Curr Genet (1995) 28:280-288

9 Springer-Verlag 1995

Norio Gunge 9Kohsai Fukuda 9Shigemasa Takahashi Friedhelm Meinhardt

Migration of the yeast linear DNA plasmid from the cytoplasm into the nucleus in Saccharomycescerevisiae

Received: 20 January / 16 February 1995

The Kluyveromyces linear plasmids, pGKL1 and pGKL2, carrying terminal protein (TP), are located in the cytoplasm and have a unique gene expression system with the plasmid-specific promoter element termed UCS, which functions only in the cytoplasm. In this study we have developed an in vivo assay system in Saccharomyces cerevisiae which enables the detection of a rare migration of the yeast cytoplasmic plasmid to the nucleus, using a pGKLl-derived cytoplasmic linear plasmid pCLU1. pCLU1 had both the UCS-fused LEU2 gene (a cytoplasmic marker) and the native URA3 gene (a nuclear marker) and therefore its cytoplasmic-nucleo localized could be determined by the phenotypic analysis of the marker. The nuclearly migrated plasmids were often detected as linear plasmids having the telomere sequence of the host yeast at both ends, although circular plasmids were also found. The circular form was produced by the terminal fusion of pCLU 1. Insertion of a Ty element into a nuclearly migrated plasmid was observed, allowing the ROAM-regulated expression of the adjacent nuclearly silent UCS-fused LEU2 gene. The nuclearly located plasmids, whether linear or circular, were less sensitive to UV-mediated curing than pGKL and pCLU1.

copies. They have inverted terminal repeats (ITR) whose 5' ends are associated with terminal protein (TP), and presumably replicate using TP as a primer of DNA synthesis by analogy to the adenovirus or bacteriophage q)29genome (Salas 1991). pGKL1 contains four ORFs and confers the killer and immunity phenotypes on the host cell, while pGKL2 has ten ORFs and plays a role essential for plasmid replication. The pGKL plasmids have been introduced into Saccharomyces cerevisiae to study them in this genetically and biochemically well-characterized host (for reviews see Gunge 1986, 1995; Stark et at. 1990). The cytoplasmic pGKL plasmids have a unique geneexpression system, different from the nuclear promoter system with a TATA-motif. In other words, all four pGKL 1 ORF genes possess the plasmid-specific promoter with a motif ACT(A/T)AATATGA, termed UCS (upstream consensus sequence) (Romanos and Boyd 1988; Stark et al. 1990). Promoter sequences with the UCS-like motif are also found upstream of the pGKL2-genes (Tommasino et al. 1988). The LEU2 gene ofS. cerevisiae was expressed in the cytoplasm of S. cerevisiae and K. lactis by replacing its TATA promoter by the UCS-promoter from the pGKLl-toxin gene (Kfimper et al. 1991; Schaffrath et al. 1992). On the other hand, during an attempt to express the native LEU2 gene, a pGKLl-based linear plasmid pLS 1 Key words Plasmid migration 9 U C S - p r o m o t e r 9 was constructed having a TATA promoter on pGKL1 in S. Telomere - Ty cerevisiae (leu2) (K~imper et al. 1989). In contrast to pGKL1, pLS 1 replicated in the nucleus and had no TP but possessed part of the repeat sequence (Gt_3T)n of the S. Introduction cerevisiae telomere (Zakian t989) at both the ITR ends, From these observations it was strongly suggested that the The Kluyveromyces linear killer plasmids, pGKL1 and cytoplasmic pGKL plasmids could migrate into the nucleus pGKL2, are located in the cytoplasm of yeast in multiple under certain conditions. In the present paper we report on an assay system in S. N. Gunge ([]) 9K. Fukuda 9S. Takahashi cerevisiae which allows for the detection of a rare migraApplied Microbial Technology,KumamotoInstitute of Technology, tion of the cytoplasmic linear plasmid into the nucleus, usIkeda 4-22-1, Kumamoto860, Japan ing a TP-associated reporter plasmid pCLU1 with both cyF. Meinhardt toplasmic-specific (UCS-fused LEU2) and nuclear-speInstitut ft~r Mikrobiologie, Westf~ilische Wilhelms-Universitfit cific (URA3) markers. The study clearly demonstrates that Mt~nster, Corrensstrasse 3, D-48149 Mfinster,Germany the cytoplasmic plasmid can enter the nucleus at a low frequency, and that the plasmid termini are susceptible to the Communicatedby K. Esser

Abstract

281

Sa

addition of a telomere sequence from the host chromosome, although circularization of the p l a s m i d takes place in some cases. Insertion of a Ty e l e m e n t (a yeast retrotransposon) into a n u c l e a r l y located plasmid, and the resulting R O A M effect on the expression of the adjacent U C S - f u s e d L E U 2 gene, will also be discussed.

U

C

Materials and methods

S

p

B H

~

!

