A synapse is a tiny gap between two neurons, or a neuron and its target tissue. The action potential cannot cross this gap, so the electrical impulse is transformed into a chemical signal that carries the impulse along to the next cell to have some effect. The effects can be quite varied, from acting as a simple relay of the signal to altering the metabolic state of the downstream cell, or altering the regulation of its genes. The simplest of these is an excitatory postsynaptic potential, which we will examine here.
OK, back to the ganglion. In the simplest case, a neuron will extend its axon from the spinal cord and form a synapse with a cell body in the ganglion. Consider an action potential traveling down this axon. It will propagate along, the depolarization of the membrane causing a cascade of opening and closing voltage gated ion channels down the axon until it reaches the bulbous structure at the end, the terminal bouton (aka, terminal knob). Here, the makeup of the plasma membrane changes. Mixed among the voltage gated sodium channels are various other ion channels, such as a voltage gated calcium channel. As you would expect, this channel behaves in a manner similar to the voltage gated sodium channel. It opens when the membrane is depolarized and allows calcium to flow into the cell. The terminal bouton has another feature important for our discussion — It is full of vesicles. These are little membrane bound bags stored in the bouton. In each "bag" is a mixture of neurotransmitters, small molecules or peptides that will act on the postsynaptic cell in some way. In our preganglionic sympathetic neuron, these vesicles are filled predominantly with acetylcholine.
As the wave of depolarization washes over the terminal bouton, these channels open, and the concentration of calcium inside the bouton rises. This calcium influx starts a cascade. First, it binds to an small protein inside the cell, called calmodulin (CaM). When bound to calcium, CaM changes its shape, and can then bind to other proteins in the cell that act as enzymes, altering the structure and activating other proteins involved in the scaffolding support matrix of the cell. All these work together to start moving some of the vesicles towards the end of the bouton, the part facing the small gap, the synaptic cleft, between the bouton and the next cell. Once the membrane around the vesicle comes in contact with the plasma membrane, another cascade starts and the two membranes fuse, dumping the contents of the vesicle into the gap. Normally, only one vesicle will make it all the way to the membrane. The rest stop before they get there. Sometimes, if there is a lot of stimulation, two or three may fuse.
Once in the gap, the acetylcholine (ACh) is faced with a gauntlet to run. The cleft is full of an enzyme called acetylcholinesterase, bound to the membranes and floating in the cleft. This is an enzyme capable of breaking down acetylcholine into a pair of inactive products that are then reabsorbed by the neuron that just dumped it out (this is reuptake). The ACh that does make it across can then bind to a protein on the postsynaptic cell. In our scenario, this is a nicotinic cholinergic receptor (nACHhR), a new type of ion channel — a ligand gated ion channel. Instead of being opened by a change in voltage, these are opened by the presence of a small molecule that specifically binds to it, known as a ligand. In our scenario, ACh is the ligand. These proteins are among a group of proteins called receptors. Receptors are not restrcted to neurons, nor are they always ion channels, but they all cause changes in a cell in response to a ligand.
When ACh binds to this receptor, it opens and allows an influx of sodium, pushing the cell towards depolarization. If enough of these channels are stimulated at once, the membrane potential will cross the threshold and a new action potential will start. This is an example of the simplest case — an excitatory postsynaptic potential.
But, as is said frequently in biological sciences, it's never really that simple.
The changes in voltage instigated by the opening of one channel are transient, restricted in space, and small in magnitude. They will quickly dissipate, and the membrane will return to its resting potential. These also only affect a small area of the plasma membrane; the drop in polarization is over only a very small area. Therefore, these signals must be summated both spatially and temporally. There must be enough signals received in a small enough area, and over a short enough period of time to force the membrane potential across the threshold. And even that is simplifying matters a bit.
Each neuron has a large number of synapses, of wide variety. On average, each central nervous system neuron receives input from about 2000 synapses that are spread out over the cell body, dendrites, axon and bouton. Not all of these are excitatory, nor are they all postsynaptic. Some synapses express GABA or glycine receptors, which are ligand gated chloride channels on the cell. Chloride has a negative charge, and is concentrated outside the cell. Opening a chloride channel will hyperpolarize a cell due to the influx of negative charge, which will make the cell harder to depolarize for a short time. This is an inhibitory potential, and can be post or presynaptic (normally on or near the bouton).
Presynaptic signals can hyperpolarize the terminal bouton, effectively quenching the inbound action potential. Other presynaptic inhibitory potentials inhibit the calcium channels on the bouton, interrupting the signal before it is passed to CaM. Other neurotransmitters are modulatory, and affect the metabolism of the target cell. These modulatory neurotransmitters don't normally act through ion channels, and have longer lasting effects. These can alter gene regulation, cause a reduction or increase in the number of excitatory or inhibitory receptors, affect scaffolding proteins, affect the milieu of neurotransmitters in that cell's bouton, reduce the sensitivity of CaM to calcium (this is a very recent discovery), change the localization of the receptors, or a host of other effects. In our example, the target neuron in the ganglion will receive input from other spinal cord neurons and from neurons in other parts of the paraveretebral chain. This input will be summated, and the signal will be sent out, or it will not.
Our ganglionic neuron must summate all of these inputs, inhibitory, excitatory, and modulatory before transmitting a signal. In the end, this signal is binary. It is a decision whether the neuron will fire, or it will not. Whether and action potential will propagate down the axon, or not. In our example, the downstream effects are immediate. A blood vessel will constrict, or a bronchiole will dilate, or a sphincter in the GI tract will close. In the central nervous system, complexity becomes further complexity. Here, the individual signal summated from thousands will itself become one more signal among thousands to be summated at the next neuron.