Erratum:Identification of a peptide toxin from Grammostola spatulata spider venom that blocks cation-selective stretch-activated channels.

Abstract

We have identified a 35 amino acid peptide toxin of the inhibitor cysteine knot family that blocks cationic stretch-activated ion channels. The toxin, denoted GsMTx-4, was isolated from the venom of the spider Grammostola spatulata and has <50% homology to other neuroactive peptides. It was isolated by fractionating whole venom using reverse phase HPLC, and then assaying fractions on stretch-activated channels (SACs) in outside-out patches from adult rat astrocytes. Although the channel gating kinetics were different between cell-attached and outside-out patches, the properties associated with the channel pore, such as selectivity for alkali cations, conductance ( approximately 45 pS at -100 mV) and a mild rectification were unaffected by outside-out formation. GsMTx-4 produced a complete block of SACs in outside-out patches and appeared specific since it had no effect on whole-cell voltage-sensitive currents. The equilibrium dissociation constant of approximately 630 nM was calculated from the ratio of association and dissociation rate constants. In hypotonically swollen astrocytes, GsMTx-4 produces approximately 40% reduction in swelling-activated whole-cell current. Similarly, in isolated ventricular cells from a rabbit dilated cardiomyopathy model, GsMTx-4 produced a near complete block of the volume-sensitive cation-selective current, but did not affect the anion current. In the myopathic heart cells, where the swell-induced current is tonically active, GsMTx-4 also reduced the cell size. This is the first report of a peptide toxin that specifically blocks stretch-activated currents. The toxin affect on swelling-activated whole-cell currents implicates SACs in volume regulation.

Full text

583
J. Gen. Physiol.
© The Rockefeller University Press
•
0022-1295/2000/05/583/16 $5.00
Volume 115 May 2000 583–598
http://www.jgp.org/cgi/content/full/115/5/583
Identification of a Peptide Toxin from
Grammostola spatulata
Spider
Venom that Blocks Cation-selective Stretch-activated Channels
Thomas M. Suchyna,* Janice H. Johnson,
‡
Katherine Hamer,
‡
Joseph F. Leykam,
§
Douglas A. Gage,
§
Henry F. Clemo,
储
Clive M. Baumgarten,
储
and
Frederick Sachs*
From the *Department of Physiology and Biophysics, State University of New York at Buffalo, Buffalo, New York 14214;
‡
NPS Pharma-
ceuticals, Inc., Salt Lake City, Utah 84108;
储
Department of Internal Medicine and Physiology, Medical College of Virginia, Virginia Com-
monwealth University, Richmond, Virginia 23298; and
§
Department of Biochemistry, Michigan State University–National Institutes of
Health Mass Spectrometry Facility, Michigan State University, East Lansing, Michigan 48824-1319
abstract
We have identified a 35 amino acid peptide toxin of the inhibitor cysteine knot family that blocks cat-
ionic stretch-activated ion channels. The toxin, denoted GsMTx-4, was isolated from the venom of the spider
Gram-
mostola spatulata
and has
⬍
50% homology to other neuroactive peptides. It was isolated by fractionating whole
venom using reverse phase HPLC, and then assaying fractions on stretch-activated channels (SACs) in outside-out
patches from adult rat astrocytes. Although the channel gating kinetics were different between cell-attached and
outside-out patches, the properties associated with the channel pore, such as selectivity for alkali cations, conduc-
tance (
ⵑ
45 pS at
⫺
100 mV) and a mild rectification were unaffected by outside-out formation. GsMTx-4 produced
a complete block of SACs in outside-out patches and appeared specific since it had no effect on whole-cell voltage-
sensitive currents. The equilibrium dissociation constant of
ⵑ
630 nM was calculated from the ratio of association
and dissociation rate constants. In hypotonically swollen astrocytes, GsMTx-4 produces
ⵑ
40% reduction in swell-
ing-activated whole-cell current. Similarly, in isolated ventricular cells from a rabbit dilated cardiomyopathy model,
GsMTx-4 produced a near complete block of the volume-sensitive cation-selective current, but did not affect the
anion current. In the myopathic heart cells, where the swell-induced current is tonically active, GsMTx-4 also re-
duced the cell size. This is the first report of a peptide toxin that specifically blocks stretch-activated currents. The
toxin affect on swelling-activated whole-cell currents implicates SACs in volume regulation.
key words:
mechanogated • swell • astrocyte • ventricular • myocytes
INTRODUCTION
High-affinity inhibitory peptide toxins have proven to
be powerful tools for elucidating the ion channel com-
ponents of whole-cell currents. Stretch-activated chan-
nels (SACs)
1
are the only major class of ion channels
for which a specific inhibitor does not exist. Gd
3
⫹
is the
best known blocker of SACs (
K
d
s ranging from 1 to 100
␮
M) and is widely used to identify these channels. How-
ever, Gd
3
⫹
also blocks a variety of other channels such
as L- and T-type Ca
2
⫹
(Biagi and Enyeart, 1990), K
⫹
de-
layed rectifier, voltage-gated Na
⫹
(Elinder and Arhem,
1994), and Ca
2
⫹
ER release channels (Kluesener et al.,
1995). Also, a variety of “specific” blockers for voltage-
and ligand-gated channels (e.g., amiloride, cationic
antiobiotics, tetrodotoxin, tetraethylammonium, quini-
dine, diltiazem, and verapamil) exhibit low affinity
blocking activity against SACs (for review, see Hamill
and McBride, 1996; Sachs and Morris, 1998).
SACs have been implicated as either activators or
modifiers of many different cellular responses to me-
chanical stimuli, including modification of electrical
and contractile activity of muscle tissue, involvement in
volume regulatory ion fluxes, and initiation of action po-
tentials in specialized sensory cells such as inner hair
cells of the cochlea and Merkel cells in the epithelium
(Sachs, 1992; Sachs and Morris, 1998; Tazaki and Suzuki,
1998). However, it has proven difficult to definitively as-
sociate mechanically stimulated physiological responses
with specific SACs in the absence of an inhibitor.
This limitation is nowhere more evident than in vol-
ume regulation studies, where cells undergo a regulated
volume decrease (RVD) in response to hypotonic stress.
RVD is produced by efflux of cytoplasmic inorganic os-
molytes (mainly K
⫹
and Cl
⫺
) and small organic mole-
cules. K
⫹
and Cl
⫺
efflux occurs via cotransporters and
individual conductive channels that are separate, but
interdependent (for reviews, see Chamberlin and
Strange, 1989; Hoffmann and Simonsen, 1989; Sarkadi
and Parker, 1991; Pasantes-Morales, 1996). RVD in as-
trocytes has been intensely studied due to its impor-
Address correspondence to Thomas M. Suchyna, Dept. of Physiology
and Biophysics, 320 Cary Hall, SUNY at Buffalo, Buffalo, NY 14214.
Fax: 716-829-2028; E-mail: suchyna@acsu.buffalo.edu
1
Abbreviations used in this paper:
CHF, congestive heart failure;
DIDS, 4,4
⬘
-diisothiocyanatostilbene-2,2
⬘
-disulphonic acid; ICK, inhib-
itor cysteine knot; RP, reverse phase; RVD, regulated volume de-
crease; SAC, stretch-activated channel; TFA, trifluoroacetic acid.

