characterization of replication origins flanking the 23s rrna gene in tobacco chloroplast dna

Plant Molecular Biology 32: 693-706, 1996. 1996 Kluwer Academic Publishers. Printed in Belgium.

693

Characterization of replication origins flanking the 23S rRNA gene in tobacco chloroplast DNA

Z h u n Lu , M u t h u s a m y K u n n i m a l a i y a a n and B r e n t L. Nie lsen* Department of Botany and Microbiology Auburn University, Auburn, AL 36849, USA (*author for correspondence)

Received 26 January 1996; accepted in revised form 17 July 1996

Key words: chloroplast DNA replication, D-loops, primer extension, tobacco

Abstract

Using 5' end-labeled nascent strands of tobacco chloroplast DNA (ctDNA) as a probe, replication displacement loop (D-loop) regions were identified. The strongest hybridization was observed with restriction fragments containing the rRNA genes from the inverted repeat region. Two-dimensional get analysis of various digests of tobacco ctDNA suggested that a replication origin is located near each end of the 7.1 kb BamHI fragment containing part of the rRNA operon. Analysis of in vitro replication products indicated that templates from either of the origin regions supported replication, while the vector alone or ctDNA clones from other regions of the genome did not support in vitro replication. Sequences from both sides of the BamHI site in the rRNA spacer region were required for optimal in vitro DNA replication activity. Primer extension was used for the first time to identify the start site of DNA synthesis for the D-loop in the rRNA spacer region. The major 5' end of the D-loop was localized to the base of a stem-loop structure which contains the rRNA spacer BamHI site. Primer extension products were insensitive to both alkali and RNase treatment, suggesting that RNA primers had already been removed from the 5' end of nascent DNA. Location of an origin in the rRNA spacer region of ctDNA from tobacco, pea and Oenothera suggests that ctDNA replication origins may be conserved in higher plants.

Introduction

Chloroplast DNA (ctDNA) in higher plants consists of a double-stranded covalently closed circular DNA molecule of between 120 and 160 kb [22], and in some species has been shown to exist in multimeric forms [ 11 ]. In general, overall ctDNA genome structure among higher plants is highly conserved in terms of the numbers and the DNA sequences of the genes encoded. Except for some legumes [35, 36], almost all plant chloroplast genomes contain two copies of a large inverted repeat (IR), which is 20 to 25 kb in size and contains the 16S, 23S and 5S rRNA genes as well as some tRNA and ribosomal protein genes, situated in opposite orientation and separated by a large single- copy (LSC) and a small single-copy (SSC) region. The complete chloroplast genome sequence is available for tobacco, and genes and open reading frames have been identified [41 ].

Although considerable work has been done on chloroplast gene identification and expression [ 13, 41, 43], much less is known about the process of plant ctDNA replication [19]. A ctDNA replication model was proposed by Kolodner and Tewari [23] based on electron microscopy (EM) of pea and maize ctDNA molecules. Replication starts with the formation of two displacement loops (D-loops), each similar to the single D-loop of animal mtDNA [6, 7, 9], about 7 kb away from each other. It has been previously reported that ! 5-30% of ctDNA molecules contain D-loops of three size classes of 820, 1560 and 2470 bases, with a smaller number of molecules continuing replication from the D-loop regions [44]. In ctDNA, the two D-loops have been shown by EM to expand unidirec- tionally towards each other, leaving the non-template strand single-stranded, until the two D-loops join to form a double-stranded Cairns (theta) type structure. Replication then proceeds bidirectionally until comple-

694

tion of new daughter molecules [23]. In pea ctDNA, one of the D-loops (oriA) has been mapped to the spacer region between the 16S and 23S rRNA genes by EM [31 ], and recently mapped by a combination of techniques to a 0.45 kb fragment from the spacer region between a SacI site near the 3' end of the 16S rRNA gene and an EcoRI site [32]. The second D-loop maps downstream of the 23S and 5S rRNA genes in pea ctDNA [31 ]. The presence of D-loop regions associated with the rRNA genes has also been observed in Oenothera, where EM analysis has identified four D-loops, two in each IR flanking the 16S rRNA genes [8]. Takeda et al. [42] identified three putative origins in cultured tobacco cell plastid DNA, one located 3' of the 23S rRNA gene near the end of each IR, and one in the LSC region. A single replication origin was mapped near the additional 16S rRNA gene in Euglena gracilis [21, 37]. However, ctDNA replication origins have been mapped at different locations in Chlamydo- monas reinhardii [26, 47, 48], soybean [ 18], petunia [10] and maize [3].

Various methods can be used for studying ctDNA replication. According to the EM-based model [23], ctDNA D-loops of discrete sizes accumulate as paused replication intermediates, prior to fusing to form bidirectional replication structures, and contain up to 2470 bp long single-stranded DNA regions [44]. These single-stranded D-loop regions should bind strongly to benzoylated, naphthoylated DEAE (BND)-cellulose, which binds single-stranded nucleic acids much more strongly than double-stranded nucleic acids. BND cellulose has been used to enrich for DNA molecules containing replication intermediates from various organisms [12, 17]. Expanding replication forks contain small transient single-stranded DNA regions, and will also bind to BND cellulose. Since the D-loops occur at specific locations within the genome [44], it should be possible to identify these regions when the 5' ends of the nascent DNA (within the D-loop) are labeled and used as probe against a restriction digest of total ctDNA. Two-dimensional agarose gel electrophoresis is used to separate restriction fragments containing replication intermediates from linear non-replicating fragments based on differences in mass and structure [2]. In the first dimension DNA molecules are separated according to mass at low voltage in low percentage agarose gels, while the second dimension separates molecules according to shape at high voltage in a higher percentage gel. Distinct and predictable patterns are obtained for each different type of replication intermediate (i.e. replication bubbles, single- or double-forked

molecules). This technique can be coupled with S1 nuclease treatment to digest and identify the single- stranded regions of D-loops and other replication intermediates, similar to studies in other organisms [16, 45]. In addition, if replication from each ctDNA origin is initially unidirectional from specific site(s) as proposed by the D-loop model [23], it should be possible to map the 5' ends of nascent DNA strands by primer extension.

In addition to the known tobacco ctDNA sequence and information about gene expression, tobacco leaves give a high yield of chloroplasts, making this a good system for studying ctDNA replication. We report here the use of 5' end-labeled nascent strands of tobacco ctDNA as a probe to identify the D-loop regions, and the characterization of replication origins by in vivo and in vitro studies. The 5' end of the oriA D-loop has been identified by primer extension.

