|Home | About | Journals | Submit | Contact Us | Français|
Yeast can be readily transformed by single-stranded oligonucleotides (ssOligos). Previously, we showed that an ssOligo that generates a 1-nt loop containing an AP site corrected the −1 frameshift mutation in the lys2ΔA746 allele. However, these experiments had to be performed in yeast apn1 mutants lacking the major AP endonuclease. In this study, we show that bypass of an AP site can be studied in repair-proficient yeast by using ssOligos that generates a 7-nt loop containing an AP site. The bypass studies performed using the ssOligos that generate a 7-nt loop was validated by demonstrating that the result obtained are similar to those derived using ssOligos containing a 1-nt loop in an apn1 mutant. By using the 7-nt loop system, we showed that the bypass efficiencies of AP sites are dependent on the sequence context that surrounds the lesion and are apparently no affected by the direction of DNA replication. In contrast, the mutagenic specificity of an AP site is not affected by the sequence context or the direction of replication. In all cases, dC is inserted at twice the frequency of dA opposite an AP site, indicating that REV1 is mainly responsible for bypass of AP sites at all lesion sites studied.
The apurinic/apyrimidinic (AP) site has been well documented to be both cytotoxic and mutagenic, and is one of the most frequently encountered DNA damages in cells [1-4]. Although AP sites are generated spontaneously via hydrolytic depurination [3-5], formation is increased substantially when cells are exposed to chemical oxidants and genotoxicants such as alkylating agents and ionizing radiation [3, 6]. Furthermore, AP sites are intermediates of base excision repair initiated by DNA glycosylases[4, 7, 8]. AP sites are recognized efficiently by AP endonucleases and lyases, which are ubiquitous DNA repair enzymes present in all organisms[4, 7, 8]. E. coli endonuclease IV and exonuclease III are prototype AP endonucleases that nick the DNA 5’ to the lesion, generating a gapped DNA that terminates with a 3’ hydroxyl and a 5’ phosphate. On the other hand, AP lyases form Schiff base covalent intermediates with AP sites, cleaving the phosphodiester bond 3’ to an AP site. There are two major types of AP lyases. The β-AP lyase generates a gapped DNA containing a 3’ 4-hydroxylpentenal moiety and a 5’ phosphate, whereas the β,δ-AP lyase leaves behind a gap terminating with a 3’phosphate and a 5’ phosphate. Yeast contains two AP endonuclease, APN1 and APN2, with APN1 being the major AP endonuclease activity[9-11]. In addition, yeast has at least three AP lyases: NTG1, NTG2 and OGG1[12, 13].
The in vivo bypass of AP sites can be studied in yeast using plasmid gap-filling assays that require the conversion of a gapped plasmid to a fully duplex form[14-16], double stranded plasmid [17, 18] or transformation assays that employ lesion-containing short single-stranded oligonucleotides (ssOligos) that hybridize to a defined genomic region[19-21]. Using ssOligos containing an AP site that targets the lys2AΔ746 allele, we previously demonstrated that dC was incorporated preferentially opposite an AP site during lesion bypass. Bypass of the AP site in this system required the participation of REV1 and Pol ζ (a complex of the REV3 and REV7 proteins), with Pol η (encoded by RAD30) only having a minor role in the bypass. These results are in agreement with earlier observations published by other laboratories using either ssOligos or plasmid-mediated transformation assays[14-16, 20].
Our previous study used ssOligos that generate a 1-nt loop containing an AP site when annealed to the chromosomal lys2ΔA746 allele. An AP site located in a 1-nt loop was found to be recognized by yeast APN1, and thus bypass experiments were performed in a yeast apn1 mutant lacking the major AP endonuclease. Recently we showed that ssOligos that generate either a 1-nt or 7-nt loop when annealed to the yeast chromosomal LYS2 locus can efficiently correct the −1 frameshift mutation in the lys2ΔA746 allele. We further showed that transformation of yeast by these ssOligos occurred preferentially on the lagging strand of replication and is consistent with a mechanism in which ssOligos anneal to single-stranded regions of DNA at a replication fork and serve as primers for DNA synthesis.
