Genetics, Vol. 163, 527-538, February 2003, Copyright © 2003
Targeted Nucleotide Repair of cyc1 Mutations in Saccharomyces cerevisiae Directed by Modified Single-Stranded DNA Oligonucleotides
Erin E. Brachmana and
Eric B. Kmieca
a Department of Biology, Delaware Biotechnology Institute, University of Delaware, Newark, Delaware 19716
Corresponding author:
Eric B. Kmiec, University of Delaware, Delaware Biotechnology Institute, 15 Innovation Way, Newark, DE 19711., ekmiec{at}udel.edu (E-mail)
Communicating editor: M. HAMPSEY
 | ABSTRACT |
|---|
Modified single-stranded DNA oligonucleotides have been used to direct base changes in the CYC1 gene of Saccharomyces cerevisiae. In this process, the oligonucleotide is believed to hybridize to the target site through the action of a DNA recombinase and, once bound, DNA repair enzymes act to excise the nucleotide, replace it, and revert the gene to wild-type status. Nucleotide exchange exhibits a strand bias as, in most cases, a higher level of base reversal appears in cells in which the oligonucleotide is designed to hybridize to the nontemplate strand. But, in one case, a higher level was observed when an oligonucleotide complementary to the transcribed strand was used. Mutant haploid and diploid strains are reverted to wild type at this locus with approximately the same frequency and all strains take up the oligonucleotide with approximately equal efficiency. Some repair preference for certain base mismatches was observed; for example, T/T and C/C mispairs exhibited the highest degree of reactivity. Finally, we demonstrate that proteins involved in DNA pairing can enhance the repair activity up to 22-fold, while others affect the reaction minimally. Taken together, these results confirm the importance and versatility of yeast as a model system to elucidate the factors regulating the frequency of nucleotide exchange directed by oligonucleotides.
SYNTHETIC oligonucleotides can be used to cause base pair changes in eukaryotic genomes. Pioneering studies by Sherman and colleagues (YAMAMOTO et al. 1992A 
, YAMAMOTO et al. 1992B ) demonstrated that the CYC1 gene from Saccharomyces cerevisiae could be mutated by the transformation of a single-stranded DNA oligonucleotide. Transformed colonies were selected by growth in the presence of a nonfermental carbon source, and the reaction was found to be dependent on several factors. These included a strand bias for the template or transcribed strand, as a higher number of transformed colonies were obtained when the oligonucleotide was designed to hybridize to the template strand of the CYC1 gene (YAMAMOTO et al. 1992A , YAMAMOTO et al. 1992B ). The reaction was also dose dependent and the reversal of phenotype required the presence of specific oligonucleotides.
We have been examining the potential use of oligonucleotides to direct nucleotide exchanges in episomal and chromosomal genes. Original designs, such as the double-stranded RNA/DNA oligonucleotide, were found to correct mutations in a variety of mammalian, plant, and yeast cells (see BRACHMAN and KMIEC 2002 for review). Biochemical studies aimed at elucidating the mechanism of this correction reaction revealed that the all-DNA region of the RNA/DNA duplex hairpin was the active component in the vast majority of cases where the target site was successfully altered. In fact, subsequent experiments demonstrated that modified single-stranded DNA vectors lacking any RNA component were highly active in promoting targeted nucleotide exchange (TNE; GAMPER et al. 2000C ; LIU et al. 2001 ). These single-stranded DNA moieties were particularly active in extracts made from S. cerevisiae and mutant strains deficient in a single gene known to be involved in a recombination or repair pathway(s). The in vitro assay system consisted of a mutated kanamycin-resistant gene harbored in a plasmid, which upon correction in the cell-free extract, conferred kanamycin resistance after transformation into Escherichia coli (RICE et al. 2001 ). Further studies extended the versatility of the single-stranded DNA oligonucleotide to extracts made from mammalian and plant cells. In each of these systems, repair of the kanamycin mutation occurred at a higher frequency when a single-stranded oligonucleotide was used in place of a double-stranded molecule. Unmodified single-stranded DNA molecules were only 20% as active as similar molecules containing nuclease-resistant phosphorothioate linkages (GAMPER et al. 2000C ). Thus, we have routinely employed single-stranded DNA oligonucleotide "vectors" that contain phosphorothioate linkages. While this requirement is not consistent with the work of MOERSCHELL et al. 1988 and YAMAMOTO et al. 1992A , YAMAMOTO et al. 1992B , cited above, these earlier studies used a 20-fold higher level of oligonucleotide. Perhaps by a mass effect, unmodified molecules were effective in their system, whereas these levels promote cell toxicity in our systems.
An in vivo system was recently established, and repair of mutations in a hygromycin fusion gene was demonstrated as both point and frameshift mutations were corrected by the introduction of a modified single-stranded oligonucleotide (LIU et al. 2001 ). Oligonucleotides designed to hybridize to the nontemplate (nontranscribed) strand were more effective than those complementary to the template strand in directing nucleotide exchange, rendering the transformed yeast cells resistant to hygromycin. This study was followed by an examination of some of the proteins required for correction of the hygromycin mutation at both the episomal and the chromosomal levels (LIU et al. 2002A ). RAD51 and RAD54 were found to be required, while RAD52 was found not to be essential. This latter fact differentiates our work, in part, from the technique described by STORICI et al. 2001 in which single-stranded oligonucleotides assist in the insertion of a double-stranded fragment into the yeast genome; this reaction is RAD52 dependent.
To extend the nucleotide alteration technique to yeast chromosomal genes and to examine further the mechanism of modified single-stranded vectors, the CYC1 locus was chosen as a target. As stated above, this particular system was used initially by YAMAMOTO et al. 1992A to correct mutations in the CYC1 gene. Upon further development, a variety of strains containing different point mutations in the Cys-22 codon were created (HAMPSEY 1991 ). The strains contain mutations that render the mutant line unable to grow on plates containing glycerol as the carbon source. The reversion rate of each is low (1 x 10-8), and thus the repair of the specific substitutions could be detectable even if it occurred at a low level. In addition, these strains enable an investigation of base correction preference within a specific DNA context. Since each strain contains a different nucleotide substitution, experiments can be designed to test whether one mismatched base pair, created by the conjunction of the oligonucleotide with its target, is corrected more efficiently than other mismatches. In this report we demonstrate that repair of a chromosomal mutation in yeast is facilitated by modified single-stranded DNA oligonucleotides at an average frequency >100-fold higher than the reversion rate. In addition, we confirm a strand preference for nucleotide exchange and demonstrate that overexpression of certain yeast genes known to be involved in DNA pairing enhances the frequency of nucleotide exchange.
| MATERIALS AND METHODS |
|---|
Yeast strains:
Haploid strains YMH1 and YMH2 (MAT
cyc1-706::CYH2 cyc7-67 ura3-52 leu2-3, 112 cyh2) and diploid strains YMH51–YMH55 [created by crossing YMH1–YMH5 with strain B-7462 (MATa cyc1-1 cyc7-67ura3-52 his1-1 can1-100)] were a generous gift from Michael Hampsey (Rutgers University; see HAMPSEY 1991 ). Derivatives of strain S260-11B (isogenic yeast strains) differ from each other by only a single base pair substitution within codon 22 of the CYC1 gene.