584
Peptide Blocker of Mechanical Transduction
tance in controlling brain edema (Kimelberg, 1995; Pas-
antes-Morales, 1996). These cells display a fast RVD re-
sponse and possess a high resting K
⫹
flux. Within the
first minute after hypotonic swelling of neonatal astro-
cytes (
ⵑ
160 mOsM reduction), the membrane poten-
tial depolarizes by
ⵑ
50 mV (Kimelberg
et al., 1990).
The depolarization is primarily caused by a rapidly acti-
vated 4,4
⬘
-diisothiocyanatostilbene-2,2
⬘
-disulphonic acid
(DIDS)–sensitive anion current (Pasantes-Morales
et al.,
1994; Bakhramov
et al., 1995). There is also a Ca
2
⫹
in-
flux during hyposmotic swelling that is partially medi-
ated by opening of dihydropyridine-sensitive Ca
2
⫹
chan-
nels (O’Connor and Kimelberg, 1993; Bender
et al.,
1994). It is important to remember that most agents
purported to be specific, such as dihydropyridines, have
not been tested against SACs.
A number of studies have suggested that nonselective
cation-permeable SACs play a role in both membrane
depolarization (Kimelberg and Kettenmann, 1990)
and Ca
2
⫹
influx observed during RVD (Christensen,
1987; O’Connor and Kimelberg, 1993; Chen
et al.,
1996). In addition, K
⫹
-selective SACs and a curvature-
sensitive nonselective cation channel have been de-
scribed in neonatal astrocyte cultures (Bowman
et al.,
1992; Islas
et al., 1993). A cation-selective SAC has also
been identified in C6 glioma cells (Bowman and Lohr,
1996). However, none of these channels have been in-
vestigated for involvement in RVD.
SACs have also been implicated in mechanical sensi-
tivity of the heart. Mechanical stimulation of cardiac
myocytes and whole heart preparations can cause depo-
larization, extrasystoles, and arrhythmias (see Hu and
Sachs, 1997). Also, chronic hemodynamic stress that
leads to congestive heart failure (CHF) and the accom-
panying cellular hypertrophy may be initiated by
stretch- or swelling-activated currents (Sachs, 1988;
Vandenberg
et al., 1996; Clemo and Baumgarten,
1997). CHF chronically activates a whole-cell cation-
selective current previously identified with hypotonic
swelling or chronic rapid pacing (Clemo
et al., 1998). A
blocker of SACs could have clinical use.
Chen
et al. (1996) showed that crude
Grammostola
venom could block SACs in an outside-out patch from
GH3 pituitary cells. They also demonstrated the venom
can block Ca
2
⫹
uptake during hypotonic swelling, but
not during high K
⫹
depolarization, which would activate
voltage-gated Ca
2
⫹
channels. Thus, it was suggested that
Ca
2
⫹
uptake was triggered by the activation of SACs. The
primary aim of this study was to isolate and characterize
the active component from
Grammostola
venom.
To isolate this SAC-blocking component(s), fractions
of the venom were screened by perfusion onto outside-
out patches from adult rat astrocytes, a preparation in
which SACs could be maintained active. A single com-
ponent peak was identified and sequenced, revealing a
unique peptide (noted GsMTx-4) containing an inhibi-
tor cysteine knot (ICK) consensus motif (Narasimhan
et al., 1994). The toxin exhibited negligible activity
against voltage-sensitive whole-cell currents. However,
the toxin did reduce swelling-activated whole-cell cur-
rents in astrocytes and CHF model cardiac myocytes.
The effect of this new toxin on whole-cell currents, for
the first time, directly implicates specific cation-selec-
tive SACs in the response to swelling.
METHODS
Toxin Isolation
Grammostola spatulata
(Theraphosidae) spiders were obtained
from a captive population at Hogel Zoo (Salt Lake City, UT).
The
Grammostola
species have recently been reassigned to the ge-
nus
Phixotricus
(Perez
et al., 1996), but
Grammostola
is used here
to maintain consistency with earlier biomedical publications.
Venom was collected by an electrical milking procedure (Bascur
et al., 1982) and stored at
⫺
80
⬚
C. It was fractionated by high-per-
formance liquid chromatography, incorporating Beckman Sys-
tem Gold 126 solvent delivery and 168 photodiode-array detector
modules (Beckman Instruments, Inc.), and using linear gradi-
ents with a flow rate of 3.5 ml/min unless noted. Whole venom
(825
␮
l) was separated into eleven 75-
␮
l aliquots that were each
diluted to 2 ml each with 15% solvent B. Solvent A was 0.1% trif-
luoroacetic acid (TFA) in water and solvent B was 0.1% TFA in
acetonitrile. The diluted venom was fractionated on a Zorbax
RX-C8 (9.4
⫻
250 mm, 5
␮
m, 300 Å; Mac-Mod Analytical, Inc.)
reversed-phase (RP) column equilibrated in 15% solvent B. The
column was developed with a 40-min gradient (15–55% solvent
B) begun 3 min after injection of the sample with a flow of 3.5
ml/min. The effluent was monitored at 280 nm and fractions
were collected as noted on the chromatogram (see Fig. 3 A).
Similar fractions from all 11 chromatographies were combined,
lyophilized, and tested for bioactivity. The assay samples were dis-
solved in 140 mM NaCl, 10 mM HEPES, 5 mM KCl, 2 mM
MgSO
4
, to a final dilution of 1:1,000, relative to whole venom, for
testing on outside-out patches. Several of the pools showed par-
tial block of the SACs (as described below), but only pool 9 gave
consistent, complete block of the channels.
Further purification of pool 9 (see Fig. 4 B) was achieved by
RP chromatography on a Vydac C18 column (10
⫻
250 mm, 5
␮
m, 300 Å; The Separations Group) equilibrated in 10% solution
B. Lyophilized pool 9 was dissolved in 4 ml of 10% solution B and
chromatographed in 1-ml portions eluting with a 10-min gradi-
ent (10–28% solution B), followed by a 64-min gradient (28–60%
solution B). The first gradient was begun 5 min after injection of
the sample, the effluent was monitored at 220 nm, and three
fractions were collected. Corresponding fractions from the four
chromatographies were combined, lyophilized, and assayed as
described above. Only pool B showed block of the SACs.
Attempts to affect further purification of pool B by anion and
cation exchange chromatography (MonoQ and MonoS resins;
Pharmacia LKB Laboratories, Inc.) were unsuccessful. The mate-
rial was not retained on the anion exchange column and re-
tained too strongly on the cation exchange columns. The pep-
tide did not elute with salt gradient from any of the cation ex-
change resins tried (0–2 M NaCl in 50 mM sodium phosphate
buffer, pH 7.0). A broad peak of material did elute at
ⵑ
pH 11
from the MonoS column with a gradient from pH 7–12 (50 mM
sodium phosphate buffer, 0.1 M in NaCl), but no resolution from
other components was visible.

585
Suchyna et al.
Therefore, pool B was subjected to a final RP chromatography
to remove a small amount of earlier and later eluting peptides.
Pool B was diluted to 4 ml with 20% solvent B and 0.5-ml por-
tions chromatographed on the Zorbax column described above,
eluting with a 7-min gradient (20–27% solution B), followed by a
46-min gradient (27–50% solution B), and the effluent was mon-
itored at 220 nm (see Fig. 3 C). The first gradient was begun 5
min after injection of the sample. The active peptide, GsMTx-4,
eluted between 29.5 and 30.5 min. Corresponding fractions from
the eight chromatographies were pooled to give 7.5 mg of
GsMTx-4. The average yield of GsMtx-4 from several purifica-
tions was 8 mg/ml of venom fractionated, which implies that the
toxin is
ⵑ
2 mM in whole venom. The purity of the final product
used in single channel and whole cell assays was assessed by ana-
lytical chromatography on an Aquapore RP300 C8 column (4.6
⫻
220 mm, 7
␮
m, 300 Å; PE Biosystems), eluting with a 25-min
linear gradient (32–47% solution B) with a flow of 1 ml/min
monitored at 220 nm (see Fig. 3 D). Elution with a gradient of
methanol/water (0.1% in TFA) gave a similar profile with a
longer retention time, but revealed no other impurities.
Mass Spectrometry
1
␮
l of the sample solutions (intact toxin or fragments) in 0.1%
TFA (or the HPLC elution solvent) were mixed on the sample
plate with 1
␮
l of a saturated solution of 4-hydroxy-
␣
-cyanocin-
namic acid in 1:1 CH
3
CN:0.1% aqueous TFA. The solution was
allowed to air dry before being introduced into the mass spec-
trometer. Spectra were acquired on a PerSeptive Biosystems Voy-
ager Elite MALDI-TOF (matrix-assisted laser desorption ioniza-
tion–time of flight) instrument operated in linear delayed extrac-
tion mode (50–100 ns). The instrument was equipped with a
nitrogen laser (3-ns pulse). The acceleration potential was 22 kV.
Sequencing
The toxin was further purified by microbore RP-HPLC (0.8
⫻
250 mm C18 column, with a linear gradient from 0.1% TFA-15%
CH
3
CN to 0.1% TFA-70% CH
3
CN in 90 min, flow rate 40
␮
l/
min, monitored at 214 nm). The toxin peak was collected at 24.6
min. The HPLC fraction (
ⵑ
1 nmol) was dried down and taken
up in 80
␮
l 8-M guanidine HCL-100 mM Tris-5 mM tributylphos-
phine, pH 8.5, and incubated for 8 h at 55
⬚
C.
N
-Isopropyliodoac-
tamide (1 mg in 20
␮
l MeOH
⫹
80
␮
l Tris) was added and the so-
lution was incubated for an additional 2 h at room temperature.
The reduced and alkylated peptide was then desalted by HPLC
on a C18 column as described above (elution time, 30.1 min).
NH
2
-terminal sequencing was carried out on an ABI 477 after
loading the reduced and alkylated peptide on PVDF membrane.
Digestion with BNPS [(2-2-nitrophenylsulfenyl)-3-methyl-3-bro-
moindolenine]-skatole (Fontana, 1972) was carried out by dis-
solving the purified reduced and alkylated peptide in 50
␮
l 0.1%
TFA and 15
␮
l BNPS-skatole. The solution was incubated at room
temperature for 8 h. The digestion products were separated by
HPLC as described above. Two main peaks were collected and se-
quenced by Edman degradation. Asp-N digestion (Wilson, 1989)
was performed by dissolving the purified reduced and alkylated
peptide in 100 mM Tris, pH 8.0 and treating with 1% (wt/wt)
Asp-N for 20 h at 35
⬚
C. The fragments were separated and ana-
lyzed by mass spectrometry before Edman degradation.
Astrocyte Cell Culture
Activated adult astrocytes, isolated from gelatin-sponge implants
from adult Sprague-Dawley rat brains (Langan et al., 1995), were
provided courtesy of Dr. Thomas Langan (SUNY Buffalo, Buf-
falo, NY) at passages 2–4. Astrocytes were maintained in DMEM,
10% fetal bovine serum, and 1% penicillin/streptomycin and
were used in experiments between 2 and 5 d after passage. Cells
between passages 4 and 35 expressed SACs with the same proper-
ties. Both stellate and polygonal (fibroblast-like) cells were used.
Astrocyte Single-Channel Patch Clamp
Patch voltage was controlled by an Axopatch 200B (Axon Instru-
ments) and stored directly on computer disk via a Labmaster
DMA version B (Scientific Instruments) board controlled by
pClamp6-Clampex acquisition software (Axon Instruments). Cur-
rents were sampled at 10 kHz and low-pass filtered at 2 kHz
through a four-pole Bessel filter on the Axopatch 200B. Experi-
mental voltage protocols were controlled by pClamp6-Clampex.
All potentials are defined with respect to the extracellular surface.
Electrodes were pulled on a pipette puller (PC-84; Brown-
Flaming Instruments), painted with Sylgard 184 (Dow Corning
Corp.) and fire polished. Electrodes were filled with KCl saline
containing (mM): 140 KCl, 5 EGTA, 2 MgSO
4
, 10 HEPES, pH
7.3) and had resistances ranging from 3 to 8 M
⍀
. Bath saline con-
sisted of (mM): 140 NaCl, 5 KCl, 1 MgSO
4
, 1 CaCl
2
, 6 glucose,
and 10 HEPES, pH 7.3.
Pressure and suction were applied to the pipette by a pressure
clamp designed and constructed in our laboratory by Dr. Steven
Besch. Pressure values refer to pressure in the pipette; i.e., the in-
tracellular side of the membrane in outside-out patches. Suction
applied to a cell-attached patch has the same sign as pressure ap-
plied to an outside-out patch. The rise time of pressure changes at
the tip were determined by monitoring the rate of current change
when pressure steps were applied to an electrode containing 150
mM KCl solution and placed in a water bath. The
␶
10–90
was
ⵑ
5
ms, as determined by exponential fits to the current decay. Perfu-
sion of the patch was handled by a pressurized bath perfusion sys-
tem with eight separate channels (BPS-8; ALA Scientific).
Offline data analysis was performed with pClamp6 analysis
software and Origin 5.0. Maximal unitary channel currents were
determined via Gaussian fits to the peaks of the all-points ampli-
tude histograms from records containing one to three channels.
Many current records displayed more than three channel open-
ings (maximal single channel currents plus subconductance
states) and were impossible to fit using Pstat software. Some of
these records were analyzed by determining all step-like changes
in the current during the pressure application and selecting the
average maximal current level as the unitary current. The data
analyzed by this method was in good agreement with the unitary
current levels determined by analysis of all points amplitude his-
tograms from single-channel patches.
Astrocyte Whole-Cell Current Clamp
Whole-cell current was measured by the Nystatin-perforated
patch technique (Horn and Marty, 1988). Bath saline was the
same as above. Pipette saline consisted of (mM): 80 KCl, 30
K
2
SO
4
, 10 NaCl, 3 MgSO
4
, 0.13 CaCl
2
, 0.23 EGTA, and 10 HEPES,
pH 7.3. Nystatin was dissolved in pipette saline to a final concen-
tration of 200
␮
g/ml. After patch formation, access resistance was
allowed to drop to ⵑ15 M⍀ (uncompensated), after which the se-
ries resistance compensation was set at ⵑ65%, and prediction was
set to ⵑ75%. Whole-cell capacitance measurements ranged from
ⵑ25 to 50 pF. Whole-cell currents were monitored by either a
voltage-step protocol (see Fig. 9), or by 600-ms voltage ramps.
During hypotonic swelling, the cell was perfused initially with iso-
tonic saline (bath saline with 160 mM mannitol replacing 80 mM
NaCl) before switching to hypotonic saline (isotonic saline minus
140 mM mannitol). The BPS-8 perfusion system described above
was used to rapidly (⬍200 ms) change the bathing solution. Peak
currents were measured at 3–5 ms into voltage steps.

