Pfliigers Arch (1992) 4 2 1 : 6 0 6 - 6 1 2

Joumal of Physiology 9 Springer-Verlag 1992

Pressure-clamp: a method for rapid step perturbation of mechanosensitive channels Don W. McBride Jr. and Owen P. Hamill Section of Neurobiology and Behavior, Cornell University, Ithaca, New York 14853, USA Received March 4, 1992/Received after revision May 19, 1992/Accepted May 25, 1992

Abstract. Here we describe a pressure-clamp method for applying suction or pressure steps to membrane patches in order to study the activation, adaptation and relaxation characteristics of mechanosensitive (MS) channels. A description is given of the mechanical arrangement of the pressure clamp which involves a balance between negative (suction) and positive pressures. The electronic circuitry of the feedback control is described. We also describe the optimal time response (= 10ms) of the pressure-clamp, the amplitude of pressure resolution (0.2-0.5 mmHg; 2 7 - 6 7 Pa) and the factors influencing these parameters. We illustrate the applications of the clamp on the Xenopus oocyte and cultured skeletal myotubes from dystrophic mouse (mdx) muscle, both of which express MS channels. Studies with pressure/suction pulses indicate that in both muscle and oocytes MS channel activity displays adaptation. The ability to study current relaxations following step changes in pressure/suction using the pressure-clamp in combination with patch-clamp techniques provides the opportunity for analysis of the time, voltage and pressure dependence of the opening and closing of MS channels. Key words: Pressure-clamp - Mechanosensitive channels - Rapid mechanical perturbation - Xenopus oocytes Dystrophic muscle

Introduction

Mechanosensitive (MS) channels represent the most recently discovered class of membrane ion channels [6, 13, 16, 20, 23]. To date most studies on single MS channels have characterized their response to steady-state mechanical stimulation. A variety of different methods has been used to activate single MS channels including osmotic

Correspondence to: O.P. Hamill

swelling [1, 10, 24, 27] or suction applied to the patch pipette by, for example, a syringe [4, 8], mouth [17], a water aspirator [19], a thumb-wheel-controlled piston [21] or a motor-driven piston [14]. However, these techniques are not amenable to rapid and precise step changes in pressure. In the study of mechanosensory cells such as Pacinian corpuscles and hair cells a probe attached to a piezoelectric element is routinely used to illicit MS responses [2, 3, 7, 15, 22]. However, this technique is not amenable to cell-attached patch recording of single MS channel currents. Here we describe a pressure-clamp that allows steps of either negative (suction) or positive pressure to be applied to membrane patches in a rapid and precise manner. We describe the mechanical arrangement, the feedback circuitry and the kinetics of this pressure-clamp. To illustrate the use of this clamp we describe some dynamic properties of MS channel currents in two different preparations: the Xenopus oocyte and myotubes from the dystrophic (mdx) mouse.

The method In analogy with typical voltage-clamp experiments, the ideal pressureclamp would provide steps of pressure with a rapid onset, a maintained pressure and a rapid turn-off. The relaxation of the MS current following a rapid turn-on and rapid turn-off of stimulation can provide kinetic information regarding transitions between various states of the channel.

The basic principle of the pressure-clamp At least three different strategies could be used to provide step changes in pressure. One strategy is to attach the pipette to a vacuum source (e.g., vacuum pump or a water aspirator). In this case the amount of suction applied to the pipette may be controlled by a needle valve bled to the atmosphere and the application of suction gated by solenoid valves. This technique, although effective, suffers the disadvantages that it is difficult to control the amount of suction and apply complex pressure waveforms, it cannot be used to apply positive pressure and finally, because the turn-off is passive, it cannot be as fast as an active process (see below). A second strategy involves a piston or syringe-type mechanism. This technique is convenient and allows for application of positive and nega-

607

Mechanical arrangement of the pressure-clamp

tive pressures. However, because of the compressibility of the air between the piston and the pipette, it has inherent limitations with regard to the magnitude and velocity of piston displacement. Although these limitations can be removed by back-filling the pipette, pipette holder and connecting tubing with an incompressible fluid such as mineral oil [25], the remedy has practical drawbacks in terms of carrying out routine experiments. The third strategy, and the one we present here, is based on a balancing of negative (i.e., suction) and positive pressures to achieve the desired pressure. Because this technique allows rapid application of both positive and negative pressures, fast step changes in pressure can be achieved as well as complex pressure waveforms.

