nm ultraviolet [UV] detection). At a flow rate of 6 ml/min, STX-Cy5 had a retention time of 26.0 min and was isolated following lyophilization as a dark blue solid (2.7 mg, 36%).
To a solution of STX-NH3+(3.1 mg, 3.0 µmol) in 150 µl of a 3:1 mixture of MeCN/ pH 9.5 buffer (0.1 M aqueous NaHCO3/Na2CO3) was added DCDHF-NHS (3.0 mg, 5.9 µmol, 2.0 equivalent) (Figure 1B). The reaction was stirred at room temperature for 3 hr, then acidified with 40 µl of 1.0 N aqueous HCl. After concentrating this solution under reduced pressure, the solid material was purified by reverse phase HPLC (Altima C18, 10 µm, 10 × 250 mm column, eluting with a gradient flow over 30 min of 20:80 MeCN/H2O with 0.1% CF3CO2H → 80:20 MeCN/H20 with 0.1% CF3CO2H, 254 nm UV detection). At a flow rate of 6 ml/min, STX-DCDHF had a retention time of 14.6 min and was isolated following lyophilization a dark violet solid (1.3 mg, 44%).
PC12 (pheochromocytoma 12) cells were grown in DMEM phenol red-free media (GIBCO BRL, Grand Island, NY) supplemented with 10% horse serum (GIBCO BRL), 5% fetal calf serum (HyClone, Logan, UT), and 100 U/ml pen-strep (GIBCO BRL). Cells were grown in a 5% carbon dioxide, 96% relative humidity incubator at 37°C, and passaged every ~3 days. Passaging of cells was accomplished by careful aspiration of media, treatment with 1 ml of trypsin-EDTA (0.05%, Invitrogen, Carlsbad, CA) for ~5 min until full dissociation of cells from the plate surface was observed, and dilution with 4 ml of growth medium. Approximately 1 ml of this suspension was then diluted in 9 ml growth medium in a new 10 cm plate. For electrophysiological recordings or microscopy experiments involving differentiated PC12 cells, 100 ng/ml of nerve growth factor (NGF; BD Biosciences, Bedford, MA) was added 1–7 days prior to data collection. Media was replaced every 3 days.
CHO cell culture was performed as described previously (Andresen and Du Bois, 2009). Transfection with rat NaV1.4 (rNaV1.4) and eGFP was accomplished using the calcium phosphate precipitation method or the Lipofectamine method.
Sodium currents were measured using the patch-clamp technique in the whole-cell configuration with an Axopatch-200b amplifier (Axon Instruments, Union City, CA). Borosilicate glass micropipettes (Sutter Instruments, Novato, CA) were fire-polished to a tip diameter yielding a resistance of 1.0–3.0 MΩ in the working solutions. Pipettes were filled with (in mM): CsCl 140, KCl 1, EGTA 1, HEPES 10, and the pH was adjusted to 7.2 with solid CsOH (D’Arcangelo et al., 1993). The external solution had the following composition (in mM): NaCl 140, CaCl2 1, HEPES 10, and the pH was adjusted to 7.4 with solid CsOH.
The output of the patch-clamp amplifier was filtered with a built-in low-pass, four-pole Bessel filter having a cutoff frequency of 10 kHz and sampled at 100 kHz. The membrane was kept at a holding potential of −100 mV. Pulse stimulation and data acquisition used 16 bit D–A and A–D converters (Axon Instruments Digidata 1322A) controlled with the PClamp software (Axon Instruments). Leak currents were subtracted using a standard P/4 protocol of the same polarity. Access resistance was always <4 MΩ and the cell capacitance was between 4 and 20 pF, as measured by the compensating circuit of the amplifier. All measurements were done at room temperature (20–22°C). Recordings were made at least 5 min after establishing the whole-cell and voltage-clamp configuration to allow for stabilization of the voltage-dependent properties of the channels. Currents were elicited by 10 ms step depolarizations from a holding potential of −100 to 0 mV.
Stock solutions of STX-Cy5 and STX-DCDHF were prepared by serial dilution with external solution and stored at 4°C in the dark. PC12 cells were grown and differentiated on glass coverslips (5 mm #1 thickness, Warner Instruments, Hamden, CT) placed on the bottom surface of a culture dish and coverslips were removed just prior to recording. Current measurements were recorded under continuous perfusion, controlled manually by syringe addition.
Data were fitted to Langmuir isotherms according to the equation:
where I / Io is the fraction of remaining current, [toxin] is the concentration of STX or STX-fluorophore, and IC50 is the concentration of STX or STX-fluorophore that effects 50% current block. Custom software developed in the Igor Pro environment (Wavemetrics) was used to determine IC50 values expressed as mean ± 1 SD.