P

Strains. The genotypes of the S. cerevisiae and E. coli strains employed in this study are listed in Table 1. E. coli HB 101 was used for amplification of plasmids pARU 1, pSAK068 and pCYR 1-T 1. E. coIi Sure was used to clone the 1.1-kb NsiI fragment of pTLU1-E1 into pUC19.

s

S. cerevisiaeKI2-2A-CK leu2 ura3, pGKLI pGKL2 Cytoplasm

UCS U-~

Media. Yeast cells were grown in YEPD (1% yeast extract, 2% Bac-

to-peptone, 2% glucose) and SD (0.67% Difco nitrogen base w/o amino acids, 2% glucose). When necessary, SD was supplemented with 30 mg of leucine and 20 mg of uracil/1. E. coli with plasmid was grown in LB (10 g Bactotryptone, 10 g NaC1 and 5 g yeast extract/l) containing 50 mg ampicillin/1. Plasmids and construction of pCLU1, pARU1 was constructed by insertion of the 1.1-kb HindIII fragment containing the URA3 gene of S. cerevisiae at the HindIII site of ORF2 cloned in pAR1 (Schaffrath et al. 1992). The 4.3-kb PstI-SacI fragment carrying the UCSfused LEU2 (a cytoplasmic marker, K~imper et al. 1991) and URA3

(a nuclear marker) genes, flanked by parts of pGKL1-ORF1 and -ORF3, was excised from pARU1 and introduced by transformation into a pGKL-killer strain, K12-2A-CK, of S. cerevisiae, so that pCLU1 could be constructed in vivo by homologous recombination with the endogenous pGKL1 (Fig. 1). The strain carrying pCLU1 was screened among Leu + transformants grown on SD+ura. Yeast transformation and other genetic methods. Standard proce-

dures were used as described by Rose et al. (1990). Preparation and detection ofplasmids. Plasmid DNA from yeast was

prepared and analyzed as described by Gunge and Yamane (1984). Plasmids from E. coIi were extracted and purified according to Sambrook et al. (1989).

\

pGKLI 8.9kb [

ORFI

pCLUI 8.4kb

i

N

2

,

0

in vivo recombination 3 J4>

UCS r---'f

ORFI

LEU2~

2 URA3

3

4

pGKL2

Fig. 1 Construction of pCLU1, pCLU1 was constructed by inserting the UCS-fused LEU2 and intact URA3 genes into pGKL1 usig an in vivo recombination technique (see Materials and methods). ORF 1, ORF2 and ORF3 in pARU 1 are derived from pGKL 1. LEU2 * stands for the LEU2 coding region with part of the terminator sequence (Kgmper et al. 1991). Arrows on pARU1 indicate the transcriptional direction of the genes. The thin line refers to the bacterial part of plasmid. The closed circle at plasmid ends indicates a terminal protein (TP). Abbreviations: P PstI; C ClaI; Sa SalI; B BamHI; HHindIII; S SacI. The figures are not to scale

Table 1 Strains used Southern hybridization and probes. DNA hybridization was carried

S. cerevisiae

K12-2A K12-2A-CK K12-PC K12-U1 K12-U2 K12-U1-S1 K12-U1-S2 K12-U2-S1 a MKO31-6B

MATer MATt~ MATo: MATer MATa MATc~ MATa MATc~ MATa

teu2 leu2 leu2 leu2 leu2 leu2 leu2 Ieu2 leu2

ura3 ura3 ura3 ura3 ura3 ura3 ura3 ura3 ura3

rho + rho + pGKL1 pGKL2 rho + pCLU1 pGKL2 rho + pTLU1 rho + pCLUI pGKL2 pRLU1 rho + pTLU1 pTLUI-E1 rho + pTLU1-E1 rho + pRLU1 ade2 metl rho +

E. colli

HB101

F - rB- m B- RecA ara proA lacy galK str x y l 5 mtl SupE

Sure

uurC umuC sbcC recB recJ AhsdR mcrA mcrB mrr (FAtet lacI q lacZAM15)

a K12-U2-SI was constructed from K12-U2 by UV-curing ofpCLU1 and pGKL2. See text for other starins

out with a digoxygenin-labelledprobe at 42 ~ overnight as recommended by Boehringer-Mannheim Biochemicals. The 1.4-kb ClaISalI and 1.1-kb HindIII fragments of pARU 1 (Schaffrath et al. 1992) and the 0.9-kb HindIII-SalI fragment of pBDI2 (Dunn et al. 1984) were used as probes, respectively, of the LEU2, URA3 and telomere DNAs. The pBD12 fragment contains the 115-bp repeat sequence (GI_3T)n of the yeast telomere (Dunn et al. 1984). The 2.1-kb EcoRI-SaII fragment of pSAK068 (Sakai et al. 1990), as well as the 1.8-kb EcoRI-BglII fragment and the 3.5- and 2.3-kb PvuII fragments of pCYR1-T1 (Iida 1988), were used as probes of the Ty element. The 1.8-kb EcoRI-BglII fragment of pCYR1-T1 resides within the 2.1-kb EcoRI-SalI fragment of pSAK068. DNA sequence analysis. The 1.1-kb NsiI fragment of pTLU 1-E 1 was cloned into the PstI site ofpUC19 and the first 623 nucleotides were

sequenced by DNA sequencer Type 373A, Applied Biosystems Inc., using a dye primer cycle sequencing kit. UV-mediated plasmid-curing and cell-survival. The procedure described by Gunge et al. (1994) was employed.

