Scanning ion conductance microscopy (SICM) is a scanning probe microscopy particularly suitable for the investigation of living biological specimens due to its low invasivity. Recently, this technique has been used not only to perform 3D-imaging, but also to stimulate and guide neuronal growth cones. In particular, it has been demonstrated that one can guide the cone growth for tens of micrometres by means of recurrent and non-contact SICM scanning along a defined line, with a pipette having an internal hydrostatic pressure. Accurate measurements of the mechanical forces acting on the cell membrane in these stimulation protocols are essential to explain the biological mechanisms involved. Herein a setup specifically developed for this purpose, combining together SICM, atomic force microscopy (AFM) and inverted optical microscopy is described. In this configuration, a SICM pipette can be approached to an AFM cantilever while monitoring the cantilever deflection as a function of the hydrostatic pressure applied to the pipette and the relative distance. In this way, one can directly measure mechanical forces down to 20 pN. The same apparatus is thus sufficient to calibrate a given pipette and immediately use it to study the hydrostatic pressure effects on living cells.

Behavioral responses of Halobacterium salinarum appear as changes in the frequency of motion reversals. Turning on orange light decreases the reversal frequency, whereas blue light induces reversals. Light pulses normally induce the same response as step-up stimuli. However, anomalous behavioral reactions, including inverse responses, are seen when stimuli are applied in sequence. The occurrence of a prior stimulus is conditioning for successive stimulation on a time scale of the same order of adaptational processes. These prolonged conditioning effects are color-specific. The only adaptation process identified so far is methylation of the transducers, and this could be somehow color-specific. Therefore we tested for the behavioral anomalies in a mutant in which all methylation sites on the transducer have been eliminated. The results show that behavioral anomalies are unaffected by the absence of methylation processes on the transducer.

A minimal and convenient experimental set-up is described, which allows an easy characterization of crystal tuning forks, especially after the modifications introduced to exploit them as sensors in nanomechanics and in force microscopy techniques. The system uses the thermal noise of the crystal as test signal, a simple frequency converter for translating the signal itself into the audio-frequency band, and a PC sound card to acquire it and eventually perform a fast Fourier transform spectrum analysis on the noise samples. Our results show that the main decrease of the Q-factor of the tuning fork is caused by its acoustical coupling to the environment, while small masses added to either or both prongs only produce minor variations.

Experiments on the integration of blue and orange stimuli in Halobacterium salinarum were performed by using different combinations of blue and orange steps. The results show that the prevalence of the blue stimulus over the orange one depends on both the blue and the orange light intensities. A quantitative analysis of the current hypotheses on the phototransduction of orange and UV-blue light stimuli is presented, showing that the balancing between the two antagonistic stimuli should depend only on the intensity of the blue stimulus and not on that of the orange one, provided that the combination of the two stimuli occurs linearly at the photoreceptor stage. We conclude that blue and orange stimuli elicit distinct intracellular signals whose integration occurs downstream of the photoreceptor.

Here we resume the results of applying periodic forcing to a single ion channel. This approach, whose main aim was to test the existence of dynamical components in the channel switching, allowed us a direct evaluation of the rate constants of the channel transitions and showed a trend which is in keeping with stochastic resonance.

