Isolation and Characterization of Chloroplast DNA from the Marine ...

Plant Physiol. (1981) 68, 641-647 0032-0889/81/68/0641/07/$00.50/0

Isolation and Characterization of Chloroplast DNA from the Marine Chromophyte, Olisthodiscus luteus: Electron Microscopic Visualization of Isomeric Molecular Forms' Received for publication August 26, 1980 and in revised form February 24, 1981

JANE ALDRICH2 AND ROSE ANN CATTOLICO Department of Botany, AJ-JO, University of Washington, Seattle, Washington 98195 contour length (22) for liverwort, 43 to 45 ,Im for fern (22), and 38 and 46 ,im for angiosperm species (20, 26) such as corn, oats, pea, snapdragon, and evening primrose. Chloroplast DNA contour lengths of four chlorophytic algal representatives have been successfully analyzed. A mol wt of 1.5 x IO' daltons has been reported (16, 33) for Acetabularia. Although linear molecules with contour lengths as high as 200 ,um have been recovered from Acetabularia, intact ctDNA molecules have not been recovered from this organism. Among the less structurally complex unicellular Chlorophyta, the chloroplasts of Chlamydomonas reinhardtii have been shown (5) to contain a circular 62 ,um ctDNA molecule whereas Euglena gracilis which has green algal-like chloroplasts (15) contains ctDNA whose unit chromosome is 40 um in size (30). Finally, the multi-cellular coenocyte, Codium fragile (Hedberg and Hammersand, personal communication) has a unit chloroplast chromosome of approximately 28 ,um in length. This size represents the smallest ctDNA chromosome among those studied. In all Chlorophyta studied to date, the ctDNA is arranged (15, 21) in nucleoidal packets dispersed throughout the organelle. Each nucleoidal packet contains many unit chromosome sets (21). Although some members (15) of the Chromophyta also have this nucleoidal arrangement, those Chromophyta in which girdle lamella are present within the plastid have been shown (15) to contain a single ring-shaped nucleoid. In this report, the first detailed study on the isolation and characterization of the chloroplast unit chromosome from a chromophytic alga, Olisthodiscus luteus4 is reported.

ABSTRACT

Chloroplast DNA (ctDNA) from the marine chromophytic alga, Olisthodiscus luteus, has been isolated using a whole cel lysis method folowed by CsCI-Hoechst 33258 dye gradient centrifugation. This DNA, which has a buoyant density of 1.691 grams per cubic centimeter was identified as plastidic in origin by enrichment experiments. Inclusion of the nuclease inhibitor aurintricarboxylic acid in all lysis buffers was mandatory for isolation of high molecular weight DNA. Long linear molecules (40 to 48 micrometers) with considerable internal organization comprised the majority of the ctDNA isolated, whereas supertwisted ctDNA and open circular molecules averaging 46 micrometers were occasionally present. Also observed in this study were folded ctDNA molecules with electron dense centers ("rosettes") and plastid DNA molecules which have a tightly wound "key-ring" center. The ctDNA of Olisthodiscus has a contour length that is median to the size range reported for chiorophytic plants. A minor component of the total celular DNA, which originates from a DNase insensitive cellular structure, has a buoyant density of 1.694 grams per cubic centimeter. This DNA consists predominantly oflinear molecules, but open circles 11.5 micrometers in length and rare 22-micrometer dimers were also present. This study represents the first analysis of the extranuclear DNA of a chromophytic alga.

MATERIALS AND METHODS Sources of Chemicals. BSA, Sarkosyl, mannitol, ATA and Cyt c were purchased from Sigma. PEG was purchased as Carbowax 6000 from Union Carbide Corp. (New York, NY) and Pharmacia Fine Chemicals (Piscataway, NJ) was the source of Ficoll. Renografm was purchased from Squibb. DNase I that was free of RNase was purchased from Worthington Biochemical Corp. (Freehold, NJ). Calbiochem-Behring Corp. (La Jolla, CA) supplied the Hoechst 33258 dye, and optical grade CsCl was purchased from Harshaw Chemical Co. (Solon, OH). Baker Chemical Co. (Phillipsburg, NJ) provided crystallized phenol and 2-mercaptoethanol. The phenol was redistilled before use and was stored in a light-protected container at -20 C. Collodion was purchased from Ladd Laboratories, Burlington, VT. Cell Maintenance and Harvest. Olisthodiscus luteus (Carter) was grown in 800 ml of artificial sea water medium contained in 3liter Fernbach flasks. Cultures were maintained on a 12 h light/12

