RNA Polymerase II Localizes in Tetrahymena thermophila Meiotic Micronuclei When Micronuclear Transcription Associated with Genome Rearrangement Occurs

Abstract

The germ line micronucleus in Tetrahymena thermophila is transcriptionally silent in vegetatively growing cells. However, micronuclear transcription has been observed in the early (“crescent”) stages of the sexual process, conjugation. This transcription is proposed to play a central role in identifying sites for subsequent genome rearrangements that accompany development of the somatic macronucleus from the micronucleus. RPB3 (cnjC), a gene encoding a protein homologous to the third largest subunit of RNA polymerase II (RNAP II), was previously reported to be expressed specifically during conjugation, suggesting a role in micronucleus-specific transcription. Rpb3p localized in the micronucleus only during the meiotic prophase, when micronuclear transcription occurs, and its intranuclear distribution is strikingly similar to that for previously described sites of micronuclear RNA synthesis. By contrast, Rpc5p, the homologous subunit shared by RNAPs I and III, was not detectable in the micronucleus at any stage of the life cycle. However, Rpb3p is not specific to the transcribing micronucleus. Like Rpc5p, it also localizes to macronuclei in all stages of the life cycle. Rpb3p is encoded by a unique, essential gene in Tetrahymena. Thus, RNAP II is associated with both somatic transcription and crescent transcription and probably has an important role in genome rearrangement.

Tetrahymena thermophila shows nuclear dimorphism (reviewed in references 8 and 24). Each cell contains a germ line micronucleus and a somatic macronucleus. Although both nuclei are derived from the micronuclei during conjugation, the macronucleus lacks ∼15% of the micronuclear genome due to two types of sequence elimination that occur during macronuclear development. One is deletion of ∼6,000 internal eliminated sequences (IES), accompanied by ligation of flanking macronucleus-destined sequences. IES in Tetrahymena vary from 0.5 to >20 kb and account for most of the sequences missing in the macronucleus. Excision of IES at a specific site can occur reproducibly or with a limited number of alternative boundaries and is epigenetically regulated by the old macronucleus (2). The other type of elimination involves chromosome breakage at specific 15-bp chromosome breakage sequences, followed by small (<50-bp) deletions of breakage-eliminated sequences (BES) and the addition of telomeres to produce 200 to 300 macronuclear chromosomes from the 5 chromosomes in the micronuclear (haploid) genome.

The micronucleus is believed to be transcriptionally silent in vegetatively growing cells (5). However, nongenic micronuclear transcription has been detected early in conjugation (11, 23), when premeiotic micronuclei adopt an elongate crescent shape that probably is related to “bouquet” or “horsetail” stage in other eukaryotes (17). RNA hybridizing to a micronucleus-specific sequence was detected in starved and mating cells (22), and long, heterogeneous RNAs homologous to both strands of IES have been observed during conjugation (3). Other features associated with transcription have also been localized to crescent micronuclei. At meiotic prophase, TATA-binding protein also first localizes in micronuclei (21). Thus, a general transcription system probably starts localizing to micronuclei at this stage. In addition, chromatin remodeling occurs in micronuclei at this stage of conjugation. Histone variant H2A.Z (formerly called hv1) and acetylated histones, hallmarks of actively transcribed chromatin, start localizing in the micronucleus during meiotic prophase (20).

Recent studies implicate an RNA interference (RNAi)-related mechanism in genome rearrangement in Tetrahymena (reviewed in reference 14). Twi1p, a member of the PPD protein family involved in diverse RNAi processes, was shown to be required for genome rearrangement (13). Twi1p interacts with and is required for the accumulation of conjugation-specific small RNAs (13, 15). Also, injection of double-stranded RNA (dsRNA) can induce ectopic DNA elimination (25). A scan RNA (scnRNA) model has been proposed that explains how IES can be eliminated in the absence of consensus sequences by an RNAi-related mechanism and accounts for the observed epigenetic regulation (13, 14). In this model the micronuclear genome is transcribed bidirectionally in early conjugation to form dsRNAs that are processed to small scnRNAs by an RNAi-like machinery. The scnRNAs then accumulate in the (parental) macronucleus where those having homology to macronuclear DNA sequences are degraded. As a result, only scnRNAs homologous to micronucleus-specific (IES or BES) sequences remain in the old macronucleus. Finally, according to this model, these IES- or BES-homologous scnRNAs move to the developing new macronucleus. There, sequences homologous to scnRNAs are identified as IES or BES and targeted for elimination. In this model, transcripts made in the early meiotic micronucleus play central roles in the genome rearrangements that occur in the newly developed macronucleus.

