Spontaneous and ultraviolet-induced mutations on a single-stranded shuttle vector transfected into monkey cells

Mutation Research, DNA Repair, 274 (19921 135-145 © 1992 Elsevier Science Publishers B.V. All rights reserved 0921-8777/92/$05.00

135

MUTDNA 06494

Spontaneous and ultraviolet-induced mutations on a single-stranded shuttle vector transfected into monkey cells Catherine Madzak *, Januario B. Cabral-Neto, Carlos F.M. Menck # and Alain Sarasin Laboratory of Molecular Genetics, Institut de Recherches Scientifiques sur le Cancer, Villejuif (FranceJ (Received 15 October 1991 ) (Revision received 10 January 19921 (Accepted 20 January 19921

Ko'words: Single-stranded DNA; Ultraviolet light; Mutagenesis; Shuttle vector: Frameshifl

Summary The shuttle vector plasmid PCF3A, carrying the supF target gene, can be transfected into monkey COS7 cells as single-stranded or double-stranded DNA. Single strand-derived plasmid progeny exhibited a 10-fold higher spontaneous mutation frequency than double strand-derived progeny. The location of spontaneous mutations obtained after transfection of the single-stranded vector shared similarities w!th that for double-stranded vectors. However, the nature of base changes was very different. Single-stranded PCF3A DNA was used to study ultraviolet-induced mutagenesis. An earlier report (Madzak and Sarasin, J. Mol. Biol., 218 (19911 667-673) showed that single-stranded DNA exhibited a lower survival and a higher mutation frequency than double-stranded DNA after ultraviolet irradiation. In the present report, sequence analysis of mutant plasmids is presented. The use of a single-stranded vector allowed us to show the targeting of mutations at putative lesion sites and to determine the exact nature of the base implicated in each mutation. Frameshift mutations were more frequent after transfection of control o, irradiated plasmid as single-stranded DNA than as double-stranded DNA. Multiple mutations, observed at a high frequency in the spontaneous and ultraviolet-induced mutation spectra following single-stranded DNA transfection, could be due to an error-prone polymerisation step acting on a single-stranded template.

Correspondence: Dr. A. Sarasin, Laboratory of Molecular Genetics, lnstitut de Recherches Scienti~,|aes snr Ic Cancer. P.O. Box 8, F.94801 Villejuif Cedex, France. * Present address: Laboratory of Genetics. INRA. Centre de Biotechnologies Agro.industrielles, F.78850 ThivervalGrignon, France. * Present address: Department .of Biology, Institute de Bioci6ncias. CP 11461, Universidade de 3fie Paulo, Silo Paulo, 05499 SP, Brazil.

One of the most relevant questions that mutagenesis studies deal with, at tile molecular level, is to analyse the relationships between given DNA lesions and induced mutations, with a special emphasis on the possible targeting of mutations opposite lesions. We designed for this purpose a single-stranded (ss) shuttle vector which permits the unambiguous determination of the nature of each base that has given rise to a mutation.

13h

ss DNA carrying lesions cannot be processed by excision repair mechanisms and translesion synthesis should occur in order to produce fully replicated double-stranded (ds) progeny. The shuttle vector used in this study, PCF3A, has been previously described (Madzak and Sarasin, 1991a). it contains the origin of replication and late region of simian virus 40 (SV40), a chloramphenicol resistance gene and a plasmidic ColE1like origin of replication. The suppressor tRNA supF gene was used as a target gene to determine the mutation frequency. The plasmid also contains the replication origin from the ss bacteriophage fl, which allows the production in bacteria of ss DNA from the vector (Madzak et al., 1989). Monkey COS7 cells were used as the host for PCF3A vector. These cells are transformed by origin-defective SV40 (Gluzman, 1981) and provided in trans the large T antigen necessary for the replication of SV40-based vectors. We used the ss PCF3A vector to analyse the spectrum of ultraviolet-induced mutations. The major lesions formed on DNA molecules after irradiation with 254-nm ultraviolet (UV) light, pyrimidine dimers

and pyrimidine(6-4)pyrimidone photoproducts, both implicate doublets of adjacent pyrimidines. However, several authors have described minor UV-induccd lesions that implicate pyrimidincs and purines (Bos¢ et ai., 1983; Gallaghcr and Duker, 198f~; Bourre ct al., 1987; Urata et al., 1989), which could possibly play a role in UV-induced mutagenesis. One cannot exclude that a given mutation, attributed to a pyrimldin¢ doublet, could in fact be due to a UV-indueed photo. product involving a purine on the opposite strand. The use of a ss shuttle vector makes it possible to avoid these ambiguities in the study of mutation targeting.

Materials and methods

Cells COS7 monkey cells (Glu,~man, 1981) were grown in Dulbecco's modified Eagle's medium supplemented with 7% fetal calf serum and antibiotics. DNA was transfected into COS7 cells using the DEAE-dextran techniqu,: (Wilson,

1978). Extrachromosomal DNA was recovered from transfected cells by a small-scale alkaline lysis procedure adapted from the method described by Birnboim and Doly (1979).