Oligonucleotides and uptake assay:
Cyc1/70T and Cyc1/70NT are single-stranded DNA oligonucleotides 70 nucleotides long that contain three phosphorothioate linkages at both the 3' and the 5' termini (see Fig 1). Cyc1/70T targets the transcribed strand of the mutant CYC1 gene, while Cyc1/70NT targets the nontranscribed strand. Hyg3S/74T and Hyg3S/74NT (LIU et al. 2001 ) are oligonucleotides bearing no complementarity to the CYC1 target, but can direct repair of a mutant hygromycin gene. All oligonucleotides were synthesized by Integrated DNA Technologies (Iowa City, IA) and purified and processed according to GAMPER et al. 2000A . For both uptake experiments, the yeast cells were grown to an OD of 
2 x 107 cells/ml in YPD (yeast extract, peptone, dextrose) medium, incubated with dithiothreitol (DTT), washed twice with sterile H2O and once with 1 M sorbitol, and resuspended in 120 µl of 1 M sorbitol, and then aliquots (40 µl) were electroporated with oligonucleotide. An aliquot from each cyc1 strain was electroporated with 5 µg of a Texas-red fluorescent-labeled oligonucleotide vector and recovered in YPD/1 M sorbitol for 1 hr. Samples were imaged with a Zeiss 51D LSM confocal microscope. Additionally, an aliquot from YMH52, YMH53, and YMH55 was electroporated with 3 µl of 4 µM 32P-labeled 70-mer and recovered in 1 ml YPD/1 M sorbitol for 30 min, followed by two washes of 1 M sorbitol. Radioactivity inside the cells was detected and quantified using a LS6500 scintillation counter (Beckman, Fullerton, CA).
View larger version (45K):
In this window
In a new window
Download PPT slide
|
Figure 1.
Genetic targets and vectors. Mutant yeast strains used in all of the experiments are listed and the single-base substitution (mutation) is shown in boldface type; all mutations occur at codon 22, a functionally critical cysteine. Cyc1/70T and Cyc1/70NT are single-stranded oligonucleotides 70 nucleotides (nt) long, designated with a T to target the transcribed strand of the cyc1 gene or designated with an NT to target the nontranscribed strand. Hyg3S/74T and Hyg3S/74NT are oligonucleotides with a length of 74 nt bearing no complementarity to the cyc1 target site, but are active in targeting the mutation in the hygromycin gene. Phosphorothioate linkages (*) are located at the terminal three linkages (3' and 5').
|
|
Chromosomal targeting of the mutant cyc1 gene:
Five micrograms of oligonucleotide were electroporated into a variety of S. cerevisiae yeast strains—designated YMH1, YMH2, YMH51, YMH52, YMH53, YMH54, and YMH55—using the same protocol described by LIU et al. 2001 . Briefly, yeast cells were grown in YPD or selective media (SC-URA) at 30° to a density of 2 x 107 cells/ml, incubated with 1 M DTT, and then harvested and washed twice with dH2O and once with 1 M sorbitol. The cells were resuspended in 120 µl of 1 M sorbitol and 40-µl aliquots were electroporated with designated oligomers using a Gene Pulser apparatus (Bio-Rad, Gaithersburg, MD; 1.5 kV, 25 µF, 200 ohms, 1 pulse, 5 sec/pulse length). A recovery period of 16 hr in 3 ml of YPD supplemented with 1 M sorbitol took place before the cells were plated (100 µl) at 10-4 dilution on YPD (500 µl or 1 ml) and on YPG (1% yeast extract, 2% peptone, 3% glycerol, and 2% agar). YPG plates were cultured at 30° for 7 days as described by HAMPSEY 1991 .
Converted CYC1 genes were confirmed by sequencing a PCR-amplified product. Colonies were picked at random from a YPG plate and diluted in 50 µl of distilled water. One microliter of yeast cell solution was added to a PCR reaction mixture [1x PCR amplification buffer (Roche), 300 µM dNTP, primer OJW-24, and primer ORB-27 (HAMPSEY 1991 ) with Taq polymerase]. Samples were preheated at 92° for 10 sec, 52° for 30 sec, 68° for 1 min, a final elongation at 68° for 8 min, and finally held at 4°. PCR products were confirmed by DNA sequence analyses using an ABI310 sequencer.
Plasmid DNA constructs:
Plasmid pAURHyg(rep)eGFP, used in the experiment outlined in Table 3, is described in LIU et al. 2001 . Expression plasmids pYNURad55, pYNUMre11, pYNUXrs2, and pYNUß were constructed by inserting the genes RAD55, MRE11, XRS2, and 
ß into plasmid pYNU132, a derivative of pYN132 containing the URA3 marker in place of the TRP1 marker (a generous gift from W. K. Holloman, Cornell University). Gene expression is under the regulation of the constitutive TPI promoter. Briefly, RAD55, MRE11, and XRS2 were amplified from yeast strain LSY678 genomic DNA using the following primer sets: Rad55F (5'-CGACATATGTCGCTTGGTATACCACTTTCCCA) and Rad55R (5'-GATCTCGAGTTAACCTTCACTATCATAAATTATCTCCTCCT); Mre11F (5'-CAGCATATGGACTATCCTGATCCAGACACA) and Mre11R (5'-GATCTCGAGCTATTTTCTTTTCTTAGCAAGGAGACTTCCAAG); and Xrs2F (5'-CGACATATGTGGGTAGTACGATACCAGAATACATTGGAAG) and Xrs2R (5'-GATCTCGAGTCATTATGGTTTTGTTCTTTTGAACGTAAACTTCGGAC).
