Kluyveromyces lactis Sir2p Regulates Cation Sensitivity and Maintains a Specialized Chromatin Structure at the Cryptic -Locus

Corresponding author: Stefan U. Åström, Umeå Center for Molecular Pathogenesis, Umeå University, S-901 87 Umeå, Sweden., stefan.astrom{at}ucmp.umu.se (E-mail)

Communicating editor: S. SANDMEYER

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In Saccharomyces cerevisiae, transcriptional silencing of the cryptic mating type loci requires the formation of a heterochromatin-like structure, which is dependent on silent information regulator (Sir) proteins and DNA sequences, called silencers. To learn more about silencing, we characterized the mating type loci from the yeast Kluyveromyces lactis. The K. lactis MAT, HMRa, and HML loci shared flanking DNA sequences on both sides of the loci presumably acting as recombinational targets during mating type switching. HMRa contained two genes, the a1 gene similar to the Saccharomyces a1 gene and the a2 gene similar to mating type genes from other yeasts. K. lactis HML contained three genes, the 1 and 2 genes, which were similar to their Saccharomyces counterparts, and a novel third gene, 3. A dam-methylase assay showed Sir-dependent, but transcription-independent changes of the chromatin structure of the HML locus. The HML3 gene did not appear to be part of the silent domain because 3p was expressed from both MAT3 and HML3 and sir mutations failed to change the chromatin structure of the HML3 gene. Furthermore, a 102-bp silencer element was isolated from the HML flanking DNA. HML was also flanked by an autonomously replicating sequence (ARS) activity, but the ARS activity did not appear to be required for silencer function. K. lactis sir2 strains grown in the presence of ethidium bromide (EtBr) accumulated the drug, which interfered with the essential mitochondrial genome. Mutations that bypassed the requirement for the mitochondrial genome also bypassed the EtBr sensitivity of sir2 strains. Sir2p localized to the nucleus, indicating that the role of Sir2p to hinder EtBr accumulation was an indirect regulatory effect. Sir2p was also required for growth in the presence of high concentrations of Ni²⁺ and Cu²⁺.

MOST fungi have distinct cell types. In Saccharomyces cerevisiae, these cell types are the two haploid mating types, a and , and haploid cells of the opposite mating types can fuse and form the third cell type, the a/ diploid. In haploid strains of S. cerevisiae, there are three loci that encode mating type information. The mating type is determined by the allele present in the expressed MAT locus. In addition, there are two transcriptionally silent loci, one containing the a- and the other containing the -mating type information. These loci, called HMRa and HML, are kept silent despite the presence of functional promoters and structural genes (LOO and RINE 1994 ; LUSTIG 1998 ). The proteins encoded by the MAT locus are transcription factors that either repress or activate the transcription of target genes. If the a and information is expressed simultaneously, which occurs in a diploid cell, then mating pheromones and the receptors for the mating pheromones are not expressed (HERSKOWITZ et al. 1992 ). This means that a diploid cell cannot mate. Haploid cells unable to silence the transcription of the HM loci due to a mutation in a gene required for silencing behave like a diploid in these aspects and are thus sterile.

The silencing of the cryptic mating type loci is an example of a position effect on gene expression, and these loci appear to be silenced by the formation of a repressive chromatin structure functionally analogous to heterochromatin. Other genes inserted into the cryptic mating type loci can be silenced, thus ruling out gene-specific mechanisms for silencing (SCHNELL and RINE 1986 ). Moreover, histones H3 and H4 are involved in silencing (KAYNE et al. 1988 ) and the HMRa locus is resistant to DNA-modifying enzymes both in vivo and in vitro (SINGH and KLAR 1992 ; LOO and RINE 1994 ). Position effects on gene expression are a widespread phenomenon, and in yeasts position effects have also been found close to telomeres (GOTTSCHLING et al. 1990 ) and centromeres (ALLSHIRE et al. 1995 ) and in the rDNA locus (BRYK et al. 1997 ; FRITZE et al. 1997 ; SMITH and BOEKE 1997 ). In other eukaryotes, intensively studied examples include X chromosome inactivation in mammalian females and the position-effect variegation observed in Drosophila melanogaster (GRIGLIATTI 1991 ; RASTAN 1994 ; HENNIG 1999 ).

Several proteins have been identified that are required for silencing in S. cerevisiae. Some act by binding DNA sequences called silencers, which flank the cryptic mating type loci. The best-characterized silencer, HMR-E, contains binding sites for origin recognition complex (ORC), a protein complex involved in replication initiation, as well as binding sites for two widely used transcriptional activators, Rap1 and Abf1. Specific mutations in ORC2, ORC5, RAP1, and ABF1 lead to the derepression of HMRa (SUSSEL and SHORE 1991 ; FOSS et al. 1993 ; MICKLEM et al. 1993 ; LOO et al. 1995A , LOO et al. 1995B ). Other proteins seem to act on the silent loci by utilizing the silencer elements to establish a specialized chromatin structure, which is spread across the cryptic mating type loci. The most prominent members of these proteins are the silent information regulator (Sir) proteins, which are required for silencing at both the cryptic mating type loci and telomeres (IVY et al. 1986 ; RINE and HERSKOWITZ 1987 ; APARICIO et al. 1991 ). The Sir proteins are recruited to the cryptic loci by direct interaction to silencer binding proteins such as Rap1 (MORETTI et al. 1994 ). Sir2, Sir3, and Sir4 also interact with each other and histones H3 and H4 to form a repressing chromatin structure that facilitates silencing (HECHT et al. 1995 ; MOAZED et al. 1997 ). Sir proteins do not appear to interact directly with DNA. However, chromatin immunoprecipitation experiments show that both telomeric DNA and the cryptic mating loci contain Sir proteins as part of the chromatin at these loci (STRAHL-BOLSINGER et al. 1997 ).

