AJP-Renal Articles in PresS. Published on February 5, 2002 as DOI 10.1152/ajprenal.00239.2001
A Mechanogated Nonselective Cation Channel in Proximal
Tubule that is ATP-Sensitive
Craig G. Hurwitz1, Vivian Y. Hu1, and Alan S. Segal1, 2
Running title: ATP-sensitive mechanogated channel in proximal tubule
University of Vermont
1
Department of Medicine
2
Department of Pharmacology
University of Vermont College of Medicine
Burlington, VT 05405
Corresponding author:
Alan S. Segal, M.D.
208 South Park Drive, Suite 2, Colchester, VT 05446
Telephone and Fax: (802) 656-8914
E-mail: [email protected]
Copyright 2002 by the American Physiological Society.
1
ABSTRACT
Ion channels that are gated in response to membrane deformation or “stretch” are
empirically designated stretch-activated (SA) channels. Here we describe a stretchactivated nonselective cation (SA-NSC) channel in the basolateral membrane (BLM) of
the proximal tubule (PT) that is nucleotide sensitive. Single channels were studied in
cell-intact and cell-free patches from the BLM of PT cells that maintain their epithelial
polarity. The limiting inward Cs+ conductance is ~28 pS and channel activity persists
following excision into a Ca2+- and ATP-free bath. The stretch dose-response is
sigmoidal with half-maximal activation of about –19 mmHg at –40 mV, and the channel
is activated by depolarization. The inward conductance sequence is: NH4+ ~ Cs+ ~ Rb+
> K+ ~ Na+ ~ Li+ > Ca2+ ~ Ba2+ > NMDG+ ~ TEA+. The venom of the common Chilean
tarantula, Grammostola spatulata, completely blocks channel activity in cell-attached
patches. Hypotonic swelling reversibly activates the channel. Intracellular ATP ([ATP]i)
reversibly blocks the channel (Ki ~ 0.48 mM), suggesting that channel function is
coupled to the metabolic state of the cell. We conclude that this channel may function
as a Ca2+ entry pathway and/or be involved in regulation of cell volume. We speculate
this channel may be important when [ATP]i is depleted, as occurs during periods of
increased transepithelial transport or with ischemic injury.
KEYWORDS: Mechanosensitive, stretch-activated, nonselective cation channels, ATPsensitive, patch clamp, pressure, proximal tubule, kidney
2
INTRODUCTION
Stretch-activated (SA) channels, gated by changes in hydrostatic pressure and/or
cellular swelling, have been found in many tissues (25). Their function is perhaps best
understood in sensory organs, where they appear to subserve electromechanical
transduction. However, the specific physiological role of SA channels in epithelia such
as the kidney remains largely unknown.
The proximal tubule (PT) performs the critical task of bulk solute and water
reabsorption for extracellular volume balance. PT cells must also tightly regulate their
own volume status because each minute the epithelium transports three times the
amount of sodium contained in its cells (34). To do so requires a reliable feedback
system that acts to maintain cell volume when there are transient discrepancies
between apical Na+ influx and basolateral Na+ efflux. Accordingly, most evidence to
date regarding a possible physiological role of SA channels in kidney pertains to their
mechanogating as a result of changes in cell volume. SA channels in the plasma
membrane may sense and respond to changes in cell volume and act as part of this
feedback system, as has been shown after cellular swelling induced by either hypotonic
shock (9,31) or more physiological substrate induced Na+ uptake (3).
Here, we describe the properties of a stretch- and swelling-activated channel in
the basolateral membrane (BLM) of PT that exhibits nucleotide sensitivity. We conclude
that this channel may function as a Ca2+ entry pathway and/or be involved in regulation
of cell volume. Furthermore, to demonstrate the coupling of SA channel function to
cellular metabolism in renal epithelia would provide evidence that this channel may be
3
important when [ATP]i is depleted, as occurs during increased transepithelial transport
or with ischemic cellular injury.
MATERIALS AND METHODS
Solutions and drugs: Experiments were performed using bath solutions that were either
nominally Ca2+-free or contained 1.8 mM calcium. The composition of the standard NaCl
bath recording solution was (in mM): 95 NaCl, 2.5 KCl, 1 MgCl2, 1 EGTA, and 10
HEPES. Unless otherwise noted, patch pipettes were filled with the standard CsCl
pipette solution (in mM): 95 CsCl, 2.5 KCl, 1 MgCl2, 1 EGTA, and 10 HEPES. The
rationale for using Cs+ in the pipette solution is 1) the SA-NSC is the only NSC channel
in the BLM that conducts Cs+, 2) Cs+ tends to block BLM K+ channels that might
otherwise contaminate recording of the SA-NSC, and 3) the single channel conductance
for Cs+ is greater than that for Na+, providing a better signal to noise ratio. After titration
to pH 7.5 (Orion Model 710A, Boston, MA), sucrose was added to adjust the osmolality
(Vapro Model 5520, Wescor, Inc., Logan, UT) of the solutions to 200 mOsm/kg. In biionic experiments, the pipette was filled with the chloride or gluconate salt of the test
cation such that pipette [test cation] is equal to bath [Na]. Hypotonic swelling
experiments were done at constant ionic strength beginning in an isotonic NaCl bath
solution containing (in mM): 58.8 NaCl, 2.5 KCl, 1.8 CaCl2, 1 MgCl2, 10 HEPES, and 76
sucrose. Cells were then exposed to a 38% decrease in osmolality (124 mOsm/kg) by
changing to the sucrose-free solution. To test the spider venom from Grammostola
spatulata (Spider Pharm, Yarnell, AZ) as a candidate specific blocker of the channel in
cell-attached patches, pipettes were backfilled with a solution containing 10 µl whole
4
venom per ml of buffer (the tip was initially venom-free). In solutions containing ATP, the
nucleotide was added as the Mg-salt. Nucleotides were prepared fresh daily in bath
solution. Chemicals used were obtained from Sigma (St. Louis, MO), Calbiochem (La
Jolla, CA), or Biomol (Plymouth Meeting, PA).