Most repair enzymes fail to recognize DNA damage when the lesion is in single-stranded DNA. In the current study, we examined AP site bypass in repair-proficient yeast cells using ssOligos that generate an AP site within a 7-nt loop that targets correction of the genomic lys2ΔA746 −1 frameshift allele. By using ssOligos that contain different sequences surrounding the AP site and targeting to either the leading or the lagging strand, we show that bypass efficiency is significantly affected by the sequence context, and only modestly affects by the direction of replication. In contrast, bypass specificity of an AP site is not affected by either sequence context or direction of replication. Finally, we confirm that REV1 and Pol ζ (REV1/REV7), but not Pol η (RAD30), play important roles in the bypass of AP sites.
Yeast strains were grown in YPDA liquid media (YPD containing 0.0075% l-adenine hemisulfate) and on YPDA plates. Synthetic complete (SC) medium containing 2% dextrose and lacking the appropriate amino acid was used for selective growth of transformants . Yeast APN1 was a gift from Dr. Dindial Ramotar (University of Montreal). E. coli endonuclease IV is routinely prepared in the laboratory as a stock reagent[24, 25].
All oligonucleotides were obtained from Qiagen (Figure 1, Supplemental Data). The COD(+7) and NC(+7) series of 40-mers contain uracil and have the same sequence as the coding and non-coding strands of LYS2. The COD(+7)-F4 oligonucleotide contain a unique tetrahydrofuran. COD(+7)-A4 and NC(+7)-T4 are the control oligonucleotides containing no lesion; the COD(+7)-AR oligonucleotide has the same sequence as COD(+7)-A4, but is of reversed polarity. Each oligonucleotide was purified by 15% polyacrylamide gel electrophoresis as described earlier .
All strains used were derivatives of SJR922 (MATa ade2−101oc his3Δ200 ura3ΔNco lys2ΔA746). Yeast strains SJR1777 (SJR922 apn1Δ), SJR1780 (SJR922 apn1Δ apn2Δ), SJR2367 (SJR922 apn1Δ ogg1Δ), SJR1141 (SJR922 rad16Δ + inverted lys2ΔA746) and SJR1454 (SJR 922 rad16Δ) were constructed as described earlier. In SJR1141, the LYS2 locus was inverted with respect to the upstream origin of replication of SJR922, an operation led to the concomitant inactivation of the adjacent RAD16 gene. All yeast strains used in this study were obtained from Dr. Sue Jinks-Robertson (Duke University) except SJR922WS4 (SJR922 rev1Δ::kanMX), SJR922WS5 (SJR922 rev3Δ::kanMX) and SJR922WS6 (SJR922 rad30Δ::kanMX) were obtained from Dr. Wolfram Siede (University of North Texas Health Science Center, Fort Worth).
Duplex D-U4 and heteroduplex D-(+7)U4 containing uracil were prepared by mixing COD(+7)-U4 (5’-ACCAAGATTTCAAATTACCUGCAGACGAGCTCAAGCATCA) with either a complementary 40-mer (CS, 5’-TGATGCTTGAGCTCGTCTGCTGGTAATTTGAAATCTTGGT) or 33-mer (CS-7, 5’-TGATGCTTGAGCTCGTCAATTTGAAATCTTGGT, similar sequence to CS but lacking the underlined 7-nt sequence) at a 1:1 ratio. The oligonucleotide mixture was then heated to 90°C for 5 minutes and cooled down gradually to room temperature. Duplex D-S4 and heteroduplex D(+7)-S4 containing an AP site were prepared by incubating the DNA with an excess amount of uracil DNA glycosylase for 15 minute at 37°C. The AP-containing DNAs were then used as substrates for yeast APN1 and E. coli endonuclease IV. Yeast APN1 assayed in a 10 μl reaction ontaining 20 femtomoles of 5’-32P-end labeled substrates, 10 mM Tris-HCl, 50 mM KCl, 1 mM EDTA at 37°C for 10 min. Reactions were stopped by adding 10 μl of loading buffer (90% formamide, 1 mM EDTA, 0.1% xylene cyanol, and 0.1% bromophenol blue) and heating at 70°C for 10 min to avoid cleavage of the AP site. Three to five μl of the reaction mixture were then loaded onto a 12.5% denaturing polyacrylamide gel and electrophoresed at 2000V for 1.5 hr. The gel was then dried under vacuum and analyzed by a Storm PhosphoImager (Molecular Dynamics).