Materials and methods

Tobacco ctDNA isolation

CtDNA was purified from sucrose gradient-isolated tobacco chloroplasts from fresh young (2 months after germination) green leaves, without DNase I treatment of nuclei, essentially as described by Palmer [34]. Large bore pipets, tips and needles were used for all subsequent transfers to minimize shearing. After cesium chloride-ethidium bromide density gradient cent- rifugation as previously described [23, 31 ], the single band containing relaxed circular tobacco ctDNA was carefully recovered, extracted with n-butanol, dialyzed with several changes of 10 mM Tris-HC1 pH 8.0, 2.0 mM EDTA and stored in small aliquots at - 8 0 °C until use for specific experiments.

Random primed labeling and hybridization of probes

Restriction fragments from tobacco clones were labeled with [a32P]-dCTP (New England Nuclear, 3000 Ci/mmol) by random priming (Prime-a-Gene kit, Promega) following the manufacturer's instructions. DNA fragments from agarose gels were retrieved by the syringe-filter squeeze method [27] for labeling. Alternatively, low melting point agarose (SeaPlaque, FMC Corp.) was used to recover DNA fragments for in-gel labeling according to the manufacturer's instructions. Southern blot hybridization was performed using Magna-nylon membrane (MSI) as described [39].

BND cellulose column enrichment of 5 ~ end-labeled nascent DNA

Tobacco ctDNA (100 #g in 100 #1) was treated with RNase A (10 #g, DNase free) and RNase H (4 U) at 37 °C and 56 °C for two cycles of 30 min each. DNA was ethanol-precipitated and resuspended in TE buffer (10 mM Tris-HC1 pH 8.0, 1 mM EDTA). The ctDNA was dephosphorylated using calf intestinal alkaline phosphatase (CIAP) [39]. Proteinase K was added to 50 #g/ml and the sample was incubated for 1 h at 37 °C, phenol extracted and ethanol precipitated. ['y32p]-ATP (150 #Ci, New England Nuclear) and 20 U of polynucleotide kinase (PNK) were added [39], and the sample was incubated at 37 °C for 1 h, with supplementation of PNK (20 U) after 30 min. The reaction mixture was extracted with phenol and the DNA precipitated with ethanol. The end-labeled ctDNA was then restricted with BamHI, dissolved in BND buffer one (10 mM Tris-HCl pH 8.0, 0.3 M NaC1, 1 mM EDTA) and loaded onto a 1 ml benzoylated-naphthoylated DEAE (BND) cellulose (Sigma) column [ 12, 17] equilibrated previously with the same buffer. The column was washed with 10 ml of BND buffers 1, 2 and 3 respectively (BND buffer 2 :10 mM Tris-HCl pH 8.0, 0.8 M NaC1, 1 mM EDTA; BND buffer 3:10 mM Tris-HC1 pH 8.0, 1.2 M NaC1, 1.8% caffeine). Twenty 0.5 ml fractions were collected from each wash, and fractions were analyzed for radioactivity (Cerenkov counting) and absorbance (measured at 260 nm). Radioactive peak fractions from the high-salt plus caffeine step were ethanol precipitated, resuspended in TE buffer, heat- denatured and utilized for hybridization.

Two-dimensional gel electrophoresis

Two-dimensional gel analysis was carried out as described [2], after digestion of total tobacco ctDNA with an excess of restriction enzyme in a short (2 h) incubation at 37 °C. For the 7.1 kb BamHI fragment, conditions were modified to enhance resolution of larger fragments, as described [20]. The first dimension was electrophoresed for 39.5 h at 0.5 V/cm in a 0.45% agarose gel. The second dimension was carried out in a 0.8% agarose gel at 2 V/cm for 22 h. After second dimension electrophoresis for each gel, DNA was transferred to nylon membranes for hybridization with appropriate probes. S 1 nuclease treatment was carried out in solution (2 U S 1 nuclease in 30/tl reaction) for 20 min at 37 °C after restriction digestion, prior to running the first dimension.

695

In vitro DNA replication

A partially purified tobacco ctDNA replication fraction was prepared as described [32]. In vitro replication reactions were carried out using CsC1 gradient- purified pUC19 and specific ctDNA clones as shown in Fig. 5. The amount and quality (> 80% supercoiled, no detectable RNA) of plasmid DNA was standard- ized by analysis in agarose gels prior to use. DNA that did not meet the criteria was not used. In vitro replication products prepared with unlabeled dNTPs [32] were analyzed by two-dimensional agarose gel electrophoresis and Southern hybridization.

Primer extension

The oligonucleotides 5J-ATACACTGATAAGGGAT GTA-3' (20 bases) and 5 ' -CTGGCGCAGCTGGGC CA-3 t (18 bases) were purchased from National Bios- ciences (Plymouth, Minnesota) and used for primer extension. Oligonucleotides were end-labeled with [-),32p]-ATP using T4 polynucleotide kinase [39]. Tobacco ctDNA (25-50 #g) and the labeled primer (20 ng) were mixed and incubated at 95 °C for 5 min. After incubation the tubes were transferred to 54 °C for 45 min to allow annealing, after which Taq DNA polymerase buffer (1 x final conc.), dNTPs (250 #M final conc.), MgCI2 (1.5 mM final conc.) and 2.5 units of Taq DNA polymerase were added and mixed thoroughly. The tubes were incubated at 50 °C for 45 min, and the reactions terminated by adding formamide loading dye and heating the samples to 95 °C for 5 min. A control reaction incubated identically but without tobacco ctDNA was carried out for comparison. Dideoxy DNA sequencing [40] was carried out with the Sequenase kit (U.S. Biochemical), using a 4.45 kbp EcoRI tobacco ctDNA subclone containing the rRNA spacer region (see Fig. 5) as template primed with the same oligonucleotide used for primer extension. Primer extension products and sequencing reactions were electrophoresed in a 6% polyacrylamide sequencing gel containing 6 M urea at 1500 V for 2 h. The samples were denatured by boiling for 5 min before loading the gel. Dried gels were exposed to X-ray film.

DNA sequence analysis

The Intelligenetics PC/GENE DNA analysis software was employed to analyze published tobacco ctDNA sequences (EMBL accession number Z00044 [41]). Direct repeats, hairpin structures, A+T- and G+C-rich

696

Figure 1. A. Hybridization of BND cellulose-enriched replication intermediates to total tobacco ctDNA restricted with various enzymes. Tobacco ctDNA was purified by the sucrose-cesium chloride procedure as described in the text, digested with specific enzymes as shown, and hybridized with a BND cellulose enriched probe. The sizes of hybridizing fragments indicated at the sides of the lanes are rounded to the nearest tenth. B. Map of restriction fragments from the Sal 4 region which hybridized with the BND cellulose-enriched probe. Sites for each restriction enzyme are shown for only the Sal 4 restriction fragment, and sizes are given for each fragment, including the fragments which overlap the ends of Sal 4 but extend to the next adjacent site for each enzyme. Specific regions hybridized by the BND cellulose enriched probe are indicated by the thick black lines. Relevant sequence positions according to Shinozaki et al. [41] are shown on the map. These sequences are also present in IR8 within the Sal 1 fragment. Transcription of the 16S and 23S rRNA genes is from right to left, and this orientation is the same in all subsequent line drawings.

regions and other c o m m o n features o f D N A replicat ion

origins were character ized for the oriA region identified by pr imer extension.

