The persistence of maternal inheritance in ... - Semantic Scholar

Proc. Natl. Acad. Sci. USA Vol. 81, pp. 3438-3442, June 1984

Genetics

The persistence of maternal inheritance in Chlamydomonas despite hypomethylation of chloroplast DNA induced by inhibitors (DNA methylation in vivo/5-azacytidine/L-ethionine/uniparental inheritance/gametogenesis)

TENG-YUNG FENG AND KWEN-SHENG CHIANG Department of Biophysics and Theoretical Biology, University of Chicago, Chicago, IL 60637

Communicated by Hewson Swift, December 28, 1983

ABSTRACT We have used inhibitors of methylation to evaluate the proposal that the extent of methylation of chloroplast DNA (cpDNA) of the mating type-plus (mt') parent occurring during gametogenesis in wild-type Chlamydomonas renharddii is directly correlated with the uniparental transmission of chloroplast genes by this parent [Sager, R., Grabowy, C. & Sano, H. (1981) Cell 24, 41-47]. As detected by highpressure liquid chromatography, the methylation of cpDNA was at its lowest level in the vegetative stage; the mt' cells had a deoxycytidine methylation index (the percentage of deoxycytidine methylated) of 0.5, while the mating type-minus (mtF) index was lower by at least a factor of 3. This basal level of cpDNA methylation increased more than 20-fold after gametogenesis to give a methylation index of 12.1 and 4.3 for mt' and mt- gametes, respectively. Another striking increase was detected at the 7-hr-zygote stage, resulting in the methylation of nearly half of the total deoxycytidine residues. The extent of zygotic cpDNA methylation was shown to be dependent on the preexisting methylation level of both parental gametic cpDNAs. L-Ethionine and 5-azacytidine effectively inhibited cpDNA methylation during gametogenesis and ensuing zygotic development as shown by both Hpa II/Msp I digestion patterns and HPLC. The transmission of chloroplast genes was analyzed concomitantly with the inhibitor studies. The two inhibitors produced different patterns of inhibition of methylation in mtf and mt+ cells at a given developmental stage. Our overall results demonstrate that the extent of mating type-specific and gamete-specific methylation during gametogenesis is not correlated with the frequency of maternal transmission of chloroplast genes.

to occur in any bacterial methylation-restriction system. In the C. reinhardtii mutant me-i, cpDNA of vegetative cells of both mating types is heavily methylated even before gametogenesis; nevertheless, maternal inheritance occurs normally in me-i (5). As detected by restriction fragment patterns, additional methylation of cpDNA seems to occur during gametogenesis in mt+ but not in mt- me-i gametes (4), suggesting that a differential pattern of male versus female gametic cpDNA methylation might exist in the me-i mutant. However, whether such additional differential methylation during gametogenesis is responsible for the maternal inheritance of chloroplast genes in me-i has not been directly shown. Furthermore, it is not known whether the methylation of cpDNA during gametogenesis of mt+ wildtype cells shares common sites with the constitutive high methylation of cpDNA in me-i vegetative cells of both mating types. It also is not known whether the methylation of cpDNA that occurs during gametogenesis of wild-type mt+ cells is required in the regulation of maternal inheritance of chloroplast genes. We report in the present paper an evaluation of this question by using two DNA methylation inhibitors: 5-azacytidine (5-azaCyd) and L-ethionine. Our results demonstrate that extensive methylation of mt+ gametic cpDNA during gametogenesis is not required for the maternal inheritance of chloroplast genes in C. reinhardtii.

MATERIALS AND METHODS Strains, Culture Conditions, and Genetic Experiments. A C. reinhardtii mt+ strain GB-68 carrying a chloroplast mutation erythromycin resistance (er-u-37) and an mt- strain CC146 carrying nuclear mutation neamine resistance and chloroplast mutation spectinomycin resistance (mt-, nr-i-i, spru-l-6-2) were used. Vegetative cell growth, gametogenesis, mating, zygote maturation, and germination were carried out as described (7, 8). The use of 5-azaCyd (9) (Sigma) or Lethionine (Sigma) was as described in the legend to Table 1. Zygote colony isolation, replica-plating, and marker scoring were performed as described by Eves and Chiang (8). Isolation and Characterization of Chloroplast DNA. Chloroplast DNA was isolated by the method of Grant et al. (10). Conditions for endonuclease digestions, gel electrophoresis, and photography were carried out as described (11). HPLC of chloroplast DNA samples was carried out as described by Bolen et al. (5).

