Quantum Calculation of Proton and Other Charge Transfer Steps in

Subscriber access provided by Gothenburg University Library

B: Biophysics; Physical Chemistry of Biological Systems and Biomolecules

Quantum Calculation of Proton and Other Charge Transfer Steps in Voltage Sensing in the K1.2 Channel v

Alisher M Kariev, and Michael E. Green J. Phys. Chem. B, Just Accepted Manuscript • DOI: 10.1021/acs.jpcb.9b05448 • Publication Date (Web): 23 Aug 2019 Downloaded from pubs.acs.org on August 26, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Quantum Calculation of Proton and Other Charge Transfer Steps in Voltage Sensing in the Kv1.2 Channel Alisher M. Kariev and Michael E. Green* Department of Chemistry and Biochemistry City College of New York New York, NY 10011 United States of America

*Corresponding Author: [email protected]; Phone: 1-212-650-6034

Alisher M Kariev: [email protected]

1 ACS Paragon Plus Environment

The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ABSTRACT: Ion channels in cell membranes control entry and exit of ions; their gating (opening and closing) is key to their functioning. It is known that protons can pass through the voltage sensing domain (VSD) of channels such as Kv1.2; quantum calculations for a section of the VSD show the steps protons take in responding to voltage, and show no major displacement of the protein backbone with voltage change; 70 amino acids are included, 42 with side chains (9 directly in the proton path), 28 as backbone only, and 24 water molecules. Protons provide much of the gating current, the capacitative current immediately preceding channel opening, with significant additional contributions from charge transfer to other groups. Most gating models, in contrast, require major protein displacement during gating. Energy terms without classical analogues (exchange plus correlation energy, which are greater than thermal energy) show quantum calculations are required. Energy as a function of voltage for a key proton transfer leads to approximately the correct voltage for channel opening. Calculated total charge transfer (not only protons) for gating is reasonable compared to experimental values. We are also able to account, at least qualitatively, for two mutations, one with the gating current curve left shifted, one right shifted, and show the alternate proton paths that are required to account for these.


Page 2 of 32

Page 3 of 32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


INTRODUCTION: Ion channels are an extremely important class of proteins; channels of various types are present in all cells, allowing the transport of Na+, K+, and Ca2+ through the cell membrane1; other channels are responsible for transport of other substances, in particular protons2,3. Channels have even been found in viruses4,5. Voltage gated ion channels (VGIC) are a particular class of channels, and certain types of these are responsible for the nerve impulse, as well as the properties of other excitable tissue6-8. VGIC, both for Na+ and K+, have four voltage sensing domains (VSD); for the K+ channels, these are identical, while for Na+ they are similar (one having a different function). Here we will be concerned with the way a K+ channel senses voltage. Each VSD of a Nat or K+ channel has four transmembrane (TM) helical segments, linked by extraor intracellular loops that extend beyond the membrane. The fourth TM helix (labeled S4) is connected via a linker of several amino acids to a domain with 8 TM helices, 2 extending from each VSD, that together form the pore through which the ion passes. Channels of the type we are discussing are closed when there is a voltage of approximately -70 mV, (minus meaning intracellular side negative with respect to the extracellular side), that holds the channel closed; with 0 mV across the channel, it is open. Fig. 1 shows the complete channel. The channel has extensive intracellular sections, and the section closest to the membrane (called T1) is involved with gating9. S4 has arginines in every third position (except for one lysine near the intracellular end). Gating is preceded by a capacitative current (“gating current”), in which positive charges move in an extracellular direction in response to membrane depolarization. Negative charges moving in an intracellular direction could produce a similar current, but, in spite of the fact that negative charges do exist on the S2 and S3 TM segments, this possibility is not considered in the standard models of gating. These models are based on the motion of the S4 TM segment in an extracellular direction upon membrane depolarization, when its positive charges, assumed to be on three of the arginines, no longer hold it down. Standard gating models: There are variations in the details of the motion proposed in different versions of the standard model, but as a class, these models are defined by extracellular motion of S4 upon depolarization. The evidence for this type of model was first based on scanning cysteine accessibility mutagenesis (SCAM11), and has since been supplemented, primarily by molecular dynamics (MD)12-16 and FRET (fluorescence quenching)17,18 We have discussed this evidence in previous 19-22 , and concluded that it does not rule out an alternative source of gating current specifically, proton transport, with S4 remaining stationary. The SCAM method requires the mutation of an amino acid to cysteine. In this case, it was the arginines on S4 that were mutated first. Arginine has a very large side chain, cysteine a very short side chain. The large difference in the molecular volume of these side chains is critical. In the SCAM method, it is assumed that cysteine must move to the membrane surface (either intracellularly or extracellularly) in order to ionize, because the membrane dielectric Fig.1: The complete Kv1.2 channel, reproduced from the (rcsb) pdb 10

