Curr Genet (1988) 13:291-297

CU~I~

Ge~iCS

© Springer-Verlag 1988

Plasmid associations with residual nuclear structures in Saccharomyces cerevisiae Michael N. Conrad and Virginia A. Zakian Genetics Department, Fred Hutchinson Cancer ResearchCenter, 1124 Columbia Street, Seattle, WA 98104, USA

Summary. Acentric yeast plasmids are mitotically unstable, apparently because they cannot freely diffuse after replicating and therefore are not included in the daughter nucleus. This behavior could result if plasmids remain attached to structural elements of the nucleus after replicating. Since DNA replication is believed to take place on the nuclear matrix, we tested whether there was a correlation between the mitotic stability of a given plasmid and the extent to which it was found associated with residual nuclear structures. Residual nuclei were prepared from yeast nuclei by extraction with either high salt, 2 M NaC1, or low salt, 10 mM lithium diiodosalicylate (LIS). Hybridization analysis was used to estimate the fraction of plasmid molecules remaining after nuclei were extracted. We examined the extent of matrix association of three ARS1 plasmids, Trpl-RI circle (1.45 kb), YRp7 (5.7 kb) and pXBAT (45.1 kb) with mitotic loss rates ranging from 3-25%. In addition we examined the matrix binding of the endogenous 2~ma plasmid and the 2#m-derived YEpl3 which is relatively stable in the presence of 2/~m and less stable in cir ° strains. Among the ARS1 plasmids we observed a negative correlation between stability and matrix association, consistent with models in which binding to the nuclear matrix prevents passive segregation of ARS1 plasmid molecules. No such correlation was observed among the 2/am plasmids. Among all plasmids examined there is a positive correlation between size and matrix association. Key words: Yeast - Nuclear matrix - Plasmid stability

Offprint requests to." M. N. Conrad

Introduction Plasmids containing an A R S (autonomously replicating sequence) transform Saccharomyces cerevisiae at high frequency and are maintained as multicopy extrachromosomal elements (Hsiao and Carbon 1979; Struhl et al. 1979). They exist as chromatin (Livingston and Hahne 1979; Nelson and Fangman 1979; Seligy et al. 1980) and replicate under the same cell cycle controls as chromosomal DNA (Zakian et al. 1979; Zakian and Kupfer 1982; Zakian and Scott 1982; Fangman et al. 1983). However, in the absence of selection most A R S plasraids are rapidly lost from the population during mitotic growth. Pedigree analysis indicates that this instability results from the failure of plasmids to partition themselves to the bud following replication (Murray and Szostak 1983). The unequal partitioning of circular plasmids does not result from unequal distribution of nuclear mass since the mother and bud nucleus are of roughly equal size (Robinow and Marak 1966; Byers and Goetsch 1976; Gordon 1977). Therefore, it has been inferred that there are constraints on the free diffusion of plasmid molecules within the nucleus and, since the yeast nuclear membrane does not break down in mitosis, these constraints introduce segregation bias (Murray and Szostak 1983; Jayaram et al. 1985). Plasmids could be sequestered, for example, by attachment to structural components of the nucleus. The nuclear matrix, operationally defined as the largely proteinaceous structure remaining after isolated nuclei have been extracted with non-ionic detergent and highsalt buffers and digested with DNase, is one candidate for a plasmid anchoring structure. In a variety of organisms, including yeast, DNA synthesis is found to be closely associated with the matrix (Hunt and Vogelstein 1980; McCready et al. 1980; Pardoll et al. 1980; Potashkin et al. 1984). An alternative extraction method which uses the detergent lithium diiodosalicylate ( L I S ) i n