282 Fig. 2 Growth behaviour on various selective media of cells harboring pCLU1 and/or a nuclearly migrated plasmid. Top of plate, cells of the S. cerevisiae strain K12-PC growing on SD +ura but not on SD +leu and SD; right middle, colonies appearing on SD + leu with a low frequency after 6 days culture; down, cells of a Leu+Ura§ clone produced on SD + leu; left middle, UV-treated cells of a Leu§ Ura§ clone (cells having noncured pCLUI show scattered growth on SD + ura and SD)

Results Construction of pCLU1 carrying both the cytoplasmic and nuclear markers The cytoplasmic plasmid pCLU1 was constructed as in Fig. 1 (for detail see Materials and methods). Briefly, the 4.3-kb Pstl-SacI fragment of pARU1 was introduced into the pGKL1,2-bearing leu2 ura3 mutant K12-2A-CK of S. cerevisiae for homologous recombination with the resident pGKL 1. Leu § transformants were selected on SD + Ura and analyzed for the plasmids. As a result, one of the Leu + clones, termed K12-PC, was found to harbor the TP-associated 8.4-kb linear plasmid pCLU1. Restriction and Southern analyses revealed that pCLU1 contained both the UCS-fused LEU2 and URA3 genes at the expected positions on pGKL1. The cytoplasmic location of pCLU1 was strongly suggested by the ability of cells carrying pCLU1 to grow on S D + U r a (the expression of the UCS-fused LEU2 gene) but not on SD +Leu (the non-expression of the URA3 gene) and was confirmed by the replication behavior of pCLU1, which was dependent on pGKL2.

Possible transmission of pCLU1 into the nucleus The K12-PC cells did not grow on SD+leu, because the nuclear URA3 gene on pCLU1 did not function cytoplasmically. After a prolonged incubation over 5-7 days at 30 ~ however Ura + clones appeared on SD + leu at a low frequency of about 10-4 to 10-5/plated cells (Fig. 2). On further examination, some of these clones were found to grow on SD + Leu only (phenotype of Leu-Ura +) while the remainder grew on SD + ura and SD as well as on SD + leu (phenotype of Leu+Ura+). We assumed that the expression of the URA gene in the above Ura + clones was exerted from inside the nucleus rather than from the cytoplasm. Taking this into account, 23 Ura + clones produced on S D + l e u were randomly picked up and analyzed for plasmid DNA. Gel-electrophoresis revealed that a majority of Ura + clones (21/23) contained novel plasmids migrating at various positions above pCLU1 (Fig. 3A). They all hybridized to both the LEU2 and URA3 probes, whether the phenotype

was Leu Ura + or Leu+Ura +. Closer examination showed that pCLU1 was present in all of the Leu+Ura + clones but in none of the Leu-Ura + clones, demonstrating that the Leu+Ura + phenotype has resulted from the simultaneous expression of the UCS-fused LEU2 gene on pCLU1 and of the URA3 gene on the nuclearly migrated plasmid. To clarify this, cells grown on SD were picked up and exposed to UV light. They then lacked both pCLU1 and pGKL2 and were able to grow on SD + leu but no longer on SD and SD + ura (Fig. 2), supporting the above view of the LEU2 gene being expressed in pCLU 1. UV light is known to preferentially eliminate the cytoplasmic plasmids of yeast (Gunge et al. 1994). Thus, the nuclear localization of novel plasmids with the Ura + phenotype was strongly suggested. In fact, many of plasmids from Ura + clones (15 of 23 Ura + clones) hybridized to the telomere probe, indicating that they had acquired part of a telomere sequence (GI_3T)n from the host chromosome subsequent to migration to the nucleus. The nuclear location of novel plasmids was also supported by the fact that they no longer required pGKL2 for replication. The two Ura + clones (Fig. 3 A, lanes e and s) appeared to lack the novel plasmid, suggesting that the phenotype may have been expressed by chromosomally integrated URA3 DNA.

Structural analysis of novel plasmids It was of interest to determine how pCLU1 may have been modified in structure after migration to the nucleus. To study this, the two representative Ura + clones, K12-U1 (Fig. 3 A, lane d) and K12-U2 (Fig. 3 A, lane g), harboring novel plasmids termed pTLU1 (telomere-added) and pRLU1 (telomere-free), respectively, were selected and analyzed as follows:

Telomere-added pIasmid pTLU1

One of the telomere-added plasmids, pTLU 1 of 8.8 kb, was extracted from K12-U1 (Leu-Ura+), purified, and subjected to restriction and Southern analyses (Fig. 4A, lane 2). Consequently it was found that pTLU1 was linear and that the structure was essentially the same as that of pCLU 1 except for lacking TP and having termini enlarged by a telomere addition (Fig. 4B-a, b). That is, the 1.6- and 0.9-kb PstI end fragments of pCLU 1 were elongated to 1.8and 1.1-kb fragments, respectively, showing that the added

