DNA Recombination Activity in Soybean Mitochondria

doi:10.1016/j.jmb.2005.11.070

J. Mol. Biol. (2006) 356, 288–299

DNA Recombination Activity in Soybean Mitochondria Medha Manchekar1, Karyn Scissum-Gunn1, Daqing Song2, Fayaz Khazi1 Stephanie L. McLean1 and Brent L. Nielsen1,2* 1

Department of Biological Sciences, 101 Life Sciences Building, Auburn University Auburn, AL 36849, USA 2 Department of Microbiology & Molecular Biology, 775 WIDB Brigham Young University Provo, UT 84602, USA

Mitochondrial genomes in higher plants are much larger and more complex as compared to animal mitochondrial genomes. There is growing evidence that plant mitochondrial genomes exist predominantly as a collection of linear and highly branched DNA molecules and replicate by a recombination-dependent mechanism. However, biochemical evidence of mitochondrial DNA (mtDNA) recombination activity in plants has previously been lacking. We provide the first report of strand-invasion activity in plant mitochondria. Similar to bacterial RecA, this activity from soybean is dependent on the presence of ATP and Mg2C. Western blot analysis using an antibody against the Arabidopsis mitochondrial RecA protein shows cross-reaction with a soybean protein of about 44 kDa, indicating conservation of this protein in at least these two plant species. mtDNA structure was analyzed by electron microscopy of total soybean mtDNA and molecules recovered after field-inversion gel electrophoresis (FIGE). While most molecules were found to be linear, some molecules contained highly branched DNA structures and a small but reproducible proportion consisted of circular molecules (many with tails) similar to recombination intermediates. The presence of recombination intermediates in plant mitochondria preparations is further supported by analysis of mtDNA molecules by 2-D agarose gel electrophoresis, which indicated the presence of complex recombination structures along with a considerable amount of single-stranded DNA. These data collectively provide convincing evidence for the occurrence of homologous DNA recombination in plant mitochondria. q 2005 Elsevier Ltd. All rights reserved.

*Corresponding author

Keywords: plant mitochondrial DNA recombination; strand-invasion assays; soybean; two-dimensional agarose gel electrophoresis

Introduction A wide range of sizes of mitochondrial genomes exists across various classes of organisms. Higher

Present addresses: M. Manchekar, Atherosclerosis Research Unit, University of Alabama at Birmingham, AL, USA; K. Scissum-Gunn, Department of Mathematics and Science, Alabama State University, Montgomery, AL, USA; F. Khazi, Howard Hughes Medical Institute, The Children’s Hospital of Philadelphia, PA, USA; S. L. McClean, Division of Clinical Immunology & Rheumatology, University of Alabama at Birmingham, AL, USA. Abbreviations used: mtDNA, mitochondrial DNA; PFGE, pulsed field gel electrophoresis; FIGE, field inversion gel electrophoresis; EM, electron microscopy; ss, single-stranded. E-mail address of the corresponding author: [email protected]

plant mitochondrial genomes (208–2000 kbp) are much larger than their counterparts in vertebrates (16–17 kbp) or fungi (w25–80 kbp).1,2 The extensively characterized circular animal mitochondrial genome is highly conserved among species, contains no introns, and has a very limited amount of intergenic sequence.3 Plant mitochondrial DNA (mtDNA) contains introns in many genes and some additional expressed genes as compared to animal mitochondria, but most of the additional sequences in plants are not expressed and do not appear to be essential.4 The complete mitochondrial genome sequences are available for a number of higher plants, including Arabidopsis thaliana,5 and the liverwort Marchantia polymorpha.6 Over the past several years there have been many reports that the mitochondrial genomes in yeast and higher plants exist as primarily linear and branched DNA molecules of widely varied sizes

0022-2836/$ - see front matter q 2005 Elsevier Ltd. All rights reserved.

289

Soybean MtDNA Recombination

that are considerably smaller than the predicted genome size.7–12 Using pulsed field gel electrophoresis (PFGE) of in-gel lysed mitochondria from various species, we and others have reported that only about 6–12% of the molecules are circular.11,13– 16 The branched molecules observed are very similar to recombination intermediates observed for yeast mtDNA 17 or with T4 phage DNA replication.18,19 Backert & Borner reported visualization of replication and recombination intermediates in mtDNA from Chenopodium album, and many of the linear DNA molecules observed were found to have an extended single-stranded end.20 Such ends may be involved in strand invasion into homologous regions of double-stranded DNA. Recombination of animal mtDNA appears to occur very rarely.21 Restriction maps of nearly all plant mitochondrial genomes predict a master circle with subgenomic circular molecules resulting from recombination across large direct repeats.1,13,22–28 Molecules of these predicted sizes, however, are in very low abundance or not detected. This can be explained if plant mitochondrial genomes are circularly permuted, as with phage T4.18,19 Oldenburg & Bendich reported that the predominantly linear mtDNA molecules in Marchantia, a lower plant, are circularly permuted with random ends.12 There is growing evidence that replication of plant mtDNA occurs by a T4-like recombination-dependent mechanism.19 Most higher plant mitochondrial genomes contain at least one pair of large (O1 kbp) direct repeats, which may be involved in homologous recombination to generate subgenomic molecules.27,29 However, it is apparent that these large repeats are insufficient to explain the range in sizes and structures of molecules observed. Many smaller repeats, ranging from as little as nine to a few hundred base-pairs, have been identified in plant mitochondrial genomes of a number of species and may be involved in homologous recombination.29 It has been proposed that the large repeats are involved in reversible recombination, while the small repeats may result in stable rearrangements.2,29 There are three mitochondrial plasmids in broad bean, and several variant plasmids have been characterized that appear to have arisen from active double recombination across repeated elements.30 There is evidence that active mtDNA recombination occurs in at least some plant species. Some individual plant mitochondria may contain less than a full genome equivalent of DNA, giving support for the possibility of deletion by recombination between direct repeats of portions of the genome.31,32 Genome rearrangements thought to be due to mtDNA recombination have been reported in a variety of plant species including A. thaliana.24,25,33–36 Both large (6.5 and 4.2 kb) direct repeats and numerous smaller (30–560 bp, totaling 144 in number) repeats are present in A. thaliana mtDNA.5 There is evidence that recombination across the large repeats occurs often but with no loss of subsequent rearrangements,35 while the small repeats may not be active in

frequent recombination.5 Very short repeated sequences have also been reported to be involved in mtDNA recombination in animal37 and yeast cells.38 In soybean mitochondria, sequence repeats of 9, 23 and 299 bp have been characterized.33,34 Numerous rearrangements of genome sequences among various soybean cultivars have been characterized and appear to have arisen by homologous recombination across these repeats,33,39,40 or across short elements that are part of a 4.9 kb PstI fragment of soybean mtDNA.41 The 299 bp repeat has been found in several copies in mtDNA from soybean and several other higher plants, suggesting that this sequence may represent a hot spot for mtDNA recombination in multiple plant species.34,41 These previous reports suggest that active homologous mtDNA recombination occurs in at least some plant species. We recently reported the identification of a mitochondrial-targeted homologue of the Escherichia coli recA gene in A. thaliana.42 However, to date there have been no reports on the characterization of recombination activity in plant mitochondria. We provide here the first report of such activity in soybean. These findings are supported by analysis of soybean mtDNA by electron microscopy and twodimensional agarose gel electrophoresis.

