Current Genetics

Curr Genet (1990) i8:1-5

9 Springer-Verlag 1990

Original articles Genetic control of plasmid DNA doublelstrand gap repair in yeast,

Saccharomyces cerevisiae V.M. Glaser

1

A.V. Glasunov 2, G.G. Tevzadze l, j.R. Perera i, and S.V. Shestakov i

1 Department of Genetics, BiologyDivision, Moscow State University, Moscow, 119899, USSR 2 Joint Institute for Nuclear Research, Moscow, 101000, Box 79, USSR Received November 2, 1989

Summary. The repair of double-strand gaps (DSGs) in the plasmid DNA of radiosensitive mutants of Saccharomyees cerevisiae has been analyzed. The proportion of repair events that resulted in complete plasmid DNA DSG recovery was close to 100% in Rad + cells. Mutation rad55 does not influence the efficiency and preciseness of DSG repair. The mutant rad57, which is capable of recombinational DNA DSB repair, resulted in no DSG recovery. Mutation rad53 substantially inhibits the efficiency of DSG repair but does not influence the precision of repair. Plasmid DNA DSG repair is completely blocked by mutations rad50 and rad54. Key words: Double-strand gap Yeast transformation

Plasmid DNA repair

Introduction Model systems, based on recombinant plasmid vectors, were recently used for studying the mechanisms of DNA double-strand break (DSB) repair (Orr-Weaver etal. 1981; Orr-Weaver and Szostak 1983; White and Sedgwick 1985; Perera et al. 1988). Plasmid DNA with DSBs is repaired effectively in yeast cells through recombination with homologous chromosomal DNA. The mechanisms of plasmid and chromosomal DNA repair are similar. Both processes are under the same genetic control. Furthermore, a double-strand gap (DSG - a deletion in the DSB region) in a plasmid is repaired through the recombination with chromosomal DNA but only partial DSG repair events are registered (Orr-Weaver and Szostak 1983). In this paper we shall define a repair as precise, or complete, if a deleted region of a plasmid is completely recovered in homologous DNA whether chromosomal or not. Plasmids carrying DSGs of up to 2.5 kb are repaired with nearly the same efficiency as plasmids with a DSG Offprint requests to." V.M. Glaser

introduced by a restriction endonuclease (Perera et al. 1988). However, what proportion of DSGs are repaired completely, and what genes are responsible for DSG repair precision, remains unclear. The answer to these questions would make it possible to elucidate the mechanism of DNA DSB repair and the functions of the genes controlling this process. We have already presented a model system for studying DNA DSB repair (Perera et al. 1988). This system is based on the autonomously replicating plasmids (1) pLL12-BX carrying the yeast genes L E U 2 and L Y S 2 or (2) its derivate pLLI2-BX harboring a 391 bp deletion tightly linked to the unique Xhol site in the L Y S 2 gene. We studied the recombinational repair of plasmid DNA DSGs introduced into the L Y S 2 gone by XhoI. In this paper we propose a method for the quantitative estimation of plasmid DNA DSB and DSG repair efficiency. Further, we have used this system for studying the preciseness of plasmid DNA DSG repair in yeast cells of different genotypes. Plasmid DNA DSGs are generated by linearization at the XhoI site in the L Y S 2 gene of pLL12-BX. All the recombinational events in Rad § cells are accompanied by the complete recovery of DSGs. The mutation rad55 does influence either the efficiency or the precision of plasmid DNA DSG repair. Cells carrying the mutation rad57 are capable of recombinational repair of DSGs. However, the efficiency of this process is decreased compared with that in Rad § cells and is not accompanied with a precise recovery of DSGs. The efficiency of DSG repair is decreased significantly by the mutation rad53, with no influence on the precision of this process. The mutations rad50 and rad54 block plasmid DNA DSG repair completely.

Materials and methods The followinghaploid strains of the yeast Saecharomyees cerevisiae were used: DC5 a leu2-3, 112 his3; 2B-D241 a leu2-3, 112 lys2-25 his3.