292

M.N. Conrad and V. A. Zakian: Plasmid associations with yeast nuclear matrix

place of high salt has been used to prepare structures termed "nuclear scaffolds". Nuclear scaffolds are reported to bind specific DNA sequences (Mirkovitch et al. 1984; Cockerill and Garrard 1986). This low-salt extraction has recently been adapted for use with yeast nuclei (R. Hajela and J. Huberman unpublished results). We use the term "nucleoid" to describe the structure remaining after nuclei are subjected to either high salt or LIS extraction but before DNase digestion. "Matrix" will be used to refer to the structure left after most of the DNA is removed from the nucleoid. We envision that all plasmids could require attachment to the matrix for efficient partitioning at mitosis. Alternatively, all plasmids might be matrix associated and detachment at the time of nuclear division might be required to insure equipartitioning. In order to test these models we prepared residual nuclear structures by both high and low salt extractions and examined the fraction of plasmid molecules which remained associated with the residual nuclei. We then looked for a correlation between the in vivo behavior of the plasmids and the extent of their matrix attachment. We have examined the interaction of the nuclear matrix with three ARS1 plasmids. Trp 1-RI circle is generated by self-ligation of the 1.45 kb EcoRI fragment containing the yeast TRP1-ARS1 elements (Zakian and Scott 1982). This small circle is present in 1 0 0 - 2 0 0 copies per cell (Zakian and Scott 1982) and, unlike otherARS plasmids, is lost at a rate of only 0.03 per cell generation and does not exhibit segregation bias (Murray and Szostak 1983). YRp7, a 5.7 kb plasmid, contains the TRP1-ARS1 fragment in pBR322, is rapidly lost when cells are grown without selection and exhibits motherdaughter bias during loss. p)tBAT is a 45 kb circular plasmid generated by cloning YRp7 into a derivative of phage ~; it is less stable than YRp7. In addition we studied the interaction of the matrix with the endogenous yeast plasmid 2/am circle. 2 / g n DNA encodes a partitioning system which promotes efficient segregation of 2 gm-derived plasmids (Broach and Hicks 1980). The system is composed of a cis-acting region of 2/lm DNA designated REP3 or STB (Kikuchi 1983; Jayaram et al. 1985) and two 2/~m encoded transacting functions REP1 and REP2 (Jayaram et al. 1983). The REP system dramatically raises the stability of the 2/1m-based plasmid YEpl3 which contains the 2 # m A R S and REP3 regions but neither REP1 nor REP2 (Broach et al. 1979). YEp13 is at least five times more stable in cells containing 2 grn DNA (cir +) than in those which lack 2/am DNA (cir °) (Murray and Szostak 1983). The 2 gin-encoded REP1 protein has been localized to the nuclear matrix (Wu et al. 1987). Thus, it is possible that 2 gm segregation could be mediated through interaction with structural elements of the nucleus.

In this study, we find that ARS1 plasmid DNA is associated with nucleoids in a manner consistent with that expected if binding to the nuclear matrix negatively influences plasmid stability. However, the results are also consistent with preferential trapping of larger DNA molecules during the preparation of the residual nuclear structures.

Materials and methods Strains and plasmids. Yeast strains A364a (MATa adel ade2 his7 lys2 tyrl urn1 gall) and 3482-16-1 cir+ (MATa met2 his3Zx-1 Ieu2-3,112 trp1-289 urn3-52) were provided by L. HartweU. A cir° derivative of 3482-16-1 was constructed by B. Veit. RDa5 (MATa urn3-52 lys2-801 ade2-101 trplA-901) was from P. Hieter. JSY322/T (MATa urn his3A-1 trpl/Trpl-RI) has been described (Zakian and Scott 1982). Yeast cultures were grown at 30 °C under selection appropriate for the plasmid (Zakian and Kupfer 1982). Plasmids YEpl3 (Broach et al. 1979), YRp7 (Struhl et al. 1979), Trpl-RI circle (Zakian and Scott 1982) (see Fig. 1) and phBAT (R. Wellinger, unpublished) were introduced into yeast by lithium acetate mediated transformation (Ito et al. 1983). pXBAT was constructed by cloning a derivative of YRp7 conraining only a single EcoRI site into hgt2 (Pansenko et al. 1977). Additional ptasmids used for hybridization probes were pBR322 (Bolivar et al. 1977); pBR-4134 (constructed by H. Blanton), formed by recloning a 1.4 kb EcoRI fragment called 4134 (originally from yeast chromosome IV and isolated from psc4134 (Stinchcomb et al. 1982) into pBR322; pJH104 (constructed by J. Hartley), containing the 1.45 kb TRPIARS1 fragment isolated from YRp7 in mp8 (Messing and Vieira 1982); mpl8-LEU2-0.9 (this work), produced by subcloning the 0.9 kb EcoRI-SalI fragment of the yeast LEU2 gene, which lacks the middle repetitive element delta, from YEpl3 into rap18 (Yanisch-Perron et al. 1985); mp8-586, containing a 586 bp Thai fragment bearing the 2 tzm ARS (originally cloned in pBR322 by W. Fangman); and Y-m21 (Roth et al. 1983) which contains the 1.3 kb HindllI fragment of 2 ~m circle. Preparation of nuclei and nucleoids. Two independent procedures were used to prepare nuclei and nucleoids. In Method I, nuclei were isolated using an unpublished procedure of P. Baum and J. Thorner with minor modifications. Cultures (200 ml) were harvested at a density of 5-10 x 107 cells/ml, washed with water and resuspended in 10 ml S buffer (1 M sorbitol, 20 mM potassium phosphate, pH 7.5, 10 mM MgC12). Mercaptoethanol was added to 100 mM and the cells were incubated at room temperature for 10 rain. The cells were recovered by centrifugation and resuspended in 10 ml fresh S buffer with 0.5 mg/ml Zymolyase 100T (Miles) and shaken gently at 30 °C for 20-45 min. Spheroplasts were washed twice with S buffer and lysed by addition of 25 ml 0.2% Triton X-100 in N buffer (40 mM MES pH 6.4, 10 mM MgC12, 1 mM PMSF, 1 mM DTT). Lysis and all subsequent manipulations were done with ice-cold solutions. The lysate was subjected to three strokes of a Potter-Elvehjem homogenizer with a Teflon pestle rotating at approximately 1,000 rpm. After centrifugation (Sorvall SS-34, 5,000 rpm, 5 rain), the crude nuclear pellet was resuspended in 25 m150% Percoil, 0.05% Triton X-100 in N buffer using a single stroke of the rotating homogenizer. The suspension was centrifuged in the Beckman 60Ti rotor at 23,000 rpm, 4 °C, 20 rain. Nuclei usually