A standard patch-pipette holder with a suction port [11] was used and connected as shown in Fig. 1 A. Critical to the pressure-clamp is the piezoelectric valve (MV-112, Maxtek, Torrence, Calif.), which is a needle valve with an opening that is proportional to the applied voltage (0-100 V) and has a time response less than 2 ms. This valve controls the flow rate of pressurized N 2 into a chamber (i.e., "mixing area"), which is also connected to a suction source provided by a vacuum pump (541/min and 0.1 Torr) (Speedivac 2, Edwards High Vacuum, Crawley, England). In the mixing area a steady-state pressure is reached. This steady-state pressure results from a balance of positive pressure (influx) and negative pressure (efflux). By controlling the amount of inflow into a constant outflow either a negative or positive pressure can be applied to the back of the pipette. A pressure transducer (33 N-005 D, I C Sensors, Milpitas, Calif.) is placed as close to the pipette holder as is practical and monitors the pressure/suction while providing feedback to the piezoelectric valve via the pressure-clamp controller. Although we do not monitor the pressure immediately above the patch, we have attempted to compensate for geometric-access considerations by introducing a branch point with equivalent arms (i.e., with respect to volume and constrictions) to the transducer and the pipette holder. This branch point is not explicitly represented in Fig. 1A but is positioned after the mixing chamber.

A

+'

Electronic feedback circuitry Negative-feedback control of the piezoelectric valve is an essential element in controlling the steady-state pressure/suction and in producing defined steps of pressure/suction (see below). This feedback also ensures that the pressure/suction applied is not distorted by leaks, fluctuations in the vacuum source or the transient overshoot/undershoot in response to a step change in pressure/suction. The circuit as shown in Fig. 1 B can be divided into four parts. Part I is the command voltage and contains an input for an external command from the computer, which is superimposed on an internal manual control. Part II is the bridge amplifier for the pressure transducer. The instrumentation amplifier AD 624 (Analog Devices, Norwood, Mass.) is used to amplify the signal from the pressure transducer. The 100-kf~ trimpot on amplifier A4 is used to set the gain of the pressure monitor system. Part III is the driver for the piezoelectric valve, which is con-

Commamd

B 5,00V

I ! OK ,I H ~ P r e s zero ~

|tlge

adjust

//I

V I

II. Pressure Measurement

I v'

I I

p,..o.,,o,r,0 I v+,.

~, ,OOK

~ ~ gain r ~just

I

Commandfrom ComPuter

9

I Command ' Voltage

;~K

r....

_ , s v + ~ - - ~ +. A ~

I

100V

10K~ ! OWI

I ' I

0.2BF 20K

~"-I -J---I- ~ . ] I A6~ '

~

flY.ErrorSignal

+"

I

and Main Loop

i

i

I I I . P i e z 0 Driver

Fig. 1A, B. Mechanical arrangement and electronic control of the pressure-clamp. A Simplified diagram of the mechanical arrangement of the pressure clamp including the patch pipette and holder. The essential elements of the clamp include (a) a pressure source (N2), (b) a suction source provide by a vacuum pump, (c) a piezoelectric valve that regulates the pressure applied, (d) a pressure transducer that monitors the final pressure and (e) feedback via the pressure-clamp controller that controls the piezoelectric valve. B Circuit schematic of the feedback control circuit of the pressure clamp (see text for details). Except where noted, all operational amplifiers were ADOP07 from Analog Devices. The 10kf~ trimpot across the pressure transducer was used to set the zero point of the transducer (see application notes for the AD 624 in: Linear Products Databook, Analog Devices, Norwood, Mass.)

