YIAST
VOL.
7: 245 252 ( I99 1 )
Killer System of KZuyveromyces Zactis: the Open Reading Frame 10 of the pGK12 Plasmid Encodes a Putative DNA Binding Protein A
MASSIMO TOMMASINO Sduro R i w u r c h C’oitrr, c‘iu Fioruitincr
I , 53100 Sicwa. IIUIJ
Rccciwd 28 July 1990: rcvised 21 October 1990
O R F 10 ofthe K 2 plasmid from Klz?,.~,croni?.c.es luctis encodcs a small basic protcin (22.3% lysine). Thc function of its product has bccn invcstigatcd. Western blot analysis, using an antibody against MS2 RNA polymerase!ORF 10 I‘usion protein. reveals a protein band with an apparent molecular weight of 14 kDa. The protein can bind a DNA-Sepharose column. and is eluted by 350 mM-salt. lmmunoprccipitation experiments show that the O R F 10 protein coprccipitates with thc lincar gcnomic DNAs ofthe two killer plasmids ( K I and K2). From Wcstcrn;Southcrn blot data. it is possiblc t o conclude that thc interaction between protein and DNA occurs directly, rather than via other protcin(s). O R F 10 is cosily detected by Western blot and its transcript is one of the most abundant of thc K 2 plasmid. suggesting that this protein may have a structural rather than a regulatory function. This possibility is also suggested by the obscrvcd squcncc homology between thc O R F 10 protcin and the family of histonc-likc proteins, l i i Y WOWS
--
K . Itrciis; killcr system; ORF 10: DNA binding protein
I N T R 0D U C T I O N Some strains of Klu~i~c~ron?ycc.s lactis are able to secrete an olipomcric protein which inhibits the growth of other strains. This so-called killer phenotype is associated with the presence of two linear plasmids in the cell, termed pGKll (8.9 kb) and pGKI2 ( 13.4 kb) (shortened to K 1 and K2. respectively) (for rcview see Stark Pt a/.. 1990). Several lines ofevidence support theidea that the twoplasmidsarelocalized in the cytoplasm (Gunge c’t ul., 1982; Stam et a/., 1986; Wilson and Meacock, 1988). The 5‘ ends of both plasmids are blocked by two proteins with apparent molecular weights of 36 kDa (K2) and 28 kDa ( k l ) (Stam et ul.. 1986). By analogy with other linear rxtrachromosomal genomes, e.g. adenovirus (for review scc Friefeld e t a/.. 1986). these proteins may function as primers in D N A replication. K I has been sequenced and shown to contain four open reading frames (ORFs) (Hishinuma et af.. 1984; Sor and Fukuhara, 1985; Stark ct ul., 1984). Two of them code for the subunits of the toxin ( O R F s 2 and 4) (Stark and Boyd, 1986), while O R F I and O R F 3 code respectively for a plasmidspecific DNA polymerase and a protein involved in the immunity system (Fukuhara. 1987; Tokunaga ot ul.. 1987). 0749 503x Y I 030245 0 8 so5.00 D I Y Y I b! J o h n Wiley & Sons l.rd
Genetic analysis has shown that the presence of K2 in thecell isessential for themaintenanceofboth plasmids, suggesting that the products of the K2 O R F s play fundamental roles in the killer system. K2 has been recently sequenced and shown to contain ten O R F s (Tommasino Pt af., 1988). O R F 2 and O R F 6 seem to code for plasmid-specific D N A and RNA polymerases respectively (Tommasino e t al., 1988: Wilson and Meacock, 1988). It is also possible that the terminal proteins of the two plasmids are encoded by K2. Nothing is known of the possible functions of the proteins encoded by the other ORFs, although thc very high percentage of lysine in O R F 10 (22.3%) suggests that its product may be a D N A binding protein. Here I show that the protein encoded by O R F 10 is a D N A binding protein and interestingly it has homology with a family of proteins classified as histone-like proteins. MATERIALS A N D M E T H O D S
Strains K . lactis CBS 2359 was obtained from the Centraal Bureau voor Schimmelcultures. Bacterial strains used were Escherichia coliJM IOITR (Sup E.
M. TOMMASINO
246
r I R
t
ORF 10 region of K2 GENOME
t
Xhol
EcoRl I
pEMBL9-EX8
t
EcoRl
t
Sau3Al
L
/ Xhol/SalI
Hindlll
r I R
pEX34A-ORF 10
t
Sau3AVBamHI
t
Hindlll
Figure 1. Schematic representation of cloning strategies and final constructs employed. All restriction enzymes used are indicated in the figure. IR is the inverted terminal repeat of K2 plasmid. For details see Materials and Methods.
MSZRNA polymerase fragment. The E. coli strain K-12 A D H l A Dtrp(whichc0ntainsa temperaturesensitive repressor of the lambda promoter pL) was transformed with the construct pEX34/ORF 10. Single colonies containing the plasmid were amplified in liquid culture at 30°C and the pL promotor Expression of MSZjORF 10 fusion protein in E. coli was derepressed by heat shock (2 h a t 42°C) (Nicosia and antiserum preparation et al., 1987). Bacterial proteins were prepared as The fragment EcoRI-XhoI (12556-13335) cover- described(Nicosiaetal., 1987)andanalysed by SDSing the total length of O R F 10 was cloned in the PAGE (Laemmli, 1970).The band corresponding to vector pEMBL9 cut with EcoRI/SalI (Figure 1). the MS2/0RF 10 protein was electroeluted from For expression of O R F 10 the vector pEX34A, a preparative polyacrylamide gel, checked for purity, derivative of pEX29, was used (Klinkert et al., 1985; and used to immunize rabbits (Nicosia et al., 1987). Strebel et al., 1986). This vector contains a fragment of DNA coding for the N-terminus of the gene MS2- Preparation of yeast protein extracts RNA polymerase, positioned downstream of the pL Late log-phase cultures were washed in 150 mMpromoter of phage lambda. An O R F 10 fragment, corresponding to 90% of the coding region, was KC1, 10 mM-Hepes, pH 7.9, 5 mM-MgCl,, 0. I mMexcised from the pEMBL9 construct by digestion EDTA, 0.5 mM-phenylmethylsulfonyl fluoride, with SAU3AI.and Hind111 and cloned in pEX34A 0.5 mhl-benzamidine and 5 mM-mercaptoethanol previously cut with BumHI/HindIII (Figure 1). This (buffer A), resuspended in 1 volume of the same results in an in-frame fusion of ORF 10 with the buffer and broken with glass beads in a beat beater. thi, srl::tnlO, recA, AD(1ac-proAB), [F, tra D365, pro AB, IacIQm ZM151) for routine work, and E. coli K-12 A D H l ADtrp for expressing O R F 10 (Remaut et al., 1981).