M. N. Conrad and V. A. Zakian: Plasmid associations with yeast nuclear matrix

TRPI H ARSI H

REPI H ~ R E P 2

Hp A

REP3

formed a single band about 1/3 of the way from the bottom of the tube. The nuclei were withdrawn with a Pasteur pipet, diluted with 3 - 5 volumes of N buffer and centrifuged, 5,000 rpm, 5 rain. The yield of nuclei was typically 50% (estimated from recovery of radioactive DNA). One-third of the nuclei were saved and the remainder resuspended in 10 ml N buffer in a glass beaker. To prepare nucleoids an equal vohlme of 4 M NaC1

pBR322

A

293

C Fig. 1 A - D . Plasmids. Yeast strains were transformed with the following plasmids. Restriction site abbreviations: A, Ava I; E, EcoRI; H, HindlII; Hp, HpaI; P, PstI; S, SalI. Not all sites present in pBR322 are indicated in the diagrams. A 2 ~m circle. The endogenous yeast plasmid is 6.3 kb in length and is drawn to illustrate the 599 bp inverted repeat. Recombination within the inverted repeat is catalyzed by the FLP gene and produces isomeric A and B forms of the circle. The B form is drawn. The REP1 and REP2 genes are noted as well as the positions of REP3 and the ARS. B YEp13. The 2.2 kb EcoRI fragment of 2 t~m circle (B form) was inserted into pBR322 and a 4.1 k0 PstI fragment carrying the yeast LEU2 gene was cloned into the 2 #m PstI site (Broach et al. 1979). C YRp7. A 1.45 kb EcoRI fragment containing the yeast TRPIARS1 fragment was cloned into pBR322 (Struhl et al. 1979). D Trpl-RI circle. The yeast fragment was excised from YRp7 with EcoRI and self-ligated before transformation into yeast (Zakian and Scott 1982)

E

l ARs

HO D

pBR522

~ =

tkb

B Table I A - C . Retention of plasmid DNA by residual nuclei Plasmid

Size (KB)

