Copy number control and compatibility of nuclear plasmids in Dictyostelium discoideum

PLASMID

22,2 1S-223 (1989)

Copy Number Control and Compatibility of Nuclear Plasmids in Dictyostelium discoideum JOANNE E. HUGHES’ AND DENNIS L. WELKER Molecular

Biology/Biochemistry

Program, Department of Biology, Utah State University, Logan, Utah 84322-5500 Received August 2 I, 1989; revised October 25, 1989

Copy number of the endogenous nuclear plasmids of Dictyostetium discoideum is a plasmidspecific trait. Copy number is stable over time, is constant relative to ploidy level, is independent of host cell genetic background, and is independent of the presence of a second unrelated plasmid in the same nucleus. Unrelated plasmids are compatible with one another within a single nucleus. Pairwise combinations of Ddpl, Ddp2, and Ddp5 were stably maintained over many generations in the absence of selection. In contrast, one of the D. discoideum plasmids (Ddp2) was incompatible with a recombinant plasmid derived from it (p7d2). In the absence of selection for retention of p7d2, transformants contain either one or the other but not both plasmids. The plasmids are stably maintained in host cells with differing genetic backgrounds, although plasmid-free colonies were detected at a frequency of about l-2% in populations of some strains after 50 generations Cloning. Q 1989 Academic PRX, Inc. growth fOlIOWing a previous

Plasmids have been extensively characterized in bacteria with regard to copy number regulation, replication, and segregation at cell division. The knowledge gained has aided our understanding of the characteristics required by a self-replicating plasmid. Construction of prokaryotic vector plasmids is largely dependent on the availability of this type of information. Plasmids have more recently been discovered in eukaryotes. These are predominantly cytoplasmic elements, but nuclear, high copy number, circular, plasmids have been found in yeast (Broach, 1982; Utatsu et al., 1987), in the protozoan Naeglaria gruberi (Clark and Cross 1987), and in the Dictyostelium slime molds (Metz et al., 1983; Noegel et al., 1985a; Orii et al., 1987; Hughes et al., 1988). Dictyostelium presents an interesting case, as 20% of isolates examined (from four different species) contain nuclear plasmids, and these are very different from one another in their size, copy number, and restriction site maps (Noegel et al., 1985a; Hughes et al.,

’ To whom correspondence and reprint requests should be addressed.

215

1988). Our understanding of why eukaryotes have some genetic material carried on extrachromosomal elements and how these plasmids regulate their own replication and stability is minimal. The 2-pm circle of yeast is the only eukaryotic nuclear plasmid for which we have any data on these processes (Reynolds et al., 1987; Cashmore et al., 1988; Futcher, 1988; Som et al., 1988). We have been examining three of the plasmids from Dictyostelium discoideum in order to better understand the self-maintenance functions of eukaryotic, nuclear, multicopy plasmids. As a first step, we have looked at copy number and stability of these plasmids in different genetic backgrounds and compatibility of different plasmids within the same nucleus. MATERIALS

AND

METHODS

Plasmids. Wild-type plasmids were recovered from NC4 (Ddpl), WS380B (Ddp2), and WS2 162 (DdpS). Details on the construction of the G418 resistance recombinant plasmid p7d2 from the endogenous plasmid Ddp2 and restriction maps of both plasmids are described in Hughes et al. (1988). Plasmid p7d2 contains 0147-619X/89

$3.00

Copyright @ 1989 by Academic F’res, Inc. All rights of reproduction in any form reserved.

216

HUGHES

AND WELKER

the entire Ddp2 sequencewith a 2.4-kb fragment encoding the gene conferring resistance to G4 18 inserted into a unique Ddp2 Sal1 site. A restriction map of plasmid Ddpl is presentedin Noegel et al. (1985b). The integrative vector plasmid BlOS confers resistance to G418 (Firtel et al., 1985).

