SEARCH

SEARCH BY CITATION

Keywords:

  • relaxosome;
  • relaxase;
  • mobilization;
  • conjugation;
  • coupling protein;
  • origin of transfer

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Comparison of F- and P-like conjugative systems
  5. Stages in the conjugative DNA process
  6. Relaxases, an overview
  7. MOBF relaxosomes
  8. MOBP relaxosomes
  9. MOBQ relaxosomes
  10. A model for conjugative DNA processing
  11. Concluding remarks
  12. Acknowledgements
  13. References

Bacterial conjugation in Gram-negative bacteria is triggered by a signal that connects the relaxosome to the coupling protein (T4CP) and transferosome, a type IV secretion system. The relaxosome, a nucleoprotein complex formed at the origin of transfer (oriT), consists of a relaxase, directed to the nic site by auxiliary DNA-binding proteins. The nic site undergoes cleavage and religation during vegetative growth, but this is converted to a cleavage and unwinding reaction when a competent mating pair has formed. Here, we review the biochemistry of relaxosomes and ponder some of the remaining questions about the nature of the signal that begins the process.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Comparison of F- and P-like conjugative systems
  5. Stages in the conjugative DNA process
  6. Relaxases, an overview
  7. MOBF relaxosomes
  8. MOBP relaxosomes
  9. MOBQ relaxosomes
  10. A model for conjugative DNA processing
  11. Concluding remarks
  12. Acknowledgements
  13. References

Since the discovery of ‘bacterial sex’ over 60 years ago (reviewed in Tatum & Lederberg, 1953) considerable progress in understanding the mechanism of bacterial conjugation, one of the processes underlying genetic recombination in bacteria, has been made. We now understand the nature of the conjugative pore (Fronzes et al., 2009) and have an approximate idea of how the process works, although we are still far from knowing the structure encoded by ‘transfer regions’, which is responsible for the transfer of genetic material between bacteria. Furthermore, we have not reconciled conjugation with other processes in the cell such as replication, partitioning, cell division, DNA silencing or alterations in the composition of the cell envelope in response to changes in the physiology of the cell. In general, much of our current knowledge is based on in vitro studies or in vivo studies using bacteria grown in a rich medium and sampled at a point in mid-exponential growth. As we move forward in our understanding of this process, integrating conjugation into the life history of the organism will become of greater importance.

Conjugation appears to be a universally conserved process among prokaryotes and is now also being intensively studied in Gram-positive and archaeal species. Conjugative elements have expanded to include not only plasmids, but also conjugative transposons and integrated conjugative elements that are related to bacteriophages. It is also now being examined under many physiologically relevant conditions including biofilms, stress and fluctuations in temperature, osmolality, pH, etc. These new research areas have revealed many exciting and often unanticipated findings; recent reviews on bacterial conjugation in general are Zechner et al. (2000) and Lawley et al. (2004).

In Gram-negative bacteria, conjugation requires the elaboration of a pilus in the donor cell that is assembled by a type IV secretion system (T4SS), called the transferosome. This complex spans the cell envelope and is involved in pilus assembly and retraction, identification of a suitable recipient cell and signaling that DNA processing and transfer should begin in a process called mating pair formation (Mpf). Events in the recipient cell or transconjugant, newly converted to plasmid-bearing status, are less well understood but are very important for the establishment of the incoming DNA. The nucleoprotein complex involved in processing the DNA [DNA transfer replication (Dtr)] and delivering it to the transferosome is called the relaxosome (Fürste et al., 1989), which contains the relaxase (Byrd & Matson, 1997) and auxiliary transfer proteins bound to the origin of transfer (oriT), a stretch of DNA usually of several hundred base pairs (bp) in length. Within oriT is the nic site where the relaxase, a phosphodiesterase, cleaves in a site- and strand-specific manner, resulting in the transfer of a single-stranded DNA (ssDNA) in a 5′–3′ direction to the recipient cell. The ‘coupling protein’ (T4CP) is required for conjugation because it links the relaxosome to the T4SS/transferosome and is thought to form the channel, or conjugative pore, through which the DNA passes on its way to the recipient cell (Gomis-Rüth et al., 2004).

