Plastome Mutator–Induced Alterations Arise in Oenothera Chloroplast DNA Through Template Slippage

Corresponding author: Barbara B. Sears, 37 Plant Biology Bld., Department of Botany and Plant Pathology, Michigan State University, East Lansing, MI 48824, searsb{at}pilot.msu.edu (E-mail).

Communicating editor: K. J. NEWTON

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The plastome mutator of Oenothera hookeri strain Johansen causes deletions and duplications at target sites defined by direct repeats in the plastid genome. Previous studies characterized the mutations long after they had occurred and could not discriminate between the possibilities that the plastome mutator acted through unequal homologous recombination or template slippage. From the known hotspots, the rRNA spacer in the large inverted repeat was chosen for this study because it contains both direct and indirect repeats. Identical deletions were recovered from independently derived plants; the altered regions were always flanked by direct repeats. The regions in which the deletions occurred have the potential to form secondary structures that would stabilize the intervening sequence. Of the two affected regions, the one with the stronger potential secondary structure was altered more frequently. Because no duplication products or inversions were recovered, it is proposed that the plastome mutator acts through template slippage rather than through a recombination mechanism.

WHEN homozygous, the plastome mutator (pm) of Oenothera causes a 200–1000-fold increase in the spontaneous appearance of pigment-deficient sectors (EPP 1973 ; SEARS and SOKALSKI 1991 ). The mutations are transmitted to the progeny in a non-Mendelian fashion and are absolutely linked to the inheritance of the chloroplast DNA (cpDNA). Additionally, RFLPs have been observed in the cpDNA of mutant lines, as well as photosynthetically competent lines after generations of maintenance of the pm genotype (CHIU et al. 1990 ). One variable region associated with pm was localized to ORF2280 in the large inverted repeat (BLASKO et al. 1988 ). The alteration was determined to result from the loss of varying numbers of 24-base pair (bp) tandem direct repeats. It was suggested that misalignment of the 24-bp repeat segments between the opposite copies of the inverted repeat during pairing, and recombination could have resulted in deletions and duplications (BLASKO et al. 1988 ). The recovery of identical deletions in both copies of the inverted repeat was interpreted to indicate that copy correction had occurred. If copy correction between the two sides of the inverted repeat is nondirectional, both duplications and deletions of the 24-bp repeat units should have been recovered. However, under the influence of the pm, no increase in the number of copies of the 24-bp repeats was found. The majority of pm-induced alterations, both in the ORF2280 and in regions outside of the large inverted repeat, are in fact deletions (BLASKO et al. 1988 ; CHANG et al. 1996 ).

The failure to recover products of reciprocal events and the association of direct repeats with the pm- induced alterations were the basis for the proposed theory that the pm causes deletions and duplications via replication slippage (CHANG et al. 1996 ). In bacteria, slippage of the template or daughter strand during replication may result in mispairing at sites containing short, tandemly repeated sequences, resulting in deletions or duplications of the repeat units when replication continues (LEVISON and GUTMAN 1987 ). CHANG et al. 1996 found that in all of the cases where insertions and deletions (indels) were created by the pm, they were associated with short tandem or direct repeats, with the smallest repeat unit involved being a single base (a stretch of tandemly repeated adenines).

The mechanism of pm action is being reexamined in this study for two reasons. First, previous assessments considered the failure to recover reciprocal duplication and deletion products as an indication that recombination is an unlikely mechanism for alterations associated with the pm. However, the inability to find both products of an unequal crossover event within the plastome of pm-generated mutants does not preclude homologous recombination as being the cause of the variability generated by the pm. Whereas intermolecular recombination can result in an unequal crossing over, intramolecular recombination between adjacent, homologous direct repeats can lead to the deletion of one copy of the repeat and of any intervening sequence (LOVETT et al. 1993 ; BI and LIU 1996 ). This, in fact, is the result that is observed in the presence of the pm. Second, although many pm-induced mutations have been characterized in Oenothera cpDNA, most of the mutant lines were not examined until long after the mutation event; therefore, the alterations observed were never directly linked to the mutagenic effect of pm.