Cell Preparation: Dissociated proximal tubule cells were isolated from amphibian
kidneys as previously described (35). Briefly, aquatic phase Ambystoma tigrinum kept at
4°C were euthanized by submersion in 0.2% tricaine methanesulfonate. The kidneys
were rapidly removed and placed in iced HEPES-buffered NaCl (in mM: 90 NaCl, 2.5
KCl, 1 MgCl2, 1 EGTA, and 10 HEPES) at pH 7.5. Renal tissue was cut into 1-2 mm3
pieces and incubated in collagenase-dispase (0.2 U/ml of collagenase, BoehringerMannheim, Indianapolis, IN) on a shaker for 30 min at room temperature. The cells
were then mechanically dispersed into suspension by repeated trituration and
resuspended in NaCl and stored at 4°C until use. The dissociated PT cells can retain
their epithelial polarity for up to 14 days (35). Cells were used for experiments from 1 to
10 days following dissociation.
Electrophysiology: A 5 µl aliquot of Ambystoma cell suspension in NaCl was placed in
a recording chamber mounted on an inverted microscope (Olympus IX-70, Optical
Analysis Corp., Nashua, NH). Proximal tubule cells were readily recognized by their
characteristic asymmetric bi-lobated morphology. The standard configurations for single
channel patch clamp technique (11) were used to record channel currents from the
BLM. Patch pipettes were pulled from Corning 7052 borosilicate glass capillaries
(Warner Instrument Corp., Hamden, CT) on a two-step puller (Narishige PP-83, Japan),
5
coated with wax to within 200 µm of the tip, and fire-polished just prior to use. The open
tip pipette resistance was 3-8 MΩ when placed in the initial bath solution. A motorized
micromanipulator (MP-285, Sutter Instrument Co., Novato, CA) was used to guide the
patch microelectrode to the cell. The success rate of forming giga-ohm seals on the
BLM was ~90%. Data have not been corrected for liquid-junction potentials, which were
all less than 3.25 mV.
Voltage clamped membrane currents were amplified with an EPC-9 patch clamp
amplifier (HEKA Elektronik GmbH, Germany) controlled by a Pentium personal
computer running PULSE software (HEKA Elektronik GmbH, Germany) within Windows
98. Currents were low-pass filtered (at 0.4-1 kHz, LBPF-48-DG, npi electronic,
Germany) and then simultaneously printed on an analog strip chart recorder (Dash IV,
Astro-Med Inc., West Warwick, RI) and digitally sampled at 1-4 ksamples/s directly to
hard disk. Experiments were carried out at room temperature.
Application of pressure: Negative pressure was applied to the membrane patch through
the side port of the pipette holder and measured with an integrated piezo-resistive
pressure transducer (Fujikura XFPN-03PGVR, Servoflo, Lexington, MA) placed in
parallel with the pipette. Steady state pressure was applied using a pressure transducer
tester (DPM-1B, Biotek Instruments, Winooski, VT). For the application of pressure
ramps, a syringe pump (YA-12, Yale Apparatus, Wantagh, NY) under negative
feedback control (based on the output voltage from the pressure transducer) was
connected to the pressure port. For the application of pressure steps, a multi-valve
pressure system of our own design (14) was connected to the pipette holder. With this
system, up to seven different pressure steps (in addition to atmospheric pressure) can
6
be made, typically with a 20 to 80% step response time < 1 ms. Pressure step protocols
were triggered by PULSE and achieved using a parallel port interface (LPTek,
Westbury, NY) controlled by a Visual BASIC™ program written in our laboratory. The
interface includes a digital I/O board for valve selection and a counter board used for
timing voltage pulses (spike, hold, off) from one channel of the D/A board to high current
DS3658 chips that drive the valves. During pressure protocols, both the channel
currents and pressure data are recorded simultaneously by PULSE, facilitating the
analysis of the temporal relationship between pressure stimulus and channel response.
Data Analysis: PULSE datafiles were read directly by ORIGIN™ 6.0 (OriginLab,
Northampton, MA) using DataAccess™ 6.0 (Bruxton Corp., Seattle, WA). Custom
software for data analysis was written in our laboratory using Matlab™ 5.2 (The
Mathworks, Natick, MA), Lab Talk™ 6.0 (OriginLab, Northampton, MA) and Visual
BASIC™ 6.0 (Microsoft Corp, Seattle, WA). Channel activity (NPo, where N is the
minimum number of channels in the patch and Po is the channel open probability) was
calculated by applying a graphical method to binned data from amplitude histograms
representing all points from at least 30 s of continuous recording as follows: Based on
mouse clicking on the peak corresponding to the all channels closed level (ic) and then
on two adjacent peaks (to calculate the single-channel current, isc), our program bins
and numerically integrates the histogram to get the total area under the curve. The all
channels closed current is subtracted from the current of a given bin, and the number of
events in that bin is multiplied by the difference in current. The sum of these products
7
yields the open channel area of the histogram. NPo is then given by dividing the open
channel area by the total area under the histogram. That is,
NPo =
∑ [(i
bin
− ic ) ⋅ (# eventsbin )]
bins
(isc ⋅ # eventstotal )
(1)
In the text the number of observations or experiments is reported, whereas n in
the analysis denotes either the whole data set or the subset of total experiments in
which precise quantitation could be reliably applied. Statistical values for the n elements
are given as mean ± standard error. Student’s t-test was applied where appropriate.
Kinetic analysis was performed from long recordings of patches exhibiting a
single open channel level. These patches were held at –40 mV and –50 mmHg, the
data were filtered at 2 kHz, digitized at 10 ksample-s-1, and analyzed based on the
method of Sigworth & Sine (37) using TAC/Fit (Bruxton, Inc., Seattle, WA).
RESULTS
A nonselective cation (NSC) channel activated by membrane tension is highly
expressed in the BLM: More than 75% (425 out of 550) of seals made on the BLM
contained at least one NSC channel that could be activated by negative pressure,
suggesting that the SA-NSC channel protein is highly expressed in this membrane.
