INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

MECHANOSENSITIVE (MS) ion channels have been described in a wide variety of cell types and are implicated in numerous cellular processes (7, 8). Initially, the notion of MS channels as a distinct class of channels and their physiological relevance were challenged based on their ubiquity (11). However, Kung and colleagues (14, 15) recently cloned an MS channel from Escherichia coli, and studies on Caenorhabditis elegans (17) revealed a family of proteins involved in mechanosensation that includes subunits of putative MS ion channels. Therefore, it appears that MS channels do exist and that they may be physiologically relevant to a variety of cellular events related to mechanotransduction.

In aortic baroreceptor neurons (ABN), the mechanotransduction of arterial distension into an electrical signal is probably mediated by MS channels. Electrical activity in ABN has been studied in vivo by recording from single or multiple fibers in the aortic depressor nerve. Electrical activity varies with changes in arterial pressure, and the firing patterns have been well characterized (10).

However, the sensory apparatus and mechanotransduction system of the nerve terminal are not well understood. Sensory nerve endings of ABN are located in the adventitia of the aortic arch, whereas the central axon terminals are located in the nucleus of the solitary tract in the central nervous system. The complex architecture of ABN sensory nerve endings (9) and their relative inaccessibility make the sensory apparatus of the nerve terminal difficult to study at the cellular level using conventional electrophysiological techniques.

Previous studies from this laboratory examined mechanoelectrical transduction in vivo (6) and also demonstrated inward ionic currents and Ca²⁺ transients induced by mechanical stimulation in ABN soma dissociated from the nodose ganglion of the rat (1, 2, 13, 16). Stimulating ABN in culture with a glass probe (13) or by ejecting fluid onto the soma (16) results in a gadolinium-sensitive increase in cytosolic Ca²⁺. We also reported that both hypo-osmotic stretch (1) and fluid ejection onto a neurite (2) activate a whole cell current that is blocked by gadolinium. Gadolinium is a trivalent cation that has been reported to block MS channels (7, 18) and has been used in several preparations to demonstrate the existence of MS channels. In the present study we used cell-attached patch-clamp techniques 1) to determine whether ABN express single MS channels that correspond to the macroscopic currents observed in previous studies, 2) to characterize the voltage sensitivity and conductance of these channels and determine their sensitivity to gadolinium, and 3) to examine the influence of culture conditions on channel expression.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Labeling of nodose baroreceptor neurons. Putative baroreceptor neurons were labeled by applying the dicarbocyanine dye 1,1'-dioleyl-3,3,3',3'-tetramethylindocarbocyanine methanesulfonate (DiI; ⁹-DiI, Molecular Probes, Eugene, OR) to the aortic arch of adult rats and allowing at least 1 wk for transport of the dye back to cell bodies in the nodose ganglion (1, 2, 5, 16). Rats were anesthetized with a mixture of ketamine (91 mg/kg) and acepromazine (1.8 mg/kg ip) and mechanically ventilated with a small animal respirator. With the use of sterile surgical techniques, a right lateral thoracotomy was performed to expose the aortic arch. Approximately 1-3 µl of DiI (50 mg/ml) were injected onto the adventitia of the aortic arch with a glass pipette. The chest was closed, negative intrapleural pressure was reestablished by applying suction through a syringe, and the rat was allowed to recover. Buprenorphine (0.01-0.05 mg/kg sc) was administered for analgesia as needed. Care and use of laboratory animals conformed to standards established by the U.S. Department of Agriculture and by the National Institutes of Health. All protocols were approved by the University of Iowa Animal Care and Use Committee.

Neuronal cell culture. Rats were anesthetized with halothane and decapitated, and both nodose ganglia were removed. Primary cultures of nodose ganglion neurons were prepared by enzymatic and mechanical dissociation as previously described (1, 2, 13, 16) using a protocol adapted from De Koninck et al. (3). Cells were plated on polylysine-coated coverslips and cultured in modified L-15 medium for 3-8 days before electrophysiological experiments. In some cases nodose neurons were either cocultured with nonneuronal cells or placed in transwell cocultures.

Transwell coculture. Commercially available transwell tissue culture plates were used to culture nodose neurons with other types of cells. The transwell culture plates allow cells to share the same medium without cell-to-cell contact. Nodose neurons were plated on polylysine-coated coverslips and placed in the bottom of the transwell. Rabbit aortic vascular smooth muscle (ASM) and/or rabbit aortic endothelial cells (AEC) were cultured on the transwell insert.

