Lithographic Microfabrication of a 16-Electrode Array on a Probe Tip

Jan 15, 2016 - We report the lithographic microfabrication of a movable thin film microelectrode array (MEA) probe consisting of 16 platinum band elec...
0 downloads 11 Views 2MB Size
Article pubs.acs.org/ac

Lithographic Microfabrication of a 16-Electrode Array on a Probe Tip for High Spatial Resolution Electrochemical Localization of Exocytosis Joakim Wigström, Johan Dunevall, Neda Najafinobar, Jelena Lovrić, Jun Wang, Andrew G. Ewing, and Ann-Sofie Cans* Department of Chemistry and Chemical Engineering, Chalmers University of Technology, SE-412 96 Gothenburg, Sweden S Supporting Information *

ABSTRACT: We report the lithographic microfabrication of a movable thin film microelectrode array (MEA) probe consisting of 16 platinum band electrodes placed on top of a supporting borosilicate glass substrate. These 1.2 μm wide electrodes were tightly packed and positioned parallel in two opposite rows within a 20 μm × 25 μm square area and with a distance less than 10 μm from the edge of the glass substrate. We demonstrate the ability to control and place the probe in close proximity to the surface of adherent bovine chromaffin cells and to amperometrically record single exocytosis release events with high spatiotemporal resolution. The two-dimensional position of single exocytotic events occurring in the center gap area separating the two rows of MEA band electrodes and that were codetected by electrodes in both rows was determined by analysis of the fractional detection of catecholamine released between electrodes and exploiting random walk simulations. Hence, two-dimensional electrochemical imaging recording of exocytosis release between the electrodes within this area was achieved. Similarly, by modeling the current spikes codetected by parallel adjacent band electrodes positioned in the same electrode row, a one-dimensional imaging of exocytosis with submicrometer resolution was accomplished within the area. The one- and twodimensional electrochemical imaging using the MEA probe allowed for high spatial resolution of exocytosis activity and revealed heterogeneous release of catecholamine at the chromaffin cell surface. using carbon fiber disk ultramicroelectrodes (CFEs) during exocytotic activity at single cells results in current spikes from the detection of catecholamines released from individual vesicles and allows the study of the kinetic parameters of single exocytosis events, which have provided knowledge of fundamental mechanisms and modes of neurotransmitter release, fusion pore biophysics, the regulation via exocytotic protein machinery, and the effects of drug on the quantal size.9−13 In order to gain information regarding spatial chemical release at cells using electrochemical methods, several different approaches have been explored. Manually probing different locations of secretion using two separately movable carbon fiber electrodes revealed the presence of active and passive zones of exocytotic activity at bovine adrenal chromaffin cells.14 Further improvement of the electrochemical imaging resolution of secretion activity at cells has been achieved with individually addressable multielectrode arrays (MEAs) fabricated from bundles of up to 7 tightly packed carbon fibers15 or 15 pyrolyzed carbon ring microelectrodes,16 simultaneously measuring release from several locations on a cell.

R

egulated exocytosis is the fundamental process responsible for secretion of a wide range of signaling molecules including hormones, neuropeptides, and small-molecule neurotransmitters involved in neuronal communication processes and endocrine regulation. During exocytosis, neurotransmitter molecules stored in secretory vesicles are released from secretory cells to the extracellular space when a vesicle membrane fuses with the cell plasma membrane through the formation of a fusion pore. A variety of different analytical techniques have been used to provide valuable insights into the dynamics and molecular mechanisms of exocytosis. Fluorescence based techniques such as total internal reflection fluorescence (TIRF) microscopy imaging1,2 and confocal microscopy3 have been used to image the process of individual vesicles undergoing docking and membrane fusion, revealing the heterogeneous nature of the spatially distributed exocytotic sites at endocrine cells4 and offering important clues into the function and regulation of underlying molecular mechanisms. However, the temporal resolution of these techniques prevents studies of processes with very fast kinetics and are challenged by their dependence on successful labeling.5 The application of amperometric measurements and patch clamp based techniques have provided information related to the formation of the vesicle fusion pore, its lifetime, size, and fate with high temporal resolution.6−8 Amperometric recording © XXXX American Chemical Society

Received: August 29, 2015 Accepted: January 15, 2016

A

DOI: 10.1021/acs.analchem.5b03316 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

cells can be approached from the top of the cell with the MEA facing down with a few degree angle tilt by applying micromanipulation (Figure 1A). The MEA electrodes were

