brief communications

Tarantula peptide inhibits atrial fibrillation A peptide from spider venom can prevent the heartbeat from losing its rhythm.

NATURE | VOL 409 | 4 JANUARY 2001 | www.nature.com

AF inducibility (%)

c 100 80 60 40 20 0 0

b

P = 2.5 cm H2O

5 10 15 20 25 30 Intra-atrial pressure (cm H2O)

d 60 AF duration

A

a

P = 5.0 cm H2O

Baseline GsMTx-4

50 40 30 20 10 0

P = 7.5 cm H2O

0

5 10 15 20 25 30 Intra-atrial pressure (cm H2O)

e 80 P = 10.0 cm H2O

ERP (ms)

trial fibrillation is the most common sustained cardiac arrhythmia to occur in humans, secondary to valve disease, hypertension or heart failure. It is often associated with passive stretching of the atrial chamber arising from haemodynamic or mechanical dysfunction of the heart. Here we show that atrial fibrillation potentiated by dilatation in rabbit heart can be inhibited by blocking stretch-activated ion channels with a specific peptide from tarantula venom, without altering the resting action potential. Our findings open a window on cardiac arrhythmogenesis and point the way towards developing a new class of drugs. Stretch-activated ion channels (SACs) are ubiquitous and have been studied in a variety of cardiac tissues1–3. Non-selective cationic SACs are permeant to Ca2& ions as well as to Na& and K&, whereas others are selective for K& and possibly Cl1 ions1,2. According to computer models, the excitatory currents carried by SACs can generate fast arrhythmias4,5, but a lack of specific inhibitors has prevented the acquisition of matching data. GsMtx-4 is a small (relative molecular mass, 4K) peptide found in the venom of the tarantula Grammostola spatulata (Fig. 1a). It specifically blocks cationic SACs in astrocytes and inhibits volume-activated currents in both adult astrocytes and cardiocytes. As a member of the neuroinhibitory ‘cysteine-knot’ family, GsMtx-4 has a binding affinity of about 500 nM for SACs in astrocytes. The specificity of its action is indicated by its lack of effect on the resting properties of rabbit ventricular cells and rat astrocytes6. At 4 mM (about 20 times the dose applied to suppress fibrillation), GsMtx-4 had no measurable effect on the action potential of isolated atrial heart cells (C. Baumgarten and H. Clemo, personal communication). Using perfused rabbit hearts, we initiated fibrillation with a burst of high-frequency stimulation (Fig. 1b) at different rightatrial pressures7,8. Stretching the chamber increased the incidence and duration of fibrillation (Fig. 1b–d). At pressures above 12 cm H2O, the probability of sustained fibrillation (for longer than 60 s) approached unity. Perfusion with 170 nM GsMtx-4 suppressed both the incidence and duration of fibrillation in all hearts (n410). At pressures below 17.5 cm H2O, sustained fibrillation was completely inhibited in all preparations (data not shown). The gadolinium ion (Gd3&), a common but non-selective SAC inhibitor, can also suppress stretch-induced cardiac fibrillation8, but its lack of specificity and non-applicability

P = 12.5 cm H2O

70 60 50 40

Sustained AF

0 2s

5 10 15 20 Intra-atrial pressure (cm H2O)

Figure 1 Inhibition of atrial fibrillation by GsMtx-4 during stretch. a, The spider Grammostola spatulata, whose venom is the source of the inhibitory peptide. b, Bipolar atrial electrograms showing an increase in atrial fibrillation (AF) with pressure, becoming sustained at 12.5 cm H2O. The probability of inducing AF was increased by stimulating the heart with a short burst of high-frequency pacing before each measurement (arrow). c, Induction of AF lasting more than 2 s: circles, control; filled circles, in the presence of 170 nM GsMtx-4. Dashed line indicates the response after 20-min washout. d, Duration of AF (n47) as a function of pressure (mean5s.e.). GsMtx-4 (170 nM) decreased the average time to spontaneous recovery from AF (asterisks, P*0.05). e, GsMtx-4 did not block stretch-induced shortening of the refractory period (n410). For details of recording methodology, see ref. 8.

under physiological conditions9 restrict its testing on SACs. Despite their different chemical structures, however, both GsMTx-4 and Gd3+ suppress fibrillation in a similar manner without altering the stretch-dependence of the effective refractory period8 (Fig. 1e), perhaps because both reagents have a high positive-charge density. The antifibrillatory effect we observe here cannot be driven by a change in surface-charge density, however, as this would alter the shape of the action potential. A different mechanism is also indicated by the persistence of the stretch-induced shortening of the refractory period while fibrillation is inhibited. It is possible that K&-selective SACs10, which are resistant to Gd3& (ref. 11) and perhaps to GsMtx-4 as well, act to shorten the action potential under stretch. The fundamental mechanism of fibrillation is viewed as the fragmentation of excitatory wavefronts, but the role of SACs in this process has not been studied directly. © 2001 Macmillan Magazines Ltd