Following formation of a cell-attached patch there is typically no channel activity in the
absence of applied pressure at 0 mV pipette potential, although channel openings will
be evident in a minority (10-15%) of patches. Fig. 1A shows that a ramp pressure
protocol reversibly activates a NSC channel. In this representative excised inside-out
membrane patch, rare channel openings are evident at a command potential of –40 mV
8
prior to application of pressure. Channel activity increases with negative pressure.
Conversely, during the ramp back to 0 mmHg, channel activity progressively declines.
Fig. 1B shows channel records from an excised membrane patch (held at –40
mV) at variety of pressure levels, which demonstrates a direct relation between channel
activity (NPo) and pipette pressure. The relationship between applied negative pressure
and NPo shown in Fig. 1C indicates that the pressure for half-maximal channel
activation (EP50) is –19.3 mmHg at –40 mV. The EP50 tends to decrease (i.e., less
negative pressure is required to gate the channel open) with depolarization . This value
is fairly consistent, but does vary in some cell preparations. This can be manifested as
either more activity at 0 mmHg, a lower threshold pressure (e.g., ~ 5 mmHg rather than
~12 mmHg), or as a much higher threshold pressure (e.g., ~18-25 mmHg) as seen in
Fig. 1D.
The SA channel inwardly rectifies and is cation-nonselective: When BLM membrane
patches are excised in the inside-out configuration under symmetrical conditions,
measurements of the single channel current demonstrate that the SA-NSC channel is a
true weak inward rectifier (Fig. 2A and 2B) with a rectification index (gin/gout) of 1.4.
Kinetic analyses of one channel patches held at –40 mV and –50 mmHg show that the
SA-NSC channel has two open states (6 ms, 75% weight; 25 ms, 25% weight) and
three closed states (~0.6 ms, 62% weight; ~16 ms, 22% weight; and ~180 ms, 16%
weight), as shown in Fig. 2C. The channel is permeable to a wide variety of cations.
Representative channel records from patches excised under bi-ionic conditions ([pipette
cation] equal to [bath Na]) are shown in Fig. 3A, and the summarized I-V relationships
9
are plotted in Fig. 3B. Note that the reversal potential is about the same for all
conditions, so although the conductances differ, the permeabilities of the conducting
cations are virtually identical. The single channel conductance for the outward flow of
Na is the same (~21 pS) regardless of the pipette cation. The inward conductance is
divided into groups that can be correlated to the hydrated radius (or the dehydration
energy) of the cation. Larger cations (Rb+, Cs+, NH4+) with smaller hydrated radii have a
higher conductance (gin~ 47 pS) than smaller cations (gin ~ 28 pS for Li+, Na+, K+) with
larger hydrated radii. The channel also conducts Ca2+, but with different kinetics (more
flickery). Both NMDG+ and TEA+ (large cations usually considered to be impermeant)
conduct at ~6 pS and ~5 pS, respectively (inset, Fig 3A). The presence or absence of
bath calcium had negligible effects on channel behavior.
The BLM SA-NSC channel is depolarization-activated: The open probability of the
channel is voltage-dependent, and increases with depolarization (Fig. 4A). The NPo was
measured from continuous current records at least 30 s long from patches held at
constant negative pressure. Because channel activity varies among patches, NPo was
normalized to the maximum NPo in each patch. Channel activity increases e-fold for
every 49.5 mV depolarization. The experiment shown in Fig. 4B is consistent with this
notion. The pressure system was set to make a step from 0 mmHg to –40 mmHg and
back to 0 mmHg over a duration of 5 ms, and channel recordings were made at ±100
mV. Note that the opening latency time is shorter, and the closing latency time is longer,
at +100 mV compared to those at –100 mV. Pooled latency time data from 7
experiments are summarized in Fig. 4C. This result suggests that depolarization
10
shortens a closed state lifetime and prolongs an open state lifetime, both of which favor
a higher open probability.
The multi-step pressure protocol in Fig. 4D reinforces this idea. With descending
negative pressure steps, –60 mmHg opens channels at both ±80 mV, but reducing the
pressure to –40 mmHg leads to closure of the channels only at –80 mV. Channels stay
open at +80 mV even when the pressure is further decreased to –20 mmHg, but close
as expected at 0 mmHg. With ascending negative pressure steps, channels re-open at
–40 mmHg at +80 mV, but fail to open at –80 mV until the pressure increases to –60
mmHg.
This SA channel is Gd3+-insensitive but blocked by spider venom: Sensitivity to the
lanthanide gadolinium (Gd3+) is a common feature of many mechanosensitive channels,
especially those that are cation non-selective (29,45). The SA-NSC channel in the BLM
appears to be insensitive to Gd3+, as detailed in Fig. 5. Activation of the channel by
negative pressure is unaffected by 100 µM Gd3+ in the pipette and/or bath of excised
inside-out (Fig. 5A) or outside-out (Fig. 5B) patches. The experiment in Fig. 5C shows
that addition of 100 µM Gd3+ to an outside-out patch conducting Na+ has no effect on
channel activity. Other typical inhibitors of mechanogated or NSC channels, including
amiloride, flufenamic acid, and niflumic acid have no significant effect on the SA-NSC
channel (data not shown).
Spider venom from Grammostola has been shown to block SA channels (4).
Therefore, cell-attached patches were made using patch pipettes backfilled with solution
containing 10 µl whole Grammostola spatulata venom per ml of buffer (the tip was
11
initially venom-free). Figure 6B represents one of four consecutive patches that showed
activity from one to four channels just after seal formation and subsequently exhibited
complete loss of channel openings within a few minutes. Such behavior is
unprecedented in the absence of venom (Fig. 6A), as our experience with hundreds of
patches has shown that SA channel activity is maintained throughout recordings that
may last more than an hour. These results strongly suggest that this spider venom
contains a specific blocker of our channel.