The sensitivity of sharks, skates, rays, and similar animals to extremely low electric fields is a popular topic in the nonlinear analysis community. It has been considered often in general editorial comments (Tsong, 1994Tsong T.Y. Exquisite sensitivity of electroreceptor in skates.Biophys. J. 1994; 67: 1367Abstract Full Text PDF PubMed Scopus (5) Google Scholar, Glanz, 1996Glanz J. Physicists advance into biology.Science. 1996; 272: 646-648Crossref PubMed Scopus (7) Google Scholar) and very recently in the “New and Notable” section of Biophysical Journal (Moss, 1997Moss F. Stochastic resonance at the molecular level.Biophys. J. 1997; 73: 2249-2250Abstract Full Text PDF PubMed Scopus (10) Google Scholar). Indeed, the sensitivity of some fish to electric fields appears to be astounding. Behavioral evaluation of the lowest field perceived by rays, found earlier to be 10 nV/cm in water (Kalmijn, 1982Kalmijn A.J. Electric and magnetic field detection in elasmobranch fish.Science. 1982; 218 (916–818)Crossref PubMed Scopus (232) Google Scholar), has recently been reduced to 1–2 nV/cm (Kalmijn, 1997Kalmijn Ad.J. Electric and near-field acoustic detection, a comparative study.Acta Physiol. Scand. 1997; 161: 25-38Google Scholar). Such a field, if applied directly to a sensory cell, produces a change in the transmembrane potential that is absolutely negligible in comparison with spontaneous fluctuations. On this premise, the usually accepted evaluation of 10−7 for the signal-to-noise ratio (SNR) is obtained (Block, 1992Block S.M. Biophysical principles of sensory transduction.in: Coirey D.P. Roper S.D. Sensory Transduction. Rockefeller University Press, New York1992: 1-17Google Scholar)—a really astonishing value. It seems impossible to detect such a low signal by ordinary means, thus lending credence to the idea that detection requires methods qualitatively different from traditional techniques of linear analysis. It has been said that in order to reach the sensitivity of fish “artificial devices … should work at liquid nitrogen temperature” (Glanz, 1996Glanz J. Physicists advance into biology.Science. 1996; 272: 646-648Crossref PubMed Scopus (7) Google Scholar). Let us take a closer look at the actual biological situation to evaluate both the signal produced at the cellular level by the external field and the noise in the membrane potential in a receptor cell. We start with the evaluation of the potential drop induced by the external field on the membrane of the electroreceptors. To make a straightforward comparison with the former evaluation, we will consider a field of 10 nV/cm, the value already used to calculate an SNR = 10−7. This value was obtained by assuming that the potential drop through the membrane of a receptor cell was equal to that occurring over the cell length (10 μm). Actually, the weak field generated by the prey is measured over a much longer distance: the Lorenzini ampulla is a relatively insulated organ, connected to the external water through insulated canals filled with a conductive jelly (Waltman, 1966Waltman B. Electrical properties and fine structure of the ampullary canals of Lorenzini.Acta Physiol. Scand. 1966; 66: 1-60PubMed Google Scholar, Kalmijn, 1974Kalmijn A.J. The detection of electric fields from inanimate and animate sources other than electric organs.in: Handbook of Sensory Physiology. vol. 3. Springer Verlag, New York1974: 147-200Google Scholar, Murray, 1974Murray R.W. The ampullae of Lorenzini.in: Handbook of Sensory Physiology. vol. 3. Springer Verlag, New York1974: 125-148Google Scholar). A simple way to enhance the electric signal before contamination with the noise in the membrane potential of the receptor is used; elasmobranchs (fish like sharks and rays) take advantage of their extended body. The potential drop on the receptors in the sensory epithelium corresponds to that occurring between the canal pore and the ampulla (5 cm for the longest canals in medium- sized dogfish) and is therefore of the order of 0.05 μV for a field of 10 nV/cm. A factor of 5 × 103 is therefore introduced with respect to the former evaluation. The noise in the membrane potential has been evaluated by applying basic physical principles (Weaver and Astumian, 1990Weaver J.C. Astumian R.D. The response of living cells to very weak electric fields: the thermal noise limit.Science. 1990; 247: 459-462Crossref PubMed Scopus (264) Google Scholar, Block, 1992Block S.M. Biophysical principles of sensory transduction.in: Coirey D.P. Roper S.D. Sensory Transduction. Rockefeller University Press, New York1992: 1-17Google Scholar). Schematically, a cell membrane can be represented as a first approximation by an equivalent circuit in which membrane resistance (R) and capacitance (C) are combined in parallel. Considering the voltage drop across a cell membrane, it is known by the Johnson-Nyquist theorem that 〈Vn2〉 = kBT/C, where kB and T have their usual meaning. For a spherical cell of radius r (10 μm), and assuming the specific capacitance of a lipid bilayer is 1 μF/cm2, C is about 3 pF; therefore, 〈Vn2〉1/2 = 30 μV. This is a lower limit for the noise that one might expect on the voltage across the membrane of an average-sized cell. This value cannot be lowered; however, 〈Vn2〉1/2 is inversely proportional to the square root of C and therefore decreases with increasing membrane surface. Moreover, it is important to evaluate the number of cells that connect the same area in the central nervous system. Figs. 1D and 1F in Murray, 1974Murray R.W. The ampullae of Lorenzini.in: Handbook of Sensory Physiology. vol. 3. Springer Verlag, New York1974: 125-148Google Scholar show that electroreceptors in each ampulla at the bottom of a canal number about 10,000, divided among a small number of swellings; the number of fibers innervating each ampulla (five) is about the same as the number of the swellings and they project to the same brain area. Therefore it is conceivable that the output of 104 cells are averaged in certain areas of the brain, and in this case the voltage thermal fluctuations might be as low as 0.6 μV root-mean-square. Biological excess noise is to be expected because ion channels switch continuously from the open to the closed conformation also because of thermal fluctuations. However, the current flowing through the open channel is due to the electrochemical gradient through the membrane, and the nonequilibrium situation is the actual source of the excess noise. The excess noise depends on the number and types of channels that are open in the resting condition. It is possible to retrieve experimental evaluations of 〈Vn2〉1/2 in receptor cells. For instance, the amplitude for the voltage fluctuations in retinal bipolar cells (cell membrane capacitance of 11 pF) of the larval amphibian axolotl, a kind of salamander, was evaluated to be 100–300 μV (Tessier-Lavigne et al., 1988Tessier-Lavigne M. Attwell D. Mobbs P. Wilson M. Membrane currents in retinal bipolar cells of the axolotl.J. Gen. Physiol. 