Copyright 0 1992 by the Genetics Society of America

Donor Locus Selection During Saccharomyces cerevisiae Mating Type Interconversion Responds to Distant Regulatory Signals Karen S. Weiler' and James R. Broach Department of Molecular Biology, Princeton University, Princeton, New Jersey 08544-1014 Manuscript received March11, 199 1 Accepted for publication August6, 1992

ABSTRACT Mating type interconversion in homothallic strains the of yeast Saccharomyces cerevisiae results from directed transposition of a mating type allele from one of the two silentdonor loci, H M L and HMR, to the expressing locus, MAT. Cell type regulates the selection of the particular donor locus to be utilized during mating type interconversion:MATa cells preferentially select HMLa and MATa cells preferentially select HMRa. Such preferential selection indicates that the cell is able to distinguish between H M L and H M R during mating type interconversion. Accordingly, we designed experiments to identify those features perceived by the cell to discriminateH M L and HMR. We demonstrate that discrimination does not derive from the different structures of the H M L and H M R loci, from the unique sequences flanking each donor locus nor from any of the DNA distal to the HM locion chromosome ZZZ. Moreover, we find thatthe sequences flanking theMAT locus do not functionin the preferential selectionof one donor locus over the other. We propose that the positionsof the donor loci on the left and right arms of chromosomeZZZ is the characteristic utilized by the cell to distinguish H M L and HMR. This positional informationis not generated by either CEN3 or the MAT locus, but probably derives from differences in the chromatin structure, chromosome folding or intranuclear localization of the two ends of chromosomeIZZ.

M

ANY cellular processes require productive interaction of two chromosomal regions located tens of kilobases apart. Forinstance, the P-globin locus control region (LCR) is located 10-20 kb upstream of the gene cluster it activates. Developmental regulation is imposed on this long rangeinteraction to effect sequential expression of the globin gene variants (EVANS,FELSENFELD and REITMAN1990).A different cellular process, immunoglobulin heavy chain class switching, requires that sequences (switch regions) separatedby 50- 100 kb be brought into close apposition. Isotypeswitching occurs by a programmed genome rearrangement during which the DNA separating one of six constant region gene segments and the variable region gene segment is deleted UACKet al. 1988). The particular antigen encountered by B cells frequently determines which isotype will be produced, indicating thatthegenomerearrangement event is regulated by external cues. Although compaction of the DNA between two distant chromosomal regions can suffice to bring them into proximity, the regulation of this event by internal or external signals implies a genetic program. Mating type interconversion in the yeast Saccharomycescerevisiae provides an experimentally tractable example of a genetically regulatedinteraction between distant chromosomal regions. T h e mating type

' Current address: Department of Zoology NJ-15. University of Washing1o11.

Seattle. Washington 98195.

(knrtics 132: 929-942 (December, 1992)

of a haploid yeast cell is determined by which of two distinct alleles is present at the MAT locus, resident on chromosome ZZZ those carrying the MATa allele are a cells and those carrying the MATa allele are a cells. MATa and MATa encode distinct transcription factors that regulate the battery of a-specific and aspecific genes responsible for the mating phenotype of the cell (reviewed in HERSKOWITZ 1989). Single copies of the a and a matingtypegenes are also present at two other loci, HMR and HML, located near the right and left ends of chromosome ZZZ, respectively (HARASHIMA,NOGIand OSHIMA1974; STRATHERN et al. 1980; NASMYTHand TATCHELL 1980). T h e mating type genes resident at HMR and HML are transcriptionally silent and donot contribute to the phenotype of the cell (KLAR et al. 198 1; NASMYTH et al. 198 1). Rather, these loci serve as templates for gene conversion of the MAT locus during a mating type switching event, described below (OSHIMAand TAKANO 1971; HICKS,STRATHERN and HERSKOWITZ 1977; reviewed in HERSKOWITZ and OSHIMA1981). In order for interconversion, or switching, to occur, one of the two telomeric regions harboring the donor template must interact with the centrally located MAT locus. This longdistanceinteraction is genetically regulated, since cell type governsselection of the locus that will productively interact with MAT (STRATHERN and HERSKOWITZ 1979). Mating type interconversion occurs in homothallic

930

K. and S. Weiler

strains of S. cerevisiae (those carrying the dominant H O gene) as often as every generation (HICKSand HERSKOWITZ 1976; STRATHERN and HERSKOWITZ 1979). Switching is initiated by a double strand DNA cleavage within the MAT locus, catalyzed by a site et specific endonuclease encoded by HO (STRATHERN al. 1982; KOSTRIKENet al. 1983; KOSTRIKENand HEFFRON 1984). Repair ofthe double strand break at MAT by geneconversion, using either HMLa or HMRa as the template, yields unidirectional transposition of the mating type information resident at the donor locus to MAT. If the mating type allele transposed from the donor cassette is different from the one previously at MAT, then the cell switches mating type.

