Search

We present an overview of the gene content and organization of the mitochondrial genome of Dictyostelium discoideum. The mitochondria genome consists of 55,564 bp with an A + T content of 72.6%. The identified genes include those for two ribosomal RNAs (rn1 and rns), 18 tRNAs, ten subunits of the NADH dehydrogenase complex (nad1, 2, 3, 4, 4L, 5, 6, 7, 9 and 11), apocytochrome b (cytb), three subunits of the cytochrome oxidase (cox1/2 and 3), four subunits of the ATP synthase complex (atp1, 6, 8 and 9), 15 ribosomal proteins, and five other ORFs, excluding intronic ORFs. Notable features of D. discoideum mtDNA include the following. (1) All genes are encoded on the same strand of the DNA and a universal genetic code is used. (2) The cox1 gene has no termination codon and is fused to the downstream cox2 gene. The 13 genes for ribosomal proteins and four ORF genes form a cluster 15.4 kb long with several gene overlaps. (3) The number of tRNAs encoded in the genome is not sufficient to support the synthesis of mitochondrial protein. (4) In total, five group I introns reside in rnl and cox1/2, and three of those in cox1/2 contain four free-standing ORFs. We compare the genome to other sequenced mitochondrial genomes, particularly that of Acanthamoeba castellanii.

We cloned a gene (topA) encoding DNA topoisomerase II from Dictyostelium discoideum nuclear DNA using oligo probes corresponding to the consensus amino acid sequences found in the gene in other eukaryotes. The gene encoding a predicted polypeptide of 1282 amino acids with M(r) of about 146 kDa, is a single copy that is expressed as a polyadenylated 4.5 kb RNA. The predicted amino acid sequence shares similarity with those of other eukaryotes with identity between 32 and 46%. The protein is 260-300 amino acids shorter in the C-terminal region and 50-80 longer in the N-terminal region than those of other eukaryotes. In TopA of D. discoideum, the N-terminal region with stretches of charged and hydrophilic amino acids is predicted to fold into an amphiphilic alpha-helix which is characteristic of a mitochondrial targeting signal presequence. Four independent polyclonal antibodies against bacterially expressed GST fusion proteins containing four portions of the polypeptide detected a single band on Western blots at about 135 kDa. Western blots analysis of subcellular fractions revealed that this protein is localized in mitochondria. The protein and the mRNA are present in growth phase and during development, although levels of both declined as development proceeded.

The DNA sequences of cytochrome oxidase (subunits 1, 2 and 3) genes of the cellular slime mold Dictyostelium discoideum mitochondria were determined. The genes for subunits 1 and 2 have a single continuous ORF (COX1/2) which contains four group-I introns. The insertion sites of the two group-I introns (DdOX1/2.2 and DdOX1/2.3) coincide with those of fungal and algal group-I introns, as well as a liverwort group-I intron, in the cytochrome oxidase subunit 1. Interestingly, intron DdOX1/2.2 has two free-standing ORFs in a loop (L8) which have similar amino-acid sequences and are homologous to ai4 DNA endonuclease (I-Sce II) and bi4 RNA maturase found in group-I introns of Saccharomyces cerevisiae mitochondrial DNA. Two group-I introns (DdOX1/2.3 and DdOX1/2.4) also have a free-standing ORF in loop 1 and loop 2, respectively. These results show that these group-I introns and the intronic ORFs have evolved from the same ancestral origin, but that these ORFs have been propagated independently.

During the course of analysis of the mitochondrial genome of the cellular slime mold, Dictyostelium discoideum, we found a gene (rpl11) for mitochondrial ribosomal protein (RPL11), having 172 amino acid residues. Southern blot analysis revealed that the gene resided in the mitochondrial DNA as a single-copy but not in the nuclear DNA. From Northern blot experiments, one major mRNA (about 27 kb) and two minor mRNAs (about 4 and 5 kb) for the gene were detected in the mitochondria. This is the first report showing that the active gene for RPL11 still resides in the mitochondrial genome and has not been transferred to the nuclear genome in D. discoideum.