the integrative vector B 1OSor with the recombinant plasmid p7d2. Following the glycerol treatment, aliquots containing lo5 cells were transferred to the wells of a 96-well plate. Transformants were selectedby incubating the cells for 4-5 days in HL5 with 20 pg/ml G4 18. Surviving cells were harvested and plated with Transformation and parasexual genetics. Escherichia coli B/r on DM medium (PodTransformations were done using the calcium gorski and Deering, 1980) and grown at 2 1 phosphate precipitation procedure described + 1°C. Independent transformants were obby Nellen et al. ( 1984). The plasmid-free hap- tained by picking a single colony from each loid recipient used in transformation was ei- plate. ther strain AX3K or strain HUD205 (Table Diploid cells containing multiple plasmids 1). Recipient cells were grown axenically in were formed from pairs of haploid, plasmidMES-HL5 medium (Nellen et al., 1984). D. containing strains by parasexual cell fusion as discoideum plasmids, either singly or in pairs, described previously (Welker, 1986;Welker et were cotransformed into recipient cells with al., 1985). The haploids were TS12 (Ddpl),

TABLE

1

HAPLOID D. discoideum STRAINS

HUD205

NC4 derivative, plasmid-free (Jensen et al., 1989) a clonal stock of axenic mutant AX3; genotype: axeAl, uxeB1 (Poole and Firtel, 1984) NC4 derivative, plasmid-free (Jensen et al., 1989); genotype: cycA1, axeAl, acrA1823. bsgA5, whiC351, manA2, cow1351 (Welker et al., 1986) NC4 derivative, plasmid-free (Jensen et at., 1989); genotype: cycAI, axeAl, acrA1836, whiB355, bwnA1, manA2, bsgB500 (Welker, 1986) WS380B and HU1852 derivative containing Ddp2, segregant of parasexual diploid DUDl; genotype: bwnA1, bsgB500 WS380B and HU I852 derivative containing Ddp2, segregant of parasexual diploid DUD I; genotype: bwnA1 WS380B and HU I852 derivative containing Ddp2, segregant of parasexual diploid DUDI; genotype: bsgB500 WS380B and NC4 derivative containing Ddp2, segregant of parasexual diploid DUD70 (HUD12 X NP2); genotype: cycA1, axeAl, uxeB1, tsgA1 NP81 derivative, plasmid-free (Jensen et al., 1989); genotype: axeAl, axeB1, tsgA1,

HUD524

NC4 derivative HU 11 I6 (Welker and Williams,

AX3K HU1628 HU1852 HUD6 HUD7 HUD22 HUD174

bwnA1, ebrA1 1985) transformed with Ddp5; genotype:

axeAl. axeB1, tsgV1826 NC4 NP8 1 TS12 WS380B WS2 162

Wild isolate containing Ddpl (Metz et a/., 1983) NC4 derivative containing Ddpl; genotype: axeAl, uxeB1, tsgA1, bwnA1, ebrA1 (Wright et al., 1977) NC4 derivative containing Ddpl; genotype: cycA1, tsgDl2, whiA1 (Katz and Sussman, 1972) Wild isolate containing Ddp2 (Noegel et al., 1985a) Wild isolate containing Ddp5 (Noegel et al., 1985a)

Note. Phenotypes associated with the genetic markers are acrAl823 or acrA1836, methanol resistance (2%); axeA and axeB1, ability to grow in axenic culture when both mutations are present; bsgA5 or bsgB500, inability to grow using Bacillus subtihs as a food organism; bwnA I, brown pigment production during fruiting body formation; couA351, coumarin (1.3 mM) and temperature (27°C) sensitivity; cycA1, cycloheximide resistance (500 &ml); ebrAI, ethidium bromide resistance (35 pg/ml); manA2, ol-mannosidase-1 deficient; tsgAf, tsgDl2, temperature (27°C) sensitive.

