A Mechanistic Reinterpretation of Fast Inactivation in Voltage-Gated Na+ Channels

The hinged-lid model is long accepted as the canonical model for fast inactivation in Nav channels. It predicts that the hydrophobic IFM motif acts intracellularly as the gating particle that binds and occludes the pore during fast inactivation. However, the observation in recent high-resolution structures that the bound IFM motif locates far from the pore, contradicts this preconception. Here, we provide a mechanistic reinterpretation of fast inactivation based on structural analysis and ionic/gating current measurements. We demonstrate that in Nav1.4 the final inactivation gate is comprised of two hydrophobic rings at the bottom of S6 helices. These rings function in series and close downstream of IFM binding. Reducing the volume of the sidechain in both rings leads to a partially conductive “leaky” inactivated state and decreases the selectivity for Na+ ion. Altogether, we present an alternative molecular framework to describe fast inactivation.

conductive "leaky" inactivated state and decreases the selectivity for Na + ion. Altogether, 22 we present an alternative molecular framework to describe fast inactivation. 23

Introduction 24
Voltage-gated sodium (Nav) channels function as the main drivers of the depolarization 25 phase of action potentials. As such, they are ubiquitously found in a wide range of 26 excitable tissues, playing fundamental roles in many vital physiological functions, such as 27 skeletal muscle contraction, heart beating, and neuronal impulse generation and 28 propagation (1). Consisting of one long protein chain, mammalian Nav channels are 29 arranged into four similar, yet not identical domains (DI to DIV). Each domain, formed by 30 6 transmembrane helixes (S1 to S6), is then organized pseudo-symmetrically into the 31 voltage-sensing subdomain (VSD, S1 to S4) and the pore-forming subdomain (PD, S5 to 32 S6) (2). Milliseconds after activation, Nav channels rapidly enter into a non-conductive 33 state, the fast inactivated state, and only recover upon repolarization (3-6). Fast 34 inactivation serves as a negative feedback switch, responsible for the attenuation of 35 inward current during action potential. Subtle abnormalities in fast inactivation can lead 36 to severe pathological states (7-9). These diseases, though sharing similar etiology, can 37 manifest in many different types of tissues and show a wide range of symptoms due to 38 the prevalence and diversity of Nav channels (10). For instance, skeletal muscle 39 hyperkalemic periodic paralysis (11, 12), long QT syndrome in the heart (13) and 40 inherited epilepsy in brain (14) can all arise from abnormal Nav fast inactivation. 41 While examining the ionic and gating current from Nav channels in the giant squid axon, 42 Armstrong and Bezanilla formulated the "ball and chain" model (15,16), where an 43 positively charge "inactivation ball", tethered by a flexible "chain" to the intracellular gating current measurements, we demonstrate and characterize a S6-located double ring 68 of hydrophobic residues in DIII and DIV. Through site directed mutagenesis, we show 69 that the fast inactivation gate is formed by two layers of bulky, hydrophobic residues, 70 whose effectiveness as inactivation gate is dependent on side chain volume. Once 71 reduced, smaller side chain substitutions lead to a leaky inactivated state. Equivalent 72 residues in DII, on the other hand, seem to be involved in the coupling between VSD 73 activation and PD opening instead. Surprisingly, our experiments point to a previously 74 overlooked coupling pathway between the bottom of S6 and the selectivity filter (SF). 75 Altogether, our results identify a new candidate and mechanism for the fast inactivation 76 gate in Nav channels, highlighting importance of the intracellular (C-terminal) end of S6 77 for both activation and fast inactivation. and NavPas. The residues identified to form the narrowest part of the channel were highlighted in cyan for 91 NavPas and purple for Nav1.7, and they are colored in red for rNav1.4.