By examiningpedigrees of standard homothallic and HERSKOWITZ strains of yeast, STRATHERN, HICKS deduced several rules thatdefinethepattern of 1976; STRATHERN switching (HICKSand HERSKOWITZ and HERSKOWITZ 1979). First, only those cells that have previously budded arecapable of switching. Second, cells switch in pairs; i.e., if a cell undergoes mating type interconversion, then both motherand daughter produced from that cell are switched. Third, a cell capable of switching will switch to theopposite mating type around 85% of the time. The first and second rules can be explainedon the basis of restricted expression of HO. HO endonuclease is present only during GI and only in those cells that have budded (NASMYTH 1983; reviewed in BREEDEN and NASMYTH 1985). T h e third rule indicates that the selection of a donor locus for transposition is not random.If it were, then productive switching, leading to achange in mating type, would occur no more than 50% of the time. That switching usually leads to achange in mating type indicates that thecell is able to distinguish HMLa from HMRa and that it has the capability of selecting one over the other. We have attempted to define those featuresof HML and HMR thatare used by the cell to distinguish between them during donorselection in the course of mating type interconversion.Previous work from other laboratories has shown that the alleles resident at the donorloci do not influence donor locus selection. MATa strains carryingthe a allele at both donor loci or carrying thea allele at HML and theCY allele at H M R still preferentially select HML as donor during interconversion (RINEet al. 1981; KLAR, HICKSand STRATHERN 1982). These observations demonstrated that donor preferenceis independent not only of the allele resident at the donorloci but also of the mating type of the strain subsequent to the switch. That is, the cell does not continue switching until a change of mating type occurs. Rather, thecell selects one donor, on the basis of some feature of HML or H M R other

J. R. Broach than the residentallele, and uses that donorin a single mating type conversion event. We have examined avariety of features other than the resident allele in an effort to define how the cell distinguishes between HML and HMR. T h e results of these analyses, presented in this report, indicate that the cell utilizes cues well removed from the donor loci themselves. Moreover, we found that the directional signals do not reside at or near the MAT locus. We conclude that theleft and right armsof chromosome ZII possess distinct characteristics, and that it is this difference, and not sequences near the ends of the chromosome, that enables the cell to distinguish between the donor loci. MATERIALS AND METHODS Media and culture conditions: Solid and liquid medium FINKand HICKS were prepared as described by SHERMAN, (1979). Yeast strainswereusuallygrownon YEPD. For transformations and prototrophic selection during matings, the cells were grown on synthetic minimal mediumor synthetic medium supplemented with uracil and all amino acids but those omitted for selection. a-Aminoadipate plates were usedwhen transforming yeastcellswith a plasmid that disrupted the LYS2 gene, to select against the Lys2+ population of untransformed cells (CHATTOOet al. 1979). For5fluoro-orotic acid (5FOA) selection, cells were grown on synthetic medium plates supplementedwith 0.1?6 5FOA. Genetic techniques: Yeast cells were transformed by the LiAc method of ITOet al. (1983) using the one step transplacement method of ROTHSTEIN(1983) or the two step replacement methodof SCHERER and DAVIS(1 979) as modified by BOEKEet al. (1987). The matingphenotypesof strains were assayed byprinting patchesof the strains tobe tested onto a lawnof either DC14(MATa h i s l ) or DC17 (MATa h i s l ) on plates that select for diploid growth. For spore to cell matings, individual cells of the haploid strain were micromanipulated adjacent to spores dissected from the sporulated diploid. Each potential mating pair was examined microscopically over several hours the for formation of a zygote, indicating successful diploid formation. Plasmidconstructions: The plasmidsutilized in this study are diagrammed in Figure 1, and their construction is described in the following. Plasmid pKSW2 was constructed in two steps. First, the 3.8-kb EcoRI to Hind111MATa2x183 fragment from the was cloned into plasmid ax1 83 (provided by K. TATCHELL) YIp5 to yield plasmid pKSWl . Plasmid ax1 83 consists of YRp7 carrying the 4.3-kb MATaHind111 fragmentinto which has been inserted a 9-bp XhoI linker insertion mutation that does not affect MATa function (TATCHELL et d . 1981).Plasmid pKSW2 wasthen madeby filling inthe XhoI site in the plasmid pKSW 1 to create a 13-bp insertion, which causes a frame-shift withinthe mata2 coding region.A PYUI site was created at the insertion site by this procedure,which wasused to follow the a2 allele during subsequent yeast strain manipulations. PlasmidpKSW14consistsof a URA3-marked deletion derivative of HMLa in pUC19. The plasmid contains the 6.6-kb BamHI fragment spanning HML, from which the internal 4-kb ClaI to EcoRV region has been removed. In addition, the 1. l-kb URA3 gene is inserted at a BglII site that marks a small deletion of the HML E site (from plasmid pDM41; MAHONEYand BROACH1989).

B““‘

FIGURE 1 .-Plasmids used for S. cerevisiae strain constructions. The details of the construction of these plasmids is described in MATERIALS The restriction sites labeled in parentheses indicate the endpoints of the relevant DNA fragments and were not preserved in the f i n a l plasmid construct. Not all of the restriction sites for an enzyme are indicated. Thin lines represent vector sequence. ANI) MFI’HODS.