cnjC was identified as a gene expressed during conjugation but not in vegetative cells (12). Because cnjC mRNA appeared to be specifically expressed when crescent transcription occurred and CnjCp was similar to some subunits of RNA polymerases (RNAPs) (10), we set out to determine whether this gene was specifically involved in production of the transcripts in crescent micronuclei that gave rise to scnRNAs. However, we discovered that cnjC expression is not conjugation specific. Rather, it is the only gene in the Tetrahymena genome that encodes the conserved, third largest subunit of RNAP II, and we have, therefore, renamed this gene RPB3. Rpb3p is expressed in vegetative cells at low levels, localizes to macronuclei at all stages of the life cycle, and is required for vegetative growth. Nonetheless, Rpb3p localizes in the micronucleus only during meiotic prophase, where its high concentration and similar localization to ongoing RNA synthesis (11, 23) argue that RNAP II is involved in crescent transcription.

MATERIALS AND METHODS

Strains and culture conditions.

Wild-type B2086 and CU428 strains of T. thermophila (provided by P. J. Bruns, Cornell University) were grown in SPP medium [1% proteose peptone (Becton, Dickinson and Co., Sparks, Md.), 0.2% glucose, 0.1% yeast extract (Becton, Dickinson, and Co.), 0.003% EDTA iron(III) sodium salt (Sigma)] (6) at 30°C. For conjugation, log-phase cells of different mating types were washed, starved (16 to 24 h at 30°C), and mixed in 10 mM Tris (pH 7.5).

Construction of RPB3 knockout strains.

To make the targeting construct (see Fig. 2A), the 5′ region flanking the RPB3 coding sequence in genomic DNA was first amplified by PCR by using primers 5′FW (5′-GGCTCGAGCTAGAATAAAGATTGAATGAATTCAG-3′; XhoI site is underlined) and 5′RV (5′-GCGGATCCTGTTTTATTTTACTAAAAAGTACTCAG-3′; BamHI site is underlined) and cloned into the XhoI and BamHI sites of the pBlueScript SK(+) vector, resulting in pCC-5′. Then, the 3′-flanking region was amplified with primers 3′FW (5′-GCGGATCCGTCGACTATTTACAGTCAAGTTTCTAGC-3′; BamHI and SalI sites are underlined and in italics, respectively) and 3′RV (5′-GGACTAGTCATAATAAATGATGATGCAATTGTACAG-3′; SpeI site is underlined) and inserted into the BamHI and SpeI sites of pCC-5′, resulting in pCC-5′/3′. For unknown reasons, when we used a construct with the neo3 drug resistance marker flanked by the 5′ and 3′ RPB3-flanking sequences, we failed to obtain any transformants that had the construct in the micronucleus. In an alternative approach that was successful, a gene encoding green fluorescent protein (gfp10; 26 to 742 bp) (16) was inserted into the BamHI and SalI sites of pCC-5′/3′ as a spacer, and the neo3 cassette, conferring paromomycin resistance in Tetrahymena cells grown in the presence of Cd²⁺ (18), was inserted into the EcoRI site in the 3′-flanking region of RPB3 for selection of the transformants.

FIG. 2.

RPB3 is required for viability. (A) Diagram of the knockout construct and wild-type locus of RPB3. A sequence encoding green fluorescent protein was inserted into the RPB3 gene, replacing the coding sequence, and a drug resistance marker (neo3) was inserted into the 3′ nontranscribed sequence of RPB3 for selection of transformants. The arrows indicate the positions of primers used for the experiment described in panel B. (B) RPB3 is required for viability. Two germ line heterozygous RPB3 knockout heterokaryons (macronucleus is wild type, and one copy of the RPB3 gene in the micronucleus is disrupted) strains were mated, and individual pairs were isolated. The pairs were allowed to grow in SPP drops and then tested for paromomycin sensitivity. Genomic DNA was extracted from the 24 independent paromomycin-resistant cell lines, and their genotypes were analyzed by PCR by using the primers shown in panel A. All 24 clones isolated were heterozygotes, indicating that RPB3 is required for normal vegetative growth. (C) RPB3 knockout cells die after the first cytokinesis. Two germ line homozygous RPB3 knockout heterokaryon strains (macronucleus is wild type, and both copies of the RPB3 gene in the micronucleus are disrupted) were mated, and pairs were isolated in drops of SPP medium. The cells in each drop were counted before cells died. The results were categorized as follows: 2 cells, no cell divisions; 3 or 4 cells, one cell division; 5 to 8 cells, two cell divisions; 9 to 16 cells, three cell divisions; more than 16 cells, more than four cell divisions. Most of the exconjugants divided once and then died.