The minimum amount of uracil DNA glycosylase needed for the complete conversion of an uracil to an AP site was determined by using oligonucleotides that were 5’ end labeled with 32 P, treated with a known amount of uracil DNA glycosylase(Data not shown). The resulting DNA containing AP sites were then heated at 90°C under alkali condition (pH 12.0) for 30 minutes. The reaction was then analyzed by loading the reaction mix on to a 12.5% denaturing polyacrylamide gel and electrophoresed at 2000V for 1.5 hr. Complete conversion of uracils to AP sites is indicated by the total cleavage of the input oligonucleotides under this condition.
Yeast transformation was performed by electroporation following the procedure published by Kow et al.[21, 22]. In a typical experiment, 300 pmol of oligonucleotide and 5 ng of a HIS3-containing CEN plasmid (pRS313; ) which was used as an internal control for transformation efficiency, were added and mixed with 300 μl of the competent cell suspension. We demonstrated previously that 300 pmol of ssOligo was in the linear range of transformation efficiencies. The mixture was kept on ice for 5 min and transferred to an electroporation cuvette (2 mm gap, 400 μl capacity). Electroporation was performed by applying a 1.5 kV, 186 Ω Manipulator Electroporation System, BTX). The cells were then diluted immediately with 900 μl YPAD medium and allowed to recover with gentle agitation at 30° for 15 min before being spread on SC-Lys for selection of oligonucleotide transformants or SC-His for selection of plasmid transformants. Colonies were counted after 5−7 days incubation at 30°C.
All transformations were done with the same mixture of oligonucleotide and plasmid. Transformation efficiencies were normalized across strains by calculating the ratio of oligonucleotide transformants to plasmid transformants.
The bypass efficiency was calculated as the fraction of Lys+ colonies obtained with ssOligos containing AP sites to the total number of Lys+ colonies obtained with the control ssOligo that had no damage. Variation in transformation efficiencies between strains and repetitions were normalized using a HIS3-containing CEN plasmid (pRS313) included in all transformations. The bypass efficiencies are reported as the average of at least two independent transformations, and the spread of the data is reported as the modified standard deviation.
Yeast genomic DNA was extracted from the Lys+ transformants using standard yeast protocols. A region of the LYS2 gene spanning the sequence of the transforming oligonucleotide was PCR amplified using primers MO18 and Lys1359R as described in Harfe and Jinks-Robertson. The PCR products were sequenced using a commercial service (Functional Biologicals, Wisconsin). Only transformants that acquired an insertion of 7-nucleotides at the expected position in the lys2AΔ746 allele were compiled to generate the mutation spectrum associated with the bypass of each of the AP sites.
We and others have demonstrated that ssOligos containing a unique lesion can be used to study lesion bypass in yeast [21, 22]. The lys2ΔA746 allele is a −1 frameshift allele and has been shown to be especially useful for studying lesion bypass[21, 22]. A unique feature of the lys2ΔA746 allele is that a compensatory +1 frameshift can occur anywhere within an approximately 150 bp “reversion window” and, therefore, transforming ssOligos can have the correcting nucleotide or lesion anywhere within this window[21, 22, 27]. In addition, the lys2ΔA746 reversion window can accept a significant number of nucleotide insertions in steps of 3n +1 (n = 0,1,2,3...). This was demonstrated by our recent work showing that yeast can be effectively transformed with ssOligos that anneal to chromosomal DNA to generate either a 1-nt or 7-nt loop. These two distinct features make the lys2ΔA746 allele very attractive for studying mechanisms of mutagenesis, in particular with respect to the effect of sequence context on mutagenic lesion bypass.