283 Fig. 3 A gel electrophoresis and Southern analysis of bulk DNAs from K12-PC-derived Ura + clones. Lane M, size marker (HindIII fragments of A DNA); lanes a-w, Ura + clones produced on SD + leu. +/- (Leu) below lanes in gel-electrophoresis, the presence/absence of growth on SD + ura. Novel plasmids are seen at different positions above pCLU1, and hybridized to both the LEU2 and URA3 probes. Some of them further hybridized to the telomere probe. Novel plasmids were not visible with two clones (lanes e and s) probably because of the chromosomal integration of the URA3 gene. The two representative clones K12-U1 (lane d) and K12-U2 (lane g) were chosen for detailed analysis of novel plasraids, telomere-added and telomere-free. B analysis of bulk DNAs from clones carrying a Ty-inserted plasmid pTLU1-E1 (15.0 kb) with and without pTLU1. Lanes 1-3 represent K12-U1, K12-U1-S1, K12-U1$2. The 2.1 -kb EcoRI-SalI fragment from pSAK068 was used as the Ty-probe

A

B Gel l

Southern l

123

i

[Probes LEU2 I

2

URA3 3

1

2

Telomere 3

I

2

3

Ty i

2

3

chr-._ pTLUI-EI-~

pTLU1--

t e l o m e r e was about 0.2 kb in size. The 3.6- and 2.3-kb internal P s t l f r a g m e n t s r e m a i n e d intact. The t e r m i n a l e l o n g a t i o n was r e p r e s e n t e d b y s m e a r bands, due to the length h e t e r o g e n e i t y o f the a d d e d telomere. S m e a r e d elongation at the termini and the e x p e c t e d size o f internal frag-

ments were also o b s e r v e d in the B a m H I / C l a I and HindIII digestions. The LEU2 and URA3 D N A s were d e t e c t e d as the 1.4-kb B a m H I / C l a I and the 1.1-kb HindIII f r a g m e n t s (Fig. 4 A ) , respectively, at the e x p e c t e d positions in pTLU1.

284 Fig. 4 A gel-electrophoresis and hybridization with the telomere probe of restriction fragments of plasmids. Lanes 1-4 represent pCLU1, pTLU1, pTLU 1-E 1 and pRLU 1, respectively. Terminal fragments from pTLU1 and pTLU1-E1 were smeared because of the length heterogeneity of the added telomere. Asterisks indicate the average size of the telomere-added terminal fragments. Several upper bands in the hybridization with the telomere probe of the PstI fragments are due to partial enzymatic digestion. For detail, see text. B structure of plasmids: a, pCLU 1; b, pTLU 1; c, pTLU-E1; d, pRLU1. Physical maps of these plasmids were made by compiling the data from restriction analysis in Fig. 4 A and other sources. pTLU1 lacks TP but contains a telomere sequence added at the termini, pTLU-E1 was produced by an insertion of a Ty element into pTLU1, pRLU 1 was formed by the end-to-end fusion of pCLU 1. To facilitate comparison with other plasraids, the physical mpa of pRLU1 is presented in a linear form by cleaving at the supposed terminal fusion site of pCLU1. Closed circles and vertical bars at the ends of linear plasmids indicate terminal protein (TP) and added telomere sequence, respectively. Ty and pGKL sequences are shown by horizontally-hatched and open boxes, resectively. Other symbols are as for Fig. 1

B

3

pCLUI

kb)

(8.4

a

'I ,P

-]

?

"dJ~ If?

r

I I~UZ*HJURA31

L U

P }---1.6 , B/C1 3.4 II ~-.9'

b

(8.8

pTLUI

'I r

(0.2 kb) p

pTLU1-EI

= .9-t 1.9-'q 2.3-..--4

kb)

Ilii,7

C

3.6 , 2.3 ' 1.4 B 1.7 4.1 , 1.1 .

(15.0

IIIIll

J--l.8 ~

~ "? " t5:ox.NIu,., l ~ 3.6

' 1.4 e

II

.....

I-1.1--4.1

2.3

9

B/C I - - - ~ 3 . 6

1.7 1.1

,

P

,,l.l-t 2 . 1 ----.4 2.5 i

kb)

,

I

r

pRLUI

r

ItUII

(0.4 kb)

(0.4 kb) P ~--- 2 . 0 B/C ~ I, ~ 1 . 3

BII

(0.2 kb)

,.7' 3.7"

5.6 9

5,9 9.9'

'3.1 ' 2.3~ 9 1 . 4 11 1 . 7 ' 1.1 ,

1.3-r 2 . 3 .--..-4 2.7

1

(8.4 k b )

P I--1.6 B/O t ti ~.9 '

Elongated linear plasmid pTLU1-E1 Cells of the clone K12-U1 carrying pTLU1 were grown in SD + leu for 3 months and plated for single colonies. A subclone culture (K12-U1-S1) was found to harbor a novel linear plasmid, termed pTLU1-E1 (15.0 kb), in addition to p T L U I . After more repeated growth, a clone K12-U1-S2 harboring only pTLU1-E1 was isolated, pTLU1-E1 hybridized to all of the LEU2, URA3 and telomere probes (Fig. 3 B).