B2086 and CU428 cells were mated and the target construct was introduced 2.5 to 3.5 h postmixing as previously described (1). Two heterozygous transformants that had the drug resistance marker both in the macro- and micronucleus were obtained. The disrupted RPB3 loci in the macronucleus were eliminated by phenotypic assortment (7), resulting in two heterozygous heterokaryon strains. These heterozygous heterokaryons were crossed with a “star” strain, B*VII, to make homozygous heterokaryon strains by uniparental pronuclear transfer (7).

Viability and growth tests.

In an attempt to construct RPB3 homozygous knockout homokaryons, two RPB3 homozygous knockout heterokaryons (4 × 10⁶ cells each) were mated. The culture was refed by adding an equal amount of 2× SPP at 24 h postmixing, and a few hours later, 120 μg of paromomycin per ml and 1 μg of CdCl₂ per ml were added to select successful conjugant progeny. At 24 h postmixing, part of the culture was fixed and stained with DAPI (4′,6′-diamidino-2-phenylindole; Roche) to determine the number of exconjugants. Also, individual pairs of two RPB3 homozygous knockout heterokaryons were isolated into drops of medium at 8 to 10 h postmixing, and the cells in each drop were counted to determine the number of cell divisions that occurred before progeny died.

Construction of RPB3-HA and RPC5-HA strains.

To make the RPB3-HA construct (see Fig. 3A), the hemagglutinin (HA) coding sequence was inserted just before the stop codon of RPB3 by overlapping PCR. The following primers were used: RPB3-FW, 5′-GGCTCGAGCTCTTCCTGATGAATACATAGCAC-3′ (XhoI site is underlined); RPB3-RV, 5′-GCGGATCCGATGATGCAATTGTACAGCTAAGG-3′ (BamHI site is underlined); RPB3-HAC-FW, 5′-TATGATGTTCCTGATTATGCTtgaGCTATTTACAGTCAAG-3′ (HA coding sequence is underlined, and the stop codon of RPB3 is in lowercase letters); and RPB3-HAC-RV, 5′-TAATCAGGAACATCATAAGGATACTCGTTACCATATTCAA-3′ (HA coding sequence is underlined). First, part of the RPB3 open reading frame, followed by the HA coding sequence, was amplified with RPB3-FW and RPB3-HAC-RV. The HA coding sequence, followed by the stop codon, 3′ untranslated region, and some 3′-flanking sequences were amplified by using RPB3-HAC-FW and RPB3-RV. Then, these two products were joined by overlapping PCR with RPB3-FW and RPB3-RV. This product was digested with XhoI and BamHI and cloned into pBlueScript SK(+) vector by using XhoI and BamHI sites. Then the neo3 cassette was introduced into the EcoRI site in the 3′-flanking sequence.

FIG. 3.

HA tagging of RPB3. (A) Diagrams of the RPB3-HA construct and the wild-type RPB3 locus. The HA epitope was inserted just before the translational stop codon of RPB3. The neo3 cassette was inserted into the 3′ nontranscribed sequence. (B) Confirmation of the complete replacement of endogenous RPB3 genes by RPB3-HA. Total DNA isolated from RPB3-HA (lanes 1 to 4) and wild-type B2086 (W) cells was digested with EcoRV and SpeI and hybridized with the probe shown in panel A. Positions of the HA-tagged and wild-type (WT) genes are indicated. The faint, wild-type-size bands observed in RPB3-HA strains are from micronuclei that had wild-type RPB3 loci. (C) Rpb3p-HA expression analyzed on a Western blot. Total cell protein was prepared from log-phase or starved RPB3-HA cells or from the mating of RPB3-HA with wild-type strain B2086 at 2, 4, 6, 8, 10 and 12 h postmixing. Proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and blotted on the membrane. Rpb3p-HA was localized with anti-HA monoclonal antibody.