Comparison of F- and P-like conjugative systems

  1. Top of page
  2. Abstract
  3. Introduction
  4. Comparison of F- and P-like conjugative systems
  5. Stages in the conjugative DNA process
  6. Relaxases, an overview
  7. MOBF relaxosomes
  8. MOBP relaxosomes
  9. MOBQ relaxosomes
  10. A model for conjugative DNA processing
  11. Concluding remarks
  12. Acknowledgements
  13. References

The publication of the DNA sequences of the F and RP4 transfer regions in 1994 (Frost et al., 1994; Pansegrau et al., 1994) led to the realization that the gene products for F and RP4 conjugative pilus synthesis and assembly were similar to those for tumorigenesis in the Ti plasmid (Shirasu & Kado, 1993). This was followed by the identification of homologs to these plasmid transfer gene products in many bacterial mobile elements and chromosomes, which have been attributed to diverse functions including DNA and protein export and import and have been grouped into the T4SS (Chen et al., 2005). The gene products encoded by the virB, -C, -D, -E and -F operons of the Ti plasmids that encode genes for tumorigenesis in plants and are homologous to gene products in conjugative transfer systems are now used as the standard nomenclature for proteins involved in T4SS and conjugative DNA metabolism (Schröder & Lanka, 2005). Research on T4SS, especially conjugative pores involved in pilus assembly and DNA transfer, has given us a basic understanding of the structure of the pore and hints to how it functions although a detailed mechanism remains unknown. Excellent recent reviews on T4SS include Christie et al. (2005) and Lawley et al. (2003b).

In Gram-negative bacteria, conjugative T4SS fall into two main categories: P-type and F-type conjugative pores with P-T4SS having two ATPases (TrbB and TrbE; VirB11 and VirB4 families, respectively), whereas F-T4SS have one (TraC/VirB4) with the latter system also encoding a separate envelope-spanning apparatus characterized by cysteine-rich outer membrane proteins, which appears to mediate mating pair stabilization, pilus assembly and retraction Lawley et al., 2003b). In parallel to these studies, detailed biochemical characterization of relaxosomes has been carried out for representatives of self-transmissible broad (IncP, IncW) and narrow (IncF) host range plasmids as well as mobilizable plasmids (ColE1, IncQ) (Fig. 1). The key component that links the T4SS to the relaxosome is the coupling protein (T4CP; VirD4 family), an inner membrane ATPase that is the conduit for DNA transfer. T4CPs are thus crucial determinants of specificity and functionality of the conjugative pore. Examining these conjugative systems from an evolutionary perspective reveals that the T4SS and relaxosome components probably came from different sources. For instance, IncW plasmid R388 encodes an F-like relaxase coupled to a P-T4SS (Bolland et al., 1990), whereas IncH plasmid R27 has a P-like relaxase coupled to a combination of P-like pili and an F-like T4SS (Lawley et al., 2003a).

