Homeologous Plastid DNA Transformation in Tobacco Is Mediated by Multiple Recombination Events

Corresponding author: Peter Medgyesy, Biological Research Center, Hungarian Academy of Sciences, P.O. Box 521, H-6701 Szeged, Hungary., pmedgyesy{at}nucleus.szbk.u-szeged.hu (E-mail)

Communicating editor: K. J. NEWTON

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Efficient plastid transformation has been achieved in Nicotiana tabacum using cloned plastid DNA of Solanum nigrum carrying mutations conferring spectinomycin and streptomycin resistance. The use of the incompletely homologous (homeologous) Solanum plastid DNA as donor resulted in a Nicotiana plastid transformation frequency comparable with that of other experiments where completely homologous plastid DNA was introduced. Physical mapping and nucleotide sequence analysis of the targeted plastid DNA region in the transformants demonstrated efficient site-specific integration of the 7.8-kb Solanum plastid DNA and the exclusion of the vector DNA. The integration of the cloned Solanum plastid DNA into the Nicotiana plastid genome involved multiple recombination events as revealed by the presence of discontinuous tracts of Solanum-specific sequences that were interspersed between Nicotiana-specific markers. Marked position effects resulted in very frequent cointegration of the nonselected peripheral donor markers located adjacent to the vector DNA. Data presented here on the efficiency and features of homeologous plastid DNA recombination are consistent with the existence of an active RecA-mediated, but a diminished mismatch, recombination/repair system in higher-plant plastids.

THE integration of cloned plastid DNA into the plastid genome occurs exclusively through site-specific homologous recombination in genera as diverse as Chlamydomonas (BOYNTON et al. 1992 ) and Nicotiana (MALIGA 1993 ) and excludes the foreign vector DNA. This is in striking contrast with the integration of cloned DNA into the nuclear genome, which occurs randomly, and routinely involves the cointegration of vector DNA (OHL et al. 1994 ; ROCHAIX 1995 ). A characteristic feature of the plastid genome is its high recombinogenic activity, best demonstrated in the inverted repeat (IR) region as a continuous flip-flop recombination and copy correction by gene conversion (GOULDING et al. 1996 and references therein). A growing body of evidence indicates that a RecA-mediated recombination system exists in chloroplasts (CERUTTI et al. 1992 , CERUTTI et al. 1993 , CERUTTI et al. 1995 ; CERUTTI and JAGENDORF 1993 ). It is probable that this mechanism, which maintains the uniformity of the plastid DNA population, also drives the homologous integration events (MEDGYESY et al. 1985 ; FEJES et al. 1990 ).

Plastid transformation is fast becoming a routine tool for exploring the contribution of plastid genes and their regulatory regions in the processes of photosynthesis and other plastid-related biosynthetic activities. There are, however, very limited data on the mechanism and limits of DNA integration associated with plastid transformation. The apparent reason for this is that the usual strategy for introducing nonhomologous DNA into the plastid genome of the two model species (Chlamydomonas reinhardtii and Nicotiana tabacum) involves the use of completely homologous flanking regions that ensure efficient integration through double exchange events (BOYNTON and GILLHAM 1993 ; MALIGA et al. 1994 ).

In Chlamydomonas an almost complete replacement of a continuous region of the plastid genome by the corresponding DNA fragment from the donor plasmid is the most common integration event and, remarkably, a partial donor/recipient homology (plastid transformation in C. reinhardtii by C. smithii plastid DNA, and vice versa) decreased the integration frequency by 10–100 times (NEWMAN et al. 1990 ). A similarly striking dependence of the efficiency of integration on the extent of homology between the donor and target regions was observed in the case of site-specific nuclear transformation in yeast (SMOLIK-UTLAUT and PETES 1983 ; NEGRITTO et al. 1997 ) and mammalian cells (DENG and CAPECCHI 1992 ; TE RIELE et al. 1992 ) and in bacteriophage-plasmid recombinations in Escherichia coli (WATT et al. 1985 ; SHEN and HUANG 1986 ). The direct role of the mismatch repair system in limiting homeologous recombination has been demonstrated in such diverse genetic systems as nuclear transformation in mammalian cells (DE WIND et al. 1995 ), meiotic and mitotic chromosomal recombination in yeast (CHAMBERS et al. 1996 ; DATTA et al. 1996 ), conjugational and transductional recombination between bacterial species (RAYSSIGUIER et al. 1989 ), and bacteriophage-plasmid recombination in E. coli (SHEN and HUANG 1989 ). There are only scant data on homeologous transformation in higher-plant plastids. The combination of N. tabacum and N. plumbaginifolia for plastid transformation has proved unsuitable for the analysis of the effect of partial homology due to the high nucleotide sequence identity in the region usually investigated by plastid transformation (O'NEILL et al. 1993 ). In the case of an intraspecific tobacco plastid transformation, where several restriction fragment length polymorphisms (RFLPs) were introduced into the donor plastid DNA fragment (STAUB and MALIGA 1992 ), the RFLP analysis of two transformants indicated the incorporation of all or nearly all of the uninterrupted transforming DNA.

Nicotiana and Solanum are two genera belonging to different subfamilies in the family Solanaceae (D'ARCY 1979 ). There are therefore sufficient nucleotide sequence differences between the plastid genomes to facilitate the molecular analysis of recombination events (MEDGYESY 1994 ; HORVATH 1995 ). Here we demonstrate a highly efficient integration of cloned Solanum plastid DNA into the Nicotiana plastid genome mediated by multiple recombination events.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Characterization of the pSSH1 plastid transformation plasmid:
The location and cloning into pUC19 of the 7.8-kb HindIII plastid DNA fragment isolated from the Solanum nigrum StSp1 mutant was described previously (KAVANAGH et al. 1994 ). The complete nucleotide sequence of the 7844-bp insert of pSSH1 was determined (EMBL accession no. Y18934). The corresponding plastid DNA region in tobacco extends from nucleotides 98,531 to 106,198 and from nucleotides 136,424 to 144,091 in IR_B and IR_A, respectively (SHINOZAKI et al. 1986 ; GenBank accession no. Z00044). The nucleotide numbering in the 16SrDNA gene takes into consideration a correction (O'NEILL et al. 1993 ) of the published sequence.

Production of plastid transformant plants:
N. tabacum L. cv. Petit Havana was maintained as shoot cultures on Murashige and Skoog medium (Sigma, St. Louis, MO) in the light (50 µE m^-2 sec^-1, 16-hr day, 25°). PEG-mediated plastid transformation was performed as described (O'NEILL et al. 1993 ). Leaf protoplasts were isolated, DNA-treated protoplasts were cultured in liquid medium, colonies were further grown for selection in antibiotic-containing solid medium, and plants were regenerated from the resistant calli as described previously (MEDGYESY 1994 ). The selective medium contained either 1000 mg liter^-1 spectinomycin dihydrochloride or both spectinomycin dihydrochloride and streptomycin sulfate at 500 mg liter^-1 each. The same antibiotic (combination) was used during shoot regeneration as for the selection of the individual lines. Seedlings and leaf calli were tested for their resistance to the appropriate antibiotics as described by CSEPLO 1994 . In the resistance tests the two antibiotics were employed separately at 1000 mg liter^-1 each.