RPC5 was identified in the Tetrahymena genome database (scaffold 8254697; November 2003 release; http://tigrblast.tigr.org/er-BLAST/index.cgi?project=ttg). An RPC5-HA construct (see Fig. 5A) was also made by the method described above by using the following primers for PCR: RPC5-FW, 5′-GCCTCGAGAATTGAAAGAACGTTGCCCTC-3′ (XhoI site is underlined); RPC5-RV, 5′-GCGGATCCTGATAGGTCGATTATCAACGACAC-3′ (BamHI site is underlined); RPC5-HAC-FW, 5′-TTATGATGTTCCTGATTATGCTtgaGATAGATAGATAGATTTACA-3′ (HA coding sequence is underlined, and the stop codon of RPC5 is in lowercase letters); and RPC5-HAC-RV, 5′-TCAGGAACATCATAAGGATATAGGTTTAGCATAAAACTAGATGC-3′ (HA coding sequence is underlined). Then, the neo3 cassette was introduced into the AccI site in the 3′-flanking sequence. CU428 was transformed with these constructs as previously described (1). The endogenous macronuclear RPB3 or RPC5 loci were completely replaced by phenotypic assortment and selection in increasing concentrations of drug. Complete replacement was confirmed by Southern blotting.

FIG. 5.

HA tagging of RPC5. (A) Diagrams of the RPC5-HA construct and the wild-type RPC5 locus. The HA epitope was inserted just before the translational stop codon of RPC5. The drug resistance marker (neo3) was inserted into the 3′ nontranscribed sequence. (B) Confirmation of complete replacement of endogenous RPC5 genes by RPC5-HA. Total genomic DNA isolated from RPC5-HA (lanes 1 to 3) and wild-type B2086 (W) cells was digested with BglII and EcoRV and hybridized with the probe shown in panel A. WT, wild type.

Western blotting.

Whole cell protein from 2.5 × 10³ cells was separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride membranes. Blots were incubated with 1:2,000 diluted mouse anti-HA antibody (16B12; Covance, Berkeley, Calif.) in blocking solution (1% fraction V bovine serum albumin, 1% nonfat dry milk, 0.1% Tween 20 in phosphate-buffered saline [PBS]) and visualized by using a 1:10,000 dilution of horseradish peroxidase-conjugated anti-mouse immunoglobulin G (Zymed Labs Inc., South San Francisco, Calif.) in blocking solution followed by reaction with Western Lightning chemiluminescence reagent (Perkin-Elmer).

Immunofluorescence staining.

Cells were fixed in Lavdowsky's fixative (ethanol/formalin/acetic acid/water ratio of 50:10:1:39) overnight at 4°C, and the fixed cells were immobilized on poly-l-lysine (Sigma)-coated cover glasses. Samples were incubated with a 1:200 dilution of anti-HA antibody in blocking solution (3% bovine serum albumin, 10% normal goat serum [Invitrogen], and 0.1% Tween 20 in PBS) and incubated with a 1:500 dilution of fluorescein isothiocyanate-conjugated anti-mouse immunoglobulin G (Zymed Labs) in blocking solution. The samples were incubated with 10 ng of DAPI per ml in PBS, mounted, and observed.

RESULTS

RPB3 is a unique gene that encodes the third largest subunit of RNAP II.

Previous studies suggested that RPB3 (cnjC) encoded a protein related to Rpb3p, the third largest subunit of Saccharomyces cerevisiae RNAP II, and to prokaryotic RNAP (10). We identified an RPC5 homologue (the shared, fifth largest subunit of RNAPs I and III, also known as RPC40 in yeasts and RPA40 in humans) in the unpublished Tetrahymena genome database as well as in an expressed sequence tag database. Phylogenetic analysis indicated that Rpb3p and Rpc5p were related to Rpb3p and Rpc5p homologues, respectively, in other eukaryotes (Fig. 1A).

FIG. 1.

