signature=544e80f2766fbb8983249e5a44020222,TbIRK is a signature sequence free potassium channel from

Sequence analysis

Tb927.11.12490 was identified by an in silico screen with the hmmer-3.0b3 programT. brucei proteome (v4.0 from ftp.sanger.ac.uk)T. cruzi (TcCLB.511127.410) and L. major (LmjF.09.0480) exhibit at the protein level 41% (53% similarity) and 32% (42% similarity) identity, respectively (global Needleman-Wunsch alignment)Homo sapiens genome is a G-protein-activated inward rectifier potassium channel (NP_002231.1) that exhibits 17% identity and 29% similarity at the amino acid level.

A putative selectivity filter for TbIRK

Intriguingly, TbIRK does not contain the classical signature motif and we attempted to identify the alternative ion selectivity filter in this protein. Using BLAST against the TbIRK sequence, no reliable hits were found outside of the TriTryp-group, consisting primarily of T. brucei spp., T. cruzi spp. and Leishmania sp. We therefore looked for proteins with a similar arrangement of secondary structures. Using the homology model prediction algorithm of SWISS-MODELS1), similar to TbIRK. Of these eight groups two were eliminated, one for lack of coverage in the signature motif region and one for an interfering signal peptide, as predicted by Phobius1, Fig. S2). The two transmembrane domains in the studied proteins, including TbIRK are the only ones unequivocally predicted by the algorithms used. Unlike the other six groups, however, TbIRK does not contain the classical TxGYG motif but has the sequence GGYVG in this region. We hypothesized that this sequence might be the pore motif of TbIRK. By introducing point mutations into the putative selectivity filter, we could indeed alter the conductive properties of the channel (see below).

Figure 1

The unusual filter motif of TbIRK. (A) The protein sequence of TbIRK was aligned with representatives from six groups of similar proteins, as identified by HHBlits (Homology detection by iterative HMM-HMM comparison) within SWISS-MODELB) Section indicated by the red box in (A), containing pore loop and filter motif, indicated in blue. Residues mutated within this study are indicated with a red asterisk.

Other cases of proteins related to potassium channels with missing classical pore motifs include TMEM175

Electrophysiological characterization of TbIRK in Xenopus oocytes

For the functional characterization of TbIRK, we successfully expressed the gene in Xenopus oocytes and applied electrophysiological techniques. This indicates that the channel protein, at least partially, assumes a plasma membrane localization in the Xenopus oocytes. For control purpose, representative current traces in potassium and sodium medium from an oocyte injected with water are shown in Fig. 2A and B, respectively. Three days after microinjection of cRNA coding for TbIRK, we found substantially larger inward currents at the holding potential of −40 mV when the perfusion medium was changed from sodium to potassium medium as compared to water-injected oocytes. For a more detailed characterization of the current mediated by TbIRK, we performed measurements applying a voltage-step protocol from −110 mV to +50 mV in intervals of 10 mV, starting from a holding potential of −40 mV (Fig. 2E). Representative current-traces are shown in Fig. 2C and D. In potassium medium, expression of TbIRK mediated large inward currents at negative potentials and small outward currents at positive potentials. During the pulses of 300 ms no time-dependent variation of the current was observed. Figure 3A and B show averaged current-voltage relationships of TbIRK expressing oocytes and water injected oocytes, respectively, in sodium, potassium and N-methyl-D-glucamine medium (NMDG). The amplitude of the latter currents in oocytes expressing TbIRK is higher than expected. This might be due to difference in oocytes leak current and/or health conditions as compared to water-injected oocytes.

