Cloning and expression in Escherichia coli of the homoserine kinase (thrB) gene from Brevibacterium lactofermentum

Summary Five DNA fragments carrying the thrB gene (homoserine kinase E.C. 2.7.1.39) of Brevibacterium lactofermentum were cloned by complementation of Escherichia coli thrB mutants using pBR322 as vector. All the cloned fragments contained a common 3.1 kb DNA sequence. The cloned fragments hybridized among themselves and with a 9 kb BamHI fragment of the chromosomal DNA of B. lactofermentum but not with the DNA of E. coli. None of the cloned fragments were able to complement thrA and thrC mutations of E. coli. Plasmids pULTH2, pULTH8 and pULTH11 had the cloned DNA fragments in the same orientation and were very stable. On the contrary, plasmid pULTH18 was very unstable and showed the DNA inserted in the opposite direction. E. coli minicells transformed with plasmids pULTH8 or pULTH11 (both carrying the common 3.1 kb fragment) synthesize a protein with an M _r of 30,000 that is similar in size to the homoserine kinase of E. coli.Abbreviations SSC 0.15 M NaCl, 0.015 M sodium citrate - SDS sodium dodecyl sulphate - TSB tripticase soy broth - m-DAP meso-diaminopimelic acid - Sm^r, Cp^r, Km^r, Am^r, Ap^r, Tc^r, MA15^r resistance to streptomycin, cephalotin, kanamycin, amykacin, ampicillin, tetracycline and microcin A 15, respectively     

Conserved Plasmid Hydrogen-Uptake (hup)-Specific Sequences within Hup+Rhizobium leguminosarum Strains

Thirteen Rhizobium leguminosarum strains previously reported as H₂-uptake hydrogenase positive (Hup⁺) or negative (Hup⁻) were analyzed for the presence and conservation of DNA sequences homologous to cloned Bradyrhizobium japonicum hup-specific DNA from cosmid pHU1 (M. A. Cantrell, R. A. Haugland, and H. J. Evans, Proc. Natl. Acad. Sci. USA 80:181-185, 1983). The Hup phenotype of these strains was reexamined by determining hydrogenase activity induced in bacteroids from pea nodules. Five strains, including H₂ oxidation-ATP synthesis-coupled and -uncoupled strains, induced significant rates of H₂-uptake hydrogenase activity and contained DNA sequences homologous to three probe DNA fragments (5.9-kilobase [kb] HindIII, 2.9-kb EcoRI, and 5.0-kb EcoRI) from pHU1. The pattern of genomic DNA HindIII and EcoRI fragments with significant homology to each of the three probes was identical in all five strains regardless of the H₂-dependent ATP generation trait. The restriction fragments containing the homology totalled about 22 kb of DNA common to the five strains. In all instances the putative hup sequences were located on a plasmid that also contained nif genes. The molecular sizes of the identified hup-sym plasmids ranged between 184 and 212 megadaltons. No common DNA sequences homologous to B. japonicum hup DNA were found in genomic DNA from any of the eight remaining strains showing no significant hydrogenase activity in pea bacteroids. These results suggest that the identified DNA region contains genes essential for hydrogenase activity in R. leguminosarum and that its organization is highly conserved within Hup⁺ strains in this symbiotic species.     

Sequence organization of a cloned tDNA1met fragment from xenopus laevis

3.18 kb fragments of X. laevis DNA coding for tRNA₁^met have been inserted into a λ vector via Hind III termini and cloned in E. coli. The organization of one cloned fragment has been analyzed by restriction endonuclease digestion and RNA-DNA hybridization. From the distribution of sites for three enzymes, this fragment appears to be typical of the majority of λ. laevis tandem tDNA₁^met repeat units. Evidence is presented to suggest that it contains two genes coding for tRNA₁^met and at least one gene coding for a second as yet unidentified 4S RNA species. The two tRNA₁^met genes are located on the same DNA strand 0.96 and 1.38 kb from one end of the repeat unit. A detailed restriction map for 19 enzymes reveals that the spacers between these genes are not identical, and it provides no indication of short repetitive sequence elements within the spacers.     

Ribosomal DNA in Drosophila melanogaster. II. Heteroduplex mapping of cloned and uncloned rDNA.

