Localization and expression of mRNA for a macronuclear-specific histone H2A variant (hv1) during the cell cycle and conjugation of Tetrahymena thermophila

hv1 is a histone H2A variant found in the transcriptionally active Tetrahymena macronucleus but not in the transcriptionally inert micronucleus. This, along with a number of other lines of evidence, suggests that hv1 is associated with active genes. We have used a cDNA clone as a probe to study hv1 mRNA accumulation throughout the cell cycle and during conjugation. In situ hybridization to glutaraldehyde-fixed growing cells, whose position in the cell cycle was determined by size and morphology, showed that hv1 message is present throughout the cell cycle. The message was uniformly distributed in these vegetative cells. Compared with four other Tetrahymena histone genes studied to date (S. -M. Yu, S. Horowitz, and M. A. Gorovsky, Genes Dev., 1:683, 1987; M. Wu, C. D. Allis, and M. A. Gorovsky, Proc. Natl. Acad. Sci. USA 85:2205, 1988), hv1 mRNA is the only one that does not show a pattern of accumulation during the cell cycle that could explain the nuclear localization of its encoded protein. Thus, either hv1 or some molecule with which it associates contains a macronuclear-specific targeting sequence or there exists a cell cycle-regulated event that restricts its translation to the macronuclear S phase. In situ hybridization to conjugating cells revealed that hv1 message amounts increase just prior to macronuclear development and decline precipitously after the cells separate. The hv1 message showed no marked subcellular localization and is, therefore, unlikely to play a role in the cytoplasmic determination known to occur during macronuclear development.

Direct measurement of tubulin and bulk message distributions on polysomes of growing, starved and deciliated Tetrahymena using RNA gel blots of sucrose gradients containing acrylamide

A method was developed using sucrose gradients containing acrylamide which greatly simplifies the measurement of the polysomal distribution of messages. After centrifugation, the acrylamide was polymerized, forming a "polysome gel". RNA gel blots of polysome gels were used to determine the polysomal distributions of alpha-tubulin and total polyadenylated mRNA in growing, starved (nongrowing) and starved-deciliated Tetrahymena and the number of messages loaded onto polysomes was calculated. These measurements indicated that the translational efficiencies of alpha-tubulin mRNA and total polyadenylated mRNA are largely unaffected when the rates of tubulin and total protein synthesis vary dramatically. Thus, differential regulation of alpha-tubulin mRNA translation initiation does not contribute to the greater than 100-fold induction of tubulin synthesis observed during cilia regeneration and in growing cells. The major translation-level process regulating tubulin synthesis in Tetrahymena appears to be a change in message loading mediated by a non-specific message recruitment or unmasking factor.

Drosophila has a single copy of the gene encoding a highly conserved histone H2A variant of the H2A.F/Z type

The Tetrahymena histone H2A variant designated hv1 is localized exclusively in the transcriptionally active macronucleus and is absent from the quiescent micronucleus (1). A cDNA clone of the hv1 gene (2) was used to screen a Drosophila cDNA library. A cross-hybridizing clone was recovered and shown by sequence analysis to code for a protein homologous to hv1 as well as to the chicken H2A variant, H2A.F (3), the sea urchin H2A variant, H2A.F/Z (4) and the mammalian H2A variant H2A.Z (5). Southern analysis of Drosophila genomic DNA indicates that the H2AvD (H2A variant Drosophila) gene is present in one copy. In situ hybridization places the locus at 97CD on chromosome 3, while the S-phase regulated histone genes are on chromosome 2 (6). Thus the Drosophila H2A variant should be accessible to genetic analysis, which will enable its function to be determined.

A temperature-sensitive mutation affecting cilia regeneration, nuclear development, and the cell cycle of Tetrahymena thermophila is rescued by cytoplasmic exchange

A temperature-sensitive mutation was isolated that blocks cilia regeneration and arrests growth in Tetrahymena thermophila. Protein and RNA synthesis and ATP production appeared to be largely unaffected at the restrictive temperature, suggesting that the mutation is specific for cilia regeneration and growth. At the restrictive temperature, mutant cells arrested at a specific point in the cell cycle, after macronuclear S phase and shortly before micronuclear mitosis. Arrested cells did not undergo nuclear divisions, DNA replication, or cytokinesis, so the mutation appears to cause true cell cycle arrest. Surprisingly, the mutation does not appear to affect micronuclear mitosis directly but rather some event(s) prior to micronuclear mitosis that must be completed before cells can complete the cell cycle. The cell cycle arrest was transiently complemented by wild-type cytoplasm exchanged during conjugation with a wild-type cell. Each starved, wild-type cell apparently contained enough rescuing factor to support an average of six cell divisions. Thus, this mutation affects assembly and/or function of at least one but not all of the microtubule-based structures in T. thermophila.