Molecular analysis of plastid transformant plants:
Chloroplasts were isolated from aseptically grown plants according to BOOKJANS et al. 1984 . Lysis of chloroplasts and the purification of DNA followed standard protocols (SAMBROOK et al. 1989 ). Restriction enzyme digestions were performed according to the instructions of the suppliers. DNA fragments were separated by horizontal agarose-slab gel electrophoresis and visualized by ethidium bromide staining following standard protocols. Nonradioactive DNA dot-blot hybridization was performed using the DIG DNA labeling and detection kit (Boehringer Mannheim, Mannheim, Germany). Undigested plastid DNA (1 µg) was denatured and bonded to a positively charged Amersham (Arlington Heights, IL) Hybond-N+ nylon membrane according to standard protocols and probed by a mixture of DraI-SmaI double- and ApaLI-AvaII-PvuII triple-digested pUC19 vector DNA. The 7.8-kb HindIII plastid DNA fragment was cloned from several transformants as described in KAVANAGH et al. 1994 . The sequences of oligonucleotide primers that were used for PCR and nucleotide sequence analysis of plastid transformants together with their position in IR_A within the N. tabacum plastid genome are as follows:

• LRP1: [136,250–136,283] 5'-TCGTCTCTGGGTGCCTAGGTATCCACCGTAAGCC-3'

• INS100: [136,757–136,738] 5'-GGACCGACCATAGAACCCTG-3'

• SPC1: [138,552–138,576] 5'-GTATGGCTGACCGGCGATTACTAGC-3'

• INS10-21: [140,674–140,694] 5'-GAAAGAGGTTGACCTCCTTGC-3'

• NT1: [140,772–140,788] 5'-CAACAATTCATCAGACT-3'

• SN1: [140,774–140,784] 5'-ACAGTTCATCACGGAAG-3'

• SSH: [140,981–141,002] 5'-CTCGGAGACCCACCAAAGTACG-3'

• CP1: [141,612–141,596] 5'-GGACCAATTTAGTCACG-3'

• STR1: [141,729–141,753] 5'-TCTGTAGAGTGGCAGTAAGGGTGAC-3'

• INS5TT10: [143,046–143,067] 5'-GAATGGCAGAGGCAAATAGAGC-3'

• LRP2: [144,131–144,098] 5'-ACGTCAGGAGTCCATTGATGAGAAGGGGCTGGGG-3'.

The oligonucleotide primers LRP1 and INS100 were used for the PCR detection of the 103-bp-long Solanum-specific insert (Figure 2C) from total DNA of control and plastid transformants. Because both primer sequences flank the site of insertion, a diagnostic fragment of either 508 bp (N. tabacum) or 611 bp (S. nigrum) will be amplified. The primer SN1 is specific for the 10-bp-long S. nigrum insertion (Figure 2B), while primer NT1 is specific for the wild-type nucleotide sequence at the same location. Primer SN1, when used in combination with primer CP1 (based on nucleotide sequence that is identical in both N. tabacum and S. nigrum), will amplify a product if the DNA template contains the insertion. Primer NT1, when used in combination with CP1, will amplify a product only if the DNA template lacks the insert. The oligonucleotide primers LRP1 and LRP2 were used to amplify, from total DNA or plastid DNA of the transformants, the entire plastid genome region corresponding to the the 7.8-kb HindIII fragment present in pSSH1. The priming sites for both LRP1 and LRP2 are located adjacent, but external to the 7.8-kb HindIII fragment on the plastid genome. Consequently, the primers cannot amplify sequences from within pSSH1 plasmid DNA. Efficient amplification of the 7.8-kb HindIII fragment was achieved using ~200 ng total DNA (or 20 ng plastid DNA) in a 50-µl reaction using the Boehringer Expand long template PCR system and the following cycle parameters: 94°, 20 sec; 63°, 30 sec; 68°, 7 min; for 35 cycles. DNA sequencing reactions were performed using either double-stranded plasmid DNA or long-range PCR-amplified DNA templates and Perkin-Elmer (Norwalk, CT)/Applied Biosystems (Foster City, CA) fluorescent dye terminator or dye primer sequencing kits. Sequencing reactions were analyzed on an Applied Biosystems Model 370 automatic sequencer.

View larger version (15K): In this window In a new window Download PPT slide   Figure 1. Genetic map of the pSSH1 plastid transformation plasmid. The 7.8-kb S. nigrum plastid DNA insert was cloned into the HindIII site of the pUC19 vector (thick line). The plastid genes (solid bands) shown inside and outside the circle are transcribed counterclockwise and clockwise, respectively. The relative position and the direction of transcription of the ampicillin resistance gene in pUC19 are shown by a long arrow. Asterisks mark intron-containing genes. BamHI restriction sites of the insert are shown. Sites of the streptomycin (strr) and the spectinomycin (sper) resistance mutations are shown by short arrows.

View larger version (38K): In this window In a new window Download PPT slide   Figure 2. Alignment of nucleotide sequence differences between N. tabacum and S. nigrum used as physical markers of plastid DNA transformation. Portions of the full N. tabacum (NT) nucleotide sequence are shown, together with the differences detected in S. nigrum (SN), as they appear in IRB (5' to 3' direction, strand A), (A) from position 365 of the intergenic region between ndhB and rps7, (B) from positions 445 and 941 (in N. tabacum) of the intergenic region between rps12(3') and trnV, and (C) from position 510 of the intron region in the trnA gene. Selected restriction enzyme cut sites characteristic of the particular region are also shown. Identical (.) and missing (-) nucleotides are marked.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Nucleotide sequence differences between S. nigrum and N. tabacum in the region inserted into the pSSH1 plastid transformation plasmid:
The pSSH1 plasmid (Figure 1) contains the 16SrDNA and the rps12(3') genes containing mutations conferring spectinomycin and streptomycin resistance, respectively, as a 7.8-kb HindIII fragment cloned from a S. nigrum plastid mutant (MCCABE et al. 1989 ; KAVANAGH et al. 1994 ). The spectinomycin resistance mutation is a G to A transition (as in strand A of IR_B) at nucleotide 1139 of the 16SrDNA gene. This nucleotide change causes the loss of an AatII site. The streptomycin resistance mutation is an A to C transversion at the first nucleotide position in codon 88 (at position 85 of exon 2) of rps12(3'), converting a lysine to a glutamine. This nucleotide change creates a new MseI site. At position 127 of the same exon of rps12(3') there is a neutral, species-specific, nucleotide difference (C in S. nigrum instead of T in N. tabacum at the first nucleotide position in codon 74). Complete nucleotide sequence analysis of the 7.8-kb cloned region identified additional species differences. The most characteristic S. nigrum feature is the presence of several 5- to 100-bp insertions/deletions in comparison with the N. tabacum nucleotide sequence. The species-specific nucleotide sequence differences used in our analysis of the plastid transformant lines are summarized in Figure 2. These nonhomologous sites are flanked on either side by regions sharing complete homology. The distinction between wild-type and transformed Nicotiana plastid DNA was facilitated by characteristic differences in the BamHI cleavage site map (cf. Figure 1 and Figure 3C). These RFLPs involved a shortening of fragment 5 by a Solanum-specific cut site and a lengthening of fragment 13 (formerly 11; SHINOZAKI et al. 1986 ) due to Solanum-specific insertions (cf. Figure 2 and Figure 3C).