Most DNA repair enzymes, including APN1, can efficiently recognize DNA damage only when the lesion is in the context of double-stranded DNA. Our earlier study showed that when yeast is transformed with a ssOligo that generates a 1-nt loop containing an AP site, efficient transformation was only observed in yeast apn1 mutants lacking the major AP endonuclease, APN1. These data suggested that yeast APN1 can still recognize an AP site when the lesion is located in a 1-nt loop. However, it is expected that when the AP site is located in a larger single-stranded loop, APN1 will cease to recognize the AP lesion. In order to confirm this, a ssOligo that anneals to generate a 7-nt single-stranded loop containing a centrally located AP site was examined for its substrate susceptibility with purified yeast APN1 (Figure 1). Duplex D-U4 and heteroduplex D-(+7)U4 were prepared by hybridizing oligonucleotide COD(+7)-U4 to CS and CS-7, respectively. Duplex D-S4 and heteroduplex D-(+7)S4, containing an AP site located in perfect duplex or in a 7-nt loop, respectively, were then prepared by incubating duplex D-U4 and heteroduplex D-(+7)U4 with an excess amount of uracil DNA glycosylase. Duplex D-S4 and heteroduplex D-(+7)S4 were then treated with various concentrations of yeast APN1. Figure 1 shows that APN1 efficiently recognizes the AP lesion when it is in a perfect duplex DNA (lanes 3−6); however, APN1 failed to recognize the AP site when the AP site was in a 7-nt loop (lanes 7−10). Appreciable nicking of the duplex D-S4 was observed at 6 ng of APN1 (Fig. 1, lane 4), and greater than 90% nicking of the duplex D-S4 was observed at APN1 concentrations greater than 30 ng (Fig.2, lane 6). In contrast, under the same reaction conditions, no appreciable nicking of heteroduplex D-(+7)S4 was observed even at 150 ng of APN1 (Fig 1, lane 10). When ssOligo COD(+7)-S4 anneals to the genomic lys2ΔA746 allele of the wild-type APN1 yeast (SJR922), an identical 7-nt loop containing a centrally located AP site will be generated (see Figure 1, Supplemental Data). The AP site in the 7-nt loop is expected to be refractory to cellular APN1 activity, and thus it is expected that one can study the bypass of an AP site in wild-type, repair-proficient yeast. To demonstrate this, we compared the number of transformants obtained by transforming various repair-deficient mutant strains with 300 pmoles of COD(+7)-S4. We have demonstrated earlier that 300 pmoles of ssOligo is optimal and within the linear range of transformation [21, 22]. Table I shows that the average number of transformants obtained was similar for all the strains tested, including the wild-type yeast SJR922, SJR1777 (apn1Δ), SJR1780 (apn1Δ apn2Δ), and SJR2367 (apn1Δ ogg1Δ). These data suggest that neither APN1 nor other DNA repair enzymes such as APN2 and OGG1 are able to initiate repair of an AP site when the lesion is located in the middle of a 7-nt loop.