Although fabrication of these MEAs offers spatially resolved measurements, there are limitations in the geometry, fabrication success rate, and the number of individual addressable electrodes. In order to overcome these limitations and to fabricate electrodes with higher reproducibility, arrays have been developed using microfabrication. The Lindau group used simultaneously detected current spikes from cells placed at the center area between four thin film platinum electrodes by a patch pipet and was used to assign the positions of individual exocytosis events at bovine chromaffin cells.17 In recent reports, larger numbers of electrodes of MEA devices were demonstrated, where a 9-electrode boron-doped diamond MEA as well as a 16- and a 36-electrode thin film MEA were used to spatially map exocytotic activity at single or clusters of secretory cells cultured on top of the MEA surfaces.18−20 In this paper, we report a new kind of lithographically microfabricated movable thin film MEA probe allowing control of placement in close proximity to adherent single cells in culture. Because of the transparent substrate of this probe, it can easily be placed on top of a cell in a similar manner to experimental procedures for single cell recordings with carbon fiber microelectrodes. This approach avoids the need for manipulation of the cells into close contact with the MEA by a patch pipet, a procedure which may cause cells to behave differently compared to undisturbed cells.17 The MEA probe area is designed to be of similar size to single chromaffin cells and with the ability to simultaneously detect individual exocytotic release events from the 16 different electrodes. By analysis of the charge ratio for amperometric spikes codetected by multiple electrodes and originating from single exocytotic release events, electrochemical imaging of exocytotic release has been achieved with submicrometer resolution for events that can be “triangulated” between electrodes.

Figure 1. (A) Illustration of the MEA device side view (not drawn to scale), where the thin film platinum electrodes are placed to touch and probe a cell. (B) SEM image of a MEA showing the 16 exposed microelectrodes within an open window of the Si3N3 insulation layer. MEA electrodes are indicated by odd numbering (1−15) in the lower and even numbering (2−16) in the upper electrode row. Scale bar is 10 μm. (C) Light microscopy image of a MEA probe placed on top of a chromaffin. Scale bar is 30 μm.

laid out in two rows (Figure 1B), with eight 12.5 μm × 1.5 μm band electrodes in each row, separated by a 1 μm gaps and covering a total area of 20 μm × 25 μm. A contact pad area at the other end of the device was designed to match a 16 pin, gold plated, spring loaded connector (Preci-Dip AC, Delémont, Switzerland), which was pressed against the underside of the MEA device to obtain electrical connection between the electrodes and the potentiostat. MEA Fabrication and Characterization. The 16-channel MEA was fabricated in several process steps by photolithography, thin film deposition, liftoff and reactive ion etching, derived from a previous method described by Wang et al.20 This was followed by a sawing and grinding procedure to separate the devices and position the MEA within a few micrometers from the edge and at the tip of each device. The fabrication process is described in more detail in the Supporting Information. The MEA electrodes were characterized with cyclic voltammetry by performing a voltage scan between the potentials 0 and +0.5 V vs Ag|AgCl reference electrode in a solution of 100 μM FcMeOH in PBS using a 1030B multichannel potentiostat (CH Instruments, Austin, TX). Carbon Fiber Electrodes. Carbon fiber electrodes (CFEs) were prepared as previously described22 with 5 μm diameter carbon fiber inserted into glass capillaries (o.d. 1.2 mm, i.d. 0.69 mm, no filament; Sutter Instrument Co., Novato, CA) which were pulled using a micropipet puller (PE-21, Narishige, Japan) filled with epoxy (EpoTek 301, Billerica, MA) and finally polished at a 45° angle on a commercial micropipet beveller (EG-400, Narishige, Japan) and backfilled with 3 M KCl. All electrodes were tested prior to experiments in 100 μM dopamine solution. Only electrodes with stable voltammograms were used for experiments. Experimental Approach. Single cell experiments with the MEA were performed similarly to the procedure used in CFE single cell experiments8 with some notable differences. Since the MEA electrodes were parallel to the glass substrate, a low



EXPERIMENTAL SECTION Chemicals and Solutions. All chemicals were analytical grade reagents unless otherwise stated and solutions were prepared with water purified with a Purelab Classic purification system (ELGA, Sweden). The physiological saline solution (HBS) was calcium free and contained 10 mM HEPES, 154 mM NaCl, 4.2 mM KCl, 0.7 mM MgCl2, and 11.2 mM Dglucose at pH 7.4. The stimulation solution contained 2 mM BaCl2 and 0.7 mM MgCl2 (without carbonates) in Lockés solution (pH 7.4). The FcMeOH solution was made from a stock solution of 25 mM FcMeOH in acetonitrile, diluted in saline phosphate buffer (PBS) to obtain a solution of 100 μM FcMeOH and 0.4% acetonitrile. Bovine Adrenal Medullary Cell Culture. Bovine chromaffin cells were prepared from fresh adrenal medulla by collagenase digestion. Cells were purified and cultured according to protocols previously described21 with slight modifications (Supporting Information). Briefly, purified chromaffin cells were plated at a density of 10.5 × 103 cells/ cm2 on collagen type IV coated Petri dishes and incubated at 37 °C in a 5% CO2 environment. Cells were used at 37 °C after 1 or 2 days of culture. MEA Design. The 16-channel MEA device was designed as a structure in three layers composed of a borosilicate glass substrate, coated with a conductive thin film platinum pattern (see Figure S1) and a top layer of silicon nitride covering most of the structure and providing insulation except at the probing area. By placing the MEA electrode structure within a few micrometers from one corner of the device, adherent single B

DOI: 10.1021/acs.analchem.5b03316 Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry



