Neuroscience has a lot of mantras. I blame textbooks.
The concept of “one neurotransmitter per neuron” nicely streamlines any discussion of neuron types. The problem: it’s at best reflection of 80-year-old dogma, and a wild over-simplification. So the fact that evidence to the contrary is rarely found in textbooks should … not surprise you.
The 1930s gave us many things: instant coffee, trampolines, and most relevant to this post, Dale’s Principle, which states that, “the nature of the chemical function … is characteristic for each particular neurone, and unchangeable”. Although this assumption remains the default, co-release of neurotransmitters has been formally discussed (read: published about) since at least 1976. In the past decade, the idea that neurons can release more than one neurotransmitter has gained ever wider acceptance amongst neuroscientists, with the list of brain regions containing co-releasing neurons growing rapidly. For example, multiple studies have reported the co-release of glutamate and various neuromodulators, including serotonin, dopamine and acetylcholine (for those with access, see this pay-walled review).
Let’s consider a couple of hypothetical neurons. At the presynaptic terminals of a single-neurotransmitter-releasing neuron, each synaptic vesicle will be filled with, and releasing, the same neurotransmitter.
But what about neurons that co-release multiple neurotransmitters?
Although chemical markers of multiple neurotransmitters consistently co-localize within the cell bodies of co-releasing neurons, these neurotransmitters are often differentially released from the axon terminals. Differential release may be mediated by activity-dependent mechanisms: trains of action potentials induce co-release of acetylcholine and glutamate from medial habenula neurons, whereas single action potentials solely evoke glutamate release (shown here, warning: paywall). But differential release can also be achieved through morphological segregation, with neurotransmitter localization differing regionally between subsets of synaptic terminals or synaptic vesicles.
What this means is that specific branches of the co-releasing neuron’s axon might contain distinct neurotransmitters. Alternatively, neurotransmitters might co-localize within the axonal arbor as a whole, but not within specific boutons, such that individual release sites will release a single neurotransmitter exclusively. Finally, co-localization might occur at the level of single boutons, such that each synapse can release both ACh and glutamate, albeit from distinct pools of synaptic vesicles. Distinguishing between these possibilities is a job for high-resolution imaging, and is no easy task.
Do these differences matter?
Well, the functional consequences of differential neurotransmitter release will strongly depend on the projection targets of the co-releasing neurons, and the identity of the neurotransmitters involved. It remains an open question for just about every instance of co-release reported so far.
As for your second question, the short answer is, yes: every dendrite contains many, many postsynaptic zones and each zone will contain multiple copies of any given receptor for a neurotransmitter. Furthermore, the number of available neurotransmitter binding sites plays an important role in the dynamics of neuronal communication. Different receptors have different binding dynamics; the kinetics of neurotransmitter release and binding have been thrilling both molecular biologists and computational modelers for years.