K⁺-Channel Transgenes Reduce K⁺ Currents in Paramecium, Probably by a Post-translational Mechanism

Corresponding author: Yoshiro Saimi, Laboratory of Molecular Biology, University of Wisconsin, 1525 Linden Dr., Madison, WI 53706., ysaimi{at}facstaff.wisc.edu (E-mail)

Communicating editor: S. L. ALLEN

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

PAK11 is 1 of more than 15 members in a gene family that encodes K⁺-channel pore-forming subunits in Paramecium tetraurelia. Microinjection of PAK11 DNA into macronuclei of wild-type cells results in clonal transformants that exhibit hyperexcitable swimming behaviors reminiscent of certain loss-of-K⁺-current mutants. PAK2, a distant homolog of PAK11, does not have the same effect. But PAK1, a close homolog of PAK11, induces the same hyperexcitability. Cutting the PAK11 open reading frame (ORF) with restriction enzymes before injection removes this effect entirely. Microinjection of PAK11 ORF flanked by the calmodulin 5' and 3' UTRs also induces the same hyperexcitable phenotype. Direct examination of transformed cells under voltage clamp reveals that two different Ca²⁺-activated K⁺-specific currents are reduced in amplitude. This reduction does not correlate with a deficit of PAK11 message, since RNA is clearly produced from the injected transgenes. Insertion of a single nucleotide at the start of the PAK11 ORF does not affect the RNA level but completely abolishes the phenotypic transformation. Thus, the reduction of K⁺ currents by the expression of the K⁺-channel transgenes reported here is likely to be the consequence of a post-translational event. The complexity of behavioral changes, possible mechanisms, and implications in Paramecium biology are discussed.

ALL known K⁺ channels have similar features: each consists of four pore-forming subunits that form the K⁺ filter and aqueous pathway across the membrane (DOYLE et al. 1998 ). Some channels are homotetramers, especially those overexpressed heterologously from a single subunit gene, whereas others appear to be heterotetramers. JEGLA and SALKOFF 1995 first described two K⁺-channel sequences from a Paramecium tetraurelia genomic library. We have expanded this K⁺-channel subfamily by an additional five members (Y. SAIMI, unpublished data). Meanwhile, the Paramecium genome-sequencing project (DESSEN et al. 2001 ) has revealed nine more sequences, all unique. As of this writing, there is evidence for at least 16 such K⁺-channel subunit genes in P. tetraurelia, and this is clearly an underestimate, since only a small fraction (<5%) of the Paramecium genome has been sequenced. In Paramecium, at least seven K⁺-specific currents that differ in gating, kinetics, activation, and inactivation have been recognized electrophysiologically (SAIMI and KUNG 1987 ; PRESTON et al. 1990A ; R. R. PRESTON, unpublished data). How so many K⁺-channel genes are expressed and logistically deployed to different structures and with different functions, all within a single cell, will be a challenge to understand.

In an attempt to identify their role in vivo, we have examined the behavioral and electrophysiological effects of expressing individual K⁺-channel subunit transgenes in macronuclei of paramecia. In two out of the three transgenes studied, we have encountered a phenomenon resembling silencing. Injection of certain full-length Paramecium K⁺-channel genes (PAKs; JEGLA and SALKOFF 1995 ; LING et al. 1998 ) elicited hyperexcitable behaviors similar to those of pantophobiac, a K⁺-channel loss-of-function mutant (SAIMI et al. 1983 ). Surprisingly, our molecular analyses indicate a post-translational instead of a post-transcriptional mechanism.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Stocks and cultures:
P. tetraurelia stocks 51s (+/+; SONNEBORN 1970 ) and nd6 (nd6/nd6; LEFORT-TRAN et al. 1981 ) were cultured at 23° or 28° in a growth medium of buffered wheat grass inoculated with Enterobacter aerogenes (SONNEBORN 1970 ). nd6 cells are behaviorally wild type, albeit they cannot discharge their trichocysts. Since these mutant cells survive the trauma of macronuclear microinjection better than the wild-type 51s (KIM et al. 1992 ; HAYNES et al. 2000 ), only cell lines homozygous for the nd6 mutation were used as the recipient for microinjection.

Preparation of total DNA and total RNA from Paramecium:
Standard molecular biology techniques were followed (SAMBROOK et al. 1989 ; AUSUBEL et al. 1999 ). Total DNAs were extracted from 51s cultures and descendants of microinjected nd6 cells as previously described (HAYNES et al. 1995 ). To obtain RNA, cells from 51s, nd6, or descendants of uninjected and microinjected cell lines were harvested and washed once in 10 mM Tris buffer, pH 7.5. Approximately 125-µl aliquots of cells (1 x 10⁵ cells) were lysed in 375 µl of TRI-REAGENT (Molecular Research Center, Cincinnati) in the presence of RNase-free glass beads (BETERMIER et al. 2000 ). The total RNA of each sample was then purified according to the protocol provided by the manufacturer as described previously (HAYNES et al. 2000 ).

Cloning of PAK1, PAK11, and PAK2:
A complete XbaI digest of total Paramecium DNA was probed with a ³²P-labeled 0.85-kb fragment of PAK1 DNA (JEGLA and SALKOFF 1995 ) (5' ATGCTCTCAATAAAG ... ACACTCACACTTGGC 3'; see GenBank entry). DNAs corresponding in size to the 4.5-kb band detected in Southern blots were cloned into the SpeI site of Bluescript II KS⁺ (pBS; Stratagene, La Jolla, CA) to form a partial library. The pBS-PAK11 used in the present study was among the positive clones identified on colony lifts by the PAK1 probe. It was consequently purified and sequenced. When a ³²P-labeled 0.8-kb fragment of PAK2 (JEGLA and SALKOFF 1995 ; (5' GGTGGAATATATACA...CAACCTTAAAAAATT 3') was used to probe the same XbaI partial library, pBS-PAK2 was retrieved, purified, and sequenced. Sequenced portions of PAK2 have only minor differences from the partial sequence of JEGLA and SALKOFF 1995 .