93
Starting with the observation that the IFM motif is located far from the pore in the putative 94 inactivated state (Fig. 1B) (28, 30, 32), we reasoned that large, hydrophobic residues 95 residing at the intracellular end of S6 might play a dual role participating both in activation 96 and fast inactivation gating. Close examination of the electron density maps in the 97 available Nav channel structures, shows that in most cases, non-protein densities can be 98 spotted inside the pore, spanning across the bundle crossing region into the inner cavity 99 (22, 32, 33). These molecules are sometimes unidentified, however in many cases 100 assigned to be detergent molecules, clouding any interpretation regarding their functional 101 state. Here we have focused our analyses on two structures that do not present such 102 non-protein densities in the pore: NavPas (PDB:6A95 (30), Fig. 1A) and Nav1.7 M11 103 (PDB:7XVF (28), referred to as Nav1.7 from now on, Fig. 1B). Given the fact that both 104 structures are determined at 0mV and the DIV VSD is in the "up" conformation, it is 105 reasonable to assume that the pore resemble a fast inactivated state. We calculated the 106 pore profile in NavPas and Nav1.7 (HOLE) as the radius along the permeation pathway 107 (graphs in Fig.1 A, B). This profile led to easy identification of the narrowest part of the 108 pore in each case. While NavPas and Nav1.7 showed very similar pore profiles at the intracellular end, the stretch along the pore that form the narrowest part, was much larger 110 in NavPas and Nav1.7 than structures with non-proteinaceous densities inside the pore, 111 for instance in the case of Nav1.5 structure ( Supplementary Fig. 1). The narrowest region 112 in NavPas and Nav1.7 was composed by two layers of hydrophobic residues organized 113 in a diamond shape facing towards the pore at the bottom of S6, seen as two minima in 114 the radius profiles (Fig. 1A, B), located one  helix turn away from each other. An early 115 indication of the feasibility of our hypothesis is related to the fact that all the residues 116 identified were hydrophobic and 7 of them were also bulky. Additionally, the 8 residues 117 (two from each domain) identified in NavPas and Nav1.7, are highly conserved among 118 different Nav isoforms ( Supplementary Fig. 2). Based on our structural analysis, we 119 hypothesize that these S6 residues form a two-tier hydrophobic barrier in the fast 120 inactivated state and serve as the fast inactivation gate in Nav channels. 121 Double alanine mutation in DIII S6 produces a conductive "leaky" inactivated state 122 After identifying residues that could potentially form the fast inactivation gate at S6, we 123 predicted that by reducing the volume of the involved side chains in BOTH layers 124 simultaneously, we could widen the two-tier barrier in S6 and partially open the fast 125 inactivation gate. We first tested DIII and DIV due to their known role in fast inactivation, 126 and the fact that the IFM motif is spatially closer to these two domains (Fig. 1B) (34-36). 127 In DIII, isoleucine I1284 and I1288 were identified based on the sequence alignment in 128 rNav1.4, our model Nav channel ( Fig. 2A). Single alanine mutations on either position 129 (I1284A or I1288A) yielded currents similar to wild type (WT) channel (Fig. 2B). When 130 both residues were mutated to alanine simultaneously (mutant DIIIAA), significant steady 131 state current was observed (Fig 2B, bottom traces). At +60mV this steady state current  152 153 corresponds to around 20% of the peak current which clearly differs from the negligible 154 amounts (less than 3%) in the WT as well as in the I1284A and I1288A mutants (Fig. 2C).  Table  157 3). The steady state current was present exclusively in the double mutant, which cannot 158 be accounted for by a combination of the effects of the two single mutants since neither 159 of them showed steady state current. But this is expected if the two-tier barrier is acting 160 as a steric hindrance in series as is observed in the pore radius measurement where two 161 layers of hydrophobic barriers blocks the pore at the bottom of S6 and the permeation 162 pathway could only be widened by reducing both residues forming the barrier. 163 To characterize the conformational changes along the fast inactivation pathway, we 164 measured gating charge immobilization in DIIIAA. As Nav channels enter fast inactivation, 165 VSDs in DIII and DIV get trapped in the up conformation and become immobilized. As a result, upon hyperpolarization, VSDs in DIII and DIV move after the channels transitioned 167 out of fast inactivated state, manifesting as a slow component in the off-gating current 168 (16). Therefore, the amount of immobilized charge, and the time course of immobilization 169 provide a direct measurement of conformational changes along the fast inactivation 170 pathway, independent of ionic current. In WT, almost all channels are inactivated 171 milliseconds after depolarization ( Fig. 2A). Consistent with this, a second slow kinetic 172 component develops in the off-gating current milliseconds after depolarization (Fig. 2E, as a second conductive state, a "leaky" inactivated state. 198 The "leaky" inactivated state is less selective for Na + ion.

