Chlorella virus PBCV-1 (Paramecium bursaria chlorella virus-1) encodes the smallest protein (94 amino acids, named Kcv) previously known to form a functional K+ channel in heterologous systems. In this paper, we characterize another chlorella virus encoded K+ channel protein (82 amino acids, named ATCV-1 Kcv) that forms a functional channel in Xenopus oocytes and rescues Saccharomyces cerevisiae mutants that lack endogenous K+ uptake systems. Compared with the larger PBCV-1 Kcv, ATCV-1 Kcv lacks a cytoplasmic N-terminus and has a reduced number of charged amino acids in its turret domain. Despite these deficiencies, ATCV-1 Kcv accomplishes all the major features of K+ channels: it assembles into a tetramer, is K+ selective and is inhibited by the canonical K+ channel blockers, barium and caesium. Single channel analyses reveal a stochastic gating behaviour and a voltage-dependent conductance that resembles the macroscopic I/V relationship. One difference between PBCV-1 and ATCV-1 Kcv is that the latter is more permeable to K+ than Rb+. This difference is partially explained by the presence of a tyrosine residue in the selective filter of ATCV-1 Kcv, whereas PBCV-1 Kcv has a phenylalanine. Hence, ATCV-1 Kcv is the smallest protein to form a K+ channel and it will serve as a model for studying structure–function correlations inside the potassium channel pore.
Phycodnaviruses [1–4] consist of a genetically diverse, but morphologically similar, group of large dsDNA (double-stranded DNA)-containing viruses (170–560 kb) that infect algae from both fresh and marine waters [5,6]. The most studied phycodnaviruses are the plaque-forming chlorella viruses that infect certain freshwater, unicellular, eukaryotic, chlorella-like green algae [7,8]. Chlorella viruses encode up to 400 proteins of which 35–40% resemble entries in the public databases. Many chlorella virus encoded proteins are either the smallest or among the smallest representatives of their functional class. Consequently, some of the chlorella virus encoded proteins are the subject of intense physical and biochemical investigations. One example is the chlorella virus PBCV-1 (Paramecium bursaria chlorella virus-1) encoded 94-amino-acid K+ channel protein, named Kcv, which forms functional channels in Xenopus oocytes , Chinese-hamster ovary cells  and HEK-293 cells (human embryonic kidney cells) . The PBCV-1 Kcv structure is truly minimal: the channel monomer consists of two TMs (transmembrane domains), a pore loop and a 12-amino-acid cytoplasmic N-terminus. Like other K+ channels, PBCV-1 Kcv forms a functional tetramer  and electrophysiology experiments established that PBCV-1 Kcv has many properties similar to more complex K+ channels, including gating, selectivity, sensitivity to voltage and susceptibility to channel blockers . These results imply that these functional properties are inherent in the structure of the protein.
Kcv-like genes were cloned and expressed from 40 additional viruses that infect the same host, Chlorella NC64A; 16-amino-acid substitutions occurred among the 94 amino acids, producing six new Kcv-like proteins that formed functional K+ selective channels in Xenopus oocytes. However, the biophysical properties of some of these Kcv channels differed from PBCV-1 Kcv, including altered current kinetics with K+ and Rb+ and altered sensitivity to channel blockers [13,14]. Even more diversity is found between K+ channel proteins from chlorella viruses with different hosts. Virus MT325, which infects Chlorella Pbi, encodes a Kcv channel that is similar in size (95 amino acids) to PBCV-1 Kcv. However, the predicted architecture of MT325 Kcv differs significantly from PBCV-1 Kcv. Whereas PBCV-1 Kcv has a short amino acid cytoplasmic N-terminus and no cytoplasmic C-terminus, MT325-Kcv is organized differently; it lacks a cytoplasmic N-terminus but has 10 amino acids at the cytoplasmic C-terminus . Therefore the minimal structure and genetic variability of the K+ channels encoded by the chlorella viruses provide excellent model systems to study K+ channel structure–function relationships.
In this context, sequencing and annotation of another chlorella virus genome, ATCV-1, which infects Chlorella SAG 3.83 , revealed three ORFs (open reading frames) with the hallmarks of membrane transport proteins . One ORF has the features of an aquaglyceroporin (ORF Z300R), the second of a potassium transporter (ORF Z696R) and the third of a K+ channel (ORF Z585R). The remarkable feature of this putative K+ channel is that it is only 82 amino acids in size and thus 12 and 13 amino acids smaller than the other viral K+ channels [9,13,15]. In this paper, we describe the characteristics of this putative K+ channel. The results indicate that an 82 amino acid protein forms a functional K+ channel.