Sequences in the cloned Drosophila melanogaster rDNA fragments described by Dawid et al. (1978) were compared by heteroduplex mapping. The nontranscribed spacer regions in all fragments are homologous but vary in length. Deletion loops were observed at variable positions in the spacer region suggesting that spacers are internally repetitious.Many rDNA repeats in D. melanogaster have a 28 S gene interrupted by a region named the ribosomal insertion. Insertions of 0.5, 1 and 5 kb were found in repeat-length EcoRI fragments. These DNA regions, named type 1 insertions, are homologous at their right ends. Although 1 kb insertions are quite precisely twice as large as 0.5 kb insertions they do not represent a duplication of the shorter sequence. Some insertions have at least one EcoRI site and therefore yield EcoRI fragments which are only part of a repeat. The sequences in two cloned right-hand partial insertion sequences are homologous, but the sequences in two lefthand partial insertions are not. None of the EcoRI-restrictable insertion sequences has any homology to any part of type 1 insertions; they are thus grouped together as type 2. Evidence for insertion sequences of at least two types in uncloned rDNA was obtained by annealing a cloned fragment with a 1 kb insertion to genomic rDNA. About 15% of the rDNA repeats show substitution type loops between the 1 kb type 1 insertion derived from the cloned fragment and type 2 insertions in the rDNA.     

The ribosomal DNA of Calliphora erythrocephala; an analysis of hybrid plasmids containing ribosomal DNA

We have analysed the ribosomal DNA of Calliphora erythrocephala, a Dipteran fly of the same sub-order as Drosophila melanogaster, through a series of rDNA² fragments cloned in a plasmid vector. We have mapped the sites for eight restriction enzymes within these plasmids, and positioned the regions coding for the 18 S and 28 S rRNAs within the maps of selected plasmids using the S₁ endonuclease mapping procedure of Berk & Sharp (1977). This analysis establishes that some rDNA cistrons of C. erythrocephala contain an “intron” (Gilbert, 1978) which interrupts the 28 S coding region at the same position as that of D. melanogaster rDNA. Two introns of 2.85 kilobases in length and part of a longer, sequence-related variant were isolated in these cloned fragments. Restriction enzyme site analysis and preliminary hybridization data indicate that the 2.85 kb intron of C. erythrocephala is largely unrelated in sequence to the two classes of D. melanogaster rDNA introns.     

Nuclear tRNATyr genes are highly amplified at a single chromosomal site in the genome of Arabidopsis thaliana

Summary We have examined the organization of tRNA^Tyr genes in three ecotypes of Arabidopsis thaliana, a plant with an extremely small genome of 7 × 107 bp. Three tRNA^Tyr gene-containing EcoRI fragments of 1.5 kb and four fragments of 0.6, 1.7, 2.5 and 3.7 kb were cloned from A. thaliana cv. Columbia (Col-O) DNA and sequenced. All EcoRl fragments except those of 0.6 and 2.5 kb comprise an identical arrangement of two tRNA^Tyr genes flanked by a tRNA^Ser gene. The three tRNA genes have the same polarity and are separated by 250 and 370 bp, respectively. The tRNA^Tyr genes encode the known cytoplasmic tRNA_GA ^Tyr. Both genes contain a 12 by long intervening sequence. Densitometric evaluation of the genomic blot reveals the presence of at least 20 copies, including a few multimers, of the 1.5 kb fragment in Col-O DNA, indicating a multiple amplification of this unit. Southern blots of EcoRl-digested DNA from the other two ecotypes, cv. Landsberg (La-O) and cv. Niederzenz (Nd-O) also show 1.5 kb units as the major hybridizing bands. Several lines of evidence support the idea of a strict tandem arrangement of this 1.5 kb unit: (i) Sequence analysis of the EcoRI inserts of 2.5 and 0.6 kb reveals the loss of an EcoRI site between 1.5 kb units and the introduction of a new EcoRI site in a 1.5 kb dimer. (ii) Complete digestion of Col-O DNA with restriction enzymes which cleave only once within the 1.5 kb unit also produces predominantly 1.5 kb fragments. (iii) Partial digestion with EcoRI shows that the 1.5 kb fragments indeed arise from the regular spacing of the restriction sites. The high degree of sequence homology among the 1.5 kb units, ranging from 92% to 99%, suggests that the tRNA^Ser/tRNA^Tyr cluster evolved 1–5 million years ago, after the Brassicaceae diverged from the other flowering plants about 5–10 million years ago.     

