A small viral potassium ion channel with an inherent inward rectification

ABSTRACT Some algal viruses have coding sequences for proteins with structural and functional characteristics of pore modules of complex K+ channels. Here we exploit the structural diversity among these channel orthologs to discover new basic principles of structure/function correlates in K+ channels. The analysis of three similar K+ channels with ≤ 86 amino acids (AA) shows that one channel (Kmpv1) generates an ohmic conductance in HEK293 cells while the other two (KmpvSP1, KmpvPL1) exhibit typical features of canonical Kir channels. Like Kir channels, the rectification of the viral channels is a function of the K+ driving force. Reconstitution of KmpvSP1 and KmpvPL1 in planar lipid bilayers showed rapid channel fluctuations only at voltages negative of the K+ reversal voltage. This rectification was maintained in KCl buffer with 1 mM EDTA, which excludes blocking cations as the source of rectification. This means that rectification of the viral channels must be an inherent property of the channel. The structural basis for rectification was investigated by a chimera between rectifying and non-rectifying channels as well as point mutations making the rectifier similar to the ohmic conducting channel. The results of these experiments exclude the pore with pore helix and selectivity filter as playing a role in rectification. The insensitivity of the rectifier to point mutations suggests that tertiary or quaternary structural interactions between the transmembrane domains are responsible for this type of gating.


Introduction
The key properties of K + channels such as ion selectivity and gating are well understood based on high-resolution structures [1][2][3][4][5][6]. However, despite this progress many functional aspects and in particular the significance of individual amino acids (AA) for function cannot be easily obtained from static protein structures. Therefore, other approaches are still needed to understand structure/function correlates in K + channels. In recent years "assumption free" genetic methods, which combine a selection and screening of randomly mutated channel libraries with high throughput functional assays were developed [7]. By using methods such as chemical mutagenesis, DNA shuffling or error-prone PCR large mutant libraries were generated for uncovering functional alterations in channels, which are caused by distinct mutational changes.
In this context also small K + channel proteins from viruses turn out to be an interesting model system for understanding functional aspects in the pore module of K + channels. From a structural and functional point of view the viral channels represent the "'pore module'", which is present in all known K + channels [8]. Because of this common architecture any insights from the viral channels are probably relevant for the function of complex human K + channels [9]. A further advantage of the viral channels as a model system is that they are the smallest proteins known to form a functional K + channel. This combination of small size and robust function limits the complexity of the system to less than 100 AAs.
Previous experiments have established that the activity of some viral K + channels is essential for infection of their hosts [10]. In the case of viruses, which infect unicellular Chlorella algae, it is for instance known that the channels are present in the membrane of the virion [11]. In an early step of infection this membrane fuses with the host plasma membrane [12]. This depolarizes the host [13] and causes a discharge of K + -salts and water from the algal cell [10]. As a result of these events the host cell looses its high internal turgor pressure, which otherwise prevents ejection of the virus DNA into the host. These data and the experimental finding that an efficient infection of the host cells can be inhibited by a specific block of the viral channels [10] implies that the viral genes are under evolutionary pressure and that their gene products need to form functional channels [14]. This assumption has been supported by experimental data, which have shown that the AA sequences of viral K + channels are variable and that the gene products are still functional in different test systems [15][16][17]. The sequence variability of viral K + channels, which can be isolated from various environmental samples, results in a large library of variable K + channel sequences with functional variability. We have exploited this structural diversity and have identified interesting functional differences, which are rooted in the sequence variability in these channels. The power of this unbiased approach is best illustrated by the fact that even very conservative AA exchanges caused significant functional differences. In the Kcv channel from chloroviruses; e.g. an exchange of Phe for Val or Leu for Iso in the first transmembrane (TM1) domain drastically altered the Cs + sensitivity of the channel as well as its voltage dependency [15,16]. The results of these experiments underscored the importance of the outer TM domain for K + channel function, which had largely been ignored. In the Kcv channel scaffold from SAG chloroviruses it was found that a mutation of Gly versus Ser in the inner transmembrane helix (TM2) affected the open probability of the channel. A closer investigation of these mutations uncovered a new type of gating mechanism, which is based on an intra-helical hydrogen bond between the critical Ser and an upstream partner AA in the alpha helix [17].
