The Cloning by Complementation of the pawn-A Gene in Paramecium

Corresponding author: Ching Kung, Laboratory of Molecular Biology and Department of Genetics, University of Wisconsin-Madison, 1525 Linden Dr., Madison, WI 53706, chung{at}facstaff.wisc.edu (E-mail).

Communicating editor: S. L. ALLEN

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The genetic dissection of a simple avoidance reaction behavior in Paramecium tetraurelia has shown that ion channels are a critical molecular element in signal transduction. Pawn mutants, for example, were originally selected for their inability to swim backward, a trait that has since been shown to result from the loss of a voltage-dependent calcium current. The several genes defined by this phenotype were anticipated to be difficult to clone since the 800-ploid somatic macronucleus of P. tetraurelia is a formidable obstacle to cloning by complementation. Nonetheless, when the macronucleus of a pawn mutant (pwA/pwA) was injected with total wild-type DNA or a fractional library of DNA, its clonal descendants all responded to stimuli like the wild type. By sorting a fractional library, we cloned and sequenced a 2.3-kb fragment that restores the Ca²⁺ current and excitability missing in pawn-A. Data from RNase protection assays, followed by the sequencing of mutant alleles and cDNA clones, established an open reading frame. The conceptually translated product suggests a novel protein that may be glycophosphatidylinositol anchored. We also discuss the general usefulness of this method in cloning other unknown DNA sequences from Paramecium that are functionally responsible for various mutant phenotypes.

Avariety of signal transduction mechanisms have been shown to provide protists with a capacity to sense and respond to chemical and physical changes in the environment. All eukaryotic cells use ion channels to mediate rapid changes within the cell in response to external stimuli. In protists, the activity of these channels often coincides with overt behavioral responses. These behavioral responses provide both a sensitive assay for changes in electrical states of the membrane and a method for isolating mutations in genes associated with signal transduction and ion channel physiology.

For decades Paramecium tetraurelia has been used to study the effect of several ion currents on the cells' swimming behavior (KUNG 1975 ; SAIMI and KUNG 1987 ; PRESTON et al. 1992B ). There were several advantages for using this particular organism to study the molecular basis of their behavioral responses. The relatively large size of the cells allowed for electrophysiological techniques to be used which showed that mutant behavior was often the direct result of changes in particular ion currents (KUNG and ECKERT 1972 ). Sexually mature cells will undergo autogamy and therefore homozygous cultures can be established in one generation after being chemically mutagenized (SONNEBORN 1970 ; KUNG 1971B ). Finally, clones could be examined by transmission genetics which showed that a variety of selectable mutant behavioral phenotypes were often the result of mutations in one of several distinct complementation groups (KUNG 1971A ; CHANG and KUNG 1973A ; CHANG and KUNG 1973B ; CHANG et al. 1974 ).

The pawn behavioral phenotype defined one of the first isolated and characterized mutations (KUNG 1971A ; CHANG and KUNG 1973A ). Pawn cells cannot swim backward in response to a depolarizing stimulus as a result of the absence of an inward voltage-dependent Ca²⁺ current (KUNG and ECKERT 1972 ). Normally, this current leads to an increased concentration of Ca²⁺ within the wild-type cell, which leads to a reversal in the power stroke of the ciliary beat and consequently backward swimming (KUNG and NAITOH 1973 ).

While mutants like pawn allowed for a thorough description of the relationship between membrane electrophysiology and swimming behavior, several qualities of Paramecium make this organism challenging to investigate using standard techniques of molecular biology. Because these cells have an 800–1000-ploid macronucleus, cloning unknown genes has been difficult. In order to clone by complementation, the phenotype produced by multiple endogenous copies must be overcome by the injected DNA. Because the genomic DNA is extremely rich in A·T base pairs (on average 65% within coding regions for 30 described genes) and these cells use only one of the three standard stop codons (TGA), researchers have often had to modify the standard molecular biology techniques used for other organisms.

One method used to circumvent these problems is to produce mass cultures, identify, purify, and eventually microsequence the proteins of interest. Degenerate oligonucleotides can be synthesized based on peptide sequences and used to clone the gene. Using this technique, the abundant and heat resistant Ca²⁺-regulatory molecule, calmodulin, was found to regulate the activity of at least two classes of ion channels (KINK et al. 1990 ; SAIMI and LING 1995 ). However, some of the mutant cell lines established in our laboratory affect proteins that have proven to be relatively difficult to biochemically identify or purify (ADOUTTE et al. 1983 ; HAGA et al. 1984 ). A second method is to design degenerate oligonucleotides based on highly conserved domains. This approach has recently led to the cloning of several potential K⁺ ion channel genes and a plasma membrane pump (JEGLA and SALKOFF 1995 ; ELWESS and VAN HOUTEN 1997 ). This method, however, requires the gene to have been described in other organisms.

Several studies have shown that plasmids injected into the macronucleus could effect a transformation of the phenotype in the clonal descendants (GODISKA et al. 1987 ; GILLEY et al. 1988 ; KANABROCKI et al. 1991 ; HAYNES et al. 1995 ). This suggested to us that one might be able to clone unknown or nonconserved genes that complement mutant defects by simply injecting wild-type genomic fragments into mutant macronuclei. We chose pawn mutants as the first candidate for such a strategy because they result from recessive mutations and they have a well-characterized and striking phenotype (KUNG 1971A ; KUNG and ECKERT 1972 ; CHANG et al. 1974 ; SATOW and KUNG 1980 ; HAGA et al. 1984 ). Previously, we showed that mutant pawn cells could be transformed with gel-purified clonable fractions of genomic DNA digested to completion with endonucleases (HAYNES et al. 1996 ). In the present study we isolate unique restriction fragments that complement pwA mutations in Paramecium. We present our analysis of a 2.3-kb fragment from a BglII digestion of wild-type genomic DNA that complements the pawn-A mutant phenotype. This fragment complements only one of three pawn complementation groups tested. Further development and improvement of cloning by complementation in Paramecium and other protists should allow for the isolation of genes that affect a variety of biological phenomena (DYNES and FIRTEL 1989 ; RYAN et al. 1993 ; ZHANG et al. 1994 ; VASHISHTHA et al. 1996 ; WILSON and SEEBECK 1997 ). By combining the ability to select for mutant phenotypes with complementation cloning and the characterization of the gene products, protists should provide new and fundamental information about the nature, regulation, and evolution of eukaryotic signal transduction mechanisms.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Stocks and cultures:
P. tetraurelia stock 51s (+/+) (SONNEBORN 1970 ), nd6 (nd6/nd6) (LEFORT-TRAN et al. 1981 ), pawn-A {d4-94} (both nd6/nd6, pwA/pwA and +/+, pwA/pwA), pawn-A {d4-132} (+/+, pwA/pwA), pawn-A {d4-513} (+/+, pwA/pwA), pawn-B {d4-95} (nd6/nd6, pwB/pwB), and pawn-D (nd6/nd6, pwD/pwD; KUNG 1971A ; KUNG 1971B ; Y. SAIMI, unpublished results) were cultured at 20° or 28° in a growth medium of buffered wheat-grass extract inoculated with Enterobacter aerogenes (SONNEBORN 1970 ). Cells with the nd6 mutation are behaviorally normal but unable to fire their trichocysts. Lacking the discharge reaction, these mutant cells survive the trauma of macronuclear microinjection better than wild type. Therefore all the cell lines used for microinjection were homozygous for the nd6 mutation. These cells were considered wild type for behavioral and other analyses.