The Chromophyta represent a plant division whose members possess both Chl a and carotenoid pigments but lack Chl b. A reference (5) to the unpublished data of Hennig and Kowallik suggests that chloroplasts of the Xanthophycean alga Vaucheria sessilis contain circular DNA molecules of 37 um contour length. This limited information exists as the only data on the chloroplast chromosome unit size for any member of this important evolutionary line of plants. In contrast to the paucity of information on ctDNA3 structure among the Chromophyta, considerable data is available from a wide variety of Chlorophyta, those plants which contain both Chl a and b. Chloroplasts of the Chlorophyta contain multiple DNA copies (4, 18) whose unit chromosome is a circular molecule (20, 26). Land plant representatives of this supertaxa display a ctDNA unit chromosome which ranges from 38.5 ,um in

'Supported by National Science Foundation Grant PCM7624440 to RAC and United States Public Health Grant HDO7183 from the NICHD to J. A. 2Present address: Standard Oil of Ohio, Cleveland, Ohio. 3 Abbreviations: ctDNA, chloroplast DNA; Sarkosyl, sodium n-laurol sarcosinate; ATA, aurintricarboxylic acid.

4 The taxonomic affmity of 0. luteus is disputed. The organism has been described as a Xanthophyte (7) a Chrysophyte (14) and a Chloromonad

