1 Department of Biology and
3 Departments of Obstetrics and
Gynecology and of Cellular and Molecular Medicine, Neurons are
mechanically robust. During prolonged swelling, molluscan neurons can
triple their apparent membrane area. They gain surface area and
capacitance independent of extracellular Ca concentration
([Ca]e), but it is
unknown if an increase in intracellular Ca concentration
([Ca]i) is
necessary. If Ca for stimulating exocytosis is unnecessary, it is
possible that swelling-induced membrane tension changes directly
trigger surface area readjustments. If, however, Ca-mediated but not
tension-mediated membrane recruitment is responsible for surface area
increases, swelling neurons should sustain elevated levels of
[Ca]i. The purpose of
this investigation is to determine if the
[Ca]i in swelling
neurons attains levels high enough to promote exocytosis and if any
such increase is required. Lymnaea
neurons were loaded with the Ca concentration indicator fura 2. Calibration was performed in situ using 4-bromo-A-23187 and Ca-ethylene
glycol-bis(
membrane tension; cell volume; molluscan neuron
IT IS WIDELY RECOGNIZED that cells regulate their
volume in response to osmotic perturbations, metabolic loads, and
cytomorphological changes. Much less consideration is given to the fact
that cell surface area, too, must be continually regulated. Because
surface area and cytoplasmic volume changes are not linearly related, we presume that regulatory mechanisms for surface area and for volume
use distinct feedback systems. For surface area regulation, membrane
tension changes may be the parameter that signals the need for changes
in surface area (J. Dai, M. P. Sheetz, and C. E. Morris, unpublished
data; Refs. 33, 42). This would constitute mechanosensitive membrane
disposition.
Molluscan neurons are well suited for studying mechanosensitive
membrane disposition because their large size facilitates the
monitoring of many parameters. We showed that osmotically swelling
Lymnaea neurons are astonishingly
compliant, increasing their apparent surface area threefold over 1 h
(40). Robustness under osmomechanical stress thus relies partly on an
ability to recruit membrane during swelling. It is likely that this
compliance is normally not used to deal with osmotic stresses but to
accommodate cell morphology changes in response to varying mechanical
stresses (4, 23, 39).
The existence of a mechanosensitive disposition of membrane between the
cell surface and internal stores is suggested by various lines of
evidence (Dai et al., unpublished data; Refs. 7, 13, 23, 33, 40, 42).
Ca-mediated exocytosis (31) and endocytosis, well-known processes by
which cells alter plasma membrane area, may be modulated by tension.
However, both rely on finely coordinated cellular chemistry and
cytoarchitecture, which could be compromised by stresses that create an
urgent demand for surface area increase (e.g., abrupt exposure to
anisosmotic media or to large mechanical perturbations). Area increases
triggered directly by the increased membrane tension might be more
fail-safe.
In osmotically perturbed molluscan neurons (see Fig.
1a), capacitance
increases and decreases are accompanied by apparent changes in surface
area (assuming a specific membrane capacitance of ~0.7-1
µF/cm2). Moreover, swelling
and shrinking not only alter total plasma membrane area, they cause
measurable membrane tension increases and decreases (Dai et al.,
unpublished data) and elicit visually dramatic rearrangements of
membrane (33). Less than 1 min after they are returned to normal saline
(NS), previously swollen neurons develop vacuole-like dilations (VLDs)
that can be 10 µm across. In neurons made to reswell, the VLDs
rapidly reverse (disappear), as if their membrane is drawn back to the
general plasma membrane surface. Membrane tension variations in
differentially adhesive regions of membrane are probably instrumental
in VLD dynamics. VLDs are initially invaginations at the substrate, but
some become fully internalized as true vacuoles (L. R. Mills and C. E. Morris, unpublished data). We interpret eventual VLD vacuolation as a sign of a particularly vigorous surface area regulation engendered by
episodes of swelling and then shrinking.
Swelling neurons may recruit lumenless infoldings of membrane to the
surface. Electron micrographs of cultured
Aplysia neurons reveal such infoldings
(12) that, by their ultrastructural appearance, are unlikely to
contribute measurable capacitance except when opened out at the
surface. Fejtl et al. (12) postulated that they constitute "hot
spots" of reserve membrane for area increases.
Osmotically induced capacitance changes and VLD dynamics occur with
equal facility in normal and in very low (0.5 µM) extracellular Ca
concentration
([Ca]e) (33, 40).
Swelling neurons increase their capacitance even with ethylene
glycol-bis( Vacuole-like membrane has been linked to cell surface area regulation
in various other neuronal preparations (5, 9, 14) and in skeletal
muscle (19). Cheng and Reese (5) showed that lumenless membranous disks
close to the plasma membrane in chick growth cones are part of an
internal membrane pool for augmenting growth cone surface area. Dailey
and Bridgman (9) observed vacuole-like organelles that are confluent
with the surface in rat superior cervical ganglion neurons; they
appeared spontaneously and then disappeared several minutes later, as
if they were participating in surface membrane recycling. On the basis
of noninvasive capacitance measurements, Sukhorukov et al. (38)
postulated that internal membrane reserves are instantaneously
available to the plasma membrane when myeloma, hybridoma, and
fibroblast cells swell osmotically. In frog muscle cells, Krolenko et
al. (19) observed vacuolar dilations, which, like
Lymnaea VLDs, formed in shrinking
cells and readily reversed with swelling.