Expression in Saccharomyces cerevisiae cloning
The ATCV-1 kcv gene was cloned into a yeast expression vector pYES2/CT at BamHI and XhoI sites using primer sequences: 5′ primer CGGGGATCCATGTTGCTGCTTATCATA and 3′ primer ATTCTCGAGCTACCACGGAAACGTGAA. pYES2/CT (Invitrogen) carries a URA3 gene as a selectable marker for positive transformants in ura negative hosts. The exogenous gene is controlled by the GAL1 promoter and is expressed in the presence of galactose.
Both the empty vector and the vector that carries the ATCV-1 kcv gene were transformed into a potassium uptake-deficient yeast strain SGY1528 with the lithium acetate procedure. SGY1528 has the following genotype: MATα ade2-1 can1-100 his3-11,15 leu2-3,112 trp1-1 ura3-1 trk1::HIS3 trk2::TRP1 (a gift from Dr Daniel Minor, University of California at San Francisco). Transformants were selected on low-potassium medium using the methods described by Minor et al.  and Chatelain et al. . For the drug inhibition experiments, CsCl was added to the medium to produce the expected concentration before the medium was solidified with agar; BaCl2 was added to the 1 cm sterile paper discs at the indicated concentrations and the discs were placed on top of yeast cells spread on the agar plates. Plates were kept at 30 °C for 72 h or until the yeast had reached adequate growth.
Pichia pastoris transformation
The ORF encoding the ATCV-1 kcv gene was amplified from ATCV-1 kcv-pSGEM by the PCR (forward primer: 5′-CAAGGACCGAGCAGCCCCTCCTTGCTGCTTATCATACATATC-3′; reverse primer 5′-ACCACGGGGAACCAACCCTTATTACCACGGAAACGTGAAGCC-3′). The PCR fragment was subcloned into a modified P. pastoris expression vector pPICZ A (Invitrogen) containing a Kozak consensus sequence, a His7 tag, a proteolytic site for the H3C protease and an LIC (ligation independent cloning) site on the N-terminus of the protein sequence. The construct was confirmed by DNA sequencing (Cogentech, IFOM-IEO, Milano, Italy).
P. pastoris cells (SMD 1163 strain) were transformed with 3 μg of the PmeI linearized construct by using the Pichia Easy Comp™ Invitrogen kit as described by the manufacturer (Invitrogen). Positive colonies were selected from YPDS (10 g/l bacto yeast, 20 g/l bacto peptone and 20 g/l dextrose) agar plates containing 50 μg/ml zeocin.
Single colonies were grown in MGYH medium [1.34% (w/v) yeast nitrogen base, 1% (v/v) glycerol, 4×10−5% (w/v) biotin and 0.004% histidine] to an attenuance (D) at 600 nm of 4. After centrifugation at 3000 g for 10 min at 4 °C, the pellet was resuspended to a D600 of 1 in MMH medium [1.34% (w/v) yeast nitrogen base, 4×10−5% (w/v) biotin, 0.004% histidine and 0.5% (v/v) methanol] and grown for 24 h.
P. pastoris membrane preparation
Cells (1 g) were suspended in 22.5 ml of breaking buffer (50 mM NaH2PO4, pH 7.4, 1 mM EDTA, 5% glycerol and 5 mM dithiothreitol) containing 1 mM PMSF, 0.1 mg/ml trypsin inhibitor type II-O, 1 mM benzamidine and 0.1 mM Pefabloc®. An equal amount (1 g) of ice-cold, acid washed glass beads (0.25–0.5 mm diameter) was added and mixed vigorously on a vortex mixer for 20 min by alternating 30 s of vortex-mixing and 30 s on ice. The cell homogenate was centrifuged at 3000 g for 10 min at 4 °C to remove glass beads, unbroken cells and cell debris. The supernatant was removed and centrifuged at 30000 g for 45 min at 4 °C. The resulting crude membrane pellet (microsomes) was suspended in 0.3 ml of ice-cold breaking buffer.
