Stephen W. Kuffler

Presynaptic inhibition at the crayfish neuromuscular junction

There are three possible ways in which synaptic inhibition may reduce excitation. (1) Inhibition may reduce the amount of excitatory transmitter that is released from nerve terminals when they are activated by a nerve impulse; this is presynaptic inhibition. (2) The reaction of the postsynaptic receptors to a constant amount of released transmitter may be reduced, for instance by competition. (3) The post-synaptic membrane may be altered in such a way that the depolarizing action of the excitatory transmitter is opposed. There exists good evidence for the last alternative only, namely, that the membrane conductance of the post-synaptic region is increased, with the membrane potential staying near the resting potential. This inhibitory mechanism of a specific post-synaptic permeability increase (to K+ and/or Cl-) seems to be widespread in different species and has been demonstrated in the vertebrate heart, the mammalian central nervous system, in crustacea at neuromuscular junctions and nerve cell synapses, as well as in various other preparations (Fatt & Katz, 1953; Coombs, Eccles & Fatt, 1955; Kuffler & Eyzaguirre, 1955; Trautwein & Dudel, 1958; for a recent review see Kuffler, 1960). In this study on the crayfish neuromuscular junction we present evidence for a presynaptic mechanism. The inhibitory nerve impulse acts on the excitatory nerve terminals and decreases the probability of release of quanta of excitatory transmitter. A prelimiInary report has appeared (Dudel & Kuffler, 1960).

The number of transmitter molecules in a quantum: an estimate from iontophoretic application of acetylcholine at the neuromuscular synapse.

1. The sensitivity of the subsynaptic membrane of twitch muscles of the frog and snake to iontophoretically applied acetylcholine (ACh) was determined. Optimal placement of ACh micropipettes on to the postsynaptic membrane resulted in potentials that were similar, though not identical, to the miniature excitatory post‐synaptic potentials (min e.p.s.p.s). A sensitive bio‐assay was developed to measure the output of ACh from micropipettes; this allowed an estimate to be made of the upper limit of the number of ACh molecules in a quantum of transmitter that is released from the nerve to produce a min e.p.s.p. 2. The assay to calibrate the output of ACh from micropipettes used the end‐plate of the snake muscle as an ACh concentration detector. The end‐plate was situated within a few mum of an oil‐water interface, and a 0‐6 nl. droplet of Ringer solution containing a known concentration of ACh (1 muM or less) was formed in the oil phase. The droplet was brought to the interface and, upon touching it, discharged its contents into the Ringer phase immediately above the end‐plate. This resulted in a membrane depolarization that was recorded with an intracellular microelectrode. By applying droplets containing various known ACh concentrations a standard curve was constructed. To measure the ACh output of micropipettes a 0‐6 nl. droplet of Ringer solution was suspended in the oil. The ACh pipette tip was inserted into the droplet and several thousand pulses of ACh were then delivered. The ACh content of the test droplet was measured by comparing its effectiveness in depolarizing the end‐plate with the standard curve. In this manner the number of ACh molecules released in a single pulse was determined as a function of charge passed through the pipette. The output of ACh was linear and an average of 30,000 molecules of ACh were released per pC. 3. The sensitivity of the subsynaptic membrane to iontophoretically applied ACh, using the linear slopes of dose‐response curves, in preparations from frog and snake treated with anticholinesterases was usually about 5 mV/pC. It follows that 6000 molecules of ACh are sufficient to produce a depolarization of 1 mV in the subsynaptic membrane. 4. The mean min e.p.s.p.s of muscle fibres treated with anticholinesterase range from 1 to 3 mV. Since the ACh released from an iontophoretic pipette is less effective than the same amount released from the nerve, it is concluded that a quantum of transmitter consists of less than 10,000 molecules of ACh. 5. It is calculated that for each molecule of ACh released in a quantum there results a minimum net flow of 3000 univalent ions across the synaptic membrane.