Article

RESULTS AND DISCUSSION MEA Probe Characterization. An SEM image of the microfabricated MEA probe is shown in Figure 1B displaying the arrangement of the 16 individual band electrodes that are tightly packed within the insulating Si3N3 layer. The electrodes in each of the two rows had a width of 1.16 ± 0.15 μm (n = 48), placed at 2.5 μm intervals. The distance between the two sets of 8 opposing electrodes was 1.3 ± 0.5 μm. Electrochemical characterization of the MEA Pt electrodes was performed in an aqueous solution of 100 μM FcMeOH by a voltage scan between 0 V and +0.5 V vs a Ag|AgCl reference electrode. A representative cyclic voltammogram from testing the electrode array is shown in Figure S2 in the Supporting Information. Steady state currents of about 80 pA are observed with an onset of oxidative current at about +0.2 V, which is comparable to previously reported current versus voltage scans at Pt multiarray electrodes.20 Electrochemical Detection of Exocytotic Release with 16 Different Electrodes at Single Bovine Chromaffin Cells. The ability of the MEA probe to access individual bovine adrenal chromaffin cells (see Figure 1C) and detect exocytotic release of catecholamine originating from multiple locations is demonstrated in Figure 2A. Here amperometrically recorded traces, detected in parallel by the 16 individually addressed MEA electrodes, show spikes signifying single exocytotic release detetected by each of the numbered electrodes. In these amperometric recordings from bovine chromaffin cells (n = 13), many occurrences of simultaneous detection were observed. Figure 2B shows an excerpt of the amperometric

approach angle (5 deg) had to be used compared to the angle used for CFE experiments (45 deg). To accommodate the low angle of approach, a part of the wall of the Petri dish containing the cells was removed with a wire cutter prior to the experiment. The metal plate supporting the Petri dish on the microscope stage also had a hole cut out to accommodate the space required by the spring loaded connector connecting the MEA (from the underside) to the potentiostat. An Olympus IX-81 microscope (Olympus, Melville, NY) with a 40× objective (Olympus, UApo/340 40×, NA 0.96) was used to monitor the experiments. Olympus SC20 digital color camera interfaced to a personal computer with the cellSens software (Olympus, Hamburg, Germany) was used for visualization of the experiments. Single Cell Amperometry. Amperometric measurements were performed at room temperature and electrodes were kept at +0.7 V vs Ag|AgCl pseudoreference electrode (World Precision Instruments, Inc., Sarasota, FL). The MEA exocytotic recordings at single cells were performed using a Triton+ 48channel patch-clamp amplifier (Tecella, Foothill Ranch, CA). The carbon fiber experiments were performed using an Axopatch 200B patch clamp amplifier (Axon Instruments, Foster City, CA). The signal was displayed in real-time (AxoScope 10.4, Axon Instruments) and stored digitally. The signal was digitized at 10 kHz and filtered with an internal 4pole Bessel low-pass filter at 2 kHz. Stimulation injection pipets were made using a commercial micropipet puller (PE-21, Narishige, Japan), and the opening at the tip was adjusted with a scalpel to around 5−8 μm. All experiments were performed in a Faraday cage. Electrochemical Data Analysis. Initial spike analysis of amperometric data was performed by in-house developed software written in MATLAB (Mathworks Inc., Natick, MA). First baseline fluctuations and offset were subtracted from the recorded trace to obtain a baseline averaged around zero. In the next step, spike candidates were found based on detecting data points exceeding a threshold (5 × RMS noise). Further processing and data analysis related to kinetic parameters, statistical analysis, grouping of spikes, and spatial assignment were performed using in-house written computer programs and the freely available software tools SciPy23 and matplotlib.24 On the basis of previous work, events detected by separate adjacent electrodes with a temporal shift of 10 ms or less were assumed to be simultaneously originating from the same exocytotic release event.18 All exocytotic events detected were sorted in different categories based on the number of electrodes detecting a single exocytotic event and the relative geometrical placement of these electrodes (adjacent vs opposing electrodes) of the MEA probing area. Random Walk Simulations. The method for determination of position for individual exocytotic release events was derived from the method used by the Lindau research group.17 Here a random-walk based computer model based on individual electron micrographs (SEM) of each array were used to determine the position of individual exocytotic release events in the gap space between individually addressable electrodes and based on the experimental recording of the amount of neurotransmitter molecules detected by each electrode. All simulations aiming for spatial assignment of exocytotic release were performed by locally written computer programs, written in MATLAB and Python.

Figure 2. (A) Amperometric recording of quantal release events from a chromaffin cell using the 16-channel MEA stimulated by a pulse of Ba2+-solution. (B) A closeup amperometric trace with red circles grouping exocytotic events codetected by two adjacent electrodes. C

DOI: 10.1021/acs.analchem.5b03316 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Table 1. Summary of Amperometric Spike Parameters from Exocytotic Events Detected at Chromaffin Cells by Different Numbers of MEA Electrodes or by a 5-μm Carbon Fiber Microelectrodea category

n (events)%

Qcell %

Qspike (zmol)

Ipeak (pA)

t1/2 (ms)

one electrode two adjacent electrodes three adjacent electrodes four adjacent electrodes five adjacent electrodes all opposing events all events carbon fiber microelectrode