247
KILLER SYSTEM OF KLUYVEROMYCES LAC'TIS
After centrifugation (12 000 x g for 20 min), the supernatants were precipitated in ammonium sulphate (0.35 giml). The pellet was resuspended in 50 mM-Tris;HCI, pH 7.9, 20% glycerol. 0.1 mMEDTA, 100 mM-KCI (buffer C), dialysed overnight against the same buffer and frozen in liquid nitrogen. Iniriiimoprec.ipitatiori
K . lac(is was grown to late log phase and protoplasted as described (Gunge and Sakaguchi, 1981). Protoplasts were resuspended in 3 volumes of buffer A and after 10 min the suspension was centrifuged at 12000 x g for 15 min. The supernatant was recovered and precipitated with ammonium sulphate (0.35g!ml). The pellet was resuspended in buffer C and dialysed overnight against the same buffer. 3 mg of protein extracts were diluted in 250 pl of 50 mMTris.HC1, pH 7.4. 100 mwKCI, 5 pgiml Aprotinin and incubated for I h at 25'C with preimmune serum and protein A-Sepharose. After centrifugation, the supernatant was incubated for 3 h at 25'C with immune serum (anti-MS2iORF 10 protein) and protein A-Sepharose. The immunocomplex was centrifuged and washed four times with 1 ml of 50 mM-Tris/HCI, pH 7.4, 100 mM-KCI. The final pellet was resuspended in 1 0 0 ~ 1TE ( I0 mu-Tris:'HCI. pH 8.0, 1 mM EDTA), 1'10 SDS, I0pg tRNA. treated with proteinase K and extracted with phenolichloroform 111. DNA was precipitated in ethanol in sodium acetate 0.3 M, resuspended in 20 pl T E and applied to an agarose gel. In some cases the concentration of KCl in the immunoprecipitation buffer was brought up to 250 mM; no difference was noted (not shown). Other tnrrhod.v Manipulation of DNA was performed by standard procedures (Maniatis et al., 1982). RNA was prepared and analysed as previously described (Maundrell rt ul., 1985: Tommasino rt al., 1988). Double-stranded DNA probes were labelled using random primers (Feinberg and Vogelstein, 1984). Western blotting was performed by standard procedures and Western-Southern blotting was carried out as previously described (Bowen et al., 1980). Protein concentration was determined with BCA Protein Assay Reagent (Pierce).
RESULTS Identifificcrtiono f t h e ORF 10 lranscript Northern blot analysis of the K2 transcripts, using the entire plasmid as a radioactive probe,
ORF 10
Figure 2. Northern blot analysis of mRNA transcribed from K 2 and ORF 10. Total RNA from two different yeast strains was analysed by Northern blotting. Lanes a, c: 10 pg total RNA from K . luciis CBS 2359 (killer strain); lane b: 10 pg total RNA from K . /ucii.s WM 37 (non-killer strain). Theentire K 2 plasmid (lanes a, b) or the 0.39 kb SAU3AI-Xhol fragment (lane c), which cover 90% of ORF 10. were used as radioactive probes. The positions of the 26s and 18s rRNA of yeast are indicated by the arrows.
revealed at least eight bands in the range of about 3.2-0.4 kb, which were consistent with the predicted sizes of the ORFs of K2 (Figure 2, lane a). In order to ic'cntify which of the eight transcripts corresponds to O R F 10. a fragment of K2 (SALi3AIXhol) which covers 90% of the entire O R F 10 was used. Figure 2 shows that the O R F 10 transcript corresponds to the band of approximately 0.4 kb (lane C ) .one of the most abundant transcripts of K2 (lane a). No hybridization was detected when total RNA o f a non-killer strain was examined (Figure 2, lane b). E.rpression of ORF 10 in E. coli
We expressed O R F 10 in E. coli as a fusion protein. A portion of the gene (900/, of the coding region) was cloned in the vector pEX34A (see Materials and Methods and Figure 1 ) and transformed into E. coli. After induction, the total cell
248
Figure 3. Detection of MS2jORF 10 protein synthesized in E. c d i . Total cellular protein extracts from E. coli transformed with pEX34A or pEX34A;ORF 10 were subjected to SDS-PAGE, stained with Coomassic blue or electroblotted and screened with a monoclonal antibody against the N-terminus ofthe MS2 RNApolymerase. Panel A: Coomassie staining. Lane a: 100 pg of total protein extract of [pex34A] E. coli; lane b: 100 pg of total protein extract of[pEX34A;ORF lo] E. coli.Panel B: Western blot. Lane a: IOOpg of total protein extract of [pEX34A] E. coli; lane b: 100 pg of total protein extract of (pEX34a:ORF 101 E. coli.The sizes in kDa of the markers (Rainbow. Amersham) are indicated on the left.
lysate was analysed by SDS-PAGE. Coomassie blue staining revealed a major band with an apparent molecular weight corresponding to that of the fusion product (Figure 3. panel A. lane b). This protein band was able to react. in a Western blot, with a monoclonal antibody raised against the N-terminus of the MS2-RNA polymerase (Figure 3, panel B, lane b). The fusion protein was purified as described in Materials and Methods and used to immunize rabbits. To check the specificity of the polyclonal antibody for O R F 10. I performed Western blot analysis with total yeast protein extracts from killer and nonkiller strains. The antiserum reacted with a single
M. TOMMASINO
Figure 4. Detection of the O R F 10 products in K. lmri.~.Total protein extracts (100 pg), prepared as described in Materials and Methods, of K. lacris WM 37 (killer strain) (lanes a. c) and K. lacris CBS 2359 (killer' strain) (lanes b. d ) were subjected to SDS-PAGE, electroblotted and screened with two different antibodies. Lanes a. h: monoclonal MS2 antibody; lanes c. d: polyclonal MS2:ORF 10 antibody. The size in kDa of the markers (Rainbow. Amersham) are indicated on the left.
protein band of apparent molecular weight 14 kDa, present in the protein extract of the killer+ strain ( K . lucfisCBS 2359) (Figure 4. lane d), while no reaction was detected in the extract of the killer- strain ( K . lucfis WM 37) (Figure 4, lane c). As a control, both extracts were probed with the MS2 monoclonal antibody and no signal was detected in either strain (Figure 4, lanes a. b). ORF loprotein b i n h to Sepharose-DNA column The highly basic character of the O R F 10 product (22.3% lysine + 2.9% arginine) suggests that this protein could be a DNA binding protein. In order to test this possibility, the total protein extract of K . lactis CBS2359 (see Materials and Methods)
249
KILLER SYSTEM OF KLL'YVEROMYCES LACTIS
d 0
FRACTION NUMBER Figure 5. Binding of the yeast O R F 10 protein to DNA-Sepharose. About 90 mg o f total cellular proteins of K. bcris CBS 2359. prepared as described in Materials and Methods. were applied to a DNA-Sepharose column (7 x 2 cm) and eluted by a linear NaCI gradient (04.8 M). An aliquot of each fraction was subjected to SDS-PAGE, electroblotted and screened with polyclonal MS2; ORF 10 antibody. The fractions containing O R F 10 protein are shown in the inset figure.