PLASMID COPY NUMBER AND COMPATIBILITY

HUD 174 (Ddp2), and HUD524 (Ddp5). The origins of these strains are given in Table 1. DNA isolation and characterization. Large amounts of D. discoideum plasmids and total nuclear DNA were isolated using published techniques (Welker et al., 1985; Hughes et al., 1988). A plasmid miniscreen technique (Hughes and Welker, 1988)was used to screen individual colonies of D. discoideum cells for plasmids. The copy number of the plasmids was determined by comparing the plasmid band in a restriction enzyme-digested nuclear DNA sample with the neighboring bands from the ribosomal DNA (an 88-kb palindrome present at about 90 copies per haploid genome (Kimme1 and Firtel, 1982)) using a densitometer (E-C Apparatus Corp.). The silver grain density was compared either from a photographic negative (Polaroid Type 665) of the ethidium bromide-stained gel or from an autoradiogram (Kodak X-Omat AR) following Southern blotting of the gel and hybridization to a mixture of 32P-labeledprobes containing a cloned fragment of the plasmid of interest and a cloned rDNA fragment of similar size. The rDNA fragment (EcoRI fragment V; Kimmel and Firtel, 1982) carried both coding sequences and noncoding, spacer DNA. Enzymes were obtained from BethesdaResearch Laboratories and were usedin accordancewith the manufacturer’s instructions. Plasmid compatibility and maintenance. To determine if the plasmids were compatible with each other, strains containing multiple plasmids were grown on DM medium in the absence of selection for approximately 200 generations (basedon a doubling time of about 4 h on solid medium, in the presenceof a bacterial food source). The cells were passaged10 times, each passageconsisting of 3 days growth (about 18 generations), at which time a single colony was transferred to a fresh plate and streaked to produce single colonies. Plasmid miniscreen samples obtained before, during, and after this period of growth were analyzed on an agarosegel to check for the continued presence of the plasmids.

217

To examine whether individual plasmids were lost from various strains, independent colonies were obtained from a previously cloned stock of the plasmid-containing strain and screened for the presence of the plasmid on agarosegelsand by hybridization of Southern blots (Hughes and Welker, 1988). Approximately 50 generations separated the original cloned cell from the single cells which were subsequently grown and screenedfor the presence of a plasmid. The one exception to this was the transformants of strain HUD205, in which case approximately 18-24 generations separated the original cloned cell from the single cells which were grown and screened for their plasmid content. RESULTS Plasmid Copy Number Is Actively Regulated

If plasmid copy number is regulated, and is an inherent characteristic of the plasmid itself, a consistent average copy number should be observed for each plasmid within clones of a particular haploid strain, in haploid strains with different genetic backgrounds, and in diploids. Table 2 indicates that this is the case. For plasmids Ddp2, Ddp5, and p7d2 the plasmid DNA bands were clearly visible on a photograph of an ethidium bromide-stained gel (haploid wild isolates and transformants containing Ddp2 and Ddp5 are shown in Fig. 1). Plasmid copy number was determined by comparing the density of the plasmid band(s) with that of nearby rDNA bands on a photographic negative of the gel using a densitometer. The copy number of plasmid Ddpl was low enough to be difficult to make out the plasmid bands on the gel itself. In this casethe densitometer was usedto compare the density of the bands on an autoradiogram following hybridization of a Southern blot. The copy number of plasmid Ddp5 was determined by both methods to compare the results. As can be seenfrom Table 2, copy number appeared slightly higher based on the Southern blots. The copy number was determined for each plasmid in the haploid wild isolate in which

218

HUGHES AND WELKER TABLE 2 COPY NUMBER PER HAPLOID GENOME OF D. discoideum

Plasmid (kb)

All samples mean f SD (n)

Nontransformants mean + SD (n)

Transformants mean f SD (n)

Ddpl” (13) Dd~2~ (5.8) DdpS”(15) Ddp5b p7d2b (8.2)

58 k 28 (9) 225 k 70 (17) 161 rt41 (4) 104 + 30 (15) N.A.

50 + 23 (5) 268 k 74 (8) 159 + 50 (3) 91 f 30 (8) N.A.