586 Peptide Blocker of Mechanical Transduction
Cardiac Myocyte Preparation
Ventricular myocytes were freshly isolated from New Zealand
white rabbits with aortic regurgitation-induced congestive heart
failure (Magid et al., 1988) using a collagenase–pronase disper-
sion method (Clemo and Baumgarten, 1997). Cells were stored
in a modified Kraft-Bruhe solution containing (mM): 132 KOH,
120 glutamic acid, 2.5 KCl, 10 KH
2
PO
4
, 1.8 MgSO
4
, 0.5 K
2
EGTA,
11 glucose, 10 taurine, 10 HEPES, pH 7.2. Myocytes were used
within 6 h of harvesting and only quiescent cells with no evi-
dence of membrane blebbing were selected for study.
Cardiac Myocyte Electrophysiology and Volume Determination
For a detailed explanation of methods, see Clemo and Baumgar-
ten (1997). In brief, electrodes were pulled from glass capillaries
to give a final tip diameter of 3–4 ␮m and a resistance of 0.5–1
M⍀ when filled with the standard electrode filling solution con-
taining (mM): 120 K aspartate, 10 KCl, 10 NaCl, 3 MgSO
4
, 10
HEPES, pH 7.1. Whole-cell currents were recorded using an
Axoclamp 200A. Pulse and ramp protocols, voltage-clamp data
acquisition, and offline data analysis were controlled with custom
programs written in ASYST. Both step and ramp voltage-clamp
protocols were applied with a holding potential of ⫺80 mV. Cur-
rents were digitized at 1 kHz and low-pass filtered at 200 Hz.
Whole-cell currents were recorded using the amphotericin per-
forated-patch technique. Solution changes were performed by
bath perfusion that was completed within 10 s. The standard
bath solution contained (mM): 65 NaCl, 5 KCl, 2.5 CaSO
4
, 0.5
MgSO
4
, 10 glucose, and 10 HEPES, pH 7.2, and 130 (1T) or 283
(1.5T) mannitol to control the osmolarity. Isotonic osmolarity
was taken as 296 (1T) and 444 (1.5T) mosm for hypertonic solu-
tion. Myocyte volume was determined by visualization with an in-
verted microscope (Diaphot; Nikon Inc.) equipped with Hoff-
man modulation optics and a high-resolution TV camera cou-
pled to a video frame grabber. Images were captured online each
time a ramp or step voltage-clamp protocol was performed using
a program written in C and assembler and linked to ASYST volt-
age-clamp software. A combination of commercial (MOCHA;
SPSS Inc.) and custom (ASYST) programs were used to deter-
mine cell width, length, and area of the image.
RESULTS
SAC Properties in Activated Adult Rat Astrocytes:
Cell-attached Versus Outside-Out
The most efficient method for screening multiple
venom samples is to use outside-out patches. However,
first it was necessary to show that SAC function is main-
tained in the outside-out patch configuration. To date,
all studies investigating the properties and function of
SACs in astrocytes have focused on neonatal prepara-
tions and C6 glioma cell lines (Bowman et al., 1992;
Bowman and Lohr, 1996; Islas et al., 1993) due to diffi-
culties in culturing adult astrocytes. Langan et al.
(1995) showed that adult-activated rat astrocytes from
gel foam implants could be isolated and cultured
through ⬎10 passages, while maintaining all type 1 as-
trocytic markers [glial fibrillary acidic protein (GFAP),
vimentin, 7B11, and RAN2], contact inhibition, and
cell cycle kinetics. In cell-attached patches, activated
adult astrocytes express primarily one type of SAC that
can be activated by both pressure and suction (Fig. 1 A,
only pressure data shown). Observation from ⬎100
patches typically showed two to five channels/patch.
Consistent with the results of Hamil and McBride
(1992) and Small and Morris (1995), SAC activity in
cell-attached patches was sensitive to the level of suc-
tion used in seal formation. Channels were rarely ob-
served when ⬎10 mmHg of suction was used during
seal formation, whereas ⬎90% of patches showed chan-
nel activity with ⬍10 mmHg. With 140 mM KCl in
the electrode, the single-channel conductance inwardly
rectified being 46 pS at ⫺100 mV, but only 21 pS at
⫹100 mV (Fig. 1 C). Channel activity was normally ini-
tiated by applying between 25 and 35 mmHg of suc-
tion. However, rundown did occur so that increasing
levels of suction were required to activate the channels
over the 5–10 min during which data was acquired.
The open probability (P
o
) was time and voltage de-
pendent, displaying a fast adaptation (within 100 ms at
hyperpolarized potentials) similar to that reported for
Xenopus oocytes (Hamill and McBride, 1992). The time
dependence of P
o
can be described by an initial phasic
period followed by a tonic period, as defined in Bow-
man and Lohr (1996). Both the duration of the phasic
period and P
o
during the tonic period showed a steep
voltage dependence, decreasing as the potential be-
came more negative (Fig. 1 A, see average currents). Of
16 cell-attached patches analyzed, 12 displayed adapta-
tion at hyperpolarized potentials. In addition to adapta-
tion, multiple voltage-dependent substates were visible
at ⫺100 mV, compared with only one at depolarizing
potentials. Kinetic analysis of these substates is cur-
rently in progress. One major subconductance state ob-
served at positive potentials (not shown in this figure)
was 80–90% of the maximal conductance state, similar
to that reported in Xenopus oocytes (Silberberg and
Magleby, 1997). Adaptation was also observed in C6
glioma cells (Bowman and Lohr, 1996). However, in
contrast to the rapid adaptation observed in adult as-
trocytes, adaptation in C6 cells occurred over 1–2 s and
was not as strongly voltage dependent.
Channel activity in outside-out patches was generally
similar to that in cell-attached patches, but they had dif-
ferent adaptation properties (Fig. 1 B). The SACs
opened in response to both pressure and suction. With
140 mM KCl in both the pipette and the bath, the I-V
profile (44 pS at ⫺100 mV, and 21 pS at ⫹100 mV, cyto-
plasmic side) was nearly identical to that observed for
cell-attached patches (Fig. 1 D). In this configuration,
the channels were initially activated by between 30 and
40 mmHg of pressure. The similarities between the
conductance and pressure sensitivity in the two patch
configurations suggest that these channel properties
have not been significantly modified by outside-out
patch formation. However, of 12 outside-out patches,
only one displayed the fast adaptation property ob-

587 Suchyna et al.
served in cell-attached patches. Instead, two showed no
change in P
o
with respect to time or voltage, while the
remaining nine patches exhibited a slow increase in
current at both positive and negative voltages, where
the number of active channels increased during the
500-ms pressure step (Fig. 1 B, 100 mV, and see Fig. 5
A, average control current). The rate of increase was
greater for pressure steps at positive voltages due to an
increase in P
o
at positive potentials. Similar responses
are observed in Xenopus oocytes when large pressure
stimuli are applied to eliminate the adaptation prop-
erty (see Figure 2 in Hamill and McBride, 1992). Out-
side-out formation is likely to disrupt the associations
between membrane and cytoplasmic attachments in a
similar way to the high pressure stimuli applied to cell-
attached patches of the Xenopus oocyte.
The single channel conductance and inward rectifica-
tion observed here were similar to the properties re-
ported for other members of the family of nonselective
cation SACs (for review, see Yang and Sachs, 1993).
Figure 1. Cell-attached and outside-out patches from adult activated astrocytes showing stretch-sensitive channels with similar unitary
conductance profiles but different gating properties. Representative single-channel current recordings are shown above average patch
currents from a cell-attached patch (A) containing a single channel, and an outside-out patch (B) containing two to three channels. Cell-
attached patch recordings were made with 140 mM KCl pipette saline, and outside-out patch recordings are with symmetrical 140-mM KCl
pipette solutions. Pressure steps (indicated by the bar at the top) were applied to the patches at different holding potentials shown to the
left of each recording. Voltages are relative to the extracellular side. Average current records were calculated from multiple pressure steps
(ranging from 5 to 15 steps) at each voltage. In cell-attached mode, channel adaptation, lower P
o
, and multiple subconductance states are
apparent at negative potentials. Channels in outside-out patches from astrocytes show slow voltage-dependent activation and lower P
o
at
negative potentials. Unitary current–voltage plots were fitted with a second-order polynomial and show inward rectification for channels in
cell-attached (C, n ⫽ 11) and outside-out (D, n ⫽ 16) patches. Voltages for cell-attached data points were corrected for the average resting
membrane potential measured in the whole cell configuration. Each point represents an average current calculated by applying multiple
pressure steps to a single patch.