3.4'

3.6 B " 2 . 3 ...... .9 q , 1.4~1.7~1.9 .---4 4 . 1 ~ 1 . 1 ' ,2.3----4

pTLU1-E1 was extracted from K12-U1-S2 and purified. The structural analysis (Fig. 4 A, lane 3) revealed that pTLU1-E1 was linear and that the termini were elongated by the telomere addition, pTLU-E1 yielded PstI terminal fragments of about 2.0 and 1.3 kb, showing that the average size of attached telomere was 0.4 kb, somewhat larger than that of pTLU1. Study indicated that pTLU1-E1 was produced from pTLU1 by an insertion of a 5.8-kb D N A sequence in front o f the UCS-fused LEU2 gene (Fig. 4 B - c ) ; that is, the 3.6-kb CIaI left-end fragment (3.4-kb flag-

285 Fig. 5 A Southern hybridization of the pTLU1-E1 CIaI fragments to the Ty probes A, B and C. Physical maps of pTLU1-E1 and pCYR1-T1 are also shown. The 1.1-kb NsiI fragment was cloned into the PstI site of pUC19 for sequence analysis. Abbreviations,

--I1 Pv X Bg Pv Bg i

C

B

r T

-- A

"1 ---] PvX E

Bg

Ty

Probe

~

7

X XhoI; Bc BclI; Pv PvuII; Bg BgllI; E EcoRI; N NsiI. Oth-

er symbols are as for Fig. 1. B nucleotide sequence around the Ty insertion site. Soruces of sequence data: pTLU1 UCSfused LEU2 gene (Kfimper et al. 1991), Tyl-912 (Clare and Farabaugh 1985), Ty2-117 (Warmington et al. 1985). Identical nucleotides are marked with an asterisk. Note, the NsiI site at the start and second codons of the LEU2 gene was created during the construction of the UCS-fused LEU2 gene (K~impfer et al. 1991): a DNA fragment containing the 5' noncoding region and the start codon of pGKL 1-ORF2 (the pGKL 1-kilter toxin gene) was connected in-frame to the second codon of the S. cerevisiae LEU2 gene by converting its original serine codon AAT (Andreadis et al. 1984) to the histidine codon CAT through in vitro mutagenesis, resulting in the creation of the NsiI site (ATGCAT) at the front of the UCSfused LEU2 gene

5.9 kb

3.7 kb

3.1 kb C

C

Bc

!1111

I

1

I

,

-

N

N

r

I I I

(0.4 kb)

IN

BH

t I

1.1 kb cloned

UCS

2.3 kb C

,'3r r"

P

HI

~I

P ,

(0.4 kb)

LEU2*

o..TATAGAAAATiAGACCGTGAiAGCT~AGTTTTTATAAiAATTATAA+TGCAT.. " Ty ************************* ATACTAGTTAGTAGATGATAGTTGATTTCTATTCCAACA~STTTTTATAATAATTATAA~IATGCA,. pTLUI-EI *************************************** Ns__~il TyI-912 ...ATACTAGTTAGTAGATGATAGTTGATTTCTATTCCAACA pTLUI

m e n t + 0 . 2 - k b telomere) of pTLU1 was elongated to a 9.6-kb sequence (3.7- and 5.5-kb fragments +0.4-kb telomere). The inserted sequence of 5.8 kd did not hybridize to any part of pTLU1 (data not shown), indicating that the elongation in pTLU1-E1 was due to a genetic event other than a molecular rearrangement within pTLU1, pTLU1El-like plasmids of about 15.0 kb were also detected among Ura + clones selected from K12-PC carrying pCLU 1 (Fig. 3 A, lanes a and n), suggesting that the insertion event has occurred not only in a long-term incubation of K12U1 but also at an early stage of the migration of pCLU1 into the nucleus.

Ty insertion in p T L U 1

The question arose as to where the 5.8-kb insertion had originated from. In a preliminary study, it was implied that the insert might be related to a Ty element (a retrotransposon ofS. cerevisiae) (Boeke and Sandmeyer 1991), because the 2.1-kb E c o R I - S a l I fragment of pSAK068, which contains the 5' long-terminal repeat (the 5 ' a domain) and the upstream one-fourths of the internal element (the e domain) of Ty, hybridized to pTLU1-E1 but not to pTLU1 (Fig. 3 B). pTLU-E1 contained three ClaI sites (one within