Preparation of wild-type genomic samples for pawn injections:
Standard molecular biology techniques were used (SAMBROOK et al. 1989 ; AUSUBEL et al. 1993 –97). Genomic DNA was obtained as previously described (HAYNES et al. 1995 ) and digested with BclI, BglII, HindIII or XbaI to completion as judged by gel electrophoresis and the constancy of small fragments over time. Each total genomic DNA sample used for injection was bound to Wizard DNA clean-up resin (Promega, Madison, WI), washed in columns, and eluted with double-distilled water. The samples were then precipitated and resuspended in 10 mM Tris (hydroxymethyl)-aminomethane hydrochloride and 1 mM ethylenediaminetetraacetic acid, pH 8.0 (TE) at >1 µg/µl. Restriction digests were electrophoresed and fractionated using agarose gels (0.5% Sea Plaque GTG; FMC, Rockland, ME), where overloading of the gel was carefully avoided. Fractions were excised from the gel and extracted with Agarase (Epicentre Technologies, Madison, WI). This process was repeated until several micrograms of DNA were obtained for each of the fractions. Samples were finally precipitated and resuspended in TE at 1–5 µg/µl for direct injection.

Preparation of RNA for Paramecium:
Cells were harvested and washed in Dryl's solution (DRYL 1959 ). Approximately 250 µl aliquots of cells (2 x 10⁵ cells) were lysed in 750 µl of TRI-REAGENT (Molecular Research Center, Cincinnati, OH). The protocol for RNA isolation, which was provided by the company, included the following additional steps. Prior to the addition of chloroform to separate the aqueous and organic phases, the tubes were centrifuged and the supernatant transferred to a new tube. Isopropanol (1/10 volume) was added and the tubes centrifuged to precipitate some of the contaminating DNA. In addition a salt precipitation was also done to eliminate any potential contaminating proteoglycans. RNA was sometimes further purified using a PolyATtract mRNA isolation system (Promega).

Microinjection:
Five to 10 pl of DNA solutions at various concentrations were injected into the macronucleus of each recipient cell as previously described (KANABROCKI et al. 1991 ; HAYNES et al. 1995 ; HAYNES et al. 1996 ). Recipients were injected between two and six fissions after autogamy. Each DNA sample was injected into at least six cells. The descendants of the individual recipients were cultured as individual clones.

Behavioral assay:
The pawn mutant cells injected with wild-type genomic DNA or plasmids carrying genomic inserts were cultured for four to seven fissions before their behavior was tested. Cells were incubated in adaptation solution (4 mM KCl, 1 mM CaCl₂, 1 mM HEPES, 0.01 mM EDTA, pH 7.2) for 10 min and then individually transferred into a K⁺-test solution (30 mM KCl, 1 mM CaCl₂, 1 mM HEPES, 0.01 mM EDTA, pH 7.2). The duration of continuous backward swimming of each cell immediately upon transfer was monitored using a stereomicroscope and recorded (KUNG 1975 ).

Cloning of the transforming factor:
A gel-purified fraction of Paramecium DNA digested with BglII which transformed the behavior of the clonal descendants from the injected mutant cells was incubated with BamHI methylase and cloned into the BamHI site of pBluescript II KS (-) (Stratagene, La Jolla, CA). The ligations were sabotaged with BamHI, to eliminate self-ligated plasmids, before transforming electrocompetent Sure cells (Stratagene). Resistant bacterial clones were screened by preparing plasmid DNA from groups of bacterial colonies and microinjecting into Paramecium. Duplicate plates and nitrocellulose lifts were used to help sort the colonies until an individual colony was isolated (Figure 2).

View larger version (27K): In this window In a new window Download PPT slide   Figure 1. —The duration of backward swimming of microinjected cells. The descendants of wild-type (nd6/nd6) cells injected with the empty pBluescript II KS (-) plasmid (Ø) were compared with those of pawn-A cells injected with empty plasmid (Ø) or total Paramecium DNA digested with various restriction enzymes. The interval of backward swimming was recorded upon transfer to a K+-test solution after 10 min in adaptation solution (mean ± SD; n 60 cells; 10 cells from each of six separately injected clones).

View larger version (26K): In this window In a new window Download PPT slide   Figure 2. —A procedure for cloning the pwA-complementing DNA. After two rounds of gels fractionation (upper left), a complete BglII digestion of total DNA from Paramecium was cloned into BamHI-digested pBluescript II KS (-) (pB). Using nitrocellulose lifts (bottom left) a transforming sector (center right) was isolated and, after subdividing groups of colonies, a single transforming plasmid was isolated (bottom right). The transforming activity was followed by microinjecting cells with DNA purified from each step and testing for a behavioral transformation in the clonal descendants (as in Figure 1).

