Chloroplast lysates support directed mutagenesis via modified DNA and chimeric RNA/DNA oligonucleotides

The Plant Journal (2001) 27(3), 267±274

TECHNICAL ADVANCE

Chloroplast lysates support directed mutagenesis via modi®ed DNA and chimeric RNA/DNA oligonucleotides Eric B. Kmiec1, Christina Johnson1 and Gregory D. May2,* 1 Department of Biological Sciences, University of Delaware, Newark, Delaware 19716, USA, and 2 Plant Biology Division, The Samuel Roberts Noble Foundation, 2510 Sam Noble Parkway, Ardmore, Oklahoma 73401, USA Received 8 January 2001; revised 9 May 2001; accepted 15 May 2001. * For correspondence (fax +1 580 221 7380; e-mail [email protected]).

Summary Chimeric RNA/DNA and modi®ed DNA oligonucleotides have been shown to direct gene-conversion events in vitro through a process involving proteins from several DNA-repair pathways. Recent experiments have extended the utility of these molecules to plants, and we previously demonstrated that plant cell-free extracts are competent to support oligonucleotide-directed genetic repair. Using this system, we are studying Arabidopsis DNA-repair mutants and the role of plant proteins in the DNArepair process. Here we describe a method for investigating mechanisms of plastid DNA-repair pathways. Using a genetic readout system in bacteria and chimeric or modi®ed DNA oligonucleotides designed to direct the conversion of mutations in antibiotic resistance genes, we have developed an assay for genetic repair of mutations in a spinach chloroplast lysate system. We report genetic repair of point and frameshift mutations directed by both types of modi®ed oligonucleotides. This system enables the mechanistic study of plastid gene repair and facilitates the direct comparison between plant nuclear and organelle DNA-repair pathways. Keywords: chloroplasts, oligonucleotides, mutagenesis, chimeric, gene targeting.

Introduction Chimeric RNA/DNA (chimeras) and modi®ed DNA oligonucleotides direct site-speci®c base changes in episomal and chromosomal targets in mammalian and plant cells (for reviews see Kmiec, 1999; May and Kmiec, 2000). These molecules are designed to pair with homologous sequences within target sites, and therefore are situated to introduce a base change in genomic DNA. Utilizing this approach, Rice et al. (2000) described the development of a cell-free extract system to study the mechanism of targeted gene correction in plants, and as a tool to investigate plant DNA-repair pathways. This work provided a mechanistic explanation for the cell-based results of Beetham et al. (1999), Zhu et al. (1999) and Zhu et al. (2000). Here we extend oligonucleotide-directed gene targeting into a plastid DNA recombination and repair system. The chloroplast genome (plastome) of eukaryotic algae and higher plants exists as double-stranded DNA, closed ã 2001 Blackwell Science Ltd

circular molecules ranging from 80 to 200 kbp in size. Much of the chloroplast (ct) DNA synthesis occurs in young leaf cells, with copy numbers as high as 22 000 per cell during various stages of development. ctDNAs are redistributed to daughter organelles during plastid division. To maintain integrity of the plastome in mature leaves that are routinely subjected to high levels of UV irradiation, ef®cient DNA-repair pathways must exist in these organelles. Chloroplast DNA homologous recombination and repair activities have been well documented. The endosymbiont theory (Palmer and Delwiche, 1996) anticipates that chloroplast DNA recombination/repair pathways are related to their eubacterial origin. Supporting this suggestion is the report of cloning an Arabidopsis RecA protein with 53% identity to Escherichia coli RecA (Cerutti et al., 1992). Inhibition of ctDNA recombination and repair has 267

268

Eric B. Kmiec et al.