1988; 91: 49-72Crossref PubMed Scopus (58) Google Scholar), whereas the voltage noise of an insect photoreceptor (cell membrane capacitance of 14 pF) is ∼100 μV (Stephenson, 1988Stephenson R.S. On the interpretation of voltage noise in small cells.J. Neurosci. Methods. 1988; 26: 141-149Crossref PubMed Scopus (3) Google Scholar). We made numerical simulations of the equivalent electric scheme of a cell membrane to obtain qualitative insight into the dependence of voltage noise on cell parameters. The obvious result was that voltage fluctuations due to the switching of ion channels decrease when the ion channel kinetics are fast compared to the membrane time constant, suggesting a possible way to minimize nonequilibrium noise in specialized structures. As for the fundamental thermal noise, a decrease of the excess voltage noise with the square root of the cell surface is to be expected. A reasonable estimate of the voltage noise due to ion channel switching in a single electroreceptor is therefore 200 μV (based on the data by Stephenson, 1988Stephenson R.S. On the interpretation of voltage noise in small cells.J. Neurosci. Methods. 1988; 26: 141-149Crossref PubMed Scopus (3) Google Scholar, normalized for the difference in cell radii); by the same reasoning used for thermal noise, we consider that averaging this noise over 104 cells results in a final root-mean-square value of 2 μV. However, it is worth recalling that the responses to voltage steps lasting 0.5 s in the excised ampullary organ have a dynamic range between −100 and 20 μV (Lu and Fishman, 1994aLu J. Fishman H.M. Interactions of apical and basal membrane ion channels underlies electroreception in ampullary epithelia of skates.Biophys. J. 1994; 67: 1525-1533Abstract Full Text PDF PubMed Scopus (30) Google Scholar, Lu and Fishman, 1994bLu J. Fishman H.M. Operational properties of voltage-clamped electroreceptive ampullary organ excised from Raja.Biol. Bull. 1994; 187: 257-258PubMed Google Scholar), and that a variation in the discharge of the nerve fibers has been measured for voltage steps of 3 μV, which suggests that the noise at low frequencies is only a fraction of a microvolt (and each nerve fiber averages over just 2 × 103 electroreceptor cells). Clearly, only direct experimental measurements can reveal how much noise originates in the electroreceptors and how its spectral distribution is shaped. The evaluation of SNR using the estimates of voltage noise (2 μV) and voltage drop (50 nV) reported above yields an SNR value of ≈10−2 for a canal of 5 cm. This value, however, should be divided by 10 according to the lowest evaluation of the threshold field, 1–2 nV/cm (Kalmijn, 1997Kalmijn Ad.J. Electric and near-field acoustic detection, a comparative study.Acta Physiol. Scand. 1997; 161: 25-38Google Scholar). This evaluation, very different from the usually accepted value of 10−7 for the SNR, is in agreement with considerations reported by Weaver and Astumian, 1990Weaver J.C. Astumian R.D. The response of living cells to very weak electric fields: the thermal noise limit.Science. 1990; 247: 459-462Crossref PubMed Scopus (264) Google Scholar (see their Table 1 and the explanation reported in their note 2). Could ordinary techniques detect signals with such SNR? By “ordinary techniques” we mean not only averaging of different units with independent noise, but also Fourier analysis or similar processing. Fourier analysis is a powerful tool for the detection of small signals, sorting them out of the noise contributions at different frequencies. This sorting out is in principle only limited by the time of observation. In a cell Fourier analysis may be implemented if the transduction system acts as a selective amplifier around a resonance frequency, as seems to occur in the excised electric organ (Lu and Fishman, 1994aLu J. Fishman H.M. Interactions of apical and basal membrane ion channels underlies electroreception in ampullary epithelia of skates.Biophys. J. 1994; 67: 1525-1533Abstract Full Text PDF PubMed Scopus (30) Google Scholar). As for further averaging, about 1000 canals of different length are present in a single animal (A. J. Kalmijn, personal communication) and the longest ones (about 20) are nearly parallel; i.e., to a reasonable approximation 20 equivalent units can be averaged. This could also help reduce the noise originating in the ohmic resistance of the canals. Thus with the present evaluation of SNR we propose that ordinary tools can work, whereas the former evaluation of 10−7 seemed to rule out this possibility. Only the experimental study of the system under physiological conditions can give direct answers to questions about how electroreception operates. However, there are insufficient reasons to assume that traditional techniques such as linear analysis do not work. The hope that systems exhibiting stochastic resonance can improve SNR has recently been denied by a clear-cut note in Nature (Dykman and McClintock, 1998Dykman M.I. McClintock P.V.E. What can stochastic resonance do?.Nature. 1998; 391: 344Crossref PubMed Scopus (79) Google Scholar). We thank Ad. J. Kalmijn for the afternoon he spent with us and for the many things that we learned from him that afternoon.

