INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Chimeric RNA/DNA oligonucleotides (chimeras) have been used to direct single base changes in episomal and chromosomal targets in mammalian cells (Cole-Strauss et al., 1996; Yoon et al., 1996; Alexeev and Yoon, 1998; Kren et al., 1998; Lai and Lien, 1999). These molecules have also been effective in mediating similar reactions in plant cells (Beetham et al., 1999; Zhu et al., 1999). Independently, these two groups demonstrated that mutations can be corrected in tobacco (Nicotiana tabacum) and maize (Zea mays) cells. Using a marker gene system, Zhu et al. (1999) corrected a mutation in vivo and showed that the targeted base was in fact altered as directed by the chimeric oligonucleotide. Although Beetham et al. (1999) were able to recover herbicide-resistant plant cells due to the action of the chimera, the targeted base in tobacco cells was not changed. Instead, the nucleotide located at the 5' side of the target base was mutated. Because such a change would also produce an amino acid alteration that would confer herbicide resistance, the plant cells were recoverable under the appropriate selection. These results were the first to demonstrate degeneracy in targeted gene repair because experiments in mammalian cells produced only precise nucleotide alterations.

The process by which these nucleotide conversions are made is still undefined, but recent evidence suggests that mismatch repair plays a critical role in mammalian cells. Using cell-free extracts from HuH7 cells, Cole-Strauss et al. (1999) demonstrated that these chimeras can correct both point and frameshift mutations and that the reaction is reduced significantly in extracts that lack a functional mismatch repair system. In addition, the presence of antibodies directed against hmsh2, the human homolog of the MutS protein from Escherichia coli, significantly decreases the efficiency of the chimera-based reaction. Although a large body of information exists for bacterial, yeast, and mammalian DNA repair systems, there is a paucity of experimental evidence for defining similar reactions in plant cells (Britt, 1996). Given that DNA repair processes impact broad areas of basic and applied plant research including cell cycle control and aspects of recombination, this field of plant biology is rapidly expanding. The current lack of knowledge may be due, in part, to plant model systems being less genetically tractable than more thoroughly studied organisms. As described above however, chimeric RNA/DNA oligonucleotide-directed gene conversion is applicable to plant systems and a growing body of evidence suggests that these molecules may have wide applications in plant functional genomics and varietal improvement (for commentary, see Hohn and Puchta, 1999).

To address several of these issues, we investigated the possibility that these molecules would be active in plant cell-free extracts. We were particularly interested to see if the degenerate repair behavior in tobacco could be confirmed in a biochemical system. Fundamentally, however there were two objectives in this investigation. First, recapitulating the reaction in a controlled environment and establishing a genetically tractable system would be helpful in understanding the mechanism(s) of chimera-directed gene correction in plants. Second, such a system could be useful in identifying certain DNA repair pathways in plant cells that act differently than those described thus far in mammalian or bacterial cells.

In this manuscript, we use cell-free extracts from monocotyledonous (banana) and dicotyledonous (tobacco) plant species and an embryonic source (maize) in experiments engaging the chimeric oligonucleotide to direct correction of point and frameshift mutations. The results of our studies indicate that these cells possess the machinery necessary to catalyze correction of both types of mutations.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

A model system was developed for the detection of gene correction using plasmid molecules containing point or frameshift mutations in the coding regions of antibiotic-resistant genes. The plasmid pK^Sm4021 (Fig. 1A) contains the mutated kanamycin (kan) gene and a wild-type ampicillin-resistant gene; plasmid pT^Sm153 contains a mutated tetracycline (tet) gene and the same ampicillin-resistant gene (Fig. 1B). Expression of the ampicillin (amp) gene permits normalization of the E. coli transformation process so that direct comparisons can be made among reaction mixtures. The plasmid and appropriate chimera are mixed with the extract. After a defined time, the plasmid DNA is extracted and transformed into competent E. coli cells harboring a mutation in the RECA gene. Previous results established the need for functional RecA protein in the bacterial system (Metz et al., 1998). Hence, the use of cells deficient in RecA function ensures that any correction observed after the phenotypic readout had occurred in the cell-free extract. These correction events are scored by selection on agar plates containing kanamycin or tetracycline depending on the plasmid assayed. A dilution from the same transformation was plated in duplicate and selected on plates containing ampicillin to normalize the efficiency of electroporation. Frequencies were calculated as kanamycin/tetracycline revertant colonies relative to ampicillin colonies selected from the same reaction sample.

