Volume 6 Number 6 1979 Volume 6 Number 6 1979

Research Nucleic Nucleic Acids Acids Research

Propagation of restriction fragments from the mitochondrial DNA of Saccharomyces cerevisiae in E. coli by means of plasmid vectors

Patricia E.Berg*, Alfred Lewin+, Thomas Christianson and Murray Rabinowitz

Departments of Medicine, Biochemistry, and Biology, University of Chicago, and The Franklin McLean Memorial Research Institute, Chicago, IL 60637, USA

Received 24 January 1979

ABSTRACT Some of the EcoRI fragments of yeast (Saccharomyces cerevisiae) mitochondrial DNA were cloned into E. coli using plasmid pMB9. The five smallest fragments in molecular weight appeared to be preferentially retained by E. coli; partial fragments derived from larger mitochondrial DNA fragments were also found. One of the fragments, R7 (2.4 kb), may contain the OIJ gene. Cloned R7 DNA was stable under a variety of growth conditions, but showed some changes in molecular weight after transfer to different E. coli strains. Fragment R7 is transcribed in minicells, producing RNA that hybridizes specifically to mitochondrial DNA. Both DNA strands are transcribed, in contrast to the asymmetric transcription found in mitochondria. No new polypeptides were observed in minicells containing cloned fragment 7.

INTRODUCTION Most studies of mitochondrial genetics have been carried out with the yeast Saccharomyces cerevisiae.1-4 The capacity of yeast to survive and grow in the absence of aerobic respiration has facilitated the isolation of mutations which affect mitochondrial respiratory function and mitochondrial protein synthesis. While the genetic analysis of the mitochondrial genome of Saccharomyces cerevisiae is convenient, the mitochondrial DNA (mtDNA) presents several obstacles for molecular biology. First, yeast mtDNA is never isolated intact. Although the 25-micron circular DNA molecules observed in yeast lysates correspond to the estimated mitochondrial genome size,5 yeast mtDNA is isolated as randomly broken fragments corresponding in length from one-third to onehalf of the total genome size.6'7 Furthermore, this DNA contains numerous single-stranded scissions (nicks) as well as single-stranded regions (gaps), so that after denaturation of mtDNA, fragments with a molecular weight of approximately 3 kb remain. This degradation of isolated mtDNA has made restriction mapping of yeast mtDNA difficult,7'8 as well as complicated the use

C) Information Retrieval Limited 1 Falconberg Court London Wl V 5FG England

2133

Nucleic Acids Research of other high-resolution mapping techniques (e.g., heteroduplex analysis, "R-loop" mapping, DNA sequencing). Cytoplasmic petites which contain amplifications of localized regions of the mitochondrial genome may be regarded as a "natural" cloning vector for mtDNA; however, the mtDNA sequences retained are often highly heterogeneous,9'10 and may contain mtDNA segments with different redundancies.11'12 Furthermore, petite mutants are generally not stable, but continue to delete markers and mtDNA sequences.13 Because of these characteristic properties of grande and petite mtDNA, we have cloned segments of the yeast mtDNA in E. coli K12 by using plasmid vectors. In contrast to cytoplasmic petites, this has provided stable, homogeneous amplification of short sequences of mtDNA which are suitable for analysis. Although the largest restriction fragment from grande mtDNA which we have cloned is 3.5 kb, it seems likely that more of the mitochondrial genome can be propagated in E. coli if alternate hosts and vectors as well as multiple endonuclease digestions are used. In addition to the analysis of mtDNA, we have examined the feasibility of studying the transcription and translation of this DNA in E. coli minicells. We give particular attention to the difficulties of cloning yeast mtDNA, since other mitochondrial DNA has been cloned with relative ease.14,15,16

MATERIALS AND METHODS Bacterial strains and plasmids. Strain C600 (r-m ) was obtained from S. Cohen. Roy Curtiss III provided the minicell strain x984, and H. Adler, lon min. Plasmid pMB9 was a gift from I. Dawid and pBR322, from H. Boyer. The former plasmid contains the gene for tetracycline resistance17,18 and the latter codes for both tetracycline and ampicillin resistance.19 DNA and RNA preparation. Plasmid DNA was isolated and purified as described by Berg et al.20 Yeast protoplasts and mitochondria were prepared according to the method of Grivell et al.,21 as modified by Hendler et al.22 DNA was prepared from isolated mitochondria as described by Lewin et al.10 Mitochondrial RNA was isolated from lysed mitochondria by a phenol-chloroformisoamyl alcohol extraction procedure.23 The DNA contamination of the RNA was removed by CsCl guanidium chloride equilibrium centrifugation.24 DNA was not detected by the DABA reaction.25 Copy RNA was synthesized using E. coli RNA polymerase as previously described.26 Construction of recombinant plasmid DNA molecules. Purified plasmid DNA and mtDNA were cut with the restriction endonuclease EcoRI or with HindIII/ EcoRI, mixed in a 2:1 molar ratio of mtDNA:plasmid DNA, and then ligated with 2134