(28). 641

642

ALDRICH AND CATTOLICO

Plant Physiol. Vol. 68, 1981

h dark regime as previously described (8). Cells were harvested in when 1 mm ATA (17) was included in all buffer systems. Preparative CsCI Hoechst 33258 Dye Density Centrifugation. the linear phase of growth (5 x 105 cells/mi) by centrifugation using a GSA rotor in a Sorvall RC-5 centrifuge for 15 min at Five-ml gradients contained CsCl at a density of 1.69, 3% w/v 3,000 rpm at 4 C. Large culture volumes were harvested using an Sarkosyl, a maximum of 200 ,ug DNA and 1 mg Hoechst dye SS-34 rotor with a Szent-Gyorgi-Blum continuous flow attach- (stock solution was 10 mg/ml distilled H20). Although increased ment. Routine analysis for fungal and bacterial contamination amounts of DNA precluded DNA species separation, dye amounts could vary from 1 to 20 times the amount of DNA present. was stringently maintained. Buffers. Buffer A contains 0.23 M mannitol, 20 mm Tris, 0.1% Gradients were centrifuged at 40,000 rpm for 24 to 36 h using a w/v BSA, and 0.01% v/v 2-mercaptoethanol (pH 7.6), whereas Beckman centrifuge and Ti 65 rotor. An easily removable floating buffer P is composed of buffer A to which 10%o w/v PEG and 1% green mat was present on the gradient after the first centrifugation. w/v Ficoll were added. These two buffers were used in cellular These initial gradients were fractionated manually to remove bulk disruption steps. Buffer R contains 10 mM Tris, 70 mm sucrose nuclear DNA. Extranuclear DNA species were centrifuged to (pH 7.0) and was used in sucrose-Renografm gradient fractiona- homogeneity and fractionated using an ISCO fractionator (Instrution of chloroplasts. Buffer D contains 5% w/v Sarkosyl, 0.15 M mentation Specialties Co., Lincoln, Nebr.) at a flow rate of 0.375 Na3C6H5O7.2H20, 50 mm EDTA, and 2 mM ATA (pH 7.0) and ml/min. No additional dye was required for these subsequent purification steps. Dye removal was effected by eight extractions is used in DNA extraction. of the recovered DNA with CsCl-saturated isopropanol. To reChloroplast Isolation. Protocol I. A total of 1 to 2 x 109 cells were collected by move CsCl, the DNA solution was dialyzed against two liter centrifugation. Pellets, resuspended in 80 ml buffer A, were dis- changes of a pH 8.0 buffer containing 100 mm NaCl, 50 mm Tris rupted at 750 psi using an automatic French Press. The cell and 10 mm EDTA. This treatment was followed by dialysis with homogenate was centrifuged at 2,700 rpm for 15 min at 5 C using two 1-liter changes using a pH 8.0 buffer which contained 0.1 mm Tris, 1 mm EDTA. Buffer changes were made every 6 or 12 h a Sorvall RC-5 centrifuge and an HB-4 rotor. Alternatively, a DNase treatment (see below) occurred before this centrifugation after which the DNA was either used immediately for electron step. The chloroplast pellet was resuspended in buffer R which microscopy, analytical CsCl gradient pycnographic analysis accontained 50%o w/v Renografm. This suspension became the most cording to the method of Cattolico (9), or was stored at -20 C. Electron Microscopy. DNA obtained from CsCl-Hoechst dye dense solution in a 0, 10, 20, 25, and 50% sucrose-Renografm gradient (6). The gradient was centrifuged at 10,000 rpm for 1 h gradients was prepared for electron microscopy by the Kleinschat 5 C using a Beckman L5-50 centrifuge and SW 27 rotor. midt (25) method. DNA spread on a Cyt c monolayer was picked Chloroplasts which banded at the 10 to 20%/o interface were col- up using collodion coated grids. The grids were stained with lected by use of a pasteur pipette and diluted by adding 100 uranyl acetate, rotary shadowed with platinum/paladium (80/20) volumes of half strength buffer R. The diluted suspension of and examined using a JEOL 100 B electron Microscope. Negatives plastids was centrifuged at 10,000 rpm for 15 min using an SS-34 of electron micrographs of DNA molecules were enlarged and the rotor. The pellets were stored at -20 C or were used immediately molecules were traced using an Electronic Graphics calculator (Numonics Corp., Lansdale, Pa.). A grating replication calibration for ctDNA isolation. grid (E. Fullam, Inc., Schenectady, N. Y.) of 2157 lines/mm Protocol II. Depending on the experiment, a total of 1 to 20 x 109 cells were collected by centrifugation. Pellets, resuspended in provided a magnification reference. The molecule linear density, 100 ml buffer P were disrupted in a French Press at 1,000 psi. 2.08 x 106 daltons/,um, was used (40) to calculate the mol wt from DNase treatment followed and was identical to that described the contour length. In later experiments, 4X174 was included as below except that the final chloroplast washes were with buffer P an internal reference DNA as Stuber and Bujard (40) have shown rather than buffer A. The homogenate was centrifuged at 2,700 that this DNA is adequate as an internal reference for higher mol wt DNA contour length determinations. rpm at S C for 15 min in an SS-34 rotor and the pellet was resuspended in buffer P. The DNase-treated, washed pellet was RESULTS placed on a sucrose pad (0.8 M sucrose, 0.15 M NaCl, 0.2 M EDTA, 50 mM Tris, 1 mm ATA, pH 8.0) and centrifuged at 5,000 rpm for Chloroplast DNA Enichment. Two DNA species are visible 1 h at 5 C. The pellet was used immediately for ctDNA isolation. when whole cell DNA is analyzed by isopycnic centrifugation DNase Treatment. Following cellular disruption, both DNase using Model E ultracentrifugation. The nuclear species has a (100 ,ug/ml) and MgCl2.6H20 (10 mM final concentration) were buoyant density of 1.702 ± 0.001 g/cm3, whereas a maior satellite added to the homogenate. This solution was then incubated at 0 C species has a buoyant density of 1.691 ± 0.001 g/cm (Fig. IA). for 1 h. After enzymic treatment, the chloroplasts were washed 5 DNA extracted from isolated chloroplasts recovered from a sutimes using buffer A which contained 50 mM EDTA and 1 mm crose-Renografm gradient displays a Model E in which 25 ATA. Centrifugation was at 2,700 rpm at 5 C in an SS-34 rotor to 30%o of the nucleic acid is of the 1.691 g/cm?rofile species (Fig. 1B). for each wash step. DNA Extraction. Whole Cell. Chilled pellets of approximately 4 x 108 cells were lysed in 6 ml of buffer D. After 5 mm, when lysis was complete, solid CsCl was added and the lysate was then prepared for CsCl density gradient centrifugation in the presence of Hoechst 33258 dye (see below). Chloroplast. Chloroplasts prepared by Protocol II were lysed in an equal volume of buffer D at 5 C for 20 to 40 min. Upon lysis completion, which was monitored by light microscropy, solid CsCl was added and Hoechst 33258 eye-density centrifugation followed. 13 . 1711.731 1.61 Alternatively, chloroplasts prepared by Protocol I were lysed as described above and the DNA was purified by the phenol extracFIG. 1. Buoyant density distribution of Olisthodiscus DNA species. A, tion technique of Cattolico (9). Occasionally, in this phenol ex- whole cell DNA; B, DNA from isolated chloroplasts; C, DNA from traction technique, a pronase digestion step was included before isolated chloroplasts which have been DNase-treated. 1.691 = chloroplast centrifugation. DNA species of much higher mol wt were obtained DNA; 1.702 = nuclear DNA; 1.731 = Micrococcus luteus marker DNA.