A 547-bp group-I intron belonging to subgroup IA1 was found near the 3' end of the large subunit ribosomal RNA-encoding gene (LSUrRNA) in the mitochondrial (mt) DNA of the cellular slime mold Dictyostelium discoideum. This intron was inserted in a highly conserved stretch within the sequence that encodes the peptidyl transferase center domain V in the corresponding region of the Escherichia coli LSUrRNA. Interestingly, the insertion site of the intron is the same as that of the So.LSU.2 intron of the green alga, Scenedesmus obliquus, mt DNA and the Pw.LSU.2 intron of the colorless alga, Prototheca wickerhamii, mt DNA. The intron could self-splice in vitro at a concentration higher than 20 mM MgCl2. Polymerase chain reaction analysis showed the possible existence of an intron similar to that of D. discoideum LSUrRNA in another cellular slime mold, Polysphondylium pallidum (CK-8), but not in D. mucoroides (Dm7 and Dm11).

We have sequenced a region (7,376-bp) of the mitochondrial (mt) DNA (54 kb) of the cellular slime mold, Dictyostelium discoideum. From the DNA and amino-acid sequence comparisons with known sequences, genes for ATPase subunit 9 (ATP9), cytochrome b (CYTB), NADH dehydrogenase subunits 1, 3 and 6 (ND1, ND3 and ND6), small subunit rRNA (SSU rRNA) and seven tRNAs (Arg, Asn, Cys, Lys, f-Met, Met and Pro) have been identified. The sequenced region of the mtDNA has a high average A + T-content (70.8%). The A + T-content of protein-genes (73.6%) is considerably higher than that of RNA genes (61.3%). Even with the strong AT-bias, the genetic code employed is most probably the universal one. All seven tRNAs are able to form typical clover leaf structures. The molecular phylogenetic trees of CYTB and SSU rRNA suggest that D. discoideum is closer to green plants than to animals and fungi.

A glycoprotein, gp138, is implicated in the sexual cell fusion of Dictyostelium discoideum. We previously cloned and sequenced the two genes encoding the gp138 protein, GP138A and GP138B (Fang et al. (1993) Dev. Biol. 156, 201-208). Here, we have constructed a vector producing antisense RNA for the gp138 genes and have transformed the vector into Dictyostelium cells. The transformed cells showed a reduction in the amounts of gp138 mRNA and protein and their sexual cell fusion activity was considerably repressed.

The sexual development of cellular slime molds is initiated by acquisition of sexual fusion competence of myxamoeboid cells. During the acquisition of fusion competence in NC4 and HM1 strain cells of Dictyostelium discoideum, opposite in mating types, a glycoprotein gp138, relevant to sexual cell fusion, is expressed on the cell surfaces of both the strains. To investigate the mechanisms controlling cell fusion, gp138 was purified and its N-terminal amino acid sequence was determined. An oligonucleotide with sequence predicted from the amino acid sequence was synthesized to isolate the gene encoding gp138. Two genes, referred to as GP138A and GP138B, were then cloned. They contain the region which encodes the N-terminal amino acid sequence of the gp138 protein and their structure is very similar on the whole. The C-terminal portion of the predictive polypeptides produced by the genes is highly hydrophobic and proline rich. It has homology with a portion of gp80, a glycoprotein for cell-cell adhesion, and PsA, a glycoprotein specific for prespore cells, reported previously in D. discoideum. The expression of GP138A is greater than that of GP138B. The mRNA of GP138A is expressed at the time of acquisition of fusion competence of cells during cultivation but the mRNA of GP138B is not.

Sexual cell fusion occurs between NC4 and HM1, heterothallic strains of Dictyostelium discoideum. Several cell surface proteins relevant to the process have been identified. One of them, gp138, exists in fusion-competent cells of both NC4 and HM1, and is considered to be more concerned with membrane fusion than gamete recognition. In this study, we raised monoclonal antibodies against gp138 and examined gp138 distribution among strains and species of cellular slime moulds to confirm its importance in sexual cell fusion. All heterothallic and bisexual D. discoideum strains examined were found to possess gp138, while asexual and homothallic strains lacked it. The anti-gp138 monoclonal antibody detected several distinct proteins in homothallic strains and one in an asexual strain. Some of the former proteins appeared together with the increase in binucleated cells. Cells of Dictyostelium mucoroides and Polysphondylium pallidum did not possess proteins reactive to the monoclonal antibody. These results indicate that gp138 is common among, but restricted to, cross-matable strains of D. discoideum. Our results also support previously published molecular phylogenetic studies which suggest that homothallic and asexual strains of D. discoideum are remote from other strains of D. discoideum but are less distantly related to them than other species are.