Cellular swelling reversibly activates the SA-NSC channel: It is reasonable to
hypothesize that it is involved in cellular volume regulation, and the experiment shown
in Fig. 7A is consistent with such a role. After baseline channel activity is established in
a cell-attached patch under isotonic conditions, the cell is exposed to hypotonic
conditions by removal of sucrose at a constant ionic strength. The response of the
channel is biphasic, increasing significantly within a few seconds and reaching a peak
(a 70% increase over baseline NPo) within a minute, and then decreasing over the next
two minutes. Channel activity decreases or abruptly ceases if cellular swelling is
stopped or reversed, respectively. Such behavior suggests that the channel may be
involved in the regulatory volume decrease (RVD) response of the cell. A summary of
six similar swelling experiments are shown in Fig 7B.
Sensitivity to adenine nucleotides: Nucleotides have been shown to inhibit some NSC
channels (17). The SA channel is sensitive to ADP and ATP, as shown in Fig. 8. In this
experiment, a constant negative pressure of –17 mmHg activates the channel. Addition
12
of 5 mM UDP to the cytoplasmic face of the inside-out patch does not affect channel
activity, but application of 5 mM ADP significantly and reversibly inhibits the channel.
Exposure to 5 mM ATP demonstrates a more potent inhibition that also reverses
(indeed, greater than that prior to ATP) after washout. That ADP also blocks the
channel indicates that inhibition by adenine nucleotides neither involves a typical protein
kinase nor requires hydrolysis by an ATPase. Therefore, channel inhibition most likely
occurs when an appropriate adenine nucleotide binds to a cytoplasmic domain of the
channel protein. The dose-response curve for ATP is sigmoidal and has a Ki of 0.48
mM (Fig. 8B).
The increase of NPo upon washout suggests that ATP refreshes some
channels during the period of inhibition. The degree of inhibition by ATP does not
appear to be dependent on the magnitude of negative pressure applied to the patch.
The block achieved using 2 mM ATP was compared at two levels of applied negative
pressure in the same patch; one at which the baseline Po was ~0.20 and another at
which baseline Po was ~0.70. There was no significant difference in the potency of
block (88 ± 3.8% vs. 87 ± 1.1%, n=3) at the two levels of negative pressure and open
probability (Fig. 8C).
DISCUSSION
The primary objective of this study was to characterize the electrophysiological
properties of a SA-NSC channel found at a very high frequency in the BLM of
Ambystoma PT cells. Previous work on SA channels in PT has demonstrated regulation
by cell volume changes induced by hypotonic shock and uptake of substrate
(3,9,16,30). Our data demonstrate that a Ca2+-independent, mechanogated NSC
13
channel in the BLM of PT is sensitive to ATP. The likelihood of finding at least one SANSC channel in a membrane patch is nearly 80%, suggesting that this integral
membrane protein is important to the cellular and molecular physiology of the proximal
tubule.
Mechanosensitivity:
Of the cell surface proteins involved in “mechanotransduction”
(e.g., integrins and ion channels), mechanosensitive ion channels can theoretically
provide a rapid and reversible signal in response to membrane deformation. The
application of steady state negative pressure to membrane patches has been shown to
both inhibit (40) and activate (30) mechanosensitive channels. Given the sigmoidal
relationship between negative pressure and channel open probability shown in Fig. 1,
the NSC channel described in this study is clearly a stretch-activated channel. Across
multiple species, stretch sensitivity of basolateral SA channels is similar. For example,
the half maximal activation of our channel (–19.3 mmHg at –40 mV) correlates well with
that reported for SA channels in Xenopus PT cells (16), and is slightly less stretch
sensitive than that described for the BLM SA channels in Rana temporaria (12) and
Necturus (30), where the half maximal activation occurs around –11 mmHg. An osmolal
gradient of 1 mOsm/kg results in a hydrostatic pressure of ~17 mmHg, suggesting these
channels operate over a physiologically relevant range. While there is some variation in
the mechanosensitivity of the channel in some of our cell preparations, the majority of
patches fall into the range detailed in Figs. 1B and 1C. It is unclear what factors are
responsible for occasional variations in the mechanosensitivity between preparations.
14
Kinetic analyses of one channel patches show that this SA-NSC channel has two
open states and three closed states. As reviewed by Sachs and Morris (29) and Sackin
(32), mechanosensitive channels have at least one stretch-sensitive closed state with a
very long dwell time in the absence of applied pressure. Negative pressure probably
lowers the energy barrier separating the closed and open states of the channel.
Membrane tension, perhaps by altering the membrane deformation energy, decreases
the probability of this closed state. Although it is likely that the longest closed state of
the SA-NSC is stretch-sensitive, additional experiments are necessary to determine if
other states are also stretch-sensitive.
Ion Conductance, Rectification, and Selectivity:
SA channels in amphibian kidney
demonstrate a fairly consistent conductance of about 25 to 30 pS (12,13,16), consistent
with that of the SA channel in the present study. Higher conductance SA channels have
been described in Necturus (9) and Rana pipiens (3). Irrespective of the cation tested
(including TEA+), under symmetrical conditions the SA-NSC channel is a weak inward
rectifier. In the present study, Mg2+ was always present on the cytosolic side of the
patch and therefore the possibility that the inward rectification is due to Mg2+ block, as is
the case for some K+ channels (20), cannot be excluded.
The inward conductance sequence suggests grouping correlated to the hydrated
radius (or the dehydration energy) of the cations. Therefore, the selectivity sequence
favors ions like Cs+ and Rb+ due to the relatively low energy required to shed their
waters (-71 and –79 kcal/mol, respectively). This implies that conductance increases as
atomic radii increase and as hydration energies decrease (Eisenman sequences I and
15
II) (6). The reduced conductance for Ca2+ may be a reflection of its large hydration
energy (-397 kcal/mol).
Voltage dependence: Of the SA channels described in renal epithelia, those found in
Necturus
and
Xenopus
are
K-selective
hyperpolarization activation (16,33).
and
have
uniformly
demonstrated
The SA-NSC channel described here is
depolarization activated, which is similar to the SA-NSC channel in Rana temporaria
(12). The observation that depolarization increases NPo even at 0 mmHg suggests that
the stretch-sensitive closed state is voltage-dependent. Using our high-speed pressure
step system to investigate the relationship between voltage and latency times, the
depolarization activation is seen from a novel perspective. That is, rather than the
steady state measurement of NPo, we see an instantaneous reflection of the
depolarization activation (on the order of milliseconds) in the form of a shortened
opening latency and a longer closing latency, both of which favor an increased open
probability. This important finding is in sharp contrast to the very slow depolarization
activation found in the SA channels of Xenopus oocytes (10) and leech central neurons
(23), which occurs on the order of seconds to tens of seconds.
