MGG

MolGenGenet (1988) 213:421-424

© Springer-Verlag 1988

Repair of double-strand breaks in plasmid DNA in the yeast Saccharomyces cerevisiae Joseph R. Perera 1, Alexander V. Glasunov 2, V adim M. Glaser 1, and Alia V. Boreiko 2

1 Department of Genetics and Selection, Biology Division, Moscow State University, Moscow 119899, USSR 2 Joint Institute for Nuclear Research, Moscow 101000, box 79, USSR

We studied the repair of double-strand breaks (DSB) in plasmid D N A introduced into haploid cells of the yeast Saceharomyces cerevisiae. The efficiency of repair was estimated from the frequency of transformation of the cells by an autonomously replicated linearized plasmid. The frequency of "lithium" transformation of Rad + cells was increased greatly (by I order of magnitude and more) compared with that for circular D N A if the plasmid was initially linearized at the XhoI site within the L YS2 gene. This effect is due to recombinational repair of the plasmid DNA. Mutations rad52, rad53, rad54 and rad57 suppress the repair of DSB in plasmid DNA. The kinetics of DSB repair in plasmid D N A are biphasic: the first phase is completed within 1 h and the second within 14-18 h of incubating cells on selective medium. Summary.

Key words: Double-strand break - Plasmid D N A repair - Yeast transformation

Orr-Weaver etal. (1981) and Orr-Weaver and Szostak (1983) have proposed the use of genetic transformation of yeast cells by plasmid linearized with a restriction endonuclease in order to study double-strand break (DSB) and double-strand gap (DSG) repair in DNA. Repair of the plasmid in a marked proportion of the transformed cells proceeds by recombination with homologous chromosomal D N A sequences and results in complete restoration of the plasmid genetic material. However, the question of whether the repair pathways of chromosomal D N A DSB and plasmid D N A DSB are the same remains open. We performed a more detailed analysis of the molecular mechanisms and genetic control of DSB and DSG repair in plasmid D N A in yeast. The following haploid strains of the yeast Saccharomyces cerevisiae were used: DC5 a leu2-3, 112 his3; 2G-D252 a leu2-3, 112 lys2-29 his3; 2B-D241 a leu2-3, 1121ys2-25his3. The strains were kindly supplied by D.A. Gordenin (Leningrad State University, USSR). DC52 c~rad52-1 leu2-3, 112 arg4 ilv2 his3, DC53 ~ rad53-1 leu2-3, 112 his3, DC54 ~ rad54-3 leu2-3, 112 his3 trp2, and DC57 a rad57-1 leu2-3, 112 his3 were obtained in this study by mating strain DC5 with haploid strains of radiation-sensitive mutants from the collection of R.K. Mortimer (Berkeley, Calif., USA). The bifunctional yeast vector pLL12 Offprint requests to." J.R. Perera

(Fig. 1) was constructed by K. Sasnauskas (All-Union Research Institute of Applied Enzymology, Vilnus, USSR). Plasmid pLLI2-BX was derived from this vector by M.V. Trofimova (Leningrad State University) by deleting 391 bp in the region of the XhoI site of the L YS2 gene. In this construction the XhoI site was retained and the BamHI site was lost. Thus, pLLI2-BX has one and pLL12 has two BamHI sites. Table 1 shows the results obtained in transformation experiments with yeast cells of various genotypes when either circular or linearized plasmids pLL12 and pLLI2-BX were used. Yeast transformation was performed by the "lithium" method (Ito et al. 1983). Five micrograms of plasmid D N A and 50 ~tg of carrier D N A (sonicated calf thymus DNA) were used for each transformation. For each strain transformation by circular or linearized plasmid was performed using the same competent culture. In Table 1 the ratio (L) of the transformation frequency for the circular plasmid to the corresponding value for the linearized molecule is presented. We found that plasmid pLL12 linearized at the unique XhoI site of the L YS2 gene transforms DC5 cells much more effectively than does the circular form. Similarly plasmid pLLI2-BX linearized at the XhoI

EcoR!