Halobacterium salinarum cells from 3-day-old cultures have been stimulated with different patterns of repetitive pulse stimuli. A short train of 0.6-s orange light pulses with a 4-s period resulted in reversal peaks of increasing intensity. The reverse occurred when blue light pulses were delivered as a finite train: with a 3-s period, the response declined in sequence from the first to the last pulse. To evaluate the response of the system under steady-state conditions of stimulation, continuous trains of pulses were also applied; whereas blue light always produced a sharply peaked response immediately after each pulse, orange pulses resulted in a declining peak of reversals that lasted until the subsequent pulse. An attempt to account for these results in terms of current excitation/adaptation models shows that additional mechanisms appear to be at work in this transduction chain.

We analyzed the motor photoresponses of Halobacterium salinarium to different test stimuli applied after a first photophobic response produced by a step-down of red-orange light (prestimulus). We observed that pulses given with a suitable delay after the prestimulus produced unusual responses. Pulses of blue, green, or red-orange light, each eliciting no response when applied alone, produced a secondary photophobic response when applied several seconds after the prestimulus; the same occurred with a negative blue pulse (rapid shut-off and turning on of a blue light). Conversely, no secondary photophobic response was observed when the test stimulus was a step (a step-up for red-orange light, a step-down for blue light) of the same wavelength and intensity. When the delay was varied, different results were obtained with different wavelengths; red-orange pulses were typically effective in producing a secondary photophobic response, even with a delay of 2 s, whereas the response to a blue pulse was suppressed when the test stimulus was applied within 5 s after the prestimulus. The secondary photophobic response to pulses was abolished by reducing the intensity of the prestimulus without affecting the primary photophobic response. These results, some of which were previously reported in the literature as inverse effects, must be produced by a facilitating mechanism depending on the prestimulus itself, the occurrence of reversals being per se ineffective. The fact that red-orange test stimuli are facilitated even at the shortest delay, whereas those of different wavelengths become effective only after several seconds, suggests that the putative mechanism of the facilitating effect is specific for different signaling pathways.