Cell-cycle regulation as a mechanism for targeting proteins to specific DNA sequences in Tetrahymena thermophila

Transcriptionally active macronuclei and transcriptionally inert micronuclei of the ciliated protozoan Tetrahymena thermophila contain similar DNA sequences but have very different histones associated with the linker regions of chromatin. In situ hybridization showed that a gene coding for micronuclear linker histone is expressed only in association with micronuclear DNA replication, whereas the gene for macronuclear H1 histone is expressed during macronuclear (but not during micronuclear) S phase. These results indicate that cell-cycle regulation plays an important role in directing proteins to the appropriate nucleus in Tetrahymena and that the replication-expression model [Gottesfeld, J. & Bloomer, L. S. (1982) Cell 28, 781-791; Wormington, W. M., Schlissel, M. & Brown, D. D. (1983) Cold Spring Harbor Symp. Quant. Biol. 47, 879-884] for establishing appropriate transcriptionally active or repressed chromatin complexes during DNA replication is generally applicable.

Sequence and properties of the message encoding Tetrahymena hv1, a highly evolutionarily conserved histone H2A variant that is associated with active genes

hv1 is a histone H2A variant found in the transcriptionally active Tetrahymena macronucleus, but not in the transcriptionally inert micronucleus. hv1 also contains antigenic determinants conserved in the histone complements of representatives of all four eukaryotic kingdoms. A cDNA clone encoding hv1 has been isolated and sequenced. Comparison of the derived protein sequence of hv1 with that of the chicken variant H2A.F and the sea urchin variant H2A.F/Z reveals remarkable homology in all but the extreme amino- and carboxy-termini and a small region in the conserved core. Putative regions of conserved antigenicity are discussed. Evidence is presented that suggests that hv1 is a single-copy, intron-containing gene that encodes a polyadenylated message. Unusual features in the 3' flanking sequence and in codon usage are also described. Evidence is also presented showing that hv1 message amounts are ten-fold greater in growing cells than in starved cells.

Scheduled and unscheduled DNA synthesis during development in conjugating Tetrahymena

Autoradiography has been used to confirm and to extend previous microspectrophotometric studies (Doerder and DeBault, 1975) on the timing of DNA synthesis during conjugation in Tetrahymena thermophila. The majority of DNA synthesis occurs at the expected periods preceding gamete formation and the two postzygotic divisions and during macronuclear development. DNA in new macronuclei is endoreplicated in an extremely discontinuous fashion. Under starvation conditions, the first endoreplication (2C to 4C) occurs immediately after the second postzygotic division when both new macronuclei and new micronuclei replicate. The second endoreplication (4C to 8C) does not occur until after separation of conjugants. If mating cells are kept under prolonged starvation conditions (20-24 hr), refeeding induces a partially synchronous division, after which an unexpectedly high percentage of cells incorporate tritiated thymidine into both macro- and micronuclei. Two previously undescribed periods of DNA synthesis were observed in the micronuclei of conjugating Tetrahymena. The first occurs during the early stages of meiotic prophase, before full crescent elongation. The second takes place in an extended period corresponding to macronuclear anlagen development, before conjugants have separated. CsCl gradient analyses indicate that, in micronuclear fractions, only main band DNA is being synthesized in both of these periods. However, in macronuclear fractions from both stages, a significant fraction (approximately 20%) of the DNA being synthesized has the buoyant density of ribosomal DNA. The finding that macro- and micronuclear DNA can be synthesized simultaneously in a single cell, both during conjugation and after refeeding starved exconjugants, raises interesting questions of how macro- or micronuclear-specific histones are targeted to the appropriate nuclei.

A single histone acetyltransferase from Tetrahymena macronuclei catalyzes deposition-related acetylation of free histones and transcription-related acetylation of nucleosomal histones

A salt-extracted histone acetyltransferase activity from Tetrahymena macronuclei acetylates mostly histone H3 and H4 when free histones are used as substrate. Free histone H4 is acetylated first at position 11 (monoacetylated) or positions 11 and 4 (diacetylated). This activity strongly resembles in vivo, deposition-related acetylation of newly synthesized histones. When acetylase-free mononucleosomes are used as substrate, all four core histones are acetylated by the same extract, and H4 is acetylated first at position 7 (monoacetylated) or positions 7 and 4 (diacetylated). In this respect, the activity of the extract is indistinguishable from postsynthetic, transcription-related histone acetylation that occurs in vivo or in isolated nuclei. Heat inactivation curves with both substrates are indistinguishable, and free histones compete with chromatin for limiting amounts of enzyme activity. These results argue strongly that two distinct, biologically important histone acetylations, one deposition related and one transcription related, are carried out by a single acetyltransferase.