been accomplished through the use of dominant negative mutants of E. coli RecA (Cerutti et al. (1995). Excision repair-pathway enzyme activities have also been reported in chloroplasts of higher plants (Howland et al., 1975; McLennan, 1988). Furthermore, in vivo gene conversion of an episomal target using shuttle plasmids in chloroplasts has been demonstrated (Staub and Maliga, 1995). Although the plastome encodes many of the proteins required for plastid function (for review see Palmer, 1985), no DNA-repair proteins are reported to be encoded by the plastid genome (Ohyama et al., 1986; Shinozaki et al., 1986). Two Arabidopsis cDNAs that encode putative plastid-targeting domains have been shown to complement an ruvC recG double mutant of E. coli which is incapable of resolving cross-strand recombination intermediates (Pang et al., 1993). Based on these results, one might predict that a bioinformatics approach utilizing putative plastid targeting domains could be used to sort plant DNA recombination and repair enzymes. Such information could help identify candidate proteins and their homologues that are common or unique to plastid repair processes, or uncover repair apparatus shared between the plastid and the nucleus. The work described here provides an opportunity to evaluate plastid DNA-repair activities that maintain the integrity of the plastid and perhaps the nuclear genomes by establishing a functional assay system. The results of our studies will also elucidate both plastid- and nuclear

oligonucleotide-directed repair and homologous recombination mechanisms. Results and discussion Gene-repair assay Gene repair directed by chimeric oligonucleotides (COs) and modi®ed single-stranded oligonucleotides (MOs) can be monitored using an in vitro system. In this assay, cellfree extracts are mixed with a plasmid containing a speci®c mutation (point or frameshift mutations) and an oligonucleotide designed to correct the error. Such a system was ®rst developed for mammalian cell extracts (Cole-Strauss et al., 1999; Gamper et al., 2000a,b), and was recently extended to plants (Rice et al., 2000). The mutated gene cannot confer antibiotic resistance when introduced into E. coli unless the mutation is corrected. Quanti®cation of plasmid-repair events is scored by plating bacteria on agarose containing the appropriate antibiotic. Extracts of two types of chloroplast preparations were mixed with plasmid pKansm4021, which bears a point mutation at nucleotide 4021 located with coding region of the kanr gene. As shown in Figure 1, in the wild-type T has been changed to a G, and speci®c oligonucleotides (Figure 1) are designed to convert the G residue to a C residue. This conservative change enables functional protein production and confers kanamycin resistance to bacteria. The

Figure 1. Targeted plasmids, DNA sequences and oligonucleotides. Plasmids pKSm4021 and pTSD208 have been used previously (Cole-Strauss et al., 1999; Gamper et al., 2000a,b; Rice et al., 2000). pKSm4021 contains an intact ampicillin gene and a mutated kanamycin gene; nucleotide 4021 is altered from a T to a G, disabling antibiotic resistance. pTSD208 has a deleted C residue at position 208 while also containing a wild-type ampr gene. Chimeric oligonucleotide Kan4021C is designed to convert the point mutation in pKSm4021 from a G to a C, re-establishing the capacity to confer antibiotic resistance in E. coli. Kan 4021G forms a perfect match with the target sequence. Single-stranded vector 3S/25G converts the G residue to C in pKSm4021, while 3S/25A forms a perfect match. Chimeric oligonucleotide TetD208T is designed to insert a T residue at position 208, as does the single-strand vector 3S/ 28A. SC1 is a non-speci®c chimeric oligonucleotide (Cole-Strauss et al., 1996). The boxed base illustrates the position within the oligonucleotide that mismatches with the target sequence. The asterisks between bases in 3S/25G and 3S/28A, respectively, indicate the positions of the phosphorothioate linkages.

ã Blackwell Science Ltd, The Plant Journal, (2001), 27, 267±274

Chloroplast lysates support targeted mutagenesis Table 1. Oligonucleotide-directed gene repair in chloroplast extracts Plasmid (1 mg)

Oligonucleotide Pre-extract Post-extract Number (1.5 mg) (mg) (mg) observed

Chimeric oligonucleotide 1. pKSm4021 Kan 4021C 2. pKSm4021 Kan 4021C 3. pKSm4021 Kan 4021C 4. pKSm4021 Kan 4021C 5. pKSm4021 Kan 4021C 6. pKSm4021 Kan 4021C 7. pKSm4021 Kan 4021C 8. pKSm4021 Kan 4021C

2.5 5 10 20 ± ± ± ±

± ± ± ±

Modi®ed oligonucleotide 1. pKSm4021 Kan 3S/25G 2. pKSm4021 Kan 4021C 3. pKSm4021 Kan 4021C 4. pKSm4021 Kan 4021C 5. pKSm4021 Kan 4021C 6. pKSm4021 Kan 4021C 7. pKSm4021 Kan 4021C 8. pKSm4021 Kan 4021C