PCR and RT-PCR were performed using Advantage cDNA polymerase mix (Clontech, Palo Alto, CA). A SpeI- and KpnI-linker primer-pair specific for the PAK1 open reading frame (ORF) (5' oligo, GACTAGTATAAATGATACCCAAACTCCAAGGG; 3' oligo, GGGGTACCTCATATAATTTTTTACACGACTTATC), truncated PAK1 3' (the same 5' oligo and 3' oligo: GGGGTACCTCATTGGTATGATAAGCAGCCA), or PAK11 ORF (5' oligos, GACTAGTATAAATGATACCCGGACTCTAAAAG and GACTAGTATAAATGATAACCCGGACTCTAAAAGATTAGAT for in-frame and frameshift, respectively; and 3' oligo, GGGGTACCTCATATAATGTTTTATACCACTTATC for both; Operon, Alameda, CA) was used to amplify from total DNA or the first-strand cDNA. Additional primers up- and downstream of the ORFs had also been used in various RT-PCR experiments to finalize the boundaries of the ORFs. The first-strand cDNA was synthesized from purified 51s total RNA using oligo/poly(dT) with a SuperScript preamplification system (Life Technologies, Rockville, MD). All reactions were performed in a programmable thermal controller 100 (MJ Research, Watertown, MA). Reactions using genomic DNA as template yielded molecules of 2.4 kb. RT-PCRs yielded cDNAs, each with three introns spliced from their corresponding PAK1 and PAK11 genomic sequences, respectively.

The PCR-amplified genomic ORFs and RT-PCR-amplified cDNA ORFs of PAK1 and PAK11 were digested with the linker restriction enzymes and cloned into Paramecium expression vector pPXV at their corresponding sites (HAYNES et al. 1995 ). At least two clones of each category were amplified and tagged for automatic sequencing using the ABI PRISM Big Dye Terminator cycle sequencing kit (Applied Biosystems, Foster City, CA). The reactions were then sequenced at UW Biotechnology Center (Madison, WI). Sequencing results were analyzed using EditView (Applied Biosystems) and SeqMan II (DNAStar, Madison, WI).

Preparation of plasmid DNAs for microinjection:
Plasmids were digested with various restriction enzymes to completion as judged by gel electrophoresis. After digestion the plasmids were phenol-chloroform extracted, ethanol precipitated, and washed twice with 75% ethanol. All samples were resuspended in Tris-EDTA buffer pH 8.0 at 0.5–5 µg/µl for macronuclear injection. Standard molecular-biological techniques were used (SAMBROOK et al. 1989 ; AUSUBEL et al. 1999 ).

Microinjection for transgene expression in nd6 cells:
An aliquot of 5–10 pl of the above-mentioned linearized plasmid at various concentrations was microinjected into the macronucleus of each recipient cell as previously described (HAYNES et al. 1995 ). Recipients were cells from young clones of less than eight fissions after autogamy. Each plasmid sample was injected into six or more cells. The descendants of the individual recipients were cultured as individual clones and subjected to behavioral, electrophysiological, and molecular analyses.

Behavioral assay:
Uninjected and microinjected cells were cultured for at least 24–48 hr before their behavior was tested. Cells were transferred into an adaptation solution [4 mM KCl, 1 mM CaCl₂, 1 mM N-(2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid; HEPES), 0.01 mM EDTA, pH 7.2] for 10 min to 1 hr. Then they were individually transferred into various testing solutions, including a Na⁺-test solution (10 mM NaCl, 1 mM CaCl₂, 1 mM HEPES, 0.01 mM EDTA, pH 7.2), where the duration of continuous backward swimming of each cell was monitored using a stereomicroscope and a stopwatch (KUNG et al. 1975 ).

Electrophysiology:
Membrane currents were recorded from cells under voltage clamp using the established techniques (PRESTON et al. 1992 ). The intracellular capillary microelectrodes used for membrane potential recording and current injection contained 3 M KCl and had tip resistances of 20–40 M. The cells were bathed in 1 mM KCl, 1 mM CaCl₂, 1 mM HEPES, and 0.01 mM EDTA, pH 7.2. Currents shown were filtered at 1 kHz and have been corrected for linear leak current as described (PRESTON et al. 1990A , PRESTON et al. 1992 ).

Analysis of total DNA by Southern blot and total RNA by Northern blot:
Gel electrophoresis of total DNA isolated from descendant clones of uninjected nd6 controls and various microinjected transformants for Southern blots and their total RNA for Northern blots was performed as previously described (HAYNES et al. 1995 , HAYNES et al. 2000 ). The probes, PAK11 ORF and calmodulin CAM ORF, were labeled using the RediprimeII random prime labeling system with [-³²P]dCTP (Amersham Life Science, Arlington Heights, IL) as directed by the manufacturer. DNA marker was from GIBCO/Life Technologies (Alameda, CA) and RNA markers were from GIBCO/Life Technologies or Promega (Madison, WI). The radioactive signals were recorded on Phospho-Imager cassettes and then digitized and analyzed using ImageQuant 1.2 (Molecular Dynamics, Sunnyvale, CA).

Sequence comparison and secondary structure prediction:
The protein sequence from the expected ORFs of PAK1 and PAK11 were used to do a homology search in the most recent databases employing several algorithms available at the National Center for Biotechnology Information web site. Secondary structures were analyzed using the PROTEAN program (DNAStar).