214
In certain voltage ranges, we noticed that the ionic currents went from inward to outward 215 during the depolarizing pulse in DIIIAA. This was shown as a fast inward and a steady- pulse. Indeed, the reversal potential for the peak current, which is related to the first 220 component, is significantly more positive than the one for the steady state current, related 221 to the second component in bionic environment (Fig. 4 A). When external Na + was 222 increased from 57.5 to 90 mM both components were shifted to more depolarized 223 voltages: from 35 to 46 mV for the peak component and from 18 to 30 mV for the steady 224 state component (Fig 4 A, B). When only K + is used as permeant ion, there is no crossing 225 in the ionic currents and the reversal potential for the peak and steady state are the same 226 ( Fig. 4C). Once K + was exchanged by Na + in the external solution, the crossing reappears, 227 showing that this process is reversible ( Supplementary Fig. 5). Therefore, surprisingly 228 enough, the relative permeability between Na + and K + is changing during the time course 229 of depolarization in DIIIAA.
To further explore this change in selectivity, we measured the relative permeability at 231 different time points during a depolarizing voltage pulse (60 mV). This was done by 232 measuring the instantaneous I-V curve at different times during depolarization, computing 233 the reversal potential as a function of time, and using the GHK equation (Equation 6) as 234 a framework to analyze the data (Fig. 4D, E, F Methods). Under bionic conditions we can 235 calculate the relative permeability between Na + and K + ions (PNa/PK) from the reversal 236 potential measurements. At the beginning of the pulse, the selectivity of DIIIAA towards 237 Na + was high, similar to the WT (PNa/PK ≈ 11) but almost three time less selective later in 238 the pulse (PNa/PK ≈ 4) (Fig. 4G). The fast time course of the permeability changes also 239 followed the time course of the apparent fast inactivation, indicating that the 240 conformational changes associated with fast inactivation might also trigger a change in 241 selectivity (Fig. 4H). These results point to the existence of an allosteric coupling pathway 242 between S6 and the SF. Independent of the underlying coupling mechanism, these 243 results show, unequivocally, that the bulky hydrophobic residues identified form part of 244 the inactivation gate. 245 "Leaky" inactivated state can be accessed from closed state and can be prevented 246 by removing fast inactivation in DIIIAA 247 So far, we find no evidence to suggest that the reduction in the side chain volume in 248 DIIIAA has any effect on the "leakiness" of the closed state at rest. One interesting feature 249 of the present results is that, at hyperpolarized voltages (from -55mV to -30mV), the 250 DIIIAA mutant displays slower ionic currents kinetics but no apparent inactivation, a 251 phenomenon not observed for the single mutants and WT (Fig. 4A). When compared to 252 to the left (calculated at the steady-state), whereas the two single alanine mutants were 254 right shifted (+15.2mV for I1284A; +9.6mV for I1288A) (Fig. 4B, Supplementary Table 2).   framework where we demonstrate that the residues at S6 forms the fast inactivation gate, 283 we interpret the IFM motif, instead of being the final effector, acts more likely as a 284 transducer that couples the VSD movement in DIV, which triggers the fast inactivation, to the pore where the inactivation gate ultimately closes. By mutating F1304 to Q, the 286 energetic barrier becomes so high that effectively, the conformational changes associated 287 with fast inactivation is terminated before reaching the pore. By preventing fast 288 inactivation, we expect to remove the features associated with the leaky inactivated state, 289 namely: the slow component on the tail currents, the change in selectivity and the shift in 290 the G-V curve. In WT Nav, IQM removes most of the fast inactivation without significant 291 shifts in the G-V curve (Fig. 4E, F, Supplementary Table 2). In contrast, in IQM_DIIIAA, 292 the G-V curve was shifted to the right (+9.6 mV) compared to the