Here we further exploit the diversity of viral genes by screening viral K + channels from a marine habitat. We have previously reported that some viruses, which infect unicellular marine algae, also encode genes with the hallmarks of K + channels [18]. An initial functional testing of some of these proteins revealed that they have noncanonical architectures in their TM domains, but that they still form functional K + channels [18]. Here we perform a comparative examination of K + channels, which are similar in their structure but fundamentally different in their voltage dependency. While one channel generates an ohmic conductance the other two proteins exhibit a typical Kir-like inward rectification in which large inward currents occur only at membrane voltages negative to the K + equilibrium potential. The data show that this rectification is an inherent property of the protein and does not require Mg 2+ or polyamines as a blocker. By mutational studies we identify the TM domains as a crucial elements for an inherent inward rectification of this channel.

Materials and methods
The electrical properties of the putative viral channels in HEK293 cells were recorded as reported previously [18]. Constructs of Kmpv 1 and Kmpv SP1 were transiently expressed as fusion proteins with GFP on the C-terminus using the liposomal transfection reagent GeneJuice® (MERCK KGaA, Darmstadt, Germany). Measurements were performed at room temperature in a bath solution containing: 1.8 mM CaCl 2 , 1 mM MgCl 2 and 5 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES, pH 7.4) and either 50 mM KCl or 50 mM NaCl; different concentrations of BaCl 2 were added to the K + containing media to block the channels. The pipette solution contained 130 mM potassium-D-gluconic acid, 10 mM NaCl, 5 mM HEPES, 0.1 mM guanosine triphosphate (Na salt), 0.1 µM CaCl 2 , 2 mM MgCl 2 , 5 mM phosphocreatine and 2 mM adenosine triphosphate (Na salt, pH 7.4). The osmolarity of all solutions was adjusted with mannitol to 330 mosmol/kg. For the standard solutions we used JPCal software [19] to calculate a liquid junction potential of 15 mV, which was subtracted from the clamp voltages.
The Kmpv PL1 , Kmpv SP1 and Kmpv 1 proteins were synthetized cell-free according to the manufacturer's instructions (MembraneMax TM HN Protein Expression Kit (Invitrogen, Carlsbad (CA) USA) as reported previously [21]. In vivo synthesis occurred on a shaker with 1000 rpm at 37°C for 1.5 h in the presence of nanodiscs (ND) with 1,2-dimyristoyl-snglycero-3-phosphocholine (DMPC) lipids. The scaffold protein of the ND were His-tagged, which allowed purification of channel/ND-complexes via metal chelate affinity chromatography. The concentration His-tagged NDs in the reaction mixture was adjusted to 30 µM. For purification of the channel/ ND-complexes, the crude reaction mixture was adjusted to 400 µL with equilibration buffer (10 mM imidazole, 300 mM KCl, 20 mM NaH 2 PO 4 , pH 7.4 with KOH) and then loaded on an equilibrated 0.2 mL HisPur Nickel-Nitrilotriacetic acid Agarose (Ni-NTA) spin column (Thermo Scientific). For binding of the His-tagged NDs to the Ni-NTA resin the columns were incubated for 45 min at RT and 200 rpm on an orbital shaker. In the subsequent step the buffer was removed by centrifugation. To eliminate unspecific binders, the column was washed three times with 400 µL of a 20 mM imidazole solution. Finally, the His-tagged NDs were eluted in three fractions with 200 µL of a 250 mM imidazole solution. All centrifugation steps were performed at 700 x g for 2 minutes. After purification the elutions were stored at 4°C. For reconstitution of channel proteins into the lipid bilayer, a small amount (~2 µL) of the purified channel/ND-conjugates was added directly below the bilayer in the trans compartment.
Data are generally presented as mean ± standard deviation (sd) of n independent experiments. Statistical significances were evaluated by one-way ANOVA and student T-tests.

Similar channels with different functional properties
We screened a library of small and structurally similar K + channels from marine phycodnaviruses [18] for interesting functional differences. In this search we found three channels with similar lengths and high AA identity. Two of the channels, Kmpv PL1 and Kmpv SP1 , are 75% identical; both channels differ from the third channel Kmpv 1 in 29 AAs of which 16 are conservative or semiconservative AA exchanges (Figure 1(a)). In spite of the structural similarities a comparative functional analysis of Kmpv 1 and Kmpv SP1 in HEK293 cells uncovered striking functional differences with respect to gating and Ba 2+ sensitivity.