608 nected (indicated by the bold arrow) to the pressure transducer by the arrangement shown in Fig. 1A. This valve requires a voltage input of 0 - 1 0 0 V. A power MOSFET (IRF330, Motorola, Phoenix, Ariz.) is used to control the voltage to the valve. The amplifier A7 drives the gate of the MOSFET and is configured with a gain of 10. A 10-kf~, 10-W resistor is used to limit the current to the piezoelectric valve. Part IV contains amplifier A6, which monitors the error between the measured and desired signal and provides overall negative feedback by feeding the inverted error into the valve driver section. The series combination of the 0.2-1xF and the 20-kf~ variable resistor immediately around A6 forms the compensation feedback, which controls the frequency response of the clamp. The value of the compensation capacitor was adjusted to give the best response (see Fig. 2), The optimal value of the capacitor is dependent on the physical dimensions of the mixing area and connecting tubing as well as the vacuum flow rate. The DC feedback loop contains the driver section and the pressure transducer section. Speed limitations of the clamp There are some basic physical limitations to the speed with which the pressure/suction can be changed. As a first approximation in determining the factors that would affect the speed, we use the ideal gas law (Eq. 1) P V = nRT

(1)

where P is the pressure, Vis the volume of the mixing area and connecting tubing, n is the number of moles of gas (i.e., predominantly N 2) in the volume V, R is the gas constant and T is temperature. Rearranging Eq. (1) we get P = a Xn

(2)

where a = RT/If. To determine what factors influence the time response we differentiate Eq. (2) with respect to time and, if we assume a is constant with respect to time, we get dP/dt = a x d n / d t

(3)

where dn ~dr = I - E . dn/dt represents the difference between the flux into volume V (influx = I, i.e., N2) and the flux out of volume V (efflux = E, i.e. vacuum). According to Eq. (3), in order to change the pressure it is necessary that d n / d t is made non-zero momentarily and returned to zero as the desired pressure is attained. The rate at which the pressure changes for a given change in flux (dn/dt) is determined by a. Since a is inversely

A --I

B l

;v I

I

I P1

P1

/10 mmHg~ 10 ms

200 ms

Fig. 2. Illustration of the speed of the pressure-damp for two different pulse durations. In each panel the top trace is the voltage eommand and the three lower are the pressures ( P I - P 3 ) with different amounts of feedback compensation. For both traces the pressure response was filtered at 400 Hz with an eight-pole Bessel filter and digitized at 2 kHz (A) and 1 kHz (B)

proportional to the volume, for a given change in the flux, smaller volumes increase dP/dt while larger volumes decrease dP/dt. Another factor that may limit the time response of the clamp is the speed of sound in air, which is 330 m/s. For example, when the piezoelectric valve opens, a pressure wave is delivered into the system. This wave can propagate no faster than the speed of sound. In the case of our system the distance from the valve to the pipette holder is 10-15 cm (i.e., this includes the mixing area and connecting tubing). Neglecting reflections, this puts an absolute limit of 0.3-0.5 ms for a pressure step to occur, which is more than an order of magnitude faster than our observed response time of 10 ms. To increase the speed of the clamp the compensation feedback around amplifier A6 was added. The variable resistor can be adjusted to optimize the time response as shown in Fig. 2 for a 50-ms (Fig. 2A) and a l-s (Fig~ 2B) pulse. The top trace in each panel is the voltage command. Undercompensation results in overshoots at the transition of the step change as shown in Pi, while overcompensation results in oscillation as shown in P3. Optimal adjustment is shown in P2. We have found that this optimization varies with the setting of the bleed screw used to adjust the vacuum efflux (see Fig. 1A). It may be that the performance of the clamp can be improved by adding a second piezoelectric valve between the vacuum source and the mixing area to control the efflux. With two valves the influx and efflux could be controlled reciprocally (i.e., to increase pressure the influx would be momentarily increased while the efflux would be momentarily decreased). However, we have found a single valve on the input side to be sufficient. P r e s s u r e r e s o l u t i o n o f t h e clamp." n o i s e s o u r c e s Another important parameter of the pressure-clamp is its pressure resolution, which determines the amplitude of the smallest step that can accurately be applied. The transducer itself is not limiting and, depending on its gain, can be sensitive to pressure changes on the order of 0.05 mmHg (6.7 Pa). The main factor determining the resolution of the clamp is the intrinsic fluctuations (or noise) of the pressure associated with balancing the fluxes. One source of extraneous noise was the cycleto-cycle fluctuations of the vacuum pump, which were easily detected. These fluctuations could be reduced to an acceptable low level ( < 0.05 mmHg) by filtering (or damping) the vacuum line with a large volume (~21) interposed between the pump and the pressure-clamp (see Fig. 1A). Another source of noise is related to the sensitivity of the pressure to a change in flux. This can be illustrated by differentiating Eq. (2) with respect to n to get dP /dn = a.