After an NdeI and XhoI (New England Biolabs, Beverly, MA) digest of the PCR products, they were ligated into pYNU132. -Phage ß was amplified from plasmid pTP248 (a kind gift from Anthony Poteete, University of Massachusetts) using the following primer set: ßF (5'-GCCTAAGCTTCACCATGAGTACTGCACTCGCAACGCTG-3') and ßR (5'-CTTACTCGAGCTATCACGTTGTGAACTTCTGAAGC-3'), designed to add flanking restriction sites. The PCR product was cut with HindIII and XhoI (New England Biolabs) and ligated into pYNU132. pYNURad51, pYNURad52, and pYNURad54 plasmid construction is identical to that described in LIU et al. 2002B except that the pYNU132 plasmid was used in place of pYN132. Insertion and orientation of all the clones was confirmed by DNA sequencing.
Enhancement of gene repair:
Yeast strain YMH55 was transformed with the following plasmids separately: pYNU132, pYNURad51, pYNURad52, pYNURad54, pYNURad55, pYNUMre11, pYNUXrs2, and pYNUß. Cells were plated on SC-URA media and the presence of the plasmids was confirmed by colony PCR. After electroporation, selection took place by spreading 1 ml or 500 µl of yeast cells (undiluted) onto YPG plates (1% yeast extract, 2% peptone, 3% glycerol, and 2% agar). In addition, 0.1 ml of the yeast cells, diluted 10-4, was plated on YPD plates. Colony counts of selected (YPG) and nonselected (YPD) yeast were determined using an AccuCount TM 1000 (Biologics).
| RESULTS |
|---|
The yeast strain S260-11B (MAT cyc1-706::CYH2cyc7-67 ura3-52 leu2-3 112 cyh2) was used by HAMPSEY 1991 to create a series of isogenic yeast strains differing only in a single base mutation at codon 22 (Cys-22) of the CYC1 gene. We used a subset of these, listed in Fig 1, to examine targeted repair; YMH1 and YMH2 are haploid strains, while YMH51, YMH52, YMH53, YMH54, and YMH55 are diploid. The mutated base is indicated in boldface type and, in each case, corresponds to either the first or the second position in codon 22. These mutations change the amino acid from Cys to Arg (YMH2 and YMH52), Ser (YMH53 and YMH55), or Gly (YMH54). These mutations render each strain unable to utilize glycerol as a carbon source and, thus, these strains will not grow on agar plates containing glycerol instead of dextrose. While some of the indicated strains are diploid, they are actually hemizygous (cyc1/cyc1-1) at the CYC1 locus. The mutant strains have a spontaneous reversion rate somewhere between 1.3 x 10-8 and 1 x 10-9 (HAMPSEY 1991 ) and, in general, low background enables the use of these strains for experiments involving nucleotide exchange. The oligonucleotides used in the correction events are also depicted in Fig 1. Cyc1/70T is a 70-mer that is complementary to the transcriptional or template (T) strand, except for the targeted nucleotide, which is located at the center of the molecule. Cyc1/70NT is also a 70-mer but is designed to target the nontemplate strand of the cyc1- gene in the same fashion. Hyg3S/74T is a nonspecific oligonucleotide bearing no complementarity to the target. An additional oligonucleotide, identical to Hyg3S/74T, also containing a fluorescent, Texas-red dye conjugated to the 5' end (not shown), was used to test for cellular uptake. The mechanism by which the repair of a mutation is facilitated is not fully elucidated; thus we have chosen to refer to the overall process as targeted nucleotide exchange (TNE).
A critical parameter for achieving TNE is the efficiency of the electroporation process. Previously, we established conditions for the electroporation of oligonucleotides into yeast, attaining 
80% uptake efficiency (LIU et al. 2001 ). Hence, we employed the same test system to examine the efficiency of electroporation of the Texas-red conjugated oligonucleotide into each yeast strain under investigation, visualizing the cells by confocal microscopy. As shown in Fig 2, each strain exhibits a high degree of transformation as the majority of cells in each field of vision appear red, indicating the presence of the oligonucleotide. The photographs were generated using a Zeiss 63XC-Apochromate water immersion lens; thus certain layers of yeast may not appear red because they do not fall within the specific focal plane. In a second assay, we quantified the uptake efficiency in three specific cell lines, because they exhibit the most interesting examples of strand bias (see below). Electrocompetency was measured by transforming strains YMH52, YMH53, or YMH55 with a 32P-labeled 70-mer oligonucleotide. The cells were grown to a density of 107 cells/ml in 40 ml of YPD medium and the cells were processed for electroporation. The cells were then recovered in 1 ml of YPD for 30 min, and the uptake efficiency was detected using a scintillation counter. As shown in Table 1, all three strains exhibit comparable electrocompetency as measured by this more quantitative assay. On the basis of these results and comparable ones previously analyzed (LIU et al. 2001 ), we conclude that each yeast strain has a comparable capacity to be transformed with the modified single-stranded DNA oligonucleotides.
View larger version (110K):
In this window
In a new window
Download PPT slide
|
Figure 2.
Transformation of various mutant strains. Five micrograms of a Texas-red fluorescent-labeled oligonucleotide vector was electroporated into the indicated strains. Samples were loaded into a Lab-Tak II chambered coverglass system and imaged with a Zeiss 63 X C-Apochromate water immersion lens. Cells were imaged on a Zeiss inverted 100 M Axioskop equipped with Zeiss 51D LSM confocal microscope and a kryptonargon laser (488 and 568 excitation lines) by using the 568-nm excitation line with a 590-nm excitation line with a 590-nm longpass emission filter for fluorescence detection.