was fractionated on a Sepharose-DNA column (prepared as described by Scheidereit er ul., 1987). The bound proteins were eluted from the column by a linear gradient of NaCl(0-800 mM). All fractions were analysed by SDS-PAGE and Western blot using the MS2/0RF 10 antibody. More than 70% of the O R F 10 product was bound to the column and it was eluted at an ionic strength of about 350 mM, demonstrating that it binds to DNA with a good affinity (Figure 5). ORF 10 protein i s ussociuted with killer plusmid D:VA
Preliminary cell fractionation experiments showed that the O R F 10 protein and the two killer plasmids were present in the same subcellular fraction, suggesting a possible association of this protein to K I and K2. To test this, I have immunoprecipitated the O R F 10 protein with MS2iORF 10 polyclonal antibody. The pellet was treated with proteinase K. phenol extracted and its DNA content analysed by agarose gel electrophoresis. Both killer plasmids were immunoprecipitated by the MS2!ORF 10 polyclonal antibody (Figure 6, lane d). N o DNA was found in the immunoprecipitated pellet when a total protein extract of K. Iuctis WM 37 (killer- strain) (Figure 6, lanes a, b) of
preimmune serum were used (Figure 6. lane c). I t was therefore of interest to find out if the interaction between the O R F 10 protein and the killer plasmids occurs directly or whether it requires other protein(s). To address this question I performed a number of protein/DNA blotting experiments (Bowen et al., 1980). Cell extracts obtained from the killer- strain (K. luctis WM 37) and from the killer ' strain (K. 1ucti.s CBS 2359) were subjected to SDSPAGE and transferred slowly to nitrocellulose by capillarity to allow the renaturation of the proteins. The filter was incubated with the entire "P-labelled K2 DNA. After washing in 200 mM-salt, the filter was exposed to detect the protein bands which retained radiolabelled DNA. A protein band with an apparent molecular weight of 14 kDa was visible only in the killer' protein extract (see Figure 7). The apparent molecular weight of this protein band is the same as that predicted for the O R F 10 gene product. DISCUSSION O R F 10 of the K 2 plasmids has been expressed in E. coli as a fusion protein. A specific antibody has been obtained.
250
M. TOMMASINO
M
a
b
c
d
a
b
kDa 92 69 46 +
+ +
kb
+ 9.4+
23.0
6.6 +
30 +
4.3.)
21 +
+
14 +
2.3 2.0 .I
Figure 6 . lmmunoprecipitation of ORF 10 protein-DNA complexes in protein extracts of K . lucris. Total protein extracts of two different strains, K . /uc/isWM 37 (killer - strain) (lanes a, b) and K . 1ucri.t CBS 2359 (killer-) (lanes c, d) were immunoprecipitatcd with preimmune serum (lanes a. c) or with MS2;ORF 10 fusion protein antiserum (lanes b, d). DNA was prepared from each sample as described in Materials and Methods and analysed by agarox gel electrophoresis stained with ethidium bromide. M: LHindlll marker: the sizes in kb are indicated on the left.
I present several data which suggest that the O R F 10 protein is a DNA binding protein and is associated directly with the killer plasmids. Comparison of the O R F I0 protein sequence with known protein sequences reveals a similarity with several proteins classified as histone-like proteins (for review see Drlica and Rouviere-Yanic, 1987). Three blocks of homology can be detected (Figure 8A). Proteins of this family are small, abundant in the cell and some are basic. The same
Figure 7. DNA binding activity of O R F 10 protein. Total protein extracts of two different strains, K. lacris WM 37 (killer strain) and K . lac/is CBS 2359 (killer' strain) were subjected to SDS-PAGE and transferred to nitrocellulose filter by capillarity to allow the renaturation of the proteins. The filter was probed with the two '*P end-labelled EumHI~Xholfragments of K2. Lane a: 100 pg protein extract of K . luc/is WM 37; lane b: 100 pg protein extract of K. lucris CBS 2359. The arrow indicates the protein band of 14 kDa which is specific to the killer plus strain. The sizes in kDa of the markers (Rainbow, Amersham) are indicated on the left. Some of the other proteins able to bind DNA, present in both extracts (killer minus and plus) can be identified as histones on the basis of their molecular weights (16 18 kDa).
characteristic properties are also shown by the O R F 10 protein. One of the histone-like proteins, the HB protein from Bacillus stearothermophilus, has been crystallized and analysed by X-ray diffraction (Tanaka et al., 1984). The deduced structure of the protein shows that two identical monomers interlock. Each monomer has a long, flexible and conformationally non-ordered arm forming in the dimer complex a concave surface which is responsible for the interaction with DNA. Interestingly, one block of homology, amino acids 48-53 of the O R F 10 protein, corresponds to the primary sequence of the long arm of the HB protein, and occurs in a region which
25 1
KILLER SYSTE,M OF KLUYVEROMYCES LACTIS
A 1
IQLIGFGN
1
1
I11 50 44
50
B 1
MANKQAEKLITAIKK-DYLK
I 11
::
.. . .. . ..
8 3 HM.NK.TQR.NEY.RLLKSVD
1 1 1 1 ):I
2
:::
I:
:
I.
HM.NKRTNR.QDWHRA.DIVA
N-ORFlO C-ORF10 DBP D108
Figure X. Amino acid sequence homology between the ORF 10 protein and some histone-like proteins. Panel A: comparison of the predicted amino acid sequence of ORF 10 with the amino acid sequence of the DNA binding protein H B from Buci/lussubri/is(DBP HB) and DNA binding protein NS2 E. coli(DBPNS2). Panel B: comparison olthe N and C terminus region ofORF 10 protein with the amino acid sequence of DNA binding protein NER from bacteriophage DlO8 (DBP D108). Numbers in parentheses indicate the positions of the amino acids in the protein .sequence
does not show any preference for either a helix or b sheet or b turn. Thus, it is reasonable to hypothesize that the structure of the O R F 10 protein in this region can mimic the flexible arm domain of the histone-like proteins. The N-terminus of O R F 10 contains an I8 amino acid motif which is repeated in the C-terminal region (Figure 8B). Interestingly, the C-terminus repeated region of the O R F 10 product shows a very high homology with the N-terminus of the D N A binding protein NER from bacteriophage D108 (Figure 8B). The O R F 10 protein shows two characteristic features not found in the histone-like proteins: (a) the presence of four cysteine residues, (b) a region which is present in both N and C terminus. These features may contribute to the conformational properties which allow O R F 10 to exist possibly as a monomer.
