A robust inducible-repressible promoter greatly facilitates gene knockouts, conditional expression, and overexpression of homologous and heterologous genes in Tetrahymena thermophila

Abstract

The Cd²⁺-inducible metallothionein (MTT1) gene was cloned from Tetrahymena thermophila. Northern blot analysis showed that MTT1 mRNA is not detectable in the absence of Cd²⁺, is induced within 10 min of its addition, is expressed in proportion to its concentration, and rapidly disappears upon its withdrawal. Similarly, when the neo1 gene coding region flanked by the MTT1 gene noncoding sequences was used to disrupt the MTT1 locus, no transformants were observed in the absence of Cd²⁺, and the number of transformants was proportional to increased Cd²⁺ concentration. The neo3 cassette, in which the MTT1 promoter replaced the histone gene HHF1 promoter of the previously used neo2 cassette, transformed cells at much higher frequencies than neo2 and produced germ-line knockouts where neo2 had failed. Rescuing the progeny of a mating of γ-tubulin gene, GTU1, knockout heterokaryons with a GTU1 gene inserted into the MTT1 locus yielded >75 times more transformants than rescuing with the wild-type GTU1 gene itself. When cells rescued with the MTT1-GTU1 chimeric gene were transferred to medium lacking Cd²⁺, they stopped growing and had phenotypic changes indistinguishable from cells containing only disrupted GTU1 genes. Thus, it is now possible to create conditional lethal mutants and study the terminal phenotypes of null mutations for essential genes by replacing the endogenous gene with one under the control of the MTT1 promoter. The MTT1 promoter also resulted in ≈30 times more overexpression of the IAG48[G1] surface antigen gene of the ciliate fish parasite Ichthyophthirius multifiliis than the highly expressed BTU1 promoter, accounting for ≈1% of the total cell protein. Thus, the MTT1 promoter should enable routine over-expression of endogenous and foreign genes in Tetrahymena.

Tetrahymena thermophila is a ciliated protozoan that grows rapidly to high densities. Its relatively large size, nuclear dimorphism, and well developed techniques for genetic analyses, cytological studies, and cell fractionation make it a useful eukaryotic model to study diverse molecular, cellular, and developmental processes (see Ref. 1). Several fundamental and evolutionarily conserved phenomena were first identified in T. thermophila, including the discovery of dynein (2), the structure of telomeres (3), the identification of telomerase as a ribonucleoprotein enzyme (4), self-splicing RNA (5), and the relationship between transcription factors and histone modification (6).

In recent years, experimental analyses in Tetrahymena have been augmented by the development of methods for DNA-mediated transformation of the somatic macronucleus, first by microinjection of individual cells (7), then en masse by electroporation (8) or biolistically (9), and by the discovery that transformation in Tetrahymena occurs largely, if not entirely by homologous integration (10). Coupled with the introduction of heterologous drug-resistance gene markers (11, 12), these methods enabled facile gene replacement, allowing gene disruption (13) and insertion of mutant genes with flanking selectable markers (14). They also led to techniques for transformation of the germ-line micronucleus which, again, occurs by homologous integration (15). Germ-line transformation led, in turn, to the creation of germ-line knockout heterokaryon strains (16), in which both copies of an essential gene are disrupted in the transcriptionally silent micronucleus whereas the transcriptionally active, somatic macronucleus contains wild-type genes. When two strains that are knockout heterokaryons for the same essential gene are mated, their progeny die unless they are transformed with a version of the gene that supports growth. Although these methods have enabled many studies (12, 17–19), molecular genetic analyses in Tetrahymena were still limited by relatively low efficiencies of transformation, precluding cloning genes by complementation of function or suppression of mutations, and by the absence of a regulatable expression system. To help overcome these drawbacks, we sought to identify an inducible-repressible promoter in Tetrahymena.

Metallothioneins (MTTs) are highly conserved, low molecular weight, cysteine-rich metal-binding proteins whose primary function is unknown, but who are generally considered to play a role in the homeostasis of metals such as zinc and copper and in the detoxification of cadmium (20). The synthesis of many metallothioneins, including those of Tetrahymena pyriformis and T. thermophila, can be induced by heavy metals, such as zinc, copper, and cadmium (21). Metal-responsive metallothionein promoters have been used successfully to regulate gene expression in other systems (22, 23).