Electrophysiology:
The techniques used to record Ca²⁺ currents from Paramecium with a two-electrode voltage clamp have been described previously (PRESTON et al. 1992A ). Briefly, membrane potential was clamped using intracellular capillary microelectrodes filled with 3 M CsCl. Cells were incubated in adaptation solution for 10 min before being placed into a standard solution of 10 mM tetraethylammonium chloride, 0.25 mM Ca(OH)₂, 0.75 mM CaCl₂, 0.01 mM EDTA, an 1 mM HEPES, pH 7.2 for electrical recording. Currents were elicited by stepwise increases in membrane potential from -40 mV and were filtered at 2 kHz. Leakage currents were estimated from the membrane response to small (3–12 mV) hyperpolarizing steps from rest and were subtracted from active membrane responses prior to analysis.

Sequencing and RT-PCR of pwA alleles:
Sequencing of the original complementing genomic fragment and subsequent subcloned PCR products was done using a Prism sequencing kit (PE Applied Biosystems, Foster City, CA). All PCR reactions were done in a Programmable Thermal Controller 100 (MJ Research Inc., Watertown, MA) using Taq polymerase (Promega) and several oligonucleotide primers (Operon Technologies Inc., Alameda, CA). PCR of the 1.6-kb (1585 bp) genomic fragment was for 90 sec at 94°, then 60 sec at 47°, followed by 120 sec at 72° using the following two oligonucleotides: A = TCATGGGAGGATCTGGTATG; B = TTCTTCGTTTATTAAGGTACTTTA. RT-PCR of the pwA cDNA used various cycle times and temperatures similar to the previous reaction with the following oligonucleotides (base pairs are in reference to the distance from the putative starting methionine): 1 = TAAGTATATTGTAATTTGGCATCGTGA (sense 16–42 bp); 2 = AATTACTTGCGAACAATATTATCACG (sense 267–292 bp); 3 = CAGAATATGATAAAAAAGCCAAAGCCAAC (sense 296–324); 4 = GATCAAATGCGATTTTAAATTCATATTA (antisense 525–552); 5 = ACAGTGATCCTTAACTATATTTGTTTTTATGAT (antisense 550–582 bp); 6 = TTTAAGGACATCTCCAAAACAGTG (antisense 577–600 bp; 63 bp from stop); 7 = AAACATCCTTTTTCTATATTTTCTATAATC (antisense 658–688; 144 bp downstream of stop codon).

RNase protection assays:
Total or (oligo)dT purified RNA was hybridized to riboprobes polymerized from various subcloned fragments of the original transforming plasmid using either T7, T3, or SP6 polymerase and incorporating either [³²-P] CTP or UTP as described in Current Protocols (AUSUBEL et al. 1993 –97). Due to the A·T richness, the hybridizations were done at high temperature (60–70°) in a high salt buffer (10 mM Tris-HCl, 1.2 M NaCl, and 5 mM EDTA, pH 7.5; BREWER and ROSS 1990 ). RNase ONE (Promega) was used to digest the unprotected probe. Sequencing gels with a known DNA ladder (G, A, T, and C dideoxy sequencing reactions) were used to size the protected fragments to within several bases. The radioactive signals were digitized and analyzed with a Phospho-Imager (Molecular Dynamics, Sunnyvale, CA).

Sequence comparison and secondary structure prediction:
The protein and nucleotide sequence from the expected open reading frame (ORF) were used to search for homologues in the most recent databases employing several different algorithms (BLAST, BLASTP, BEAUTY, BLITZ, FASTA, FASTA-SWAP, FASTA-PAT, MPSRCH, PROPSEARCH). Additional searches were done with the percentages of amino acids using the program PROPSEARCH (European Molecular Biology Laboratory, Heidelberg, Germany). Periodically, CD-ROM recorded databases (DNAStar, Madison, WI) were also searched. A statistical analysis of the protein was performed using the SAPS program (Stanford University, CA). Secondary structure was analyzed using methods available in the programs PROTEAN (DNAStar), PHD (European Molecular Biology Laboratory), PSA (Biomolecular Engineering Research Center), PSSP (Baylor College of Medicine, Houston, TX) and COILS (Swiss Institute for Experimental Cancer Research, Epalinges, Switzerland). Searches for potential signal sequences and domains were done using PROSITE (University of Geneva, Switzerland), PSORT (National Institute for Basic Biology, Osaka, Japan), and BLOCKS (Fred Hutchinson Cancer Research Center, Seattle, WA).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Isolation of the pawn-A transforming fragment:
Total DNA harvested from wild-type cells was partially digested with SauIIIA and microinjected into the macronucleus of pawn mutant cells at approximately 5 µg/µl. Pawn mutants are characterized by an inability to swim backward, but several clonal descendants of these injected cells responded to a K⁺-induced depolarization by swimming backward for several seconds (Y. SAIMI and R. R. PRESTON, unpublished results). The success of these initial experiments suggested that pawn mutations could be complemented by the injected wild-type DNA fragments. We then tested total wild-type genomic DNA that had been digested to completion with one of four restriction enzymes for the ability to transform. DNA digested with HindIII, XbaI, and BglII retained the ability to transform, but DNA digested with BclI did not (Figure 1). The strongest transforming DNA (BglII digest in Figure 1) was next separated into four size fractions by agarose gel electrophoresis (Figure 2, upper left). When these fractions were microinjected, one fraction alone contained the transforming activity (Figure 3). This 1.5–3-kb fraction was further separated into three fractions and the activity was followed to a subfraction containing 2–2.5-kb fragments. This fraction was then methylated with BamHI methylase and ligated into plasmids linearized with BamHI endonuclease (Figure 2, center left). Bacteria were then transformed with the remaining intact plasmids and incubated at 37° for approximately 40 min before being frozen at -80° in 15% glycerol. Small aliquots of the frozen stocks were plated out and the number of colony-forming units per microliter was determined. Aliquots containing 500 colonies of the transformed bacteria were grown on individual plates under selective conditions. A large reaction (>60%) of the bacterial colonies appeared to contain plasmids with inserts as determined by blue-white selection and gel electrophoresis. The colonies from each plate were lifted onto nitrocellulose and grown in Luria-Bertani broth for one hour before the plasmid DNA was isolated by alkaline lysis. In our serial test, the plasmid DNA prepared from one of the first eight plates contained the pawn-A transforming activity. Colonies from the transforming plate were again lifted onto a nitrocellulose filter. The nitrocellulose filter was cut into eight sectors. Plasmid DNA was isolated from a liquid culture of each individual sector and separately injected (see Figure 2, middle right). All 48 colonies from the only transforming sector were individually handpicked onto fresh plates. Plasmids to be used in injections were prepared from liquid cultures each inoculated with 10 individual colonies. Separate plasmid preparations were finally made from the only group that transformed and a single transforming plasmid with a 2.3-kb insert was isolated and sequenced. The plasmids isolated from this colony, called pPwnA, were further tested in a dilution series (Figure 4). The number of copies needed to transform, as anticipated from the crude fractions, appeared to conform to the assumed number of copies in the genomic digests that we were originally injecting (Figure 4; see PREER 1986 ). We also injected the plasmid into mutant cells from two additional separate complementation groups, pwB and pwD (pwC is temperature sensitive and was not tested). The insert had no effect on the behavioral phenotype of these mutants (Figure 4).

