Role of Micronucleus-Limited DNA in Programmed Deletion of mse2.9 during Macronuclear Development of Tetrahymena thermophila

Abstract

Extensive programmed DNA rearrangements occur during the development of the somatic macronucleus from the germ line micronucleus in the sexual cycle of the ciliated protozoan Tetrahymena thermophila. Using an in vivo processing assay, we analyzed the role of micronucleus-limited DNA during the programmed deletion of mse2.9, an internal eliminated sequence (IES). We identified a 200-bp region within mse2.9 that contains an important cis-acting element which is required for the targeting of efficient programmed deletion. Our results, obtained with a series of mse2.9-based chimeric IESs, led us to suggest that the cis-acting elements in both micronucleus-limited and macronucleus-retained flanking DNAs stimulate programmed deletion to different degrees depending on the particular eliminated sequence. The mse2.9 IES is situated within the second intron of the micronuclear locus of the ARP1 gene. We show that the expression of ARP1 is not essential for the growth of Tetrahymena. Our results also suggest that mse2.9 is not subject to epigenetic regulation of DNA deletion, placing possible constraints on the scan RNA model of IES excision.

The ciliated protozoan Tetrahymena thermophila exhibits nuclear dimorphism, with a mostly transcriptionally silent diploid germ line nucleus (micronucleus) and a polyploid, transcriptionally active somatic nucleus (macronucleus) contained within the same cell. When two cells of different mating types undergo sexual development (conjugation), the micronucleus in each divides meiotically and mitotically to generate a haploid gametic nucleus that is reciprocally exchanged and fuses with that of its partner to form a zygotic nucleus. This zygotic nucleus divides, and from one of the products develops a new macronucleus. Macronuclear development involves extensive programmed DNA rearrangements, including chromosome fragmentation, DNA amplification, and the site-specific interstitial DNA deletion of internal eliminated sequences (IESs) (11, 42). Possible functions of IESs and the reasons for their elimination from the micronucleus remain unclear. One model describing the possible function of micronucleus-limited sequences suggests that they may participate in events unique to the micronucleus, such as mitosis and meiosis (11, 21). The macronucleus divides during vegetative growth by an amitotic process that is devoid of obvious chromosome condensation, and it differs from the micronucleus in the timing of DNA replication (27). Another model (the two models are not necessarily mutually exclusive) proposes that micronucleus-limited DNA is derived from mobile genetic elements (23).

In Tetrahymena, interstitial DNA deletion is responsible for the elimination of approximately 10 to 15% of the germ line genome, involving >5,000 single and multicopy elements (41, 42). The sizes of IESs in Tetrahymena range from 0.6 kb (3) to over 22 kb (39). Different IESs are generally not conserved in sequence and are AT-rich, and most are flanked by short nonconserved direct repeats (42). Alternate forms of rearrangement have been suggested to exist for approximately 25% of IESs (9), and varying degrees of microheterogeneity are observed at macronuclear junctions (1, 24, 33). IESs have not yet been found in the coding sequence of Tetrahymena, although two are located within introns (10, 20).

Several Tetrahymena IESs have been characterized at the molecular level. The deletion of the tightly linked M and R elements, which map to micronuclear chromosome 4 (4), has been shown to be controlled by flanking cis-acting sequences (5, 19). The M element uses an alternative left boundary, resulting in either a 0.6- or 0.9-kb deletion (3). A 10-bp A₅G₅ tract is present ∼45 bp outside of M on both sides of the macronucleus-retained sequence (19). This sequence is also present at the same distance from the alternative left junction that results in the smaller deletion. The sequence has been shown to be necessary and sufficient for M element deletion and to control deletion boundaries at a distance (18). To date, this A₅G₅ polypurine tract has not been found flanking any other IES. However, cis-acting sequences in flanking DNA have been shown to be necessary for the efficient and accurate deletion of mse2.9 (14, 24) and for the accurate deletion of Trl1 (34). The controlling sequences for R deletion also flank the IES on both sides (5). Although the exact identity of these sequences is unclear, they function in a manner similar to the polypurine tract of M to specify deletion boundaries at a distance, suggesting that a similar mechanism is utilized for the deletion of both M and R (5). Flanking sequences to the right of R can substitute for those flanking the left, although there is no extensive similarity between them (5).

There is evidence that the micronucleus-limited sequence of the M element contains cis-acting elements for programmed elimination (41). Yao (41) has discussed the importance of the role of internal promoting sequences (IPS) in M and R element excision. During programmed DNA deletion, the proposed function of an IPS is to target an element for deletion while cis-acting sequences in macronucleus-retained flanking sequences control the placement of deletion boundaries (11, 41). Wuitschick and Karrer (40) have extended this model by demonstrating that multiple micronucleus-limited fragments of the multiple-copy Tlr elements target their own programmed excision, suggesting that multiple redundant elimination targeting signals are distributed through the entire >22-kb element.

To examine possible functions of micronucleus-limited DNA, Chalker and Yao (8) placed the micronucleus-limited sequence of the M or R element into high-copy ribosomal DNA (rDNA) in the macronucleus of vegetatively growing cells. When they mated these strains to initiate a new round of macronuclear development, they observed an inhibition of the ability of exconjugants to delete M or R at its normal chromosomal locus. This effect was demonstrated to be sequence specific in that little interference was seen with the deletion of other IESs. Epigenetic effects on programmed DNA deletion have also been observed for Paramecium (28).

A 2.9-kb IES, mse2.9 (20), is present within the second intron of the ARP1 locus, a gene encoding a highly acidic protein of unknown function containing numerous internal repeats. mse2.9 has 81% AT content and does not appear to contain an open reading frame (ORF), and its termini are located within TTAT direct repeats. The extensive microheterogeneity found at mse2.9 macronuclear junctions initially suggested the possibility that mse2.9 could be excised by a different mechanism than that for M and R (24). However, our previous analysis suggested that the mechanism of action of the cis-acting sequences in macronucleus-retained DNA flanking mse2.9 is conserved with that of M and R elimination, and we have suggested that the same molecular mechanism is used to delete mse2.9, M, R, and likely the majority of IESs (14).

We have further analyzed the role of micronucleus-limited DNA during mse2.9 deletion and have identified a 200-bp region of the mse2.9 micronucleus-limited DNA that contains a cis-acting sequence element which is necessary for the efficient targeting of programmed excision. In an attempt to extend mechanistic links with the programmed excision of other IESs, we assessed the ability of a series of chimeric IESs, with the mse2.9 flanking sequence bounding a variety of single- and multiple-copy micronucleus-limited sequences, to be processed. The differing efficiencies of programmed deletion of these chimeric IESs, in combination with an experiment in which we removed an important cis-acting sequence from macronucleus-retained DNAs of several of the chimeras, demonstrate that the cis-acting elements in both the eliminated micronucleus-limited sequence and the macronucleus-retained flanking sequences stimulate programmed excision to different degrees, depending upon the IES in question. In addition, we have investigated the potential epigenetic regulation of mse2.9 deletion. Our data suggest that mse2.9 may not be subject to this type of regulation.