In this article, we describe the cloning and regulation of MTT1, a gene encoding a Cd²⁺-inducible metallothionein in T. thermophila. We demonstrate that this promoter can greatly increase the efficiency of many aspects of DNA-mediated transformation in Tetrahymena, including somatic and germ-line gene disruption and rescue of knockout heterokaryons. The MTT1 promoter also can be used to overexpress homologous or heterologous genes and to create a conditional lethal mutation of an essential gene.

Materials and Methods

Strains and Culture Conditions.

Wild-type strain CU428 and paclitaxel-sensitive strain CU522 were kindly provided by P. J. Bruns (Cornell University). Knockout heterokaryon strains (GTU1-KO5 and GTU1-KO6) of the GTU1 gene encoding the single γ-tubulin of Tetrahymena were constructed as described by Hai and Gorovsky (24). Successful creation of germ-line knockout heterokaryons of the GTU1 gene was demonstrated by the fact that no viable conjugation progeny were obtained when the GTU1 heterokaryons were mated and that progeny could be rescued by transforming with a wild-type GTU1 gene. Strain cTTMG was generated by rescuing the mating GTU1 knockout heterokaryons with a MTT1-driven GTU1 gene. Cells were grown routinely in SPP (1% protease peptone/0.2% glucose/0.1% yeast extract/0.003% EDTA ferric sodium salt) (25). To starve cells, a log-phase culture was washed and resuspended in 10 mM Tris⋅HCl (pH7.5) and incubated for approximately 24 h at 30°C without shaking. To study induced expression of the MTT1 promoter, different amounts of CdCl₂ were added to growing, starved, or mating cells for the indicated times.

Cloning the MTT1 Gene.

Based on the amino acid sequence of the T. pyriformis MTT protein (26), two degenerate oligonucleotides were synthesized and used to PCR amplify a 300-bp product encoding a fragment of the coding region of the MTT1 gene of T. thermophila. This PCR product was used to probe a genomic Southern blot of T. thermophila DNA digested with TaqI, HindIII, or EcoRI. A single band was observed with each enzyme, indicating that the MTT protein is encoded by a single gene in T. thermophila. Double digestion with HindIII and EcoRI gave a fragment of about 3.2 kb. Size-selected HindIII and EcoRI double-digested genomic DNA was ligated into digested pBluescript KS(+) plasmid (Stratagene) and transformed into DH5α bacteria. Colony lifts were hybridized with the 300-bp product. A positive colony was subsequently cloned and sequenced, revealing an ORF of 163 amino acids that had very high similarity with MTT from T. pyriformis. This construct, pTTMet, contains ≈2.5 kb of 5′ flanking, 489-bp coding, and ≈375 bp of 3′-flanking region of the MTT1 gene of T. thermophila.

Construction of Transformation Vectors.

The reporter construct pTTMN was obtained by replacing the MTT1 coding region with the neo1 coding sequence. Plasmid p4T2–1 is a pBluescript KS (+) derivative containing the neo2 cassette, a chimeric HHF1/neo1/BTU2 gene, with a HindIII site after the neo1 start codon (27). The neo1 coding region was PCR-amplified from p4T2–1, and the fragment was purified after first treating with T4 DNA polymerase followed by HindIII. Using pTTMet and an inverse PCR primer that added a HindIII site after the ATG start codon and a second primer starting ≈120 bp 3′ of the TGA stop codon, the 5′-flanking MTT1-pBluescript vector–3′-flanking MTT1 sequence was amplified. This fragment was treated with T4 DNA polymerase, then digested with HindIII and ligated to the neo1 fragment to create pTTMN.

The pMNBL plasmid, which contains the neo3 cassette, a hybrid MTT1/neo1/BTU2 gene, was constructed as follows. Plasmid pTTMN was digested with KpnI, which cleaves in the multiple cloning site of the pBluescript vector, blunted by T4 DNA polymerase, and digested with HindIII. The 2.5-kb fragment containing the MTT1 5′-flanking sequence was gel purified. p4T2–1 was also digested with KpnI, blunted with T4 DNA polymerase, and digested with HindIII, and the large fragment containing the neo1-coding region–BTU2 3′-flanking-pBluescript vector sequence was isolated. The two fragments were then ligated to construct the pMNBL plasmid. To construct pMNBM, which contains only ≈600 bp of MTT1 5′-flanking sequence upstream of the ATG start codon, the EcoRV-HindIII fragment containing the HHF1 promoter in the p4T2–1 plasmid was replaced by the 600-bp AflIII-HindIII fragment from the pTTMN plasmid.