EcoRl B a m H ~

Pstlz ~ I ~

stI

~EcoRl BamHl

Fig. 1. Map of plasmid pLLI2-BX: ~ Chromosomal DNA of Saccharomyces cerevisiae; ~ fragment of 2 mkm DNA (B-form); ~wrr PstI-EeoR[ fragment of pBR322; A the deleted sequence (391 bp) tightly linked to the XhoI site

These strains were kindly supplied by D. A. Gordenin (Leningrad State University, USSR). DC53 e rad53-1 DC54 c~tad54-3 DC55 c~rad55-3 DC57 a rad57-1

leu2-3, ii2 his3; leu2-3, 112 his3 trp2; leu2-3, 112 his3; leu2-3, I12 his3

These strains were constructed by mating strain DC5 with haploid strains of radiation-sensitive mutants from the collection of R. K. Mortimer (Berkeley, USA). The following diploid strains were used: H1

the homologous chromosome of the host cells (Perera et al. 1988). Two different unique sites in the plasmid pLL12-BX used here allowed us to estimate quantitatively the effectiveness of D N A DSB or D S G repair. The plasmid linearized in the region of homology at the XhoIsite can be repaired: (1) through recombinational repair with the homologous chromosome (recombinational repair); (2) through non-recombinational repair involving the ligation of the sticky ends, or other possible mechanisms, e.g., the formation of inverted dimers (Kunes et al. 1985); in this event, the transformation frequency for the plasmid linearized at the XhoI site is proportional to the sum of the following p values: fXhoI ~ P ~ +PL where PR and p/~ are the probabilities of recombinational and non-recombinational repair, respectively. If plasmid D N A is linearized at the B a m H I site (in the non-homologous region), only non-recombinational repair of the plasmid D N A is possible, i.e., fBomm ~PL 9

Hence, the proportion of molecules repaired by the nonrecombinational mechanism can be defined from the following equation: L = PL/(PR + PL ) ----f, amHI/fXho,"

(1)

Thus, the relative efficiency of recombinational repair (RERR) is proportional to the PR value:

a/~ RAD/RAD leu2-3, 112/leu2-3, 112 adel/ADE1 trpl/TRP1;

H50 a/~t rad50-1/rad50-1 leu2-3, 112/leu2-3, 112 adel/ADE1; H55 a/u rad55-3/rad55-3 leu2-3, 112/1eu2-3, 112 his3/HIS3.

PR ~ R E R R =fXhoI/ft~amnI

These strains were kindly supplied by N. A. Koltovaya (Joint Institute for Nuclear Research, Dubna, USSR). The bifunctional yeast vector pLLI2-BX (Fig. 1), a derivative ofpLL12 (Gordenin et al. 1988), was used. Plasmid pLLI2-BX was constructed by M. V. Trofimova (Leningrad State University) by deleting 391 bp in the region of the XhoI site of the LYS2 gene. In this construction the XhoI site was retained and the BamHI site was lost. Thus, pLL12-BX has two different unique sites, XhoI (in the LYS2 gene) and BamHI (in the tet gene of pBR322 sequence). The plasmid transformation of yeast was performed by the "lithium" method (Altherr et al. 1983), using carrier DNA, with modifications described previously (Perera et al. 1988). Leu + transformants were checked for stability by cloning on complete YEPD medium; thereafter 20 subclones of each transformant were replica plated on selective medium without leucine. Plasmid DNA extraction from yeast cells was performed by the modified method of Toh-E et al. (1982). The E. coli strain HB101 (recA) cells were transformed with DNA obtained by the "calcium'method (Maniatis et al. 1982). Plasmid DNA extraction from E. coli was performed by the alkaline method (Maniatis et al. 1982). The treatment of DNA was carried out under standard conditions (Maniatis et al. 1982) by the restriction endonucleases XhoI and BamHI obtained from commercial sources. Plasmid DNA was analysed by electrophoresis in a 0.8% agarose gel.