The sequence data presented in this article have been submitted to the EMBL/GenBank Data Libraries under accession nos. AF424539 (PAK1), AF432226 (PAK2), and AF424540 (PAK11).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The sequences for two very closely related K⁺-channel genes, PAK11 and PAK1 (LING et al. 1998 ), as well as a more distantly related PAK2, are germane to the present article (JEGLA and SALKOFF 1995 ; LING et al. 1998 ). Corrections to the full-length sequences of PAK11 and PAK1 were provided to GenBank. The ORFs of PAK11 and PAK1 were determined through successful RT-PCR amplification from the wild type using different gene-specific primer pairs. Within the very large superfamily of K⁺ channels with six transmembrane -helices, the PAKs studied here most closely resemble the AKT/EAG/ERG group of K⁺ channels that have potential cyclic nucleotide-binding sites in their C-terminal cytoplasmic domains (ROBERTSON et al. 1996 ; GANETZKY et al. 1999 ).

PAK11 transgenes cause hyperexcitable behaviors:
We transformed wild-type paramecia by microinjecting plasmids into individual macronuclei and examined the clonal descendants of the recipients (HAYNES et al. 2000 ). The insert in the injected plasmid (pBS-PAK11) contained the entire PAK11 ORF as well as 0.5 kb of the 5' and 1.5 kb of the 3' untranslated region (UTR) flanked by the restriction sites XbaI and XhoI (Fig 1). The descendants of wild-type cells injected with an XbaI-XhoI-double digest of pBS-PAK11 invariably showed hyperexcitable behavior readily recognized even in the culture medium: they spontaneously swim backward for longer than tens of seconds in contrast to the occasional 1 sec or less for the uninjected controls. When tested with the Na⁺-test solution (see MATERIALS AND METHODS), they invariably swim backward continuously for 1 min or longer, distinctly different from the transient avoiding reactions of 1 sec observed in uninjected cells. Repeated experiments with intact PAK11 consistently produce this effect in numerous separate injections. Usually this hyperexcitable phenotype was observed after the first post-injection fission. In cases where the first fission was delayed for a variety of reasons, the injected cells themselves showed hyperexcitability after 12 hr. The effect has complete penetrance, having been observed in all descendant cells in each clone and at all clonal ages up to 20 fissions, but was lost after autogamy. This hyperexcitable behavioral phenotype is striking but unexpected. K⁺ channels normally pass outward currents that repolarize the membrane after depolarization. Therefore, the overexpression of K⁺ channels is expected to terminate rather than to prolong membrane depolarization (excitation).

View larger version (25K): In this window In a new window Download PPT slide   Figure 1. Diagram of injected PAK11 and PAK1 plasmids. Solid lines represent the pBluescript or the pPXV vector sequence. Hatched bars represent the ORF of PAK11 and open bars that of PAK1. Light-shaded bars represent the 5' and 3' UTRs of PAK11 and dark-shaded bars that of CAM. Open arrows represent the telomeres.

Intact PAK11 ORF in the transgene is required for the effect:
It is not necessary to release the PAK11 fragment from the vector to effect the clonal behavioral transformation. The same behavioral effect was observed by injecting pBS-PAK11 linearized with BamHI or HindIII, which left the 3-kb vector sequence still attached to the uncut PAK11 fragment (Fig 1). In contrast, pBS-PAK11 linearized with StyI or EarI, which have cut sites within the PAK11 ORF (Fig 1), was completely ineffective in the phenotypic transformation.

Artificially promoted PAK11 ORF is also effective, but frameshift PAK11 ORF is not:
The flanking sequences and UTRs of the PAK11 ORF may be involved in causing the peculiar phenotypic change described above. To test this possibility and to prevent possible multimerization of the ORF in the recipient macronucleus, we subcloned the PAK11 ORF into pPXV, a Paramecium expression plasmid with the constitutive calmodulin promoter (CAM 5' UTR) and CAM 3' UTR (HAYNES et al. 1995 ). The plasmid pPXV contains telomeric sequences and has been shown to not multimerize in the macronucleus (HAYNES et al. 1995 ; MCCORMICK-GRAHAM et al. 1997 ). Injection of empty pPXV, linearized with SfiI, did not affect the behavior of paramecia (Fig 2). On the other hand, SfiI-linearized pPXV-PAK11 induces the same hyperexcitable behavior in the clonal descendants of the injected cells as observed above (Fig 2). Molecular evidence verifies that the injected plasmids indeed replicated autonomously in the transformed macronuclei with little multimerization (see below).

View larger version (18K): In this window In a new window Download PPT slide   Figure 2. Behavioral response of transgenic Paramecium. The descendants of wild-type cells injected with empty plasmid (pPXV), frameshift pPXV-PAK11 (PAK11-fs), or in-frame pPXV-PAK11 (PAK11) were compared to uninjected wild-type cells (uninjected). The duration of continued backward swimming upon transfer from an adaptation solution into the Na+-test solution was monitored. Each value presented is the mean ± standard deviation of five cells from each of six separate clones.

A shift in the reading frame should allow us to discern whether it is the nucleotide sequence in the PAK11 RNA or the amino acid sequence of the translated PAK11 protein that causes the hyperexcitable phenotype. Accordingly, we engineered a pPXV plasmid with a single base-pair insertion downstream of the start codon of the PAK11 ORF, frameshifted (fs) pPXV-PAK11, and tested for its transformation effect. This insertion (5' ATGATAC ... to 5' ATGATAAC ...) was expected to shift the reading frame and to terminate the translation at amino acid residue 32. As shown in Fig 2, this plasmid, which was also linearized with SfiI, is entirely ineffective in producing hyperexcitability.