To construct the GTU1 knockout plasmid, pΔGN, a 0.7-kb fragment of the GTU1 5′-flanking sequence (from a BglII site to the ATG start codon) was PCR-amplified from genomic DNA with use of a forward primer that introduced a KpnI site at its end and a reverse oligo that introduced a NotI site at its end. The PCR fragment was blunted with T4 DNA polymerase, digested with KpnI, and inserted into the 5′ polylinker region (between KpnI and EcoRV) of p4T2–1. A 1.0-kb fragment of the GTU1 3′-flanking sequence was PCR-amplified from Tetrahymena genomic DNA by using a forward primer and a reverse primer that introduced a XhoI and SacI sites at their ends, respectively. This PCR product was blunted with T4 DNA polymerase, digested with SacI, and inserted into the 3′ polylinker region (between SmaI and SacI) of p4T2–1. The plasmid pΔGMN, which contains 2.5 kb of MTT1 5′-flanking sequence upstream of the ATG start codon, or pΔGMM, which has only 600 bp of MTT1 5′-flanking sequence, was constructed by subcloning the NotI-XhoI fragment containing the MTT1/neo1/BTU2 gene from either pMNBL or pMNBM into pΔGN between the NotI and XhoI sites. To construct pΔGMMII, which contains the 900 bp of MTTI 5′-flanking sequence directly 5′ of the ATG start codon, pΔGMN was digested with AccI to release the distal 1.6-kb MTT1 5′-flanking sequence with ≈120 bp of GTU1 3′-flanking sequence. The large fragment obtained from this restriction digestion was self-ligated. In the ngoA gene knockout constructs, pΔNgoAH4 or pΔNgoAMT, either the neo2 or neo3 cassette was inserted between the ngoA 5′- and 3′-flanking regions using similar procedures.

To create the gene expression construct pMTT-BICH3, site-directed mutagenesis was used to introduce a BglII site about 490 bp upstream of the ATG start codon within the 5′-flanking region of pBICH3 (31), which contains the coding region of the IAG48[G1] surface antigen gene of the ciliate Ichthyophthirius multifiliis inserted between the flanking sequences of the BTU1 gene of T. thermophila. A HindIII site exists a few base pairs downstream of the ATG start codon in pBICH3; and a BglII site is very close to the 5′ end of the MTT1 5′-flanking region. The proximal part of the BTU1 promoter was removed by digestion with HindIII and BglII, and replaced by the 2.5-Kb BglII-HindIII fragment of the MTT1 promoter from the pTTMN plasmid.

To create plasmid pTTMG to rescue the progeny of matings of GTU1 knockout heterokaryons, the γ-tubulin coding region was PCR-amplified from pGTU (28), which contains a wild-type version of GTU1. This amplified fragment was ligated to the inverse PCR fragment from pTTMet, containing the pBluescript vector sequence and the 5′- and 3′-flanking sequences of MTT1, as described above. pGTU-E is pGTU with an EcoRI site added by mutagenesis directly 5′ of the TGA and an additional 1.5 kb of 5′-flanking sequence added by inverse PCR.

Northern Blot Analysis.

RNA was isolated with Trizol (Life Technologies, Grand Island, NY), electrophoresed in 2.2 M formaldehyde-1.2% agarose gels, blotted, and hybridized with [α-³²P]dATP-labeled, randomly primed probes (29). The probe for rRNA was a 2-kb HindIII fragment from pBS26S encoding the Tetrahymena 26S RNA (30). The GTU1 probe was synthesized from a 1.0-kb StyI-NsiI fragment from pBL-GTU4 (28). The MTT1 probe was a 300-bp PCR product amplified from T. thermophila genomic DNA with coding region primers. Hybridizations were done at 42°C in 50% formamide, 5× SSC, 1× SPED (0.1% Ficoll/0.1% polyvinylpyrrolidone/0.1% BSA/6 mM SDS/2 mM sodium pyrophosphate/2 mM EDTA), 1% SDS, and 100 μg/ml salmon sperm DNA.

Southern Blot Analysis.