This study contains the characterization of silencing of the cryptic -locus from the budding yeast, Kluyveromyces lactis. Comparing the Kluyveromyces cryptic mating type loci with their Saccharomyces counterparts revealed both similarities and differences. Among the differences, K. lactis HML contains an additional gene missing in S. cerevisiae. This 3 gene showed -specific expression, but was expressed from both HML and MAT in wild-type strains. A silencer element flanking HML was found, and this element did not contain any apparent binding sites for Rap1, Abf1, or ORC. As in S. cerevisiae, Sir proteins controlled the chromatin structure at HML. In the case of Sir2p, this protein appeared to control functions unrelated to mating type, telomeres, or rDNA.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Cloning and sequencing of HML and HMRa:
Escherichia coli strain DH5 was transformed with a genomic K. lactis library in plasmid pAB24 (2µm, URA3). Approximately 20,000 transformants were screened by colony hybridization using a PCR fragment corresponding to the K. lactis 1 open reading frame (ORF) as probe. Several plasmids were recovered from colonies that hybridized to the probe. The DNA sequence of the insert of one of these isolates (M1) and also in part two overlapping isolates obtained from S. Fields (A1 and B1; YUAN et al. 1993 ) was determined on both strands by primer walking. Clones encoding K. lactis HMRa were isolated in a similar fashion using a probe that was a PCR fragment that contained sequences from the 1 gene 3'-flanking region including the R sequence (Fig 1). This probe corresponded to nucleotides 74–700 downstream of the 1 stop codon. The sequence of two overlapping clones (B and K) was partially determined on both strands by primer walking. Sequencing was performed using a Prism sequencing kit (Applied Biosystems Inc., Foster City, CA) and a DNA sequencer (model 373; Molecular Dynamics, Sunnyvale, CA). These sequences have been submitted to the GenBank database under the accession nos. AF195066 (HML) and AF195067 (HMRa).

View larger version (10K): In this window In a new window Download PPT slide   Figure 1. Genomic organization of the K. lactis HML and HMRa loci. Arrows indicate the direction the genes are transcribed and the boxes labeled L and R correspond to the repeated sequences that were shared between the HML, HMR, and MAT loci. The star shows the position of the silencer.

Plasmid constructions:
The pRS306 1::KanMX plasmid (p350) was made in three steps. First a 3428-bp MunI fragment from HML containing clone M1 was cloned into the EcoRI site of pRS306 (SIKORSKI and HIETER 1989 ) forming p345. Plasmid 345 was subjected to in vitro mutagenesis (KUNKEL 1985 ) using an oligonucleotide (5'-GGAGCATTCGATTGCATGCTTTGGCGTTAC-3') that introduces a SphI site at the 1 gene start codon, which generated plasmid p348. A 900-bp PCR fragment containing the KanMX gene (WACH et al. 1994 ) with a SphI site at the start codon and an EcoRI site downstream of the ORF was cloned into the corresponding sites of p348. The resulting vector, p350, thus exchanged amino acids 1–122 of the 1 gene for a KanMX gene. The pRS306 3::LEU2 plasmid (p332) was made in three steps. A 4383-bp ClaI-EcoRI fragment from M1 containing the entire 3 ORF was cloned into the corresponding sites of pRS306, resulting in vector p322. Then p322 was digested with MunI, resulting in a deletion of internal 3 sequences corresponding to amino acids 138–678, and the vector fragment was ligated to an oligonucleotide (5'-AATTAGATCT-3'), resulting in the conversion of the MunI site into a BglII site. Into the resulting vector, a 2.9-kb BglII fragment containing the LEU2 gene from vector pCXJ20 (CHEN 1996 ) was cloned, thus resulting in vector p332 (pRS306 3::LEU2). The pRS306 hmlp (p400) construct was generated in a single step by cloning two PCR fragments corresponding to the 1 and 2 genes into pRS306, exchanging the promoter region in between the two genes for a BamHI-AlwI site. A 1747-bp XhoI-BamHI PCR fragment containing the 3' end of the 3 gene plus the 2 ORF and an 801-bp BamHI-XbaI PCR fragment containing the 1 ORF was combined with a XhoI-XbaI-digested pRS306 in a three-factor cloning, generating p400. To generate the pCXJ20-GFP-SIR2 construct, in which the fusion gene is driven by a glycerol phosphate dehydrogenase gene promoter (GPD), a 2.0-kb MunI-XbaI PCR fragment containing the entire SIR2 gene (CHEN and CLARK-WALKER 1994 ) with the MunI site inserted at the start ATG codon was cloned into the EcoRI-XbaI sites of vector pJW192 (2µm TRP1 GPD-GFP; WHISTLER 1996 ), thus generating an in-frame GFP-SIR2 fusion gene. Then a 3.4-kb HindIII-XbaI fragment from the resulting vector was cloned into the corresponding sites of pCXJ20(CEN ARS LEU2), resulting in p280. Plasmid 280 complemented strains containing the sir2::URA3 allele with respect to mating deficiency and ethidium bromide (EtBr) sensitivity. Plasmid 291 was a 3.2-kb PstI-NsiI fragment from A1 cloned into the PstI site of pCXJ18, and p300 was generated by cloning a 4.7-kb PstI-BglII fragment from A1 into the PstI-BamHI sites of pCXJ18. PCR fragments with a BamHI site in the 5' end and a KpnI site in the 3' end were cloned into the corresponding sites of p291. The PCR fragments corresponded to different parts of the 1 gene 3'-flanking DNA [the first base after the stop codon of the 1 gene was defined as nucleotide (nt) 1], p324 (nt 521–1078), p339 (nt 521–902), p311 (nt 726–1571), p386 (nt 521–818), p357 (nt 563–1078), and p414 (nt 521–622). Some PCR fragments were also cloned into the BamHI-KpnI sites of pRS306 to assay for ARS activity, p302 (nt 356–1571), p316 (nt 521–1078), p333 (nt 521–902), and p305 (nt 726–1571). A 1.6-kb HindIII-XbaI fragment containing the E. coli dam⁺ gene from vector pJR1830 (J. RINE, unpublished results) was cloned into the corresponding sites of pCXJ20 (LEU2) and pCXJ18 (URA3), generating vectors p365 and p401, respectively.