Plasmid pKSW2O consists of pUC19, with the 6.6-kb BamHI fragment spanning HMLa inserted at the BamHI site and the2.2-kb LEU2 XhoI-Sal1 fragment inserted at the SmaI site. Plasmid pKSW23 is identical to pKSWPO except for the deletion of regions W and 22. T h e deletions were made by oligonucleotidemutagenesis using the Bio-Rad MUTA-GENE in vitro mutagenesis system. T h e HML fragment was cloned intoM 13, each deletion was made individually and the two deletions were combined upon recloning thefragmentintopUC19. T h e oligonucleotide used to delete theW domain ofHML was 5”CCTTCTTGAACAAGAGTTTGAATGACGATT-3’ and thatused to delete the 2 2 domain of HML was 5”AAGAGCAGTGAAAGATTTCTATATGAGTGTATAAACAAAC-3’. Plasmids pKSWGLEU, pKSW8LEU and pKSW9LEU are similar plasmids constructed from plasmid pBR322. Each hasthe 2.2-kb LEU2 SalI-XhoI sequenceinsertedas a HindIII fragment into the HindIIIsite of pBR322. Plasmid pKSW6LEU was constructed by inserting the 5.2-kbBamHI to HindIII fragment spanning HMLa into the PvuII site of the vector, a process that reformed the BamHI site. T h e HML fragment is oriented such that W is nearer the @lactamase gene. Plasmids pKSW8LEU and pKSW9LEU were constructed by inserting the 4.1-kb EcoRI to HindIII fragment spanning HMRa, derived from plasmid pJA82.6~268 (ABRAHAM et al. 1984), at the PvuII site of pBR322, a process that reformed theEcoRI site. T h e HMR fragment in pKSW8LEU is oriented such that the X region is nearer the@-lactamasegene; pKSW9LEU has HMR in the reverse orientation. Plasmid pKSW45 was constructed by first inserting the 3.2-kb matalx65 EcoRI to Sau3AI fragment from theplas-

mid ax65 (provided by K. TATCHELL)into the EcoRI to BamHI sites of the pUCl8polylinker. Plasmid ax65 consists of YRp7 carrying the 4.3-kb matalx65 HindIII fragment (TATCHELL et al. 1981). The matalx65mutation is a 213-bp duplication marked with an XhoI linker insertion in the a 1 coding region.T h e 1.1-kb URA3 HindIII fragmentwas then inserted into the HindIIIsite of the pUC18 polylinker. Plasmid pKSW32 is derived from plasmid pKSW27, which consists of the 3.8-kb EcoRI to HindIII fragment spanningmata2insertedintopUC18. T h e 0.8-kb NdeI fragment extending from within the Y a region to the Z1/ 22boundary was removedfrom plasmid pKSW27, and replaced with the 2.2-kb LEU2 SaZI-XhoI sequence. Plasmid pKSW34 contains the 3.8-kb EcoRI to HindIII fragment carrying a MATaincallele in pUC18. T h e MATai,, allele was constructed by replacing the 1.2-kb ScaI fragment carrying theY a region of MAT from plasmid pKSW27 with the ScaI HMLai,, fragment from theplasmid pAKlO (kindly provided by A. KLAR).Plasmid pAK 10 contains the HMLai., BumHI fragment in YEp13, created by retrieval of the inc mutation by in vivo gap repair using the HMLain, strain K673. Plasmid pKSW34 also contains a 4.5-kb lys2 inverted cassette, inserted at the NarI site of pUC18, for directing recombination of the plasmid into the LYS2 coding region (VOLKERT and BROACH1986). This cassette consists of the lys2 3’ region from the internal XhoI site to the BglII site followed directly by the lys2 5‘ region from the BgZII site to the internal XhoI site. T h e linear fragment produced by cleavage of theplasmid with BglII is essentially the Zys2 gene interrupted by pUC andMATain,. Plasmid pKSW37 contains a URA3-marked deletion derivative of MAT in pUC18. T h e plasmid consists of the 4.3-

K. S. Weiler and J. R. Broach

932

TABLE 1 Experimental yeast strains Genotype Strain

Additional markers

Chromosome 111

uru3-52 HO HMRa HMLa MATa" KSW4 MATa HMLa HMRa

KSW7 KSW 10 KSW 17 KSW17-211T

trpl-289 Leu2-3,112 his3Al HO ura3-52 leu2-3,112 his3Al trpl-289

"

MATa HMLaAWAZ2 HMRa MATO HMLnAWAZ2 HMRa

"

A5 KSW4

MATa (hmla-ter)A:HMLa (hmra-ter)A::HMRa-stk (hmla-ter)A::HMLa (hmra-ter)A::HMRa-stk MATa mfal-lacZ MATa-stk (hmla-ter)A::HMRa-stk (hmra-ter)A::HMLa (hmla-ter)A::HMRa-stk MATa (hmra-ter)A::HMLa mfal-lacZ MATa (hmla-ter)A::HMRa-stk (hmra-ter)A:HMLa (hmla-ter)A::HMRa-stk MATa (hmra-ter)A::HMLa mfal-lacZ

As KSW4; mfcul-lacZ

As KSW4; @!&? As KSW4; mfclil-lacZ

KSW 1 7 a l

MATa-stk (hmla-ter)A::HMRa-stk (hmra-ter)A::HMLa matcul (hmla-ter)A::HMRa-stk (hmra-ter)A::HMLa mfal-lacZ

As KSW4; mforl-lacZ

KSW19

matA::URA3 HMLu HMRa matA::URA3 HMLa HMRa

A~ K S W ~ lysPA::MATa,LEU2 ; lysZA::MATa,LEU2

KSW24

MATaHMLa HMRa cen3A::CEN3,URA3 " MATa HMLa HMRa cen3A::CEN3,URA3

As KSW4;

"