O R F IOisoneofthemost highly transcribed genes of the killer plasmids, as indicated by the Nothern blot data (Figure 2), and the O R F 10 product can be easily detected in total yeast protein by Western blot (Figure 4). This suggests that this protein is abundant in the cell, and it seems likely therefore that: (a) the O R F 10 protein may have a structural rather than a regulatory function, o r (b) it is a regulatory protein existing in two forms: active and inactive. These hypotheses are under investigation. ACKNOWLEDGEMENTS
I thank Dr C. L. Galeotti for assistance in the beginning of this work, Dr K. Maundrell for continuous useful advice and discussions. I a m also grateful to Drs F. Cavalieri, M. Melli and R. Rappuoli for critical reading of the manuscript, Dr Mario Domenighini for computer assistance. A. Ruspetti
252 and L. Fini for plates and media, G. Corsi for artwork, and C. Mallia and A. Mori for secretarial assist a nce.
M. TOMMASINO
the five subunits of pertussis toxin. In/ecr. Immun. 55, 963-967. Remaut, E.. Stansscns, P. and Fiers. W. (1981). Plasmid vector for high-efficiency expression controllcd by pL promoter ofcoliphage lambda. Gene IS,81 93. REFERENCES Schcidercit, C., Heguy. A. and Roedcr, R . G. (1987). Boucn. B., Steinbcrg, J . . Lammli, C. K. and Wcintraub, Identification and purification of human lymphoidH. (1980). The detcction of DNA-binding proteins by specific octamer-binding protein (OTF-2) that actiprotein blotting. Nicclcic Acids Res. 8, 1-20. vates transcription of an immunoglobulin promoter in Drlica. K . and Rouviere-Yaniv. J . (1987). Histone-like v i m . Cell 51, 783-793. proteins of bacteria. Microhiol. Rev. 51,301- 319. Sor, F. and Fukuhara. H. (1985). Structure of a linear Fcinbcrg, A. P. and Vogelstein. B. (1984). A technique for plasmid of the yeast K1ujwromjw.s Iuctis. Compact radiolabelling DNA restriction nuclease fragments to organization of the killer genome. Curr. Gener. 9, 147- 1 55. high specific activity. Addcndum. Anul. Biocliem. 137, 266267. Stam. J . C., Kwakman, J.. Meijer, M.and Stuitje, A. R. Fricfcld. B. R., Licky. J., Gronostajski. R. M., (1986). Efficient isolation of the linear DNA killer Guggcnheimer. R. A,. Krevolin. M. D.. Nagata, K.. plasmid of K1uj~reromjc'e.s1ucri.s: cvidence for location Hurwitz, J . and Horwitz, Ivl. S. (1986). The in rirro and cxprcssion in the cytoplasm and characterisation of replication of adcnovirus DNA. Curr. Top. Microhiol. their terminally bound proteins. :Vuclik Acid7 Res. 14, Imniunol. 110,221-255. 687 I-- 6884. Fukuhara. H. (1987). The RFI gcnc of thc killer DNA of Stark. M. J . R.. Milcham. A. J.. Romanos. M. A. and yeast may encode a DNA polymerasc. Nucleic Acidy Boyd. A. (1984). Nuclcotide sequences and transcripRrs. IS, 10046. tion analysis of a linear DNA plasmid associated with Gunge, N . and Sakaguchi. K. (1981). lntcrgcnic transfer the killer character of the ycast Kluj-rerotnyci..s 1acti.s. of dcoxyribonuclcic acid killer plasmids pGKll and Nucleic Acids Res. 12,601 1 6030. pG K 12, from Kluy veromj~ce.~ lac'tis in to Saccharotnj~ci~s Stark, M . J . R. and Boyd, A. (1986). The killer toxin of cererisiae by cell fusion. J . Ructeriol. 147, 155 160. K l u j w r o m ~ c e slactis: characterisation of the toxin subGungc. N.. Murata. K. and Sakaguchi, K. (1982).Transunitsand identification ofthegenes which encode them. formation of Saccliarom~cesccwvi.siciewith linear DNA E M B O J. 5, 1995-2002. killer plasmids from K1uj.ri~ronij.c.c.sluctis. J . Bucrctcv%)l. Stark. M. J. R.. B0yd.A. J.. Mileham. A. J . and Romanos. 151,462464. M. A. (1990). The plasmid-encoded killer system of I Iishinuma, F., Nakamura, K., Hirai. K.. Nishizawa. R.. K l u y i w w m j w s lacris: a review. Ycasr 6, I . -29. Gungc, N. and Maeda. T. (1984). Cloning and nucleic Strcbel, K.. Beck. E.. Strohmaier. K. and Shallcr. H. acid sequences of the linear DNA plasmids from yeast. (1986). Characterization of foot-and-mouth disease Nucleic acid.^ Res. 12, 758 1-7597. virus gene products with antiscra against bacterially Klinkert. M. Q..Herrmann. R. and Schaller. H. (1985). synthesized fusion proteins. J . Virol. 57,983 991. Surface protcins of Mjwplusniu l i y p i w n o n i a e ident- Tanaka, 1.. Appelt. K.. Dijk. J., White. S. and Wilson. K. ified from an Eschericlriu c d i expression plasmid (1984). 3-A resolution structure of a protein with library. 1njix.t. Imniun. 49,329-355. histone-likc propcrtics in prokaryotcs. Nuture 310, Lacmmli. U. K. (1970). Cleavage of structural proteins 376 381. during the assembly of the head of bactcriophage T4. Tokunaga, M.. Wada, N. and Hishinuma. F. (1987). R'rrtitre 227, 680-685. Expression and identification ofimmunity determinants Maniatis. T.. Fritsch. E. F. and Sanibrook. J . (1982). of linear DNA killer plasmids pGKll and pGKl2 in Moleculur Cloning. A Labororor? Manual. Cold Spring K l u j w r o t n j c w 1ucti.s. Nuclcic Acids Res. 15, I03 I - 1046. Harbor Laboratory. Cold Spring Harbor. New York. Tommasino. M.. Ricci, S. and Galcotti. C. L. ( 1988). Maundrell. K.. Nurse, P., Schonholzer. F. and Gcnomc organisation of the killer plasmid pGKI2 from Schwcingruber, E. M. (1985).Cloning and charactcrizKluj~veromj*ces lucris. N d e i c Acids Rcis. 16,5863 5878. ation of two genes restoring acid phosphatase activity Wilson, D. W. and Meacock, P. A. (1988).An RNA polyin pho mutants of S,pombe. Gene 39,223-230. merase subunit of novel structure encodcd by a lincar Nicosia. A,. Bartoloni, A,. Perugini. M. and Rappuoli. R. yeast plasmid: evidence for cytoplasmic transcription. ( 1987). Expression and immunological properties of Nu(.lc.ic acid.^ RCS.14, 8097-81 12.