Figure 2

Electrophysiological characterization of TbIRK expressed in X. laevis oocytes. The 2-electrode voltage-clamp was used to study currents in oocytes injected with cRNA coding for TbIRK or with water. All recordings were done by applying the voltage-step protocol depicted in (E). The holding potential was −40 mV. With a frequency of 1 Hz voltage-steps of 300 ms duration were applied from −110 mV to 50 mV in 10 mV intervals. Currents recorded from a water-injected oocyte in potassium medium (KME) and sodium medium (NaME) are shown in (A) and (B). The corresponding representative current traces of a TbIRK-expressing oocyte in potassium medium and sodium medium (NaME) are shown in (C) and (D). (F) Recording of the same oocyte in potassium medium in presence of 10 mM CsCl.

Figure 3

Current-voltage relationship. (A) Current-voltage relationship of TbIRK expressing oocytes in sodium medium (squares), potassium medium (circles) and N-methyl-D-glucamine medium (open triangels) (mean ± S.D.; n = 31 for sodium and potassium medium and n = 23 for N-methyl-D-glucamine medium). (B) The corresponding I/V relationship of water-injected oocytes in sodium medium (squares), potassium medium (circles) and N-methyl-D-glucamine medium (open triangels) (mean ± S.D.; n = 25 for sodium and potassium medium and n = 19 for N-methyl-D-glucamine medium). (C) Comparison of the I/V relationship of TbIRK expressing oocytes in potassium medium (circles) or sodium medium (squares) in presence (open symbols) or absence (closed symbols) of 10 mM CsCl (mean ± S.D.; n = 31 and 16, respectively for TbIRK-expressing oocytes and n = 25 an 14, respectively for water-injected control oocytes). (D) Inhibition by caesium was determined by changing from sodium medium to potassium medium containing different CsCl concentrations at −80 mV. The observed currents normalized to the response elicited by potassium medium in the absence of caesium. The inhibition curve was fitted with an IC50 of 0.74 ± 0.11 mM (mean ± S.E.M.; n = 7).

Alternatively to the voltage-step protocol we applied a voltage ramp (Fig. S3). The two measurement strategies resulted in a similar current-voltage relationship, indicating the absence of time dependence of the TbIRK currents.

TbIRK-mediated currents in potassium medium were substantially reduced by caesium. To describe the inhibition of the channel by Cs+ in more detail, we measured the current induced by changing from sodium medium to potassium medium containing increasing concentrations of Cs+ at a holding potential of −80 mV. The inhibition of the current by Cs+ was fitted with an IC50 of 0.74 ± 0.11 mM (mean ± S.E.M.; n = 7; Fig. 3D). Endogenous conductances were not significantely affected by high (10 mM) concentrations of Cs+ (Fig. 3C). Ba2+ (10 mM), Cd+ (1 mM), tetraethylammonium (20 mM) and amantidine (100 μM), did not significantly influence the current. We are not aware of a potassium selective inward rectifier channel that is insensitive to 10 mM Ba2+. The activator for Slo2.2 channels

Additionally, a set of nine compounds known to block members of the inward rectifier potassium channel family were tested (Table S1). Except for 400 μM rosiglitazone that showed an inhibition of 39 ± 7% (mean ± S.D.; n = 3) none of the compounds had a significant effect at the tested concentrations.

The reversal potential of oocytes expressing TbIRK was determined in medium containing Na+, K+ or NMDG as major cations. The reversal potential in sodium medium was −50 ± 2 mV (mean ± S.E.M.; n = 30), in potassium medium −10 ± 1 mV (mean ± S.E.M.; n = 30) and in NMDG medium −67 ± 2 mV (mean ± S.E.M.; n = 23). Using the Goldman-Hodgkin-KatzK/pNa of the plasma membrane of TbIRK-expressing oocytes to be 7.3 (see also Materials and Methods). For water-injected control oocytes we found a relative permeability ratio pK/pNa of the plasma membrane of about 1.3. As the currents through the plasma membrane are composed to a minor part from endogenous contributions and a major part mediated by TbIRK, the relative permeability ratio pK/pNa of the channel is somewhat larger than 7.3. As inward rectifiers do not necessarily follow Hodkin-Huxley kinetics, we quantified currents observed at −110 mV in potassium (IK) and sodium medium (INa) in TbIRK-expressing oocytes and subtracted those observed in control oocytes. IK/INa was 8.3. This value is similar to the one obtained using reversal potentials for the estimation of the selectivity.