Characterizations of RPB3 (cnjC) and RPC5. (A) Phylogenetic relationships of Rpb3p and Rpc5p homologues. Amino acid sequences of the third largest subunits of RNAP II and the fifth largest, shared subunit of RNAPs I and III were used to construct a phylogenetic tree by using the neighbor-joining method. RpoD, an archaebacterial (Methanothermobacter thermautotrophicus) homologue, was used as an outgroup. Accession numbers are as follows: Rpb3p (CnjCp, Tt), S12807; RPB36A, NP_179145; CG7885-PA, NP_477419; hRPB33, NP_116558; Rpb3p (Sc), NP_012243; Rpb3p (Sp), NP_588324; At1g60620, NP_176261; Rpc40p (Sp), 094616; Rpc40p (Sc), NP_015435; CG3756-PA, NP_608885; RPA40, 015160. Rpc5p was identified in the Tetrahymena genome database. At, Arabidopsis thaliana; Dm, Drosophila melanogaster; Hs, Homo sapiens; Sc, S. cerevisiae; Sp, Schizosaccharomyces pombe; Tt, T. thermophila. (B) Expression of RNAP subunits. Total RNA from log-phase, stationary-phase, starved, and conjugating (at 2, 4, 6, 8, 10, 12, and 14 h postmixing) cells was analyzed by Northern hybridization. rpL21 encodes a ribosomal subunit and is used as a hybridization control.

We failed to identify other Rpb3p homologues in the Tetrahymena genome database, which should be nearly complete and has not failed to contain any previously known gene sequence. Thus, RPB3 is likely to be the only Tetrahymena gene encoding the third largest subunit of RNAP II.

RPB3 expression is not restricted to conjugation.

RPB3 (cnjC) was described as a gene specifically expressed during early conjugation (12). However, this result is puzzling in light of our finding that RPB3 encodes the only third largest subunit of RNAP II in Tetrahymena. Therefore, we reexamined RPB3 mRNA expression (Fig. 1B). RPB3 mRNA was expressed strongly at early conjugation stages as previously reported. RPB3 expression was similar to that of TWI1 (Fig. 1B), which is involved in small RNA accumulation and genome rearrangement (13) and is identical to a cDNA (cnjA) (GenBank accession number AY129082) identified in the same screen that identified cnjC (12). RPB3 mRNA also was detected in growing cells and at very low levels in stationary phase and starved cells (Fig. 1B). This is not likely to be due to cross-hybridization with other genes because, as described below, Rpb3p was also observed in vegetatively growing cells (see Fig. 3C and 4). Thus, RPB3 mRNA expression is not restricted to conjugating cells but also occurs in vegetative cells where its low level of expression probably explains its failure to be detected in the original studies.

FIG.4.

Localization of Rpb3p-HA. Log-phase (A) and 24-h starved (B) cells of strain RPB3-HA or mating pairs of RPB3-HA and B2086 (wild-type) cells in stage I (leptotene; C), stage II (D), stage III (zygotene; E), stage IV (F and G), and stage V (H) of meiotic prophase, pronuclear exchange (I), macronuclear anlagen (J), and nuclear alignment stage (K) were processed for indirect immunofluorescence staining. Rpb3p-HA was localized by using anti-HA monoclonal antibody. Green, anti-HA staining; blue, DAPI staining; Ma,macronucleus; Mi, micronucleus; An, anlagen or new macronucleus. Arrows indicate the core region in micronuclei (see text). See references 4 and 23 for the stages of conjugation and of meiotic prophase, respectively.

RPB1, which encodes the largest subunit of RNAP II (19), was also expressed in log-phase and stationary-phase cells and, like RPB3, was expressed at an increased level during conjugation (Fig. 1B). Thus, the expression patterns of RPB3 and RPB1 are similar, except that RPB3 is not up-regulated in stationary-phase cells, and likely reflect the pattern of expression of all RNAP II subunits. In contrast, RPC5 expression was detected in log-phase growing vegetative cells and weakly in early stages of conjugation. Thus, de novo synthesis of RNAPs I and III should be low in conjugating cells.

RPB3 is essential for vegetative growth.

If RPB3 is the only gene encoding a subunit of RNAP II in Tetrahymena, it is expected to be essential for vegetative growth. To address this question, the RPB3 gene was disrupted. The knockout construct (Fig. 2A) was introduced into conjugating Tetrahymena strains to give germ line heterozygous RPB3 knockout heterokaryon strains with micronuclei in which one of the two copies of the RPB3 gene was replaced with the knockout construct by homologous recombination (Fig. 2A). Two germ line heterozygous heterokaryon strains were mated, and the genotypes of the paromomycin-resistant progeny were analyzed by PCR (Fig. 2B). If RPB3 is not essential, one-third of the paromomycin-resistant progeny should be homozygous RPB3 knockout cells (Fig. 2B) (one-fourth of the progeny should be wild type, but they are paromomycin sensitive). However, all 24 progeny analyzed were heterozygous RPB3 knockout cells (Fig. 2B), suggesting that RPB3 is essential for vegetative growth.

