INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Retroviruses are currently under intensive study because they can elicit malignant tumors and cause AIDS. Furthermore, at least 0.1 to 0.6% of the human genome is composed of endogenous retroviruses and long-terminal-repeat (LTR)-containing retrotransposons, which resemble retroviruses in their structural organization and mode of transposition (38). The life cycle of these elements begins with the transcription of an integrated DNA copy of the element and the incorporation of the transcribed RNA into a viruslike particle composed of element-encoded proteins, including the capsid protein, protease, reverse transcriptase, and integrase. The RNA is then reverse transcribed into cDNA, which integrates into a new chromosomal location in the host cell (for a review, see reference 5). Such integration is required for retroviruses to induce disease, for example, by activating a nearby cellular proto-oncogene (70). Likewise, insertions of the yeast Ty1 element (5) can alter the regulation of nearby cellular genes.

Common laboratory yeast (Saccharomyces cerevisiae) strains contain five types of Ty elements composed of central regions of DNA flanked by LTRs. Structural proteins and enzymatic activity are encoded within the central region (for reviews, see references 5 and 22). Ty1, Ty2, Ty4, and Ty5 elements are members of the copia class of retrotransposons, while Ty3 elements belong to the gypsy class. Ty1 and Ty2 elements are respectively present at about 30 to 35 and 5 to 15 copies per genome, contain the same LTR sequences (called delta elements), and have very similar internal regions. The more distantly related retrotransposons Ty3, Ty4, and Ty5 have different sets of LTRs, are present in lower copy numbers, and have not been observed to be insertional mutagens.

Yeast retrotransposons provide an attractive model system in which to define host functions required for retroviral transposition. Mutations that reduce Ty1 and Ty2 transcription levels also reduce transposition (6, 18, 19, 73, 74), since Ty elements transpose through RNA intermediates (3). Transcription of Ty1 and Ty2 elements is also inhibited in mating-incompetent cells (52) and by growth on glycerol rather than glucose (69). Also, both the Ty1 and Ty2 RNA levels and transposition rates are increased by DNA damage (7). In addition, a number of conditions and mutations alter Ty1 transposition via posttranscriptional mechanisms, including translation of Ty-encoded open reading frames (ORFs) (22), reverse transcription of the mRNA into a cDNA copy (10, 11), and growth at 20°C (51, 52).

The ubiquitin-conjugating enzyme Rad6p (Ubc2p) also affects Ty1 transposition. Mutations in RAD6 increase the overall rate of Ty1 transposition into CAN1 and SUP4 without causing an increase in the level of total Ty1 RNA (8, 31, 54). However, when genetically marked Ty1 elements were used, it was determined that the deletion of RAD6 affected the transposition of some, but not other, elements, and these effects were at the transcriptional level (8). The recent findings that mutations in RAD6 release transcriptional silencing at telomeres and HM loci (28) and that rad6 mutations enhance the transcription of marked Ty1 elements located in silent rDNA chromatin regions (8) are consistent with the hypothesis that mutations in RAD6 cause alterations in chromatin structure in certain chromosomal regions, making it easier for Ty1 elements to integrate. Ubiquitination has also been shown to affect Ty3 transposition. Cellular stress, induced by growth at high temperature or in ethanol, inhibits transposition of Ty3 without affecting Ty3 transcription. This inhibition can be reversed by overexpression of a protease that cleaves ubiquitin off proteins (43).

In this paper, we describe a genetic screen designed to identify mutations that increase the rate of transposition of Ty1 and/or the closely related Ty2 elements (referred to henceforth together as Ty1 for simplicity). Using this screen, we have uncovered a synthetic interaction between mutations in CAC3 and HIR3 that leads to a dramatic increase in Ty1 transposition rates without affecting the overall levels of Ty1 transcription. Simultaneous loss of both genes also increases methyl methanesulfonate (MMS) and UV sensitivities and reduces telomeric silencing and the growth rate. Since CAC3 encodes a component of chromatin assembly factor I (CAF-I) (34) and HIR3 controls the levels and balance of histone mRNAs (45, 50, 57), our results suggest that alterations in chromatin structure can increase the efficiency of integration of Ty1 cDNA into the host genome.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Strains, cultivation conditions, and scoring for markers. Standard yeast cultivation conditions were used (66). Cells were grown on organic complete medium (yeast extract-peptone-dextrose [YPD]) or synthetic complete medium (SC) lacking nutrients (e.g., uracil [Ura]). Ura colonies were selected on medium containing 1 g of 5-fluoro-orotic acid and 12 mg of uracil per liter (+FOA) (4).