For immunoblotting, SDS/PAGE-separated proteins were electro-transferred to a nitrocellulose (0.2 μm; Hybond) membrane, probed with anti-polyhistidine monoclonal antibody and visualized with alkaline phosphatase-coupled secondary antibody.
Expression of ATCV-1 kcv and PBCV-1 kcv genes in oocytes
Kcv cDNAs were cloned into a pSGEM vector (a modified version of pGEM-HE, courtesy of M. Hollmann, Max Planck Institute for Experimental Medicine, Göttingen, Germany). in vitro transcription was performed on linearized plasmids using T7 RNA polymerase (Promega) and cRNAs were injected (50 ng per oocyte) into Xenopus laevis oocytes, prepared according to standard methods . Electrophysiological measurements were made 2–4 days after injection.
Site-directed in vitro mutagenesis
Point mutations were created using the QuikChange® site-directed mutagenesis kit (Stratagene, La Jolla, CA, U.S.A.) and confirmed by DNA sequencing (Cogentech, IFOM-IEO).
Two-electrode voltage clamp experiments were performed using a GeneClamp 500 amplifier (Axon Instruments) and digitized at 50 kHz with a Digidata 1200 (Axon Instruments). Data acquisition and analysis were done using the pCLAMP8 software package (Axon Instruments). Electrodes were filled with 3 M KCl and had a resistance of 0.2–0.8 MΩ in 50 mM KCl. The oocytes were perfused at room temperature (25–27 °C) with a bath solution containing 50 mM KCl (or RbCl, CsCl, NaCl or LiCl as indicated in the Figure legend and text), 1.8 mM CaCl2, 1 mM MgCl2, 5 mM Hepes, adjusted to pH 7.4 with KOH (or RbOH, CsOH, NaOH or LiOH) at a rate of 2 ml/min. Mannitol was used to adjust the osmolarity of the solution to 215 mosmol/l. BaCl2 and CsCl diluted from 1 M stocks were added to the various solutions as indicated. The standard clamp protocol consisted of steps from the holding voltage of −20 mV to voltages in the range +80 mV to −160 mV; tail currents were measured at −80 mV. Instantaneous and steady-state currents were sampled after 10 ms and at the end of the voltage step (usually 600 ms) respectively.
Single-channel measurements by patch-clamp
Patch pipettes were pulled from thin-walled borosilicate glass capillaries, coated with Sylgard (Dow Corning) and fire-polished to a final resistance of 8–15 MΩ for single-channel experiments. Single-channel recordings were made in a cell-attached configuration. Pipette and bath solutions contained 100 mM KCl, 1.8 mM CaCl2, 1 mM MgCl2 and 10 mM Hepes (pH adjusted to 7.4 with KOH). When indicated, 100 mM RbCl was used instead of KCl. Experiments were performed at room temperature after removal of the vitelline membrane in a hyperosmotic solution (200 mM NaCl). Currents were recorded with a Dagan 3900 amplifier; data were low-pass-filtered at 2 kHz and digitized at a sampling rate of 10 kHz. Single-channel analysis was done using pCLAMP 9 (Axon Instruments) with the threshold-based algorithm.
Permeability ratios (PX/PK) were calculated according to the equation ΔErev=ErevX−Erev,K=RT/zF ln PX [X]o/PK[K]o, where Erev is the value in mV of the current reversal potential measured in the presence of the different univalent cations in the external solution (either X+ or K+); [X]o and [K]o are the cation concentrations in the external solution; and R, T, z and F have their usual thermodynamic meanings.
RESULTS AND DISCUSSION
Virus ATCV-1 codes for a channel protein with 82 amino acids
The chlorella virus ATCV-1 genome was recently sequenced and annotated . Among the 860 ORFs, one (ORF Z585R) had all the hallmarks of a K+ channel. The derived amino acid sequence of Z585R, termed ATCV-1 Kcv, is compared with the sequence of the prototype viral K+ channel, PBCV-1 Kcv in Figure 1(A). The ATCV-1 Kcv sequence is only 82 amino acids in size, 12 residues smaller than PBCV-1 Kcv. In spite of ATCV-1 Kcv's small size, the protein has the essential features of a K+ channel. Structural prediction programs TMHMM (http://www.cbs.dtu.dk/services/TMHMM) and Phobius (http://phobius.cbr.su.se) suggest two TMs, TM1 and TM2, and a pore helix (Phobius) preceding the filter region; it has a canonical selectivity filter sequence (TTVGYGD) and the aromatic couple, YF, in the pore helix. The two proteins have considerable sequence identity and similarity (Figure 1A) but presumably differ in structure as shown by the superimposed structural models in Figure 1(B). ATCV-1 Kcv lacks the ‘slide helix’ at the N-terminus (amino acids 1–14 in PBCV-1 Kcv) and most of the external loop connecting TM1 to the pore helix, the ‘turret’ (amino acids 36–48 in PBCV-1 Kcv). The C-terminal transmembrane segments, TM2, of ATCV-1 Kcv are longer and the intracellular mouth of this channel is probably narrower in comparison with PBCV-1 Kcv. The finding that ATCV-1 Kcv lacks the slide helix is of particular interest since this domain is essential for PBCV-1 Kcv activity .