(4)

Since a is inversely proportional to V,, a given change in the flux will produce a small change in P for a large V and vice versa for a small V. Eqs. (3) and (4) indicate there is a conflict in attempting to optimize both the time response and pressure resolution o f the system. For example, attempts to decrease the noise fluctuations by increasing the volume will slow down the time response and vice versa. Our preference has been to optimize the time response since oocyte and muscle MS channels are activated by pressures/suctions of 10 mmHg (1.3 kPa) or more when using standard-size pipettes. However, when studying channels that are more sensitive to pressure, one could improve pressure resolution by increasing the system volume thereby reducing the pressure fluctuations.

Mechanosensitive channel responses in different cell types The pressure-clamp described here has been applied to t w o d i f f e r e n t cell t y p e s t h a t h a v e b e e n p r e v i o u s l y d e m o n strated to express mechanosensitive (MS) channels, n a m e l y t h e X e n o p u s o o c y t e [12, 14, 18, 26] a n d c u l t u r e d m y o t u b e s [5, 8]. T h e o o c y t e h a s a n a d v a n t a g e t h a t it expresses a h i g h d e n s i t y o f M S channels. T h e exact role(s)

609

MS channels play in oocytes is unknown. However, they have been implicated in diverse functions including volume regulation, fertilization and embryogenesis [23, 26]. In dystrophic mouse myotubes both stretch-activated and stretch-inactivated channels have been reported [5] but the relationship between these apparently separate classes of MS channels has not been determined. The only requirement for application of the pressureclamp to a cell is the ability to make tight (i.e., giga) seals between the pipette glass and the membrane. Standard patch-clamp recording techniques were used [11]. The patch pipette tip diameter was varied between 1 ~m and 3 ~m to change the number of channels in the patch. In all figures, inward currents and suctions are represented as downward deflections.

A

B

C

Preparations Xenopus oocytes. The procedures to isolate Xenopus oocytes and remove their follicular and vitelline layers for patch-clamp recording have been described previously [18]. Frogs were anaesthetized by being placed for approximately 20 min in a beaker containing 300 mg ethyl 3-aminobenzoate methanesulfonic acid (Aldrich) in 200 ml distilled water. Sterile surgical procedures were used to remove oocytes. After isolation, type V and VI oocytes were sorted and stored overnight in Barth's medium [in mM: NaC1 88, KC1 1, MgSOa 0.82, Ca0NO3) 2 0.33, CaC1z 0.41, NaHCO 3 2.4, TRIS/HC1 5, pH 7.4, osmolarity 200mOsm containing 75 ~g/ml gentamicin sulfate] at 17~ In all oocyte recordings the bath solution contained (in mM) 115 NaC1, 2.5 KC1, 1.8 CaCI/, 10 HEPES (NaOH), pH 7.2 and the pipette solution contained (in mM) 100 KC1, 10EGTA (KOH), 10 HEPES (KOH) pH 7.2. Cultured myotubes from the dystrophic (mdx) mouse. Dystrophic (C 57 BL/10 ScSn-mdx/mdx) mice were obtained from Jackson Laboratories (Barr Harbor, Me.). The method for preparing mouse myotubes was essentially that described in [5]. The growth medium (Dulbecco's modified Eagle medium+10% fetal calf serum) was changed on day 3 and on day 6 after plating of the cells on 18-mm coverslips placed into culture dishes. After 3 days the myotubes fused and by days 4 - 5 were showing spontaneous contractions. In muscle recordings the bath solution contained (in mM) 145 NaC1, 2 CaCI2, 1.5 KC1, 1 MgC12, 10 HEPES (NaOH) pH 7.2. The pipette solution was the same as that used in oocyte recordings (see above).