588 Peptide Blocker of Mechanical Transduction
Like other nonselective SACs, the astrocyte channels
were not blocked by 130 mM Cs
⫹
(data not shown).
The channels were cation selective based on external
ion substitution experiments with outside-out patches.
When gluconate (Fig. 2, 䉬) was substituted for Cl
⫺
(䊐),
we saw no change in the I-V profile. However, substitut-
ing NMDG
⫹
(䉲) for K
⫹
in the bath produced an 88%
reduction in current at ⫺100 mV. The channel dis-
played a weak selectivity for K
⫹
over Na
⫹
since the cur-
rent was reduced 24% at ⫺100 mV when Na
⫹
was sub-
stituted for K
⫹
in the bath (Fig. 2, 䉭). The properties of
this SAC in adult astrocytes were similar to the cation-
selective SACs described for many other cell types, includ-
ing C6 glioma cells. SACs in C6 cells display a unitary
conductance of 40 pS in equimolar 100-mM KCl are in-
wardly rectifying and show voltage-dependent adapta-
tion (Bowman and Lohr, 1996). However, SACs in our
adult astrocyte differed from the highly K
⫹
-selective
channels (Islas et al., 1993) and from the ⵑ50-pS curva-
ture-sensitive channels (Bowman et al., 1992) reported
in neonatal astrocytes preparations.9
Astrocyte SAC Pharmacology
To characterize the SACs of adult astrocyte, we exam-
ined a number of compounds purported to be active
against SACs in other systems (Table I). Gd
3⫹
, the most
commonly used SAC reagent, completely blocked the
channels in adult astrocytes at 50 ␮M. Amiloride, which
blocks epithelial Na
⫹
channels with nanomolar affinity,
and the endogenous SACs in oocytes and audiovestibu-
lar hair cells (Jorgensen and Ohmori, 1988; Lane et al.,
1991), had no effect at 20 ␮M.
Although NMDA-type glutamate receptors have not
been observed in astrocytes, these channels do show
stretch sensitivity in macropatches from neonatal neu-
rons (Paoletti and Ascher, 1994). In neonatal astro-
cytes, the potent (nanomolar affinity) NMDA open-
channel blocker MK801 inhibits glutamate-stimulated
K
⫹
influx (Bender et al., 1998). Hence, we tested
glutamate and MK801 on the adult astrocyte SACs.
Glutamate at 10 ␮M did not activate SACs in un-
stretched patches that contained SACs, and 20 ␮M
MK801 did not block SAC opening when stretched.
Some reports indicate that L-type Ca
2⫹
channels are
stretch sensitive (Langton, 1993; Ben-Tabou et al.,
1994) and that blockers of these channels are active
against SACs (Ruknudin et al., 1993; Small and Morris,
1995). Neither 100 ␮M diltiazem nor 50 ␮M nifedipine
showed any blocking activity against the SACs in adult
astrocytes. The anion channel blocker DIDS produced
a significant reduction in the amount of swelling-acti-
vated anion current in astrocytoma cells (Bakhramov et
al., 1995), but 50 ␮M DIDS had no effect on stretch-
activated currents in the patch.
Identification and Characterization of a SAC Blocking Toxin
from Grammostola Venom
HPLC fractions of Grammostola (Gs) were lyophilized,
redissolved at a 1:1,000 dilution and perfused onto out-
side-out patches. Fraction 9, on the Gs whole-venom
Figure 2. External ion substitution shows SACs are cation selec-
tive in outside-out patches. The pipette saline was 140 mM KCl.
Each data point represents the average current from multiple
pressure steps to a single patch. The data sets were fitted with sec-
ond-order polynomials. Data for symmetrical 140 mM KCl (䊐,
solid line fit) are the same as shown in Fig. 1 D. Switching from
140 mM KCl to 140 mM NaCl (䉭, dotted line fit) in the bath re-
duced the conductance at ⫺100 mV from 44 to 33 pS, while there
w
as no change in conductance at ⫹100 mV (n ⫽ 10 patches). This
suggests a weak selectivity for K
⫹
over Na
⫹
. When external Cl
⫺
is
replaced with a less permeable anion gluconate (䉬, dashed line
fit), there was a negligible reduction in conductance at ⫺100 mV
(41 pS), and no change in conductance at ⫹100 mV (n ⫽ 4
patches). However, when external K
⫹
is replaced with the imper-
meant cation NMDG (䉲, dash-dot line fit), the conductance was
reduced to 5 pS at ⫺100 mV (n ⫽ 3 patches).
TABLE I
Pharmacology of Astrocyte SACs
Compound n Concentration Activity
␮
M
Gd
⫹3
2 50 Block
Amiloride 2 20 No block
MK801 2 20 No block
Glutamate 2 10 No activation
Diltiazem 3 100 No block
Nifedipine 3 50 No block
DIDS 3 50 No block
Effects of selected ion channel ligands on adult astrocyte SACs.
Compounds listed were perfused, at the concentration shown, onto
outside-out patches held at ⫺50 mV, while channels were activated by
pressure. n, number of trials.

589 Suchyna et al.
chromatogram (Fig. 3 A) blocked the SACs. This frac-
tion was further resolved on two slower gradients (Fig.
3, B and C) at 0.5% change in acetonitrile/min until a
single peptide peak was identified containing the activ-
ity (Fig. 3 D). The amount of isolated peptide was de-
termined by weight and three different protein spectro-
scopic methods, showing that a 1:1,000 dilution corre-
sponds to a concentration of 8 ␮g/ml whole venom.
The peak was determined to have a molecular weight
of 4,093.90 by mass spectrometry and designated
GsMTx-4 (for Grammostola mechanotoxin #4). Other
peptides have been isolated from Gs venom that are ac-
tive against SACs (Sachs, F., unpublished observations,
and GsMTx-1 U.S. Patent #5756663); however, GsMTx-4
showed the most consistent and potent activity. At this
concentration, the block was complete, and occurred
rapidly upon superfusion of the patch, as shown by rep-
resentative current traces in Fig 4.
The association rate of the toxin was determined by
applying toxin to an outside-out patch while the chan-
nels were activated by stretch. In the absence of GsMtx-4,
channel activity increased over time at constant pres-
sure (compare Fig. 1 B with 5 A). When 5 ␮M toxin was
perfused onto the patch, 1 s after the initiation of the
Figure 3. Reverse-phase HPLC chromatograms showing sequential purification steps for identification of GsMTx-4 peptide. The per-
cent acetonitrile that corresponds to specific venom peaks is indicated by the dotted line shown overlaying each chromatogram. A chro-
matogram of Grammostola whole venom (A) produced by a 40-min linear gradient from 15 to 55% acetonitrile at a flow rate of 3.5 ml/min.
1–11, labeled at the bottom, designate fractions pooled for testing on outside-out patches. The lines within the chromatogram mark the
boundaries of each fraction. Fraction 9 (A) contained SAC blocking activity and was further fractionated in B. (B) Only fraction B showed
SAC blocking activity and was further purified in C. 10 ␮g of the final material used in all experiments was run on a 25-min 32–47% linear
gradient of acetonitrile at a flow rate of 1 ml/min (D).