Nucleic Acids Research T4 DNA ligase.27 In some experiments, we pretreated the cloning vehicle with bacterial alkaline phosphatase (BAP) to remove 5' terminal phosphates and to prevent recirculation of the plasmid.28 The BAP was obtained from Worthington (BAP-F). It was further treated before use for removal of nucleases29 and dialyzed for removal of ammonium sulfate, which inhibits transformation into bacteria (K. Agarwal, personal conmmunication). Bacterial cells were made competent for transformation by the procedure of Lederberg and Cohen.30 After transformation, cells were grown in L broth containing 10% glycerol and 2.5 ig/ml tetracycline to induce tetracycline resistance. Selection was on plates containing either 12.5 or 25 ig/ml tetracycline. Colonies were screened by means of the colony hybridization technique described by Grunstein and Hogness.31 32P-cRNA was made against either grande mtDNA or various petite mtDNAs for use as the probe. Colonies showing hybridization (positive colonies) were treated in one of two ways. Some were grown in 15 ml cultures with chloramphenicol amplification. The cells were lysed as described in the section on DNA isolation, phenol-extracted, dialyzed, and subsequently analyzed by agarose gel electrophoresis. In other cases, a rapid screening method was used which was a modification of Telford's technique32 of direct electrophoresis of colony lysates. Incorporation into minicell RNA and protein. Minicells were purified according to a modification of Levy's method33 in which two 5-20% sucrose gradients are used before penicillin treatment and one 5-20% sucrose gradient after treatment. For RNA studies, isolated minicells were preincubated for 20 minutes at 370 in M9 minimal medium,34 supplemented with adenine and 0.5% casamino acids. Next, 5-3H-uracil (26 Ci/mmole), to a final concentration of 200 pCi/ml was added and the minicells incubated for 20 minutes. The suspension was chilled and the RNA extracted as described by Roozen et al , except that 3 minutes in boiling water were used instead of 15 minutes at room temperature. We had found that the heat treatment gave better minicell lysis. To determine the polypeptides synthesized in minicells, we preincubated purified minicells in the presence of methionine for 30 minutes at 370, then centrifuged and resuspended them in methionine-free assay medium.33 We added 35S-met (1245 Ci/mmole) to 1.0 OD620 unit of minicells to obtain a final concentration of 100 pCi/ml. The incubation time was 30 minutes. The minicells were chilled, washed with BSG buffer,36 frozen in ethanol-dry ice for storage at -200C, resuspended in electrophoresis stacking-gel buffer (0.14 M Tris, pH 6.8, 5% SDS, and 0.38 M I-mercaptoethanol), and boiled 3 to 5 minutes prior 2135

Nucleic Acids Research to loading on an SDS-polyacrylamide gel. Approximately 25,000 hot-TCAprecipitable cpm were loaded per gel slot. Electrophoresis. For DNA, agarose gel electrophoresis was performed as previously described. 7 For proteins, polyacrylamide gel electrophoresis was carried out according to the conditions of Laemmli,37 except that 15% acrylamide was used as provided by the procedure of Blattler et al.38 Gels were fixed and stained with 50% TCA, 5% methanol, and 0.1% Coomassie blue, and destained in 7.5% acetic acid, 5% methanol. The fluorography technique of Bonner and Laskey39 was used for detection of labeled peptides. Hybridization. Denatured DNA (20 ,ug) was loaded on 47-mm-diameter nitrocellulose filters (Schleicher and Schuell Bac-T Flex, type B6). Quarter filters (5 pg of DNA) were used for hybridization. The filter hybridizations40 were performed in 2 x SSC (SSC is 0.15 M NaCl, 0.015 M sodium citrate, pH 7.0) and 36% formamide at 35°C for 16 hours. Labeled minicell RNA was hybridized to total mitochondrial RNA at a R0t of 1200 mole-sec 1-1. 3H-minicell RNA (20,000 cpm) was mixed with 100 pg of total mtRNA in a final volume of 15 ,ul. The RNA mixture was heated at 100IC for 5 minutes, and then the samples were rapidly cooled. Samples to be hybridized were placed at 600C for 24 hours, whereas control samples were frozen. Half of each sample was precipitated and counted, and the other half was treated with Tl RNAse (3 pg/ml) and pancreatic RNAse (20 ig/ml) for 30 minutes at 370C. After the RNAse treatment, the samples were precipitated and counted. The control counts (cpm of unhybridized samples) were subtracted from the counts of samples which were hybridized. Physical and biological containment conditions. The above experiments were performed in a P2 facility with EKl hosts, as required under the NIH

guidelines. RESULTS We have stably cloned five of the ten EcoRI fragments of yeast mtDNA from Saccharomyces cerevisiae. These fragments, designated R6, R7, R8, R9, and RIO, have molecular sizes of 3.5 kb, 2.4 kb, 1.7 kb, 0.9 kb, and 0.2 kb, respectively (Fig. 1). Fragments R7, R8, R9, and R10 were cloned from strain MH41, while R6, R7, R8, R9, and R10 were cloned from strain 19D. The fragment numbers used are those given to MH41. According to current physical and genetic mapping data, these two strains are very similar or identical in these regions.4' The fragments which are cloned (Fig. 1) are not clustered on the 2136