View larger version (44K): In this window In a new window Download PPT slide   Figure 3. BamHI and PstI-AatII restriction fragment patterns and partial physical maps of the N. tabacum plastid transformants. (A) The gel electrophoretic separation of BamHI-digested plastid DNA of wild-type N. tabacum (lane 1) and several plastid transformants (lanes 2–4) reveals different combinations of the Solanum-specific cut sites. Only fragments >2.6 kb are shown. The transformant-specific fragments are marked by arrowheads. On the left a HindIII digest of lambda DNA is also shown. (B) The fragment patterns obtained with PstI-AatII double-digested plastid DNA in wild-type N. tabacum (lane 1) and several plastid transformants (lanes 2–4) demonstrate the identity of the copies of the inverted repeated region in the plastid transformants for the spectinomycin resistance site. Only fragments >5 kb are shown. The transformant-specific fragments are marked by arrowheads. On the left a HindIII digest of DNA is also shown. (C) BamHI and PstI-AatII cleavage site maps covering a part of the inverted repeat (solid bars) and the entire small single copy region, as they appear in the wild type (WT) and a N. tabacum plastid transformant (TR) possessing the full-length S. nigrum insert (the 7.8-kb area of integration is labeled by bordered lines). All the BamHI, PstI (P), and AatII (A) sites of the region displayed are marked (short vertical lines). Numbers refer to the corresponding digestion fragments ordered by their size.

Efficient genetic transformation of the Nicotiana plastid genome by Solanum plastid DNA:
Following PEG-mediated direct transformation of protoplasts using the pSSH1 plasmid, putative plastid transformant colonies were selected on the basis of their green color in a medium containing either spectinomycin or both spectinomycin and streptomycin. Tests for the presence of the antibiotic resistance markers were performed at three stages: the primary calli, the regenerated plants, and their progeny. Each line exhibited stable retention of the initial resistance phenotype and proved to be completely homoplasmic (Table 1). The frequency of double-resistant calli (~1 in 10⁴ viable colonies or in 10⁵ protoplasts treated), considered as plastid transformants, was very similar after either initial single or double resistance selection (Table 1). A single spectinomycin-resistant spontaneous mutant (not included in Table 1) was identified by its streptomycin sensitivity and the presence of the diagnostic, Nicotiana-specific, AatII site.

All but 1 of the 11 double-resistant Nicotiana plants characterized contained at least one of the unselected BamHI markers diagnostic for S. nigrum (for representative RFLP patterns see Figure 3A). Apart from the Solanum-specific RFLPs, no other deviation from the expected plastid DNA fragment patterns was observed in the transformants. This indicated the site-specific integration of the targeting plastid DNA in each case and a lack of integration of the pUC19 vector DNA. The latter conclusion was verified by a complete lack of dot-blot hybridization of the vector DNA to plastid DNA of each of the transformants (data not shown). The polymorphic DNA regions (revealed as RFLPs between the Solanum and the Nicotiana plastid DNA) were shown to be homozygous in all but one of the transformants, demonstrating complete intraorganellar plastid DNA segregation after transformation. This conclusion was corroborated by PCR analysis of Solanum- and Nicotiana-specific regions in several transformants (data not shown). In one of the primary regenerants a single locus, the peripheral Solanum-specific BamHI marker, proved to be mixed, i.e., both the Nicotiana- and the Solanum-specific BamHI fragment 5 were observed in the RFLP pattern (data not shown). The Solanum-specific fragment gradually faded over successive rounds of vegetative propagation and was absent in the sexual progeny of this plant (12 plants investigated, data not shown). A diagnostic test for the presence of the Solanum spectinomycin resistance site was based on a concomitant loss of a Nicotiana-specific AatII site that resulted in distinctive changes in the PstI-AatII restriction fragment pattern diagnostic for IR_A and IR_B (Figure 3C). The RFLP patterns demonstrated the identity of the copies of the inverted repeated region with respect to the spectinomycin resistance site in all of the plastid transformant plants analyzed (for representative RFLP patterns see Figure 3B).

Efficient cointegration of homologous and heterologous regions of the S. nigrum plastid DNA into the N. tabacum plastid genome:
A preliminary compilation of the BamHI RFLPs (for representative RFLP patterns see Figure 3A) in the 11 double-resistant, pSSH1-mediated plastid transformants indicated that the typical event was a full-length integration of the targeting plastid DNA. This was found in 8 of the transformants, while in 2 transformants only the peripheral Solanum-specific BamHI marker was missing. In a single transformant, despite using so few markers, multiple recombination events were detected. In this particular transformant both the peripheral and the internal (located between the two resistance markers) Solanum-specific BamHI markers were missing. These observations were corroborated by another experiment (HORVATH et al. 1997 ; E. M. HORVÁTH, S. O. PETER, T. JOET, D. RUMEAU, T. KAVANAGH, C. SCHÄFER, G. PELTIER and P. MEDGYESY, unpublished results) in which the plasmid pSSH1M was used as the donor DNA and spectinomycin alone was used to select for plastid transformants. In this latter experiment the pSSH1 plasmid was modified (pSSH1M) to include a SmaI site (Figure 4) near that end of the HindIII plastid DNA insert that possessed no RFLP in the original construct. A Solanum-specific XhoI site was also included in the analysis of the pSSH1M-mediated plastid transformants. This Solanum-specific XhoI site is the result of a single nucleotide difference (ctcGag instead of ctcCag) at position 1411 (in N. tabacum) of the intergenic region between rps12(3') and trnV (data not shown). The use of these additional markers confirmed the previously observed dominance of uninterrupted integration and a bias against the integration of peripheral donor markers (Figure 4). However, in one-third of the 18 transformants investigated in this experiment, integrations via quadruple recombination events were also revealed (Figure 4). The frequency of recombination events between the individual markers, calculated as a percentage of the transformants recombined in a particular interval, was in the range of 10–50% for internal regions. Furthermore, 50 and 60% of the transformants possessed the peripheral donor marker located 113 bp (SmaI) and 231 bp (BamHI) from the cloning site, respectively (Figure 4), demonstrating that the peripheral recombination events occurred very close to the vector-insert junction in these particular transformants. It was notable that in spite of the multiple exchange events the integration of the XhoI site (marking a single nucleotide difference), the internal BamHI site (marking 10- to 21-bp insertions), and the peripheral BamHI site (marking the 103-bp insertion) of the donor DNA (cf. Figure 2 and Figure 3) occurred similarly (in 70, 70, and 60% of the transformants, respectively). The remarkably frequent recombination events via a 231-bp homologous peripheral region located between the 103-bp Solanum-specific insertion and the pUC19 vector and the occurrence of multiple recombination events in several transformants prompted a more detailed analysis of heterologous sites across the region.