590 Peptide Blocker of Mechanical Transduction
pressure step, the current decayed exponentially (Fig.
5 A, GsMTx-4). When the control and GsMTx-4 aver-
age current records are superimposed, before GsMTx-4
application, the currents are nearly identical (Fig. 5 B).
The difference current was calculated (Fig. 5 C), and
the period of GsMTx-4 application was fitted with a sin-
gle exponential (Fig. 5 D), yielding a time constant of
594 ⫾ 10 ms. Assuming a 1:1 binding, this gives an asso-
ciation constant, k
A
, of 3.4 ⫻ 10
5
M
⫺1
s
⫺1
.
To determine the dissociation rate, we fit the increase
in average patch current (n ⫽ 7 patches) during toxin
washout. Fit to a single exponential, the washout time
constant was 4.7 ⫾ 1.7 s (Fig. 6 B). From this dissocia-
tion constant (k
d
⫽ 0.21 s
⫺1
) and the association con-
stant determined above (k
a
⫽ 3.3 ⫻ 10
5
M
⫺1
s
⫺1
), the
calculated equilibrium constant, K
d
⫽ k
d
/k
a
⫽ 631 ⫾
240 nM (standard error calculated from the first-order
approximation using the errors of k
a
and k
d
). Using the
ratio of rate constants to evaluate K
d
minimizes errors
caused by rundown. However, the K
d
calculated from
the mean currents was similar. The mean SAC current
was 2.04 ⫾ 0.14 pA (SEM) over 11 pressure steps before
GsMTx-4 application (Fig. 6 A), and fell to 0.17 ⫾ 0.02
pA during toxin perfusion. (The average current over
the last eight pressure steps, 10 s after GsMTx-4 wash-
out, returned to the initial current level of 2.28 ⫾ 0.17
pA.) For a single binding site, Michaelis-Menten kinet-
ics predicts the ratio of the blocked to the unblocked
current is I/I
0
⫽ 1(1 ⫹ K
d
/S), where S is the substrate
(toxin) concentration and K
d
is the equilibrium dissoci-
ation constant. Using the data from Fig. 6, I/I
0
⫽ 0.083,
which gives a binding constant K
d
⫽ 415 nM, consistent
with the value calculated from the ratio of association
and dissociation rates.
Determining the specificity of a pharmacologic agent
is an unending project, and defined only for the sys-
tems tested. We tested the pseudo–steady state I-V rela-
tionship as it related to voltage-sensitive channels. Us-
ing the perforated patch technique, 5 ␮M GsMTx-4 did
not significantly change the I-V profile (Fig. 7 A), sug-
gesting that it did not interact with slow, voltage-depen-
dent, channels. By comparison, 5 mM CsCl, which was
shown not to effect SACs in patches, produced a signifi-
cant decrease in current at hyperpolarized potentials
(Fig. 7 B), where a Cs
⫹
-sensitive inward-rectifying K
channel is known to be activated (Sontheimer, 1992).
Sequence and Disulfide Structure of GsMTx-4
MALDI-MS analysis showed the molecular weight of
the native toxin was 4,093.90 (MH⫹ ion). The alky-
lated and reduced toxin displayed a peak at mass/
charge 4,690, indicating three disulfide bonds or six
cysteine residues were present. NH
2
-terminal sequenc-
ing was followed by sequencing of two different
COOH-terminal fragments produced by enzymatic di-
gests with BNPS-skatole and Asp-N. The peptide se-
quence produced had a predicted mass 4,022.86, which
is 71.04 D less than that measured for the intact toxin.
This difference supports the presence of a COOH-ter-
minal alanine (71.09 D), even though alanine was
clearly absent in the last cycle of the Asp-N digestion
product. The mass accuracy of the MALDI-MS analysis
is approximately ⫾0.5 D with internal calibration. The
final sequence shown in Fig. 8 is 35 amino acids in
length with the C-term alanine added.
The six cysteine residue included in boxes form an
ICK motif (CX
3-7
CX
3-6
CX
0-5
CX
1-4
CX
4-13
C) commonly ob-
served in many other peptide toxins from both terrestrial
and aquatic animal venoms (Narasimhan et al., 1994;
Norton and Pallaghy, 1998). GsMTx-4 shows ⬍50% ho-
mology to any other peptide toxin. Other tarantula tox-
ins that block voltage-gated Ca
2⫹
and K
⫹
channels show
Figure 4. GsMTx-4 blocks SACs in outside-out patches. SAC ac-
tivity in response to pressure steps applied to an outside-out patch
before and during GsMTx-4 application, and after washout are
shown. The patch was held at ⫺50 mV, and the pressure pulse is
shown above the records. The entire experiment is comprised of
60 pressure steps: steps 1–20 occur before GsMTx-4 application,
21–38 while GsMTx-4 is being perfused, and 39–60 occur during
w
ashout. Each 500-ms pressure step was separated by 1.5 s at 0
pressure. Four representative records from each stage of the ex-
periment are displayed.

591 Suchyna et al.
the highest percentage of similarity to GsMTx-4, as illus-
trated by the amino acid alignment in Fig. 8. A K
⫹
chan-
nel toxin labeled protein 5 from Brachypelma smithii (Kai-
ser et al., 1994; Norton and Pallaghy, 1998) shows ⵑ50%
total sequence similarity. The most significant regions of
homology occur within the cysteine motif. Besides the
conserved cysteine motif, there are three other residues
(F4, D13, and L20) that are conserved in all five toxins.
Like the positively charged ␻-conotoxin and ␻-agatoxin
families of Ca
2⫹
channel blockers, GsMTx-4 carries an
overall positive charge (⫹5).
GsMTx-4 Effects on Astrocyte Whole-Cell
Swelling-activated Currents
A large conductance increase occurs upon hypotonic
swelling of neonatal astrocytes. Part of this current may
Figure 5. GsMTx-4 rate of blocking determined by superfusion of activated SACs in outside-out patches. Average SAC currents calcu-
lated from 3-s pressure steps are indicated by the bars above the traces (A). The control trace was generated from 37 pressure steps applied
to seven different patches held at ⫺50 mV, with pressure levels ranging from 35 to 70 mmHg. The current increased exponentially over
the 3-s pressure application. The GsMTx-4 response was produced by applying 5 ␮M toxin 1 s after the onset to the pressure step indicated
by the GsMTx-4 bar. The GsMTx-4 current record was averaged from 29 pressure steps to six different patches held at ⫺50 mV, with the
steps ranging between 38 and 80 mmHg. Currents were nearly identical over the first second of the average current records, as shown
w
hen the two are superimposed in B. Subtracting the control current trace from the GsMTx-4 trace produced the difference current in C.
The current trace during GsMTx-4 application was fitted with a single exponential yielding a time constant of 594 ⫾ 10 ms (D). The fit is
shown displaced from the data for clarity.
Figure 6. The GsMTx-4 disso-
ciation rate was determined from
the recovery rate of SAC current
on washout. SAC currents were
activated by 500-ms pressure
steps at 2-s intervals in outside-
out patches held at ⫺50 mV. (A)
average current (䊏 ⫾ SEM)
from seven different patches.
Channel current drops to the
noise level rapidly upon applica-
tion of toxin, and shows a slow
recovery to the initial current
level upon toxin washout. The
recovery kinetics were fitted to a
single exponential with a time
constant of 4.7 ⫾ 1.7 s (B).

592 Peptide Blocker of Mechanical Transduction
be due to nonselective cationic SACs (Kimelberg and
Kettenmann, 1990). However, a large rapidly develop-
ing DIDS-sensitive anionic current dominates the pas-
sive membrane current (Pasantes-Morales et al., 1994;
Bakhramov et al., 1995). After 30-s exposure to hypo-
tonic conditions, adult astrocytes display a similar large-
conductance increase that slowly inactivated at large
depolarizing voltages [compare Fig. 9 A (resting cur-
rent) to B (swelling-activated current)]. During hypo-
tonic exposure, cells were held at ⫺50 mV before I-V
test voltage steps to reduce the influence of voltage-
gated Ca
2⫹
channels on Ca
2⫹
influx. The swelling-acti-
vated current has a large anionic component since 50
␮M DIDS produced a significant reduction in current
(especially at depolarized potentials) and a ⫺33-mV
shift in reversal potential (Fig. 10, n ⫽ 6). A residual
current with a reversal potential shifted toward E
K
re-
mained. Applying 5 ␮M GsMTx-4 while hypotonically
swelling the cell significantly reduced the peak current
response at 30 s after hypotonic exposure (Fig. 9 C). Af-
ter washout of GsMTx-4, a hypotonic stimulus pro-
duced larger swelling-activated currents, although less
than the original control stimulus (Fig. 9 D). This re-
duced response after washout is not due to lingering
toxin effects, since ⬎3 min of washout separated suc-
cessive hypotonic stimuli. The response to successive
hypotonic exposures slowly decreased over time (Fig. 9
F, a and b), probably due to RVD mechanisms. Repre-
sentative peak-current responses from two different
cells displayed a roughly linear decrease in swelling-
activated current (Fig. 9 F, 䊏). GsMTx-4 always reduced
the swelling-activated current from the control re-
sponse (Fig. 9 F, 䉫). In light of the slowly degrading hy-
potonic response, to estimate the amount of GsMTx-4
block, we had to correct for the “rundown” by linear in-
terpolation. The I-V profiles for the swelling-activated
difference currents (Fig. 9 G) show a clear difference
between the before (䊏) and after (䊉) responses. The
percent block produced by GsMTx-4 (䉫) relative to
each of the control curves is shown to the right. The es-
timated reduction in swelling-activated current pro-
duced by 5 ␮M GsMTx-4 was similar at both hyperpo-
larizing and depolarizing potentials (ⵑ48% at ⫺100
mV and ⵑ38% at ⫹100 mV). Furthermore, unlike
DIDS, which produced a large (⫺33 mV) shift in rever-
sal potential due to the specific loss of anionic current,
GsMTx-4 produces almost no change in reversal poten-
tial (⫹2 mV, statistically indistinguishable from 0 mV).
GsMTx-4 Effects on CHF Model Ventricular Myocytes
Whole-Cell Currents
Swelling-activated currents in rabbit and dog cardiac
myocytes are persistently activated in CHF and may
play a role in the development of congestive heart fail-
ure (Clemo and Baumgarten, 1997; Clemo et al.,
1998). Both an inwardly rectifying cation selective cur-
Figure
7
. GsMTx-
4
d
oes not
significantly affect voltage-sensi-
tive currents. Whole-cell cur-
rents are shown from astrocytes
voltage clamped using the perfo-
rated-patch technique. (A) Aver-
age whole-cell current from six
cells, produced by a 600-ms ramp
from ⫺120 to 80 mV. There is
no significant difference between
whole-cell currents in isotonic sa-
line (䉬) and currents measured
between 30 and 120 s after perfu-
sion with 5 ␮M GsMTx-4 (䉫).
(B) Peak currents from a single
cell stepped in 20-mV increments
between ⫺120 and 120 mV in
the absence (䉬) or presence (䉫)
of 5 mM CsCl. In contrast to
GsMTx-4, CsCl produces a significant decrease in current at hyperpolarized potentials. After the first application of CsCl, the cell was
w
ashed and the experiment was repeated. Thus, the I-V plot shows two sets of data points for both control and CsCl.
Figure 8. Sequence of GsMTx-
4
showing homology to other ion
channel peptide toxins. Cysteine
motif residues are included in
boxes. Dark shaded residues in
the comparison peptide sequences are identical to GsMTx-4, while lighter shaded residues are similar. TXP5, K
⫹
channel blocker (Kaiser
et al., 1994): 40% identity, 54% similarity; SNX-482, blocks E-type Ca
2⫹
channels (Newcomb et al., 1998): 40% identity, 49% similarity;
␻-GramTX S1A, blocks N-, P-, and Q-type Ca
2⫹
channels (not L-type) (Lampe et al., 1993): 34% identical, 46% similarity; Hanatoxin, K
⫹
channel blocker (Swartz and MacKinnon, 1995): 28% identical, 37% similarity.