|
|
The yeast strains presented in Fig 1 were tested for nucleotide repair activity. YMH1 and YMH51 served as controls to ensure that electroporation of the oligonucleotide vector did not cause a cellular toxicity. As shown in Table 2, both wild-type strains can grow heartily on YPD and YPG plates, as predicted. Unlike the wild-type strains, the mutant strains cannot grow in YPG unless the specific base change in each mutant strain is corrected. Transformation of the nonspecific oligonucleotide Hyg3S/74T does not produce colonies on YPG plates. Colonies do appear after the cells have been treated with either Cyc1/70T or Cyc1/70NT (Table 2). The correction efficiency (C.E.) is calculated by dividing the number of colonies able to grow on YPG by the total number of colonies on YPD (adjusted for plating volume); this value normalizes for transformation and survival variability. Note that in the haploid strain YMH2 only a modest difference is observed when the nontemplate (NT; C/A mismatch) and the T (G/T mismatch) strands are compared. The other strains are also amendable to TNE and exhibit a strand bias favoring the nontemplate strand in most cases. The highest overall level of gene repair appears in YMH55, where the largest degree of strand bias (toward the NT strand) is also seen. In contrast, YMH53 exhibits a significant level of gene repair, yet the strand bias is reversed in comparison to YMH55. In this case, targeting the template strand elicits a higher level of repair than targeting its nontemplate counterpart. It is also noteworthy that the hemizygous target, including YMH52, exhibits only a slightly higher level of correction than a haploid counterpart (YMH2). In addition, the efficiency of mismatch repair with regard to specific unpaired bases varies somewhat from a previously established hierarchy of preferential base (mismatch) repair (BISHOP et al. 1989B ). This hierarchy, however, came from an experimental protocol in which a preformed partial heteroduplex containing a designed, mismatched base pair was transformed into yeast. Our system does not utilize preformed complexes, nor does it involve the transformation of a partially duplex circular molecule. In our experiments, the oligonucleotide is electroporated directly into the cell and must first pair at a specific site and then direct the repair of a mutation in a chromosomal gene. Thus, differences in repair frequencies between the components of each system are not totally unexpected.
The two strains exhibiting the highest level of nucleotide exchange, YMH53 and YMH55, also produce opposite results in the preference for strand bias. Thus, to explore this result in greater detail and to test the limits of the system, a dose curve was carried out. As seen in Fig 3, each strain responds to increasing levels of oligonucleotide by displaying a corresponding increase in the number of converted colonies. Furthermore, the observation that YMH55 exhibits a strand bias of repair toward the nontemplate strand, while YMH53 shows the opposite bias, holds true at each level of oligonucleotide transformed into the cells.
View larger version (58K):
In this window
In a new window
Download PPT slide
|
Figure 3.
Nucleotide exchange occurs in a dose-dependent fashion. Strains YMH55 and YMH53 were electroporated with varying amounts of the oligonucleotide vectors Cyc1/70NT and Cyc1/70T, and gene repair activity was quantified after plating. Average YPG colony counts and YPD colony counts are presented as well as YPG colony counts per 105 YPD colonies.
|
|
Strain YMH53 exhibits a decided bias for repair with increased correction seen on the transcribed strand of the target. Several factors may contribute to preferential correction, including the possibility that strand bias is dictated by an inherent biochemical process throughout the cell. To examine this issue, we utilized a plasmid containing a mutated hygromycin gene, pAURHyg(rep)eGFP, which had previously been shown to be a target for gene repair by single-stranded oligonucleotides (LIU et al. 2001 ). Plasmid pAURHyg(rep)eGFP confers resistance to hygromycin and permits growth on agar plates upon correction. Thus, we introduced this plasmid into YMH53 and maintained it under aureobasidin selection. Two types of oligonucleotides were electroporated into this strain: Hyg3S/74NT or Hyg3S/74T and Cyc1/70NT or Cyc1/70T. The intent was to promote a dual repair reaction in which Hyg3S/74NT or Hyg3S/74T (see Fig 1) targeted the episomal hygromycin mutation, while Cyc1/70NT or Cyc1/70T directed correction of the chromosomal (cyc1) mutation. As shown in Table 3, strand bias was consistently observed in the repair of the chromosomal mutation, independent of which Hyg3S vector was used to repair the episomal mutation. As in all other examples, gene repair in YMH53 occurs at a higher frequency when the transcribed strand is targeted. In contrast, correction of the hygromycin mutation in the episome does not reveal preferential selection of the transcribed strand. These results are not entirely consistent with the results of LIU et al. 2001 in which a distinct strand bias for episomal correction was seen. One clear difference here is that the YMH53 cell line grows more slowly in the presence of the plasmid and in comparison to LSY678, the strain used by LIU et al. 2001 . In addition, the number and levels of oligonucleotides were different; in this case, two separate oligonucleotides were coelectroporated, whereas the previous experiments used a single oligonucleotide experiment. Taken together, these data suggest that factors determining strand bias in YMH53 do not regulate preferential strand selection during episomal targeting. And thus, preferential strand correction at the CYC1 locus may be controlled by the chromosomal environment.
The data presented in Table 2 also suggest that a difference in TNE activity may exist between the isogenic haploid strain YMH2 and the diploid strain YMH52, which, in fact, is hemizygous at the CYC1 locus. We compared the level of repair in each of these strains more closely by increasing the level of oligonucleotide vector. As shown in Table 4, YMH52 is generally more proficient in gene repair at various doses, but the difference is less than twofold, even when 10 µg is used. This result is comparable to the data presented in Table 1 (at 5 µg) and, since only a modest difference is observed, we believe that the haploid or diploid state of the cell does not contribute significantly to repair activity in this system.
A DNA repair gene likely to be involved in the pairing phase of the TNE reaction is RAD51 (RICE et al. 2001 ; LIU et al. 2002A ), a member of the RAD52 epistasis group. RAD51 is activated by DNA damage and catalyzes DNA strand transfer, a critical step in the rejoining of broken ends. This function has led some to propose that Rad51p may participate in the general process of DNA pairing. Mechanistically, Rad51p can assimilate a single strand of DNA into a helix, search for homology, and begin the strand exchange process. Because the oligonucleotide used in our system is single stranded and the first step in the repair reaction is likely to involve DNA pairing, we wanted to examine the possibility that elevating the levels of Rad51p in YMH55 would increase the level of DNA pairing, thereby leading to a higher level of TNE. Hence, the RAD51 gene from S. cerevisiae was cloned into the overexpression vector, pYNU132, placing it under the control of the constitutive promoter, TPI. This vector contains CEN and ARS elements maintaining it in low copy number. The plasmid was then transformed into yeast, and cells harboring the overexpression vector were selected by growth in SC-URA media (see MATERIALS AND METHODS). As a control, plasmid pYNU132, containing no yeast gene, was introduced into YMH55 and maintained under selection. This strain was transformed with 5 µg of Cyc1/70NT and repair activity was assessed by colony growth on YPG. As shown in Table 5, colonies appear on YPG, producing an average correction efficiency of 0.04. This value is lower than that obtained in YMH55 bearing no pYNU plasmid, a difference likely attributable to the double selection growth conditions on glycerol and in SC(URA-) media (see DANHASH et al. 1991 ). Even under these conditions, the presence of pYNURad51 led to a 15-fold enhancement of gene repair compared to the control strain containing only pYNU132 (Table 5). DNA sequence analyses of the targeted region revealed that the specific base was changed in a precise manner (data not shown).