Yeast media and methods:
Rich media, sporulation media, and conditions for matings were as described (CHEN and CLARK-WALKER 1994 ; ADAMS et al. 1998 ). Specifically, matings were performed by mixing 2 x 5 µl of suspensions (A600 = 2) of the two mating partners on a 2 x 2-cm area on a YEPD plate. The mating plate was incubated for 18–24 hr at 30°, replica-plated onto selective medium, and incubated for 72 hr before the results were scored. Preparation of yeast RNA/DNA and conditions for RNA and DNA blots were described before (ADAMS et al. 1998 ). Quantification of signals was performed using a Phosphorimager (Molecular Dynamics) and Imagequant software. In the dam-methylase assay, an [-³²P]dCTP-labeled PCR fragment corresponding to the 1 ORF was used as probe. To detect the AlwI site in the 1 gene, the DNA (2 µg) was digested with EcoRV and AlwI for 4 hr at 37° and the 572-bp band (cut) was compared to the 957-bp (uncut) band. The AlwI site in between 1 and 2 genes was detected by digesting with ClaI, HindIII, and AlwI for 4 hr at 37° and the 520-bp (cut) and 760-bp (uncut) bands were compared. For the BclI site in the 3 gene, the DNA was digested with ClaI (2 hr, 37°) and BclI (4 hr, 50°) and the 2.7-kb (cut) and 4.3-kb (uncut) bands were compared. This procedure was repeated with independent DNA preparations with very similar results. EtBr staining was performed by adding the drug at 0.5 µg/ml to logarithmically growing cultures (A600 = 0.5). The cells were harvested after 1–4 hr and prepared for fluorescence microscopy by staining with DAPI. Cation sensitivity assays were performed by adding a drug disc, to which 6 µl of a specific cation solution had been added previously, to a YEPD plate with a uniform lawn of the tester strain. The plate was incubated for 48 hr and the inhibition zone caused by the drug disc was measured. The following solutions were tested: saturated CuSO₄, 2 M NiCl₂, 2 M MnCl₂, 0.5 M CoCl₂, and saturated PbSO₄.

Strain constructions:
The strains used in this study are listed in Table 1. Strain SAY119 was generated by crossing strain CK213 with SAY102 followed by tetrad analysis. Strains with mat1::KanMX or hml1::KanMX alleles were generated by a two-step gene disruption procedure (SCHERER and DAVIS 1979 ) by transforming SAY119 (MAT) and CK213 (MATa) with PacI-linearized p350. This procedure generated strains carrying either the mat1::KanMX or the hml1::KanMX disruptions and the allele present in a particular strain was identified by DNA blot hybridizations, generating strains SAY128 (mat1::KanMX), SAY129 (MAT hml1::KanMX), and SAY130 (MATa hml1::KanMX). An identical procedure was used to generate the mat3::LEU2 and hml3::LEU2 alleles except that strain SAY119 was transformed with PacI-linearized p332, generating strains SAY120 (hml3::LEU2) and SAY121 (mat3::LEU). The double mutant mat3::LEU2 hml3::LEU2 strain was generated in two steps. Crossing strain SAY120 with CK213 generated a MATa hml3::LEU2 strain (SAY122). SAY122 was crossed to SAY121 and from this cross the double mutant mat3::LEU2 hml3::LEU2 strain (SAY124) was recovered. Since HML and MAT were tightly linked, no nonparental ditypes with respect to the LEU2 gene were found in this cross so the double mutant strain was identified by a DNA blot hybridization using genomic DNA from haploids arising from a meiosis with a tetratype segregation. Plasmid 400 was linearized with PacI and transformed into CK213 to generate the MATa hmlp strain SAY186 by a two-step gene disruption procedure. Crossing SAY186 to CK57-7A (MAT sir2::URA3) and SAY 96 (MAT sir4::LEU2), respectively, generated the hmlp sir2::URA3 (SAY189) and hmlp sir4::LEU2 (SAY191) double mutant strains. The hml1::KanMX sir2 double mutant strains, SAY155 and SAY156, were spontaneous G418-resistant isolates from strains SAY129 and SAY130, respectively. These isolates behave as sir2 null alleles with respect to mating defects and EtBr sensitivity and are complemented by a single copy plasmid carrying SIR2.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The K. lactis cryptic -locus (HML) contains three genes:
To learn more about mating type in K. lactis, we cloned and characterized a locus encoding -mating type information. We used a functional homologue of S. cerevisiae 1 from Kluyveromyces (YUAN et al. 1993 ) to screen a K. lactis genomic plasmid library to identify a complete -locus. This locus contained three putative genes (Fig 1). The 1 gene, encoding a protein of 261 amino acids sharing 30% identity with S. cerevisiae 1p, can functionally complement a S. cerevisiae 1^- mutant (YUAN et al. 1993 ). A second gene was found that encoded a protein that shared 31% identity with Saccharomyces 2p. As in S. cerevisiae, the 1 and 2 mating type genes in K. lactis were divergently transcribed from a common promoter region. On the 3' end of the coding strand for 2, a third large open reading frame was found, which we called 3. The putative peptide encoded by this gene was 897 amino acids and lacked all significant homology to other proteins in GenBank. Since we previously determined that K. lactis contains both an expressed and a silent -locus (ASTROM and RINE 1998 ), we had to determine which of these loci we had cloned. The restriction map of our sequence combined with DNA blot analysis of MATa and MAT strains showed that the locus sequenced by us was the cryptic -locus (data not shown), which we thus call K. lactis HML.

To compare the sequences shared between MAT and HML, as a prelude to identifying a silencer, DNA blots were performed with various DNA probes from the 1, 2, and 3 genes. The blots showed that all three genes were duplicated in the K. lactis genome and thus present at both the MAT and HML loci. Furthermore, these data revealed that MAT and HML shared at least 304 bp, but <885 bp of flanking DNA upstream of the 3 start codon (as drawn) and 370 bp, but <750 bp downstream of the 1 stop codon (as drawn).

The K. lactis cryptic a-locus contains two genes:
When a probe corresponding to the 1 3'-flanking DNA was used on DNA blots from genomic K. lactis DNA, three bands were present. Two bands were of invariant size, but one band's size was cell type specific, and thus varied between MATa and MAT strains (data not shown). We speculated that this flanking sequence was shared among the MAT, HML, and HMR loci and that the variable sized band corresponded to a MAT locus containing restriction fragment. This speculation was later confirmed when the silent a-locus (HMRa) was cloned from a genomic library, using a probe corresponding to the 1 3'-flanking DNA. The sequence of HMRa revealed that HML and HMRa shared common flanking sequences on both sides (Fig 1), and we call these sequences L (left) and R (right; Fig 1). The L sequence on the left side of both HML and HMRa (as drawn) was 250 bp long and the R sequence on the right side of both HML and HMRa (as drawn) was 360 bp long. Presumably the L and R sequences act as homologous blocks for resolving recombination intermediates during mating type interconversion. Further sequence analysis of K. lactis HMRa revealed the presence of two ORFs. One of these ORFs (228 amino acids) shared 46% identity in the last 50 carboxyl-terminal amino acids with the Saccharomyces a1 protein. This domain corresponded to the homeodomain of Saccharomyces a1p. The sequence of the K. lactis a1 gene indicated the presence of an intron in the gene. The intron splice donor/acceptor sites were in frame with the potential unspliced message, so further experiments are required to determine the length of the a1 protein. The other ORF (256 amino acids) was not similar to the Saccharomyces a2 gene. Rather this gene shared similarities to genes from the mating type loci of other yeasts, such as the sporulation minus regulator 2 (SMR2) gene from Podospora anserina (DEBUCHY et al. 1993 ) and the MAT 1-3 gene from Gibberella fujikuroi. The proteins encoded by these genes all contain a high mobility group (HMG) DNA binding motif, characteristic of proteins that bind DNA in the minor groove.