his3-200

lys2::his3-4

lys2::his3-4

View larger version (10K): [in this window] [in a new window]   FIG. 1.   Ty1 transposition assay. The most common way for the His strains containing the promoterless his3-4 allele diagrammed above to revert to His+ is by the insertion of an upstream Ty1 element. Such insertions were detected by the appearance of 400- to 500-bp PCR amplicons when the HIS3 (1His3) and delta element (U5-in) primers were used. No insertions were detected when the HIS3 (1His3) and delta element (U3-in) primers were used, indicating that the Ty1 elements had to be in the indicated orientation to cause a His+ reversion. Arrows in boxes indicate the direction of transcription.

Plasmids. The URA3-integrative plasmid pCB1 (9) contains a 1.37-kb SmaI-BamHI fragment from pAB100 (63) carrying his3-4, which replaces the internal HpaI-BamHI fragment of a LYS2 gene on the plasmid. The YCp50 (61)-based CEN-URA3 plasmids pHIR3 (which contains HIR3) and pCAC3 (which carries CAC3) were partially digested with EcoRI or HindIII and self-ligated to generate the deletion plasmids diagrammed in Fig. 2B. To generate CAC3 deletion plasmid pZJ13, a BglII-BamHI fragment containing hisG-URA3-hisG (1) (kindly supplied by E. Alani), cloned into the BglII site of the XbaI-SalI subclone of pCAC3 in Bluescript II KS, was partially digested with ClaI and self-ligated.

View larger version (17K): [in this window] [in a new window]   FIG. 2.   Characterization of the htr1 mutant. (A) Complementation of htr1 MMS sensitivity. Transformants of L1356 (wild type [WT]) and L1561 (htr1) with the indicated plasmids (YCp50, pCAC3, and pHIR3) were spotted from Ura onto YPD and YPD plus 0.02% MMS. (B) Localization of htr1-complementing activity. Restriction maps of the cloned fragments in plasmids pCAC3 (top) and pHIR3 (bottom) and subcloned DNA fragments described in the text are shown. +, MMS sensitivity was complemented by the fragment; , no complementation occurred. Arrows indicate the direction of transcription. Restriction sites are abbreviated as follows: B, BamHI; Bg, BglII; Bs, BssHII; C, ClaI; E, EcoRI; H, HindIII; K, KpnI; S, SalI; Sn, SnaBI; Sp, SphI; P, PstI; X, XhoI; and Xb, XbaI. (C) Deletion of HIR3 does not affect levels of Ty1 mRNA. RNA blot analysis of Ty1 (and Ty2) versus TEF1 (internal control) mRNAs from L1356 and the isogenic hir3::LEU2 deletion strain, L1675.

Screen for mutants with high rates of Ty1 transposition. L1356 mutagenized with ethyl methanesulfonate to 20 to 40% survival was incubated on YPD at 20°C for 2 weeks. Colonies consisting of approximately 5 × 10⁷ to 1 × 10⁸ cells were spread on His (or His containing 0.0125% Casamino Acids so that His cells could divide for a few generations). After 3 days at 20°C, which increased the transposition rate to a level more easily measured, the plates were incubated at 30°C for another week and then the number of His⁺ revertants per plate was counted.