643

ctDNA IN A CHROMOPHYTIC ALGA

If chloroplasts are DNase-treated before the Renografm sucrose gradient step to remove contaminating nuclear DNA, 75 to 80%o of the DNA in the resulting Model E profile is represented by the 1.691 g/cm3 buoyant density component (Fig. IC). These enrichment experiments provide strong evidence that the 1.691 g/cm3 component is chloroplast in origin. Extranuclear DNA: the Presence of Two Satellites. Highly refractile chloroplasts give indication that the membrane system of the organelle is intact (39). This factor is critical in reducing plastid DNA loss resulting from nuclease penetration of the organelle. Although chloroplasts of excellent quality (Fig. 2) were isolated from Olisthodiscus, recovery of only 1% of the theoretical yield of ctDNA was obtained. This low yield was not a result of DNase penetration, for high mol wt ctDNA was isolated (see below) from these DNase-treated organelle preparations. The low ctDNA recovery values obtained in this experimental approach were probably due to the presence of PEG in the chloroplast isolation buffer. This reagent, although critical to chloroplast maintenance in early DNase experiments caused extensive chloroplast clumping (Fig. 2B) when the DNase cofactor Mg2+ was added. Thus, PEG was virtually impossible to remove and tenaciously protected the chloroplasts from disruption. For this reason, a whole cell lysis technique (see "Materials and Methods") was developed. This method represented a significant advance in the isolation of plastid DNA from Olisthodiscus, for ctDNA recoveries near 75% of that theoretically present within the cell were obtained. DNA isolated from these whole cell lysates was fractionated on CsCl gradients. The gradients contained the dye, Hoechst 33258, which preferentially binds to AT-rich regions of the DNA (31). Three DNA species were separable using this dye-gradient combination (Fig. 3). The buoyant density of these DNA species by Model E analysis were 1.702 ± 0.001 g/cm3 (nuclear), 1.691 + 0.001 g/cm3 (chloroplast), and 1.694 ± 0.001 g/cm3 (origin unknown). When DNA was isolated from chloroplast preparations and analyzed by the Hoechst dye-CsCl method, the 1.691 g/cm3 and 1.694 g/cm3 DNA species were present in highly enriched quantities relative to 1.702 g/cm3 nuclear DNA. It should be noted that the origin of the cryptic 1.694 g/cm3 DNA species is unknown. Inasmuch as this satellite is seen in DNase-treated chloroplast preparations, it must originate from a nonnuclear, DNase-insensitive organelle. Although not directly measured, this DNA species represents approximately 0.05 to 0.1% of the total DNA (2.2 x 10-12 g) within the cell; hence, isolation of this satellite required enormous quantities of cells.

qpC

A

1731 1702

B

1731 1.694 1691

C

1.731 1691

FIG. 3. Separation of whole cell DNA into component species using a Hoechst dye-CsCl gradient. Model E ultracentrifugation of recovered bands (lower scans) revealed that band A is the nuclear 1.702 buoyant density component, band B is a satellite of unknown origin having buoyant density of 1.694 and band C is chloroplast DNA which has a buoyant density of 1.691. Micrococcus luteus marker DNA has a buoyant density of 1.731.

Size Analysis: 1.694 g/cm3 Satellite DNA. Six centrifugation cycles were required to obtain a homogeneous 1.694 g/cm3 satellite species before it was spread for electron microscopic analysis. The predominant proportion of this DNA species was present in linear form. However, 5 to 15% of the molecules were observed to occur (Fig. 4) in both relaxed and highly twisted configurations. The presence of the nuclease inhibitor ATA greatly increased the number of intact circular molecules recovered. The relaxed molecules of the 1.694 g/cm3 DNA species had a contour length of 11 ± 0.7 ,um (23.0 ± 1.5 x 106 daltons). A total of 60 molecules obtained from both cell lysates and isolated chloroplast preparations were measured. Occasionally, larger relaxed and highly twisted molecules (Fig. 4-2) were observed (less than 0.1% of the population). The relaxed molecules (12 measured) had a contour length of 21.9 ± 1.2 tim (45.5 ± 2.5 x 106 daltons) and probably represent a dimer of the smaller size class. Size Analysis: 1.691 g/cm3 Chloroplast DNA. The 1.691 g/cm3 buoyant density component was centrifuged to homogeneity (four rebandings) and spread for electron microscopic analysis. Although the majority of molecules observed were linear with a mol wt of 80 to 100 x 106 daltons, a very low proportion of the molecules (0.5% or less) were either in supertwisted (Fig. SA) or b A relaxed (Fig. 5B) form. A total of 15 circular molecules were FIG. 2. Isolated chloroplasts from Olisthodiscus luteus. A, chloroplasts measured and a contour length of 45.7 ± 2.5 ,um (94.9 ± 5.2 x 106 are discrete refractile bodies following cellular disruption. B, clumping of daltons) was obtained. Five of these molecules were from DNasechloroplasts results from addition of Mg2" required for DNase treatment, treated chloroplasts (94.5 ± 4.7 x 106 daltons) and 10 were from whole cell lysates (95.1 ± 5.5 x 106 daltons). but refractility is maintained.

644


V~~~ .1

FIG. 4. Circular DNA molecules from 0. luteus. These were obtained from the 1.694 g/cm3 DNA satellite. 23 x

106 daltons

and

(2),

(1), indicates supertwisted and relaxed

indicates

supertwisted

and relaxed 46 x

106

daltons DNA.

0.5

FIG. 5. Circular DNA molecules from 0. luteus. These molecules were from the 1.691

relaxed form.

g/cm3 (chloroplast)

DNA satellite.