Results Strand transfer activity in partially purified soybean mitochondrial protein fractions Soybean mitochondria isolated on Percoll gradients42 were used for preparation of a total lysate, which was enriched for DNA recombination activity by DEAE-cellulose and heparin-Sepharose chromatography as described in Materials and Methods. Fractions 6–8 from the heparin-Sepharose column supported formation of joint molecules, indicating strand-exchange between the linear labeled double-stranded DNA and homologous circular single-stranded DNA molecules (Figure 1).

Figure 1. Detection of homologous strand exchange activity in soybean mitochondrial proteins eluted from the heparin-Sepharose column. Lane E, E. coli RecA positive control. KP, minus protein contol. L, loading fraction (contains nuclease activity). Lanes 2–12, elution fractions from the column. Peak activity was observed in fractions 6–8. Nucleases that degrade the substrate DNA appear to be present in lanes 9–12.

290 Later fractions (9–12) appeared to contain significant nuclease activity that led to degradation of the substrate. The control assay without any mitochondrial protein fraction (lane KP) showed that the formation of joint molecules is protein-dependent. Optimum strand exchange activity was found to be dependent upon the presence of ATP and Mg2C (Figure 2), which are conserved properties of bacterial RecA proteins. In the absence of both Mg2C and ATP, activity was only 15% of that obtained in the complete reaction (Figure 2(b)). Some activity was consistently observed in the absence of only ATP (Figure 2(a), lane 6). The amount of activity in the absence of ATP was found to average just over 33% (Figure 2(b)). This activity may be due to the presence of some ATP still bound to the enzyme. It may also be possible that this plant enzyme may retain some activity in the absence of ATP. An antibody against a unique region at the N-terminal end of the Arabidopsis mtRecA protein42 was used to determine if the soybean mitochondrial RecA activity is related to Arabidopsis RecA. This antibody cross-reacts with a protein of about 44 kDa in the heparin-Sepharose fraction containing soybean mitochondrial RecA activity (Figure 3, lane 2), which represents the mature protein with the mitochondrial targeting N-terminal presequence

Soybean MtDNA Recombination

Figure 3. Western blot analysis of mitochondrial RecA proteins in soybean and Arabidopsis. Lane 1, molecular mass markers, with sizes shown at left; lane 2, total protein (10 mg) from the soybean mitochondrial heparinSepharose fraction used for activity assays in Figure 1; lane 3, total protein (120 mg) from soybean leaves; lane 4, total protein (100 mg) from Arabidopsis leaves.

removed. Lane 3 shows total protein from soybean leaf tissue, including both the precursor and mature proteins. A very faint band of the same size as the precursor protein is also observed in lane 2, which may represent some contamination of precursor protein in the heparin-Sepharose preparation or protein that is within some stage of the protein import process. The mature soybean protein appears to be slightly larger than the Arabidopsis RecA protein (lane 4), while the precursor soybean protein in this species, prior to removal of the N-terminal sequence, is substantially larger than the Arabidopsis RecA precursor protein (compare lanes 3 and 4). This antibody does not react with chloroplast RecA, which is about 4 kDa smaller than the mitochondrial RecA activity that we observed.43 In addition, to confirm mitochondrial origin of this activity and rule out contamination by the chloroplast homologue, Western blot analysis was performed using an antibody against chlorophyll a/b binding protein, which indicates absence of chloroplast protein contamination in our mitochondrial preparations (data not shown here).42 Field-inversion gel electrophoresis separates plant mtDNA molecules

Figure 2. Soybean mitochondrial DNA strand invasion assay. (a) Lane 1, RF dsDNA only; lane 2, ssDNA only; lane 3, E. coli RecA control; lane 4, non-concentrated soybean mt protein. Lanes 5–8, concentrated soybean mt protein; lane 5, complete reaction; lane 6, no ATP; lane 7, no Mg2C; lane 8, no ATP or Mg2C. Arrows indicate the locations of the ssDNA and dsDNA substrates and the migration of joint molecules (labeled JM). (b) The chart shows the relative percentage of joint molecules produced under each set of reaction conditions. The results shown are an average from three independent experiments.

We have utilized field-inversion gel electrophoresis (FIGE) to separate populations of plant mtDNA.16 With this method the mitochondria are purified, embedded in agarose, and lysed in the agarose just prior to electrophoresis to minimize handling and shearing of the mtDNA.16 The markers used show the relative migration of linear versus circular DNA molecules under FIGE conditions. Depending on the specific FIGE conditions used mtDNA populations were separated differently within the gel (Figure 4). In addition, a significant proportion of mtDNA molecules remained well-bound under any conditions. At higher voltages (180 V, 30–35 mA), separation of long linear molecules was obtained, such as for the lambda concatamer (Figure 4, lane 4). However, the mtDNA samples under these conditions yielded only the fast moving band that migrates roughly at

291

Soybean MtDNA Recombination

monomer molecules can be seen), while separation of lambda monomer, dimer and trimer (lanes 7 and 12 as compared to lane 5) molecules was enhanced. The supercoiled plasmid markers also showed good separation over the size range from 8 to 165 kbp under these conditions (lanes 6 and 11). The 8 kbp plasmid marker migrates very near the location of the linear lambda monomer DNA in the gel (compare lanes 6 and 7, and lanes 11 and 12). In addition, under these conditions the turnip mitochondrial linear 11.3 kbp plasmid (lane 14) was very distinct.44 Bands were also observed near the compression zone for both soybean (lane 9) and turnip (lane 14), with a diffuse smear between this area and the faster moving thick band. These results show that similar electrophoresis patterns are obtained for mtDNA from both of these plant species. Branched molecules observed by electron microscopy Figure 4. FIGE analysis of mtDNA molecules. FIGE was performed under different conditions (see the text). Lanes 1, 2 and 9 contain soybean mtDNA; lane 14 contains turnip mtDNA, and the other lanes are either empty (lane 3) or contain marker DNA molecules. Markers include lambda concatamer (lanes 4, 8 and 13, the monomer band is quite faint in some lanes), lambda monomer DNA (lanes 5, 7 and 12; monomer and some dimer and trimer bands can be seen), and 1 kbp ladder (lane 10). Lanes 6 and 11 contain supercoiled circular plasmid DNA (Epicentre BAC marker with sizes of 165, 120, 95, 55, 38, 28 and 8 kbp, with bands of unequal intensity). Lanes 1–5, conditions (180 V, 35 mA) that show greater separation of the Lambda concatamer ladder (lane 4) and a fast diffuse band of linear mtDNA (lanes 1 and 2), while only a diffuse smear with no distinct bands of higher molecular mass DNA are observed in these lanes. Lanes 11–14, conditions (120 V, 25 mA) that do not allow separation of the lambda concatamer (lanes 8 and 13), with this marker concentrated in a compression zone (top arrow). Smaller linear molecules resolve well, as seen in lane 10 (1 kbp ladder). The fast-migrating band in lanes 1, 2 and 9 (lower arrow in lane 9) was found to contain linear DNA molecules ranging in size from about 4–45 kbp (see Figure 5(e)). The upper two arrows indicate regions of the gel from which DNA was recovered for electron microscopy (Figure 5). The arrows in this lane correspond from top to bottom to populations 1, 2 and 3, respectively (see Table 1). Lane 14, turnip mtDNA shows a similar pattern with the presence of higher molecular mass DNA near the compression zone and the distinct linear 11.3 kbp plasmid present in turnip mitochondria (bottom arrow in lane 14).44