This result prompted us to test several other members of the RAD52 epistasis group, including RAD52, RAD54, MRE11, XRS2, and RAD55, for the same enhancement capacity. The experimental protocol was identical to that described above for testing the effect of overexpressing RAD51. All of these genes were cloned individually into pYNU132 and again expressed under the control of the TPI promoter. Expression of each gene was confirmed by increased resistance to methyl methanesulfonate (data not shown, but see LIU et al. 2001 ). As shown again in Table 4, a wide range of effects are seen. Overexpression of RAD52 has a modest, yet positive, effect on TNE, while the overexpression of MRE11, XRS2, and RAD54 has little or no effect. One gene, RAD55, appears to elevate the frequency substantially, raising the level 22-fold (±5.7). Rad55p is often complexed with RAD57, forming a heterodimer that can catalyze DNA pairing through strand assimilation (see VASQUEZ et al. 1999 ; KREJCI et al. 2001 ). But, Rad55p can also act as a checkpoint protein, regulating the type of genetic exchange taking place in the cell (gene conversion or crossover; BASHKIROV et al. 2000 ). It was this activity that prompted us to test RAD55 for the potential of enhancing the repair reaction.
Recently, ß-protein, of the -bacteriophage red-gam locus, was shown to catalyze homologous recombination and gene targeting of exogenously added DNA in E. coli (YU et al. 2000 ). The biochemical activity of ß-protein, DNA annealing, as described initially by KMIEC and HOLLOMAN 1981 , is apparently critical for the targeting reaction. DNA annealing differs from strand assimilation in that proteins, such as ß-protein, promote simple hybridization of complementary single strands of DNA, but do not and cannot catalyze strand assimilation. From a purely biochemical standpoint, the S. cerevisiae protein possessing the most similar activity to that of ß-protein is Rad52p (MORTENSEN et al. 1996 ). Since overexpression of RAD52 led to some enhancement of gene repair activity (see above), we postulated that the same activity may be seen with ß-protein. Hence, the gene encoding ß-protein was cloned into pYNU132 and the plasmid was introduced into YMH55. (The plasmid was maintained stably under URA3 selection.) The oligonucleotide vector Cyc1/70NT was transformed into this strain and gene repair activity was assessed by colony growth on YPG plates. As shown in Table 5, overexpression of ß-protein enhances gene repair ninefold (±1.7). We also examined the stimulatory effect of each of these proteins on TNE directed by Cyc1/70T. While the same type of general enhancement is observed, the total number of colonies growing on YPG is, as expected, quite limited (data not shown). These results support the notion that elevating the levels of proteins involved in DNA pairing can increase gene repair frequency, even if these proteins come from other unrelated species.
| DISCUSSION |
|---|
Targeted nucleotide exchange directed by modified single-stranded oligonucleotides can be used to introduce changes into the genome at specific sites. In this article, we examine this process in yeast strains differing primarily by a single point mutation in the CYC1 gene (HAMPSEY 1991 ). These mutations, created at codon 22, shift a cysteine codon to codons for arginine, serine, or glycine. The resulting strains cannot grow on media containing glycerol as the sole carbon source. Here, we demonstrate repair of these mutations in all of these strains through the introduction of specific oligonucleotides capable of directing such conversion followed by selection in media containing glycerol.
The frequency of repair is likely to be affected by the efficiency with which the molecule enters the cell and the nucleus. By utilizing an oligonucleotide conjugated to a Texas-red fluorescent marker, we visualized cellular uptake by confocal microscopy. On the basis of these criteria, we found that each yeast strain in this study was amendable to electroporation rendering similar efficiencies of transformation. While this is only a rough evaluation of this critical parameter (see BANDYOPADHYAY et al. 1998 ), no striking differences among the strains were noticed. A second, more quantitative, assay confirmed that each strain used in this study had approximately the same level of electrocompetency. It is still unclear, however, how many molecules must actively enter the nucleus to ensure a robust level of TNE.
The mismatches created between the oligonucleotide and the target sequence differ in each strain. Strains YMH2, YMH52, and YMH54 exhibit a lower level of TNE activity than do YMH55 and YMH53 and contain heterologous mismatches when paired with their respective oligonucleotide vectors (T or NT). For example, the template mispair in YMH53 is T/T, while the nontemplate vector produces an A/A mismatch. TNE in YMH55 produces either a C/C mismatch with the NT vector or a G/G mismatch with the T vector. The results indicate that targeting the NT strand in YMH55 produces the highest level of correction in our system, even though this particular mismatch (C/C) has, in other assay systems, been inefficiently repaired. Earlier studies, aimed at defining a hierarchy for mismatched base repair, relied on a preformed partial heteroduplex (BISHOP et al. 1989A ) or conformation consisting of a two-stranded structure at the site of the mismatch. A likely reaction intermediate acted upon by the TNE machinery in our system is a three-stranded D-loop structure, a configuration in which the oligonucleotide hybridizes to its complement and displaces its homolog. In this case, one might predict that the cell's initial response is dictated more by the presence of a displacement loop than by the particular type of mismatched base enclosed within the D-loop, but the most obvious difference between the two systems is that the preformed templates containing a mismatch were episomal targets, and here we target a chromosomal mutation. Thus, it is likely that a unique hierarchy might exist for the TNE reaction on an integrated gene. Support for this notion comes from structural analyses of the various mismatched base pairs (CORNELIS et al. 1979 ; HO et al. 1985 ; BROWN et al. 1986 ; HARE et al. 1986 ; HUNTER et al. 1986 ). Mispairs that are corrected most efficiently appear to induce a more rigid deformation of the helix. Those that are more poorly corrected deform the helix into a so-called dynamic state resulting from cooperative hydrogen binding and reduced interhelical stacking. The latter configuration is most likely the one adopted by a C/C mismatch and, as such, it can escape recognition by the mismatch repair enzymes (HUNTER et al. 1986 ; WERNTGES et al. 1986 ). But complexes involving a D-loop containing heterologous inserts, such as a single mismatched base pair, are likely to adopt an "antirotational lock" conformation (BELOTSERKOVSKII et al. 1999 ). Thus, the mismatched base pair within the D-loop intermediate in our system may exist in a more rigid helical environment and would be more efficiently recognized by enzymes of the mismatch repair pathway. While this system is very limited and the analyses are far from complete, a tentative order of preferential repair based on our results can be put forward: G/A = G/T = A/A = C/T < G/G = C/A < T/T < C/C (note that the equals sign is an approximate). This list assumes that other factors are equivalent and that other forces in the cell contribute equally to each repair event.