A,

supertwisted,

and

B,


Chloroplast DNA: Internal Organization. Although fully supertwisted or relaxed molecules were observed in the electron microscopic analysis of ctDNA, other topological forms in which a high degree of internal structural organization occurred were also evident. The linear molecule of Figure 6C, which represents the major size class of ctDNA isolated displays distinct "organizational centers" from which loops of twisted or untwisted DNA extend. Organizational centers connected by several strands of untwisted DNA from which supertwisted or relaxed DNA loops emanate are also present within the relaxed ctDNA dimer of Figure 6E. An electron opaque core (Fig. 6, A and D) was occasionally located at the organization center. These rosette forms (core + looped or linear DNA) were often clustered (Fig. 6D) and frequently joined one another via strands of DNA. Rosettes of different sizes were observed to occur. For example, the two rosette structures of Fig. 6D were estimated to be 22 and 39-x 106 daltons, whereas the molecule displayed in Figure 6A has a mol wt of 80 x 106 daltons, which approaches unit chromosome length. A most interesting topological form (Fig. 6B) was seen only twice in all samples prepared for electron microscopic analysis. Although this structure (approximately 80 x 106 daltons) lacked a core region, it did contain a "key-ring" arrangement (27) of tightly wound DNA (see Fig. 6B inset) from which loops of DNA extend. To eliminate the possibility that rosette-like structures were an artifact of spreading, ammonium acetate concentration in the hypophase and hyperphase was varied in DNA spreads prepared for electron microscopy. Variations were also made in Cyt c age and concentration. Although these factors were found (11) to affect the quality of DNA spreading, rosette molecules remained under all conditions tested. Several additional observations suggest that rosette structures were not artifactual. (a) It is known (3) that intercalation of ethidium bromide into DNA decreases the buoyant density of the molecule. When Olisthodiscus ctDNA was fractionated on CsCl-ethidium bromide gradients, rosette molecules were present only within the most dense portion of the banded DNA. The gradient fraction in which the rosettes appeared (1) never contained supertwisted DNA molecules and migrated lower than those fractions which contained either long linear or relaxed circular DNA. This observation is consistent with limited ethidium bromide uptake imposed (3) by a conformationally constrained molecule. (b) Enhanced recovery of rosette molecules was always obtained when ctDNA was minimally cycled through CsCl-Hoechst dye gradients, suggesting that unfolding of the ctDNA organization centers and DNA purification were simultaneous events. (c) Although very long, linear molecules of nuclear DNA (1.702 g/cm3 buoyant density) with nucleosomal structures were obtained when minimal CsCl gradient centrifugation was performed, no rosette structures were seen in this DNA species. Further cycling of the nuclear DNA through CsCl gradients produced only linear DNA which was devoid of nucleosomal particles. (d) As a control, 4X174 DNA was co-spread with ctDNA. No rosette structures were induced by the experimental conditions of DNA spreading in this exogenous DNA source. DISCUSSION The highly fluorescent dye, Hoechst 33258, which binds preferentially to AT-rich DNA sequences, has been employed successfully in CsCl gradients to accentuate the buoyant density differences of cellular DNA (12, 31, 38). Excellent resolution of the three 0. luteus cellular DNA species present in crude cell lysates is obtained in a one-step purification/separation procedure. A significant variation in DNA to dye ratio (1- to 10-fold) is easily tolerated, thus this preparative technique offers a valuable method for recovery of DNA from crude cell lysates where DNA quantity may only be estimated. The fact that the minor 1.694 g/cm3 satellite was revealed and successfully purified from the chloro-



645

*.D°-e.,,'> e A+,;eCa,,A, s @ * ° j 8 .lk 1 + A s 4 t X *~ ~ '

* ," 47 ..* j.. .:.y>'§;-... -.

1.

¼'.' 'C

-v.r.44

4$'

444* V~

I 7./I.

'...***..'%:

4.

NV

at

,-

.,

A

4

.4'.. a.;