Results

Hybridization of BND-enriched DNA with tobacco ctDNA and clones

Nascent strands o f tobacco c t D N A were labeled at the

5' ends, fo l lowed by restr ict ion digest ion and enrich-

ment by B N D cel lu lose chromatography, and fract ions

were moni to red by absorbance and radioactivity. Mos t

B

124307 Sal I Small Slnale

I co ,, ron I odB

Sal 4 19,5 kb

Inverted Repa¢

or~

BamH I out 140,992 130,856 137,723 ,43 kb

17,5kb I 7,1kb I 3'27kb I t"1 5,2kb ....

9,g kb

Pvu II cut

144,374 Sail

I

1~t2~ 1371202 1391~2

,~c I cut

0,4kb 16.3kb I 3.gkb ll,.,2kb I 3,4kb I 9,1kb

17.4 kb

Slu I cut

0,gkb .12.7 kb I 5.9kb I I 11,2kb

Figure 1. Continued.

(about 98%) of the D N A , as measured by absorb-

ance, was r e m o v e d with the 0.3 M and 0.8 M NaC1

washes. In contrast, more than ha l f o f the radioact ivi ty

was recovered in the 1.2 M NaCI + caffeine elut ion

(data not shown). The latter fract ion was used as probe

for hybr idizat ion with various digests of total tobacco

c t D N A (Fig. 1A). Based on the publ ished tobacco

c t D N A sequence [41], when diges ted with BamHI, 26 f ragments o f at least 0.4 kb are generated, some

of which are dupl icated in the inverted repeat. Only a

few of these showed hybr idizat ion with the B N D cel-

lulose probe. The 7.1 kb and 3.27 kb BamHI f ragments

showed strong hybridizat ion. Two other f ragments (ca.

4 - 5 kbp bands) also showed lower levels o f hybridization, and based on the publ ished tobacco c t D N A

sequence there are only 5 BamHI f ragments in this size range. The larger o f these two bands l ikely repres-

ents the 4.8 or 5.0 kb BamHI f ragments located near the 7.1 and 3.27 kb f ragments over lapping the ends of the

IRs and the large s ingle-copy region, a l though a 5.2 kb

BamHI f ragment is also located in the IR (Fig. 1B). The

smaller hybridizing fragment probably represents one of two BamHI fragments of 4.47 and 4.54 kb located at the opposite side of the circular chloroplast genome at the putative replication termination site [23]. In this figure, fragments less than 1.5 kb have run off the gel. With PvulI digestion, 4 kb and 2.4 kb fragments were hybridized, while for SacI restriction, 9.1, 3.4, 2 kb, and some higher-molecular-weight partial digest fragments were hybridized. With StuI digestion, the 11.2 kb fragment showed hybridization. Nearly all of the strongly hybridized fragments are localized in the same region of the Sal 4 fragment (Fig. 1B). Other restriction fragments of tobacco ctDNA showed lower levels of hybridization only upon longer exposure. It is expected that because replication proceeds around the entire genome and each replication fork contains transient single-stranded regions, the probe should show some hybridization to other fragments.

Hybridization of 5' end-labeled BND cellulose enriched fragments with tobacco ctDNA SalI clones showed similar results, with only restriction fragments from the Sal 4 region showing strong hybridization (not shown). Further hybridization analysis with specific tobacco ctDNA subclones showed that the probe hybridizes not only with the rRNA gene sequences, but also with the rRNA spacer region (Fig. 2A, lane 1) and the region downstream of the rRNA genes (Fig. 2A, lanes 2 and 3). This suggests that there may be more than one D-loop region, with one on each side of the 23S rRNA gene. The largest fragment in lane 1 retains 121 bp of ctDNA sequence along with the vector, and shows hybridization with the probe, while the size difference is not sufficient to cause a detectable difference in mobility as compared to the vector alone in lanes 2 and 3. The faint hybridized band in lane 2 which is not visible in the gel photograph is an artifact, possibly due to EcoRI* activity. This hybridization data indicates that the probe does not simply represent end-labeled rRNA, tRNA or mRNA molecules. This is further supported by the observation that hybridization remains after treatment with RNase (Fig. 1 and 2) under conditions where RNA is completely degraded in control reactions (not shown). Similar results were obtained with 5' end-labeled pea ctDNA [32] and tobacco ctDNA which had not been subjected to BND enrichment, although greater background hybridization was observed and longer exposure times were required (data not shown). This suggests that BND cellulose enrichment removes the majority of unlabeled double-stranded DNA molecules which may compete

697

with the less abundant 5' end-labeled population for hybridization.

Two-dimensional gel electrophoresis

Using the results from BND cellulose probe hybridization, specific fragments from the rRNA operon region were used as probes for two-dimensional gel analysis. The CsCI gradient protocol used to isolate the ctDNA has yielded intact circular pea and maize ctDNA molecules as visualized by electron microscopy [23, 31 ], and avoids the use of phenol extraction and ethanol precipitation. Figure 3H shows a schematic representation of common patterns observed in two-dimension gels [2, 16]. The drawing at the bottom of the figure shows the probes used. Figure 3A (without S 1 treatment) and B (with S 1 treatment) show two-dimensional gels of total tobacco ctDNA restricted with BamHI and probed with the 7.1 kb BamHI fragment, under conditions for resolution of large fragments [20, 25]. The 7.1 kb BamHI fragment generated a simple Y pattern [2, 28, 29] with an extended arc (E-arc) past the dimer size [ 16], and an additional spike representing X-shaped molecules [2, 29]. E-arcs are larger than dimer size because of the inability of restriction enzymes to cleave single-stranded regions as found in D-loops, resulting in a longer molecule consisting of two double-stranded segments joined by a single- stranded segment (Fig. 4E, see also Discussion). The simple Y pattern was more easily observed when the 2-D gel was electrophoresed under conditions for resolution of large fragments, although the E-arc became diffused and was difficult to visualize after the extended period of electrophoresis (Fig. 3A). The faint E- arc became undetectable after SI treatment (Fig. 3B). The faint arc in Fig. 3A and 3B (large filled arrows) superimposing the simple Y pattern and extending to the recombination spike may represent low-abundance double-stranded bubble or double Y forms, as this arc is resistant to S 1 nuclease. Spots on the simple Y arcs (Fig. 3A and B, open arrows) may indicate accumu- lated replication intermediates generated as a result of pausing at specific locations during replication. The simple Y forms observed for all probes used in Fig. 3 may result from breakage of bubbles located near the ends of linear fragments (see Fig. 4B and 4C), as has been observed in other systems [25], or from lagging strand synthesis at each D-loop to generate double- stranded DNA which can be digested by the restriction enzyme (Fig. 4F). The near-vertical spikes originating from the dimer spot for most of the probes in Fig. 3 are