the same rate as lambda monomer DNA (lanes 5, 7 and 12) and a diffuse continuum above this band that gets fainter towards the lanes (Figure 4, lanes 1 and 2). This diffuse band was found by EM to contain linear molecules of various sizes (Figure 5(e)). At lower voltages (120–130 V, 20– 25 mA), resolution of the lambda concatamer was lost and most of this marker DNA ran in a compression zone (Figure 4, lanes 8 and 13; some

DNA was recovered from different regions of the FIGE gel and analyzed by electron microscopy to determine conformation of individual molecules. Representative results are shown in Figure 5 (arrows in lane 9 indicate the areas of the gel from which DNA was recovered for analysis). About 75–80% of the total number of molecules observed were linear, while the remainder were either branched (Figure 5(a)), circular (Figure 5(b) and (c)) or contain complex networks (Figure 5(d)), with sizes of individual molecules ranging from about 16 kbp to 35 kbp. Approximately 14–18% of the molecules from the well or from the slower migrating populations 1 and 2 contained branched or complex networks as shown in Figure 5(d) (summarized in Table 1), which are similar to recombination intermediates.13,17,20 Network molecules were consistently observed over many samplings, suggesting that they are not artifacts of isolation. Some of the circular molecules appear to have tails that may be a result of recombination or rolling circle replication (Figure 5(b) and (c)). The circular mtDNA molecules are much smaller (about 18 kbp in (b) and 6 kbp in (c)) than the expected total mitochondrial genome size and represent only 4–7% of the total number of mtDNA molecules observed from within the gel (Table 1). The fastest moving, diffuse population (Figure 4, lanes 1, 2 and 9) was found to consist mostly of linear DNA molecules ranging in size from approximately 4–45 kbp (Figure 5(e)). The fast-migrating population 3 was also found to contain networks, but these were of much smaller size and complexity as compared to those from the other parts of the gel. The well-bound fraction (observed in Figure 4, lanes 9 and 14, but lost in lanes 1 and 2) was found to consist of circular, rosette-type or linear DNA molecules ranging in size from about 150 kbp or larger (Table 1 and Figure 5(f) and (g)). To ensure that the molecules observed by EM were not an artifact of the FIGE or subsequent DNA

292

Soybean MtDNA Recombination

Figure 5. Electron microscopy of FIGE-separated mtDNA fractions. (a)–(d) From slower migrating regions of the gel (Figure 4, lane 9, upper arrow for molecules in (a) and (c), middle arrow for molecules in (b) and (d)); (e) from fast migrating, diffuse band (bottom arrow in Figure 4, lane 9). (f) and (g) Well-bound mtDNA, showing typical long linear DNA molecules, with some rosette-like nature in the center of the molecule in (f). The bar in each panel represents 1 mm except for in (c), where the bar represents 0.5 mm.

recovery, total mtDNA from gently lysed soybean mitochondria were directly analyzed without agarose gel separation. Highly branched molecules were observed (Figure 6(a)), as well as circular molecules with tails (Figure 6(b)) and linear and circular molecules (Figure 6(c) shows a circle of about 51 kbp and a separate linear molecule of about 20 kbp). Analysis of mtDNA recombination intermediates by two-dimensional gel electrophoresis A schematic diagram of patterns expected from neutral/neutral two-dimensional agarose gel electrophoresis of recombination and replication intermediates is shown in Figure 7. The various types of DNA structures expected for each gel pattern are indicated. In general, the more complex the molecule, the greater the effect on migration. In addition to the typical structures expected for replicating DNA, the region where complex recombination structures migrate is indicated (labeled R), and this is the region where multiply branched molecules migrate.45,46 Soybean mtDNA was restricted with PstI and separated by two-dimensional gel electrophoresis. Faint simple Y patterns (labeled Y in Figure 8) were observed in 2-D gels when probed with a 3.0 kbp (Figure 8(a) and (b)), 1.8 kbp (Figure 8(c) and (d)), 5.35 kbp (Figure 8(e) and (f)) or other (not shown) soybean mtDNA PstI restriction fragments. In some cases the simple Y arc was very faint or smaller than

expected for the probe used, and difficult to resolve from the background due to a strong signal associated with the monomer fragment (Figure 8). Complex recombination intermediates (labeled R) were detected with each probe, and these signals were not sensitive to S1 digestion, in agreement with the inferred complexity of the structures. In addition to the other patterns, an intersecting arc was also observed with each probe (labeled S in Figure 8). This arc is sensitive to S1 nuclease digestion (Figure 8(b), (d) and (f)) and likely represents single-stranded (ss) DNA or fragments containing ssDNA. Double Y-like patterns were also observed and generally became much more apparent after digestion with S1 nuclease. Fractionation of soybean mtDNA molecules through BND-cellulose provided further support for the presence of single-stranded regions.47 The high salt elution step from BND-cellulose shows a Table 1. Summary of electron microscopy analysis of soybean mtDNA molecules recovered from FIGE gels Percentage of molecules

Well-bound Population 1 Population 2 Population 3

Networks

Linear

Circular

Rosette

Total no. of samples

18 17.6 14 23.5

66 75.8 82 74.5

12.5 6.5 4 2

3.5 0 0 0

56 153 206 98

Soybean MtDNA Recombination

293

Figure 6. Electron microscopy of soybean mtDNA molecules from gently lysed isolated mitochondria to minimize shearing. (a) Complex branched molecule containing networks; (b) circular molecule with short tail; (c) linear and circular DNA molecules. These molecules are all smaller than the expected total genome size. The bar in each panel represents 1 mm.

Figure 7. Schematic of potential 2-D gel electrophoresis patterns.45,55,69 The directions of electrophoresis in the first and second dimensions are shown, as well as migration of linear monomer (M) and dimer (D) molecules. Other abbreviations: Y, simple Y forks; DY, double Y molecules; L, arc of linear molecules; X, simple recombination structures; R, complex recombination intermediates; S, arc of single-stranded DNA molecules. Sketches of molecules that contribute to each pattern are also shown (not to scale; complex molecules may range widely in size).