46. ± 9.3 29 ± 6.1 11 ± 3.5 3.0 ± 3.2 0.6 ± 1.0 11 ± 6.1 100 100

17 ± 7 30 ± 11 21 ± 5 8±8 2±3 23 ± 12 100 100

2927 ± 1276 8404 ± 3516 17203 ± 6984 20358 ± 12294 11597 ± 13695 17996 ± 7411 8350 ± 2855 4340 ± 1700

28.8 ± 17.2 62.8 ± 33.7 106.0 ± 46.8 70.3 ± 43.7 32.5 ± 44.3 100.0 ± 52.8 57.6 ± 25.5 79.3 ± 39.1

16.4 ± 4.6 14.2 ± 4.0 15.0 ± 4.1 15.1 ± 9.2 8.2 ± 9.5 14.2 ± 4.8 15.3 ± 4.0 12.2 ± 3.2

trise (ms) 2.2 1.9 2.1 2.1 1.5 2.1 2.1 1.9

± ± ± ± ± ± ± ±

0.6 0.4 0.6 1.2 2.1 0.6 0.5 0.5

The pooled data is presented as the mean ± standard deviation per cell (13 cells for each electrode). The MEA current spikes are categorized by the number of electrodes (one, two, three, four, or five) and the spatial relationship between these electrodes (adjacent or opposing) detecting a single exocytosis event. All events codetected by electrodes in opposing rows are placed into one group. The fraction of the total number of release events (n%), the fraction of the total amount of molecules detected (Qcell%), the amount of molecules per event (Qspike), the current spike amplitude (Ipeak), the half-width of the current spike (t1/2), and the peak rise time (trise) are listed. The total events per cell were 365 ± 296 and 86 ± 63 for the array and CFE, respectively. a

exocytosis event. Here the spike parameters analyzed from the recordings by the MEA are also compared to spike parameters obtained from amperometric recording at chromaffin cells (n = 13) using a 5 μm carbon fiber electrode (CFE) (n = 13). It is noteworthy that the kinetic parameters of the current spikes of the MEA and the CFE are in good agreement. As each electrode in the array is small, one-electrode events constitute 46% of the total events observed (see Table 1) but only 17% of the total molecules detected. Nearly as many events are those observed at adjacent electrodes (43%) as these represent much of the electrode array surface. The events detected by two or more opposing electrodes represent 11% of the total number of spikes (23% of the total detected amount of molecules). These events at opposing electrodes are likely to originate close to the gap region between the two rows of electrodes, which covers about 10% of a chromaffin cell contour (∼20 μm diameter) during the amperometric recordings. Comparing the average amount of molecules detected from adjacent electrodes, codetecting single vesicle release events (Table 1) shows that with an increasing number of electrodes codetecting a single exocytosis event an average larger amount of molecules is detected. This appears to be a detection threshold issue as more molecules are needed to reach more electrodes (vide infra). In Figure 3, the distribution of the number of molecules detected from single vesicle release events by adjacent and opposing electrodes are presented. Here, the histogram shows

recording trace in Figure 2A and reveals that some single exocytosis events are simultaneously codetected by neighboring electrodes (red circles). In agreement with previously reported results,17,18,25 we attribute those simultaneously detected spikes to single exocytotic events originating in the gap regions between the electrodes. Following the method of Gosso et al.,18 spikes codetected at neighboring electrodes with a temporal shift of less than 10 ms were considered simultaneous and thus ascribed to the same exocytotic event. All the events recorded were thereafter sorted and categorized by the number of electrodes detecting each event. In this study, the exocytotic events detected by one, two, three, or four electrodes are categorized as one-, two-, three-, four-, and five-electrode detection events, respectively. In addition, the exocytotic release events were labeled as adjacent if all the detecting electrodes were located adjacently in the same electrode row or opposing if any of the detecting electrodes were located in the opposite row of electrodes (see Figure S3). The number of exocytosis events detected at the chromaffin cells was on average 365 ± 296. After sorting the events in terms of number of electrodes detecting each single exocytotic event, the fraction of the total number of exocytosis events detected was determined for each category (n%). Thereafter characteristics of the amperometric current spikes were analyzed. The statistical data from the analysis of current spike parameters are summarized with the average values for each category of the event in Table 1. In this analysis, the time integral of each current spike generating the total charge detected was used to calculate the number of molecules released per exocytosis event (Qspike). The total amount of molecules for each type of detection event is the sum of the amount neurotransmitters all electrodes together are codetected from a single exocytosis event. From these quantification values, the fraction of neurotransmitter release detected by each electrode category was compared to the total amount of catecholamine released per cell (Qcell%). The maximum amplitude of the current spike (Ipeak), the half-width of the spike (t1/2), and the spike rise time from 25% to 75% of the spike amplitude (trise) were analyzed to achieve information on the kinetic parameters for the exocytosis events. The kinetic parameters have been determined from the spikes recorded by the electrode detecting the largest contribution of catecholamine released, which is likely to be the electrode closest to the

Figure 3. Number of exocytosis events detected by one (blue), two adjacent (red) or three adjacent (green) or opposing (black) electrodes versus the number of molecules detected per exocytosis event are presented in pooled data from recordings at chromaffin cells (n = 13). D