Mat1p is required for -mating proficiency:
To learn more about the function of the genes encoded by the -locus, we generated strains in which the 1 gene at MAT and HML was replaced by a KanMX gene, which encodes a protein that mediates resistance against the aminoglycoside Geneticin. The replacement was constructed such that the promoter of the 1 gene transcribed the KanMX gene (see MATERIALS AND METHODS). Strains with mat1::KanMX or hml1::KanMX alleles were then tested for mating ability with a MATa tester strain. The mat1::KanMX mutant exhibited a large mating defect, whereas the hml1::KanMX mutant was indistinguishable in mating efficiency from the wild-type control strain (Fig 2A). Thus, Mat1p was required for -mating proficiency and the HML1 gene did not contribute to the 1 protein function. It was interesting to note that in quantitative mating determinations with a large surplus of the MATa strain, the mat1::KanMX strain did not show a severe mating defect (data not shown). This behavior is fundamentally different from the corresponding Saccharomyces mutant strains, in which 1^- null mutants are unable to mate even in the presence of a large excess of MATa cells.

View larger version (55K): In this window In a new window Download PPT slide   Figure 2. (A) The MAT1 but not the HML1 gene was required for -mating proficiency. Patch matings of strains SAY119 (MAT), SAY128 (mat1::KanMX), and SAY129 (hml1::KanMX) are shown. The MATa mating type tester strain was CK213 and diploids were selected on synthetic dextrose plates supplemented with uracil, leucine, tryptophane, and methionine (SD + Ura, Leu, Trp, and Met). (B) The silencing of the hml1::KanMX gene required Sir2p. Patches of strains SAY129 (hml1::KanMX) and SAY155 (hml1::KanMX sir2) were replica-plated onto rich media (YEPD) containing 25 µg/ml Geneticin and incubated for 72 hr at 30°.

Silencing at HML is not specific to -genes:
The phenotypic difference between 1::KanMX insertions at MAT1 and HML1 was most likely due to the transcriptional silencing of the HML locus. Consistent with this model, the hml1::KanMX strain was sensitive to Geneticin (Fig 2B), thus implying that the KanMX gene was subject to repression when integrated at the HML locus. An alternative hypothesis was that the hml1:: KanMX allele was nonfunctional, due perhaps to the PCR procedure employed to generate the disruption construct. To distinguish between these two possibilities we combined the hml1::KanMX allele with a mutation in the sir2 gene and compared the growth of the hml1::KanMX strain with the double mutant hml1:: KanMX sir2 strain on media containing Geneticin. The double mutant strain grew on plates containing 25 µg/ml of Geneticin, but the hml1::KanMX strain did not (Fig 2B). This result thus confirmed that the KanMX gene present at HML was functional, yet not expressed. Thus, the silencing at HML was not gene specific but was Sir2p dependent.

Either Mat3p or Hml3p is required for -mating proficiency:
The 3 gene at K. lactis HML is missing from both the MAT and HML loci of S. cerevisiae, which prompted us to investigate the function and regulation of the 3 gene. A 3 null allele was made, in which approximately two-thirds of the 3 ORF was exchanged for a functional LEU2 gene. This construct was exchanged for the wild-type 3 gene at both the MAT and HML loci, thus generating mat3::LEU2 and hml3::LEU2 strains. By analogy to the results presented above, we expected that the hml3::LEU2 strain would require the addition of leucine to the medium for optimal growth due to silencing at HML, whereas the mat3::LEU2 strain would not. Both the mat3::LEU2 and hml3::LEU2 strains were complete leucine prototrophs (data not shown), indicating either that the HML3 gene was not silenced or that the LEU2 promoter was resistant to silencing.

To distinguish between these two possibilities, we first tested the mating proficiency of both mat3::LEU2 and hml3::LEU2 strains (Fig 3A; data not shown). Surprisingly, neither strain showed a mating defect. Given that the HML3 and MAT3 genes are identical in sequence we investigated the possibility that both genes were expressed. We thus generated a double mutant mat3::LEU2 hml3::LEU2 strain by mating two single mutant strains. The resulting diploid, homozygous for leu2, was sporulated and tetrad analysis revealed linkage between MAT and HML of 10 cM based upon 31 parental ditypes, 8 tetratypes, and no nonparental ditypes from a total of 39 tetrads. We then tested the double mutant mat3::LEU2 hml3::LEU2 strain in a mating assay and found that this strain mated with a low efficiency (Fig 3A). Thus, the 3 gene appeared to be required for efficient mating as an -cell and both the MAT3 and HML3 genes seemed to be expressed.

View larger version (83K): In this window In a new window Download PPT slide   Figure 3. (A) The MAT3 and HML3 genes were redundant for a function that ensured optimal -mating proficiency. Patch matings of strains SAY121 (mat3::LEU2) and SAY124 (mat3::LEU2 hml3::LEU2) are shown. The MATa mating type tester strain used was CK213 and diploids were selected on SD + Ura, Lys, Trp, and Met. (B) The 3 gene was an -specific gene that required both Sir2p and Sir4p to remain transcriptionally repressed. RNA blot hybridization of K. lactis total RNA from strains WM52 (MAT), CK213 (MATa), CK213XWM52 (MATa/), SAY99 (MATa sir2::URA3), and SAY90 (MATa sir4::LEU2). The blot was probed with a radiolabeled PCR fragment corresponding to the middle portion of the 3 gene (top), and then the blot was stripped and reprobed with an actin probe (bottom). The migration of the 3 transcript indicated a size of 2.5–3.0 kb, consistent with the size of the 3 ORF.

To investigate the regulation of the 3 gene we performed RNA blots, probing for the 3 transcript, on RNA from different cell types (Fig 3B). The experiment showed that the 3 transcript was present in both the - and a/-cell types, but not in the a-cell type. Furthermore, MATa strains that contained mutations in either sir2 or sir4 also expressed the 3 transcript. We noted that the 3 transcript was a doublet band and in the sir2 strain the slower migrating band was more abundant than in the wild-type MAT strain, but at the moment we do not understand the significance of this observation. Thus, 3 was an -specific gene expressed from both MAT and HML.