MATERIALS AND METHODS

Cell strains.

T. thermophila strains CU428 [Mpr/Mpr (VII, Mp^s)] and B2086 [Mpr⁺/Mpr⁺ (II, Mp^s)] from inbreeding line B were provided by J. Gaertig, University of Georgia. Cells were cultured axenically in 1× SPP (1% proteose peptone, 0.2% glucose, 0.1% yeast extract, 0.003% EDTA:ferric sodium salt) at 30°C as described previously (32).

Tetrahymena transformation.

Transformation of the Tetrahymena macronucleus by microinjection was done essentially as previously described (6, 8, 35, 38). Whole-cell DNA containing the micronucleus-specific sequence cloned into mature rDNA was injected directly into the macronucleus of the cells at a concentration of 1 μg/ml under a Zeiss Axiomat 35 inverted microscope and by use of an Eppendorf micromanipulator and microinjection system. After injection, individual cells were transferred into 200 μl of 1× SPP in a 96-well microtiter plate and were grown at 30°C to saturation. For the screening of transformants, 25 μl from each well was replicated into 1× SPP or 1× SPP plus 100 μg of paromomycin/ml. This pattern was continued for several days, and transformants were identified by growth to saturation in the presence of paromomycin. The procedure of Gaertig and Gorovsky (15) was used to electroporate conjugating Tetrahymena as previously described (14).

Tetrahymena conjugation.

Wild-type and transformed strains were mated as described previously (25). After 8 to 10 h of mating, individual mating pairs were transferred into 30 μl of SPP. Cells were grown to saturation and then screened for both sensitivity to paromomycin (100 μg/ml) (Pm^s) and resistance to 6-methylpurine (15 μg/ml) (Mp^r). Whole-cell DNAs were harvested from 10-ml cultures that were both Pm^s and Mp^r. We also performed this experiment by adding SPP and 6-methylpurine to individual mating pairs at 24 h postmixing. Directly selected Mp^r clones were screened for Pm^s and then were grown as described above for extraction of whole-cell DNA.

DNA purification and analysis.

Whole-cell DNA for microinjection was purified essentially as described by Gaertig and Gorovsky (15) as modified by Li and Pearlman (24). For all other experiments, the extraction of whole-cell DNA was done essentially according to the method of Gaertig et al. (17), scaled down for 10-ml cultures (14). Standard molecular biology techniques were performed as described by Sambrook et al. (36) or by following the supplier's instructions. Probes for Southern blot analysis were labeled by random priming (36) with [α-³²P]dATP (Amersham). DNA-modifying enzymes were obtained from New England Biolabs.

Plasmid construction.

The plasmids heh2.2 and heh2.2-9R have been described previously (14, 24). To make DNA constructs containing mse2.9 flanking macronucleus-retained sequences surrounding different micronucleus-limited DNA fragments, we used inverse PCR of heh2.2 in pHSS6 (24), using the primers msef/r and heh4 (14) to generate a product that was then gel purified, digested with EcoRV, and treated with alkaline phosphatase. This was ligated to the following EcoRV-digested PCR products amplified with the following primer sets (Table 1) and template DNAs: MMICF(0.6 kb) or MMICF(0.9 kb) and MMICR with pCA455-3; RMICF and RMICR with pMY404-21 (plasmids pCA455-3 and pMY404-21 were provided by Douglas Chalker, Washington University, St. Louis, Mo.); and H1F and H1R, as well as RTF and RTR, both with Tetrahymena whole-cell DNA. All pHSS6-based subclones used for this study were subsequently cloned into the NotI site of pD5H8 or pD5H8N1, as previously described (14). To generate chimeric IES clones lacking the A-rich flanking sequence, we amplified the respective pHSS6 chimeric IES subclone with the primers HEH2(ApaI) and HEH6(ApaI) (14) (Table 1). The resulting PCR products were gel purified, digested with ApaI, and ligated at a dilute concentration. To generate internal deletions in the heh2.2 micronucleus-limited sequence, we performed inverse PCR, using heh2.2:pHSS6 as the template, with the indicated primers (Table 1) to replace 200 bp of sequence with EcoRV sites. The ARPKO construct was generated by amplifying the plasmid E-15, which carries a 5.5-kb EcoRI fragment containing the entire ARP1 locus (20), with primers ARP1KOF and ARP1KOR (Table 1), digesting the amplified product with EcoRV, and ligating it to the 1.4-kb SmaI/EcoRV-digested product of p4T2-1.

TABLE 1.

Oligonucleotides used for amplification in this study

Primer name	Sequence (5′→3′)
RTRII(RV)	CATGATATCCTTAGTCTGAGTGAGTCC
5R	AATAAGATGCAAAGCAGC
3R	GCTTAAACACAACTATTC
RU4	CCATTTTCTAATTTTATAGTTAAGAAA
5J	ATTATAGGTACCATAAAC
3J	GAATTGGTTTATATATTG
MMICF(0.9 kb)	CATGATATCTAATTAGTATGGAATAAATTA
MMICF(0.6 kb)	CATGATATCTAATTGAAAGGAGGTTGCTAT
MMICR	CATGATATCAATTATTCATTCATTTTATAAT
RMICF	CATGATATCGTGATTCAAAAAAATGGT
RMICR	CATGATATCAAGGAAGAAATTTGAGAA
H1IESF	CATGATATCGTACAAAAACGGATTATTAAT
H1IESR	CATGATATCTTTTGGTCATAATATATTTAA
HEHI1	CATGATATCGAATTCTTTAAGTTTGTACTT
HEHI1REVERSE	CATGATATCAAGTACAAACTTAAAGAATTC
HEHI2	CATGATATCTCTTAATTTTAGAAAAGTAAG
HEHI2REVERSE	CATGATATCCTTACTTTTCTAAAATTAAGA
HEHI3	CATGATATCAGCTTGCTATTTTAAAGATTG
HEHI3REVERSE	CATGATATCCAATCTTTAAAATAGCAAGCT
HEHI4	CATGATATCTGGTTCTTAGTGCTCATGAAT
HEHI4REVERSE	CATGATATCATTCATGAGCACTAAGAACCA
HEH2(ApaI)	CATGGGCCCCAATATATAAACCAATTCAAT
HEH6(ApaI)	CATGGGCCCTTAACACGTTTAAAATAAAAC

PCR analysis.