Plasmid and bacteria PCF3A construction has been described previously (Madzak and Sarasin, 1991a). Bacterial transformations were performed with the method of Hanahan (1983). The E. coli strain JM105 carries the F' episome and is therefore permissive for ss bacteriophage infection. PCF3A ss DNA was prepared from plasmid-carrying strain JM105 as previously described (Madzak et al., 1989) using M13K07 helper phage (Pharmacia-PL, Biochemicals inc.). A DNAase I digest was performed on the ss pseudophage preparation in order to remove any contaminating ds naked DNA (Madzak et al., 1989). The E. coli strain MBM7070 (Seidman et al., 1985) was used for the screening of supF- mutants among PCF3A plasmid progeny. A functional supF gene suppresses an amber mutation in the lacZ gene of the MBM7070 indicator strain, leading to the formation of bright blue colonies in the presence of X-Gal and IPTG. Mutations in this gene yield white or light blue colonies. Experimental procedure The irradiation of ss or ds PCF3A DNA was performed with a germicidal lamp (producing UV light mainly at a wavelength of 254 nm). The UV fluence at 254 nm was measured with a Vilber Lourmat radiometer. Control or UV-irradiated ss or ds plasmids were transfected into COS7 cells (0.1-1 /zg DNA per 10-ram ceU dish). After 3 days, extrachromosomal DNA was purified, digested with Dpnl in order to remove input ds DNA, and shuttled to MBM7070 strain. Transformants were plated in the presence of ehloramphenicol, X-Gal and IPTG. The mutation frequency was defined as the ratio of white or light blue colonies (containing supF- plasmid) to the total amount of bacterial colonies obtained. The DNA of the supF ~ mutants was sequenced with the Sanger method using the Sequenase Version 2,0 (United States Biochemical Corporation) kit.

137

Results

Following transfection of control or UV-irradiated ss or ds PCF3A DNA into COS7 cells, the relative survival of the plasmid progeny and the mutation frequency on the supF target gene were determined after shuttling to strain M~M7070. Numerous supF- mutations screened in bacteria were further analysed by DNA sequencing. Only independent mutants were integrated in the mutation spectrum (mutants carrying the same modification(s) were considered independent only if they were isolated from different transfection experiments in COS7 cells). The Dpnl treatment being unable to degrade some input ss DNA that would eventually remain after the passage in COS7 cells, we found it necessary to determine if this could introduce a bias in our results. We previously showed that input ss DNA was extensively degraded in the few hours following transfection into monkey cells (Madzak and Sarasin, 1991b): we were unable to detect unreplieated ss DNA after only 8 h in monkey cells. Moreover, the efficiency of the ss PCF3A for the transformation of MBM7070 bacteria was very low: 0.3% of that found with the ds vector. Any minor contamination of the DNA recovered from COS7 cells would therefore be unable to yield bacterial colonies. At last, we ensured that the mutation frequencies observed after direct transformation of bacteria with control or UV-irradiated ss vector were much lower than those obtained after passage into COS7 cells, which was indeed the case: less than 3.3 x 10 -4 (0 mutant/3014 colonies) for the control and less than 10 =4 (0 mutant/9787 colonies) for the UV-irradiated (200 J / m " ) ss PCF3A, the survival compared to control being only 3.9% in this latter case. The results of these experiments ensure that our observations indeed reflected what occurred in monkey cells.

Spontaneous mutation spectrum We previously reported a spontaneous mutation frequency of 2.4 x 10 -3 following transfection of 0.1/.tg ss PCF3A DNA per COS7 cell dish (Madzak and Sarasin, 1991a). We found it necessary to check if this mutation frequency was susceptible to variation within the range of ss DNA

quantities used for the recovery of mutants (0.1-1 /zg), We found that mutation frequencies obtained after transfection of 0.1, 0.5 or ! # g ss PCF3A DNA were not significantly different (chi-square statistical test). The spontaneous mutation frequency obtained for ss PCF3A in COS7 cells, for the total of these experiments, was 2,8 x 10 -3 (152 white or light blue colonies out of a total of 54,006 colonies). We sequenced 30 spontaneous mutants; four of them carried deletions (from = 20 to = 150 bp) and the other 26 carried a total of 32 point mutations. These point mutations comprised ( - 1) frameshifts and substitutions, and are presented in Table 1. Of these point mutations, 22 (69%) were previously found in the spontaneous spectrum on the sup F gene of the ds pZ189 vector (RoUides et al., 1988; Moraes et al., 1990). The 10 mutations that were not previously described with ds vectors were, however, located at sites already found in the ~pontaneous spectrum on ds vectors (Roilides et al., 1988; Moraes et al., 1990): the differences lay essentially in the nature of the base changes (see below) but not in the location of the mutations. Among these mutations, 3 were part of multiple mutations (and are therefore putative silent mutations, which can explain their absence from other spectra). Spontaneous hot spots were observed, after transfection of ss PCF3A, at positions 65 and (to a lesser extent) 98, 101 and 102 (respectively 34%, 12.5%, 12.5~, and 15.5e/~ of the point mutations). Among 32 spontaneous point mutations found after transfection of ss PCF3A, 28 (87.5%) concerned a C nucicotide, 2 a (3 and 2 a T. it is worth noting that these 4 latter mutations, not targeted on a C nucleotidc, were all part of multiple mutations (thus associated with at least 1 mutation on a C nucleotide). Among the C-targeted mutations, 11 ( 4 1 ~ ) w e r e C to A transversions, 8 (29.5%) were C to G transversions and 8 (29.5%) were C to T transitions. These differences between the 3 possible base changes are not significant in a chi.square statistical test. The presence of hot spots did not introduce a bias in these results, since the numbers of mutation sites for the 3 possible base changes were also similar: 5 (C to A transversions), 6 (C to G transversions) and 4 (C to T transitions). The cytosines therefore appeared to

TABLE 1 SPONTANEOUS MUTATIONS OBSERVED AFTER TRANSFECTION OF ss PCF3A DNA INTO COS7 CELLS

teristics of these mutations were similar to those of frameshifts found after UV irradiation of the ss vector (see below).