HMRa HMLa matA::LEUP KSW25

kb HindIII fragment spanning M A T , from which the internal 3.3-kb Sty1 fragment completely encompassing MAT has been removed,with a 1.1-kb URA3 BglII fragment replacing the deleted region. Plasmid pKSW55 was constructed by first inserting the 2.2-kb LEU2 SalI-XhoI sequence into theSalI site of pUC18. T h e 4.5-kb lys2 inverted cassette described for pKSW34 was also inserted into theSalI site. A 2.5-kb BglII fragment that precisely spans M A T a was then insertedinto theBamHI site, deleting a small fragment of lys2 3' to the XhoI site in the process. The BglII sites precisely bracketing M A T a were created by oligonucleotide mutagenesis of the cloned M A T a gene, using the Altered Sites mutagenesis kit from Promega. pJC303-6 (CLARKE andCARBON 1983;kindly provided by J. CARBON) carries a4.8-kb region spanning CEN3, in which an internal 627-bp fragment completely encompassing CEN3 is inverted. The URA3 gene is inserted as a 1.1kb HindIII fragment adjacent to CEN3 to serve as a selectable marker for transformation. Yeast strains and strain constructions: The strains used in this study were all constructed in this laboratory and are listed in Table 1. They have identical genetic backgrounds and derive from a common HO haploid precursor, strain KSW3-20C, whichwas the parent strain for our genetic analyses of mating type interconversion in this and other studies. T o construct strain KSW3-20C, strain S150-2B (MATa HMLa HMRa leu2-3,112 his3Al ura3-52 trpl-289 gal2 ho) was made hmla2 by a two-step gene transplacement using the URA3 mata2 plasmid, pKSW2. Plasmid pKSW2 was digested with H p a l , which cuts the DNAin the W region, and transformed into strain S150-2B. The transformants were screened by Southern analysis to identify the desired strain with an integration at HML. Isolates that had excised URA3 and either HMLa or hmla2 were selected by growth on 5FOA,and hmla2 strains were identified by Southern analysis. The HO gene was introduced into this strain by crossing it with strain SFY46 (ura3 H O ) , sporulating the diploid, and identifying a spore clone having the

ly~2::MATai., AS KSW4;

desired genotype. Strain KSW3-20C was haploid, since the a 2 mutation at HML, when transposed to MAT, resulted in a sterile phenotype. Consequently, strain KSW3-20C was composed of a mixture of haploid MATa and mata2 cells (TANAKA et al. 1984). Strains KSW4 and KSW7 were made by first replacing the hmla2 locus of strain KSW3-20C with URA3, by transformation with the BamHI hmlA::URA3 fragment fromplasmid pKSW14, to yield strain KSW3-20C-hmlA. The hmlA::URA3 allele was replaced by either HMLa or HMLaAWAZ2, using plasmids pKSW2O or pKSW23, respectively, in two steps. Strain KSW3-20C-hmlA was transformed to Leu+with pKSW2O or pKSW23 DNA, linearized at the unique StuI site in the HML flanking sequence to direct its integration next to URA3. The wild type a allele at HML, introduced into the genome by plasmids pKSW20 and pKSW23, as well as similar plasmids employed subsequently for strain constructions, served as a template for gene conversion of the mata2 locus, during a normal switching event. Consequently, the a 2 allele was permanently lost in these strains. With this exception, strain KSW3-20C is isogenic to strain KSW4. One transformant of each plasmid was retained and isolates of thetransformants inwhich URA3, LEU2 and vector sequences were excised by recombination, were identified by growth on 5FOA, and confirmed by Southern analysis. Strain KSW 10 was constructed through the sequential deletion and replacement of each end of chromosome 111 with the HM locus of interest. The success of each step of the construction was confirmed by Southern analysis. The left arm of chromosome 111 was truncated by transforming strain KSW3-2OC with the fragmentation plasmid pAR65 (ROSE 1990; derived from the plasmid YCF3 described in VOLLRATH et al. 1988), linearized with EcoRI. This telomeric plasmid contains a DNA fragment centromere-proximal to H M L , -0.9 kb away from the 22 region, for targeting the recombination event, and the URA3 gene for selection. The resulting strain, pAR65-KSW3-20C, was transformed