Nucleic Acids Research \.~22-J

EcoRI Bond

tFi'A,41

KB

I@9 > > 2

23.7 2 7.3 3 10.0

OXI-3

254 ~~~~~~~~78*

\

{;/

~~~~~~ 0 ~ ~

78 ~~~~~5 ~

Ntz

/ l

~

~

~

~

9

X

Figure 1. Physical and genetic map of yeast mtDNA from Saccharomyces cerevisiae strain MH41-7B. The cloned EcoRI fragments are designated by double lines, and on the right by asterisks. Genetic markers are as previously described.26

genome. Although the comigration of two EcoRI fragments is presumptive evidence for their identity, it is important to establish sequence homology by other criteria. Therefore, DNA from an EcoRI digest of mtDNA was transferred to nitrocellulose according to the technique of Southern.42 The transferred DNA was hybridized to cRNA made with the chimeric plasmids serving as templates. As the autoradiograms in Figure 2 show, cRNA from cloned EcoRI fragments R6-R9 does hybridize to the corresponding mtDNA fragments. This establishes that we have cloned yeast mtDNA corresponding in molecular size to known EcoRI fragments of mtDNA, but does not eliminate the possibility of rearrangements. Tracings of agarose gels of EcoRI-cleaved mtDNA have shown that the lower molecular weight fragments are present in molar ratios; higher molecular weight fragments are present in submolar ratios since the mtDNA is not intact when isolated.4l143 Therefore, one would expect that R6, R7, R8, R9, and R10 would have equal probabilities of being cloned. This was not the case, however (Table 1). R10 was excluded from this analysis because its detection by colony hybridization is less reliable due to its low molecular weight. Fragment R8 accounted for more than half of the clones and R9 for most of the 2137

Nucleic Acids Research Figure 2. Hybridization of cRNA from chimeric plasmids to mtDNA. 32P-cRNA made against the hybrid plasmids was hybridized to EcoRI digests of mtDNA. The DNA was transferred to nitrocellulose by the method of Southern.42 Number 1 is pBLT727, an R6 clone; 2 is pBLT778, an R7 clone; 3 is pBLT715, an R8 clone; and 4 is pBLT966, an R9 clone.

rest; R6 and R7 were inserted at much lower frequencies. Several cloned segments of mtDNA retained both EcoRI ends, as shown by the fact that pMB9 could be recovered, but the cloned DNA did not correspond in molecular weight to any mitochondrial EcoRI fragments. Several other mtDNA clones retained only one EcoRI site, giving rise to a single fragment larger than pMB9 (Fig. 3). A number of these plasmids, including pBLTB117, pBLTB8107, and pBLTB7350, were analyzed in more detail. MH41-7B mtDNA was digested with EcoRI and two other restriction enzymes, HhaI and XbaI, which were selected for their overlapping fragments. These Table 1. Frequency of occurrence of cloned EcoRI fragments. Of 68 mtDNA clones which were investigated in detail, 7 were anomalous and 61 were identified as EcoRI fragments. The three R10 clones are exclude because of the difficulty of detecting all RIO clones by Grunstein-Hogness colony hybridization. EcoRI fragment R6 R7 R8 R9

2138

Frequency of Occurrence (%) (number)

2 5 35 16

3.4 8.6 60.3 27.6

Nucleic Acids Research Figure 3. Electrophoresis of EcoRI digests of cloned yeast mtDNA on 1% agarose gels. Slots 3-6 contain EcoRI digests of four cloned DNAs; these fragments comigrate with EcoRI digests of the MH41 mtDNA, slot 2. Slots 9-11 are three examples of cloned mtDNA fragments which do not comigrate with mtDNA EcoRI fragments. In 9, only one fragment larger than pMB9 appears, while 10 and 11 show anomolously migrating fragments. The bottom diagram suggests how these non-comigrating fragments might be generated. The single band larger than pMB9 could arise from a deletion spanning one restriction site, and the anomolous fragments could arise from a deletion wholly within the inserted fragment.

digests were run on 1% agarose gels and transferred to nitrocellulose strips, as described by Southern.42 3H-cRNA made to the chimeric plasmids was hybridized to this DNA. The plasmid pBLTB117 proved to have both ends of EcoRI fragment 1, and pBLTB8107 and pBLTB7350 (Fig. 4) had a single end of EcoRI fragments 5 and 3, respectively. Physical mapping data indicated that the gene oliII maps in the vicinity of R7.9,26'44 For this reason, further studies were performed with plasmid pBLT778, which contains R7. Stability of inserted DNAs. It was important to verify the stability of R7 in pBLT778 before we undertook additional studies. Stability was assayed in two strains under a variety of conditions. The original recipient, C600, was grown in complex medium for up to 23 generations before chloramphenicol was added for plasmid amplification. Plasmid DNA was isolated, cleaved with EcoRI, and run on 1.0% agarose gels. A microdensitometer tracing of the gel was made, and the molar ratio of R7 to pMB9 DNA was calculated relative to the molar ratio of R7 to pMB9 at three generations (zero time). As shown in Table 2, no significant change in the amount of R7 was detected when this method was used. Since we planned to study transcription and translation products of R7