View larger version (23K): In this window In a new window Download PPT slide   Figure 3. —Fractionating the pwA-transforming activity by gel electrophoresis. The descendants of wild-type cells injected with empty plasmid (Ø), compared to pwA cells injected with empty plasmid (Ø) or fractions of DNA after the first round of fractionation by agarose gel electrophoresis as diagramed in Figure 2. The transforming activity was unambiguously found to be in fraction 3 (fraction 1 = 5–12 kb; 2 = 3–5 kb; 3 = 1.5–3 kb; 4 = 0.7–1.5 kb). The duration of backward swimming as tested in K+-test solution (see Figure 1 and MATERIALS AND METHODS) (mean ± SD; n 60 cells; 10 cells from each of six separately injected clones).

View larger version (27K): In this window In a new window Download PPT slide   Figure 4. —The cloned pwA-transforming fragment is specific and does not transform pwB or pwD. The descendants of three complementation groups of pawn mutations compared to wild-type cells after injection with the transforming plasmid at various concentrations. Plasmid DNA (5–10 pl) injected at 5300 copies/pl effected a full rescue of pwA but not pwB or pwD. Even at 50 copies/pl transformation is evident for pwA by the duration of backward swimming.

Restoration of the voltage-dependent Ca²⁺ current by pPwnA:
As mentioned earlier, the inability of pawn mutants to swim backward is due to the loss of an inward transient Ca²⁺ current. Electrophysiological examination of the clonal descendants of pawn-A cells injected with the plasmid, pPwnA, clearly shows that the inward Ca²⁺ current had been restored (Figure 5A). The magnitude of the peak current is plotted against membrane potential in Figure 5B. While the transformed cells express a significantly smaller peak current than the wild type, the voltage dependence of the inward peak is similar.

View larger version (30K): In this window In a new window Download PPT slide   Figure 5. —Ca2+ currents in wild-type and mutant paramecia examined under a two-electrode voltage clamp. (A) Currents elicited from pawn-A, transformed pawn-A, and wild-type cells in Ca2+-TEA solution. Currents were elicited by 20-msec voltage steps to -10 mV from -40 mV. The voltage-activated Ca2+ current is apparent as an inward peak during step depolarization in the wild type and in transformants, but not in pawn-A mutant cells. The capacitive transients have been suppressed. (B) Peak inward Ca2+ currents (ICa peak) are plotted against the membrane potential at which they were elicited (Vm). Points are means ± SD determinations from three wild-type (), three pwA mutants () or eight cells descended (10 to 15 fissions after injection) from plasmid-injected pwA (). The injected plasmid was at a concentration of 530 copies/pl.

Identification the ORF and putative product:
pPwnA was digested with restriction enzymes and smaller fragments were subcloned. We found that a subcloned 1.6-kb HindIII fragment still transformed the pwA cells but not when digested with PstI or BclI (Figure 6). The presence of a BclI site in the pwA coding sequence correlates with our observation that BclI destroys the transforming factor in total DNA (see Figure 1). A Northern blot of total RNA showed a weak signal that was approximately 700–1000 bp long using a 432-bp probe (data not shown; see Figure 6). The location of the ORF and the direction of transcription was determined by RNase protection assays. Riboprobes polymerized from several plasmids containing smaller subcloned fragments of the 1.6-kb HindIII fragment showed that a species of RNA molecule in both total and oligo dT purified RNA was protected when probes were polymerized in one direction but not the other (Figure 6). The sequence protected from RNase digestion began approximately 16 bp upstream of a potential starting methionine (Figure 6; filled arrows) and the assay showed that an intron, similar in size and sequence to other known Paramecium introns, was spliced out of the middle of the mRNA (Figure 6, gap in middle arrow; RUSSELL et al. 1994 ). It is possible that an intron located near the 5' end of the mRNA would allow for the protection of a fragment too small to be detected by the assay (80 bases). However, all of the potential introns based on the consensus sequence and length would have either shifted the frame to ones with premature stop codons or would have resulted in substantially longer fragments than those detected by our probes.

View larger version (19K): In this window In a new window Download PPT slide   Figure 6. —A diagram of the pwA locus. The hatched bar marks the location of the northern probe relative to the smallest transforming fragment tested. Two endonuclease cutting sites that eliminated the ability to transform are also represented (BclI and PstI). The arrows represent the location and direction of polymerization of the riboprobes used in RNase protection assays. The position of the intron is indicated by an unprotected gap in one of these riboprobes. Several additional probes not shown were also tested to help verify the protected region. The lower gray bar corresponds to the deduced ORF of the pwA gene; its sequence is shown in Figure 7.

RT-PCR products from (oligo)dT purified RNA were cloned and sequenced. All the cDNAs cloned and sequenced confirmed the lack of the intron sequence. The two longest cDNAs had a stop codon in frame with the putative start codon (primer pairs 1 and 4 or 1 and 5; see MATERIALS AND METHODS and Figure 7). Two additional oligonucleotides based on the genomic sequence further downstream from the putative stop codon consistently failed to produce RT-PCR products (see Figure 7 for primer 6 and MATERIALS AND METHODS for primer 7). However, they did polymerize products when plasmids containing the 1.6-kb genomic sequence were introduced into the RT-PCR reactions at low concentrations (data not shown). The RT-PCR products combined with the RNase protection data allow for only one interpretable ORF and conceptually translated product (Figure 7). Computer searches for similarity to other proteins and hypothetical translations as suggested by the pwA ORF failed to reveal any definitive homology with any other known protein (see MATERIALS AND METHODS).