PAK2 does not have the same effect as PAK11 and PAK1:
While PAK1 and PAK11 are 96% identical, PAK2 and PAK1 have only 24% identity between their transmembrane amino acid sequences (JEGLA and SALKOFF 1995 ) except in the regions that are suggested to encode the channel K⁺ filter where sequence identity is high. We constructed pBS-PAK2 with a 4.5-kb insert encompassing the PAK2 ORF and its UTRs. Injection of BamHI-digested pBS-PAK2 did not generate any hyperexcitable clones. We subcloned the entire genomic PAK1 ORF into pPXV to form pPXV-PAK1. SfiI-linearized pPXV-PAK1 produced hyperexcitable clones indistinguishable from those transformed by SfiI-linearized pPXV-PAK11 (Fig 2). We also constructed a pPXV bearing a truncated version of PAK1 ORF, pPXV-PAK1 3' (Fig 1), that encodes all of the six transmembrane domains, including the filter and the pore, as well as the first 179 amino acids of the cytoplasmic C-terminal end followed by a TGA stop. Injection of a SfiI digest of this construct was just as ineffective as the SfiI digest of frameshifted pPXV-PAK11 described above (data not shown).

Fate of the injected transgenes in PAK11 and PAK1:
Our previous work demonstrated that the linearized pPXV expression plasmid and its insert consistently replicate faithfully and autonomously in Paramecium macronuclei (HAYNES et al. 1995 ; HAYNES et al. 1998 ). Here, the molecular status of the PAK transgenes and the endogenes in the transformed clones was also examined. The entire PAK11 ORF was used to probe a complete HindIII digest of total DNA extracted from the various clones. In all the phenotypically transformed clones, the major signal was at 6.3 kb (Fig 3, lanes 1, 3, and 4), matching exactly the size of the SfiI-cut in-frame and frameshifted pPXV-PAK11 plasmids originally used for injection (Fig 3, lanes 5 and 6). Within the resolution of the blot, there were no signs of deletion, insertion, or multimerization, indicating faithful autonomous replication of the linearized plasmid over the many fissions after injection. The 6.3-kb major signal is absent in DNA from uninjected clones (Fig 3, lane 2). All DNA samples have a band >12 kb in size, assumed to be the result of hybridization to the genomic DNA. This >12-kb band may represent multiple endogenes since the PAK11 probe apparently can detect both PAK1 and PAK11 sequences (see below).

View larger version (68K): In this window In a new window Download PPT slide   Figure 3. Southern analysis of microinjected clones for PAK11 endogene(s) and transgenes. Total DNAs extracted from various clones were digested with HindIII before being electrophoresed in a 0.7% agarose gel at 1 µg per lane alongside lanes loaded with SfiI-digested plasmids used in the injection experiment. The plasmids have no HindIII site. The blot was probed with random-primed 32P-labeled PAK11 ORF. Clones from cells injected with pPXV-PAK11 (lane 1), frameshifted pPXV-PAK11 (lane 3), and pPXV-PAK1 (lane 4) all have the same prominent 6.3-kb band present in the SfiI-digested pPXV-PAK11 (lane 5) and frameshifted pPXV-PAK11 (lane 6), indicating the transgenes have replicated in the descendants. All clones, including the uninjected control (lane 2), have a faint band of >12 kb.

PAK11 trangenes are transcribed and translated:
We also used the entire PAK11 ORF to probe total RNAs from the various descendant clones. A 2.4-kb hybridization signal is evident in the clones transformed with SfiI-linearized pPXV-PAK1 or pPXV-PAK11 that show the hyperexcitable phenotype (Fig 4, lanes 3 and 4, top). The same 2.4-kb Northern signal is present in the clones from cells injected with the pPXV bearing the frameshifted PAK11 ORF, which, as stated above, shows no hyperexcitability (Fig 4, lane 2, top). These signals match the expected size of the PAK1 and PAK11 RNA and indicate that the transgenes are being transcribed whether they are cloned in-frame or off-frame. Taking into consideration the loading control (Fig 4, bottom) and the Southern signals of PAK transgenes (Fig 3), the failure of the frameshifted pPXV-PAK11 to transform was clearly not because of underinjection. Clones transformed with an empty pPXV (Fig 4, lane 5, top) and those of uninjected cells (Fig 4, lane 1, top), which do not show any phenotypic changes, have no recognizable signal, 2.4-kb or otherwise. Thus, the transcripts from the endogenes are far fewer compared to those from the large number of transgenes.

View larger version (71K): In this window In a new window Download PPT slide   Figure 4. Northern analysis of microinjected clones. Total RNAs purified from various clones were loaded at 25–30 µg/lane and separated in a 1.5% formaldehyde-agarose gel. The blot was first probed with random-primed 32P-labeled PAK11 ORF (top) and then reprobed with random-primed 32P-labeled CAM ORF (bottom). Clones descended from cells injected with frameshifted pPXV-PAK11 (lane 2), in-frame pPXV-PAK1 (lane 3), and pPXV-PAK11 (lane 4) have a major band of 2.4 kb, which is not detectable in the uninjected clone (lane 1) and pPXV-injected control clone (lane 5). But all clones have a calmodulin message band of 0.5 kb.

