Multiple physical differences in the genome structure of functionally related bacteriophages P1 and P7

RecBCD enzyme and Chi recombination hotspots as determinants of self vs. non-self: Myths and mechanisms

Bacteria face a challenge when DNA enters their cells by transformation, mating, or phage infection. Should they treat this DNA as an invasive foreigner and destroy it, or consider it one of their own and potentially benefit from incorporating new genes or alleles to gain useful functions? It is frequently stated that the short nucleotide sequence Chi (5' GCTGGTGG 3'), a hotspot of homologous genetic recombination recognized by Escherichia coli's RecBCD helicase-nuclease, allows E. coli to distinguish its DNA (self) from any other DNA (non-self) and to destroy non-self DNA, and that Chi is "over-represented" in the E. coli genome. We show here that these latter statements (dogmas) are not supported by available evidence. We note Chi's wide-spread occurrence and activity in distantly related bacterial species and phages. We illustrate multiple, highly non-random features of the genomes of E. coli and coliphage P1 that account for Chi's high frequency and genomic position, leading us to propose that P1 selects for Chi's enhancement of recombination, whereas E. coli selects for the preferred codons in Chi. We discuss other, previously described mechanisms for self vs. non-self determination involving RecBCD and for RecBCD's destruction of DNA that cannot recombine, whether foreign or domestic, with or without Chi.

Diversity of P1 phage-like elements in multidrug resistant Escherichia coli

The spread of multidrug resistance via mobile genetic elements is a major clinical and veterinary concern. Pathogenic Escherichia coli harbour antibiotic resistance and virulence genes mainly on plasmids, but also bacteriophages and hybrid phage-like plasmids. In this study, the genomes of three E. coli phage-like plasmids, pJIE250-3 from a human E. coli clinical isolate, pSvP1 from a porcine ETEC O157 isolate, and pTZ20_1P from a porcine commensal E. coli, were sequenced (PacBio RSII), annotated and compared. All three elements are coliphage P1 variants, each with unique adaptations. pJIE250-3 is a P1-derivative that has lost lytic functions and contains no accessory genes. In pTZ20_1P and pSvP1, a core P1-like genome is associated with insertion sequence-mediated acquisition of plasmid modules encoding multidrug resistance and virulence, respectively. The transfer ability of pTZ20_1P, carrying antibiotic resistance markers, was also tested and, although this element was not able to transfer by conjugation, it was able to lysogenize a commensal E. coli strain with consequent transfer of resistance. The incidence of P1-like plasmids (~7%) in our E. coli collections correlated well with that in public databases. This study highlights the need to investigate the contribution of phage-like plasmids to the successful spread of antibiotic resistant pathotypes.

Function of the N-terminal segment of the RecA-dependent nuclease Ref

The bacteriophage P1 Ref (recombination enhancement function) protein is a RecA-dependent, HNH endonuclease. It can be directed to create targeted double-strand breaks within a displacement loop formed by RecA. The 76 amino acid N-terminal region of Ref is positively charged (25/76 amino acid residues) and inherently unstructured in solution. Our investigation of N-terminal truncation variants shows this region is required for DNA binding, contains a Cys involved in incidental dimerization and is necessary for efficient Ref-mediated DNA cleavage. Specifically, Ref N-terminal truncation variants lacking between 21 and 47 amino acids are more effective RecA-mediated targeting nucleases. We propose a more refined set of options for the Ref-mediated cleavage mechanism, featuring the N-terminal region as an anchor for at least one of the DNA strand cleavage events.

Biochemical activities of the parA partition protein of the P1 plasmid

The unit-copy P1 plasmid depends for stability on a plasmid-encoded partition region called par, consisting of the parA and parB genes and the parS site. ParA is absolutely required for partition, but its partition-critical role is not known. Purified ParA protein is shown to possess an ATPase activity in vitro which is specifically stimulated by purified ParB protein and by DNA. ParA is responsible for regulation of expression of parA and parB, and purified ParA has an ATP-dependent, site-specific DNA binding activity which recognizes a sequence that overlaps the parA promoter. The role of the ATP-dependence of the binding activity, as well as other possible functions of the ATPase activity in partition, is discussed.