VOL.
7: 245 252 ( I99 1 )
Killer System of KZuyveromyces Zactis: the Open Reading Frame 10 of the pGK12 Plasmid Encodes a Putative DNA Binding Protein A
MASSIMO TOMMASINO Sduro R i w u r c h C’oitrr, c‘iu Fioruitincr
I , 53100 Sicwa. IIUIJ
Rccciwd 28 July 1990: rcvised 21 October 1990
O R F 10 ofthe K 2 plasmid from Klz?,.~,croni?.c.es luctis encodcs a small basic protcin (22.3% lysine). Thc function of its product has bccn invcstigatcd. Western blot analysis, using an antibody against MS2 RNA polymerase!ORF 10 I‘usion protein. reveals a protein band with an apparent molecular weight of 14 kDa. The protein can bind a DNA-Sepharose column. and is eluted by 350 mM-salt. lmmunoprccipitation experiments show that the O R F 10 protein coprccipitates with thc lincar gcnomic DNAs ofthe two killer plasmids ( K I and K2). From Wcstcrn;Southcrn blot data. it is possiblc t o conclude that thc interaction between protein and DNA occurs directly, rather than via other protcin(s). O R F 10 is cosily detected by Western blot and its transcript is one of the most abundant of thc K 2 plasmid. suggesting that this protein may have a structural rather than a regulatory function. This possibility is also suggested by the obscrvcd squcncc homology between thc O R F 10 protcin and the family of histonc-likc proteins, l i i Y WOWS
--
K . Itrciis; killcr system; ORF 10: DNA binding protein
I N T R 0D U C T I O N Some strains of Klu~i~c~ron?ycc.s lactis are able to secrete an olipomcric protein which inhibits the growth of other strains. This so-called killer phenotype is associated with the presence of two linear plasmids in the cell, termed pGKll (8.9 kb) and pGKI2 ( 13.4 kb) (shortened to K 1 and K2. respectively) (for rcview see Stark Pt a/.. 1990). Several lines ofevidence support theidea that the twoplasmidsarelocalized in the cytoplasm (Gunge c’t ul., 1982; Stam et a/., 1986; Wilson and Meacock, 1988). The 5‘ ends of both plasmids are blocked by two proteins with apparent molecular weights of 36 kDa (K2) and 28 kDa ( k l ) (Stam et ul.. 1986). By analogy with other linear rxtrachromosomal genomes, e.g. adenovirus (for review scc Friefeld e t a/.. 1986). these proteins may function as primers in D N A replication. K I has been sequenced and shown to contain four open reading frames (ORFs) (Hishinuma et af.. 1984; Sor and Fukuhara, 1985; Stark ct ul., 1984). Two of them code for the subunits of the toxin ( O R F s 2 and 4) (Stark and Boyd, 1986), while O R F I and O R F 3 code respectively for a plasmidspecific DNA polymerase and a protein involved in the immunity system (Fukuhara. 1987; Tokunaga ot ul.. 1987). 0749 503x Y I 030245 0 8 so5.00 D I Y Y I b! J o h n Wiley & Sons l.rd
Genetic analysis has shown that the presence of K2 in thecell isessential for themaintenanceofboth plasmids, suggesting that the products of the K2 O R F s play fundamental roles in the killer system. K2 has been recently sequenced and shown to contain ten O R F s (Tommasino Pt af., 1988). O R F 2 and O R F 6 seem to code for plasmid-specific D N A and RNA polymerases respectively (Tommasino e t al., 1988: Wilson and Meacock, 1988). It is also possible that the terminal proteins of the two plasmids are encoded by K2. Nothing is known of the possible functions of the proteins encoded by the other ORFs, although thc very high percentage of lysine in O R F 10 (22.3%) suggests that its product may be a D N A binding protein. Here I show that the protein encoded by O R F 10 is a D N A binding protein and interestingly it has homology with a family of proteins classified as histone-like proteins. MATERIALS A N D M E T H O D S
Strains K . lactis CBS 2359 was obtained from the Centraal Bureau voor Schimmelcultures. Bacterial strains used were Escherichia coliJM IOITR (Sup E.
M. TOMMASINO
246
r I R
t
ORF 10 region of K2 GENOME
t
Xhol
EcoRl I
pEMBL9-EX8
t
EcoRl
t
Sau3Al
L
/ Xhol/SalI
Hindlll
r I R
pEX34A-ORF 10
t
Sau3AVBamHI
t
Hindlll
Figure 1. Schematic representation of cloning strategies and final constructs employed. All restriction enzymes used are indicated in the figure. IR is the inverted terminal repeat of K2 plasmid. For details see Materials and Methods.
MSZRNA polymerase fragment. The E. coli strain K-12 A D H l A Dtrp(whichc0ntainsa temperaturesensitive repressor of the lambda promoter pL) was transformed with the construct pEX34/ORF 10. Single colonies containing the plasmid were amplified in liquid culture at 30°C and the pL promotor Expression of MSZjORF 10 fusion protein in E. coli was derepressed by heat shock (2 h a t 42°C) (Nicosia and antiserum preparation et al., 1987). Bacterial proteins were prepared as The fragment EcoRI-XhoI (12556-13335) cover- described(Nicosiaetal., 1987)andanalysed by SDSing the total length of O R F 10 was cloned in the PAGE (Laemmli, 1970).The band corresponding to vector pEMBL9 cut with EcoRI/SalI (Figure 1). the MS2/0RF 10 protein was electroeluted from For expression of O R F 10 the vector pEX34A, a preparative polyacrylamide gel, checked for purity, derivative of pEX29, was used (Klinkert et al., 1985; and used to immunize rabbits (Nicosia et al., 1987). Strebel et al., 1986). This vector contains a fragment of DNA coding for the N-terminus of the gene MS2- Preparation of yeast protein extracts RNA polymerase, positioned downstream of the pL Late log-phase cultures were washed in 150 mMpromoter of phage lambda. An O R F 10 fragment, corresponding to 90% of the coding region, was KC1, 10 mM-Hepes, pH 7.9, 5 mM-MgCl,, 0. I mMexcised from the pEMBL9 construct by digestion EDTA, 0.5 mM-phenylmethylsulfonyl fluoride, with SAU3AI.and Hind111 and cloned in pEX34A 0.5 mhl-benzamidine and 5 mM-mercaptoethanol previously cut with BumHI/HindIII (Figure 1). This (buffer A), resuspended in 1 volume of the same results in an in-frame fusion of ORF 10 with the buffer and broken with glass beads in a beat beater. thi, srl::tnlO, recA, AD(1ac-proAB), [F, tra D365, pro AB, IacIQm ZM151) for routine work, and E. coli K-12 A D H l ADtrp for expressing O R F 10 (Remaut et al., 1981).