View larger version (31K): [in this window] [in a new window]   Figure 1.   Targeted plasmid sequences and chimeras designed to repair indicated mutations. The plasmids displayed above contain mutations at the indicated sites. These mutations are in the coding regions of genes that confer antibiotic resistance. Plasmids pKSm4021 (A) and pTSm153 (B) contain point mutations, whereas pTS208 (C) harbors a frameshift mutation. The sequences of the wild-type, mutant, and converted bases are listed below the specific chimeric oligonucleotide designed to correct the mutation. DNA bases are presented in uppercase and RNA nucleotides are presented in lowercase. Targeted bases are highlighted in red. D, The chimeric oligonucleotide, SC1, is a non-specific control bearing no sequence homology to any of the gene targets.

A final, but important feature of plasmid pK^Sm4021 is the target sequence itself. Wild-type sequence conferring antibiotic resistance contains a T residue at position 4,018. This base was mutated to a G, disabling functional gene activity. To avoid the possibility of positive results emanating from contaminating sources, the chimera was designed to convert the G residue to a C, instead of a T. This switch still generates a functional protein thereby preserving the phenotypic readout as kanamycin resistance. Figure 1 illustrates the series of chimeric oligonucleotides used in this study. Kan4021C should direct correction (Fig. 1A); SC1 is a non-specific chimera and should not elicit any change (Fig. 1D).

Figure 1 also displays a second substitutory system utilizing a point mutation in the gene responsible for tetracycline resistance at position 153 (Fig. 1C). Similar to the kan system, the chimera Tet153T was designed to correct the mutation. To ensure that the mutated plasmid was corrected to enable tetracycline resistance and to avoid artifactual results due to contamination of wild-type plasmids, a silent mutation (AG) was engineered at position 325. DNA sequence verification of antibiotic-resistant colonies was performed so that sites 153 and 325 were analyzed simultaneously.

A model system for the mechanism of correction directed by these chimeric oligonucleotides (Fig. 2) has been postulated (Gamper et al., 2000). Upon target recognition, the chimeric oligonucleotide forms a joint molecule known as a complement-stabilized D-loop. Once stabilized, the complex is recognized by the cell's inherent DNA repair activities and through interaction with the appropriate factors, sets in motion a cascade of events that directs nucleotide exchange or insertion at the specific site on the target strand. We hypothesize that the repair activity takes place through a pathway that includes homologs of the MutS protein (Cole-Strauss et al., 1999).

View larger version (12K): [in this window] [in a new window]   Figure 2.   Proposed model for chimera-directed gene repair. The chimeric oligonucleotide interacts with the DNA target, in a process known as homologous base pairing, forming a double D-loop juncture. The joint molecule complex is recognized by the cell's DNA repair machinery. The mis-paired bases are corrected through the activity of the mismatch repair pathway. The overall process is envisioned as a concerted series of steps using enzymes involved in the process of homologous recombination for the pairing phase, followed by the repairing phase, which is dependent at least on the protein, msh2, or its analogs.