To determine the number of copies of the MTT1 gene in T. thermophila, total genomic DNA was isolated from log-phase CU428 cells, digested, electrophoresed, blotted onto Magnagraph nylon membrane (Osmonics Inc., Westborough, MA), and hybridized with [α-³²P]dATP-labeled, randomly primed probes using standard protocols (29).

Western Blotting.

Immunoblots of whole-cell protein extracts from Tetrahymena transformed with pBICH3 or pMTT-BICH3 reacted with antibodies to I. multifiliis were prepared as described (31).

Tetrahymena Transformation.

CU428 cells were starved overnight in 10 mM Tris⋅HCl (pH 7.5) and biolistically transformed with KpnI and SacI digested plasmid pTTMN, pΔGMN, pΔGMMII, or pΔGMM using the DuPont Biolistic PDS-1000/He particle delivery system (Bio-Rad) (9). After bombardment, cells were resuspended in 50 ml of SPP medium with varying concentrations of CdCl₂, incubated at 30°C for 3–5 h, and then plated in 96-well microtiter plates in the presence of 120 μg/ml paromomycin sulfate. Plasmid pMTT-BICH3 was linearized with SacI and SalI and used to transform the CU522 strain by biolistic bombardment, as described (31). Transformants were selected by growth in 20 μM paclitaxel for about 2 weeks. For rescue experiments, plasmid pGTU-E or pTTMG was biolistically transformed into unfed exconjugants from the cross between the GTU1 knockout heterokaryon strains GTUKO5 and GTUKO6, followed by refeeding with SPP medium with 1.0 μg/ml CdCl₂ and plating.

Results

Induction Characteristics of the MTT1 Promoter in T. thermophila.

We cloned the metallothionein (MTT1) gene of T. thermophila (GenBank accession no. AY061892). It encodes a protein (MTT1p) containing 162 amino acids, which is very similar to cadmium-metallothioneins from T. pyriformis and Tetrahymena pigmentosum (21, 26), except that it contains a duplication corresponding to residues 3–55. The MTT1 gene is present in a single copy gene in T. thermophila (data not shown), as in T. pyriformis (26).

To evaluate the inducibility of the MTT1 promoter in T. thermophila, RNA isolated from log-phase wild-type strain CU428 cells grown overnight in SPP medium containing the indicated concentrations of CdCl₂ was analyzed by Northern blotting. T. thermophila can grow in up to 2.0 μg/ml CdCl₂, and the growth rate in 1.0 μg/ml CdCl₂ is indistinguishable from that in CdCl₂-free medium (data not shown). The expression of the MTT1 gene is not detectable in the absence of CdCl₂, can be induced to high levels, and can be regulated by the level of CdCl₂ in growing (Fig. 1A) and starved (Fig. 1B) cells. Starved cells were more sensitive to CdCl₂, and MTT1 expression in these cells could be induced at lower concentrations than in growing cells. The mating process was also delayed in the presence of CdCl₂. Three hours after mixing, about 61% of the cells were paired in 0.06 μg/ml CdCl₂, compared with 81% in CdCl₂-free Tris. Induction in mating cells was similar to starved cells (Fig. 1B). Reduction of MTT1 induction in starved cells after 24 h exposure to CdCl₂ (compared with 2 h in Fig. 1B) is reproducible and does not occur in starved-mating cells, but was not investigated further.

Figure 1.

Northern blot analyses of MTT1 induction. (A) Transcription of the MTT1, but not the GTU1 gene is induced by cadmium in growing cells. Total RNA was analyzed from wild-type CU428 cells grown overnight in SPP medium containing the indicated concentrations of CdCl₂. (B) Transcription of the MTT1 gene is induced by cadmium in starved and mating cells. To starve cells, a culture of log phase CU428 cells was washed twice and then incubated at 30°C without shaking in 10 mM Tris with the indicated concentration of CdCl₂. To obtain mating cells, CU428 and B2086 cells were starved overnight as described above, mixed to initiate mating, and incubated at 30°C without shaking in 10 mM Tris with the indicated concentration of CdCl₂. Total RNA was isolated either 2 or 24 h after starvation (starved cells) or mixing (mating cells). (C) Induction and repression of the MTT1 promoter occur rapidly. In +CdCl₂, RNA was analyzed from log-phase wild-type CU428 cells incubated in SPP containing 1.0 μg/ml CdCl₂ for the indicated times. In −CdCl₂, wild-type CU428 cells were grown in SPP with 1.0 μg/ml CdCl₂ overnight and then washed twice with SPP medium. Total RNA was isolated at the indicated times. All Northern blots were probed with a MTT1 or GTU1 coding sequence probe or with a 25S rRNA probe as a control for loading (data not shown).