247
KILLER SYSTEM OF KLUYVEROMYCES LAC'TIS
After centrifugation (12 000 x g for 20 min), the supernatants were precipitated in ammonium sulphate (0.35 giml). The pellet was resuspended in 50 mM-Tris;HCI, pH 7.9, 20% glycerol. 0.1 mMEDTA, 100 mM-KCI (buffer C), dialysed overnight against the same buffer and frozen in liquid nitrogen. Iniriiimoprec.ipitatiori
K . lac(is was grown to late log phase and protoplasted as described (Gunge and Sakaguchi, 1981). Protoplasts were resuspended in 3 volumes of buffer A and after 10 min the suspension was centrifuged at 12000 x g for 15 min. The supernatant was recovered and precipitated with ammonium sulphate (0.35g!ml). The pellet was resuspended in buffer C and dialysed overnight against the same buffer. 3 mg of protein extracts were diluted in 250 pl of 50 mMTris.HC1, pH 7.4. 100 mwKCI, 5 pgiml Aprotinin and incubated for I h at 25'C with preimmune serum and protein A-Sepharose. After centrifugation, the supernatant was incubated for 3 h at 25'C with immune serum (anti-MS2iORF 10 protein) and protein A-Sepharose. The immunocomplex was centrifuged and washed four times with 1 ml of 50 mM-Tris/HCI, pH 7.4, 100 mM-KCI. The final pellet was resuspended in 1 0 0 ~ 1TE ( I0 mu-Tris:'HCI. pH 8.0, 1 mM EDTA), 1'10 SDS, I0pg tRNA. treated with proteinase K and extracted with phenolichloroform 111. DNA was precipitated in ethanol in sodium acetate 0.3 M, resuspended in 20 pl T E and applied to an agarose gel. In some cases the concentration of KCl in the immunoprecipitation buffer was brought up to 250 mM; no difference was noted (not shown). Other tnrrhod.v Manipulation of DNA was performed by standard procedures (Maniatis et al., 1982). RNA was prepared and analysed as previously described (Maundrell rt ul., 1985: Tommasino rt al., 1988). Double-stranded DNA probes were labelled using random primers (Feinberg and Vogelstein, 1984). Western blotting was performed by standard procedures and Western-Southern blotting was carried out as previously described (Bowen et al., 1980). Protein concentration was determined with BCA Protein Assay Reagent (Pierce).
RESULTS Identifificcrtiono f t h e ORF 10 lranscript Northern blot analysis of the K2 transcripts, using the entire plasmid as a radioactive probe,
ORF 10
Figure 2. Northern blot analysis of mRNA transcribed from K 2 and ORF 10. Total RNA from two different yeast strains was analysed by Northern blotting. Lanes a, c: 10 pg total RNA from K . luciis CBS 2359 (killer strain); lane b: 10 pg total RNA from K . /ucii.s WM 37 (non-killer strain). Theentire K 2 plasmid (lanes a, b) or the 0.39 kb SAU3AI-Xhol fragment (lane c), which cover 90% of ORF 10. were used as radioactive probes. The positions of the 26s and 18s rRNA of yeast are indicated by the arrows.
revealed at least eight bands in the range of about 3.2-0.4 kb, which were consistent with the predicted sizes of the ORFs of K2 (Figure 2, lane a). In order to ic'cntify which of the eight transcripts corresponds to O R F 10. a fragment of K2 (SALi3AIXhol) which covers 90% of the entire O R F 10 was used. Figure 2 shows that the O R F 10 transcript corresponds to the band of approximately 0.4 kb (lane C ) .one of the most abundant transcripts of K2 (lane a). No hybridization was detected when total RNA o f a non-killer strain was examined (Figure 2, lane b). E.rpression of ORF 10 in E. coli
We expressed O R F 10 in E. coli as a fusion protein. A portion of the gene (900/, of the coding region) was cloned in the vector pEX34A (see Materials and Methods and Figure 1 ) and transformed into E. coli. After induction, the total cell
248
Figure 3. Detection of MS2jORF 10 protein synthesized in E. c d i . Total cellular protein extracts from E. coli transformed with pEX34A or pEX34A;ORF 10 were subjected to SDS-PAGE, stained with Coomassic blue or electroblotted and screened with a monoclonal antibody against the N-terminus ofthe MS2 RNApolymerase. Panel A: Coomassie staining. Lane a: 100 pg of total protein extract of [pex34A] E. coli; lane b: 100 pg of total protein extract of[pEX34A;ORF lo] E. coli.Panel B: Western blot. Lane a: IOOpg of total protein extract of [pEX34A] E. coli; lane b: 100 pg of total protein extract of (pEX34a:ORF 101 E. coli.The sizes in kDa of the markers (Rainbow. Amersham) are indicated on the left.
lysate was analysed by SDS-PAGE. Coomassie blue staining revealed a major band with an apparent molecular weight corresponding to that of the fusion product (Figure 3. panel A. lane b). This protein band was able to react. in a Western blot, with a monoclonal antibody raised against the N-terminus of the MS2-RNA polymerase (Figure 3, panel B, lane b). The fusion protein was purified as described in Materials and Methods and used to immunize rabbits. To check the specificity of the polyclonal antibody for O R F 10. I performed Western blot analysis with total yeast protein extracts from killer and nonkiller strains. The antiserum reacted with a single
M. TOMMASINO
Figure 4. Detection of the O R F 10 products in K. lmri.~.Total protein extracts (100 pg), prepared as described in Materials and Methods, of K. lacris WM 37 (killer strain) (lanes a. c) and K. lacris CBS 2359 (killer' strain) (lanes b. d ) were subjected to SDS-PAGE, electroblotted and screened with two different antibodies. Lanes a. h: monoclonal MS2 antibody; lanes c. d: polyclonal MS2:ORF 10 antibody. The size in kDa of the markers (Rainbow. Amersham) are indicated on the left.
protein band of apparent molecular weight 14 kDa, present in the protein extract of the killer+ strain ( K . lucfisCBS 2359) (Figure 4. lane d), while no reaction was detected in the extract of the killer- strain ( K . lucfis WM 37) (Figure 4, lane c). As a control, both extracts were probed with the MS2 monoclonal antibody and no signal was detected in either strain (Figure 4, lanes a. b). ORF loprotein b i n h to Sepharose-DNA column The highly basic character of the O R F 10 product (22.3% lysine + 2.9% arginine) suggests that this protein could be a DNA binding protein. In order to test this possibility, the total protein extract of K . lactis CBS2359 (see Materials and Methods)
249
KILLER SYSTEM OF KLL'YVEROMYCES LACTIS
d 0
FRACTION NUMBER Figure 5. Binding of the yeast O R F 10 protein to DNA-Sepharose. About 90 mg o f total cellular proteins of K. bcris CBS 2359. prepared as described in Materials and Methods. were applied to a DNA-Sepharose column (7 x 2 cm) and eluted by a linear NaCI gradient (04.8 M). An aliquot of each fraction was subjected to SDS-PAGE, electroblotted and screened with polyclonal MS2; ORF 10 antibody. The fractions containing O R F 10 protein are shown in the inset figure.