2139

Nucleic Acids Research

~.

:'. C

:.: x

].'.

.

..

.U,

...

'.%

.-

.,..:

'..' :..L.

"

....

....

..

... .......

'

........ ........ ................ .. -X

S........ .... ....

Figure 4. Hybridization of 3H-cRNA synthesized from two cloned yeast mtDNAs, pBLTB8107 and pBLTB7350. These two clones exhibit only one fragment, larger than pMB9, after EcoRI digestion. The cRNA probe was hybridized to Southern transfers42 of EcoRI, HhaI, and XbaI digests of MH41 mtDNA. On the left, probe from pBLTB8107 hybridizes to fragments 5 of EcoRI, 2 of HhaI, and 1 of XbaI. On the right, probe from pBLTB7350 hybridizes to fragments 3 of EcoRI, HhaI, and XbaI. Below is a restriction map26 for these enzymes, with shading illustrating fragments with hybridization to respective clones.

Table 2.

Molar ratios of cloned fragment 7 DNA.

Generationsb Strain

Medium

CAMa

3

13

23

C600

Complex Minimal

+

1.0 1.0 1.0 1.0 1.0

1.3 1.1 1.1 1.0 1.0

1.0 1.0

X984

+

Complex

+

aChloramphenicol

1.2 1.1 1.2

addition is indicated by (+). of growth before DNA isolation or before addition of chloramphenicol.

bNumber of generations 2140

Nucleic Acids Research in minicells, it was important to test the stability of pBLT778 in x984, a minicell-producing strain. Cells were grown as above, in either minimal or complex medium with or without the addition of chloramphenicol. Apparently the inserted fragment is stable under all conditions (Table 2). Another sensitive method of detecting loss of cloned DNA is that of cutting pBLT778 with the restriction enzyme BamHI, which makes one cut in pMB9 and none in R7.43 Any change in the molecular weight of the chimeric plasmid would be easily detected by the appearance of a DNA band with altered mobility during agarose gel electrophoresis. Such experiments showed no change in the molecular weight of pBLT778 for cells grown for 23 generations in the complex medium whether treated or not treated with chloramphenicol (data not shown). To test the nature of R7 further, we redigested R7 purified either from a mtDNA digest or from the plasmid pBLT778 (Fig. 5) by using HpaII. Both DNAs migrated to the same position on the gel, indicating that the recognition site for HpaII is in the same place. However, we observed heterogeneity in the size of several cloned fragments when they were initially transferred to x984. The plasmid DNA was isolated from C600, which lacks the E. coli K12 restriction and modification system and therefore was unmodified. This DNA was used to transform x984, which has the E. coli K12 restriction-modification system. The number of

.R,r4i'

Figure 5. Digestion of purified R7 mtDNA with EcoRI and HpaII. Samples were visualized using 2% agarose gel electrophoresis. In slot 1 is R7 DNA isolated from pBLT778; slot 2 contains R7 DNA from mitochondria; and slot 3 contains total mtDNA.

2141

Nucleic Acids Research transformants in X984 was reduced by three orders of magnitude compared with C600, similar to the observation of Chang et al.'4 When two colonies were picked at random from cells transformed with plasmids containing R6, R7, R8, or R9, only one containing R8 and both containing R9 were transferred to give their original molecular weight (Fig. 6). Additional colonies were tested until one that had a band with the correct molecular weight was found for each fragment. The plasmid containing an EcoRI fragment corresponding to EcoRI R7, pBLT778, was the colony which we used to test long-term stability (Table 2). Thus, when intact R7 was successfully transferred, it appeared to be quite stable. We attribute the initial transfer changes to the restriction system of x984. Another minicell strain, lon min, was used as a recipient for pBLT778 DNA grown in X984 (r m+). Of 13 colonies tested, none exhibited the correct molecular weight for R7 (data not shown). This necessitated a more detailed study of the transfer stability of R7. PBLT778 plasmid DNA isolated from X984 was used to transform both C600 (r-m ) and x984. Ten clones were grown from each transformation, and in each case R7 had the correct molecular weight. Some possible reasons for the instability of R7 in lon min are discussed below. Cloned R7 thus appears to be the same as the original R7 with respect to two restriction enzymes, EcoRI and HpaII, shares sequence homology with R7, and is stable under the conditions which we used for long-term growth. It may be unstable upon transfer, depending on the bacterial recipient strain. Once in the cell, however, it appears to be stable. With this knowledge, we studied the RNA and protein made by pBLT778 in minicells of x984. Transcription of R7 in minicells. Minicells are small, anucleate cells

-w S

2142

Figure 6. Transfer by transformation of chimeric plasmids which contain mtDNA fragments R6-R9 from C600 (rRmR) to X984 (rtmi). DNA was isolated from two X 84 clones after each transformation, digested with EcoRI, and visualized by 1% agarose gel electrophoresis. In this sample, R6 failed to transfer correctly (slots 4, 5), one R7 (slot 7) exhibited several bands including one comigrating with fragment 7, one R8 (slot 9) exhibited a band of submolar intensity comigrating with fragment 8, and R9 transferred correctly in both cases (slots 10, 11). Slots 1, 2, and 3 and 12 contain EcoRI digests respectively of X, pMB9, and MH41 mtDNA.