¼

~~~~

- ;s6'

4,

,

tt-A, s"2''"

4)

I

"'

"'-r

gsA'

...½.. 7-:0'

of.4-

*'*

444

½ 4n.1&. 'rc, FIG. 6. Highly organized molecules present in ctDNA preparations. A, folded molecule of near unit length (80 x 106 daltons) with core body attached at organization center. B, folded molecule of near unit length (80 x 106 daltons) lacking core body. Inset shows compact coiling of center keyring. C, linear molecule of approximately 100 x 10 daltons; arrows indicate centers of organization. D, smaller folded rosette-like molecules joined at their centers by DNA strands. The single center part of this joint structure is 39 x 106 daltons and the double center part is 22 x 106 daltons. E, dimer of approximately 200 x 106 daltons with super-twisted regions (long arrow), and organization centers (2 small arrows) from which large twisted and untwisted loops extend.

plast and nuclear DNA species, demonstrates the power of the method. A mol wt of 23 x 106 daltons obtained from contour length

measurements of circular molecules in the 1.694 g/cm3 cryptic satellite DNA is in agreement with values obtained (Aldrich et al, manuscript in preparation) by restriction enzyme analysis of this

646


DNA. Although the origin of this DNA is presently unknown, it is probably not a nuclear component. Enhanced recovery of this satellite is routinely obtained after the DNase treatment step of the isolation procedure. Only a DNA present within a DNase insensitive structure wouldshow this refractility to destruction. Since the chloroplast preparations are contaminated by mitochondria as evidenced from electron microscopy, it is possible that the 1.694 g/cm3 species originates from this extranuclear source. Although experiments are in progress, the isolation and characterization of mitochondrial DNA from Olisthodiscus has been extremely difficult. Contour length measurements of the open circular molecules of the 1.691 g/cm3 ct DNA of 0. luteus provide a mol wt of an average of 95x 106 daltons. Restriction enzyme analysis indicates that this DNA is a homogeneous population of molecules of an average of 99 x 106 daltons (Aldrich et al., manuscript in preparation). In addition, reassociation kinetic analysis (Ersland et al., manuscript in preparation) using phage X and T4 DNA as internal standards gives an Olisthodiscus ctDNA genome size of 91 x 106 daltons and 105 x 106 daltons, respectively. Therefore, by three independent measurements, an average value of 97 x 10W daltons may be assigned to the ctDNA. Whether this unit chromosome is circular or linear remains to be confirmed by restriction enzyme

mapping. It has been reported (9) that a change in the growth conditions under which an 0. luteus culture is maintained will induce the organism to alter its chloroplast complement from 38 to 14 plastids. Both chloroplast and nuclear DNA amount per cell remain constant during this transition. As these earlier studies demonstrated, a reciprocal relationship exists between chloroplast number and DNA content per plastid. Knowledge of the size of the 0. luteus chloroplast chromosome permits assignment of a copy number to these data. A cell which has a 14 plastid complement potentially contains 34 ctDNA molecules per chloroplast whereas a cell with 38 plastids would contain 12. The lower limit of DNA needed to maintain a fully functional chloroplast (one which increases mass and replicates) is still an unanswered question. The isolation of structural isomers of ctDNA may provide insight into the possible organization and localization of the unit chromosome within the chloroplast nucleoid. Highly organized internal molecular arrangement of the ctDNA is consistently observed. Frequently, molecules containing a double-stranded break unwind in only a restricted area, while maintaining other local domains in a supertwisted configuration. Organization centers or areas from which loops of DNA emanate in both a supertwisted or untwisted form were also frequently noted. The organization centers seen in the 0. luteus ctDNA molecules (Fig. 6, C and E) strikingly resemble the complexes formed by catenation of the plasmid HM 456 and SV 40 viral DNA by topoisomerase activity of Xenopus germinal vesicle extracts (2). The rare "key-ring" arrangement seen in Figure 6B is similar to the key-ring structure observed (27) in the enormous Col El plasmid catenates generated by Escherichia coli gyrase. These observations imply that topoisomerase activity may exist within the chloroplast. DNA supercoiling which results from the double stranded DNA break-and-reunion activity of these enzymes is suggested to be important to the process of DNA replication, recombination and transcription (10). The fact that ctDNA replication in Euglena is inhibited by naladixic acid (29), an inhibitor which specifically interferes (13) with subunit A activity (nicking-closing) of topoisomerase, supports this hypothesis. Folded chromosomes with twisted or relaxed loops attached to a core structure seen in 0. luteus have also been reported for ctDNA of Spinacia (19, 41), Antirrhinum (19), and Sphaerocarpos (22). These molecules shows a remarkable similarity in appearance to both the looped folded E. coli chromosome (24) and to the histone-depleted HeLa cell metaphase chromosomes (35) wherein