698

Figure 2. A. Hybridization of BND cellulose-enriched replication intermediates to the rRNA spacer and downstream regions. Left, ethidium bromide stained gel; right, Southern blot. Exposure times were adjusted as indicated to obtain optimal visualization. Lane M, HindlII-digested bacteriophage lambda DNA as molecular weight markers (3 d exposure for blot). Lane 1, BamHI/HindHI digest of the 3.73 kb EcoRI/BamHI clone (3 d exposure). Lane 2, PstI/EcoRI digest of the 1.42 kb SspI/EcoRI clone (12 d exposure). Lane 3, EcoRI digest of the 1.37 kb EcoRI clone (3 d exposure). B. Restric- tion map of the rRNA gene region and clones used for hybridization, including flanking and spacer regions. Sites for restriction enzymes used in panel A are shown (H =- HindlII). All genes shown except ORF75 and trnN are transcribed from right to left according to the map. The trnN gene is within the ORF75 reading frame, which are both transcribed from left to right. For reference, the border of the inverted repeat and the small single copy region is between positions 130,505 and 130,506. The tRNA genes in the rRNA spacer region shown at the right of the 23S rRNA gene both contain large introns.

likely due to recombination intermediates [2]. Recom- bination in chloroplasts is known to be involved in copy correction of the IRs [36], and recombination activity has been characterized in pea [4, 5]. EcoRI- digested ctDNA probed with the 4.45 kb EcoRI fragment generated a simple Y pattern and a strong E-arc of greater than dimer size (Fig. 3C). This is predicted from expansion of the D-loop past one EcoRI site, with one Strand remaining single-stranded and thus insensitive to EcoRI digestion, while the nascent strand and the two adjacent EcoRI sites are double-stranded and cleaved, but held together by the single-stranded region (see Fig. 4E). As predicted, upon S1 nuclease treatment the E-arc was removed (Fig. 3D), and was not observed even after longer exposure (not shown). The simple Y pattern also showed some decrease in signal upon S 1 treatment, while the monomer spot containing double-stranded DNA remained at similar intensity. PvulI-digested ctDNA probed with the 7.1 kb BamHI fragment (Fig. 3E, see also Fig. 4F) showed a very strong recombination spike and typical simple Y as well as faint double Y patterns originating from the 4.0 kbp PvulI fragment. In contrast, the adjacent 2.4 kb PvulI fragment showed only simple Y and recombination patterns (Fig. 3F). It is not clear why the spike in Fig. 3E is so strong, but it may be due to the similarity of some nearly completed double-Y replication structures to recombination structures, which would both have an X-shape. Analysis of the 4.29 kb SspI fragment containing the putative oriB region downstream of the rRNA genes in the IR showed the presence of an E-arc pattern of greater than dimer size [16], a recombination spike, an additional higher-molecular-weight spike of unknown origin (per- haps a more complex recombination intermediate), and simple Y-type patterns (Fig. 3G). The E-arc observed with SspI digestion is analogous to the EcoRI E-arc, with only the double-stranded SspI sites away from the D-loop being digested, and the single-stranded D 7 loop region holding together the two double-stranded regions (Fig. 4H), producing a molecule of consider- ably greater than dimer size (small arrow in Fig. 3G). Of these patterns, only the E-arc was sensitive to S 1 nuclease (Kunnimalaiyaan et al., manuscript in preparation), as predicted by the model (Fig. 4H). The schematic representation of tobacco ctDNA replication presented in Fig. 4 summarizes the double D-loop c tDNA replication model based on EM [8, 23, 31 ] and the results of two-dimensional gel analysis of in vivo and in vitro (see below) replication intermediates in tobacco ctDNA (see discussion). D-loops have been

shown to exist independently of each other, with some ctDNA molecules containing only a single D-loop rather than two as illustrated in Fig. 4B. Expansion of individual D-loops is described in Fig. 4D and E (or/A region) and Fig. 4G and H (oriB region), although both may expand simultaneously as shown in Fig. 4F and bidirectional replication occurs after fusion of the two D-loops as illustrated in Fig. 41.

In vitro DNA replication

Several cloned tobacco ctDNA fragments from the rRNA region were tested as templates for in vitro DNA replication using a partially purified tobacco chloroplast enzyme preparation. The pUC 19 vector alone and a 1.37 kb EcoRI clone from the region between the two putative origins were analyzed as controls. However, no replication intermediate patterns were observed for either of these two templates (Fig. 5, panels A and F, respectively). Templates incubated in the absence of the tobacco chloroplast enzyme fraction showed no replication intermediate patterns (data not shown, see [32], indicating that the observed forms are due to enzyme activity in the protein preparation. Simple Y form patterns were observed with EcoRI- restricted 4.45 kb EcoRI clone containing the putative oriA region (Fig. 5B, see discussion). The 3.27 kb BamHI and 3.73 kb BamHI-EcoRI clones which overlap with this EcoRI fragment, but not with each other (see drawing at the bottom of Fig. 5), showed barely or no detectable replication intermediate patterns (Fig. 5C and D, respectively). Similar results were observed for pea ctDNA replication [32]. In vitro replication with the 4.29 kb SspI fragment from the oriB region also generated simple Y patterns (Fig. 5E). Clones from several other regions of the tobacco chloroplast genome showed no replication intermediates from in vitro replication reactions (not shown). Thus, among those templates tested, only those which contain putative ori regions generated replication intermediate patterns by two-dimensional gel analysis of in vitro replication products.

Primer extension

Based on the results from two-dimensional gel electrophoresis, the region between the PvulI site at position 137,202 of IRA (105,326 of IRB) and the 3 t end of the 16S rRNA gene (see Fig. 4A) was targeted for primer extension mapping of the start site of DNA synthesis. Initially, an 18 b primer containing the PvulI site

699

at position 137,202 within the IRA region (105,326 of IRB) was used. The extension products were ca. 500 bp, and were too large to characterize further. A 20 b primer located about 400 bp closer to the 16S rRNA gene, near the BamHI site, resulted in products which could be analyzed (Fig. 6, lanes 1-3). A control reaction was incubated with the labeled primer but no ctDNA, and produced no specific fragments of a size larger than the primer, indicating that the primer extension products were not a result of self-priming (lane 4, the labeled primer is seen at the bottom of the gel for each lane). In order to determine the exact site where DNA synthesis begins in this region, primer extension products were analyzed along with the sequence reaction products of the cloned 4.45 kb EcoRI fragment (see Fig. 5) using the same 20 bp primer. From the autoradiograph, a single major primer extension product was observed (Fig. 6 lanes 1-3), which cor- responds to the 'T ' at position 137,692 (104,838 of IR~), at the base of a potential stem-loop structure (Fig. 7). This band is not due to the inability of the Taq polymerase to read through the strong secondary structure, as the DNA sequencing reactions carried out with the same enzyme does not show any evid- ence of blockage (Fig. 6, sequencing lanes). Upon longer exposure, a few fainter primer extension bands could be observed, but these bands were 50-100 times less intense than the major primer extension product, suggesting that these bands represent minor extension products. To show that the primer extension product is not due to RNA contamination of the ctDNA preparation, tobacco ctDNA was treated with alkali (0.2 M NaOH final conc) at 65 °C for 30 min (lane 2), or the DNA was treated with RNase-H and RNase-A at 37 °C for 30 min (lane 3) prior to primer extension. All three experiments (no treatment, alkali or RNase treatment) resulted in identical products (lanes 1-3), indicating that the extended product represents DNA without any remaining RNA primer at the 5 t end.