Unusual features of transcribed and translated regions of the histone H4 gene family of Tetrahymena thermophila

The complete DNA sequence is presented of H4-II, the second of the pair of histone H4 genes of the ciliated protozoan, Tetrahymena thermophila. Both H4 genes code for the same protein. Codon usage in these and other Tetrahymena genes is severely restricted and is similar to that in yeast. Flanking regions are AT-rich (greater than or equal to 75%), relative to coding sequences (approximately 45% GC). Except for small, similarly positioned homologies, flanking sequences of the two genes are different. Canonical sequences in higher eukaryotic promoters are not obvious in these genes. Instead, short, localized, base composition eccentricities characterize the 5' flanking sequences of all Tetrahymena genes analyzed. The consensus, P yP u(A)3-4 ATGG initiates translation in these and all other known Tetrahymena genes. Nuclear transcripts and messages of both growing and starved cells begin at multiple sites, mainly at the first or second A residue following a pyrimidine. The palindrome typical of histone message 3' termini in higher organisms is not present. Downstream of both genes are sequences similar to the processing/polyadenylation signal of higher eukaryotes, although the unique 3' ends are not those predicted by the location of the signals.

Enzyme activity dot blots: a rapid and convenient assay for acetyltransferase or protein kinase activity immobilized on nitrocellulose

Methods are described for assaying (Tetrahymena) histone acetyltransferase activity and (Drosophila) casein kinase II activity by spotting extracts on nitrocellulose filters. The methods are quantitative over a wide range of enzyme concentrations and are almost as sensitive as liquid assays. Examples are presented for illustrating the use of these methods for enzyme purification, concentration, and desalting, as well as for electrophoretic blotting from agarose gels. A simple method for autoradiographic enhancement of nitrocellulose filters is also described.

An intervening sequence in an unusual histone H1 gene of Tetrahymena thermophila

An intervening sequence of 254 base pairs interrupts the coding region of the single gene for macronuclear histone H1 of the ciliated protozoan, Tetrahymena thermophila. The intervening sequence has splice junctions similar to those found in RNA polymerase II genes of other organisms. No obvious similarities are observed between this intron and the self-splicing intervening sequence of the Tetrahymena ribosomal gene. The derived amino acid sequence describes a small extremely basic H1 protein missing most of the central hydrophobic domain that is conserved in all other H1 proteins. Macronuclei divide amitotically, without chromosome condensation, suggesting the conserved globular domain of H1 plays a role in higher-order chromatin structure.

In situ dot blots: quantitation of mRNA in intact cells

A rapid, simple and reproducible dot blot method is described for quantitating the amounts of specific messages in small numbers of intact cells. The method has been used to accurately determine the number of histone H4 mRNA molecules in growing (approximately 40,000) and in starved (approximately 1600) Tetrahymena thermophila, and to measure the amount of message contributed by an E. coli plasmid containing part of the S10 ribosomal operon. Use of the method is illustrated to optimize in situ hybridization protocols and to measure mRNA amounts in cell lysates. Preliminary studies also indicate that the method can be used to detect mRNA in intact yeast cells.

Formation of stable chromatin structures on the histone H4 gene during differentiation in Tetrahymena thermophila

The relationship between chromatin structure and the transcriptional activity of the histone H4-I gene of Tetrahymena thermophila was explored. Indirect end-labeling studies demonstrated that major DNase I- and micrococcal nuclease-hypersensitive sites flank the active macronuclear genes but not the inactive micronuclear genes. Runon transcription experiments with isolated macronuclei indicated that histone gene transcription rates decreased when cells were starved. However, macronuclear nuclease-hypersensitive sites persisted upon starvation. Thus, one level of transcriptional control of the H4-I gene results in altered chromatin structure and is established during nuclear differentiation. The rate of transcription is also controlled, but not through hypersensitive site-associated structures.

hv1 is an evolutionarily conserved H2A variant that is preferentially associated with active genes