PCRs were used to amplify junction sequences of processing constructs obtained by transformation as previously described (14, 24). Long-range PCRs were performed with the Expand Long Template PCR system (Boehringer Mannheim) under the conditions specified by the supplier.

DNA sequencing.

Automated cycle sequencing was done with dye-labeled dideoxy terminators and a PE/ABI 373a or 377 sequencer at the Core Molecular Biology Facility, York University, Toronto, Ontario, Canada. PCR products of mse2.9 junctions were sequenced with the RU4 primer (Table 1).

RESULTS

A region of mse2.9 micronucleus-limited DNA is required for excision.

We previously characterized the role of flanking DNA in an mse2.9 deletion (14) by utilizing the mse2.9-based construct heh2.2 (Fig. 1A). The heh2.2 construct lacks the internal 1.9-kb EcoRI fragment of mse2.9, but when transformed into conjugating cells, is processed in a manner identical to that for wild-type mse2.9 (14, 24). Mechanistic analyses of the programmed deletion of mse2.9 have previously been limited to the cis-acting elements in macronucleus-retained flanking sequences. We were therefore interested to determine if any micronucleus-limited DNA of mse2.9 is required for efficient and accurate programmed excision. The fact that heh2.2 is efficiently and accurately processed while recapitulating the microheterogeneity found at the mse2.9 macronuclear locus demonstrates that the internal 1.9-kb EcoRI fragment of mse2.9 is not required for programmed deletion (24). We generated five heh2.2-based constructs, with each containing a 200-bp internal deletion within the 1-kb heh2.2 micronucleus-limited sequence, and tested them for their ability to be processed. Four of the constructs were processed with high efficiencies, as assayed by Southern blotting (Fig. 2A to D), indicating that the 800 bp from the right boundary of the micronucleus-limited sequence of heh2.2 is dispensable for efficient programmed mse2.9 elimination. The processing was accurate in these transformants, and we observed junctional microheterogeneity both within and between transformants (Table 2; data not shown). In contrast, the deletion of 200 bp of micronucleus-limited sequence from the left heh2.2 boundary significantly decreased the efficiency of programmed excision (Fig. 2E). Two of the transformants completely failed to process the DNA (Fig. 2E, lanes 4 and 6), and the other four were significantly altered in the ability to efficiently process micronucleus-limited DNA (Fig. 2E, lanes 1 to 3 and lane 5). We determined by sequencing of the deletion junctions that the low level of processing observed for these transformants does represent accurate processing (Table 2; data not shown).

FIG. 1.

Constructs and transformation assay. (A) Organization of T. thermophila DNA at the mse2.9 locus and construction of heh2.2 and heh2.2Δ19. MAC, macronuclear DNA; MIC, micronuclear DNA; ARP1, acidic repetitive protein gene. The area inside the dotted lines indicates micronucleus-limited DNA. Outside of these lines, the macro- and micronuclear sequences are identical. Filled and open boxes denote ARP1 exons and introns, respectively. The construction of heh2.2 and heh2.2Δ19 has been described previously (14, 24). The solid line under the unprocessed heh2.2 indicates the probe used for Southern blot analysis of transformants. Intron boundaries that flank mse2.9 are indicated by asterisks. (B) Transformation assay for IES excision. The heh2.2-based sequences described in this study were cloned into pD5H8 and transformed into conjugating Tetrahymena as previously described (14, 15). H, HindIII; X, XbaI; E, EcoRI; R, EcoRV.

FIG. 2.

Southern blot analysis of transformants of five heh2.2-based clones with 200-bp internal deletions in micronucleus-limited DNA (the deleted sequence is represented in the diagram of the respective construct). Whole-cell DNAs from transformants were digested with HindIII. Southern blots were probed with the 597-bp HindIII-XbaI fragment (Fig. 1A). (A) heh2.2Δ200internal-1. (B) heh2.2Δ200internal-2. (C) heh2.2Δ200internal-3. (D) heh2.2Δ200internal-4. (E) heh2.2Δ200internal-5. H, HindIII; X, XbaI; R, EcoRV; E, EcoRI.

TABLE 2.

Sequences of macronuclear deletion junctions from transformants of heh2.2 constructs containing 200-bp micronucleus-limited internal deletions

Construct	Sequencea
heh2.2Δ200 internal-1 1-3, 6	TTTCTAGATttatttattcaa.......ataaaaGTACAAACTTAAAGAATTCGATATCGTTTT
heh2.2Δ200 internal-1 5	TTTCTAGATTTATttattcaa.......ataaaagtACAAACTTAAAGAATTCGATATCGTTTT
heh2.2Δ200 internal-2 3-5	TTTCTAGATTTATTtattcaa.......ttattattatattattAAAAAAATATT
heh2.2Δ200 internal-3 4	TTTCTAGATTTATTtattcaa.......ttattattatattattAAAAAAATATT
heh2.2Δ200 internal-3 6	TTTCTAGATTTATTtattcaa.......ttattattatattATTAAAAAAATATT
heh2.2Δ200 internal-4 2	TTTCTAGATTTATttattcaa.......ttattattatatTATTAAAAAAATATT
heh2.2Δ200 internal-4 3	TTTCTAGATTTATttattcaa.......ttattattatattATTAAAAAAATATT
heh2.2Δ200 internal-4 4	TTTCTAGATTTATTtattcaa.......ttattattatattATTAAAAAAATATT
heh2.2Δ200 internal-4 6	TTTCTAGATttatttattcaa.......ttattattATATTATTAAAAAAATATT
heh2.2Δ200 internal-5 5	TTTCTAGATGATatctcttaa.......ttattattatattATTAAAAAAATATT
heh2.2-9R background
heh2.2Δ200 internal-1 3, 6	TTTCTAGATTTATttattcaa.......ataaaagtacAAACTTAAAGAATTCGATATCGTTTT
heh2.2Δ200 internal-1 2, 4, 5	TTTCTAGATttatttattcaa.......ataaaaGTACAAACTTAAAGAATTCGATATCGTTTT

^aThe right junctions are shifted into mse2.9 micronucleus-limited sequences by ∼19 bp in the heh2.2Δ200internal-1 transformants, as the deletion constructs were synthesized by using the same right primer as the heh2.2Δ19 construct (14). XbaI restriction sites are in italics, and underlined nucleotides indicate repetitions of the bases at either end of the micronucleus-limited DNA which make the precise deletion boundaries unclear. Uppercase sequences represent the macronucleus-retained sequence at deletion junctions, while lowercase letters represent the eliminated micronucleus-limited sequence.