In this study we used extracts from monocots and dicots, as well as embryonic tissue. Embryos were obtained from maize seeds; the source of the monocot extract was banana; and tobacco Nt-1 cell suspensions served as the source for the dicot extract. Cell-free extracts were prepared using the strategy of Cole-Strauss et al. (1999) with slight modifications as outlined in "Materials and Methods." Central among the changes was the use of liquid nitrogen to freeze the samples for grinding with a mortar and pestle. The extract was mixed with plasmid DNA and the chimeric oligonucleotide in a reaction buffer containing NTPs and dNTPs. After incubation, the samples were extracted with phenol/chloroform and precipitated with ethanol. The plasmid DNA was then electroporated into a mutant strain of E. coli, containing a mutation in the RecA gene (DH10B). The bacteria were plated on agar containing the appropriate antibiotic and allowed to grow for 18 h at 37°C.

Kanamycin-resistant colonies are present in samples containing the maize embryo, banana, and tobacco extracts (Table I). There appears to be no significant difference in the frequency of resistant colonies between monocot and dicot extracts, but a reduced number of colonies appear when the conversion reaction occurs in the embryo extract. The number of colonies in each table reflects colony count per 10⁷ ampicillin-resistant colonies. Similar results were obtained in the tetracycline system with all three extracts enabling correction. The same trend is apparent as the monocot and dicot extracts contain a significantly higher level of repair activity when compared to the maize embryo extract. The conversion required for kan resistance is GC, whereas tetracycline resistance is conferred upon AT transversion, and the base pair mismatches created by the respective chimeras are G/G and A/A, respectively. Both are purine-purine mismatches and are among the most efficiently repaired, as judged by mammalian cell experiments (Lahue et al., 1989; Holmes et al., 1990). The response in both systems was dose-dependent and successful correction relied on the presence of the designated extract. The maximal frequency of conversion observed in these experiments was approximately 0.08%.

A series of control experiments was performed to establish which components of the reaction were necessary for repair (Table II). Complete reaction mixtures produce a similar number of colonies compared to the data presented in Table I, whereas the absence of plasmid, chimera, or extract resulted in no antibiotic-resistant colonies. Requirements were the same for the tetracycline system (data not shown). Important controls are represented in lines 10 through 12 of Table II, wherein the plasmid and the chimera were incubated separately with the extract, and the DNA was purified and mixed prior to electroporation. With these reaction parameters no colonies were observed, reinforcing the fact that the measured correction events are occurring in the plant cell extract and not in the bacterial cells.

Conversion at the DNA level was measured by sequencing plasmids isolated from antibiotic-resistant bacterial colonies. DNA sequence analyses indicate that the kanamycin-sensitive mutant base G has been converted to the base C (Fig. 3). A total of 60 randomly selected colonies were sequenced (30 from banana and 30 from maize) and gave the same results when either the maize embryo extract or the banana extract catalyzed the reaction (Table III). Sequencing of the non-coding strand confirmed that both strands were repaired. Similar results were obtained in the tetracycline system (Fig. 4), wherein an A to T conversion is observed. Figure 4 also presents the coding strand of the tetracycline gene harboring the silent mutation, and in all cases, the G residue was present. Hence, these results suggest that the change from antibiotic sensitivity to antibiotic resistance is the result of a unique nucleotide exchange at position 4,021 (kan^r) or 153 (tet^r).

View larger version (30K): [in this window] [in a new window]   Figure 3.   Correction of the 4021 kan mutation. The targeted plasmid and sequence are displayed, as well as the DNA sequence of the resulting clones exhibiting resistance to kanamycin. The indicated extract is listed in the left side of the panel and the altered base from the coding strand of the target is highlighted in yellow. Without treatment, the G residue is observed. The chimeric oligonucleotide used in the reaction is listed with the source of the cell-free extract. The term 4021 mix indicates the presence of a multiple base readout at the target site, in this case printed as an "N" within the sequence.

View this table: [in this window] [in a new window]   Table III.   DNA sequence analyses of mutant antibiotic resistance gene corrections Underlined nucleotides represent targeted  observed changes, respectively. Nos. indicate the no. of a particular conversion type observed.  indicates missing nucleotide in target sequence. Conversion types in which more than one conversion species are present are indicated by [mix].