To determine the rapidity with which the MTT1 promoter can be activated by CdCl₂, log-phase wild-type cells were incubated in standard culture medium (SPP) with 1.0 μg/ml CdCl₂, and expression of MTT1 was measured after different periods. Fig. 1C indicates that MTT1 expression can be induced very rapidly. MTT1 mRNA was detected within 10 min after adding the CdCl₂ and reached a maximum level at about 45 min. Cessation of MTT1 expression also occurs rapidly after CdCl₂ is removed. In cells that had been growing with 1.0 μg/ml CdCl₂ overnight and were transferred into medium without CdCl₂, a decrease in MTT1 mRNA was evident by 30 min, and by 60 min the MTT1 mRNA could not be detected (Fig. 1C), which suggests that transcription of the MTT1 gene can be turned off very rapidly simply by depleting the medium of cadmium.

The neo Reporter Gene, neo1, Can Be Inducibly Expressed at the MTT1 Locus.

To test whether the MTT1 promoter can induce the expression of other coding regions in response to cadmium, a reporter construct was made by replacing the MTT1 coding region with the neo1 coding region which confers paromomycin (pm) resistance when expressed in Tetrahymena macronuclei (11). Because this reporter gene is flanked by MTT1 noncoding sequences, it integrates into the MTT1 locus by homologous recombination when biolistically transformed into Tetrahymena. After transformation, cells were selected in medium containing paromomycin and increasing amounts of CdCl₂. Fig. 2 indicates that transformation efficiency with this construct is highly dependent on cadmium concentration. No pm-resistant cells were obtained in the absence of cadmium, and the number of transformants was proportional to cadmium concentration. This transformation experiment was repeated three times with similar results, and up to 25,000–30,000 transformants per μg DNA were obtained at 2.0 μg/ml CdCl₂.

Figure 2.

Transformation of Tetrahymena with a neomycin resistance gene driven by the MTT1 promoter. (A) The structure of the endogenous MTT1 gene and the insert in plasmid pTTMN which contains the neo1 gene flanked by MTT1 5′ and 3′ noncoding sequences. (B) The effect of cadmium concentration on biolistic transformation of Tetrahymena by pTTMN. Wild-type CU428 cells were transformed (see Materials and Methods) and plated at the concentrations of CdCl₂ indicated on the abscissa. The percentage of paromomycin-resistant transformants per 96-well plate was plotted relative to the number obtained with 2.0 μg/ml CdCl₂.

To test whether MTT1p is required for survival of Tetrahymena in CdCl₂, transformants were transferred in SPP containing 0.5 μg/ml cadmium with increased concentration of paromomycin for a month. Both genomic PCR and Northern blot analysis showed that the endogenous MTT1 gene was completely replaced (data not shown). Cells lacking MTT1 can barely grow in 2.0 μg/ml CdCl₂, whereas their growth rate in 1.0 μg/ml CdCl₂ or in cadmium-free medium is indistinguishable and similar to that of wild-type control cells (data not shown).

The MTT1 Promoter Enables Higher Frequency Gene Disruption in Somatic Cells than the HHF1 Promoter.

For gene disruption studies in Tetrahymena we had developed the neo2 cassette in which the promoter of the HHF1 gene (encoding histone H4) and the termination region of the BTU2 gene (encoding β-tubulin) were used to express the neo1 coding region (27). To determine whether the MTT1 promoter would allow higher transformation efficiencies if used instead of the HHF1 promoter, the neo3 cassette was created by replacing the HHF1 gene 5′ region of neo2 with 2.5 kb of MTT1 5′-flanking sequence. To test this new cassette, the knockout plasmid pΔGMN was constructed in which the neo3 cassette is flanked by the 5′ and 3′ sequences of the GTU1 gene encoding the single γ-tubulin gene of Tetrahymena. Somatic biolistic transformation was performed by using either linearized pΔGMN or pΔGN, which contains the neo2 cassette (Fig. 3A). Approximately 1,800 pm-resistant transformants per μg DNA were obtained by using neo3, whereas fewer than 10 transformants per μg DNA were obtained with neo2. Therefore, the new MTT1/neo1/BTU2 cassette enables gene disruption at much higher frequency than the HHF1/neo1/BTU2 cassette.