Identification of a genomic region that complements a temperature-sensitive, high CO2-requiring mutant of the cyanobacterium, Synechococcus sp. PCC7942

Summary In a temperature-sensitive, high CO₂-requiring mutant of Synechococcus sp. PCC7942, the ability to fix intracellularly accumulated inorganic carbon was severely impaired at non-permissive temperature (41° C). In contrast, inorganic carbon uptake and ribulose-1,5-bisphosphate carboxylase activity in the mutant were comparable to the respective values obtained with the wild-type strain. The mutant was transformed to the wild-type phenotype (ability to form colonies at non-permissive temperature under ordinary air) with the genomic DNA of the wild-type strain. A clone containing a 36 kb genomic DNA fragment of the wild-type strain complemented the mutant phenotype. The complementing activity region was associated with internal 17 kb SmaI, 15 kb HindIII, 3.8 kb BamHI and 0.87 kb Pstl fragments. These 4 fragments overlapped only in a 0.4 kb HindIII-PstI region. In the transformants obtained with total genomic DNA or a plasmid containing the 3.8 kb BamHI fragment, the ability to fix intracellular inorganic carbon was restored. Southern hybridization and partial nucleotide sequence analysis indicated that the cloned genomic region was located approximately 20 kb downstream from the structural genes for subunits of ribulose-1,5-bisphosphate carboxylase/oxygenase. The cloned region was transcribed into a 0.5 kb mRNA. These results indicate that the cloned genomic region of Synechococcus sp. PCC7942 is involved in the efficient utilization of intracellular inorganic carbon for photosynthesis.     

Cloning of human mitochondrial DNA in Escherichia coli

In order to determine its nucleotide sequence, human mitochondrial DNA (mtDNA) purified from term placentae was cloned in Escherichia coli using the plasmid vector pBR322. The products of an mtDNA MboI digestion (23 fragments ranging in size from 2800 to 25 base-pairs (bp)) were ligated with BamHI-cut pBR322. The ampicillin-resistant tetracycline-sensitive colonies obtained upon transformation of E. coli χ1776 were screened by agarose gel electrophoresis of colony lysates, colony hybridization and restriction analysis. All but MboI fragment 2 were obtained in this way. MboI fragments 5 and 8 were each found only once among the 705 clones screened. All other MboI fragments were approximately equally represented in the population of clones except for a slight bias towards smaller fragments. MboI fragment 2 overlaps with the mtDNA BamHI/EcoRI (1.7 kb³) and the 0.9 kb HinIII fragments. These were cloned in similarly restricted pBR322 to provide a set of clones covering most of the mtDNA molecule. Clones representative of each MboI fragment were shown to be complementary to mtDNA by hybridization to Southern blots of mtDNA digests and were thereby partially mapped. Further mapping was obtained by restriction analysis of mtDNA sequentially degraded by exonuclease III. A collection of recombinant clones has thus been obtained using the mtDNA isolated from a single placenta and is now being used to obtain a complete nucleotide sequence of human mtDNA.     

Molecular cloning and characterization of the chicken gene homologous to the transforming gene of rous sarcoma virus

Four molecular clones containing DNA homologous to the Rous sarcoma virus transforming gene (src) have been isolated from a random library of normal chicken DNA. The four clones are distinct overlapping isolates, which together span approximately 33 kb of cellular DNA. The cloned locus appears to represent the major region of chicken DNA homologous to src, since src-containing restriction fragments of this locus account for the fragments detected by hybridization of src-specific probe to restriction digests of total chicken DNA. Analysis of the cloned chicken src locus by restriction and heteroduplex mapping indicates that the locus contains 1.6-1.9 kb of DNA homologous to the viral src gene. The chicken DNA sequences homologous to viral src are interrupted by five or six nonhomologous regions, totaling approximately 6 kb, which presumably represent introns in the cellular src gene.     