2.5 5 10 20 ± ± ± ±

± ± ± ±

2.5 5 10 20

2.5 5 10 20

61 135 199 235 53 103 191 273 77 123 229 315 94 379 584 612

Each reaction contained 1.0 mg plasmid DNA and 1.5 mg oligonucleotide mixed with the indicated amounts of extract. Genetic readout took place in DH10B (recA1); colony counts re¯ect the average of ®ve independent experiments. Variations among samples were less than 15%. Kanamycin-resistant colonies are per 107 ampicillin-resistant colonies that were quanti®ed by duplicate plating.

conversion of G®C rather than G®T ensures that resistance developed through conversion directed by the oligonucleotide, not through wild-type plasmid contamination. Figure 1 also illustrates plasmid pTSD208 in which a C residue has been removed, rendering the plasmid incapable of providing tetracycline resistance, and the two oligonucleotides used to direct the correction of pTsD208. Two control oligonucleotides are presented: one sequence matches perfectly the target region in pKansm4021 (Kan4021G), while the other (SC1) is a nonspeci®c targeting molecule bearing no sequence complementarity to the target. Two types of oligonucleotide are used in these reactions (Gamper et al., 2000a; Gamper et al., 2000b). The CO consists of complementary RNA and DNA residues folded into a double-hairpin con®guration resistant to cellular nucleases due to the 4T residues at each hairpin end. The MOs consist of a 25-mer and a 28-mer composed of all DNA residues, but having phosphorothioate linkages between the terminal four bases on each end. These molecules are also resistant to nuclease digestion in the cell-free extract. After the initial reaction mixture has been incubated at 37°C, the plasmids are isolated and electroporated into ã Blackwell Science Ltd, The Plant Journal, (2001), 27, 267±274

269

Table 2. Gene correction requires reaction components Plasmid (1 mg)

Oligonucleotide Pre-extract Post-extract Number (1.5 mg) (mg) (mg) observed

Chimeric oligonucleotide 1. pKanSm4021 Kan4021C 2. pKanSm4021 ± 3. pKanSm4021 ± 4. pKanSm4021 ± 5. pKanSm4021 Kan4021C 6. pKanSm4021 Kan4021C

± 10 ± ± 10 ±

± ± 10 ± ± 10

5 2 2 1 123 258

Modi®ed oligonucleotide 1. pKanSm4021 3S/25G 2. pKanSm4021 ± 3. pKanSm4021 ± 4. pKanSm4021 ± 5. pKanSm4021 3S/25G 6. pKanSm4021 3S/25G

± 10 ± ± 10 ±

± ± 10 ± ± 10

1 1 0 6 141 437

Reaction mixture contained the indicated components. Plasmid DNA was electroporated into DH10B cells and colony counts determined by antibiotic resistance. Results represent an average of ®ve independent reactions. Variations among samples were less than 15%.

E. coli strain DH10B. This strain is de®cient in RECA activity that is known to participate in recombination events in E. coli; previous work con®rmed that the repair reaction takes place in cell-free extract from plants (Rice et al., 2000). The plasmids also have an intact copy of an ampicillin-resistance gene, and therefore colonies arising on the kanamycin or tetracycline plates should be resistant to ampicillin. The ampicillin colonies also provide a way to normalize potential variations in colony counts due to electroporation. Two types of chloroplast preparations were utilized, and were prepared mechanically using standard methods. The ®rst, a `pre-gradient' preparation, was obtained by gentle resuspension of the pelleted chloroplasts following lowspeed centrifugation. Using a simple Percoll step gradient further puri®ed the second, `post-gradient' preparation. As shown in Table 1, both types of extract promote gene repair of plasmid pKansm4021. This table presents the average colony count of ®ve independent reactions for each type of extract at varying levels. A direct comparison is made between the chimeric oligonucleotide and the single-stranded vector, 3S/25G (Gamper et al., 2000b). In all four sets of reactions a dose-dependent response of extract is seen, but a maximal number of colonies is generated using 10±20 mg extract. The more puri®ed chloroplast extract supports a higher level of repair and, as predicted (Gamper et al., 2000b), 3S/25G is more ef®cient in directing the conversion reaction. Table 2 illustrates that all reaction components must be present for antibiotic-resistant colonies to arise. Spurious

270

Eric B. Kmiec et al.