To better characterize the pm-specific mutagenesis process, this study was designed to examine the mutational events as they occurred. The region between the 16S rRNA gene (16S) and the tRNA isoleucine gene (trnI) of the plastome was chosen as a focal point because it was known to be a hotspot for pm-induced alterations (CHIU et al. 1990 ). It also was an ideal site to study pm-derived variability because the repeat units that compose the spacer are highly amplified (HORNUNG et al. 1996 ; SEARS et al. 1996 ), and both direct and indirect repeats are present. This study set out to use these features to determine if the mode of mutagenesis by the pm is through recombination or template slippage. Our strategy was to examine the DNA as it was being altered in the plastome of Oenothera hookeri and to use the observations to propose a model to explain how the genetic defect, which results in the pm phenotype, is responsible for the deletions and duplications that it creates.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Plant material:
Although the I-D plastome type had been present in the original pm line (EPP 1973 ), generations of self-pollination resulted in an accumulation of deletions (CHIU et al. 1990 ; CHANG et al. 1996 ). To regenerate the homozygous pm genotype with the I-D plastome, a homozygous wild-type line of strain Johansen with plastome I-D was crossed as the female parent with pollen obtained from a periclinal chimera of a pm/pm plant, thereby allowing the mutant plastid contributed by the male parent to be recognized and culled. The F₁ plants were backcrossed to a pm/pm line, again with selection against plants having mutant plastids in the cotyledons or the first several true leaves. Among the progeny, half were expected to be homozygous for pm and all would contain the I-D plastome type. To recover new pm variants, the plants were crossed as the female parent to a plant with the O. hookeri strain Johansen nuclear background containing a plastome IV cytoplasm.

Isolation of DNA:
In a procedure adapted from DOYLE and DOYLE 1987 , total DNA was isolated from soil-grown Oenothera leaves. Leaves were removed from the plants and ground into a fine powder with liquid nitrogen. CTAB buffer (3 ml; 0.1 M Tris-HCl, pH 8.0, 1.4 M NaCl, 0.02 M EDTA, 2% cis-trimethylammonium bromide) was added for every 1 g of fresh tissue. The grindate was incubated in a water bath at 60° for 30 min. Once it had cooled to room temperature, the solution was mixed with one volume of phenol/chloroform. This was repeated until no interphase was visible. To remove the prevalent secondary metabolites, 5 M NaCl was added to a final concentration of 3 M, and the DNA was precipitated with 0.75 volumes of isopropanol (FANG et al. 1992 ).

PCR amplification:
The primer within the 16S gene is the forward 17-mer oligonucleotide 5'-TCGTAACAAGGTAGCCG-3', and the reverse primer on the edge of the trnI gene is a 15-base oligonucleotide, 5'-CGTTAATAGTCCCCG-3'. DNA (5 ng) was added to individual 100-µl reaction mixtures containing 200 µmol of each dideoxynucleotide, 2.5 units of Taq polymerase (GIBCO BRL, Gaithersburg, MD), 70 ng of each primer, and 1x PCR reaction buffer (GIBCO BRL). Conditions for the 30 cycles of PCR were as follows: 1 min denaturation at 94°, 1.5 min at 40° to anneal the primer with the template, and 2 min at 72° for DNA extension.

Figure 4 was digitally scanned using Microtek's Microscan software (Redondo Beach, CA). The image was then displayed using Adobe Pagemaker software (San Jose, CA).

View larger version (55K): In this window In a new window Download PPT slide   Figure 1. —PCR visualization of alterations within the 16S-trnI spacer region of cpDNA. D, the progenitor plastome type; IV, an evolutionarily distinct plastome type; C1 and C2, two previously isolated pm-induced deletion variants, Cornell-1 and Cornell-2; P1–P3, pooled cpDNA from pm/pm plants before visible mutant sectors; V1–V10, individual plants after sectors were observed.