DNA sequence analysis

The entire tobacco ctDNA sequence has been published [41] and is available from the EMBL data- base (accession number Z00044). PC/GENE software (Intelligenetics) was utilized to search for characterist- ic features of DNA replication origins such as direct repeats, stem-loop structures, and G + C and A + T rich regions in the oriA region identified by primer extension. The major primer extension product maps at the base of a relatively strong stem-loop structure (free

700

Figure 3. Two-dimensional gel electrophoresis of tobacco ctDNA. Total tobacco ctDNA was digested with specific enzymes and subjected to two-dimensional gel electrophoresis, blotted to nylon membranes and hybridized with specific probes shown on the next page, with panel showing hybridization for each probe indicated at the left. In each panel, M indicates the migration of linear monomer molecules, and D indicates the migration of dimer-sized molecules. Sizes were determined by comparison with lambda DNA digested with HindlII in the first dimension gel, which was run from left to right. The second dimension was from top to bottom. Small filled arrowheads indicate extended arc (E-arc) intermediates. Panels A ( -S1 nuclease treatment) and B (+S1 nuclease treatment) show a BamHI digest, electrophoresed using conditions to enhance resolution of large restriction fragments as described in the text, probed with the 7.1 kbp BamHI fragment. Large filled arrows indicate the faint arc patterns which superimpose the simple Y pattern. Large open arrows indicate putative paused replication intermediates. The rectangular signal above the 'D' in panel B is an exposure artifact. Panels C ( -S1 nuclease treatment) and D (+S1 treatment nuclease), an EcoRI digest probed with the 4.45 kbp EcoRI fragment. Panel E, a PvulI digest hybridized with the 7.1 kbp BamHI fragment. Y indicates simple Y form molecules, while DY indicates double Y forms. Panel F, a PvulI digest probed with the 2.4 kbp PvulI fragment. Panel G, a SspI digest probed with the 4.29 kbp SspI fragment. Panel H shows a schematic of two-dimensional gel patterns based on Brewer et al. [2] and Han and Stachow [16].

e n e r g y o f ca. - 8 5 k J /mo l ) w h i c h is c en t e r ed at posi -

t ion 137,711 o f IRA (104 ,818 o f IRB) in a G + C r ich

r eg ion f lanked by A + T - r i c h r eg ions o f 3 0 - 4 0 b p each

(Fig. 7). Th i s s t ruc ture a lso con ta ins the B a m H I site

in the r eg ion requ i red for in v i t ro rep l i ca t ion act iv-

ity (Fig. 5). W h i l e o the r s econda ry s t ruc tures can be

f o r m e d in the r R N A space r reg ion , this is the on ly

s t ruc ture w h i c h m a p s in the c lose v ic in i ty o f the m a j o r

p r i m e r ex t ens ion product . Two cop ies o f an 8 bp d i rec t

r epea t are loca ted ad j acen t to the s t em- loop s t ructure.

Discussion

B N D ce l lu lose was used to en r i ch for t o b acco c t D N A

f r a g m e n t s c o n t a i n i n g rep l i ca t ion D- loops , w h i c h were

used as p r o b e s o f S o u t h e r n b lo t s to iden t i fy pu ta t ive

701

Panel BemHI

A,B I 130,~6

C,D

E

F

G I °aB 130,50£

23S rRNA B,,mHI

I I I 7.1 kb 137,723

EcoRI EooRI I or~ I 4.~kb

133,1M14 I~,447 P~II Pwll

I I 4,okbp Pvull 1~ rRNA I:~11

I r--124kb Sspl

I 4,29 kb I~,7u


replication regions. While ctDNA fragments containing D-loops or transient single-stranded regions found at expanding replication forks will bind tightly to this matrix, RNA is single-stranded and also able to bind to BND cellulose [17]. The BND cellulose enriched probe was insensitive to RNase, indicating that this probe is not RNA. The 7.1 kb B a m H I and neighboring fragments showing the highest levels of hybridization are present in two copies in the genome because they are in the IR, but the level of hybridization observed is much greater than could be expected from random labeling. The specificity of the hybridization and the strength of the signal suggest that while the amount of DNA in the BND cellulose eluted fraction is small, this fraction contains a relatively stable DNA structure, in keeping with what is known about ctDNA and mtDNA D-loops [9, 44]. Other bands appear upon longer exposure, representing other parts of the genome as expected from expanding replication forks. These results are similar to those obtained with a 5' end-labeling approach which was previously used to localize D-loops in the rRNA gene region of pea ctDNA [32] and tobacco [42]. However, Takeda et al. enriched nucleic acids by BND cellulose from cultured tobacco cell plastids and labeled the DNA by nick translation, resulting in a difference between the signal and background of only 10-20%. In the tobacco cell report, a single replication origin was mapped in each copy of the IR in tobacco ctDNA, just downstream of the 23S rRNA gene and about 20 kbp apart [42]. The presence of a second origin in each inverted repeat as reported here but not by Takeda et al. [42] may be due to the low specificity of probes labeled by nick translation of replication intermediates after BND cellulose purification and inherent difficulties in accurately measuring DNA

7,1 kb 3,27 kl:

A - I E s P E I IIII I I I

orlB 235 rRNA oH 16S rP~

B (Rg, 3 panel= A,B)

(panels A,B}

D ~ (panels G,D) E E E

E I I t0,n,,,e,D)

P P

(panels E,F)

G ~ (panel G)