Confocal images were acquired at the Cell Sciences Imaging Facility (Stanford University) on a Zeiss LSM 510 Confocal Laser Scanning Module mounted on a Zeiss Axiovert 200M inverted microscope using either a LCI Plan-Neofluar 25 × /0.8 l mm Korr DIC objective or an A-Plan-Apochromat 63 × /1.4 Oil DIC objective. Cy5 and DCDHF chromophores were excited at 633 nm with a 5 mW HeNe laser at 10%–15% output; emission was detected through a HFT UV/488/543/633 dichroic beam splitter with a 650 nm long-pass filter. Alexa488 was excited at 488 nm with a 25 mW Argon laser at 10% output, and emission was detected through a HFT 488 dichroic beam splitter with a 500–550 nm IR bandpass filter.
PC12 cells were plated on glass bottom dishes (MatTek Corp, Ashland, MA) and induced to differentiate with NGF (unless otherwise noted) for 1–7 days. Prior to data collection, cells were washed with PBS and incubated in media consisting of phenol red-free DMEM containing 10% CCS, 1:100 v/v pen-strep, and 100 ng/ml NGF as indicated. Stock solutions of STX-Cy5 or STX-DCDHF in <100 µl PBS were added to final concentrations of 10 nM or 15 nM, respectively.
Both transmission images of the cells and epifluorescence images of single molecules were acquired using an inverted microscope (IX71, Olympus, Center Valley, PA). Bright field illumination from a condenser allowed for visualization of the edges of the cells. Fluorescence imaging of the cells was performed with wide-field epi-illumination in an area of ~42 µm × 42 µm. Because the mechanical drift of the stage (Semprex, Campbell, CA) was negligible during the recording time a fiduciary correction was not necessary. Laser illumination at 633 nm afforded an intensity of ~200 W/cm2 at the sample plane for Cy5 chromophore imaging. Laser illumination at 594 nm provided an intensity of ~40 W/cm2 at the sample plane for DCDHF chromophore imaging. Epifluorescence images were collected with a 60× magnification, 1.45 NA, oil-immersion objective (PlanApo, Olympus, Center Valley, PA) and imaged through a 640 nm long-pass filter and a 635 nm dichroic mirror (Omega Optical, Brattleboro, VT) for images acquired with 633 nm excitation, and a 595 dichroic mirror and 615 long-pass filter were used for images acquired with 594 nm excitation. An EMCCD-camera (IXonDV887, Andor, South Windsor, CT) was used for data collection. The pumping beam intensity was adjusted with neutral density filters to give an acceptable signal-to-background ratio while minimizing fluorophore bleaching. During data collection samples were housed in a Tokai Hit thermo incubator (Tokai Hit, Japan) maintained at 37°C with 5% CO2 that was fixed to a Semprex precision stage using a custom machined adaptor.
Borosilicate coverglass coated with collagen (Nalge Nunc International, Naperville, IL) was treated with 50 µg/ml fibronectin (human plasma, Calbiochem, San Diego, CA) in PBS at pH 7.4 for 1 hr at room temperature prior to deposition of cells. PC12 cells were plated and induced to differentiate with NGF for 1–7 days. Prior to imaging, cells were washed with PBS and placed in phenol red-free DMEM containing 100 ng/ml NGF and STX-DCDHF (13 nM) or STX-Cy5 (3 nM or 10 nM).
Trajectories of single conjugates (Figure 3) were extracted using a Single Particle Tracking program from ETH-Zurich. Maxplot images (Figure 4) were obtained from raw image stacks using ImageJ function Z Project with maximum intensity. Super-resolution reconstructions (Figure 5) were made possible by using super-localization fitting to determine subdiffraction single-molecule locations (Biteen et al., 2008) and final reconstructions utilized a custom macro written in ImageJ, where each single molecule was represented as a two-dimensional Gaussian having a constant amplitude and σ (SD) equal to the average statistical error of the center of the fit, i.e., 30 nm.
A.E.O., H.L., S.I., B.M.A., W.E.M., and J.D. designed research; A.E.O., H.L., S.I., and W.H.P. performed research; A.E.O., H.L., and S.I. analyzed data; and A.E.O., H.L., W.E.M., and J.D. wrote the paper. The authors thank Professor R.J. Twieg and J.C. Williams for the gift of DCDHF-NHS. We are grateful to Professor Merritt Maduke for allowing use of her electrophysiology equipment and for many helpful discussions. This work was supported in part by R01-GM086196 (to W.E.M.) and R01-NS45684 (to J.D.) from the National Institute of General Medical Sciences and the National Institute of Neurological Disorders and Stroke, respectively, and by a grant from the Tobacco-Related Disease Research Program (to J.D.). W.H.P. is the recipient of a Stanford Interdisciplinary Graduate Fellowship. J.D. is a cofounder of SiteOne Therapeutics, Inc.
Supplemental Information includes eight figures, one movie, and Supplemental Experimental Procedures and can be found with this article online at http://dx.doi.org/10.1016/j.chembiol.2012.05.021.