WT, comparable to the 293 single mutants and the shift in Q-V curve, and we saw no changes in selectivity and tail 294 currents did not change with the depolarization time (Fig. 5 G-I). Since the IQM mutation 295 itself does not change the G-V curve and removes inactivation, our results indicate that: 296 i) the "leaky" inactivated state appears downstream of IFM binding; ii) the apparent shift 297 seen in DIIIAA G-V curve was likely due to the channel entering into the "leaky" fast 298 inactivation state from a closed state which contributed to the early current seen at 299 hyperpolarized voltages; and iii) the change in selectivity is only apparent as channels 300 transition into the leaky inactivated state. Additionally, this result implies that the high 301 energy barrier associated to the sharp reduction in conductance in the closed state is 302 likely linked to a separate set of S6 residues. 303 Residues identified at DIV S6 are also part of the inactivation gate. 304 DIV, similar to DIII, has been shown to have specialized roles in fast inactivation. We 305 identified two large residues at the bottom of DIV S6 based on the sequence alignment, 306 I1587 and L1591 (Fig. 5 A). When I1587 and L1591 residues were mutated to alanine 307 simultaneously (DIVAA), significant steady-state current was observed across all voltages tested (around 13% at 60mV) (Fig. 6B, C, D). However, single alanine mutations 309 (I1587A, L1591A) did not show significant change in fast inactivation (Fig. 6B-D  immobilization is similar to the WT and DIIIAA (Fig. 6 E, F). Furthermore, in DIVAA, the 312 apparent fast inactivation process had similar time course as the WT and DIIIAA 313 ( Supplementary Fig. 6). These results show that DIVAA channels can also transit into the 314 fast inactivated state but, similarly to DIIIAA, the fast inactivation gate is partially open, 315 creating a "leaky" inactivated state.  Supplementary Fig. 6). After determining the relative 334 permeability between Na + and K + as a function of time, we observed a decrease in 335 selectivity in DIVAA. At the onset of the pulse, the selectivity of DIVAA towards Na + was 336 high (PNa/ PK ≈ 13), but later in the pulse the relative permeability for Na decreased 337 (PNa/PK ≈ 2) (Fig. 6I). The change in selectivity follows the kinetics of inactivation as 338 evidenced by their similar time constants (Fig. 6J). The gating charge movement and the 339 activation process are not drastically affected by the double alanine mutations DIVAA 340 ( Supplementary Fig. 6). Based on the many parallels between DIIIAA and DIVAA, we 341 believed that the effects produced by the double alanine mutation in DIII S6 and DIV S6 342 share the same underlying mechanism and the identified hydrophobic residues in DIII 343 together with the ones in DIV are part of the fast inactivation gate. 344 To further demonstrate the shared role of DIII and DIV as part of the inactivation gate, we 345 mutated all four residues identified to alanine (DIII_IVAA, Fig. 6A). The effect of the 346 quadruple alanine mutation was drastic. This mutant exacerbated the effects seen in DIIIAA and DIVAA alone. Firstly, significant steady state current was seen (Fig. 6B, C). 348 Secondly, the ionic selectivity became severely impaired as evidenced by the fast inward 349 and steady-state outward component at some voltages (Fig. 6B, D). The PNa/PK at the 350 onset of the pulse was already ~2.4 and at steady state ~1.5 (Fig. 7C). These results

Residues identified in DI and DII S6 contribute to the channel behavior differently 366
The effects of alanine mutations in DI and DII were drastically different from DIII and DIV. 367 In DI S6, out of the two identified residues, only one residue was large and hydrophobic 368 (L437), and the second identified residue was an alanine (A441, Fig. 7A). Therefore, we 369 only tested a single alanine mutation (L437, DIA). In DIA, a small amount of steady state 370 current was detected as well as a ~15 mV right-shift in the voltage dependence of 371 activation (Fig. 7B, C). Despite this effect on steady state current, the h-infinity curve 372 shows that channels can fully inactivate (Fig. 7D).