View larger version (52K): [in this window] [in a new window]   Figure 4.   Correction of the 153 tet mutation. Plasmid and chimera are listed as described in the legend to Figure 3 except that the target was the tet mutation with an A residue designated for alteration. The left panel represents the sequence data from colonies exhibiting resistance to tetracycline. The right panel displays the sequence of the silent marker base created within this plasmid to identify the correct plasmid construct. The specific chimeric oligonucleotide used in the reaction is listed with the source of the cell-free extract. 153 mix refers to the presence of multiple peaks appearing in the sequence at the target site.

The activity of the tobacco cell-free extract also produced kanamycin and tetracycline-resistant colonies, but the targeted nucleotide exhibited a mixed base sequence for both genes (Figs. 3 and 4). In the kanamycin system, the mixed nucleotides are C (converted) and G (mutant) (see highlighted region; C, blue peak; G, black peak) evidenced by the placement of an "N" at the targeted sequence site. In the tetracycline system, the mixture is A and T (see highlighted region), and although an "N" is not inscribed, a mixture of red (T) and green (A) peaks is apparent. Because the silent mutation (G) is present at position 325, the mixed sequence was not the result of wild-type plasmid contamination. Rather, the mixed sequence may have resulted either from correction of only one of the two strands followed by plasmid replication or from correction of only a subset of the plasmid molecule population (see "Discussion").

Thus far, the cell-free extracts have been shown to repair point mutations. Using the same genetic readout system, the capacity of these extracts to repair frameshift mutations can also be tested. As shown in Figure 1, the plasmid pT^S208 contains a deleted base at position 208 rendering the tet^r gene non-functional. Chimeric oligonucleotide Tet208C was designed to correct this mutation by directing the insertion of a C residue. The insertion event restores the reading frame enabling antibiotic resistance. Each plant extract has the capacity to repair frameshift mutations directed by the chimera (Table IV). The frequency of conversion is at least one log lower than that observed in the repair of point mutations, similar to results seen by Cole-Strauss et al. (1999) in mammalian cell-free extracts. DNA sequence analyses of mutant and converted plasmids, the latter picked from colonies growing on tet⁺ agar plates, were performed (Fig. 5). The extracts generated from maize and banana produced targeted gene repair at position 208 in all the clones sequenced. The results from the tobacco extract experiments were, however, different. Although some clones contained the designed T insertion at the correct site, another population of tetracycline-resistant colonies was found. These molecules contained either a mixture of G and A residues at the correct position or a C residue six bases downstream from the targeted site (Table III). Interestingly, this alternative insertion site was the only one found in the clones and, as in the case of the point mutation conversion, was only present in experiments using the tobacco extract.

View larger version (30K): [in this window] [in a new window]   Figure 5.   Correction of the 208 tet mutation. Plasmid pTS208 contains a frameshift mutation at nucleotide position 208 (note triangular marker in mutant sequence listing). Sequence data from resistant colonies resulting from treatment with the indicated cell-free extract are displayed with the targeted site highlighted in yellow inserted base C, 208 mix refers to the presence of multiple peaks appearing in the sequence at the site of insertion, here depicted by an "N". In the bottom panel, nucleotide position 214 is highlighted due to the non-specific insertion of a C residue at that site.

The concept of gene repair using the chimeric oligonucleotide relies on the presence of RNA in the molecule. Recent evidence has confirmed the importance of this RNA region in stabilizing the conjunction of the chimera with the target site (Gamper et al., 2000). To test the activity of an oligonucleotide having the same structure as Kan4021C, but replacing the RNA residues with DNA bases, we utilized the banana cell-free extract as it has routinely demonstrated the highest level of repair activity. A direct comparison was made between Kan4021C and Kan4021C-DNA in correcting the mutation in pK^Sm4021 (Fig. 1). The action of both oligonucleotides produced antibiotic-resistant colonies with the chimeric oligonucleotide (Kan4021C) demonstrating a 2-fold higher level of activity (Table V). Colonies from each plate were selected, the plasmid DNA was extracted, and the sequence was analyzed. As predicted, all kanamycin-resistant colonies tested from reactions containing the chimeric oligonucleotide and banana cell-free extracts gave the targeted correction (Table III). In contrast, only six of 16 colonies from the reaction containing the all DNA oligonucleotide harbored plasmid molecules with the targeted sequence alteration. The other 10 colonies contained altered sequence variations (Table V). The all DNA oligonucleotide is designated KAN4021C-DNA. Hence, 62.5% of the colonies tested contained plasmids with an incorrect, non-targeted base change. We have never observed random, non-targeted correction using the chimeric oligonucleotide and in sequencing over 150 plasmid molecules from kanamycin-resistant colonies from this and previously published work (Cole-Strauss et al., 1999).