Figure 3. DNA sequence of converted pTSD208 plasmids. Con®rmation of sequence alteration in isolated plasmid DNA, directed by the appropriate CO or MO. The correction involves the repair of a frameshift mutation through the insertion of a T (boxed).

Figure 2. DNA sequence of converted pKSm4021 plasmids. Con®rmation of sequence alteration in isolated plasmid, directed by the indicated chimeric oligonucleotide (CO) or the indicated modi®ed singlestranded oligonucleotide (MO), is displayed. This represents the repair of a point mutation (G®C); the altered residues are boxed.

colonies are occasionally found in some of the control plates, but after DNA sequencing analyses these few colonies do not harbor the corrected, targeted base and may appear due to random reversion. Plasmid DNA harbored in three colonies from each reaction point was isolated and processed for DNA sequencing. As shown in Figure 2, the DNA sequences indicate that chimeric oligonucleotides and singlestranded vectors direct precisely targeted gene repair. While only ®ve sequencing reactions are shown, all samples tested produced the same result. The complementary strand of the repaired plasmid target was also sequenced and was found to contain the proper complementary base at the correct position (data not shown). As the post-gradient extract was more highly puri®ed and likely to more closely re¯ect the contents of the

chloroplast fraction, we used only this source of extract for reactions to search for the optimal dosage of oligonucleotide. This experiment is also important because the singlestranded, 3S/25G, oligonucleotide is approximately 50% smaller that the chimeric oligonucleotide (70 nucleotides). In terms of molecules, a unit amount (mg) of MO would contain more correction vehicles than the same amount of CO. We adjusted the dose curve so that approximately the same numbers of molecules are present in each reaction. The results (Table 3) display a dose-dependency for both vectors, and con®rm that MOs are more ef®cient in directing gene repair even when the number of correction vehicles is the same. Finally, oligonucleotides were tested that either form a perfect match (Kan4021G) or are nonspeci®c (SC1) for the target site. No antibiotic-resistant colonies were generated at several different dosages. Correction of frameshift mutation These data indicate that a puri®ed chloroplast fraction can promote repair of a point mutation. Plasmid pTSD208 (Figure 1) contains a frameshift mutation at position 208 in the coding region of the tetracycline-resistance gene. This plasmid was mixed with the appropriate oligonucleotides, and the reaction initialized by the addition of the extract. As shown in Table 4 (reactions 1±10), correction of the frameshift mutation was enabled by the extract and either CO or MO. The colony number was reduced compared to the numbers found when the point mutation was targeted for repair. We have previously reported similar results when mammalian or plant cell extracts are employed (Cole-Strauss et al., 1999; Rice et al., 2000). The level of correction is again dependent on the amount of extract ã Blackwell Science Ltd, The Plant Journal, (2001), 27, 267±274

Chloroplast lysates support targeted mutagenesis

271

Table 3. Dosage dependence of gene repair Oligonucleotide (mg) Plasmid

Chimeric

Modi®ed

Post-extract (mg)

Number observed

1. pKanSm4021 2. pKanSm4021 3. pKanSm4021 4. pKanSm4021 5. pKanSm4021 6. pKanSm4021 7. pKanSm4021 8. pKanSm4021 9. pKanSm4021 10. pKanSm4021 11. pKanSm4021 12. pKanSm4021 13. pKanSm4021 14. pKanSm4021

3.0 ± 0.25 0.5 1.5 3.0 ± ± ± ± Kan4021G(0.5) SCI(0.5) Kan4021G(3.0) SCI(3.0)

± 3.0 ± ± ± ± 0.08 0.175 0.52 1.05 ± ± ± ±

± ± 10 10 10 10 10 10 10 10 10 10 10 10

1 0 30 47 116 271 81 212 317 399 0 0 0 0

Reactions contained plasmid (1 mg) and oligonucleotides at the indicated amounts and were initialized by addition of 10 mg post-extract. The number of kanr colonies was determined after electroporation in DH10B E. coli and counting on kanamycin plates. Kanr colonies are per 107 ampr colonies. Colony numbers represent three independent reactions. Variations among samples were less than 15%.