384
Furthermore, there is no evidence of a time dependent change in the selectivity, as 385 evidenced by the lack of two components in the ionic currents ( Supplementary Fig. 7). 386 On the other hand, when we performed the double alanine mutation on the identified 387 residues in DII S6 (L792A_L796A, DIIAA. Fig. 7A), the effects were different to what was 388 described for DI, DIII and DIV. In DIIAA, the ionic current showed robust and complete 389 fast inactivation across all voltages tested (Fig. 7E). Despite the lack of steady state 390 current, at the end of the depolarizing pulse a large tail current was observed. The most 391 likely explanation for the origin of these "tail" currents is that they are gating currents. This 392 correlates with the observation of two distinct current components during the depolarizing 393 pulses (Fig. 7F). One component that is fast and always outward and another that follows 394 the reversal potential for sodium and is similar to the ionic currents seen in the WT 395 channels. After external application of TTX, the first component persisted while the 396 second diminished (Fig. 7H). Clearly, the first component originates from gating charge 397 movement, and the second component is ionic current. These results indicate that there 398 exists significant uncoupling between PD and VSD in DIIAA. Altogether, the results of the 399 DIA and DIIAA despite their possible relevance do not show unequivocally that these 400 residues form part of the inactivation gate, unlike DIIIAA and DIVAA. 401

402
The controversies on fast inactivation gate: location and identity 403 Fast inactivation is one of the defining features in Nav channels and has implications in 404 many physiological processes. The IFM motif has long been considered the inactivation 405 particle that blocks the pore during a "ball and chain"-type fast inactivation (18, 21). 406 Manipulations to the IFM motif can lead to severe impairments in fast inactivation, for 407 instance, by mutating IFM to QQQ, fast inactivation can be removed completely (18, 38). 408 Moreover, adding a small peptide that contains the IFM sequence into the intracellular 409 side of the membrane can partially rescue the fast inactivation in the QQQ mutant, which 410 suggests IFM could block directly the open pore (39). However, when the first structure 411 was solved for mammalian Nav channels(22), it clearly contradicted the status quo model. 412 Several Nav structures showed, not only that the IFM motif was in principle far away from 413 the pore in the putative inactivated state, but also that some residues previously identified 414 to influence fast inactivation did not interact directly with IFM at all, L437 and A438 being 415 the prime example (40). 416 An important "missing link" in the Nav fast inactivation saga, relates to the fact that, while 417 there are multiple high resolution Nav cryoEM structures, actually assigning a precise 418 functional state to individual structures is among the current major challenges in structural 419 biology. While it is possible that the structures were resolved in the fast inactivated state, 420 it is also possible that they may have been captured in a slow inactivated state, a different 421 process from fast inactivation (37, 41) with putatively significantly different structural 422 features. Yet, if we assume that some of the existing structures represent a fully or 423 partially fast inactivated conformation, the S6 helices and associated bulky hydrophobic 424 side chains logically emerge as strong candidates to block ion flow during fast inactivation 425 (Figure 1).