The most interesting finding is that Kmpv 1 generates under these conditions an ohmic conductance in HEK293 cells (Figure 1(c,e)), while Kmpv SP1 produces a strong inward rectification ( Figure 1(f,h)). Even though both channels exhibit a similar selectivity for K + over Na + (Figure 1(b-h); Table 1) they differ in their sensitivity to the canonical K + channel blocker Ba 2+ (Figure 2). While the ohmic Kmpv 1 channel is moderately sensitive to Ba 2+ (Figure 2(a-c)) the inward rectifier Kmpv SP1 exhibits a stronger voltage dependent block ( Figure 2(d-f)). The dose dependency for a Ba 2+ block of Kmpv SP1 (Figure 2(g,h)) has at −115 mV a half maximal inhibition at 14 µM Ba 2+ ; Kmpv 1 requires a ca. 100-fold higher concentration for the same degree of inhibition.

The inward rectification is an inherent function of Kmpv sp1
The voltage dependency of Kmpv SP1 (Figure 1(h), 2 (f)) resembles the strongly inward rectifying Kir channels [22]. To test if the voltage dependency in Kmpv SP1 is like Kir channels depending on the driving force for K + ions, we recorded currents over a range of K + concentrations in the bath. The representative current responses to voltage ramps show that the rectification shifts like in Kir channels with the K + reversal voltage (Figure 3(a)). The channel shows no appreciable conductance at voltages positive of the K + reversal voltage (V K+ ). These data suggest that Kmpv SP1 functions similar to Kir channels.
A hallmark of many Kir channels is that the slope conductance of the inward current has a square root dependency on the extracellular K + concentration [23][24][25]. To determine the respective dependency for Kmpv SP1 we measured the slope conductance of Kmpv SP1 between −135 and −95 mV in solutions with K + concentrations between 5 and 50 mM. The values from 4 measurements are shown in a double logarithmic plot (Figure 3(b)). The data can be jointly fitted with Equation (1) where g is the slope conductance, g the reference slope conductance (5 mM external K + and constant 130 mM internal K + ). K þ ½ o gives the variable external K + -concentration and n is a constant. The best fit is obtained with a n value of 0.46 (Figure 3(b)), which indicates that the slope conductance is like in Kir channels approximately proportional to the square root of the external K + concentration. Same experiments with cell expressing Kmpv SP1 in 50 mM K + (f) and 50 mM Na + (g) plus corresponding I/V relations in (h). The currents for I/V relations were sampled at end of voltage pulses (marked by arrows). Symbols at current traces correspond to symbols in respective I/V curves. In both I/V relations open symbols report measurements in K + and closed symbols report measurements in Na + . Note that Kmpv 1 generates an ohmic conductance in K + and Kmpv SP1 exhibits a strong inward rectification. Asterisks (*) indicate conserved residues, colons (:) indicate residues with very similar properties, and periods (.) indicate residues with weakly similar properties.
The data so far indicate that the viral channel shares the functional features of Kir channels. To examine the mechanism, which allows rectification, Kmpv SP1 and Kmpv PL1 were synthesized in vitro and reconstituted in planar lipid bilayers. For comparison the ohmic Kmpv 1 channel was examined in the same manner.  Reversal voltages (V rev , mean ± sd; number of experiments in brackets) were measured in HEK293 cells expressing either Kmpv SP1 or Kmpv 1 with 130 mM K + and 10 mM Na + in the pipette and either 50 mM K + or 50 mM Na + in the bath medium. The relative permeability ratio P K+ /P Na+ was calculated with Goldman equation based on measured V rev values (numbers in brackets denote number of independent recordings). For comparison the permeability ratio P K+ /P Na+ ratio of Kir channel ROMK1 is > 100 [46]. The representative data in Figure 4(a,c) show that Kmpv 1 generated current fluctuations with a small unitary conductance of ca. 6 pS and a high voltage independent open probability. The time averaged I/V relation, e.g. the product of unitary conductance and open probability, is largely linear over a voltage window of ±100 mV. This is consistent with the view that this channel is generating the ohmic I/V relation of Kmpv 1 measured in HEK293 cells (Figure 1(c,e)). The causal relationship between the Kmpv 1 protein and the small channel fluctuations is furthermore supported by control measurements. The respective channel fluctuations were not observed when nanodiscs, prepared as in Figure 4(a), but with no channel protein, were administered to the bilayer (Figure 4(b)).