Location"

UV-induced mutation spectra

Mutationh

Previouslyfound with ds vector c

11

AT(I)

II

I T (2)

13 50 53 57 65 65 65 65 65 65 65 65 65 f15 65 98 98

CtoA(3) G to A (4) C to A (4) C to G C to A (3) C to A (4) C to A C to A C to A C to A C to A C to G (I) C to G C to T C to T C to A (2) C to G

08

C to G

98

C to T (5) C to G C to T C to T C to T C to A C to G C to T C Io T AC C It) G G to T(5)

no no yes yes yes no yes yes yes yes yes yes yes yes yes yes yes yes no no yes no yes yes yes

IOI 101 Illl fill 102

102 1l)2 102 102

1(16 123

yes

When the vector was UV-irradiated and transfected as ss DNA, the plasmid survival was lower and the increase in mutation frequency higher than when ds DNA was used (Figs. 1 and 2). Fig. 3 compares the UV-induced mutation spectrum following transfection of ss PCF3A (69 mutants were found to carry 117 point mutations and 1 deletion), and the UV-induced spectrum following transfection of ds PCF3A (24 mutants were found to carry 36 point mutations and 1 deletion). The 2 deletions were found at low UV doses (50 and 100 J / m 2) and could therefore be spontaneous mutations; they concerned respectively 92 and 11 bp. Table 2 presents the different types of spontaneous or UV-induced point mutations obtained with ss or ds PCF3A. The UV-induced point mutations comprised frameshifts (essentially deletions of 1 nucleotid¢) and substitutions concerning a single base, 2 adjacent bases (tandem), 2 base:+ separated by an unchanged nucleotide (double close) or several - 2-4 bases separated by as many as 70 nucleotides

y,zs

(multiple),

no

Th0 framcshift mutations observed following transfection of control or UV-irradiated ss PCF3A DNA shared similar characteristics: they wore located in (or at the border of) tracks of pyrimidines. Most of those frameshifts were associated with substitutions (tandem, double close or multiple mutations) as shown in Fig. 3 (and in Table 1 for spontaneous frameshifts}. The frequencies of multiple mutations found in plasmids derived from UV-irradiated ss and ds DNA were similar (Table 2), These multiple mutations wore also found in the spontaneous mutation spectrum of ss PCF3A. As previously discussed (Madzak and Sarasin, 1991a), the frequency of multiple mutations cannot always be explained by an equal number of UV-induced lesions. The use of a shuttle vector as ss DNA is particularly interesting for the study of mutation targeting: we observed that 92% of the mutations obtained after UV irradiation of ss PCF3A were

no no

yes no

" Location or the mutations on the sequence of the supF gene (the numeration is that shown in Fig, 3),

~' The bases indicated are those present on the transfected DNA strand: A indicates the deletion of a nucleotide: mutations found in the same mutant (multiple mutations) are marked by identical numbers, " Indicates if the mutation was previously d~scribcd in th¢

spontaneous mut;ttion spectrum of the ds supF-contuining pZ189 vector (Roilides el al,, i988; Moraus ut u[,, 1990).

be mutated at similar frequencies to the 3 possible other bases, after transfection of ss DNA into monkey cells. Frameshift mutations (deletions of I base) were found at a rather high frequency (11.5%, Table 2) following transfection of ss PCF3A. The charae-

139

20 u)

ss PCF3A ---a-..-

._>

ds PCF3A

I0

I I I I I I I I I 11 0 501C0 *d$

e S$

2°0

: 1 ', : I

~

S0O uvDOSE

I¢00

0

2oo

4oo

6oo

000

1000

UV dose ( J / m Z )

(Jim2)

Fig. 1, Survival was determined 3 days after transfcction into COS7 cells of 0,1 p,g of ss or ds PCF3A D N A by transformation of MBM7070 bacteria, The 100% level corresponds to about 50 ng D N A / c e l l dish after transfection of ds PCF3A and i0 ng/cell dish for ss PCF3A,

targeted on pyrimidine doublets (only 39% of mutations would correspond to pyrimidine doublets if they had occurred randomly). This result confirms the targeting of UV-induced mutations at the level of lesions, and is free of the ambiguities present in previous studies using ds templates (Bourre and Sarasin, 1983; Protic-Sabljic et al., 1986; Brash et al., 1987). The base substitutions observed after transfection of UV-irradiated ss PCF3A are presented in

Fig. 2. The mutation frequency on the supF target gene was determined, 3 days after transfection into COS7 cells of 0.1 pg of ss or ds PCF3A DNA, by analysis in MBMT070 bacteria. The enhancement factor is the ratio of the UV-induced mutation frequency to the spontaneous one. The ratio of the 2 slopes is 1.6.