In this study, our focus is on events during swelling. Specifically, we
ask if, over the time frame used to elicit VLDs, sustained swelling-induced intracellular Ca concentration
([Ca]i) increases occur that might be capable of driving exocytotic membrane area increases. We find that on average (but not in all neurons) swelling is
accompanied by a small rise in cytosolic Ca but never to a level
sufficient to stimulate exocytosis. We also show that neurons can swell
and then shrink and make VLDs under various conditions of enforced low
[Ca]i. Swelling
neurons need not, we conclude, rely on Ca-driven exocytosis as a means
of avoiding rupture.
Membrane tension therefore continues to be a good contender as a
primary signal for swelling-induced recruitment of membrane from
internal stores. Beyond enabling neurons to adjust during osmomechanical perturbations, tension-mediated membrane disposition could be important during passive stretch of neurites and during their
growth and retraction.
Cells, solutions, and flow-through chambers.
Circumesophageal ganglia were dissected from adult snails
(Lymnaea stagnalis) and placed
immediately into NS containing (in mM) 50 NaCl, 1.6 KCl, 3.5 CaCl2 (or 0.5 µM
CaCl2 for low-Ca experiments), 2.0 MgCl2, 5.0 N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid, and 5.0 glucose, pH adjusted to 7.6 with 1 M NaOH. Gentamicin was
added to 50 µg/ml. Osmolarity was 126 mosM, as determined by the
Advanced MicroOsmometer (model 3MO). Ganglia were digested for 30 min
in reduced Ca (0.5 mM CaCl2) NS
with 0.25% type XIV protease (Sigma, St. Louis, MO) and washed twice
in NS.
ABSTRACT
Top
Abstract
Introduction
Methods
Results
Discussion
References
-aminoethyl
ether)-N,N,N',N'-tetraacetic acid (EGTA), with free Ca concentration ranging from 0 to 5 µM. Swelling perturbations (medium osmolarity reduced to 25% for 5 min)
were done at either a standard
[Ca]e or very low
[Ca]e level (0.9 mM or
0.13 µM, respectively). In neither case did the
[Ca]i increase to
levels that drive exocytosis. We also monitored osmomechanically driven
membrane dynamics [swelling, then formation and reversal of
vacuole-like dilations (VLDs)] with the
[Ca]i clamped below 40 nM via
1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid (BAPTA). [Ca]i
did not change with swelling, and VLD behavior was unaffected,
consistent with tension-driven,
[Ca]i-independent surface area adjustments. In addition, neurons with
[Ca]i clamped at 0.1 µM via an ionophore could produce VLDs. We conclude that, under
mechanical stress, neuronal membranes are compliant by virtue of
surface area regulatory adjustments that operate independent of
[Ca]i. The findings
support the hypothesis that plasma membrane area is regulated in part
by membrane tension.
INTRODUCTION
Top
Abstract
Introduction
Methods
Results
Discussion
References
-aminoethyl
ether)-N,N,N',N'-tetraacetic acid (EGTA) in the cytosol (40). Although this does not rule out the
possibility that swelling stimulates exocytosis, it suggests that
membrane tension may trigger the surface area increases.
METHODS
Top
Abstract
Introduction
Methods
Results
Discussion
References
Testing for VLD formation. Because the ratio imaging system did not allow simultaneous visualization of cell morphology (specifically, VLD formation and reversal) and measurement of [Ca]i, VLD dynamics were checked in parallel experiments, using the test solutions on fura 2-loaded, BAPTA-loaded, and/or Ca-permeabilized cells. The flow-through chambers and the solutions were the same as for the [Ca]i measurements, and fura 2 loading was identical (see below). What we took note of was whether VLD formation occurred, was reversible, and was repeatable at discrete locations (see Fig. 1, a and b), as in previous work (33) without fura 2.
|
Fura 2 measurements of [Ca]i. About 2 h before an experiment, neurons in flow-through chambers were loaded for 45 min with 2 µM fura 2-AM (Molecular Probes) in 0.1% DMSO plus 0.02% Pluronic F-127, a dispersing agent. At 45 min and again at 75 min, chambers were washed with NS (5 × 50 µl), enabling neurons to rid themselves of untransformed fura 2-AM (intracellular fura 2 is generated when endogenous esterases cleave the AM) either by hydrolysis to fura 2 or by washout. A field with one or more neurons bearing a lamella was located (VLDs are easier to see in thin extensions than under the soma). The objective was focused not at the substrate but at a rounded region of the cell body, and the focal plane was maintained throughout. Unless noted, pairs of fluorescent images (at two excitation wavelengths, see below) of the field were generated every 30 s; each image was the average of 32 frames collected at 30 frames/s for a total duration of 1.07 s. A background image pair was taken at the end of each experiment in a field of view containing no neurons. Neurons exhibiting abnormally high resting [Ca]i were not used in the experiments.