View this table: [in this window] [in a new window]   Table V.   DNA sequence analyses of pKsm4021 corrections directed by chimeric and all-DNA oligonucleotides Comparison of the types (and nos.) of DNA sequence changes directed by either chimeric or all-DNA oligonucleotides. Underlined nucleotides represent targeted  observed changes, respectively.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

This manuscript reports the use of chimeric oligonucleotides and DNA hairpins for correction of point and frameshift mutations in cell-free extracts from plants. Extracts from both monocot and dicot cells, as well as from embryonic tissue support the repair of mutations. By using mutant strains of E. coli lacking RecA protein activity as a genetic readout system, the results establish sustained inheritance and clonal expansion of corrected DNA templates. Sequence analyses of these clones confirm genetic repair at the DNA level.

Although such repair systems have been developed in mammalian cells (Lahue et al., 1989; Holmes et al., 1990; Thomas et al., 1991), to our knowledge, the plant system described herein is the first of its kind. Based on this study, many similarities appear to exist between mammalian and plant cells in chimera-directed gene repair, including the capacity to correct multiple types of mismatches, the facile reaction conditions, and the time of correction. There are, however, some interesting differences. First, the frequency of correction is lower in plant cell-free extracts due, we believe, in large part to the increased DNA degradation by endogenous plant nucleases. In pilot studies, we have observed a high level of nuclease activity on plasmid DNA, particularly in the maize embryo extract (data not shown). Surprisingly, the chimeric oligonucleotide remains stable and intact during the course of all reactions. We are now developing extract preparation protocols and reaction conditions that reduce the inherent nuclease activity. Nuclease activity may influence the type of repair pathway used to correct the mismatched base. Second, the tobacco cell-free extract exhibited incomplete and promiscuous gene conversion events, whereas no non-targeted mutagenesis was found with banana or maize embryo extracts. Preliminary experiments using Arabidopsis cell-free extracts have not revealed targeted, non-specificity in repair events (M. Rice, unpublished observations). Hence, this may not be a universal property of dicots, but rather particular to cultured Nt-1 tobacco cells, although much more work is required to establish this concept. In any case, these results differ significantly from those obtained using mammalian cell-free extracts where all clones sequenced revealed precise targeted base conversion.

Beetham et al. (1999) have recently reported non-targeted slippage in gene correction reactions. Using chimeric oligonucleotides, these workers modified codon 196 of the acetolactate synthase gene in cultured tobacco cells, conferring resistance to high levels of the sulfonylurea herbicide chlorsulfuron (Glean, DuPont-Dow Elastomers L.L.C., Wilmington, DE). Although the targeted gene conversion was successful in generating the desired phenotype, DNA sequence analysis revealed that the converted base was consistently one position 5' to the targeted nucleotide. In an accompanying paper, Zhu et al. (1999) reported precise targeted correction using chimeric oligonucleotides in maize. To a lesser extent, imprecise targeted corrections were also observed. Our results agree with both observations with regards to the degree of targeted precision in both maize and tobacco. The system described herein provides a biochemical means to examine the differences in precise versus imprecise chimera-directed gene conversion.