View larger version (33K): In this window In a new window Download PPT slide   Figure 7. —The sequence and translation of the putative ORF of the pawn-A gene (GenBank accession no. AF050753). The intron sequence is presented along with the location of several oligonucleotides used to generate cDNA. The primer pairs 1 and 4, or 1 and 5 generated cDNA with the intron removed along with the putative stop codon. Primer 6 along with another oligonucleotide (see MATERIALS AND METHODS) further downstream generated only genomic clones. The location of a central highly charged hydrophilic region () and the terminal hydrophobic domains () are marked by bars. The base farthest 5' of the putative start codon preserved in the RNase protection assay is indicated by an arrow (). A fortuitous SwaI site located in the intron, used to distinguish genomic from cDNA PCR products is noted. Two additional restriction sites marked correspond to endonucleases which eliminated the ability of the injected DNA to transform (BclI and PstI).

Sequencing of mutant alleles:
Total DNA was prepared from three separate pwA mutant cell lines (d4-94, d4-132 and d4-513). These allelic variants were kept in homozygous cell lines in our laboratory since they were produced in three independent N-methyl-N'-nitro-N-nitrosoguanidine mutagenesis experiments (CHANG and KUNG 1973A ). The products from at least three independent PCR reactions using oligonucleotides matching the ends of the 1.6-kb HindIII fragment were cloned from each cell line. A minimum of three clones from each of the PCR reactions were sequenced. All the clones consistently showed that each allele had only a single point mutation in the region of the ORF. The fact that each of the three pawn-A isolates has a substitution in this ORF strongly suggests that the complementations observed are not extragenic (G383A in d4-94; C365T in d4-132; A287G in d4-513). The location and amino acid substitution predicted by the mutations (G128D in d4-94; P122L in d4-132; Y94C in d4-513) are indicated on a Kyte-Doolittle hydropathy plot averaged over a window of seven amino acids (Figure 8). The significance of particular secondary structures predicted by various algorithms is difficult to evaluate given the lack of similarity to other known proteins. However, the hydrophilic region preceding two of the mutations is a highly charged stretch of amino acids (11 out of 21 from H 97 to K 117; see Figure 7 and Figure 8). The fact that the two clustered mutations are of amino acids that typically influence flexibility and secondary structure suggests that this region may be important for the function of this protein.

View larger version (15K): In this window In a new window Download PPT slide   Figure 8. —A Kyte-Doolittle hydrophilicity plot (averaged over seven amino acids). The position and the deduced amino acid substitutions found in the three independently isolated pawn mutant alleles are indicated. The highly charged region 5' of two of the pawn mutant alleles is also represented ().

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

This paper presents the first successful cloning of an ion-channel regulatory gene by complementation in Paramecium. It is a result of a number of preliminary investigations that indicated this method might work (GODISKA et al. 1987 ; GILLEY et al. 1988 ; KANABROCKI et al. 1991 ; ENDOH et al. 1995 ; HAYNES et al. 1995 ). The technique developed in our laboratory had also been used successfully to clone an unrelated gene (SKOURI and COHEN 1997 ). We have shown that by microinjecting cloned fragments of wild-type Paramecium DNA, a genomic fragment can be isolated that functionally complements the pawn mutation (HAYNES et al. 1996 ). Because the cloned fragment in pPwnA complements pwA and not pwB or pwD and since there are base substitutions present in the same ORF of three pwA allelic variants, the transformation is unlikely to be the result of extragenic complementation. The electrophysiology confirms that the behavioral transformation is associated with the return of an inward Ca²⁺ current (Figure 5A). The voltage-sensitivity of the restored current is almost identical to that of the wild-type current (Figure 5B). This argues against the possibility that backward swimming is regained in pawn-A transformants through expression of a novel Ca²⁺ current of over-expression of another Ca²⁺-permeable conductance. While the transformed pawn-A cells show a reduction in current compared with wild-type values, a full investigation of this reduction is beyond the scope of this article. The possible reduction in the transformed current does not conflict with behavioral observations (Figure 1). Previous observations of wild-type cells indicates that backward swimming duration is not strictly correlated with Ca²⁺-current magnitude (R. R. PRESTON, unpublished observations), although this had not been investigated systematically. The size of a Northern blot signal, the results from several repeated RNase protection assays, as well as the sequences of several cDNA clones are all consistent with the ORF and putative translation product shown in Figure 7. While it is possible that the transcript was not fully protected at the very end of the molecule in the RNase protection assay, it is unlikely that there would be a substantial change in the ORF and most of the translated amino acid sequence of the putative product would remain unchanged.

The analysis of the pawn-A product is limited by the fact that there is no significant primary sequence homology with any currently described or hypothetically translated protein. The amino acid sequence has a predicted molecular weight of 23.5 kD; a pI of 4.36; and a relatively high percentage of cysteines (5.9%) and tyrosines (9.8%) (95% and 99% quantile, respectively, by the SAPS algorithm; BRENDEL et al. 1992 ). The hydropathy plot suggests a possible association with membrane (KYTE and DOOLITTLE 1982 ). While the disulfide bridges alone could link this molecule to another membrane-bound molecule, the amino- and carboxy-terminal hydrophobic domains are also long enough for some types of membrane spanning structures. Additional empirical information of this protein is provided by the mutations found in the three pawn-A allelic variants. Two of the mutations indicate that flexibility and secondary structure may be important in the middle of the protein (G128D in d4-94; P122L in d4-132). This is in a region immediately downstream of a highly charged sequence predicted to be alpha helical by several different algorithms (see MATERIALS AND METHODS). The other allele substitutes a cysteine for a tyrosine (Y94C in d4-513) suggesting that the substitution might affect secondary or tertiary structure, perhaps involving disulfide bridges or beta structures.