[Ca]i was measured ratiometrically with a quantitative imaging fluorescence videomicroscopy system built around a Zeiss Axiovert inverted epifluorescence microscope. Fluorescence excitation was via a grating-type monochromator (Photon Technologies International, New Brunswick, NJ) with an electronic shutter (Uniblitz, Vincent Associates, Rochester, NY). Images were detected with a solid-state video camera (CCD72, Dage-MTI, Michigan City, IN) fitted with a Geniisys image intensifier (Dage-MTI). The emission wavelength was selected by a 510-nm bandpass filter. All of these were controlled by Isee and dsp/os software and imaging hardware (Inovision, Durham, NC) running on a Silicon Graphics Indigo2 (R4400, Extreme Graphics) Unix workstation (Silicon Graphics, Mountain View, CA). The excitation wavelengths were 350 nm (for detecting Ca-bound fura 2) and 380 nm (for Ca-free fura 2). For a circular area covering most of the soma of each neuron (preswelling), the ratio of the intensity at 350 nm divided by that at 380 nm was calculated pixel-by-pixel after background was subtracted using the Inovision Isee software, and the mean ratio was recorded. The quantitized areas were representative of the image of the entire ratio of the neurons, which were very homogeneous in appearance. Occasionally, however, it was evident from these images that a neuron had moved between data acquisition at 350 and 380 nm; striking crescents of low and high intensity appeared at opposite edges of the neuron. Data from these neurons were discarded.Calibration of fura 2, a free Ca concentration indicator, in Lymnaea neurons. Instead of calibrating by only determining emission ratios for extremely low (zero Ca) and extremely high (saturating) Ca levels, as is normally done (15), we calibrated fura 2 in neuronal cytoplasm (Fig. 2) over a range of extreme and intermediate concentrations. Thirty minutes before the experiment, 20 µM of the Ca ionophore 4-bromo-A-23187 (a nonfluorescent form) in 0.6% DMSO was added to the series of calibration solutions. Neurons were exposed to the 0 µM Ca-NS-A-23187 solution for 25 min. Fluorescent image pairs were taken, and then neurons were exposed for 10 min each to successively higher Ca-A-23187 solutions, with image pairs obtained at the end of each incubation. Background fluorescent images were subtracted before analysis. The calibration curve for [Ca]i was constructed by plotting the fluorescence ratios vs. free Ca concentration and fitting to an arbitrary polynomial equation using a nonlinear least squares fit. In essence, this is a computer-assisted "by eye" fit. The fitted line was used to read off interpolated Ca concentrations. To verify the accuracy of our in vivo calibration, the process was repeated months apart by different workers, using newly made solutions. The in vivo half-maximal ratios for fura 2 from the calibration curves of the two workers were at Ca concentrations of 0.13 and 0.14 µM.
|
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
VLDs: a phenomenon that occurs at submicromolar [Ca]e. Figure 1c demonstrates that VLDs were elicited in the same neuron with both normal and low-Ca media (see METHODS regarding osmotic dilutions) and, furthermore, confirms that [Ca]e does not affect the ability of neurons to form VLDs. Osmotic perturbations (swell, shrink, swell, i.e., 0.25× NS, 1× NS, 0.25× NS) were first done with normal Ca and then repeated with low-Ca solutions. VLDs formed when swollen neurons shrank (in NS) and reversed during the second swell (0.25× NS). During the low-Ca repeat runs, VLDs formed at the same sites as happened previously, reversed with a second swelling, and, finally, reformed when a second reshrinking occurred (in low-Ca NS).
Clearly, the VLD dynamics (formation, reversal, and reformation at discrete sites) are not dependent on [Ca]e levels but, even when swelling is stimulated in 0.13 µM [Ca]e, free [Ca]i may become elevated as a result of swelling-induced release of Ca from intracellular stores. The following experiments address this possibility.Does [Ca]i change during swelling and shrinking? Fura 2-loaded cells were used to measure [Ca]i during swelling. In preliminary trials using Hoffman modulation optics (which emphasize membranous structures), we confirmed that fura 2-loaded neurons retain their ability to repeatedly make VLDs, as above. For the Ca-monitoring experiments, inspection of the cells by transmitted light at the end of the experiment usually revealed the VLDs that had formed, but the optics were not set up for transmitted light, so during fluorescence image acquisition only fluorescence imaging was done.