View larger version (53K): In this window In a new window Download PPT slide   Figure 2. —Deletions in plastids of green and white tissues within the 16S-trnI spacer. A Southern blot of EcoRI-digested plant DNA probed with an EcoRI clone containing the 16S-trnI spacer. Lane D shows the homoplasmic DNA of plastome I-D; lane C1 contains DNA of the Cornell-1 plastome type, both in wild-type nuclear background. The remaining lanes depict the heteroplasmic state of the cpDNA for plants when the pm allele was homozygous, as determined by the creation of sectors. The V-numbered lanes use the nomenclature in Figure 1. The letters below indicate the type of tissue from which the cpDNA was isolated (G, green only; W, green and white tissue).

View larger version (69K): In this window In a new window Download PPT slide   Figure 3. —Evidence of pm-specific mutations occurring in the plastome of O. hookeri. (Lane 1) Low-range DNA molecular weight marker (Amresco, Solon, OH). (Lane 2) PCR amplification product of the 16S-trnI spacer from the Ann Arbor line of O. hookeri strain Johansen. (Lanes 3–13) PCR products from 10 different plants of the Düsseldorf line of O. hookeri.

View larger version (28K): In this window In a new window Download PPT slide   Figure 4. —PCR-amplified products of pm-induced variants after vegetative segregation. Each lane represents the 16S-trnI spacer from separate progeny that inherited different cpDNA from their maternal parent. V numbers indicate the maternal parent, as shown in Figure 1.

Sequencing of pm-variants:
DNA was PCR amplified with both the 16S and the trnI primer, as described above. The amplified product (10 µl) was subsequently reamplified with only one of the primers to create a single-stranded PCR product. The product was then treated as described by INNIS et al. 1990 . Subsequently, 2 µg of PCR-amplified DNA was incubated with 70 ng of primer (as listed in CHANG et al. 1996 ) and 1x sequencing buffer at 100° for 3 min. Immediately after the 3-min incubation, the DNA and primer were flash frozen in liquid nitrogen. The components of the sequencing reaction were mixed in the amounts specified in the Sequenase kit (United States Biochemical, Cleveland, OH; dATP labeled with [³⁵S]dATP was from New England Biolabs, Wilmington, DE), and the method for annealing of primers and sequencing was modified from CASANOVA et al. 1990 as follows: The mixture was incubated for 45 sec at room temperature to allow the primer to anneal to the template. One quarter of the mixture was added to each individual termination mix (United States Biochemical), each prewarmed at 37°. Incubation occurred at 37° for 2 min and was terminated with the addition of stop solution (United States Biochemical). Before loading onto an 8% acrylamide gel, the sequencing reactions were denatured by heating for 3 min at 80°.

The GCG program from the Wisconsin Computer Group was used to identify and analyze potential secondary structures within the 16S-trnI spacer. STEMLOOP was used to identify all possible intrastrand pairing; MFOLD and FOLDRNA were used to quantify the free energy of the best secondary structures.

Southern analysis:
Total DNA, isolated as described above, was electrophoresed into a 0.6% agarose gel to remove the excess secondary metabolites. The band of DNA was excised from the gel, and the agarose was removed with the DNA extraction kit from Qiagen (Chatsworth, CA). DNA (5 µg) was digested with 10 units of EcoRI (GIBCO BRL) and electrophoresed in a 1.0% TBE gel at 100 V for 7 hr. The DNA was then transferred overnight to a nylon membrane according to the method of SAMBROOK et al. 1989 . The membrane was probed with a radioactively labeled EcoRI clone (pOjor1) containing the 16S-trnI spacer (Multiprime DNA Labeling Kit from Amersham, Amersham, UK, and [³²P]dCTP from New England Nuclear/Dupont, Boston). The filter was washed in high salt at room temperature and twice in low salt at 50° according to the method of SAMBROOK et al. 1989 .