concentric DNA loops are attached to scaffold proteins. In 0. luteus folded ctDNA chromosomes with core structures are recovered most frequently in first run CsCl-Hoechst dye gradients. In these gradients, virtually no supercoiled or relaxed ctDNA molecules are seen. Subsequent centrifugation results in loss of the folded ctDNA form and a low yield recovery of supertwisted or relaxed ctDNA molecules. This low recovery of circular DNA forms might result from a transition from the folded molecule directly to a nicked linear form which is induced by high salt core stripping. The fact (20) that 80%o of the total plastid DNA of spinach is in the open circular form following repeated CsCl gradient centrifugation suggests that the affinity of the core body to DNA varies among plant types. The molecular composition of a"Gcore"l structure is not known. Work with spinach (41) has demonstrated that prolonged proteinase K treatment is necessary to remove core structures from ctDNA indicating the presence of a "masked" or of a "heterogeneous" protein component. This enzymic treatment causes conversion of the folded spinach ctDNA to the linear form. Spinach ctDNA cores are refractile to diethylpyrocarbonate, SDS, chloroform, and RNase treatment. Similar refractility is seen in 0. luteus ctDNA, where core bodies survive mild pronase treatment, phenol extraction and RNase digestion. Folding of ctDNA may not be entirely dependent upon the presence of a core structure. When membrane-free E. coli chromosomes are isolated, an RNase sensitive site is apparent (36). Pettijohn and Hecht (36) have suggested that DNA-RNA interaction is involved in the maintenance of folded chromosome integrity. Whether this mode of chromosome folding occurs within the chloroplast is unknown. The replicon model of Jacob et al. (23) postulates that DNA membrane attachment sites are involved in DNA replication. The fact that (a) ctDNA is closely associated with thylakoid membranes in higher plant (14, 37) and algal (14) cells, and (b) [3HJthymidine incorporation into ctDNA is closely associated with granal thylakoids, suggests that membrane ctDNA interaction might also be of importance in plastid DNA replication. Recently, Pardoll et al. (34) have presented a model of DNA replication in which DNA is attached to a nuclear matrix at numerous fixed replication complexes. Looped DNA is reeled through these sites and replicated. Isolation ofjoined rosette structures from 0. luteus (Fig. SD) supports this model. The presence of one or more attachment sites per chloroplast chromosome of 0. luteus may be logistically advantageous in segregation of ctDNA within the chloroplast which divides (Cattolico, unpublished) its ring-shaped nucleoid by central fission. In summary, the maintenance, replication, and segregation of ctDNA may be dependent upon interaction of the DNA with thylakoid membranes, intermolecular catenation/decatenation promoted by topoisomerase(s) and possible DNA-RNA interactions. The 0. luteus chloroplast chromosome (98 x 106 daltons) is mid-range in size when compared to the chloroplast chromosome of the Chlorophyta. It is interesting to note that the organellar DNA of these two plant supertaxa are similar (32) in size to the symbiotic, nonfreeliving cyanelles of Cyanophoraparadoxa. These cyanelles, which have a genome size of 115 x 106 daltons, have been postulated to represent an intermediate between blue green algae and chloroplasts. The mechanism and significance of maintaining the remarkable size constancy among ctDNA populations of different plant types remains an intriguing question. Acknowledgment-We would like to thank D. McCabe and J. Voss for their

excellent technical support and H. Lyman and L. McIntosh for helpful discussions. LITERATURE CITED

1. ALDRICH J 1980 Extranuclear DNA from the marine chromophyte, Olisthodiscus luteus. PhD thesis, University of Washington, Seattle, Washington 2. BALDI MI, P BENEDETTI, E MATTOCCIA, GP TocCHINI-VALENTINI 1980 In vitro catenation and decatenation of DNA and a novel eukaryotic ATP-dependent