£

J

BamttI Fig. 1. Map ofplasmid pLL12.1 [ Chromosomal DNA ofSaccharomyces cerevisiae, ~ fragment of 2 ~tm DNA (B-form), I--~_--_-_-_-] PstI--EcoRI fragment of pBR322

422 Table 1. Efficiency of transformation of yeast cells of various genotypes by circular and linearized plasmids Strain

Plasmid

Site of linearization

La

Proportion of unstable Leu + cells among Leu + transformants circular

linearized

DC5

pLLI2 pLLI2-BX pLL12-BX

XhoI XhoI BamHI

0.090 -}-0.004 b 0.092 + 0.017

20/20 -

44/55 -

1.0 c

-

-

2B-D241 (lys2-25)

pLL12

XhoI

0.081 _+0.019

24/24

35/48

2G-D252 (Iys2-29)

pLL12

XhoI

0.970 _+0.09

24/24

23/24

DC52 (rad52)

pLL12

XhoI

1.0 c

-

-

DC53 (rad53)

pLLI 2

XhoI

0.730 ± 0.04

23/24

21/24

DC54 (rad54) 23° C 34° C DC57 d (rad57)

pLL12 pLL12 pLL12

XhoI XhoI

0.140___0.02 1.0 ° 1.0 c

22/24 -

18/24 22/24 -

XhoI

a Ratio of transformation (to Leu + phenotype) frequency for the circular plasmid to the corresponding value for the linearized molecule b Error of the mean for L averaged over 3-5 experiments ° Transformation frequencies for circular and linearized plasmids are not significantly different d Transformation mixture was incubated on selective agar at 23° C

site is, by an order of magnitude, more effective than circular D N A in transformation of DC5 and 2B-D241 cells. It should be pointed out that the latter strain carries a deletion, 1ys2-25, of about 2.1 kb located at a distance of about 1 kb from the X h o I site of the L Y S 2 gene (Chernoff et al. 1984). A similar effect has been observed before for lithium transformation (White and Sedgwick 1985). These authors have demonstrated that the effect is R A D 5 2 dependent. However the underlying mechanisms remain obscure. It should be pointed out here that the absolute values for transformation frequency varied greatly from one experiment to another. Circular plasmid pLL12 D N A and D N A linearized at the X h o I site transformed DC5 cells with frequencies ranging from 7 x 101 to 3 x 102 and from 5 x 10 z to 3 x 103 transformants per 10 gg D N A per 10 v cells, respectively. At the same time the corresponding values for transformation frequencies for pLL12-BX were in the range 2 x 101-8 x 101 and 1 x 10z-5 x 102. Nevertheless the value of L showed much lower variation. Table 1 presents L values averaged over several experiments; they are similar for the DC5 and 2B-D241 strains, with transformation by pLL12 and p L L I 2-BX being practically identical. If strain 2G-D252 (lys2-29) deleted for the L Y S 2 gene (Chernoff et al. 1986) is used as a recipient, the transformation frequencies obtained with circular pLL12 and p L L I 2 linearized at the J(hoI site practically coincide (L~-1). The same is true for pLLI2-BX. The frequencies of transformation of DC5 cells by circular pLL12-BX and B a m H I - l i n e a r ized pLL12-BX (that is linearized in the region bearing no homology with chromosomal D N A ) do not differ essentially one from another. It is thus apparent that the increased transformation frequency resulting from plasmid linearization depends on the presence in the recipient cell of chromosomal D N A with homology to plasmid D N A in the DSB region. The transformation frequencies obtained with circular p L L I 2 D N A and the same D N A linearized at the X h o I site were similar for the strains carrying mutations which block DSB repair in chromosomal D N A , that is in rad52,

(Resnick and Martin 1976; Budd and Mortimer 1982; Glaser et al. 1985). The experiments with strain DC54 deserve special discussion. Mutation rad54-3 is temperature sensitive (it is expressed at the restrictive temperature 34°C whereas at 2 3 ° C this strain possesses the wild-type phenotype) (Budd and Mortimer 1982). It follows from Table 1 that when the transformation mixture is incubated on solid selective medium at the permissive temperature, the transformation efficiency obtained with p L L l 2 linearized at the X h o I site increases greatly as compared with that of the circular plasmid. This phenomenon is absent at 34 ° C. Another important feature of our data is that generally more than 7 0 % - 8 0 % of clones resulting from transformation by the linearized plasmid are unstable for the Leu + phenotype (see Table 1). This means that the bulk of plasmid recovered during the transformation is present in the cells in an autonomous state, The transformants were checked for stability by cloning on complete Y E P D medium; thereafter 20 subclones of each transformant were replica plated on selective medium without lencine. We analysed the plasmids isolated from 19 DC5 clones transformed by pLL12-BX linearized at the X h o I site. Plasraids were isolated from yeast cells by the method of Crabeel et al. (1981). Thereafter Escherichia coli HB101 cells were transformed by these plasmid probes (Maniatis et al. 1982). Plasmid D N A was isolated from E. coli cells by the alkaline method (Birnboim 1983), treated with B a m H I endonuclease and analysed by electrophoresis in a 0.8% agarose gel. It was found that 18 out of 19 Leu + transformants of DC5 cells contained the plasmid with two B a m H I sites (as in p L L I 2 , and not one as in the original pLLI2-BX). Only a recombinational mechanism would make it possible to restore the second B a m H I site to the plasmid D N A . Thus our data indicate that the observed increase in transformation efficiency resulting from plasmid linearization is associated with the recovery of plasmid D N A with DSB or D S G by recombinational mechanisms. The majority of R a d + cells transformed by linearized plasmid conrad53, rad54, r a d 5 7 m u t a n t s