Sexual cell fusion occurs between NC4 and HM1, the heterothallic strains in Dictyostelium discoideum. Cells of these strains are fusion incompetent when cultured on agar plates in the light and become fusion competent upon cultivation in a liquid medium in darkness. Two cell-surface components, gp70 and gp138, have been identified and characterized as being relevant to sexual cell fusion. Both are glycoproteins, and the former is detected only in fusion-competent HM1 cells, while the latter is detected both in fusion-competent HM1 and fusion-competent NC4 cells. We therefore suspect gp 70 to be responsible for cell recognition and gp138, for membrane fusion. Therefore, NC4 cells are expected to possess specific surface molecule(s) that can be recognized by HM1 cells. In the present study, we raised monoclonal antibodies (mAbs) against membrane fractions of NC4 cells and selected fusion-inhibiting mAbs to identify novel molecules related to sexual cell fusion in D. discoideum. Out of the five mAbs we obtained three, DE1, GG6, and HH9, were characterized. DE1 recognized antigens that specifically existed in fusion-competent NC4 cells but not in fusion-incompetent NC4 or HM1 cells. GG6 recognized cell-surface proteins with approximate molecular weights of 125 and 32 kDa in both fusion-competent NC4 and fusion-competent HM1 cells. In addition GG6 also recognised other proteins commonly present in fusion-incompetent cells. The 125 kDa protein appeared to be the same as gp138. The epitope recognized by HH9 was sodium dodecyl sulfate (SDS)-sensitive.(ABSTRACT TRUNCATED AT 250 WORDS)

Sexual cell fusion is an initial step of macrocyst formation in Dictyostelium discoideum and requires environmental conditions such as darkness, plenty of water and the presence of calcium ions. We have been analyzing the mechanism of sexual cell fusion between HM1 and NC4, heterothallic strains in D. discoideum. Cells of these strains have been shown to be fusion competent when cultured in a liquid medium in darkness, but not so when cultured on agar plates or in a liquid medium in the light. Two cell-surface proteins, gp70 and gp138, have been identified as target molecules for fusion-blocking antibodies and therefore as relevant to sexual cell fusion. In the present study, gp70 was shown to be present in HM1 cells cultured in the light, and fusion incompetent. Intact HM1 cells cultured in the light were unable to absorb the fusion-blocking activity of antibodies against membrane components of fusion-competent HM1 cells, whose activity had been shown to be absorbed by gp70, but they did so after separation of proteins in the SDS-PAGE. In addition, fusion-competent HM1 cells were found to lose their fusion competence by subsequent cultivation in the light. This loss of competence was cycloheximide sensitive, indicating that de novo synthesis of proteins was necessary for this inhibition. From these results, we presume that light induces a protein that hinders the interaction of gp70 in HM1 cells with its receptor on the NC4 cell surface and thereby inhibits the sexual process between these strains.

Sexual cell fusion occurs between HM1 and NC4, heterothallic strains in Dictyostelium discoideum. A membrane component of HM1 cells with a molecular mass of 70 kDa (70K protein) has been shown to be implicated in cell fusion (Urushihara et al. (1988) Cell Differ. Dev. 25, 81-88). In the present study, 70K protein was partially purified using affinity Sepharose on which membrane proteins of NC4 cells were immobilized. Through this process, involvement of Ca2+ in the interaction of 70K protein with its receptor was suggested. Lectin staining of partially purified 70K protein indicated it to be a glycoprotein containing D-mannose and/or D-glucose residues.

Macrocyst formation in Dictyostelium discoideum is initiated by the fusion of cells between two opposite mating-type strains, such as NC-4 and HM1. When these cells become fusion-competent under appropriate environmental conditions, a specific protein involved in sexual cell fusion, 138 kDa, appears on the cell surface of both NC-4 and HM1 strains, as reported previously. The present study was carried out to purify and characterize this protein. The 138 kDa protein was shown to be a glycoprotein (gp138) which binds specifically to lectins, WGA, Con-A or LCA, but not to PNA, PHAE4, RCA60 or RCA120. The isoelectric point of the gp138 was determined as pH 4.5-4.9. To confirm again the previous results, a Coomassie blue-stained gel containing the 138-kDa band was cut out following the SDS-polyacrylamide gel electrophoresis, and antiserum against this band protein(s) was obtained. Fab fragments of this antiserum caused the complete inhibition of sexual fusion between NC-4 and HM1 cells.