structure 3Lut ; the red and blue planes mark the membrane boundaries, blue intracellular. The entire blue section below the blue membrane boundary is intracellular; the red helices constitute the VSD and pore. There are four voltage sensing domains, each with four transmembrane segments, between these boundaries; the pore is hidden behind the front VSD. Extracellular loops connecting alternating transmembrane segments are above the red extracellular membrane boundary. The T1 moiety (arrow) constitutes the upper 9

part of the intracellular section, and is involved in gating



constant is too low to allow ionization; only the ionized form can react with a methanethiosulfonate (MTS) reagent; the test for whether the cysteine had moved to the membrane surface really indicated that the S4 had moved enough as a fairly rigid body to make the cysteine accessible, then one would have evidence for the standard model. However, if the cysteine could react in situ, without moving, then the conclusion would have to be that the ability of the MTS reagent to penetrate the membrane had changed, not that the S4 had moved. For the cysteine to ionize, at least two or three water molecules are needed near the cysteine; the presence of water molecules would make the low dielectric assumption no longer valid. The difference in side chain volume is in fact adequate for two or three water molecules. Once ionization occurs, the reactive head of an MTS reagent could replace the water, so that it could react with the ionized cysteine. Therefore, the interpretation of the SCAM data can be reconsidered. FRET data sometimes show appreciable motion. While we cannot discuss all the data that have been adduced to support the standard model, we note that FRET data showing small S4 motions have also been reported18, 23-25. FRET data may not have high resolution, as the probe is typically approximately 10 Å, and the orientation of the probe is important and hard to determine. Finally, most of the computational literature on the VGIC uses molecular dynamics (MD), which is strictly classical. We will discuss below reasons for doubting that MD simulations of voltage gating can properly describe this phenomenon. Proton transport in similar systems: We have proposed proton transport as the main source of gating current instead of S4 motion, in Kv channels, especially for channels in the Shaker and Kv1 families, not only in the reviews just cited, but in earlier work as well26-28. Gating current includes charge transfer on heavy atoms (oxygen and nitrogen), in addition to the proton transfers. These must be added to get the complete gating current. Charge transfer from water to ions is known to be non-negligible, and to require quantum calculations to determine it29, 30. Proton transport is known in a number of systems that are analogous to the Kv voltage gated potassium family of channels. The discovery that the Hv1 proton channel is very similar to the VSD31-33 helps show that proton transport could exist in the VSD. Other channels with similarities to the Kv channels also show proton transport through membranes: Woelke et al give MD calculations on cytochrome c (cyt c), in which the analogy to the Kv channel is discussed explicitly34. Cyt c shows a very similar sequence to what is present in the Kv VSD35, 36; this paper shows a tyrosine ionization essentially similar to what we report here. Additional analogous systems, not all protein, include: the M2 channel in the influenza virus 37-40 ; a guanidinium-ammonium competition with a crown ether 4 1 ; an amino acid-polyether/polyamine macrocycle 4 2 ; bacteriorhodopsin (a case involving tyrosine)43; and arginine-carboxylate interactions44. Han et al have shown how protonation states control function in a Na+,K+,ATPase45, in this case relevant to Na+ binding. Mutated