Sequence alignment, homology modelling and protein expression of ATCV-1 Kcv in yeast
ATCV-1 Kcv forms a functional K+ channel
The first indication that ATCV-1 Kcv forms a functional K+ channel was obtained by complementation of a S. cerevisiae Δtrk1, Δtrk2 mutant that lacks endogenous K+ uptake systems [18,19]. Expression of ATCV-1 Kcv protein in the mutant rescued yeast growth in low K+ (0.5 mM), whereas yeast transformed with the empty vector only grew in high K+ (100 mM) (Figure 1C). These results indicate that ATCV-1 Kcv forms a functional K+ channel presumably at the yeast plasma membrane. Growth of the ATCV-1 Kcv expressing yeast was inhibited by Ba2+ and Cs+, two known K+ channel blockers (Figure 4A and Supplementary Figure S1A at http://www.BiochemJ.org/bj/420/bj4200295add.htm).
ATCV-1 Kcv forms a stable tetramer
Since functional K+ channels are tetrameric , the oligomeric state of ATCV-1 Kcv protein was examined. The channel was expressed in the yeast P. pastoris with an N-terminal His tag . The microsomal proteins were analysed by SDS/PAGE; one band, with an apparent molecular mass of 32.5 kDa, was detected by anti-histidine antibody (Figure 1D, lane 1). The observed mass is roughly four times that of the monomer (6.5 kDa) obtained after heating the sample at 100 °C for 3 min (Figure 1D, lane 2); this result indicates that ATCV-1 Kcv can form a stable tetramer.
ATCV-1 Kcv has characteristic electrical properties
ATCV-1 Kcv was expressed in Xenopus oocytes and subjected to electrophysiological investigation by the two-electrode voltage clamp procedure. Macroscopic currents occurred (Figure 2A) that clearly differ in amplitude and kinetics from the endogenous currents (water-injected oocyte). The ATCV-1 Kcv-induced currents have features typical of other chlorella virus Kcv channels [9,13–15]. The voltage clamp protocol (holding potential −20 mV, steps from 60 to −200 mV) evoked large instantaneous currents, which resulted in the current–voltage (I/V) relationship shown in Figure 2(B). The channel conductance is approximately linear between +/−80 mV but shows a pronounced negative slope at extreme negative voltages. Figure 2(A) shows, as an example, that the instantaneous current level recorded at −200 mV is smaller than that at −120 mV. Macro currents of Figure 2(A) also show a slow kinetic component at negative potentials that is interpreted as an inactivation of ATCV-1 Kcv conductance.
Properties of ATCV-1 Kcv currents at the macroscopic and single channel level
ATCV-1 Kcv is K+ selective
ATCV-1 Kcv channel currents were recorded in bath solutions with different ionic compositions to examine its selective properties. An increase of KCl concentration from 6 to 100 mM resulted in a progressive negative shift of the current reversal voltage. The inset in Figure 2(B) shows the semi-logarithmic relationship between the measured current reversal potential and the external K+ concentration. Linear regression analysis resulted in a straight line with a slope of 53.9±1.4 mV (n≥3). This value is close to the theoretical value of 59.2 mV calculated with the Nernst equation  and indicates that ATCV-1 Kcv is a K+ channel.