To confirm this result, germ line homozygous RPB3 knockout heterokaryon strains were constructed (see Materials and Methods) with micronuclei in which both of the RPB3 genes were replaced with the knockout construct but with the wild-type RPB3 gene in the macronuclei, and their progeny were analyzed. Because both copies of the RPB3 loci in the micronucleus are disrupted in the germ line homozygous RPB3 knockout heterokaryons, all progeny derived from this mating are homozygous homokaryons whose RPB3 genes are disrupted in both macro- and micronuclear genomes. As expected if RPB3 is an essential gene, no progeny were obtained from ∼2 × 10⁶ exconjugants from a mating of two germ line homozygous RPB3 knockout heterokaryon strains, although we could easily obtain progeny from wild-type cells under the same conditions. Individual pairs of a mating between two germ line homozygous RPB3 knockout heterokaryon strains were isolated into drops of culture medium, and the number of progeny cells produced was determined. Most of the exconjugants divided once but eventually died (Fig. 2C). Because maternally expressed Rpb3p can be detected in the zygotic macronucleus (see Fig. 4), the cell division(s) observed in the progeny of the germ line homozygous RPB3 knockout heterokaryon strains probably reflects utilization of maternal Rpb3p. These results confirm that RPB3 is essential for vegetative growth, consistent with the view that RPB3 is the only gene that encodes the third largest subunit of RNAP II in Tetrahymena.

Expression and localization of Rpb3p.

To observe the localization of Rbp3p, we replaced the endogenous RPB3 gene with a transgene that encodes Rpb3p tagged with HA at the C terminus (RPB3-HA) (Fig. 3A). All endogenous RPB3 genes in the macronucleus could be replaced by RPB3-HA (Fig. 3B), demonstrating that RPB3-HA is functional.

By Western blotting (Fig. 3C), a protein whose molecular mass (36 kDa) was similar to that predicted for Rpb3p-HA (37.1 kDa) was detected by using anti-HA antibody. Rpb3p-HA was observed at all stages of the life cycle examined, consistent with the essential function of RNAP II in transcription. The amount of Rpb3p did not appear to increase greatly during conjugation and clearly did not increase to the same extent as the mRNA, possibly because the macronuclear Rpb3p (see below) synthesized by vegetative cells is stable.

Rpb3p-HA was detected by immunofluorescence staining in macronuclei in all stages of the life cycle examined (Fig. 4). Staining of micronuclei was not detected in log-phase growing cells or in starved cells (Fig. 4A and B). Strikingly, in the early stages of conjugation, Rpb3p-HA rapidly localized in the micronucleus (Fig. 4). It was first observed in early crescent micronuclei (stage I, leptotene) (Fig. 4C) just before elongation and continued to be localized in micronuclei (Fig. 4C to G) until they were fully elongated (stage IV, pachytene) (Fig. 4G). Because an RPB3-HA strain and a wild-type strain were mated in this experiment, staining was first observed only in one cell of the pair (Fig. 4C). However, Rpb3p-HA appeared rapidly in the micronucleus (Fig. 4C and D) and gradually appeared in the old macronucleus (Fig. 4E and F) of the untagged cell. In the micronucleus, a core spot lacking Rpb3p-HA staining was observed. This probably corresponds to the chromatin-dense region described previously (23). At stage III (zygotene; micronucleus elongated to almost the same length as cell length), Rpb3p-HA staining began to disappear from the central part of the crescent (Fig. 4E) until, at stage IV, only a small portion of the crescent stained at one end (Fig. 4F and G). Staining became undetectable when chromosome condensation occurred prior to meiosis I (Fig. 4H). These localizations are quite similar to the sites where crescent transcription has been observed (11, 23), indicating that RNAP II is involved in crescent transcription.

Rpb3p-HA localized in parental macronuclei throughout conjugation (Fig. 4C to J) until the new macronuclei developed, at which time it localized to newly developed macronuclei, while staining of old macronuclei rapidly disappeared (Fig. 4K).

RNAPs I and III are not detected in micronuclei.