S S (pmel G) S

Figure 4. Schematic representation of two replication origins near the rRNA genes in each inverted repeat (A) as localized by two- dimensional gel electrophoresis. The gel panels from Fig. 3 which correspond to the different structures in the schematic drawing are indicated at the right of each structure. The initial stage of replication from each ori is unidirectional (B), and molecules may contain either single or double D-loops of discrete sizes (either one or both of the D-loops in B may be present in an individual ctDNA molecule). For fragments with D-loops near the ends, breakage may occur to generate a simple Y type pattern at each end but with one side of the Y being single-stranded (C). An EcoRl fragment containing a single oriA D-loop is shown in D. Once this D-loop has expanded, the single-stranded side of the D-loop region is resistant to EcoRI digestion and sensitive to S1 nuclease, as shown in E. Also in E, an adjacent EcoRl site is indicated which could be cleaved to generate the higher molecular weight E-arc pattern in Fig. 3C. The single- stranded region would be removed by S 1 nuclease, as indicated in E, releasing 2 linear double-stranded molecules which would migrate with the arc of linears in Fig. 3D. Digestion with PvulI should generate double Y patterns as seen in Fig. 3E if some lagging strand synthesis occurs in the D-loop regions (F). A situation similar to that observed with the EcoRI digest of oriA is seen with the SspI digest for the oriB region. An SspI fragment containing a single D-loop is shown in G, and the expanded D-loop past one SspI site is shown in H. The next adjacent SspI site in the direction of oriB expansion is 6.5 kbp away (H), creating a longer E-arc fragment in the two-dimension gel (Fig. 3G). Some lagging strand synthesis may occur in the D-loops (F), and is necessary for primer synthesis for bidirectional replication after fusion. Alternatively, the displaced strands may remain single-stranded until the two D-loops fuse to form a single Cairns intermediate (I) which will then continue to replicate bidirectionally. Restriction sites: B, BamHl; E, EcoRI; P, PvulI; S, SspI.

702

Panel EcoRI EcoRI B I orlA I 4,4Skb

EcoRI 235 rRNA BsmHI

C [ I I 3.73kb

BamHI 1~ rRNA D EcoRI EcoRI S.irkbp I 1 - - 1 F I J 1.37kb tar,ra

131,124 l:kl,2U Pstl Sapl Kpnl ~lpl

E II °~B II 4,~kb l~r~ 1~,71~


I 140,992

Figure 5. Two-dimensional gel electrophoresis of in vitro DNA replication products. In vitro DNA replication reactions were carried out as described in the text, and reaction products were analyzed by two-dimensional agarose gel electrophoresis (see Fig. 3 legend). The cloned fragments used as templates are shown at top right, with the panel indicated for each when used as probe indicated at the left. Reaction products were visualized by hybridization with labeled insert (or pUC 19 for panel A). Monomer (M) and dimer (D) fragments are indicated in each panel. Panel A, pUC19 vector alone, restricted with PstI. Panel B, 4.45 kbp EcoRI template restricted with EcoRI. Panel C, 3.27 kbp BamHI template restricted with BarnHI. Panel D, 3.73 kbp BamHI-EcoRI template restricted with BamHI and EcoRI. Panel E, 4.29 kbp SspI template restricted with Pstl and KpnI (removes all but 216 bp of the insert, sites are shown in the bottom line of the drawing; necessary because the SspI fragment was cloned into the Sinai site of pUC19, disrupting both the SspI and SmaI sites). Panel F, 1.37 kbp EcoRI template restricted with EcoRI.

molecules in electron micrographs, or to differences in the preparation of the plastid DNA.

Two-dimensional agarose gel electrophoresis is a technique which has been used to map and charac-

Figure 6. Primer extension mapping of the 5' end(s) of nascent DNA strands at oriA. Primer extension was carried out as described in Materials and methods with total tobacco ctDNA, using 5 ~ end- labeled primers. Lane 1, untreated ctDNA sample; lane 2, ctDNA treated with alkali prior to primer extension; lane 3, ctDNA treated with RNase A and RNase H prior to primer extension; lane 4, labeled primer control (no tobacco ctDNA). An empty lane was left between lanes 3 and 4. The filled arrowhead indicates the major primer extension product. The arrowhead labeled P at the bottom shows the location of the labeled primer. The circular spot between lanes 1 and 2 is an exposure artifact. For mapping, the same primer was used to carry out sequencing reactions with the 4.45 kbp EcoRI clone (see Fig. 5) as template. DNA sequencing lanes are labeled TCGA.

terize bidirectional DNA replication origins in animal and yeast cells [1, 2, 12, 25, 46], and unidirectional replication in bacterial plasmids [28, 29] and sea urchin mtDNA [30]. However, this technique has only recently been applied to the analysis of ctDNA replication [ 18, 32]. Some of the patterns observed with ctDNA are unique, as expected for a molecule which replicates by a D-loop mechanism. Han and Stachow [ 16] recently proposed a theoretical model for molecules which replicate by a single or double D-loop mechanism. According to the model, in the absence of lagging strand synthesis of the displaced strand, an atypical 2-D gel pattern will be generated with an intermediate larger than dimer size. This is due to the inability of most restriction enzymes to cut single stranded DNA (see Fig. 4E, H). According to the Kolodner and Tewari model [23] of ctDNA replication based on EM analysis, individual ctDNA molecules may contain only one or both of the D-loops, and expansion of each may occur independently, as outlined in Fig. 4. In those molecules in which both D-loops expand, once the two D-loops have joined replication proceeds bidirectionally with both leading and lagging strand synthesis [23] (Fig. 4I), which would generate only simple Y intermediates in adjacent fragments in the genome. Simple Y patterns with extended arcs (E-arcs) were observed for the 4.45 kb EcoRI fragment which contains the rRNA spacer region (Fig. 3C) and the 4.29 kb SspI fragment which contains the 23S rRNA gene and the oriB region downstream (Fig. 3G). The EcoRI fragment E-arc was removed with S1 nuclease digestion (Fig. 3D), indicating that this part of the structure contains at least some single-stranded DNA, as would be present in the displaced strand of a D-loop (Fig. 4E). Some lagging strand synthesis may occur in the single-stranded D- loop intermediates prior to fusing, which could result in the presence of the faint population of double Y patterns as observed with PvulI digestion (Fig. 3E and 4F). In addition, lagging strand synthesis could also generate simple Y forms typical of expanding replication forks for each of the fragments tested by creating double-stranded DNA capable of being restricted at the appropriate sites in these regions. DNA primase has been shown to exist in chloroplasts [33], and is capable of generating primers for lagging strand synthesis, which is known to occur after fusion of the D-loops. Another means of generating simple Y patterns may be by breakage of replication intermediates containing bubbles near the ends of the restriction fragments (see Fig. 4C), as has been proposed [25]. Adjacent fragments showed only simple Y forms without E-arcs

703

(Fig. 3F), compatible with the presence of expanding replication forks through these fragments. This suggests that the E-arcs are unique to the EcoRI and SspI fragments containing the putative replication origins. A population of paused intermediates was observed on some of the arcs, consistent with previous EM observation of paused D-loops of a uniform size in pea ctDNA [23, 44]. A significant population of paused D-loops has also been found in animal and sea urchin mtDNA, and these structures may represent parental molecules primed for replication [9, 30].