In Dictyostelium discoideum, there exist genes which regulate the timing of developmental events. In one class of these rapidly developing (rde) mutants, the rate of development is accelerated, resulting in the formation of spores and stalk cells in about 2/3 of the time required for the parent. Linkage analysis of one of the rde strains, HTY506, demonstrated that rdeC1850, carried by the mutant is located on linkage group III. Therefore HTY506 differs genetically from the strain FR17, the first rde mutant described which carries a mutation rdeA1 on linkage group IV. While strain HTY506 has a pleiotropic phenotype, the altered characteristics of strain HTY506 appear to be caused by a single mutation, rdeC1850.

Rapidly developing (rde) mutants of Dictyostelium discoideum, in which cells precociously differentiated into stalk and spore cells without normal morphogenesis, were investigated genetically and biochemically. Genetic complementation tests demonstrated that the 16 rde mutants isolated could be classified into at least two groups (groups A and C) and that the first described rde mutant FR17 (D. R. Sonneborn, G. J. White, and M. Sussman, 1963, Dev. Biol. 7, 79-93) belongs to group A. Morphological studies revealed several differences in development and final morphology between group A and group C mutants. In group A mutants, the time required for cell differentiation from vegetative cells to aggregation competent cells is reduced, whereas the time required for spore and stalk cell differentiation following the completion of aggregation is shortened in group C mutants. This suggests that group C mutants represent a new class of rde mutants and that there exist at least two mechanisms involved in regulating the timing of development in D. discoideum. Measurements of cell-associated and extracellular phosphodiesterase activities, and intracellular and total cAMP levels revealed that cAMP metabolism in both groups is significantly altered during development. Group A mutants showed precocious and excessive production of phosphodiesterase and cAMP during the entire course of development; intracellular cAMP levels in group C mutants were extremely low, and spore and stalk cell differentiation occurred without an apparent increase in these levels. Thus, while cAMP metabolism is abnormal in all the rde mutants studied, there exist several distinct types of derangement, not necessarily involving the overproduction of cAMP.

A factor was discovered that markedly enhances the degree of fusion activity between cells of the opposite mating-type strains, MI1 and NC4, in the sexual cycle of Dictyostelium discoideum. This factor, designated cell-fusion inducing factor (CFIF), was detected initially in supernatants taken from 24 h dark-grown mixed cultures of HM1 and NC4 cells, and subsequently shown to be secreted by giant cells - the fusion products of HM1 and NC4 cells. HM1 cells, cultured in the dark at appropriate temperature, normally acquire fusion-competence specific to NC4 cells. The addition of CFIF to such dark-grown HM1 cultures results in a marked increase in their fusion-competence. In addition, when CFIF is added to light-grown HM1 cultures, in which cells normally do not acquire the ability to fuse with NC4 cells, fusion-competence is induced. The fusion-competence of NC4 cells is unaffected by CFIF, being quite high under the culture conditions used here. Experiments, using actinomycin D, daunomycin and cycloheximide showed that the secretion of CFIF from giant cells requires synthesis of RNA and protein. Possibly, cell fusion triggers production of CFIF, which is rapidly released into the external medium.

The sexual cycle of Dictyostelium discoideum is initiated by the fusion of cells that are of opposite mating types (e.g. NC4- and HM1-type cells). Cells grown in light on agar plates are not capable of sexual cell fusion, but become capable when cultured in the dark in a liquid medium. Cells in the incapable state are called fusion-incompetent cells, and cells in the latter state, fusion-competent cells. To gain some understanding of the mechanism of cell fusion, cell ghosts prepared by freeze-thawing intact cells were incubated with intact cells. The cell ghosts killed the intact cells by directly fusing with them, the extent of fusion depending on the particular strains employed and the fusion-competency of the intact cells and of the cells from which the cell ghosts had been prepared. A detailed examination revealed that fusion-competent NC4 cells were always more easily killed by cell ghosts than fusion-incompetent NC4 cells. It also became apparent that cell ghosts prepared from fusion-competent NC4 cells killed all cell types far more efficiently than did those prepared from fusion-incompetent NC4 cells. However, fusion-competent and fusion-incompetent HM1 cells were equally sensitive to cell ghosts, and cell ghosts prepared from fusion-competent HM1 cells had the same ability to kill as those prepared from fusion-incompetent HM1 cells. From these findings, it thus appears that opposite mating-type cells have distinct membrane properties related to sexual cell fusion.

The development of Dictyostelium discoideum may proceed by two pathways, macrocyst or fruiting-body formation, the former being the sexual and the latter the asexual cycle. The pathway of development depends on the presence or absence of zygote giant cells which are produced through fusion of opposite mating-type cells in a population, in heterothallic strains. During the early stages of macrocyst development the patterns of developmentally regulated proteins were noted to differ considerably from those during fruiting-body development. Furthermore, the haploid cells around zygote giant cells synthesized a large number of specific proteins for macrocyst development through the influence of giant cells.