View larger version (14K): In this window In a new window Download PPT slide   Figure 4. Schematic interpretation of the postulated recombination events following transformation of N. tabacum plastids with the pSSH1M plasmid. The 7.8-kb donor insert of the pSSH1M plasmid is shown at the top. Arrows mark the location of the Solanum-specific resistance and RFLP sites scored (a restriction endonuclease in parentheses means the absence of a Nicotiana-specific cut site). Asterisks mark intron-containing genes. Diagrams are presented for the origin of each class of plastid transformants. Each diagram is a single line with horizontal segments representing Solanum-type (top, thick line) and Nicotiana-type (bottom, dotted line) regions that are specified by a particular distribution of the donor genetic and physical markers (short vertical lines) and are separated by transverse segments (thin line) representing the regions of the postulated recombination events. The number of transformants belonging to a certain class is indicated beside each diagram. In this experiment only spectinomycin selection has been used.

Nucleotide sequence analysis of all of the initial 11 pSSH1-transformed plants was performed, using cloned plastid DNA or long-range PCR products covering the region of insertion, to confirm the presence or absence of the heterologous sites and the resistance mutations. In most of the transformants showing full-length integration on the basis of the genetic and RFLP analyses, multiple recombination events were revealed (Figure 5). These integrations involved mainly quadruple but also sextuple or octuple recombination events. On the basis of the 14 distinct nucleotide sequence markers scored, only one of the 11 transformants proved to be homogeneously Solanum type, showing uninterrupted full-length integration. Within the limitations of the analysis due to the relatively low number of transformants and the complete homology of large regions of the donor and target plastid DNA, no polarity in marker recombination/incorporation was detected along the targeted region. Nucleotide sequence analysis did not reveal any examples of illegitimate recombination, inaccurate exchange, or spontaneous mutation in the DNA regions investigated. Natural plastid DNA polymorphism as a potential source of the differences was excluded by the nucleotide sequence analysis of five independently isolated clones of the 7.8-kb HindIII plastid DNA fragment obtained from the original S. nigrum StSp1 mutant. Where a recombination site consisted of only a single nucleotide difference, a similar approach was used to exclude PCR-generated errors. The selection conditions had little or no influence on the distribution of the transformants among the different recombination classes (Figure 5). For example, of the two plants in which the neighboring Solanum-specific sites are missing on both sides of the streptomycin resistance site, one was regenerated after selection on spectinomycin alone, and the other after double antibiotic selection. Recombination events were revealed in between the resistance markers in four out of six double-selected and two out of five spectinomycin-selected transformants. The frequencies of cointegration between the spectinomycin resistance and the different types of unselected Solanum markers, calculated as a percentage of transformants possessing both markers, were comparable: 45–90% for single nucleotide differences, 45–90% for 5- to 21-bp-long insertions/deletions, and 75% for the 103-bp-long insertion.

View larger version (16K): In this window In a new window Download PPT slide   Figure 5. Schematic interpretation of the distribution of postulated recombination events, based on nucleotide sequence analysis of the N. tabacum plastid transformants, following transformation of N. tabacum plastids with the pSSH1 plasmid. The 7.8-kb donor insert of the pSSH1 plasmid is shown at the top. Arrows mark the location of the resistance mutations. Short vertical lines mark the regions of nucleotide sequence differences characteristic of the donor plastid DNA. The capital letters in parentheses above them refer to the appropriate nucleotide sequence alignments shown in Figure 2. The resistance mutations (and the Solanum-specific nucleotide adjacent to the streptomycin resistance site) are described in the first section of RESULTS. The XhoI site is described in the third section of RESULTS. Asterisks mark intron-containing genes. Diagrams are presented for the origin of each class of plastid transformants. Each diagram is a single line with horizontal segments representing Solanum-type (top, thick line) and Nicotiana-type (bottom, dotted line) regions that are specified by a particular distribution of the donor genetic and physical markers (short vertical lines) and are separated by transverse segments (thin line) representing the regions of the postulated recombination events. The number of transformants (spectinomycin-selected plus double-selected) belonging to a certain class is indicated beside each diagram.

Marked position effects resulted in a nonrandom distribution of the recombination events on the targeted plastid DNA region:
The recombination events were restricted in several cases to short (41–141 bp) homologous regions (cf. Figure 2 and Figure 5). No recombination was detected in the three locations where the Solanum-specific sites are separated by 1–22 nucleotides. The recombination frequency in the individual internal sections, calculated as a percentage of the transformants recombined in the particular interval, was in a similar range when, for example, the five 41- to 290-bp (10–45%) and the five 415- to 2140-bp (10–55%) internal sections were compared. The recombination frequency in the entire 6.8-kb region between the peripheral Solanum markers (35%) or between one of the peripheral Solanum markers and the spectinomycin resistance marker (30% in both cases) was also in this range. This suggests no linkage between the markers separated by >40 bp. The distribution of the recombination events, however, was not random. In 90% of the transformants the peripheral recombination events were observed in one of the two short 762- and 231-bp peripheral regions, and in 50% of the cases they were found in both (Figure 5). These data were corroborated by similar results (70 and 40%) obtained by RFLP analysis in the other tobacco plastid transformation experiment (Figure 4) or in N. plumbaginifolia (15 transformants investigated; data not shown). A dependence of the recombination frequency on the length and the location of the particular intervals determined by the Solanum-specific sites was further addressed in both the pSSH1- and the pSSH1M-mediated plastid transformation experiments (Figure 6). The observed frequency of recombination (calculated as percentage of the intervals recombined in the region) was compared with the expected frequency (calculated to be proportional to the length of the individual intervals in the region). The recombination frequency of the individual sections calculated in this way also showed a hot spot in the shortest peripheral sections between the vector DNA and the first marker. The observed frequency in the 762-, 231-, and 113-bp peripheral sections was 2, 7, and 10 times higher than the expected frequency, respectively. The frequent recombination events at the short homologous regions adjacent to the heterologous vector DNA resulted in a very frequent cointegration of (at least one of) the peripheral donor markers. Integration events, however, never expanded into the vector DNA, demonstrating that several kilobases of completely nonhomologous DNA were an efficient barrier to recombination. The recombination frequencies were relatively low in the intervals adjacent to the selected spectinomycin resistance site, although this was pronounced only in the pSSH1 experiment. The lack of a conclusively strong effect of the selection was also indicated by the recombination frequencies in the intervals adjacent to the streptomycin resistance site, which were very similar in the double-selected (13 and 6%) and the spectinomycin-selected (14 and 4%) transformants of the pSSH1 experiment. All these position effects were independent of the length of the nonhomologous sites scored, indicating that the plastid DNA heterologies, at least up to 100 nucleotides, did not hamper their own integration.