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Evidence of plastome mutator–induced deletions occurring within the 16S-trnI spacer:
Previous studies have noted the existence of large deletions between the 16S gene and the trnI gene in several pm-derived lines (CHIU et al. 1990 ; CHANG et al. 1996 ). The progenitor plastome type, represented by the plastome in the Düsseldorf line of O. hookeri strain Johansen (I-D), has a 16S-trnI spacer of 1100 bp, but significantly smaller spacers were recovered in all the pm lines analyzed (CHANG et al. 1996 ). To assess the speed at which the deletions occur, a wild-type I-D plastome type was crossed into the pm line (see MATERIALS AND METHODS). From plants grown on soil, leaf samples were pooled and analyzed by PCR to detect deletion or insertion events. Several of the known pm hotspots were analyzed by PCR, but no alterations were observed (data not shown). The lack of variability was also evident initially in the 16S-trnI spacer (Figure 1, lanes P1–P3) with the lack of any evident white sectors on the leaves. Once the plants had developed flowering shoots, white sectors were observed and the plants were analyzed individually by PCR amplification. At this sampling, almost every individual that was screened showed heteroplasmy within the 16S-trnI spacer (Figure 1, lanes V1–V10).

To determine if the spacer alterations occurred only in the white tissues, DNA was extracted separately from white and green tissue. PCR amplifications indicated that the white tissue was not the only cell type with changes in the 16S-trnI spacer; green tissues sampled from the pm/pm plants also showed deletions (data not shown). An independent Southern analysis (Figure 2) was conducted to better visualize the population of plastome types within each type of tissue. All the detectable changes in size revealed by the Southern blot correspond to deletions. The sizes of deleted sequence visible from the two procedures are comparable, ranging between 180 and 600 bp.

CpDNA deletions are not observed in wild-type lines of Oenothera:
To determine if the cpDNA deletions were a result of natural variation or a direct result of the pm, DNA from 10 separate plants containing the Düsseldorf plastome type and from one plant from the identical Johansen strain maintained at Ann Arbor were analyzed with PCR by amplifying through the 16S-trnI spacer (Figure 3). None of these plants had the pm/pm genotype, but they were all exposed to the same environmental conditions as the mutants analyzed earlier in this study. PCR amplification showed that all the plants contain 16S-trnI spacers equivalent to the I-D plastome type, indicating that no alterations occur without the presence of pm.

Common deletions in the 16S-trnI variants newly derived from the pm:
The plants containing cpDNA deletion variants were crossed as the female parent to remove the plastids from the pm nuclear background and to allow the variant plastomes to sort out. PCR amplification of the 16S-trnI spacer from 181 progeny showed that almost all were homoplasmic. Of these, only five or six discrete size classes were observed (data not shown). A subset (Figure 4) was maintained for further analysis, including six plants with a spacer of wild-type size.

PCR amplification products from the size variants shown in Figure 4 were purified and sequenced. Figure 5 displays the DNA sequences from some of those plastome variants. Figure 5A shows the sequence of this region from the progenitor Düsseldorf plastome line, with a shorthand notation that describes the direct repeats that compose the spacer (SEARS et al. 1996 ). Figure 5B displays the shorthand form of the sequences from representatives of the sequenced plastome variants. Cornell-1, -2, and -3 have been described previously (CHANG et al. 1996 ) and are included for comparison in this study. Variants derived from the same parent plant are depicted with the same V numbers. These numbers correspond to the V numbers in Figure 1 and Figure 2 as well. A number of deletion events were found (depicted as black lines in Figure 5B). When the identical cpDNA deletion was recovered from the same parent plant, the sequence is only shown once because the deletions were probably derived from the same event. However, many identical deletions were recovered from different plants as a consequence of independent events. Of the 21 variants sequenced, seven contained a 16S-trnI spacer of wild-type size. No inversions were identified.

View larger version (65K): In this window In a new window Download PPT slide   Figure 5. —Sequence comparison between pm-derived variants. (A) The sequence for the progenitor plastome, D, through the region of the 16S-trnI spacer in which the deletions occurred. The letters A–G represent different repeat families and the numbers after the letters identify subtle differences found within that specific repeat. (B) A schematic of the unique 16S-trnI variants collected. The thick lines locate the regions in which deletions were found. D, the progenitor sequence from the Düsseldorf line (GenBank accession number U41046); C1–C3, deletion variants characterized from the Sears pm mutant collection (CHANG et al. 1996 ); V1–V7, deletion variants isolated for this study. Plants with the same V number were derived from the same parental plant (numbers correspond to the numbers in previous figures). The lowercase letters after the V numbers indicate plastomes that sorted out in the subsequent generation; among related plants, only those with different spacer sequences are shown.