Table 3. The C to "1' transition represented 65% of the substitutions obtained from UV-irradiated ss DNA. in the UV-induced spectrum of ss PCF3A, 79% of the mutations concerned a cytosine and only 15% concerned a thymine (Table 3). However, these bases were equally represented in the sequence of the supF gene in ss PCF3A (see Fig. 3). The amounts of C-containing and T-containing pyrimidine doublets were also

TABLE 2 TYPES OF M U T A T I O N S IN ss (CONTROL OR U V - I R R A D I A T E D ) A N D ds ( U V - I R R A D I A T E D ) PCF3A REPLICATED IN COS7 CELLS PCF3A DNA Number of independent mutant phlsmids sequenced Number of mutants carrying: deletions ( = 20-150 bp) point mutations frameshifts substitutions d: single base tandem double close multiple

ss conlrol

ss UV "

ds UV i,

30

fig

24

4 26 ( 100% ) 3 (11.5~)

I 69 ( 10(Ir;; ) 7 (10~) c

! 23 ( 101)c~ ) -

20 (77~) _ 5 (19C/rl

31 18 l0 14

16 (7ql~) 4 (17r~) I (4~;) 3 (13c;)

(45r/~) (25'~/~) (16r/~) (19~)

The percentages correspond to the fraction of plasmids with pc)int mutations that carry each particular type of mutation. The total is greater than I()()~ because one mutant can carry, for example, a multiple mutatkm in which a tandem mutalion is included. " UV dose to ss PCF3A [number of plasmids sequenced]: 50 J / m 2 [24], 100 J / m 2 [10], 2(}{)J / m : [20], 500 J / m : [8] and 101)0J / m :

[71. h UV dose to ds PCF3A [number of plasmids sequenced]: 200 J / m : [6]. 500 J/m" [7] and 10(}{)J / m 2 [S]. " One ( + I) insertion was found after 200 J / m : UV treatment and one ( - 2 ) deletion after 500 J/m-" UV trealmcnt: all others were ( - I) deletions. d Substitutions can involve a single base. 2 adjacent bases (tandem}. 2 bases separated by an unchanged nucleotide (double oh}sol or several bases distant from each tither (multiple).

140 TABLE 3

BASE CHANGES OF 103 SUBSTITUTIONS IN MUTANTS DERIVED FROM U V - I R R A D I A T E D ss PCF3A Transhions ( 70r~;) C~ T T--'C A -'*G G~A

Transt'ersions (30%) C ---.G C-~A T--*A T--,G G -~(" G-.T A "--.C A-*T

65e/r 3% 2%

6~ 8c; 8% 4¢; 2% I¢~ lC/~

similar: the sequence of supF contains 13 TT, 12 CT, 10 CT and 12 CC pyrimidine doublets. Moreover, the frequency of formation of UV-induced lesions was found to be several times higher for thymine doublets than for cytosine ones (Bourre et al., 1985). Therefore, thymine appeared to bc under-rcprescnted in the UV-induced mutation sites, when compared to the number of lesion sites.

Discussion Spontaneous mutation ]}'t'quenO' The spontaneous mutation frequency observed after transfeetion of ds PCF3A is very low for a transient shuttle vector: 2.3 × 10 -4. in contrast, the spontaneous mutation frequency after transfection of the same vector as ss DNA is 10-fold higher: 2.8 × 10-'k We were able to produce s~ DNA from the supF-containing pZ189 vector (Madzak and Sarasin, 1991b) and to determine the spontaneous mutation frequency after its transfection into monkey cells. Similarly, this frequency was 10-fold higher when pZ189 was used as ss DNA, compared to ds DNA (data not shown). A higher spontaneous mutation frequency following transfection of ss vectors can bc attributed to: (1) the fact that ss DNA can be more subject to the formation of spontaneous lesions during transfer

through the cytoplasm than ds DNA (in vitro, ss DNA was found to be 4-fold more liable to depurination and 250-fold more liable to cytosine deamination than ds DNA; Lindahl and Nyberg, 1972, 1974; Friedberg, 1985) a n d / o r (2) the absence of repair of spontaneous lesions on a ss template. In contrast, in previous experiments using another shuttle vector (~-SVF1) carrying the lacO sequence as a target gene, we found a similar mutation frequency for ss and ds DNA (Madzak et al., 1989). However, the very small size of the lacO target (26 bp) should considerably lower the range of recoverable mutations and could introduce a bias in mutagenesis frequencies. Moreover, the spontaneous mutation frequency per nucleotide was higher for ds ~-SVF1 (2 × 10 -5) than it is for ds PCF3A (1.5 × 10-6). The lower spontaneous mutagenesis of this latter vector makes it possible to detect some phenomena that were not evident with the other vector. However, we cannot rule out the hypothesis that the nature of the target gene could play a role in the relationships between the DNA structure and the mutation frequencies.