Donor Locus Recognition Signals

k

to Leu+ with Sa I-digestedpKSW6LEUDNA(HMLa LEU2), which int grated at the new left terminus next to the URA3 marker, directed by homology to pBR322 sequences. T h e transformant was able to diploidize, through HO-mediated mating type interconversion, using HMLa as a donor locus. In order to select for isolates in which URA3, LEU2 and vector sequence, but not HMLa, had excised by intramolecular recombination, the diploid was sporulated and sporeswereplated on 5FOA. One such isolate was retained as strain 6LEU6, which, due to HO-mediated mating type interconversion, persisted as a homozygous diploid. T h e right arm of one chromosome ZZZ homolog in strain 6LEU6 was then truncated by transforming the strain to Ura+ with NotI-digested pAR63 DNA (ROSE 1990; derived from theplasmid YCF3 described in VOLLRATH et al. 1988). This URA3 telomeric plasmid containsaDNA fragment centromere-proximal to HMR, -1.7 kb away from the X region,to specify its recombination site. T h e resulting strain, pAR63-6LEU6, was transformed to Leu+ with Sal1 digested pKSW9LEU DNA(HMRa LEUZ), which integrated through homology to pBR322 at the new chromosome ZIZ right arm terminus, next to URA?. One transformant was then sporulated and a clone homozygous for the plasmid integration at the right arm terminus was retained as strain 9/6-3D. As before, strain 9/6-3D persisted as a homozygous diploid due toHO-mediated mating type interconversion. An isolate having excised URA3, LEU2 and vector sequences by intramolecular reqmbination was selected by sporulating 9/6-3D and plating spores on 5FOA, and was retained as strain KSW 10. Strain KSW 10 carried HMLa oriented with the 22 region most centromere-proximal and HMRa oriented with the X region most centromere-proximal. Strain KSW 17 was constructed in a manner analogous to that used to construct strain KSW 10. In order to position HMRa at the end of the left arm of chromosome ZZZ, strain pAR65-3-20C was transformed with plasmid pKSW8LEU (HMRa LEUZ) to yield strain 8LEU8. Since the HMR locus containedthe a mating type allele, strain 8LEU8 was haploid. After truncating the right arm of chromosome ZZI of strain 8LEU8 with plasmid pAR63, the HML locus was introduced by transformation of plasmid pKSW6LEU (HMLa LEUZ). T h e final strain, KSWl7, carriedHMRa on the left arm with the Z1 region centromere-proximal, and HMLa on the right arm with the 22 region centromereproximal. The pheromone sensitivity exhibited by strain KSWl7 spores during pedigree analyses (typically, 2:2 afactor sensitive/a-factor resistant:a-factor resistant/a-factor sensitive) confirmed that HML and HMR were still subject to transcriptional silencing in their new positions. Strain KSW 17-21 1T was constructed fromstrain KSW 17 by replacing the MATa locus, bearing the stk mutation, with the wild type MATa allele. We determined that the HMRa locus carried on plasmid pKSW8LEU8 harbored the stk mutation, an A at position 11 of the Z1 region instead of a T (RAY,WHITEand HABER 1991b). This Z11A allele was transposed to MATa through HO-mediated gene conversion during the strain construction, and resulted in a reduced efficiency of HO cleavage of MATain strain KSW17, as revealed during subsequent genetic analyses of this strain. In order to remove the stk allele at MATa, and restore switching to wild type, we integrated the wild type MATa locus (Z1 l T ) , carried as a Hind111 fragment in vector YIPS, at the MATa locus of strain KSW17. We transformed strain KSW 17 to Ura+with plasmid YIp5-MATa, linearized at the unique BglII site in the Ya region. A transformant having the plasmid integratedatthe MAT locus, as shown by Southern analysis, was grown on 5FOA medium in order to

933

isolate recombinants in which the vector sequence and one copy of the MAT locus had excised by intramolecular recombination. We sequenced the remaining single copy of MATa present in these isolates by amplifying the locus by polymerase chain reaction, and retained one isolate bearing a T at position Z11 as strainKSW17-Z11T. Strain KSW17Z1 I T is isogenic with strain KSW 17 except for the single base substitution at position Z 1 1 of MATa. Strain KSW 1 7 a l was produced from strain KSW 17 by introducing the matalx6.5 mutation atMAT. Strain KSW 17 was transformed to Ura+with plasmid pKSW45 DNA linearized at the unique EugI site in the W region of MAT. A transformant in which the plasmid had integrated at the MATa locus was identified by Southern analysis. Isolates of this transformant that had excised URA3 and either MATa ormatal by intrachromosomalrecombination, were selected by growth on 5FOA. One isolate that retained the matal allele was identified by Southern analysis and saved as strain KSW 17a 1. Strain KSWl9 was a derivative of strain KSW3-20C by way of intermediate strains constructed for otherpurposes, and consequently had a circuitous history. Transformation of strain KSW3-20C with the matA::LEU2 Hind111 fragment from pKSW32 yielded strain KSW3-20C-matA::LEU2, in which part of the mata2locus including the a l - a 2 promoter and the HO endonuclease cleavage site was replaced with LEU2Strain KSW3-20C-matA::LEU2 was then transformed to a-amino adipate resistance with BglII-digested pKSW34 DNA togenerate strain KSW12, which had MATa,,, (and vector sequence)inserted intoLYSZ, rendering the strain Lys-. We then mated strain KSW 12 toan a spore from strain KSW4, by spore to cell mating. T h e KSW4/ KSW 12 diploid was sporulated and Leu+ Lys+ progeny (i.e., matA::LEUZLYS2) were screened by Southern analysis for those harboring the HMLa locus. One such strain, KSW4/ KSWl2 -1OC, was retained. T h e deletion of MAT in strain KSW4/KSW12 -lOC was then replaced with a larger deletion by transforming it to Ura+ with the Hind111 fragment encompassing matA::URA? from plasmid pKSW37. We obtained Ura+ Leu- transformants and these were shown to have the correct structure by Southern analysis. One such transformant was retained as strain KSW 14. Since this deletion removes the C terminal end of the essential TSMl gene (as determined by RAY,WHITEand HABER 1991a)we also transformed diploid cells with the same DNA fragment from plasmid pKSW37. Two diploid transformants were sporulated and tetrads from both produced four viable spores that segregated 2:2 for Ura+. We conclude that either the C terminus of the TSMl protein is not essential for its function or the gene is not essential in our strain background. T h e final step in the construction of strain KSW 19 was the insertion of MATa at LYS2. Plasmid pKSW55 contained, in place of part of the coding region of LYS2, the LEUZ gene and a fragment encompassing MATa, whose endpoints were internal to the deletion present in strain KSW14. Strain KSW 14 was transformedtoLeu+ using BglII-digested pKSW55 DNA and one Leu+ Lys- transformant was retained as strain KSW 19. Strain KSW24 was created by transforming strain KSW4 to Ura+ using the EcoRl fragment from plasmid pJC303-6, which carried the inverted CEN3 region, its flanking sequence and a URA? marker. One transformant was sporulated and a Ura+ clone was retained as strain KSW24. T h e inversion ofCEN? in strain KSW24 was confirmed by Southern analysis. Strain KSW24 was a diploid, homozygous for the CEN? inversion, as a resultof HO-mediated mating type interconversion. Strain KSW25 was derivedfromthe diploid, KSW4/