was fractionated on a Sepharose-DNA column (prepared as described by Scheidereit er ul., 1987). The bound proteins were eluted from the column by a linear gradient of NaCl(0-800 mM). All fractions were analysed by SDS-PAGE and Western blot using the MS2/0RF 10 antibody. More than 70% of the O R F 10 product was bound to the column and it was eluted at an ionic strength of about 350 mM, demonstrating that it binds to DNA with a good affinity (Figure 5). ORF 10 protein i s ussociuted with killer plusmid D:VA
Preliminary cell fractionation experiments showed that the O R F 10 protein and the two killer plasmids were present in the same subcellular fraction, suggesting a possible association of this protein to K I and K2. To test this, I have immunoprecipitated the O R F 10 protein with MS2iORF 10 polyclonal antibody. The pellet was treated with proteinase K. phenol extracted and its DNA content analysed by agarose gel electrophoresis. Both killer plasmids were immunoprecipitated by the MS2!ORF 10 polyclonal antibody (Figure 6, lane d). N o DNA was found in the immunoprecipitated pellet when a total protein extract of K. Iuctis WM 37 (killer- strain) (Figure 6, lanes a, b) of
preimmune serum were used (Figure 6. lane c). I t was therefore of interest to find out if the interaction between the O R F 10 protein and the killer plasmids occurs directly or whether it requires other protein(s). To address this question I performed a number of protein/DNA blotting experiments (Bowen et al., 1980). Cell extracts obtained from the killer- strain (K. luctis WM 37) and from the killer ' strain (K. 1ucti.s CBS 2359) were subjected to SDSPAGE and transferred slowly to nitrocellulose by capillarity to allow the renaturation of the proteins. The filter was incubated with the entire "P-labelled K2 DNA. After washing in 200 mM-salt, the filter was exposed to detect the protein bands which retained radiolabelled DNA. A protein band with an apparent molecular weight of 14 kDa was visible only in the killer' protein extract (see Figure 7). The apparent molecular weight of this protein band is the same as that predicted for the O R F 10 gene product. DISCUSSION O R F 10 of the K 2 plasmids has been expressed in E. coli as a fusion protein. A specific antibody has been obtained.
250
M. TOMMASINO
M
a
b
c
d
a
b
kDa 92 69 46 +
+ +
kb
+ 9.4+
23.0
6.6 +
30 +
4.3.)
21 +
+
14 +
2.3 2.0 .I
Figure 6 . lmmunoprecipitation of ORF 10 protein-DNA complexes in protein extracts of K . lucris. Total protein extracts of two different strains, K . /uc/isWM 37 (killer - strain) (lanes a, b) and K . 1ucri.t CBS 2359 (killer-) (lanes c, d) were immunoprecipitatcd with preimmune serum (lanes a. c) or with MS2;ORF 10 fusion protein antiserum (lanes b, d). DNA was prepared from each sample as described in Materials and Methods and analysed by agarox gel electrophoresis stained with ethidium bromide. M: LHindlll marker: the sizes in kb are indicated on the left.
I present several data which suggest that the O R F 10 protein is a DNA binding protein and is associated directly with the killer plasmids. Comparison of the O R F I0 protein sequence with known protein sequences reveals a similarity with several proteins classified as histone-like proteins (for review see Drlica and Rouviere-Yanic, 1987). Three blocks of homology can be detected (Figure 8A). Proteins of this family are small, abundant in the cell and some are basic. The same
Figure 7. DNA binding activity of O R F 10 protein. Total protein extracts of two different strains, K. lacris WM 37 (killer strain) and K . lac/is CBS 2359 (killer' strain) were subjected to SDS-PAGE and transferred to nitrocellulose filter by capillarity to allow the renaturation of the proteins. The filter was probed with the two '*P end-labelled EumHI~Xholfragments of K2. Lane a: 100 pg protein extract of K . luc/is WM 37; lane b: 100 pg protein extract of K. lucris CBS 2359. The arrow indicates the protein band of 14 kDa which is specific to the killer plus strain. The sizes in kDa of the markers (Rainbow, Amersham) are indicated on the left. Some of the other proteins able to bind DNA, present in both extracts (killer minus and plus) can be identified as histones on the basis of their molecular weights (16 18 kDa).
characteristic properties are also shown by the O R F 10 protein. One of the histone-like proteins, the HB protein from Bacillus stearothermophilus, has been crystallized and analysed by X-ray diffraction (Tanaka et al., 1984). The deduced structure of the protein shows that two identical monomers interlock. Each monomer has a long, flexible and conformationally non-ordered arm forming in the dimer complex a concave surface which is responsible for the interaction with DNA. Interestingly, one block of homology, amino acids 48-53 of the O R F 10 protein, corresponds to the primary sequence of the long arm of the HB protein, and occurs in a region which
25 1
KILLER SYSTE,M OF KLUYVEROMYCES LACTIS
A 1
IQLIGFGN
1
1
I11 50 44
50
B 1
MANKQAEKLITAIKK-DYLK
I 11
::
.. . .. . ..
8 3 HM.NK.TQR.NEY.RLLKSVD
1 1 1 1 ):I
2
:::
I:
:
I.
HM.NKRTNR.QDWHRA.DIVA
N-ORFlO C-ORF10 DBP D108
Figure X. Amino acid sequence homology between the ORF 10 protein and some histone-like proteins. Panel A: comparison of the predicted amino acid sequence of ORF 10 with the amino acid sequence of the DNA binding protein H B from Buci/lussubri/is(DBP HB) and DNA binding protein NS2 E. coli(DBPNS2). Panel B: comparison olthe N and C terminus region ofORF 10 protein with the amino acid sequence of DNA binding protein NER from bacteriophage DlO8 (DBP D108). Numbers in parentheses indicate the positions of the amino acids in the protein .sequence
does not show any preference for either a helix or b sheet or b turn. Thus, it is reasonable to hypothesize that the structure of the O R F 10 protein in this region can mimic the flexible arm domain of the histone-like proteins. The N-terminus of O R F 10 contains an I8 amino acid motif which is repeated in the C-terminal region (Figure 8B). Interestingly, the C-terminus repeated region of the O R F 10 product shows a very high homology with the N-terminus of the D N A binding protein NER from bacteriophage D108 (Figure 8B). The O R F 10 protein shows two characteristic features not found in the histone-like proteins: (a) the presence of four cysteine residues, (b) a region which is present in both N and C terminus. These features may contribute to the conformational properties which allow O R F 10 to exist possibly as a monomer.