Mechanistically, voltage dependence can be due to voltage-induced changes in
1) channel protein(s), 2) surrounding structures, or 3) membrane deformation. Gil et al.
hypothesize that voltage across the patch influences membrane tension and gates SA
channels (10). Because depolarization activation did not occur with soft glass pipettes
and was absent in outside-out patches, they concluded that voltage dependence was
an artifact due to specific interactions between the glass pipette and membrane patch in
16
Xenopus oocytes.
In contrast, the SA channel described here is activated by
depolarization regardless of the excised patch configuration, similar to that described for
mechanogated channels in leech neurons (23). If the same process of voltage-induced
membrane deformation were occurring in Ambystoma epithelia, Xenopus oocytes, and
leech neurons, such drastic differences in times to activation would not be expected.
However, we are unable to rule out the possibility that voltage-dependence may be an
intrinsic property of the channel.
Blockade: Since first being described to inhibit the beating of a perfused frog heart (24),
lanthanides have become widely used in the study of cellular physiology. Sensitivity to
the lanthanide gadolinium (Gd3+) is a common feature of many mechanosensitive
channels and is reported in a number of different preparations including yeast MID1
(15), Necturus PT cells (8), renal A6 cells (43), Xenopus oocytes (45), and human
neurons (28). The SA channel described in this paper is unaffected by 100 µM Gd3+ in
the pipette and/or bath of excised inside-out or outside-out patches. However, given the
broad range of cross-reactivity, Gd3+ inhibition is currently viewed as a less reliable
marker for SA channels (29). Other typical inhibitors of mechanogated or NSC
channels, including amiloride, flufenamic acid, and niflumic acid have no significant
effect on this SA-NSC channel. Unfortunately, specific pharmacological inhibitors of SA
channels are not readily available, but a natural toxin found in the venom of the
common Chilean tarantula, Grammostola spatulata, has been shown to block stretchactivated channels in GH3 pituitary cells (4). Our results indicate that this venom
completely blocks channel activity in cell-attached patches containing SA-NSCs. The
17
block may be due to a 35 amino acid peptide (GsMTx-4) in the venom recently identified
by Sachs and coworkers (39). GsMTx-4 is the first peptide toxin that specifically blocks
stretch-activated channels. They showed that 5 µM GsMTx-4 completely blocks
mechanosensitive cation channels in astrocytes and cardiac myocytes. Based on a
molecular weight of 4093.9 and a concentration of 1.95 mM GsMTx-4 in whole venom
(39), channels in cell-attached patches were exposed to ~20 µM GsMTx-4 toxin. To our
knowledge, the SA-NSC in the proximal tubule is the first epithelial channel blocked by
this venom.
Swelling Activation: Swelling activated mechanogated channels have been described in
many tissues including the proximal tubule (3,30).
The SA channel in the BLM of
Ambystoma PT cells is also activated by hypotonic challenge. Unlike activation induced
by stretch, swelling activation of the SA-NSC results in a biphasic response. Based on
studies implementing simultaneous cell volume and patch recording techniques, it
appears that the biphasic response of SA channels during swelling can be directly
related to the cell volume changes during RVD (9). Whether or not the SA channel in
Ambystoma is actually a component of the RVD remains unknown and, in the absence
of having a specific channel blocker, is difficult to ascertain.
From a physiological
standpoint, hypotonic swelling likely does not reflect cellular volume changes as they
occur during actual transepithelial transport. That is, hypotonic stress may influence a
number of cellular signaling systems (e.g., the cytoskeleton and intracellular
messengers), making interpretation of these data more complex (29).
18
Nucleotide Sensitivity and Possible Physiological Role: An important feature of the
channel is inhibition by cytoplasmic ATP. That the nucleotide diphosphate ADP also
blocks the SA-NSC channel indicates that inhibition by adenine nucleotides probably
does not require a typical protein kinase or hydrolysis by an ATPase. Therefore,
channel inhibition most likely occurs when an appropriate adenine (but not uridine)
nucleotide binds to a cytoplasmic domain of the channel protein. While ATP (41) or
cyclic nucleotides (7) regulate certain members of the NSC channel superfamily, to our
knowledge the SA-NSC channel is the only mechanogated channel that is also
nucleotide-sensitive. Reminiscent of ATP-sensitive K+ channels (26), the tendency of
NPo to increase upon washout of ATP (see Fig. 8) suggests that some degree of
channel refreshment by ATP occurs at a second site, most likely due to phosphorylation
of the channel or an associated protein.
As is the case with many mechanosensitive channels, its precise physiological
function remains uncertain. Indeed, only within the past two years have data been
presented to support the hypothesis that epithelial SA channels in cell-attached patches
underlie the whole-cell currents elicited by cellular swelling (44). One mechanism that
may be involved in the regulatory volume decrease (RVD) of epithelial cells exposed to
hypotonicity involves Ca2+ entry via stretch-activated NSC channels, which in turn
activates Ca2+-activated K+ channels in the same membrane (5). The BLM of PT cells
appears to contain at least two K+ channels (27), one of which is a KATP channel that is
inhibited by Ca2+ (21,22). If the other K+ channel is also regulated by calcium, then Ca2+
entry via the SA-NSC channel may contribute to RVD through secondary effects on K+
channels also in the BLM (18).
19
Recently cloned mammalian TRP-like Ca2+-permeable NSC channels—OTRPC4
(38) and VR-OAC, the vanilloid receptor-related osmotically activated channel (19)—are
both activated by hypotonicity and found at high levels in the renal tubules.