68 f 34 (4) 186 f 38 (9) 119 + 24 (7) 139 f 30 (8)

Note. N.A., not applicable; n, number of samples. a A densitometer was used to compare the hybridization on a Southern blot. b A densitometer was used to compare the amount of DNA in the plasmid band and that in nearby rDNA bands directly from a photographic negative of the ethidium bromide-stained gel.

it was originally found, in transformants of the haploid, axenic, plasmid-free strain AX3K, and in diploids constructed from various haploid strains by parasexual cell fusion. In all casescopy number fell within a characteristic range for each plasmid. Although the mean copy number differed for transformants and nontransformants in Ddp2-containing and DdpS-containing haploid strains, the significance of this is not clear as there was considerable overlap in the range in each case. In diploids, plasmid copy number is twice that found in haploids, consistent with a doubling of total nuclear DNA. The copy number of plasmid p7d2 was monitored in transformant clones over a period of growth corresponding to about 200 generations in the absenceof any form of selection and was found to be constant (mean 147 f 30, six samples). Transformants contain appropriate plasmid copy numbers (Table 2), indicating that copy number must be actively regulated. This was confirmed by transforming AX3K with a lOOO-fold range of plasmid p7d2 DNA amounts (1 ng-1 pg). Regardless of the amount of p7d2 DNA used, the plasmid copy number in each transformant was the same, indicating that final copy number was independent of the number of molecules which entered the cell during the transformation procedure. The number of transformants obtained with 1 ng of p7d2 DNA was lower (5

of 20 wells contained transformants, eachwell initially containing lo5 cells) than with 10 ng to 1 pg of p7d2 (20 of 20 wells for 10 ng, 100 ng and 1 pg DNA). Transformation frequencies were thus 1.2 X 103/pg with 1 ng p7d2 DNA and a minimum of 5 X 103/Fgwith 10 ng p7d2 DNA. Characteristic plasmid copy numbers are not determined by a maximum carrying capacity of the D. discoideum cell. The copy number of a particular plasmid is the same whether it is the only plasmid present in the cell or is one of two different plasmids in the same cell. This is true in transformants of AX3K (seeFig. 1) and in diploids constructed via parasexual genetics. A cell with Ddp2 and Ddp5 contains an average of 330 plasmid molecules per haploid genome, twice as much plasmid DNA as a cell containing either plasmid alone. Plasmid Stability in D@erent Chromosomal Backgrounds The stability of the plasmids in different genetic backgrounds was examined by screening up to 300 colonies of various haploid strains containing each plasmid. Lab strains and transformants were tested. Among strains carrying Ddpl , 220 colonies of strain NP8 1 and 150 colonies from three transformants derived from HUD205 were screened,and 4 plasmid-

PLASMID COPY NUMBER AND COMPATIBILITY

free colonies were found, all from strain NP8 1. Among strains carrying Ddp2,242 colonies of strain HUD6,66 colonies of strain HUD7,48 colonies of strain HUD22, 107 colonies of strain HUD 174,309 colonies from four transformants of strain HUD205 and 56 colonies of a transformant of strain AX3K were screened with 25 plasmid-free colonies being detected, 24 from strain HUD6, and 1 from strain HUD7. Thirty-four colonies of a transformant of strain AX3K carrying Ddp5 were screened and all contained Ddp5. Therefore in the majority of strains examined, a cell which contained a plasmid produced descendants which still carried that plasmid after 50 generations. All 29 casesof plasmid loss occurred in strains NP81, HUD6, and HUD7 from a total of 522 colonies. It may be significant that the three strains which lost plasmids all carry the same linkage group IV marked with the bwnAI mutation. Closely related strains, HUD22 and HUD 174,that lacked this version of linkage group IV did not lose plasmids. The data from HUD6 may represent an atypically high frequency of plasmid-free cells. When colonies from one population of strain HUD6 were tested 22 casesof plasmid loss were seen in a total of 156 colonies (14%). However, when 84 colonies derived from two newly purified clones were examined, 2 plasmid-free sampleswere seen(2.4%), consistent with the frequency of plasmid-free colonies seen with strains NP8 1 (1.8%) and HUD7 (1.5%). Unexpectedly, a plasmid-free derivative of NP8 1 (HUD205) used as a recipient in transformations with Ddpl or Ddp2 or with Ddpl and Ddp2 (cotransformed with the integrating vector B 10s) showed no plasmid loss. This indicates that strains that previously have lost a plasmid are not incapable of supporting that plasmid and suggeststhat plasmid loss is not solely dependent on strain genotype but also on other, as yet unknown, factors. In this system, as in others, the number of elapsed generations may be important in the detection of plasmid loss. In the case of the transformants of strain HUD205, fewer generations had passedbetween the original plasmid-con-