Table 4. Gene repair of frameshift and point mutations in a mutated tetr gene Plasmid (1 mg)

CO (1.5 mg)

MO (0.52 mg)

Post-extract (mg)

Number observed

1. pTSD208 2. pTSD208 3. pTSD208 4. pTSD208 5. pTSD208 6. pTSD208 7. pTSD208 8. pTSD208 9. pTSD208 10. pTSD208

TetD208T ± TetD208T TetD208T TetD208T TetD208T ± ± ± ±

± 3S/28A ± ± ± ± 3S/28A 3S/28A 3S/28A 3S/28A

± ±

0 0 13 29 42 71 19 37 59 93

2.5 5 10 20 2.5 5 10 20

Reactions contained the indicated components with varying amounts of extract. Reactions 1±10 contained pTS D208 and the appropriate oligonucleotide. Colony counts were determined by genetic readout in E. coli (DH10B). Number of kanr colonies is per 107 ampr colonies and represents the average of three independent reactions. Variations among samples were less than 15%.

added, and the number of colonies is higher when the single-stranded vector is used. However, in this case the difference in repair ef®ciency between the two types of vector is modest. This may re¯ect the dif®culty in repairing a frameshift mutation as opposed to a point mutation, as each type of event requires different members of the repair protein family. Three colonies from each set of reactions were selected and the plasmid DNA sequenced around the target site. Figure 3 illustrates a representative sequence from each set, and the speci®ed nucleotide (T) has been inserted at the targeted location. An important consideration is the potential for contaminating nuclear protein factors present in the chloroplast fractions possibly contributing to conversion activities. ã Blackwell Science Ltd, The Plant Journal, (2001), 27, 267±274

Figure 4. Gel electrophoretic and immunoblot analyses of whole-cell and chloroplast protein extracts. Proteins were separated from spinach whole-cell (lanes 2 and 4) and chloroplast (lanes 3 and 5) protein extracts. Lane 1 shows molecular weight markers. The gel was stained for protein with Coomassie blue (lanes 1 through 3) or immunoblotted with mouse antihistone H1 antibody.

Although further enrichment of the chloroplast fraction through the use of Percoll gradients resulted in increased repair activities with the modi®ed oligonucleotides, we wanted to determine if contaminating nuclear proteins

272

Eric B. Kmiec et al.

could be detected in the chloroplast fractions. To this end, immunoblot analyses were performed on spinach wholecell and chloroplast protein fractions using an antihistone H1 antiserum known to cross-react with plant histone proteins. No histone protein was detected in 20 mg of the chloroplast protein fraction (Figure 4). Additional experiments using a threefold excess (60 mg protein) of the chloroplast fraction over that of the whole-cell extract were likewise unable to detect the presence of histone protein in the chloroplast protein extracts (data not shown). The use of chimeric (Rice et al., 2000) and modi®ed (Gamper et al., 2000b) oligonucleotides in plant whole-cell extract DNA-repair assays has been reported previously. On average, both whole-cell and chloroplast protein fractions give similar levels of gene-correction activities on a per mg protein basis. Our results, and those reported by Gamper et al. (2000b), suggest that modi®ed oligonucleotides have greater repair activity than chimeras in both whole-cell and chloroplast protein extracts. The system and the results reported herein provide an opportunity to compare DNA-repair pathways that maintain the integrity of the plastid and nuclear genomes. As no DNA damage-repair proteins have been reported to be encoded by the plastid genome (Ohyama et al., 1986; Shinozaki et al., 1986), targeting domains may suggest which nuclear encoded DNA repair proteins are destined to the plastid. The ability to compare different and physically separate DNA-repair pathways between organelles within the same cell may help to elucidate factors effecting fundamental differences in homologous and illegitimate recombination mechanisms observed between plastid and nuclear genomes. Experimental procedures Plant materials Spinach (cv. Trias) was grown in a growth chamber from seed in MetroMix 350 for 4 weeks under 12 h, 20°days.