DOI: 10.1021/acs.analchem.5b03316 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

number of molecules for each electrode was recorded as the fraction of the total number of released molecules. These fractions are unitless and can therefore be compared to experimental data expressed as either fractional number of molecules or fractional charges detected, by multiple codetecting electrodes. It should be noted that the fractional charges detected in simulated exocytotic release are independent of any particular diffusional coefficients. Initial simulations focused on detection of exocytosis events detected by two adjacent electrodes (see electrode “a” and “b” in Figure 4A) along a horizontal line “x1” between these electrodes. In this model the relative charge detected was predicted at the various positions between the two parallel bands of electrodes as illustrated in Figure 4B. Here the fraction of number of molecules detected by each of the two electrodes is shown to gradually shift depending on the spatial position of exocytotic release, favoring detection by the closest electrode. Hence, the dependency of fractional detection can be used to determine the position for exocytosis events along the distance “x1”. However, as shown in Figure 4B the fractional detection is also dependent on the cell-to-MEA distance as simulated for distances between 200 and 1000 nm, with 200 nm intervals. Thus, determining the distance between the cell and the MEA is crucial for an accurate assignment of the position for the originating exocytosis event. Complementary simulations (data not shown) at positions closer to the short end of the electrodes (see x2 and x3 in Figure 4A) show that the position along the axis of parallel electrodes does not have a significant effect on the detected ratios and therefore the entire gap region between two parallel electrodes can be considered as a onedimensional system. The optimal detection efficiency for an electrode is achieved for exocytotic release that occurs directly under the center of the electrode surface (Figure 4B) and is also affected by the cell-to-MEA distance. At short distances (h < 200 nm) the detection efficiency of the catecholamine release is close to 100% for an electrode placed directly above a release event. However, increasing the distance between the cell surface and the electrode leads to a larger fraction of catecholamine released escaping detection and results in a reduced amount of secreted substance being detected by that electrode. This distance dependency of fractional detection by neighboring electrodes for single vesicle release events was utilized in our analysis to determine the distance between the surface of the cell and the MEA. The fraction of number of molecules detected by an electrode closest to an exocytotic release event measured never exceeded ∼95% when codetected by two adjacent electrodes (see Figure S4). For exocytotic release events above a certain value, this limit is independent of the total detected number of molecules for the single release events and is thereby unrelated to the detection limit. Comparing this fraction to the maximum detection in the model displayed in Figure 4B suggests a distance between the cell and the MEA to be in the range of ∼500 nm. This estimated distance is ∼30 times larger than the estimated gaps of 15−20 nm separating chromaffin cells in tissue,27,28 and about 2 times larger than reported cellto-electrode distances in amperometric recordings at PC12 cells26,29 or at an artificial cell model for exocytosis.30 Two-Dimensional Electrochemical Imaging of Exocytotic Release. For exocytotic release that originates in the zone, or “window”, near the gap separating the two rows of MEA electrodes, electrochemical imaging in two-dimensions

that detection of a release event by one electrode has a detection threshold of about 60 zmol, which is lower compared to events codetected by two- and three-electrodes. During an exocytotic release event, the catecholamines are released from a pore that is much smaller than the size of both electrodes as well as the gaps separating the electrodes. The secreted catecholamine is distributed over the MEA electrodes, resulting in amperometric current spikes favoring the electrode closest to the exocytosis release site and detects a larger fraction of the molecules released.17,26 Thus, exocytotic release events occurring directly under an electrode are predominantly detected by that electrode, compared to more distant electrodes where a significantly smaller fraction of molecules reach the electrodes and become more difficult to detect when approaching the limits of detection. Events that occur from locations between electrodes are codetected by multiple electrodes. These events are recorded as smaller individual amperometric current spikes and hence result in a higher detection threshold. Using Random Walk Simulations to Estimate the Distance between the Cell Surface and the MEA. The model presented here extends previous work by Hafez et al.17 by also considering the effect of the distance between the cell and the electrode surface. In the experiments performed in the Lindau laboratory, the exact distance between the cell and the MEA was not considered important, due to the much larger distances between the four electrodes, and the distance was set to a fixed value of 100 nm. In the work presented here, the electrode dimensions are considerably smaller (∼1 μm compared to ∼5 μm) and therefore simulations were performed to determine the MEA-to-cell distance. At various MEA−cell distances, simulations of fractional detection of exocytotic release at positions placed with 50−100 nm intervals long the distance between codetecting electrodes was uniquely based on models generated from individual SEM images for each MEA to account for small variations in electrode geometry after fabrication (see Figure 4A). From each position, release of 10 000 neurotransmitter molecules was simulated and the number of molecules reaching each of the 16-channel MEA electrodes was counted. Thus, simulation shows the calculated number of catecholamine molecules codetected by the MEA electrodes for single exocytotic release events. The detected