O R F IOisoneofthemost highly transcribed genes of the killer plasmids, as indicated by the Nothern blot data (Figure 2), and the O R F 10 product can be easily detected in total yeast protein by Western blot (Figure 4). This suggests that this protein is abundant in the cell, and it seems likely therefore that: (a) the O R F 10 protein may have a structural rather than a regulatory function, o r (b) it is a regulatory protein existing in two forms: active and inactive. These hypotheses are under investigation. ACKNOWLEDGEMENTS
I thank Dr C. L. Galeotti for assistance in the beginning of this work, Dr K. Maundrell for continuous useful advice and discussions. I a m also grateful to Drs F. Cavalieri, M. Melli and R. Rappuoli for critical reading of the manuscript, Dr Mario Domenighini for computer assistance. A. Ruspetti
252 and L. Fini for plates and media, G. Corsi for artwork, and C. Mallia and A. Mori for secretarial assist a nce.
M. TOMMASINO
the five subunits of pertussis toxin. In/ecr. Immun. 55, 963-967. Remaut, E.. Stansscns, P. and Fiers. W. (1981). Plasmid vector for high-efficiency expression controllcd by pL promoter ofcoliphage lambda. Gene IS,81 93. REFERENCES Schcidercit, C., Heguy. A. and Roedcr, R . G. (1987). Boucn. B., Steinbcrg, J . . Lammli, C. K. and Wcintraub, Identification and purification of human lymphoidH. (1980). The detcction of DNA-binding proteins by specific octamer-binding protein (OTF-2) that actiprotein blotting. Nicclcic Acids Res. 8, 1-20. vates transcription of an immunoglobulin promoter in Drlica. K . and Rouviere-Yaniv. J . (1987). Histone-like v i m . Cell 51, 783-793. proteins of bacteria. Microhiol. Rev. 51,301- 319. Sor, F. and Fukuhara. H. (1985). Structure of a linear Fcinbcrg, A. P. and Vogelstein. B. (1984). A technique for plasmid of the yeast K1ujwromjw.s Iuctis. Compact radiolabelling DNA restriction nuclease fragments to organization of the killer genome. Curr. Gener. 9, 147- 1 55. high specific activity. Addcndum. Anul. Biocliem. 137, 266267. Stam. J . C., Kwakman, J.. Meijer, M.and Stuitje, A. R. Fricfcld. B. R., Licky. J., Gronostajski. R. M., (1986). Efficient isolation of the linear DNA killer Guggcnheimer. R. A,. Krevolin. M. D.. Nagata, K.. plasmid of K1uj~reromjc'e.s1ucri.s: cvidence for location Hurwitz, J . and Horwitz, Ivl. S. (1986). The in rirro and cxprcssion in the cytoplasm and characterisation of replication of adcnovirus DNA. Curr. Top. Microhiol. their terminally bound proteins. :Vuclik Acid7 Res. 14, Imniunol. 110,221-255. 687 I-- 6884. Fukuhara. H. (1987). The RFI gcnc of thc killer DNA of Stark. M. J . R.. Milcham. A. J.. Romanos. M. A. and yeast may encode a DNA polymerasc. Nucleic Acidy Boyd. A. (1984). Nuclcotide sequences and transcripRrs. IS, 10046. tion analysis of a linear DNA plasmid associated with Gunge, N . and Sakaguchi. K. (1981). lntcrgcnic transfer the killer character of the ycast Kluj-rerotnyci..s 1acti.s. of dcoxyribonuclcic acid killer plasmids pGKll and Nucleic Acids Res. 12,601 1 6030. pG K 12, from Kluy veromj~ce.~ lac'tis in to Saccharotnj~ci~s Stark, M . J . R. and Boyd, A. (1986). The killer toxin of cererisiae by cell fusion. J . Ructeriol. 147, 155 160. K l u j w r o m ~ c e slactis: characterisation of the toxin subGungc. N.. Murata. K. and Sakaguchi, K. (1982).Transunitsand identification ofthegenes which encode them. formation of Saccliarom~cesccwvi.siciewith linear DNA E M B O J. 5, 1995-2002. killer plasmids from K1uj.ri~ronij.c.c.sluctis. J . Bucrctcv%)l. Stark. M. J. R.. B0yd.A. J.. Mileham. A. J . and Romanos. 151,462464. M. A. (1990). The plasmid-encoded killer system of I Iishinuma, F., Nakamura, K., Hirai. K.. Nishizawa. R.. K l u y i w w m j w s lacris: a review. Ycasr 6, I . -29. Gungc, N. and Maeda. T. (1984). Cloning and nucleic Strcbel, K.. Beck. E.. Strohmaier. K. and Shallcr. H. acid sequences of the linear DNA plasmids from yeast. (1986). Characterization of foot-and-mouth disease Nucleic acid.^ Res. 12, 758 1-7597. virus gene products with antiscra against bacterially Klinkert. M. Q..Herrmann. R. and Schaller. H. (1985). synthesized fusion proteins. J . Virol. 57,983 991. Surface protcins of Mjwplusniu l i y p i w n o n i a e ident- Tanaka, 1.. Appelt. K.. Dijk. J., White. S. and Wilson. K. ified from an Eschericlriu c d i expression plasmid (1984). 3-A resolution structure of a protein with library. 1njix.t. Imniun. 49,329-355. histone-likc propcrtics in prokaryotcs. Nuture 310, Lacmmli. U. K. (1970). Cleavage of structural proteins 376 381. during the assembly of the head of bactcriophage T4. Tokunaga, M.. Wada, N. and Hishinuma. F. (1987). R'rrtitre 227, 680-685. Expression and identification ofimmunity determinants Maniatis. T.. Fritsch. E. F. and Sanibrook. J . (1982). of linear DNA killer plasmids pGKll and pGKl2 in Moleculur Cloning. A Labororor? Manual. Cold Spring K l u j w r o t n j c w 1ucti.s. Nuclcic Acids Res. 15, I03 I - 1046. Harbor Laboratory. Cold Spring Harbor. New York. Tommasino. M.. Ricci, S. and Galcotti. C. L. ( 1988). Maundrell. K.. Nurse, P., Schonholzer. F. and Gcnomc organisation of the killer plasmid pGKI2 from Schwcingruber, E. M. (1985).Cloning and charactcrizKluj~veromj*ces lucris. N d e i c Acids Rcis. 16,5863 5878. ation of two genes restoring acid phosphatase activity Wilson, D. W. and Meacock, P. A. (1988).An RNA polyin pho mutants of S,pombe. Gene 39,223-230. merase subunit of novel structure encodcd by a lincar Nicosia. A,. Bartoloni, A,. Perugini. M. and Rappuoli. R. yeast plasmid: evidence for cytoplasmic transcription. ( 1987). Expression and immunological properties of Nu(.lc.ic acid.^ RCS.14, 8097-81 12.