Transposition rate measurements. Most transposition rates were measured by the his3-4 assay. In this case, rates of mutation to His⁺ were calculated by fluctuation tests, using the equation P₀ = e^Nµ (40), or by the method of the median, using the equation µ = f/ln(Nµ) (17). Cells spread on Ura (to retain plasmids) or YPD were incubated at 20°C until the colonies consisted of about 10⁸ cells (on YPD) or 5 × 10⁷ cells (on Ura). For each rate measurement, 10 (or in some cases 20) colonies of equal size were scooped up and each entire colony was spread on a His plate. Two or three additional colonies of equal size were used to determine cell viability. We used PCR to estimate how many of the His⁺ colonies resulted from a Ty1 transposition (Fig. 1). Representative His⁺ colonies were analyzed with the HIS3 primer 1His3 (TGTAATACGCTTTACTAGG) and Ty primers U5-in (ATTGTTGGGATTCCATT) and U3-in (ATATTATCATATACGGTGTT). Since Ty1 insertions were found in 90% or more of the His⁺ revertants, the His⁺ revertant rate was used as an estimate of the Ty1 transposition rate.

Analyses of DNA, RNA, and protein. Restriction fragments to be sequenced were subcloned into Bluescript II KS (Stratagene) and sequenced by the dideoxy chain termination method, using Sequenase version 2.0 (U.S. Biochemical Corp.). The cac3 alleles from L1356 and L1561 were cloned for sequencing by gap repair (48) of BssHII- and SnaBI-digested pCAC3.

Indirect immunofluorescence and immunogold labeling. Indirect immunofluorescence was performed as previously described (21), using a 1:5,000 dilution of anti-HA mouse monoclonal antibody, rabbit polyclonal anti-Rap1, and anti-nuclear pore (BAbCo, Inc.) as well as a fluorescein isothiocyanate-conjugated secondary antibody. Processing and immunogold labeling were performed as described elsewhere (2) with the changes noted below. Cells fixed overnight at 4°C in buffer (4% sucrose in 100 mM sodium phosphate [pH 7.5]) containing 1% acrolein and 4% paraformaldehyde were washed twice with this same buffer for 10 min and then dehydrated at room temperature in an ethylene glycol concentration gradient (25, 50, 75, 90, and 90% for 15, 15, 15, 10, and 10 min, respectively). Infiltration was carried out over a period of 3 days. The primary antibody, HA mouse monoclonal antibody, was used at a 1:1,000 dilution. The secondary antibody was 10-mm gold particles attached to goat anti-mouse antibody (Nanoprobe, Inc., no. CG1001). Grids stained with 1% (wt/vol) uranyl acetate (aqueous) for 10 min and then washed three times for 5 min each in distilled water were stained for 4 min with lead citrate.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

An assay for Ty1 transposition. The assay used to determine Ty1 transposition rates was based on the observation that His⁺ revertants of a plasmidborne his3-4 allele that lacks a promoter result either from insertions of Ty1 into the region upstream of his3-4 or from plasmid rearrangements (3, 63). To eliminate the occurrence of His⁺ revertants due to plasmid rearrangements, we made a haploid strain, L1356, that bears a large deletion at the normal HIS3 locus and an insertion of the his3-4 allele in its genomic LYS2 locus (Fig. 1). By using DNA blot and PCR analyses, we found that 38 of 41 independent His⁺ revertants derived from L1356 colonies grown at 20°C contained Ty1 insertions in the 5' region of his3-4. As expected, all Ty1 elements detected among His⁺ revertants were in the orientation which enabled the enhancer sequences in this element to activate the promoterless his3-4 gene (Fig. 1). Furthermore, as reported for other assay systems (51, 52), the transposition rate in our assay was increased at low temperature (data not shown).

Isolation of mutants with a high Ty1 transposition rate phenotype and characterization of one mutant strain, L1561. To isolate mutations that increase the rate of Ty1 transposition, we screened mutagenized L1356 (his3-4) cells for increased levels of His⁺ revertants (see Materials and Methods). Colonies which gave rise to more than five His⁺ revertants (on average, there was about 0.75 His⁺ revertant/colony) were retested. Of 1,000 colonies screened, 7 candidates repeatedly gave rise to an increased number of His⁺ revertants. Of these seven, two were MMS sensitive and two were temperature sensitive. Each of these four was crossed with SL1005-54, which also contains the his3-4 transposition assay system. The MMS or temperature sensitivity failed to segregate with the high transposition rate phenotype in three of these crosses.