593 Suchyna et al.
rent (I
Cir,swell
) and an outwardly rectifying Cl
⫺
current
(I
Cl,swell
) are involved in volume regulation in cardiac
myocytes. In ventricular myocytes from rabbits with
aortic regurgitation-induced congestive heart failure,
these two currents are constitutively activated in iso-
tonic media (Fig. 11 A, a) and are inactivated when the
cell is perfused with hypertonic saline (b). At 0.4 ␮M,
GsMTx-4 produced a nearly complete block of the in-
ward I
Cir,swell
, but had no effect on the outward I
Cl,swell
(Fig. 11 B, c). However, whole-cell current was unaf-
fected by GsMTx-4 when swelling-activated currents
were inactivated by 1.5T hypertonic saline (Fig. 11 B,
d). The difference currents in Fig. 11 C show that
GsMTx-4 blocked only inward swelling-activated cur-
rent [compare Fig. 11 C, 1.0T
C
⫺ 1.5T
C
(total I
Cir,swell
)
with 1.0T
C
⫺ 1.0T
Tx
(toxin sensitive I
Cir,swell
)]. The re-
Figure 9. GsMTx-4 reduces whole-cell swelling-activated current in astrocytes exposed to hypotonic saline. Whole cell currents (A–D)
from perforated patches on astrocytes. (A) Resting whole-cell currents in isotonic saline produced by the waveform shown in E (isotonic
saline is normal bath saline with 80 mM NaCl replaced by 160 mM mannitol). Current scale bar is shown (right). Swelling-activated cur-
rents were recorded after the cell had been exposed for 30 s to hypotonic saline (B, isotonic saline minus 140 mM mannitol). (C) Perfu-
sion of hypotonic saline with 5 ␮M GsMTx-4 produced an ⵑ75% reduction in the peak swelling-activated current at 30 s, after subtracting
resting current. Swelling currents partially recovered ⵑ4 min after washout of GsMTx-4 (D). Peak swelling-activated currents at 100 mV (F,
䊏) from two different cells (a and b) decreased over successive exposures to hypotonic solution. (F, 䉫) Peak currents measured during hy-
potonic exposures with GsMTx-4 present were reduced from the control. (G) I-V plot of the average swelling-activated peak currents from
six cells measured 30–40 s after hypotonic exposure. The data points represent difference currents calculated by subtracting the resting
current from hypotonic current. Control hypotonic current (䊏), hypotonic currents in the presence of GsMTx-4 (䉫), and hypotonic cur-
rents after ⵑ5 min of washout (䊉). The hypotonic current in the presence of GsMTx-4 is ⵑ38% lower than control swell currents at ⫹100
mV and ⵑ48% lower at ⫺100 mV.

594 Peptide Blocker of Mechanical Transduction
maining inward current is largely I
Cl,swell
. The toxin pro-
duced no further current reduction in the presence of
hypertonic saline, which inactivates the swelling-acti-
vated current (Fig. 11 C, 1.5T
C
⫺ 1.5T
Tx
). This result
again supports the specificity of GsMTx-4. When vol-
ume changes are examined (Fig. 11 D), GsMTx-4 pro-
duced a cell volume reduction that is ⵑ40% of that
produced by 1.5T hypotonic saline.
DISCUSSION
A New Tool
We have found the first specific reagent for mecha-
nosensitive ion channels (MSCs). It is surely only a pro-
totype, with many reagents to follow. While studies are
needed to establish the cross reactivity of GsMtx-4 for
different types of MSCs, as well as for cross reactivity
with different types of channels, GsMtx-4 is a unique
agent. It can be used to test the involvement of SACs in
physiological processes in situ or in vivo. Its ability to
suppress cardiac arrhythmias (Bode et al., 1999) sug-
gests a clinical significance. Its ability to suppress cell
swelling may also be clinically useful for dealing with
edema. GsMtx-4 may be derivatized to make affinity col-
umns for purifying the channel protein and for making
histochemical markers to localize the channels.
Figure 10. DIDS reduces swelling-activated currents in adult as-
trocytes. Whole-cell currents produced by a 600-ms ramp in volt-
age from ⫺120 to 80 mV at 30 s after exposure to hypotonic saline
are shown. Hypotonic and hypotonic ⫹ 50 ␮M DIDS difference
current are shown, produced by subtracting the whole-cell current
under isotonic conditions from the current observed during hypo-
tonic exposure. A large reduction in swelling-activated current is
observed at hyperpolarized potentials compared with the reduc-
tion observed at negative potentials. The reversal potential shift
caused by DIDS is approximately ⫺30 mV. The average reversal
potential shift from six cells exposed to DIDS during hypotonic
swelling was approximately ⫺33 ⫾ 3.5 mV.
Figure 11. Ionic currents (A–
C) and cell volumes (D) mea-
sured during perforated patch
voltage clamp (E
hold
⫽ ⫺80 mV)
of ventricular myocytes from rab-
bits with aortic regurgitation-
induced CHF. Myocytes were ex-
posed to 1.0T and 1.5T solution
in the absence [1.0T
C
(dashed
line), 1.5T
C
(solid line)] and
presence [1.0T
Tx
(dashed line),
1.5T
Tx
(solid line)] of 0.4 ␮M
GsMTx-4. (A) Osmotic shrinkage
in the control solution reduced
both inward and outward cur-
rents. (B) Toxin reduced the in-
ward currents in 1.0T, but the
currents in 1.5T were unaffected
(compare A and B). (C) Dif-
ference currents: 1.0T
C
–1.5T
C
(dash-dot line), shrinkage-sensi-
tive current due to inhibition of
cationic SACs and anionic
swelling currents; 1.0T
C
–1.0T
Tx
(dashed line), inwardly rectify-
ing toxin-sensitive current. The
toxin-sensitive current was simi-
lar to the Gd
3⫹
-sensitive currents
recorded in the same model (data not shown, Clemo and Baumgarten, unpublished observations). 1.5T
C
–1.5T
Tx
(solid line), toxin did not
affect membrane currents when SACs were inhibited by osmotic shrinkage. (D) Consistent with block of ion influx via cationic SACs, toxin
reduced cell volume by 7% in 1.0T solution. In contrast, cell volume was unaffected after osmotic shrinkage in 1.5T solution, conditions
under which SACs are closed. I-V curves elicited by ramp clamps (28 mV/s) and cell volume data are averages from five cells. Lower case
letters (a, b, c, and d) designate time of current measurements in I-V profiles.

595 Suchyna et al.
It is surprising that GsMtx-4 can act across different
tissues in similar concentrations. This implies a strong
homology between the cationic SACs of these tissues
and may help to define a family of channels. The teleo-
logical significance of why a rather mild tarantula
venom would have the ability to block SACs in a rabbit
heart or a rat brain is unclear. Perhaps insects, the spi-
der’s normal prey, have similar channels.
The mechanism of action of GsMtx-4 remains to be
determined. We do know that it can act on closed chan-
nels, but we don’t know if this occurs because the acti-
vation curve is shifted to higher tensions or because
GsMtx-4 blocks the permeation path. These studies are
now in progress.
GsMTx-4 Structure
GsMTx-4 possesses an ICK consensus cysteine motif with
the basic structure defined by three cysteine pairs (C
1
–
C
4
, C
2
–C
5
, and C
3
–C
6
) that stabilize a core region com-
posed of a triple-stranded antiparallel ␤ sheet (for re-
view, see Norton and Pallaghy, 1998). Enzymatic digest
fragments are currently being analyzed to confirm this
structure. Examples of toxin families active against ion
channels that possess ICK motifs are: the ␮-agatoxins
and ␦-atracotoxins, which block voltage-activated Na
⫹
channels; the ␻-agatoxins, ␻-conotoxins, and ␻-atrac-
otoxins, which block voltage-gated Ca
2⫹
channels; and
hanatoxin, ␬-conotoxin, and TXP-5, which block volt-
age-gated K
⫹
channels. GsMTx-4 shows the most se-
quence similarity to K
⫹
and Ca
2⫹
channel blocking tox-
ins from tarantula venoms, the highest being TXP-5
from the Brachypelma smithii tarantula, where the similar-
ity is 54%. However, sequence homology between ICK
containing peptide toxins is a poor indicator of func-
tional similarities. For example, the N-type Ca
2⫹
channel
blocker ␻-conotoxin MVIIA has ⬎80% sequence simi-
larity with P/Q-type Ca
2⫹
channel blocker ␻-conotoxin
MVIIC, while sharing only 45% sequence similarity with
the N-type Ca
2⫹
channel blocker ␻-conotoxin GVIA.
We have recently produced a recombinant GsMTx-4
peptide in bacteria that in initial experiments blocks
SACs in outside-out patches from astrocytes. This re-
moves the possibility of a copurified contaminant along
with GsMtx-4 from raw venom.
GsMTx-4 Binding Affinity in Astrocyte and Cardiac
Myocyte Assays
The equilibrium constant for toxin binding was calcu-
lated to be ⵑ600 nM. While many peptide toxins are
highly specific for their receptor, having affinities (IC
50
or K
d
) in the 0.1–100 nM range, GsMTx-4 binds 5–50⫻
more tightly than any other antagonist tested to date
on any stretch channel. It appears to be specific for cat-
ion SACs since it did not effect voltage-sensitive cur-
rents in astrocytes or I
Cl,swell
in myocytes. The complete
block of I
Cir,swell
at 0.4 ␮M indicates that toxin affinity
for its binding site in cardiac myocytes may be even
stronger than in astrocytes. Preliminary results from
atrial-induced fibrilation experiments with Langen-
dorff-perfused rabbit hearts suggest that GsMTx-4
doesn’t block normal electrical activity of the heart
(eliminating many possible sites of cross reactivity) and
the toxin may have a higher affinity for these cells in
situ (Bode et al., 1999). In these studies, 0.17 ␮M
GsMtx-4 was capable of inhibiting the atrial fibrillation
associated with dilatation.
Changes in SAC Properties between Cell-attached and
Outside-Out Patches
The properties of SACs in activated adult astrocytes, in-
cluding ion selectivity, conductance, inward rectifica-
tion, and adaptation, are similar to cation-selective
SACs observed in other systems (Yang and Sachs, 1993).
Outside-out patches were used in only one previous
study to determine the effects of Gd
3⫹
(Yang and Sachs,
1989) on SACs in Xenopus oocytes. However, in that re-
port, the effects of outside-out formation on channel
properties was not rigorously assessed because it was dif-
ficult to maintain activity in the patch. Patches from
most cell types rapidly lose SAC activity with excision.
In the adult astrocytes, SAC adaptation is lost during
outside-out patch formation, or if ⬎10 mmHg suction
is used for cell-attached seal formation. At voltages
where adaptation should rapidly reduce channel P
o
(⫺50 mV), in outside-out patches we observed a de-
layed activation instead (Figs. 1 B, ⫹100 mV average
current, and 5 A, average current). This selective loss of
adaptation is similar to the two stages of decoupling de-
scribed by Hamill and McBride (1992) in Xenopus oo-
cytes. The mechanisms for these changes in gating as a
function of patch history remain to be determined.
However, the intrinsic permeation properties of the
channels, such as channel conductance, rectification,
and ion selectivity, seem less likely to be affected by
cytoskeletal attachments and appear less sensitive to
patch history, as shown in Fig. 1, C and D. Even while
more pressure/suction is required over time to activate
the channel in either configuration (decoupling of the
tonic gating element), channel conductance and rec-
tification remain unchanged. Furthermore, although
ion selectivity was not rigorously compared between the
two patch configurations, channel conductance was 46
pS with 130 mM CsCl substituted for KCl in the pipette,
demonstrating the channel is nonselective for cations
in the cell-attached mode. Thus, outside-out patches
are an adequate representation of the activity in cell-
attached patches and a much more flexible prepara-
tion for screening.