the insert and two on the right part of pTLU1); so, in order to delimit the site of the putative Ty insertion, the four ClaI fragments (5.9, 3.7, 3.1 and 2.3 kb) ofpTLU1-E1 were hybridized to each of the Ty probes, A, B and C, from pCYR1-T1 (Fig. 5 A). Probes A and B together encompass the 5'~domain and the 5' two-thirds of the ~ domain, while probe C includes the 3'6 domain and the 3' one-third of the domain. The result showed that the 3.7-kb ClaI fragment covering the right two-thirds of the insert hybridized specifically to both the A and B probes, while the 5.9-kb ClaI fragment containing the left one-third of the insert hybridized to probe C. Thus a Ty element of 5.8 kb was shown to be inserted into pTLU1 upstream of the L E U 2 gene in a reversed transcriptional direction. The 3.1- and 2.3-kb ClaI right fragments did not hybridize to any Ty probes. To determine the exact Ty insertion site, the 1.1-kb NsiI fragment, covering from the first five nucleotides of the L E U 2 gene to an upstream region of Ty, was cloned into the PstI site of pUC19 and the first 623 nucleotides were sequenced. The result is partly given in Fig. 5 B, showing that the Ty element was inserted 22 bp upstream of the L E U 2 start codon,just 1 bp downstream from the UCS motif. Sequence comparison with the two well-studied Ty elements, Ty 1 (Ty 1-912) and Ty2 (Ty2-117), revealed that the inserted Ty had a perfect homology with Tyl-912 for the

286 SD + ura

SD + ura + leu

SD + leu

1

10 ~

= o

g

'~ 10 "~

o

10-~

12 Time

(hrs)

12

24 Time

24

(hrs)

12 Time

24

10-4

(hrs)

Fig. 6 Repressed expression in MATa/MAT~ diploid cells of the UCS-fused LEU2 gene on pTLU1-E1. Symbols: 9 haploid K12U1-$2; e, diploid K12-U1-S2xMK031-6B

first 173 bp and 90% homology for the total 596 bp of the analyzed Ty sequence. Homology with Ty2-117 was lower: there were three bp mismatches in the first 173 bp and 82% homology for the total analyzed sequence (the total sequence data will be published elsewhere).

10-5

I

I

I

100

200

300

uv

close

(g/m 2)

Fig. 7 Differential UV-sensitivity of pCLU1 and the derived nuclear plasmids. Symbols: e , cell survival in K12-2A; o, pCLUl-curing in K12-PC; n pTLUl-curing in K12-U1; O, pTLU1-El-curing in K12-U1-S2; A, pRLUl-curing in K12-U2-S1. All data were obtained in the isogenic host strain K12-2A

Telomere-free plasmid pRLU1 DNA of the telomere-free plasmid pRLU1 was purified from K12-U2. pRLU1 was digested into three fragments of 2.5, 3.6 and 2.3 kb with PstI, of 5.3, 1.4 and 1.7 kb with BamHI/ClaI, and of 3.2, 4.1 and 1.1 kb with HindIII, respectively (Fig. 4 A, lane 4), demonstrating that the total plasmid size is 8.4 kb, identical to that of pCLU1. Comparison of the constructed restriction maps led to the conclusion that pRLU1 was circular and produced by the endto-end fusion of pCLU1 without carrying a TP (Fig. 4Bd), although the details of the terminal fusion site remains unsettled. K 12-U2 initially contained pCLU 1 and pGKL2, besides pRLU1, and grew on all of SD + leu, SD + ura and SD. When cured of pCLU1 and pGKL2 by UV irradiation, cells grew only on SD + leu, consistent with the view that the Ura § phenotype was due to the nuclear location of pRLU1.

ing-type locus of the host cell: this phenomenon is called the ROAM (regulated overproducing alleles responding to mating type) effect. To examine whether the ROAM effect can be seen for the above Ty-mediated expression of the UCS-fused LEU2 gene, K12-U1-S2 (MATa leu2 ura3 pTLU1-E1) was crossed to MKO31-6B (MATa leu2 ura3 ade2 met1) and the resulting diploid cells were compared with the haploid parent K12-U1-S2 for LEU2 gene expression (the ability to grow in SD + ura). As shown in Fig. 6, the diploid cells grew much slower in SD + ura than did K12-U1-S2, indicating the decreased expression in diploid cells of the LEU2 gene. The growth of diploid cells in SD + leu was at the same level as that in a complete medium (SD + leu + ura), showing that the expression of URA3 gene was not affected by cell-type. Thus, the ROAM effect on the UCS-fused LEU2 gene was clearly demonstrated.

UV sensitivity Ty-mediated LEU2 expression on pTLU1-E1 As described previously, the clone K12-U1 carrying pTLU1 did not grow on SD + ura owing to the non-expression in the nucleus of the UCS-fused LEU2 gene. Interestingly, however, we noticed that the clone K12-U1-S2 carrying pTLU1-E1 grew on S D + u r a and SD as well as on SD + leu (phenotype of Leu+Ura+). When cured of pTLU 1E 1, cells failed to grow on any of the selective media. Thus the expression of the LEU2 gene in K12-U1-S2 was pTLU1-El-dependent and certainly ascribable to the Ty insertion upstream of the LEU2 gene. Ty insertion is generally known to affect the expression of a neighboring gene by modifying the function of the 5'regulatory sequence (Boeke and Sandmeyer 1991). Ty-mediated activation of a neighboring gene is usually under the control of the mat-

Previously we have reported that the UV-sensitivity of yeast plasmids varied according to the cellular location of the plasmids (Gunge et al. 1994). This was also true of pCLU1 and pCLUl-derived nuclear plasmids: the cytoplasmic plasmid pCLU1 was hundreds of times more curable than the nuclearly located plasmids, both linear (pTLU1 and pTU1-E1) and circular (pRLU1) (Fig. 7).