Results

MS channel activation in oocytes Figure 3 illustrates the rapid activation by suction pulse (A) of MS channels on two different cell-attached patches (B, C) from Xenopus oocytes. Note that in this figure and in all subsequent figures the pressure was initially held at 0 mmHg (i.e., with respect to atmospheric pressure) and stepped to the indicated pressure/suction. The left-hand

1 s

20 ms

Fig. 3. Illustration of the activation of mechanosensitive channels by a suction pulse (A) on two cell-attached patches of different sizes (B, C) from frog oocytes. Note the different time scales in the left and right

panels

panel indicates there was a rapid activation of MS channels followed by closure of all channels, with only an occasional subsequent opening even i n the presence of maintained suction. This adaptation phenomenon has generally been overlooked in studies in which MS channels were activated by steady-state pressure application [20, 23]. However, a number of studies have reported single MS channel activity consistent with adaptive behavior (reviewed in [6]). Some of the characteristics of this adaptation behavior in Xenopus oocytes have been described elsewhere [12]. The right-hand panels of Fig. 3 show a time expansion of the left-hand panels at the onset of the suction step and show individual MS channel openings. The delay between the onset of the suction pulse and channel openings presumably reflects a combination of the time course of the development of membrane tension and the tension dependence of MS channel gating. The expanded time scale of MS channel currents in Fig. 3 C, indicates that once a threshold suction is reached many channels (~- 10) can be opened in less than 200 ~s (channel currents recorded with a cut-off of 10 kHz). Figure 4 illustrates the use of the pressure-clamp in determining the pressure/response relationship of MS channels. In analogy with voltage-clamp experiments, a family of increasing pressure (A) or suction (B) steps (step size = 2.5 mmHg; 0.33 kPa) of 2.5 s duration was applied to a single patch and the resulting channel activity recorded. Notable features of these responses are (a) a threshold pressure/suction for MS channel activation, (b) rapid adaptation of the MS channel current for both pressure and suction steps, (c) a progressive increase in the number of MS channels activated with increasing pressure/suction and (d) saturation in the peak current response. The independence o f the above features on the polarity of pressure stimulation reinforces the concept that MS channels are gated by membrane tension rather

610

A f

B

'

" ' t !,,,.' ' .

I

V''

--!l

"II

f

I

|

500 ms

A

__1 20 rnmHg t

B/

--I

1

25 pA

1 $

Fig. 5. The MS current response of a cell-attached patch of a frog oocyte to a single positive pressure pulse (10 mmHg; 1.3 kPa) (A) and a double step pressure pulse (10&20 mmHg; 1.3 &2.6 kPa) applied 30 s later (B). This experiment indicates that MS channels that have adapted can be reactivated by increasing the stimulation. In this patch it was subsequently determined that the peak MS current shown in the three responses was close to the maximum current that could be activated on this patch, thus indicating that the second response in B was not due to channels that had not been activated by the first step in B