A consistent characteristic of TNE that pervades all of the experimental results is the discrimination between the two DNA strands in the target. Strand bias in yeast using similar oligonucleotides was first established by MOERSCHELL et al. 1988 . These workers found that the template or transcribed strand was corrected more readily than the nontranscribed strand by synthetic vectors. Work from our laboratory confirmed that a strand bias for the correction of simple base mutations in yeast does exist (LIU et al. 2001 ), but we found that the nontranscribed strand was preferentially repaired in an episomal system containing a mutated hygromycin resistance gene. And we have recently shown that this bias is likely due to the action of RNA polymerase in which the oligonucleotide hybridized to the template strand is displaced by the movement of the polymerase (LIU et al. 2002A ). These two studies differed in many ways, including amounts of oligonucleotides being used, target site, and the strains in which the experiments were conducted. In this study, however, we employ strains and a target (CYC1 gene) that is similar to those used by YAMAMOTO et al. 1992A , YAMAMOTO et al. 1992B . And we find that, while the NT strand bias does dominate (YMH52, YMH54, and YMH55), there appears to be some reliance on the type of yeast strain used to determine preferential activity. For example, YMH53 exhibits the same strand bias reported by YAMAMOTO et al. 1992A , albeit to a lesser extent, while the other strains reveal a nontemplate preference for correction. But, the observation from YMH53 reconciles, to some degree, the observations reported by YAMAMOTO et al. 1992B and LIU et al. 2001 .
Since the level of transcription of the corrected gene and codon usage is likely to be the same for YMH53 and the other strains, the possibility exists that the new DNA sequence created either by the initial mutation or by its correction in YMH53 changed the position of a replication origin, activated a latent origin, or switched an early origin to a late origin. On the basis of the available data, however, the mutant or corrected DNA sequence does not correlate with any known yeast sequence perceived to be active in origin recognition, and we have recently shown experimentally that a new origin is not created by specific mutations or corrections in the CYC1 gene (E. E. BRACHMAN and E. B. KMIEC, unpublished data). But, this does not exclude replication as an important factor in determining strand bias. The CYC1 gene is positioned between two origins of replication, as defined by RAGHURAMAN et al. 2001 . A separate group (WYRICK et al. 2001 ) defined an inactive origin sequence located within the same region. If this origin became activated during the creation of either strain, YMH53 or YMH55, then the leading and lagging strands of replication passing through the CYC1 gene might be reversed. Recent evidence from ELLIS et al. 2001 suggests that the lagging strand in DNA replication is more amenable to gene correction as a function of the annealing of a single-stranded vector in E. coli. Thus, the large difference in repair activity when the T and NT strands are targeted in YMH53 or YMH55 may result from the activation of a silent origin near the target gene region. Experiments to determine the direction of fork movement through the CYC1 gene sequence are underway.
The overexpression of enzymes involved in DNA pairing can also affect the frequency and perhaps the bias of TNE. We have observed that the overexpression of Rad51p stimulates targeted repair of an integrated yeast gene (LIU et al. 2002B ). In this report, we show that TNE in YMH55 is robust and enhanced by the overexpression of genes involved in DNA recombination and/or repair. The fact that Rad51p and Rad55p can stimulate TNE is consistent with the fact that proteins involved in strand assimilation, the step uniting the oligonucleotide with the target DNA, is important and perhaps rate limiting in the overall process. The same is true for the overexpression of ß-protein, an enzyme involved in simple DNA annealing, consistent with the data of ELLIS et al. 2001 , but extending it to eukaryotes. While this protein acts in DNA pairing, it does so by annealing complementary strands of DNA but does not promote detectable levels of D-loops. DNA annealing could promote the association of the oligonucleotide with the target site on the lagging strand of a DNA replication fork or a partially gapped molecule (ELLIS et al. 2001 ). The lagging strand provides a natural gap between the Okazaki fragments, making the parental strand available to a single-stranded vector by a simple annealing reaction. Hence, this targeting path would require only the activity of proteins that catalyze DNA hybridization and not necessarily DNA strand transfer. Experiments testing the linkage of annealing proteins to "lagging strand targeting" and strand transferases to "leading strand targeting" are underway.
An alternative interpretation of these results involving homologous recombination centers on the concept that the oligonucleotides used as templates in any repair-based model may actually be promoting DNA insertion as described by STORICI et al. 2001 . In this "delitto perfecto" strategy, oligonucleotides are used to prime the insertion of a core DNA fragment, resulting in an inactivation or mutagenic event at a specific DNA site. While such a model could explain the results presented in this work, several important facts diminish this possibility. First, if the oligonucleotides were recombining into the genome, then one would presume that all types of mutations, replacement and frameshift, would be "corrected" with equal frequencies. All published works examining this issue have shown that replacement mutations are repaired at a higher frequency than frameshift mutations (COLE-STRAUSS et al. 1999 ; GAMPER et al. 2000B , GAMPER et al. 2000C ; LIU et al. 2001 , LIU et al. 2002A ). Second, an insertion/recombination model would be unlikely to exhibit the type of strand bias reported herein or in earlier studies (LIU et al. 2001 , LIU et al. 2002A , LIU et al. 2002B ). Third, mutations in mismatch repair genes, such as MSH2 or MSH3, should have minimal impact on this reaction or arguably their absence should elevate the frequency. In fact, a deletion of MSH2 reduces the frequency of repair significantly (COLE-STRAUSS et al. 1999 ). But MSH2 is also known to be involved in certain recombination pathways; thus there remains a possibility that, in MSH2-deficient strains, the reduction in TNE is a result of a lack of recombination activity. Support for this alternative explanation comes from elegant work done by Haber and colleagues (LEUNG et al. 1997 ). In this study, a long, single-stranded DNA molecule was found to pair with a chromosome and direct allelic repair by a recombination event; this reaction was blocked substantially by the mismatch repair protein Pms1p. This possibility cannot be excluded and biochemical results from cell-free extract experiments do reveal an elevated level of TNE when extracts deficient in Pms1p were used to catalyze the reaction (RICE et al. 2001 ). But in vivo analyses of TNE in strains deficient in MSH6 reveal a >60% reduction in activity (L. LIU, M. BRUNER and E. KMIEC, unpublished data). Since MSH6 is known to be involved only in mismatch repair, a compelling argument against a recombination model and in favor of a repair model can be made. More experiments are needed to establish the details of the initiation of TNE.