Discussion

The previous study of the telomere-added linear plasmid pLS 1 (K~imper et al. 1989) has suggested that the cytoplasmic pGKL plasmids could migrate into the nucleus under

287 certain conditions. To obtain more information, we have, of the Debaryomyces pDHL plasmids was confirmed by a in this paper, constructed a p G K L l - b a s e d linear plasmid cell fractionation analysis (our unpublished result). pCLU1 with indicator genes capable of expression only in the cytoplasm or in the nucleus and have revealed a low- Acknowledgements We thank Dr. B. Dunn for providing pBD12, Dr. A. Sakai for pSAKO68 and pCYR1-T1, Dr. M. Tokunaga for S. frequency migration to the nucleus of pCLU1, which nor- cerevisiae MKO31-6B, and Dr. K. Shimizu for E, coli Sure. We are mally resides in the cytoplasm. Nuclearly migrated plas- also grateful to Dr. K. Esser for his interest in this study. raids were mostly detected as linear plasmids having a telomere repeat sequence at their termini, although some of them existed in a circularized form. References The exact mechanism whereby p C L U t migrates into the nucleus and gains the telomere sequence is not known. Andreadis A, Hsu Y-R Hermodson M, Kohlhaw G, Schimmel P Considering that the yeast nucleus does not break down (1984) Yeast LEU2: repression of mRNA levels by leucine and primary structure of the gene product. J Biol Chem 259: during mitosis, one may argue that pCLU1 may first need 8058-8062 to settle outside a nuclear pore prior to the access to a chroBoeke JD, Sandmeyer SB (1991) Yeast transposable elements. In: mosome end. Having entered the nucleus, pCLU1 may The molecular and cellular biology of the yeast Saccharomyces have needed the telomere addition at the termini to replivol 1. Genome dynamics, protein synthesis, and energetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, cate in a linear form, because the pGKLl-based TP would pp 191-261 probably not be utilized in the nucleus as a primer of DNA Clare J, Farabaugh P (1985) Nucleotide sequence of a yeast Ty elesynthesis. Unless the telomere was added, pCLU1 must ment: evidence for an unusual mechanism of gene expression. have been circularized, as pRLU1, for the replication. A Proc Natl Acad Sci USA 82:2829-2833 high-frequency addition of the telomere sequence to the Company M, Errede B (1988) A Tyl cell-type-specific regulatory sequence is a recognition element for a constitutive binding factor. plasmid termini was noticed in this study (Fig. 3 A ) This Mol Cell Biol 8:5299-5309 result was in marked contrast to the previous observation Cong YS, Yarrow D, Li YY, Fukuhara H (1994) Linear DNA plasthat attempts to integrate a nuclear gene into pGKL mostly raids from Pichia etchellsi, Debaryornyces hansenii and Wingea led to telomere-free circular plasmids locating in the nurobertsiae. Microbiology 140:1327-1335 cleus (De Louvencourt et al. 1983; Fujimura et al. 1987; De Louvencourt L, Fukuhara H, Heslot H, Wesolowski M (1983) Transformation of Kluyveromyces lactis by killer plasmid DNA. K~imper et al. 1989; Tanguy-Rougeau et al. 1990) The reaJ Bacteriol 154:737-742 son for these contrasting results is unclear, but one pos- Dunn B, Szauter R Pardue ML, Szostak JW (1984) Transfer of yest sible explanation may be that the novel plasmids in the telomeres to linear plasmids by recombination. Cell 39:191-201 present study were obtained through a natural migration to Errede B, Company M, Hutchison CA (1987) Tyl sequence with enhancer and mating-type-dependent regulatory activities. Mol Cell the nucleus of the already marker-inserted linear plasmid Biol 7:258-265 pCLU1. Fujimura H, Hishinuma F, Gunge N (1987) Terminal segment of A Ty element was inserted in the nuclearly migrated Kluyveromyces lactis linear DNA plasmid pGKL2 supports auplasmid pTLU1 22 bp upstream of the LEU2 coding retonomous replication of hybrid plasmids in Saccharomyces cerevisiae. Curr Genet 12:99-104 gion of the UCS-fused gene, allowing cell-type-regulated Gunge N (1986) Linear DNA killer plasmids from the yeast Kluyveexpression of the LEU2 gene. Considering that the LEU2 romyces. Yeast 2:153-162 coding region was separated far away from the UCS ele- Gune N (1995) Plasmid DNA and the killer phenomenon in Kluyvement by the Ty insertion (Fig. 5 B), we presume that, as in romyces. In: Kack U (ed) Genetics and biotechnology. The Mycota vol II. Springer Verlag, Berlin Heidelberg, pp 189-209 many reported cases of Ty-mediated gene activation, the ROAM-controlled LEU2 gene expression was wholly Gunge N, Yamane C (1984) Incompatibility of linear DNA killer plasmids pGKL1 and pGKL2 from Kluyveromyces lactis with under the control of Ty-based transcription regulatory mitochondrial DNA from Saccharomyces cerevisiae. J Bacteriol sequences, such as TAS1 (Rathjen et al. 1987) or Ty 159:533-539 enhancer-like elements (Errede et al. 1987; Company and Gunge N, Fukuda K, Morikawa S, Murakami K, Takeda M, Miwa A (1993) Osmophilic linear plasmids from the salt-tolerant yeast Errede 1988) located downstream from the 5' domain. Debaryomces hansenii. Curt Genet 23:443-449 Ty-mediated gene activation occurs when these Ty-based Gunge N, Takahashi S, Fukuda K, Onishi T, Meinhardt F (1994) UVregulatory sequences are positioned proximal to the target hypersensitivity of yeast linear plasmids. Curr Genet 26:369-373 gene. Such a positional relationship was also the case for Iida H (1988) Multistress resistance of Saccharomyces cerevisiae is generated by insertion of retrotransposon Ty into the 5'coding rethe direction of Ty in pTLU1-E1. gion of the adenylate cyclase gene. Mol Celt Biol 8:5555-5560 The UV-hypersensitivity of yeast linear plasmids was K~imper J, Meinhardt F, Gunge N, Esser K (1989) In vivo construcobserved with the cytoplasmic pCLU1 and has been retion of linear vectors based on killer plasmids from Kluyveromyported also for naturally-occurring linear plasmids such as ces lactis: selection of a nuclear gene results in attachment of telomeres. Mol Cell Biol 9:3931-3937 pDHL1,2,3 (Gunge et al. 1993) and pDH1A,1B (Cong K~imper J, Esser K, Gunge N, Meinhardt F (1991) Heterologous gene et al. 1994) from Debaryomyces hansenii, pPacl 1,2 from expression on the linear DNA killer plasmid from KluyveromyPichia acaciae (Worsham and Bolen 1990), pWR1A,1B ces lactis. Curr Genet 19:109 118 from Wingea robertsii and pPE1A,1B from Pichia etchell- Rathjen PD, Kingsman AJ, Kingsman SM (1987) The yeast ROAM sii (Cong et al. 1994), which are all believed to locate in mutation - identification of the sequences mediating host-gene activation and cell-type control in the yeast retrotransposon, Ty. the cytoplasm. Thus a UV-senitivity test appears useful to Nucleic Acids Res 15:7309-7324 predict the cellular location of yeast plasmids as to whether Romanos M, Boyd A (1988) A transcriptional barrier to expression cytoplasmic or nulear. Recently, the cytoplasmic location of cloned toxin genes of the linear plasmid kl of Kluyveromyces