View larger version (22K): In this window In a new window Download PPT slide   Figure 6. Distribution of recombination frequencies in the individual sections of the targeted region in the N. tabacum plastid transformants. The 7.8-kb donor insert of the pSSH1 and the pSSH1M plasmids is shown at each top. Arrows mark the location of the Solanum-specific resistance and RFLP sites scored. Short vertical lines mark the regions of nucleotide sequence differences characteristic of the donor plastid DNA. The capital letters in parentheses above them refer to the appropriate nucleotide sequence alignments shown in Figure 2. Asterisks mark intron-containing genes. Dotted horizontal arrows mark the cotranscription units in the region (MATSUBAYASHI et al. 1987 ; HILDEBRAND et al. 1988 ; IRATNI et al. 1997 ). At the bottom of the figures the line is sectioned to show all intervals between the Solanum-specific sites. The minute intervals in which no recombination was detected are not shown. The expected and the observed recombination frequencies characteristic for the individual intervals are shown above and below the lines, respectively. The observed recombination frequency is expressed as the percentage of the recombination events per interval. A 100% value represents the total number of recombined intervals of the entire region counted in all transformants. The expected recombination frequency is calculated as proportional to the length of the intervals, considering 100% for the entire region between the cloning sites.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

This article demonstrates efficient transformation of the N. tabacum plastid genome by plastid DNA of S. nigrum, which belongs to a separate subfamily in the Solanaceae. The phylogenetic difference between these species is sufficient to prevent the production of green cybrids with new nucleus-plastid combinations (THANH et al. 1988 ). The previously reported production of a normal green plastid DNA recombinant plant between tobacco and potato through protoplast fusion (THANH and MEDGYESY 1989 ), although predicting the feasibility of direct plastid gene transfer in such a taxonomically distinct combination, left open the question of efficiency. In this study the frequency of PEG-mediated plastid transformation using the incompletely homologous (homeologous) 7.8-kb segment of Solanum plastid DNA was ~10 transformants per 10⁶ protoplasts treated, using either N. tabacum or N. plumbaginifolia (data not shown) as recipient in several independent experiments. This gure is similar to that observed in other experiments, in which either a 3- or 4-kb segment of completely homologous plastid DNA (GOLDS et al. 1993 ; O'NEILL et al. 1993 ) or a foreign gene flanked by 1–4 kb of completely homologous plastid DNA was used (KOOP et al. 1996 ). In a recent independent experiment (B. UIJTEWAAL, personal communication) similar plastid transformation frequencies were obtained in tobacco irrespective of whether completely homologous (pTB-116; O'NEILL et al. 1993 ) or homeologous (pSSH1; this article) transforming plasmid was used (both plasmids confer streptomycin and spectinomycin resistance and target the same IR region). Therefore, in contrast to interspecific plastid DNA transformation in Chlamydomonas, in Nicotiana no reduction in the transformation frequency was observed using partially homologous donor DNA. This observation may reflect fundamental differences in the mechanism of integration in the plastid genome of the two organisms. In Chlamydomonas plastid transformants the integration of the donor plastid DNA occurred exclusively via one double exchange event (NEWMAN et al. 1990 ), even if a recombination hot spot was located in the recipient DNA region (NEWMAN et al. 1992 ). Consequently, complete or partial, but uninterrupted, DNA integration was always observed, even if 6- to 7-kb donor fragments labeled with six or seven RFLP/genetic markers were used (NEWMAN et al. 1990 ). The Nicotiana plastid transformants reported here, however, exhibited multiple internal exchange events in 33–73% of the transformants (depending on the number of markers scored), resulting in a mosaic-type integration of the donor DNA. It was notable that, in spite of the multiple recombination events, a larger extent of integrative recombination was also observed. A peripheral diagnostic marker located 762, 231, or 113 bp from the vector-insert junction was cointegrated in 73, 61–73, and 50% of the Nicotiana transformants, respectively. In Chlamydomonas, however, a decrease in the distance between the vector DNA and the first donor marker from 1.35 kb to 150 bp resulted in a decrease in the ratio of transformants possessing that marker from 34–61% to 2–9% (NEWMAN et al. 1992 ). In our experiments, therefore, the mechanism resulting in multiple recombination events was believed to facilitate both the frequency and the efficacy of the cointegration of homologous and heterologous plastid DNA regions of the targeting DNA.

Recombination between homeologous nucleotide sequences is influenced by the extent and degree of DNA homology that is monitored by the recombination/repair enzymes. A central component of the system that promotes homologous recombination in bacteria is the RecA protein and its functional homologs in eukaryotes (KOWALCZYKOWSKI and EGGLESTON 1994 ; CAMERINI-OTERO and HSIEH 1995 ). RecA protein is broadly responsible for facilitating recombination steps involving DNA pairing, strand transfer, and branch migration. The presence and activity of a RecA homolog has also been observed in higher plant plastids (CERUTTI et al. 1992 , CERUTTI et al. 1993 ; CERUTTI and JAGENDORF 1993 ). Furthermore, the expression of the wild type and a dominant negative mutant of the E. coli recA gene located on the Chlamydomonas plastid genome substantially facilitated and diminished plastid DNA recombination, respectively (CERUTTI et al. 1995 ). Data presented in this article are fully compatible with a RecA-mediated recombination mechanism. The observed local recombination frequencies near the vector-insert junction were up to 10 times higher than would be predicted. This phenomenon was most probably due to the stimulation of homologous recombination adjacent to regions of extensive heterology. An enhanced neighboring recombination frequency as a result of nonhomology has been reported in crosses between bacteriophages (LIEB et al. 1984 ). It was concluded that large heterologous blocks in the recombining molecules can present a barrier to branch migration of Holliday structures that, consequently, would lead to an increased probability of resolution by cleavage in the adjacent region (LIEB et al. 1984 ). A central element of current recombination models is random walking of the branch point of a Holliday intermediate along the homologous region, which is then resolved by either integration before or dissociation at the end of the homology (ALANI et al. 1994 ; FUJITANI et al. 1995 ). It is notable, however, that the RecA-mediated DNA strand-exchange reaction in vitro readily bypasses short internal regions of heterology up to ~100 bp (MOREL et al. 1994 ). The lack of internal nonhomologous regions longer than 100 bp that might have posed a barrier to branch migration and, in contrast, the blocking effect of the extensive nonhomologous vector DNA, is the most plausible explanation of the high frequency of integration of both the internal and peripheral donor plastid DNA markers observed in our experiments. In this context the mechanism promoting multiple recombination events is likely to involve random resolution of branch migration along the entire targeted region. Unfortunately the inherent difficulty in detecting each type of segregant after plastid recombinations in higher plants mitigates against a detailed investigation of the recombination mechanism (occurrence and ratio of gene conversions and crossovers). The above considerations, however, have taken into account only the effects of the RecA-mediated recombination/repair system.