Table 1 shows the values of R E R R and L for yeast cells of different rad genotypes. It should be pointed out that the absolute values for transformation frequency varied greatly from one experiment to another. Nevertheless, the value of L showed much lower variation. Table 1 presents L values averaged over several experiments. As can be seen, L = 0 . 1 2 - 0 . 1 5 for Rad + cells, i.e., 8 5 - 8 8 % of Rad + clones transformed with p L L I 2 - B X and linearized at the XhoI site were produced as a result of recombinational repair of plasmid DNA. R E R R for Rad + cells ranged from 6.0 to 7.3. F o r the cells hearing the rad55-3 mutation the R E R R of the plasmid pLL12-BX slightly exceeds the respective value for Rad + cells. Mutations rad53-i and rad57-i decrease the efficiency of plasmid D N A D S G recombinational repair. However, the influence of mutation rad57-i is less pronounced ( R E R R = 3.5 _+0.7, compared with R E R R from 6 up to 7 for wild type cells) than that for rad53-i ( R E R R = 1.3 ___0.2). The mutations rad50-1 and rad54-3 block plasmid D N A recombinational repair completely ( R E R R in both events is close to zero). The mutation rad54-3, which blocks DSB repair in chromosomal DNA, is temperature-sensitive (Budd and Mortimer 1982). In cells bearing this mutation the linearized plasmid is repaired effectively during the incubation of the transformation mixture at the permissive conditions (23 ~ R E R R = 6 . 1 +1.2. The mutations rad50, rad53, rad55, rad57 were first obtained through their capability to increase yeast cell sensitivity to ionizing radiation (Game and Mortimer 1974). They increase the radiosensitivity of diploid yeast

Results

It was shown recently (White and Sedgwick 1985; Perera et al. 1988) that linearization of a plasmid in a region of homology resulted in a significant increase in the Rad + cell transformation frequency compared with the respective values for the circular molecule. This effect is due to plasmid D N A DSB repair through recombination with

-

-

1.

(2)

Table 1. The efficiency of plasmid DNA DSG repair in yeast ceils of various genotypes Recipient strain

Genotype

L~

RERR 2

DC5 2B-D241 H1 H50 DC53 DC54 DC54 DC55 H55 DC57

RAD L Y S 2 RAD l y s 2 - 2 5 RAD/RAD rad50-1/rad50-1 rad53-1 tad54-3 (23 ~ rad54-3 (32~ rad55-3 rad55-3/rad55-3 rad57-1 (23 ~

0.13_+0.033 0.14+_0.03 0.12+_0.02 14 0.44_+0.09 0.14_+0.03 14 0.11 -+0.02 0.10-+0.02 0.22-+0.04

6.2_+1.2 6.0_+1.2 7.3 +_1.5 0 1.3_+0.2 6.0_+ 1.2 0 8.1 + 1.6 9.0+1.8 3.5_+0.7

L 1= fBamHi/['Xhoi ' where fxhol,fBamHI a r e the cell transformation frequencies by the plasmid pLLI2-BX linearized at the XhoI and BamHI sites, respectively RERR 2= relative efficiency of recombinational repair of plasmid DNA, defined by equation (2) - see text 3 Error of mean for L averaged over 3-5 experiments 4 The transformation frequencies for the plasmids linearized at the XhoI and the BamHI sites are not significantly different

Table 2. Estimation of the precision of DSG repair in plasmid pLL12-BX, linearized at the XhoI site, in cells strain 2B-D241 No. of L s experiment

1 2 3

Proportion of Leu + transformants with a Lys- phenotype

Proportion of LeuLys+ plasmidIess segregants

0.09+0.026 16/200 (9.5%) 0/10 0.11+0.03 16/141 (11.3%) 0/29 0.07+0.02 10/79 (12.7%)-

Proportion of unstable Leu + cells among Leu + transformants 18/24 34/48 15/16

s See footnote 1 to Table 1 6 The error for L was calculated as the sum of relative errors for the transformation efficiency of plasmid pLL12-BX linearized at the XhoI and the BamHI sites

