MT  >   3. Why are Action Potentials initiated at AIS?

It has been well-established that the threshold for generating action potentials at the axon initial segment (AIS) is lower than at somatodendritic domain (Royeck et al., 2008; Kole and Stuart, 2008). This feature explains why action potentials initiate at AIS. However, there is a major disagreement on the mechanism underlying the lowest threshold at AIS.

Assuming identical sodium channel properties in all regions of a pyramidal neuron, computational studies found that the sodium channel density at AIS must be orders of magnitude higher than at somatodendritic domain to simulate the lowest AIS threshold (Mainen at al., 1995). Surprisingly, patch clamp recordings revealed little difference in sodium channel density between AIS and the soma of pyramidal neurons (Colbert and Johnston, 1996; Colbert and Pan, 2002). It was argued that patch clamp could distort measurement of true channel densities (Kole et al., 2008). To eliminate this possibility, Fleidervish et al. (2010) employed high-speed fluorescence Na+ imaging, and found that the action potential-associated Na+ influx was only threefold larger at AIS than at the soma, and no difference between AIS and the first node of Ranvier.

This chapter aims to resolve the controversy by showing that the microtubules at AIS has capacity to amplify the effects of depolarizing fields on channel gating such that small depolarization at AIS is sufficient to increase the open probability of sodium channels dramatically.

Less Depolarization Is Required to Activate AIS Sodium Channels

By using patch recordings, Colbert and Pan (2002) demonstrated that in pyramidal neurons the AIS sodium channels were activated by 7 mV less depolarization than somatic sodium channels. To put it another way, the voltage dependence of sodium channels at AIS is shifted toward hyperpolarization by about 7 mV compared to those at the soma. This biophysical property, rather than high density, of AIS sodium channels may account for the lowest threshold for action potential generation. The finding is supported by the studies of Katz et al. (2018) using Na+ imaging, which also revealed hyperpolarization (leftward) shift in the voltage dependent of sodium channels in pyramidal neurons. Importantly, this feature is independent of sodium channel subtypes, whether they are Nav1.2 or Nav1.6 (Figure 1).


Figure 1. Normalized transient Na+ influx (ΔF) in AIS (blue) and soma (black) elicited by 2-second-long voltage ramps from −70 to 0 mV.
Left: in control neurons where Nav1.6 is the dominant sodium channel subtype in AIS.
Right: in conditional knockout (cKO) neurons where Nav1.6 is removed and compensated by Nav1.2 in AIS.
Note that in both control and cKO neurons the AIS Na+ influx began to change at significantly more negative voltages than somatic ΔF. The sharp rise in AIS could occur at the moment when the depolarizing field is sufficient to disrupt the binding between Ankyrin-G and microtubules, thereby allowing microtubules to move away from the membrane, and consequently activating sodium channels, whether they are Nav1.6 or Nav1.2. The soma lacks Ankyrin-G. Its sodium channels were activated directly by the depolarizing field, without involving microtubules.
[Adapted from: Katz et al., 2018]

On the basis of their data, Katz et al. (2018) suggested that the hyperpolarization shift probably reflects interactions of sodium channels with the proteins (e.g., Ankyrin-G) that characterize AIS. Recalling that microtubules are highly negatively charged (Chapter 1). Their negative electric fields could have significant impact on voltage-gated sodium channels. Furthermore, AIS has ample space for microtubules to move toward and away from the membrane (Chapter 2). Therefore, the open probability of Nav1.2 or Nav1.6 could depend on the distance between microtubules and AIS membrane. Strikingly, Ankyrin-G is known to anchor microtubules to the AIS membrane. The open probability of sodium channels would decrease when the negatively charged microtubules associate with the AIS membrane via Ankyrin-G. The bundling of microtubules into fascicles is likely to enhance the inhibitory effect.

Effects of Microtubules on Channel Gating

Fundamentally, the open probability of voltage-gated sodium channels is determined by electric fields acting on their voltage sensor (the S4 segment). In a nerve membrane, the electric fields may come from various sources, including ions in the intracellular and extracellular solutions, charged molecules on membrane surface and negatively charged microtubules near the membrane. The effects of membrane surface charges on channel gating have been reported (Cukierman et al., 1988; Ednie et al., 2015), but the contribution from microtubules was largely ignored.

