The effect of tra mutations on the synthesis of the F-pilin membrane polypeptide

Summary We had previously demonstrated that several F specific polypeptide bands could be detected in the membranes of Flac, but not F^- strains of Escherichia coli K 12, (Moore et al. 1981). One of these polypeptides co-migrated with F-pilin protein on polyacrylamide gels. We have now analyzed ³⁵[S]methionine labelled membrane preparations from a series of strains containing Flac tra mutant plasmids. The F-pilin polypeptide was absent from preparations of strains containing all traA mutants tested, confirming the importance of the traA gene in F-pilin biosynthesis. A polypeptide which migrated in the F-pilin position was still present, however, in membranes prepared from Flac strains carrying mutations in traL, traE, traK, traB, traV, traW, traC, traU, traF, traH or traG despite the inability of these mutants to elaborate F-pili filaments. Thus, all of these gene products may be concerned with F-pilus assembly and outgrowth rather than biosynthesis of the F-pilin subunit. The polar mutation tra-4 did, however, prevent the appearance of pilin polypeptide, indicating that at least one unidentified gene in the region between traE and traG must also be required in F-pilin biosynthesis.Our analysis also permitted the identification of a 100,000 dalton membrane protein as the product of traG. The appearance of an F specific 12,000 dalton protein was prevented by traD amber mutants. As expected, traJ mutants prevented the expression of all the tra operon products detected except the product of traT. The traT product band was reduced only to 50–60% of its normal intensity.     

A new activity in the Ftra operon which is required for F-pilin synthesis

Summary Membrane preparations from a series of Hfr mutant strains of Escherichia coli K12 deleted in the promoter distal end of the F transfer operon were analyzed. Deletions which extended into traG, as expected, had no discernible effect on synthesis of membrane F-pilin. A more extensive deletion in strain KI777 which eliminated traH activity similarly had no effect on F-pilin synthesis. Membranes from three other TraF⁺ TraH^- deletion strains, as well as membranes from all strains carrying deletions extending into traF or further, lacked F-pilin, however. Since traH amber mutations do not affect synthesis of membrane pilin (Moore et al. 1981 b) we conclude that a gene required for F-pilin biosynthesis is located between traF and traH. We have named this gene traQ.Further evidence for traQ and an assay for its activity was obtained by examining the products of a TraM⁺ TraJ⁺ TraA⁺ lambda transducing phage, KI13, in UV irradiated cells. Infection of F^- cells with KI13 does not result in F-pilin synthesis. Membrane pilin is synthesized as a product of the transducing phage if an Flac or Hfr irradiated host is used, however. Mutant analysis demonstrated that this synthesis is independent of host expression of traA, traL, traE, traK, traB, traV, traW, traC, traU, traF, or traH, but dependent on expression of the traF-traH region. We interpret our data to indicate that an activity encoded by traQ is required for the conversion of traA product to F-pilin.     

The spread of plasmids in model populations of Escherichia coli K12.

Comparison of R100 with its derepressed derivative R100-1 showed that the capacity to repress tra function does not significantly affect the spread by retransfer of R100. F′lac was used to investigate the contributions of growth and transfer to spread of a plasmid through a recipient population. Ability to transfer F′lac was lost rapidly when donor cultures entered stationary phase, but aggregate-forming ability was lost much more slowly. Comparison of F′lactra⁺ with F′lactraH88, which is unable to retransfer from recipients, showed the importance of retransfer. We used a mathematical model to calculate the amount of retransfer needed to explain the rate of increase of F′lac progeny. This showed that the lag between a cell receiving F′lac and being able to retransfer it was a less important constraint on this rate of increase than the inherent rate of plasmid transfer by established donors.     

Construction and characterization of multicopy plasmids containing the entire F transfer region

The restriction endonucleases HindIII and/or BamI have been used to clone the entire F transfer region into pBR322 to create a series of transfer-proficient multicopy plasmids. Despite the insertion of 40.7- to 55.9-kilobase F fragments, the plasmid copy numbers remained high, at 25–40 copies per cell. One of the chimeric plasmids contained the F replication and incompatibility region, and its high copy number confirmed that replication of the cointegrate was governed by pBR322. Despite the 30- to 40-fold increase in tra gene copy number compared to Flac, the transfer frequencies, number of pili per cell, and syntheses of the individual traT and traI proteins were increased only by about 5-fold. The level of tra mRNA in cells carrying the multicopy transfer-proficient plasmids was also increased only by about 5-fold, suggesting that its relative synthesis or stability was reduced in this situation. Nonetheless, the increased production of tra DNA, mRNA, and protein makes cells carrying the multicopy conjugative plasmids excellent sources of these products.     

