Histone H3 lysine 9 methylation is required for DNA elimination in developing macronuclei in Tetrahymena

Abstract

Genome-wide DNA elimination accompanies development of the somatic macronucleus from the germ-line micronucleus during the sexual process of conjugation in the ciliated protozoan Tetrahymena thermophila. Small RNAs, referred to as "scan RNAs" (scnRNAs), that accumulate only during conjugation are highly enriched in the eliminated sequences, and mutations that prevent DNA elimination also affect the accumulation of scnRNAs, suggesting that an RNA interference (RNAi)-like mechanism is involved in DNA elimination. Histone H3 that is methylated at lysine 9 (K9) is a hallmark of heterochromatin and, in Tetrahymena, is found only in developing macronuclei (anlagen) in association with chromatin containing the sequences undergoing elimination. In this article, we demonstrate that a mutation in the TWI1 gene that eliminates the accumulation of scnRNAs also abolishes H3 methylation at K9. We created mutant strains of Tetrahymena in which the only major H3 contained a K9Q mutation. These mutants accumulated scnRNAs normally during conjugation but showed dramatically reduced efficiency of DNA elimination. These results provide strong genetic evidence linking an RNAi-like pathway, H3 K9 methylation, and DNA elimination in Tetrahymena.

Fig. 2.

Generation of H3 K9Q mutants. (A) Restriction maps of the endogenous H3 and H4 loci in Tetrahymena and their corresponding knockout constructs. (B) Creation of strains with mutant histone H3 by somatic rescue of the two germ-line knockout heterokaryon strains ΔH3/H4. The strains were generated by crossing the appropriate single germ-line knockout strains and further genetic manipulation. The conjugation progeny of the ΔH3/H4 strains were not viable unless rescued with a construct carrying functional H3 and H4 genes.

Fig. 1.

TWI1 is required for H3 K9 methylation. TWI1 knockout cells (ΔTWI1-S2 ×ΔTWI1-S4) or wild-type cells (CU428 × B2086) were mated. Cells collected 14 h after mixing were fixed, processed for immunofluorescence staining with anti-dimethyl K9 H3 antibody, and stained with DAPI.

Fig. 3.

H3 K9 methylation is not detectable in H3 K9Q mutants. The cells from 14 h after mixing were fixed and processed for immunofluorescence analysis using anti-dimethyl K9 H3 primary antibody and FITC-conjugated secondary antibody. Cells were also stained with DAPI. AN, anlagen; OM, old macronuclei; MA, macronuclei in nonmating cells.

Fig. 4.

Absence of K9 methylation does not affect accumulation of scnRNAs or Pdd1p. (A) scnRNA accumulation in conjugating K9Q or wild-type rescued cells. Total RNA was extracted from mating cells (at the indicated hours after mixing) and separated in 12% acrylamide–urea gels and stained by ethidium bromide. Position of the scnRNAs (Small RNA) is indicated by the arrow. (B) Localization of Pdd1p in conjugating K9Q or wild-type rescued cells. Cells were fixed 14 h after mixing, processed for immunofluorescence analysis by using primary anti-Pdd1p antibody and FITC-conjugated secondary antibody, and stained with DAPI. AN, anlagen; OM, old macronuclei; MA, macronuclei in nonmating cells.

Fig. 5.

Analysis of M-element processing in the progeny of K9Q and wild-type rescued cells. (A) Alternative processing of the M element. The micronuclear-specific form of the M element can be processed into either of two macronuclear-specific forms, M-long and M-short, which occur in about equal frequency (35). Both elements have the same 3′ break site. In the M-long region, 0.6 kb of sequence 5′ of the break point is eliminated. In the M-short region, 0.9 kb of sequence 5′ of the break point is eliminated. The M element is flanked by HindIII sites. Primers 1 (M5′-1) and 2 (M3′-1) were used in PCR analysis of M processing. Dashed lines indicate the regions that are eliminated in each of the two processing alternatives. (B) Schematic representation of the rationale for detecting M-element processing in conjugating rescued cells. The parental rescued cells were assorted, and clones containing only the M-long form of the M element were used for conjugation. The appearance of the M-short form in conjugant progeny is indicative of the processing of the M element during macronuclear development. (C) Southern blot analysis of M-element processing in conjugating K9Q and wild-type rescued cells. Genomic DNA was isolated from K9Q and wild-type rescued cells at different times of conjugation and digested with HindIII. The blot was probed with the M-short PCR product (see A and D), which detects both the M-long and M-short fragments with equal affinity. (D) PCR analysis of M-element processing in conjugating K9Q and wild-type rescued cells. PCR was conducted on conjugating whole-cell lysates collected at 2-h intervals 12–36 h after mixing. Primers 1 and 2 were used to examine the relative abundance of the two alternative macronuclear-specific forms, M-long and M-short.