By analysis of restriction fragment patterns, extensive methylation of cytosine in chloroplast DNA (cpDNA) was shown to occur in Chlamydomonas reinhardtii mating type-plus (mt') or female wild-type gametes, whereas no such methylation was found in mating type-minus (mt-) or male gametes (1, 2). The gametic cpDNA of the mat-i mutant, which is closely linked to the mt- allele and increases transmission of its chloroplast genes to >50% (3), was methylated to an extent intermediate between wild-type mt- and mt+ gametes (1, 2). Based on this and other evidence (cf. refs. 4 and 5), Sager and co-workers have suggested that the extent of methylation occuring during gametogenesis directly correlates with the genetic transmission of chloroplast genes in wild-type and mat-i mutant cells (2). They proposed earlier that the maternal inheritance of cpDNA in Chlamydomonas is governed by a methylation-restriction system analogous to those in bacteria (6). However, the functional significance of extensive methylation seen in gametogensis of C. reinhardtii mt+ wild-type or mt- mat-i cells (1, 2, 4) is not understood at present because such extensive methylation was not known

RESULTS The Effect of 5-azaCyd and L-Ethionine on Cell Growth. At a concentration up to 1 mM, 5-azaCyd and L-ethionine had little effect on the cell growth rate or cpDNA content; except in stationary phase, the treated culture bleached earlier than the control culture (ref. 12; data not shown). For control ga-

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Abbreviations: cpDNA, chloroplast DNA; 5-azaCyd, 5-azacytidine;

m5dCyd, 5-methyldeoxycytidine.

3438

Proc. NatL. AcadL Sci. USA 81 (1984)

Genetics: Feng and Chiang metes, a mating efficiency of at least 90% was usually attained within 2 hr, whereas only 70-80% mating was observed for gametes treated with 0.2-1 mM 5-azaCyd or Lethionine. No other change in either the behavior of gametes or the ensuing development of zygotes was observed for either treatment. Throughout the present study, zygotes were germinated in the absence of drugs. Germination frequency and zoospore viability were not significantly different from the untreated controls (ref. 12; data not shown). The Inhibition of cpDNA Methylation by 5-azaCyd and LEthionine. The difference in cpDNA methylation between mt and mt+ vegetative cells versus gametes of wild-type and the me-i mutant has been analyzed mainly by restriction fragment patterns after digestion with Msp I and Hpa II (1, 2, 4, 5). We first repeated, with strains GB68 and CC146, the determinations reported previously under similar experimental conditions by Sager and co-workers for wild-type strains (1, 2). The mt' gametic cpDNA, in contrast to mtgametic or vegetative cpDNA of either mating type, was shown to be substantially more resistant to Hpa II (but sensitive to Msp I) than were the other three cpDNA samples (Fig. 1). We then tested the effect of 5-azaCyd and L-ethionine on this methylation of mt+ gametic cpDNA under the identical experimental conditions. Methylation of cpDNA in mt+ cells exposed to 5-azaCyd or L-ethionine during gametogenesis was inhibited by both drugs (Figs. 2 and 3). As the concentration of the drugs was increased from 0.01 to 0.30.4 mM, the resulting mt' gametic cpDNA gradually lost its resistance to Hpa II digestion; above -0.1 mM of 5-azaCyd or L-ethionine, the Hpa II and Msp I digests could no longer be differentiated. It has been reported that immediately after normal gamete fusion, zygotic cpDNA was resistant to the digestion of Msp I, Hpa II, BamHI, and EcoRI (1, 2). We found that this was not the case. However, extensive purification (e.g., three cycles of CsCl gradient purification) of zygotic DNA was necessary prior to any successful digestion with endonuclease. The purified 7-hr zygotic cpDNA could be digested by Msp I to produce an electrophoretic pattern indistinguishable from that of parental mt- or mt+ gametic cpDNA (Figs. 4 c and f and Sa). Nevertheless, the same zygotic cpDNA was substantially resistant to Hpa II digestion (Figs. 4d and Sd), as was the parental mt+ gametic cpDNA (Fig. 4g). If ab c d e f