Nucleic Acids Research which result from abnormal cell division in certain E. coli mutants. They contain no chromosomal DNA, but plasmid DNA segregates into them. Since they contain the enzymes necessary for RNA and protein synthesis, they make an ideal system for the study of transcription and translation products of plasmids.45 As described above, pBLT778 DNA was used to transform x984. Minicell RNA was labeled with 3H-uracil and isolated as described in Materials and Methods. We hybridized this RNA to 5 pg of DNA on filters to determine whether the mtDNA was being transcribed in vivo. Table 3 shows that R7 DNA is indeed transcribed, since it hybridizes to mtDNA but not to yeast nuclear DNA from a p0 petite. The RNA also hybridized to pMB9 DNA, as expected. We approached the problem of strand-specific transcription by hybridizing minicell RNA to total mitochondrial RNA. We have previously obtained evidence that at high R0t values more than 60% of a single strand equivalent of mtDNA is transcribed. 46,47 Furthermore, total mtDNA hybridizes exclusively and almost completely to one separated strand of an R7 mtDNA-pMB9 recombinant molecule. Asymmetric transcription is supported by our observation that a transcription complex of yeast mitochondrial DNA and mitochondrial RNA polymerase synthesizes RNA asymmetrically.48 If the strand transcribed in minicells is the same as that transcribed in mitochondria, no significant hybridization should occur with mitochondrial RNA. If both strands, or predominantly the "nonsense" strand, are being transcribed, significant hybridization should be observed. As shown in Table 4, about 19% of the RNA made in minicells hybridizes to mitochondrial RNA; thus there is significant transcription of the "nonsense" strand in R7 in minicells. The fact that about 27% of mitochondrial cRNA hybridized to mitochondrial RNA indicates that under the conditions used, E. coli RNA

Table 3. Hybridization of RNA from minicells containing pBLT778 or pMB9 to DNA on filters. 20,000 input counts were hybridized to filters containing 5 pg of DNA (DNA was in excess).

DNA 19D (mtDNA) p0

pMB9

Hybridization (cpm) pBLT778 RNA pMB9 RNA

3,509 88 10,044

47 78

9,368

2143

Nucleic Acids Research Hybridization of RNA to total mitochondrial RNA.

Table 4. RNA

pBLT778 (minicell) mt cRNA

Hybridization (%)a 19 27

aMinicell

RNA from cells containing pBLT778 was hybridized to mtRNA. cRNA was made to mtDNA with E. coli RNA polymerase. Input was 20,000 cpm. The background (RNA alone) of 5% was subtracted from the total hybridization.

polymerase in vitro transcribes both strands of the mtDNA. Translation of R7 in minicells. Proteins from minicells containing pMB9 and pBLT778 were labeled with 35S-methionine. Labeled proteins were displayed on a 15% SDS-polyacrylamide gel which was treated for fluorography. The results, seen in Fig. 7, show no new proteins in minicell extracts from pBLT778.

DISCUSSION The cloning of yeast mtDNA was found to be difficult. Experiments in which EcoRI fragments were cloned by use of pMB9, when pMB9 was either treated or not treated with bacterial alkaline phosphatase to prevent recirculation, yielded the five smallest EcoRI fragments. These fragments comprise 11.5% of

.... *~~~~~~~~~~~~~~o ...: _

-W

t- s@

I..

......

2144

.....

Figure 7. Autoradiogram of polypeptide profile from minicells containing pMB9 or pBLT778. Analysis was performed on a 15% SDSpolyacrylamide gel. 25,000 hot TCAprecipitable cpm were loaded in both wells.

Nucleic Acids Research the total genome. Cloning of HindIII/EcoRI fragments by use of pBR322 also did not yield any intact fragments from other regions of the genome (A. Lewin, unpublished observation). This indicates a preferential selection for certain fragments by E coli. The basis of this selection could be the GC content, since E. coli is 49% GC49 whereas yeast mtDNA is only 17% GC.50 It is unlikely that the size of the EcoRI fragments, per se, was the limiting factor. Brown et al.15 cloned the entire mouse mitochondrial genome of approximately 15 kb, and Miller et al.51 cloned Drosophila DNA of 12.75 kb. Yeast mitochondrial EcoRI fragments within this size range4l include R5, R4, and R3 (3.5 kb, 7.8 kb, and 10 kb, respectively). Probably a combination of molecular size and high AT content reduces the capacity of certain fragments to be cloned. It is also known that there are single-strand regions (nicks and/or gaps) in yeast mtDNA, which may affect the capacity of the larger fragments to be cloned. 