The activity of APN1 on an AP site might vary depending on whether the AP site is at the 5’ base or at the 3’ base of the loop. To determine whether the position of an AP site within a 7-nt loop affects APN1 activity, 7-nt loop substrates containing an AP site at various positions within the loop were treated with APN1. Figure 2 shows that when an AP site is located at positions S2 to S7 (see Figure 1, Supplemental Data) in the 7-nt loop, no appreciable nicking of the heteroduplex substrates by APN1 was observed even at 75 ng of APN1(Figure 2, Panels S2 to S7). However, when an AP site is located at the 5’ position of the loop (S1 position, Figure 1, Supplemental Data), significant cleavage of the substrate was observed at high APN1 concentrations (Figure 2, Panel S1). At 30 ng and 75 ng of APN1, 13% and 78% of the 7-nt loop substrates were cleaved (Fig.2, panel S1, lanes 2 and 3). In contrast, at 6 ng of APN1, greater than 95% of the duplex substrate cleaved (Fig.2, panel S1, lane 4). These data suggest that this system is generally suitable for studying bypass of an AP site in wild-type, repair-proficient yeast cells. When an AP site is located at the 5’ base of the 7-nt loop (S1 position, Figure 1, Supplemental Data), however, a reduced transformation efficiency is expected due to the potential repair by cellular APN1 activity. This is clearly shown in Figure 3. Figure 3 shows that the average number of transformants obtained with all COD oligonucleotides [COD(+7)-S2 to COD(+7)-S7], with the exception of COD(+7)-S1, are similar for both wild-type and apn1 yeast. For example, WT and apn1 yeast transformed with 300 pmoles of COD(+7)-S7 yielded 1144 and 1374 Lys+ transformants, respectively. In contrast, WT and apn1 yeast transformed with 300 pmoles of COD(+7)-S1 yielded 709 and 1450 Lys+ transformants, respectively. This is in agreement with the observation that when an AP site is located at the 5’ base of the 7-nt loop (S1 position, Figure 1, Supplemental Data), APN1 can still cleave the AP site lesion, thus generating a lower number of transformants for the wild-type yeast.
The above experiment clearly demonstrates that COD ssOligos containing an AP site can be used to study AP site bypass in repair-proficient yeast. To validate that ssOligos containing a7-nt loop is suitable for studying lesion bypass, we will determine the role of various DNA lesion-bypass polymerases in the bypass of an AP site. Transformation was performed using ssOligos COD(+7)-S4 and COD(+7)-F4 containing a unique AP site and tetrahydrofuran, an AP site analog, respectively. Electrocompetent wild-type yeast (SJR922) was transformed with a mixture containing 300 pmoles of ssOligo and 5 ng of pSR313 plasmid, a HIS3-containing CEN plasmid. The ratio of Lys+/His+ colonies was used for to compare the transformation efficiency of a control ssOligo [COD(+7)-A4] with that of ssOligos containing an AP site or tetrahydrofuran [COD(+7)-S4 and COD(+7)-F4], and was used to calculate the efficiency of AP site bypass (Table II). Table II shows that both an AP site and a THF were bypassed at relatively high efficiencies, at 65% and 59%, respectively. The observed bypass efficiencies were significantly higher than those previously reported by us and others [14-16, 19-21] . In earlier results using ssOligos that generate a +1 nt loop containing an AP site, the bypass efficiency in wild-type was observed to be less than 10% for both the lys2ΔA746 and cyc1−1 alleles.
We and others have shown that bypass of an AP site requires the active participation of REV1 and Pol ζ (REV3/REV7) but not RAD30 Polymerase η) [14-16, 20, 21]. In order to demonstrate that bypass of an AP site in wild-type yeast also primarily depends on the REV1 and REV3 proteins, yeast mutants lacking the REV1, REV3 or RAD30 gene were transformed with COD(+7)-S4 and COD(+7)-F4. Figure 4 shows that when COD(+7)-S4 was used to transform yeast, bypass efficiencies for the AP site decreased significantly in rev1Δ(7.4%) and rev3Δ (2.7%) mutants, and was not affected in rad30Δ mutant (88 %). Smaller decreases in bypass efficiency for THF were observed for rev1Δ, rev3Δ and rad30Δ mutants (21%, 11%, and 55 %, respectively). These data are in agreement with earlier observations that bypass efficiencies for THF in both rev1Δ and rev3Δ mutants were significantly higher than the bypass efficiency for an AP site, suggesting that a significant amount of the bypass at a THF residue is mediated by polymerases other than REV1 and REV3[20, 21]. It is likely that at least part of the residual THF bypass might be due to Polη.