Figure 4. (A) SEM image of a 16-channel MEA illustrating the distances (x1, x2, and x3) used in the random walk simulations for onedimensional electrochemical imaging and to determine the cell-toMEA distance during amperometric recordings. Scale bar: 10 μm. (B) Simulation of the fractional number of catecholamine molecules of an exocytosis event that occur in the gap region separating two codetecting adjacent electrodes “a” and “b” at various spatial position along the distance x1. The fractional charges detected by electrode “a” (solid line) and electrode “b” (dashed line) are simulated for MEA distances between 200 and 1000 nm away from the cell surface and with 200 nm intermediate distance intervals. E

DOI: 10.1021/acs.analchem.5b03316 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

experiment serve as a basis for the simulations to account for individual geometric variability of the band electrodes and is shown as an overlay in the figure. An additional five representative electrochemical two-dimensional imaging experiments of exocytotic activity are shown in Figure S6. Within the zone for two-dimensional imaging, most events detected were located in the areas with the lowest detection threshold, in general agreement with simulations of detection thresholds (see Figure S5). The two-dimensional electrochemical imaging illustrates not only the location for the exocytosis events but also displays exocytosis release as a function of time, as illustrated in the movie sequence created (Movie S1, Supporting Information) by amperometrically recording exocytosis activity over time following a single pulse of Ba2+ stimulation. Within the window of where the two-dimensional amperometric imaging was performed, the exocytotic release did not appear to be random but showed locations with pronounced exocytotic activity (hot spots) of submicrometer dimensions as well as areas devoid in detected exocytotic activity. One Dimensional Imaging of Single Vesicle Release Events with Nanometer Scale Resolution. As discussed previously, simulations show that detection of exocytotic release events in the gap region between two adjacent electrodes can be approximated with a one-dimensional system, thus supporting one-dimensional electrochemical imaging. The two dimensions of the cell surface are then projected onto a common axis where a position thus corresponds to a line across the surface. Although two-dimensional imaging is preferable from many aspects, the one-dimensional imaging between two electrodes has a lower detection threshold (see Figure 3 and Table 1) and therefore allows observation of smaller exocytotic events. The one-dimensional position of an exocytosis event along the distance “x1” between pairs of adjacent codetecting electrodes was determined by comparing the closest match of experimental data to the predicted fraction of codetection using random walk simulations (as defined in Figure 4A). A cell-toMEA distance of 500 nm was used for the simulations and interpolated values were used for positions along the “x1” line in between simulated points of location. The information gained from one-dimensional imaging of exocytosis activity is summarized in Figure 6. Here a time trail of individual exocytotic events are resolved along the one-dimensional distance between two parallel codetecting electrodes (Figure 6A). By counting and quantifying each release event, the total amount of molecules detected within 50 nm intervals (Figure 6B) and the number of events at these intervals (Figure 6C) are depicted. In each single cell experiment, one-dimensional imaging of exocytotic release is performed by amperometric recording from 14 pairs of adjacent MEA electrodes (n = 13 cells) and where each image created represents exocytosis activity at a fraction of the cell surface. The one-dimensional imaging of exocytosis activity at chromaffin cells showed both active and passive locations as well as a considerable variation in the total number of events at the cell surface (representative plots shown in Figure S7). The hot spots, defined as a length of the imaging axis detecting multiple exocytosis events, were quantified and characterized with regard to their size. Hot spot selection was done manually by choosing length intervals containing spatially clustered multiple exocytosis release events (Figure 6). This manual selection is likely to contain a bias toward small intervals that cluster events. Furthermore, it might favor a selection of space

can be performed (see Figure S5). This is realized by comparing the current spikes recorded and codetecting an exocytosis event by a minimum of three MEA electrodes positioned on both sides of the two electrode rows, based on random walk computer simulations. By finding the closest match between the experimental data collected by the electrodes and simulated detection efficiencies of codetecting electrodes, the position of each exocytotic event can be spatially determined. Simulations were performed uniquely for each MEA using spatial models generated from SEM images where release was simulated at coordinates located in 100 nm intervals, and with a cell-to-MEA distance of 500 nm, based on the estimations made from detection of release by two electrodes. The area near the gap separating two rows of electrodes covers roughly 10% of the upper surface of a chromaffin cell when the MEA is placed centered and in close proximity to the cell surface and hence constitutes a “window” of this MEA probe to carry out two-dimensional electrochemical imaging. As only a partial amount of the neurotransmitters released reaches each electrode during codetection of a single release event, the detection limit is generally higher for exocytosis events detected by multiple electrodes in the gap region between the two rows of electrodes compared to events detected by one single electrode (Figure 3). Thus, it is important to note that the detection threshold for electrochemical imaging of the exocytotic activity might be biased toward larger release events as discussed above. In addition, according to simulations as displayed in Figure S5, the limit of detection varies within the two-dimensional imaging region. Detection by at least 2 electrodes is required for 1D determination of events between the electrodes, which can be done with extremely good spatial resolution, and at least 3 electrodes for spatial imaging in 2D. Therefore, only events that are sufficiently above the detection threshold to be detected at 3 electrodes can be “triangulated.” Thus, sites in between electrodes can be imaged with high resolution, while events under individual electrodes or those where the amount released is low cannot be imaged in practice. Performing two-dimensional electrochemical imaging of exocytotic activity at bovine chromaffin cells, single exocytotic release events can be displayed as a collection of release sites by their position and spatial distribution over time. This type of two-dimensional illustration of exocytotic activity at chromaffin cells stimulated by a single pulse of Ba2+ is presented in Figure 5. Here the SEM image of the MEA electrodes used in this