tra cistrons and proteins encoded by the Escherichia coli antibiotic resistance plasmid R6-5

Summary Hybrid plasmids obtained by cloning individual EcoRI and HindIII fragments of the conjugative plasmid, R6-5, were analyzed for their ability to complement transfer-deficient point mutations of Flac. As a result, the locations of 10 tra cistrons were defined on the physical map of R6-5. Two cistrons, traE and traG, are interrupted by EcoRI restriction sites and one cistron, traC, probably contains a HindIII restriction site. The origin of DNA transfer, oriT, was also localized. Surprisingly the hybrid plasmid carrying oriT is mobilized by the F factor as well as by R6-5. The surface exclusion cistrons, traS and traT, were mapped and their biological expression analyzed. A total of 18 proteins encoded by cistrons within the tra region were detected by SDS polyacrylamide gel electrophoresis of proteins synthesized in minicells; they represent about 53% of the coding capacity of the cloned DNA. R6-5 DNA fragments containing the cistrons traC, traE, and traT directed the synthesis of proteins which comigrated during SDS gel electrophoresis with the F-coded proteins previously characterized as TraCp, TraEp, and TraTp. A further two proteins encoded by R6-5 comigrated with F-encoded (but genetically unidentified) proteins whose cistrons map in the corresponding part of the tra region. In contrast, no R6-5 proteins corresponding to F proteins TraAp, TraDp, TraJp, TraMp, 6a or 6c were detected. These results are discussed in relation to known DNA sequence homologies between the F and R6-5 plasmids. A preliminary physical map of the tra region of R6-5 is presented and compared with that of F.     

Towards a structural biology of bacterial conjugation

Horizontal DNA transfer among bacteria (conjugation) requires the formation of secure intercellular contacts before DNA transfer can occur. The formation of such contacts among Gram-negative bacteria is mediated by conjugative pili. Like other pili, conjugative pili are filaments extending from the surface of donor cells. They are composed, in so far as is known, entirely of conjugative pilin subunits. Here, we review the structure of F-pilin, the subunit of which F-pili are composed. We emphasize recent studies suggesting a specific domain organization for F-pilin and present a model of how these domains might be arranged in filament subunits.     

Non-random distribution of transduction termini in transductants from the integrated R plasmid, R100-1

Summary Tra ⁺and tra ^–derivatives of drug resistance plasmid, R100-1, were isolated by phage P1 from an Hfr donor with integrated R100-1 and then analyzed by complementation tests with tra ^–point mutants of Flac. Tra ⁺derivatives of R100-1 carrying tetracycline resistance alone and those carrying all six drug-resistance genes could support transfer of tra ^–point mutants of Flac except Flac traJ, whereas all of tra ^–derivatives of R100-1 failed to complement any one of tra ^–point mutants of Flac. This suggests that these tra ^–derivatives of R100-1 carrying tetracycline resistance gene are deleted for all the transfer genes impaired in the Flac point mutants tested. We assume a hot point, probably a specific base sequence similar to an IS element, at the left of the tetracycline gene (Fig. 1) becomes a transduction terminus in transduction of the integrated R100-1 by phage P1. Complementation analysis of tra ^–derivatives carrying five resistance genes except the tetracycline gene led us to a supposition that a gene(s), probably analogous to traJ of the F plasmid, is located on R100-1 near the tetracycline gene which plays an important regulatory role for self-transfer as well as for the complementation of tra ^–Flac mutants.     

Overproduction, purification and characterization of the F traT protein

Summary A lambda transducing phage (ED110) which carries the sex factor F surface exclusion genes, traS and traT, was characterized by both genetic and physiochemical techniques. The transducing segment consists of 5.2 kilobases of F tra DNA, and carries the carboxy-terminal onehalf of the upstream traG gene, as well as traS, traT, and the adjacent downstream gene traD. These tra proteins could be identified in infected UV-irradiated cells, and the major part of their synthesis was found to occur from the phage's late promoter p_R under Q control.Lysogens for ED110 were induced and found to greatly overproduce the traT gene product (TraTp), an outer membrane protein normally found in about 20,000 copies per cell, to levels which exceeded the major outer membrane proteins. This led to the development of a simple purification procedure for TraTp, the most important step of which was the construction of an appropriate ompB derivative to eliminate the major outer membrane porin proteins, which have several physical properties in common with TraTp.Purified TraTp was added to mixtures of donor and recipient cells and found to inhibit mating. The specificity of this assay was demonstrated by using an R100-1 donor, which responds to a heterologous surface exclusion system, and by using an altered TraTp containing a missense amino acid substitution.A mechanism by which TraTp mediates surface exclusion is proposed.     