Chimeric IESs with mse2.9 flanking sequences excise with different efficiencies.

We previously analyzed the processing of a series of chimeric IESs that substituted flanking DNA from a variety of IESs for 103 bp of macronucleus-retained DNA flanking the right side of heh2.2 (the removal of this sequence eliminated accurate programmed excision of heh2.2 [14]). The differential ability of the added sequence to restore processing suggests that different classes of IES exist in Tetrahymena (14). To extend this observation, we asked whether mse2.9 macronucleus-retained flanking sequences are able to effect the efficient processing of other micronucleus-limited sequences. We replaced the 1-kb micronucleus-specific sequence of the heh2.2-based heh2.2Δ19 clone (14) with a variety of micronucleus-limited sequences. We used the 0.6- and 0.9-kb forms of the M element (3), the 1.1-kb R element (2), the 1.0-kb histone H1 IES (22), and 1.0 kb of micronucleus-limited DNA that encodes a portion of the reverse transcriptase (RT) domain of the REP element, a multiple-copy Tetrahymena non-long-terminal-repeat retrotransposon (14a). The heh2.2Δ19 internal deletion clone is missing 25 bp of macronucleus-retained DNA that is normally found at the right boundary of the heh2.2 micronucleus-limited sequence (the net loss is 19 bp since an EcoRV site replaces the 25 bp of deleted sequence [Fig. 1A]). This clone has previously been shown to undergo efficient programmed excision, demonstrating that this sequence does not function in programmed mse2.9 elimination. Due to the fact that cis-acting elements in macronucleus-retained DNA control the placement of deletion boundaries at a specific distance, heh2.2Δ19 deletion junctions are shifted to a micronucleus-limited sequence by a distance that is approximately equal to the amount deleted (14). Microheterogeneity was observed at macronuclear deletion junctions in cells transformed with heh2.2Δ19 (14).

The results of the transformations with heh2.2Δ19+Mmic(0.6 kb) and heh2.2Δ19+Mmic(0.9 kb) are shown in Fig. 3A. heh2.2Δ19+Mmic(0.6 kb) did not process efficiently. Only one of the six transformants analyzed showed efficient and accurate programmed deletion (Fig. 3A, lane 4), while four of the six had mostly unprocessed DNA (Fig. 3A, lanes 1, 3, 5, and 6). One of the transformants showed a combination of accurately processed DNA and an aberrantly processed smaller fragment (Fig. 3A, lane 2). Thus, the mse2.9 flanking sequences are able to direct accurate, but inefficient, processing of the 0.6-kb M element. A similar result was obtained for the heh2.2Δ19+Mmic(0.9 kb) chimera (Fig. 3A, lanes 7 to 12). Five of the six transformants contained a low level of accurately processed construct along with significant amounts of unrearranged DNA (Fig. 3A, lane 7 and lanes 9 to 12). One transformant showed almost exclusively unrearranged DNA (Fig. 3A, lane 8). Therefore, mse2.9 flanking sequences are able to direct a low level of processing of both forms of the M element. We verified that the excision products observed at the expected size represented accurate deletions by amplifying and directly sequencing macronuclear junctions from transformants (Table 3). The sequence data indicated that while the proper left macronuclear boundary was utilized, the right macronuclear boundary in these transformants was shifted into a micronucleus-limited sequence of the chimeric IES by a distance of 16 to 21 bp (Table 3) (this is a consequence of the internal deletion in the construct [14]). These transformants exhibited a low level of accurate processing, and microheterogeneity of the macronuclear junctions was observed between different transformants (Table 3). The results of transformations with the heh2.2Δ19+H1IES (Fig. 3B) construct echoed those of the M element constructs. The H1 IES was processed with a low efficiency in several transformants (Fig. 3B, lanes 1 to 3 and lane 6) and was completely unprocessed in two others (Fig. 3B, lanes 4 and 5). We verified that the excision products observed at the expected size represented accurate deletions (Table 3). Microheterogeneity was observed between different transformants and within the same transformants (Table 3; data not shown). The heh2.2Δ19+Rmic IES was processed more efficiently than the two previous chimeric IESs, with all transformants displaying efficient processing (Fig. 3C, lanes 1 to 6). Programmed excision was accurate as well as efficient, and junctional microheterogeneity was observed between as well as within transformants (Table 3; data not shown).

FIG. 3.

Southern blot analysis of transformants of heh2.2Δ19-based IES chimeras. Whole-cell DNAs from transformants were digested with HindIII, except those of the heh2.2Δ19+Rmic series, which were digested with NotI. Southern blots were probed as described for Fig. 2. (A) heh2.2Δ19+M(0.6 kb) and heh2.2Δ19+M(0.9 kb). (B) heh2.2Δ19+H1IES. (C) heh2.2Δ19+Rmic. (D) heh2.2Δ19+REP. H, HindIII; X, XbaI; R, EcoRV.

TABLE 3.

Sequences of macronuclear junctions from transformants of chimeric IESs

Construct	Sequencea
heh2.2Δ19+M (0.6kb) 4	TTTCTAGatgatatc.......attaaTTATAAAATGAATGAATAATTGATATCGTTTTA
heh2.2Δ19+M (0.9kb) 1	TTTCTAGATgatatc.......attaattATAAAATGAATGAATAATTGATATCGTTTTA
heh2.2Δ19+M (0.9kb) 2	TTTCTAGATGAtatc.......attaattatAAAATGAATGAATAATTGATATCGTTTTA
heh2.2Δ19+M (0.9kb) 3	TTTCTAGATGatatc.......attaattaTAAAATGAATGAATAATTGATATCGTTTTA
heh2.2Δ19+M (0.9kb) 4	TTTCTAGatgatatc.......attaaTTATAAAATGAATGAATAATTGATATCGTTTTA
heh2.2Δ19+M (0.9kb) 5	TTTCTAGATGatatc.......attaaTTATAAAATGAATGAATAATTGATATCGTTTTA
heh2.2Δ19+Hl IES 2	TTTCTAGATGAtatc.......gtcgtttaaATATATTATGACCAAAAGATATCGTTTTA
heh2.2Δ19+Hl IES 4	TTTCTAGatgatatc.......gtcgtTTAAATATATTATGACCAAAAGATATCGTTTTA
heh2.2Δ19+Hl IES 5	TTTCTAGATGatatc.......gtcgtttaAATATATTATGACCAAAAGATATCGTTTTA
heh2.2Δ19+R element 2	TTTCTAGATGatatc.......atctaTTTTTCTCAAATTTCTTCCTTGATATCGTTTTA
heh2.2Δ19+R element 4	TTTCTAGATGAtatc.......atctattttTCTCAAATTTCTTCCTTGATATCGTTTTA
heh2.2Δ19+R element 5	TTTCTAGATGATatc.......atctatttttCTCAAATTTCTTCCTTGATATCGTTTTA
heh2.2Δ19+R element 6	TTTCTAGatgatatc.......atctatTTTTCTCAAATTTCTTCCTTGATATCGTTTTA
heh2.2Δ19+REP 2	TTTCTAGATGatatc.......tacatAAAGGACTCACTCAGACTAAGGATATCGTTTTA
heh2.2Δ19+REP 3	TTTCTAGATGATatc.......tacatAAAGGACTCACTCAGACTAAGGATATCGTTTTA