As shown in Fig. 3a, [Ca]i for Lymnaea neurons in NS before any osmotic insult was 96 ± 9 nM (n = 19), which is in keeping with values for other molluscan neurons [40 nM in Helix aspersa (17), 90 nM in Helix pomatia (18), and 10-100 nM in Helix pomatia (25)]. Neurons were swollen in [Ca]e of 0.9 mM (Fig. 3a). Of 19 neurons tested, 11 showed a [Ca]i increase, 4 did not change, and 4 decreased; the average [Ca]i for a swollen neuron was 105 ± 11 nM, amounting overall to an 11 ± 0.5% increase over normal [Ca]i (P = 0.016, n = 19, within-neuron paired t-tests). After 5 min of swelling, neurons were returned to NS. At this point, they shrank and made VLDs and their average [Ca]i was 100 ± 10 nM, with three increased, seven decreased, and nine unchanged (a 9 ± 0.3% decrease but not significant, P = 0.08). Pairwise t-tests (n = 19) also showed that [Ca]i values in NS before swelling and in NS after the swelling episodes were statistically indistinguishable (P = 0.05). Thus, with [Ca]e at 0.9-3.5 mM during osmotic perturbations, there were at best marginal increases in intracellular Ca.
|
|
VLDs form with [Ca]i clamped to 0.1 µM using a Ca ionophore. Having determined (Fig. 3a) that resting [Ca]i (preswell, 3.5 mM [Ca]e) for these neurons was ~0.1 µM, we clamped [Ca]i at equilibrium with this level of external Ca and then tested the neurons' ability to form VLDs in response to osmotic perturbations. Under these conditions, [Ca]i should be unable to fluctuate above 0.1 µM Ca during perturbations.
After a brief wash in 0.1 µM Ca-NS, neurons were incubated for 25 min in 0.1 µM Ca NS containing 20 µM of the Ca ionophore, 4-bromo-A-23187 (Molecular Probes) dissolved in 0.6% DMSO. While doing the fura 2 calibration procedures, we found that Ca clamping with this ionophore requires the ionophore to be continuously present in the medium. In the continued presence of the ionophore, therefore, neurons were swollen in 0.25× NS for 4 min and then reshrunk in NS (note that for 0.1 µM Ca-NS this results in 0.025 µM Ca during swelling). The osmotic perturbation protocol was repeated to test if VLDs could reverse and then reform. Only about one-third of the neurons remained healthy during the 25-min preincubation in 0.1 µM [Ca]e. The other two-thirds retracted processes and/or became grainy. Nevertheless, those neurons with a tolerably normal appearance after the low-Ca preincubation withstood the osmotic perturbation cycles and formed and then reversed VLDs. Figure 5 shows two such neurons with VLDs in a lamellar region. The important point here is that VLDs can form in neurons that are prevented from experiencing a Ca perturbation during the swelling stimulus.
|
Are VLD dynamics possible when [Ca]i is held at very low levels by BAPTA? To clamp the [Ca]i to a lower level, cells were loaded with a membrane-permeant fast Ca chelator, BAPTA, by incubating them in 30 µM BAPTA-AM in 0.3% DMSO for 1 or 3 h. The ester (i.e., the AM form) freely crosses the lipid bilayer (6); in the cytoplasm, esterases cleave the ester, trapping BAPTA in the cytoplasm. Although the cytoplasmic BAPTA concentration is not known because loading is not an equilibration process, intracellular BAPTA concentration ([BAPTA]i) should have been at least as high at 3 h as at 1 h. At 1 or 3 h, the neurons were swollen in low-Ca 0.25× NS (0.13 µM Ca) for 4 min and then reshrunk in NS with 0.5 µM Ca. All neurons formed VLDs, which were then reversed and reformed twice more with the same osmotic shock solutions. Figure 6a illustrates VLDs made by this swell/shrink protocol in a control neuron, and b and c show neurons preloaded with BAPTA for 1 and 3 h (respectively) and treated in the same way.
|
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Resting [Ca]i in Lymnaea neurons was about one-half the normal level when [Ca]e was changed to 0.5 µM instead of the normal 3.5 mM (Figs. 3a and 4). Evidently, neuronal Ca homeostasis mechanisms cannot effectively counteract extremely low extracellular Ca levels by drawing on internal Ca stores. Because submicromolar external Ca would not be experienced in vivo, this is not surprising. It is, moreover, advantageous here because the induced low [Ca]i facilitates the study of putatively [Ca]i-independent events.
Swelling coincided with a small average increase in [Ca]i for media of both low and high [Ca]e. Although they are statistically significant, these increases are not physiologically significant in the context of the Ca levels normally thought to drive exocytosis (see below). The increases (from 96 to 105 nM and from 53 to 61 nM for the 0.9 mM and 0.13 µM [Ca]e + 0.25× NS swelling solutions, respectively) amount to absolute changes of <10 nM. Moreover, as seen in individual [Ca]i traces (see Figs. 3b and 4), some neurons experienced either no increase or even a decrease during swelling. The between-neurons variation in baseline [Ca]i is easily in the 10 nM range, as is the noise level for some neurons. This noise is probably cellular rather than instrumentational, since BAPTA-loaded neurons (Fig. 6a) were markedly quieter.