Rate of loss per cell generation a

Host strain b

3H Plasmid probe

32p Reference probe

Average fraction of plasmid retained (individual values)

cir + cir° RDa5

pBR322 pBR322 mp8-586

RDa5

X

RDa5 RDa5

YRp7 pJH104

pJH104 pJH104 pBR-4134 mp18-LEU2 mpl8-LEU2 pBR-4134

0.95 0.81 0.47 1.30 0.55 0.13

(0.73, (0.77, (0.54, (1.02, (0.58, (0.10,

0.62, 0.81, 0.42, 1.59) 0.66, 0.15,

cir + cir ° RDa5 RDa5 RDa5 RDa5

pBR322 pBR322 mp8-586 k pJH104 pJH104

pJH104 pJH104 pBR-4134 mpl 8-LEU2 pBR-4134 pBR-4134

0.90 0.98 0.67 1.07 0.69 0.26

(1.18, (1.20, (0.89, (0.91, (0.63, (0.33,

0.63) 0.75) 0.45) 1.23) 0.75) 0.17)

cir + cir° RDa5 RDu5

pBR322 pBR322 mp8-856 X pJH104 pJH104

pJH104 pJH104 pBR-4134 mp18-LEU2 pBR-4134 pBR-4134

0.94 1.00 1.05 1.08 0.94 0.33

(0.95, (1.06, (1.32, (1.21, (0.95, (0.34,

0.95) 0.95) 0.78) 0.96) 0.94) 0.32)

A Method I nuclei, 2 M NaC1 extraction YEp13

10.7

2 ~m pkBAT YRp7 Trpl-RI

6.3 45.1 5.8 1.45

0.05 0.48 -10 - 4 0.25 0.20 0.03

1.08, 1.36) 0.59, 1.07) 0.44) 0.42) 0.14)

B Method II nuclei, 10 mM LIS extraction YEpl3 2 #m pM3AT YRp7 Trpl-RI C Method II nuclei, 2 M NaC1 extraction YEp13 2/~m pXBAT YRp7 Trpl-RI

RDa5 RDc~5

a The stability of YEpl3 was determined by the method of Murray and Szostak (1983). The value for 2 tzm loss was taken from Futcher and Cox (1983)and the Trpl-RI stability is that reported by Murray and Szostak (1983). The rates of loss of YRp7 and pLBAT were estimated from the fraction of ceils retaining the plasmid after growth under non-selective conditions as described by Dani and Zakian (1983) b Cir + and cir ° refer to strain 3482-16-1. Note that strain RDa5 contains a deletion of the chromosomal TRP1-ARS1 region

294

M.N. Conrad and V. A. Zakian: Plasmid associations with yeast nuclear matrix

in N buffer was added slowly with gentle swirling. The nucleoids were recovered in the pellet after centrifugation (SS34, 10,000 rpm, 10 rain). Approximately 1/3 of the DNA present in the nuclei was recovered with the nucleoid pellet. The loss of DNA during the salt extraction presumably results from lysis of nueleoids, since the DNA isolated from nucleoids appears undegraded (>23 kb) when examined on agarose gels (data not shown). Method It was developed by Hajela and Huberman and allows yeast nuclei to be extracted using the low salt procedure of Mirkovitch et al. (1984). Briefly, spheroplasts were prepared without the use of sulfhydryl reducing agents and then lysed and homogenized in cold 0.2% Triton X-100 in Nuclear Isolation Buffer (Nil]) containing 20 mM MOPS (Sigma) pH 7.4, 0.84 mM spermine, 2.16 mM spermidine, 10 mM EDTA, 20 mM KC1, 1% thiodiglycol, 1 mM aminoacetonitrile, I mM PMSF. Large debris was removed by low speed centrifugation (SS34, 1,500 rpm, 5 min), and the nuclei were recoverd from the supernatant by centrifugation at 5,000 rpm for 5 rain. Nuclei were washed with NIB containing 0.005% Triton X-100. Nuclei were stabilized by incubation in NIB with 2.5 mM sodium tetrathiohate on ice for 1 h, followed by incubation in NIB without EDTA at 37 °Cfor 25 rain. To prepare nucleoids, nuclei were suspended in 0.5 ml of NIB and added to i0 ml of cold 2 M NaC1 in NIB or to 10 ml room temperature 10 mM LIS in 5 mM HEPES pH 7.4, 0.025 mM spermidine, 2 mM EDTA, 2 mM KC1, 0.1% digitonin. Nucleoids were recovered by centrifugation at 3,000 rpm for 10 rain in a Beckman TJ-6 centrifuge. Recovery of DNA was comparable to that obtained with Method I. To extract DNA, nucleoids (or nuclei) were resuspended in 10 mM Tris-HCl pH 8, 10 mM EDTA, 0.05% SDS and the mixture incubated at 65 °C, I0 min. RNase A was added to 10 t~g/ ml and the mixture incubated at 37 °C, 30 min. Proteinase K (50 t~g/ml) was added and the mixture incubated 37°C, 2 h. Lysates were extracted twice with phenol-chloroform-isoamyl alcohol, twice with chloroform-isoamyl alcohol, and ethanol precipitated.