In the resting state, a few microtubules (bundled into fascicles) are presumably associated with AIS membrane via Ankyrin-G. The membrane thickness is about 7 nm while the thickness of the submembrane coat varies in the range 3–11 nm (Jones et al., 2014). The N-terminal side of Ankyrin-G is associated with the submembrane coat, while its C-terminal side extends away from the plasma membrane, ~ 35 nm deeper in the inner AIS shaft where it may bind microtubules (Leterrier et al., 2015). Therefore, when the microtubule associates with the AIS membrane via Ankyrin-G, its distance with the voltage sensor of sodium channels is about 40 nm. As calculated below, translocation of microtubules within AIS can have significant influence on channel gating.


Figure 2. The electric field in the membrane produced by a tubulin dimer. At the distance of 40 nm, the tubulin dimer produces an electric field, E1 ~ 107 V/m. When the tubulin dimer translocates to a distance of 120 nm, the produced field in the membrane (E2) decreases to 106 V/m.

A microtubule is made up of tubulin dimers. The electric field in the membrane produced by a tubulin dimer is given by

Et = kQ/rm2

where k is the Coulomb's constant, Q represents the effective charge on a tubulin dimer and rm denotes the distance between a tubulin dimer and the middle of the membrane. Assuming Q = 12 e (Chapter 1) and rm = 40 nm (when the microtubule is anchored to the membrane by Ankyrin-G), we have Et ~ 107 N/C = 107 V/m. As the tubulin dimer translocates to a distance of 120 nm, Et will decrease to 106 V/m.

On the other hand, the membrane potential field, Em, can be obtained from the following formula,

Em = Vm / d

where Vm is the membrane voltage and d is the membrane thickness (~ 7 nm). At the resting membrane voltage (~ 70 mV), Em ~ 107 V/m. Therefore, the electric field produced by microtubules can be as strong as membrane potential fields. Translocation of microtubules within AIS should have significant impact on voltage-gated sodium channels. Generally speaking, microtubule movement toward the membrane is equivalent to membrane hyperpolarization while away from the membrane has the same effects as depolarization. A microtubule fascicle comprises many tubulin dimers, most AIS sodium channels would be closed when a fascicle associates with the membrane. As the fascicle moves away from the membrane longer than 200 nm, it would have little influence on channel gating.

Signal Amplification by Microtubules

In pyramidal neurons, activation of AIS sodium channels requires less depolarization (~ 7 mV ) than somatic sodium channels (discussed above). The smaller depolarizing field may not activate channels by direct interaction with voltage sensors, but by breaking the binding between Ankyrin-G and microtubule fascicle, thereby allowing the fascicle to move away from the membrane. Therefore, the highly negatively charged microtubule fascicle can be used to amplify the effects of depolarizing fields on channel gating, provided that the subthreshold depolarization is sufficient to disrupt the binding between Ankyrin-G and fascicle. In Figure 1, the sharp rise of the Na+ influx probably reflects the moment when their binding is broken, thus allowing microtubules to dissociate from the AIS membrane and consequently activating sodium channels in AIS.

In retinal ganglion cells, the sodium channel density at AIS is indeed very high (Wollner and Catterall, 1986). In spinal cord motor neurons, the density of sodium channels at AIS was estimated to be about 7.4-fold greater than at the soma (Catterall, 1981). In these cases, high density of sodium channels could be a determining factor for the lowest threshold at AIS. Microtubules may not be required to amply the effects of electric fields in these neurons. Theoretically, if the binding between Ankyrin-G and microtubule fascicle is so strong that the subthreshold depolarization cannot dissociate microtubules from the membrane, then microtubules would not be able to dynamically modulate channel gating. This mechanism predicts that Ankyrin-G should also play an important role in channel gating, which has been demonstrated experimentally (see this article).

What kind of neurons require microtubules to amplify the effects of electric fields? The answer could be: the neurons participating in long-range synchronization of neural oscillations that give rise to brain waves. This is because the electric fields coming from distant brain areas are very weak. They cannot influence neuronal excitability by direct interaction with the voltage sensor of sodium channels. Pyramidal neurons, but not retinal ganglion cells or motor neurons, are the major type of neurons involved in long-range communication within the brain. Consistently, a fascicle in pyramidal neurons may contain up to 22 microtubules while in the motor neurons of the spinal cord, it contains only 3 - 5 microtubules (Chapter 2). The higher number of microtubules per fascicle in pyramidal neurons may increase the sensitivity of channel gating to fascicle location within AIS.

The next chapter will describe the electric fields at the extracellular space. The extracellular electric fields, together with the microtubules at AIS, could play a central role in the generation of brain waves.


Author: Frank Lee
First published: August, 2018
Last updated: July 8, 2019