934

K. S. Weiler and J. R. Broach

KSWl2. It was identified as a Leu+ Lys- clone (matA::LEU2 ly~2::MATa,,~)that harbored the HMLa allele, as shown by Southern analysis. Pedigree analysis: Pedigree analysis was performed on a spores in order to determine the frequency ofMATa to MATa switches as follows. Diploid cells were sporulated and the asci were placed on a YEPD plate that had been preincubated overnight with a thick bisecting line of the a tester strain, DC17. Individual tetrads were micromanipulated to within a few hundred micrometers of the mass of (y cells and dissected. The spores were placed at 1-mm intervals along the a cell line with 2 mm separating the four spores of each tetrad, and the positions of each spore on the plate were noted. The cells were examined microscopically at frequent intervals after about 4 hr of incubation at 30". MATa spores germinated but arrested and changed morphology in response to a pheromone (a-factor)produced by the DC17 a cells; MATa spores germinated and divided. The a spore progeny were then continually monitored through the first two cell divisions, and the daughter cells were micromanipulated apart as early as possible. The first switching event occurred in the mother cell and was scoreable at the four cell stage. If the mothercell switchedmating type, then two a cells were produced along with the two a cells from the daughter line. Therefore, a successful switch was evident by the production of twoa-factor sensitive cells. If the mother cell did not switch mating type, then all four cells remained as a cells. Therefore, four a-factor resistant cells indicated that a switch had not occurred or that the MATa allele was replaced with a new a allele. Pedigree analysis was performed on a spores in order to determine the frequency of MATa to MATa switches by a similar procedure. Diploid cellswere sporulated and the asci were placed on aYEPD plate outside of a region onto which a cell pheromone (a-factor) had been applied. The a-factor was obtained from the liquid media of a culture of a cells and was prepared according to the method of MICHAELIS and HERSKOWITZ (1988). We applied 120 pl of a 90-fold concentrate to an area of ca. 4 cm2. Individual tetrads were micromanipulated into the a-factor region and dissected. The spacing between spores was as described above except they were placed in rows spaced 2 mm apart in the a-factor region. The position of each spore was noted. DC17 a cells were applied in a thick line bisecting the plate, as before. The plate was incubated at 30 O , and microscopic examination was initiated after about 4 hr. The MATa spores were identified by their sensitivity to the a-factor: they germinated but arrested and changed morphology. The MATa spores were resistant to the a-factor and divided. The progeny of the a spores were removed from the a-factor region as soonas the mother cellbegancelldivision and were placed adjacent to the DC 17 a cells by micromanipulation. T h e four progeny cells were then examined for sensitivity to a-factor, as described above. A change in cell type was again evident as two a-factor resistant cells (mother line) that divided and twosensitivecells (daughter line) that arrested. However, if the cells did not switch but remained MATa, all of the cells were a-factor sensitive. When determining the frequencies of switching, only those spores that derived from tetrads with four viable spores that segregated 2:2 MATa:MATa were counted. Pedigree analysisof matal spores was initiated in the same manner as that for a spores, since, like a spores, matal spores are resistant to a-factor. However, the spore progeny were not movedaway from the a-factor-treated area but rather were examined for the acquisition of a-factor sensitivity, which indicated a switch to MATa. We found that the morphological change exhibited by MATa cells responding

to a-factor was not as distinctive or persistent as that exhibited byMATa cells in response to a factor, and was often difficult to score. Therefore, KSW4 a spores were simultaneouslyassayedas a positive control for switching from MATa (a-factor resistant) to MATa (a-factor sensitive). Nevertheless, a small degree of uncertainty existed in the switching values obtained. A few tetrads that gave rise to two a-factor sensitive spores were presumed to have resulted from meiotic reversion of the a1 mutation, and were not analyzed. Mother cell viability: Tetrads from strain KSW 19 yielding four viable spores were examined for the viability ofthe mother cell produced by the firstcelldivision. The two progeny of the first mother cell were micromanipulated away from the daughter cell line and placed in a known position on theplate. After two daysand then again at about a week, the cell pairs were scored for colony formation. Those that did not form colonies consisted of a cluster of cells that did not appear to change after the second day. Cell rescue experiment:The rescue of cellsthat appeared to haveswitched from a to a was initiated in the same manner as a standard MATa to MATa pedigree analysis. The progeny from a spores that exhibited a-factor sensitivity were identified and scored as a. These cells were micromanipulated toanother sectionof the plate and placed adjacent to strain KSW25 cells, for cell-to-cell matings. The mating pairs were microscopically examined over the next few hours to determine if a mating had taken place. If the cells had mated, the zygote was isolated and moved again to its final position, which was noted. In this manner, seven cells were successfully rescued as diploids. For subsequent experiments, strain KSW25 was transformed with the vector YEp57 in order to make it His+. This enabled us to set up matings between each a cell to be rescued and several a cells of strain KSW25 and then later isolate the diploid cells that resulted from each successful mating by selection for His+ Uraf clones. We observed His+ Ura+ colonies only at the positions on the plate that corresponded to the arranged matings. We rescued 13 cells that appeared to have switched from MATa to MATaby this procedure. All 20 diploids were then tested for their mating phenotypes. RESULTS In the process of mating type interconversion in the yeast S. cereuisiae, one of the two donor loci residing near the left and right telomeres of chromosome ZZZ must be brought into spatial proximity of the mating type locus, MAT, as shown in Figure 2. A unidirectional gene conversionof the MAT allele to the allele residing at the donorlocus follows. While an interactionbetween MAT a n d t h e donor locus on either chromosome arm is possible, cell type regulates the relative probabilities that each onewill be utilized for interconversion. In an a cell, HML is nearly always selected and in an CY cell, HMR,is nearly always selected. Implicit in the deliberate selection of HML or H M R as thepreferred donor locus during a switching event is therequirementthatthe cell beableto distinguish these two loci. As a means of identifying the feature(s) that distinguish the two loci, we deleted or otherwise manipulated sequences at the donor loci or at other specific sites within chromosome ZZZ t o reveal any recognition