Figure 5. Overlay of an SEM image illustrating the zone near the two rows of MEA electrodes displaying a two-dimensional electrochemical imaging with dots marking the position for individual amperometrically recorded exocytosis events at a chromaffin cell following a pulse of Ba2+ stimulation. Codetection by a minimum of three electrodes positioned across the two rows is required for two-dimensional imaging, hence only exocytosis events located near the gap separating the two electrode rows may be imaged. Scale bar: 10 μm. F

DOI: 10.1021/acs.analchem.5b03316 Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article



CONCLUSIONS We present the development of a new kind of lithographically microfabricated tightly packed MEA probe, positioned at the tip of a glass substrate, that has the capability to be placed in close proximity to single adherent cells in culture. This allows quantification and imaging of exocytotic release with high temporal and spatial resolution. The array probe approach allows multiple recordings at cells in the same culture. A total of 16 individual simultaneous amperometric recordings of exocytosis release have been carried out at multiple locations at the surface of a cell. In the zone near the gap between two rows of MEA electrodes, corresponding to 10% of the MEA probe surface area, two-dimensional electrochemical imaging of single vesicle release events was achieved. In the remaining surface area of the MEA probe corresponding to the space between pairs of adjacent MEA electrodes, one-dimensional imaging of exocytosis activity was performed. Through analysis of the number of molecules codetected by the electrodes, hot spots in exocytotic activity of submicrometer dimensions were observed. The MEA probe provides another means to gain quantitative information about exocytosis release and has been used to demonstrate the potential of electrochemical imaging of exocytosis activity at the top surface of adherent cells in culture with a spatial resolution comparable to optical methods such as TIRF imaging, which are limited to the basal region of cells.

Figure 6. One-dimensional electrochemical imaging of exocytotis events from a chromaffin cell, codetected by two adjacent electrodes. The exocytotic activity along the distance between the two electrodes is presented as (A) the position of individual exocytosis events as they occur over time, (B) the total amount of molecules detected per event quantified in 50 nm intervals, and (C) the total number of detected events summarized in 50 nm intervals.

with scarce exocytotic activity, since multiple hot spots imaged in two-dimensions on the cell surface might overlap when imaged in one dimension, thus appearing as a single broader spot. From the one-dimensional imaging of exocytosis release at chromaffin cells, a total of 72 hot spots were selected that were composed of between 2 and 22 exocytosis events and quantified to between 2.8 and 240 million molecules (Figure S8). Comparing the number of catecholamine molecules detected from these hot spots with previously reported31,32 average quantification of catecholamine content within a single chromaffin vesicle, corresponding to 4.3 × 106 molecules, suggests that a majority of the recorded hot spots constitute release from multiple vesicles. To determine the size of a selected hot spot, the standard deviation relative to the position of the hot spot center was calculated. Two times the standard deviation, a value containing ∼95% of the events, was used as a measure of hot spot size. The size distribution obtained by this calculation is presented in Figure S9 showing that the measured hot spots size is smaller than 120 nm. As previously mentioned, it is important to realize that this population is biased toward hot spots with smaller spatial distribution and might not represent a general average hot spot size at chromaffin cells. However, this result implies that the resolution for one-dimensional imaging between pairs of adjacent MEA electrodes is comparable with optical methods, with considerably higher temporal resolution, and has the ability to quantify the numbers of molecules released. It is interesting to speculate that two-dimensional imaging can be developed with a spatial resolution in a similar range, given the caveat that exocytosis events can only be imaged when the number of molecules per event is large enough to be detected by three or more opposing electrodes. The results showing hot spots with sizes of submicrometer dimensions supports previous observations by TIRF imaging at INS-12 and chromaffin cells.1 However, in these experiments additional information is offered regarding kinetics and quantitative information on the amounts of catecholamine molecules released, which is important as insight into the regulation and modes of exocytosis.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.5b03316. Materials, methods, and additional experimental data (PDF) Movie sequence illustrating the recorded two-dimensional spatial location of exocytosis release events at a chromaffin cell as the events occur over time (AVI)



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We thank The Swedish Research Council (349-2007-8680), the Knut and Alice Wallenberg Foundation, the European Research Council (Advanced Grant), and the National Institutes of Health for supporting this work. The authors acknowledge Dalsjöfors Kött AB (Dalsjöfors, Sweden) for donation of bovine adrenal glands.



REFERENCES

(1) Allersma, M. W.; Wang, L.; Axelrod, D.; Holz, R. W. Mol. Biol. Cell 2004, 15, 4658−4668. (2) Yuan, T.; Lu, J.; Zhang, J.; Zhang, Y.; Chen, L. Biophys. J. 2015, 108 (2), 251−260. (3) Qian, W.; Kennedy, R. T. Biochem. Biophys. Res. Commun. 2001, 286, 315−321. (4) Koval, L. M.; Yavorskaya, E. N. Neuroscience 2000, 96 (3), 639− 649. (5) Keighron, J. D.; Ewing, A. G.; Cans, A.-S. Analyst 2012, 137 (8), 1755.