Speculation on the biological role of this protein can be guided by a summary of the empirical data from earlier studies on various pawn-A mutant cell lines. Electrophysiology of leaky pawn-A mutants suggested that these mutations influenced the number of functional voltage-dependent Ca²⁺ channels present in the membrane (SATOW and KUNG 1974 ; SATOW and KUNG 1980 ). Transfer of cytoplasm by microinjection from a wild-type cell to a pawn-A mutant cell resulted in the temporary restoration of a voltage-dependent Ca²⁺ current (HAGA et al. 1982 ). It was later shown that the transferred curing agent was likely a protein (not DNA or RNA) and that the injected protein could effect a transformation without further protein synthesis (HAGA et al. 1982 ). Biochemical fractionation of protein from the soma of Paramecium suggested that the transforming element was abundant in a microsomal fraction, while protein fractions prepared from cilia and ciliary membrane vesicles could not transform cells (HAGA et al. 1984 ). These results do not rule out the possibility that a fraction of the transforming protein was targeted to the ciliary membrane and that post-translational modifications eliminated the ability to transform. It was also reported that the transforming pawn-A factor was inactivated by both elevated temperatures and by treatment with N-ethyl maleimide which suggested the presence of modifiable thiol groups (HAGA et al. 1984 ). Their conclusion from these studies was that the transforming protein was membrane bound and involved in the assembly, transport, or functional expression levels of an ion channel required for the described voltage-dependent Ca²⁺ current.

Auxiliary membrane-bound subunits currently known to be associated with the pore-forming subunit of voltage-dependent ion channels in other organisms have some of these same characteristics (CATTERALL 1996 ; GURNETT and CAMPBELL 1996 ). Both the Ca²⁺ channel and Na⁺ channel auxiliary membrane-bound subunits have been shown to be sensitive to the modification of thiol groups with N-ethyl maleimide, and they are also involved with the expression levels of these channels in the plasma membrane. However, unlike the pwA gene product, all of these membrane-bound subunits have long hydrophobic regions bounded by hydrophilic domains. Other genetic based investigations using Drosophila and Caenorhabditis have yielded subunits similar to those already found by homology or biochemical means (SCHAFER and KENYON 1995 ; MARYON et al. 1996 ), as well as additional molecules that do not have strong homology to previously cloned subunits (e.g., TipE; WARMKE et al. 1997 ); but all of them have generally had predicted alpha helical transmembrane domains. The product from the pwA gene has hydrophobic domains but does not have the flanking hydrophilic regions generally associated with transmembrane domains. Instead, the locations of the hydrophobic domains of the pwA protein are indicative of a class of proteins found in the secretory pathway known as glycophosphatidylinositol (GPI)-anchored proteins.

The PSORT algorithm (National Institute for Basic Biology, Osaka, Japan) predicts that the hydrophobic domains of the pwA protein may be amino- and carboxy-terminal sequences of GPI-anchored proteins including characteristic amino acids at the putative cleavage sites (NAKAI and KANEHISA 1992 ; UDENFRIEND and KODUKULA 1995 ). This possible interpretation correlates well with the fact that a search of protein databases using simply the percentage of amino acids found in this translation generally matches proteins that are either secreted or expressed in the plasma membrane (PROPSEARCH; EMBL; HOBOHM and SANDER 1995 ). A similar search with the only extensively studied surface membrane protein in Paramecium, the GPI-anchored immobilization surface antigen, yields a similar list of plasma membrane proteins or proteins in the secretory pathway (W. J. HAYNES, unpublished results). The existence of a GPI anchor on the pawn-A protein would likely mean that it is eventually expressed on the surface of the cell. This is appealing given the fact that all of the major ciliary proteins in Paramecium are apparently GPI anchored (CAPDEVILLE and BENWAKRIM 1996 ) and since the channels responsible for the voltage-dependent Ca²⁺ current are apparently localized on the cilia (MACHEMER and ECKERT 1973 ; MACHEMER and OGURA 1979 ; OGURA and TAKAHASHI 1976 ; DUNLAP 1977 ).

It is interesting that a previous study showed that a polyclonal antisera against a GPI-anchored surface immobilization antigen had a significant effect on the voltage-dependent Ca²⁺ current (RAMANATHAN et al. 1983 ). It is possible that the polyclonal antibodies used in their study were cross-reactive with other GPI-anchored proteins since several 40-kD proteins now thought to be GPI anchored were also immunoprecipitated by the same antisera (EISENBACH et al. 1983 ; CAPDEVILLE and BENWAKRIM 1996 ). A protein of approximately 19 kD was also immunoprecipitated with the same antisera and this closely corresponds to the expected 18.8-kD size of a processed pwA GPI-anchored protein (EISENBACH et al. 1983 ). A recent report of another potential GPI-anchored protein, a serine protease, influencing the function of an amiloride-sensitive Na⁺ channel now establishes a precedent for the possibility that there is a group of uncharacterized GPI proteins that modify or influence ion channel expression and function (VALLET et al. 1997 ). We are now attempting to biochemically characterize the protein with both antibodies and fusion constructs.

Although the conceptual translation does not match the primary sequence of any other protein, it is possible that the pwA product is a member of a known group of proteins that have secondary structural requirements that do not involve the conservation of primary sequence. After assuming that this protein could be GPI anchored, a close inspection of most of the known GPI proteins revealed a group of mono(ADP-ribosyl)transferases (mADPRTs) that do not require extensive primary sequence conservation (BAZAN and KOCH-NOLTE 1997 ). Many of the secondary structural requirements described among mADPRTs are predicted to be present in the pwA product by the Garnier algorithm including a very conserved catalytic glutamic acid residue (see Figure 7; E 152; GARNIER et al. 1978 ). This glutamate is predicted to be at the amino-terminal end of a second pair of beta structures in the carboxy-terminal half of the molecule exactly as in all other known mADPRTs (BAZAN and KOCH-NOLTE 1997 ). Furthermore these mADPRT proteins have been implicated in calcium-mediated signal transduction (ROBINSON 1997 ).

Besides the pwA mutation, there are three other known pawn loci. We have already started to isolate two of these additional genes by complementation cloning. A similar group of mutations exists in Paramecium caudatum (CNR loci) which can probably be cloned, since microinjection of digested genomic DNA can cure the CNR mutants (ENDOH et al. 1995 ). Injection of cytoplasm between these mutants has shown that all interspecies and intraspecies cytoplasms complement (HAGA et al. 1983 ). This suggests that there may be as many as eight separate genes that can be cloned by this method.

The technique presented here has allowed us to clone a novel gene, which when mutated greatly reduces or prevents the expression of a voltage-dependent Ca²⁺ current in a unicellular organism. The most important finding is that at least some of the genes responsible for the many known and well-characterized Paramecium behavioral mutants can be cloned by this method. While there are many reasons why a particular gene might not be cloned by this technique and would require other established strategies, e.g., cloning by protein purification and microsequencing, or cloning by homology, we believe that effective cloning by complementation combined with the sensitivity of the behavioral assay and the power of genetics will allow us to discover novel elements involved with ion channel signal transduction and a variety of other biological phenomena in Paramecium.