The most potent inhibitor of recombination between moderately diverged nucleotide sequences is the mismatch repair system (MODRICH and LAHUE 1996 ). Mispaired bases in heteroduplex regions formed in the recombination reaction trigger long-patch mismatch repair leading to dissociation of the heteroduplex and abortion of the recombination reaction. In vitro studies demonstrated that mismatch repair proteins can directly inhibit the branch-migration stage of RecA-catalyzed strand transfer when mismatched base pairs occur within the heteroduplex (WORTH et al. 1994 ). Even 1% nucleotide sequence difference causes an order of magnitude reduction in the frequency of DNA recombination in various genetic systems (PETIT et al. 1991 ; TE RIELE et al. 1992 ; DATTA et al. 1997 ; ZAHRT and MALOY 1997 ; ELLIOTT et al. 1998 ). A similar observation was reported in the case of interspecific plastid transformation in Chlamydomonas (NEWMAN et al. 1990 ), although the involvement of a recombination hot spot can conceal the effect (NEWMAN et al. 1992 ). The most direct proof of the role of the mismatch repair system is that mutations impairing defined mismatch repair genes increase the frequency of homeologous recombination by one to two orders of magnitude (see Introduction). Remarkably, homeologous recombination in different organisms typically results in the incorporation of continuous tracts (MATIC et al. 1994 ; SWEETSER et al. 1994 ; PORTER et al. 1996 ; DOONER and MARTINEZ-FEREZ 1997 ; YANG and WALDMAN 1997 ; ELLIOTT et al. 1998 ). Our results revealed recombination frequencies and efficiencies typical of those found in homologous plastid DNA transformation despite the 2.4% nucleotide sequence divergence between Nicotiana and Solanum plastid DNA in the transformed region. Furthermore, we have demonstrated a primarily discontinuous homeologous integration (multiple recombination events involving closely located markers). All these data suggest that the mismatch repair system might be naturally suppressed in higher-plant plastids.

If further investigations support this hypothesis the phenomenon can be related to the lack of plastid DNA recombination in sexual crosses in higher plants. Meiotic recombination between chromosomes is a fundamental feature of the sexual cycle (ROEDER 1990 ; RADMAN and WAGNER 1993 ), and homologous recombination is an inherent part of bacterial gene transfer processes (CLARK and LOW 1988 ). Recombination occurs between parental plastid DNAs in Chlamydomonas (BOYNTON et al. 1992 ), even in interspecific crosses (LEMIEUX et al. 1981 ), because plastid fusion and the occasional mixing of parental plastid DNAs is a natural part of mating (CAVALIER-SMITH 1970 ; KUROIWA et al. 1982 ). In interspecific Chlamydomonas plastid recombinants the frequency and distribution of recombination events are determined by the plastid genome region and the species combination (LEMIEUX and LEE 1987 ; LEMIEUX et al. 1990 ; NEWMAN et al. 1992 ). In most higher plants, however, the particular mechanisms of plastid transmission during the sexual cycle appear to prevent the mixing of parental plastids (MOGENSEN 1996 ). Even in those species where plastids are inherited biparentally, recombination between the parental plastid genomes cannot be demonstrated, suggesting that plastid fusion itself is precluded in higher plants (HAGEMANN 1992 ). Whatever the evolutionary significance of the asexuality of plastids in higher plants (BIRKY 1995 ), we hypothesize that it may have resulted in the deterioration of a mismatch repair system that was no longer required to defend the plastid genome against the potentially destabilizing consequences of homeologous recombination.

In this study the primary plastid transformant shoots regenerated following antibiotic selection proved to be homoplasmic. The regeneration of homoplasmic shoots in a single step occurs routinely when selection is based on antibiotic insensitivity mutations located in plastid genes (MEDGYESY 1990 , MEDGYESY 1994 ; DIX and KAVANAGH 1995 ). In the case of a multicellular, multiplastidic organism in which each plastid contains a multicopy genome, homoplasmy is achieved by a complex segregation process. This initially involves copy correction of the changes introduced into the inverted repeat region and segregation of the plastid DNA copies during multiplication of the individual plastids. Under selective conditions the individual resistant plastids have a selective advantage during partitioning at the subsequent cell divisions and gradually come to dominate the intracellular plastid population in the daughter cells. A sufficiently strong selection pressure results in early and complete segregation of plastids to the pure transformed type in the cell colonies prior to shoot regeneration. In our experiments a single plant was regenerated from an individual cell line, i.e., from a protoplast-derived primary resistant callus. Therefore the individual plants investigated represent independent primary transformation events.

There is an increasing recognition that the plastid genome is a particularly attractive target for the introduction of genes carrying agriculturally valuable traits in various crop species (DIX and KAVANAGH 1995 ), a prospect exploited till now only in tobacco (MCBRIDE et al. 1995 ; DANIELL et al. 1998 ). This poses the question as to whether it will be necessary to develop species-specific plastid transformation vectors for each particular crop species to ensure efficient and site-specific integration of transgenes. The gene order and the coding sequences in the ribosomal RNA gene operon of the plastid genome are highly conserved in higher plants (DELP and KOSSEL 1991 ). The data presented in this article demonstrate that the nucleotide sequence divergence in the targeted region (at least up to 2.4%) does not hamper the integration of the donor plastid DNA. These considerations indicate that plastid transformation vectors directed to this region are not required to be species specific. Further investigations are necessary, however, to determine the limits of nucleotide sequence heterogeneity, allowing efficient integration. The use of the Solanum plastid rRNA gene-based plasmid pSSH1M in tobacco proved to be efficient for the targeted inactivation of a photosynthetic gene linked at a 5-kb distance to the marker gene (HORVATH et al. 1997 ; E. M. HORVÁTH, S. O. PETER, T. JOET, D. RUMEAU, G. V. HORVÁTH, T. KAVANAGH, C. SCHÄFER, G. PELTIER and P. MEDGYESY, unpublished results).

	FOOTNOTES

¹ Present address: Institute of Biotechnology, National Center for Natural Science and Technology, Nghia Do, Tu Liem, Hanoi, Vietnam.
² Present address: Botanisches Institut, Universität Bayreuth, D-95440 Bayreuth, Germany.

	ACKNOWLEDGMENTS

We gratefully acknowledge support from the following sources: from the Országos Tudományos Kutatási Alap (T016995 and T019759), the Volkswagen-Stiftung (I70961) and the Országos Müszaki Fejlesztési Bizottság (EU-98-D8-011) to P.M., from Forbairt (Basic Science Research Grant SC/95/1270) to T.A.K., and a fellowship to S.O.P. from the Daimler-Benz-Stiftung.

Manuscript received August 21, 1998; Accepted for publication March 15, 1999.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ALANI, E., R. A. G. REENAN, and R. D. KOLODNER, 1994  Interaction between mismatch repair and genetic recombination in Saccharomyces cerevisiae.. Genetics 137:19-39[Abstract].

BIRKY, C. W., 1995  Uniparental inheritance of mitochondrial and chloroplast genes: mechanisms and evolution. Proc. Natl. Acad. Sci. USA 92:11331-11338[Abstract/Free Full Text].

BOOKJANS, G., B. M. STUMMANN, and K. W. HENNINGSEN, 1984  Preparation of chloroplast DNA from pea plastids isolated in a medium of high ionic strength. Anal. Biochem. 141:244-247[Medline].