SACs convert gradients of stress into gradients of electrical activity, and sustained mechanical gradients can develop, particularly during repolarization, when the systolic pressure and the wall tension are high while contraction is waning. These gradients are exacerbated in regions of the heart weakened by myocardial infarction, for example. Stretch sensitivity is not unique to the atrium, so GsMtx-4 should act similarly on all chambers12. As an investigative tool, this peptide should be useful for studying mechanical transduction at the molecular and whole-organism levels. We believe that GsMtx-4 could be the first of a new class of antiarrhythmic agents to be directed against the causes rather than the symptoms of fibrillation. Frank Bode*†, Frederick Sachs‡, Michael R. Franz* *Department of Pharmacology, Georgetown University, 3900 Reservoir Road, Washington DC 20007, USA 35

brief communications

Photonic engineering

Aphrodite’s iridescence he most intense colours displayed in nature result from either multilayer reflectors or linear diffraction gratings1–3. Here we investigate the spectacular iridescence of a spine (notoseta) from the sea mouse Aphrodita sp. (Polychaeta: Aphroditidae). The spine normally appears to be deep red in colour, but when light is incident perpendicular to the axis of the spine, different colours are seen as stripes running parallel to the axis of the spine; over a range of smaller incident angles, the complete visible spectrum is reflected with a reflectivity of 100% to the human eye. The simple structure responsible for this effect is a remarkable example of photonic engineering by a living organism. Photonics is a branch of optics that concerns the control of photons, much as electronics depends on electrons, and photonic structures have been described in butterfly wings4. The sea mouse has iridescent spines situated at the base of the dorsal surface, where they can be viewed laterally, with less strongly coloured threads being seen more readily from above (see http:// www.physics.usyd.edu.au/~nicolae/submitted.html). The biological function of the iridescence is unknown, but it may be related to species recognition or to courtship. We prepared a spine with strong reflectance in the red for electron microscopy by mounting it in resin and slicing it with a microtome into sections transverse to the long axis. These sections were of constant thickness and were mounted on grids for transmission microscopy. The micrograph in Fig. 1a shows an array of hollow cylinders, with the long axis of the cylinders along the spine. The image is of the wall region of the spine, the centre of the spine being hollow. The hollow cylinders have thicker walls near the edge of the spine, possibly for mechanical rigidity, but after a few layers their wall thickness decreases to a constant value. The cylinders are closepacked, with each cylinder having six nearest neighbours. The spacing of adjacent layers of cylinders is 0.51 mm, and remains constant

T

36

throughout the wall cross-section, which contains 88 layers of close-packed voids. The cylinder walls are made of pure a-chitin5,6 (refractive index, 1.54; ref. 2). To calculate the expected optical properties of the spine, we used a formulation for layered stacks of cylinder gratings7. Figure 1c shows the reflectance of a 500-layer stack of gratings with period d40.51 mm, for radiation incident in the plane of Fig. 1a, a

b

kd/π

1. Ruknudin, A., Sachs, F. & Bustamante, J. O. Am. J. Physiol. 264, H960–H972 (1993). 2. Hagiwara, N., Masuda, H., Shoda, M. & Irisawa, H. J. Physiol. (Lond.) 456, 285–302 (1992). 3. Sato, R. & Koumi, S. J. Membr. Biol. 163, 67–76 (1998).

with its electric vector perpendicular to the plane (E||). The grating consists of hexagonally packed cylindrical holes of radius a40.2 mm, filled with sea water (refractive index, 1.33), in a matrix of chitin. The bulk reflectance of chitin immersed in sea water is 0.54%, so the strong reflectance evident in Fig. 1c can only be achieved through the coherent stacking of scores of layers to form the crystal shown in Fig. 1a. Were the layers to be of irregular separation or composition, the peak reflectance for a given number of layers would be much lower7 and the iridescence less pronounced. We calculated the total reflectance of this structure, for normal incidence, for free-space wavelengths in the visible spectrum (from 0.4 to 0.7 mm). Figure 1c shows a strong maximum in reflectance around the wavelength l40.633 mm. To investigate the origin of this narrow-band reflectance, we calculated the photonic band diagram, which gives the frequencies

4. Sachs, F. in Cell Mechanics and Cellular Engineering (eds Mow, V. C. et al.) 308–328 (Springer, New York, 1994). 5. Zabel, M., Koller, B. S., Sachs, F. & Franz, M. R. Cardiovasc. Res. 32, 120–130 (1996). 6. Suchyna, T. M. et al. J. Gen. Physiol. 115, 583–598 (2000). 7. Ravelli, F. & Allessie, M. Circulation 96, 1686–1695 (1997). 8. Bode, F., Katchman, A., Woosley, R. L. & Franz, M. R. Circulation 101, 2200–2205 (2000). 9. Caldwell, R. A., Clemo, H. F. & Baumgarten, C. M. Am. J. Physiol. 275, C619–C621 (1998). 10. Small, D. L. & Morris, C. E. Br. J. Pharmacol. 114, 180–186 (1995). 11. Kim, D. J. Gen. Physiol. 100, 1021–1040 (1992). 12. Nazir, S. A. et al. J. Physiol. (Lond.) 494P, 111–112 (1996).

M

K

Γ

Γ

...M

c 1.0

0.8

0.6 Reflectance

‡Department of Physiology and Biophysics, State University of New York, 124 Sherman Hall, Buffalo, New York 14214, USA e-mail: [email protected] †Present address: Medizinische Klinik II, Universitaetsklinikum Luebeck, Ratzeburger Allee 160, 23538 Luebeck, Germany

0.4

0.2

0 0.40

0.45

0.50

0.55 λ (µm)

0.60

0.65

0.70

Figure 1 Structure and optical properties of a sea mouse hair from Palm Beach, New South Wales, Australia. a, Electron micrograph of a section of the spine (the dark regions are chitin). b, E|| polarization photonic band structure for a hexagonal crystal of voids in chitin, near the four-fold degenerate point on the G-axis (dashed lines indicate a partial band gap). G, M and K are the corners of the irreducible part of the first Brillouin zone; k is the free space wavenumber. c, E|| polarization total reflectance for normally incident radiation for a hexagonal structure having 500 layers of voids in chitin (dashed lines indicate the partial bandgap of b). Reflectance measurements using a Zeiss MMS spectrometer confirm a 100% reflectance in the red. © 2001 Macmillan Magazines Ltd

NATURE | VOL 409 | 4 JANUARY 2001 | www.nature.com