Polyclonal antibodies to the Tetrahymena macronuclear-specific histone variant hv1 cross-react with histone-like molecules from yeast, wheat, and mouse. A novel purification scheme has allowed isolation of sufficient hv1 to enable determination of the sequence of 61 amino-terminal residues as well as 27 additional internal residues. These data clearly demonstrate that hv1 shares a number of conserved sequence elements with the H2A family of histones. Comparison of hv1 with H2A.F (= H2A.Z = M1), another evolutionarily conserved H2A variant whose sequence is known, reveals that they share an unblocked amino-terminal alanine (instead of acetylserine) and a distinctive structure in a "variant box" region that distinguishes them from major H2As. In addition, 10 residues have been identified which are identical (or highly similar) in hv1 and H2A.F, but are different from residues conserved in the major H2As. Therefore, in many ways hv1 resembles chick H2A.F more than the major Tetrahymena H2A. The sites of acetylation of hv1 also differ from those of the major Tetrahymena H2As. In spite of their similarities, hv1 and H2A.Z differ significantly in their amino termini, and antibodies against hv1 do not react with H2A.Z. Interestingly, the nucleolar staining pattern reported with anti-hv1 serum is similar to that reported for an antiserum to another H2A variant, mouse testes-enriched H2A.X. Since both H2A.Z and hv1 appear to be enriched in transcriptionally active chromatin, these results suggest that there may be a number of different, functionally distinct, nonallelic variants in the H2A family of histones and that hv1 is a hybrid H2A variant with properties of both vertebrate H2A.Z and H2A.X.

An unusual genetic code in nuclear genes of Tetrahymena

We have cloned and partially sequenced two histone H3 genes of Tetrahymena thermophila. The DNA sequences strongly suggest that both genes are active in the vegetatively growing cell. Comparison of the derived amino acid sequences of these two genes with the actual sequence of Tetrahymena histone H3 results in the surprising conclusion that TAA codes for glutamine. This represents the first demonstration of a coding function for this termination codon of the "universal" code. This observation has important implications for the evolution of ciliates and of the genetic code.

Changes in chromatin structure accompany modulation of the rate of transcription of 5S ribosomal genes in Tetrahymena

The chromatin structure of a single cluster of six tandemly repeated 5S ribosomal RNA genes (5S genes) in Tetrahymena thermophila has been characterized. Indirect end labeling experiments indicate that the actively transcribed 5S genes in macronuclei are rapidly cut by DNAse I near the putative internal promotor and just 5' to the transcribed region. When cells are starved to reduce 5S gene transcription rates, the DNAse I sensitivity of the intragenic site is reduced relative to the 5' site. In the nontranscribed 5S genes in micronuclei, neither of these sites is hypersensitive to DNAse I. Thus structural alterations accompany both the activation of transcription during macronuclear development and physiological changes in the rate of transcription of the 5S genes. These DNAse I data together with studies using Staphylococcal nuclease suggest that rapidly transcribed 5S genes may not be associated with histones as nucleosomes. In contrast, the genes in starved cell macronuclei appear to be associated with one nucleosome per 280 base pair tandem repeat.

Sequence organization within and flanking clusters of 5S ribosomal RNA genes in Tetrahymena

Macro- and micronuclei of Tetrahymena thermophila each contain approximately 30 clusters of 5S genes per haploid genome. Structural changes in DNA sequences associated with some of these clusters occur during the development of the transcriptionally active macronucleus from the transcriptionally inert micronucleus. Exonuclease digestion indicates that DNA fragmentation is not responsible for these changes, making it likely that sequence rearrangements occur near some 5S genes during macronuclear development. These rearrangements appear to be random in location with respect to the 5S genes and do not alter either the tandem repeat organization of the genes, the average number (five) or the range in number (one to about 16) of genes per cluster. The 5S gene clusters are not closely linked and are not flanked by common repeating elements large enough to cross-hybridize. Sequence analysis of one tandem repeat suggests that Tetrahymena 5S genes have intragenic promoters. These observations indicate that features other than DNA rearrangements or DNA sequence per se are responsible for the transcriptional activation of 5S genes during macronuclear development.

Tetrahymena H4 genes: structure, evolution and organization in macro- and micronuclei

The ciliated protozoan Tetrahymena thermophila contains two types of H4 histone genes (H4-I and H4-II). Southern blotting and analysis of DNA from nullisomic strains indicate that H4-I and H4-II are on different chromosomes and that only H4-II is closely linked to an H3 gene. No DNA sequence rearrangements are observed for either of the H4 genes when the transcriptionally inert, germ line, micronucleus is compared to the transcriptionally active, somatic macronucleus. Comparison of the H4-I gene and its flanking sequences to H4 gene sequences of other organisms indicates that there are evolutionary constraints on coding nucleotides that are unrelated to their protein coding function and that these evolutionary pressures operate at the level of translation.