The currents generated by Kmpv SP1 are different from those of Kmpv 1 . The representative single channel recordings show that this protein exhibits a high flicker type activity at negative voltages with only some longer lasting resolvable closed times (Figure 5(a)). These fluctuations are only evident at negative but not at positive voltages resulting in a strong inward rectifying I/V relation of the mean current ( Figure 5(b)). The I/ V relation from the bilayer measurements is similar to that recorded in HEK293 cells under comparable ionic conditions ( Figure 5(c)). In contrast to Kir channels, the viral inward rectifier does not require the PIP 2 phospholipid for activity [26]. The data furthermore underscore that the flicker type channel fluctuations in the bilayer are indeed responsible for the Kmpv SP1 currents in HEK293 cells. This assumption is further supported by the finding that 1 mM Ba 2+ greatly reduced the flickering channel activity in bilayers at negative voltages ( Figure 5(a,b)).
To estimate the steepness of the voltage dependency in the Kmpv SP1 channel we plotted the relative chord conductance (G rel ) of the mean current from Figure 5(d) as a function of voltage. The data were fitted to a Boltzmann function (dotted line) of the form G = (1 + e zF(V -V 1/2 )/RT ) −1 where z is the effective charge (= charge * electrical distance), V 1/2 the half activation voltage, F, R and T have their usual thermodynamic meaning. This yields values of −48 mV and 0.9 for V 1/2 and z, respectively. The results of this analysis indicate that the inherent voltage dependency of the Kmpv SP1 channel is more shallow than that induced by spermidine in canonical Kir channels but in the range of the voltage dependency generated by Mg 2+ block in these channels [27,28,29].
While expression of Kmpv PL1 generated no current in HEK293 cells it was successfully reconstituted in bilayers using the same procedure used for Kmpv SP1 . The data in Figure 1 supplement show that this protein generated the same inward rectifying I/V relation as the related Kmpv SP1 protein. From these data we conclude that the amino acid differences between these two   (Figure 1(a)) can be excluded as the origin of rectification.
In Kir channels inward rectification is generated by a voltage dependent block by cytosolic Mg 2+ or polyamines [4,30]. The finding that Kmpv SP1 rectifies in a bath solution with only KCl and HEPES in the same manner as in cells suggests that the channel may harbor an inherent mechanism of rectification. To exclude the possibility of HEPES [31] or any contamination of divalent cations in the bath solution as a source of rectification, we performed experiments as in Figure 5(a) in an unbuffered pure KCl solution. We also measured channel activity in a KCl/HEPES buffer plus 1 mM EDTA in the cis and trans chamber to chelate any possible divalent ion contaminant. In both cases Kmpv SP1 generated the same inward rectifying I/V relation as in the standard recording condition ( Figure 5(e)). In additional experiments channel activity was measured in a 100 mM KCl/ HEPES buffer plus 1 mM EDTA (in the same buffer as before) for more than 0.5 h. We reasoned that the affinity of a binding site in the channel to a putative blocker might be very high and that it might take time to dissociate [32]. However, when we compared the I/V relations from measurements at the beginning of the experiments with those recorded 0.5 h later, we found no difference; such a long incubation time is sufficient to release blockers from the channel pore of canonical Kir channels [29] (Figure 2 supplement). Collectively these data suggest that Kmpv SP1 has an inherent mechanism of gating, which generates a phenotype similar to that of Kir channels.

Structural basis of rectification
After finding functional similarities and differences between the viral inward rectifiers and canonical Kir channels we aligned Kmpv SP1 with representatives (Kir1.1, Kir2.1, Kir3.1 and Kir6.1) of the four major functional groups of Kir channels [22]. The alignment in Figure 3 supplement shows that except for the selectivity filter, the viral channel has essentially no similarity with canonical Kir channels. Most importantly the viral channel contains none of the AAs, which are crucial for the function of canonical Kir channels, i.e. the AAs, which determine the steepness of rectification or polyamine sensitivity.
After discovering that inward rectification of the viral channel differs from Kir channels and that rectification is an inherent property of Kmpv SP1 , we created a chimera with Kmpv 1 to identify domains that might be responsible for rectification. First, we addressed the question of whether the functional differences in gating could be caused by the minor AA variations in the selectivity filter. Notably the inward rectifying virus channels do not include any of the specific AAs, which are typical for the filter region of Kir channels (Figure 3 supplement). Neither the conserved disulfide bridge nor the pair of charged AAs, which are known to stabilize the filter of Kir channels [33][34][35], are present in Kmpv SP1 .