It is worth pointing out that the different Nav structures do not offer a clearcut consensus 427 regarding the exact residues that could form the fast inactivation gate. Different S6 428 residues have been proposed based on the different structures (Nav1.5 (33) and 429 Nav1.7(23, 28)). However, detergents were found lodging at the bundle crossing region 430 of many Navs, potentially distorting the conformation of the protein around that region 431 and further clouding potential interpretations. Here, our experiments and interpretations 432 are based on two structures that are free from this problem, NavPas and Nav1.7 M11. 433 Even though NavPas has never been functionally characterized and Nav1.7 M11 has 11 434 stabilizing mutations incorporated, our functional data support the idea that the pore 435 profiles from these two channels are likely to be similar to the ones under physiological 436 conditions and represent the fast inactivated pore. Therefore, they serve as a framework 437 to analyze the fast inactivation gate. 438 In DIII and DIV, none of the single alanine mutation yielded significant change to the fast 439 inactivation. Only when both residues in the same domain or when quadruple residues 440 were mutated to alanine, did we start to see significant amount of steady state current 441 from a leaky inactivated state (Fig. 2, 5, 6). This result clearly demonstrates that an S6-442 based fast inactivation comprises at least two layers of residues with side chains pointing 443 into the pore, making the hydrophobic region longer along the pore. The requirement of 444 two layers to make the gate might also be the reason why many previous single alanine 445 scanning studies on S6 did not show a clear effect and therefore were not able to identify 446 the fast inactivation gate (42): the two rings of identified bulky hydrophobic residues act 447 in series. interpretation of these results is that the double alanine mutations DIIIAA and DIVAA did 457 not affect the sequence of conformational changes that lead to inactivation but rather 458 make the inactivated state conductive. 459

Potential coupling between DIII, DIV S6 helices and the SF 460
An unexpected result from our work was that the ion selectivity changed concomitantly 461 following the kinetics of fast inactivation in DIIIAA and DIVAA. Na + ions were more 462 permeable than K + ions during the early component (before fast inactivation set in) while 463 the relative K + permeability increased over time. Given that TTX blocks both components 464 of the current completely, Na + and K + are clearly conducted through the pore domain of 465 the channel. It is worth noting that the general behavior of DIIIAA and DIVAA show a 466 striking resemblance with that of batrachotoxin (BTX) modified Nav channels. BTX 467 modifies Nav channels in three major ways: i) it removes fast inactivation; ii) it shifts the 468 GV curve to the left; iii) it alters the selectivity of the channel (43). Parallels could be found 469 in DIIIAA and DIVAA corresponding to all three modifications. More intriguingly, BTX is 470 hypothesized to bind to the neurotoxin receptor site 2 which is close to the mutated residues (I433, N434, L437 in DI S6 and F1579 N1584 in DIV S6 have been shown to be 472 crucial for BTX binding) (44). Therefore, it is likely that they share a similar underlying 473 mechanism, so that BTX evolved as a toxin to disrupt the inactivation gate at S6 even 474 more effectively than DIIIAA and DIVAA. 475 Asymmetry in the Nav channel pore 476 Due to the asymmetrical assembly of Nav channels, each domain is only similar, not 477 identical to the others, each one playing slightly different roles in the activation and 478 inactivation processes (34, 45). In this work, we also observed the effects of this 479 asymmetry. In DI, only one of the identified residues was bulky and hydrophobic (L437) 480 while the second position was an alanine (A441). L437A produces slight changes to fast 481 inactivation. The apparent conclusion would be that either the residues at DI S6 did not 482 contribute enough to the fast inactivation or our analysis of the structures was not able to 483 identify the correct residues in DI for fast inactivation gate. However, we cannot rule out 484 the importance of residue L437 since previous work has demonstrated the double 485 L437C/A438W mutant (CW) removes close to 90% of Nav fast inactivation (40). Our data 486 show that residues located at DI S6 are able to influence fast inactivation, however, the 487 effects observed are not sufficient to unequivocally assign it as part of the fast inactivation 488