Spontaneous mutation spectrum on ss PCF3A A major (position 651 and 3 minor (positions 98, I01 and 1(12) hot spots were observed in the spontaneous mutation spec.mtm of the ss PCF3A vector. Position 102 was represented only by I mutation in the spontaneous spectrum of ds pZ189, but the 3 others were also hot spots with this ds vector (Roilides et al., 1988; Moraes ctal., 1990). However, some of the main hot spots observed with ds pZ189 (positions 7, 45, 53, 92,106 and 111; Roilides et al., 1988; Moraes et al,, 1990) were not represented (or only by I mutation) in our spectrum on the ss PCF3A, Therefore, the location of spontaneous mutations on ss and ds DNA presents similarities, but some hot spots could be characteristic of a given type of DNA structure. Differences in hot spot patterns could also be attributed to specific cellular factors (PCF3A was transfected into COS7 cells and pZl89 into CV1 cells), as was exemplified by Seetharam et al, (1990). The point mutations observed in the spontaneous mutation spectrum on ss PCF3A were es-

141

~

-=

-i ~ .~'~ _~ p,

!

= ~.~ .~ " o g.~u

<~.=N ~.,~

b

IE~.i1¢¢~: .

='=..~

E~='E= ~. . ~ ' ~

~.. __ .¢= ~ ~ ¢ ~ ' ' i

|

~ £, ~.-

",~ .~ = .= ,= " '7~'~ =

e=

'~w

.~

=E

sentially (87.5~) located at C nucleotides. Since mutations concerning other nucleotides (T or G) were all part of multiple mutations, 100% of single-base substitutions were, in fact, targeted at C nucleotides. This result is consistent with that found with ds pZ189 vector, for which 99% of spontaneous mutations were targeted at G : C base pairs (Moraes et al., 1990). We did not find any preferential mutation in the spontaneous spectrum of ss PCF3A: the C nucleotides were mutated at similar frequencies to the 3 possible other bases. There was no preferential insertion of a particular nucleotide in front of putative C-targeted spontaneous lesions. This result is in striking contrast with that obtained with ds vectors, for which G : C to A : T transition was the main spontaneous mutation observed (Roilides et al., 1988; Moraes et al., 1990). Transitions of G : C to A : T were also found as the major spontaneous mutation occurring in mammalian cells by several authors using various systems (Miller et al., 1984; De Jong et al., 1988; Phear et al., 1989). This type of transition was classically attributed to the spontaneous deamination of cytosine, leading to the formation of a uracil residue in the DNA molecule. The spontaneous deamination of cytosine to uracil is 250-fold more frequent in ss than in ds DNA. under physiological conditions (Lindahl and Nyberg, 1974; Friedberg, 1985). The presence of U nucleotides in a ss molecule is expected to give C to T transitions during replication. However, this mutation was not predominant in the spontaneous mutation spectrum of ss PCF3A. UraciI-DNA glycosylasc is able to remove uracil from a s s DNA molecule. and this process is even 3-fold more efficient than on a d s one (Krokan and Wittwer, 1981). Such a repair process is expected to lead to the nicking of the DNA strand (Loeb and Preston, 1986), which would be lethal for a s s molecule. In the absence of complete repair of uracil, it could be envisaged that some blockage occurs at an intermediate step (possibly an abasic site) wl, ich would be the mutagenic substrate, We did ~ot observe any incorporation bias during in vi,o experiments using a vector containing a unique abasic site in monkey cells: no preferential insertion of a particular base was observed in front of the non-coding abasic site (Gentil et al., 1990). A non-coding

142

spontaneous lesion, such as an abasic site, targeted on C nucleotides could therefore be responsible for the mutation spectrum obtained. Spontaneous depurination is 20-fold more frequent than depyrimidination on a ss DNA molecule, under physiological conditions (both being 4-fold more frequent than on ds DNA; Lindahl and Nyberg, 1972; Friedberg, 1985). Mutations due to spontaneously formed abasic sites are therefore expected to be targeted on purines, which was not the case in our experiments.

Mutagenesis foUowing transfection of UV-irradiated ss PCF3A No system was available, up to now, to compare, in mammalian cells, the increase in mutation frequency obtained after UV irradiation of the same target gone as ss or ds DNA. The lower survival and the higher mutagenesis levels observed after transfection of irradiated ss DNA can be attributed to the lack of excision repair a n d / o r post-replicative recombination on a s s template. Mutations observed after transfection of the UV-irradiatcd vector were essentially targeted at putative lesion sites, namely pyrimidine doublets. As discussed in our earlier report (Madzak and Sarasin, 1991a), untargetcd mutations found after transfection of UV-irradiated ss DNA were el. ther obtained at low L)V do~cs (2 putative spon. taneous mutations at nuclcotidcs = 33 and 83) or associated with targeted mutations in the same plasmid (part of double close or multiple mutations). We proposed a schema of 'semi-targeted' mutagenesis to illustrate the formation of double close mutations (Madzak and Sarasin, 1991a). It is interesting to note that one of the untargeted mutation sites (the adenine at position 43) we observed in the UV-induced spectrum on ss PCF3A was also found in that on ds PCF3A and was also present in the UV-induced spectrum in a human xeroderma pigmentosum (XP) variant cell line (Wang et al,, 1991). This mutation site could be due to some rare UV lesion or to 'semi-targeted' mutagenesis, since all the 3 mutations observed were tandem substitutions implicating the adenine and the cytosine in 5' or in 3', The UV-induced mutations we described here for the ds PCF3A vector (Fig. 3) were a subset of