Figure 8. Two-dimensional gel analysis of in vivo replication intermediates in PstI-restricted soybean mtDNA. In each pair of panels, the second was treated with S1 nuclease to identify patterns specific to single-stranded DNA molecules. Orientation of electrophoresis is the same as shown in the schematic. (a) 3.0 kbp PstI fragment as probe, KS1 nuclease treatment. (b) Same as (a) CS1 nuclease. (c) 1.8 kbp PstI fragment as probe, KS1 nuclease. (d) Same as (c) CS1 nuclease. (e) 5.35 kbp PstI fragment as probe, KS1 nuclease. (f) Same as (e) CS1 nuclease. S1 nuclease-sensitive single-stranded arcs are labeled with an S, simple Y patterns are labeled Y, potential double Y arcs are labeled DY, and complex recombination structures are labeled R. M, location of linear monomer molecules hybridizing with the probe; D, dimers.

294

Figure 9. Two-dimensional gel electrophoresis of BND cellulose-enriched soybean mtDNA. (a) Salt wash fraction after binding of restricted soybean mtDNA to the BND cellulose column. (b) High salt plus caffeine elution fraction of mtDNA from BND cellulose. Blots were probed with the 3.0 kbp soybean mtDNA fragment used in Figure 8. Abbreviations are the same as in Figure 8.

very strong double-stranded monomer spot but does not show the single-stranded arc (Figure 9(a)), while DNA from the high salt plus caffeine elution shows a strong single-stranded arc in addition to a faint and smaller than expected (see above) simple Y pattern (Figure 9(b)). This fraction also shows a very strong enrichment of complex recombination intermediates (labeled R in Figure 9(b)). This pattern is very similar to observations of recombination intermediates in malarial parasite mtDNA,45 while it is absent in the BND wash fraction (Figure 9(a)).

Discussion Homologous DNA strand invasion activity in soybean mitochondria We have optimized in vitro DNA recombination assay conditions using E. coli RecA as a positive control, and have applied these conditions to test for strand exchange activity in plant mitochondrial extracts. These assays have confirmed the presence of homologous strand exchange activity in partially purified soybean mitochondrial protein fractions (Figure 1) and we have shown that this activity is dependent on ATP and Mg2C (Figure 2). In the absence of ATP about 1/3 of the activity was retained as compared to the complete reaction (Figure 2, lane 6). This level of activity in the absence of ATP was consistent with all our preparations, and may be due to the presence of some residual ATP remaining in the protein preparation. Alternatively, this enzyme may retain some activity in the absence of ATP, although this would distinguish it from bacterial RecA. Nearly

Soybean MtDNA Recombination

undetectable activity is observed in the absence of Mg2C only or both ATP and Mg2C (lanes 7 and 8). The general properties of this enzyme are similar to those of bacterial RecA proteins and the previously reported chloroplast RecA homologue.43,48 This is consistent with our recent report of an Arabidopsis nuclear gene that encodes a mitochondrial targeted homologue of E. coli RecA distinct from chloroplast RecA.42 Western blot analysis using an antibody specific for a unique region at the N-terminal end of the Arabidopsis mtRecA42 homologue confirms homology with the soybean mitochondrial RecA-like activity. The antibody recognizes both the precursor protein (higher Mr band) including the mitochondrial targeting sequence and the mature protein from mitochondria that has had the targeting sequence cleaved off (lower Mr band w44 kDa). The soybean mitochondrial RecA homologue appears to have a considerably longer targeting sequence of w10 kDa as compared to w3 kDa for the Arabidopsis mtRecA.42 The DNA sequence for the soybean gene is not yet available for analysis, but when it is it will be interesting to compare the targeting and mature sequences between the Arabidopsis and soybean homologues. Analysis of mtDNA structure by FIGE and EM analysis Pulsed field gel electrophoresis (PFGE) is used for separation of very large DNA molecules that do not resolve in traditional agarose gels. Various types of PFGE separations have been developed, including CHEF, in which DNA molecules migrate in alternating diagonal angles down the gel, and FIGE, which utilizes a complete reversal of polarity at a specific forward/reverse time ratio during electrophoresis. FIGE has been shown to be useful for separating large open circle molecules, as it allows these molecules to disentangle from the agarose fibers during each reversal phase of the run.49 FIGE was previously used in our laboratory to separate different populations of turnip mtDNA.17 This method has been shown to have some advantages over CHEF pulsed field gel electrophoresis for separation of non-linear DNA molecules.50 It appears from our work that both methods have advantages and limitations. Alteration of FIGE conditions can be used to modify the resolution of non-linear DNA molecules. Under certain FIGE conditions (i.e. lower voltage) the mtDNA from gently lysed purified mitochondria separates into some distinct bands but with diffuse areas above and below the bands (Figure 4). The bands that form between the well and the fastest migrating band may represent the compression zone that is commonly observed for large molecules separated under some PFGE conditions, and is one potential advantage of FIGE. With FIGE performed at higher voltages the mtDNA does not form distinct lower mobility bands, but rather separates as a continuum of

Soybean MtDNA Recombination

slower mobility molecules and a faster migrating, diffuse band, suggesting a range in sizes and shapes of mtDNA molecules. The patterns are very similar for both soybean and turnip (Figure 4), except for the additional linear plasmid band observed in turnip mitochondrial samples. This suggests that the mtDNA in these two species may be quite similar but may consist of a wide range of size and complexity of conformation. However, due to the wide variety of molecules recovered from the gel and observed by EM (Figure 5) it is not possible to predict the sizes of the mtDNA molecules by comparison with either linear or circular marker DNA molecules included in the FIGE gel. Only the linear plasmid in turnip mtDNA migrates as predicted based on size (Figure 4). DNA populations recovered from the FIGE gels were found to contain a small proportion of branched, circular, or other complex molecules of various sizes (Figure 5), which were similar to molecules observed upon examination of total mtDNA spreadings from carefully purified mitochondria to minimize damage due to shearing (Figure 6). This suggests that the branched and circular molecules are not an artifact of the FIGE separation or gel recovery, and these findings are similar to those observed for turnip mtDNA.16 From these results it appears that branched and complex molecules constitute about 14–18% of the molecules that separate in the FIGE gel. A significant proportion of mtDNA molecules remain trapped in the well, and were found to mostly consist of long linear or complex networked molecules. These observations are consistent with those reported for yeast and other plant mtDNA.10,11,15,51 While most of the molecules observed by EM are linear, a small proportion of circular and branched molecules were observed in both total lysed mitochondrial samples and in FIGE-separated fractions (slower migrating fractions). The circular and branched molecules observed are much smaller than the predicted genome size, and resemble recombination intermediates similar to those reported for yeast17 and Plasmodium falciparum parasite45 mtDNA. Strand invasion of small circular molecules by linear molecules could generate the rolling circle-like structures observed by EM (Figure 5), or the circles could be generated by the recombination of one end of a linear molecule to a homologous region within the molecule. It may also be possible that some of these circular molecules are mitochondrial plasmids, which are known in some plant species, though their presence has never been reported in soybean mitochondria. In Brassica species, there is a linear mtDNA plasmid, and a very small circular plasmid in Chenopodium mitochondria has been characterized by Backert et al.52 The circles observed in soybean mitochondria vary widely in size, suggesting either the presence of a number of plasmids, or more likely that they have arisen by some mechanism such as recombination.