The molecular mechanism of sexual cell fusion in Dictyostelium discoideum was studied using the heterothallic strains HM1 and NC4. Monovalent antibodies (Fab) prepared from rabbit antiserum against a crude membrane preparation of fusion-competent HM1 cells inhibited fusion between HM1 and NC4 cells. Six out of 43 antigenic proteins were found in fusion-competent HM1 cells but not in fusion-incompetent cells. Among them, only one protein with a molecular mass of 70 kDa was able to neutralize the fusion-inhibiting activity of Fab, suggesting its possible participation in sexual cell fusion.

Dictyostelium purpureum S5 and S6, mating type strains, form fruiting-bodies in a monoclonal culture, but product macrocysts in a mix culture. The effects of Concanavalin A (Con A) on both fruiting-body formation and macrocyst formation, and changes of Con A-mediated cell agglutinability during development were studied. It was found that Con A inhibits macrocyst formation but not fruiting-body formation, and that macrocyst-forming cells are much more susceptible to Con A agglutination than are fruiting-body-forming cells during the aggregation stages. When fruiting-body-forming cells are treated with either trypsin or alpha-chymotrypsin, their Con A agglutinability is enhanced to the same extent as that of macrocyst-forming cells. It was also found that when S6 cells are treated with proteases they sometimes produce normal macrocysts even in a monoclonal culture. The results obtained in these experiments showed that the surface properties of fruiting-body-forming cells and macrocyst-forming cells are different, and that the cell surface might play an important role in determining the two developmental courses.

In Dictyostelium discoideum, cytological and physiological studies on macrocyst formation revealed that this process consists of at least two steps: the production of giant cells, which are believed to be formed from the fusion of cells of two opposite mating types, and the subsequent induction of macrocyst development by the giant cells. The conditions that had been considered formerly to be required for macrocyst formation, such as darkness at the presence of two cells of complementary mating types in heterothallic strains, were actually required only for the production of the giant cells. Once giant cells are produced, the surrounding cells can aggregate and form macrocysts even in the light. Furthermore, it was demonstrated that giant cells can switch the developmental mode of the surrounding cells to macrocyst formation. That is, if a critical number of the isolated giant cells are introduced into a cell population of a single strain of NC4, which normally would produce only fruiting-bodies, macrocysts are formed instead. When in the presence of giant cells, the development of macrocysts may be initiated by starvation. Therefore, if all cells are made to starve simultaneously development begins and proceeds synchronously. Using this technique of synchronous development, the developmental kinetics of enzyme activities were assayed during macrocyst and fruiting-body formation. Considerable differences in the patterns of those enzyme activities were demonstrated between the two developmental modes of D. discoideum.

The early events of macrocyst development in Dictyostelium discoideum have been investigated using a new culturing system. When cells of opposite mating-types, HM1 and NC4, are cultured together at the appropriate temperature in the dark, giant cells appear, ingest the surrounding amoebae, and develop into macrocysts. Although these giant cells have been assumed to be the products of the fusion of opposite mating-type cells, no experimental evidence to prove this assumption has been obtained using such mixed-culture systems. In order to avoid the complexities involved in mixed-culturing, we have developed a new system involving the separate culture, and later mixing, of opposite mating-type cells. This has enabled us to obtain direct evidence that giant cells are produced by fusion between opposite mating-type cells. Cell fusion occurs immediately after mixing and is completed within 30 min. As a number of cells fuse simultaneously, giant cells produced by this method are very large multinucleated cells, and not binucleated zygotes. Using this system we also discovered the following facts related to giant cell formation: (1) cells can acquire their fusion competence without the presence of cells of the opposite mating-type; (2) only HM1 cells require darkness to acquire their fusion competence; (3) the restrictive temperature, 25 degrees C, inhibits the induction of fusion competence in HM1 cells, but not in NC4 cells.