BOYNTON, J. E. and N. W. GILLHAM, 1993  Chloroplast transformation in Chlamydomonas.. Methods Enzymol. 217:510-536[Medline].

BOYNTON, J. E., N. W. GILLHAM, S. M. NEWMAN and E. H. HARRIS, 1992 Organelle genetics and transformation of Chlamydomonas, pp. 3–64 in Cell Organelles, edited by R. G. HERRMANN. Springer-Verlag, Vienna.

CAMERINI-OTERO, R. D. and P. HSIEH, 1995  Homologous recombination proteins in prokaryotes and eukaryotes. Annu. Rev. Genet. 29:509-552[Medline].

CAVALIER-SMITH, T., 1970  Electron microscopic evidence for chloroplast fusion in zygotes of Chlamydomonas reinhardtii.. Nature 228:333-335[Medline].

CERUTTI, H. and A. T. JAGENDORF, 1993  DNA strand-transfer activity in pea (Pisum Sativum L.) chloroplasts. Plant Physiol. 102:145-153[Abstract].

CERUTTI, H., M. OSMAN, P. GRANDONI, and A. T. JAGENDORF, 1992  A homolog of Escherichia coli RecA protein in plastids of higher plants. Proc. Natl. Acad. Sci. USA 89:8068-8072[Abstract/Free Full Text].

CERUTTI, H., H. Z. IBRAHIM, and A. T. JAGENDORF, 1993  Treatment of pea (Pisum Sativum L.) protoplasts with DNA-damaging agents induces a 39-kilodalton chloroplast protein immunologically related to Escherichia coli RecA. Plant Physiol. 102:155-163[Abstract].

CERUTTI, H., A. M. JOHNSON, J. E. BOYNTON, and N. W. GILLHAM, 1995  Inhibition of chloroplast DNA recombination and repair by dominant negative mutants of Escherichia coli RecA. Mol. Cell. Biol. 15:3003-3011[Abstract].

CHAMBERS, S. R., N. HUNTER, E. J. LOUIS, and R. H. BORTS, 1996  The mismatch repair system reduces meiotic homeologous recombination and stimulates recombination-dependent chromosome loss. Mol. Cell. Biol. 16:6110-6120[Abstract].

CLARK, A. J., and K. B. LOW, 1988 Pathways and systems of homologous recombination in Escherichia coli, pp. 155–215 in Genetic Recombination, edited by R. KUCHERLAPATI and G. R. SMITH. American Society for Microbiology, Washington, DC.

CSÉPLÓ, A., 1994 Transfer of lincomycin resistance through somatic and sexual cybridization in Nicotiana, pp. 394–404 in Somatic Hybridization in Crop Improvement I, Biotechnology in Agriculture and Forestry, Vol. 27, edited by Y. P. S. BAJAJ. Springer-Verlag, Berlin.

DANIELL, H., R. DATTA, S. VARMA, S. GRAY, and S.-B. LEE, 1998  Containment of herbicide resistance through genetic engineering of the chloroplast genome. Nature Biotechnol. 16:345-348[Medline].

D'ARCY, W. G., 1979 The classification of the Solanaceae, pp. 3–47 in The Biology and Taxonomy of the Solanaceae, edited by J. G. HAWKES, R. N. LESTER and A. D. SKELDING. Academic Press, London.

DATTA, A., A. ADJIRI, L. NEW, G. F. CROUSE, and S. JINKS-ROBERTSON, 1996  Mitotic crossovers between diverged sequences are regulated by mismatch repair proteins in Saccharomyces cerevisiae.. Mol. Cell. Biol. 16:1085-1093[Abstract].

DATTA, A., M. HENDRIX, M. LIPSITCH, and S. JINKS-ROBERTSON, 1997  Dual roles for DNA sequence identity and the mismatch repair system in the regulation of mitotic crossing-over in yeast. Proc. Natl. Acad. Sci. USA 94:9757-9762[Abstract/Free Full Text].

DELP, G., and H. KÖSSEL, 1991 rRNAs and rRNA genes of plastids, pp. 139–167 in The Molecular Biology of Plastids, Cell Culture and Somatic Cell Genetics of Plants, Vol. 7A, edited by L. BOGORAD and I. K. VASIL. Academic Press, San Diego.

DENG, C. and M. R. CAPECCHI, 1992  Reexamination of gene targeting frequency as a function of the extent of homology between the targeting vector and the target locus. Mol. Cell. Biol. 12:3365-3371[Abstract/Free Full Text].

DE WIND, N., M. DEKKER, A. BERNS, M. RADMAN, and H. TE RIELE, 1995  Inactivation of the mouse Msh2 gene results in mismatch repair deficiency, methylation tolerance, hyperrecombination, and predisposition to cancer. Cell 82:321-330[Medline].

DIX, P. J. and T. A. KAVANAGH, 1995  Transforming the plastome: genetic markers and DNA delivery systems. Euphytica 85:29-34.

DOONER, H. K. and I. M. MARTÍNEZ-FÉREZ, 1997  Recombination occurs uniformly within the bronze gene, a meiotic recombination hotspot in the maize genome. Plant Cell 9:1633-1646[Abstract].

ELLIOTT, B., C. RICHARDSON, J. WINDERBAUM, J. A. NICKOLOFF, and M. JASIN, 1998  Gene conversion tracts from double-strand break repair in mammalian cells. Mol. Cell. Biol. 18:93-101[Abstract/Free Full Text].

FEJES, E., D. ENGLER, and P. MALIGA, 1990  Extensive homologous chloroplast DNA recombination in the pt14 Nicotiana somatic hybrid. Theor. Appl. Genet. 79:28-32.

FUJITANI, Y., K. YAMAMOTO, and I. KOBAYASHI, 1995  Dependence of frequency of homologous recombination on the homology length. Genetics 140:797-809[Abstract].

GOLDS, T., P. MALIGA, and H.-U. KOOP, 1993  Stable plastid transformation in PEG-treated protoplasts of Nicotiana tabacum.. Bio/Technology 11:95-97.

GOULDING, S. E., R. G. OLMSTEAD, C. W. MORDEN, and K. H. WOLFE, 1996  Ebb and flow of the chloroplast inverted repeat. Mol. Gen. Genet. 252:195-206[Medline].

HAGEMANN, R., 1992 Plastid genetics in higher plants, pp. 65–96 in Cell Organelles, edited by R. G. HERRMANN. Springer-Verlag, Vienna.

HILDEBRAND, M., R. B. HALLICK, C. W. PASSAVANT, and D. P. BOURQUE, 1988  Trans-splicing in chloroplasts: the rps12 loci of Nicotiana tabacum.. Proc. Natl. Acad. Sci. USA 85:372-376[Abstract/Free Full Text].

HORVÁTH, E. M., S. O. PETER, S. CUINÉ, G. GUEDENEY, D. RUMEAU et al., 1997 Targeted inactivation of the plastidic ndhB gene in tobacco reveals its role in higher-plant photosynthesis, abstr. 1198 in Abstracts of the 5th International Congress of Plant Molecular Biology. The International Society for Plant Molecular Biology, Singapore.