cells similarly (Petin and Kabakova 1981; Saeki etal. 1980). These mutations block the repair of chromosomal D N A DSBs (Vishnevetskaya etal. 1983; Glaser et al. 1985). However, as shown here, they differ significantly by their influence on plasmid D N A recombinational repair efficiency. The purpose of the following experiments was to estimate the precision of plasmid D N A D S G repair. The plasmid was linearized at the unique XhoI site of the L Y S 2 gene. Strain 2B-D241 was transformed by the linearized plasmid. This strain bears the deletion lys2-25, about 2.1 kb long, situated about 1 kb from the XhoI site (Chernoff et al. 1984). Thus, plasmid D N A and chromosomal D N A have non-overlapping deletions in the L Y S 2 gene. Leu + transformants were tested on a Lys § phenotype, replica plating their subclones on selective medium without lysine. Table 2 shows the results of three independent experiments. The proportions of Lys- transformants was 10%, 11%, and 12%, respectively. The value o f L was

equal tO 9_+2%, 11 _+3%, and 7 + 3 % , respectively. As mentioned above, the value of L corresponds to the proportion of transformants which appeared as a result of non-recombinational repair of the linearized plasmid. Thus, the results obtained prove that practically all Leu + transformants with a Lys + phenotype resulted from plasmid D N A recombinational repair. Thus, one can conclude that, in Rad + cells, recombinational repair of the plasmid is accompanied with a complete recovery of DSGs and this event occurs with a high probability. It must be noted, that Lys + transformants can appear after D S G repair in plasmid D N A as well as a consequence of the recovery of a deletion in the chromosomal DNA. In all studied events (ten tested transformants in the first experiment and 29 in the second), after losing the plasmid during incubation on the non-selective medium Leu+Lys + transformants lost the Lys + as well as the Leu + phenotype. Subsequently, D N A repair (i.e., complete recovery of the deleted region) in the chromosomal DNA, as a result of recombination with the homologous plasmid DNA, is a low probability event. The reason would be either the great length of the deletion (2.1 kb) in the chromosomal gene, or the large distance between the deletion and the DSB (about i kb from the XhoI site). It was of interest to define the preciseness of plasmid D N A D S G recombinational repair in radiosensitive yeast mutants. Yeast cells were transformed by plasmid pLL12-BX, linearized at the XhoI site. Transformants unstable for the Leu + phenotype constitute about 70% of the total (see Table 2 and Perera et al. 1988). Plasmid D N A was extracted from these unstable transformants and analysed for the existence of the B a m H I site in the L Y S 2 gene. The deletion in the plasmid L Y S 2 gene covers the B a m H I site (see Fig. 1). Thus, based on the number (1 and 2) of B a m H I sites in the plasmid p L L I 2 - B X extracted from Leu + transformants, one can judge the preciseness of DSG repair. The correctness of such a method of estimation is proved by the following observation: all the Leu + Lys + clones of strain 2B-D241, transformed by the linearized plasmid pLL12-BX, carry the repaired plasmid with two BamHI sites. In Table 3 the data on D S G repair in the linearized plasmid for Rad + cells and for radiosensitive mutants are shown. All the Rad + transformants (haploid as well as diploid) carry a plasmid with a second recovered B a m H I site in all tested cases. For example, Fig. 2 shows the plasmid DNA, extracted from three transformants of strain 2B-D241. Every plasmid is represented on two lines: circular, and linearized by treatment with the B a m H I restriction endonuclease. The repaired plasmid carries two B a m H I sites. The same results were obtained for strains DC5, and H1. The present data prove the high precision of plasmid D N A D S G repair in Rad + cells and is consistent with the results given previously (see Table 2). The recovery of plasmid D N A DSGs for haploid and diploid rad55 mutants is also characterized by a high degree of precision (see Table 3). All tested transformants of the DC55 and H55 strains carry the plasmid with two BarnHI sites. The values of R E R R in plasmid D N A for

Table 3. The recovery ofplasmid DNA DSGs after the transformation of Saccharomyces cerevisiae cells by plasmid pLL12-BX, linearized at the XhoI site Recpient strain

Genotype

Number of clones analysed

Numberofclones carrying plasmid pLL12-BX with a recovered second BamHI site

DC5 2B-D241

RAD LYS2 RAD lys2-25 RAD/RAD radSO-1/rad50-i rad53-1 tad54-3 (32 ~ rad54-3 (2 h at 23 ~ tad54-3 (23 ~ tad55-3 rad55-3/rad55-3 rad57-1(23 ~