596 Peptide Blocker of Mechanical Transduction
SAC Activity during Astrocyte Cell Swelling
The sensory processes for RVD have not been deter-
mined, but these experiments strongly suggest that cat-
ionic SACs play a role. While dilution of internal K
⫹
and
an increase Na
⫹
flux could contribute to membrane de-
polarization, this is an ineffective stimulus under voltage
clamp and it has been demonstrated in multiple studies
on different cell types that an increase in anionic cur-
rent is the major contributor to membrane depolariza-
tion during hypotonic swelling (Pasantes-Morales et al.,
1994; Bakhramov et al., 1995). The trigger for the vol-
ume-activated anion current is still unclear. In our adult
astrocytes, anion current accounts for nearly the entire
reversal potential shift that occurs after hypotonic expo-
sure. GsMTx-4 produces a 40–50% reduction in swell-
ing-activated anion current, suggesting that cationic
SACs contribute to activation of the current. However,
in contrast to DIDS blockade of the anionic current, the
reduction in swelling-activated current produced by
GsMTx-4 showed no change in reversal potential. This
could be due to a reduction of SAC cation current or an
overall decrease in all swelling-activated currents. Dose
response studies with GsMtx-4 are in order.
Although anion current dominates the whole-cell
conductance during RVD, swelling-activated K
⫹
cur-
rents that generally develop more slowly are rate limit-
ing for Cl
⫺
efflux. Increasing the cation flux with gram-
icidin can circumvent the rate-limiting effect of the
slowly increasing K
⫹
current (Pasantes-Morales et al.,
1994). Activation of cation-selective SACs, like gramici-
din, would increase the flux of cations (increasing Cl
⫺
efflux), and thus the rate of RVD.
RVD in astrocytes has been reported to be a Ca
2⫹
-
dependent process (O’Connor and Kimelberg, 1993;
Bender et al., 1994). Thus, GsMTx-4 may block Ca
2⫹
in-
flux through SACs’ reducing Ca
2⫹
-sensitive swelling-acti-
vated currents. Contributions to Ca
2⫹
influx through
voltage-gated Ca
2⫹
channels were reduced by clamping
the cell at ⫺50 mV before recording the whole-cell cur-
rents. A number of swelling-activated anion currents
have been identified that are induced by secondary mes-
senger systems such as Ca
2⫹
, calmodulin, and various ki-
nases. Although it has been shown that the DIDS-sensi-
tive Cl
⫺
current is not Ca
2⫹
dependent in neonatal astro-
cytes (Pasantes-Morales et al., 1994; Bakhramov et al.,
1995; Crepel et al., 1998), Ca
2⫹
-activated Cl
⫺
currents
are partially responsible for the anion efflux during RVD
in Ehrlich ascite cells (Hoffmann et al., 1986; Lambert et
al., 1989). It is possible that anion currents between neo-
natal and adult rat astrocytes may differ, especially since
different SACs are observed between separate neonatal
preparations (Bowman et al., 1992; Islas et al., 1993) and
adult astrocytes. Further characterization of GsMTx-4’s
effects on membrane currents and Ca
2⫹
influx is re-
quired to more clearly define the role of SACs in RVD.
GsMTx-4 Effect on Rabbit CHF-model Cardiac Myocytes
In cardiac myocytes, stretch/swell-induced currents
may play a critical role in the development of dysrhyth-
mias and hypertrophy, and may alter contractile func-
tion. Cationic (I
Cir,swell
) and anionic (I
Cl,swell
) swelling-
activated currents have been identified in hypotonically
swollen rabbit cardiac myocytes (Clemo and Baumgar-
ten, 1997), and these may also be activated by the cell
swelling that occurs during ischemia. Cardiac myocytes
isolated from dog hearts with Tachychardia-induced
CHF (Clemo et al., 1998) maintain a cell volume 1.24⫻
greater than normal cells. Both I
Cir,swell
and I
Cl,swell
are
constitutively active in these CHF-model cardiomyo-
cytes under isotonic conditions. Moreover, swelling-acti-
vated currents are persistently active in rabbit myocytes
from an aortic regurgitation model. GsMTx-4 at 0.4 ␮M
specifically blocked I
Cir,swell
in CHF cardiomyocytes. The
cation current I
Cir,swell
might represent activity of cat-
ionic SACs. They are both inwardly rectifying, blocked
by Gd
3⫹
, and poorly selective for cations (P
K
/P
Na
⫽ 6).
The fact that GsMTx-4 blocks SACs in rat astrocytes
and I
Cir,swell
(properties similar to cation-selective SACs)
in rabbit cardiac myocytes suggests that many cell types
incorporate SACs as part of the volume-regulatory pro-
cess. Furthermore, the common toxin sensitivity sug-
gests that at least some cation channels opened by di-
rect mechanical stimulation are also opened by cell
swelling. GsMTx-4 will be useful in elucidating the
function of SACs in a variety of systems under physio-
logically normal and stressed conditions.
This work was funded by grants to Dr. Frederick Sachs from the
National Institutes of Health, the United States Army Research
Office, and NPS Pharmaceuticals, Inc. Studies on myocytes were
supported by HL46764 to Dr. Baumgarten.
Submitted: 9 July 1999
Revised: 2 March 2000
Accepted: 6 March 2000
REFERENCES
Bakhramov, A., C. Fenech, and T.B. Bolton. 1995. Chloride current
activated by hypotonicity in cultured human astrocytoma cells.
Exp. Physiol. 80:373–389.
Bascur, L., I. Yevenes, and P. Barja. 1982. Effects of Loxosceles laeta
spider venom on blood coagulation. Toxicon. 20:795–796.
Ben-Tabou, S., E. Keller, and I. Nussinovitch. 1994. Mechanosensi-
tivity of voltage-gated calcium currents in rat anterior pituitary
cells. J. Physiol. 476:29–39.
Bender, A.S., L.L. Mantelle, and M.D. Norenberg. 1994. Stimula-
tion of calcium uptake in cultured astrocytes by hypoosmotic
stress: effect of cyclic AMP. Brain Res. 645:27–35.
Bender, A.S., A. Schousboe, W. Reichelt, and M.D. Norenberg.
1998. Ionic mechanisms in glutamate-induced astrocyte swelling:
role of K
⫹
influx. J. Neurosci. Res. 52:307–321.
Biagi, B.A., and J.J. Enyeart. 1990. Gadolinium blocks low- and
high-threshold calcium currents in pituitary cells. Am. J. Physiol.
259:C515–C520.
Bode, F., F. Sachs, and M.R. Franz. 1999. Atrial fibrillation is inhib-