VSD of a Shaker potassium channel46-48, which is essentially the same channel as we are considering, also transmits protons. The key for this case is an RH mutation at the end of the path, while the section we calculate here is part of the same path the H+ is expected to follow in that case. In WT, the transfer to the extracellular solution is presumably blocked by arginine. Thus, our calculated proton path is consistent with previous experimental results. A proton current through the VSD has also been measured by Zhao and Blunck49, who showed that the VSD forms a cation channel, with a preference for protons. This appears distinct from the “omega current” 50, 51 Motion of the S4 segment would seemingly produce approximately the correct gating current (with Kv1.2, about 10 charges move across the electric field52, compared to approximately 13 for Shaker); many standard models assume that there were full +1 charges on the arginines. However, given the distribution of charge, this would not be the case. As we show below, charge on the arginines is 50 kJ). Experimentally, the Q –V curve is left shifted by over 40 mV; this suggests that the energy change is not unreasonable. All the structural details of this calculation are shown in the figure, so we do not go through them here; however, the important point is the direction of change of the energy, and the fact that an alternate path exists. Again, we can only claim qualitative agreement, as there is not enough information to allow a quantitative comparison. Summary of mutations: Y266F: We should expect that if the proton cannot transfer from the Y266, it should make gating more difficult, and it does; experimentally the Y266F mutation leaves a still functional channel, but right shifted by approximately 8.5 mV for the gating current curve. Fig. 5 showed the missing –OH and the possible alternate paths through which the channel could still function. The discussion that accompanied Fig. 5 showed how plausible alternate proton paths could be found. R303cit: The second mutation, R303cit, in which a putative (in standard models) gating charge is replaced by the isosteric, but uncharged, citrulline side chain, showed a left shift75. The energy for the proton transfer dropped in the mutant by approximately 10 kJ, equivalent to 4kBT, so that the calculation is consistent with a left shift, as in the experiment; the >50 kJ drop is the total for 0 mV, while the 10 kJ refers to the change in the barrier, that is, the energy at which the two curves in Fig. 7 would cross. in this case, the magnitude is quite reasonable. In the wild type, the energy at this point rises, so the mutant favors the open state. Fig, 6 shows the mutant, with the relevant shifts in position of the aromatic rings marked. For lack of computer resources, we did not calculate the R297cit mutation, nor the R300cit mutation. Experimentally, R297cit kills the channel; from Fig, 4 it is fairly easy to see why this should be, as both paths for the proton must go through that arginine, and the citrulline does not have the requisite NH2 group to transfer a proton; unlike arginine it is not amphoteric. In addition, that is the location where the largest part of the electric field drops64. The R300cit mutation right shifts the Q-V curve. Given that R300 is part of the Y266,E183,R300 triad, it is not surprising that R300 behaves like Y266, in that the R300cit mutation removes the same path. Thus, removing that path right shifts the curve, similarly to removing Y266; experimentally, it is somewhat larger, but not so much so as to make the interpretation implausible75 Bond orders of hydrogen bonds: The bond orders of certain hydrogen bonds differ in open and closed states. Bond orders were calculated using NBO61,62, and the results for some hydrogen bonds in WT are shown in Table 2.