Finally, STORICI et al. 2001 state that delitto perfecto is dependent on RAD52; the reaction pathway(s) catalyzing our reaction can occur quite productively in the absence of RAD52 (LIU et al. 2002A ). While overexpression of RAD52 in the CYC1 system does elevate nucleotide exchange, strains deficient in RAD52 promote the reaction at levels that rival and exceed wild-type levels (RICE et al. 2001 ; LIU et al. 2002B ). We have postulated that the abundance of Rad52p as a result of overexpression may actually promote some degree of strand transfer (LIU et al. 2002B ), and a weak strand transfer activity promoted by Rad52p has been previously reported (KAGAWA et al. 2001 ).
Finally, the role of Rad55p in stimulating TNE is intriguing. While normally complexed with Rad57p in a heterodimer, reports have indicated that Rad55p has an important, individual role in directing the repair of damaged DNA toward a gene conversion pathway and away from a pathway involving crossover (BASHKIROV et al. 2000 ). Overexpression of the RAD55 gene in the current experimental protocol might promote the gene conversion pathway, as suggested by BASHKIROV et al. 2000 . This scenario would naturally elevate repair, since the process of gene conversion has some molecular features in common. Thus, Rad55p may be a key factor in enhancing the frequency simply because it promotes entry into a pathway that leads to gene conversion without crossover. In contrast, Mre11p, Rad54p, or Xrs2p, which may be involved in the processing of the terminal phases of recombinational repair, acting in a complex with Rad50p, may not influence this choice. Overexpression of these individually apparently does not facilitate elevated frequencies of TNE due perhaps to the inability to form these complexes at proper stoichiometry; additional experiments are underway to evaluate how the simultaneous expression of these genes influences the reaction. Studies such as these will lead to the formulation of a protein-driven nucleotide exchange strategy that may be translatable into higher eukaryotes.
| ACKNOWLEDGMENTS |
|---|
We thank Li Liu, Michael Rice, Anja van Brabant, and Hetal Parekh-Olmedo for helpful discussions during the course of this work. We also thank one of the reviewers for highlighting alternative explanations and suggesting specific experiments to challenge our hypotheses. We are grateful to Elizabeth Feather and Eric Roberts for manuscript preparation and editing. Support from the National Institutes of Health (1 RO1 CA89325-01A1) and NaPro BioTherapeutics is acknowledged.
Manuscript received July 15, 2002; Accepted for publication November 14, 2002.
| LITERATURE CITED |
|---|
BANDYOPADHYAY, P., B. T. KREN, X. MA, and C. J. STEER, 1998 Enhanced gene transfer into HuH-7 cells and primary rat hepatocytes using targeted liposomes and polyethylenimine. Biotechniques 25:282-292.[Medline]
BASHKIROV, V. I., J. S. KING, E. V. BASHKIROVA, J. SCHMUCKLI-MAURER, and W. D. HEYER, 2000 DNA repair protein Rad55 is a terminal substrate of the DNA damage checkpoints. Mol. Cell. Biol. 20:4393-4404.[Abstract/Free Full Text]
BELOTSERKOVSKII, B. P., G. REDDY, and D. A. ZARLING, 1999 DNA hybrids stabilized by heterologies. Biochemistry 38:10785-10792.[Medline]
BISHOP, D. K., J. ANDERSEN, and R. D. KOLODNER, 1989a Specificity of mismatch repair following transformation of Saccharomyces cerevisiae with heteroduplex plasmid DNA. Proc. Natl. Acad. Sci. USA 86:3713-3717.[Abstract/Free Full Text]
BISHOP, D. K., J. ANDERSEN, and R. D. KOLODNER, 1989b Specificity of mismatch repair following transformation of Saccharomyces cerevisiae with heteroduplex plasmid DNA. Proc. Natl. Acad. Sci. USA 86:3713-3717.
BRACHMAN, E. E. and E. B. KMIEC, 2002 The "biased" evolution of targeted gene repair. Curr. Opin. Mol. Ther. 4:171-176.[Medline]
BROWN, T., W. N. HUNTER, G. KNEALE, and O. KENNARD, 1986 Molecular structure of the G.A base pair in DNA and its implications for the mechanism of transversion mutations. Proc. Natl. Acad. Sci. USA 83:2402-2406.[Abstract/Free Full Text]
COLE-STRAUSS, A., H. GAMPER, W. K. HOLLOMAN, M. MUNOZ, and N. CHENG et al., 1999 Targeted gene repair directed by the chimeric RNA/DNA oligonucleotide in a mammalian cell-free extract. Nucleic Acids Res. 27:1323-1330.[Abstract/Free Full Text]
CORNELIS, A. G., J. H. HAASNOOT, J. F. DEN HARTOG, M. DE ROOIJ, and J. H. VAN BOOM et al., 1979 Local destabilisation of a DNA double helix by a T–T wobble pair. Nature 281:235-236.[Medline]
DANHASH, N., D. C. GARDNER, and S. G. OLIVER, 1991 Heritable damage to yeast caused by transformation. Biotechnology 9:179-182.[Medline]
ELLIS, H. M., D. YU, T. DITIZIO, and D. L. COURT, 2001 High efficiency mutagenesis, repair, and engineering of chromosomal DNA using single-stranded oligonucleotides. Proc. Natl. Acad. Sci. USA 98:6742-6746.[Abstract/Free Full Text]
GAMPER, H. B., A. COLE-STRAUSS, R. METZ, H. PAREKH, and R. KUMAR et al., 2000a A plausible mechanism for gene correction by chimeric oligonucleotides. Biochemistry 39:5808-5816.[Medline]