Expression of 2p inhibits mating of MATa strains:
In Saccharomyces, simultaneous expression of a and information leads to the formation of the a1/2 heterodimer, which represses the transcription of haploid- specific genes (HERSKOWITZ et al. 1992 ). Among the haploid-specific genes are those encoding the trimeric G-protein required for mating pheromone receptor signaling, whose repression leads to the sterility of cells expressing the a1/2 repressor. As in Saccharomyces, introduction of a plasmid encoding the 1 and 2 genes into K. lactis MATa strains inhibited mating, indicating that K. lactis mating was also subject to a/ repression (see below). Plasmids carrying only the 1 gene did not inhibit mating of MATa strains, indicating that it was 2p and not 1p that inhibited mating. To confirm this notion, we tested the mating proficiency of two double mutant strains. One strain contained the hml1::KanMX allele and the other strain contained a promoter deletion of the entire region between the 1 and 2 genes, hmlp (see MATERIALS AND METHODS), and thus expressed neither 1p nor 2p. These two hml mutant alleles were combined with sir2 mutations in MATa strains and the mating proficiencies of the resulting strains were determined (Fig 4). Only the hmlp allele suppressed the mating defect caused by the sir2 mutation, thus confirming that it was the expression of 2p and not 1p that inhibited mating of the MATa strain. Since the HMRa locus contained an a1-like gene (Fig 1) we propose that diploid K. lactis strains also contain an a1/2 repressor.

View larger version (38K): In this window In a new window Download PPT slide   Figure 4. Expression of 2p inhibited mating of MATa strains. Patch matings of strains CK213 (MATa), SAY124 (MATa sir2::URA3), SAY156 (MATa sir2 hml1::KanMX), and SAY189 (MATa sir2::URA3 hmlp) are shown. The MAT mating type tester strain used was WM52. Diploids were selected on SD + Ura plates.

A silencer and an ARS flanked K. lactis HML:
In Saccharomyces, the transcriptional silencing of HML and HMRa requires flanking DNA elements called silencers. To investigate whether DNA elements flanking the K. lactis HML locus were required for silencing, we developed a plasmid-based assay for silencing. This assay measured the mating proficiency of a MATa strain containing plasmids with HML sequences (Fig 5A). Since the simultaneous expression of a and information inhibits mating in K. lactis, mutant plasmids that inhibited mating of the MATa strain contained expressed 1 and 2 genes. The HML DNA fragments tested contained the entire 1 and 2 genes, but included only half of the 3 gene (Fig 5B). A plasmid (p291) that contained only 78 bp of DNA downstream of the 1 gene could inhibit mating of the MATa strain, but a plasmid (p300) that included 1.6 kb of flanking DNA did not affect the mating ability of the tester strain. Thus, within these 1.6 kb there was a DNA sequence that inhibited transcription of HML in a cis-dominant manner. Furthermore, this putative silencer was not likely to be located within the first 370 bp downstream of the 1 gene, since this flanking sequence was shared between the MAT and HML loci. To pinpoint further the potential silencer element, PCR fragments of the DNA flanking HML were cloned into p291, and the resulting plasmids were tested for their ability to inhibit mating of MATa strains. This analysis showed that a 102-bp fragment (p414), corresponding to nucleotides 521–622 downstream of 1 gene, completely retained silencing activity in this assay. An additional deletion of 50 bp of this minimal silencer (p357) completely abolished silencing (Fig 5B).

View larger version (74K): In this window In a new window Download PPT slide   Figure 5. (A) A plasmid-based assay for silencer function. Patch matings of strain CK213 transformed with the indicated plasmids. Vector indicates the pCXJ18 (CEN ARS URA3) plasmid, p291 is pCXJ18 with the HML1 and 2 genes without flanking DNA, p300 was p291 + 1.6 kb of flanking DNA, p386 was p291 + 297 bp of flanking DNA, and p414 was p291 + the 102-bp minimal silencer. The MAT mating type tester strain used was WM52 and diploids were selected on SD plates. (B) A silencer and an ARS activity flanked HML. Schematic drawing of the HML locus indicating the PCR fragments used to pinpoint the localization of the silencer and ARS activity. - means that the plasmid abolished mating of MATa strains CK213 (Silent) or that the DNA fragment was unable to confer a high transformation efficiency to pRS306 (ARS). + means that the plasmid did not affect mating of MATa strain CK213 (Silent) or that the DNA fragment conferred a high transformation efficiency to pRS306 (ARS). nd, not determined. The names of the plasmids are indicated in parentheses.

The S. cerevisiae HMR-E and HMR-I silencers are also origins of replication. To investigate if this feature of a silencer was conserved between yeasts, we tested if the DNA flanking the K. lactis HML locus contained autonomously replicating sequence (ARS) activity. Different DNA fragments from the HML flanking DNA were cloned into a yeast vector lacking origins of replication (pRS306) and transformed into K. lactis. The transformation frequency of vector without insert was very low due to the inability of the transformants to replicate the plasmid DNA. Several plasmids containing K. lactis DNA increased the transformation frequency by several orders of magnitude (Fig 5B), which indicated that an ARS was present close to HML. These transformants grew on selective media, but lost the selectable marker following nonselective growth. Interestingly, the ARS was found close to the minimal silencer element, but was not necessary for silencing activity (Fig 5B). Thus, K. lactis HML was flanked by a silencer element and by a functionally separable ARS activity.