7

9

CAC3 and ORF YJR140c were cloned by functional complementation of MMS sensitivity. The L1561 mutant was transformed with a genomic S. cerevisiae library made in the CEN-URA3 vector YCp50 (61). Transformants obtained on Ura or Ura supplemented with 0.01% MMS were tested for resistance to 0.025% MMS in YPD, and 21 resistant transformants that regained MMS sensitivity when the plasmid was forced out by growth on +FOA medium were recovered. Plasmid pCAC3, isolated from one of these transformants, was retransformed into L1561, in which it complemented both MMS sensitivity (Fig. 2A) and the high Ty1 transposition rate (see below). Restriction mapping and partial sequencing of the insert in pCAC3 identified a fragment from chromosome II containing six ORFs, including CAC3. By deletion analysis, CAC3 was identified as the complementing gene (Fig. 2B). PCR analysis showed that 16 of the complementing plasmids contained CAC3.

Deletion of HIR3, but not CAC3, in parental strain L1356 causes MMS sensitivity and a high rate of transposition. L1675, a hir3::LEU2 deletion mutant created in our original L1356 strain, was MMS sensitive (see Fig. 4) and had an average Ty1 transposition rate of 1.3 × 10⁷, approximately 20-fold higher than the rate in L1356. In contrast, when CAC3 was disrupted in L1356, the null mutant did not display any increase in MMS sensitivity or Ty1 transposition rate (data not shown).

The htr1 mutation is an allele of HIR3. Since a deletion of HIR3 reproduced the phenotype of the htr1 mutation, we suspected that htr1 was allelic to HIR3. This hypothesis was confirmed by genetic complementation. SL1006-1B (htr1) was crossed with both W303 (HIR3) and W3033 (hir3::HIS3) (kindly supplied by M. A. Osley), and SL106-1D was crossed with W3033. The diploids were then examined for the hir3 phenotype. Mutations in HIR3 cause cells to lose their ability to repress the transcription of histone HTA1-HTB1 mRNA in response to the DNA replication-inhibiting drug hydroxyurea (HU) (50). As shown in Fig. 3, htr1 failed to complement hir3, since the hir3/htr1 diploid (as well as the htr1 haploid) constitutively expressed HTA1 in the presence of HU while transcription of HTA1 was repressed by HU in the hir3/HTR1 and HIR3/htr1 diploids. This identified htr1 as an allele of hir3, now called hir3-2.

View larger version (39K): [in this window] [in a new window]   FIG. 3.   The htr1 mutation depressed transcription of HTA1-HTB1 and failed to complement hir3. RNA blot analysis was performed after 30 min of growth in the absence () or presence (+) of HU. The diploids were formed by crossing SL1006-1B (htr1) and SL1006-1D (HTR1) with W303 (HIR3) and W3033 (hir3). The constitutively transcribed gene PRT1 was used as an internal loading control since its abundance is not affected by HU.

The MMS sensitivity and high Ty1 transposition rate caused by hir3 are dependent on an additional mutation. Transformation with pCAC3 complemented the MMS sensitivity (Fig. 2A and 4A) and high Ty1 transposition rate (Table 2) phenotypes associated with both the L1561 hir3-2 mutant strain and strain L1675, which contains a hir3::LEU2 deletion made in the parental strain L1356. We originally thought that pCAC3 was an extra-copy suppressor of hir3-2. However, to our surprise, we found that the L1356 parental strain contained a mutation in CAC3. The CAC3 alleles in L1356 and L1561 were cloned by gap repair and sequenced. The CAC3 gene cloned from these strains contained, relative to the database sequence, a G-to-C change at position 877 and an A inserted between 877 and 878, resulting in a frameshift causing the loss of about 30% of the Cac3p. Additional differences were silent or neutral. The sequence of the PCR-amplified CAC3 gene from SL1005-54 (used in the original cross with L1561 to score for segregation of htr1) indicated the presence of the same frameshift allele, which we have called cac3-1.