The observations described above were made using a selectable genetic system necessitated by the frequency of repair. A similar situation exists using mammalian cell-free extracts (Cole-Strauss et al., 1999; Gamper et al., 2000) and poses relevant questions. Is random or non-targeted mutagenesis occurring in chimera-directed gene repair? Because, in most cases, the colonies or clones isolated for study have been identified as a function of selective pressure or phenotypic change, other events that do not confer resistance could occur, but go undetected. Beetham et al. (1999) were provided a window to observe such events in the tobacco system because, fortuitously, the second base of codon 196 was chosen as a target and the alteration of the proximal 5' base (a non-targeted event) also confers resistance to chlorsulfuron. It may not simply be a characteristic of tobacco, but rather a general phenomenon that merits further investigation. The results reported herein confirm the observation of Beetham et al. (1999) at the biochemical level. To this end, we have constructed chimeric oligonucleotides with a wide variety of structural modifications and have tested them for gene repair activity in mammalian cell-free extracts. The results indicate that two of these new designs create non-targeted mutations that confer antibiotic resistance as measured by the genetic readout system, but do not direct targeted correction (Gamper et al., 2000). It is possible that certain molecular structures, occurring as the chimeric oligonucleotide falls into homologous register with its target sequence, enable mis-corrections or mutations.

Degeneracy in targeted correction was also observed when an all DNA oligonucleotide, designed to adopt the same double hairpin configuration as the chimera, was used to convert the kanamycin mutation in banana cell-free extracts. Over 60% of the isolated plasmid molecules had a variety of altered bases within the specific codon. Again, based on the design of the genetic readout system, only non-targeted changes that enable antibiotic resistance will be observed. Sequencing 200 bases upstream or downstream from the targeted codon revealed no non-specific, non-targeted mutations. We cannot, however, rule out such mutagenic behavior on plasmids that would not confer kanamycin resistance. This phenomenon is also observed when extracts from maize embryo and tobacco are used (data not shown); the same distribution of mutations is present. Hence, this second type of mutagenic activity may be a function of the all DNA oligonucleotide rather than a property of a particular type of plant extract, as opposed to the results obtained with the tobacco extract and the chimeric oligonucleotide. Because work in the mammalian cells (Cole-Strauss et al., 1996; Yoon et al., 1996; Kren et al., 1998) established that DNA hairpins could not repair mutations, these results are particularly intriguing and indicate the presence of potentially different repair pathways in plants.

The mechanism of chimera-directed gene repair is far from elucidated, but mutational studies in mammalian cells indicate that the repair phase of the reaction is mediated by a type of mismatch repair system (Cole-Strauss et al., 1999; Gamper et al., 2000). Among the strongest evidence is the inhibition of gene repair by antibodies directed against the hmsh2 protein. This protein is critical for the establishment of the repair complex, enabling the binding of other mismatch repair proteins. Similar studies using antibodies directed against a plant homolog of hmsh2 in conjunction with the plant cell-free extract readout system are currently under way. Whether or not this pathway will turn out to be active in the chimera-directed targeted conversion or participate in the non-targeted reaction awaits further experimentation.

The development of this system enables the examination of structure-function relationships for plant cells. Simple experiments can be conducted to measure the capacity of different plant types or tissues to catalyze gene conversion events. One may also be able to identify chimera structures that direct specific correction and not non-targeted mutation or vice versa for a specific plant gene. This information would likely be of importance in the selection of plant materials for genetic alteration. For example, cell culture conditions, enhancing environmental stimuli, cell cycle position, and meiotic versus mitotic tissue are parameters or conditions that can be addressed using the cell-free extracts described in this paper. Finally, this approach will form the basis for the isolation and characterization of plant DNA repair proteins by providing a reproducible and simple assay system for their activities, and for assaying plant mutants with a DNA repair deficient phenotype.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Plant Materials