	ACKNOWLEDGMENTS

We thank KIT-YIN LING and LYNN HAYNES for their comments on the manuscript. We also thank the other members of our lab who indirectly contributed to this work. This research was founded by National Institutes of Health grants GM-22714, GM-36386, and GM-51498 to R.R.P.

Manuscript received December 11, 1997; Accepted for publication February 26, 1998.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ADOUTTE, A., K.-Y. LING, S. CHANG, F. HUANG, and C. KUNG, 1983  Physiological and mutational protein variations in the ciliary membrane of Paramecium. Exp. Cell Res. 148:387-404[Medline].

AUSUBEL, F. M., R. BRENT, R. E. KINGSTON, D. D. MOORE, J. G. SEIDMAN et al., 1993–97 Current Protocols in Molecular Biology. John Wiley & Sons, New York.

BAZAN, J. F. and F. KOCH-NOLTE, 1997  Sequence and structural links between distant ADP-ribosyltransferase families. Adv. Exp. Med. Biol. 419:99-107[Medline].

BRENDEL, V., P. BUCHER, I. NOURBAKHSH, B. E. BLAISDELL, and S. KARLIN, 1992  Methods and algorithms for statistical analysis of protein sequences. Proc. Natl. Acad. Sci. USA 89:2002-2006[Abstract/Free Full Text].

BREWER, G. and J. ROSS, 1990  Messenger RNA turnover in cell-free extracts. Methods Enzymol. 181:202-209[Medline].

CAPDEVILLE, Y. and A. BENWAKRIM, 1996  The major ciliary membrane proteins in Paramecium primaurelia are all glycosylphosphatidylinositol-anchored proteins. Eur. J. Cell Biol. 70:339-346[Medline].

CATTERALL, W. A., 1996  Molecular properties of sodium and calcium channels. Bioenerg. Biomembr. 28:219-230[Medline].

CHANG, S. and C. KUNG, 1973a  Genetic analysis of heat sensitive pawn mutants of Paramecium aurelia.. Genetics 75:49-59[Abstract/Free Full Text].

CHANG, S. Y. and C. KUNG, 1973b  Temperature-sensitive pawns: conditional behavioral mutants of Paramecium aurelia.. Science 180:1197-1199[Abstract/Free Full Text].

CHANG, S. Y., J. VAN HOUTEN, L. J. ROBLES, S. S. LUI, and C. KUNG, 1974  An extensive behavioral and genetic analysis of the pawn mutants in Paramecium aurelia.. Genet. Res. 23:165-173[Medline].

DRYL, S., 1959  Effect of adaptation to the environment on chemotaxis of Paramecium caudatum.. Acta. Biol. Exp. 19:83-93.

DUNLAP, K., 1977  Localization of calcium channels in Paramecium caudatum.. J. Physiol. 271:119-133[Abstract/Free Full Text].

DYNES, J. L. and R. A. FIRTEL, 1989  Molecular complementation of a genetic marker in Dictyostelium using a genomic DNA library. Proc. Natl. Acad. Sci. USA 86:7966-7970[Abstract/Free Full Text].

EISENBACH, L., R. RAMANATHAN, and D. L. NELSON, 1983  Biochemical studies of the excitable membrane of Paramecium tetraurelia. IX. Antibodies against ciliary membrane proteins. J. Cell Biol. 97:1412-1420[Abstract/Free Full Text].

ELWESS, N. L. and J. L. VAN HOUTEN, 1997  Cloning and molecular analysis of the plasma membrane Ca2+-Atpase gene in Paramecium tetraurelia.. J. Eukaryot. Microbiol. 44:250-257[Medline].

ENDOH, H., K. MIKAMI, M. HORI, and M. TAKAHASHI, 1995  Injection of total genomic DNA and restoration of wild type phenotype in trichocyst non-discharge mutant TND of Paramecium caudatum.. Jpn. J. Genet. 70:633-642.

GARNIER, J., D. J. OSGUTHORPE, and B. ROBSON, 1978  Analysis of the accuracy and implications of simple methods for predicting the secondary structure of globular proteins. J. Mol. Biol. 120:97-120[Medline].

GILLEY, D., J. R. PREER, JR., K. J. AUFDERHEIDE, and B. POLISKY, 1988  Autonomous replication and addition of telomerelike sequences to DNA microinjected into Paramecium tetraurelia macronuclei. Mol. Cell. Biol. 8:4765-4772[Abstract/Free Full Text].

GODISKA, R., K. J. AUFDERHEIDE, D. GILLEY, P. HENDRIE, and T. FITZWATER et al., 1987  Transformation of Paramecium by microinjection of a cloned serotype gene. Proc. Natl. Acad. Sci. USA 84:7590-7594[Abstract/Free Full Text].

GURNETT, C. A. and K. P. CAMPBELL, 1996  Transmembrane auxiliary subunits of voltage-dependent ion channels. J. Biol. Chem. 271:27975-27978[Free Full Text].

HAGA, N., M. FORTE, Y. SAIMI, and C. KUNG, 1982  Microinjection of cytoplasm as a test of complementation in Paramecium. J. Cell Biol. 92:559-564[Abstract/Free Full Text].

HAGA, N., Y. SAIMI, M. TAKAHASHI, and C. KUNG, 1983  Intra- and interspecific complementation of membrane-inexcitable mutants of Paramecium. J. Cell Biol. 97:378-382[Abstract/Free Full Text].

HAGA, N., M. FORTE, Y. SAIMI, and C. KUNG, 1984  Characterization of cytoplasmic factors which complement Ca2+ channel mutations in Paramecium tetraurelia.. J. Neurogenet. 1:259-274[Medline].

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., K.-Y. LING, Y. SAIMI, and C. KUNG, 1996  Toward cloning genes by complementation in Paramecium. J. Neurogenet. 11:81-98[Medline].

HOBOHM, U. and C. SANDER, 1995  A sequence property approach to searching protein databases. J. Mol. Biol. 251:390-399[Medline].