The c4 repressors of bacteriophages P1 and P7 are antisense RNAs

The c4 repressors of P1 and P7 inhibit antirepressor synthesis and are solely responsible for heteroimmunity of the phages. We show that c4 is a new type of antisense RNA acting on a target, ant mRNA, that is transcribed from the same promoter. Interaction depends on complementarity of two pairs of short sequences encompassing the ribosome binding site involved in ant expression. We demonstrate that heteroimmunity of P1 and P7 is due to just two substitutions in each of the complementary sequences of c4 and ant mRNA. Based on P1-P7 sequence comparison and a mutant analysis, we propose a secondary structure model for c4 RNA, with the complementary regions in loops as important sites for antisense control.

Plasmid-partition functions of the P7 prophage

The sequences responsible for the proper partition of the P7 plasmid prophage to daughter cells lie within a discrete block of non-similarity between P7 and its close relative P1. The DNA sequence of the P7 region was determined. A segment with near identity to the replication (rep) region of P1 is followed by sequences (P7 par) that are clearly related to but very divergent from the P1 partition region. Subcloning was used to define the ends of the functional P7 partition region. It begins with a transcription promoter followed by two large open reading frames, parA and parB, that overlap by a single base and are read in the same direction. The genes direct the synthesis of two proteins, P7 ParA and ParB, with apparent Mr of 44,000 and 37,000. Specific frameshift mutations were introduced into the two genes. Each mutation blocked plasmid partition and both were complemented when the P7 ParA and ParB proteins were supplied in trans. The amino acid sequences of the P7 proteins show strong similarities to the P1 ParA and ParB proteins. However, the DNA sequences of the P7 and P1 open reading frames are remarkably divergent, largely caused by variability at the third positions in the codons. Interspecific complementation tests showed that the P7 proteins are unable to complement P1 parA or parB mutants, and the P1 proteins fail to complement the P7 mutations. Downstream from the P7 parB open reading frame is a sequence that conserves 27 of the 34 base-pairs of the P1 partition site parS. Unlike the P1 parS site, the P7 equivalent does not contain as extensive an inverted repeat. The heptamer sequences that define ParB binding sites within P1 parS are represented in P7 but differ from it by one base. A related sequence that coincides with the secondary ParB binding site within the P1 incB sequences is present nearby. Other sequences within the P7 incB region are rather different from their P1 counterparts. The basis for the major differences in specificity of the P1 and P7 par components is discussed. Comparison of the P1 and P7 sequences, and the nature of the junctions between similar and different sequences, suggest that the phages could have evolved by the pickup of divergent cassettes by recombination.

Bacteriophage P1 tail-fibre and dar operons are expressed from homologous phage-specific late promoter sequences

Two plasmid systems, containing the easily assayable galK and lacZ functions, were employed to study the regulation of the bacteriophage P1 tail-fibre and dar operons. Various P1 DNA fragments carrying either the 5' end of lydA (the 1st gene in the dar operon) or the tail-fibre gene 19 precede the promoterless coding region of galK or were fused, in-frame, to the lacZ gene. In the presence of an induced P1 prophage, GalK and LacZ activities were both detected after a 20 to 30 minute lag period, indicating that the dar and tail-fibre operons are expressed from positively regulated, late promoters. The corresponding DNA operons are expressed from positively regulated, late promoters. The corresponding DNA region of the closely related p15B plasmid exhibits comparable promoter properties. Deletion analysis mapped the promoter of a gene 19-lacZ fusion to a DNA region upstream from gene R, an open reading frame that precedes the coding frame of gene 19. The tail-fibre gene thus forms the second gene in a three gene operon (genes R, 19 (S) and U). Sequence comparison between this promoter region, upstream sequences of the lydA gene and the corresponding portions of the p15B genome allowed the identification of a highly conserved 38 base-pair sequence, which most likely represents a P1-specific late promoter. This was confirmed by 5' mapping of P1 mRNA. Transcription of both the tail-fibre and dar operons is initiated at sites five and six base-pairs, respectively, downstream from the first conserved nucleotide of this sequence. The conserved motif consists of a standard Escherichia coli -10 region followed by a nine base-pair palindromic sequence located centrally about position -22.

Organization of the bacteriophage P1 tail-fibre operon

The revised sequence of a bacteriophage P1 DNA fragment containing the 5' end of the tail-fibre gene, gene 19, revealed that this gene is closely preceded by another open reading frame (ORF) of 432 bp. We have designated this ORF as gene R. The tail-fibre gene and gene R are transcriptionally and translationally coupled. Thus, the tail-fibre operon of bacteriophage P1 consists of three genes: gene R, gene 19 (or gene S) and gene U.