Copy number estimates. To prepare dot blots for copy number estimates, DNA samples were diluted to 0.5-1.0 ~g/100 ~l. After addition of 25 t~l 1.5 M HC1 the samples were incubated at room temperature for 3 rain. DNA was denatured by the addition of 25 tA 4 M NaOH and the samples were incubated at 65 °C for 30 rain. Samples were chilled and neutralized with HC1. A 3-fold dilution series with approximately 100 ng of DNA in the most concentrated spot was prepared in 20 x SSC (1 x SSC = 0.15 M NaC1, 0.015 M Na Citrate, pH 7) and applied to nitrocellulose using a BRL Hybridot apparatus. The filter was airdried and baked in vacuo at 80 °C for 1-2 h. Prehybridization and hybridization was carried out as described by Wahl et al. (1979) with the addition of 50 #g/ml herring sperm DNA (Sigma #D2251). Hybridizatio~ was conducted with a mixture of two nick-translated probes (2-5 ~g each): a 3H plasmid-specific probe and a 32p probe specific for a single copy chromosomal reference sequence (see Table 1). Individual dots were cut out, and the 3H/32p ratio was determined by liquid scintillation counting. Background was determined from dots containing no DNA from various places on the filter. As a control, DNA from strains lacking the plasmids was included on each filter. The fraction of plasmid in the nucleoid pellet relative to that in the nucleus was determined from the ratio:

3H/32P nucleoid 3H/32p nucleus.

Results

Newly replicated DNA is closely associated with the nuclear matrix Potashkin et al. (1984) have shown that newly synthesized DNA is more closely associated with the yeast nuclear matrix than is bulk DNA, consistent with the notion that DNA replication occurs on the nuclear matrix. We conducted similar pulse-labeling experiments using Method I nuclei with virtually identical results, i.e., when approximately 90% of total DNA is digested from the matrix with DNase I the matrix-bound DNA is 2.5 to 3-fold enriched for newly synthesized DNA (not shown). This extent o f enrichment is comparable to that obtained with similar experiments in higher eukaryotic cells (PardoU et al. 1980) and in yeast (Potashkin et al. 1984). Yeast plasmids and chromosomal DNA appear to share the same replication functions (Zakian et al. 1979; Zakian and Kupfer 1982; Zakian and Scott 1982; Fangman et al. 1983). Therefore, if our preparations accurately reflect the in vivo nuclear organisation, plasmids are also likely to replicate in association with the nuclear matrix.

Plasmid interaction with residual nuclear structures The extent of plasmid association with residual nuclei was assayed by determining the fraction of plasmid molecules present in nuclei which was retained by nucleoids. Assays were performed on five plasmids (Fig. 1). The endogenous 2 p m circle, 6.3 kb, is an extremely stable plasmid being lost at a rate of roughly 10-4/cell generation (Futcher and Cox 1983). YEp13, containing the 2/am A R S and REP3 was introduced into a cir + strain, where it is relatively stable and into an isogenic cir ° strain where it is unstable. YRp7 consists of a 1.45 kb yeast fragment containing TRPIARS1 in pBR322, producing a plasmid which is about the same size as 2 p m but 5 0 0 - 1 0 0 0 times less stable (Struhl et al. 1979; this work, Table 1). Self ligation of the TRPIARS1 fragm e n t produces the T r p l - R I circle which is 5 - 1 0 times more stable than YRp7 (Zakian and Scott 1982; Murray and Szostak 1983). Finally, YRp7 was cloned into Xgt2 to form a 45 kb plasmid called pXBAT (not shown). The ARS1 plasmids were transformed into a strain having a chromosomal deletion o f the TRPIARS1 region. In an initial experiment (Fig. 2), DNA was prepared from nuclei and nucleoids as well as from a spheroplast lysate o f a strain containing b o t h 2 p m circle and Trp 1-RI circle. DNAs were digested with I-IindlII, and a Southern blot was hybridized to a probe mixture containing a 615 bp HindlII-EcoRI ARS1 fragment and a 1.3 kb