PAK11 transformants have reduced K⁺ currents:
While swimming behaviors reflect the activities of multiple ion channels and pumps, currents through individual channels can be identified by their charge carrier, kinetics, and other biophysical characteristics. To determine which of the previously described activities might be responsible for the hyperexcitable phenotype of the pPXV-PAK11-transformed wild-type paramecia, we examined individual ion currents using a two-electrode voltage-clamp analysis. Membrane currents were examined under conditions optimal for Ca²⁺ and K⁺ conductance (see MATERIALS AND METHODS). Four principal K⁺ currents have been described to date (SAIMI and KUNG 1987 ; PRESTON et al. 1990A ). The first two are activated upon membrane depolarization and were elicited using 1500-msec steps from rest (-40 mV) to potentials between -35 mV and 0 mV. Examples of currents elicited by a step from -40 mV to -5 mV are shown in Fig 5, left. The voltage step (V_m) first elicited a rapid inward Ca²⁺ transient (asterisk in upper current trace in Fig 5), followed by an outward, voltage-gated K⁺ current (I_K(d)). The magnitude of this delayed rectifier could be approximated from the peak current at 50–100 msec (Table 1). A second, Ca²⁺-dependent K⁺ current (I_K(Ca,d)) developed slowly during step depolarizations toward a maximum at 1–1.5 sec (Fig 5, single arrow) and yielded a slowly decaying tail current upon returning to -40 mV (Fig 5, double arrows, and Table 1). We found no significant difference in the amplitude of I_K(d) between uninjected control cells and clonal descendants of cells injected with frameshift and in-frame PAK11 plasmids (Table 1). We did record a significant decrease in I_K(Ca,d) in pPXV-PAK11 transformants, however, as reflected in tail current amplitude at 20 msec (Table 1). This finding is consistent with the behavioral observation suggesting that the hyperexcitable phenotype results from suppression of the same K⁺ current that is affected in pantophobiac mutants. We also examined the two K⁺ currents that are activated specifically upon hyperpolarization. The currents shown in Fig 5 (right) were elicited by 500-msec steps to -120 mV from rest. The inward peak (asterisk) consisted largely of a hyperpolarization-induced Ca²⁺ current and a voltage-gated K⁺ current (I_K(h)). A Ca²⁺-dependent K⁺ current (I_K(Ca,h)) activated more slowly during the voltage step and was the principal contributor to the late inward current at 500 msec. The tail current elicited upon returning to -40 mV (Fig 5, double arrows) decayed biexponentially, reflecting the time course of deactivation of the two K⁺ currents. By fitting these tails to two exponentials, it was possible to determine the individual contributions of the two K⁺ currents to the total current; their amplitudes are given in Table 1. Again, we found no significant difference in amplitude of the voltage-gated K⁺ current among the three populations of cells, but the Ca²⁺-dependent current was significantly reduced by transformation with pPXV-PAK11.

View larger version (15K): In this window In a new window Download PPT slide   Figure 5. Ion currents in control and injected cells. Membrane currents were elicited under two-electrode voltage clamp from cells bathed in 1 mM K+. The top indicates the step changes in membrane potential (Vm) used to elicit the currents shown below. Traces at left were elicited using 1.5-sec steps from -40 to -5 mV. In uninjected control cells, this elicits a rapid inward Ca2+ transient (asterisk) followed within a millisecond by a voltage-dependent K+ current. A Ca2+-dependent K+ current develops more slowly during the step toward a maximum at 1.5 sec (arrow) and manifests as a prominent outward tail current that decays slowly upon a return to rest (double arrows). A similar sequence of currents is observed upon step depolarization of clonal descendants of cells injected with frameshifted pPXV-PAK11, but in-frame transformants lack the slow Ca2+-dependent K+ component noted in the controls. Note that this manifests both as a reduced outward current during the step and as loss of the corresponding outward tail. Traces at right were elicited using 500-msec steps to -120 mV. This elicits an inward peak (asterisk) that consists of both a hyperpolarization-activated Ca2+ current and a second voltage-dependent K+ current. A Ca2+-dependent K+ current develops over a period of 50–400 msec during step hyperpolarization. The outward tail current elicited upon returning to -40 mV decays biexponentially, reflecting the coincident decay of the two K+ currents (indicated by the two arrows). Both tail components are apparent in frameshifted pPXV-PAK11 transformants but descendants of in-frame transformants lack the Ca2+-dependent currents.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Transgene-induced loss of K⁺ currents is not a common "gene silencing" phenomenon:
One common strategy to study eukaryotic channels is to express them in Xenopus oocytes, which can be readily manipulated and examined using electrophysiological and other experimental techniques. Since Paramecium uses UAA and UAG to encode glutamines (PREER et al. 1985 ) and PAK1 or PAK11 each has 21 such codons, they needed to be changed before they could be expressed heterologously. To investigate their corresponding currents, we therefore expressed them as transgenes in Paramecium itself. We might have expected that overexpression of the PAK genes would enhance cell permeability to K⁺, but instead we observed a behavior reminiscent of the pantophobiac mutants that are characterized by their deficits in Ca²⁺-activated K⁺ currents (HINRICHSEN et al. 1986 ; PRESTON et al. 1990B ). Pantophobiac mutants were first isolated for their hyperexcitable behavior that correlated with the loss or reduction of the Ca²⁺-dependent K⁺ current activated upon depolarization (SAIMI et al. 1983 ) as well as that upon hyperpolarization (PRESTON et al. 1990A ). The mutations were eventually traced to the calmodulin gene (KINK et al. 1990 ; SAIMI and KUNG 1994 ). Direct measurements of the various membrane currents under voltage clamp confirmed that the Ca²⁺-dependent K⁺ currents were indeed attenuated in the PAK11 transformants (Fig 5 and Table 1). The phenomenon is robust, as PAK1 or PAK11 transgene invariably causes the dramatic hyperexcitable behavior (Fig 2) in the descendant clones that persists until the next autogamy, a process in which the transgene-bearing macronuclei are replaced by new macronuclei derived from the germ nucleus. At this level of phenotypic observation, the phenomenon resembles most cases of gene silencing in that the transgene leads to a loss of the activity of the gene product, such as the cases in Paramecium reported by RUIZ et al. 1998 . In most cases of sequence-specific gene silencing in other species currently under study, the loss of the activity is due to a reduction in the targeted mRNA (see BAULCOMBE 2000 ; COGONI and MACINO 2000 ; SHARP 2001 for reviews.) In the present studies, transgene-derived RNA was easily detected although the RNA from the endogene is below the level of detection of our hybridization (Fig 4). The fact that we can clone intron-free cDNAs of PAK11 and PAK1 endogenes by RT-PCR from uninjected wild-type cells demonstrates the presence of both specific RNA templates in Paramecium. In addition, clonal hyperexcitability is inducible by pBS-PAK11. In it, the PAK11 transgene is identical to the endogene flanked by 2 kb of flanking UTR sequences. We therefore reason that the PAK11 endogene is also transcribed in the hyperexcitable transformants. Preliminary investigation with pPXV-PAK11-green fluorescent protein fusion constructs elicited both clonal hyperexcitability and a green fluorescent signal that are absent in the uninjected controls. While this indicates that the transgene message was actually translated, further study is underway to characterize the nature of this translated product.