Two regions in the 16S-trnI spacer were found to be deleted frequently. The one to the left is found in 10 of the 13 unique lines in this study. In every one of these variants, the deletion begins within the first B1 repeat unit and ends within the second B1 repeat unit. The other deletion is located closer to the trnI gene. This deletion is common to five of the 13 represented lines. However, just as in the deletion to the left, in each of these five variants, the deletion appears to begin and end at the exact same repeat units. In the case of the right deletion, it begins within an E-G repeat unit and ends at the E-G repeat immediately downstream. Three of the 13 deletion variant lines represented in Figure 5B have alterations other than those described. The deletions found in the spacer of the Cornell-1 and Cornell-2 lines have been described previously (CHANG et al. 1996 ). The remaining variant, V5b, has a deletion at the right end of the 16S-trnI spacer that is similar to the deletion found in the other five lines in that it begins at the E-G repeat unit but ends at a more distant E-G repeat unit. In summary, identical deletions occur independently, identifying sequences that are highly susceptible to pm-mediated mutagenesis.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In the presence of the pm allele, deletions between direct repeats accumulate within the plastome of Oenothera after an initial delay (Figure 1). By screening the plastome as the plants were growing and the alterations were occurring, we were able to visualize this ongoing process once the pm/pm plants were identified by the presence of newly generated sectors. One of the hotspots for mutagenesis by pm is within the 16S-trnI spacer, a region containing many direct as well as some indirect repeats. Both the PCR-amplified products (Figure 1) and Southern analysis (Figure 2) clearly show that deletions were occurring frequently at that site. Deletions do not occur unless the pm is present (Figure 3). It is important to note that the deletions generated within the 16S-trnI spacer will have no obvious deleterious effects on the plastid. The deletions have occurred within a stretch of sequence that was amplified in an evolutionary time frame in the genus Oenothera (SEARS et al. 1996 ; HORNUNG et al. 1996 ). Although some sequences in the spacer are doubtlessly needed for processing the rRNA transcript, these repeats are theorized to be dispensable because they are not found in other species. Furthermore, deletions were recovered in both green and white tissue, so photosynthetic competence is not compromised by the deletions. We conclude that the white sectors contain plastomes that must have suffered at least one additional pm-induced alteration responsible for the white phenotype.

The Southern blot depicts a population of plastomes in which half of the 16S-trnI spacers have suffered deletions (according to PhosphoImager quantifications). Insertion products were neither observed nor recovered. This is an important observation because a misaligned reciprocal recombination event would be expected to produce both deletions and duplications. Although we have found no evidence of intermolecular recombination, it is possible that the mutations are caused by intramolecular recombination. Intramolecular recombination between adjacent direct repeats causes the loss of one copy of the repeat and any intervening sequence (BI and LIU 1996 ). If the pm is causing deletions through intramolecular recombination, then recombination between inverted repeats should also occur. Subsequently, this event would be visualized as a sequence inversion between the repeat units. Within the 16S-trnI spacer, which contains indirect as well as direct repeats (Figure 5A as well as SEARS et al. 1996 ), we found no inversions within any of the sequences analyzed. Therefore, because there is no evidence of either intermolecular or intramolecular recombination, we conclude that the pm is causing alterations in the plastome through template slippage.