Maize (Zea mays) seeds (line B3733, Pioneer Hi-Bred, Des Moines, IA) were imbibed in a sealable plastic box containing paper towels saturated with distilled water. Seeds were incubated at 25°C for 18 to 36 h. Embryos were dissected from imbibed seed at 18, 24, and 36 h post-imbibition, frozen in liquid nitrogen, and stored at 80°C. Tobacco (Nicotiana tabacum) Nt-1 cell suspensions were maintained as shaker cultures (27°C, 200 rpm in a 250 mL flask) and transferred weekly to fresh CSM suspension medium containing: Murashige and Skoog salts (Gibco-BRL, Grand Island, NY), 500 mg L¹ MES, 1 mg L¹ thiamine, 100 mg L¹ myoinositol, 180 mg L¹ KH₂PO₄, 2.21 mg L¹ 2,4-diclorophenoxyacetic acid (2, 4-D), and 30 g L¹ Suc (pH 5.7). Banana (Musa acuminata cv Rasthali) cell suspensions (the kind gift of T.R. Ganapathi) were maintained as shaker cultures (27°C, 80 rpm in a 125 mL flask) and transferred every 10 d to fresh M2 cell suspension medium (Cote et al., 1996). Dense Nt-1 and banana cell suspensions were centrifuged in 50-mL disposable centrifuge tubes at 700g for 5 min at room temperature. Following centrifugation, the liquid medium was decanted, and the pelleted cells were frozen in liquid nitrogen and stored at 80°C.

Preparation of Cell-Free Extracts

Cell-free extracts were prepared from imbibed maize embryos and tobacco Nt-1 and banana cell suspensions by a modification of Cole-Strauss et al. (1999). Plant samples were ground under liquid nitrogen with a mortar and pestle. Three milliliters of the ground plant tissue were extracted in 1.5 mL of extraction buffer (20 mM HEPES, pH 7.5, 5 mM KCl, 1.5 mM MgCl₂, 10 mM DTT, 10% [v/v] glycerol, and 1% [w/v] PVP). Samples were then homogenized with 15 strokes of a Dounce homogenizer. Following homogenization, samples were incubated on ice for 1 h and centrifuged at 3,000g for 5 min to remove plant cell debris. Protein concentrations of the supernatants were determined by Bradford assay. Extracts were dispensed into 100-µg aliquots, frozen in a dry ice-ethanol bath, and stored at 80°C.

Plasmids

Kanamycin and tetracycline selectable markers were used in two substitutory systems to determine nucleotide exchange in the cell-free extracts. The kanamycin-sensitive plasmid pK^Sm4021 contains a single base transversion (TG), which creates a TAG stop codon in the kan gene at codon 22. A tetracycline-sensitive plasmid pT^Sm153 carries a single TA nucleotide change at position 153 in the pBR322 plasmid, which creates a stop codon in the tet gene at codon 23. A nucleotide insertional system with a tetracycline-sensitive plasmid, pT^S208, was used to analyze repair of single base deletions in cell-free extracts. The plasmid carries a single nucleotide deletion at position 208, which creates a frameshift in the tet gene of pBR322 at codon 41. The plasmids also contain a wild-type ampicillin gene used for propagation and normalization (Cole-Strauss et al., 1999).

Oligonucleotides

Synthetic oligonucleotides were used to direct reversion of kan^S and tet^S genes to restore resistance to their respective antibiotics. Chimeric RNA/DNA oligonucleotides, Kan4021C and KanGGrv, which can direct conversion of the kan^S gene in pK^Sm4021 at codon 22 from TAG to TAC (stopTyr), were synthesized as previously described (Cole-Strauss et al., 1999). Chimeric RNA/DNA oligonucleotides Tet153C and Tet208C were used to revert the tet^S genes of plasmids pT^Sm153 and pT^S208, respectively at their mutated bases. An all DNA oligonucleotide Kan4021-DNA and the non-specific chimera SC1 (Cole-Strauss et al., 1996) were used for comparison and as controls.

In Vitro Assays

Reaction Conditions

1

1

Electroporation, Plating, and Selection

1

1

1

1

1

4