To further examine the roles of REV1 and Polζ in Polη the bypass and of an AP site generated by COD(+7)-S4, transformants were sequenced and the mutation specificity was obtained for each of the yeast strains (Table III). As expected, dC(68%) followed by dA(32%) was the major nucleotides inserted across from an AP site is wild-type yeast (Table III). Deletion of REV1 or REV3 gene led to a reduction of dC and an increased frequency of incorporating dA across from an AP site (insertion of dA in wild type is 31% as compared to 64% and 82% for Δrev1 and Δrev3, respectively). The reverse pattern was evident in the Polη-defective rad30 mutant, with dC incorporation increasing to 81% and dA incorporation decreasing to 19%. This is consistent with the interpretation that Polη plays a minor role in inserting dA opposite an AP site.
In order to examine the effect of sequence context on the bypass of an AP site, ssOligos containing AP sites embedded within different sequence context were used for the transformation assay (Figure 1, Supplemental Data). The COD and NC series of oligonucleotides are ssOligos that have same sequence as the coding and non-coding strands of the LYS2 gene, respectively. In the context of chromosome replication, LYS2 gene is replicated primarily from an upstream origin, making the coding and non-coding strands the lagging and leading strands of replication, respectively. Table IV shows that bypass efficiencies of AP sites observed in experiments using the eight COD ssOligos ranged from 7% to as high as 74%. Similarly, when yeast transformations were performed with the NC ssOligos, bypass efficiencies for AP sites also varied widely, and depending on the position of the AP site within the 7-nt loop, ranged from 8% to 35%. These data clearly support the notion that bypass efficiency of an AP site is significantly affected by the sequence context that surrounds the lesion. To further confirm the effect of sequence context on AP site bypass, we compared the bypass efficiencies of two ssOligos that have an AP site in the same position in the loop, but of different sequence context. Table IV showed that bypass efficiency of an AP site observed with COD(+7)-S4’ was significantly lower than COD(+7)-S4 (16 % versus 74%), supporting the notion that the sequence context surrounding an AP site can influence the bypass efficiency of an AP site. There is the possibility, however, that AP sites at or near the bottom of the loop are recognize by APN1, thus leading to an apparent lower efficiency of bypass. As indicated earlier, only when an AP site was located at the 5’ base of the loop did we observe significant recognition by APN1 (Figure 2, Panel S1), and this might explain the lower efficiency of AP bypass observed at the 5’ base of the loop [COD(+7)-S1 and NS(+7)-S7 ssOligos, Table IV] as compared to the 3’ base of the loop [COD(+7)-S7 and NC(+7)-S1 ssOligos, Table IV].
To determine the effect of sequence context on REV1, Polζ and Polη activities, ssOligo-mediated transformations were performed in rev1Δ, rev3Δ and rad30Δ mutants (Table IV; Table I in Supplemental Data shows the actual bypass data the two independent transformations performed for these mutant strains). As expected, in the absence of the REV1 or REV3 protein, bypass efficiencies for AP sites were greatly diminished to a range of 1% to 11% (Table IV). The bypass efficiencies for an AP site were consistently lower in the rev3Δ than in the rev1Δ mutant, suggesting that bypass is not completely dependent on REV1. As indicated in Table IV, bypass efficiencies in the absence of REV1 or REV3 were also dependent on the sequence context in a manner that is similar to the wild-type yeast. In the absence of REV1 or REV3, bypass of an AP site is potentially carried out by Pol α/δ, suggesting that bypass of an AP site by Pol α/δ may be similarly affected by the sequence context. As expected, RAD30 only played a minor role in the bypass of AP sites (Table IV), with bypass efficiencies of AP sites observed with the NC ssOligos in the rad30Δ mutant being similar to those observed in wild-type yeast (compare rad30Δ column with NC column in Table IV). Taken together, these data suggest that the sequence context surrounding an AP site can exert a significant effect on the ability of these polymerases to insert or extend a nucleotide across from an AP site.
When transformants derived from the COD series of ssOligos in the wild-type background were sequenced, dC was observed to be the preferred nucleotide inserted opposite the AP site (range from 52% to 66%, Table V); dA was the second most frequently inserted nucleotide (range from 12% to 44%, Table V). The mutation specificity is consistent with REV1 being the major enzyme responsible for nucleotide insertion across from all AP sites irrespective of the sequence context.