A single amino acid difference between Rep proteins of P1 and P7 affects plasmid copy number

P1 and P7 are closely related plasmid prophages which are members of the same incompatibility group. We report the complete DNA sequence of the replication region of P7 and compare it to that of P1. The sequence predicts a single amino acid difference between the RepA proteins of these two plasmids, no differences in methylation sites or regions where dnaA protein is expected to bind, and no difference in the spacing of the major features of the two replicons. A P1 replicon with a mutation in repA, the gene that encodes an essential replication protein, is complemented for replication by providing either the P1 RepA protein (RepA1) or the P7 RepA protein (RepA7) in trans. Furthermore, when either of these proteins is supplied in trans, the plasmid copy number of P1 cop mutants drops to that of P1 cop+. However, when RepA7 is supplied, the copy number of P1 cop and P1 cop+ is higher than that when RepA1 is supplied. This indicates that the single amino acid difference between the two versions of the RepA protein plays an important role in determining the plasmid copy number.

The plasmid-maintenance functions of the P7 prophage

The region responsible for the maintenance of the prophage of bacteriophage P7 as a stable, unit-copy plasmid was isolated in a lambda att vector which lysogenizes Escherichia coli as a stable unit-copy plasmid under the control of the P7 replication origin. The P7 plasmid-maintenance region was shown to consist of adjacent replication and partition regions capable of functioning independently. The isolated replication region could support plasmid maintenance but the resulting plasmids were highly unstable unless the partition region was also included. Stable composite plasmids were isolated containing the putative P7 partition region and the origin of replication of the unrelated plasmid F, indicating that P7 encoded an active partition mechanism. The replication regions of P7 and P1 were shown to be highly homologous but the partition regions of the two plasmids appear to be unrelated in sequence. The incompatibility determinants associated with the two replication regions showed the same specificity, whereas the partition-region incompatibility determinants were different, showing no cross-specificity.

Expression and proteolytic processing of the darA antirestriction gene product of bacteriophage P1

The darA gene coding for one of the two bacteriophage P1 antirestriction functions is expressed late after infection or induction. The protein is made as a high-molecular-weight soluble precursor. This is proteolytically cleaved to the mature form, which is a structural component of the phage head. Defective mutants of the phage have been found in which the synthesis of gpdarA is normal but processing does not take place. These mutations all map to the same region of the P1 genome and we propose that they lie in the structural gene for the processing protease.

Partition functions of unit-copy plasmids can stabilize the maintenance of plasmid pBR322 at low copy number

The maintenance of plasmid pBR322 is highly unstable in a polA12 strain of Escherichia coli at 29 degrees C due to severely reduced copy number. Under these conditions, introduction of the par (partition) locus of plasmid P1 or the par (sop) region of F into pBR322 stabilizes it. A region with similar activity was detected in the P7 plasmid. The activity of the P1 par locus was dependent on the P1 parA gene product and was sensitive to par-specified incompatibility.

Sequence relations among the IncY plasmid p15B, P1, and P7 prophages

Electron microscopic analysis of heteroduplex molecules between the 94-kb plasmid p15B and the 92-kb phage P1 genome revealed nine regions of nonhomology, eight substitutions, and two neighboring insertions. Overall, the homologous segments correspond to 83% of the P1 genome and 81% of p15B. Heteroduplex molecules between p15B and the 99-kb phage P7 genome showed nonhomology in eight of the same nine regions; in addition, two new nonhomologous segments are present and P7 carries a 5-kb insertion representing Tn902. The DNA homology between those two genomes amounts to 79% of P7 DNA and 83% of p15B. Plasmid p15B contains two stem-loop structures. One of them has no equivalent structure on P1 and P7 DNA. The other substitutes the invertible C segments of P1 and P7 and their flanking sequences including cin, the gene for the site-specific recombinase mediating inversion.

IS1-dependent generation of high-copy-number replicons from bacteriophage P1 Ap Cm as a mechanism of gene amplification

Mutant P1 Ap Cm lysogens were isolated in which the drug resistance genes resident on the plasmid prophage P1 Ap Cm are amplified by a novel mechanism. The first step required for amplification is IS1-mediated rearrangement of the P1 Ap Cm prophage. The drug resistance genes are amplified from the rearranged P1 Ap Cm prophage by the formation of a plasmid (P1dR) which contains the two resistance genes. The P1dR plasmid is an independent replicon about one-half the size of P1 Ap Cm that can be maintained at a copy number eightfold higher than that at which P1 Ap Cm can be maintained. It contains no previously identified replication origin and is dependent on the Rec+ function of the host.