10 10

10 10

10 15

10 0

28 12 v 11 12 10

17 0 3 6 10

H1

H50 DC53 DC54 DC55 H55 DC57

5

5

20

0

Fig. 2. Electrophoresis of repaired plasmids extracted from Leu + transformant~ of strain 2B-D241. 1 - Circular form; 2 - plasmid treated with BamHt; I - I I l - Numbers of clones from which repaired plasmids were extracted

7 The transformation mixture was held on selective agar at 23 ~ and then replaced by non-permissive conditions (34 ~

these strains are similar to that for wild type cells. Therefore, one can conclude that the rad55-3 mutation does not influence plasmid D N A D S G recombinational repair. The situation with mutation rad53-1 differs from that previously described. The efficiency o f p l a s m i d D N A recombinational repair is decreased significantly by this mutation ( R E R R = 1.3.+0.2), but the rare events of repair are accompanied by the recovery of DSGs. In Fig. 3 the plasmid D N A extracted from three Leu § transformants of strain DC53 is shown. Every plasmid is represented on three lines: circular, and linearized by treatment with the XhoI and the B a m H I restriction endonucleases, Two out of the three transformants carry a repaired plasmid with a second B a m H I site. The XhoI site is retained in all plasmids. The average value of L over several experiments for strain DC53 is 0.43, i.e., a b o u t 57% of Leu + transformants, chosen randomly, carry a plasmid repaired by recombination. In our case, t 7 out o f 28 clones of DC53 transformants (60%) carried the plasmid with two BarnHI ,sites (the results of three independent experiments are summarized). Consequently, one can conclude that the recombinational repair of plasmid D N A in the m u t a n t rad53 proceeds with low efficiency, but in the m o s t cases is accompanied by the precise recovery of DSGs. The results of experiments with strain DC57, carrying the mutation rad57-1, are of particular interest. This mutant has a relatively high level of linearized plasmid recombinational repair efficiency ( R E R R = 3.5 _+0.7). However, not a single plasmid extracted from 20 transformants included the second B a m H I site. Only 20% of these transformants could appear after a recombinational repair of plasmid D N A (L = 0.22). Figure 4 shows the plasmid D N A extracted from three Leu + transformants of strain DC57. All plasmids retained the XhoI site, but none of them contained the second B a m H I site. The same result was obtained in the other experiments

Fig. 3. Electrophoresis of repaired plasmids extracted from Leu + transformants of strain DC53 (rod53). 1 Circular form; 2 - plasmid treated with XhoI; 3 plasmid treated with BamHI; I - I I I Numbers of clones from which repaired plasmids were extracted

Fig. 4. Electrophoresis of repaired plasmids extracted from Leu + transformants of strain DC57 (tad57). 1 - Circular plasmid; 2 plasmid treated with XhoI; 3 - plasmid treated with BamHI; I - I I I - Numbers of clones from which repaired plasmids were extracted

with strain DC57. Thus, these data show that the recombinational repair of plasmid D N A in the rad57 m u t a n t is not accompanied by the complete recovery of DSGs. It is hard to imagine a mechanism for a such repair. However, the possibility of D S G partial recovery, even in the wild type cells, was shown by Orr-Weaver and Szostak (1983) though from their data the frequency of these events remained unclear, The results of the experiments with strain DC54 (temperature-sensitive mutation tad54-3) are most informative. Previously, using the m u t a n t rad54-3, we have

shown the biphasic kinetics ofplasmid D N A DSB repair: the first, fast phase, is completed in one h; a second, slow phase, requires 18 h of cell incubation under the permissive conditions (23 ~ Here, we studied the preciseness o f plasmid D N A D S G recovery after the end of both phases o f repair. The cells of strain DC54 were transformed by plasmid pLL12-BX, linearized at the XhoI site. The transformation mixture was incubated at 23 ~ and, after the different intervals of time, was then transferred to restrictive conditions (34~ Plasmid D N A was extracted from the transformants and analysed for the presence of a second BamHI site. The plasmids extracted from 12 tested clones grown under the non-permissive conditions did not contain the second BamHI site. After holding the transformation mixture during two h under the permissive conditions about 30% of analysed transformants (three out of 11) carried the plasmid with a second BamHI site. When cells were held at 23 ~ for 18 h, the proportion of plasmids with two BamHI sites increased to 50% (six out of 12). The recovery of the second BamHI site in only some of the plasmids is likely to result from the incomplete recovery of the Rad § phenotype of DC54 cells at the permissive temperature. There is a high probability that plasmids without DSGs in DC54 cells were repaired through recombination ( R E R R = 6,0 + 1.2). A second BamHI site was not recovered in plasmids extracted from transformed diploid H50 cells homozygous for radSO-t. This is to be expected, since the R E R R ofplasmid D N A for this strain is about zero, and almost all transformants appeared as a result of non-recombinational repair. It must be noted that the XhoI site was retained in all tested plasmids extracted from the different strains, whether or not the second BamHI site was recovered.