Interestingly, TRP-like channels have not been reported in patch clamp experiments in
native kidney, raising the possibility that the properties of recombinant channels
expressed in vitro differ from those in situ. If so, it is possible that the SA-NSC channel
is related to one of these TRP-like channels.
Nucleotide sensitivity is a property that may provide important new insight to the
physiological role of this channel because of the possibility that SA channel function is
linked to cellular metabolism in the proximal tubule. However, care must be exercised
regarding interpretation of the physiological significance within the possible range for
the changes in [ATP] and [ADP] because we do not yet know what these ranges are,
whether the channel responds to the absolute level of these nucleotides or their ratio,
and/or if key modifiers are lost after patch excision. On the one hand, the Ki of 0.48 mM
for ATP-inhibition in excised patches suggests that the channel may not be open given
PT [ATP]i levels in the range of 2 – 5 mM under healthy conditions (2,42). Rather, it is
possible that the channel becomes important when intracellular ATP is depleted, as
occurs during ischemic cell injury. During such a stress, SA-NSC channels would open
providing a cation entry pathway for Ca2+ or Na+. In the case of the latter, as opposed to
Na+ entry across the apical membrane, Na+ entering across the BLM would decrease
net Na+ reabsorption and induce further cellular swelling. This raises the possibility that
if the channel conducts Na+ in vivo, it may be to promote nonspecific cell death, as has
been proposed for liver cells (1).
20
On the other hand, as has been shown for KATP channels, protein-lipid
interactions may modify ATP sensitivity, essentially permitting channel function in the
presence of what would otherwise be inhibitory concentrations of ATP (36). For this
reason, the SA-NSC channel described may play a role during transepithelial transport
under healthy conditions. In this case, the SA-NSC channel could become active in
response to both cellular swelling (44) and reductions in [ATP]i allowing Na+ or Ca2+ to
enter, or (less likely) K+ to exit. We have previously shown that intracellular Ca2+ is an
important regulator of the hyperpolarization-activated KATP channel in the same
membrane (22). With these considerations, it is possible that Ca2+ entry or a
depolarizing Na+ current may be important to optimize the pump-leak coupling that
occurs between the Na,K-ATPase and KATP channels in the BLM of PT.
We conclude that an ATP-sensitive mechanogated non-selective cation channel
is highly expressed in the BLM of PT. This mechanogated channel could be functionally
coupled to the metabolic state of the cell and may be important for cellular volume
regulation and/or as a calcium entry pathway during transepithelial transport or in the
response to ischemic injury.
21
ACKNOWLEDGMENTS
We thank Dr. J. Brayden, Dr. S. Lidofsky, and the anonymous reviewers for valuable
suggestions and comments on the manuscript. We are grateful for generous support
from the Department of Medicine at the University of Vermont. Dr. C. Hurwitz is a
recipient of an Individual National Research Service Award from the NIH (DK10061).
22
REFERENCES
1. Barros LF, Stutzin A, Calixto A, Catalan M, Castro J, Hetz C and Hermosilla T.
Nonselective cation channels as effectors of free radical-induced rat liver cell
necrosis. Hepatology 33: 114-122., 2001.
2. Beck JS, Breton S, Mairbaurl H, Laprade R and Giebisch G. Relationship
between sodium transport and intracellular ATP in isolated perfused rabbit proximal
convoluted tubule. Am J Physiol 261: F634-639, 1991.
3. Cemerikic D and Sackin H. Substrate activation of mechanosensitive, whole cell
currents in renal proximal tubule. Am J Physiol 264: F697-714, 1993.
4. Chen Y, Simasko SM, Niggel J, Sigurdson WJ and Sachs F. Ca2+ uptake in GH3
cells during hypotonic swelling: the sensory role of stretch-activated ion channels.
Am J Physiol 270: C1790-1798, 1996.
5. Christensen O. Mediation of cell volume regulation by Ca2+ influx through stretchactivated channels. Nature 330: 66-68, 1987.
6. Eisenman G and Dani JA. An introduction to molecular architecture and
permeability of ion channels. Annu Rev Biophys Biophys Chem 16: 205-226, 1987.
7. Fesenko EE, Kolesnikov SS and Lyubarsky AL. Induction by cyclic GMP of
cationic conductance in plasma membrane of retinal rod outer segment. Nature 313:
310-313, 1985.
8. Filipovic D and Sackin H. A calcium-permeable stretch-activated cation channel in
renal proximal tubule. Am J Physiol 260: F119-129, 1991.
9. Filipovic D and Sackin H. Stretch- and volume-activated channels in isolated
proximal tubule cells. Am J Physiol 262: F857-870, 1992.
23
10. Gil Z, Magleby KL and Silberberg SD. Membrane-pipette interactions underlie
delayed voltage activation of mechanosensitive channels in Xenopus oocytes.
Biophys J 76: 3118-3127., 1999.
11. Hamill OP, Marty A, Neher E, Sakmann B and Sigworth FJ. Improved patchclamp techniques for high-resolution current recording from cells and cell-free
membrane patches. Pflugers Arch 391: 85-100, 1981.
12. Hunter M. Stretch-activated channels in the basolateral membrane of single
proximal cells of frog kidney. Pflugers Arch 416: 448-453, 1990.
13. Hurst AM and Hunter M. Stretch-activated channels in single early distal tubule
cells of the frog. J Physiol (Lond) 430: 13-24, 1990.
14. Hurwitz CG and Segal AS. Application of pressure steps to mechanosensitive
channels in membrane patches: a simple, economical, and fast system. Pflugers
Arch 442: 150-156., 2001.
15. Kanzaki M, Nagasawa M, Kojima I, Sato C, Naruse K, Sokabe M and Iida H.
Molecular identification of a eukaryotic, stretch-activated nonselective cation
channel. Science 285: 882-886, 1999.
16. Kawahara K. A stretch-activated K+ channel in the basolateral membrane of
Xenopus kidney proximal tubule cells. Pflugers Arch 415: 624-629., 1990.
17. Korbmacher C, Volk T, Segal AS, Boulpaep EL and Fromter E. A calciumactivated and nucleotide-sensitive nonselective cation channel in M-1 mouse cortical
collecting duct cells. J Membr Biol 146: 29-45, 1995.