a

b

c

219

defghi

PIG. I. Plasmid maintenance in D. discoideum transformants. Nuclear DNA from plasmid-bearing strains was digestedwith the restriction enzyme EcoRI and separated on a 0.8% agarose gel. The presence of plasmids Ddp2 and Ddp5 waseasily detected under such conditions. Plasmid copy number was determined by comparing the density of the plasmid band with that of neighboring rDNA bands. DNA from wild isolates WS380B (a) and WS2 162 (b). DNA from transformants of AX3K containing plasmids Ddp2 (c), Ddp5 (d), and Ddp2 plus Ddp5 (e). Plasmid-free strain AX3K (f). Purified, EcoRI-digested Ddp2 DNA (g) and Ddp5 DNA (h), and phage X DNA digested with Hind111(i) as size markers (23, 9.6, 6.6, 4.4, 2.3, 2.0 kb).

taming cell and the descendantsused to set up populations for screening (about 20 generations instead of 50). Unrelated Plasmids Are Compatible In transformants of AX3K containing pairwise combinations of Ddp 1, Ddp2, and Ddp5 both plasmids were stably maintained for at least 200 generations in the absenceof selection. For an example of such transformants seeFig. 1. This was also true of diploids carrying pairwise combinations of these three

220

HUGHES AND WELKER

plasmids. The fact that different plasmids are compatible within the same nucleus suggests that these three plasmids have independent copy number control mechanisms. We interpret this to mean that the origins of replication of these plasmids are recognized independently of one another.

B 10s. A transformant containing the plasmid Ddp2, but lacking BlOS, was chosen. This strain was then transformed a second time with p7d2 and G4 18 resistant colonies selected (negative controls in which this strain was transformed with pUC 18 gave no G4 18 resistant colonies). Following 40 generations in the absenceof selection the plasmid p7d2 was not detected in any transformant derived from a Related Plasmids Are Incompatible recipient cell which already contained Ddp2. To test whether plasmids carrying the same In contrast, when a similar transformant conorigin of replication would interfere with one taining Ddp5 but lacking B 10s was subseanother, AX3K was transformed with a mix- quently transformed a secondtime with p7d2, ture of Ddp2 and the recombinant plasmid plasmid p7d2 wasdetectedin all transformants derived from it, p7d2 (G418 resistance). In- in addition to Ddp5. Plasmid p7d2 was unable dependent transformants from three transfor- to establish itself and oust a resident populamation experiments contained one or the tion of Ddp2 even when conditions initially other plasmid, not both. Ddp2 appeared to be favored maintenance of p7d2 (selection for favored (of 36 independent transformants ex- resistance to G4 18). This suggeststhat in the amined, 27 contained Ddp2). In these trans- Ddp2/p7d2 cotransformation, the colonies formation experiments the transformants were shown to contain only p7d2 may never have initially selected for G418 resistance (p7d2 contained Ddp2. Clearly, all of the transforpresence)and then spent about 40 generations mants recovered must have contained p7d2 in the absence of selection before being ex- originally, as selection was for resistance to amined for their plasmid complement, ap- G418. parently sufficient time for incompatible plasmids (those sharing the same origin) to DISCUSSION segregate.Southern blots of digests of chromosomal DNA samplesfrom a subsetof these The D. discoideum nuclear plasmids are transformants confirmed that only one plasmaintained at characteristic copy numbers, in mid was present extrachromosomally. In concommon with prokaryotic plasmids (Scott, trast, AX3K transformed with a mixture of 1984), eukaryotic nuclear plasmids such asthe p7d2 and the unrelated plasmid Ddp5 showed 2-pm circle of Saccharomyces cerevisiae no evidence of incompatibility ( 19 of 19 in(Futcher, 1988), the extrachromosomal rDNA dependent transformants contained both elements found in a number of protozoans plasmids). (including D. discoideum; Kimmel and Firtel, These data show that Ddp2 and p7d2 are incompatible and, in the absenceof selection 1982), and latent DNA viruses maintained as for G4 18 resistance,Ddp2 has someadvantage extrachromosomal circles, such as bovine over the recombinant p7d2, with 75% of the papillomavirus (BPV2; Mecsas and Sugden, transformants ultimately containing only 1987). The copy number of each plasmid is Ddp2. We carried out an experiment in which maintained within a narrow range even when Ddp2 was deliberately given an advantage. In the plasmid is transferred to a strain different from that in which it is found in nature. Inthis caseAX3K was cotransformed with Ddp2 appropriate copy numbers can be corrected, and B 1OS.Independent G4 18 resistant transas must occur during the transformation of formants were passagedin the absence of selection for about 60 generations and then * Abbreviation used: BPV, bovine papillomavirus. tested for the presenceof the integrating vector