Analysis of in vitro replication products by two- dimensional agarose gel electrophoresis indicated that clones from the putative oriA and oriB regions were capable of supporting replication, while templates from other regions did not. The 4.45 kb EcoRI clone which showed highest in vitro replication activity generated only simple Y replication forms. It has recently been reported that pea ctDNA clones which contain a single replication origin replicate in vitro by a rolling circle mechanism with simultaneous replication of the tail by lagging strand synthesis, while a clone containing both origins replicates by a mechanism similar to the in vivo replication model, as determined by EM analysis [38]. Thus it may be expected that for in vitro replication, templates containing a single ctDNA origin may generate only simple Y form intermediates by two-dimensional gel electrophoresis [ ! 6]. The in vitro replication studies indicate that sequences on both sides of the BamHI site are necessary for optimal replication activity, while clones ending at this site retain only minimal or no replication activity (Fig. 5C, D). In pea, sequences on both sides of the BamHI site in the rRNA spacer region were also shown to be required for in vitro DNA replication activity, with the presence of only an additional 47 bp from the BamHI site towards the 23S rRNA gene in the pCSE0.45 clone being required to obtain a high level of replication activity [32]. This BamHI site is common to both pea and tobacco ctDNA, and lies within the stem-loop structure shown in Fig. 7.

From these results we propose that there is a replication origin at each end of the 7.1 kbp BamHI fragment within the IR of tobacco ctDNA. The locations of two origins flanking the 23S rRNA gene appears to be the same as locations determined for the pea ctDNA replication origins [31 ].

The primer extension experiment provides the first determination of the 5 ~ end of a D-loop associated with ctDNA replication. This experiment identified a single major 5 ~ end mapping to the base of a relatively

704

G G G G G CG CG TA GC GC

C A GC GC

A GC CG CG

Free Energy: -20.5 kcal/mol

AT 137,770 GC 137,692 137,650 5' GC ~ 3' CGGGTGAGATCCAATGTAGATCCAACTTTCGATTCACTCGTG TCTCTTCTCGAGAATCCATACATCCCTTATCAGTGTATGGACA

direct repeats 3' oligo primer 5'

Tobacco ctDNA oriA region

Figure 7. DNA sequence and structure of the oriA synthesized strand. The tobacco ctDNA sequence from positions 137,650 to 137,770 of IRA (104,779 to 104,879 of IRB ) as designated by Shinozaki et al. [41 ] is shown. A strong stem-loop structure containing the BamHI site is shown, with the location of the 5 r end of the major primer extension product indicated with a filled arrow. The location of the oligonucleotide used for the primer extension data in Fig. 6 is shown by underlining, as are the direct repeats 5 r of the stem-loop structure. The large open arrow indicates the direction of replication from oriA. For reference, the Y end of the 16S rRNA gene is at 138 283 in IRA (104,246 in IRB), and the 5' end of the 23S rRNA gene is at 136,204 in IRA (106,137 in IRB).

stable stem-loop structure for the oriA region. The functional role of these sequences in ctDNA replication has not yet been determined. However, secondary structures have been implicated in RNA primer transition to DNA synthesis at replication origins in animal mtDNA [9] and bacteriophage DNA [24]. Some fainter bands were observed upon longer exposure of the primer extension gels, and may represent minor extension products. Multiple start sites for animal mtDNA replication have been reported [9, 14, 15], and may be a result of imprecise transition from RNA to DNA synthesis at the origin. The primer extension products were identical after pretreatment of tobacco ctDNA with alkali or RNase, indicating that the products represent DNA and that any RNA primer for DNA synthesis had been removed prior to or during isolation of ctDNA. The DNA sequence of the tobacco ctDNA oriA region is similar to the oriA region of pea ctDNA [32]. This suggests that the location of this origin is the same in tobacco, Oenothera and pea ctDNA, and may represent a conserved location for this origin in other plant species as well. The origin downstream of the 23S rRNA gene (oriB) is in the same general location as the one in the IR identified by Takeda et al.

[42]. The oriB region lies within a longer stretch of DNA (ca. 2570 bp) which contains several dispersed stem-loop structures, and is currently under study.

Acknowledgments

This work was supported by grants from the USDA NRICGP (91-37301-6366 and 95-37301-2079), and by the Alabama Agricultural Experiment Station. We are thankful to Pal Maliga and Jeff Staub, Rutgers Uni- versity, for providing the tobacco ctDNA SalI library, and for helpful discussions. The authors are also grate- ful to Henry Daniell for helpful discussions. Z.L. and M.K. contributed equally to this work.

References

1. Brewer B J, Fangman WL: A replication fork barrier at the 3 r end of yeast ribosomal RNA genes. Cell 55:637-643 (1988).

2. Brewer B J, Sena EP, Fangman WL: Analysis of replication intermediates by two-dimensional agarose gel electrophoresis. Cancer Cells 6:229-234 (1988).

3. Carritlo N, Bogorad L: Chloroplast DNA replication in vitro: site-specific initiation from prefen'ed templates. Nucl Acids Res 16:5603-5620 (1988).

4. Cerutti H, Osman M, Grandoni P, Jagendorf AT: A homolog of Escherichia coli RecA protein in plastids of higher plants. Proc Natl Acad Sci USA 89:8068-8072 (t 992).

5. Cerutti H, Jagendorf AT: DNA strand-transfer activity in pea (Pisum sativum L.) chloroplasts. Plant Physiot 102:145-153 (1993).

6. Chang DD, Clayton DA: Priming of human mitochondrial DNA replication occurs at the light-strand promoter. Proc Natl Acad Sci USA 82:351-355 (1985).

7. Chang DD, Hauswirth WW, Clayton DA: Replication priming and transcription initiate from precisely the same site in mouse mitochondrial DNA. EMBO J 4:1559-1567 (1985).

8. Chiu W-L, Sears BB: Electron microscopic localization of replication origins in Oenothera chloroplast DNA. Mot Gen Genet 232:33-39 (1992).

9. Clayton DA: Replication and transcription of vertebrate mitochondrial DNA. Anna Rev Cell Biol 7:453-478 (1991).

10. de Haas JM, Kool A J, Overbeeke N, van Brug W, Nijkamp H J J: Characterization of DNA synthesis and chloroplast DNA replication initiation in a Petunia hybrida chloroplast lysate system. Curr Genet 12:377-386 (1987).

11. Deng X-W, Wing RA, Gruissem W: The chloroplast genome exists in multimeric forms. Proc Natl Acad Sci USA 86: 4156- 4160 (1989).

12, Dijkwel PA, Vaughn JP, Hamtin JL: 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 (1991).

13. Diogenes IH, Weissbach A: Organellar DNA replication in Nicotiana tabacum cultured cells. Plant Mol Biol 14:891-897 (1990).

t4. Gillum AM, Clayton DA: Displacement loop replication initiation sequence in animal mitochondrial DNA exists as a fam- ily of discrete lengths. Proc Natl Acad Sci USA 75:677-681 (1978).

