INTRODUCTION

The most common mechanism for genetic exchange in bacteria is by conjugation.  Bacteria transfer DNA by conjugation not only between different species, but also into higher organisms such as plants and yeast.

 Antibiotic-resistance genes on plasmids are generally disseminated by conjugation. Transfer is so widespread that many antibiotics have now become therapeutically useless.

 Conjugation-like mechanisms are also involved in secretion (for example, the Bordetella pertussis toxin) and in plant tumors caused by Agrobacterium tumefaciens.

    In our lab, we are investigating the DNA processing that occurs during conjugation, with the objective of developing rational strategies to control conjugal transmission of genes in nature.

 Processing during transfer entails:

1)     Cleavage of one of the two DNA strands at a unique site called the origin of transfer (oriT). 

2)     Transport of this strand, in the 5’ to 3’ direction, through an intercellular pore.

3)     Rejoining of the ends of the linear strand in the recipient cell.

4)     Replacement synthesis of the complementary DNA strand in the recipient and donor.

    A single protein, MobA, is involved in the major DNA processing steps.  MobA is a strand transferase that remains covalently bound to the 5’ end

    The MobA protein both “nicks” double-stranded oriT DNA in the donor, and rejoins single-stranded oriT DNA in the recipient.

    Our work has revealed an underlying unity in the structure of the DNA substrate in both cases.  The substrate is partly double-stranded and partly single-stranded in character.

    In the donor cell, a single-stranded domain is created because MobA, along with another protein, separates the DNA strands.  This can be detected by increased sensitivity of the DNA bases in this region to oxidation by permanganate.  The DNA is actually cleaved in the single-stranded region.

    At the end of a round of transfer, the DNA is already single-stranded.  The double-stranded domain is recreated by base-pairing between the arms of an inverted repeat, part of the oriT DNA.

    What are the rules for recognition of oriT DNA by MobA, and for cleavage and rejoining of the transferred strand by this protein?  We have used biochemical and genetic approaches to answer these questions.

    Binding of MobA to an oriT oligonucleotide causes decrease in mobility of the DNA during electrophoresis.  We have used a partially degenerate oligonucleotide (ie with a mixture of different base sequences) to identify the subpopulation able to bind MobA.

    Our results show that only the inverted repeat and the adjacent sequence TAA is required for binding.

What is required to initiate and terminate a round of transfer – to cleave the DNA and then rejoin the ends?

   We have developed a genetic assay to answer these questions.  A plasmid was constructed with two copies of oriT separated by the lac operator.  Because of the large number of copies of the lac operator, the lac operon is induced, and colonies become blue in the presence of the indicator X-Gal. When transfer is initiated at oriT(2) and terminated at oriT(1), the lac region is deleted, and transconjugants are now white in the presence of X-Gal.  If mutations placed in oriT(2) affect initiation of transfer, then they will reduce the proportion of white colonies formed after transfer.  Similarly, mutations that affect termination will also reduce the proportion of white colonies when placed in oriT(1).

   We have found the consensus sequence for termination and initiation of transfer.  This consensus can be shifted from the inverted repeat domain by one base:  thus, there is some flexibility in the positioning of the single- and double-stranded domains of oriT.

The mechanics of strand separation:

   MobC is a small protein that assists MobA in separating the DNA strands at oriT.  How does it do this?  Although MobC does not cause a mobility shift in a gel, in the cell it increases the supercoiling of plasmid DNA.  We believe this is because it does bind to oriT, causing a localized decrease in underwinding of the strands.  This is sensed by DNA gyrase in the cell, which then adds more supercoils.  As a result, when the DNA is extracted from the cell, it is more supercoiled than normal.  We are now testing this interpretation.

Recognition of the conjugal transporter:

    Supercoiled DNA is cumbersome to work with biochemically, and MobA does not bind to linear, double-stranded DNA.  We have found, however, that MobA will bind to a small, double-stranded oriT oligonucleotide if it has several base pair mismatches.  We are using such a substrate to study in detail how MobA can “unzip” the two strands, allowing it to then cleave one strand, and how MobC helps in this process.

      How does plasmid DNA recognize and contact the molecular machine that transports it from cell to cell during conjugation?  Productive recognition is very specific:  different plasmids interact efficiently with some transporters but not others.  We have obtained genetic evidence suggesting that a domain of MobA protein is involved in this recognition.  We will display protein fragments of the transporter machine on phage, and then identify the site of interaction by biopanning against MobA.  Phage particles that display proteins allowing them to bind to MobA are enriched.  The cloned gene fragment responsible can then be identified by sequencing.

 Replication and transfer:

 Since only one DNA strand is transferred, the complementary strand must be synthesized in the recipient cell, and there is probably replacement strand synthesis in the donor cell as well.  What are the properties of conjugation-specific DNA replication?  This has been a difficult question to answer, since there is ongoing vegetative replication in the cell.

We have developed a means to look specifically at replication during conjugation, by having plasmid transfer take place under conditions where vegetative replication is strictly inhibited.  This involves electroporation of DNA into cells that can support transfer but not vegetative replication, and then immediately mating these cells with a strain that encodes that lambda integrase.  After transfer, the incoming plasmid is rescued by integration into the chromosome, and there is no vegetative replication in donor or recipient cells. 

Using this system, we have found that a plasmid-encoded primase is required for transfer.  Presumably, this protein is required for complementary strand synthesis in the recipient cell, but surprisingly, it must be provided in the donor cell.  This hints at some new and unexpected role of replication during transfer.  Possibly, replication is part of the “trigger” that directs DNA into the transfer machinery.