View larger version (55K): [in this window] [in a new window]   FIG. 4.   Double mutations of hir and cac cause MMS sensitivity. (A) Deletion of HIR3 causes MMS sensitivity in the presence of cac3 or cac2. L1356 was found to contain a relevant mutation, cac3-1. L1356 (cac3 HIR3) was transformed with YCp50 or pCAC3, and L1675 (cac3 hir3) was transformed with YCp50, pCAC3, or pHIR3. Cells were grown on Ura medium and were then spotted onto YPD and YPD plus 0.025% MMS. UV irradiation was at 90 J/m2. (B) Deletion of HIR3, HIR1, or HIR2 in L1356 causes MMS sensitivity. MMS sensitivity was scored by spotting 10-fold serial dilutions of cells on YPD with 0.02% MMS (incubation at 30°C for 2 days).

Mutations in other HIR and CAC genes also cause MMS sensitivity and high rates of Ty1 transposition when combined with cac3 and hir3 mutations, respectively. Since mutations in HIR1, HIR2, and HIR3 all have the same effects on the regulation of histone expression (50, 68), we tested whether deletions of HIR1 and HIR2 in our parental cac3-1 strain, L1356, would increase the MMS sensitivity and the Ty1 transposition rate as did mutations in HIR3. Indeed, in the cac3-1 strain, deletion of either HIR1 or HIR2 led to an increase in the sensitivity to MMS (Fig. 4B) and caused a dramatic increase in the Ty1 transposition rate (Table 3). Likewise, cac3 hir1 and cac3 hir2 double deletions in UCC4543 caused increased sensitivity to MMS, while single hir deletions did not (data not shown).

Simultaneous deletion of HIR3 and CAC3 has a dramatic effect on telomeric silencing and growth rate. To examine the effects of mutations in HIR3 and CAC3 on telomeric silencing, we used strain UCC4543, which contains the RNA polymerase III (RNAP III)-transcribed gene SUP4-o placed close to telomere VII-L (28). Telomeric silencing inhibits the expression of SUP4-o, a tyrosyl-tRNA ochre suppressor, thereby causing the loss of suppression of the ade2-101 ochre marker, leading to red colony color and the absence of growth on Ade. As shown in Fig. 5A, silencing of SUP4-o was reduced in cac3 hir3 double mutants, while deletions of CAC3 alone reduced silencing to a much lesser extent and deletions of HIR3 alone had no effect on telomeric silencing. Six, three, and two independent cac3 hir3, cac3, and hir3 deletion strains were examined, respectively. Similar data was obtained with single hir1 or hir2 and double hir1 cac3 or hir2 cac3 deletions (data not shown).

View larger version (48K): [in this window] [in a new window]   FIG. 5.   Silencing and growth defects in hir3 cac3 double-deletion mutants. Strain UCC4543 (wild type [WT]) and its deletion derivatives L1678 (cac3), L1687 (hir3), and L1688 (cac3 hir3) were used. (A) Silencing of telomere VII-L-located SUP4-o (SUP4-o-TEL) is released in hir3 cac3 double-deletion mutants. Expression of SUP4-o-TEL was assayed by spotting 10-fold serial dilutions on SC (control) and Ade and incubating the plates at 30°C for 4 days. The slight increase in growth of the hir3 strain shown on Ade was not reproducible. (B) Deletion of both HIR3 and CAC3 retards cell growth. The indicated strains were diluted, plated on YPD, and incubated at 20°C.

Effects of cac3 hir3 double mutations on Ty1 transposition at other loci. Above we showed that cac3 hir3 double mutations increase the rate of Ty1 transposition into his3-4, located at the LYS2 locus. To determine the effect of cac3 hir3 double mutations on Ty1 transposition into other loci, we used two previously published assays. In the his3AI assay, the generation of His⁺ prototrophs is indicative of the transposition of a single marked Ty1 element to any position in the genome (14, 15). Since the majority of transpositions occur in a few genomic hot spots (16, 30), this assay essentially measures the frequency of transposition into these hot spots. In strain JC364 (CAC3 HIR3), the frequency (mean ± standard deviation) of His⁺ prototrophs was 5.07 × 10⁷ ± 0.15 × 10⁷. Deletion of CAC3 alone did not increase the frequency of His⁺ prototrophs. However, the simultaneous deletion of CAC3 and HIR3 (L1690 [cac3::hisG-URA3-hisG hir3::LEU2]) resulted in a His⁺ frequency of 2.27 × 10⁶ ± 0.43 × 10⁶, representing a three- to fivefold increase in transposition rate relative to that of the CAC3 HIR3 and cac3 HIR3 parents. Transposition into CAN1 increased about 10-fold in L1675 (cac3 hir3) relative to that in L1356 (cac3 HIR3), from 1.0 × 10⁷ (duplicate, 1.05 × 10⁷) to 1.15 × 10⁶ (duplicate, 9.0 × 10⁷).