15. Gillum AM, Clayton DA: Mechanism of mitochondrial DNA replication in mouse L-cells: RNA priming during the initiation of heavy-strand synthesis. J Mot Biol 135:353-368 (1979).

16. Han Z, Stachow C: Anlaysis of Schizosaccharomyces pombe mitochondrial DNA replication by two dimensional gel electrophoresis. Chromosoma 103:162-170 (1994).

17. Hay RT, DePamphilis ML: Initiation of SV40 replication hi vivo: location and structure of 5 ~ ends of DNA synthesized in the ori region. Cell 28:767-779 (1982).

18, Hedrick LA, Heinhorst S, White MA, Cannon GC: Analysis of soybean chloroplast DNA replication by two-dimensional gel electrophoresis. Plant Mol Biol 23:779-792 (1993).

19. Heinhorst S, Cannon GC: DNA replication in chloroplasts. J Cell Sci 104:1-9 (1993).

20~ Hyrien O, Mechali M: Plasmid replication in Xenopus eggs and egg extracts: a 2D gel electrophoretic analysis. Nucl Acids Res 20:1463-1469 (1992).

21. Koller B, Delius H: Origin of replication in chloroplast DNA of Euglena gracilis located close to the region of variable size. EMBO J 1:995-998 (t982)~

22. Kolodner RD, Tewarl KK: The molecular size and conform- ation of the chloroplast DNA from higher plants. Biochim B iophys Acta 402:372-390 (1975).

705

23. Kolodner RD, Tewari KK: Chloroplast DNA from higher plants replicates by both the Cairns and rolling circle mechanism. Nature 256:708-711 (1975).

24. Kornberg A, Baker TA: DNA Replication, 2nd ed., pp. 523- 526, 533-548, 615-617, 710-714. W.H. Freeman & Co, New York ( 1992).

25. Linskens MHK, Huberman JA: Ambiguities in results obtained with 2D gel replicon mapping techniques. Nucl Acids Res. 18: 647-652 (1990).

26. Lou JK, Wu M, Chang CH, Cuticchia AJ: Localization of a r-protein gene within the chloroplast DNA replication origin of Chlamydomonas. Curr Genet 11 : 537-541 (1987).

27. Lu Z, Templer M, Nielsen BL: Rapid method for recovery of DNA from agarose gets. Biotechniques 1 6 : ~ 0 2 (1994).

28. Martin-Parras L, Hernandez P, Martinez-Robles ML, Schvartz- man JB: Unidirectional replication as visualized by two- dimensional agarose gel electrophoresis. J Mol Bio1220: 843- 853 (1991).

29. Martin-Parras L, Hernandez P, Martinez-Robles ML, Schvartz- man JB: Initiation of DNA replication in ColE 1 plasmids containing multiple potential origins of replication. J Biol Chem 267:22496-22505 (1992).

30. Mayhook AG, Rinaldi AM, Jacobs HT: Replication origins and pause sites in sea urchin mitochondrial DNA. Proc R Soc Loud B 248:85-94 (1992).

31. Meeker R, Nielsen B, Tewari KK: Localization of replication origins in pea chloroplast DNA. Mol Cell Biol 8:1216-1223 (1988).

32. Nielsen BL, Lu Z, Tewari KK: Characterization of the pea chloroplast DNA oriA region. Plasmid 30:197-211 (1993).

33. Nielsen BL, Rajasekhar VK, Tewari KK: Pea chloroplast DNA primase: characterization and role in initiation of replication. Plant Mol Biol 16:1019-1034 (1991).

34. Palmer J: Isolation and structural analysis of chloroplast DNA. Meth Enzymol 118:167-186 (1986).

35. Palmer JD: Comparative organization of chloroplast genomes. Annu Rev Genet 19:325-354 (1985).

36. Palmer JD, Osorio B, Aldrich J, Thompson WF: Chloroplast DNA evolution among legumes: Loss of a large inverted repeat occurred prior to other sequence rearrangements. Cure Genet 11:275-286 (1987).

37. Ravel-Chapuis P, Heizmann P, Nigon V: Electron microscopic localization of the replication origin of Euglena gracilis chloroplast DNA. Nature 300:78-81 (1982).

38. Reddy MK, Choudhury NR, Kumar D, Mukherjee SK, Tew- ari KK: Characterization and mode of in vitro replication of pea chloroplast oriA sequences. Eur J Biochem 220:933-941 (1994)~

39. Sambrook J, Fritsch EF, Maniatis T: Molecular Cloning: A Laboratory Manual, 2rid ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY (1989).

40. Sanger F, Nicklen S, Coutson AR: DNA sequencing with chain- terminating inhibitors. Proc Natl Acad Sci USA 74:5463-5467 (1977).

41. Shinozaki K, Ohme M, Tanaka M, Wakasugi T. Hay- ashida N, Matsubayashi T, Zaita N, Chunwongse J, Obokata J, Yamaguchi-Shinozaki K, Ohto C, Torazawa K, Meng BY, Sugita M, Dent H, Kamogashira T, Yamada K, Kusuda J, Takaiwa F, Kato A, Tohdoh N, Shimada H, Suguira M: The complete nucteotide .~quence of the tobacco chloroplast gen- tree: its gene organization and expression. EMBO J 5: 2043- 2049 (1986).

706

42. Takeda Y, Hirokawa H, Nagata T: The replication origin of proplastid DNA in cultured cells of tobacco. Mol Gen Genet 232:191-198 (1992).

43. Taylor WC: Regulatory interactions between nuclear and plastid genomes. Annu Rev Plant Physio140:211-233 (1989).

44. Tewari KK: Replication of plant organelle DNA. In: Bryant JA, Dunham VL (eds) DNA Replication in Plants, pp. 69-116. CRC Press, Boca Raton, FL (1988).

45. van't Hof J, Lamm SS: Single-stranded replication intermediates of ribosomal DNA replicons of pea. EMBO J 10: 1949- 1953 (1991).

46. Vaughn JP, Dijkwel PA, Hamlin JL: Replication initiates in a broad zone in the amplified CHO dihydrofolate reductase domain. Cell 61:1075-1087 (1990).

47. Waddell J, Wang X-M, Wu M: Electron microscopic localization of the chloroplast DNA replication origins in Chlamydo- monas reinhardii. Nucl Acids Res 12:3843-3856 (1984).

48. Wu M, Lou JK, Chang DY, Chang CH, Nie ZQ: Structure and function of a chloroplast DNA replication origin of Chlamydo- monas reinhardii. Proc Natl Acad Sci USA 83:6761-6765 (1986).

characterization of replication origins flanking the 23s rrna gene in tobacco chloroplast dna

Documents