To examine the role of the filter in rectification we substituted the domain between the end of TM1 and the beginning of TM2 in Kmpv SP1 , with the equivalent domain from Kmpv 1 (ChA, Figure 6(a)). When this chimera was expressed in HEK293 cells it exhibited the inward rectification of the parental channel that donated the TM domains ( Figure 6(b,d)).
To quantitatively compare the rectification properties of the wild type (wt) channel and the chimera we calculated the rectification index (R.I.) as the ratio of the chord conductance at +25 mV (G 25 ) and −75 mV (G −75 ) according to Ozawa et al. [36],with Equation 2 A comparison of the R.I. value between wt channel and chimera ChA shows that both channels exhibit no apparent difference in their voltage dependency ( Figure 7). This suggests that the filter domain is not the origin of rectification. Further experiments showed that the chimera ChA maintained the low Ba 2+ sensitivity of the parental filter domain (Figure 6(c-f)). A scrutiny of the sequence of the two channels highlights a single AA difference in the vicinity of the selectivity filters with a Ser in Kmpv SP1 and a Phe in Kmpv 1 (Figure 1(a)). To test if this Ser is crucial for the high Ba 2+ sensitivity of Kmpv SP1 it was mutated into a Phe. Functional testing of this mutant showed that the protein retained inward rectification but had lost its high sensitivity to Ba 2+ (Figure 6(e,f)); the dose response curve became similar to that of Kmpv 1 . The results of these experiments show that Ba 2+ sensitivity is a function of the selectivity filter domain and that this property is not causally related to the mechanism of inward rectification.
The results so far suggest that the TM segments rather than the filter region confer inward rectification. To test this hypothesis we identified the AAs, which are different in the N-and C-termini between the ohmic Kmpv 1 channel on one side and the two rectifiers Kmpv SP1 and Kmpv PL1 on the other side ( Figure 1). The relevant AAs in Kmpv SP1 were mutated to those of Kmpv 1 and the mutants expressed in HEK293 cells to search for a loss of inward rectification. All mutants except for F72I and L76A generated a functional channel in which the mean current significantly exceeded the mean current of untransfected cells; the R.I. values of all the functional mutants show that they all maintained an inward rectification (Figure 7). The results of these experiments suggest that inward rectification is not conferred by a singular AA. Instead it seems to require tertiary or quaternary structural interactions of the TM domains.

Discussion
The present study shows that small viral K + channels from environmental samples provide a rich protein library, which promotes understanding of basic structure/function correlates in K + channels. Because of high evolutionary pressure on viral genomes these proteins exhibit an unbiased structural and functional diversity [15][16][17][18]. The main outcome of the present comparative analysis of three viral channel proteins is that one is a simple ohmic conductor while the other two represent inward rectifiers. The latter resemble the function of Kir-type channels [22]. Like their larger relatives from bacteria and animals they conduct no substantial outward current at membrane potentials positive to the K + equilibrium voltage, they are highly selective for K + and undergo a voltage-dependent block by Ba 2+ . In spite of their functional similarities to strong rectifying Kir channels the mechanisms of gating in the viral channels must be fundamentally different. Unlike Kir channels the viral channels exhibit strong inward rectification without contributions from any cytosolic domains; the latter are important determinants of rectification in Kir channels [30]. More importantly, while rectification in Kir channels is achieved by a voltage dependent block of the pore by cytosolic Mg 2+ and polyamines, the viral channels rectify even in a pure KCl solution. The latter comprise AAs in which Kmpv1 differs from the inward rectifiers KmpvPL1 and KmpvSP1 (see Figure 1). The relevant AAs were mutated individually in KmpvSP1 to match the sequence of the ohmic Kmpv1 channel. In the mutant dNSP1 three AAs (T2-I4) in the N-terminus of KmpvSP1 were deleted. All constructs were expressed in HEK293 cells as in Figures 1 and 2. For mutant channels, which conducted significantly more inward current than un-transfected cells, the rectification index R.I. was calculated. R.I. is the ratio of the chord conductance at +25 mV (G 25 ) and −75 mV (G -75 ) (R. I. = G 25 /G-75). The mean R.I. values (± sd; number of recordings in brackets) are plotted. The R.I. value for Kmpv1 (V) is significantly (P < 0.001; ***) larger than that of all other constructs. a one way anova test showed that the r.i. values of kmpvsp1 (sp) and all mutants/chimera are not significantly different (p > 0.1); none of the mutations converted a rectifier into an ohmic conductor.