24
18. Lang F, Busch GL, Ritter M, Volkl H, Waldegger S, Gulbins E and Haussinger
D. Functional significance of cell volume regulatory mechanisms. Physiol Rev 78:
247-306, 1998.
19. Liedtke W, Choe Y, Marti-Renom MA, Bell AM, Denis CS, Sali A, Hudspeth AJ,
Friedman JM and Heller S. Vanilloid receptor-related osmotically activated channel
(VR-OAC), a candidate vertebrate osmoreceptor. Cell 103: 525-535, 2000.
20. Matsuda H, Saigusa A and Irisawa H. Ohmic conductance through the inwardly
rectifying K channel and blocking by internal Mg2+. Nature 325: 156-159, 1987.
21. Mauerer UR, Boulpaep EL and Segal AS. Properties of an inwardly rectifying ATPsensitive K+ channel in the basolateral membrane of renal proximal tubule. J Gen
Physiol 111: 139-160, 1998a.
22. Mauerer UR, Boulpaep EL and Segal AS. Regulation of an inwardly rectifying
ATP-sensitive K+ channel in the basolateral membrane of renal proximal tubule. J
Gen Physiol 111: 161-180, 1998b.
23. Menconi MC, Pellegrini M and Pellegrino M. Voltage-induced activation of
mechanosensitive cation channels of leech neurons. J Membr Biol 180: 65-72.,
2001.
24. Mines GR. The action of beryllium, lanthanum, yttrium and cerium on the frog's
heart. J Physiol 40: 327-345, 1910.
25. Morris CE. Mechanosensitive ion channels. J Membr Biol 113: 93-107, 1990.
26. Noma A. ATP-regulated K+ channels in cardiac muscle. Nature 305: 147-148, 1983.
25
27. Noulin JF, Brochiero E, Lapointe JY and Laprade R. Two types of K(+) channels
at the basolateral membrane of proximal tubule: inhibitory effect of taurine. Am J
Physiol 277: F290-297, 1999.
28. Quasthoff S. A mechanosensitive K+ channel with fast-gating kinetics on human
axons blocked by gadolinium ions. Neurosci Lett 169: 39-42., 1994.
29. Sachs F and Morris CE. Mechanosensitive ion channels in nonspecialized cells.
Rev Physiol Biochem Pharmacol 132: 1-77, 1998.
30. Sackin H. Stretch-activated potassium channels in renal proximal tubule. Am J
Physiol 253: F1253-1262, 1987.
31. Sackin H. A stretch-activated K+ channel sensitive to cell volume. Proc Natl Acad
Sci U S A 86: 1731-1735, 1989.
32. Sackin H. Mechanosensitive channels. Annu Rev Physiol 57: 333-353, 1995.
33. Sackin H and Palmer LG. Basolateral potassium channels in renal proximal tubule.
Am J Physiol 253: F476-487, 1987.
34. Schultz SG. Membrane cross-talk in sodium-absorbing epithelial cells. In: The
Kidney: Physiology and Pathophysiology (2nd ed.), edited by Seldin DW and
Giebisch G. New York: Raven Press, Ltd., 1992, p. 287-299.
35. Segal AS, Boulpaep EL and Maunsbach AB. A novel preparation of dissociated
renal proximal tubule cells that maintain epithelial polarity in suspension. Am J
Physiol 270: C1843-1863, 1996.
36. Shyng SL and Nichols CG. Membrane phospholipid control of nucleotide sensitivity
of KATP channels. Science 282: 1138-1141., 1998.
26
37. Sigworth FJ and Sine SM. Data transformations for improved display and fitting of
single-channel dwell time histograms. Biophys J 52: 1047-1054, 1987.
38. Strotmann R, Harteneck C, Nunnenmacher K, Schultz G and Plant TD.
OTRPC4, a nonselective cation channel that confers sensitivity to extracellular
osmolarity. Nat Cell Biol 2: 695-702, 2000.
39. Suchyna TM, Johnson JH, Hamer K, Leykam JF, Gage DA, Clemo HF,
Baumgarten CM and Sachs F. Identification of a peptide toxin from Grammostola
spatulata spider venom that blocks cation-selective stretch-activated channels. J
Gen Physiol 115: 583-598, 2000.
40. Suzuki M, Sato J, Kutsuwada K, Ooki G and Imai M. Cloning of a stretchinhibitable nonselective cation channel. J Biol Chem 274: 6330-6335., 1999.
41. Thorn P and Petersen OH. Activation of nonselective cation channels by
physiological cholecystokinin concentrations in mouse pancreatic acinar cells. J Gen
Physiol 100: 11-25, 1992.
42. Tsuchiya K, Wang W, Giebisch G and Welling PA. ATP is a coupling modulator of
parallel Na,K-ATPase-K-channel activity in the renal proximal tubule. Proc Natl Acad
Sci U S A 89: 6418-6422, 1992.
43. Urbach V, Leguen I, O'Kelly I and Harvey BJ. Mechanosensitive calcium entry and
mobilization in renal A6 cells. J Membr Biol 168: 29-37, 1999.
44. Vanoye CG and Reuss L. Stretch-activated single K+ channels account for wholecell currents elicited by swelling. Proc Natl Acad Sci U S A 96: 6511-6516, 1999.
45. Yang XC and Sachs F. Block of stretch-activated ion channels in Xenopus oocytes
by gadolinium and calcium ions. Science 243: 1068-1071, 1989.
27
FIGURE LEGENDS
FIGURE 1.
Mechanosensitivity of the SA-NSC channel.
A) Slow application of
negative pressure reversibly activates a NSC channel. In this original recording from an
inside-out patch containing a single SA-NSC channel, the pipette contains CsCl, the
bath contains Na-Gluconate, and the command potential is –40 mV.
B) The open
probability of the SA-NSC channels increases with steady state negative pressure
(conditions as in A). The all channels closed is level (c) is indicated. C) Summary of the
stretch dose response of the SA-NSC channel for five excised patches (Boltzmann fit).