9h

a b c d e f

H

M

H MH

H

H

H

H

M

Fig. 2. Restriction fragment patterns of 5-azaCyd-treated gametic mt+ cpDNA digested with Hpa II (lanes H) and Msp I (lane M). Addition of 5-azaCyd (mM) at the onset of gametogenesis: 0 (lane a), 0.01 (lane b), 0.03 (lane c), 0.1 (lane d), and 0.3 (lanes e and f).

mt+ gametes were pretreated with 5-azaCyd or L-ethionine during gametogenesis, prior to fusion with control untreated mt- gametes, the resulting zygotic cpDNA became only partially resistant to Hpa II digestion. Numerous poorly digested and inadequately resolved large fragments were detected on top of the gel (Figs. 4b and 5b). Since prior to gametic fusion both untreated mt- and drug-treated mt+ gametic cpDNAs were sensitive to Hpa II, the partial resistance to Hpa II digestion of 7-hr zygotic cpDNA appeared to have resulted from additional methylation after zygote formation, which occurred in the presence of -0.1 mM 5-azaCyd or Lethionine (resulting from mixing of the 0.2 mM treated mt' with an approximately equal volume of untreated mt- gametic culture). A similar partial resistance to Hpa II digestion of 7-hr zygote DNA was observed if the mt- gametes were pretreated with L-ethionine prior to mating with untreated mt' gametes (Fig. Sc). These results indicate that (i) the zygotic methylation of cpDNA that occurred within 7 hr after gametic fusion is not inhibited effectively by 0.1 mM 5azaCyd or L-ethionine under our experimental conditions, and (it) the zygotic cpDNA methylation is dependent on the normal unaltered methylation pattern of both parental gametic cpDNAs. This conclusion is also borne out by the HPLC studies discussed below. The Persistence of the Maternal Inheritance Pattern After Inhibitor Treatment. Having established that 5-azaCyd and L-ethionine caused profound changes in cpDNA methylation during mt' gametogenesis and early zygotic development, we then analyzed the effect of these drugs on the inheritance a

H

3439

e lf

b c d

H H

M

H

FIG. 3. Restriction fragment patterns of L-ethionine-treated mt' gametic cpDNA digested with Hpa II (lanes H)

M

FIG. 1. Restriction fragment patterns of cpDNA digested with Msp I (lanes- M) or Hpa II (lanes- H4) .nLnes: a and b, mt' vegetative; c and d, mt- vegetative; g and h, mt+ gametic; i and j, mt gametic; e, 0.2 mM 5-azaCyd-treated mt- vegetative; f, 0.2 mM L-ethioninetreated mt- vegetative.

and Msp I (lane M). Addition of L-ethionine (mM) at the onset of gametogenesis: 0 (lane a), 0.01 (lane b), 0.03 (lane c), 0.1 (lane d), 0.2 (lane e), and 0.4 (lanes f and

H H

H

H

ff H M

g).

3440 a

Proc. NatL Acad Sci. USA 81

Genetics: Feng and Chiang

b cde

f

a

h

FIG. 4. Restriction fragment patterns of cpDNA digested with Msp I (lanes M) and Hpa II (lanes H). Lanes: a and b, 0.2 mM 5-azaCyd-treated 7-hr zygotic cpDNA; c and d, untreated 7-hr zygotic cpDNA; e and h, untreated mt- gametic cpDNA; and f and g, untreated mt+ gametic cpDNA. genes. Over the entire range of concentrations tested (i.e., 0.01-10 mM), 5-azaCyd and L-ethionine did not alter the normal maternal inheritance pattern, whether the inhibitors were added at the onset of mt' gametogenesis or at the time of gametic fusion. The same is true even if both 5-azaCyd and L-ethionine were added together during mt' gametogenesis (Table 1). This persistence of normal inheritance pattern of chloroplast genes is in striking contrast to the marked inhibition of cpDNA methylation by these drugs (see Figs. 2 and 3). The transmission pattern of chloroplast genes was not affected by treating male or female gametes alone or both gametes during gametogenesis and subsequent zygote development. Here again the normal maternal inheritance of chloroplast genes persisted irrespective of the various combinations of treatment (Table 2). As expected, a few percentages of spontaneous exception-

pattern of chloroplast

b

c

(1984)

d

FIG. 5. Restriction fragment patterns of zygotic cpDNA digested with Hpa II (lanes H) or Msp I (lanes M). Lanes: a, untreated 7-hr zygotic cpDNA; zygotic cpDNA resulting from a cross of 0.2 untreated mtL-ethionine-pretreated gametes; c, 7-hr zygotic cpDNA resulting from a cross of 0.2 mM L-ethionine-pretreated mt- and ungametes; and d, same as lane a. treated b,