HORVÁTH, G. V., 1995 Plastid DNA recombination is facilitated by an SOS-like repair mechanism in taxonomically distinct species combination. Ph.D. Thesis, József Attila University, Szeged, Hungary.

IRATNI, R., L. DIEDERICH, H. HARRAK, M. BLIGNY, and S. LERBS-MACHE, 1997  Organ-specific transcription of the rrn operon in spinach plastids. J. Biol. Chem. 272:13676-13682[Abstract/Free Full Text].

KAVANAGH, T. A., K. M. O'DRISCOLL, P. F. MCCABE, and P. J. DIX, 1994  Mutations conferring lincomycin, spectinomycin, and streptomycin resistance in Solanum nigrum are located in three different chloroplast genes. Mol. Gen. Genet. 242:675-680[Medline].

KOOP, H.-U., K. STEINMÜLLER, H. WAGNER, C. RÖSLER, and C. EIBL et al., 1996  Integration of foreign sequences into the tobacco plastome via polyethylene glycol-mediated protoplast transformation. Planta 199:193-201[Medline].

KOWALCZYKOWSKI, S. C. and A. K. EGGLESTON, 1994  Homologous pairing and DNA strand-exchange proteins. Annu. Rev. Biochem. 63:991-1043[Medline].

KUROIWA, T., S. KAWANO, S. NISHIBAYASHI, and C. SATO, 1982  Epifluorescent microscopic evidence for maternal inheritance of chloroplast DNA. Nature 298:481-483[Medline].

LEMIEUX, B., M. TURMEL, and C. LEMIEUX, 1990  Recombination of Chlamydomonas chloroplast DNA occurs more frequently in the large inverted repeat sequence than in the single-copy regions. Theor. Appl. Genet. 79:17-27.

LEMIEUX, C. and R. W. LEE, 1987  Nonreciprocal recombination between alleles of the chloroplast 23S rRNA gene in interspecific Chlamydomonas crosses. Proc. Natl. Acad. Sci. USA 84:4166-4170[Abstract/Free Full Text].

LEMIEUX, C., M. TURMEL, and R. W. LEE, 1981  Physical evidence for recombination of chloroplast DNA in hybrid progeny of Chlamydomonas eugametos and C. moewusii.. Curr. Genet. 3:97-103.

LIEB, M., M. M. TSAI, and R. C. DEONIER, 1984  Crosses between insertion and point mutations in gene cI: stimulation of neighboring recombination by heterology. Genetics 108:277-289[Abstract/Free Full Text].

MALIGA, P., 1993  Towards plastid transformation in flowering plants. Tibtech 11:101-107.

MALIGA, P., J. STAUB, H. CARRER, I. KANEWSKI and Z. SVÁB, 1994 Homologous recombination and integration of foreign DNA in plastids of higher plants, pp. 83–93 in Homologous Recombination and Gene Silencing in Plants, edited by J. PASZKOWSKI. Kluwer Academic Publishers, Dordrecht, The Netherlands.

MATIC, I., M. RADMAN, and C. RAYSSIGUIER, 1994  Structure of recombinants from conjugational crosses between Escherichia coli donor and mismatch-repair deficient Salmonella typhimurium recipients. Genetics 136:17-26[Abstract].

MATSUBAYASHI, T., T. WAKASUGI, K. SHINOZAKI, K. YAMAGUCHI-SHINOZAKI, and N. ZAITA et al., 1987  Six chloroplast genes (ndhA–F) homologous to human mitochondrial genes encoding components of the respiratory chain NADH dehydrogenase are actively expressed: determination of the splice sites in ndhA and ndhB pre-mRNAs. Mol. Gen. Genet. 210:385-393[Medline].

MCBRIDE, K. E., Z. SVÁB, D. J. SCHAAF, P. S. HOGAN, and D. M. STALKER et al., 1995  Amplification of a chimeric Bacillus gene in chloroplasts leads to an extraordinary level of an insecticidal protein in tobacco. Bio/Technology 13:362-365[Medline].

MCCABE, P. F., A. M. TIMMONS, and P. J. DIX, 1989  A simple procedure for the isolation of streptomycin-resistant plants in Solanaceae. Mol. Gen. Genet. 216:132-137.

MEDGYESY, P., 1990 Selection and analysis of cytoplasmic hybrids pp. 287–316 in Plant Cell Line Selection, edited by P. J. DIX. VCH Publishers, Weinheim, Germany.

MEDGYESY, P., 1994 Cybrids—transfer of chloroplast traits through protoplast fusion between sexually incompatible Solanaceae species, pp. 72–85 in Somatic Hybridization in Crop Improvement I, Biotechnology in Agriculture and Forestry, Vol. 27, edited by Y. P. S. BAJAJ. Springer-Verlag, Berlin.

MEDGYESY, P., E. FEJES, and P. MALIGA, 1985  Interspecific chloroplast recombination in a Nicotiana somatic hybrid. Proc. Natl. Acad. Sci. USA 82:6960-6964[Abstract/Free Full Text].

MODRICH, P. and R. LAHUE, 1996  Mismatch repair in replication fidelity, genetic recombination, and cancer biology. Annu. Rev. Biochem. 65:101-133[Medline].

MOGENSEN, H. L., 1996  The hows and whys of cytoplasmic inheritance in seed plants. Am. J. Bot. 83:383-404.

MOREL, P., A. STASIAK, S. D. EHRLICH, and E. CASSUTO, 1994  Effect of length and location of heterologous sequences on RecA-mediated strand exchange. J. Biol. Chem. 269:19830-19835[Abstract/Free Full Text].

NEGRITTO, M. T., X. WU, T. KUO, S. CHU, and A. M. BAILIS, 1997  Influence of DNA sequence identity on efficiency of targeted gene replacement. Mol. Cell. Biol. 17:278-286[Abstract].

NEWMAN, S. M., J. E. BOYNTON, N. W. GILLHAM, B. L. RANDOLPH-ANDERSON, and A. M. JOHNSON et al., 1990  Transformation of chloroplast ribosomal RNA genes in Chlamydomonas: molecular and genetic characterization of integration events. Genetics 126:875-888[Abstract].

NEWMAN, S. M., E. H. HARRIS, A. M. JOHNSON, J. E. BOYNTON, and N. W. GILLHAM, 1992  Nonrandom distribution of chloroplast recombination events in Chlamydomonas reinhardtii: evidence for a hotspot and an adjacent cold region. Genetics 132:413-429[Abstract].

OHL, S., R. OFFRINGA, P. J. M. VAN DEN ELZEN and P. J. J. HOOYKAAS, 1994 Gene replacement in plants, pp. 191–217 in Homologous Recombination and Gene Silencing in Plants, edited by J. PASZKOWSKI. Kluwer Academic Publishers, Dordrecht, The Netherlands.

O'NEILL, C., G. V. HORVÁTH, E. M. HORVÁTH, P. J. DIX, and P. MEDGYESY, 1993  Chloroplast transformation in plants: polyethylene glycol (PEG) treatment of protoplasts is an alternative to biolistic delivery systems. Plant J. 3:729-738[Medline].