Effect of tra mutations on F factor-specified immunity to lethal zygosis

Summary Hfr, F⁺, and F-prime cells are, unlike F^– cells, insensitive to an excess of Hfr donor cells, indicating that there is an F factor mediated immunity to lethal zygosis (Ilz). Results with Flac episomes carrying traJ, traS or various polar mutations in the tra region indicate that this immunity is independent of surface exclusion, of traJ control, and of all known genes within the tra operon. However, analysis of a series of strains with deletions in the F factor, extending from the right into the tra region, suggests that a gene for immunity to lethal zygosis is located within the tra region. We therefore conclude that Ilz is genetically complex, and present a hypothesis to account for these results.     

Incompatibility between Flac, R386, and F:pSC101 recombinant plasmids: the specificity of F incompatibility genes.

The specificity of F incompatibility genes (inc⁺) has been studied with the Flac and R386 plasmids, members of the IncFI incompatibility group. Recently, two inc⁺ regions, incA (46.4–49.3F) and incB (43.1–46.4F) were identified by cloning these F sequences onto pSC101 and subsequently demonstrating incompatibility of the recombinants with Flac. It is shown here that the FincA⁺ recombinant is incompatible with both Flac and R386 while the FincB⁺ recombinant is incompatible only with Flac. Also, a plasmid mutant is described that has reduced incompatibility against Flac and R386. The mutation is located on the BamHI restriction fragment that contains the FincA region. These genetic findings are consistent with the deduction of Palchaudhuri and Maas, based on heteroduplex analysis of IncFI plasmids, that placed the IncFI determinant in the 46.4–48.6F region. The findings also indicate that the FincB⁺ gene product, which has been implicated in negative control of F copy number, is specific for the F replicon.     

Characterization of F-pilin as an inner membrane component of Escherichia coli K12.

Antipeptide antibodies were used to detect, purify, and characterize nonfilament F-pilin in the cell envelope of an F'tra+ strain of Escherichia coli. Affinity-purified goat antibodies raised against a peptide corresponding to the amino-terminal 14 amino acids of F-pilin detected F-pilin in immuno-overlay ("Western") blots of electrophoretically separated inner and outer membrane proteins. As expected, the molecule was absent from inner membrane preparations of F- or F'traA[Am] strains. Immunoreactive material was purified from inner membrane fractions and shown to be F-pilin by amino acid analysis. The anti-peptide antibodies also detected membrane forms of F-pilin produced by cells containing plasmid pTG801 (Grossman, T. & Silverman, P. (1989) J. Bacteriol. 171, 650-656). Most cell envelope pilin was in the inner membrane fraction, but a significant quantity fractionated with the outer membrane as well. The hydropathy profile of F-pilin suggested that the molecule is an integral membrane protein with two membrane-spanning domains. In confirmation, F-pilin and pTG801 pilins in inner membrane preparations were solubilized by a single extraction with the nonionic detergents Nonidet P-40 (2%) or Triton X-100 (2%), but not by 2 M KCl or 0.1 M NaOH. Moreover, analysis of traA'-'phoA constructs indicated that both the amino and carboxyl termini of F-pilin face the periplasm. The periplasmic location of the amino terminus was confirmed by immunoelectron microscopy of spheroplasts from F' and pTG801 strains, using a monoclonal antibody that recognizes an amino-terminal epitope. These data suggest a specific structure for membrane F-pilin. We discuss that structure in relation to the probable structure of filament F-pilin.     

Host range of conjugation and replication functions of the Escherichia coli sex plasmid Flac. Comparison with the broad host-range plasmid RK2

The Escherichia coli conjugative plasmid Flac has a restricted host range, in that transfer to Pseudomonas aeruginosa is not detectable. The molecular basis for this host-range restriction was studied by a separate comparison of the replication and conjugation systems of Flac with those of the broad host-range plasmid RK2. The origin of transfer of Flac (oriT_F) was cloned onto a small RK2 replicon. The hybrid plasmid, pDG2906, could be transferred efficiently by both the Flac and RK2 conjugation systems to an E. coli recipient. The Flac conjugation system was able to transfer pDG2906 to P. aeruginosa, but only at a frequency of 10^?4 of that of the RK2 conjugation system. A second hybrid plasmid, containing the replication region of Flac with the transfer region of RK2, could not be established in P. aeruginosa. These results show that Flac is able to mediate low frequency transfer to P. aeruginosa, and that the lack of replication in Pseudomonas is ultimately responsible for the restricted host range.     