The effects are likely due to the channel protein:
A single nucleotide insertion at the 5' end of the ORF should not greatly affect the structure and function of RNA. On the other hand, this insertion would shift the reading frame and abolish translation. Our frameshifted pPXV-PAK11 construct produces RNA of the same size as the in-frame plasmid (Fig 4). The fact that it is entirely ineffective in producing the hyperexcitable phenotype (Fig 2) supports the hypothesis that the effect is not due simply to the presence of transgenic PAK11 RNA but to the production of the protein. The possibility that the frameshifted RNA is somehow specifically destroyed is refuted because the level of plasmid-derived PAK11-specific RNA is approximately the same in both cases (Fig 4). We also found that the PAK ORF in the plasmids needs to be intact for the hyperexcitability to occur: truncating or bisecting the ORF invariably removes the phenotypic transformation. Furthermore, the effect does not depend on the specificity of the untranslated portions of the transgenic PAK gene since the ORF placed behind its own promoter and 3' UTR or the CAM promoter and 3' UTR have the same effect. These results indicate that the hyperexcitability is the consequence of the translational product of the PAK11 transgenes in the macronucleus. The failure of the truncated PAK1 transgenes to transform indicate that the C-terminal region of this K⁺-channel subunit could be critical in assembly and/or functioning. Direct localization and quantification of PAK11 proteins await further investigation with fusion constructs and antibodies.

Meanwhile, we have used the specific electric currents through these channel proteins to follow their individual activities. Because each type of channel has its own ion permeability, gating parameters, activation and inactivation kinetics, as well as its own set of specific blockers, each can be studied separately under a voltage clamp in vivo (HILLE 1992 ). Our previous investigations have shown the loss of two types of Ca²⁺-dependent K⁺ currents in the pantophobiac mutants (PRESTON et al. 1990B ). The present studies have suggested that the presence of either PAK1 or PAK11 transgenes yield pantophobiac-like hyperexcitability by suppressing these same Ca²⁺-dependent currents (Fig 5). Since both I_K(Ca,d) and I_K(Ca,h) require an increase in the intracellular free Ca²⁺ concentration to be activated, we have to consider the possibility that overexpression of the PAK genes attenuates these currents through suppression of the Ca²⁺ currents activated upon depolarization and hyperpolarization, respectively. This is clearly not a plausible explanation for the phenotype, because suppressed Ca²⁺ currents would have decreased the duration of backward swimming and inward Ca²⁺ currents are clearly observed upon both step depolarization and hyperpolarization (asterisks in Fig 5). A complete inventory of all known membrane currents in the various transformant clones is far beyond the scope of the present work. Therefore, we cannot be certain that currents other than the four K⁺ currents listed in Table 1 are not also affected by PAK1 or PAK11 injection. We have noted variations in the amplitudes of several currents between control and injected cells, but they may prove to be statistically insignificant upon further scrutiny. Regardless, they are unlikely to be the cause of the obvious deficit in I_K(Ca,d) and I_K(Ca,h) (Table 1) and the hyperexcitable behavioral phenotype described here.

Possible mechanism(s):
There are now over 15 known ORFs, each capable of specifying one type of K⁺-channel subunit, at least conceptually, in Paramecium. If we assume that Paramecium K⁺ channels are heterotetramers, and each type of channel has a different subunit composition, the overabundance of one type of subunit from the expression of a certain PAK transgene may disrupt tetramer assembly. Misassembled nonfunctional tetramers may compete with the functional ones and perturb channel production, secretion, and deployment into limited surface sites. In addition, nonfunctional proteins may trigger degradation that removes even the functional proteins. The observation that the two Ca²⁺-dependent K⁺ channels are strongly affected by the PAK1 or PAK11 transgenes but not by PAK2 indicates that these two channels are most sensitive to ill effects of overproduction of PAK1-like subunits. It seems possible that these two channels normally use PAK1 subfamily members as some or all of their subunits and the natural proclivity of accepting these subunits into the tetramers is being abused here by the excessive abundance of one type of subunit produced by the transgene. In Drosophila, both the Sh (shaker) and Eag (ether-a-go-go) loci encode different K⁺-channel subunits. WU et al. 1983 and ZHONG and WU 1991 showed that, in larval muscles, mutant eag produces multiple types of K⁺ currents and different alleles have quantitatively different effects. Coexpression of Sh and Eag in oocytes produces K⁺ currents that are different from the expression of Sh alone, and the interaction among the different channel subunits depends on both the amount of RNA injected and the length of time after the injection (CHEN et al. 2000 ). These results were interpreted as the effects of coassembly of different subunits into the K⁺-channel tetramers. Among the superfamily of cation channels with six transmembrane -helices, PAKs are most similar to the Eag group. It therefore seems possible that they also enter heterotetrameric assembly. The C-terminal cytoplasmic domain of the Eag subunit has a cyclic nucleotide-binding site (GUY et al. 1991 ) and the Eag current expressed in oocytes is modulated by cAMP (BRUGGEMANN et al. 1993 ). One of the PAK-like genes revealed in the Paramecium genome sequencing project (DESSEN et al. 2001 ) apparently encodes a polypeptide, a portion of which has all the canonical features of the same cyclic nucleotide-binding domain. PAK1 and PAK11 polypeptides seem to have a similar domain but also have substitutions of two conserved residues in the nucleotide (MCCUE et al. 2000 ). These changes are not likely artifacts of our DNA manipulations, since they are identical in PAK1 and PAK11 and in multiple PAK11 sequences derived from different Paramecium stocks. The significance of these changes is unknown, but may be related to the suppression-of-K⁺-currents effect discussed here.