M. N. Conrad and V. A. Zakian: Plasmid associations with yeast nuclear matrix

Fig. 2. Plasmicl content of nuclei and nucleoids. Yeast strain JSY322/T carrying both 2 um circle and Trpl-RI circle was grown in YC-trp medium (Zakian and Scott 1982). DNA was prepared from nuclei and nucleoids using Method I and from spheroplast iysates (Davis et al. 1980). Following digestion with HindlII, DNAs were subjected to electrophoresis in a 0.7% agarose gel and blotted to nitrocellulose. The blot was hybridized to a mixture of 32p nick-translated probes including the 615 bp EcoRI-HindlII ARS1 fragment of YRp7 and a 1.3 kb HindlII fragment from within the large unique region of 2 ~m circle. Hybridization is to a 2.1 kb fragment of chromosomal ARS1, a 1.4 kb fragment representing linearized Trpl-RI circle and a 1.3 kb fragment of 2/~m circle. Note that because a mixture of two probes was used the absolute copy number of 2 ~m circle cannot be estimated from this blot

HindlII fragment of 2 pm circle which hybridizes to bands at 2.1 kb (chromosomal ARS1), 1.4 kb (plasmid A R S 1 ) and 1.3 kb (2 pm circle). By densitometry, we estimate that approximately 40% of the plasmid DNA present in DNA extracted from total ceils is present in the nuclei. Although a certain fraction of the plasmids are lost during the preparation of the nuclei, both plasmids are lost to an equal extent. In contrast, when nuclei were extracted with 2 M NaC1 to prepare nucleoids, Trpl-RI circle was lost to roughly a four-fold greater extent than was 2 gm circle.

295

Although Southern blot analysis indicated that different plasmids were differentially recovered with the nucleoid, difficulty in reliably obtaining complete restriction digests of DNA from nuclei and nucleoids precluded extensive use of this method. Therefore, we devised the following procedure to assay relative plasmid content in DNA purified from nuclei and nucleoids. DNA samples were denatured, and a dilution series was applied to nitrocellulose filters in the form o f dot blots. Each filter was then hybridized to a mixture of two nicktranslated probes, one a 32P-probe for a sequence found in single copy in chromosomal DNA, and the other a 3H probe specific for plasmid sequences. Individual dots were cut out of the filter and radioactivity determined by liquid scintillation counting. The 3H/32P ratio provides a measure of plasmid copy number. The 3H/32P ratio in the nucleoids relative to the ratio in nuclei gives the fraction of plasmid DNA retained after salt extraction. The ratio was constant for all dots in a dilution series, confirming that there was an excess of probe in the hybridization mixture. The results for Method I nucleoids (Percoll nuclei) are presented in Table 1A. Among the ARS1 plasmids the most highly retained is pXBAT. YRp7 is recovered with 50% efficiency and Trpl-RI circle is recovered at an efficiency of 10-15%. YEp13 is retained to a slightly higher degree by the cir + nucleoid (95%) than the cir ° nucleoid (81%). However, recombination between 2 / a n and YEp13 generates a larger plasmid Broach et al. (1979) which comprises 10-20% of the plasmid population in our strain (data not shown). Since larger plasmids appear to be retained by the nucleoids more efficiently than smaller ones (see below), the presence of the hybrid plasmid can account for the somewhat higher retention of Y E p l 3 in cir + nucleoids. YRp7 and 2/am DNA, despite a large difference in plasmid stabilities, are both recovered with the nucleoid at about 50% efficiency. Separate experiments determined 2 #m content relative to YRp7 or pXBAT by using a 3H 2 grn-specific probe and a 32p YRpT- or pXBAT-specific probe. 2 / a n was found to be nucleoid associated 1.04 times as well as YRp7 (individual values: 1.19, 1.22, 0.70) and pXBAT was nucleoid associated 1.4-2.5 times as well as 2/.an, entirely consistent with the measurements made relative to chromosomal DNA. Since it is possible that high-salt extraction may disrupt specific protein-DNA interactions, we repeated the copy number determinations using a different nuclei isolation procedure (Method II) which permits extraction with LIS. Those results are presented in Table lB. In general, the LIS extracted residual nuclei retained a higher percentage of the plasmids than the nuclei prepared by Method I and extracted with high salt. For example, 26% of the Trpl-RI circle is now retained compared with 13% in Percoll nuclei extracted with high salt