52 In yeast, mtDNA is frequently deleted spontaneously, resulting in the formation of cytoplasmic petite mutants.9'53 This inherent instability of yeast mtDNA may be attributable in part to the nonrandom distribution of AT "spacer" regions throughout the mitochondrial genome.50 Because of these characteristics of yeast mtDNA and the particular problems of cloning of this DNA, we investigated the stability of our cloned DNA. Long-term growth of cloned R7 in two E. coli strains, whether in complex or minimal medium, amplified or unamplified, showed that this fragment was quite stable (Table 2). Experiments using BamHI, which cuts the joint plasmid once, were in agreement with these results. One additional factor regarding stability of cloned DNA should be considered, the recA genotype of the E. coli recipient. There is good evidence that spontaneous recombination causing deletion of all or part of cloned eucaryotic DNA sequences is independent of the recA character of the host cell. When Chang et al.14 transferred cloned mouse mitochondrial DNA from strain C600 (r+m+ recA+) to strain HB101 (rjm- recA ), they observed extensive recombination. Similarly, Hershey et al.58 found that cloned Drosophila melanogaster 5S DNA was unstable in strain HB101 and concluded that excision of tandemly repeated 5S units of DNA from chimeric plasmids was recA independent. Brutlag et al.59 observed marked instability of cloned D. melanogaster simple sequence satellite DNA in a recA host. However, Carroll and Brown60 found that cloned tandem repeats of the 5S DNA from Xenopus laevis were stable in a recA_ host. In contrast, we have observed that fragment R7 was stable after long term growth in two recA+ hosts, although some other 2145

Nucleic Acids Research EcoRI fragments were not stably cloned. Thus, there is increasing evidence that "illegitimate" recombination, which is recA independent, can cause gross instability of cloned DNA (for reviews of illegitimate recombination, see Saedler6l and Kleckner62). In our study, a recA+ strain was used in most experiments. However, internal deletion of inserted fragments was also observed when we used a recA strain (HB101 r m recA ). It is of note that Bernardi63 has shown that spontaneous deletion of yeast mtDNA appears to be caused by illegitimate recombination at high GC clusters interspersed within the 95% AT "spacer" sequences that comprise about 50% of yeast mtDNA. The transfer of pBLT778 from C600 (r-m ) to X984 (r+m ) usually resulted in changes in the molecular weight of R7, i.e., it was not stably transferred (Fig. 6). Cloned R7 DNA was stably transferred from X984 to either C600 or X984 (data not shown). In this case, then, the changes were probably due to cutting of the unmodified DNA by the restriction system of x984. R7 DNA which has been modified by X984 is not stably transferred to lon min, the minicell strain that harbors the lon (capR) mutation.. The reason for the selective instability of R7 was not investigated (the plasmid alone, pMB9, transferred intact), but two reasons can be suggested. Two groups of investigators have observed transfer instability into strain P678-54; Chang et al.14 used cloned mouse mtDNA and Miller et al.51 used cloned Drosophila DNA. The fact that lon min is a capR6 (lon) derivative of P678-5454 could account for our observations. In addition, since mutations in the lon gene may cause derepression of a nuclease,55 this nuclease could preferentially recognize some sequence or sequences present in R7, but not in pMB9. The use of lon min may still prove valuable in some cases, since it should have fewer proteases due to the lon mutation.56'57 The protein profile of pMB9 in lon min had fewer peptides than pMB9 in X984. Those peptides present, however, had similar mobilities (P. Berg, unpublished data). We can conclude that, due to the apparent dependence of transfer stability on the recipient strain, it is necessary to test each cloned DNA when it is transferred to a new strain. We detected transcription of cloned R7 in minicells (Table 3), in agreement with transcriptional studies of other cloned eukaryotic DNAs.14,51,64,65 Our experiments do not differentiate between transcriptional read-through for a pMB9 promoter, non-specific initiation on R7, or specific initiation on R7. The evidence tends to rule out the possibility of specific initiation at a true mitochondrial promoter, since RNA from minicells containing pBLT778 2146

Nucleic Acids Research hybridizes to mitochondrial RNA from yeast. The amount of this hybridization was probably underestimated because total pBLT778 RNA was used, and R7 DNA accounts for only about one-third of the total plasmid DNA. If minicell transcription is equal over the plasmid, up to 60% of the R7 minicell RNA may hybridize to mitochondrial RNA. This would mean transcription is occurring about equally from both DNA strands in minicells. In contrast, yeast mtDNA is apparently transcribed asymmetrically in vivo)46 No new polypeptides were produced in X984 minicells containing pBLT778. Although new polypeptides have been observed for several cloned eucaryotic DNAs, including mouse mtDNA, cauliflower mosaic virus DNA, and Drosophila DNA,14.66,67 thus far functional expression has been detected for cloned yeast