Point mutations in the predicted selectivity filter alter ion selectivity

The glycine residues in the predicted selectivity filter GGYVG at positions 131, 132 and 135 were mutated individually to alanine. These three point mutations had drastic effects on the survival rate of injected oocytes. While injection with water or mRNA coding for TbIRK resulted in little mortality, most oocytes injected with mRNA coding for one of the mutant channels showed damage, as indicated by pigmentation changes and low membrane resistance. Fewer than 10% of the oocytes expressing mutant channels could be analyzed using electrophysiological techniques. After 48 h in modified Barth medium 50–60% of the oocytes showed a white spot at the animal pole, presumably the nucleus floating to the surface membrane

Surviving oocytes expressing mutant receptors were functionally characterized. One of the notable changes as compared to wild type receptors was a shift of the reversal potential determined in potassium medium (reversal potential of about −11 mV) close to the corresponding value determined in sodium medium (about −16 mV). Figure 4 shows original current traces in sodium and potassium medium and potassium medium supplied with 10 mM CsCl recorded from oocytes expressing wild type and mutant G132A TbIRK obtained using a voltage-step protocol depicted in Fig. 2E. The corresponding averaged current-voltage relationships of the mutant channel are shown in Fig. S4. Interestingly and for reasons that are far from clear, inward rectification was converted into an outward rectification upon mutation. Current amplitudes determined at −80 mV in sodium medium relative to those in potassium medium amounted to 22 ± 3% (mean ± S.E.M., n = 10) and 73 ± 7% (mean ± S.E.M., n = 6) in wild type and mutant G132A, respectively. Inhibition by 10 mM Cs+ was determined similarly for wild type channels and the mutant G132A. Residual current amounted to 23 ± 3% (mean ± S.E.M., n = 10) and 81 ± 11% (mean ± S.E.M., n = 6) in wild type and mutant channels, respectively. As compared to wild type channels, the mutant showed enhanced conduction of Na+ and decreased inhibition by Cs+. Similar observations were made with the mutants G131A and G135A, but the number of oocytes measured was too small to reliably report quantitative values. We conclude that at least the mutation G132A has a large impact on the ion selectivity of the channel, reducing the selectivity for potassium over sodium ions. Moreover, this mutation affects the recification properties of the channel.

Figure 4

Loss of selectivity upon mutation of G132A located in the suspected selectivity filter. All recordings were done by applying the voltage-step protocol depicted in Fig. 2E. Representative current traces obtained from TbIRK-expressing oocytes in sodium medium (NaME), potassium medium (KME) and potassium medium supplemented with 10 mM CsCl are shown in the upper row. Recordings obtained with a oocyte expressing the G132A-mutant are shown in the lower row.

Localization of TbIRK in procyclic and bloodstream form parasites

To determine the localization of the protein in the parasites we generated stably transfected cell lines expressing tetracycline-inducible C- and N-terminal hemagglutinin (HA)-tagged versions of TbIRK. Each HA-construct was verified by Western blot (Fig. S5). Cells were analyzed by immunofluorescence microscopy 24 h post-induction for procyclic and 48 h post-induction for bloodstream form parasites. TbVP1T. brucei, was used as acidocalcisome marker. C-terminally tagged TbIRK in procyclic form parasites (Fig. 5A) and N-terminally tagged TbIRK in bloodstream form parasites (Fig. 5B) localized to acidocalcisomes. In procyclic form parasites N-terminal tagging did not yield a visible signal. In bloodstream form parasites C-terminally tagged versions partially localized to acidocacisomes and partially to the endoplasmic reticulum (Fig. 5C). It has previously been shown that a tagged protein may be retained in the endoplasmic reticulum on its way to the final destination