296

M.N. Conrad and V. A. Zakian; Plasmid associations with yeast nuclear matrix

(Table 1A). However, the same rank order of plasmid retention by the nucleoids is observed in the LIS extracted preparations. When the Method II nuclei were extracted with 2 M NaC1, a slightly different result was obtained (see Table 1C). While Trpl-RI circle was still retained least efficiently, pXBAT, YEpl3, YRp7 and 2 gm were retained at roughly 100% efficiency. We believe that the Method II high salt nucleoids simply retain all plasmids more efficiently such that all plasmids above a certain size are 100% bound. This might result if the Method II high salt nucleoids are contaminated with DNA binding components which are removed by purification on Percoll gradients. Despite some quantitative differences in the results obtained with different extractions, we find that the less stable ARS1 plasmids are the more tightly bound plasmids. However this correlation does not hold true for 2 pm and YEp13. When the data from the ARS1 and 2 pm plasmids are taken together there is a strong correlation between size and matrix association.

Discussion The genetic behavior of yeast plasmids suggests that they do not freely diffuse within the nucleus and that efficient partitioning requires a mechanism either for actively segregating plasmid molecules or for permitting their passive diffusion. As a first step towards identifying the physical interactions which influence plasmid segregation, we tested for a correlation between mitotic stability and the fraction of plasmid found associated with operationally defined residual nuclear structures. We reasoned that attachment to the nuclear matrix or scaffold might be a requirement for stability or, conversely, that plasmid stability requires detachment from points of anchorage at the time of nuclear division. Although there are no objective standards for assaying the integrity of extracted nuclei, by several criteria the yeast nuclear matrix is comparable to that of higher organisms. By light and electron microscopy yeast residual nuclei exhibit the general appearance of extracted nuclei from other cells (Potashkin et al. 1984; Wu et al. 1987; our unpublished results). In addition, as is the case with other organisms, newly replicated DNA is found in close association with the yeast nuclear matrix (Potashkin et al. 1984; our results). If one takes the data from the ARS1 plasmids alone, there is a negative correlation between stability and matrix binding. However, one must also remember that, among the acentric ARS1 plasmids considered here, the least stable plasmids are also the largest. Thus for ARS1

plasmids the extent of association with the matrix is also positively correlated with plasmid size. In the case of 2/xm-based plasmids, we found no correlation between plasmid stability and attachment to residual nuclei. This fact is best illustrated by the behavior of YEpl3 which is nucleoid-associated to approximately the same degree in isogenic cir + and cir ° ceils despite a five-fold difference in stability between the two strains. This is perhaps surprising since the 2 grnencoded protein REP1 which enhances the stability of YEp13 has been localized to the nuclear matrix (Wu et al. 1987). Our results imply that if REP1 mediates the attachment of 2/am DNA to the nuclear matrix, those attachments must be transient. Potashkin and Huberman (1986) have noted that 2 tam sequences are less matrix associated at times in the cell cycle when 2 #m should have completed replication. We did note a strong correlation between plasmid size and recovery with nucleoids when the results from the ARS1 and 2 gm plasmids are combined. For example, YRp7 and 2 grn are of roughly equal size, and even though they differ 100-fold in stability they are recovered with the nucleoid with about the same efficiency. pXBAT is consistently associated with the nucleoids to the greatest extent, while Trpl-RI circle is the least associated. The correlation between size and matrixassociation was seen with both high and low salt extraction procedures. Current models of nuclear structure propose that specific DNA sequences mediate attachment of chromosomes to the matrix or scaffold (Mirkovitch et al. 1984; Cockerill and Garrard 1986). In addition, interaction with the nuclear matrix or a similar structure has been proposed as a way to explain the nonrandom segregation of certain yeast plasmids (Murray and Szostak 1983; Jayaram et al. 1985; Wu et al. 1987). Our data are consistent with a model in which DNA molecules interact with structural components of the nucleus through random or frequently occurring sequences such that larger plasmids are found more tightly and thus are less likely to segregate passively. An active segregation system, either a centromere or the 2 pm REP system would be required to overcome the attachments. Nevertheless, the size-dependent attachment is also expected from artifactual trapping of large DNA molecules during the extractions. Thus at this point we cannot exclude the possibility of artifacts and must urge caution in interpreting the biological significance of these structures in yeast and possibly in other biological systems. Acknowledgements. We thank Liebe Wetzelfor technical assistance, Raymund WeUinger for p~.BAT, and Peter Baum and Jeremy Thorner for nuclei isolation Method I. We are especially grateful to Joel Huberman and Ravindra Hajela for providing Method II