^aThe right deletion junctions are shifted into micronucleus-limited DNA by approximately 19 bp, as the chimeric IESs contain flanking sequence derived from the heh2.2Δ19 construct (14). XbaI and EcoRV restriction sites are in italics and macronuclear-retained IES sequences not originating from mse2.9 are bold. Underlined nucleotides indicate repetitions of the bases at either end of the micronucleus-limited DNA which make the precise deletion boundaries unclear. Uppercase sequences represent the macronucleus- retained sequence at deletion junctions, while lowercase letters represent the eliminated micronucleus-limited sequence.

To test the effect of placing a multiple-copy micronucleus-limited sequence between mse2.9 flanking sequences, we generated a chimeric IES with the heh2.2Δ19 flanking sequence and 1.0 kb of an RT domain of ORF2 of the REP element, a micronucleus-limited multiple-copy non-long-terminal-repeat retrotransposon (14a). The heh2.2Δ19+REP transformants excised the micronucleus-limited sequence as efficiently (Fig. 3D, lanes 1 to 6) as wild-type heh2.2Δ19 (14) and heh2.2Δ19+Rmic (Fig. 3C). The excision was also accurate, and we observed microheterogeneity both between different transformants (Table 3) and within the same transformant (data not shown). In addition to the correctly excised DNA, each transformant contained a significant amount of a smaller product that could represent the aberrant deletion of a larger fragment than was expected (Fig. 3D).

Removal of an A-rich sequence in mse2.9 flanking sequence affects excision of heterologous IESs.

We previously identified an ∼10-bp A-rich sequence in the DNA flanking the right side of chromosomal mse2.9 that is essential for accurate mse2.9 processing (14). We replaced this A-rich region, 47 to 61 bp to the right boundary of chromosomal mse2.9, with an ApaI site in several of the chimeric constructs to generate the heh2.2Δ19Δcis series. As a control, we removed the sequence from heh2.2Δ19, and as expected (14), abolished all accurate processing (Fig. 4A). A low level of aberrantly processed DNA was observed in these transformants (Fig. 4A, lanes 2 to 4 and lane 7). In addition, when the A-rich sequence was removed from heh2.2Δ19+Mmic(0.9 kb) (Fig. 4B), very little processing was observed. We did not observe accurate R element excision from heh2.2Δ19Δcis+Rmic (Fig. 4C). However, three of the six transformants displayed significant levels of aberrantly processed DNA (Fig. 4C, lanes 3 to 5) which we confirmed did not represent accurate processing by PCR amplification and sequencing of transformant deletion junctions. Similarly, we observed no accurate processing of heh2.2Δ19Δcis+REP for the four transformants analyzed (Fig. 4D, lanes 1 to 4). Only two of the four contained a significant amount of unrearranged DNA (Fig. 4D, lanes 3 and 4). A variety of fragments were observed in all of the transformants that likely represent aberrantly processed DNAs (Fig. 4D). In addition, two of the four transformants contained significantly less hybridizing DNA than was expected for rDNA-based heh2.2 constructs (Fig. 4D, compare lanes 1 and 2 with lane C, which contains an equivalent amount of NotI-digested DNA from the heh2.2Δ19Δcis+Rmic transformant shown in Fig. 4C, lane 5).

FIG. 4.

Southern blot analysis of transformants of heh2.2Δ19Δcis-based IES chimeras. Whole-cell DNAs from transformants were digested with HindIII, except those of the heh2.2Δ19+RmicΔ47-61 and heh2.2Δ19+RTMΔ47-61 series, which were both digested with NotI. All Southern blots were probed as described for Fig. 2. (A) heh2.2Δ19Δ47-61. (B) heh2.2Δ19+M(0.9kb)Δ47-61. (C) heh2.2Δ19+RmicΔ47-61. (D) heh2.2Δ19+REPΔ47-61. Lane C contains an equivalent amount of NotI-digested DNA from the heh2.2Δ19Δcis+Rmic transformant of panel C, lane 5. H, HindIII; X, XbaI; R, EcoRV; E, EcoRI; A, ApaI.

Expression of ARP1 is not essential for growth of Tetrahymena.

The mse2.9 micronucleus locus is within the second intron of ARP1, a gene of unknown function (20). We previously demonstrated that all necessary macronucleus-retained cis-acting sequences necessary for mse2.9 excision are contained within intronic sequences (14). Since mse2.9 is located in the second intron of ARP1, there is a possibility that the inhibition or interference of programmed mse2.9 deletion could affect Arp1p functioning. We addressed this by generating an ARP1 knockout strain to determine if its expression is essential for the growth of Tetrahymena. Using gene replacement by homologous recombination, we generated somatic ARP1 knockout strains with a construct containing a neomycin resistance cassette that replaces almost the entire ARP1 coding sequence (Fig. 5A). The wild-type allele was replaced entirely with the knockout allele by phenotypic assortment (Fig. 5B), indicating that ARP1 expression is not essential for the vegetative growth of Tetrahymena. Simple modular architecture research tool (SMART) analysis of the ARP1p predicted protein sequence indicated the presence of a hydrophobic signal peptide in the N-terminal 20 amino acids, suggesting that Arp1p may be a secreted protein.

FIG. 5.

ARP1 expression is not essential for growth. (A) Design of an ARP1 knockout construct. The probe for Southern blot analysis is indicated by a solid bar. The solid arrow represents the 1.4-kb EcoRV/SmaI neo cassette from p4T21 (16). (B) Southern blot analysis of HaeIII-digested whole-cell DNAs purified from wild-type CU428 (C) and two ARP1 knockout strains. Transformed strains were grown for >100 generations in 100 μg of paromomycin [lanes 1(+) and 2(+)]/ml, at which point single cells from each transformant were isolated and grown for ∼100 generations in 1× SPP [lanes 1(−) and 2(−)]. H, HindIII; X, XbaI; E, EcoRI.