Page 14 of 32

Table 2 Bond Orders of Certain Hydrogen Bonds (WT)**

H bond* R300 E183 R303 E226 Y266 E226 Y266 E266 Y266 R303

H shift No H (right) shift V(mV) -70 mV H1-O1 0.16 H2-O1 0.24 H1-O1 -H2-O2 O-H-O 0.31

No H shift 0 mV

R300 E183 -70 mV

R300 E183 0 mV

Y266 E183 -70 mV

Y266 E183 0 mV

Y266F Y266F mutation mutation -70 mV 0 mV

0.16 0.24 --

0.13 0.09 -0.22

0.13 0.12 0.22 0.30 --

0.16 0.10 0.22 0.29 --

0.16 0.21 --

0.31

0.10 0.11 0.24 0.31 --

--

0.16 0.23 0.19 0.31 --

ONE(H) O-NH

--

--

--

-0.31+

-0.30+

--

--

--

--

--

--

--

0.12

0.11

--

--

*first column shows the two amino acids involved, second the specific atoms

+: Note that these are fairly strongly anti-bonding: there is at least one node in the electron density between the two possible positions of the hydrogen atom **There is also a strong H-bond (0.40 bond order) between the NE of R300 and the –OH of Y266 in one case only: (-70 mV, R300 proton transferred to E183) --Blanks: there is no H-bond Not surprisingly there is a relation between bond order and bond length. The small side chain displacements strongly affect the bond strength, since 0.2 Å is enough to change the bond strength appreciably, especially in H-bonds. Where there is a blank in Table 2, there is no interaction, generally because the side chain has rotated in such a way that the relevant heavy atoms are either too far apart to form an H-bond, or oriented so as to make the bond impossible by reason of intervening atoms—examination of Fig. 3 through Fig. 6 also shows this. The strongest H-bond bond order, 0.40, almost covalent, (see footnote, Table 2), is between the NE of R300 and the –OH of Y266, when the proton has shifted from R300 to E183, neutralizing the salt bridge; this practically locks the structure in the closed state, and has the lowest energy (Fig. 7), about 20 kBT (≈50 kJ) below the crossing point of the two positions, within the error of the calculation and the experimental error of what is found for the activation energy of gating from the temperature dependence of the current. There may be a complication, in that the steps from the calculated section to the gate, not represented in the calculations done so far, might also be responsible for some limited temperature dependence76, 77. It would be surprising if this voltage independent section contributed appreciably. Removing the voltage allows the proton to transfer so that the R300-Y266 H-bond no longer exists, and the side chains can change their local structure. Second, there are hydrogen bonds with bond order near 0.3 between E226 and Y266, and E226 and R303. There is a net of one extra bond with V= -70 mV. It appears that there are no other particularly remarkable H-bonds in this region. However, there are side chain rotations. Since the closed state has lower energy than the open state (with the aid of the voltage) it appears that these strong H-bonds are in the range that could contribute to the lower energy of the closed state. The 14 ACS Paragon Plus Environment

Page 15 of 32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


complication of the system, with multiple possibilities, suggests great caution in interpreting temperature dependence as a consequence of individual steps. It is possible for an individual step to be calculated to require so large an energy step as to be incompatible with observation, but our result is not incompatible with the observed temperature dependence. The exchange and correlation energies were also calculated; in general, their differences with voltage, as well as proton position, are on the order of tens of kJ (see Table 3), so that they cannot be neglected.

Table 3 Exchange and correlation energies (kJ/mole)+ for the same six cases as Table 1 Nominal Y266 charge and energy 

R300

E183

Exchange energy

Correlation energy

Sum

-70 mV

0

+

-

0

2.5

2.5

0

0

+

-

0

0

0

-70

0

0

0

75.8

20.2

96.0

0C

0

0

0

37.9

22.7

60.6

-70

-

+

0

65.7

37.9

103.6

0

-

+

0

12.6

12.6

25.2

+ Absolute value of energy of the lowest energy state (2nd row) is subtracted from each term. The energies shown are the energy differences between this state and the energy of the configuration in that row. 0,+,in columns 2-4 are the nominal charge states of the amino acid shown for that column. To get absolute values of energy, add 6635.04 kJ to exchange energy, 430.78 kJ to correlation energy for all cases.

Translate the exchange + correlation energies into units of kBT for T=300 K, to understand the physiological significance. For the (0,0,0) cases that total 24 kBT for 0 field, 38 kBT with field, there is a difference of 14 kBT; for the (-,+,0) cases, the energy is 10 kBT (0 field) compared to 41.5 kBT (with field), a difference of approximately 31 kBT. In a classical calculation, without the exchange and correlation terms, the wrong ionization states would be assigned. The difference of 17 kJ, or about 7 kBT, makes a considerable difference in the voltage at which the crossing of the two states occurs. These “quantum terms”, (i.e., those with no classical analogue; while the values of electrostatic etc. energy also are quantum terms in this calculation, classical values exist for these) also bring the “normal” ionization states somewhat closer to the others, although the sum of potential plus kinetic terms still leaves them slightly higher in energy. The nominal charges in Table 3 show the location of the proton, while the actual charges, from NBO calculation, are in Table 4.



Page 16 of 32

Table 4a Charges on Y266 and glutamates* as determined by NBO calculations 61, 62

Field, R300 charge**

V=0, R300(+)

V=0, R300(0)

V

Quantum Calculation of Proton and Other Charge Transfer Steps in

Recommend Documents