Figure 3.

The MTT1 promoter improves the efficiency of DNA-mediated, biolistic transformation of Tetrahymena. (A) The MTT1 promoter-driven neo3 cassette gives higher somatic transformation rates than the HHF1 promoter-driven neo2 cassette. Four different GTU1 knockout constructs are shown. All cassettes use the same BTU2 3′-flanking region, but they differ in the 5′-flanking region. In pΔGMN, neo3 expression is driven by 2.5 kb of MTT1 5′-flanking region, and in pΔGN, neo2 is driven by the HHF1 promoter. In pΔGMMII, neo3 is driven by 900 bp of MTT1 5′-flanking sequence. In pΔGMM, neo3 is driven by 600 bp of MTT1 5′-flanking sequence. Three micrograms of DNA were used in each transformation. After transformation, the CU428 cells were refed in 1× SPP with 1.0 μg/ml CdCl₂ for 3 to 6 h before adding 120 μg/ml paromomycin followed by plating. The number of transformants was obtained by counting the number of wells with viable transformants in 96-well plates at known dilutions. (B) The MTT1 promoter-driven neo3 cassette enables both somatic and germ-line knockout of the ngoA gene where the HHF1-driven neo2 cassette fails. Two different ngoA knockout constructs are shown. Both contain the same ngoA-flanking sequences. pΔNgoAMT contains the neo3 cassette driven by the MTT1 promoter, and pΔNgoAH4 contains neo2 driven by the HHF1 promoter. Wild-type CU428 cells were mated with B2086 cells, and 3 μg DNA was used in each transformation. After transformation, cells were starved in 10 mM Tris overnight and then refed in SPP containing 1.2 μg/ml CdCl₂ for 3 to 6 h before addition of 80 μg/ml paromomycin and plating. After 3 days, more than 700 transformants were obtained. Paromomycin-resistant transformants were tested for sensitivity to 6-methylpurine in SPP. The Pm/6-methylpurine double-resistant transformants were further tested by Southern blotting to determine whether the neo cassettes were in the correct locus and for ability to sexually transmit the knockout phenotype (data not shown). Two neo3 transformants were actual germ-line knockout transformants. (C) The MTT1 promoter increases the rescue efficiency of knockout heterokaryons. Two constructs were used to rescue GTU1 germ-line knockout heterokaryons. Mating cells were transformed with 3 μg DNA. Transformants were selected in 60 μg/ml paromomycin for 4 days and the number of transformants calculated by counting the number of wells with viable transformants in 96-well plates at known dilutions.

We next examined the amount of MTT1 5′-flanking sequence required for biolistic gene disruption by using neo3. Plasmid pΔGMMII, which contains a ≈900-bp 5′-flanking sequence of the MTT1 gene, transformed Tetrahymena at high frequencies, close to those obtained with pΔGMN, containing ≈2.5 kb of flanking sequence. In contrast, pΔGMM, which contains only a 600-bp flanking sequence, failed to transform Tetrahymena (Fig. 3A). Thus, an important, Cd-responsive promoter element is likely to be in the region 600–900 bp upstream of the MTT1 coding sequence.

The neo3 Cassette Enables Gene Disruption in the Germ Line in Cases Where neo2 Failed.

In the course of studies to be described in detail elsewhere, we encountered two genes that we were unable to disrupt either in the somatic macronucleus or in the germ-line micronucleus. These failures likely represent extreme cases of a more general, position-effect phenomenon where the neo2 cassette inserted into different loci produces transformants differing in their resistance to paromomycin by greater than 100-fold (unpublished observations). We therefore tested the ability of the neo3 cassette to disrupt these loci. ngoA is a gene of unknown function specifically expressed in conjugating cells (32). Although the neo2 knockout construct failed to produce any transformants (Fig. 3B), with the neo3 cassette, more than 700 pm-resistant transformants per 12 μg DNA were obtained, two of which were shown by subsequent analyses to be true germ-line knockout transformants (X.S. and M.A.G., unpublished observations). Similarly, for the BLT1 gene encoding a highly divergent β-tubulin present in very low abundance (K. Clark, B. Li, and M.A.G., unpublished observations), a disruption construct containing the neo3 cassette produced numerous somatic knockouts and a low number of germ-line knockouts (data not shown). Thus, in both cases we have encountered where neo2 failed to produce somatic or germ-line transformants, neo3 was successful.