The bypass efficiencies of AP sites in wild-type yeast determined using the COD ssOligos [COD(+7)-S3, -S4, -S6 and -S7] were approximately two-fold higher than the corresponding NC ssOligos [NC(+7)-S3, -S4, -S5 and -S7]. COD and NC ssOligos hybridize to the leading and lagging strands of replication, respectively, in wild-type SJR922 (Table IV). AP sites derived from COD and NC ssOligo series will thus be on the lagging- and leading-strand templates during the subsequent round of DNA synthesis, and bypass of the AP site will occur on the lagging and leadings, respectively (see Figure 2A, Supplemental Data). These data suggest that the bypass of AP sites may be more efficient on the lagging as compare to the leading strand of DNA replication. Even Because we have demonstrated that bypass of an AP site is affected significantly by the sequence context surrounding an AP site, the difference in bypass efficiencies between the NC and COD ssOligos could potentially be due solely to the differences in sequence context that surround the AP sites. In order to directly determine whether bypass of a given AP site is affected by the direction of DNA replication, bypass of AP sites was performed in two yeast strains, SJR1454 and SJR1141, using the same ssOligos(Table VI). In SJR1141, the LYS2 gene was inverted relative to the upstream replication origin, which reverses the identity of the leading and lagging strands. Because the flipping of the LYS2 gene also led to the inactivation of RAD16 gene, a rad16Δ derivative of SJR922 (SJR1454) was used in these experiments. Thus, NC ssOligos will anneal to the leading and lagging strands of LYS2 in SJR1141 and SJR1454, respectively, and bypass of an AP site in these strains will occur on lagging and leading strands, respectively (Figure 2B, Supplemental Data). Even though our earlier data indicated that yeast transformation by single stranded oligonucleotide is significantly more efficient when the oligonucleotide hybridizes to the lagging strand (; see also Table II, Supplemental Data), the differences in the leading and lagging strand transformation efficiencies are corrected when the bypass efficiencies of AP sites are normalized against oligonucleotides that do not contain AP sites (See Table II for how the bypass efficiency is calculated). Table VI shows that bypass efficiencies of AP sites observed with NC ssOligos NC(+7)-S4, NC(+7)-S5 and NC(+7)-S6 appears to be two-fold higher in SJR1141 as compared to SJR1454, and no significant differences in bypass efficiencies were observed with other ssOligos, it is therefore likely that bypass of AP sites is not affected by the direction of DNA replication. This observation is in agreement with the recent observations of Pages et al.
We have demonstrated that the 7-nt loop system allows one to perform detailed analysis of the genetic and cellular factors that influence the bypass efficiency and specificity of DNA lesions without imposing additional repair defects. This is demonstrate by the observation that similar to the 1-nt loop system, bypass efficiencies of AP site is dependent on REV1 and Rev3, but not Rad30 (Figure 4). Furthermore, the insertion specificity opposite an AP site in a wild-type, repair-proficient yeast is similar to an apn1 mutant. In agreement with our earlier studies and those of others, dC is clearly the preferred nucleotide incorporated opposite an AP site, followed by dA. Incorporation of dC opposite an AP site is consistent with an enzymatic role of REV1, which has been shown to have deoxycytidyl transferase activity. Consistent differences between the rev1 and rev3 strains suggest that not all all Polζ activity is REV1-dependent. Incorporation of dA could potentially be carried out by Pol α/δ, although Pol η is thought to be primarily responsible for extending the inserted nucleotide across from an AP site[33, 34].