Small‐nerve junctional potentials. The distribution of small motor nerves to frog skeletal muscle, and the membrane characteristics of the fibres they innervate*

The ventral roots in the frog contain two distinct groups of motor nerve fibres, which subserve two different functions in their innervation of skeletal muscles. Repeated stimulation of nerve fibres of small diameter, about 5 u, causes slow, graded muscular contractions which are accompanied by non-propagated muscle potentials of small amplitude and relatively long duration (Tasaki & Mizutani, 1944; Tasaki & Tsukagoshi, 1944). The larger nerve fibres, in the region of 12 p diameter, set up the familiar motor unit twitches accompanied by fast propagated muscle action potentials. Kuffler & Gerard (1947) demonstrated that when one or a few small-diameter motor nerve fibres were isolated by dissection and stimulated, the resulting muscle potentials were confined to the regions of the neuromuscular junctions; they were accordingly named Small-nerve junctional potentials (s.j.p.s). The mechanical effects of stimulation of the small motor nerves were studied in detail by Kuffler, Laporte & Ransmeier (1947) who found that quite large tensions could be developed in numerous muscles widely distributed over the body. Differences in the organization of the reflex responses of small and large-diameter motor fibres were also demonstrated. It was concluded that the small motor nerves are part of a specialized neuromuscular system with a specific holding or tonic function in the maintenance of the frogs posture. In the latter two studies it could not be decided whether the two groups of nerves went to different types of fibres within the muscle or whether both innervated the same single muscle fibres. The possibility existed that a single muscle fibre could give two different types of response, according to the type

Synaptic excitation and inhibition resulting from direct action of acetylcholine on two types of chemoreceptors on individual amphibian parasympathetic neurones

1. Synaptic transmission was studied in visually identified parasympathetic ganglion cells that modulate the heart beat of the mudpuppy Necturus maculosus).

Synaptic Transmission and Its Duplication by Focally Applied Acetylcholine in Parasympathetic Neurons in the Heart of the Frog

Synaptic transmission has been analysed in parasympathetic nerve cells that lie in the transparent interatrial septum of the heart of the frog. Using Nomarski interference optics, one can see much cellular detail, including synaptic boutons in living preparations. 1. On each ganglion cell, the 10 to 20 synaptic boutons are usually derived from a single vagal nerve fibre. These fibres branch extensively to innervate a number of septal ganglion cells. 2. The chemical transmitter, acetylcholine (ACh), liberated by a presynaptic impulse survives for up to 40 ms, setting up an excitatory postsynaptic potential (e.p.s.p.) which triggers one and sometimes two action potentials in the postsynaptic cell. The e.p.s.p. is made up of quantal components, as at the neuromuscular junction. 3. Nerve-evoked e.p.s.p.s can be well matched in amplitude and time course by iontophoretic application of ACh to selected areas of the neuronal membrane. In particular, the miniature e.p.s.p., which is due to the focal release of a small quantity of transmitter, was accurately mimicked by iontophoretic application of ACh. By grading the amount of ACh released from an electrode one could also duplicate the wide variety of nerve-evoked postsynaptic discharges of ganglion cells. 4. The permeability changes initiated in the postsynaptic membrane by applied ACh and the synaptic transmitter appear identical, since the ionic fluxes for both responses have the same equilibrium potential. Also, the receptors which react with the synaptic transmitter are desensitized by applied ACh. 5. Cholinesterase inhibitors (Tensilon and Eserine) have a variable action on different cells, with respect both to nerve-evoked and Ach evoked potentials. The reasons for this variation are unclear, and need further study. 6. Miniature e.p.s.p.s resemble analogous potentials at nerve-muscle junctions and other synapses. A significant proportion of the min e.p.s.p.s is released as multiple units. This proportion is increased in high Ca2+, while single units alone occur in a low Ca2+-high Mg2+ environment. 7. The experiments provide information about the release of ACh from nerve terminals and its action on the postsynaptic membrane of neurons. They are in good agreement with analogous studies on skeletal neuromuscular junctions

Visual identification of synaptic boutons on living ganglion cells and of varicosities in postganglionic axons in the heart of the frog.