The MTT1 Promoter Increases Rescue of Knockout Heterokaryons.

Rescue of progeny of matings of knockout heterokaryons with in vitro mutagenized genes has proved to be a rapid method for in vivo mutagenesis of essential genes in Tetrahymena (12, 19). We compared the frequencies of rescue of mating GTU1 knockout heterokaryon strains (Y.S. and M.A.G., unpublished observations) by a wild-type GTU1 gene targeted to the GTU1 locus with that of wild-type GTU1 coding region targeted to the MTT1 locus (Fig. 3C). Clearly, rescue by cloning into the MTT1 locus is 2–3 orders of magnitude more efficient than integration into the GTU1 locus itself.

The MTT1 Promoter Enables Analysis of the Hypomorphic or Terminal Phenotype of an Essential Gene by Converting It into a Conditional Mutant.

Because knocking out essential genes usually results in dead cells, it is frequently difficult to obtain information on the null phenotype of these genes. This difficulty is a particularly acute problem in Tetrahymena, where many new mutations are the product of reverse genetic analysis of gene knockouts. This problem is often addressed by creation of conditional lethal mutations, whose properties can be studied under the nonpermissive condition. To determine whether placing an essential gene under the control of the MTT1 promoter would lead to a conditional, Cd-regulatable mutant, we examined the phenotype of a mutant strain (cTTMG) in which the only expressed GTU1 gene was under MTT1 control. The growth rate of such cells is indistinguishable from that of wild-type cells in the presence of CdCl₂ (data not shown), but they cease growing when resuspended in CdCl₂-free medium (Fig. 4A), and cell shape and the organization of the tubulin cytoskeleton become extremely abnormal (Fig. 4C). This phenotype is similar to the phenotype observed in progeny cells of mating GTU1 knockout heterokaryons, which have only disrupted copies of GTU1. Although the detailed phenotype of these cells in the absence of CdCl₂ will be presented elsewhere (Y.S. and M.A.G., manuscript in preparation), it is clear from these studies that the wild-type GTU1 gene exhibits a conditional phenotype in the absence of CdCl₂ that is likely caused by γ-tubulin depletion.

Figure 4.

An essential gene regulated by the MTT1 promoter behaves as a conditional mutation. Wild-type Cu428 (●) or cTTMG (⋄) cells, in which the GTU1-coding sequence was regulated by the MTT1 promoter, were resuspended in SPP medium without cadmium. Cells were counted at various times after suspension. Growth of the cTTMG cells without cadmium slowed at about 11 h relative to wild-type cells. Growth of cTTMG cells in the presence of cadmium is indistinguishable from that of wild-type cells (data not shown). (B and C) The shape and microtubule distribution of cTTMG cells was disrupted after depletion of cadmium. Wild-type cells (B) or cells containing the MTT1-GTU1 chimeric gene (C) were grown in normal SPP media for 24 h and then were fixed and stained with anti-α-tubulin antibody. The shape and microtubule distribution of cTTMG cells grown in the presence of cadmium is indistinguishable from that of wild-type cells (data not shown).

Overexpression of an Ich Surface Antigen Gene Driven by the MTT1 Promoter.