Movement is inherent in swelling and shrinking neurons. Swelling is most dramatic in the first minute after any osmotic perturbation in Lymnaea neurons (24), as in any cell; movement artifacts would have been most problematic as solution changes were implemented, and in most experiments [Ca]i was not monitored over this period. Spatially, movement artifacts are most pronounced at the cell perimeter; as mentioned (METHODS), we discarded data if such artifacts were evident. The large size of Lymnaea neurons (~20-80 µm diameter) should enhance the reliability of our measurements because, the larger a cell, the smaller its perimeter-to-cross-sectional ratio. Moreover, measurements were not made of the entire cytoplasm but made from a circular region slightly less than the minimum diameter of the soma, established for the control condition and then used throughout. This largely obviated problems at the perimeter of the cell related to inflation and deflation. In future experiments, it would be desirable to monitor the immediate submembrane region, but movements here will always introduce error. In any case, we know that transmembrane influxes of Ca are not needed for cells to produce VLDs or to increase capacitance as they swell (33, 40), so a sustained Ca release into the bulk of the cytoplasm from internal stores is what we were trying to detect.
Our data do not rule out a minor release of Ca from stores in swelling neurons, but comparing the swelling-induced increases of [Ca]i in Lymnaea to the [Ca]i increases in various cells undergoing exocytosis puts the swelling-induced changes in perspective. Eosinophil [Ca]i rose from 240 nM to 1.5 µM on exocytotic stimulation with guanosine 5'-O-(3-thiotriphosphate) (29), a difference of ~1,300 nM. In squid giant nerve terminals, presynaptic [Ca]i rose from 100 to 1,000 nM during trains of 250 action potentials (2), a 10-fold increase or, in absolute terms, an increase of 900 nM. Neher and Augustine (26) measured the [Ca]i in chromaffin cells at rest to be ~100 nM, and, once depolarized, this measurement was found to be ~550 nM at the plasma membrane and 500 nM deep in the cytoplasm. In the same type of preparation, Penner and Neher (32) found levels in stimulated cells in the range of 400-1,500 nM. Others have confirmed these results, showing that the [Ca]i at rest ranges from 10-100 nM and [Ca]i, when depolarization occurs, rises to 300-1,000 nM (3).
Even in normal [Ca]e, the largest swelling-related [Ca]i increases in Lymnaea neurons should not have been able to drive Ca-dependent exocytosis. Nevertheless, membrane capacitance and area increases in swelling Lymnaea neurons (40) and volume increases in swelling Aplysia neurons (12) indicate that the membrane is recruited by swelling. A simple possibility is that lumenless membrane folds or mechanically accessible membranous cisternae (12) are drawn into the plasma membrane by elevated tension. We have measured membrane tension in swelling neurons using plasma membrane tethers pulled by optical tweezers (Dai et al., unpublished data); Lymnaea neurons swelling in a 50% medium experience a tension transient that relaxes to a sustained 0.12 mN/m, compared with 0.04 mN/m before swelling. This is consistent with acquisition of new membrane in response to swelling tension. The sustained threefold increase in tension is not dangerously high, since 0.12 mN/m is about two orders of magnitude below membrane lytic tension in Lymnaea neurons. In fact, 10-12 mN/m has been determined as the lytic pressure in many biological membrane preparations (27).
Nichol and Hutter (28) showed that [Ca]i of 2 µM but not 0.8 µM is mechanically treacherous; a sharp decline in membrane strength occurs between these concentrations. Because the "danger zone" of [Ca]i overlaps that of neurons undergoing Ca-driven exocytosis, it may be imprudent for a mechanically stressed neuron to rely on Ca-driven exocytosis as a means of increasing membrane area. The small swelling-induced Ca concentration elevations in Lymnaea neurons (even in the higher external Ca concentrations) are well below the danger range.
The specific capacitance of Lymnaea neurons before swelling is consistent with a relatively nonwrinkled membrane surface (40), and fluorescent aqueous phase markers examined by confocal microscopy (33) reveal no dye-accessible infoldings. Thus, to prevent rupture, major options for a swelling neuron are 1) to retract processes and/or 2) to recruit new membrane. Brief swelling episodes elicit VLDs yet seldom cause retraction of processes, so we assume recruitment of membrane occurs. In cell culture, some cell types (20, 41) generate discrete spherical blebs as they swell. In contrast, swelling neuronal cells, including Lymnaea neurons and rat hippocampal neuron cultures (Dai et al., unpublished data), seldom bleb. Neurons made unhealthy by overdigestion during isolation, by age in culture, or by Ca-loading generate large discrete blebs (unpublished observations) consistent with the availability of membrane in this form. Perhaps the subplasmalemmal reserves discussed in Fejtl et al. (12) are the source of bleb membrane in weakened neurons. However, even when swelling is stimulated in near-distilled water (40), healthy neurons remain bleb-free. Contractile resistance to hydrostatic stress may help protract the duration of swelling and capacitance increase (40). Overall, neurons seem to maintain good control of membrane reserves, playing them out only as needed to keep membrane tension low (Dai et al., unpublished data). When stress is excessive, rupture, not blebbing, results. In initial trials for the present work, healthy Lymnaea neurons exploded as distilled water abruptly washed through the double coverslip chambers; previously (33), stepwise exchanges to distilled water were used, presumably allowing sufficient time for augmentation of membrane area.
Any reserve membrane contiguous with the plasma membrane (5, 9, 19, 38) could play a role in mechanosensitive membrane disposition. In addition, cells undergo constitutive exocytosis and endocytosis; a swelling-induced increase in exocytosis and/or decrease in endocytosis would result in net recruitment.