Molecular Cloning of the Fujinami Sarcoma Virus Genome and Its Comparison with Sequences of Other Related Transforming Viruses

Full-length proviral DNA of Fujinami sarcoma virus (FSV) of chickens was molecularly cloned and characterized. An analysis of FSV DNA integrated in mammalian cells showed that restriction endonuclease SacI has a single cleavage site on FSV DNA. Unintegrated closed circular FSV DNA obtained from newly infected cells was linearized by digestion with SacI and cloned into λgtWES·λB. The following three different molecules were isolated: FSV-1 (4.4 kilobases [kb]) and FSV-2 (4.7 kb), which appeared to be full-length FSV DNA molecules containing either one or two copies of the long terminal repeat structure, and FSV-3 (6 kb), which consisted of part FSV DNA and part DNA of unknown origin. An analysis of the structure of cloned FSV-1 and FSV-2 DNA molecules by restriction endonuclease mapping and hybridization with appropriate probes showed that about 2.6 kb of the FSV-unique sequence called FSV-fps is located in the middle of the FSV genome and is flanked by helper virus-derived sequences of about 1.3 kb at the 5′ end and 0.5 kb at the 3′ end. The long terminal repeats of FSV were found to have no cleavage site for either EcoRI or PvuI. Upon transfection, both FSV-1 DNA and FSV-2 DNA were able to transform mammalian fibroblasts. Four ³²P-labeled DNA fragments derived from different portions of the FSV-fps sequence were used for hybridization to viral RNAs. We found that sequences within the 3′ half of the FSV-fps gene are homologous to RNAs of PRCII avian sarcoma virus and the Snyder-Theilen strain of feline sarcoma virus, both of which were previously shown to contain transforming genes related to FSV-fps. These results suggest that the 3′ portion of the FSV-fps sequence may be crucial for the transforming activity of fps-related oncogenic sequences.     

A novel arrangement of the 18S and 28S sequences in a repeating unit of Drosophila melanogaster rDNA.

The sequences corresponding to the 18S and 28S rRNAs have been mapped within a cloned 17 kilobase (kb) fragment formed by Eco R1 cleavage of Drosophila melanogaster rDNA. This fragment, Dm103, represents the longer of two major types of repeating units that are present in the rDNA of this fly, and was cloned as a hybrid plasmid, pDm103, consisting of Dm103 inserted at the Eco R1 site of the pSC101 vector (Glover et al., 1975). Mapping of the 18S and 28S rDNA in Dm103 was accomplished by quantitative determination of the amount of these rDNAs in each member of an ordered set of restriction fragments obtained by Hind III and Eco R1 ccleavage of pDm103. The amounts of 18S and 28S rDNAs were determined by hybridization of the rRNAs to fragments that were purified by cloning, and an unambiguous order of the fragments within pDm103 was established by heteroduplex mapping and from the stoichiometry of the fragment lengths. The resulting map revealed that the 4 kb of 28S rDNA within the long repeating unit represented by Dm103 is divided into two blocks that are separated by 5.4 kb of DNA of unknown function. It is this unusual arrangement of the 28S rDNA that distinguishes the long repeating units (17 kb) from the short units (11.5) kb), whose 4 kb of 28S rDna is confined to a single block, as is shown in the accompanying paper (White and Hogness, 1977). The remainder of the DNA in this long unit appears to be typically arranged, with the 2 kb of 18S rDNA confined to a single block that is separated by about 1 kb from the closest block of 28S rDNA.     

Isolation and mapping of ribosomal RNA genes of Caulobacter crescentus

Ribosomal DNA fragments of 1.0, 3.4, 3.7 and 6.1 kb² produced by EcoRI digestion of the Caulobacter crescentus genome were identified by hybridization to a labeled ribosomal RNA probe. These genomic sequences were further characterized by the isolation of 13 hybrid λ Charon 4 phages with rDNA inserts, and two of the recombinant phages, Ch4Cc773 and Ch4Cc1880, were examined extensively. The Cc773 insert contains EcoRI fragments of 1.0 kb, 3.4 kb and 3.7 kb and the Cc1880 insert contains EcoRI fragments of 1.0 kb, 3.4 kb and 6.1 kb that hybridized to ³²P-labeled rRNA. Thus, the two clones contain different DNA inserts which together account for all of the rDNA fragments detected in digests of the C. crescentus genome. Hybridization with isolated transfer RNA and individual rRNA species indicated that the arrangement of genes in both units is 16 S-spacer tRNA(s)-23 S-5 S, tRNA(s). Homology between the DNA inserts is largely restricted to the rRNA coding regions, which suggests that the two rDNA units are located in different regions of the chromosome. Results of quantitative hybridization experiments are most consistent with a single Cc1880 and Cc773 unit per genome equivalent of 2.7 × 10⁹ daltons. The relatively simple organization of rDNA sequences in the C. crescentus chromosome compared to Escherichia coli is discussed.     