Reversion of a truncated gene for ampicillin resistance by genetic rearrangements in Escherichia coli K12

The composite transposon Tn2672 is a derivative of the Tn3-related transposon Tn902 whose bla gene providing ampicillin resistance had been inactivated by the insertion of the IS1-flanked multiple drug resistance transposon Tn2671. Most ampicillin resistant revertants of Tn2672 are due to precise excision of Tn2671. However, a rare Bla+ revertant which still retains all the previously acquired drug resistance markers was isolated. On this revertant, the 5' part of the split bla gene on Tn2672 has converted to an intact, active bla gene, and the entire Tn902 is structurally restored. In contrast, the adjacent IS1b element belonging to Tn2671 has its terminal 142 base pairs deleted. Despite of this rearrangement, the split 3' part of bla and its adjacent sequences have remained unchanged. Models are presented to explain the observed DNA rearrangements, and their similarity with gene conversion events is discussed.

Crossover sites cix for inversion of the invertible DNA segment C on the bacteriophage P7 genome

The bacteriophage P7 genome contains an invertible DNA segment called C which determines its host range. P7 C(+) phages produce plaques on Escherichia coli K12. The C segment consists of a 3-kb unique sequence and 0.62-kb inverted repeats of which one carries an internal 0.2-kb deletion. This deletion has been mapped within the right inverted repeat in the C(+) orientation. The crossover sites cix for inversion of the C segment do not map at the inside boundaries of the inverted repeats, as had been proposed. They are localized at the external ends of these repeats. Thus organization of the C segment in phage P7 is analogous to that in the related phage P1.

The partition functions of P1, P7, and F miniplasmids

The partition regions of P1, P7, and F miniplasmids are discrete DNA sequences of about 3 kb in length that will promote accurate partition of hybrid plasmids independent of the source of replication functions or the position or orientation of the elements. Each of the par regions seems to be very similarly organized, with open reading frames for essential proteins and a terminal site which appears to be analogous to the centromere of eukaryotic cells. When cloned, these terminal sites exert incompatibility against their respective parent plasmids presumably because they can compete with the parent plasmids as substrates for partition. We have determined the complete DNA sequence of the P1 par region. In addition to the open reading frame for the essential parA protein (42-44 kd), the region contains a second open reading frame which could encode a 38-kd protein. The 2 large open reading frames appear to form an operon that is negatively regulated from a site adjacent to the promoter and responds to the par gene products in trans. Both this site and the downstream "centromere" site, incB, contain blocks of extremely AT-rich sequences, which are postulated to be binding sites for par proteins. The incB and upstream AT-rich regions both contain 20-bp imperfect inverted repeats. Further downstream from the minimal incB sequence (172 bp) lies an additional region which is essential for partition. The further analysis of the P1 par region should be greatly facilitated by the finding that it can function in cis to stabilize pBR322 vectors under conditions where the copy number of pBR322 is reduced.

Characterization of generalized transducing phage phi w39 heteroimmune to phage P1 in Escherichia coli W39

Generalized transducing phage similar to phage P1 in Escherichia coli was isolated from E. coli W39, an antigenic test strain of the O121 group. This phage, designated phi w39, was reciprocally heteroimmune to phages P1 and P7, but nonreciprocally heteroimmune to phage D6. Transduction experiments using various R plasmids with different molecular weights suggested that phage phi w39 could transduce at least 65 megadaltons DNA. As in the case of P1 prophage, phi w39 prophage existed as a plasmid belonging to incompatibility group Y and carried a dnaB-like function. The molecular weight of phi w39 plasmid was nearly the same as that of plasmid, i.e., 58.6 megadaltons. Despite the pronounced structural and functional similarity of phages phi w39 and P1, restriction cleavage patterns of their genomes differed considerably.

Physical analysis of the genomes of hybrid phages between phage P1 and plasmid p15B

The genomes of three plaque-forming recombinant phages between phage P1 and plasmid p15B were characterized by restriction cleavage analysis and electron microscopic heteroduplex studies. The structure of all three P1-15 hybrid genomes differs from that of P1 DNA in the res mod region coding for restriction and modification systems EcoP15 and EcoP1, respectively. P1-15 hybrid 2 shows an additional major difference to P1 around the site of the residential IS1 element of P1 and it does not carry an IS1 in its genome.