Single-channel recording. Cell-attached patch-clamp recordings of single ion channels were obtained from DiI-labeled cells. To stimulate MS channels, negative pressure (suction) was applied to patches using a glass tuberculin syringe attached via tubing to the pipette holder. A mercury manometer was used to measure pressure. Unless otherwise stated, channels were recorded with the membrane patch at the same potential as the cell resting potential. Resting potential was assumed to be 55 mV, which was the average value obtained in whole cell patch-clamp experiments (1, 2). Currents were recorded with an Axopatch 200A amplifier (Axon Instruments), low-pass filtered at 5 kHz, and recorded digitally on videocassette recorder tape. The recordings were analyzed off-line using a 486 IBM-compatible computer with pCLAMP software (Axon Instruments). The composition of the bath solution was (in mM) 120 NaCl, 5 KCl, 2 CaCl₂, 1 MgCl₂, and 10 HEPES; osmolarity was adjusted to 285-295 mosM with 30-40 mM mannitol. The composition of the pipette solution was (in mM) 140 KCl, 5 NaCl, 2 CaCl₂, 1 MgCl₂, and 10 HEPES. For some experiments 1 µM tetrodotoxin (TTX, Sigma, St. Louis, MO), 40 mM tetraethylammonium (TEA), 4 mM 4-aminopyridine (4-AP), 200 nM charybdotoxin, and/or 1 µM -conotoxin GVIA were added to the pipette solution to block voltage-gated currents. The trivalent cation gadolinium (20 µM; Aldrich, Milwaukee, WI), a blocker of mechanosensitive channels, was also added to the pipette solution in some experiments.

Data analysis. The open probability of a single MS ion channel could not be determined directly because the number of channels in each patch was not known. Therefore NP_o was used as a measure of open probability, where N is the number of channels in each patch and P_o is the open probability of a single channel. For each patch, NP_o was calculated in two ways. First, each digitized point in the record was binned to create an all-points histogram of point count as a function of current amplitude. A major peak occurred at the baseline, and a small peak occurred at the current amplitude corresponding to the open channel level. The histogram was fit using a simplex least-squares routine, with the area under the small peak equal to NP_o. Second, NP_o was estimated by generating an events list. Each opening and closing event was identified, and the sum of the open intervals was divided by the total length of the recording period. A patch was excluded from analysis if the two methods of determining NP_o yielded results that differed by >20% or if the patch showed continuous spontaneous activity before application of negative pressure.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Cell-attached patch-clamp recordings demonstrated the presence of discrete opening events that were activated by negative pressure applied through the patch pipette. MS channels were rapidly activated by negative pressure, and they quickly turned off as soon as the negative pressure was released (Fig. 1). Figure 2A shows traces from a patch in which 0-, 10-, 30-, and 50-mmHg negative pressures were applied through the recording electrode. Channel openings became more frequent as the negative pressure was increased. Aggregate data are shown in Fig. 2B, in which NP_o is plotted as a function of applied suction (negative pressure). Open probability increased significantly when pipette suction was 30 mmHg. In all cases, NP_o values calculated using the all-points histograms were slightly higher than events-list estimates. The events-list method requires identification of all opening events and thus is more sensitive to bandwidth limitations. Extremely fast opening events may be missed, leading to underestimation of NP_o.

View larger version (23K): [in this window] [in a new window]   Fig. 1.   Single-channel recording showing rapid activation and turning off of a mechanosensitive (MS) ion channel on application and release of 30-mmHg suction. Traces represent a single continuous recording. Downward deflections are inward currents caused by channel activation.

View larger version (20K): [in this window] [in a new window]   View larger version (14K): [in this window] [in a new window]   Fig. 2.   A: representative single-channel recording showing graded responses to changes in negative pressure. Channel openings become more frequent with increased negative pressure (suction). B: averaged data showing how channel open probability (NPo, where N is number of channels in each patch and Po is open probability of single channel) becomes larger as the suction is increased. For each patch, NPo was calculated from both all-points histogram and events list. Each point is average of 3-20 cells. Error bars represent SE.

To examine the properties of single MS channels, recordings were made at different patch potentials. The current-voltage relationship (Fig. 3) appears to be linear, with a slope conductance of 114 pS. The reversal potential for the MS channel was ~0 mV, and channels were not affected by addition of the K⁺ channel blockers TEA, 4-AP, or charybdotoxin to the pipette solution. These results are consistent with the properties expected of a nonspecific cation conductance. Channel amplitude did not vary with applied suction up to 60 mmHg (not shown).

View larger version (11K): [in this window] [in a new window]   Fig. 3.   Current-voltage relationship obtained by plotting amplitude of single-channel current as function of patch potential. MS channel activity was induced by application of 60-mmHg suction. Data are means ± SE from 6 cells. Slope conductance is 114 pS. Reversal potential is ~0 mV.

To determine whether channel activation was voltage dependent, NP_o was plotted as a function of patch potential (Fig. 4). Channel opening was not significantly affected by changes in voltage over the range tested.