topoisomerase. Cell 20: 461467 3. BAUER W, J VINOGRAD 1970 The interaction of closed circular DNA with intercalative dyes. III. Dependence of the buoyant density upon superhelix density and base composition. J Mol Biol 54: 281-298 4. BEDBROOK JR, R KOLODNER 1979 The structure of chloroplast DNA. Annu Rev Plant Physiol 30: 593-620 5. BEHN W, RG HERRMANN 1977 Circular molecules in the ft-satellite of Chiamydomonas reinhardii Molec Gen Genet 157: 25-30 6. BROWN RD, R HASELKORN 1972 The isolation of Euglena gracilis chloroplasts uncontaminated by nuclear DNA. Biochim Biophys Acta 259: 1-4 7. CARTER N 1937 New or interesting algae from brackish water. Arch Protistenk N 90: 1-68 8. CATrOLICO RA, JC BooTHRoYD, SP GIBBs 1976 Synchronous growth and plastid replication in the naturally wall-less alga Olisthodiscus luteus. Plant Physiol 57: 497-503 9. CATTOLICO RA 1978 Variation in plastid number: effect on chloroplast and nuclear deoxyribonucleic acid complement in the unicellular alga Olisthodiscus luteus. Plant Physiol 62: 558-562 10. CozzARELLi NR 1980 DNA gyrase and the supercoiling of DNA. Science 207: 953-960 11. DAvis RW, M SIMON, N DAVIDSON 1971 Electron microscope heteroduplex methods for mapping regions of base sequence homology in nucleic acids. Methods Enzymol 21: 413-428 12. FENNEWALD M, W PREVATT, R MEYER, J SHAPIRO 1978 Isolation of Inc P-2 plasmid DNA from Pseudomonas aeruginosa. Plasmid 1: 164-173 13. GEQuiRT M, K MizuucHI, MH O'DEA, T ITOH, J ToMiZAWA 1977 Nalidixic acid resistance: a second genetic character involved in DNA gyrase activity. Proc Natl Acad Sci USA 74: 4772-4776 14. GIBBS SP 1962 Nuclear envelope-chloroplast relationships in algae. J Cell Biol 14: 433-444 15. GIBBS SP 1970 The comparative ultrastructure of the algal chloroplast. Ann NY Acad Sci 175:454-473 16. GREEN BR 1976 Covalently closed minicircular DNA associated with Acetabularia chloroplasts. Biochim Biophys Acta 447: 156-166 17. HALLICK RB, BK CHLEM, PW GRAY, EM OROZCO JR 1977 Use of aurintricarboxylic acid as an inhibitor of nucleases during nucleic acid isolation. Nucleic Acids Res 4: 3055-3064 18. HERMANN RG, KV KOWALLIK 1970 Multiple amounts of DNA related to the size of chloroplasts. II. Comparison of electron microscopic and autoradiographic data. Protoplasma 69: 365-372 19. HERMANN RG, KV KOWALiK, HJ BOHNERT 1974 Structural and functional aspects of the plastome L. The organization of the plastome. Port Acta Biol 14: 91-110 20. HERRMANN RG, HJ BOHNERT, KV KOwALLIK, JM SCHMITT 1975 Size, conformation, and purity of chloroplast DNA of some higher plants. Biochim Biophys Acta 378: 305-317

647

21. HERRMANN RG, JV POSSINGHAM 1980a Plastid DNA-The Plastome. In J Reinert, ed, Results and Problems in Cell Differentiation, Vol 10. Springer/ Verlag, New York, pp 45-96 22. HERRMANN RG, HK PALTA, KV KOWALLIK 1980b Chloroplast DNA from three Archegoniates. Planta 148: 319-327 23. JACOB F, S BRENNER, F CUZIN 1963 The regulation of DNA replication in bacteria. Cold Spring Harbor Symp Quant Biol 28: 329-348 24. KAVENOFF R, OA RYDER 1976 Electron microscopy of membrane-associated folded chromosomes of Escherichia coli. Chromosoma 55: 13-25 25. KLEINSCHMIDT, AK 1968 Monolayer techniques in electron microscopy ofnucleic acid molecules. Methods Enzymol 12: 361-377 26. KOLODNER R, KK TEWARI 1975 The molecular size and conformation of the chloroplast DNA from higher plants. Biochim Biophys Acta 402: 372-390 27. KREuzER KN, NR COZZARELLI 1980 Formation and resolution of DNA catenanes by DNA gyrase. Cell 20: 245-254 28. LOEBLICH AR, KE FINE 1977 Marine chloromonads: more widely distributed in necritic environments than previously thought. Proc. Biol Soc Wash 90: 388399 29. LYMAN H, AS JUPP, I LARRINUA 1975 Action of nalidixic acid on chloroplast replication in Euglena gracilis. Plant Physiol 55: 390-392 30. MANNING JE, DR WOLSTENHOLME, RS RYAN, JA HUNTER, OC RIcHARDS 1971 Circular chloroplast DNA from Euglena gracilis. Proc Natl Acad Sci USA 68: 1169-1173 31. MANUELIDIS L 1977 A simplified method for preparation of mouse satellite DNA. Anal Biochem 78: 561-568 32. MucKE H, W LOFFELHARDT, JH BOHNERT 1980 Partial characterization of the 'endosymbiotic' cyanelles from Cyanophoraparadoxa FEBS Lett 11 1: 347-352 33. PADMANABHAN M, B GREEN 1978 The kinetic complexity of Acetabularia chloroplast DNA. Biochim Biophys Acta 521: 67-73 34. PMADoLL DM, B VOGELSTEIN, DS COFFEY 1980 A fixed site of DNA replication in eucaryotic cells. Cell 19: 527-536 35. PAULSON JR, UK LAEmmLi 1977 The structure of histone-depleted metaphase chromosomes. Cell 12: 817-828 36. PETrLoHN D, R HECHT 1973 RNA molecules bound to the bacterial genome stabilize DNA folds and segregate domains of supercoiling. Cold Spring Harbor Symp Quant Biol 38: 31-41 37. RoSE RJ 1979 The association of chloroplast DNA with photosynthetic membrane vesicles from spinach chloroplasts. J Cell Sci 36: 169-183 38. SI,SON L 1979 Isolation of maxicircle component of kinetoplast DNA from hemoflagellate protozoa. Proc Natl Acad Sci USA 76: 1585-1588 39. STILL CC, CA PRICE 1967 Bulk separation of chloroplasts with intact membranes in the zonal centrifuge. Biochim Biophys Acta 141: 176-178 40. STUBER D, H BUJARD 1977 Electron microscopy of DNA: determination of absolute molecular weights and linear density. Molec Gen Genet 154: 299-303 41. YOSHIDA Y, JP LAULHERE, C ROZIER, R MACHE 1978 Visualization of folded chloroplast DNA from spinach. Biol Cell 32: 189-193