Preparation of chloroplast lysates Chloroplasts were mechanically isolated based on previously published methods (Whitehouse and Moore, 1993) Brie¯y, 50 g freshly harvested young spinach leaves were rinsed in ice-cold water, blotted dry to remove excess water, and deribbed. Leaf materials were ®nely sliced, placed in a chilled beaker containing 150 ml ice-cold isolation medium (330 mM sorbitol, 10 mM Na2P4O7, 5 mM MgCl2, 2 mM Na-isoascorbate adjusted to pH 6.5 with HCl), and disrupted into a slurry using short bursts of a Polytron tissue homogenizer. The resulting slurry was ®rst squeezed through two layers of muslin, and subsequently passed through a muslin cotton wool sandwich into a 250 ml beaker on ice. The ®ltrate was divided equally, centrifuged for 1 min at 3000 g, the supernatants decanted, and the pellets resuspended in 1.0 ml resuspension medium (330 mM sorbitol, 50 mM Hepes±

KOH pH 7.6, 2 mM EDTA, 1 mM MgCl2, 1 mM MnCl2). Samples were washed in a total 150 ml of resuspension medium and centrifuged as above. To enhance the percentage of intact chloroplasts, one half of the sample was gently resuspended in 1 ml ice-cold resuspension medium, layered onto a 6 ml cushion of 40% (v/v) Percoll containing osmoticum/buffer, and centrifuged at 3000 g for 1 min. The pellets of the pre-gradient and postgradient samples were lysed in 300 ml lysate buffer (20 mM Hepes pH 7.5, 5 mM KCl, 1.5 mM MgCl2, 10 mM DTT, 10% [v/v] glycerol, 1% [w/v] PVP). Samples were then homogenized with 20 strokes of a Dounce homogenizer. Following homogenization, samples were incubated on ice for 1 h and centrifuged at 3000 g for 5 min to remove debris. Protein concentrations of the supernatants were determined by Bradford assay. Extracts were dispensed into 100 mg aliquots, frozen in a dry ice±ethanol bath, and stored at ± 80°C.

Plasmids Kanamycin- and tetracyclin-resistance genes were used in two substitutory systems to determine nucleotide exchange in chloroplast lysates. The kanamycin-sensitive plasmid pKsm4021 contains a single base transversion (T®G), that creates a TAG stop codon in the kan gene at codon 22. A tetracycline sensitive plasmid pTSm153 carries a single T®A nucleotide change at position 153 in the pBR322 plasmid, which creates a stop codon in the tetracycline (tet) gene at codon 23. A nucleotide insertional system with a tetracycline-sensitive plasmid, pTSD208, was used to analyze repair of single base deletions in chloroplast lysates. 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 kanS and tetS genes to restore resistance to their respective antibiotics. The chimeric RNA/DNA oligonucleotide Kan4021C, which can direct conversion of the kanS gene in pKSm4021 at codon 22 from TAG to TAC (stop®tyrosine), was synthesized as previously described (Cole-Strauss et al., 1999). Chimeric RNA/DNA oligonucleotide TetD208T was used to revert the tetS genes of plasmid pTSD208, respectively, at the mutated bases. A non-speci®c chimera SC1 (Cole-Strauss et al., 1996) was used for comparison and as a control. Single-stranded oligonucleotides, 3S/25G and 3S/28A, were synthesized with the appropriate modi®cations using phosphoramidites or controlled pore glass supports. After deprotection and removal from the solid support, all oligonucleotides were gel-puri®ed (Gamper et al., 2000a; Gamper et al., 2000b), and concentrations were determined spectrophotometrically (33 or 40 mg ml±1 per A260 unit).

In vitro assays Reaction mixtures consisted of 1 mg substrate plasmid pKSm4021 and either 1.5 mg of chimeric oligonucleotide Kan4021C and the non-speci®c CO SCI, or 0.55±1.5 mg of modi®ed oligonucleotide Kan 3S/25G or 3S/25A for the kanS system. 1 mg of substrate plasmid pTSm153 or pTSD208, and 1.5 mg of effector oligonucleotide Tet153 or TetD208T, or 0.55 mg of the modi®ed oligonucleotide 3S/28A were used for the tetS system. These components ã Blackwell Science Ltd, The Plant Journal, (2001), 27, 267±274

Chloroplast lysates support targeted mutagenesis were mixed in a buffer of 200 mM Tris pH 7.5, 100 mM MgCl2, 1 mM DTT, 0.2 mM spermidine, 25 mM ATP, 1 mM each CTP, GTP, UTP, 0.1 mM each dNTPs, and 10 mM NAD. The reaction was initialized by adding 0±20 mg chloroplast lysates in 100 ml reaction volumes. The reactions were incubated at 30°C for 30 min and stopped by placing on ice. The substrate plasmid was then isolated by phase partition with 1 : 1 phenol : chloroform extraction, followed by ethanol precipitation on dry ice for 2 h or overnight and centrifugation at 4°C for 30 min. Samples were then washed with 70% ethanol and centrifuged for 15 min and resuspended in 50 ml TE.