1. Parasympathetic neurons were studied in the transparent interatrial septum of the frog (Rana pipiens) with light- and electron-microscopic techniques. The aim was to identify visually cellular and subcellular details in a living preparation, especially synaptic boutons on ganglion cells and the varicosities in postganglionic axons supplying the muscles of the heart. 2. The interatrial septum contains the following nervous elements: unipolar parasympathetic ganglion cells, their preganglionic vagal innervation, postganglionic sympathetic axons and sensory fibres. These structures and the nuclei of their related Schwann cells can be viewed with various optical systems, especially differential interference contrast optics. The same neural elements identified in the live preparation can be sectioned for electron microscopy. 3. Most ganglion cells are innervated by a single presynaptic axon, terminating in up to 27 synaptic boutons which on the average cover about 3.0 % of the surface of nerve cell bodies. A few scattered boutons also occur on the initial axonal portion of the ganglion cells. 4. Synaptic boutons on ganglion cells were recognized in the living unstained preparation. Their identity was confirmed by electron microscopy and by light microscopy combined with methylene blue, zinc iodide and osmium, and cholinesterase staining methods. 5. The terminal branches of postganglionic axons have numerous varicosities along their course. Some are as close as a few hundred angstroms (10 Å = 1 nm) to muscle fibres, others are many pm away. There are two types of varicosities: (i) those which contain predominantly granular vesicles characteristic of neurons releasing catecholamines, and (ii) those with predominantly agranular vesicles which belong to the cholinergic axons of septal ganglion cells. Regardless of their distance from muscle fibres, the cholinergic varicosities have the same fine structural features, including membrane thickenings, as synaptic boutons on the ganglion cells. These findings support earlier suggestions that the varicosities along postganglionic axons are a series of transmitter release sites. 6. Varicosities were observed in the live septum; their identity was confirmed by subsequent electron microscopy. Many live varicose axons were traced back to the vicinity of individual septal ganglion cells. Additional evidence that they belonged to a particular ganglion cell, and were therefore cholinergic, was obtained by injecting Procion yellow into the cell body and observing the neuron with a fluorescence microscope after the dye had spread into the axonal processes. Time lapse photography of up to 24 h showed no ‘ peristaltic ’ movement of varicosities. 7. Granular or agranular vesicles also occur along cylindrical axons within nerve bundles many pm away from muscle fibres. Like the vesicles in varicosities, they are clustered close to ‘thickenings’ in the surface membrane and belong to postganglionic nerve fibres. 8. Ganglion cells in isolated septa survive for 2 weeks or longer, still giving membrane potentials and impulses. Time lapse cinematography for up to 2 weeks after removing the septum showed that the organelles within the neurons were in motion and that a two-way traffic takes place between the cell body and axon, as commonly found in cultured neurons.

The Ferrier Lecture: Neuroglial Cells: Physiological Properties and a Potassium Mediated Effect of Neuronal Activity on the Glial Membrane Potential

Neuroglial cells constitute a separate class of cells in the nervous system; they have been studied intensively since their original description by Virchow in 1846. As a rule anatomists find no difficulty in recognizing them by their staining properties, their shape and configuration as well as by their characteristic location between and around neurons. Electron microscopy has in recent years added much important subcellular detail and has shown how intermingled neurons and glial cells are, being separated from each other by narrow clefts 100 to 200 Å wide (figures 1 A, B and 5, plates 1, 2 and 4). These studies have not changed the well-established grouping of mammalian glial cells into two main classes, the oligodendrocytes and the astrocytes. It is customary to state that glial cells outnumber neurons by 10 to 1 in the vertebrate nervous sytem. They are, however, smaller and according to some rough estimates they make up as much as 50% of the volume of mammalian brains. That glial cells differ significantly from neurons was clear from the beginning because they do not possess axons and, unlike mammalian neurons, they retain their ability to divide throughout life. The possible role of the large mass of glial cells in our nervous system has been of continued interest. During the past decade this interest in the physiology of neuroglia has been reinforced, largely under the stimulus of electron-microscopic and chemical studies of the nervous system. Among the numerous recent reviews and symposia only a few will be mentioned (Windle 1958; Nakai 1963; Mugnaini & Walberg 1964). The recent studies of the physiology of neuroglial cells have been reviewed by Kufller & Nicholls (1966) and a bibliography on neuroglia has been compiled by Little & Morris (1965).