423 Table 2. Efficiency of transformation of yeast cells by plasmid pLL12-BX linearized at the XhoI or BamHI sites and treated by $1 nuclease Recipient strain

DC5

2G-D252 (1ys2-29)

DC5

Plasmid

Site of linearization

$1 treatment

Circular Linearized Linearized

-

-

XhoI XhoI

+

Circular Linearized Linearized

-

-

XhoI XhoI

+

Circular Linearized

-

+

BamHI

Number of Leu + transformants per 10 lag DNA p e r i 0 7 cells

20 10 8

6

4

53 _+10 a (5.1 _+0.9) x 102 (3.9 _+0.8) x 102 (1.7 _+0.3) x 102 (2.2_+0.4) x l02 0b

(1.5_+0.3) x 102 < 0.6

a The error was calculated as the sum of the standard deviation normalized per mean number of transformants per plate and the relative error for the determination of cell concentration and plasmid DNA concentration b NO transformants were observed

0

I

I

~

1

2

3

g

I

j,

18

g

L

72

Fig. 2. The efficiency of transformation of DC54 (rad54-3) by plasmid pLLI2 linearized at the XhoI site with time of incubation of cells at the permissive temperature 23 ° C. Abscissa, time in hours; ordinate, relative transformation efficiency (relative units, logarithmic scale)

Table 3. Efficiency of transformation of DC5 cells by plasmid pLLI2 linearized at the XhoI site and treated with Bal31 nuclease Plasmid

tained plasmid recovered by recombinational interaction with the c h r o m o s o m a l D N A . In subsequent experiments plasmid pLL12-BX linearized at the X h o I site was treated with $1 nuclease in order to exclude the possibility of cohesive end ligation (Table 2) and used to transform DC5 and 2G-D252 cells. In this case the efficiency o f DC5 cell transformation was considerably higher than the corresponding value for the circular plasmid and close to that observed with the plasmid linearized at the X h o I site but not treated with $1 nuclease. In contrast, in strain 2G-D252 with the lys2-29 deletion extending over the L Y S 2 gene, no transformants were observed with SItreated plasmid. A similar drastic decrease in the transform a t i o n frequency resulting from S1 nuclease treatment o f B a m H I - l i n e a r i z e d p L L 1 2 - B X was also observed for DC5 cells. These d a t a confirm our conclusion that a plasmid linearized at a site having h o m o l o g y with c h r o m o s o m a l D N A is repaired p r e d o m i n a n t l y by a recombinational event, whereas when there is no h o m o l o g y with c h r o m o s o m a l D N A in the DSB region the recircularization o f the plasmid is p r o d u c e d by cohesive end ligation (Orr-Weaver and Szostak ~983) or other mechanisms (Kunes et al. 1985). A n o t h e r i m p o r t a n t point should be mentioned: these results show that the efficiency o f recombinational repair o f a " b l u n t " and a " r a g g e d " DSB is virtually identical. Studies on the kinetics o f DSB repair in plasmid D N A were o f considerable interest. F o r this p u r p o s e we used mutant DC54 carrying the temperature-sensitive m u t a t i o n rad54-3. Plasmid p L L 1 2 - B X linearized at the X h o I site was used for the transformation o f DC54 cells. The transformation mixture was plated on selective agar in petri dishes and incubated at the permissive temperature (23 ° C) for various periods o f time; thereafter the plates were transferred to non-permissive conditions (34 ° C). Figure 2 shows the dependence o f the t r a n s f o r m a t i o n frequency to Leu + (in relative units) on the length o f time the cells were incubated at 23 ° C. It can be seen that the frequency o f transform a t i o n of the cells by linearized p L L 1 2 - B X increased 2 -