Gene silencing and reverse genetics in Paramecium:
In most cases of sequence-specific gene silencing currently under study in other species, the loss of the activity is due to a reduction in the targeted mRNA. This occurs either by reducing the amount of the gene being transcribed (transcriptional silencing) or by degrading the RNAs produced (post-transciptional silencing; BAULCOMBE 2000 ; COGONI and MACINO 2000 ; SHARP 2001 ). RUIZ et al. 1998 showed clearly that transgenes reduced gene expression in a sequence-specific manner by a post-transcriptional mechanism in Paramecium. They showed that this phenomenon can effect the expression of both unique genes as well as members in gene families. The discovery of a silencing phenomenon in Paramecium is an important development. Combined with the sequence information yielded by the ongoing Paramecium genome project (DESSEN et al. 2001 ), the physiological functions of individual gene products can be investigated in vivo. This article shows that the actual mechanism for a particular loss-of-function phenotype, induced by injection of plasmids, may not always be a reduction of transcripts. Our observation of the reduction of two separate Ca²⁺-dependent K⁺ currents upon the expression of PAK11 or PAK1 is apparently a consequence of a post-translational rather than a post-transcriptional mechanism. The results of the present work do not, however, preclude the possibility of inducing post-transcriptional silencing of the PAK genes using a different protocol. RUIZ et al. 1998 not only examined a different set of genes, but also used a different protocol. During our initial attempts to characterize the physiological role of the many homologs to metazoan K⁺ and Na⁺/Ca²⁺ channels uncovered in the Paramecium genome project, silencing will be a valuable technique. At the same time we must improve our understanding of the diversity of mechanisms to allow for more specific and effective investigations of any one particular gene.

	FOOTNOTES

¹ These authors contributed equally to this article.
² Present address: Laura Oesterle, USDA-ARS, CCRU, Barley and Malt Laboratory, 501 N. Walnut St., Madison, WI 53705.

	ACKNOWLEDGMENTS

We thank Lynn Haynes for critical comments on the manuscript. We also thank Franziska Brendt, Benjamin B. Smalley, and Brian Vaillant for technical assistance. This research was funded by National Institutes of Health grants GM-22714 (C.K.), GM-36386 (Y.S.), and GM-51498 (R.R.P.).

Manuscript received April 13, 2001; Accepted for publication July 5, 2001.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AUSUBEL, F. M., R. BRENT, R. E. KINGSTON, D. D. MOORE, J. G. SEIDMAN et al., 1999 Current Protocols in Molecular Biology. Greene Publishing Associates & Wiley-Interscience, New York.

BAULCOMBE, D. C., 2000  Molecular biology: unwinding RNA silencing. Science 290:1108-1109[Free Full Text].

BETERMIER, M., S. DUHARCOURT, H. SEITZ, and E. MEYER, 2000  Timing of developmentally programmed excision and circularization of Paramecium internal eliminated sequences. Mol. Cell. Biol. 20:1553-1561[Abstract/Free Full Text].

BRUGGEMANN, A., L. A. PARDO, W. STUHMER, and O. PONGS, 1993  Ether-a-go-go encodes a voltage-gated channel permeable to K⁺ and Ca²⁺ and modulated by cAMP. Nature 365:445-448[Medline].

CHEN, M. L., T. HOSHI, and C. F. WU, 2000  Sh and eag K⁺ channel subunit interaction in frog oocytes depends on level and time of expression. Biophys. J. 79:1358-1368[Abstract/Free Full Text].

COGONI, C. and G. MACINO, 2000  Post-transcriptional gene silencing across kingdoms. Curr. Opin. Genet. Dev. 10:638-643[Medline].

DESSEN, P., M. ZAGULSK, R. DROMADKA, H. PLATTNER, and R. KISSMEHL et al., 2001  Paramecium genome survey: a pilot project. Trends Genet. 17:306-308[Medline].

DOYLE, D. A., J. MORAIS-CABRAL, R. A. PFUETZNER, A. KUO, and J. M. GULBIS et al., 1998  The structure of the potassium channel: molecular basis of K⁺ conduction and selectivity. Science 280:69-77[Abstract/Free Full Text].

GANETZKY, B., G. A. ROBERTSON, G. F. WILSON, M. C. TRUDEAU, and S. A. TITUS, 1999  The Eag family of K⁺ channels in Drosophila and mammals. Ann. NY Acad. Sci. 868:356-369[Abstract/Free Full Text].

GUY, H. R., S. R. DURELL, J. WARMKE, R. DRYSDALE, and B. GANETZKY, 1991  Similarities in amino acid sequences of Drosophila eag and cyclic nucleotide-gated channels. Science 254:730[Free Full Text].

HAYNES, W. J., K.-Y. LING, Y. SAIMI, and C. KUNG, 1995  Induction of antibiotic resistance in Paramecium tetraurelia by the bacterial gene APH-3'-II. J. Eukaryot. Microbiol. 42:83-91[Medline].

HAYNES, W. J., B. VAILLANT, R. R. PRESTON, Y. SAIMI, and C. KUNG, 1998  The cloning by complementation of the pawn-A gene in Paramecium. Genetics 149:947-957[Abstract/Free Full Text].

HAYNES, W. J., K.-Y. LING, R. R. PRESTON, Y. SAIMI, and C. KUNG, 2000  The cloning and molecular analysis of pawn-B in Paramecium tetraurelia. Genetics 155:1105-1117[Abstract/Free Full Text].

HILLE, B., 1992 Ionic Channels of Excitable Membranes, Ed. 2. Sinauer Associates, Sunderland, MA.

HINRICHSEN, R. D., A. BURGESS-CASSLER, B. C. SOLTVEDT, T. HENNESSEY, and C. KUNG, 1986  Restoration by calmodulin of a Ca²⁺-dependent K⁺ current missing in a mutant of Paramecium. Science 232:503-506[Abstract/Free Full Text].

JEGLA, T. and L. SALKOFF, 1995  A multigene family of novel K⁺ channels from Paramecium tetraurelia. Receptors Channels 3:51-60[Medline].