PLASMID COPY NUMBER AND COMPATIBILITY

plasmids into a new strain or the construction of diploid cells by parasexual cell fusion. Plasmids are found at their characteristic copy number, regardlessof the presenceof a second, high copy number plasmid in the same nucleus. Thus, as with the 2-pm circle and BPV, the D. discoideum plasmids are actively involved in controlling their own copy number. The mechanism(s) by which the D. discoideum plasmids regulate copy number is as yet unknown. It is unlikely that they use an FLPmechanism like that of the 2-pm circle, since there is no evidence for the presence of isomerit forms of the D. discoideum plasmids. A plasmid isolated from a Dictyostelium species by Orii et al. (1987) does contain inverted repeats, but they make no mention of seeing isomeric forms of this plasmid. The data of Ahern et al. ( 1988) suggestthat several regions of plasmid Ddpl may be involved in copy number control. Deletion of their “cassettes” A and G markedly reduce copy number of the derivative plasmids in transformants. Both of these plasmid regions are transcribed (Noegel et al., 1985b) and could potentially encode factors which act in tram to regulate replication initiation, a common mechanism of copy number control in prokaryotic plasmids (Scott, 1984).Similarly, most of the transcripts of the yeast 2-pm circle have been shown to encode trans-acting factors involved in copy number control and partitioning (Reynolds et al., 1987; Som et al., 1988). In most strains the plasmids are maintained stably. However, loss of Ddpl or Ddp2 from strains NP81, HUD6, and HUD7 was observed. Frequencies of about 1 to 2% plasmidfree colonies in most experiments were seen with these strains. Other plasmid-free strains that have lost Ddpl are also known; these include AX3 (Metz et al., 1983) as well as AX3K, HU1628, and HU1852 (Jensen et al., 1989). This high frequency of loss in a few strains may suggestbreakdown in a system to actively partition the plasmids between daughter cells at cell division. If the plasmids were free to diffuse in the nucleus and were segregatedrandomly at cell division, their high

221

copy number would make plasmid loss an extremely rare event (Futcher and Cox, 1983), for example 7 X 1O-37for Ddp 1 at a copy number of 60. Therefore active partitioning of the D. discoideum plasmids at cell division may be taking place, as is the casein both the 2-pm circle and the BPV. That either native plasmid Ddpl or Ddp2 can be lost implies that partitioning of these plasmids may use a common chromosomally encoded gene function. NP8 1, HUD6, HUD7, and HU1852 sharethe samelinkage group IV, marked with the bwnAZ mutation. Perhaps a mutation on this linkage group increases the frequency of plasmid loss. It is of interest to note that no loss of plasmids was seen with transformants generated from HUD205, a plasmid-free derivative of NP8 1. This may be a consequence of slight changes in experimental protocol or environmental effects. The experiments with HUD205 transformants took place approximately 1 year after those with NP8 1, HUD6, and HUD7, and we utilized smaller, newly cloned colonies as the source of cells to establish the HUD205 transformant colonies to be screened. In prokaryotes plasmid incompatibility is thought to indicate that two plasmids share one or more elements of their replication or partitioning systems;i.e., they are recognized by the DNA replication and/or partitioning machinery as being a single population (Novick, 1987). Our results suggestthat the same is true in D. discoideum. The lack of detectable cross-hybridization between plasmid sequences(Noegel et al., 1985a),and their ability to coexist stably in the same nucleus, each at their own appropriate copy number, indicates that they are recognized as independent populations of molecules for purposes of DNA replication and partitioning. Attempts to have two distinguishable plasmids with the same origin of replication coexist were unsuccessful, although either of these two plasmids could coexist stably with an unrelated plasmid. The size difference in the two plasmids could explain the advantage enjoyed by the smaller Ddp2, since a fasterreplication rate would lead