Close linkage in Pseudomonas stutzeri of the structural genes for respiratory nitrite reductase and nitrous oxide reductase, and other essential genes for denitrification

Summary The structural gene, nirS, for the respiratory nitrite reductase (cytochrome cd ₁) from Pseudomonas stutzeri was identified by (i) sequencing of the N-terminus of the purified protein and partial sequencing of the cloned gene, (ii) immunoscreening of clones from a lambda gt11 expression library, (iii) mapping of the transposon Tn5 insertion site in the nirS mutant strain MK202, and (iv) complementation of strain MK202 with a plasmid carrying the insert from an immunopositive lambda clone. A mutation causing overproduction of cytochrome c ₅₅₂ mapped on the same 8.6 kb EcoRI fragment within 1.7 kb of the mutation affecting nirS. Two mutations affecting nirD, which cause the synthesis of an inactive cytochrome cd ₁ lacking heme d ₁, mapped 1.1 kb apart within a 10.5 kb EcoRI fragment contiguous with the fragment carrying nirS. Nir^– mutants of another type that had low level synthesis of cytochrome cd ₁, had Tn5 insertions within an 11 kb EcoRI fragment unlinked to the nirS ⁺ and nirD ⁺ fragments. Cosmid mapping provided evidence that nirS and nirD, and the previously identified gene cluster for nitrous oxide respiration are closely linked. The nirS gene and the structural gene for nitrous oxide reductase, nosZ, are transcribed in the same direction and are separated by approximately 14 kb. Several genes for copper processing are located within the intervening region.     

Precise Detection and Tracing of Trichoderma hamatum 382 in Compost-Amended Potting Mixes by Using Molecular Markers

Randomly amplified polymorphic DNA (RAPD) analysis and the PCR assay were used in combination with dilution plating on a semiselective medium to detect and enumerate propagules of Trichoderma hamatum 382, a biocontrol agent utilized in compost-amended mixes. Distinct and reproducible fingerprints were obtained upon amplification of purified genomic DNA of T. hamatum 382 with the random primers OPE-16, OPH-19, and OPH-20. Three amplified DNA fragments of 0.35 (OPE-16_0.35), 0.6 (OPH-19_0.6), and 0.65 (OPH-20_0.65) kb were diagnostic for T. hamatum 382, clearly distinguishing it from 53 isolates of four other Trichoderma spp. tested. Some isolates of T. hamatum shared these low-molecular-weight fragments with T. hamatum 382. However, RAPD analysis of isolates of T. hamatum with all three random primers used in consecutive PCR tests distinguished T. hamatum 382 from other isolates of T. hamatum. These three RAPD amplicons were cloned and sequenced, and pairs of oligonucleotide primers for each cloned fragment were designed. Use of the primers in the PCR assay resulted in the amplification of DNA fragments of the same size as the cloned RAPD fragments from genomic DNA of T. hamatum 382. A combination of dilution plating on a semiselective medium for Trichoderma spp. and PCR, with the RAPD primers OPH-19, OPE-16, and OPH-20 or the three sequence-characterized primers, was used successfully to verify the presence of T. hamatum 382 propagules in nine different soil, compost, and potting mix samples. All 23 Trichoderma isolates recovered on semiselective medium from commercial potting mixes fortified with T. hamatum 382 were identified as T. hamatum 382, whereas 274 Trichoderma isolates recovered from the other nine samples were negative in the PCR assay. Thus, this highly specific combination of techniques allowed detection and enumeration of propagules of T. hamatum 382 in fortified compost-amended potting mixes. Sequence-characterized amplified region markers also facilitated the development of a very simple procedure to amplify DNA of T. hamatum 382 directly from fortified compost-amended potting mixes.     