G

DOI: 10.1021/acs.analchem.5b03316 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry (6) Borges, R.; Camacho, M.; Gillis, K. D. Acta Physiol. 2008, 192, 173−184. (7) Lindau, M. Biochim. Biophys. Acta, Gen. Subj. 2012, 1820 (8), 1234−1242. (8) Wightman, R. M.; Jankowski, J. A.; Kennedy, R. T.; Kawagoe, K. T.; Schroeder, T. J.; Leszczyszyn, D. J.; Near, J. A.; Diliberto, J. E. J.; Viveros, O. H. Proc. Natl. Acad. Sci. U. S. A. 1991, 88, 10754−10758. (9) Schroeder, T. J.; Borges, R.; Finnegan, J. M.; Pihel, K.; Amatore, C.; Wightman, R. M. Biophys. J. 1996, 70 (2), 1061−1068. (10) Mellander, L. J.; Trouillon, R.; Svensson, M. I.; Ewing, A. G. Sci. Rep. 2012, 2, 907. (11) Staal, R. G. W.; Mosharov, E. V.; Sulzer, D. Nat. Neurosci. 2004, 7 (4), 341−346. (12) Amatore, C.; Oleinick, A. I.; Svir, I. ChemPhysChem 2010, 11, 159−174. (13) Omiatek, D. M.; Dong, Y.; Heien, M. L.; Ewing, A. G. ACS Chem. Neurosci. 2010, 1 (3), 234−245. (14) Schroeder, T. J.; Jankowski, J. A.; Senyshyn, J.; Holz, R. W.; Wightman, R. M. J. Biol. Chem. 1994, 269 (25), 17215−17220. (15) Zhang, B.; Adams, K. L.; Luber, S. J.; Eves, D. J.; Heien, M. L.; Ewing, A. G. Anal. Chem. 2008, 80 (5), 1394−1400. (16) Lin, Y.; Trouillon, R.; Svensson, M.; Keighron, J. D.; Cans, A.; Ewing, A. G. Anal. Chem. 2012, 84 (6), 2949−2954. (17) Hafez, I.; Kisler, K.; Berberian, K.; Dernick, G.; Valero, V.; Yong, M. G.; Craighead, H. G.; Lindau, M. Proc. Natl. Acad. Sci. U. S. A. 2005, 102 (39), 13879−13884. (18) Gosso, S.; Turturici, M.; Franchino, C.; Colombo, E.; Pasquarelli, A.; Carbone, E.; Carabelli, V. J. Physiol. 2014, 592 (15), 3215−3230. (19) Wang, J.; Ewing, A. G. Analyst 2014, 139, 3290−3295. (20) Wang, J.; Trouillon, R.; Lin, Y.; Svensson, M. I.; Ewing, A. G. Anal. Chem. 2013, 85, 5600−5608. (21) O'Connor, D. T.; Mahata, S. K.; Mahata, M.; Jiang, Q.; Hook, V. Y.; Taupenot, L. Nat. Protoc. 2007, 2 (5), 1248−1253. (22) Kelly, R. S.; Wightman, R. M. Anal. Chim. Acta 1986, 187, 79− 87. (23) Jones, E.; Oliphant, T.; Peterson, P.; et al. SciPy: Open Source Scientific Tools for Python, 2001, http://www.scipy.org/. (24) Hunter, J. D. Comput. Sci. Eng. 2007, 9, 90−95. (25) Wang, J.; Trouillon, R.; Dunevall, J.; Ewing, A. G. Anal. Chem. 2014, 86, 4515−4520. (26) Zhang, B.; Heien, M. L. A. V; Santillo, M. F.; Mellander, L. J.; Ewing, A. G. Anal. Chem. 2011, 83 (2), 571−577. (27) Kuramoto, H.; Kondo, H.; Fujita, T. Cell Tissue Res. 1987, 247, 309−315. (28) Coupland, R. E.; Weakley, B. S. J. Anat. 1968, 102, 425−455. (29) Trouillon, R.; Lin, Y.; Mellander, L. J.; Keighron, J. D.; Ewing, A. G. Anal. Chem. 2013, 85, 6421−6428. (30) Cans, A.; Wittenberg, N.; Eves, D.; Karlsson, R.; Karlsson, A.; Orwar, O.; Ewing, A. G. Anal. Chem. 2003, 75, 4165−4175. (31) Koval, L. M.; Yavorskaya, E. N.; Lukyanetz, E. A. Gen. Comp. Endocrinol. 2001, 121, 261−277. (32) Dunevall, J.; Fathali, H.; Najafinobar, N.; Lovric, J.; Wigström, J.; Cans, A.-S.; Ewing, A. G. J. Am. Chem. Soc. 2015, 137, 4344−4346.

H

DOI: 10.1021/acs.analchem.5b03316 Anal. Chem. XXXX, XXX, XXX−XXX