those previously obtained on the supF gene of the ds pZ89 shuttle vector (Bredberg et al., 1986; Keyse et al., 1988). In our earlier report, the location and intensity of the mutation hot spots obtained after irradiation of ss PCF3A were compared to these data. The UV-induced substitution spectra obtained on ss and ds DNA appeared to be globally quite similar (Madzak and Sarasin, 1991a). The C to T transition was the major (65%) substitution observed after irradiation of the ss vector. The nature of the substitutions found after UV irradiation of ss PCF3A were therefore markedly different from those of the spontaneous spectrum (in which all 3 types of C-targeted substitutions were equally observed m see above). The G : C to A: T transition was previously found to be the major UV-induced mutation in several ds systems (Bredberg et al., 1986; Hauser et al., 1986; Brash et al., 1987; Drobetsky et al., 1987; Hsia et al., 1989; Madzak and Sarasin, 1991a). Our results with ss DNA allow us to attribute more specifically this transition to UV-irradiated C-containing pyrimidine doublets. The under-

representation of thymine nucleotides in the UV-induced mutation spectrum, compared to the high number of T-containing lesions, was also observed in several systems (see references cited above). The low frequency of mutations involving thymines (and also the predominance of the C to T transition) are classically explained by the 'A rule': polymerases would present an incorporation bias towards adenine, in front of UV-induced photoproducts.

Frameshif¢ mutations Frameshifts were found at rather high frequencies after transfection of control or UVirradiated ss PCF3A into monkey cells: 3 frameshifts out of 26 mutants with point mutations (11,5%) for control ss DNA and 7 out of 69 (10%) for UV-irradiated ss DNA. in contrast, only 3% of frameshifts were found after transfec. tion of ds pZI89 into monkey cells (Roilides et al,, 1988; Moraes et al., 1990) and only 2.5% after transfection of UV-irradiated ds vectors into monkey cells (Hauser et al., 1986; Keyse et al., 1988; Madzak and Sarasin, 1991a). The differ-

143

ence between UV-irradiated ss and ds DNA is significant in a chi-square statistical test. It is interesting to note that frameshifts found after transfection of control or UV-irradiated ss DNA share the same characteristics: they were essentially ( - 1 ) deletions and they all occurred in (or at the border of) a track of 3-8 pyrimidines. Indeed, they all occurred at a C or T nucleotide in a sequence 5' CnT m 3' where n and m varied from 1 to 3 (in one case, an AC doublet was deleted in a 5' A C T I T 3' sequence). Despite these similarities, the frameshifts found after UV irradiation are probably not simply spontaneous mutations: the frequency of frameshifts among mutants remained constant with increasing UV dose and their overall frequency among the total plasmid population recovered from monkey cells showed a dose-dependent increase following UV irradiation (the number of frameshifts is, however, too low to allow a statistical study of this data). Frameshifts found after UV irradiation could therefore be related to UV-induced lesions, but it is difficult to assess if they occurred in front of a lesion (which could be the case since all but one - at position 139 - were targeted at pyrimidine doublets) or if they were the consequence of some error-prone mechanism: indeed, 6 UV-induced frameshifts out of 8 were associated with another targeted mutation (putative error-prone pathway leading to semi-targeted double close mutations or to multiple mutations). Similarly, 2 spontaneous frameshifts out of 3 were part of multiple mutations (putative error-prone pathway linked to one spontaneous lesion). Some reasons for an increase in the frequency of frameshifts after transfection of UV-irradiated ss DNA in monkey cells could possibly be differences in the structure of the UV-induced lesions a n d / o r of the template. Some speculations about this problem can be made on the basis of a recent work comparing the effect of a single cis-syn or trans-syn T-T cyclobutane dimer in MI3 DNA (Banerjee et al., 1990). The cis-syn dimer-containing molecule produced only targeted substitutions in E. coli (Banerjee et al., 1988). In striking contrast, the trans-syn dimer-containing molecule leads exclusively to targeted single T deletions in uninduced E. coli (Banerjee et al., 1990). In SOS-induced E. coli, the mutations included tar-

geted substitutions and near-targeted single base additions as well as the T deletions. Considering that the trans-syn isomer, rare in ds DNA, represents 10% of the dimers when ss DNA is treated with UV (Patrick and Rahn, 1976), we propose that this difference could contribute to a higher frequency of frameshifts following irradiation of the ss vector.

Multiple mutations A high frequency of multiple mutations was found after transfection of UV-irradiated ss PCF3A (19%). Multiple mutations were also observed in the spontaneous mutation spectrum of the ss plasmid (17% of mutants). A similar frequency of multiple base changes in the spontaneous and UV-induced mutation spectra of ds shuttle vectors propagated in monkey or human cell (Hauser et al., 1986; Hsia et al., 1989) was attributed to an error-prone replication step during excision repair (Bredberg et al., 1986; Seidman et al., 1987). As excision repair cannot be performed on a ss template, multiple mutations were not expected to occur after transfection of ss DNA. in E. coil, multiple mutations were also observed in the UV-induced spectrum of the ss bacteriophage M13 (Le Clerc et al., 1984). They were attributed to an error-prone polymerase activity in SOS-induced bacteria. The multiple mutations we obtained after transfection of ss DNA can be due to an errorprone replication pathway acting on the ss template, resembling that postulated in SOS-induced E. coll. Such a pathway may enter the register of putative inducible SOS functions in mammalian cells (Sarasin, 1985). DNA replication on a UVirradiated ss template could possibly lead to the formation of gaps opposite lesions, if UV lesions block DNA synthesis. Such a structure could be a substrate for an error-prone polymerase. It could also be a substrate for a putative post-replication repair process which should occur ir~ such a case between 2 distinct plasmid molecules. This putative pathway should be less efficient than regular post-replication repair occurring at the replication fork between 2 daughter molecules. We cannot, however, rule out the hypothesis that the multiple mutations observed are the result of an

144

excision repair process working on ds DNA after copying the ss template. Analysis of the mutation spectrum of ss DNA in XP cells would be helpful to clarify this point.