Fig. 4. Activation of mechanosensitive (MS) channels in a cell-attached patch on a frog oocyte by families of incrementally increasing pressure (A) and suction 0l) pulses. These results indicate that MS channel adaptation can occur with both pressure and suction pulses. We have noticed that, with regard to the application of positive pressure, the patch is more vulnerable to rupture than at an equivalent negative pressure. Furthermore, with large positive pressures ( > 50 mmHg), patch breakdown is often preceded by large spike-type artefacts in the electrical recording (not shown here)

than pressure itself [8, 9]. For example, when we refer to a threshold pressure for MS channel activation, presumably what the channel experiences is a threshold membrane tension that is related to pipette tip size and pressure according to Laplace's law [8, 9]. With the pressure-clamp, complex pressure/suction waveform protocols (as defined by computer control) can be applied to the pipette, including multilevel pressure steps of defined amplitude and duration, rapid changes from suction to pressure and vice versa, and double pulse protocols for measuring recovery from adaptation. The clamp can also be driven with sine, triangular, saw-tooth or ramp waveforms up to ~-.20Hz. However, our focus has mainly been on rapid step perturbations in stimulation. Figure 5 illustrates a control step (A) and a double pressure step (B) applied to the same patch. This experiment indicates that the MS channel adaptation seen here is due to a relaxation in either the gating tension or the sensitivity to that tension rather than to complete inactivation of the MS channel. In voltage-clamp experiments, the analysis of voltagesensitive tail current relaxations following a voltage step has provided useful kinetic information. Similarly, the pressure clamp enables the rapid turn-off of a pulse and can be used to investigate MS tail current relaxations. Figure 6 illustrates tail currents that follow a brief (100 ms) suction pulse applied to a patch at potentials of + 150 mV and - 150 mV. In this case the application of a brief pulse is critical to minimize the effects of adaptation seen at negative potentials [12]. Aside from the single MS channel rectification [14, 26], the most striking feature of Fig. 6 is the strong voltage dependence of the tail current relaxation.

611 20 mmHg

lO pA

after the removal of pressure stimulation (i.e,, see traces 4 and 5). This type of current behavior was never seen in the oocyte at negative potentials. It remains to be determined whether such channel behavior is a unique feature of dystrophic muscle and whether it contributes to the pathogenesis of this disease. Conclusion

-150 mV

20 pA 500 ms

Fig. 6. MS channel tail currents following a 100-ms suction pulse measured at patch potentials + 150 mV and - 150 mV. Note the different current scales and the pronounced voltage dependence of the tall-current relaxation. The larger baseline current noise seen at + 150 mV reflects, aside from the difference in gain, a general instability of the membrane patch (i.e., more "leaky" and fragile) at strongly positive potentials

The pressure-clamp technique described here provides the opportunity to study kinetic phenomena of MS channels that cannot be studied by steady-state pressure/suction application. The application of this technique to both nonsensory and sensory cells that express MS channels should provide new insights into the kinetics of MS channels and mechanotransduction. Acknowledgements: We thank the Muscula-Dystrophy Association and the Cystic Fibrosis Foundation for financial support.

References

MS channel activation in mdx myotubes A recent report has implicated abnormal expression of MS channels in the pathogenesis of muscular dystrophy [5]. We were therefore interested in applying the pressure clamp to dystrophic muscle membrane. Our general finding is that stretch-activated MS channel activity in muscle displays qualitatively similar adaptive behavior to that seen in oocytes. This channel behavior may complicate the interpretation of a separate class of stretch-inactivated channels. However, as illustrated in Fig. 7, occasionally MS channel activity is maintained for several seconds

"•

t

20 mmHg I

1 7

IV

Ii

-

'

3~

5

6__ ls

Fig. 7. Activation of MS channels on a cell-attached patch measured at - 100 mV from an mdx mouse myotube. Note in trace 4 the prolonged opening of MS channels after the removal of stimulation. In general we have found that adaptive behavior is less complete in dystrophic muscle than in oocytes