Holding [Ca]i constant does not prevent VLDs. Control neurons and BAPTA-loaded neurons were similar in appearance and produced VLDs. BAPTA liberated intracellularly has low membrane permeability, and so [BAPTA]i may have continued to increase during 3 h of loading. Experiments with neurons loaded for 45 min with BAPTA plus fura 2 showed that [Ca]i was buffered below 40 nM and that no [Ca]i increase was elicited by swelling. Therefore, we conclude that <40 nM [Ca]i did not impair the ability of neurons to swell without rupturing or to form VLDs repeatedly at the same sites. Morán et al. (21) also showed that BAPTA plus fura 2-loaded cerebellar granule neurons exposed to 50% medium for 90 s maintain a fixed [Ca]i; no tendency for low-Ca neurons to rupture was reported.
In theory, the ideal experiment is to clamp [Ca]i and [Ca]e to the resting [Ca]i level (0.1 µM, see Fig. 3a) and then attempt to elicit VLDs so that [Ca]i during swelling and subsequent VLD formation is known and invariant. However, Ca ionophores used to produce the clamp do not limit their effects to the plasma membrane. Ca released from intracellular stores (16) during A-23187 treatment probably caused the general cellular trauma noted, since, in BAPTA experiments, low [Ca]i per se was benign. A-23187 may also have sufficiently increased the proton permeability of plasma membrane and/or endomembranes to alter the intracellular pH. The granular appearance of A-23187-treated Lymnaea neurons echoes that of rat peripheral nerve (35) and dog spinal neurons (10). Neurons that survived and withstood the first osmotic perturbations were fragile, and recurrent shocks resulted in rounding up. Emery and Lucas (10) noted, as did we, varying degrees of damage with A-23187. As A-23187 releases Ca from stores, the Ca may stimulate exocytosis. BAPTA does not stimulate release of Ca from stores. Thus, even though we assume that A-23187-treated neurons swelled and then made VLDs with their cytoplasm clamped to [Ca]i of 0.1 µM, the ionophore clearly had unwanted side effects and we consider the BAPTA experiments a better indication that neurons can swell (presumably recruiting membrane) and subsequently form VLDs with no change in [Ca]i. In particular, the VLD formation experiments in low [Ca]e with BAPTA-AM-loaded neurons rule out the possibility that an undetected, fast, swelling-evoked Ca transient produced a net increase in plasma membrane area, either through Ca-mediated stimulation of exocytosis or through Ca-mediated inhibition of endocytic plasma membrane recycling.[Ca]i and swelling in other preparations. In neurons, resting [Ca]i falls between 10 and 300 nM (17). This range easily encompasses all our findings, even those for swollen neurons. We did, however, reject at the outset, as potentially unhealthy, cells whose initial [Ca]i was abnormally high. O'Connor and Kimelberg (30) measured [Ca]i in swollen astrocytes using fura 2; [Ca]i increased rapidly to 600 nM in ~25 s, then decreased to 250 nM (over 5 min). The same pattern was seen without extracellular Ca. The astrocyte Ca transient is thought to be a component of the signaling pathway for regulatory volume decrease (RVD).
Like many vertebrate and invertebrate neurons (1, 12), Lymnaea neurons do not exhibit RVD (24). In contrast, Pasantes-Morales and co-workers (34), using 0.2-0.5× medium, observed partial RVD in isolated neurons. Nevertheless, in their most recent work, they show that RVD is not dependent on [Ca]i. Cerebellar granule cells were loaded with fura 2, and osmotic changes were made at high time resolution. In 8 of 10 cells, swelling-induced changes were absent or inconsequential (21). BAPTA buffered [Ca]i to levels comparable to what we observed and abolished any hint of swelling-induced effects. Swelling did not detectably release Ca from internal stores. In a [Ca]e of 0.13 µM, the swelling-associated increase of [Ca]i (from 53 to 61 nM) of 0.008 µM in Lymnaea neurons could have been supplied by the external medium or by stores. The diminutive increases in Lymnaea neurons do not reflect insensitivity of the imaging system; in the same setup, we (36) showed that mouse zygotes (which exhibit RVD) loaded with fura 2 experience small swelling-induced increases in [Ca]i. As in the neurons, BAPTA clamped [Ca]i to a lower level. In summary, swelling does not elicit in Lymnaea neurons a sustained increase in [Ca]i that could support membrane recruitment via Ca-driven exocytosis. Swelling-induced increases in surface area and capacitance may, therefore, be driven directly by elevated membrane tension. For a neuron changing shape in a manner that demands increased surface area, tension-induced recruitment could locally augment plasma membrane area independent of any cytoplasmic signal. In vivo, both Ca-mediated (31, 37) surface area regulation and a separate tension-mediated form of regulation may operate. Coexisting tension- and Ca-dependent membrane recruitment may be widespread; plant protoplasts exhibit both Ca-mediated and (with ~30 nM [Ca]i; Ref. 42) pressure-stimulated capacitance increases, and it is suggested that the membrane pools involved are different. ![]() |
ACKNOWLEDGEMENTS |
---|
This work was supported by grants to C. E. Morris from the Heart and Stroke Foundation of Ontario and Natural Sciences and Engineering Research Council of Canada and by a grant to J. M. Baltz from the Medical Research Council (MRC), Canada. J. M. Baltz is an MRC Scholar.