View larger version (10K): [in this window] [in a new window]   Fig. 4.   NPo as a function of patch potential at 50-mmHg suction. Only averaged data calculated from all-points histograms are shown. Channel opening was not significantly affected by changes in voltage applied to patch.

Additional experiments were performed with 20 µM gadolinium added to the pipette solution to determine whether gadolinium, a known blocker of MS channels, blocked the response to applied suction. We first recorded from several patches on a single coverslip using neurons cocultured with a mixture of ASM and AEC. The pipette solution was then changed to one containing gadolinium, and additional patches were obtained from different cells on the same coverslip. This procedure was then repeated for each coverslip studied. Before gadolinium, 11 of 15 patches contained MS channels. With 20 µM gadolinium in the pipette, MS channels were observed in only 2 of 19 patches during application of up to 60-mmHg negative pressure. Representative responses are shown in Fig. 5. These results suggest that gadolinium does indeed block MS channels in ABN.

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Recordings obtained with 20 µM Gd3+ in pipette solution showed little or no channel activity in response to application of 60-mmHg negative pressure. Control and Gd3+ traces were obtained from 2 different aortic baroreceptor neurons on same coverslip.

Expression of the MS channel in cultured ABN was significantly influenced by culture conditions. When ABN were cultured in the absence of other cells, MS channels were observed in only 1 of 10 dissociations and 3 of 16 patches (19%) from this dissociation. When ABN were cocultured with a mixture of ASM and AEC in transwells, however, 14 of 28 dissociations yielded MS channels and 39 of 65 patches (60%) from these dissociations exhibited MS channels. These observations suggest that a diffusible factor released from ASM and/or AEC may have increased the functional expression of MS channels in ABN. This diffusible factor was apparently secreted by AEC but not by ASM. When ABN were cultured with ASM alone, MS channels were found in only 2 of 20 patches (10%). With AEC alone in the transwell plates, MS channels were observed in 29 of 46 patches (63%).

Expression of MS channels was also compared in DiI-labeled and nonlabeled neurons. In one group of cells cocultured with AEC, channels were found in 5 of 36 patches (14%) from labeled cells and 1 of 20 patches (5%) from nonlabeled cells.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The single-channel currents characterized in this report most likely represent the opening of single MS ion channels in ABN. The channels have a reversal potential of ~0 mV, suggesting a nonspecific cation conductance, and a slope conductance of 114 pS, which is similar to the conductance of MS channels in chick cardiac cells (4, 12). Channel activity increased rapidly on application of negative pressure >30 mmHg and decreased on release of the pressure. This observation cannot be explained simply on the basis of nonspecific effects on the membrane or membrane breakdown, because channel activity was not seen in other cells not thought to be mechanosensitive (BC₃H1 cell line, both rapidly dividing and differentiated cells; R. Wachtel, unpublished observations). Channel activity was also relatively infrequent in ABN cultured in the absence of AEC and in nonlabeled nodose ganglion neurons. It is also unlikely that channel activity was caused by activation of voltage-gated conductances, because channels were seen in the presence of the N-type Ca²⁺-channel antagonist -conotoxin GVIA or blockers of voltage-gated K⁺ or Na⁺ channels. Channels were, however, blocked by 20 µM gadolinium, a known blocker of MS channels (7, 18). These observations are all consistent with our previous observations on mechanically induced whole cell currents (1, 2) and calcium transients (13, 16).

MS channels in ABN were not expressed at a high rate under standard culture conditions, making these recordings difficult to acquire. Coculturing the ABN with a mixture of ASM and AEC in a manner that allowed them to share the same medium appeared to increase the functional expression of these channels, suggesting that a diffusible factor may have been responsible for the increase in expression. When the ABN were cocultured with either ASM or AEC individually, results indicated that the AEC alone were capable of inducing expression of the MS channel. Others have previously reported that MS channel activity in embryonic heart cells is also sensitive to culture conditions (12).

A related issue is the distribution of channels on the surface of the neuron. All of our recordings were obtained from patches of membrane located on the soma of ABN, whereas in vivo the sensory apparatus that normally transduces mechanical signals is located in sensory nerve endings in the adventitia of the aortic arch. MS channels may be distributed nonuniformly across the cell membrane; they may be concentrated in the sensory nerve ending and expressed only at low levels throughout the rest of the cell. This could explain the small number of patches that contained the MS channel under standard culture conditions. It is also conceivable that differences in the cytoskeletal infrastructure between soma and nerve ending could influence the function of the MS channels, because our previous work with whole cell recordings suggests that the MS current is affected by alteration of the cell cytoskeleton (2).

In summary, our results show that ABN exhibit MS channels that are gadolinium-sensitive, nonspecific cationic conductances. These channels could participate in the process of mechanotransduction through which distension of the aortic arch is converted into a neuronal signal for relay to the central nervous system.