Dictyostelium discoideum was used as a model system for elucidating the molecular mechanism of sexual cell fusion. In heterothallic strains NC4 and HM1 of D. discoideum, complements in mating type, amoeboid cells acquire fusion competence only under certain environmental conditions, such as the presence of excess water and a certain period of darkness, to fuse sexually. The surface of cells which acquired fusion competence was found to possess specific antigens. Monovalent antibodies prepared from rabbit antiserum against fusion-competent NC4 cells inhibit the sexual cell fusion of these cells completely. Two specific antigenic proteins, 39 and 138 k Da in relative molecular mass and specific for fusion-competent cells, were detected. Only one, the 138-k Da protein, was capable of neutralizing the fusion-inhibitory activity of the monovalent antibody. These results show that the 139-k Da protein is the one involved in the sexual cell fusion of NC4 and HM1 strains in D. discoideum.

We have determined the complete nucleotide sequence of pDG1, a plasmid found in a wild isolate of Dictyostelium. The 4439-bp long pDG1 contains only one, 2718-bases-long, open reading frame (ORF) and nearly perfect inverted repeats of 551 bp and 552 bp. Northern-blot analysis showed that only one 2.7-kb poly (A)+ RNA transcript was expressed at a maximum level, 2 h (early aggregation stage) after the onset of development. The expression of this transcript was suppressed by the addition of cAMP. In the upstream region of the ORF, there are several putative consensus sequences, e.g. (1) TGACTTAGAA-AAATT which is a putative site for cleavage by topoisomerase I, and (2) TGACGACA which may be a cAMP-responsive element, found in several genes that are regulated by cAMP at the level of transcription. A possible mechanism of the partitioning of pDG1 into daughter cells is discussed.

In Dictyostelium discoideum, 16 mutants in which cells differentiate into spores and stalk cells without normal morphogenesis were isolated. All these mutants are rapidly developing and capable of differentiating in a shaken suspension of phosphate buffer. The developmental kinetics of specific activities of enzymes in one of the mutants, HTY 1851, cultured in the suspension was compared with that in the parental strain, X2, developed on a solid surface. Most of the enzyme activities appeared much earlier and the peaks of the activities were lower in HTY1851 than X2, but the order or appearance of the activities was the same in both the strains cultured under the conditions described above. These results suggest that the biochemical steps in the development of the mutant in a shaken suspension are essentially the same as those of the parental strain X2 on a solid surface. It was also found that addition of cyclic AMP (2.5 X 10-5 M to 1 X 10-4 M) to the mutant cell suspension 6-8 h after the initiation of development induced an increase in the number of spores and the specific activities of some enzymes to values twice as high as those of an untreated control.

A circular plasmid having high copy number was found in a wild isolate of Dictyostelium species. Gel electrophoresis, electron microscopy and Southern blot hybridization revealed that the plasmid, named pDG1, is 4.5 Kb (1.5 micron) in size with closely situated long inverted repeats. The plasmid seems to be located in the nuclei. It was not a derivative of ribosomal DNA. The possible correlation of the plasmid with the putative intermediate DNA of retrotransposon DIRS-1 found in Dictyostelium discoideum is discussed.

In Dictyostelium discoideum cyclic AMP (cAMP) metabolism during macrocyst development, i.e., the sexual cycle of this organism, and in giant cells, i.e., fusion products from opposite mating-type cells, was investigated. The pattern of change in cAMP levels during macrocyst development differed considerably from that observed during fruiting-body formation, i.e., the asexual cycle. Giant cells produced and excreted considerable amounts of cAMP. Adenylate cyclase activity catalyzing cAMP production in giant cells was comparable to that of unfused cells. However, the activity of membrane-bound phosphodiesterase in giant cells was extremely low, and no extracellular phosphodiesterase was excreted. A phosphodiesterase inhibitory protein was secreted in excess by giant cells.

We constructed a recombinant plasmid containing the 2.1 kb HindIII fragment of plasmid pDG1, isolated from the cellular slime mold (Dictyostelium sp. strain GA11), and using pAG60 as cloning vector. We found that deletions of the recombinant plasmid took place frequently in Escherichia coli wild-type cells. However, the deletion was not observed when the plasmid was introduced into a strain that was an isogenic temperature-sensitive mutant of the gyrA gene. These results suggest that E. coli DNA gyrase is involved in the mechanisms of the deletion formation. It was shown that the 1.0 kb deletant derived from the 2.1 kb HindIII insert was produced by elimination of a 1.1 kb region. Sequence analysis of the deletants showed that cutting and rejoining took place between two out of the six nearly perfect direct repeats [21 bp palindromic sequences; AAAAAA(T/C)GGC(G/C)GCC(A/G)TTTTTT], located near the distal ends of the inverted repeats, preserving one copy of the repeats. These sequences consist of local short inverted repeats, where cutting and rejoining occur at one of the two regions.