222

HUGHES AND WELKER

to a higher copy number, and lower probability of random loss at cell division. Four different plasmids have been detected in wild isolates of D. discoideum, a situation unlike that in S. cerevisiae where all strains contain the same plasmid. Other species of Dictyostelium carry additional different plasmids (Hughes et al., 1988; Orii et al., 1987), again unlike in yeasts, where all of the plasmids studied have the same basic structure as the 2-pm circle. Copy number of the D. discoideum plasmids is an inherent property of the plasmid itself, but the mechanism by which copy number is regulated may be different from that employed by the 2-pm circle. We thus have a situation in which a number of plasmids may have evolved independently within the same species. This should allow us to expand our knowledge of how extrachromosomal elements in eukaryotes originate, how copy number of such elements is regulated, and how stability is ensured. A comparison of the origins of replication in the D. discoideum plasmids should contribute much information on consensus features required for initiation of DNA replication in this organism. ACKNOWLEDGMENTS Some preliminary experiments on plasmid compatibility were carried out by D.L.W. with Dr. A. Noegel (Max Planck lnstitut fuer Biochemie). We thank M. Largent for typing the manuscript. This research was supported by grants from the National Institutes of Health (GM35 167) and from the Utah Agricultural Experiment Station, Utah State University, Logan, Utah 84322-4845. Approved as journal paper No. 3849.

CLARK, C. G., AND CROSS,G. A. M. (1987). rRNA genes of Naegleria gruberi are carried exclusively on a 14kilobase-pair plasmid. Mol. Cell. Biol. 7, 3027-303 1. FIRTEL, R. A., SILAN, C., WARD, T. E., HOWARD,P., METZ, B. A., NELLEN,W., AND JACOBSON, A. (1985). Extrachromosomal replication of shuttle vectors in Dictyostelium

discoideum. Mol. Cell. Biol. 5, 3241-

3250. FUTCHEK,A. B. (1988). The 2 pm circle plasmid of Saccharomyces cerevisiae. Yeast 4, 21-40.

FUTCHER,A. B., AND Cox, B. S. (1983). Maintenance of the 2 pm circle plasmid in populations of Saccharomyces cerevisiac J. Bacterial. 154, 6 12-622.

HUGHES,J. E., ASHKTORAB,H., AND WELKER, D. L. (1988). Nuclear plasmids in the Dicryostelium slime molds. Dev. Genet. 9,495-504. HUGHES,J. E., AND WELKER,D. L. (1988).A mini-screen technique for analyzing nuclear DNA from a single Dictyostelium colony. Nucleic Acids Res. 16, 2338. JENSEN,S. L., ASHKTORAB,H., HUGHES, J. E., AND WELKER,D. L. (1989). Gene amplification associated with the dominant cob-354 cobalt resistance trait in Dictyostelium discoideum. Mol. Gen. Genet., in press. KATZ, E. R., AND SUSSMAN,M. (1972). Parasexual recombination in Dictyostelium discoideum: Selection of stable diploid heterozygotes and stable haploid segregants. Proc. Natl. Acad. Sci. USA 69, 495-498. KIMMEL, A. R., AND FIRTEL,R. A. (1982). The organization and expression of the Dictyostelium genome. In “The Development of Dictyostelium discoideum” (W. F. Loomis, Ed.), pp. 233-324. Academic Press,New York. MECSAS,J., AND SUGDEN,B. (1987). Replication of plasmids derived from Bovine Papilloma Virus type 1 and Epstein-Barr Virus in cells in culture. Annu. Rev. Cell. Biol. 3, 87-108. METZ, B. A., WARD, T. E., WELKER,D. L., AND WILLIAMS, K. L. (1983). Identification of an endogenous plasmid in Dictyostelium discoideum. EMBO J. 2,5 15519. NELLEN,W., SILAN,C., ANDFIRTEL,R. A. (1984). DNAmediated transformation in Dictyostelium discoideum: Regulated expression of an actin gene fusion. Mol. Cell. Biol. 4,2890-2898.