Integration of hup Genes into the Genome of Chickpea-Rhizobium Through Site-Specific Recombination

The hup gene fragment of cosmid pHU52 was integrated into the genome of chickpea-Rhizobium Rcd301 via site-specific homologous recombination. Two small fragments of genomic DNA of strain Rcd301 itself were provided to flank cloned hup genes to facilitate the integration. The hup insert DNA of cosmid pHU52 was Isolated as an Intact 30.2 kb fragment using EcoRI, and cloned on partially restricted cosmid clone pSPSm3, which carries a DNA fragment of strain Rcd301 imparting streptomycin resistance. One of the recombinant cosmid clones, pBSL 12 thus obtained was conjugally transferred to the strain Rcd30₁. The integration of hup gene fragment into the genomic DNA through site-specific homologous recombination, was ensured by introducing an incompatible plasmid, pPH1 JI. The integration was confirmed by Southern hybridization. The integrated hup genes were found to express ex plants in two such constructs BSL 12–1 and BSL 12–3.     

β-Lactamase genes of Streptomyces badius,Streptomyces cacaoi and Streptomyces fradiae: Cloning and expression in Streptomyces lividans

Genes encoding extracellular β-lactamases (EC 3.5.2.6) of Gram-positive Streptomyces badius, Streptomyces cacaoi and Streptomyces fradiae have been cloned into Streptomyces lividans. The β-lactamase gene of S. badius was initially isolated on a 7 kb BamHI fragment and further located on a 1300 bp DNA segment. An 11 kb BamHI fragment was isolated encompassing the S. cacaoi β-lactamase gene, which was subcloned to a 1250 bp DNA fragment. The β-lactamase gene of S. fradiae was cloned on an 8 kb BamHI fragment and mapped to a 4 kb DNA segment. Each of the three BamHI fragments encompassing the β-lactamase genes hybridized to a BamHI fragment of the corresponding size in chromosomal DNA from the respective strain used for cloning. The activities of the three β-lactamases were predominantly found to be extracellular in the S. lividans recombinants. The S. badius and S. cacaoi β-lactamases exhibited a 10–100-times lower activity in S. lividans, whereas the S. fradiae β-lactamase showed an approximately 10-fold higher activity in the cloned state, compared with the activities found in the original strains.     

Homology of pCS1 Plasmid Sequences with Chromosomal DNA in Clavibacter michiganense subsp. sepedonicum: Evidence for the Presence of a Repeated Sequence and Plasmid Integration

Restriction fragments of pCS1, a 50.6-kilobase (kb) plasmid present in many strains of Clavibacter michiganense subsp. sepedonicum (“Corynebacterium sepedonicum”), have been cloned in an M13mp11 phage vector. Radiolabeled forms of these cloned fragments have been used as Southern hybridization probes for the presence of plasmid sequences in chromosomal DNA of this organism. These studies have shown that all tested strains lacking the covalently closed circular form of pCS1 contain the plasmid in integrated form. In each case the site of integration exists on a single plasmid restriction fragment with a size of 5.1 kb. Southern hybridizations with these probes have also revealed the existence of a major repeated sequence in C. michiganense subsp. sepedonicum. Hybridizations of chromosomal DNA with deletion subclones of a 2.9-kb plasmid fragment containing the repeated sequence indicate that the size of the repeated sequence is approximately 1.3 kb. One of the copies of the repeated sequence is on the plasmid fragment containing the site of integration.     

Genomic clones of a wild-type allele and a transposable element-induced mutant allele of the sucrose synthase gene of Zea mays L

In an attempt to isolate the transposable genetic element Ds from Zea mays L., we cloned DNA fragments hybridizing to a cDNA clone derived from the sucrose synthase gene in a λ vector (λ::Zm Sh). The fragments cloned from wild-type and from the Ds-induced mutant sh-m5933 (λ::Zm sh-m5933) share a segment 6 kb long while a contiguous segment of 15 kb of λ::Zm sh-m5933 (mutant-derived DNA) does not hybridize to the DNA segment cloned from the wild-type. Restriction maps are given, and the junction point between the two DNA segments in the mutant clone was determined. Hybridization of DNA fragments, present in the wild-type DNA of λ::Zm Sh, but not in the mutant clone, λ::Zm sh-m5933, to genomic DNA of sh-m5933 showed that no part of this DNA is deleted. It cannot be said whether the DNA found in the mutant, but not in the wild-type clone, has been brought there by Ds insertion or by another Ds-dependent DNA rearrangement. The mutant-derived DNA was hybridized to genomic DNA of various maize lines digested by several restriction endonucleases. Approximately 40 bands were detected. The mutant-derived DNA contains two pairs of inverted repeats several hundred nucleotide pairs long, one of which is located at the junction to wild-type-derived DNA.