![]() |
FOOTNOTES |
---|
Address for reprint requests: C. E. Morris, Neurosciences, Loeb Institute, Ottawa Civic Hospital, 1053 Carling Ave., Ottawa, Ontario, Canada K1Y 4E9.
Received 24 July 1997; accepted in final form 10 October 1997.
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
1.
Andrew, R. D.,
M. E. Lobinowich,
and
E. P. Oshehobo.
Evidence against volume regulation by cortical brain cells during acute osmotic stress.
Exp. Neurol.
143:
300-312,
1997[Medline].
2.
Augustine, G. J.,
E. M. Adler,
M. P. Charlton,
M. Hans,
D. Swandulla,
and
K. Zipser.
Presynaptic calcium signals during neurotransmitter release: detection with fluorescent indicators and other calcium chelators.
J. Physiol. (Lond.)
86:
129-134,
1992.
3.
Cheek, T. R.,
and
V. A. Barry.
Stimulus secretion coupling in excitable cells: a central role for calcium.
J. Exp. Biol.
184:
183-196,
1993
4.
Chen, C. S.,
M. Mrksich,
S. Huang,
G. M. Whitesides,
and
D. E. Ingber.
Geometric control of cell life and death.
Science
276:
1425-1428,
1997
5.
Cheng, T. P. O.,
and
T. S. Reese.
Recycling of plasmalemma in chick tectal growth cones.
J. Neurosci.
7:
1752-1759,
1987[Abstract].
6.
Cohan, C. S.,
J. A. Connor,
and
S. B. Kater.
Electrically and chemically mediated increases in intracellular calcium in neuronal growth cones.
J. Neurosci.
7:
3588-3599,
1987[Abstract].
7.
Dai, J.,
and
M. P. Sheetz.
Axon membrane flows from the growth cone to the cell body.
Cell
83:
693-701,
1995[Medline].
9.
Dailey, M. E.,
and
P. C. Bridgman.
Vacuole dynamics in growth cones: correlated EM and video observations.
J. Neurosci.
13:
3375-3393,
1993[Abstract].
10.
Emery, D. G.,
and
J. H. Lucas.
Ultrastructural damage and neuritic beading in cold stressed spinal neurons with comparisons to NMDA and A23187 toxicity.
Brain Res.
692:
161-173,
1995[Medline].
11.
Fabiato, A.,
and
F. Fabiato.
Calculator programs for computing the composition of the solutions containing multiple metals and ligands used for experiments in skinned muscle cells.
J. Physiol. Paris
75:
463-505,
1979[Medline].
12.
Fejtl, M.,
D. H. Szarowski,
D. Decker,
K. Buttle,
D. O. Carpenter,
and
J. N. Turner.
Three dimensional imaging and electrophysiology of Aplysia neurons during volume perturbation: confocal light and high voltage electron microscopy.
J. Microsc. Soc. Am.
1:
75-85,
1995.
13.
Fink, R. D.,
and
M. S. Cooper.
Apical membrane turnover is accelerated near cell-cell contacts in an embryonic epithelium.
Dev. Biol.
174:
180-189,
1996[Medline].
14.
Fujimoto, Y.,
and
K. Ogawa.
Retrieving vesicles in secretion-induced rat chromaffin cells contain fodrin.
J. Histochem. Cytochem.
37:
1589-1599,
1989[Abstract].
15.
Grynkiewicz, G.,
M. Poenie,
and
R. Y. Tsien.
A new generation of Ca2+ indicators with greatly improved fluorescence properties.
J. Biol. Chem.
260:
3440-3450,
1985[Abstract].
16.
Itoh, T.,
Y. Kanmura,
and
H. Kuriyama.
A23187 increases calcium permeability of store sites more than of surface membranes in the rabbit mesenteric artery.
J. Physiol. (Lond.)
359:
467-484,
1985[Abstract].
17.
Kennedy, H. J.,
and
R. C. Thomas.
Effects of injecting calcium buffer solutions on [Ca2+]i in voltage-clamped snail neurons.
Biophys. J.
70:
2120-2130,
1996[Abstract].
18.
Kostyuk, P. G.,
S. L. Mironov,
A. V. Tepikin,
and
P. V. Belan.
Cytoplasmic free Ca2+ in isolated snail neurons as revealed by fluorescent probe fura-2: mechanisms of Ca2+ recovery after Ca2+ load and Ca2+ release from intracellular sites.
J. Membr. Biol.
110:
11-18,
1989.
19.
Krolenko, S. A.,
W. B. Amos,
and
J. A. Lucy.
Reversible vacuolation of the transverse tubules of frog skeletal muscle: a confocal fluorescence microscopy study.
J. Muscle Res. Cell Motil.
16:
401-411,
1995[Medline].