Cointegrational transduction and mobilization of gentamicin resistance plasmid pWP14a is mediated by IS140

The structures of two R-plasmids pWP14a and pWP12a (Tra-, Ap, Gm; 21 kb) and of several cointegrates they form with bacteriophages P1Cm and P1-15 were analyzed. In each case, replicon fusion was mediated by the element IS140 (about 0.8 kb), one copy of which resides on both plasmids adjacent to the gentamicin resistance determinant (AAC(3)-III). pWP14a cointegrated preferentially into or near the invertible C-loop structure of the P1 genome. Cointegrational mobilization of pWP14a was observed also with several conjugative R-factors. The process of replicon fusion is independent of the host's rec+ functions. Sequences homologous to IS140 are constituents of many R-factors, including RA1, R40a, R124, R144, Rts1, N3, and pJR255. IS140 also shows homology to two other sequences, IS15 delta and Tn2680, but not to other, well studied transposable elements. The ampicillin resistance determinant of pWP14a is within a Tn3-like transposon, Tn3651.

Coliphage P1 morphogenesis: analysis of mutants by electron microscopy

We used electron microscopy and serum blocking power tests to determine the phenotypes of 47 phage P1 amber mutants that have defects in particle morphogenesis. Eleven mutants showed head defects, 30 showed tail defects, and 6 had a defect in particle maturation (which could be either in the head or in the tail). Consideration of previous complementation test results, genetic and physical positions of the mutations, and phenotypes of the mutants allowed assignment of most of the 47 mutations to genes. Thus, a minimum of 12 tail genes, 4 head genes, and 1 particle maturation gene are now known for P1. Of the 12 tail genes, 1 (gene 19, located within the invertible C loop) codes for tail fibers, 6 (genes 3, 5, 16, 20, 21, and 26) code for baseplate components (although one of these genes could code for the tail tube), 1 (gene 22) codes for the sheath, 1 (gene 6) affects tail length, 2 (genes 7 and 25) are involved in tail stability, and 1 (gene 24) either codes for a baseplate component or is involved in tail stability. Of the four head genes, gene 9 codes for a protein required for DNA packaging. The function of head gene 4 is unclear. Head gene 8 probably codes for a minor head protein, whereas head gene 23 could code for either a minor head protein or the major head protein. Excluding the particle maturation gene (gene 1), the 12 tail genes are clustered in three regions of the P1 physical genome. The four head genes are at four separate locations. However, some P1 head genes have not yet been detected and could be located in two regions (for which there are no known genes) adjacent to genes 4 and 8. The P1 morphogenetic gene clusters are interrupted by many genes that are expressed in the prophage.

Sequence homology and recombination between the genomes of morphologically dissimilar bacteriophages LP 52 and theta

A restriction fragment map of Bacillus licheniformis temperate phage LP 52 DNA (molecular weight 38.5 X 10(6)) was established, using restriction endonucleases BamHI (8 target sites), BglI (10 sites,) BglII (13 sites) and EcoRI (22 sites). The map is linear, with well-defined ends, without any signs of circular permutation. The DNA of a related phage, LP 51, produced identical restriction fragments. At least 62% DNA of LP 52 has been found homologous to the DNA of the recently discovered, morphologically quite dissimilar, phage I, as demonstrated by hybridization of electrophoretically separated restriction fragments of DNA. Under the same conditions, the DNAs of LP 52 and of the morphologically similar Bacillus subtilis phage phi 105 did not cross-hybridize. The homologous regions in the genomes of phages LP 52 and I have been shown to be colinear. Comparison of the cleavage maps of phages LP 52 and I has shown that, within the regions of homology, not a single restriction fragment and few restriction sites have been conserved during divergent evolution. Three major regions of heterology were defined; the longest one, covering the right-hand end of the map (73 +/- 2.75% up to 100% LP 52 genome length) appeared to contain genes coding for structural proteins of the virions; a shorter region at the left-hand end of the map (coordinates zero to 10.3 +/- 3.3% LP 52 genome length) and a very short central region (coordinates 41.8-43.9%) could be identified, the latter apparently containing a regulatory locus responsible for the heteroimmune behavior of the two phages. Recombinants between phages LP 52 and I were isolated. Mapping of recombinant genomes has indicated mutual substitution of allelic pieces of LP 52 and I DNAs upon strict conservation of overall genome length.