295 Two-dimensional agarose gel electrophoresis Two-dimensional gel analysis of replication intermediates in total soybean mtDNA identified simple Y patterns, as well as putative complex recombination forms for each fragment used as probe (Figures 8 and 9). Some Y arcs are smaller than expected, and may represent molecules with unequal fork lengths as would be expected for recombining DNA molecules that share homology to the probe (such as the forked molecule observed by EM in Figure 5(a)). Treatment with S1 nuclease removed the ssDNA arc, but also appeared to result in a significant concentration of double Y (DY) forms. These may represent recombination intermediates where two linear molecules are joined by a third molecule, which are not as pronounced in the presence of ssDNA regions within the same molecules. The upper ends of the DY patterns extend into and overlap the complex R region and may contribute to the greater intensity observed in the R regions in Figure 8(b) and (d). The R structures could be formed by recombination between large or small repeats present within the mtDNA.24,53 The strength of the complex molecule signal may suggest that only a small percentage of mtDNA molecules undergo recombination at any given time. These recombination forms are similar to those observed for recombination-dependent replication in the malarial parasite,45 and for yeast where the presence of unresolved recombination intermediates appears to influence the segregation of mtDNA molecules.54 Strong single-stranded DNA arcs that are susceptible to S1 nuclease digestion were observed for each probe. Single-stranded arcs are not commonly observed in 2-D gels for systems that replicate by most mechanisms. When T4 phage DNA is nicked it can be induced to replicate by a rolling circle mechanism in vitro, and in the absence of DNA primase generates single-stranded arcs in 2-D gels similar to the patterns that we observed.55 The presence of extended single-stranded regions in soybean mtDNA was further supported by the strong binding of a significant population of mtDNA fragments to BND-cellulose (Figure 9(b)), which has a very strong interaction with singlestranded nucleic acids. 56 This treatment also resulted in a very strong concentration of complex recombination forms, suggesting that these molecules contain a considerable amount of singlestranded regions. Complex recombination structures would be expected to contain some singlestranded regions and thus bind to BND-cellulose, as we observed. BND-cellulose has been used to enrich replicating DNA molecules from total DNA in other nuclear and organelle systems, based on the presence of short single-stranded regions at replication forks and at the junctions of Okazaki fragments before they are processed and fused together.57–59At the same time, at least some of the complex molecules appear to be resistant to S1 digestion as shown in Figure 8. This indicates that these

296 molecules likely contain a mixture of both doublestranded and single-stranded regions. A significant amount of single-stranded DNA has been observed in mtDNA preparations from various plant species8,10 and from yeast, where single-stranded regions affect PFGE patterns.51,60 Single-stranded regions of DNA molecules may be directly involved in strand-exchange steps of recombination and may be a result of DNA unwinding after introduction of nicks by either a random or sequence-specific process. A doublestranded break, followed by preferential degradation of one strand by a 5 0 -3 0 exonuclease, would lead to single-stranded regions with a free 3 0 end, which is the preferred element for strand invasion.20 Similar fragments could also be generated by exonuclease activity on linear DNA molecules, which appear to be the prevalent form of mtDNA.9,11,16 Alternatively, single-stranded DNA may result from rolling circle DNA replication if the tail does not undergo lagging strand synthesis.8,52,60,61 Taken together, the results presented here provide strong support for the occurrence of active DNA recombination in soybean mitochondria. The relatively low amounts of recombination structures observed in 2-D gels (faint R patterns relative to strong linear monomer signals) and by EM analysis suggest that only a small portion of mtDNA molecules are undergoing recombination at any given time. The next question to be considered is whether mtDNA recombination is essential for mitochondrial function and plant growth. Analysis of mutants in mtDNA recombination proteins in Arabidopsis is currently underway.

Materials and Methods Isolation of soybean mitochondrial strand invasion activity Mitochondria from young soybean seedlings were purified on Percoll gradients as described.62 The mitochondrial band was recovered and the mitochondria were pelleted by centrifugation and washed. The pellets were resuspended in a total of 10 ml of water containing 0.1% Triton X-100 and incubated on ice for 20 min to lyse the mitochondria. The sample was centrifuged at 15,000g for 20 min, and the supernatant was recovered and loaded on a 30 ml DEAE cellulose column equilibrated with buffer A (50 mM Tris–acetate (pH 7.5), 20% (v/v) glycerol, 50 mM 2-mercaptoethanol, 0.2 mM PMSF, 0.1% Triton X-100) containing 50 mM sodium acetate. The column was then washed with 200 ml buffer A, followed by elution with buffer A containing 500 mM sodium acetate. Selected elution fractions were collected, pooled and dialyzed in buffer A overnight with two buffer changes. The dialyzed proteins were then loaded onto an equilibrated heparin Sepharose column (w5 cm3 column volume) at 0.3 ml/min. The column was washed with 50 ml of buffer A at 0.33 ml/min. Proteins were eluted with buffer A containing 0.6 M sodium acetate at 0.5 ml/ min, and fractions assayed for strand invasion activity.