7-hr

mt' and

mM

.M

H

H

H

mt+

al zygotes, in which paternal chloroplast genes are transmitted, were observed for the untreated control zygote cultures. In drug-treated crosses, a trend toward a slight increase in the frequencies of paternal and decrease in biparental zygotes, especially when either gametic parent was treated with 5-azaCyd, was observed (Tables 1 and 2). When both gametic parents were treated with either 5-azaCyd or L-ethionine, an increase of paternal zygotes relative to the control was apparent (Table 2). The Effect of 5-azaCyd and L-Ethionine on 5-Methyldeoxycytidine Content in cpDNA. The Hpa II/Msp I digestion analysis is limited to the detection of internal cytidine methylation in 5' C-C-G-G 3' sequences. To examine other cytidine residues, HPLC quantitation of 5-methyldeoxycytidine (m5dCyd) was used (13). A basal methylation index of 0.50 for dCyd in cpDNA was determined for mt+ (GB68) vegetative cells. An increase of >20-fold was observed after gametogenesis to give a methylation index of 12.1 at the gametic

Table 1. Effect of 5-azaCyd and L-ethionine concentration on the inheritance pattern of non-Mendelian genes in C. reinhardtii Drug Concentration, mM 1.0 0.3 0.1 498 622 673 95.38 98.89 95.15 4.62 4.75 1.22 0.1 0.1 0.2

5-azaCyd

Time of addition Gametogenesis*

Total zygotes Maternal Paternal Biparental

5-azaCyd

Gametic fusiont

636 97.01 2.36 0.63

237 97.89 2.11 0.4

578 95.68 4.32 0.2

428 94.86 4.67 0.47

690 97.83 2.17 0.1

265 96.98 3.02 0.37

261 98.47 1.53 0.38

Total zygotes Maternal Paternal Biparental

L-Ethionine

Gametogenesis*

696 97.56 2.01 0.43

395 98.39 2.00 0.25

-

587 98.43 1.40 0.17

235 98.11 1.89 0.04

765 95.95 3.92 0.13

398 99.03 0.97 0.25

307 90.55 9.12 0.33

Total zygotes Maternal Paternal

L-Ethionine

636 97.01 2.36 0.63

747

-

765 98.43 1.57 0.01

451 98.11 1.71 0.18

690 97.83 2.17 0.14

841 97.70 2.23 0.07

448 95.98 3.35 0.67

541 97.08 1.44 0.84

-

580 96.77 3.03 0.20

Type of zygote Total zygotes Maternal Paternal

Drug

Biparental

Gametic fusiont

Biparental Total zygotes Maternal Paternal

5-azaCyd and L-ethionine

Gametogenesis*

0 696 97.56 2.01 0.43

0.01 745 97.31 2.42 0.27

696 97.95 2.59 0.14

0.03

96.85 3.15 0.10

3.0 1138 96.05 3.95 0.09

10

355 95.59 1.41 0.3

Biparental The data were pooled from three separate experiments. *Drug was added to fresh nitrogen-free medium at the onset of gametogenesis of mt' vegetative cells and kept in the culture during subsequent mating with untreated mt- gametes and zygote development until just prior to germination. tDrug was added to the medium at the onset of gametic fusion and kept in the culture during zygote development until just prior to germination.

Proc. Natl. Acad. Sci. USA 81 (1984)

Genetics: Feng and Chiang Table 2. Effect of 5-azaCyd and L-ethionine on the inheritance pattern of non-Mendelian genes in C. reinhardtii Type of zygote, % Treatment Total mt' mt- zygotes MZ PZ BPZ PZ+BPZ Drug 1735 97.35 2.07 0.58 2.65 5-azaCyd + 3.44 2205 96.56 3.40 0.04 + 1.51 1586 98.49 1.45 0.06 + + 7.68 638 92.32 5.96 1.72 L-Ethio1.26 636 98.74 1.05 0.21 nine + 2.40 745 98.39 2.00 0.40 + 1.17 603 98.83 1.00 0.17 + + 94.55 5.11 0.34 5.45 587 The data were pooled from four separate experiments. Cells were exposed to 0.2 mM 5-azaCyd or L-ethionine at the onset of gametogenesis and kept in the culture during subsequent mating and zygote development. MZ, maternal zygotes; PZ, paternal zygotes; BPZ, biparental zygotes.