To examine the single-channel activity that underlies the ATCV-1 Kcv conductance, we measured the current in the plasma membrane of oocytes by the patch clamp method in cell-attach configuration. A typical example of single-channel fluctuations monitored in oocytes expressing the viral protein is shown in Figure 2(C). The I/V relationship of the unitary conductance (Figure 2D) has a pronounced outward rectifying shape and closely resembles that of the macro currents reported in Figure 2(B). The nonlinearity of the single channel I/V curve is a typical feature of this as well as of other Kcv channels and is due to a voltage-dependent reduction in the apparent channel conductance. At voltages more negative than −50 mV, the I/V curve progressively bends because the unitary currents become smaller even though the driving force is increasing. The reason for this voltage-dependent decrease in unitary conductance was not analysed further; however, since the lifetime of the channel openings becomes shorter with negative voltages it is likely due to a fast gating mechanism. Note that the full openings are not recorded at very negative voltages. In the linear range ±50 mV, the estimated single channel conductance was 80 pS. This value is similar to that of PBCV-1 Kcv, which, under the same ionic conditions (100 mM symmetric K+), has a unitary conductance of approx. 100 pS ( and A. Abenavoli, unpublished work).
In summary, the membrane conductance generated by ATCV-1 Kcv is due to the activity of a canonical ion channel that displays a characteristic voltage-dependent decrease in the unitary channel conductance. ATCV-1 Kcv thus resembles PBCV-1 Kcv that also exhibits a voltage-dependent single-channel I/V. This result indicates that the structural elements that underlie the gating mechanism are shared by both channel proteins.
Permeability to univalent cations
To study the permeability of ATCV-1 Kcv, we performed two-electrode voltage clamp experiments with different univalent cations in the external solution. Replacing 50 mM K+ with Rb+, Cs+, Na+ or Li+ resulted in the macroscopic currents shown in Figure 3(A) from which the corresponding I/V relations were produced (Figure 3B). Only Rb+ permeates the channel among the tested cations, as shown by the presence of an inward current. Rubidium induced a negative shift in the current reversal potential, Erev, of 16 mV (more clearly visible in Figure 5A: Erev K+=−20±1.08 mV and Erev Rb+=−36.2±0.86 mV, n=7). A qualitatively similar result was obtained when the permeability to rubidium was tested at the single channel level. Figure 3(C) shows examples of single-channel currents of ATCV-1 Kcv recorded in a cell-attach configuration with 100 mM KCl or RbCl in the pipette. Replacing external K+ with Rb+ resulted in a reduction in inward current amplitude and a negative shift of the reversal potential (Figure 3D). These results indicate that Rb+ has a lower permeability than K+ and confirm the results obtained with macro currents. Taking the constant field model as a basis for explaining selectivity , we have used macroscopic current Erev to estimate the permeability ratios reported in Figure 3(E) (PX/PK, where X is the tested cation) by using a simplified version of the Goldman equation (see the Experimental section). ATCV-1 Kcv is compared with PBCV-1 Kcv to highlight its peculiar behaviour; ATCV-1 Kcv is indeed the only Kcv channel found so far with a PRb/PK<1 . The permeability sequence of ATCV-1 Kcv is K+>Rb+>Cs+>>Na+≈Li+, whereas that of PBCV-1 Kcv is Rb+>K+>Cs+>>Li+≥Na+ ( and Table 1). A closer view of the currents recorded with different cations indicates that the cation composition of the medium affects other aspects of the I/V relations of the channel in addition to the reversal potential. Replacement of K+ with Na+ or Li+ not only reduces the inward current of the ATCV-1 Kcv generated conductance but also abolishes its outward current. This result is consistent with a blocking effect of external Na+ and Li+ on the outward K+ current. A reasonable explanation for this effect is that Na+ or Li+ enters the filter in the absence of K+ and that this entry results in a destabilization of the filter geometry . A similar effect by Na+, but not Li+, was previously reported for the prototype PBCV-1 Kcv channel . The present data indicate that Li+ behaves like Na+ in the ATCV-1 Kcv channel, meaning that structural differences between the two channels could allow the small Li+ ion to enter the pore. In contrast, Cs+ does not have the same effect on the outward K+ currents, which resemble the control condition (with external K+).
Permeability and conductance of ATCV-1 Kcv to different univalent cations
|ATCV-1 Y49F||PBCV-1 F66Y|
|ATCV-1 Y49F||PBCV-1 F66Y|
The aforementioned results imply that the cations interact with the channel protein. Hence permeation through Kcv channels is more complex than predicted by constant field theory. The nonlinearity of the Ii/V relations, i.e. in Rb+ and K+, furthermore implies that the permeability ratios could be voltage dependent, as shown by Eisenman and Horn . For this reason the Erev value alone is not sufficient to provide an absolute measure of Kcv channel permeability. We also estimated the conductance of the channel for different ions away from equilibrium . This relative conductance, GX/GK, provides additional information on the selectivity of the channels beyond the constant field theory.