image

Figure 1.  Overview of the organization of IncQ (R1162), IncF (F), and IncP (RP4) oriT regions. ORFs (italicized) are indicated by open arrowheads within boxes. Large arrows indicate the direction of transfer. Vertical wedges indicate the nic site. (a) Organization of the R1162 mobilization and replication regions (Scholz et al., 1989). The region encoding mobB that overlaps mobA is indicated by a black box. The origin of vegetative replication (oriV) has two arrowheads indicating bidirectional replication. Below is the sequence of the 38-bp (only the nicked strand is shown) oriT region. The inverted arrows indicate the site of MobA recognition, binding and termination. (b) An abbreviated F transfer region illustrating the positioning of the members of the relaxosome based on Frost et al. (1994). The site of binding for TraI (sbi, black box), TraY (sbyA, -B, -C, clear boxes), TraM (sbmA, -B, -C; speckled boxes) and TraJ (sbj, checked box) are indicated. The major promoters are indicated by hooked arrows that represent the transcriptional start sites. The slanted lines indicate an interruption in the sequence of 21.5 kb (traA to traT). The trbH gene between traD and traI, and traX and finO after traI are not shown. A detailed view of the oriT region (bracket) is given in Fig. 2. (c) A depiction of the oriT region of RP4 based on Pansegrau et al. (1994). The binding sites for TraK are shown as speckled boxes with black indicating strong binding and grey indicating binding at high TraK concentrations. The three promoters associated with expression of the relaxosome components of RP4 are shown as hooked arrows. A detailed view of the oriT region (bracket) is given in Fig. 3.

Download figure to PowerPoint

The step-wise process of DNA transfer has a number of checkpoints to ensure the efficiency and specificity that is the hallmark of Gram-negative conjugation. Under optimum conditions, nearly 100% of a recipient cell culture can be converted to transconjugants within 30 min after the addition of donor cells at 37 °C. Mpf requires about 5–7 min and DNA transfer occurs at 45 kb min−1 under these conditions (Lawley et al., 2004). Thus, one can imagine that active growth would favor conjugation by supporting the energy expenditure required by this process. The timing of conjugative transfer must involve coordination with vegetative replication and partitioning of the plasmid to defined positions relative to the mid-point in the cell. A signal that a mating pair has formed must be transmitted through the transferosome and coupling protein to the relaxosome to initiate transfer. Following this signal, a cascade of processes occurs that results in the transfer and recircularization of the ssDNA in the recipient cell, which will be discussed in detail below. After decades of effort, no one has been able to trigger DNA release into the medium using a conjugative system, suggesting that a single mutation will not be sufficient to allow this to happen. In this review, we wish to unite the data from many studies over the last 20 years on the biochemistry of relaxosomes, including their interaction with the T4CP in order to better predict whether a T4SS is capable of conjugative DNA transport, thereby contributing to lateral gene transfer.

Compared with many prokaryotic and viral models of replication initiation, elucidation of the molecular mechanisms initiating conjugative DNA processing and replication has lagged behind. One reason is that many conjugation systems do not encode a helicase enzyme dedicated to the unwinding of the DNA in preparation for transfer but do encode an ATPase (VirB11 family) that is involved in processing the DNA for transfer (Schröder & Lanka, 2005). A notable exception is the IncF–IncW family of DNA mobilizing systems. The fusion of a helicase to the site- and strand-specific DNA relaxase domains within TraI (F) and TrwC (R388) presents these proteins with intriguing regulatory challenges. Since phosphodiesterase nicking and religation activities are ongoing within the cell and helicase activity is initiated only in response to a signal that mating should begin, a coordinated progression of events within the conjugative initiation pathway must exist. Thus, these systems appear to share similarities with mating systems that lack a conjugation-specific helicase but also must differ from them in important ways.

Stages in the conjugative DNA process

  1. Top of page
  2. Abstract
  3. Introduction
  4. Comparison of F- and P-like conjugative systems
  5. Stages in the conjugative DNA process
  6. Relaxases, an overview
  7. MOBF relaxosomes
  8. MOBP relaxosomes
  9. MOBQ relaxosomes
  10. A model for conjugative DNA processing
  11. Concluding remarks
  12. Acknowledgements
  13. References

Initiation of chromosome replication requires controlled access to the DNA duplex to produce an unwound region for entry of the replication machinery. Usually, this task is performed by initiator proteins, typically in a multistep process. Initiator melting reactions share features including multiple DNA-binding sites and regions of low thermodynamic stability. Initiator proteins occupying these sites loop, wrap, encircle and bend DNA to provoke the required duplex distortion. During F-type DNA transfer initiation, nick cleavage must occur as well as recruitment of the conjugative helicase to the transferosome. These steps enable an unwinding of the plasmid genome, which in turn renders the transferred DNA-strand (T-strand) ready for translocation in its single-stranded form and exposes the retained complementary (R-strand)-strand as a template for replacement strand synthesis by DNA polymerase III. Accordingly, an initiator complex for conjugative replication has several basic functions to fulfill:

  • 1
    recognition: preparation of oriT by binding auxiliary proteins;
  • 2
    relaxation: nicking by the oriT-specific relaxase;
  • 3
    prereplication complex assembly: loading the DNA helicase; and
  • 4
    commitment to translocation: assembly of additional proteins to start T-strand transfer and replacement strand synthesis.

Origin recognition is conferred by the specific binding of dedicated oriT-specific auxiliary factors within relaxosomes. Typically multiple binding sites for one, or frequently more, plasmid-specific initiator proteins are encoded near an oriT region that is characterized additionally by regions of thermal instability. Genetic and biochemical approaches have discerned that relaxosomes of closely related plasmids interact most productively with their cognate oriT regions. Functional specificity is predicted to involve differential affinity among related auxiliary proteins for noncognate DNA-binding sites. In the case of the MOBF family, optimized protein–DNA interactions have been shown to be important determinants of specificity during relaxosome formation (Klimke et al., 1998; Kupelwieser et al., 1998; Fekete & Frost, 2000; Zechner et al., 2000; Harley & Schildbach, 2003).

Coevolution of a suite of effectors offers several advantages as a general strategy for ensuring specificity. These include gaining or maintaining large surface areas on relaxosome proteins to facilitate multiple protein–protein and protein–DNA interactions, generating reciprocal mutations within interaction domains to increase specificity, and regulating the concentrations of these effectors to take advantage of affinity parameters. Coevolution of the relaxase active site and its cognate nic site would further refine specificity within these systems. The quaternary structure of the relaxosome also confers recognition features that allow for selection by its cognate T4SS.

The DNA phosphodiesterase reaction of conjugative relaxases (VirD2 family) has been intensively studied. Biochemical reconstitution of relaxosomes in vitro has provided a wealth of insights into the contribution of auxiliary factors to the nicking reaction. Seminal studies carried out with MOBP and MOBQ prototype plasmids were key to early advances in the field (Pansegrau et al., 1988; Scherzinger et al., 1992). Developments of these systems were closely followed by work in other paradigms. However, progress on understanding the relaxosome in vitro has not been matched by an understanding of events in vivo. The problem is manifest on both genetic and biochemical levels. Genes essential to conjugative DNA processing are rapidly identified through genetic approaches, but distinguishing between a minimal and an optimal initiator complex remains challenging. Moreover, since relaxases universally catalyze nick cleavage reactions in vitro on supercoiled oriT DNA in the absence of auxiliary proteins, the activity of host- and plasmid-encoded factors required during the initiation cascade in vivo will not necessarily be detectable in vitro. Currently, translating the detailed information on relaxosomes in the MOBF,P,Q families in vitro to the in vivo situation is the biggest challenge in the field. Here, we summarize progress in understanding relaxosomes in these families. Our detailed knowledge of these systems belies our lack of understanding about triggering DNA unwinding, with or without a dedicated helicase, in vivo, as well as understanding the role of the T4CP and VirB11 homologs in DNA processing, T4SS recognition and transfer.