Sir2p and Sir4p were required for maintaining a specialized chromatin structure at HML in the absence of transcription:
Unlike other fungi, budding and fission yeasts lack endogenous DNA methylation. This absence allows one to investigate the chromatin structure at different loci, by expressing foreign DNA methylases and assaying the accessibility of specific sequences to these methylases (SINGH and KLAR 1992 ). We used such an assay to determine the effect of sir2 and sir4 mutations on the chromatin structure of HML. Others have shown (SINGH and KLAR 1992 ) that actively expressed genes are more accessible to the E. coli dam-methylase than are silenced genes. Since both sir2 and sir4 mutations result in a partial derepression of HML1 expression (ASTROM and RINE 1998 ) we generated a transcriptionally compromised HML allele to distinguish between effects on transcription from effects on chromatin structure that were independent from transcription. We thus generated an HML allele, in which the entire promoter region of the divergently transcribed 1 and 2 genes was replaced with an AlwI site. Then we combined this hmlp allele with sir2 and sir4 mutations and introduced the dam⁺ gene on a plasmid. Dam-methylase accessibility was determined by digesting chromosomal DNA from these strains with restriction enzymes that were sensitive to a methyl group at its target site. We investigated such sites at three different positions (Fig 6), an AlwI site in the middle of the 1 open reading frame, an AlwI site between the 1 and 2 genes, and a BclI site present in the 3 gene. Bands generated from DNA molecules in which AlwI or BclI were unable to cut the chromosomal DNA were normalized to the total DNA in each lane. We could thus determine the methylation levels in sir2 and sir4 mutant strains relative to the level in a wild-type strain. Both the sir2 and sir4 strains showed increased accessibility of the AlwI site present in the 1 gene compared to the wild-type strain. This was expected since both of these genes are required for complete silencing of HML1 transcription. More surprising was the effect observed at the AlwI site in between the 1 and 2 genes. In this case only the sir2 strain showed increased accessibility, whereas the sir4 mutant strain showed similar accessibility compared to the wild-type strain. The BclI site found in the 3 open reading frame showed less than twofold differences in accessibility between all strains tested. These data thus indicated that the silencing was weak in the 3 region of HML, which was expected since the genetic analysis indicated that 3p was expressed from the HML locus in silencing-proficient strains. Moreover, these data indicated that sir2 and sir4 mutations affected the chromatin structure to different extents at HML.

View larger version (37K): In this window In a new window Download PPT slide   Figure 6. Sir2p and Sir4p maintained a specialized chromatin structure at HML. Dam-methylase accessibility assay of chromosomal DNA from strains SAY186 (hmlp), SAY186 + p365 (pCXJ20-dam+), SAY189 (hmlp sir2::URA3) + p365, and SAY191 (hmlp sir4::LEU2) + p401 (pCXJ18-dam+). A BclI site in the 3 gene (left), an AlwI site between the 1 and 2 genes (middle), and an AlwI site in the 1 gene (right) were assayed for the efficiency of enzyme digestion. For comparisons, the bands generated from when the enzyme did cut are shown (cut, 12-hr exposure) above the panels in which the site was protected by dam-methylation (uncut, 24-hr exposure). The HML locus is shown schematically, indicating the deletion of the promoter region between the 1 and 2 genes, and the star indicating the silencer. The methylation levels relative to strain SAY186 + p365 are shown below.

Sir2p regulated cation sensitivity in K. lactis:
K. lactis sir2 strains are hypersensitive to the DNA intercalating drug EtBr (CHEN and CLARK-WALKER 1994 ). To investigate what this sensitivity might be due to, we grew sir2 and wild-type strains in the presence of EtBr. Microscopic examination of sir2 cells grown in the presence of EtBr revealed that the drug accumulated inside of the cells in a punctate pattern, similar to mitochondrial staining (Fig 7A). This accumulation was rapid, being visible by 1 hr after the addition of the drug to logarithmically growing cultures. At identical EtBr concentrations the wild-type strain showed only diffuse low-level staining (Fig 7A), showing that Sir2p function blocked accumulation of EtBr inside the yeast cells. EtBr preferentially intercalates into mitochondrial DNA in living cells. Thus, the role of Sir2p in preventing EtBr accumulation is consistent with Sir2p's role in maintaining the mitochondrial genome in the presence of EtBr (CHEN and CLARK-WALKER 1994 ). K. lactis is a so-called "petite-negative yeast," meaning that, unlike S. cerevisiae, respiratory-deficient K. lactis strains cannot be isolated under normal circumstances because the mitochondrial genome is essential in K. lactis (CLARK-WALKER and CHEN 1996 ) even for growth on fermentable carbon sources. There are, however, a group of genes called mitochondrial genome integrity (mgi) genes, that, when mutated, allow growth of K. lactis in the absence of the mitochondrial genome (CHEN and CLARK-WALKER 1993 ). Thus, if the sole reason sir2 mutants in K. lactis were EtBr sensitive was because EtBr caused the loss of the mitochondrial genome, then mgi1 sir2 double mutant strains should grow in the presence of EtBr. Indeed, the mgi1-1 mutation completely suppressed the EtBr sensitivity of sir2 strains (Fig 7B). The mgi1-1 sir2 double mutant strain was, in fact, more resistant to EtBr than a wild-type strain. The mgi1-1 sir2 strain still accumulated EtBr and consequently all mgi1-1 sir2 cells grown in the presence of EtBr were converted into respiratory-deficient petites (data not shown). Taken together these data showed that the toxic effect of EtBr in sir2 strains was solely due to the interference of the mitochondrial genome. The molecular explanation for how Sir2p stops EtBr accumulation remains unclear, but since a functional Sir2-Gfp fusion protein was localized exclusively to the nucleus of K. lactis (Fig 7C), it would appear that the effect of Sir2p in this respect is mediated by a role of that protein outside of the mitochondria, such as regulating a transport function. We tested whether this potential transport defect could affect the growth of sir2 strains in the presence of other compounds. It was found previously that the DNA intercalating drug Berenil, which like EtBr is a cation, inhibited growth of sir2 strains (CHEN and CLARK-WALKER 1994 ). Therefore we tested the relative sensitivities of a sir2 strain either transformed with plasmid alone or transformed with a plasmid encoding a functional SIR2-GFP fusion gene to several different cations. The sir2 strain showed increased sensitivity to Cu²⁺ and Ni²⁺ ions, but not to Mn²⁺, Zn²⁺, or Pb²⁺ ions (Fig 7D; data not shown). Thus, Sir2p appeared to regulate the transport of several cations in K. lactis.

View larger version (200K): In this window In a new window Download PPT slide   Figure 7. (A) Sir2 mutant strains accumulated EtBr. Nomarski and fluorescence micrographs of viable cells from strains CK56-7A (mgi1-1) and CK164 (mgi1-1 sir2::URA3) grown in the presence of 0.5 µg/ml EtBr. (B) A mgi1-1 mutation completely suppressed the EtBr sensitivity of sir2 strains. Serial dilutions (10-fold) of overnight cultures of strain CK213 (wild-type), SAY102 (sir2::URA3), CK56-7A, and CK164 were spotted onto either YEPD or YEPD plates containing 0.5 or 3 µg/ml EtBr. (C) Sir2p localized to the nuclei of K. lactis cells. Micrographs of strain SAY102 (sir2::URA3) transformed with p279 (pCXJ20-GFP-SIR2). Nomarski, DAPI, and GFP-SIR2 fluorescence of viable cells are indicated. (D) Sir2p was required for growth in the presence of high concentrations of Ni2+ and Cu2+ ions. Solutions of the indicated cations were spotted on a lawn of strain SAY102 (sir2::URA3) transformed with pCXJ20 or pCXJ20-GFP-SIR2. The relative sensitivities to the cation was determined by growing the cells on synthetic complete (SC) plates lacking leucine for 36 hr at 30°. Growth inhibition zones were: Cu2+: SIR2 13 mm, sir2 18 mm; Ni2+: SIR2 9 mm, sir2 12 mm.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