Effects of F-encoded components and F-pilin domains on the synthesis and membrane insertion of TraA'-'PhoA fusion proteins

F-pilin, the 70-amino-acid F-pilus subunit, accumulates in the cell envelope of F⁺strains in a process that requires interactions between its precursor (the traA gene product) and other host and F-encoded proteins. Here, we have used a set of (traA-phoA) genes to explore the effects of different TraA domains on the synthesis and membrane insertion of TraA-PhoA fusion proteins, particularly in relation to other F-encoded gene products. The 51-amino-acid TraA leader peptide fused directly to alkaline phosphatase was synthesized at comparable rates and incorporated rapidly and efficiently into the inner membrane in F' and F^? cells. A second fusion gene encoded the TraA leader peptide and the first 51 amino acids of F-pilin itself fused to PhoA (TraA'-'PhoA-102 polypeptide). Alkaline phosphatase activities and patterns of pulse-labelled polypeptides indicated that TraA'-'PhoA-102 was synthesized at comparable rates in F' and F^? cells, but in neither was the TraA'-'PhoA-102 polypeptide efficiently processed as a membrane protein. A third gene encoded the entire 121-amino-acid TraA polypeptide fused to PhoA (TraA-'PhoA-121 polypeptide). About 70% of the pulse-labelled TraA-'PhoA-121 polypeptide was rapidly processed in F'cells, where it accumulated in the cell envelope as active alkaline phosphatase, whereas in F- cells, >5% of the pulse-labelled polypeptide was processed. Additionally, the apparent rate of TraA-'PhoA-121 polypeptide synthesis was threefold higher in F'cells. The traQ gene alone could not substitute for F in restoring TraA-'PhoA-121 (or wild-type F-pilin) accumulation.     

Selective,reversible inhibition of the lactose phosphotransferase system of Staphylococcus aureus by dodecyl sulfate and deoxycholate

Low concentrations of sodium dodecyl sulfate (0.015%) and sodium deoxycholate (0.33%) completely inhibit phosphorylation of β-galactosides by the lactose phosphotransferase system of Staphylococcus aureus. Inhibition is reversible, even after prolonged detergent treatment. Phosphorylation of methyl-α-glucoside by the same preparations is only slightly inhibited by 0.015% dodecyl sulfate. The membrane-bound component, Enzyme IF^lac, is not solubilized by 0.015% dodecyl sulfate, nor is its ability to bind [¹⁴C]lactose affected. The results are consistent with hypotheses of selective binding of anionic detergent to Enzyme II^lac or to Factor III^lac, the detergent serving in the latter case as a membrane analog.     

The Escherichia coli K-12 F plasmid gene traX is required for acetylation of F pilin.

The Escherichia coli F plasmid gene required for amino-terminal acetylation of F-pilin subunits was identified. Using Western blots (immunoblots), we assayed the reaction of monoclonal antibodies with F-pilin polypeptides in inner membrane preparations from various F mutant strains. It was known that JEL92 recognizes an internal pilin epitope and JEL93 recognizes the acetylated amino-terminal sequence (L.S. Frost, J.S. Lee, D.G. Scraba, and W. Paranchych, J. Bacteriol. 168:192-198, 1986). As expected, neither antibody reacted with inner membranes from F- cells or Flac derivatives that do not synthesize pilin. Mutations that affected the individual activities of F tra genes traA, -B, -C, -D, -E, -F, -G, -H, -I, -J, -K, -L, -M, -N, -P, -R, -U, -V and -W or trb genes trbA, -B, -C, -D, -E, -G, -H, and -I did not prevent JEL92 or JEL93 recognition of membrane pilin. However, Hfr deletion mutants that lacked the most-distal transfer region genes did not express pilin that reacted with JEL93. Nevertheless, all strains that retained traA and traQ did express JEL92-reactive pilin polypeptides. Analysis of strains expressing cloned tra segments showed that traA and traQ suffice for synthesis of JEL92-reactive pilin, but synthesis of JEL93-reactive pilin is additionally dependent on traX. We concluded that the traX product is required for acetylation of F pilin. Interestingly, our data also showed that TraA+ TraQ+ cells synthesize two forms of pilin which migrate at approximately 7 and 8 kDa. In TraX+ cells, both become acetylated and react with JEL93. Preparations of wild-type F-pilus filaments contain both types of subunits.     