Soybean MtDNA Recombination

DNA substrates M13 replicative form (RF) dsDNA and M13 singlestranded DNA were prepared as described.63 DNA concentrations were determined by spectrophotometry (BioRad SmartSpec 3000). Replicative form DNA was linearized by digestion with BamHI (BioLab), separated in an agarose gel and recovered using a QIAquick kit (QIAGEN). Soybean mtDNA strand invasion assays Protein fractions were assayed for homologous strand invasion activity as described.48,64 The reaction mixtures contained: 30 mM Tris–acetate (pH 7.5), 2 mM DTT, 50 mg/ml BSA, 20 mM magnesium acetate, 2.5 mM ATP, 15 ml protein fraction (for control, added 2 mg E. coli RecA (New England BioLab)), 15 mM ssDNA and 15 mM RF dsDNA, 1.8 mg E. coli single-stranded binding protein (SSB) (Promega) and dH2O to 25 ml final volume. In the first step of the reaction, Tris–acetate (pH 7.5), DTT, BSA, MgCl2, ATP, protein fraction (RecA for control), and dH2O were combined and incubated at 37 8C for 5 min. In the second step, ssDNA was added and incubation continued at 37 8C for 10 min. In the third step, SSB was added with incubation at 37 8C for another 10 min. For the fourth step, RF dsDNA was added and the samples incubated at 37 8C for 60 min. Stop solution (0.5% (w/v) SDS, 50 mM EDTA) was then added followed by incubation for 5 min at 37 8C and 65 8C for 5 min. The DNA was recovered and separated by agarose gel electrophoresis, and visualized after staining by ethidium bromide. Isolation of soybean mitochondria for DNA analysis Approximately 500 g of 1.5 to 2 week old etiolated soybean (Glycine max Merr. var Essex) seedlings were homogenized in 150–200 g batches with two to three volumes (ml/g) of cold STM buffer (0.5 M sucrose, 50 mM Tris–HCl (pH 8.0), 5 mM MgCl 2, 5 mM b-mercaptoethanol, 0.2 mM PMSF) in an ice-cold blender jar with three 5 s bursts.65 Subsequent steps were carried out at 4 8C. The resulting homogenate was filtered through four layers of cheesecloth and three layers of Miracloth (Calbiochem). The filtrate was centrifuged at 2500 rpm in a GSA (Sorvall) rotor for 20 min to remove chloroplasts and nuclei, and the resulting supernatant was centrifuged at 12,000 rpm in a GSA rotor for 30 min to pellet mitochondria. The pellets were resuspended in a total of 200 ml cold STM buffer, both centrifugation steps were repeated, and the final pellets were resuspended for DNA purification or for direct FIGE or EM analysis of DNA molecules. FIGE analysis of soybean mtDNA The mitochondrial pellet from above was resuspended in 1 ml of wash buffer, mixed with an equal volume of 1.2% (w/v) low-melting point agarose prepared in 0.5X TBE and allowed to solidify in plug-molds. Mitochondrial membranes were digested by incubating plugs in lysis buffer (1% SDS, 0.1 M EDTA, 0.5 mg/ml proteinase K) at 50 8C for 3 h overnight. The plugs were washed in several volumes of TE (10 mM Tris–HCl (pH 8.0), 1 mM EDTA). Field inversion gel electrophoresis was conducted essentially as described.16 The mtDNA from lysed mitochondria was resolved by FIGE for 40 h at 4 8C, with pulse

297

Soybean MtDNA Recombination

times ramped from 6 s to 24 s in a 3:1 TF:TR ratio.66 Two sets of conditions were used, the first at 35 mA (180 V), the second at 25 mA (125 V). Lambda DNA concatemers were used as high molecular mass linear size markers. After electrophoresis, a portion of the gel was stained with 0.3 mg/ml ethidium bromide and photographed. The remainder of the gel was used for recovery of DNA for EM. Electron microscopy analysis of soybean mtDNA FIGE-separated mtDNA samples were excised from the unstained portion of the gel and dialyzed in Gelase (Epicenter) buffer for 1 h at room temperature. Gel pieces were then heated at 70 8C until completely melted, cooled to 45 8C and digested for 1 h with one unit of Gelase. For EM analysis, 20 ml of the Gelase-digested sample or total soybean mtDNA was mixed in 200 ml of buffer containing 0.25 M ammonium acetate, 0.01 M Tris–HCl (pH 8.0), 0.001 M EDTA and 3 mg/ml cytochrome c, and 45 ml of this mixture was deposited on a Teflon block.67 The DNA was allowed to diffuse for 20 min, and then transferred to parlodion-coated, 200 mesh copper grids by touching the grid to the drop surface. The grids were washed in dH2O for 5 s, flooded with 140 ml of 0.4% Kodak Photoflo and stained for 30 s in 30 ml of 1% uranyl acetate.68 The grids were washed in 90% (v/v) ethanol for 10 s, dried and viewed with a Zeiss transmission electron microscope operating at 60 kV. In addition, total mtDNA molecules were observed directly from lysed mitochondria without heating or Gelase treatment. Purification of soybean mtDNA Mitochondrial pellets from above were resuspended and treated with DNase I,16 and mtDNA was purified by cesium chloride gradient centrifugation.16,27 The purity of the mtDNA was analyzed by probing a Southern blot63 of restriction-digested DNA with a tobacco nuclear 18 S rRNA gene and a chloroplast rpl16 gene labeled by random priming (Prime-A-Gene, Promega), with no detectable hybridization (not shown). 2-D gel analysis of restricted soybean mtDNA Total soybean mtDNA was restricted with PstI prior to gel separation. For some experiments DNA samples were enriched for molecules containing single-stranded regions by passage through BND-cellulose.47,56 The high salt plus caffeine elution fraction was concentrated by precipitation with ethanol and separated by 2-D gel electrophoresis. Samples were loaded in a 0.5% agarose gel without EtBr, and electrophoresed in the first dimension at 1 V/cm overnight in 1X TBE buffer.69 The gel was stained with 0.3 mg/ml EtBr and photographed, and the sample lanes were cut apart and placed horizontally at the top of a second gel. The second dimension gel contained 1% agarose and 0.3 mg/ml EtBr in 1X TBE buffer.69 The gel was electrophoresed at 5 V/cm for 6 h with buffer recirculation, after which the DNA was transferred to nylon membrane by Southern blotting.63 Hybridization with specific soybean mtDNA PstI fragments was followed by exposure to X-ray film to visualize replication intermediates specific for each probe. To identify patterns resulting from single-stranded DNA molecules or regions, some samples were treated with S1 nuclease (two units in a 30 ml reaction) for 20 min at 37 8C

after restriction digestion, prior to electrophoresis of the sample in the first dimension.59

Acknowledgements This work was supported in part by grants from the NIH (GM066787), the Mentoring Environment Grants Program (BYU), and by the Office of the Vice President for Research, Auburn University. K.S.G. was supported in part by an NIH-MARC faculty fellowship (GM15475). We thank John Cupp, Benn Fronk and Joel Hall for assistance in data analysis and preparation of some of the Figures.