Developmental rearrangements associated with a single type of expressed alpha-tubulin gene in Tetrahymena

Southern blotting analyses with diverse heterologous probes indicate that there is only a single type of alpha-tubulin gene in the ciliated protozoan Tetrahymena thermophila. Comparisons of the germ-line configuration of alpha-tubulin sequences in the transcriptionally inert micronucleus with the somatic configuration in the transcriptionally active macronucleus indicate that rearrangements involving breakage and rejoining of sequences flanking both sides of the alpha-tubulin gene accompany macronuclear differentiation.

Multiple, independently regulated, polyadenylated messages for histone H3 and H4 in Tetrahymena

Heterologous probes for yeast H4 and H3 histone genes have been used to study the corresponding histone mRNAs in growing and starved Tetrahymena. Histone mRNAs in both physiological states are polyadenylated. Two types of H4 protein and two types of H3 protein have previously identified in Tetrahymena. Two size classes of H4 messages and three classes of H3 messages have been detected by northern analyses. Southern blot analysis indicate that the number of different kinds of H3 and H4 genes is the same or slightly greater than the number of different messages, suggesting that each message is derived from a different gene. Growing cells have -30 times more histone mRNA than starved cells, even though their total mRNA content is only 4 times greater. The relative abundance of different H4 and H3 messages in growing and starved cells is different, demonstrating that the different messages for a particular type of histone are regulated non-coordinately. In starved cells the presence of a single size class of H3 messages correlates with the preferential synthesis of a previously described macronuclear-specific H3 variant. The fraction of histone messages loaded in growing and starved cells is the same as for bulk mRNAs, and the relative concentrations of the multiple messages for H4 and H3 are the same in polysomal and total RNAs of each cell type. These observations suggest that histone synthesis in Tetrahymena is controlled largely at the level of message abundance, and that very little, if any, control occurs at the translational level.

Regulation of protein synthesis in Tetrahymena. RNA sequence sets of growing and starved cells

The complexity of messenger RNA in growing or starved Tetrahymena thermophila is similar and unusually high (approximately 4.5 X 10(7) nucleotides). The complexity of nuclear RNA in growing cells (approximately 7.8 X 10(7) nucleotides) is only about 1.7 times that of mRNA. The concentration of complex class (rare) messages (approximately 53 copies/growing cell and approximately 11 copies/starved cell) is low in comparison to the size of the cell. The concentration of complex nuclear transcripts is also very low (approximately 0.7 copies/growing cell nucleus and approximately 2.6 copies/starved cell nucleus) considering that the macronucleus contains 45 to 90 copies of each single copy sequence. The complex sequence sets found on polysomes of growing and starved cells overlap about 80% and about 60% of the complex nuclear transcripts appear to be held in common. About 60% of macronuclear single copy DNA is transcribed in one or both physiological states. Although growing and starved cells have extremely different fractions of their messages loaded onto polysomes, within each cell type the complex messages in polysomal and nonpolysomal cytoplasmic fractions are indistinguishable, suggesting that exchange may occur between loaded and unloaded messages. Although T. thermophila DNA has an unusually low G + C content (23%), sequences coding for complex RNAs have base ratios similar to those of total DNA. Therefore, codon usage in Tetrahymena must be extremely biased towards adenine- and uridine-rich codons.

Regulation of protein synthesis in Tetrahymena. Quantitative estimates of the parameters determining the rates of protein synthesis in growing, starved, and starved-deciliated cells

The calculated rate of protein synthesis for growing Tetrahymena is 360 pg/h, whereas starved cells synthesize only about 3 pg of protein/h. Within 50 min after deciliation of starved cells, the rate of protein synthesis increases to about 60 pg/h. The major mechanism to accomplish these large and rapid changes in the rate of bulk protein synthesis involves regulation of the number of messages loaded on polysomes. Logarithmically growing cells contain about 3.2 X 10(7) mRNA molecules/cell, of which approximately 60% is loaded on polysomes. Starved cells contain about 0.8 X 10(7) messages and the percentage of messages loaded is reduced to 4%. Thus, the number of loaded messages is approximately 60-fold lower in starved cells than in growing cells, although the total message content of cells in these two physiological states differs by only a factor of 4. After deciliation of starved cells, message loading increases about 10-fold. It seems likely that much of the message loaded after deciliation is derived from the large pool of nonpolysomal message in starved cells. Although large differences in message loading exist for growing, starved, and starved-deciliated cells, measurements of the rate of polypeptide elongation and the rate of message initiation indicate the translational efficiency of loaded messages (pg of protein synthesized per pg of message/unit time) is very similar under all conditions.