Further study of efferent small-nerve fibres to mammalian muscle spindles. Multiple spindle innervation and activity during contraction

In a previous paper (Kuffler, Hunt & Quilliam, 1951), it was shown that only about two-thirds of the nerve fibres in the lumbosacral ventral root outflow in the cat are directly concerned with the production of muscle tension. These are the large-nerve fibres with diameters of 8-18,., and conduction velocities of about 50-110 m./sec.; they set up the well-known motor unit twitch response. The remaining fibres, forming a distinct group of smaller diameter (3-8,u.) and conducting at rates between 15 and 50 m./sec., were found to increase the sensory discharge originating in muscle spindles (Matthews A-type receptors). They serve as a nervous control mechanism of such stretch receptors. Thus, afferent discharges from muscle spindles can be set up by external muscle stretch or by excitation through the efferent small-nerve group. The term spindle receptor was applied in this study to all stretch receptors giving an A-type response (see Hunt & Kuffler, 1951). During the present investigation, it was found that the effect of small-nerve stimulation on the discharge from muscle spindles could be recorded during muscle contraction, and in this way the influence of specific spindle innervation on afferent activity was analysed during the changing tension conditions of muscle contraction. The present findings also enabled an approach to additional problems of spindle function. While in the previous investigation stimulation was limited to single efferent fibres to individual spindles, the present experiments permitted stimulation of several efferent fibres to the same spindle and provided physiological evidence for multiple innervation of spindles by smallnerve fibres. The response from a spindle could be analysed under conditions of full or at least near-maximal nervous activation. It was shown that the afferent discharge pattern in a single sensory fibre could be significantly modified,

The Development of Chemosensitivity in Extrasynaptic Areas of the Neuronal Surface after Denervation of Parasympathetic Ganglion Cells in the Heart of the Frog

1. The vagosympathetic trunks innervating the heart of the frog were cut on both sides. Two to 28 days following this denervation the chemosensitivity of the surface of the denervated neurons was explored with iontophoretic microapplication of ACh to restricted areas. 2. While in normally innervated neurons the synaptic areas alone are highly chemosensitive, after denervation the whole cell surface becomes sensitive to ACh. Synaptic transmission fails on the second day (30 to 40 h) after denervation (in frogs kept at 22 to 24 °C) and at the same time new chemoreceptive areas start to appear. After 4 to 8 days this development of chemosensitivity has reached a peak and remains at the same level for 4 weeks (the longest period of the present tests). 3. Measurements of chemosensitivity from different cells in different animals were compared. The finely localized chemosensitivity at synaptic areas in normally innervated neurons was of similar magnitude as the uniformly distributed sensitivity in denervated neurons.

Dark adaptation, absolute threshold and purkinje shift in single units of the cat's retina

Granit (1943, 1944) observed that the threshold of retinal ganglion cells in the cat drops when they are allowed to dark-adapt after exposure to bright lights, and that there is a shift in spectral sensitivity analogous to the Purkinje shift in the human. The drop in threshold was rather small compared to that found in human dark adaptation, and the final threshold reached was not nearly as low as one would expect in an eye adapted for nocturnal vision. Pirenne (1954) has emphasized that there is a big discrepancy (a factor of 103-106) between the lowest threshold obtained by electro-physiological methods and the much lower thresholds found by psychophysical methods in humans (Hecht, Shlaer & Pirenne, 1942). In addition, the existence of a random maintained discharge from retinal ganglion cells (Kuffler, FitzHugh & Barlow, 1957) raises some doubts about the relative parts played by the retina and the central nervous system in determining thresholds. We therefore set out to measure the absolute threshold of the large type of ganglion cell (Rushton, 1949) isolated by metallic micro-electrodes of the type used by Granit & Svaetichin (1939). The opportunity was taken of following threshold through the period of dark adaptation, and some observations of the Purkinje shift were made.