stage. An additional 4.0-fold increase was observed at the zygotic stage, resulting in the methylation of nearly half of the total dCyd residues in cpDNA (Table 3). The methylation of cpDNA in mt- vegetative cells was lower than the limit of our HPLC detection limit (i.e., 1-2 m5dCyd residues in 1000 total dCyd). In contrast, mt- gametes contained 43 m5dCyd residues per 1000 total dCyd. Thus, a >20-fold increase in cpDNA methylation also occurred during gametogenesis of mt- vegetative cells (Table 3). 5-azaCyd and L-ethionine exhibit different specificities of inhibition: 5-azaCyd administered during gametogenesis reduced methylation of mt+ gametic cpDNA to one-third but had little inhibitory effect on the methylation of the mtcounterpart. In contrast, L-ethionine appears to have a broader specificity of inhibition and reduced the methylation of the mt- gametic cpDNA more than that of the mt+ cpDNA during gametogenesis (Table 3). The mating-typedependent differential inhibition of 5-azaCyd continued into the zygotic stage. A 66% reduction of cpDNA methylation was detected if the mt+ gametic parents were pretreated with 5-azaCyd during gametogenesis, whereas the same treatment for mt- gametic parents reduced the resulting zyTable 3. Effect of 5-azaCyd and L-ethionine on m5dCyd content in cpDNA at various development stages of C. reinhardtii m'dCyd content*

Mating type

Drug, Vegetative Drug

+

5-azaCyd L-Ethionine -

5-azaCyd

mM

cells

Gametes

Zygotest

0 0.2 0.2 0 0.2 0.2

0.5

12.1 4.0§

47.3

ND: ND** ND

1.3§

16.11 21.2$

4.3

47.3

3.8§ ND§

39.911

L-Ethionine 18.211 ND, not detected. (The detection limit under our experimental conditions is about 0.1-0.2.) *The m5dCyd content values were expressed as the dCyd methylation index-i.e., the percentage of m5dCyd over total dCyd (dCyd + m'dCyd) residues in cpDNA. tcpDNA was obtained from 7-hr zygotes. tCells were exposed to 5-azaCyd for eight doublings. §Drug was added at the onset of gametogenesis. $Zygotes were obtained from the fusion between 0.2 mM drug treated mt+ and untreated mt- gametes. The drug was not removed from the fusion mixture. 1Zygotes were obtained from the fusion between 0.2 mM drug treated mt- and untreated mt+ gametes. The drug was not removed from the fusion mixture. **Cells were exposed to L-ethionine for 24 doublings.

3441

gotic cpDNA methylation by only 16%. This mating-typedependent differential inhibition of zygotic cpDNA methylation was less apparent with L-ethionine (Table 3).