Localization of Hir3p. To determine the cellular location of Hir3p, we fused three repeats of the HA epitope tag to the 3' end of the HIR3 ORF and placed the tagged gene on a multicopy 2µm plasmid, pHIR3-HA. pHIR3-HA transformants of the MMS-sensitive strain L1675 (cac3-1 hir3::LEU2) became resistant to 0.02% MMS (data not shown), demonstrating that the tagged Hir3p (Hir3p-HA) was at least partially functional. The size of Hir3p-HA, as determined by Western blot analysis, was consistent with the predicted size of 191.7 kDa (data not shown). YJB2306 and L1675 transformed with pHIR3-HA (or pHIR3 as a control) were respectively examined by indirect immunofluorescence microscopy and immunoelectron microscopy, using antibodies specific for the HA epitope. Both techniques localized Hir3p-HA to the nucleus (Fig. 6). About 50 and 10%, respectively, of the YJB2306 and L1675 pHIR3-HA transformants were labeled with the HA antibody, and Hir3p-HA was clearly present in the nucleus, in all of the labeled cells. In contrast, only background immunofluorescence and immunogold labeling were observed in cells transformed with the pHIR3 plasmid and stained with the HA epitope antibody. In double-labeling immunofluorescence experiments with HA antibody and either Rap1p antibody (21) or nuclear pore antibody (MAB414, BAbCo, Inc.), the nuclear Hir3p-HA signal did not colocalize with either Rap1p or the nuclear pores (data not shown).

View larger version (66K): [in this window] [in a new window]   FIG. 6.   Nuclear localization of HA-tagged Hir3p. (A) Immunogold labeling of HA-Hir3p in L1675. (B) Indirect immunofluorescence and 4',6-diamidino-2-phenylindole (DAPI) staining of HA-Hir3p in YJB2306.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

During the analysis of a genetic screen for mutations that increase the rate of transposition of Ty1 into a promoterless his3-4 target, we discovered that the parental strain in which the search was undertaken contained a frameshift mutation in a gene (CAC3) encoding a component of yeast CAF-I (yCAF-I). Since the cac3 mutant phenotype is very subtle, it is unknown how widespread this mutation is in laboratory stocks. Our search revealed that a newly induced hir3 mutation in this strain caused a dramatic increase in the Ty1 transposition rate. Mutations in HIR3 are known to affect the regulation of histone synthesis (45, 50, 68). The observed effect on Ty1 transposition was dependent on the simultaneous presence of both the cac3 and hir3 mutations. The two mutations also caused a synthetic increase in growth inhibition as well as in MMS, UV, and cold sensitivity but did not cause gamma-ray sensitivity. The combined presence of mutations in CAC3 and HIR3 had no effect on the overall level of Ty1 mRNA, so the increase in the Ty1 transposition rate must result from another mechanism. We propose that this mechanism is an alteration in chromatin structure which makes the DNA more available for Ty1 integration. In support of this hypothesis, silencing of a telomere-proximal RNAP III-transcribed gene was reduced by a combination of cac3 and hir3 mutations much more than by the single cac3 mutation alone. Likewise, in the accompanying paper (32), mutations in CAC and HIR genes are shown to have a synergistic effect on the release of telomeric silencing of an RNAP II-transcribed gene.