I

Circular Linearized Linearized and treated with Ba131

Size of double strand gap (DSG) in kb

Number of Leu + transformants per 10 lag DNA per 10 v cells

-

(1.8_+0.3)" x 102 (2.3__0.4) x 103

1.6___0.3 2.3_+0.3 2.9 _+0.4 4.0 _+0.7

(1.8_+0.3) x 103 (1.5_+0.3) x 103 (7.7 _+0.9) x 102 (5.4 _+0.7) x 10 z

-

" See footnote a to Table 2

3 times during 1 h o f incubation; this was followed by a small plateau and subsequently the transformation frequency increased again, attaining a value practically identical to that for cells grown under permissive conditions after 18 h. The transformation frequencies o f both DC5 ( R a d +) and DC54 cells by circular plasmid D N A and the transform a t i o n frequency o f DC5 by linearized plasmid D N A were constant in these experiments. Thus, the kinetics o f plasmid D N A DSB repair are biphasic, i.e. they have two c o m p o n ents: the fast phase is completed by at least I h and the second phase by 18 h o f cell incubation at 23 ° C. We next studied the effect of D S G size on the efficiency o f its repair. Plasmid p L L I 2 linearized at the X h o I site was treated for various periods o f time by Ba131 nuclease. The size o f the resulting D S G was estimated by comparing the electrophoretic mobility o f nuclease-treated plasmid in a 0.6% agarose gel with the corresponding value for m a r k e r fragments ( H i n d I I I digest o f 2 D N A ) . Plasmids with D S G of various sizes were used to transform DC5 cells. It was found (Table 3) that the efficiency o f recombinational repair o f plasmids carrying D S G o f up to 2.0-2.5 kb is approximately similar to that observed for the plasmid linearized at the X h o I site (without DSG). Indeed, since Bal31 treatment excludes cohesive end ligation, the transformation frequency is correlated with the efficiency o f recombinational repair o f plasmid D N A .

424

Acknowledgements. The authors are grateful to D.A. Gordenin for kindly providing yeast strains and to K. Sasnauskas and M.V. Trofimova for providing the plasmids. References Birnboim HC (1983) A rapid alkaline method for isolation of plasmid DNA. Methods Enzymol 100:243-255 Budd M, Mortimer RK (1982) Repair of double-strand breaks in temperature conditional radiation-sensitive mutant of Saccharomyces cerevisiae. Murat Res 103 : 19-29 Chernoff YO, Kidgotko OV, Demberelijn O, Luchnikova IL, Soldatov SP, Glaser VM, Gordenin DA (1984) Mitotic intragenic recombination in the yeast Saccharomyces cerevisiae: markereffects on conversion and reciprocity of recombination. Curr Genet 9:31-37 Chernoff YO, Kidgotko OV, Demberelijn O, Gordenin DA (1986) Mapping of mutations in L YS2 gene of yeast. In: Inge-Vechtomov SG (ed) Issledovanija po genetike (Researches in genetics; in Russian) vol. 10. Leningrad, pp 92-104 Crabeel M, Messenguy F, Lacroute F, Glansdorff N (1981) Cloning arg3, the gene for ornithine carbamoyl transferase from Saceharomyces cerevisiae; expression in Escheriehia coli requires secondary mutations; production of plasmid/Mactamase in yeast. Proc Natl Acad Sci USA 78 : 5026-5030 Glaser VM, Samadashwili MP, Vishnevetskaya OV, Soldatov SP, Shestakov SV (1985) On the relationship of DNA double-

strand break repair to recombination and chromosomal aberrations in yeast. Mutat Res 147 : 296-297 Ito H, Fukuda Y, Murata K, Kimura A (1983) Transformation of intact yeast cells treated with alkali cations. J Bacteriol 153:163-168 Kunes S, Botstein D, Fox MS (1985) Transformation of yeast with linearized plasmid DNA. Formation of inverted dimer and recombinant plasmid products. J Mol Biol 184:375-387 Maniatis T, Fritsch EF, Sambrook J (1982) Molecular cloning. A laboratory manual. Cold Spring Harbor Laboratory, New York Orr-Weaver TL, Szostak JW (1983) Yeast recombination: the association between double-strand gap repair and crossing-over. Proc Natl Acad Sei USA 80:4417-4421 Orr-Weaver TL, Szostak JW, Rothstein RJ (1981) Yeast transformation: a model system for the study of recombination. Proe Natl Acad Sci USA 78:6354-6358 Resnick MA, Martin P (1976) The repair of double-strand breaks in the nuclear DNA of Saccharomyces cerevisiae and its genetic control. Mol Gen Genet 143:119-129 White CI, Sedgwick SG (1985) The use of plasmid DNA to probe DNA repair function in the yeast Saccharomyces cerevisiae. Mol Gen Genet 201:99 106 Communicated by B.J. Kilbey Received October 29, 1987 / March 3, 1988