Presence of deletion clone heh2.2-9R in the macronucleus of conjugating cells does not affect programmed mse2.9 deletion.

We previously established that mechanistic links exist in the programmed excision of several different IESs (14). The results of the present study suggest that mechanistic links may extend to include a requirement for sequence elements within micronucleus-limited DNA. Chalker and Yao (8) showed that by ectopically loading the macronucleus of growing cells with micronucleus-limited DNA and then forcing them to mate, it was possible to interfere with the subsequent deletion of the homologous IES in the next round of macronuclear development. Similarly, we attempted to interfere with programmed mse2.9 deletion. Since this effect required only micronucleus-limited DNA, not macronucleus-retained flanking sequences, we performed an experiment with heh2.2Δ19+Mmic(0.6 kb), which we have shown is not processed efficiently when introduced into conjugating cells (Fig. 3A). We purified whole-cell DNA from a transformant that completely failed to delete the IES and microinjected it into the macronucleus of growing CU428 and B2086 strains of Tetrahymena (Fig. 6). The transforming DNA was mature, linear rDNA that carried unprocessed heh2.2Δ19+Mmic(0.6 kb) DNA. CU428 is a heterokaryon that carries a homozygous dominant allele for resistance to 6-methylpurine (Mp^r) in its micronucleus. Since CU428 does not carry the Mp^r allele in its macronucleus, it is sensitive to 6-methylpurine (Mp^s), allowing the direct selection of all exconjugants by their Mp^r phenotype. Southern blotting of the whole-cell DNA isolated from the microinjected transformants showed that the transformants carried the injected DNA at high copy numbers in the macronucleus (Fig. 7A). Individual mating pairs were cloned from conjugating B2086:heh2.2Δ19+Mmic(0.6 kb) and CU428:heh2.2Δ19+Mmic(0.6 kb) transformants. Whole-cell DNAs were isolated from exconjugant cultures that successfully completed conjugation and were subjected to Southern blotting to analyze their ability to excise the M element (Fig. 7B). The cells containing the heh2.2Δ19+Mmic(0.6 kb) construct in their macronuclei produced exconjugants that were significantly inhibited in the ability to excise the M element (Fig. 6B). The M element is always excised efficiently from exconjugants of wild-type matings (3). This result confirmed that the epigenetic inhibition of DNA deletion is not a strain-specific phenomenon in Tetrahymena and that the presence of a nonhomologous mse2.9 flanking sequence does not interfere with the M-induced inhibition of M deletion. Chalker and Yao (8) previously showed that although this epigenetic effect was largely sequence specific, the programmed deletion of unrelated IESs could be affected to a degree. To see if this was the case for the heh2.2Δ19+Mmic(0.6 kb) construct, we examined the processing of the closely linked R element (Fig. 7C) and of unlinked mse2.9 (Fig. 7D). Although programmed excision of the linked R element was affected to a small degree (Fig. 7C, lanes 3, 4, 8, 9, 13, and 14), that of unlinked mse2.9 was not affected (Fig. 7D). Although the clone used for this assay contained mse2.9 flanking sequence, there was no effect on programmed mse2.9 deletion.

FIG. 6.

Schematic outline of epigenetic experiment. See the text for details.

FIG. 7.

The M element within mse2.9 flanking sequence inhibits its own programmed excision. (A) Southern blot analysis of HindIII-digested whole-cell DNAs from strains of B2086 (B) and CU428 (C) microinjected with whole-cell DNA containing heh2.2Δ19+M(0.6 kb) in rDNA and probed as described for Fig. 2. (B) Southern blot analysis of whole-cell DNAs purified from exconjugant clones digested with HindIII and probed for M element excision. (C) Southern blot analysis of whole-cell DNAs purified from exconjugant clones digested with EcoRI and BglII and probed for R element excision. *, aberrant processing. (D) Southern blot analysis of HaeIII-digested whole-cell DNAs probed for mse2.9 excision as described for Fig. 2.

To examine whether mse2.9 is subject to the epigenetic effect on DNA deletion, we performed the experiment using the heh2.2-9R deletion clone that was previously shown not to process (24). We could not use wild-type heh2.2 for this experiment, as unprocessed heh2.2 is never observed in transformants (24; J. Fillingham and R. Pearlman, unpublished data). In contrast, Tetrahymena cells transformed with the heh2.2-9R construct contain in their macronucleus the full micronuclear sequence of heh2.2 (minus 9 bp at the right boundary) cloned into their rDNA. We generated Tetrahymena strains B2086 and CU428, which carry the heh2.2-9R clone at high copy numbers in the macronucleus (Fig. 8A). Individual mating pairs from conjugating B2086:heh2.2-9R and CU428:heh2.2-9R transformants were cloned as described above. Whole-cell DNAs were isolated from exconjugant cultures and subjected to Southern blotting for an analysis of their ability to excise mse2.9 (Fig. 8B), the M element (Fig. 8C), and the R element (Fig. 8D). In contrast to the results observed for the heh2.2Δ19+Mmic(0.6 kb) transformants, we saw no significant effect on the deletion of the mse2.9 locus (Fig. 8B) or the unlinked M and R elements (Fig. 8C and D).

FIG. 8.

Presence of the heh2.2-9R clone in parental strains does not interfere with programmed DNA deletion. (A) Analysis of microinjected parental strains as described for Fig. 6A. (B) Analysis of mse2.9 structure as described for Fig. 9. (C) Analysis of M element as described for Fig. 6B. (D) Analysis of R element as described for Fig. 6C.

Presence of heh2.2-9R in the macronucleus of conjugating cells does not affect programmed deletion of heh2.2Δ200internal-1 in a processing assay.

To address the possibility that its chromosomal position could influence the potential epigenetic regulation of mse2.9 deletion, we used conjugating B2086:heh2.2-9R and CU428:heh2.2-9R (Fig. 8A) in a transformation assay (Fig. 9A). If the underlying mechanism of the epigenetic effect was common to all IESs, we might have observed an epigenetic effect on heh2.2 deletion in a different chromosomal environment, in this case the ∼21-kb rDNA minichromosome. We electroporated the heh2.2Δ200internal-1 clone, which processes accurately and efficiently in wild-type B2086 and CU428 conjugations, into conjugating B2086:heh2.2-9R and CU428:heh2.2-9R (Fig. 2A). In this case, transformants were selected with paromomycin and 6-methylpurine to ensure the exclusive survival of exconjugants (Fig. 9A). An analysis of these transformants showed that they accurately processed the heh2.2Δ200internal-1 constructs with the same efficiency and accuracy as wild-type cells (Fig. 9B; Table 2). This result indicates that the programmed deletion of mse2.9 may not be subject to epigenetic effects on its programmed deletion.