References 1. Mackenzie, S. & McIntosh, L. (1999). Higher plant mitochondria. Plant Cell, 11, 571–585. 2. Schuster, W. & Brennicke, A. (1994). The plant mitochondrial genome: physical structure, information content, RNA editing, and gene migration to the nucleus. Annu. Rev. Plant Physiol. Plant Mol. Biol. 45, 61–78. 3. Larsson, N.-G. & Clayton, D. A. (1995). Molecular genetic aspects of human mitochondrial disease. Annu. Rev. Genet. 29, 151–178. 4. Binder, S., Marchfelder, A. & Brennicke, A. (1996). Regulation of gene expression in plant mitochondria. Plant Mol. Biol. 32, 303–314. 5. Unseld, M., Marienfeld, J., Brandt, P. & Brennicke, A. (1997). The mitochondrial genome of Arabidopsis thaliana contains 57 genes in 366,924 nucleotides. Nature Genet. 15, 57–61. 6. Oda, K., Yamato, K., Ohata, E., Nakamura, Y., Takemura, M., Nozato, N. et al. (1992). Gene organization from the complete sequence of liverwort, Marchantia polymorpha, mitochondrial DNA: a primitive form of plant mitochondrial genome. J. Mol. Biol. 223, 1–7. 7. Backert, S., Dorfel, P. & Borner, T. (1995). Investigation of plant organellar DNA by pulsed-field gel electrophoresis. Curr. Genet. 28, 390–399. 8. Backert, S., Lurz, R., Oyarzabal, O. A. & Borner, T. (1997). High content, size and distribution of singlestranded DNA in the mitochondria of Chenopodium album (L.). Plant Mol. Biol. 33, 1037–1050. 9. Bendich, A. J. & Smith, S. B. (1990). Moving pictures and pulsed-field gel electrophoresis show linear DNA molecules from chloroplasts and mitochondria. Curr. Genet. 17, 421–425. 10. Bendich, A. J. (1996). Structural analysis of mitochondrial DNA molecules from fungi and plants using moving pictures and pulsed-field gel electrophoresis. J. Mol. Biol. 255, 564–588. 11. Oldenburg, D. J. & Bendich, A. J. (1996). Size and structure of replicating mitochondrial DNA in cultured tobacco cells. Plant Cell, 8, 447–461. 12. Oldenburg, D. J. & Bendich, A. J. (2001). Mitochondrial DNA from the liverwort Marchantia polymorpha: circularly permuted linear molecules, head-to-tail concatemers, and a 5 0 protein. J. Mol. Biol. 310, 549–562.

298 13. Backert, S., Nielsen, B. L. & Borner, T. (1997). The mystery of the rings: structure and replication of mitochondrial genomes from higher plants. Trends Plant Sci. 2, 477–483. 14. Jacobs, M. A., Payne, S. R. & Bendich, A. J. (1996). Moving pictures and pulsed-field gel electrophoresis show only linear mitochondrial DNA molecules from yeasts with linear-mapping and circular-mapping mitochondrial genomes. Curr. Genet. 30, 3–11. 15. Oldenburg, D. J. & Bendich, A. J. (1998). The structure of mitochondrial DNA from the liverwort, Marchantia polymorpha. J. Mol. Biol. 276, 745–758. 16. Scissum-Gunn, K. D., Gandhi, M., Backert, S. & Nielsen, B. L. (1998). Separation of different conformations of plant mitochondrial DNA molecules by field inversion gel eletrophoresis. Plant Mol. Biol. Rep. 16, 219–229. 17. Sena, E. P., Revet, B. & Moustacchi, E. (1986). In vivo homologous recombination intermediates of yeast mitochondrial DNA analyzed by electron microscopy. Mol. Gen. Genet. 202, 421–428. 18. Mosig, G. (1987). The essential role of recombination in phage T4 growth. Annu. Rev. Genet. 21, 347–371. 19. Mosig, G. (1998). Recombination and recombinationdependent DNA replication in bacteriophage T4. Annu. Rev. Genet. 32, 379–413. 20. Backert, S. & Borner, T. (2000). Phage T4-like intermediates of DNA replication and recombination in the mitochondria of the higher plant Chenopodium album (L.). Curr. Genet. 37, 304–314. 21. Sato, A., Nakada, K., Akimoto, M., Ishikawa, K., Ono, T., Shitara, H. et al. (2005). Rare creation of recombinant mtDNA haplotypes in mammalian tissues. Proc. Natl Acad. Sci. USA, 102, 6057–6062. 22. Andre, C. P. & Walbot, V. (1995). Pulsed-field gel mapping of maize mitochondrial chromosomes. Mol. Gen. Genet. 247, 255–263. 23. Bendich, A. J. (1993). Reaching for the ring: the study of mitochondrial genome structure. Curr. Genet. 24, 279–290. 24. Fauron, C., Casper, M., Gao, Y. & Moore, B. (1995). The maize mitochondrial genome: dynamic, yet functional. Trends Genet. 11, 228–235. 25. Fauron, C. M. R., Casper, M., Gesteland, R. & Albertsen, M. (1992). A multi-recombination model for the mtDNA rearrangements seen in maize cmsT regenerated plants. Plant J. 2, 949–958. 26. Palmer, J. D. & Herbon, L. A. (1988). Plant mitochondrial DNA evolves rapidly in structure, but slowly in sequence. J. Mol. Evol. 28, 87–97. 27. Palmer, J. D. & Shields, C. R. (1984). Tripartite structure of the Brassica campestris mitochondrial genome. Nature, 307, 437–440. 28. Palmer, J. D., Makaroff, C. A., Apel, I. J. & Shirzadegan, M. (1990). Fluid structure of plant mitochondrial genomes: evolutionary and functional implications. In Molecular Evolution (Clegg, M. T. & O’Brien, S. J., eds), pp. 85–96, Alan R. Liss, New York. 29. Andre, C. P., Levy, A. & Walbot, V. (1992). Small repeated sequences and the structure of plant mitochondrial genomes. Trends Genet. 8, 128–132. 30. Flamand, M. C., Duc, G., Goblet, J. P., Hong, L., Louis, O., Briquet, M. & Boutry, M. (1993). Variant mitochondrial plasmids of broad bean arose by recombination and are controlled by the nuclear genome. Nucl. Acids Res. 21, 5468–5473. 31. Satoh, M., Nemoto, Y., Kawano, S., Nagata, T., Hirokawa, H. & Kuroiwa, T. (1993). Organization of

Soybean MtDNA Recombination

32. 33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44. 45.

46.