The role of CAC3 in Ty1 transposition is as a component of CAF-I. yCAF-I (composed of Cac1p, Cac2p, and Cac3p) and human CAF-I (hCAF-I) show a high degree of conservation. Both yCAF-I and hCAF-I preferentially assemble nucleosomes on replicating DNA (33, 34). hCAF-I is thought to be involved in the first step of the nucleosome assembly process, bringing newly synthesized, acetylated isoforms of H3 and H4 tetramers to replicating DNA. Nucleosome assembly is then completed by the addition of H2A-H2B dimers. Mutations in any of the yeast CAC genes cause UV but not X-ray sensitivity and a reduction in telomeric silencing (21, 34, 44). Furthermore, the localization of Rap1p, a DNA binding protein which is a significant component of telomeric chromatin, is altered in cac1 mutants (21). Our finding that a cac2 deletion (in a hir3 background) causes increases in Ty1 transposition and MMS sensitivity similar to those observed in the cac3 hir3 strains suggests that these phenotypes are caused by inactivation of the CAF-I activity. In addition, our finding that mutations in CAC3 or CAC2 (synergistically with hir3) cause an increase in transposition of Ty1 into a nontelomeric region (lys2::his3-4 is, respectively, 141 and 58 centimorgans from its telomere and centromere) supports the hypothesis that CAF-I has a general role in chromatin assembly that is not restricted to the formation of heterochromatin.

30

18

17

The effect of mutations in HIR3 on Ty1 transposition is mimicked by mutations in HIR1 and HIR2. The HIR1, HIR2, and HIR3 genes are required for the proper balance of the core histones H2A, H2B, H3, and H4 (45). These histones are encoded by four pairs of divergently transcribed genes, HTA1-HTB1 and HTA2-HTB2 (each encoding H2A and H2B) and HHT1-HHF1 and HHT2-HHF2 (each encoding H3 and H4). Often the consequence of an imbalance of histones is altered chromatin structure (47), which appears to lead to altered transcription of numerous genes (12, 25, 26) and altered chromosome segregation (42). Hir1p, Hir2p, and Hir3p are required for a feedback control system which autogenously regulates transcription of the HTA1-HTB1 locus in response to intracellular H2A and H2B levels. Consistent with its role as a transcriptional repressor, Hir2p was previously shown to be located in the nucleus (68). Here we showed that overexpressed Hir3p-HA also localizes to the nucleus. Despite the presence of seven hypothetical transmembrane domains in Hir3p (60), there was no indication from either the immunofluorescence or immunogold labeling studies that Hir3p-HA is associated with a membrane.

his4-912

The dramatic increase in Ty1 transposition associated with cac3 hir double mutations is not a property of other DNA repair mutations. Each of the cac3 hir double mutants is sensitive to MMS and therefore deficient in DNA repair. It is thus possible that the increase in Ty1 transposition associated with cac3 hir mutations is a secondary effect of the DNA repair deficiency and results from unrepaired lesions in the DNA. However, using the promoterless his3 assay, we found that combinations of cac3 and deletions of genes from the excision repair pathway (RAD1), the error-prone repair pathway (RAD6), or the recombination repair pathway (RAD52) cause only marginal increases in Ty1 transposition. Since rad6 and rad52 mutants are very MMS sensitive, this suggests that the increase in Ty1 transposition caused by the hir cac3 double mutations is not simply due to unrepaired DNA lesions.

Possible mechanisms for the synergistic effects of cac and hir mutations on silencing and the rate of Ty1 transposition. Wild-type cells use CAF-I to assemble nascent histones into new nucleosomes in replicated DNA. In cac mutant strains, the assembly of new nucleosomes must be accomplished by another pathway. One way to explain the finding that cac mutants reduce silencing (21, 34, 44) is to propose that the assembly of new nucleosomes is slower in the absence of CAF-I, allowing time for silencing factors to diffuse away, for acetylases to act on recycled nucleosomes, and for activators to bind to newly replicated DNA (20). Likewise, the synergistic interaction between cac and hir mutations could be explained if the nucleosome assembly activity substituting for CAF-I in cac mutants were further slowed by the hir mutations (possibly due to an imbalance in histones) while the bona fide CAF-I activity in CAC strains was relatively resistant to the effects of the hir mutations. A longer delay in chromatin reassembly in cac hir strains would increase the chance that dissociated proteins diffuse away and that recycled histones are modified. After several generations, the result would be a steady-state structure in cac hir strains that differed from that in CAC HIR strains in both histone modification levels and DNA binding proteins. In silent regions, this could cause a reduction of silencing; in other regions of the genome, it could result in increased availability for Ty transposition.