The data above indicate that RNAP II is localized in the early meiotic micronucleus and could be involved in crescent transcription. To determine whether this is specific for RNAP II or general for RNAPs, we examined the localization of HA-tagged Rpc5p, the fifth largest subunit shared in RNAPs I and III (Fig. 5A). RPC5-HA could completely replace the endogenous RPC5 loci in the macronucleus (Fig. 5B). We have not tested whether RPC5 is an essential gene in Tetrahymena. However, because RPC5 is the only gene encoding the eukaryotic Rpc5 or Rpc40 homologue in the Tetrahymena genome and Rpc40 is indispensable for cell viability of budding yeast (9), it is likely that RPC5 is essential for cell viability in Tetrahymena. Thus, Rpc5p-HA should reflect the localization of endogenous Rpc5p.

As expected, Rpc5p-HA was localized in macronuclei in vegetative cells (Fig. 6). When an RPC5-HA strain was crossed with a wild-type strain, Rpc5p-HA remained localized in the macronucleus in only one of the paired cells and was not transferred to the other cell (Fig. 6). Rpc5p-HA was not detectable in micronuclei in any stage of the life cycle examined (Fig. 6). Thus, RNAP II is probably the only known RNAP that can be involved in micronuclear transcription during conjugation. While we cannot rule out a possibility that RNAPs I and III are localized in the micronucleus at low levels that cannot be detected by our method, the relative intensities of Rpb3-HA in macro- and micronuclei (micronuclear staining > macronuclear staining) compared to those of Rpc5p-HA (easily detectable in macronuclei but not detectable at all in micronuclei) argue that polymerase II is likely to be responsible for the RNA synthesis observed autoradiographically in crescent micronuclei, which occurs at levels comparable to that for macronuclear transcription (11, 23).

FIG. 6.

Localization of Rpc5p-HA in vegetative cells. Log-phase growing RPC5-HA cells (A) and mating RPC5-HA and wild-type (B2086) pairs in stage I (leptotene; B), stage III (zygotene; C), and stage IV (D) of meiotic prophase, first meiosis (E), second meiosis (F), prezygotic mitosis (G), and nuclear alignment stage (H) were processed for indirect immunofluorescence staining. Rpc5p-HA was localized by using anti-HA monoclonal antibody. Green, anti-HA staining; blue, DAPI staining; Ma, macronucleus; Mi, micronucleus; An, anlagen or new macronucleus; FITC, fluorescein isothiocyanate.

DISCUSSION

RPB3 (cnjC) is not conjugation specific and is required for vegetative growth.

Although RPB3 (cnjC) was first identified as a gene expressed only during conjugation, our analysis revealed that this gene is expressed at a low level in vegetative stages (Fig. 1B). RPB3 is essential for vegetative growth (Fig. 2) and is probably the only gene encoding the third largest subunit of RNAP II. Thus, we conclude that RPB3 expression is not specific to conjugation. Rather, RPB3 is expressed throughout the life cycle as a general subunit of RNAP II.

RNAP II is probably involved in crescent transcription.

Rpb3p-HA was localized in macronuclei in all stages of the life cycle examined (Fig. 3C and 4). This is also consistent with the view that RPB3 is a general RNAP II subunit. Strikingly, Rpb3p-HA appeared in micronuclei in meiotic prophase (Fig. 4). While it is possible that Rpb3p (but not all RNAP II subunits) is localized in the meiotic micronucleus and that micronuclear Rpb3p is part of a transcriptionally inactive RNAP II, this seems highly unlikely. The fact that the localization of Rpb3p-HA in micronuclei was similar to the sites where RNA synthesis was observed autoradiographically by using [³H]uridine (11, 23) argues that Rpb3p in the meiotic micronucleus is in active RNAP II. In contrast, Rpc5p-HA, the fifth largest subunit of RNAPs I and III, was not detected in the micronucleus (Fig. 6). These results suggest that RNAP II, but not RNAPs I and III, is involved in crescent transcription. Transcription of several micronucleus-specific sequences has been detected during conjugation (3). These transcripts are heterogeneous in size and probably are not polyadenylated (3), making them different from typical, macronuclear RNAP II transcripts. Although it is reasonable to expect that RNAP II localized in the micronucleus must contain subunits essential for transcription, it is not known if micronuclear RNAP II has exactly the same subunit composition as RNAP II in the macronucleus. Given that micronucleus-specific transcripts of specific sequences are extremely heterogeneous (3), it seems likely that micronucleus-specific RNAP II subunits, posttranslational modifications, or associated transcription factors might be involved in micronuclear transcription.