Acknowledgements We thank Dr R. Tyrrell (ISREC, Epalinges, Switzerland) and Dr M. Seidman (Otsuka America Pharmaceuticals, MD, U.S.A.) for the gift of E. colt MBM7070 strain and pZ189 shuttle vector. C.M. held a fellowship from the Association pour ia Recherche sur le Cancer (Villejuif, France). J.B.C.-N. held a doctoral fellowship from the Brazilian National Research Council (CNPq). This work was supported by grants from A.R.C., from the Minist~re de la Recherche et de la Technologie (Paris, France) and from the Commission of the European Communities (No. BI7 0034-C, Brussels, Belgium). References Banerjee, S.K,, R,B, Chrislensen, C.W. Lawrence and J.E. LeClerc (1988) Frequency and spectrum of mutations prodoted by a single cis.syll thymine-thymine cyclobutane dimer in a single-stranded vector, Proc. N;HI, Acad. Sci. (U,S,A,). 85, 8141=8145. Banerjee, S.K.. A, Borden, R.B. Christenscn, J.E. LeClerc and C,W, Lawrence (1991)) SOS-dependent replication past :l single tr.tl,~'.,~yt) T-T cyclobulune dimer gives a different mutation spectrum and increased error rate compared with replication past this lesion in uninduced cells, J, Bact., 172, 21115=2112, Birnboim, H.C.. and J. Doly (19791A rapid alkaline extraction procedure for screening plasmid DNA, Nucleic Acids Res., 7, 1513=1523, Bose, S.N.. RJ,H. Davies, S,K. Sethi and J.A, McCIoskey (1983) Formation of an adenine-thymine photoadduct in the deoxydinucleoside monophosphute d(TpA) and in DNA, Science, 220, 723=725, Bourre, F,, and A, Surasin (1983) Targeted mutagenesis of SV40 D N A induced by UV-light, Nature (London), 305, 68-70. Bourre. F.. (3. Renault, P.C. Seawell and A. Sarasin (19851 Distribution of ultraviolet.induced lesions in simian virus 40 DNA, Biochimie, 67, 293-299. Bourre, F,, G, Renault and A. Sarasin (1987)Sequence effect on alkali-sensitive sites in UV-irradiated SV4(I DNA, Nucleic Acids Res., 15. 8861-8875. Brush. D,E., S, Seetharam, K.H. Kraemer, M,M. Seidman and A. Bredberg (19871 Photoproduct frequency is not the major determinant of UV base substitution hot spots or cold spots in human cells, Proc, Natl. Acad. Sci. (U,S.A.), 84, 3782-3786.

Bredberg, A., K.H. Kraemer and M.M. Seidman (19861 Restricted ultraviolet mutational spectrum in a shuttle vector propagated in xeroderma pigmentosum cells, Proc. Natl. Acad. Sci. (U.S.A.), 83, 8273-8277. De Jong, P.J., A.J. Grosovsky and B.W. Glickman (19881 Spectrum of spontaneous mutation at the aprt locus of Chinese hamster ovary cells: an analysis at the DNA sequence level, Proc. Natl. Acad. Sci. (U.S.A.), 85, 34993503. Drobetsky, E.A., A.J. Grosovsky and B.W. Glickman (1987) The specificity of UV induced mutations at an endogenous locus in mammalian ceils, Proc. Natl. Acad. Sci, (U.S.A.), 84, 9103-9107. Friedberg, E.C. (1985) in: E.C. Friedberg and P.C. Hanawalt (Eds.), DNA Repair, Freeman, New York, pp. 3311-340. Gallagher, P,E,, and N.J. Duker (1986) Detection of purine photoproducts in a defined sequence of human DNA, Mol. Ceil. Biol,, 6, 707-709. Gentil, A., G. Renault, C. Madzak, A. Margot, J.B. CabralNet(), J.J. Vasseur, B. Rayner, J,L. lmbach and A. Sarasin (1990) Mutagenic properties of a unique abasie site in mammalian cells, Biochem. Biophys. Res. Commun., 173, 704-710. Gluzman, Y. (1981) SV40-transformed simian cells support the replication of early SV40 mutants, Cell, 23, 175-182. Hanahan, D, (1983) Studies on transformation of Escherichia colt with plasmids, J. Moi. Biol., 166, 557-580. Hauser, J., M.M. Seidman, K. Sidur and K. Dixon (1986) Sequence spccificiw of point mutations induced during passage of a UV-irradiated shuUle vector plasmid in monkey cells. Mol, Cell, Biol., 6, 277-285. ltsia, H,C,, J,S, Lehkowski, P,-M, Leong, M,P. Calos and J,II, Miller (1989) Comparison of ultraviolel-induced mulagencsis of the lacl gent in Escherichiu colt and in human 293 cells, J, Mol, Biol,, 2115, 103=113. Ke~,~e, S,M,, F. Amaudruz and R,M. Tyrrell (1988) Determination of the spectrum of mutations induced by definedwavelength solar UVB (313 nm) radiation in mammalian cells by use of a shuttle veclor, Mol. Cell. Biol., 8, 54255431, Krokan, H., and C,U. Wittwer (1981 ) Urac~I-DNA glycosylase from HeLa cells: general properties, substrate specificity and effect of uracil analogs, Nucleic Acids Res,, 9, 25992613. Le Clerc, J,E,, N.L, Istock,B.R, Saran and R. Allen Jr, (1984) Sequence analysis of ultraviolet.induced mutations in Ml31acZ hybrid phage DNA, J, Mol, Biol,, 180, 217-237. Lindahl, T,, and B, Nyberg (19721 Rate of depurination of native deoxyribonucleic acid, Biochemistry, I I, 3610-3618, Lindahl, T,, and B. Nyherg (19"/41 Heat-induced deamination of cytosine residues in deoxyribonucleic acid, Binchemistw, 13, 3405-3410, Loeb, L.A., and B,D, Preston (1986) Mutagenesis by apudnic/apyrimidinic sites, Annu, Rev. Genet., 20, 201230, Ma(v,ak, C,, and A. Sarasin (1991a) Mutation spectrum following transfection of ultraviolet-irradiated single-stranded or double-stranded shuttle vector DNA into monkey cells, J. Mol. Biol,, 218, 667-673.