This study presented the first characterization of the HMRa and HML loci in the yeast K. lactis. The HML locus contained three genes that were transcriptionally silent in MATa cells and silencing of all three genes required Sir2p and Sir4p (Fig 3; ASTROM and RINE 1998 ). Given the sequence conservation between the Saccharomyces and Kluyveromyces 1 and 2 genes and the fact that the K. lactis 1 gene also is functionally interchangeable between yeasts (YUAN et al. 1993 ), these two genes probably have similar or identical roles in the two yeasts. Consistent with this notion, the 1 gene was required for mating proficiency of -strains in K. lactis, just as 1p is required for mating proficiency of -strains in Saccharomyces. K. lactis is likely to form an a1/2 heterodimer that inhibits haploid-specific gene expression in a/ diploids, since MATa and HMRa contained a gene homologous to the Saccharomyces a1 gene and expression of 2p inhibited mating in MATa strains. The third gene at K. lactis HML did not show significant similarities to other known genes. Single mutant hml3:: LEU2 or mat3::LEU2 strains did not show a mating defect, but a hml3::LEU2 mat3::LEU2 double mutant strain did. This indicated that the 3 gene promoted the -mating phenotype and was expressed from both HML and MAT. Moreover, a DNA methylase assay, which measured transcriptionally independent effects on the chromatin structure of HML, revealed that a BclI site present in the 3 gene was much less affected by the absence of Sir2p or Sir4p than an AlwI site in the 1 gene. Thus, the 3 gene did not appear to be part of the silent domain at HML. The reason for the absence of the 3 transcript in MATa cells is most likely due to the lack of 1p, since preliminary observations indicated that the 3 transcript was absent in a mat1:: KanMX strain.

Our analysis suggested that the 3 protein was required for optimal -mating, but the exact function of 3p was not revealed. Based on the knowledge of -specific genes in Saccharomyces, it seems reasonable to assume that 3p was either involved in -factor maturation or a-factor receptor activity. Since the identity of neither -factor nor the a-factor receptor is known in K. lactis we have been unable to test these ideas. MAT loci from several other yeasts have been identified and most of the genes encoded by these loci correspond to transcriptional regulators. Recently however, the sequence of the mating type-like (MTL) loci from the asexual yeast Candida albicans revealed the presence of genes encoding oxysterol binding proteins, phosphatidylinositol kinases, and poly(A) polymerases (HULL and JOHNSON 1999 ). In Candida, it is not known if the genes in the MTL loci are involved in a mating type-like process, but this observation illustrates that the genes encoded by MAT loci in yeasts cannot always be immediately connected to mating.

The silencer we identified downstream of the 1 gene was apparently sufficient to silence the 1 and 2 genes in a plasmid assay. At face value these data indicate that in K. lactis HML is flanked by only one silencer and not two, like HML in S. cerevisiae. It should be noted that at HML in S. cerevisiae either silencer is sufficient to silence HML. In Saccharomyces, boundary elements have been described close to HMRa that protect euchromatic genes close to HMRa from heterochromatin (DONZE et al. 1999 ). However, more experiments are required to determine whether or not K. lactis has discrete boundary elements and if so, where they are located at HML. The first long ORF downstream of the K. lactis HML1 gene was found 3.0 kb from the 1 stop codon. An RNA blot analysis using a probe corresponding to our partial sequence of this gene revealed mRNA expression in both SIR⁺ and sir^- cells, indicating that the silent domain did not extend into the promoter of this gene (data not shown). Since this gene was transcribed on the opposite strand from the HML1 gene we did not determine the exact distance to its promoter.

The dam-methylase assay that measured transcriptionally independent effects on the chromatin structure of HML revealed a difference between strains lacking Sir2p and Sir4p. Both strains showed increased dam accessibility in the HML1 gene, but only the sir2 strain showed increased accessibility at a site in between the HML1 and HML2 genes. This result was consistent with sir2 mutations derepressing HML1 transcription to a higher degree than sir4 strains do (ASTROM and RINE 1998 ). One explanation for this effect is that Sir4p does not spread along the chromatin, starting from the silencer, to the same extent as Sir2p. Others showed previously that in Saccharomyces Sir3p could spread from telomeres further than either sir2p or Sir4p, indicating that not all Sir proteins may be part of the same complex (HECHT et al. 1996 ; STRAHL-BOLSINGER et al. 1997 ).

Silencing in both Saccharomyces and Kluyveromyces requires silencer elements. Other similarities are that Sir proteins are required for silencing, and marker genes integrated at HML can be silenced in a Sir-dependent manner, in both yeasts. Downstream of the HML1 gene we identified a 102-bp fragment that was sufficient to mediate silencing. In the same assay a 300-bp DNA fragment from the HMRa locus, corresponding to a region located downstream of the a1 gene, could also mediate silencing of the HML locus in the plasmid assay (data not shown). Thus, both silent loci appear to be flanked by silencers. Comparisons of the DNA sequences of these two regions should facilitate the identification of DNA sequence motifs that are important for silencing. An ARS activity close to the cryptic mating type loci appeared to be evolutionarily conserved between yeasts, suggesting that ORC binding was important for silencing also in Kluyveromyces. However, we could delete the sequences that mediated the ARS activity and still retain full silencer activity. This observation did not exclude the possibility that ORC binding plays a role in silencing. For example, at the S. cerevisiae HMR-E silencer, the deletion of the Rap1p, Abf1p, or ORC binding sites individually does not abolish silencing, but the deletion of two of the sites simultaneously does (BRAND et al. 1987 ). Perhaps the K. lactis HML silencer shows similar redundancy. The consensus DNA binding sites for K. lactis Rap1p and Abf1p are known. However, the 102-bp silencer did not contain any close matches to the consensus binding sites for Rap1p or Abf1p. Thus, silencers in K. lactis must be quite different compared to the Saccharomyces silencers, and studying these silencers should thus lead to a deeper understanding of both the architecture of a silencer and the mechanism of silencing.