Electroporation, plating and selection 5 ml of resuspended reaction precipitates were used to transform 20 ml aliquots of electrocompetent DH10B bacteria using a CellPorator apparatus (Life Technologies, Rockville, MD, USA) as described by the manufacturer. Each mixture was transferred to a 1 ml SOC culture, incubated at 37°C for 1 h, then converted plasmids were ampli®ed by adding kanamycin to 50 mg ml±1 or tetracycline to 12 mg ml±1, and an additional incubation for 3 h at 37°C. 100 ml aliquots of undiluted cultures were then plated onto Luria-Bertani (LB) agar plates containing 50 mg ml±1 kanamycin or 12 mg ml±1 tetracycline, respectively. 100 ml aliquots of a 104 dilution of the cultures were also plated onto LB agar plates containing 100 mg ml±1 ampicillin. Plating was performed in duplicate using sterile Pyrex beads. Both sets of plates were incubated for 16±18 h at 37°C, and colonies were counted using an Accucount 1000 plate reader (Biologics Inc., Gainsville, VA, USA). Targeted conversion of the kanS or tetS gene was determined by normalizing the number of kanamycin-resistant or tetracycline-resistant colonies by dividing by the number of ampicillin-resistant colonies, as all plasmids contain a wild-type amp gene. Resistant colonies were con®rmed by selecting isolated clones for mini-preparation of plasmid DNA, followed by sequencing using an ABI Big Dye Terminator kit on an automated ABI 310 capillary sequencer.

Protein gel-blot analysis Protein samples (20 mg) were resolved by 4±12% SDS±PAGE, then blotted onto a nitrocellulose membrane (Bio-Rad, Hercules, CA, USA). The membrane was blocked with 5% (w/v) skimmed milk in phosphate-buffered saline with Tween 20 (PBST) at room temperature overnight, and incubated with mouse anti(histone H1) serum (Calbiochem, San Diego, CA, USA) in PBST with 5% (w/v) skimmed milk for 1 h. The membrane was washed with PBST and incubated with alkaline phosphatase-conjugated goat anti-mouse polyclonal antibodies (Promega, Madison, WI, USA) at a 1 : 7500 dilution for 1 h. Signals were detected with chemiluminescent reaction reagents (CSPD, TROPIX, Bedford, MA, USA) according to the manufacture's recommendations.

Acknowledgements We thank members of the Kmiec and May laboratories for their comments on this manuscript. This work was supported by The Samuel Roberts Noble Foundation.

References Beetham, P.R., Kipp, P.B., Sawycky, X.L., Arntzen, C.J. and May, ã Blackwell Science Ltd, The Plant Journal, (2001), 27, 267±274