145 Madzak, C., and A. Sarasin (1991b) Conversion of a singlestranded SV40-based shuttle vector to its double-stranded form does not require the SV40 large T antigen in monkey cells, J. Gen. Virol., 72, 3091-3093. Madzak, C., C.F.M. Menck, J. Armlet and A. Sarasin (1989) Analysis of single-stranded DNA stability and damage-induced strand loss in m~mmalian cells using SV40-based shuttle vectors, J. Mol. Biol., 205, 501-509. Miller, J.H., J.S. Lebkowski, K.S. Greisen and M.P. Calos ( 19841 Specificity of mutations induced in transfected DNA by mammalian cells, EMBO J., 3, 3117-3121. Moraes, E.C., S.M. Keyse and R.M. Tyrreil (19901 Mutagenesis by hydrogen peroxide treatment of mammalian cells: a molecular analysis, Carcinogenesis, i I, 283-293. Patrick, M.H., and R.O. Rahn 0976) Photochemistry of DNA and polynucleotides: photoproducts, in: S.Y. Wang led.), Photochemistry and Photobiology of Nucleic Acids, Vol. II, Academic Press, New York, pp. 35-95. Phear, G., W. Armstrong and M. Meuth (19891 Molecular basis of spontaneous mutation at the aprt locus of hamster cells, J. Mol. Biol., 209, 577-582. Protic-Sabljic, M., N. Tuteja, P.J. Munson, J. Hauser, K.H. Kraemer and K. Dixon (19861 UV light-induced cyclobutane pyrimidine dimers are mutagenic in mammalian cells, Mol. Cell, Biol., 6, 3349-3356. Roilides, E., J.E. Gielen, N. Tuleja, A,S. Levine and K. Dixon (19881 Mutational specificity of benLo[a]pyrene diolepoxide in monkey cells, Mutation Res., 198, 199-206. Sarasin, A. (19851 SOS response in mammalian cells, Cancer Invest., 3, 163-174,

Seetharam, S., K.H. Kraemer, H.L. Waters and M.M. Seidman (199(|) Mutational hotspot variability in an ultraviolet-treated shuttle vector plasmid propagated in xeroderma pigmentosum and normal human lymphoblasts and fibroblasts, J. Mol. Biol., 212, 433-436. Seidman, M.M., K. Dixon, A. Razzaque, RJ. Zagursky and M.L. Berman (19851 A shuttle vector plasmid for studying carcinogen-induced point mutations in mammalian cells, Gene, 38, 233-237. Seidman, M.M., A. Bredberg, S. Seetharam and K.H. Kraemer (19871 Multiple point mutations in a shuttle vector propagated in human cells: evidence for an error-prone DNA polymerase activity, Proc. Natl. Acad. Sci. (U.S.A.), 84, 4944-4948. Urata, H., K. Yamamoto, M. Akagi, H. Hiroaki and S. Uesugi (19891 A 2-deoxyribonolactone-containing nucleotide: i~lation and characterization of the alkali-sensitive photoproduct of the trideoxyribonucleotide d(ApCpA), Biochemistry, 28, 9566-9569. Wang, Y.-C., V.M. Maher and J.J. McCormick (1991) Xeroderma pigmentosum variant cells are less likely than normal cells to incorporate dAMP opposite photoproducls during replication of UV-irradiated plasmids, Proc. Nat!. Acad. Sci. (U.S.A.), 88, 7810-7814. Wilson, J.H, 0978) Interference in SV40 DNA infections: a possible basis for cellular competence, Virology, 91,380388.