We observed a core region where Rpb3p-HA staining was not detected in the micronuclei during early conjugation (Fig. 4). Does this mean that not all of the micronuclear genome is transcribed? This issue bears on the nature of the micronuclear transcripts and their role in the process of DNA rearrangement. If the answer is yes, micronuclear transcription is specific, and the nature of the transcribed and untranscribed sequences and their relationship to IES and BES require additional study. However, three other explanations are also possible. First, the core region, though detectable cytologically, may not correspond to specific chromosomal regions; rather, it could represent regions that have not yet (or have already) been transcribed during meiotic prophase. Second, because the core region is a chromatin-dense region, the unstained spot may be an artifact caused by poor accessibility of the antibody. Third, the core could be aggregates of specific chromosomal regions, such as micronuclear telomeres, which are not transcribed because telomeres are added de novo during macronuclear development. Unfortunately, resolution of the sites of RNA synthesis determined by using the incorporation of [³H]uridine and light microscopic autoradiography was not sufficient to determine whether the core spot was transcribed or not.

What enables Rpb3p localization in the micronucleus?

Because the amount of RPB3 mRNA is greater in conjugating cells than in vegetative cells (Fig. 1B), one possible explanation for the accumulation of Rpb3p in micronuclei is that macronuclear sites become saturated and the excess accumulates in the micronucleus in the conjugating cells. However, two observations argue that Rpb3p is actively localized to the micronucleus. First, compared to the RPB3 mRNA level, expression of the protein, Rpb3p, was not increased significantly during conjugation (compare Fig. 1B and 3C). Second, when the RPB3-HA strain was mated with a wild-type (nontagged) strain, Rpb3p appeared in the micronucleus of the wild-type partner much faster than in the macronucleus (Fig. 4C to F). Thus, Rpb3p is probably actively transported into micronuclei or micronuclei have stronger affinity for Rpb3p than the macronuclei during meiotic prophase. The increase in RPB3 mRNA in early conjugation and its rapid appearance in micronuclei suggest that Rpb3p is translated during early conjugation. In contrast, levels of RPC5 mRNA remain low at this time, and in the absence of growth, it is likely that there is little or no translation of Rpc5p in early conjugation. Thus, it is possible that de novo synthesis of proteins is required for their localization in the micronucleus as well as for their transfer to the macronucleus of the mating partner cell. The seemingly contradictory observations that the up-regulation of RPB3 mRNA (Fig. 1B) occurs without significant increase of Rpb3p (Fig. 3C) during early conjugation may simply reflect the fact that the total amount of newly synthesized micronuclear Rpb3p is small relative to the amounts that preexist in the macronucleus.

Is micronuclear transcription required for the genome rearrangement?

In the scnRNA hypothesis, we proposed that micronuclear transcripts were processed by an RNAi-related mechanism to produce small RNAs involved in genome rearrangement. Because RPB3 is required for vegetative growth and probably for conjugation, we could not test whether RPB3 is required for the micronuclear transcription that leads to genome rearrangement. To rigorously test this hypothesis will require a mutation that prevents RNAP II from localizing in the micronucleus without affecting the transcriptional activity of RNAP II in the macronucleus. Nonetheless, for several reasons we think it highly likely that RNAP II is responsible for the micronuclear transcription that leads to IES elimination. First, as described here, RNAP II, but not RNAP I or RNAP III, appears in micronuclei when crescent transcription starts. Second, a Dicer-like protein is also localized to the micronucleus at early meiotic stages (our unpublished results). Third, double-stranded transcripts, a required substrate for the formation of small RNAs by Dicer-like enzymes, have been detected (3). Fourth, small RNAs homologous to IES have been detected (13). Fifth, a PPD protein (Twi1p) homologous to proteins involved in RNAi-like systems in other organisms is required both for the accumulation of small RNAs and for IES and BES elimination (13). Finally, dsRNA injected into conjugating cells can lead to IES elimination (25). Thus, the evidence linking RNAP II transcription in crescent micronuclei to scnRNA-mediated DNA rearrangement in Tetrahymena, though circumstantial, is compelling (14).