The effective pressure required for half-maximal activation (EP50) is –19.3 mmHg.
Symbols represent the mean („) ± SEM (bars). D) This multi-step protocol for a patch
containing one channel shows that although –20 mmHg is able to maintain the opened
channel in the open state, even –40 mmHg is not able to gate the closed channel open.
A subsequent step to –60 mmHg re-opens the channel (conditions as in A).
FIGURE 2.
The BLM SA-NSC channel is a weak inward rectifier. A) Representative
current records at various command potentials from an excised inside-out BLM patch
containing at least three SA-NSC channels. Negative pressure is held constant at –10
mmHg and conditions are symmetrical Na-Gluconate. The all channels closed level (c)
is indicated. B) Current-voltage relation for the BLM SA-NSC channel shown in A. The
rectification index (gin/gout) is 1.4. C) Closed and open dwell time histograms for a SANSC channel held at –40 mV with –50 mmHg applied. The dwell-time histogram is
logarithmically binned and plotted on a log abscissa and a square-root ordinate. The
theoretical curves represent probability density functions for a distribution of weighted
28
exponentials fit using the method of maximum likelihood (37). Channel behavior is
kinetically described with three closed states: τC1 = 0.628 ms (61.6% weight), τC2 = 15.9
ms (22.2% weight), τC3 = 177 ms (16.2% weight), and two open states: τO1 = 6.17 ms
(75.1% weight) and τO2 = 25.1 ms (24.9% weight).
FIGURE 3.
Ion selectivity of the SA-NSC channel.
A) Original current records
showing that the SA channel is nonselective for monovalent cations and also conducts
calcium. All traces are from excised inside-out patches in a Na-Gluconate bath at –40
mV with steady-state negative pressure applied. (inset) Current record from an excised
patch in symmetrical TEA-Cl at Vm = –140 mV with stretch applied (same scale as A).
Similar results from salt dilution experiments indicate that the channel has a finite
permeability to TEA+ (and NMDG+), albeit with a substantially lower conductance. B)
Steady-state I-V relationships for cation from each group are shown (NH4+, Li+, and
Ca2+).
FIGURE 4.
Open probability of the SA-NSC channel is voltage dependent.
A)
Normalized NPo (i.e., NPo/maximum NPo) increases with depolarization. Data are from
four inside-out membrane patches with CsCl in the pipette and a NaCl bath. Solid line
is a single exponential fit with a voltage constant of 49.5 mV. Symbols represent the
mean („) ± SEM (bars). B) Representative current records at command potentials of
±100 mV during a fast (5 ms) pressure step from 0 to –40 mmHg. Conditions are
inside-out basolateral membrane patch with Cs-Gluconate in the pipette and NaGluconate in the bath. C) Pooled latency data from 7 experiments as described in B.
29
Depolarization exerts a stronger effect on the lengthening of closing latencies (tLC) than
on the shortening of opening latencies (tLO). In both cases, the change in latency time is
significant. D) Current records at command potentials of ±80 mV during a mutli-step
pressure protocol. Conditions as described in B. See text for details.
FIGURE 5.
The SA-NSC channel is insensitive to gadolinium (Gd3+). A) Gd3+ (100
µM) in both the bath and pipette exerts no inhibitory effect on the stretch response of
the SA channel in this inside-out patch. Conditions are Cs-Gluconate pipette, NaGluconate bath and command voltage of –40 mV. B) As in A, except patch configuration
is outside-out and positive pressure is applied. C) Addition of 100 µM Gd3+ to an
outside-out patch conducting Na+ has no effect on channel activity. Channel currents
are markedly reduced when bath Na+ is replaced by TEA+. Conditions are CsCl pipette,
test cation bath ± 100 µM Gd3+, and command potential of –40 mV. Baseline pressure
is +25 mmHg.
FIGURE 6.
Spider venom from Grammostola spatulata completely inhibits SA-NSC
channels. Representative channel records from two cell-attached patches held at –40
mV with –50 mmHg applied. The bath was Na-Gluconate and the pipette contained
CsCl ± 10 µl venom per ml buffer. A) In the absence of pipette venom, channel activity
in cell-attached patches persists for the duration of the recording. Channel activity over
time (Po measured every 10 seconds) for this patch is summarized in the bar graph. B)
In the presence of pipette venom (the tip was front-filled and therefore initially venomfree), an SA-NSC channel with a high open probability over the first 90 seconds (top
30
trace) begins to exhibit longer dwells in a closed state (second trace) and closes
completely after 150 seconds (lower traces and bar graph), presumably coinciding with
diffusion of the venom to the membrane patch. Increasing negative pressure fails to reopen the channels. Similar results were observed in four consecutive cell-attached
experiments with patches containing up to four SA-NSC channels.
FIGURE 7.
Hypotonic stress activates the SA channel. A) Basolateral single channel
currents during 38% reduction in bath tonicity (by removal of sucrose) at constant ionic
strength (see Methods) and –40 mV command potential. A “priming” pressure of –20
mmHg is applied throughout the protocol.
The insets show representative current
traces from before, during, and after the hypotonic challenge. All channels closed level
(c) and open channel levels (o) are indicated. B) Summary of 6 experiments as in A.
Open probabilities are corrected for the number of channels in each patch.
31
FIGURE 8.
Adenine nucleotides reversibly inhibit the BLM SA channel.
A) In the
inside-out experiment depicted, inward currents carried by Cs+ at –60 mV are activated
by application of –17 mmHg, maintained throughout the experiment. After 5 mM UDP
fails to inhibit channel activity, 5 mM ADP reversibly blocks the channels. Subsequent
inhibition by 5 mM ATP is considerably more potent. The inhibitory effect of adenine
nucleotides is completely reversible. B) Dose-response curve for ATP inhibition shows a
Ki of ~0.5 mM. Symbols represent the mean (z) ± SEM (bars). C) Summary of 3
experiments illustrating Po independence of ATP block on excised inside-out patches.
Inhibition at Po ~0.2 is 88 ± 3.8% vs. 87 ± 1% block at Po ~ 0.7.