The overall inward current in the presence of Rb+ is smaller than in K+, as expected from the lower permeability of the channel in Rb+. This results in a low conductance ratio (Figure 3F and Table 1: GRb/GK=0.3, conductance estimated in the linear portion of the I/V relation) and again confirms a difference in behaviour between ATCV-1 Kcv and PBCV-1 Kcv, which has a conductance ratio value largely higher than 1 (Table 1: GRb/GK=1.8). Finally it has to be mentioned that the kinetics of the channel change as a consequence of the transported substrate. While the inward current in K+ has a slow inactivating component, the current activates in a slow manner when Rb+ is present in the medium. This kinetic effect was not analysed further in the present study.
ATCV-1 Kcv is blocked by conventional K+ channel blockers
As mentioned previously, the yeast complementation experiments indicate that ATCV-1 Kcv is sensitive to the two K+ channel blockers Ba2+ (Figure 4A) and Cs+ (Supplementary Figure S1A). Figure 4(A) shows that, in selective conditions (0.5 mM KCl), growth of yeast expressing ATCV-1 Kcv is strongly reduced by 1 mM Ba2+. To confirm and further study the Ba2+ block, ATCV-1 Kcv macro currents were recorded in oocytes with and without Ba2+ in the external solution. Figure 4(B) shows the current response of an oocyte expressing ATCV-1 Kcv before and after addition of 1 mM Ba2+ to an external solution containing 50 mM K+. Barium produces a voltage-dependent block of the ATCV-1 Kcv conductance. It abolishes the inward current and also slightly affects the outward current. The overall effect of Ba2+ is summarized in the mean I/V curves obtained with four different Ba2+ concentrations, 0.01, 0.1, 1 and 10 mM (Figure 4C); ATCV-1 Kcv is partially blocked by sub-millimolar concentrations of Ba2+. The dose–response relationship of the Ba2+ block was obtained at different test potentials and plotted in Figure 4(D). The plotted values of fractional current, If, could be fitted by the Hill equation as follows:
where Kd is the concentration of Ba2+ for half inhibition and n is the Hill coefficient. Since the best fit of n was between 1.03 and 1.07, we fixed n at 1. The resulting Kd for Ba2+ changed with voltage and was 31, 40, 55 and 85 μM at −100, −80, −60 and −40 mV respectively. Owing to the voltage-dependent nature of the block, the apparent Kd obtained from the dose–response curves varies in a linear fashion with voltage when plotted in a semi-logarithmic scale (Figure 4E). The extrapolated value of Kd at 0 mV was 0.16 mM and decreased e-fold by 120 mV hyperpolarization, confirming that the Ba2+ block is facilitated by hyperpolarization. Collectively these data show that the smaller version of a Kcv-type channel behaves similarly to its larger orthologues. Other Kcv channels are also blocked by Ba2+ in a voltage-dependent manner with Kd values in the range of tens of micrometers (; A. Moroni, unpublished work).
ATCV-1 Kcv is blocked by external Ba2+
The same procedure was used to examine Cs+-sensitivity of the ATCV-1 Kcv channel. Addition of micromolar levels of Cs+ inhibits yeast growth when expressing the ATCV-1 Kcv channel (Supplementary Figure S1A). The electrophysiology data in Supplementary Figures S1(B)–S1(E) can be summarized as follows: ATCV-1 Kcv is inhibited by sub-millimolar concentrations of Cs+ in a strong voltage-dependent manner. The Kd value is strongly voltage-dependent and decreases e-fold for 31 mV hyperpolarization.
Structural basis for the difference in Rb+ selectivity
The macroscopic and microscopic I/V data in Figure 3 reveal that the ATCV-1 Kcv-generated channel is better at conducting K+ than Rb+. This behaviour is unusual for Kcv-type channels since all Kcv-like channels characterized to date  conduct Rb+ better than K+. In search of a structural explanation for this difference we noted that ATCV-1 Kcv has a tyrosine in the conserved selectivity sequence of the filter (GYG), whereas PBCV-1 Kcv has a phenylalanine (GFG) (Figure 1A). To determine whether this substitution in the filter region is responsible for the difference in Rb+ conductance and permeability, we prepared mutant channels in which we exchanged the two residues, i.e. ATCV-1 Kcv Y49F and PBCV-1 Kcv F66Y.