Figure 5

TbIRK co-localizes with a marker for acidocalcisomes. Co-localization of hemagglutinin (HA)-tagged TbIRK (green) with the acidocalcisomal marker TbVP1 (red) for C-terminally tagged TbIRK in procyclic form parasites (PCF, A) and N-terminally tagged TbIRK in bloodstream form parasites (BSF, B). (C) With the HA-tag at the C-terminus of TbIRK, its signal (green) partially co-localizes with TbVP1 (red, top) and partially with the ER marker BiP (red, bottom). Cells were counterstained with DAPI, shown in blue, visualizing the nuclear and kinetoplast DNA. DIC, differential interference contrast. Scale bars indicate 10 µm.

Pearson’s correlation coefficients (PCC) characterizing co-localization of TbVP1 with TbIRK for three biological replicates of C-terminally HA-tagged TbIRK in procyclic form parasites were 0.71 ± 0.08 (n = 13), 0.75 ± 0.08 (n = 12), 0.78 ± 0.04 (n = 10) (mean ± S.D., n = number of cells). In bloodstream form parasites, PCC for two biological replicates of N-terminally HA-tagged TbIRK were 0.70 ± 0.04 (n = 11), 0.69 ± 0.07 (n = 3) (mean ± S.D., n = number of cells). For the C-terminal clone partially co-localizing, it was 0.54 ± 0.08 (n = 7). PCC in two biological replicates for the ER-marker BiP

We were thus able to localize HA-tagged TbIRK to acidocalcisomes. To our knowledge, this is the first time a potassium channel has been localized to the acidocalcisomes. The role of potassium in these organelles is not clear yet. Reported K+-concentrations in acidocalcisomes of trypanosomatids are dependent on the species and cultivation conditionsT. brucei acidocalcisome-rich fractions

Acidocalcisomes are found in eukaryotic as well as prokaryotic species. In trypanosomatids the acidocalcisomes have been studied in most detail. The cellular functions of the acidocalcisomes described so far are polyphosphate (polyP) and cation storage2+ signalinget al.T. cruzi. The mRNA coding for the TbIRK ortholog in T. cruzi (TcCLB.509029.20) was down-regulated only transiently by 22%T. cruzi may only play a minor role in osmoregulation.

RNAi-mediated downregulation of TbIRK

To investigate whether the protein is essential, we down-regulated its expression using inducible RNAi in both procyclic and bloodstream form parasites. Induction of RNAi by tetracycline did not affect the growth of procyclic form parasites (Fig. 6). To verify the efficiency of the RNAi constructs, we isolated total RNA from the tested clone 48 h after induction. The change in mRNA level of TbIRK was determined by qPCR. We found a down-regulation of the TbIRK mRNA of 80 ± 5% (n = 3; mean ± S.D.) in procyclic form parasites. For bloodstream form parasites, the target mRNA was not efficiently down-regulated (<50%, n = 3).

Figure 6

RNAi-mediated downregulation of TbIRK in procyclic form parasites. (A) Growth of non-induced (open circles) and induced (triangles) was monitored over 10 days (mean ± S.D., n = 3; error bars are smaller than the symbols). Induction of RNAi against TbIRK did not result in a growth phenotype. Changes in target mRNA levels were confirmed by qPCR (B). The black bar represents the mRNA level of non-induced cells and the white bar represents the mRNA level of TbIRK in induced cells. TbIRK mRNA was downregulated by 80 ± 9% (mean ± S.D., n = 3).

Summary

In summary, we describe here an ion channel in T. brucei, TbIRK, with a moderate selectivity for potassium ions, but lacking the classical signature sequence forming the selectivity filter. Except for its insensitivity to Ba2+, the channel is structurally and functionally reminicent of inward rectifier potassium channels. TbIRK localizes to acidocalcisomes. It remains to be described how this potassium channel is involved in the acidocalcisome functions including polyphosphate (polyP) and cation storage2+ signaling