Relaxases, an overview

  1. Top of page
  2. Abstract
  3. Introduction
  4. Comparison of F- and P-like conjugative systems
  5. Stages in the conjugative DNA process
  6. Relaxases, an overview
  7. MOBF relaxosomes
  8. MOBP relaxosomes
  9. MOBQ relaxosomes
  10. A model for conjugative DNA processing
  11. Concluding remarks
  12. Acknowledgements
  13. References

Short history

Conjugative relaxase proteins have the capacity not only to interact with the origin nick site carried by their cognate plasmids with remarkable affinity but also to cleave that site in the absence of additional proteins. Nonetheless, independent initiation of conjugative plasmid strand transfer by a relaxase alone appears to occur either very rarely or not at all. Instead the majority of plasmid prototypes assemble a relaxosome known to contain two or three proteins of host or plasmid origin in addition to the relaxase. The designation of a conjugation gene to Dtr, or the assignment of auxiliary proteins to a relaxosome, is usually achieved through assays detecting the efficiency of intracellular relaxase-catalyzed nic cleavage. Seminal observations revealing the existence of (nic-cleaved) ColE1 plasmid relaxation complexes made in the laboratory of Donald Helinski at the onset of the 1970s opened the field to biochemical and molecular genetics investigation of conjugative DNA processing (Clewell & Helinski, 1969; Lovett & Helinski, 1975). Alternative assays for detecting intracellular oriT strand cleavage led also to the identification of auxiliary factors acting in the MOBF and MOBQ nicking reactions (Everett & Willetts, 1980; Zhang & Meyer, 1995, 1997; Zechner et al., 1997; Karl et al., 2001).

Studies on ColE1 peaked in the 1970s with the report that a relaxation complex had been isolated with proteins of molecular mass comparable to those of MbeA (57.7 kDa), MbeB (19.5 kDa) and MbeC (12.9 kDa) (Lovett & Helinski, 1975), with the sizes being determined by DNA sequence analysis (Boyd et al., 1989). This study also revealed that the sequences for MbeB and a new mobilization protein, MbeD, overlapped with that for MbeA. MbeA was also shown to be associated with the 5′ end at a nic site, presaging investigations into relaxases of large self-transmissible plasmids (Guiney & Helinski, 1975).

ColE1 and pMB1, a closely related plasmid, were destined to become the basis for some of the most popular cloning vectors ever developed. In order to reduce the size of the vector, the mobilization regions of these plasmids were deleted. While interest in ColE1 mobilization waned (but see below), a new group of plasmids that were extremely broad in their host range were discovered – the IncQ plasmids. The two best-known representatives are RSF1010 (Derbyshire & Willetts, 1987; Derbyshire et al., 1987; Scholz et al., 1989) and R1162 (Brasch & Meyer, 1987), which are virtually identical. Like ColE1, the mobilization region of IncQ plasmids is tightly organized with three essential mobilization proteins and a mob (nic) site that involves overlapping ORFs (Fig. 1a). The MobA protein, encoding the relaxase at its proximal end, was found to overlap with MobB and encodes a primase domain at its C-terminal end that was also translated separately to give RepB′. Although it functions in vegetative replication, it is also active in mobilization (Henderson & Meyer, 1999).

Two paradigms for relaxosomes of self-transmissible plasmids were well established by the early 1990s – F and RP4. However, both F and RP4 encoded complex transfer regions with many genes encoding products of unknown function (Fig. 1b and c). A simpler T4SS was seen as highly desirable. Two plasmids in particular filled the bill: R388, an IncW plasmid with a straightforward T4SS and relaxosome (Bolland et al., 1990), and the Ti plasmid, which also had a compact T4SS and supported Dtr from a bacterium to a plant cell during tumorigenesis (reviewed in Christie & Cascales, 2005). The latter system does not have a particularly simple Dtr process but does benefit from the ability to easily distinguish the donor bacterium from the recipient plant cell. Thus, R388 has provided an excellent system for relaxosome function whereas the Ti vir system has been instrumental in understanding the structure and function of the T4SS.

Relaxase activities

Table 1 shows a list of relaxases representative of enterobacterial plasmids. They are ordered in families that represent the most studied mechanisms of conjugative DNA processing. For the purpose of this review, we concentrate most of our discussion on R388 TrwC and F TraI as representatives of MOBF relaxases, and RP4 TraI and RSF1010 MobA as representatives of MOBP and MOBQ relaxases, respectively (see Garcillán-Barcia et al., 2009; Francia et al., 2004, for reviews on relaxase diversity). Among the six relaxase protein families, MOBF and MOBQ constitute the only two families for which detailed structural and biochemical data are available whereas extensive biochemical data exist for the MOBP family. The other three families are still relatively unknown in their molecular mechanisms.