K. lactis sir2 strains grow slowly compared to wild-type strains (data not shown), a phenotype that is not observed in Saccharomyces sir2 strains. Thus, the SIR2 gene has different and perhaps more important functions in K. lactis compared to S. cerevisiae. One of these differences was that K. lactis Sir2p was required for hindering accumulation of EtBr inside of the cell. A mutation that bypassed the need for the mitochondrial genome also relieved the EtBr sensitivity of sir2 strains. The effect that Sir2p had on EtBr accumulation was likely to be indirect, since Sir2p appeared to be exclusively localized to the nucleus. Sir2 strains were also more sensitive to both Cu²⁺ and Ni²⁺ ions compared to a wild-type strain, suggesting that Sir2p regulated either the intake or efflux of a subset of cations in K. lactis. Perhaps a gene required for the transport of cations is close to a telomere in K. lactis and the absence of Sir2p may derepress such a gene, similar to the effect sir2 mutations have on telomeric position effect (TPE) in S. cerevisiae. However, sir4 mutations in K. lactis are slightly more resistant to EtBr compared to the wild type. Mutations in sir4 affect the telomere length in K. lactis and are thus likely to affect TPE. If Sir2p has a role in cation transport in K. lactis, it would appear to perform this function independently of Sir4p. The EtBr sensitivity of sir2 strains cannot be an indirect effect of the simultaneous expression of a and information, since MATa sir2 hmlp strains still were sensitive to EtBr. A more likely model is thus that Sir2p has a different regulatory role in K. lactis, perhaps as a part of a complex with other molecules. Such distinct Sir2p-containing complexes have been observed already in Saccharomyces, in which Sir2p has a Sir4p-independent role in regulating the chromatin structure at the rDNA locus (SMITH et al. 1998 ; STRAIGHT et al. 1999 ).

Since SIR2-like genes are present in organisms ranging from archea to mammals, it is likely that Sir2p-like molecules are involved in many different processes. In this study we have demonstrated one such role in what appears to be regulation of cation transport. Further studies of Sir proteins and the cryptic mating type loci in K. lactis are likely to reveal more interesting features of heterochromatin and new dimensions to the role of silencing proteins.

	ACKNOWLEDGMENTS

We thank all members of the Rine laboratory for interesting discussions, Dr. Clark-Walker and Dr. Chen for K. lactis strains and plasmids, Dr. S. Fields for plasmids containing the K. lactis -locus, and Dr. A. Johnson and Dr. C. Hull for sharing the sequence of the K. lactis 2 gene. This study was supported by a European Molecular Biology Organization postdoctoral fellowship, the Swedish Natural Science Research Council (B-AA/BU 11279-306 to S.U.Å.), and the National Institutes of Health (GM-31105 to J.R.).

Manuscript received January 4, 2000; Accepted for publication May 22, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ADAMS, A., D. E. GOTTSCHLING, C. A. KAISER and T. STEARNS, 1998 Methods in Yeast Genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

ALLSHIRE, R. C., E. R. NIMMO, K. EKWALL, J. P. JAVERZAT, and G. CRANSTON, 1995  Mutations derepressing silent centromeric domains in fission yeast disrupt chromosome segregation. Genes Dev. 9:218-233[Abstract/Free Full Text].

APARICIO, O. M., B. L. BILLINGTON, and D. E. GOTTSCHLING, 1991  Modifiers of position effect are shared between telomeric and silent mating-type loci in S. cerevisiae.. Cell 66:1279-1287[Medline].

ÅSTRÖM, S. U. and J. RINE, 1998  Theme and variation among silencing proteins in Saccharomyces cerevisiae and Kluyveromyces lactis.. Genetics 148:1021-1029[Abstract/Free Full Text].

BRAND, A. H., G. MICKLEM, and K. NASMYTH, 1987  A yeast silencer contains sequences that can promote autonomous plasmid replication and transcriptional activation. Cell 51:709-719[Medline].

BRYK, M., M. BANERJEE, M. MURPHY, K. E. KNUDSEN, and D. J. GARFINKEL et al., 1997  Transcriptional silencing of Ty1 elements in the RDN1 locus of yeast. Genes Dev. 11:255-269[Abstract/Free Full Text].

CHEN, X. J., 1996  Low- and high-copy-number shuttle vectors for replication in the budding yeast Kluyveromyces lactis.. Gene 172:131-136[Medline].

CHEN, X. J. and G. D. CLARK-WALKER, 1993  Mutations in MGI genes convert Kluyveromyces lactis into a petite-positive yeast. Genetics 133:517-525[Abstract].

CHEN, X. J. and G. D. CLARK-WALKER, 1994  sir2 mutants of Kluyveromyces lactis are hypersensitive to DNA-targeting drugs. Mol. Cell. Biol. 14:4501-4508[Abstract/Free Full Text].

CLARK-WALKER, G. D. and X. J. CHEN, 1996  A vital function for mitochondrial DNA in the petite-negative yeast Kluyveromyces lactis.. Mol. Gen. Genet. 252:746-750[Medline].

DEBUCHY, R., S. ARNAISE, and G. LECELLIER, 1993  The mat- allele of Podospora anserina contains three regulatory genes required for the development of fertilized female organs. Mol. Gen. Genet. 241:667-673[Medline].

DONZE, D., C. R. ADAMS, J. RINE, and R. T. KAMAKAKA, 1999  The boundaries of the silenced HMR domain in Saccharomyces cerevisiae.. Genes Dev. 13:698-708[Abstract/Free Full Text].

FOSS, M., F. J. MCNALLY, P. LAURENSON, and J. RINE, 1993  Origin recognition complex (ORC) in transcriptional silencing and DNA replication in S. cerevisiae.. Science 262:1838-1844[Abstract/Free Full Text].

FRITZE, C. E., K. VERSCHUEREN, R. STRICH, and R. EASTON ESPOSITO, 1997  Direct evidence for SIR2 modulation of chromatin structure in yeast rDNA. EMBO J. 16:6495-6509[Medline].

GOTTSCHLING, D. E., O. M. APARICIO, B. L. BILLINGTON, and V. A. ZAKIAN, 1990  Position effect at S. cerevisiae telomeres: reversible repression of Pol II transcription. Cell 63:751-762[Medline].

GRIGLIATTI, T., 1991  Position-effect variegation—an assay for nonhistone chromosomal proteins and chromatin assembly and modifying factors. Methods Cell Biol. 35:587-627[Medline].

HECHT, A., T. LAROCHE, S. STRAHL-BOLSINGER, S. M. GASSER, and M. GRUNSTEIN, 1995  Histone H3 and H4 N-termini interact with SIR3 and SIR4 proteins: a molecular model for the formation of