1. Christensen O (1987) Mediation of cell volume regulation by Ca 2+ influx through stretch-activated channels. Nature 330:66-68 2. Corey DP, Hudspeth AJ (1980) Mechanical stimulation and micromanipulation with piezoelectric bimorph elements. J Neurosci Methods 3:183-202 3. Crawford AC, Evans MG, Fettiplace R (1991) The actions of calcium on the mechano-electrical transducer current of turtle hair cells. J Physiol (Lond) 434:369-398 4. Erxleben C (1989) Stretch-activated current through single ion channels in the abdominal stretch receptor organ of the crayfish. J Gen Physiol 94:1071 - 1083 5. Franco Jr A, Lansman JB (1990) Calcium entry through stretch inactivated ion channels in mdx mouse. Nature 344:670-673 6. French AS (1992) Mechanotransduction. Annu Rev Physiol 54:135-152 7. Gray JAB, Malcolm JL (1950) The initiation of nerve impulses by mesenteric Pacinian corpuscles. Proc R Soc Lond [B] 137:96-140 8. Guharay F, Sachs F (1984) Stretch-aeStivated single ion channel currents in tissue cultured chick skeletal muscle. J Physiol (Lond) 352:685-701 9. Gustin MC, Zhou X-L, Martinac B, Kung C (1988) A mechanosensitive ion channel in the yeast plasma membrane. Science 242:762- 765 10. Hamill OP (1983) Potassium and chloride channels in red blood cells. In: Sakmann B, Neher E (eds) Single channel recording. Plenum, New York, pp 451-471 11. Hamill OP, Marry A, Neher E, Sakmann B, Sigworth FJ (1981) Improved patch clamp techniques for recording from cells and cellfree membrane patches. Pfliigers Arch 391:85-100 12. Hamill OP, McBride DW Jr (1992) Rapid adaptation of single mechanosensitive channels in Xenopus oocytes. Proc Natl Acad Sci USA 89(16):7462-7466 13. Hudspeth A J (1989) How the ear's works work. Nature 341: 397 - 404 14. Lane JW, McBride DW Jr, Hamill OP (1991) Amiloride block of the mechanosensitive cation channel in Xenopus oocytes. J Physiol (Lond) 441:347- 366 15. Loewenstein WR, Mendelson M (1965) Components of receptor adaptation in a pacinian corpuscle. J Physiol (Lond) 177:377-397 16. Martinac B (1992) Mechanosensitive ion channels: biophysics and physiology. In: Jackson MB (ed) Thermodynamics of cell surface receptors. CRC, Boca Raton (in press)

612 17. Martinac ]3, Buechner M, Delcour AH, Adler J, Kung C (1987) Pressure-sensitive ion channels in Escherichia coil Proc Natl Acad Sci USA 84:2297-2301 18. Methfessel C, Witzemann V, Takahashi T, Mishina M, Numa S, Sakmann B (1986) Patch clamp measurements on Xenopus laevis oocytes: currents through endogenous channels and implanted acetylcholine receptor and sodium channels. Pflilgers Arch 407:577- 588 19. Moody WJ, Bosma MM (1989) A nonselective cation channel activated by membrane deformation in oocytes of the ascidian Boltenia villosa. J Membr Biol 107:179-188 20. Morris CE (1990) Mechanosensitive ion channels. J Membr Biol 113:93 - 107 21. Morris CE, Horn R(1991) Failure to elicit neuronal macroscopic mechanosensitive currents anticipated by single channel studies. Science 251:1246-1249

22. Ohmori H (1985) Mechano-electrical transduction currents in isolated vestibular hair cells of the chick. J Physiol (Lond) 359: 189-217 23. Sachs F (1988) Mechanicai transduction in biological systems. CRC Crit Rev Biomed Eng 16:141-169 24. Sackin H (1989) A stretch-activated K + channel sensitive to cell volume. Proc Natl Acad Sci USA 86:1731-1735 25. Sokabe M, Sachs F (1990) The structure and dynamics of patchclamped membranes: a study using differential interference contrast light microscopy. J Cell Biol 111:599-606 26. Taglietti V, Toselli M (1988) A study of stretch-activated channels in the membrane of frog oocytes: interactions with Ca ++ ions. J Physiol (Lond) 407:311 - 328 27. Ubl J, Murer H, Kolb H-A (1988) Ion channels activated by osmotic and mechanical stress in membranes of opossum kidney cells. J Membr Biol 104:223-232