Role of surface exclusion genes in lethal zygosis in Escherichia coli K12 mating

Summary We find that diaminopimelic acid in the recipient membrane is released into the medium during bacterial matings, indicating that membrane damage was inflicted on the recipient by the donor, probably for forming a channel for DNA transfer. When the damage is extensive, as in matings with an excess of Hfr bacteria, the F^- bacteria are killed (lethal zygosis). The transfer of a large amount of DNA in Hfr matings appears to enhance the killing. In analogous F⁺xF^- (Nal^r) matings, on the other hand, killing of F^- bacteria does not occur unless F plasmid transfer is inhibited by a substance like nalidixic acid. The F^- bacteria are killed, suggesting that F plasmids contain genes that express immunity to lethal zygosis in the recipient. For example, bacteria containing surface exclusion-deficient mutants of F plasmids, such as traS ^- and traT ^-, induce lethal zygosis in F^- bacteria and are susceptible to it. Various tra ^- polar mutants that abolish surface exclusion are also susceptible to lethal zygosis when mated with Hfr bacteria. Kinetic experiments indicate that in F⁺ (wild type) x F^- matings, immunity to lethal zygosis is expressed in the F^- recipient within 1/4 division time, whereas a complete expression of surface exclusion requires more than 1 division time. Thus, a complete change in all receptor sites seems to be required for the expression of surface exclusion.     

Identification and characterization of four new tra cistrons on the E. coli K12 sex factor F.

Thirty-eight amber-suppressible transfer-deficient mutants of the Flac element F13-1 have been analyzed in conjugational complementation tests. Many of the mutations were in tra genes that have been identified previously (including traI), but several were in four new genes, traN, traU, traV, and traW. Results for representative mutations in the new genes were confirmed by complementation tests using λtra transducing phages and these also allowed the genes to be mapped quickly and simply. Approximate physical locations were found by complementation with a series of chimeric plasmids carrying various EcoRI fragments of the F transfer region. All four genes are located in the transfer operon, the order being traA … traBtraVtraWtraCtraUtraNtraF … traD. While traU, traV, and traW mutants were resistant to male-specific phages, traN mutants were sensitive.     

Growth and reactivation of single stranded DNA phage phiX174 in E. coli undergoing "thymineless death".

Summary Flac maintenance was aberrant at permissive temperature in a temperature-sensitive dnaC mutant of Salmonella typhimurium when the normally resident pLT2 plasmid was present. Flac was, however, efficiently transferred into the dnaC pLT2⁺ strain and the resulting Flac derivative was almost as efficient in transferring Flac as were dnaC ⁺ pLT2⁺ Flac strains indicating that aberrant Flac maintenance was not associated with appreciable inhibition of transfer replication. A range of F-like plasmids behaved like pLT2 in causing aberrant Flac maintenance when present in the dnaC pLT2^- strain. Flac was, however, stably maintained in the dnaC strain in the absence of other plasmids. Although the F-like plasmids destabilized Flac, each was stably maintained when introduced into strain 11G dnaC pLT2⁺ and pLT2 was also apparently stable under these conditions. The destabilizing effect of pLT2 and other fi ⁺ plasmids was not consequent upon their inhibiting the formation of a repressible F transfer component needed for Flac replication in the dnaC strain. Incompatibility between Flac and the other plasmids induced by the dnaC lesion also appeared unlikely to be a cause of the aberrant Flac maintenance. The possibility is discussed that the initiation of Flac replication differs from that of pLT2 and the F-like plasmids with F competing less effectively than the others for the DnaC gene product.     

Cloning of F DNA fragments carrying the origin of transfer oriT and the fertility inhibition gene finP

The two SalI fragments derived from the F transfer region that are bounded at one end by the SalI cleavage site in traM were cloned into pBR322. From these, smaller (540 bases) SalI-BglII fragments were subcloned to give plasmids containing the origin of transfer oriT (pED806) and finP (pED812), respectively, but no entire tra genes. All four plasmids were characterized by genetic tests and by restriction endonuclease analysis. pED806 could not be used to search for an F oriT-related “relaxation complex” because of its unexpected instability in the presence of Flac, and extensive efforts to prepare such a complex using other oriT⁺ plasmids were unsuccessful. We therefore suggest that a cell-free F relaxation complex does not exist. Protein synthesis directed by pED812 in minicells allowed the finP product to be tentatively identified as a 4000 M_r, protein.