FIG. 9.

Presence of heh2.2-9R in the macronucleus of parental strains does not interfere with processing of heh2.2Δ200internal-1 in processing assay. (A) Cartoon showing the modified transformation assay. (B) Southern blot analysis of HindIII-digested whole-cell DNAs from transformants of heh2.2Δ200internal-1 probed as described for Fig. 2.

DISCUSSION

A cis-acting sequence in micronucleus-limited DNA is required for targeting of efficient programmed mse2.9 deletion.

We have identified a 200-bp region of micronuclear sequence at the left side of the micronucleus-limited sequence in heh2.2 that contains an important cis-acting element for the programmed deletion of mse2.9. The fact that a low level of accurate programmed elimination occurs in cells transformed with this construct suggests that the sequence does not function to control the placement of deletion boundaries, as has been shown for cis-acting sequences in the DNA flanking mse2.9 (14, 24). Yao (41) has discussed the importance of the role of IPSs in IES excision. IPSs are described as sequence elements existing within micronucleus-limited DNA that are required in cis for programmed deletion. We suggest that the 200-bp stretch of mse2.9 comprises or contains an IPS. The 200 bp of micronucleus-limited sequence at the left side of heh2.2 has also been shown to contain a cis-acting sequence element that, in the absence of a wild-type heh2.2 macronucleus-retained flanking sequence, directs alternate processing (24). It will be informative to examine a possible relationship between these two functionally different cis-acting elements to see, for example, if they are genetically separable. In addition, the fact that a low level of processing was observed in some transformants using this construct indicates the possible presence of an additional IPS within the remaining 800 bp of micronucleus-limited heh2.2 sequence.

cis-Acting elements in flanking sequences and micronucleus-limited DNA have different IES-specific abilities to promote programmed excision.

The current model of IES excision suggests that IPSs target elements for deletion while the cis-acting sequences in macronucleus-retained flanking sequences control the proper placement of deletion boundaries (11, 41). The results of experiments described here and in our previous study (14) suggest that the programmed excision of mse2.9 follows an identical pattern and that both of these functionally distinct classes of cis-acting sequences are required for efficient and accurate programmed excision. Using chimeric IESs with mse2.9 macronucleus-retained flanking sequences surrounding different micronucleus-limited sequences, we observed variability in the efficiency of programmed excision. Chimeric IESs with the mse2.9 flanking sequence surrounding versions of the M element and the histone H1 IES are processed with relatively low efficiencies, while the R element, mse2.9 (24), and the RT domain from the REP element are processed with relatively high efficiencies. We suggest that the variable efficiencies of processing of these constructs in transformants are a direct result of cis-acting elements within the micronucleus-limited sequence of the different IESs. Specifically, we suggest that the micronucleus-limited sequence targets its programmed deletion with an efficiency that depends on the particular micronucleus-limited sequence in question. For example, relative to mse2.9, both forms of the M element and the H1 IES do not strongly target their own excision. Alternatively, the R element, mse2.9, and the REP element target their programmed excision relatively strongly.

The relative strength of cis-acting signals in micronucleus-limited DNA is not the only determinant of the efficiency of processing. The efficiency of processing also depends on the strength of cis-acting sequences in macronucleus-retained flanking DNA that are utilized to determine deletion boundaries. For example, it has been shown that the R element does not process with a high efficiency when using its own flanking sequences in the processing assay (5) and occasionally exhibits some unrearranged DNA at its wild-type locus (Fillingham and Pearlman, unpublished data). Since we suggested that the R micronucleus-limited sequence targets its deletion with a relatively high efficiency, we hypothesize that the stimulation of programmed deletion by the micronucleus-limited sequence of a particular IES is modified by the relative strength of the cis-acting sequences in its flanking macronucleus-retained DNA. Thus, the R element itself strongly targets its excision, but it has relatively weak cis-acting signals in its macronucleus-retained flanking sequences. The relative weakness of the R element flanking sequence is demonstrated by the fact that the R flanking sequence only weakly rescues a processing-deficient heh2.2 construct (14). This may also explain the functional redundancy observed in the R flanking sequence (5): since R micronucleus-limited DNA strongly enhances its own programmed deletion relative to M and H1 IESs, it is more compelled to use the cryptic cis-acting sequences in macronucleus-retained flanking DNA as they become available. Consistent with this, we observed a larger amount of aberrant processing in heh2.2Δ19+RmicΔcis transformants (Fig. 4C) than in heh2.2Δ19+M(0.9 kb)Δcis transformants (Fig. 4B). Thus, IESs with relatively strong targeting cis-acting signals in their micronucleus-limited DNA are more likely to utilize cryptic cis-acting signals in the macronucleus-retained sequence.

Relative to those of mse2.9 and the R element, the micronucleus-limited sequences of the M element and the H1 IES do not contain strong targeting cis-acting signals. However, our previous results suggested that the relative strengths of the cis-acting elements in their flanking sequences may differ (14). The H1 IES is predicted to contain stronger cis-acting sequences than is the M or R element (14). One inconsistency of this model for describing the behavior of cis-acting elements in IES excision is the in vivo behavior of the M element. The M element has not been observed in its unprocessed form at its chromosomal locus in wild-type strains (3, 8). However, wild-type M element constructs are frequently not processed with high efficiencies in processing assays (18, 19), a fact that is in agreement with our interpretation of the relative strengths of the cis-acting elements in the M element DNA. Thus, there may be an additional degree of regulation of M element excision at its chromosomal locus.