Discussion The present data show that mutation rad55-3 does not influence either the efficiency or the preciseness of the recombinational repair of DSGs; the mutation rad53-1 decreases the efficiency with no influence on the precision of repair. Plasmid D N A is repaired effectively through recombination in the mutant rad57-1, but here repair is not accompanied by the recovery of DSGs. The mutations rad50-1 and rad54-3 completely block plasmid D N A repair. The system used here, and previously (Perera

et al. 1988), indicated significant differences in the modes of repair of D N A DSBs and DSGs in the radiosensitive mutants of yeasts whereas the method of sedimentation analysis, used previously, was not capable of distinguishing these different modes. It must be emphasized that the character of plasmid D N A DSB repair in the system employed can differ from that in chromosomal D N A when the DSB is induced by ionizing radiation. For transformant viability the repair of a single plasmid (one DSB or D S G per cell) is enough. It corresponds to an X-ray dose of up to I 0 - 1 0 0 Gy (Frankenberg-Schwager et al. 1980). In the experiments on chromosomal D N A DSB repair using sedimentation analysis the doses of radiation used, and the levels of DSBs formed, per cell are far higher due to the lower sensitivity of the method. Thus, one must take this into account when comparing the efficiency of plasmid and chromosome D N A DSB repair in yeast. References Altherr MR, Quinn LA, Kado CI, Rodrigues RL (1983) In: Lurquin PF, Klienhofs A (ed) Genetic engineering in eukaryotes. Washington, p 33 36 Budd M, Mortimer RK (1982) Mutat Res 103:19-24 ChernoffYO, Kidgotko OV, Demberelijn O, Luchnikova IL, Soldatov SP, Glaser VM, Gordenin DA (t984) Curr Genet 9:31-37 Frankenberg-Schwager M, Frankenberg D, Blocher D, Adamczyk C (1980) Int J RadioI Biot 37:207-212 Game JC, Mortimer RK (1974) Mutat Res 24:281-292 Glaser VM, Samadashvili MP, Vishnevetskaya OV, Soldatov SP, Shestakov SV (1985) Mutat Res 147:296-297 Gordenin DA, Trofimova MV, Shaburova ON, Pavlov Yi, Chernoff YO, Chekuolene YU, Proscyavichus YY, Sasnauskas KV, Janulaitis AA (1988) Mol Gen Genet 213:388-393 Kunes S, Botstein D, Fox MS (1985) J Mol Biol 184:375-387 Maniatis T, Fritsch EF, Sambrook J (1982) Molecular cloning. A laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York Orr-Weaver TL, Szostak JW, Rothstein RJ (1981) Proc Natl Acad Sci USA 78:6354-6358 Orr-Weaver TL, Szostak JW (1983) Proc Natl Acad Sci USA 80:4417-4421 Perera JR, Glasunov AV, Glaser VM, Boreiko AV (1988) Mol Gen Genet 213:421-424 Petin VG, Kabakova NM (1981) Mutat Res 82:285-294 Toh-E A, Tada S, Oshima Y (1982) J Bacteriol 151:1980-1990 Saeki T, Nashida I, Nakai S (1980) Mutat Res 73:251-265 Vishnevetskaya OYu, Luchnik AN, Arutunova LS, Shestakov SV (1983) Genetika 19:26-32 White C, Sedgwick S (1985) Mol Gen Genet 201:99-106 Communicated by P. P. Slonimski