GAMPER, H. B., Y. M. HOU, and E. B. KMIEC, 2000b Evidence for a four-strand exchange catalyzed by the RecA protein. Biochemistry 39:15272-15281.[Medline]
GAMPER, H. B., H. PAREKH, M. C. RICE, M. BRUNER, and H. YOUKEY et al., 2000c The DNA strand of chimeric RNA/DNA oligonucleotides can direct gene repair/conversion activity in mammalian and plant cell-free extracts. Nucleic Acids Res. 28:4332-4339.[Abstract/Free Full Text]
HAMPSEY, M., 1991 A tester system for detecting each of the six base-pair substitutions in Saccharomyces cerevisiae by selecting for an essential cysteine in iso-1-cytochrome c. Genetics 128:59-67.[Abstract]
HARE, D., L. SHAPIRO, and D. J. PATEL, 1986 Extrahelical adenosine stacks into right-handed DNA: solution conformation of the d(C-G-C-A-G-A-G-C-T-C-G-C-G) duplex deduced from distance geometry analysis of nuclear Overhauser effect spectra. Biochemistry 25:7456-7464.[Medline]
HO, P. S., C. A. FREDERICK, G. J. QUIGLEY, G. A. VAN DER MAREL, and J. H. VAN BOOM et al., 1985 G.T wobble base-pairing in Z-DNA at 1.0 Å atomic resolution: the crystal structure of d(CGCGTG). EMBO J. 4:3617-3623.[Medline]
HUNTER, W. N., T. BROWN, N. N. ANAND, and O. KENNARD, 1986 Structure of an adenine-cytosine base pair in DNA and its implications for mismatch repair. Nature 320:552-555.[Medline]
KAGAWA, W., H. KURUMIZAKA, S. IKAWA, S. YOKOYAMA, and T. SHIBATA, 2001 Homologous pairing promoted by the human Rad52 protein. J. Biol. Chem. 276:35201-35208.[Abstract/Free Full Text]
KMIEC, E. and W. K. HOLLOMAN, 1981 Beta protein of bacteriophage lambda promotes renaturation of DNA. J. Biol. Chem. 256:12636-12639.[Abstract/Free Full Text]
KREJCI, L., J. DAMBORSKY, B. THOMSEN, M. DUNO, and C. BENDIXEN, 2001 Molecular dissection of interactions between Rad51 and members of the recombination-repair group. Mol. Cell. Biol. 21:966-976.[Abstract/Free Full Text]
LEUNG, W., A. MALKOVA, and J. E. HABER, 1997 Gene targeting by linear duplex DNA frequently occurs by assimilation of a single strand that is subject to preferential mismatch correction. Proc. Natl. Acad. Sci. USA 94:6851-6856.[Abstract/Free Full Text]
LIU, L., M. C. RICE, and E. B. KMIEC, 2001 In vivo gene repair of point and frameshift mutations directed by chimeric RNA/DNA oligonucleotides and modified single-stranded oligonucleotides. Nucleic Acids Res. 29:4238-4250.[Abstract/Free Full Text]
LIU, L., S. CHENG, A. J. VAN BRABANT, and E. B. KMIEC, 2002a Rad51p and Rad54p, but not Rad52p, elevate gene repair in Saccharomyces cerevisiae directed by modified single-stranded oligonucleotide vectors. Nucleic Acids Res. 31:2742-2750.
LIU, L., M. C. RICE, M. DRURY, S. CHENG, and H. GAMPER et al., 2002b Strand bias in targeted gene repair is influenced by transcriptional activity. Mol. Cell. Biol. 22:3852-3863.[Abstract/Free Full Text]
MOERSCHELL, R. P., S. TSUNASAWA, and F. SHERMAN, 1988 Transformation of yeast with synthetic oligonucleotides. Proc. Natl. Acad. Sci. USA 85:524-528.[Abstract/Free Full Text]
MORTENSEN, U. H., C. BENDIXEN, I. SUNJEVARIC, and R. ROTHSTEIN, 1996 DNA strand annealing is promoted by the yeast Rad52 protein. Proc. Natl. Acad. Sci. USA 93:10729-10734.[Abstract/Free Full Text]
RAGHURAMAN, M. K., E. A. WINZELER, D. COLLINGWOOD, S. HUNT, and L. WODICKA et al., 2001 Replication dynamics of the yeast genome. Science 294:115-121.[Abstract/Free Full Text]
RICE, M. C., M. BRUNER, K. CZYMMEK, and E. B. KMIEC, 2001 In vitro and in vivo nucleotide exchange directed by chimeric RNA/DNA oligonucleotides in Saccharomyces cerevisae. Mol. Microbiol. 40:857-868.[Medline]
STORICI, F., L. K. LEWIS, and M. A. RESNICK, 2001 In vivo site-directed mutagenesis using oligonucleotides. Nat. Biotechnol. 19:773-776.[Medline]
VASQUEZ, K. M., G. WANG, P. A. HAVRE, and P. M. GLAZER, 1999 Chromosomal mutations induced by triplex-forming oligonucleotides in mammalian cells. Nucleic Acids Res. 27:1176-1181.[Abstract/Free Full Text]
WERNTGES, H., G. STEGER, D. RIESNER, and H. J. FRITZ, 1986 Mismatches in DNA double strands: thermodynamic parameters and their correlation to repair efficiencies. Nucleic Acids Res. 14:3773-3790.[Abstract/Free Full Text]
WYRICK, J. J., J. G. APARICIO, T. CHEN, J. D. BARNETT, and E. G. JENNINGS et al., 2001 Genome-wide distribution of ORC and MCM proteins in S. cerevisiae: high-resolution mapping of replication origins. Science 294:2357-2360.[Abstract/Free Full Text]
YAMAMOTO, T., R. P. MOERSCHELL, L. P. WAKEM, D. FERGUSON, and F. SHERMAN, 1992a Parameters affecting the frequencies of transformation and co-transformation with synthetic oligonucleotides in yeast. Yeast 8:935-948.[Medline]
YAMAMOTO, T., R. P. MOERSCHELL, L. P. WAKEM, S. KOMAR-PANICUCCI, and F. SHERMAN, 1992b Strand-specificity in the transformation of yeast with synthetic oligonucleotides. Genetics 131:811-819.[Abstract]
YU, D., H. M. ELLIS, E. C. LEE, N. A. JENKINS, and N. G. COPELAND et al., 2000 An efficient recombination system for chromosome engineering in Escherichia coli. Proc. Natl. Acad. Sci. USA 97:5978-5983.[Abstract/Free Full Text]
TOP
ABSTRACT
MATERIALS AND METHODS