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

KANABROCKI, J. A., Y. SAIMI, R. R. PRESTON, W. J. HAYNES, and C. KUNG, 1991  Efficient transformation of cam2, a behavioral mutant of Paramecium tetraurelia, with the calmodulin gene. Proc. Natl. Acad. Sci. USA 88:10845-10849[Abstract/Free Full Text].

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., 1971a  Genic mutants with altered system of excitation in Paramecium aurelia. II. Mutagenesis, screening and genetic analysis of the mutants. Genetics 69:29-45[Free Full Text].

KUNG, C., 1971b  Genic mutations with altered system of excitation in Paramecium aurelia. I. Phenotypes of the behavioral mutants. Z. Vgl. Physiol. 71:142-164.

KUNG, C., 1975  Genetic dissection of the excitable membrane of Paramecium. Genetics 79:423-431.

KUNG, C. and R. ECKERT, 1972  Genetic modification of electric properties in an excitable membrane (Paramecium-calcium conductance-electrophysiological measurements-membrane mutant). Proc. Natl. Acad. Sci. USA 69:93-97[Abstract/Free Full Text].

KUNG, C. and Y. NAITOH, 1973  Calcium-induced ciliary reversal in the extracted models of "Pawn", a behavioral mutant of Paramecium. Science 179:195-196[Abstract/Free Full Text].

KYTE, J. and R. F. DOOLITTLE, 1982  A simple method for displaying the hydropathic character of a protein. J. Mol. Biol. 157:105-132[Medline].

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].

MACHEMER, H. and R. ECKERT, 1973  Electrophysiological control of reverse ciliary beating in Paramecium. J. Gen. Physiol. 61:572-587[Abstract/Free Full Text].

MACHEMER, H. and A. OGURA, 1979  Ionic conductances of membranes in ciliated and deciliated Paramecium. J. Physiol. 296:49-60.

MARYON, E. B., R. CORONADO, and P. ANDERSON, 1996  68 encodes a ryanodine receptor involved in regulating C. elegans body-wall muscle contraction. J. Cell Biol. 134:885-893[Abstract/Free Full Text].

NAKAI, K. and M. KANEHISA, 1992  A knowledge base for predicting protein localization sites in eukaryotic cell. Genomics 14:897-911[Medline].

OGURA, A. and K. TAKAHASHI, 1976  Artificial deciliation causes loss of calcium-dependent responses in Paramecium. Nature 264:170-172[Medline].

PREER, J. R., 1986 Surface antigens of Paramecium, pp. 301–339 in The Molecular Biology of Ciliated Protozoa, edited by J. G. GALL. Academic Press, Orlando, FL.

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

PRESTON, R. R., Y. SAIMI, B. MARTINAC, and C. KUNG, 1992b  Genetic analysis of ion channels of prokaryotes and lower eukaryotes. Curr. Opin. Genet. Dev. 2:780-784[Medline].

RAMANATHAN, R., Y. SAIMI, J. B. PETERSON, D. L. NELSON, and C. KUNG, 1983  Antibodies to the ciliary membrane of Paramecium tetraurelia alter membrane excitability. J. Cell Biol. 97:1421-1428[Abstract/Free Full Text].

ROBINSON, P. J., 1997  Signal transduction via GPI-anchored membrane proteins. Adv. Exp. Med. Biol. 419:365-370[Medline].

RUSSELL, C. B., D. FRAGA, and R. D. HINRICHSEN, 1994  Extremely short 20–33 nucleotide introns are the standard length in Paramecium tetraurelia.. Nucl. Acids Res. 22:1221-1225[Abstract/Free Full Text].

RYAN, K. A., L. A. GARRAWAY, A. DESCOTEAUX, S. J. TURCO, and S. M. BEVERLEY, 1993  Isolation of virulence genes directing surface glycosyl-phosphatidylinositol synthesis by functional complementation of Leishmania. Proc. Natl. Acad. Sci. USA 90:8609-8613[Abstract/Free Full Text].

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

SAIMI, Y. and K.-Y. LING, 1995  Paramecium Na+ channels activated by Ca(2+)-calmodulin: calmodulin is the Ca2+ sensor in the channel gating mechanism. J. Membr. Biol. 144:257-265[Medline].

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

SATOW, Y. and C. KUNG, 1974  Genetic dissection of active electrogenesis in Paramecium aurelia.. Nature 247:69-71[Medline].

SATOW, Y. and C. KUNG, 1980  Membrane currents of pawn mutants of the pwA group in Paramecium tetraurelia.. J. Exp. Biol. 84:57-71[Abstract/Free Full Text].

SCHAFER, W. R. and C. J. KENYON, 1995  A calcium-channel homologue required for adaptation to dopamine and serotonin in Caenorhabditis elegans.. Nature 375:73-78[Medline].

SKOURI, F. and J. COHEN, 1997  Genetic approach to regulated exocytosis using functional complementation in Paramecium—identification of the Nd7 gene required for membrane fusion. Mol. Biol. Cell 8:1063-1071[Abstract].

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.

UDENFRIEND, S. and K. KODUKULA, 1995  How glycosylphosphatidylinositol-anchored membrane proteins are made. Annu. Rev. Biochem. 64:563-591[Medline].

VALLET, V., A. CHRAIBI, H. P. GAEGGELER, J. D. HORISBERGER, and B. C. ROSSIER, 1997  An epithelial serine protease activates the amiloride-sensitive sodium channel. Nature 389:607-610[Medline].

VASHISHTHA, M., G. SEGIL, and J. L. HALL, 1996  Direct complementation of Chlamydomonas mutants with amplified YAC DNA. Genomics 36:459-467[Medline].

WARMKE, J. W., R. A. REENAN, P. WANG, S. QIAN, and J. P. ARENA et al., 1997  Functional expression of Drosophila para sodium channels. Modulation by the membrane protein TipE and toxin pharmacology. J. Gen. Physiol. 110:119-133[Abstract/Free Full Text].

WILSON, W. and T. SEEBECK, 1997  Identification of a profilin homologue in Trypanosoma brucei by complementation screening. Gene 187:201-209[Medline].

ZHANG, H., P. L. HERMAN, and D. P. WEEKS, 1994  Gene isolation through genomic complementation using an indexed library of Chlamydomonas reinhardtii DNA. Plant Mol. Biol. 24:663-672[Medline].