Since the control experiments exclude contaminations in the solution as a source of channel blockers, we must assume that rectification is an inherent property of the viral channels. This interpretation is in line with previous reports, which had already shown that Kir mutants can acquire an intrinsic inward rectification. In the case of the ROMK1 channel it was found that the N171D mutation converted the channel into a strong inward rectifier even in the absence of Mg 2+ [37]. This electrostatic effect however cannot explain the mechanism of rectification in the viral channels. The equivalent of an anionic AA equivalent to residue 171D in the ROMK1 mutant is not present in the viral channels ( Figure 3 supplement). If we consider that the alignment between Kir channels and Kmpv SP1 may not be correct the residue E74 of Kmpv SP1 could be in the same position of D171 in the ROMK1 mutant. This however provides no explanation for rectification in the viral channels since this Glu is also conserved in the Ohmic conducting Kmpv 1 channel (Figure 1).
A Kir channel with an intrinsic rectification was also found in response to the E224G mutant in Kir2.1 [38]. In this case it was proposed that charges around the cytoplasmic pore may generate a local electrostatic potential at the channel entry for rectification. Since this domain is not present in the viral K + channels (Figure 3 supplement) it cannot account for their inherent rectification properties.
In the case of the intrinsic inward rectification of the viral channels we cannot explain the molecular mechanism of this voltage dependency. The finding that the pore loops of rectifying and nonrectifying channels can be swapped without any effect on rectification indicates that the pore domains with the selectivity filter architecture is not crucial for this voltage dependent phenomenon. This interpretation is supported by a mutation close to the selectivity filter, which alters the Ba 2+ sensitivity of the inward rectifier. A reduction of Ba 2+ sensitivity by nearly two orders of magnitude has in this case no impact on the rectifying properties of the channel. Hence the selectivity filter or the fine-tuning of the filter geometry seems to be irrelevant for an inherent inward rectification. This finding is unexpected considering the fact that inherent rectification is voltage and K + sensitive and that the filter is the domain, which provides the selectivity [1] for the channel and where the majority of the voltage drop in a K + channel occurs [39,40].
In the context of the differential Ba 2+ sensitivity of the channels it is interesting to mention, that several previous structural and functional studies have identified the 4th binding site, S4, in the filter of K + channels as the primary site of the Ba 2+ block [41][42][43]. The importance of S4 for the Ba 2+ block was confirmed by experiments in which a Thr in this site, which is part of the consensus sequence of the filter, was mutated to a Ser. This mutation caused a drastic decrease in the sensitivity of another viral K + channel but also in Kir channels to the Ba 2+ block [42,44]. In the three channels studied here the corresponding AA is already a Ser, which does not prevent a Ba 2+ block with µM affinity in Kmpv SP1 . Hence a Ser in this position is not solely responsible for a low Ba 2+ affinity. As a main structural determinant of Ba 2+ sensitivity the present data identify an AA in Kmpv SP1 , which is on the opposite side of the selectivity filter that is exposed to the external medium. A single point mutation S53F is sufficient to shift the high Ba 2+ sensitivity of Kmpv SP1 into that of Kmpv 1 . At this point we cannot answer the question of whether these data indicate an additional binding site on the extracellular side of the pore for Ba 2+ or whether the mutation S53F affects the entry of Ba 2+ into the filter and hence into its binding site deep in the electrical field. It is interesting to mention that an interplay of two residues on both sides of the selectivity filter is proposed for the Ba 2+ block in the Kir2.1 channel [45].
While the experiments exclude the filter as an important structure for inherent inward rectification in viral channels, they underscore the importance of the TM domains. The data also indicate that the conversion between a rectifying and ohmic conductance is not a matter of a single AA. The effective switch in voltage dependency, which was achieved by the exchange of the TM domains could neither be mimicked by the deletion of the first 3 AAs, which are lacking in Kmpv 1 nor by point mutations of the AAs, which are different between the respective channels. Taken together these data suggest that the fold of the TM is crucial for generating an inherent inward rectification in Kmpv SP1 . The question on how this is achieved and where the significant voltage dependency comes from remains unanswered.