DNA repeats are often associated with mutations that occur in Escherichia coli (FARABAUGH et al. 1978 ). In the misalignment theory, GLICKMAN and RIPLEY 1984 identified not only the involvement of repeats in the creation of many deletions, but the cooperation between repeats and palindromes in the deletion event. They proposed that the presence of a palindrome allows single-stranded DNA to form stemloops or cruciform structures, and that these secondary structures stabilize the misalignment across repeats. GLICKMAN and RIPLEY found that if the occurrence of deletions involved repeat units, most often one copy of the repeat was lost, as well as all of the intervening sequence, while the other repeat was maintained. Upon sequencing the spacer in more than 20 variants, pm-induced deletions were always found to be flanked by direct repeats. Figure 6 shows the potential structure that can form for the two regions in which deletions most often occurred. In both locations, one copy of the repeat, as well as all intervening sequences, would be enclosed in the secondary structure. The potential for secondary structure helps explain the likelihood with which some regions are deleted. The structure shown in Figure 6A has a much higher free energy (G = -71 ) than the structure in Figure 6B (G = -25 ) or any other pm hotspot identified thus far. This region also shows the highest frequency of deletions in the pm lines. This correlation, along with the low rate of mutations at the other known pm hotspots in these plants, suggests that the presence of secondary structures increases the probability of the pm causing deletions.

View larger version (27K): In this window In a new window Download PPT slide   Figure 6. —Secondary structures that predict how deletions could have occurred in the 16S-trnI spacer. (A and B) The potential secondary structures that can form within the two common deleted regions of the 16S-trnI spacer in plastome I-D of Oenothera. (A) Secondary structure proposed to delimit the more common right-hand deletion (G = -71 ). (B) Stemloop proposed to define the left-hand deletion (G = -25 Kcal/mole).

From all the data accumulated thus far (in this paper and in EPP 1973 ; BLASKO et al. 1988 ; WOLFSON et al. 1991 ; CHANG et al. 1996 ), we can make some predictions as to what the pm locus might encode. If one assumes that the mutations arise because of replication slippage and that slippage is enhanced by the presence of secondary structures from intrastrand pairing, one could predict that the misfunction of any one of a number of proteins involved in DNA metabolism could contribute to template slippage. Because the mutation frequency in the heterozygote is comparable to that of the wild-type plant, we believe that mutations arise not because of the malfunctioning of a protein, but rather because of the complete absence of a protein or protein function. Thus it is unlikely that the mutant allele encodes an essential protein, such as a DNA polymerase, which is necessary for cpDNA replication. We propose that the pm locus encodes a protein that directly affects the helicity or rigidity of the plastome, such as a topoisomerase, helicase, or single-stranded binding protein. If supercoiling is not relaxed during replication or the helix is not unwound efficiently, the probability that the replication complex will stall is increased. That, in turn, increases the likelihood of the template slipping (BRILL et al. 1987 ). If single-stranded binding protein is absent, intrastrand pairing will not be prevented and the resulting secondary structures could be passed over during replication (MEYER and LAINE 1990 ). In the majority of instances, the polymerase would be repositioned on the template properly, and no alterations in the daughter strand would result. However, in regions that contain tandem repeats—which is the case for every one of the alterations created by the pm—the replication complex could become misaligned, resulting in a DNA indel in the daughter strand as replication continues.

While the identity of the nonfunctional protein in the pm-homozygous plants has yet to be established, we are convinced that the pm causes alterations within the plastome via template slippage and that the occurrence of slippage by the template is enhanced in the presence of secondary structures, thereby implicating a topoisomerase or other protein, such as a helicase, involved in altering the cpDNA topology or progression of the replication complex.

	ACKNOWLEDGMENTS

The authors thank Dr. PAL MALIGA and Dr. DAVID JARRELL for their critical reading of the manuscript and Drs. GORDON CANNON and SABINE HEINHORST for their insightful discussions. This work was supported by the grant DCB-9019488 from the National Science Foundation and the Michigan Agricultural Experiment Station.

Manuscript received December 31, 1996; Accepted for publication January 21, 1998.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BI, X. and L. F. LIU, 1996  RecA-independent DNA recombination between repetitive sequences: mechanisms and implications. Prog. Nucleic Acid Res. Mol. Biol. 54:253-292[Medline].

BLASKO, K., S. A. KAPLAN, K. G. HIGGINS, R. WOLSON, and B. B. SEARS, 1988  Variation in copy number of a 24-base pair tandem repeat in the chloroplast DNA of Oenothera hookeri strain Johansen. Curr. Genet. 14:287-292[Medline].