DNA68-70 and catabolic dehydroquinase of Neurospora crassa.71 In these cases, chimeric plasmids were used to complement host mutations of E. coli. Synthesis of an ovalbumin-like protein, of somatostatin, and of proinsulin by E. coli has been accomplished by fusion of the genes coding for these peptides into suitable E. coli genes.72-74 Meagher et al.66 contend that cloning of DNA into the EcoRI site of pMB9 does not alter the polypeptide profile directed by this plasmid in minicells. We also observed no change in the polypeptide profile after cloning into the EcoRI site of pMB9, although the actual polypeptide pattern which we observed for pMB9 proteins differs significantly from that of Meagher et al. One possible explanation for this difference is the use of different minicell strains. There may well be a significant difference in proteolysis in these different strains. In a preliminary experiment with lon min/pMB9, which expresses normal tetracycline resistance, we no longer observed some of the minor bands present in extracts from X984/pMB9. It would be useful to compare the conditions affecting the number and molecular weights of the protein bands made in minicells, since a discrepancy was also observed between Miller et al.51 and Chang et al.l4 for polypeptides coded by pSC101.

ACKNOWLEDGEMENTS Research was conducted under the Departments of Medicine, Biochemistry, and Biology, The University of Chicago, and The Franklin McLean Memorial Research Institute (operated by The University of Chicago for the U. S. Department of Energy under Contract No. EY-76-C-02-0069), Chicago, Illinois 60637. This study was supported in part by Grants HL09172 and HL04442 from the National Institutes of Health, U. S. Public Health Service, and the Louis Block Fund of The University of Chicago. A.L. was supported by USPHS Training Grant 5-TOl-GMOOO90, and T.C. by USPHS Training Grant T32-GM7197. 2147

Nucleic Acids Research *Current address: Bethesda Research Laboratories Institute of Molecular Biology, Rockville, Maryland 20850.

tCurrent

address: C/O Dr. G. Schatz, Biozentrum der Universitat Basel, CH4056, Basel, Khugelbergstrasse 70, Switzerland.

REFERENCES

1. 2. 3. 4. 5.

6. 7. 8. 9.

10.

11. 12. 13.

14. 15. 16. 17.

18. 19. 20. 21.

2148

Giliham, N.W. (1974) Ann. Rev. Genet. 8, 347-391. Saccone, C. and Kroon, A.M., Eds. (1976) The Genetic Function of Mitochondrial DNA. Elsevier/North-Holland, Amsterdam. Bucher, 1., Neupert, W., Sebald, W. and Werner, S., Eds. (1976) Genetics and Biogenesis of Chloroplasts and Mitochondria. Elsevier/North-Holland, Amsterdam. Bandlow, W., Schweyen, R.J., Wolf, K. and Kaudewitz, F., Eds. (1977) Mitochondria 1977: Genetics and Biogenesis of Mitochondria. Walter de Gruyter, New York. Hollenberg, C.P., Borst, P. and van Bruggen, E.F.J. (1970) Biochim. Biophys. Acta 209, 1-15. Locker, J., Rabinowitz, M. and Getz, G.S. (1974) J. Mol. Biol. 88, 489502. Morimoto, R., Lewin, A. and Rabinowitz, M. (1977) Nucl. Acid Res. 7, 2331-2351. Sanders, J.P.M., Heyting, C., Verbeet, M.P., Meijlink, F.C.P.W. and Borst, P. (1977) Molec. gen. Genet. 157, 239-261. Lewin, A., Morimoto, R., Merten, S., Martin, N.C., Berg, P., Christianson, T., Levens, D. and Rabinowitz, M. (1977) in Mitochondria 1977: Genetics and Biogenesis of Mitochondria, Bandlow, W., Schweyen, R.J., Wolf, K. and Kaudewitz, F., Eds., pp. 272-289. Walter de Gruyter, New York. Lewin, A., Morimoto, R., Rabinowitz, M. and Fukuhara, H. (1978) Molec. gen. Genet. 163, 257-275. Gordon, P., Casey, J. and Rabinowitz, M. (1974) Biochemistry 13, 10671075. Casey, J., Gordon, P. and Rabinowitz, M. (1974) Biochemistry 13, 10591067. Faye, G., Fukuhara, H., Grandchamp, C., Lazowska, J., Michel, F., Casey, J., Getz, G.S., Locker, J., Rabinowitz, M., Bolotin-Fukuhara, M., Coen, D., Deutsch, J., Dujon, B., Netter, P. and Slominsky, P.P. (1973) Biochimie 55, 779-792. Chang, A.C.Y., Lansman, R.A., Clayton, D.A. and Cohen, S.N. (1975) Cell 6, 231-244. Brown, W.M., Watson, R.M., Vinograd, J., Tait, K.M., Boyer, H.W. and Goodman, H.M. (1976) Cell 7, 517-530. Miller, D.L., Andreone, T.L. and Donelson, J.E. (1977) Biochim. Biophys. Acta 477, 323-333. Cohen, S.N., Chang, A.C.Y., Boyer, H.W. and Helling, R.B. (1973) Proc. Nat. Acad. Sci. 70, 3240-3244. Cohen, S.N. and Chang, A.C.Y. (1977) J. Bacteriol. 132, 734-737. Bolivar, F., Rodriguez, R.C., Greene, P.J., Betlach, M.C., Heyneker, H.L., Boyer, H.W., Crosa, J.H. and Falkow, S. (1977) Gene 2, 95-113. Berg, P.E., Gayda, R., Avni, H., Zehnhauer, B. and Markovitz, A. (1976) Proc. Nat. Acad. Sci. 73, 697-701. Grivell, L.A., Reijnders, L. and Borst, P. (1971) Biochim. Biophys. Acta 247, 91-103.

Nucleic Acids Research 22.

23. 24. 25. 26. 27.

28. 29. 30. 31.

32. 33. 34. 35. 36. 37. 38. 39. 40. 41.

42. 43. 44. 45. 46.

47. 48.

49. 50. 51.