DISCUSSION The original aim of this study was to examine whether the mating type-specific gamete-specific methylation of cpDNA during gametogenesis of mt' cells is responsible for maternal inheritance of chloroplast genes in C. reinhardtii (1, 2). If such extensive methylation of cpDNA as detected originally by Hpa II digestion resistance (1, 2, 4) indeed regulates maternal inheritance of chloroplast genes, then a marked hypomethylation should elicit an equally marked change of the inheritance pattern of the chloroplast genes. The results presented above clearly demonstrate that this is not the case. Since the extent of mating type-specific and gamete-specific methylation that occurs during mt+ gametogenesis is not correlated with the frequency of maternal transmission of chloroplast genes (Figs. 2 and 3 and Table 1), the significance of the observation that gametic cpDNA of the mat-i mutant (3) was methylated to an extent intermediate between wild-type mt- and mt' gametes (2) also becomes obscure. There is no requirement for extensive methylation of the mt' gametic cpDNA during gametogenesis in order to observe maternal inheritance of chloroplast genes in C. reinhardtii. This conclusion is consistent with observations made on the methylation mutant me-i (5). An unexpected finding here on the basis of HPLC was that cpDNA methylation occurred during mt- as well as mt+ gametogenesis. The absence of detectable cpDNA methylation in mt- gametes of a wild-type strain, 5177D, was shown by an antibody binding method (2). Unless the methylation level in mt gametes is substantially different in different mtstrains, it would appear that for methylation HPLC is a more sensitive detector than restriction enzyme digestion or the antibody binding method (2). It is clear that 5-azaCyd and L-ethionine act differently in inhibiting cpDNA methylation during gametogenesis (Table 3). Nevertheless, maternal inheritance occurs in the presence of either or both of these inhibitors (Table 1). These results cast doubt on the functional significance of the extensive cpDNA methylation that occurs during gametogenesis of mt+ cells; after all, gametogenesis, mating, zygote development, and maternal inheritance all proceed virtually unperturbed under the severe hypomethylation conditions of mt+ gametic cpDNA as shown here or under hypermethylation of vegetative cpDNA of either mating type as shown in the me-i mutant (5). Whether a small amount of specific cpDNA methylation unaffected by both 5-azaCyd and Lethionine still occurs preferentially during formation of mt+ gametes, as in the case of me-i (4), is not known at present. Given the difference in mode of action of the two drugs (9), it seems unlikely that the same cytidine residues could remain methylated after both treatments. Two methyl transferase activities of Mrs 200,000 and 60,000 have been identified in whole-cell extracts of wildtype C. reinhardtii (14, 15). It was suggested that the Mr 200,000 activity is specific to gametic cells and regulates maternal inheritance, whereas the other activity (Mr 60,000) in gametes is inactive in vivo (15). However, this is inconsistent with the existing experimental evidence because the mt+ vegetative cells were shown by the same investigators to contain 0.6 unit of the Mr 200,000 species per 109 cells, which represents 14% of the total methyltransferase activities present in these cells (see table 2 of ref. 15). Moreover, our result showing considerable cpDNA methylation in mt- gametes (Table 3) challenges the contention that the smaller methyltransferase was inactive in vivo because it was the only activity found in mt- gametes (15). Thus, a proper appreciation

3442

Genetics: Feng and Chiang

of the biological significance of these two methyltransferases with respect to maternal inheritance must await further research. The site specificity of the cpDNA methylation that occurs during gametogenesis of the mt- and mt' gametes appears to be different (2, 4, 5). A great majority of the internal cytidine in 5' C-C-G-G 3' sequences of the mt' cpDNA became methylated during gametogenesis under our experimental conditions, conferring substantial resistance to Hpa II digestion. In contrast, very few, if any, of the same sequences in the mt- gametic cpDNA were methylated during gametogenesis as indicated by the continued sensitivity of this cpDNA to Hpa II digestion even though 4.3% of the dCyd residues were methylated (Table 3 and Figs. 2 and 3). Moreover, about 33% of cpDNA methylation that occurs regularly during gametogenesis of mt' gametes was not inhibited by 5-azaCyd (Table 3), whereas by the Hpa II/Msp I digestion analysis, the cpDNA of the drug-treated mt+ gametes could not be differentiated from mt+ vegetative cells (Figs. 1 and 2). Approximately two-thirds of the remaining 33% mt+ cpDNA methylation was shown to be inhibited by L-ethionine (Table 3), indicating that L-ethionine has a considerably broader inhibition specificity than that of 5-azaCyd. Taken together, these results suggest that in mt+ gametes the sequence specificity of cpDNA methylation is heterogeneous in nature. Inhibition by 5-azaCyd appeared biased toward the 5' C-C-G-G 3' sequence in mt+ gametes, whereas the virtual absence of methylation of such sequences in the mtgametic cpDNA rendered 5-azaCyd ineffective as a methylation inhibitor during gametogenesis of mt- gametes. However, methylation did occur at other sequences that were sensitive to L-ethionine inhibition (Table 3). Consistent with earlier observations (1, 2), within 7 hr after gametic fusion, a massive additional cpDNA methylation occurred (Table 3). We found, however, that the reported resistance of the resulting zygotic cpDNA to digestion by Msp I (and by inference a few other endonucleases; refs. 1 and 2) is an artifact, which could be eliminated by extensive purification of the DNA (Fig. 4). By both Hpa II/Msp I digestion and HPLC analyses, the amount of early zygotic cpDNA methylation appeared to be dependent on the preexisting methylation of both gametic parental cpDNAs (Figs. 4b and 5 b and c and Table 3). In fact a 2.5-fold difference in inhibition of zygotic cpDNA methylation was detected by HPLC depending on whether mt' or mt- gametogenic methylation was inhibited by 5-azaCyd prior to gametic fusion (Table 3). These results imply that, 7 hr after the gametic fusion, at least some of the cpDNA molecules of both parental origins are still present and play a role in the further methylation of zygotic cpDNA in the young zygotes because, if all cpDNA molecules of paternal origin had been degraded completely by this stage, then a 5-azaCyd pretreatment of either of the parental gametes should not make any difference in the ensuing 5-azaCyd-induced inhibition of cpDNA methylation in the early zygotic stage. Thus, this result is in agreement with our earlier finding that paternal gametic cpDNA molecules were partially conserved and further replicated in zygotes (16). Actually, partial conservation and