Signals Recognition Locus

Donor

c'--,-3A

935

being favored over the other is a logical extension of the known involvement of homology in recombination. HA TsJ - B 88% (41) T o test if donor preference dependsupon the extra homologous regions shared between HML and MAT, we precisely deleted these regions fromHML in vitro 84% (86) (as described in MATERIALS AND METHODS). The wildtype HMLa locus on chromosome ZZZ of the parent FIGURE2.-Mating type interconversion. (A) The relative posistrain, KSW4, was replaced with the double deletion tions andorientations of the mating type locus, MAT, and the donor derivative, HMLaAWAZ2, to createstrain KSW7, loci, HML and H M R , are indicated on a diagram of chromosome /I/.The boxes represent regions of sequence identity shared bewhose structure is depicted in Figure 3A. In order to tween the loci (see text). Thearrows indicate the direction of allele determine if this alteration of HML affected the cell's transfer during an interconversion event, which is dependent upon ability to distinguish it from HMR, we measured the the mating type allele expressed at MAT. (B) The frequencies of frequencies at which a and a cells were able to change tnating type interconversion are shown for strain KSW4, the parent mating type. This was done by pedigree analysis, as o f all of the strains used in this study. The value for a to a switching represents the percentage of switching competent MATa cells that described in detail in MATERIALS AND METHODS. Basiproduced two MATa cells at the next cell division. Similarly, the cally, we first identified a and a spores by their sensivalue for to a switching represents the percentage of switching tivity to the pheromone producedby the opposite cell competent MATa cells that produced two MATa cells at the next type. By analyzing the spore progeny for pheromone cell division, The number in parentheses indicates the number of sensitivity following two rounds of cell division, we cells assayed in each case. assessed whether a matingtype interconversion event, elements to which the cell might respond. The design indicative of proper donor locus selection, had ocof each experiment was influenced by the fact that curred. donor locus preference is not absolute but appears to Results of pedigree analysis from the parentHMLa be a competition between the two loci for a productive strain, KSW4, are presented in Figure 2B and those interaction with MAT. When the donor locus that is from the HMLaAWAZ2 strain, KSW7, are presented preferred as the source of mating type information of mating in Figure 3B. We found that the frequencies for MAT is deleted,the remaining donor locus is type interconversion for strainKSW7 were equivalent utilized efficiently (KLAR, HICKS and STRATHERN to thoseobtainedforstrain KSW4. These results 1982). Even in normal strains, the non-preferred dounequivocally demonstratethatWand 2 2 are not nor locus is selected in a small percentage of all HOHML to distinguish it from HMR during required at initiatedevents, yielding a phenotypically invisible donor locus selection. Thus, the difference in donor replacement of the MAT allele with the same genetic locus structure and accordingly, the extent of homolinformation. Thus, in the experimentswe performed, ogy between MAT and each donor locus, is not imporwe avoided introducing any alteration that might intant for thedistinction of HML and HMR. advertently place one donor locus at a competitive The chromosomeZZZ telomeres and telomere proxdisadvantage: all perturbationstothestructure of imal sequences do not provide signals for differenchromosome ZZZ were limited to the feature that we tiating HML and HMR: As a next step toward localizwished to examine. Finally, we used pedigree analysis ing the sequence(s) recognized by the cell during to test the effect of each deletion or rearrangement donor locus selection, we asked if chromosomal seon the selection of HML and on theselection of HMR, quences distal to the HM loci were required for disby determiningthefrequencies of both MATa to crimination. HML is located approximately 10 kb M A T a and M A T a to MATa switching. The results of from the left telomere of chromosome ZZZ and HMR these experiments aredescribed in the following. is located approximately 25 kb from the right telomThe difference in locus structure between HML (YOSHIKAWAand ISONO1990). No genes essential ere and HMR is not required to distinguish them during for vegetative growth reside in either distal region. donor locus selection:Physical analyses of MAT and T h e two chromosome ZZZ telomeres have the same the HM loci have shown that they all possess an allelebasic structure:embedded in CI-SA repeats are X specific region (a or a) bracketed by regions desiget al. 1980; NASMYTH regions but no Y' repeats. We deleted both ends of nated X and Z 1 (STRATHERN and TATCHELL chromosome ZZZ, leaving only 1.8 kb distal to W at 1980). However,HML and MAT have sequence identity in addition to regions X and Z1, HML, and 0.8 kb distal to Z 1 at HMR (see Figure 4A). referred to as regions W and 22,which is not shared Telomeric sequences from Tetrahymena substituted by HMR. We wished to test whether HML and H M R for the natural telomeres of chromosome HZ. The are distinguished through these additional regions of frequencies with which this strain, KSW10, was able homology present only at HML and MAT. That the to switch mating type were determined by pedigree degree ofhomology with MAT might lead to onelocus analysis. The results from these analyses, presented in ~