The BTU1 gene is one of two coexpressed genes encoding the major β-tubulin of T. thermophila (33, 34). BTU1 mRNA is highly abundant and β-tubulin makes up about 2–3% of the total Tetrahymena cell protein (35). It was shown previously that this highly active promoter could drive high-level expression of the IAG48[G1] surface antigen gene of the parasite ciliate I. multifiliis inserted into the Tetrahymena BTU1 locus in place of the BTU1 coding region (31). To determine whether the MTT1 promoter would be useful for overexpression of heterologous genes in Tetrahymena, we compared the expression levels of the IAG48[G1] surface antigen gene driven by the MTT1 promoter with that driven by the BTU1 promoter. In the pMTT-BICH3 construct, the 2.5-kb MTT1 5′-flanking region was inserted upstream of the coding region of the IAG48[G1] surface antigen gene in the previously described pBICH3 plasmid which also contains BTU1 3′- and 5′-flanking sequences (Fig. 5A). Both constructs were then (separately) transformed biolistically into the BTU1 locus of strain CU522, which contains a dominant, paclitaxel-sensitive BTU1 gene (31, 36). Replacement of the paclitaxel-sensitive gene by the insertion of either plasmid results in paclitaxel-resistant cells, which are easily selected. Expression of Ich surface antigen gene in both transformed strains was measured by Western blotting with and without treating the cells with CdCl₂ (Fig. 5B). The expression level of the Ich surface antigen driven by the MTT1 promoter was much higher than that driven by the BTU1 promoter. After optimization for CdCl₂ concentration (5 μg/ml) and time of treatment (9 h), quantitative analyses of similar gels indicated that the expression of the Ich surface antigen driven by the MTT1 promoter was 18–30 times greater than from the BTU1 promoter and constituted up to ≈1% of the total cell protein (data not shown). Similarly, the γ-tubulin can be overexpressed in cells in which the GTU1 gene is regulated by the MTT1 promoter (data not shown). Clearly, the MTT1 promoter is now the promoter of choice for overexpression of genes in Tetrahymena.

Figure 5.

The MTT1 promoter allows overexpression of a foreign gene in Tetrahymena. (A) Schematic maps of the target taxol-sensitive BTU1-K350M locus of T. thermophila and two transforming plasmid inserts in which the BTU1-coding region was replaced by IAG48[G1] sequences encoding a surface antigen from the fish parasite, Ich. The expression of the IAG48[G1] gene is driven by either the BTU1 (pBICH3) or the MTT1 (pMTT-BICH3) promoter. (B) Western blot with an anti-Ich surface antigen antibody shows that, in the presence of cadmium, expression of the IAG48[G1] gene driven by the MTT1 promoter is much higher than that driven by the BTU1 promoter.

Discussion

We have cloned the T. thermophila metallothionein gene (MTT1) and demonstrated that its highly regulatable promoter can be used to increase the efficiency of most of the commonly used types of DNA-mediated transformation in this organism. The MTT1 promoter can be expressed in a graded fashion in proportion to CdCl₂ concentration in growing, starved, and conjugating cells. This promoter can be turned on and off rapidly, suggesting it may be possible to use it to study the site and kinetics of incorporation and turnover of MTT1-regulated tagged genes by treating cells briefly with CdCl₂. The MTT1 promoter is able to highly overexpress both homologous and heterologous genes and the fact that the MTT1 coding region is not essential offers the possibility that Tetrahymena might be useful as an inexpensive, easy-to-grow, eukaryotic expression system for foreign genes.

Many lines of evidence suggest that the MTT1 promoter is tightly regulated. In the absence of CdCl₂, MTT1 expression was not detected in growing, starved, and mating cells, and no transformants were obtained when the neo1 coding region was used to disrupt the MTT1 gene. In addition, the growth and microtubule organization of cells whose GTU1 gene was regulated by the MTT1 promoter was markedly altered when CdCl₂ was removed from the medium. These observations indicate that the MTT1 promoter enables fine control of gene expression-induction over a wide range in the presence of inducer, and tight repression in its absence. However, although depletion of cadmium resulted in the cessation of growth in cells containing MTT1-GTU1 chimeric genes, cells left in growth medium in the absence of cadmium eventually recovered normal morphology and resumed growth. We were able to detect very low expression of GTUp in these recovered cells on Western blots (data not shown). We also have observed leaky expression after the depletion of cadmium when MTT1 promoter-driven genes are inserted into the BTU1 locus (data not shown). The GTU1 gene is only weakly detected on Northern blots in wild-type cells, suggesting that it provides an especially sensitive test of leaky expression. On the other hand, cells containing a MTT1-BTU1 chimeric gene as their only major β-tubulin gene do not resume growth, even when maintained for a week after being resuspended in cadmium-free medium (J. Duan and M.A.G., unpublished data). These observations suggest that the tightness and consequences of MTT1 promoter silencing in the absence of cadmium may show some locus-specific effects. In summary, we have described a robust, inducible promoter derived from the T. thermophila metallothionein gene that greatly enhances molecular genetic analyses in this organism.

Abbreviations

MTT

metallothionein

pm

paromomycin