Proc. NatL. Acad Sci. USA 81 (1984)

further methylation of paternal gametic cpDNA in zygotes 6 and 24 hr after gametic fusion were also unequivocally demonstrated by the cesium chloride gradient and HPLC profiles contained in an earlier publication (see figures 1 and 3 of ref. 17) but were dismissed as not real by the authors. Due to the repetitive nature of cpDNA molecules, a serious but often neglected dilemma exists in the identification of the molecular mechanism of cpDNA inheritance because only a single properly protected cpDNA molecule is needed for postchloroplast fusion propagation to maintain chloroplast inheritance. The modification of this single genetically competent molecule out of a large population may not be identical to that of the majority, whose properties are determined experimentally. Underscoring this dilemma is the fact that after zygote germination, cpDNA methylation routinely undergoes an '100-fold decrease to the basal level of the vegetative cells (Table 3). Thus, either most of the heavily methylated zygotic cpDNA molecules are not replicated or their methylation sites are not "maintained" in the ensuing development of the 7-hr zygotes (Table 3), suggesting that a possible preferential protective mechanism(s) other than the methylation of the genetically competent mt+ gametic cpDNA molecule still cannot be ruled out. We thank Drs. J. Boynton, E. Harris, and N. Gillham for providing GB68 and CC146 strains, Patricia Dwyer-Hallquist and Kan L. Agarwal for donating Hpa II and R. Haselkorn and Mr. E. Friedman for critical reading of the preliminary form of the manuscript (T.Y.F.'s thesis). This work was supported by National Institutes of Health Grant HD 05804. 1. Royer, H.-D. & Sager, R. (1979) Proc. Natl. Acad. Sci. USA 76, 5794-5798. 2. Sager, R., Grabowy, C. & Sano, H. (1981) Cell 24, 41-47. 3. Sager, R. & Ramanis, Z. (1974) Proc. Natl. Acad. Sci. USA 71, 4698-4702. 4. Sager, R. & Grabowy, C. (1983) Proc. Natl. Acad. Sci. USA 80, 3025-3029. 5. Bolen, P. L., Grant, D. M., Swinton, D., Boynton, J. E. & Gillham, N. W. (1982) Cell 28, 335-343. 6. Sager, R. & Lane, D. (1972) Proc. Natl. Acad. Sci. USA 69, 2410-2413. 7. Chiang, K.-S., Kates, J. R., Jones, R. F. & Sueoka, N. (1970) Dev. Biol. 22, 655-669. 8. Eves, E. & Chiang, K. S. (1982) Genetics 100, 35-60. 9. Jones, P, A. & Taylor, S. M. (1981) Nucleic Acids Res. 9,

2933-2947. 10. Grant, D. M., Gillham, N. W. & Boynton, J. E. (1980) Proc. Natl. Acad. Sci. USA 77, 6067-6071. 11. Grant, D. M. & Chiang, K.-S. (1980) Plasmid 4, 82-96. 12. Feng, T.-Y. (1981) Dissertation (University of Chicago, Chicago). 13. Singhal, R. P. (1974) Sep. Purif. Methods 3, 339-398. 14. Sano, H. & Sager, R. (1980) Eur. J. Biochem. 105, 471-480. 15. Sano, H., Grabowy, C. & Sager, R. (1981) Proc. Natl. Acad. Sci. USA 78, 3118-3122. 16. Chiang, K.-S. (1976) in Genetics and Biogenesis ofChloroplast and Mitochondria, eds. Bucher, T., Neupert, W., Sebald, W. & Werner, S. (North-Holland, Amsterdam), pp. 305-312. 17. Burton, W. G., Grabowy, C. T. & Sager, R. (1979) Proc. Natl. Acad. Sci. USA 76, 1390-1394.