(Y

936

K. S. Weiler andJ. R. Broach FIGURE3.-Donor preferencedoes not depend ondifferences in locus struc-

A ,.

w. x

-11

-

w x

21.2 - / / " " - .

i

z12

the R

x

21

at

a t o a sw it-

B

87% ( 7 8 ) to a swltchlna

A

HMLa

NA T

CEN3

(88)

HHRa

HMRa

C//U//"//-//~//+

- 10

- 95

HHHHH -110

-100

89%

B

NAT

-25

a

to

(L

s w l t u

KB

88% ( 7 7 )

c3

e

I .8

0.8

84% (63) KB

Figure 4B, demonstrate that strainKSW 10 can switch bothfrom a to a andfrom a to a with the same efficiency as can the parental strain, KSW4. T h e fact that strain KSWlO switched cell type at wild type efficiency demonstrated that (1) truncating the two ends of chromosome ZZZ did not hinder the switching process in general and (2)the features thatdistinguish H M L and H M R did not reside either in the telomeric sequences or in theunique sequences distal to the donor loci on chromosome ZZZ. Unique sequences flanking HML and HMR do not function to discriminate the donor loci: Since neither the difference in locus structure between HML and H M R nor the sequences distal to the donor loci provide cues for distinguishing between them, we tested the possibility that the unique sequences flanking the loci contained the critical distinctive features. If this were the case, then the transposition of each donor locus and its flanking sequences to another location o n the chromosome would not affect the donor specificity of mating type interconversion. In other words, the loci would be identifiable despite their new positions. T o make the least drastic chromosome rearrangement possible, we simply interchangedthe donor loci, along with a substantial amount of flanking sequence from each. This reconstruction retained the donor loci on chromosome ZZZ and maintained their telomere proximal positions; only the chromosome arms upon which they reside was altered. The result-

chromosome 111 for strain KSW7. HMLAWAZP is a deletion allele of HMLa in which theextra regionsofhomology between H M L and M A T are precisely excised. (B) T h e frequencies of mating type ,. interconversion obtained for strain KSW7, as explained in the legend to Figure 2.

p-+ l/-/l/-

HNLaAWAZ2

ture two donor loci. (A) Upper line: diagram of chromosome I I I for strain KSW4. Lower line: diagram of -

FIGURE4.--Sequences distal to the donor loci are not required for donor preference. (A) Upper line: diagram of chromosome 111 for strain KSW4, under which are indicated the approximate distances in kilobases between the various landmarks (YOSHIKAWA and ISONO 1990). Lower line: diagram of chromosome 111 for strain KSW10, in which sequences 1.8 kb distal to HML and 0.8 kb distal to H M R have been truncated and replaced with telomeric sequencefrom Tetrahymena (arrows). Hatched regions indicate positions of pBR322 sequences. (B) T h e frequencies of mating typeinterconversion obtained for strain' KSW10, as explained in the legend to Figure 2.

ing strain, KSW 17, was similar in structure to strain KSW10, used in the previous experiment, differing only in the relative placements of HML and H M R (Figure 5A). Two possible outcomes of this experiment were anticipated. If the donor locus recognition information were relocated along with the donor locus itself, then we would have expected that matingtype switching would occur at wild-type efficiency in strain KSW17. On the other hand, if the donor locus recognition information were not included on the interchanged fragments but remained at the normal location, then we would have expected that the frequency of mating type interconversion would be quite low. In this case an a cell would most often select the donor locus on the left arm and an a cell would most often select the donor locus on the right arm,leading only infrequently to an observable mating type switch. The frequencies of interconversionthat we obtainedforstrain KSW 17 by pedigree analysis are presented in Figure 5B. Both a to a and a to a switches were observed only infrequently. Thus, either mating type interconversion was rarely initiated in this strain or the primary result of each switching event was homologous allele replacement. Since the chromosome ZZZ end truncations and concomitant manipulations did not adversely affect the efficiency of mating type switching in strain KSW 10, we favored the latter interpretation. This conclusion was strengthened by the results of theexperimentdescribed below, in

DonorSignals Locus Recognition

937

FIGURE 5.-Signals required for donor preference do not liein the sequences flanking H M L and H M R . (A) Upper line: diagram of chromosome Iff for strain KSW4. Lower line: diagram of chromosome 111 for strain KSW17, in which H M L and the indicated flanking sequences have been interchanged with H M R and the indicated flanking sequences. Note that the H M L locus at the end of the right arm was inverted relative to the normal orientation of H M R at this position. The H M R locus at the endof the left arm was in the same orientation as the H M L locus normally is in this location. Hatched regions indicate pBR322 sequences. As described in MATERIALS AND METHODS, the strain construction w a s such that chromosome 111 sequences were not duplicated or lost proximal to H M L and H M R . (B) The frequencies of mating type interconversion obtained for strain KSWl7 (Q to a) and strain KSW17-211T (a to a),as explained in the legend to Figure 2. ( C )

Donor locus selection during Saccharomyces cerevisiae mating type interconversion responds to distant regulatory signals.

Mating type interconversion in homothallic strains of the yeast Saccharomyces cerevisiae results from directed transposition of a mating type allele f...
2MB Sizes 0 Downloads 0 Views