273

G.D. (1999) A tool for functional plant genomics: chimeric RNA/ DNA oligonucleotides cause in vivo gene-speci®c mutations. Proc. Nat. Acad. Sci. USA, 96, 8774±8778. Cerutti, H., Osman, M., Grandoni, P. and Jagendorf, A.T. (1992) A homolog of Escherichia coli RecA protein in plastids of higher plants. Proc. Natl Acad. Sci. USA, 89, 8068±8072. Cerutti, H., Johnson, A.M., Boynton, J.E. and Gillham, N.W. (1995) Inhibition of chloroplast DNA recombination and repair by dominant negative mutants of Escherichia coli RecA. Mol. Cell Biol. 15 (6), 3003±3011. Cole-Strauss, A., Yoon, K., Xiang, Y., Byrne, B.C., Rice, M.C., Gryn, J., Holloman, W.K. and Kmiec, E.B. (1996) Correction of the mutation responsible for sickle cell anemia by an RNA±DNA oligonucleotide. Science, 273, 1386±1389. Cole-Strauss, A., Gamper, H.B., Holloman, W.K., MunÄoz, M., Cheng, N. and Kmiec, E.B. (1999) Targeted gene repair directed by the chimeric RNA/DNA oligonucleotide in mammalian cellfree extract. Nucl. Acids Res. 27, 1323±1330. Gamper, H.B., Cole-Strauss, A., Metz, R., Parekh, H., Kumar, R. and Kmiec, E.B. (2000a) A plausible mechanism for gene correction in chimeric oligonucleotides. Biochemistry, 39, 5808±5816. Gamper, H.B., Parekh, H., Rice, M.C., Bruner, M., Youkey, H. and Kmiec, E.B. (2000b) The DNA strand of chimeric RNA/DNA oligonucleotides can direct gene repair/conversion activity in mammalian and plant cell-free extracts. Nucl. Acids Res. 28, 4332±4339. Howland, G.P., Hart, R.W. and Yette, M.L. (1975) Repair of DNA strand breaks after gamma-irradiation of protoplasts isolated from cultured wild carrot cells. Mutation Res. 1, 81±87. Kmiec, E.B. (1999) Targeted gene repair. Gene Ther. 6, 1±4. May, G.D. and Kmiec, E.B. (2000) Plant gene therapy: crop varietal improvement through the use of chimaeric RNA/DNA oligonucleotide-directed gene targeting. Agbiotechnet, 2, 1±4. McLennan, A.G. (1988) DNA damage, repair, and mutagensis. In DNA Replication in Plants (Bryant, J.A. and Dunham, V.L., eds). Boca Raton, Florida: CRC Press, pp. 135±186. Ohyama, K., Fukuzawa, H., Kohchi, T. et al. (1986) Chloroplast gene organization deduced from the complete sequence of liverwort Marchantia polymorpha chloroplast DNA. Nature, 322, 572±574. Palmer, J.D. (1985) Comparative organization of chloroplast genomes. In Annual Review of Genetics (Campbell, A., Herskowitz, I. and Sandler, L.M., eds). Palo Alto, CA: Annual Reviews, Inc., pp. 325±354. Palmer, J.D. and Deliwiche, C.F. (1996) Second-hand chloroplasts and the case of the disappearing nucleus. Proc. Natl Acad. Sci. USA, 93, 7432±7435. Pang, Q., Hays, J.B. and Rajagopal, I. (1993) Two cDNAs from the plant Arabidopsis thaliana that partially restore recombination pro®ciency and DNA damage resistance to E. coli mutant lacking recombination-intermediate-resolution activities. Nucl. Acids Res. 21, 1647±1653. Rice, M.C., May, G.D., Kipp, P.B., Parekh, H. and Kmiec, E.B. (2000) Genetic repair of mutations in plant cell-free extracts directed by speci®c chimeric oligonucleotides. Plant Physiol. 123, 427±437. Shinozaki, K., Ohme, M., Tanka, M. et al. (1986) The complete sequence of the tobacco chloroplast genome: its gene organization and expression. EMBO J. 5, 2043±2049. Staub, J.M. and Maliga, P. (1995) Marker rescue from the Nicotiana tabacum plastid genome using plastid/Escherichia coli shuttle vector. Mol. Gen. Genet. 249, 37±42.

274

Eric B. Kmiec et al.

Whitehouse, D.G. and Moore, A.L. (1993) Isolation and puri®cation of functionally intact chloroplasts from leaf tissue and leaf tissue protoplasts. In Methods in Molecular Biology, Vol. 19: Biomembrane Protocols: I. Isolation and Analysis (Graham, J.M. and Higgins, J.A., eds). Totowan, NJ: Humana Press, pp. 123±151. Zhu, T., Peterson, D., Tagliani, L., St Clair, G., Baszczynski, C. and

Bowen, B. (1999) Targeted manipulation of maize genes in vivo using chimeric RNA/DNA oligonucleotides. Proc. Natl Acad. Sci. USA, 96, 8768±8773. Zhu, T., Mettenburg, K., Peterson, D.J., Tagliani, L. and Baszcyzynski, C.L. (2000) Engineering herbicide-resistant maize using chimeric RNA/DNA oligonucleotides. Nature Biotechnol. 18, 555±558.

ã Blackwell Science Ltd, The Plant Journal, (2001), 27, 267±274