597 Suchyna et al.
ited by a peptide from spider venom that blocks stretch activated
ion channels. Circulation. 100:1784. (Abstr.)
Bowman, C.L., J.P. Ding, F. Sachs, and M. Sokabe. 1992. Mechan-
otransducing ion channels in astrocytes. Brain Res. 584:272–286.
Bowman, C.L., and J.W. Lohr. 1996. Mechanotransducing ion
channels in C6 glioma cells. Glia. 18:161–176.
Chamberlin, M.E., and K. Strange. 1989. Anisosmotic cell volume
regulation: a comparative view. Am. J. Physiol. 257:C159–C173.
Chen, Y., S.M. Simasko, J. Niggel, W.J. Sigurdson, and F. Sachs.
1996. Ca
2⫹
uptake in GH3 cells during hypotonic swelling: the
sensory role of stretch-activated ion channels. Am. J. Physiol. 270:
C1790–C1798.
Christensen, O. 1987. Mediation of cell volume regulation by Ca
2⫹
influx through stretch-activated channels. Nature. 330:66–68.
Clemo, H.F., and C.M. Baumgarten. 1997. Swelling-activated Gd
3⫹
-
sensitive cation current and cell volume regulation in rabbit ven-
tricular myocytes. J. Gen. Physiol. 110:297–312.
Clemo, H.F., B.S. Stambler, and C.M. Baumgarten. 1998. Persistent
activation of a swelling-activated cation current in ventricular my-
ocytes from dogs with tachycardia-induced congestive heart fail-
ure. Circ. Res. 83:147–157.
Crepel, V., W. Panenka, M.E. Kelly, and B.A. MacVicar. 1998. Mito-
gen-activated protein and tyrosine kinases in the activation of as-
trocyte volume-activated chloride current. J. Neurosci. 18:1196–
1206.
Elinder, F., P. Arhem. 1994. The modulatory site for the action of
gadolinium on surface charges and channel gating. Biophys. J. 67:
71–83.
Fontana, A. 1972. Modification of tryptophan with BNPS-Skatole
(2-2-nitrophenylsulfenyl)-3-methyl-3-bromoindolenine). In Meth-
ods in Enzymology. Vol. 25, part B. S.P. Colowick, and N.O. Ka-
plan, editors. Academic Press, Inc., New York, NY. 419–423.
Hamill, O.P., and D.W. McBride. 1996. The pharmacology of mech-
anogated membrane ion channels. Pharmacol. Rev. 48:231–252.
Hamill, O.P. and D.W. McBride, Jr. 1992. Rapid adaptation of single
mechanosensitive channels in Xenopus oocytes. Proc. Natl. Acad.
Sci. USA. 89:7462–7466.
Hoffmann, E.K., I.H. Lambert, and L.O. Simonsen. 1986. Separate,
Ca
2⫹
-activated K
⫹
and Cl
⫺
transport pathways in Ehrlich ascites
tumor cells. J. Membr. Biol. 91:227–244.
Hoffmann, E.K., and L.O. Simonsen. 1989. Membrane mecha-
nisms in volume and pH regulation in vertebrate cells. Physiol.
Rev. 69:315–381.
Horn, R., and A. Marty. 1988. Muscarinic activation of ionic cur-
rents measured by a new whole-cell recording method. J. Gen.
Physiol. 92:145–159.
Hu, H., and F. Sachs. 1997. Stretch-activated ion channels in the
heart. J. Mol. Cell. Cardiol. 29:1511–1523.
Islas, L., H. Pasantes-Morales, and J.A. Sanchez. 1993. Characteriza-
tion of stretch-activated ion channels in cultured astrocytes. Glia.
8:87–96.
Jorgensen, F., and H. Ohmori. 1988. Amiloride blocks the mech-
ano-electrical transduction channel of hair cells of the chick. J.
Physiol. 403:577–588.
Kaiser, I.I., P.R. Griffin, S.D. Aird, S. Hudiburg, J. Shabanowitz, B.
Francis, T.R. John, D.F. Hunt, and G.V. Odell. 1994. Primary
structures of two proteins from the venom of the Mexican red
knee tarantula (Brachypelma smithii). Toxicon. 32:1083–1093.
Kimelberg, H.K. 1995. Current concepts of brain edema. Review of
laboratory investigations. J. Neurosurg. 83:1051–1059.
Kimelberg, H.K., E. Anderson, and H. Kettenmann. 1990. Swelling-
induced changes in electrophysiological properties of cultured
astrocytes and oligodendrocytes. II. Whole-cell currents. Brain
Res. 529:262–268.
Kimelberg, H.K., and H. Kettenmann. 1990. Swelling-induced
changes in electrophysiological properties of cultured astrocytes
and oligodendrocytes. I. Effects on membrane potentials, input
impedance and cell–cell coupling. Brain Res. 529:255–261.
Kluesener, B.G. Boheim, H. Liss, J. Engelberth, and E.W. Weiler.
1995. Gadolinium-sensitive, voltage-dependent calcium release
channels in the endoplasmic reticulum of a higher plant mecha-
noreceptor organ. EMBO (Eur. Mol. Biol. Organ.) J. 14:2708–2714.
Lambert, I.H., E.K. Hoffmann, and F. Jorgensen. 1989. Membrane
potential, anion and cation conductances in Ehrlich ascites tu-
mor cells. J. Membr. Biol. 111:113–131.
Lampe, R.A., P.A. Defeo, M.D. Davison, J. Young, J.L. Herman, R.C.
Spreen, M.B. Horn, T.J. Mangano, and R.A. Keith. 1993. Isola-
tion and pharmacological characterization of omega-grammo-
toxin SIA, a novel peptide inhibitor of neuronal voltage-sensitive
calcium channel responses. Mol. Pharmacol. 44:451–460.
Lane, J.W., D. McBride, and O.P. Hamill. 1991. Amiloride block of
the mechanosensitive cation channel in Xenopus oocytes. J. Phys-
iol. 441:347–366.
Langan, T.J., R.J. Plunkett, H. Asada, K. Kelly, and P. Kaseloo. 1995.
Long-term production of neurotrophic factors by astrocyte cul-
tures from hemiparkinsonian rat brain. Glia. 14:174–184.
Langton, P.D. 1993. Calcium channel currents recorded from iso-
lated myocytes of rat basilar artery are stretch sensitive. J. Physiol.
471:1–11.
Magid, N.M., M.S. Young, D.C. Wallerson, R.S. Goldweit, J.N.
Carter, R.B. Devereux, and J.S. Borer. 1988. Hypertrophic and
functional response to experimental chronic aortic regurgita-
tion. J. Mol. Cell Cardiol. 20:239–246.
Narasimhan, L., J. Singh, C. Humblet, K. Guruprasad, and T. Blun-
dell. 1994. Snail and spider toxins share a similar tertiary struc-
ture and “cystine motif”. Nat. Struct. Biol. 1:850–852.
Newcomb, R., B. Szoke, A. Palma, G. Wang, X. Chen, W. Hopkins,
R. Cong, J. Miller, L. Urge, K. Tarczy-Hornoch, et al. 1998. Selec-
tive peptide antagonist of the class E calcium channel from the
venom of the tarantula Hysterocrates gigas. Biochemistry. 37:15353–
15362.
Norton, R.S., and P.K. Pallaghy. 1998. The cystine knot structure of
ion channel toxins and related polypeptides. Toxicon. 36:1573–
1583.
O’Connor, E.R., and H.K. Kimelberg. 1993. Role of calcium in as-
trocyte volume regulation and in the release of ions and amino
acids. J. Neurosci. 13:2638–2650.
Paoletti, P., and P. Ascher. 1994. Mechanosensitivity of NMDA re-
ceptors in cultured mouse central neurons. Neuron. 13:645–655.
Pasantes-Morales, H. 1996. Volume regulation in brain cells: cellu-
lar and molecular mechanisms. Metab. Brain Dis. 11:187–204.
Pasantes-Morales, H., R.A. Murray, L. Lilja, and J. Moran. 1994.
Regulatory volume decrease in cultured astrocytes. I. Potassium-
and chloride-activated permeability. Am. J. Physiol. Cell Physiol.
266:C165–C171.
Perez, F., S. Luca, P. DaSilva, and R. Bertani. 1996. Systematic revi-
sion and cladistic analysis of Theraphosinae. Mygalomorph. 1:33–
68.
Ruknudin, A., F. Sachs, and J.O. Bustamante. 1993. Stretch-acti-
vated ion channels in tissue-cultured chick heart. Am. J. Physiol.
Circ. Physiol. 264:H960–H972
Sachs, F. 1988. Mechanical transduction in biological systems. Crit.
Rev. Biomed. Eng. 16:141–169.
Sachs, F. 1992. Stretch sensitive ion channels: an update. In Sensory
Transduction. D.P. Corey and S.D. Roper, editors. Rockefeller
University Press. New York, NY. 241–260.
Sachs, F., and C. Morris. 1998. Mechanosensitive ion channels in
non specialized cells. In Reviews of Physiology and Biochemistry
and Pharmacology. M.P. Blaustein, R. Greger, H. Grunicke, R.
Jahn, L.M. Mendell, A. Miyajima, D. Pette, G. Schultz, and M.

598 Peptide Blocker of Mechanical Transduction
Schweiger, editors. Springer Verlag, Berlin, Germany. 1–78.
Sarkadi, B., and J.C. Parker. 1991. Activation of ion transport path-
ways by changes in cell volume. Biochim. Biophys. Acta. 1071:407–
427.
Silberberg, S.D., and K.L. Magleby. 1997. Voltage-induced slow acti-
vation and deactivation of mechanosensitive channels in Xenopus
oocytes. J. Physiol. 505:551–569.
Small, D.L., and C.E. Morris. 1995. Pharmacology of stretch-acti-
vated K
⫹
channels in Lymnaea neurones. Br. J. Pharmacol. 114:
180–186.
Sontheimer, H. 1992. Astrocytes, as well as neurons, express a diver-
sity of ion channels. Can. J. Physiol. Pharmacol. 70:S223–S238
Swartz, K.J., and R. MacKinnon. 1995. An inhibitor of the Kv2.1 po-
tassium channel isolated from the venom of a Chilean tarantula.
Neuron. 15:941–949.
Tazaki, M., and T. Suzuki. 1998. Calcium inflow of hamster Merkel
cells in response to hyposmotic stimulation indicate a stretch-
activated ion channel. Neurosci. Lett. 243:69–72.
Vandenberg, J.I., S.A. Rees, A.R. Wright, and T. Powell. 1996. Cell
swelling and ion transport pathways in cardiac myocytes. Cardio-
vasc. Res. 32:85–97.
Wilson, K.J. 1989. Methods in protein sequence analysis: proceed-
ings of the 7
th
international conference, Berlin, July 3–8, 1988. B.
Wittmann-Leibold, editor. Springer-Verlag, New York, NY. 310–
330.
Yang, X.C., and F. Sachs. 1989. Block of stretch-activated ion chan-
nels in Xenopus oocytes by gadolinium and calcium ions. Science.
243:1068–1071.
Yang, X.C., and F. Sachs. 1993. Mechanically sensitive, non-selec-
tive, cation channels. In Non-Selective Ion Channels. D. Siemen
and J. Hescheler, editors. Springer-Verlag, Heidelberg, Germany.
79–92.