288

lactis: evidence that native kl has novel promoters. Nucleic Acid Res 16:7333-7350 Rose MD, Winston F, Hieter P (1990) Methods in yeast genetics: a laboratory course mannual. Cold Spring Harbor Laboratory. Cold Spring Harbor, New York Sakai A, Shimizu Y, Hishinuma F (1990) Integration of heterologous genes into the chromosome of Saccharomyces cerevisiae using a delta sequence of yeast retrotransposon Ty. Appl Microbiol Biotechnol 33:302-306 Salas M (1991) Protein-priming of DNA replication. Ann Rev Biochem 60:39-71 Sambrook J, Fritsch EF, Maniatis T (1989) Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York Schaffrath R, Stark MJR, Gunge N, Meinhardt F (1992) Kluyveromyces Iactis killer system: ORF1 of pGKL2 has no function in immunity expression and is dispensable for killer-plasmid replication and maintenance. Curr Genet 21:357-363

Stark MJR, Boyd A, Mileham A, Romanos MA (1990) The plasmidencoded killer system of Kluyveromyces lactis: a review. Yeast 6:1-29 Tanguy-Rougeau C, Chen XJ, Wesolowski-Louvel M, Fukuhara H (1990) Expression of a foreign KmR gene in linear killer-DNA plasmids in yeast Gene 91:43-50 Tommasino M, Ricci S, Galeotti CL (1988) Genome organization of the plasmid pGKL2 from Kluyveromyces Iactis. Nucleic Acids Res 16:5863-5878 Warmington JR, Waring RB, Newlon CS, Inge KJ, Oliver SG (1985) Nucleotide-sequence characterization of Ty 1-17, a class-II transposon from yeast. Nucleic Acids Res 13:6679-6693 Worsham PL, Bolen PL (1990) Killer-toxin production in Pichia acaciae is associated with linear DNA plasmids. Curr Genet 18:77-80 Zakian VA (1989) Structure and function of telomeres. Annu Rev Genet 23:579-604