A site-specific, conservative recombination system carried by bacteriophage P1. Mapping the recombinase gene cin and the cross-over sites cix for the inversion of the C segment

The bacteriophage P1 genome carries an invertible C segment consisting of 3-kb unique sequences flanked by 0.6-kb inverted repeats. With insertion and deletion mutants of P1 derivatives the site-specific recombinase gene cin for C inversion) has been mapped adjacent to the C segment and the cix sites (for C inversion cross-over) have been located at the outside ends of the inverted repeats. Inversion of the C segment functions as a biological switch and controls expression of the gene(s) responsible for phage infectivity carried on the C segment. The cin gene product can promote recombination between a 'quasi- cix ' site on plasmid pBR322 and a cix site on P1 DNA. The junctions formed on the resulting co-integrate can also serve as cix sites. This observation implies a potential evolutionary process to bring genes under the control of a biological switch acting by DNA inversion.

Cointegrates between bacteriophage P1 DNA and plasmid pBR322 derivatives suggest molecular mechanisms for P1-mediated transduction of small plasmids

We characterized cointegrates formed in an Escherichia coli rec+ strain between bacteriophage P1 genomes and small plasmids related to pBR322. The partners were, on the one hand, either phage P1 DNA, which carries one copy of IS1, or phage P1-15 DNA, a derivative which lacks the IS1, and, on the other hand, plasmids containing either a split IS1 or no. In the presence of IS1 sequences on both partners, cointegrates were usually formed by reciprocal recombination between SI1 sequences. Cointegrates between P1 and a plasmid carrying no IS1 sequence were formed by transpositional cointegration mediated by IS1 of P1. Cointegrates between P1-15 and small plasmid containing a split IS1 were formed by one of three ways: (a) acquisition of an IS1 by P1-15 followed by reciprocal recombination between IS1 sequences, (b) transpositional cointegration mediated by the split IS1 element, Tn2657, or (c) involvement of the invertible segment carried on P1-15 DNA. Most cointegrates segregated into the small plasmids and phage P1 derivatives. A comparison of the phenomenon studied and of their frequencies allowed us to conclude that cointegrate formation is a molecular mechanism involved in the transduction of plasmids smaller than those packageable into P1 virions, although it does not seem to be the only process used.

Characterization of P1argF derivatives from Escherichia coli K12 transduction. I. IS1 elements flank the argF gene segment

Specialized transducing derivatives of the temperate bacteriophage P1 (P1std) are selected by transduction into recipients with deletions in the corresponding genes (Stodolsky 1973). When Escherichia coli K12 strains are used as donors in such transduction experiments, P1argF derivatives can be selected. The argF gene is unique to these strains (Glansdorff et al. 1967). Under these experimental conditions P1argF are formed with frequencies 10,000 times greater than other P1std. The majority of the P1argF derivatives that have been analyzed are indistinguishable by cleavage analyses. One such derivative, P1argF5 has been characterized in detail. Heteroduplex analysis against P1, P7, and P1CmO identified an 11 kb insertion of DNA precisely at the naturally occurring IS1 locus of P1. Cleavage analysis with EcoRI, BamHI and PstI confirmed this finding. To further define the argF insertion, a P1Cm13argF derivative was constructed having the IS1 sequences of Cm13 and argF in opposite orientation. Intrastrand annealing of P1Cm13argF5 DNA established that the argF segment is flanked by directly repeated IS1 sequences. The IS1-argF-IS1 segment is designated Tn2901. The assignment of the map position of the argF gene within the 11 kb insert of P1argF5 is discussed. The evolutionary significance of this finding and a model for P1argF formation is also presented.

Amplification of chloramphenicol resistance transposons carried by phage P1Cm in Escherichia coli

We have characterized a number of P1Cm phages which contain the resistance genes to chloramphenicol and fusidic acid as IS1-flanked Cm transposons. Restriction cleavage and electron microscopic analysis showed that these Cm transposons were carried as monomers (M) or tandem dimers (D). Lysogens of P1Cm (D) are more resistant to chloramphenicol than those of its P1Cm (M) presumably as a result of an increased gene dosage. Amplification of the Cm transposons to tandem multimers was frequently observed in P1Cm (D) lysogens grown in the presence of high concentrations of chloramphenicol or fusidic acid and was also detected in P1Cm (M) lysogens. The degree of amplification varied in different clones which suggests that cells containing spontaneously amplified Cm transposons were selected by high doses of the antibiotics. The dimeric as well as the amplified Cm transposons carried in P1Cm lysogens grown in the absence of chloramphenicol displayed considerable stability. Mechanisms for the amplification of the IS1-flanked transposons are discussed.