REFERENCES AHERN,K. G., HOWARD,P. K., ANDFIRTEL,R. A. (1988). Identification of regions essentialfor extrachromosomal replication and maintenance of an endogenousplasmid in Dictyostelium. Nucleic Acids Res. 16, 6825-6837. BROACH,J. R. (1982). The yeast plasmid 2 pmcircle.Cell 28,203-204. CASHMORE,A. M., ALBURY, M. S., HADHELD, C., AND MEACOCK,P. A. (1988). The 2 wrn D region plays a role in yeast plasmid maintenance. Mol. Gen. Genet. 212,426-43 1.

NOEGEL,A., WELKER,D. L., METZ,B. A., ANDWILLIAMS, K. L. (1985a). Presenceof nuclear associatedplasmids in the lower eukaryote Dictyostelium discoideum. J. Mol. Biol. 185, 447-450. NoEGEL,,A.,METZ,B. A.,AND~ILLIAMS, K.L.(1985b). Developmentally regulated transcription of Dictyostehum discoideum plasmid Ddpl. EMBO J. 4, 37973803. NOVICK,R. P. (1987). Plasmid incompatibility. Microbial. Rev. 51, 381-395. 0~11, H., SUZUKI, K., TANAKA, Y., AND YANAGISAWA, K. (1987). A new type of plasmid from a wild isolate

PLASMID

COPY NUMBER

of Dictyostelium species: The existenceof closelysituated long inverted repeats. Nucleic Acids Res. 15,1097-l 107.

PODGORSKI, G., ANDDEERING,R. A. ( 1980).Quantitation of induced mutation in Dictyostelium discoideum: Characterization and use of a methanol-resistance mutation assay. Mutat. Rex X,459468. POOLE,S. J., AND RRTEL, R. A. (1984). Genomic stability and mobile genetic elements in regionssurrounding two discoidin I genesof Dictyostelium discoideum. Mol. Cell. Biol. 4,67 I-680.

REYNOLDS,A. E., MURRAY, A. W., AND SZOSTAK,J. W. (1987). Roles of the 2 pm gene products in stable maintenance of the 2 pm plasmid of Saccharomyces cerevisiae. Mol. Cell. Biol. 7, 3566-3573.

SCO?T,J. R. (1984). Regulation of plasmid replication. Microbial. Rev. 48, l-23. SOM, T., ARMSTRONG,K. A., VOLKERT, F. C., AND BROACH,J. R. (1988). Autoregulation of 2 pm circle gene expression provides a model for maintenance of stable plasmid copy levels. Cell 52,27-37. UTATSU, I., SAKAMOTO,S., IMURA, T., AND TOH-E, A. (1987).Yeast plasmids resembling 2 Nrn DNA: Regional similarities and diversities at the molecular level. J. Bacterial. 169, 5537-5545.

223

AND COMPATIBILITY

WELKER,D. L. (1986). Linkage analysis of nystatin resistance mutations in Dictyostelium discoideum. Genetics 113,53-62.

WELKER,D. L., HIRTH, K. P., ROMANS,P., NOEGEL,A., FIRTEL, R. A., AND WILLIAMS, K. L. (1986). The use of restriction fragment length polymorphisms and DNA duplications to study the organization of the actin multigene family in Dictyostelium discoideum. Genetics 112, 27-42. WELKER, D. L., HIRTH, K. P., AND WILLIAMS, K. L. (1985). Inheritance of extrachromosomal ribosomal DNA during the asexual life cycle of Dictyostelium discoideum: Examination by useof DNA polymorphisms. Mol. Cell. Biol. 5, 273-280.

WELKER,D. L., AND WILLIAMS,K. L. (1985). Translocations in Dictyostelium discoideum. Genetics 109,34 l364. WRIGHT, M. D., WILLIAMS,K. L., AND NEWELL,P. C. ( 1977).Ethidium bromide resistance:A selectivemarker located on linkage group IV of Dictyostelium discoideum. J. Gen. Microbial.

102, 423-426.

Communicated by Barry Polisky