It is worth noting that the bypass efficiencies for AP sites and THF determined using ssOligo COD(+7)-S4 and COD(+7)-F4, respectively were much higher that those reported earlier(59−65% versus 5−10%)(Table II). Most of the earlier studies on AP site bypass were done using a limited sequence context imposed by the experimental system[14-16]. Our earlier study was performed in yeast Δapn1 mutant, using a 30-mer containing an AP site located precisely at the deleted A in the lys2AΔ746 allele, thus generating an AP site in a 1-nt loop . Since the bypass efficiency of an AP using ssOligo COD(+7)-S4 is similar in both wild-type and Δapn1 mutant, it is unlikely that the large difference observed in the bypass efficiencies of the AP site between these two studies is due to the lack of APN1. It is also unlikely that the difference in bypass efficiency between the current and the past studies is due to the difference in the size of the loop generated when the transforming ssOligos anneal to the replicating fork. Since bypass of an AP site only occurs during the subsequent round of DNA replication, there will be no loop when the bypass machinery encounters the lesion (see Figure 2, Supplemental Data). The simplest and most likely reason for the difference in bypass efficiencies observed between the current and earlier studies may be the difference in sequence context that surrounds the AP site. The effect of sequence context is clearly demonstrated when we transformed yeast with both the COD and NC series of oligonucleotides. Depending on the location of AP site in the 7-nt loop, we observed a range of bypass efficiency for the AP site (from 15% to as high as 74%). It appears that as the AP site is moving away from the central location of the loop, bypass efficiency for AP site gradually decreased from 74% (position S4, Figure 1) to 7% (position S1, Figure 1, Supplemental Data)(Table IV). As indicated in Figure 2, APN1 activity is severely inhibited when AP site is located within the 7-nt loop, except when the AP site is located at the 5’ base of the loop(position S1). APN1 was found to exhibit significant nicking at the AP site at the S1 position (Fig.2). It is therefore likely that the significantly lower number of transformants obtained when COD(+7)-S1 was used as the transforming ssOligo could be the result of the combine effect of APN1 nicking and sequence context. Despite the fact that only a limited number of sequence context was examined in this manner, it is clear that sequence context have a significant effect on the bypass efficiency of an AP site.
Sequence context can significantly modulate the mutagenic potential of a lesion by affecting the efficiency of repair and subsequent bypass of the lesion. Observations from many laboratories have demonstrated the effect of sequence context on the ability of a DNA repair enzyme to repair DNA damage. For example, it has been shown that removal of 7-methylguanine can vary up to 185 fold from position to position within the PGK1 coding region. Sequences context not only affects the repair efficiencies of the base excision repair enzymes[36-39], but also that of the nucleotide excision repair complex[40, 41]. The role of sequence context has also been demonstrated to exert substantial effect on the efficiency of lesion bypass carried out by DNA polymerases. The ability of a polymerase to bypass a lesion depends on the combined kinetics of the incorporation and extension reactions catalyzed by the polymerase[42-47]. In agreement with the in vitro studies, our current study clearly demonstrated that the in vivo bypass of an AP site can occur at different efficiency and is affected by the sequence context. Despite the fact that bypass efficiencies of an AP site was affected by sequence context, however, our data indicated that bypass specificity was not affected by the sequence context. Since insertion of dC opposite an AP site is predominantly carried out by REV1, our data thus suggest that that sequence context has a significant effect on the interaction of REV1 at the primer terminus across from an AP site. The mechanism by which sequence context affects the binding of REV1 at the primer terminus is thus of significant interest, and certainly a topic for future studies.
Our earlier data indicated that MutSα and MutSβ are distributed differently on the two replication strands. MutSα is more efficient on the lagging strand, but MutSβ is distributed approximately equally on the two strands, or perhaps even somewhat greater on the leading strand . Since the AP site is located within the 7-nt loop, it is expected that the loop, rather than the lesion is recognized more efficiently by the mismatch repair (MMR) system. In this case, we expect that AP sites located at different position within the 7-nt loop will not lead to significant differences in discrimination by MMR. Despite the fact that the DNA sequences on the leading and lagging strand are different, our data in Table VI suggest that the direction of replication did not appears to have a significant effect on the bypass of an AP site. This observation is in agreement with the recent observations of Pages et al .
The authors thank Dr. Sue Jinks-Robertson for critical reading and provided valuable suggestions for the preparation of this manuscript. This work was supported by NIH grants CA90860 (Y.W.K.).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.