20.
Menke, A.,
and
H. Jockusch.
Decreased osmotic stability of dystrophin-less muscle cells from the mdx mouse.
Nature
349:
69-71,
1991[Medline].
21.
Morán, J.,
S. Morales-Mulia,
A. Hernandez-Cruz,
and
H. Pasantes-Morales.
Regulatory volume decrease and associated osmolyte fluxes in cerebellar granule neurons are calcium independent.
J. Neurosci. Res.
47:
144-154,
1997[Medline].
23.
Morris, C. E.,
H. Lesiuk,
and
L. R. Mills.
How do neurons monitor their mechanical status?
Biol. Bull.
192:
118-120,
1997
24.
Morris, C. E.,
B. Williams,
and
W. J. Sigurdson.
Osmotically-induced volume changes in isolated cells of a pond snail.
Comp. Biochem. Physiol. A Physiol.
92A:
479-483,
1989.
25.
Muller, T. H.,
L. D. Partridge,
and
D. Swandulla.
Calcium buffering in bursting Helix pacemaker neurons.
Eur. J. Physiol.
425:
499-505,
1993.[Medline]
26.
Neher, E.,
and
G. J. Augustine.
Calcium gradients and buffers in bovine chromaffin cell.
J. Physiol. (Lond.)
450:
273-301,
1992[Abstract].
27.
Nichol, J. A.,
and
O. F. Hutter.
Tensile strength and dilatational elasticity of giant sarcolemmal vesicles shed from rabbit muscle.
J. Physiol. (Lond.)
493:
187-198,
1996[Abstract].
28.
Nichol, J. A.,
and
O. F. Hutter.
Ca2+ loading reduces the tensile strength of sarcolemmal vesicles shed from rabbit muscle.
J. Physiol. (Lond.)
493:
199-209,
1996[Abstract].
29.
Nusse, O.,
M. Lindau,
O. Cromwell,
A. B. Kay,
and
B. D. Gomperts.
Intracellular application of guanosine-5'-O-(3-thiotriphosphate) induces exocytotic granule fusion in guinea pig eosinophils.
J. Exp. Med.
171:
775-786,
1990[Abstract].
30.
O'Connor, E. R.,
and
H. K. Kimelberg.
Role of calcium in astrocyte volume regulation and in the release of ions and amino acids.
J. Neurosci.
13:
2638-2650,
1993[Abstract].
31.
Okada, Y.,
A. Hazama,
A. Hashimoto,
Y. Maruyama,
and
M. Kubo.
Exocytosis upon osmotic swelling in human epithelial cells.
Biochim. Biophys. Acta
1107:
201-205,
1992[Medline].
32.
Penner, R.,
and
E. Neher.
The role of calcium in stimulus-secretion coupling in excitable and non-excitable cells.
J. Exp. Biol.
139:
329-345,
1988[Abstract].
33.
Reuzeau, C.,
L. R. Mills,
J. A. Harris,
and
C. E. Morris.
Discrete and reversible vacuole-like dilations induced by osmomechanical perturbation of neurons.
J. Membr. Biol.
145:
33-47,
1995[Medline].
34.
Sanchez-Olea, R.,
H. Pasantes-Morales,
and
A. Schousboe.
Neurons respond to hyposmotic conditions by an increase in intracellular free calcium.
Neurochem. Res.
18:
147-152,
1993[Medline].
35.
Schlaepfer, W. W.
Structural alterations of peripheral nerve induced by the calcium ionophore A23187.
Brain Res.
136:
1-9,
1977[Medline].
36.
Séguin, D. G.,
and
J. M. Baltz.
Cell volume regulation by the mouse zygote: mechanism of recovery from a volume increase.
Am. J. Physiol.
272 (Cell Physiol. 41):
C1854-C1861,
1997
37.
Steinhardt, R. A.,
G. Bi,
and
J. M. Alderton.
Cell membrane resealing by a vesicular mechanism similar to neurotransmitter release.
Science
263:
390-393,
1994[Medline].
38.
Sukhorukov, V. L.,
W. M. Arnold,
and
U. Zimmermann.
Hypotonically induced changes in the plasma membrane of cultured mammalian cells.
J. Membr. Biol.
132:
27-40,
1993[Medline].
39.
Van Essen, D. C.
A tension-based theory of morphogenesis and compact wiring in the central nervous system.
Nature
385:
313-318,
1997[Medline].
40.
Wan, X.,
J. A. Harris,
and
C. E. Morris.
Responses of neurons to extreme osmomechanical stress.
J. Membr. Biol.
145:
21-31,
1995[Medline].
41.
Wilkerson, E. H.,
D. DiBona,
and
J. A. Schafer.
Analysis of structural changes during hypotonic swelling in Ehrlich ascites tumor cells.
Am. J. Physiol.
251 (Cell Physiol. 20):
C104-C114,
1986
42.
Zorec, R.,
and
M. Tester.
Rapid pressure driven exocytosis-endocytosis cycle in a single plant cell: capacitance measurements in aleurone protoplasts.
FEBS Lett.
333:
283-286,
1993[Medline].