Both mutants were functional in oocytes and produced macrocurrents that were tested for K+ selectivity; both exhibited Nernstian behaviour when tested in buffers with different KCl concentrations (see Supplementary Figure S2 at http://www.BiochemJ.org/bj/420/bj4200295add.htm). The I/V relationships obtained from wild-type and mutant channels in the presence of K+ and Rb+ are presented in Figure 5 (mean data for wild-type ATCV-1 Kcv of Figure 5A include the data set of Figure 3B). For the aforementioned reasons we again used both the shift of Erev (ΔErev=Erev Rb–Erev K) and the ratio in conductance (GRb/GK) as parameters to estimate the selectivity of the channel for K+ over Rb+. In the case of ATCV-1 Kcv, the Y49F mutation results in a dramatic increase in Rb+ conductance and a small negative ΔErev of −7mV (inset of Figure 2B shows an enlargement of the currents at Erev K+=−21.52±1.13 mV, n=5; Rb+=−28.13±1.11 mV, n=4). The conductance ratio, GRb/GK, increases from 0.33 in the wild-type to 2.8 in the mutant (Table 1), while the permeability ratio, PRb+/PK+, estimated from the shift in Erev, is increased from 0.54±0.01 in the wild-type to 0.76±0.02 in the mutant (Table 1). These results imply that phenylalanine and tyrosine serve an important role in determining the selectivity between the two similar cations Rb+ and K+. However, this amino acid alone cannot explain the entire selectivity. The reverse mutation F66Y in PBCV1-Kcv does not cause a complete reversion of Rb+/K+ selectivity. But both parameters GRb/GK and ΔErev change in the mutant in a direction that is consistent with a reduced selectivity of the channel for Rb+ over K+; while the replacement of K+ by Rb+ causes a mean shift of the reversal voltage in the wild-type of +10 mV (Erev K+=−21.39±0.70 mV; Erev Rb+=−11.23±1.06 mV, n=5), the same operation results only in a +6 mV shift in the mutant highlighted by the enlargement of Figure 5(D) (Erev K+=−19.97±0.31 mV; Erev Rb+=−14.07±0.39 mV, n=5). Also, while the GRb/GK ratio is 1.8 in the wild-type channel, it is reduced to 1.5 in the mutant (Table 1).
Effect of a mutation in the selectivity filter on conductance and permeability to rubidium
Altogether these results are consistent with the idea that phenylalanine in the selectivity filter favours Rb+ over K+ conductance. However, the data also indicate that the overall selectivity of a filter can only be understood in the context of other structural parts of the channel protein.
The present results show that a Kcv protein from chlorella virus ATCV-1, which has the hallmarks of a K+ channel, forms a functional K+ channel. Even though the monomer is only 82 amino acids long it still forms a tetramer. This tetrameric channel has many features typical of more complex K+ channels. ATCV-1 Kcv can discriminate between cations and it has some voltage-dependent features; furthermore, ATCV-1 Kcv has a distinct voltage-dependent sensitivity to typical K+ channel blockers such as Ba2+ and Cs+. All of these functional features must be inherent in the basic structure of this miniature channel protein. A comparison of the ATCV-1 Kcv functions with another viral channel, PBCV-1 Kcv, indicates that it differs significantly because it is more permeable to K+ than Rb+. This result can be partially explained by a tyrosine residue in the ATCV-1 Kcv selectivity filter; in contrast PBCV-1 Kcv has a phenylalanine.
We thank Dan Minor (University of California at San Francisco) for the mutant yeast strain SGY1528 and for help with the complementation protocol.
This work was partially supported by the Public Health Service grant number GM32441 and the National Institutes of Health grant number P20RR15635 from the COBRE program of the National Center of Research Resources to J. L. V. E.; by [grant number 2005052099 PRIN 2005] the Ministero dell' Università e Ricerca; by project EDICT (European Drug Initiative on Channels and Transporters) EU PF7 [grant number 201924] to A. M.; and by a grant from the Deutsche Forschungsgemeinschaft to G. T.