BRILL, S. J., S. DINARDO, K. VOELKEL-MEIMAN, and R. STERNGLANZ, 1987  Need for DNA topoisomerase activity as a swivel for DNA replication for transcription of ribosomal RNA. Nature 326:414-416[Medline].

CASANOVA, J. L., C. PANNETIER, C. JAULIN, and P. KOURILSKY, 1990  Optimal conditions for directly sequencing double-stranded PCR products with sequenase. Nucleic Acids Res. 18:4028[Free Full Text].

CHANG, T.-L., L. L. STOIKE, D. ZARKA, G. SCHEWE, and W.-L. CHIU et al., 1996  Characterization of primary lesions caused by the plastome mutator of Oenothera. Curr. Genet. 30:522-530[Medline].

CHIU, W.-L., E. JOHNSON, S. KAPLAN, K. BLASKO, and M. SOKALSKI et al., 1990  Oenothera chloroplast DNA polymorphisms associated with plastome mutator activity. Mol. Gen. Genet. 221:59-64.

DOYLE, J. J. and J. L. DOYLE, 1987  A rapid DNA isolation procedure for small quantites of fresh leaf tissue. Phytochem. Bull. 19:11-15.

EPP, M., 1973  Nuclear gene-induced plastome mutations in Oenothera hookeri. I. Genetic analysis. Genetics 75:465-483[Abstract/Free Full Text].

FANG, G., S. HAMMER, and R. GRUMET, 1992  A quick and inexpensive method for removing polysaccharides from plant genomic DNA. Biotechniques 13:52-54[Medline].

FARABAUGH, P. J., U. SCHMEISSNER, M. HOFER, and H. MILLER, 1978  Genetic studies of the lac repressor. VII. On the molecular nature of spontaneous hotspots in the lacI gene of Escherichia coli. J. Mol. Biol. 126:847-863[Medline].

GLICKMAN, B. W. and L. S. RIPLEY, 1984  Structural intermediates of deletion mutagenesis: a role for palindromic DNA. Proc. Natl. Acad. Sci. USA 81:512-516[Abstract/Free Full Text].

HORNUNG, S., H. FULGOSI, P. DÖRFEL, and R. G. HERRMANN, 1996  Sequence variation in the putative replication origins of the five genetically distinct basic Euoenothera plastid chromosomes (plastomes). Mol. Gen. Genet. 251:185-188.

INNIS, M. A., D. H. GELFAND and J. J. SNINSKY, 1990 PCR Protocols: A Guide to Methods and Applications. Academic Press, Boston.

LEVISON, G. and G. A. GUTMAN, 1987  Slipped-strand mispairing: a major mechanism for DNA sequence evolution. Mol. Biol. Evol. 4:203-221[Abstract].

LOVETT, S. T., P. T. DRAPKIN, V. A. SUTERA, JR., and T. J. GLUCKMAN-PESKIND, 1993  A sister-strand exchange mechanism for recA-independent deletion of repeated DNA sequences in Escherichia coli.. Genetics 135:631[Abstract].

MEYER, R. R. and P. S. LAINE, 1990  The single-stranded DNA-binding protein of Escherichia coli.. Microbiol. Rev. 54:342-380[Abstract/Free Full Text].

SAMBROOK, J., E. F. FRITSCH and T. MANIATIS, 1989 Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

SEARS, B. B. and M. B. SOKALSKI, 1991  The Oenothera plastome mutator: effect of UV-irradiation and nitroso-methyl urea on mutation frequencies. Mol. Gen. Genet. 229:245-252[Medline].

SEARS, B. B., L. L. STOIKE, and W.-L. CHIU, 1996  Proliferation of direct repeats near the Oenothera chloroplast origin of replication. Mol. Biol. Evol. 13:850-863[Abstract].

WOLFSON, R., K. HIGGINS, and B. B. SEARS, 1991  Evidence for replication slippage in the evolution of Oenothera chloroplast DNA. Mol. Biol. Evol. 8:709-720[Abstract].