M. N. Conrad and V. A. Zakian: Plasmid associations with yeast nuclear matrix and for communicating unpublished results. We also thank Jim Broach for communicating results prior to publication. Dan Gottschling, Ann Pluta and Rosemary Sweeney provided helpful comments on the mansucript. This work was supported by a grant from the National Institutes of Health to V.A.Z.M.N.C. was the recipient of an ACS postdoctoral fellowship and an NIH lxaineeship in Molecular Biology.

References Bolivar F, Rodriguez RL, Greene PJ, Betlach MC, Heynacker HL, Boyer HW, Crossa JH, Falkow S (1977) Gene 2:95-113 Broach JR, Hicks JB (1980) Cell 21:501-508 Broach JR, Strathern J, Hicks J (1979) Gene 8:121-133 Byers B, Goetsch L (1976) J Cell Bio169:717-721 Clarke L, Carbon J (1980) Nature 287:504-509 Cockerill PN, Garrard WT (1986) Cell 44:273-282 Dani GM, Zakian VA (1983) Proc Natl Acad Sci USA 80:34063410 Davis RW, Thomas M, Cameron J, St John T, Scherer S, Padgett RA (1980) Methods Enzymo165:404-411 Fangman WL, Hice RH, Chlebowicz-Sledziewska E (1983) Cell 32:831-838 Futcher AB, Cox BS (1983) J Bacteriol 154:612-622 Gordon CN (1977) J Cell Sci 24:81-93 Hsiao C-L, Carbon J (1979) Proc Natl Acad Sci USA 76:38293833 Hunt BF, Vogelstein B (1980) Nucleic Acids Res 9:349-363 Ito H, Fukuda Y, Murata K, Kimura A (1983) J Bacteriol 153: 163-168 Jayaram M, Li Y-Y, Broach JR (1983) Cell 34:95-104 Jayaram M, Sutton A, Broach JR (1985) Mol Cell Biol 5:24662475 Kikuchi Y (1983) Cell 35:487-493

297

Livingston DM, Hahne S (1979) Proc Nat1 Acad Sci USA 76: 3727-3731 McCready SJ, Godwin J, Mason DW, Brazell IA, Cook PR (1980) J Cell Sci 46:365-386 Messing J, Vieira J (1982) Gene 19:269-276 Mirkovitch J, Mirault ME, Laemmli UK (1984) Cell 39:223-232 Murray AW, Szostak JW (1983) Cell 34:961-970 Nelson RG, Fangman WL (1979) Proc Natl Acad Sci USA 76: 6515-6519 Pansenko SM, Cameron JR, Davis RW, Lehman IR (1977) Science 196:188-189 Pardoll DM, Vogelstein B, Coffey DS (1980) Cell 19:527-536 Potashkin JA, Huberman JA (1986) Exp Cell Res 165:29-40 Potashkin JA, Zeigel RF, Huberman JA (1984) Exp Cell Res 153: 374-388 Robinow CF, Marak J (1966) J Cell Biol 29:129-151 Roth GE, Blanton HM, Hager L, Zakian VA (1983) Mol Cell Biol 3:1898-1908 Seligy VL, Thomas DL, Miki BLA (1980) Nucleic Acids Res 8: 3371-3391 Stinchcomb DT, Mann C, Davis RW (1982) J Mol Biol 158: 157-179 Struhl K, Stinchcomb DT, Scherer S, Davis RW (1979) Proc Natl Acad Sci USA 76:1035-1039 Wahl GM, Stern M, Stark GR (1979) Proc Natl Acad Sci USA 76:3683-3687 Wu L-CC, Fisher PA, Broach JR (1987) J Biol Chem 262:883891 Yanisch-Perron C, Vieira J, Messing J (1985) Gene 33:103-109 Zakian VA, Brewer BJ, Fangman WL (1979) Cell 17:923-934 Zakian VA, Kupfer DM (1982) Plasmid 8:15-28 Zakian VA, Scott JR (1982) Mol Cell Biol 2:221-232 C o m m u n i c a t e d by R. J. R o t h s t e i n Received June 17, 1987 / Revised December 21, 1987