The multiple-copy RT domain of the REP fragment processes with a relatively high efficiency within the heh2.2Δ19 flanking sequence, and we suggest that it targets its own programmed excision with a high efficiency. Accordingly, transformants of the heh2.2Δ19+REPΔcis construct process their DNA aberrantly, and several of the transformants do not contain significant amounts of unrearranged heh2.2Δ19+REPΔcis DNA (Fig. 3D). We suggest that the reason for this is that targeting of cis-acting signals in the REP element DNA stimulates its processing to such a high degree that it forces the use of cryptic cis-acting sequences in macronucleus-retained flanking DNA to an even greater degree than the R element. The behavior of the heh2.2Δ19+REPΔcis construct in the processing assay echoes that of several Tlr-based constructs used by Wuitschick and Karrer (40). In their analysis of the requirement for micronucleus-limited DNA during programmed excision of the multiple-copy Tlr element, they demonstrated that Tlr DNA was efficiently deleted when its flanking sequences were replaced with DNA from regions of the genome that are not normally associated with rearrangement (40). An interpretation of this result is that Tlr element micronucleus-limited DNA contains targeting cis-acting signals that very strongly enhance programmed excision and that in the absence of strong flanking sequences it will, like REP element DNA in the context of heh2.2Δ19Δcis, force processing through the use of any cryptic flanking sequence it can find. Wuitschick and Karrer (40) interpreted their data to suggest that there is a strong correlation between sequence copy number and DNA elimination. In support of this, Yao et al. (43) demonstrated that programmed DNA elimination can be induced by increasing the copy number of a particular sequence. Using this criterion, we expect that the REP element, which is present in the Tetrahymena micronucleus at an estimated copy number of 40 to 175 (14a), similar to the Tlr element, contains targeting cis-acting signals in its micronucleus-limited DNA that strongly enhance programmed excision. Wuitschick and Karrer (40) have also demonstrated that the Trl multiple-copy family of micronucleus-limited sequences contains elimination targeting signals that are likely distributed throughout the entire sequence. We suggest that targeting cis-acting signals are similarly distributed throughout the REP element DNA. It is possible that IESs that are present as a single copy in the micronucleus, such as mse2.9, contain different numbers of IPSs that promote excision, while the entire sequence of multiple-copy micronucleus-limited DNA functions as an IPS. Thus, the multiple-copy REP element and the Tlr family strongly enhance their own deletion as a consequence of their repeated nature in the micronucleus.

Deletion of mse2.9 is not epigenetically inhibited.

The epigenetic inhibition of IES excision has been described for both Paramecium (30) and Tetrahymena (8). The effect is sequence specific in that interference is observed at the IES locus that is ectopically represented in the macronucleus. The inhibitory effect is observed with micronucleus-limited DNA with no flanking sequence, not when the macronucleus-destined flanking sequence is used by itself. Two models have been presented to describe the epigenetic effect on DNA deletion. Meyer and Duharcourt (29) have argued that the sequence specificity observed was likely the result of a trans-nuclear communication between the parental macronucleus and the developing anlagen. They pointed out that the epigenetic inhibition of DNA deletion resembled homology-dependent gene silencing in plants, at the time an enigmatic phenomenon (26), and suggested that the molecule mediating this communication was a nucleic acid that could act as a template to guide IES excision. Chalker and Yao (8) suggested that IESs play a passive role in this epigenetic effect, titrating trans-acting factors that are specific for individual IESs. The recently developed scan RNA model (31) provides a mechanistic framework to understand this epigenetic effect and brings together the two models. Transcription of an IES during meiosis results in the generation of small scan RNAs (scnRNAs) that travel to and scan the parental macronucleus for the presence of homologous sequences. In this model, the parental macronuclear sequence titrates (8) the specific trans-acting scnRNAs, leaving those that are unpaired to travel to the anlagen, where a nucleic acid, the scnRNA, guides IES excision (29).

Our results indicate that, similar to the case in Paramecium (12), the epigenetic inhibition of IES excision may not be a universal phenomenon in Tetrahymena. We were able to inhibit the programmed excision of the M element by using a chimeric M element but were unable to inhibit that of mse2.9 by using the heh2.2-9R deletion clone. The chromosomal position of ARP1 presents one potential problem in interpreting the results of this experiment. If ARP1 were an essential gene and the retention of mse2.9 in the macronucleus interfered with processing of the ARP1 mRNA, then there would be selection for any cells containing deleted mse2.9. In this case, the apparent lack of an epigenetic effect on mse2.9 processing would be artifactual. However, this is not the case, as we have demonstrated that the expression of ARP1 is not essential for growth. In addition, by using Southern blot analysis, we analyzed the whole-cell DNAs extracted after a B2086:heh2.2-9R and CU428:heh2.2-9R conjugation had finished but before selection for growth. We observed no effect on the programmed deletion of mse2.9 (data not shown).

One additional caveat to our interpretation of the lack of an epigenetic effect on programmed mse2.9 excision is that it is possible that the full 2.9-kb IES is required as a template for the titration of scnRNAs. We have attempted to perform this experiment using the full 2.9-kb IES cloned into rDNA but have had difficulty generating the required strains with appropriate macronuclear copy numbers (Fillingham and Pearlman, unpublished data). To attempt to address this issue, we changed the chromosomal context, testing the ability of the heh2.2Δ200internal-1 construct to process by using conjugating cells loaded with heh2.2-9R in the macronucleus. These cells loaded with heh2.2-9R in the macronucleus contained all of the micronucleus-limited sequence that is present in the heh2.2Δ200internal-1 construct. If scnRNAs direct mse2.9 excision, this construct should not have processed in this mutant background, as the required scnRNAs would have been titrated and therefore not available to guide processing.

There is a precedent for the differences between mse2.9 and the M and R elements. The meiosis-specific transcription of a particular IES is proposed to comprise the first step of the scnRNA model. Chalker and Yao (7) described two classes of micronucleus-limited sequence defined by their temporal transcription profiles during early conjugation. The first class, represented by the M and R elements, begins to be transcribed in early meiosis and is transcribed through to the time period corresponding to early macronuclear development (7). The other class is represented by the multicopy pTt2512 element and is transcribed strongly, not in meiosis, but in the time period corresponding to early macronuclear development. Chalker and Yao have placed mse2.9 into this latter class (7). According to the scnRNA model, if an IES is not transcribed during meiosis, it cannot be processed to a scnRNA and consequently will not be available to scan the parental macronucleus. We suggest that IESs transcribed solely during early macronuclear development may require a different targeting mechanism than that of the first class of IESs (M and R elements). Taverna et al. (37) used chromatin immunoprecipitation to show that mse2.9 is associated with the methyl-H3K9 modification as well as with Pdd1p during the time period corresponding to IES excision. This result is consistent with our previous conclusion that a common mechanism is used to delete IESs from Tetrahymena (14). We suggest that Tetrahymena uses more than one mechanism to generate the methyl-H3K9 modification, which then attracts a common IES excision machinery. The programmed deletion of mse2.9 is unaffected by the addition of the histone deacetylase inhibitor trichostatin A (TSA) during macronuclear development (13). It is tempting to speculate that the different response of mse2.9 to TSA than those of M and R elements is related to our observation that mse2.9 is not subject to epigenetic inhibition of deletion. Could scnRNAs be required at a subset of IESs to target histone deacetylation? Our results raise the possibility that mse2.9 may not require histone deacetylation for the generation of the methyl-H3K9 epigenetic marker.