This paper reports a systematic study of the step-up photophobic responses exhibited by the unicellular alga Haematococcus pluvialis when stimulated unidirectionally or bidirectionally. The stimulus-response curves, obtained at four different wavelengths, are interpreted in terms of the structure of the photoreceptive apparatus. In addition, an indirect method to obtain information about the stigma and photoreceptor(s) is reported. This method is based on the interpolation of experimental with simulated data. Our results confirm the widely accepted view that the photoreceptors are located in the stigma region, and suggest the presence of two photoreceptors.

The general modeling of dose-response curves to very low stimuli in a photosensory-effector system is critically reshaped starting from basic assumptions on the fluctuations of chemical signals inside the receptor cell, which add to those of the stimulus itself, both arising from their granular (or quantal) structure, We have shown, both through the analytical treatment of a simple kinetic scheme and by means of Monte Carlo simulations of the same, that shot noise arising from chemical transduction (''chemical shot noise'') contributes considerably to the output noise of the receptor-effector system, thus affecting both the shape and the abscissa shift of dose-response curves under these conditions; the latter phenomenon has indeed been reported in Halobacterium halobium. After evaluating the general properties of a single-step amplifying mechanism, the effects of introducing several low-amplifying steps in cascade were investigated briefly. The results obtained were qualitatively and quantitatively at variance from those of earlier models on the same phenomenon, and the discrepancies are discussed in order to highlight the fundamental contribution of chemical shot noise to the response of any kind of sensory system to very low stimuli.

The contribution of chemical steps in trasduction chains to the output noise of a receptor cell (as opposed to the noise inherent in the stimulus,e.g., a dim light flash) has been evaluated both analytically and by a computer simulation study on a simple model, using a Monte Carlo technique. The noise deriving from chemical reactions (chemical shot noise, CSN) is shown to add to the source noise thereby increasing significantly the output variation coefficient. It is proposed that this noise contribution should be taken into account in interpreting the sensitivity of a sensory system and in particular the dose-response curves at low levels of the stimulus.

We consider here a simple example of stimulated sensory neurons operating under the influence of their own internal noise: the hair mechanoreceptor of the crayfish stimulated by a weak, periodic, hydrodynamic signal. Action potential spike trains from the sensory neuron are recorded and assembled into two objects for analysis: the interspike interval histogram (ISIH) and the cycle histogram of the spike density. A time transformation is carried out on the ISIH's in order to test the hypothesis that the spike train is basically random and that the probability of coherent spike generation is related to the instantaneous stimulus amplitude. Moreover it is shown that the physiological spike train data can be qualitatively mimicked by an electronic Fitzhugh-Nagumo model, operated in the subcritical mode, driven by noise and a weak periodic signal. A discussion of how the Fitzhugh-Nagumo model is properly operated to mimic noisy data from sensory neurons is included.

The effect of a weak periodic perturbation on the firing activity of two different types of neurons is investigated. Post-stimulus and interspike interval histograms are built up from the times of occurrence of the action potentials and a comparative analysis is performed between the two preparations. A discussion of how the leaky integrate-and-fire model can mimic the experimental data is included.