Hendler, F.J., Padmanaban, G., Patzer, J., Ryan, R. and Rabinowitz, M. (1975) Nature 258, 357-359. Hirsch, M. and Penman, S.J. (1973) J. Mol. Biol. 80, 379-391. Enea, V. and Zinder, N.D. (1975) Science 190, 584-586. Kapp, U.N., Brown, S.L. and Klevecz, R.R. (1974) Biochim. Biophys. Acta 361, 140-143. Morimoto, R., Merten, S., Lewin, A., Martin, N. and Rabinowitz, M. (1978) Molec. gen. Genet. 163, 241-255. Weiss, B., Jacquemin-Sablon, A., Live, T.R., Fareed, G.C. and Richardson, C.C. (1968) J. Biol. Chem. 243, 4543-4555. Ullrich, A., Shine, J., Chirgwin, J., Pictet, R., Tischer, B., Rutler, W.J. and Goodman, H.M. (1977) Science 196, 1313-1319. Efstratiadis, A., Vournakis, J.N., Donis-Keller, H., Chaconas, G., Dougall, D.K. and Kafatos, F.C. (1977) Nucl. Acid Res. 4, 4165-4174. Lederberg, E.M. and Cohen, S.N. (1974) J. Bacteriol. 119, 1072-1074. Grunstein, M. and Hogness, D.S. (1975) Proc. Nat. Acad. Sci. 72, 39613965. Telford, J., Boseley, P., Schaffner, W. and Birnstiel, M. (1977) Science 195, 391-393. Levy, S.B. (1974) J. Bacteriol. 120, 1451-1463. Adams, M.H. (1959) in Bacteriophages, pp. 445-447. Wiley-Interscience Publishers, Inc., New York. Roozen, K.J., Fenwick, R.G. Jr. and Curtiss, R. (1971) J. Bacteriol. 107, 21-33. Frazer, A.C. and Curtiss, R. (1973) J. Bacteriol. 115, 615-622. Laemmli, U.K. (1970) Nature 227, 680-685. Blattler, D.P., Garner, F., van Slyke, K. and Bradley, A. (1972) J. Chromatography 64, 147-155. Bonner, W.M. and Laskey, R.A. (1974) Eur. J. Biochem. 46, 83-88. Gillespie, D. and Spiegelman, S. (1965) J. Mol. 12, 829-842. Morimoto, R., Lewin, A., Merten, S. and Rabinowitz, M. (1976) in Genetics and Biogenesis of Chloroplasts and Mitochondria, BUcher, T., Neupert, W., Sebald, W. and Werner, S., Eds., pp. 519-524. Elsevier/North-Holland, Amsterdam. Southern, E.M. (1975) J. Mol. Biol. 98, 503-517. Morimoto, R., Lewin, A., Hsu, H.J., Rabinowitz, M. and Fukuhara, H. (1975) Proc. Nat. Acad. Sci. 72, 3868-3872. Sanders, J.P.M., Heyting, C., DiFranco, A., Borst, P. and Slominski, P.P. (1976) in The Genetic Function of Mitochondrial DNA, Saccone, C. and Kroon, A.M., Eds., pp. 259-272. Elsevier/North-Holland, Amsterdam. Frazer, A.C. and Curtiss, R. (1975) Current Topics Microbiol. Immunol. 69, 1-84. Rabinowitz, M., Jacovcic, S., Martin, N., Hendler, F., Halbreich, A., Lewin, A. and Morimoto, R. (1976) in The Genetic Function of Mitochondrial DNA, Saccone, G. and Kroon, A.M., Eds., pp. 219-230. Elsevier/ North-Holland, Amsterdam. Jakovcic, S., Hendler, F., Halbreich, A. and Rabinowitz, M. (1979) Biochemistry, in press. Levins, D., Edwards, J., Locker, J., Lustig, A., Merten, S., Morimoto, R., Synenki, R. and Rabinowitz, M. (1979) in ICN-UCLA Symposia on Molecular and Cellular Biology, Vol. 15, Extrachromosomal DNA, Cummings, Borst, Dawid, Weisman, and Fox, Eds., Academic Press, New York, in press. Rudner, R., Rejman, E. and Chargaff, E. (1965) Proc. Nat. Acad. Sci. 54, 904-911. Bernardi, G. (1976) J. Mol. Evol. 9, 25-35. Miller, D.L., Gubbins, E.J., Pegg, R.W. and Donelson, J.E. (1977) Biochemistry 16, 1031-1038. 2149

Nucleic Acids Research 52.

53. 54. 55.

56. 57. 58.

59. 60. 61. 62. 63. 64. 65.

66.

Locker, J., Rabinowitz, M. and Getz, G.S. (1974) Proc. Nat. Acad. Sci. 71, 1366-1370. Locker, J., Lewin, A. and Rabinowitz, M. (1979) Plasmid, in press. Adler, H.I., Fisher, W.D. and Hardigree, A.A. (1969) Trans. N.Y. Acad. Sci. 31, 1058-1070. Gayda, R.C., Yamamoto, L.T. and Markovitz, A. (1976) J. Bacteriol. 127, 1208-1216. Gottesman, S. and Zipser, D. (1978) J. Bacteriol. 133, 844-851. Shineberg, J.B. and Zipser, D. (1973) J. Bacteriol. 116, 1469-1471. Hershey, N.D., Conrad, S.E., Sodja, A., Yen, P.H., Cohen, M. and Davidson, N. (1977) Cell 11, 585-598. Brutlag, D., Fry, K., Nelson, T. and Hung, P. (1977) Cell 10, 509-519. Carroll, D. and Brown, D.D. (1976) Cell 7, 477-486. Starlinger, P. and Saedler, H. (1976) Curr. Top. Microbiol. Immunol. 75, 111-152. Kleckner, N.J. (1977) Cell 11, 11-23. Bernardi, G. (1979) in ICN-UCLA Symposia on Molecular and Cellular Biology, Vol. 15, Extrachromosomal DNA, Cummings, Borst, Dawid, Weisman, and Fox, Eds. Academic Press, New York, in press. Morrow, J.F., Cohen, S.N., Chang, A.C.Y., Boyer, H.W., Goodman, H.M. and Helling, R.B. (1974) Proc. Nat. Acad. Sci. 71, 1743-1747. Kedes, L.H., Cohn, R.H., Lowry, J.C., Chang, A.C.Y. and Cohen, S.N. (1975) Cell 6, 359-369. Meagher, R.B., Tait, R.C., Betlach, M. and Boyer, H.W. (1977) Cell 10,

521-536.

67. 68. 69. 70. 71.

72. 73.

74.

2150

Rambach, A. and Hogness, D.S. (1977) Proc. Nat. Acad. Sci. 74, 5041-5045. Ratzkin, B. and Carbon, J. (1977) Proc. Nat. Acad. Sci. 74, 487-491. Struhl, K., Cameron, J.R. and Davis, R.W. (1976) Proc. Nat. Acad. Sci. 73, 1471-1475. Struhl, K. and Davis, R.W. (1977) Proc. Nat. Acad. Sci. 74, 5255-5259. Vapnek, D., Hautala, J.A., Jacobson, J.W., Giles, N.H. and Kushner, S.R. (1977) Proc. Nat. Acad. Sci. 74, 3509-3512. Mercerean-Puijalon, 0., Royal, A., Cami, B., Garapin, A., Krust, A., Gannon, F. and Kourilsky, P. (1978) Nature 275, 505-510. Itakura, K., Hirose, T., Crea, R., Riggs, A.D., Heyneker, H.L., Bolivar, F. and Boyer, H.W. (1977) Science 198, 1056-1063. Villa-Komaroff, L., Efstratiadis, A., Broome, S., Lomedico, P., Tizard, R., Naber, S.P., Chick, W.L. and Gilbert, W. (1978) Proc. Nat. Acad. Sci. 75, 3727-3731.