gate. 489
The DII S6, double alanine (DIIAA) seems to not be involved in the inactivation gate. We 490 demonstrated that in DIIAA the gating current became disproportionally larger compared 491 to the ionic current, therefore, it is likely that the double alanine mutation in DII S6 492 decreases the open probability of the channel. DII VSD has been shown to be involved 493 mostly in channel activation instead of fast inactivation(45). We suspect that this apparent decrease in PO was a result of an uncoupling between the VSD and the PD during 495 activation. Our results suggest that the bottom of S6 region in DII plays a different role 496 from the other domains. Even though, our data support an uncoupling between VSD and 497 PD, we cannot rule out the possibility of a drastic enhancement of inactivation similar to 498 the effects of W434F mutation in Shaker K + channel (46, 47). Therefore, it requires further 499 investigation to determine the underlying mechanism of this uncoupling. 500 Fast inactivation as a multistep process 501

513
Our results show that the bottom of the S6 region of the pore serves as the fast 514 inactivation gate. One implication of these results is that fast inactivation is a multistep 515 process and the IFM motif binding is only one step, albeit critical, in a whole sequence of 516 conformational changes. Previously, the fast inactivation process was largely described 517 as a two-step process: VSD in DIV activates, exposing the binding pocket for the IFM 518 motif and IFM motif binds, blocking the pore. Our experiments provide a different 519 interpretation of the mechanism of fast inactivation in Nav channels as a series of 520 conformational changes. In this new model (Fig. 8), the activation of VSD in DIV triggers 521 the movement of the IFM motif by exposing a binding pocket. The IFM motif then binds 522 to its binding pocket that is away from pore. The binding site of the IFM is far from the S6 523 regions, therefore it is expected that the binding event of the IFM is transduced to the S6 524 gate in the pore through a yet to be defined pathway. Once the movement is allosterically 525 relayed to the S6 segments, the large residues at the bottom of S6 occlude the pore and stop the permeation. The exact nature of these conformational changes and the residues 527 involved in the third step is still elusive and the underlying mechanism is yet to be 528 determined. However, either a rotation, translocation, or a slippage of the S6 helices 529 towards the pore could lead to the pore closure during fast inactivation. To minimize the loss of gating currents due to P/-4 protocol while maintaining good 593 resolution of ionic currents and good health of the oocytes, -130mV was chosen to be the 594 subtraction voltage. Voltage clamp speed, measured by capacitive transients, yielded a 595 time constant around 75µs and the kinetics of gating current was reliably resolved. 596

Internal solution dialysis for oocytes 597
To ensure the ionic composition of the oocyte cytoplasm we exchanged the internal Where A1, τ1, A2 and τ2 represent amplitudes and time constants of the first and second 630 components, respectively. ISS is the steady state current. When one exponential was 631 used, the term of the second exponential component was eliminated. 632 To obtain the weighted time constant we used the following relationship: 633 5: = * + 634 IV) Relative permeability between Na + and K + ions was determined in bionic 635 environmental with 120mM intracellular K + and different extracellular Na + concentration. 636 The reversal potential was determined by fitting the instantaneous IV curve with a linear 637 relationship and locating the intersection point on the voltage axis. The reversal potentials as a function of depolarization time were subsequently used to determine the relative 639 permeability using Equation 6, derived from GHK equation (52, 53). where Vrev(t) is the reversal potential after depolarizing pulse of duration t; PNa and PK 642 represent relative permeability of Na + and K + respectively; [Na + ]o and [K + ]I represent 643 extracellular Na + concentration and intracellular K + concentration, respectively. 644 V) The ionic conductance (G(Vm)) was calculated by dividing either the peak current or 645 steady state current by the experimentally determined driving force, at each depolarizing 646 voltage (Vm). Subsequently, the curve was normalized to the maximal conductance 647 across all voltages (Gmax) to obtain the conductance versus voltage (G-V) curves. The 648 G-V curves were fitted using a two-state model: