The importance of charge compensation in the membrane

In this post I continue my series on the omega current by discussing how a mutation that removes a charged group from the voltage-sensor domain (VSD) would be highly destabilizing and disruptive. The omega current is a leak current that passes through the VSD of mutated voltage-gated cation channels. In some cases, the mutated channels form a large pore through the VSD (the so-called omega pore) that passes ions as large as guanidinium.  The central question surrounding the formation of an omega pore through the VSD is whether the mutations reveal a cryptic gating-charge pore through the VSD (I will call this the cryptic pore hypothesis), or whether the mutations simply disrupt the native structure of the VSD resulting in ionic leak (I will call this the disruption hypothesis). As I discussed in my previous posts, these two hypotheses for omega pore formation lead to very different conclusions concerning our understanding of the mechanism of voltage gating. Evidence from naturally-occurring omega-current-causing mutations suggested that in certain cases of mutation to tryptophan the disruption hypothesis is most likely. This leads to the questions: is it possible “that all of the other mutations also disrupt the structure? Even the mutations to smaller side chains? And if so why?” In this post I will argue that it is plausible that even the omega-current-causing mutations of the S4 arginines to smaller side chains significantly disrupt the structure of the VSD and that this limits the amount we can learn about the mechanism of voltage sensing from the cryptic pore hypothesis.


The permittivity of the lipid bilayer is low.

Before I discuss the omega-current-causing mutations, I need to introduce the concept of the electrical permittivity of a medium. This is a complex, well-characterized physical property that is important in electronics. It is also plays a very important role in membrane biophysics. I will not discuss the mathematics that underly the behavior of permittivity or the related electric susceptibility of a medium, but I will try to describe an intuitive way in which to think about it.

It is commonly said that the membrane is a “low dielectric medium” or that the “dielectric constant of the membrane is small.” What this means is that the relative permittivity of the membrane is low, especially when compared to the aqueous cytoplasm and extracellular medium. The permittivity of a medium is the degree to which the medium permits the existence of an electric field. In other words, it is the degree to which media can stabilize the presence of charges. Every ion generates an electric field that influences the molecules around it. If these molecules are polar or polarizable, i.e. they contain or can easily generate an internal electrical dipole, they can rearrange themselves in such a way that stabilizes this electric field. Water, for example, is made up of polar H2O molecules that readily reorganize themselves around ions, stabilizing their charge. Essentially, water has high permittivity because the molecules can rearrange themselves in an electric field, lining up their dipoles to generate their own compensating electric field. The carbon chains of lipid fatty acids are not polar and are not significantly polarizable. Hence, the lipid membrane is really bad at stabilizing charges.

The ability of a medium to stabilize charge has real energetic consequences when moving ions between different media. To move an ion from the high permittivity medium of water into the low permittivity medium of the membrane requires a substantial amount of energy – this is why ions do not spontaneously cross the membrane. Thermal noise, even at the high temperatures experienced by hyperthermophiles, does not provide enough energy to push an ion spontaneously through the membrane. This energetic barrier is why cells evolved ion channels in the first place. Ion channels are a type of enzyme that catalyze the conduction of ions across the low permittivity membrane by providing an environment that compensates for the charge of the ion.

Although proteins themselves are polar molecules (the peptide bonds that link the amino acids along the protein main chain have a dipole moment), they are able to stably insert into the membrane because of the different properties of the amino acid side chains.  Many amino acid side chains are hydrophobic, i.e. it is more energetically favorable for them to be buried (these are the non-polar, non-charged side chains). Conversely, other side chains are hydrophilic, i.e. it is more energetically favorable for them to be exposed to water (these are the polar or charged side chains). For soluble proteins, the hydrophilic residues are generally on the outside whereas the hydrophobic residues make up the core. In membrane proteins the opposite is true, the hydrophobic residues are on the outside interacting with the low permittivity membrane while the hydrophilic residues are turned inwards towards each other where they can mediate specific interactions. Exposing a polar or charged side chain in the middle of the membrane would greatly stress the protein structure. This is, of course, different on the surface of the membrane where the animo acid residues can interact with water or the charged phosphate head groups of the lipids.

Because of the large energy cost of burying a charge in the low permittivity medium, membrane protein structures have evolved to minimize the number of charged amino acid side chains within their transmembrane segments. This tendency is so strong that it is easy to design algorithms to predict the presence of transmembrane segments simply from the relative hydrophobicity of the corresponding sequences. If – such as in the case of a voltage-sensor – the presence of a charged residue in the membrane is mechanistically required, then compensation for the charge must be provided by the protein structure. Usually, this compensation is provided by the presence of oppositely charged residues.

This charge compensation can readily be seen in the structures of VSDs.

Charge Compensation in the VSD

voltage-sensor domain charge compensation

Figure 1. Structure of the Kv1.2-2.1 paddle chimera VSD (PDB accession code 2R9R) showing the side chains of charged amino acid residues located within the membrane boundaries. The black lines delineate the approximate hydrophobic thickness of the bilayer. Charge compensation sites within the VSD are circled in pink. Side chain are colored by atom with carbon grey, oxygen red and nitrogen blue, except for the Phe233 side chain of the phenylalanine gap which is colored blue.

As you can see in Fig.1, all of the positively charged S4 gating charges that are buried within the hydrophobic membrane are paired with a negatively charged residue from one of the other transmembrane helices (Glu183 on S1, Glu226 and Glu236 on S2 and Asp259 on S3a). This pairing stabilizes the charges in the membrane, hence, stabilizing the structure of the VSD itself.

The internal charge-compensation site occupied by K5 in Fig.1 is known as the charge transfer center (Tao et al., 2010). This site is different from the external charge-compensation sites by the presence of two negatively charged side chains and the proximity to the phenylalanine gap (represented by the blue side chain of Phe233 in Fig.1). The phenylalanine gap forms a hydrophobic barrier that separates the external and internal charge compensation clusters and is considered the most constrictive point within the VSD structure.

The structure shown in Fig.1 is the depolarized conformation of the Kv1.2-2.1 paddle chimera VSD (Long et al., 2007). It is thought that, during gating, the S4 gating charge residues that occupy each of the charge compensation sites changes. For example, in the hyperpolarized state immediately preceding the depolarized state, R2 would be in the external charge compensation site occupied by R3 in Fig.1, R3 would be in the site occupied by R4 and R4 will be in the internal charge transfer center occupied by K5. This coordinated movement between the sites would allow for sustained charge compensation throughout the gating conformational changes of the VSD. In more hyperpolarized conformations, it is thought that R2 or even R1 (in the potential deeply hyperpolarized conformation, see my previous post) would occupy the charge transfer center. This would leave the external charge compensation sites unoccupied by positive charges and suggests that they would have to be solvent-exposed to prevent the large energy costs of unpaired charges in the low permittivity membrane.

In a recent paper by Hoshi & Armstrong, it was shown that the extracellular charge compensation sites are likely unoccupied and solvent-exposed in a hyperpolarize conformation of the VSD (Hoshi & Armstrong, 2012). By adding the strong acid-binding tri-valent lanthanum ion (La3+) to the extracellular solution they were able to significantly slow down the opening of the Shaker Kv channel. Indeed, they even observed a slowing of gating charge movement in the presence of extracellular La3+, indicating an inhibitory effect of the ion on S4 movement (Hoshi & Armstrong, 2012). This led them to propose that the La3+ ions bind to the solvent exposed extracellular charge-compensation sites and compete with the S4 arginines for occupancy. This slows the movement of the S4 gating charge residues into these sites and hence, gating of the channels.

I would speculate that, due to the reasonable hyperpolarization of -80 mV applied by Hoshi & Armstrong in this study, the majority of the VSDs would be in a hyperpolarized conformation corresponding to having R2 in the charge transfer center and not the deeply hyperpolarized conformation with R1 in the charged transfer center (see my previous post for more discussion on VSD hyperpolarized conformations). Tombola et al. showed that, upon stronger hyperpolarizations (>-100 mV), there is additional movement of S4 gating charge, which may correspond to the movement of  R1 into the charge transfer center (Tombola et al., 2005). Moreover, recent work by Meng-chin et al. have shown that this deeply hyperpolarized conformation can be occupied at -80 mV in mutant VSDs in the presence of Zn2+ ions (Meng-chin et al., 2011). With R2 in the charge transfer center, only the outermost external charge compensation site would be unoccupied. Hence, the entire effect of the La3+ on gating may only be due to binding to the outermost acid (Glu183 on S1 in the Kv1.2-2.1 paddle chimera). The reason that it requires such strong hyperpolarizations to move R1 into the charge transfer site may be because of the energetic costs in removing the charge compensation from the lower external site. Of course, this is only speculation, and it has yet to be shown that the additional gating charge movement seen at strong hyperpolarizations corresponds to movement of R1 into the charge transfer center and not to some other conformational rearrangement within the VSD.

Removal of a single charge from the VSD could have serious consequences

Because of the low electrical permittivity of the membrane, removal of a single membrane-buried charged group would have large energetic cost that could significantly alter the conformational stability of different states of the VSD. Given that, as the structure clearly shows, the VSD has evolved to contain specifically matched charge-compensation sites, removal of one of the charges would leave the alternate charge in the site unpaired. In order to compensate for the large energetic cost of an unpaired charge, the structure of the VSD would likely distort. One of the most readily available mechanisms to compensate for this charge would be to adopt a conformation that allowed solvent accessibility into the VSD. As discussed above, polar water molecules are able to stabilize charges, and a distortion of the VSD structure that allowed water in would be energetically favorable. In certain conformations, this may even lead to a pore forming through the VSD bridging the extra- and intracellular solutions.

Another possible mechanism to counter this missing charge would be for the VSD to adopt a conformation that allowed soluble, oppositely charged ions into the disrupted charge-compensation site. For example, in the case of a mutation of one of the S4 positive charges to a neutral amino acid, the binding of positively charged ions within the VSD would be favorable to compensate for the missing positive charge. This necessity for charge compensation would mean that, in the case of a pore being formed through the VSD, there would be selectivity for ions matching the charge of the missing side chain. If you remove a positive charge, you generate a cation selective pore due to the excess negative charge within the VSD. The opposite is true if you remove a negative charge from the VSD (see Musset et al., 2011).


I hope my reasoning as to how a mutation of an arginine to a smaller uncharged amino acid could lead to significant disruption of the VSD structure is now clear. To recap:

  • Because of the inherent instability of charges within the low permittivity membrane, the VSD has evolved a structure that contains specific charge-compensation sites to stabilize the gating charges within the membrane.
  • Disruption of one of these charge-compensation sites by removal of one of the interacting charged residues would result in significant energetic penalties that would likely result in deformation of the VSD structure.
  • Likely deformations of the structure would allow water or ions into the VSD in an attempt to stabilize the unpaired charge.
  • Depending on the conformation of the VSD and on the identity of the mutated residue, this disruption of the structure could result in the formation of pores through the VSD.

This argument supports the disruption hypothesis of the omega pore. If true, it would indicate that the leak current observed is not due to the exposure of a native cryptic-gating-charge pore, but instead due to a non-native distortion of the VSD structure.

The VSD has evolved to balance competing forces. It requires charges in order to sense the membrane potential, but these charges are inherently unstable in the low permittivity membrane where they are needed. Evolution has found a structural solution to this problem in the four-helix bundle of the VSD. This solution is both elegant and delicate. Nature has used this four-helix bundle to regulate many different types of channels and enzymes. Additionally, through mechanisms we don’t yet fully understand, the voltage sensitivity of VSDs can be tweaked to respond to membrane potentials across the physiological range. However, if you try to understand this elegant mechanism by standard mutagenesis studies, you are more likely to disrupt the structure than to reveal its secrets. This frustration is well known to all who study these domains, and it forces us to be extra careful when interpreting data from mutated VSDs.

Works Cited and Further Reading

Hoshi, T., & Armstrong, C. M. (2012). Initial steps in the opening of a Shaker potassium channel. Proceedings of the National Academy of Sciences of the United States of America, 109(31), 12800–12804. doi:10.1073/pnas.1209665109

Lin, M.-C. A., Hsieh, J.-Y., Mock, A. F., & Papazian, D. M. (2011). R1 in the Shaker S4 occupies the gating charge transfer center in the resting state. The Journal of General Physiology, 138(2), 155–163. doi:10.1085/jgp.201110642

Long, S. B., Tao, X., Campbell, E. B., & Mackinnon, R. (2007). Atomic structure of a voltage-dependent K+ channel in a lipid membrane-like environment. Nature, 450(7168), 376–382. doi:10.1038/nature06265

Musset, B., Smith, S. M. E., Rajan, S., Morgan, D., Cherny, V. V., & DeCoursey, T. E. (2011). Aspartate 112 is the selectivity filter of the human voltage-gated proton channel. Nature, 1–6. doi:10.1038/nature10557

Tao, X., Lee, A., Limapichat, W., Dougherty, D. A., & Mackinnon, R. (2010). A gating charge transfer center in voltage sensors. Science, 328(5974), 67–73. doi:10.1126/science.1185954

Tombola, F., Pathak, M. M., & Isacoff, E. Y. (2005). Voltage-Sensing Arginines in a Potassium Channel Permeate and Occlude Cation-Selective Pores. Neuron, 45(3), 379–388. doi:10.1016/j.neuron.2004.12.047

Back to top button