KIM, C. S., J. R. PREER, and B. POLISKY, 1992  Bacteriophage lambda DNA fragments replicate in the Paramecium macronucleus: absence of active copy number control. Dev. Genet. 13:97-102[Medline].

KINK, J. A., M. E. MALEY, R. R. PRESTON, K.-Y. LING, and M. A. WALLEN-FRIEDMAN et al., 1990  Mutations in Paramecium calmodulin indicate functional differences between the C-terminal and N-terminal lobes in vivo. Cell 62:165-174[Medline].

KUNG, C., S. Y. CHANG, Y. SATOW, J. VAN HOUTEN, and H. HANSMA, 1975  Genetic dissection of behavior in Paramecium. Science 188:898-904[Free Full Text].

LEFORT-TRAN, M., K. AUFDERHEIDE, M. POUPHILE, M. ROSSIGNOL, and J. BEISSON, 1981  Control of exocytotic processes: cytological and physiological studies of trichocyst mutants in Paramecium tetraurelia. J. Cell Biol. 88:301-311[Abstract/Free Full Text].

LING, K.-Y., B. VAILLANT, W. J. HAYNES, Y. SAIMI, and C. KUNG, 1998  A comparison of internal eliminated sequences in the genes that encode two K⁺-channel isoforms in Paramecium tetraurelia. J. Eukaryot. Microbiol. 45:459-465[Medline].

MCCORMICK-GRAHAM, M., W. J. HAYNES, and D. P. ROMERO, 1997  Variable telomeric repeat synthesis in Paramecium tetraurelia is consistent with misincorporation by telomerase. EMBO J. 16:3233-3242[Medline].

MCCUE, L. A., K. A. MCDONOUGH, and C. E. LAWRENCE, 2000  Functional classification of cNMP-binding proteins and nucleotide cyclases with implications for novel regulatory pathways in Mycobacterium tuberculosis. Genome Res. 10:204-219[Abstract/Free Full Text].

PREER, J. R., JR., L. B. PREER, B. M. RUDMAN, and A. J. BARNETT, 1985  Deviation from the universal code shown by the gene for surface protein 51A in Paramecium. Nature 314:188-190[Medline].

PRESTON, R. R., Y. SAIMI, and C. KUNG, 1990a  Evidence for two K⁺ currents activated upon hyperpolarization of Paramecium tetraurelia. J. Membr. Biol. 115:41-50[Medline].

PRESTON, R. R., M. A. WALLEN-FRIEDMAN, Y. SAIMI, and C. KUNG, 1990b  Calmodulin defects cause the loss of Ca²⁺ -dependent K⁺ currents in two pantophobiac mutants of Paramecium tetraurelia. J. Membr. Biol. 115:51-60[Medline].

PRESTON, R. R., Y. SAIMI, and C. KUNG, 1992  Calcium-dependent inactivation of the calcium current activated upon hyperpolarization of Paramecium tetraurelia. J. Gen. Physiol. 100:233-251[Abstract/Free Full Text].

ROBERTSON, G. A., J. W. WARMKE, and B. GANETZKY, 1996  Potassium currents expressed from Drosophila and mouse eag cDNAs in Xenopus oocytes. Neuropharmacology 35:841-850[Medline].

RUIZ, F., L. VAYSSIE, C. KLOTZ, L. SPERLING, and L. MADEDDU, 1998  Homology-dependent gene silencing in Paramecium. Mol. Biol. Cell 9:931-943[Abstract/Free Full Text].

SAIMI, Y. and C. KUNG, 1987  Behavioral genetics of Paramecium. Annu. Rev. Genet. 21:47-65[Medline].

SAIMI, Y. and C. KUNG, 1994  Ion channel regulation by calmodulin binding. FEBS Lett. 350:155-158[Medline].

SAIMI, Y., R. D. HINRICHSEN, M. FORTE, and C. KUNG, 1983  Mutant analysis shows that the Ca²⁺-induced K⁺ current shuts off one type of excitation in Paramecium. Proc. Natl. Acad. Sci. USA 80:5112-5116[Abstract/Free Full Text].

SAMBROOK, J., E. F. FRITSCH and T. MANATIS, 1989 Molecular Cloning: A Laboratory Manuel. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

SHARP, P. A., 2001  RNA interference: 2001. Gene Dev. 15:485-490[Free Full Text].

SONNEBORN, T. M., 1970 Methods in Paramecium research, pp. 241–339 in Methods in Cell Physiology, edited by D. M. PRESCOTT. Academic Press, New York.

WU, C. F., B. GANETZKY, F. N. HAUGLAND, and A. X. LIU, 1983  Potassium currents in Drosophila: different components affected by mutations of two genes. Science 220:1076-1078[Abstract/Free Full Text].

ZHONG, Y. and C. F. WU, 1991  Alteration of four identified K+ currents in Drosophila muscle by mutations in eag. Science 252:1562-1564[Abstract/Free Full Text].

This article has been cited by other articles:

				S. H. Loukin, M. M.-C. Kuo, X.-L. Zhou, W. J. Haynes, C. Kung, and Y. Saimi Microbial K+ Channels J. Gen. Physiol., May 31, 2005; 125(6): 521 - 527. [Full Text] [PDF]

				H. Lalucque and P. Silar Incomplete Penetrance and Variable Expressivity of a Growth Defect as a Consequence of Knocking Out Two K+ Transporters in the Euascomycete Fungus Podospora anserina Genetics, January 1, 2004; 166(1): 125 - 133. [Abstract] [Full Text] [PDF]

				W. J. Haynes, K.-Y. Ling, Y. Saimi, and C. Kung PAK Paradox: Paramecium Appears To Have More K+-Channel Genes than Humans Eukaryot. Cell, August 1, 2003; 2(4): 737 - 745. [Abstract] [Full Text] [PDF]