heterogeneous mitochondrial DNA molecules in mitochondrial nuclei of cultured tobacco cells. Protoplasma, 175, 112–120. Small, I., Suffolk, R. & Leaver, C. J. (1989). Evolution of plant mitochondrial genomes via substoichiometric intermediates. Cell, 58, 69–76. Kanazawa, A., Tozuka, A., Kato, S., Mikami, T., Abe, J. & Shimamoto, Y. (1998). Small interspersed sequences that serve as recombination sites at the cox2 and atp6 loci in the mitochondrial genome of soybean are widely distributed in higher plants. Curr. Genet. 33, 188–198. Kato, S., Kanazawa, A., Mikami, T. & Shimamoto, Y. (1998). Evolutionary changes in the structures of the cox2 and atp6 loci in the mitochondrial genome of soybean involving recombination across small interspersed sequences. Curr. Genet. 34, 303–312. Klein, M., Eckert-Ossenkopp, U., Schmiedeberg, I., Brandt, P., Unseld, M., Brennicke, A. & Schuster, W. (1994). Physical mapping of the mitochondrial genome of Arabidopsis thaliana by cosmid and YAC clones. Plant J. 6, 447–455. Marienfeld, J. R., Unseld, M., Brandt, P. & Brennicke, A. (1997). Mosaic open reading frames in the Arabidopsis thaliana mitochondrial genome. Biol. Chem. 378, 859–862. Schon, E. A., Rizzuto, R., Moraes, C. T., Nakase, H., Zeviani, M. & DiMauro, S. (1989). A direct repeat is a hotspot for large-scale deletion of human mitochondrial DNA. Science, 244, 346–349. Sor, F. & Fukuhara, H. (1983). Unequal excision of complementary strands is involved in the generation of palindromic repetitions of rhoK mitochondrial DNA in yeast. Cell, 32, 391–396. Grabau, E. A., Davis, W. H., Phelps, N. D. & Gengenbach, B. G. (1992). Classification of soybean cultivars based on mitochondrial DNA restriction fragment length polymorphisms. Crop Sci. 32, 271–274. Moeykens, C. A., Mackenzie, S. A. & Shoemaker, R. C. (1995). Mitochondrial genome diversity in soybean: repeats and rearrangements. Plant Mol. Biol. 29, 245–254. Hanlon, R. W. & Grabau, E. A. (1997). Comparison of mitochondrial organization of four soybean cytoplasmic types by restriction mapping. Soybean Genet. Newsletter, 24, 208–210. Khazi, F. R., Edmondson, A. C. & Nielsen, B. L. (2003). An Arabidopsis homologue of bacterial RecA that complements an E. coli recA deletion is targeted to plant mitochondria. Mol. Gen. Genet. 269, 454–463. Cerutti, H., Osman, M., Grandoni, P. & Jagendorf, A. T. (1992). A homolog of Escherichia coli RecA protein in plastids of higher plants. Proc. Natl Acad. Sci. USA, 89, 8068–8072. Palmer, J. D., Shields, C. R., Cohen, D. B. & Orton, T. J. (1983). An unusual mitochondrial DNA plasmid in the genus Brassica. Nature, 301, 725–728. Preiser, P. R., Wilson, R. J. M., Moore, P. W., McCready, S., Hajibagheri, M. A. N., Blight, K. J. et al. (1996). Recombination associated with replication of malarial mitochondrial DNA. EMBO J. 15, 684–693. Han, Z. & Stachow, C. (1994). Analysis of Schizosaccharomyces pombe mitochondrial DNA replication by two dimensional gel electrophoresis. Chromosoma, 103, 162–170.

Soybean MtDNA Recombination

47. Gamper, H., Lehman, N., Piette, J. & Hearst, J. (1985). Purification of circular DNA using benzoylated naphthoylated DEAE-cellulose. Gene, 4, 157–164. 48. Cerutti, H. & Jagendorf, A. T. (1993). DNA strandtransfer activity in pea (Pisum sativum L.) chloroplasts. Plant Physiol. 102, 145–153. 49. Levene, S. & Zimm, B. (1987). Separation of opencircular DNA using pulsed-field electrophoresis. Proc. Natl Acad. Sci. USA, 84, 4054–4057. 50. Beverley, S. M. (1988). Characterization of the “unusual” mobility of large circular DNAs in pulsed fieldgradient electrophoresis. Nucl. Acids Res. 16, 925–939. 51. Maleszka, R. (1993). Single-stranded regions in yeast mitochondrial DNA revealed by pulsed-field gel electrophoresis. Appl. Theor. Electrophoresis, 3, 259–263. 52. Backert, S., Meisner, K. & Borner, T. (1997). Unique features of the mitochondrial rolling circle-plasmid mp1 from the higher plant Chenopodium album (L.). Nucl. Acids Res. 25, 582–589. 53. Levy, A. A., Andre, C. P. & Walbot, V. (1991). Analysis of a 120-kilobase mitochondrial chromosome in maize. Genetics, 128, 417–424. 54. Lockshon, D., Zweifel, S. G., Freeman-Cook, L. L., Lorimer, H. E., Brewer, B. J. & Fangman, W. L. (1995). A role for recombination junctions in the segregation of mitochondrial DNA in yeast. Cell, 81, 947–955. 55. Belanger, K. G., Mirzayan, C., Kreuzer, H. E., Alberts, B. M. & Kreuzer, K. N. (1996). Twodimensional gel analysis of rolling circle replication in the presence and absence of bacteriophage T4 primase. Nucl. Acids Res. 24, 2166–2175. 56. Sedat, J. W., Kelly, R. B. & Sinsheimer, R. L. (1967). Fractionation of nucleic acid on benzoylated– naphthoylated DEAE cellulose. J. Mol. Biol. 26, 537–540. 57. Dijkwel, P. A., Vaughn, J. P. & Hamlin, J. L. (1991). Mapping of replication initiation sites in mammalian genomes by two dimensional gel analysis: stabilization and enrichment of replication intermediates by isolation on the nuclear matrix. Mol. Cell. Biol. 11, 3850–3859. 58. Hay, R. & dePamphilis, M. (1982). Initiation of simian virus 40 DNA replication in vivo: location and structure of 5 0 -ends of DNA synthesized in the oriregion. Cell, 28, 767–779.

299 59. Lu, Z., Kunnimalaiyaan, M. & Nielsen, B. L. (1996). Characterization of replication origins flanking the 23 S rRNA gene in tobacco chloroplast DNA. Plant Mol. Biol. 32, 693–706. 60. Maleszka, R., Skelly, P. J. & Clark-Walker, G. D. (1991). Rolling circle replication of DNA in yeast mitochondria. EMBO J. 10, 3923–3929. 61. Backert, S., Dorfel, P., Lurz, R. & Borner, T. (1996). Rolling-circle replication of mitochondrial DNA in the higher plant Chenopodium album (L.). Mol. Cell. Biol. 16, 6285–6294. 62. Hrubec, T., Robinson, J. & Donaldson, R. (1985). Isolation of mitochondria from soybean leaves on discontinuous Percoll gradients. Plant Physiol. 77, 1010–1012. 63. Sambrook, J. & Russell, D. W. (2001). Molecular Cloning: A Laboratory Manual, 3rd edit., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 64. Steffen, S. E. & Bryant, F. R. (2001). Purification and characterization of the single-stranded DNA binding protein from Streptococcus pneumoniae. Arch. Biochem. Biophys. 388, 165–170. 65. Daniell, H., Zheng, D. & Nielsen, B. L. (1995). Isolation and characterization of an in vitro DNA replication system from maize mitochondria. Bioch. Biophys. Res. Commun. 208, 287–294. 66. Heller, C. & Poll, F. (1990). Field inversion gel electophoresis with different pulsed time ramps. Nucl. Acids Res. 18, 6299–6304. 67. Weiden, M., Osheim, Y., Beyer, A. & van der Ploeg, L. (1991). Chromosome structure: DNA nucleotide sequence elements of a subset of the minichromosomes of the protozoan Trypanosoma brucei. Mol. Cell. Biol. 11, 3823–3834. 68. Serwer, P. (1978). A technique for observing extended DNA in negatively stained specimens: observation of bacteriophage T7 capsid–DNA complexes. J. Ultrastruct. Res. 65, 112–118. 69. Brewer, B. J. & Fangman, W. L. (1991). Mapping replication origins in yeast chromosomes. BioEssays, 13, 317–322.

Edited by J. Karn (Received 8 August 2005; received in revised form 12 November 2005; accepted 22 November 2005) Available online 9 December 2005