Table 1.   Classification of prototype relaxases of enterobacterial plasmids
Plasmid familyCladePrototype relaxase plasmid (incompatibility group)Representative reference
  1. The table shows the MOB classification (Francia et al., 2004; Garcillán-Barcia et al., 2009) of representative relaxases belonging to plasmids isolated from enterobacteria, which gave names to the classical incompatibility groups. Relaxases of mobilizable plasmids were in general not tested for incompatibility. Nevertheless, RSF1010, ColE1 and CloDF13 are compatible among themselves and compatible with all classical Inc groups.

MOBFF11TrwC_R388 (IncW)González-Pérez et al. (2007)
TraI_pKM101 (IncN)Paterson et al. (1999)
F12TraI_F (IncF1)Hekman et al. (2008)
TraI_R1 (IncFII)Csitkovits et al. (2004)
MOBHH11TraI_R27 (IncHI)Lawley et al. (2002a, 2003a)
H12TraI_pIP1202 (IncA/C)Welch et al. (2007)
TraI_R391 (IncJ)Böltner & Osborn (2004)
TraI_Rts1 (IncT)Murata et al. (2002)
MOBCC1MobC_CloDF13 (mobilizable)Núñez & de la Cruz (2001)
MOBQQ1MobA_RSF1010 (IncQ; mobilizable)Parker & Meyer (2007)
MOBPP11TraI_RP4 (IncP1)Pansegrau & Lanka (1996)
P12NikB_R64 (IncI1)Furuya & Komano (1991)
NikB_R387 (IncK)Tschäpe & Tietze (1980)
P13pCTX-M3 (IncL/M)Gołebiewski et al. (2007)
P14MobA_Rms149 (IncG)Haines et al. (2005)
P3TaxC_R6 K (IncX)Núñez et al. (1997)
P4VirD2_pFBAOT6 (IncU)Rhodes et al. (2004)
P5MbeA_ColE1 (mobilizable)Varsaki et al. (2003)
P6NikB_R721 (IncI2)Komano et al. (1990)
MOBVV2Mob pBBR1Szpirer et al. (2001)

In general, relaxases are large proteins composed of several protein domains. For instance, MOBF relaxases typically contain a relaxase domain at the N-terminus, and a DNA-helicase domain at the C-terminus. Relaxases are DNA phosphodiesterases (mechanistically similar to topoisomerase I; Forterre & Gadelle, 2009), which catalyze a number of in vitro reactions that mimic some physiological functions in vivo:

Phosphodiesterase activity on ssDNA (oligonucleotide cleavage and strand-transfer reactions)

These are the simplest in vitro reactions. They require only the protein, the DNA substrate and a divalent metal ion. The result of the reaction is the formation of a protein–DNA adduct in which the reaction products do not dissociate. The phosphodiesterase activity is visualized on gels usually only after incubation with a protease and sodium dodecyl sulfate (SDS) that destroy the complexes, suggesting that the protein is tightly bound to the DNA, even after cleavage. This is confirmed by the crystal structures of protein–substrate complexes (Guasch et al., 2003). The adduct is formed when an activated tyrosine in the catalytic centre of the protein breaks the scissile bond and becomes covalently bound to the 5′-moiety of the oligonucleotide. The 3′-moiety of the oligonucleotide remains bound by strong noncovalent interactions. Interestingly, two different tyrosines (Y18 and Y26) can catalyze these reactions in R388 TrwC. These reactions are thought to occur in the termination step of DNA mobilization, as explained below and in González-Pérez et al. (2007).

Binding to supercoiled DNA containing a nic site

This can be detected by electrophoretic mobility shift assay and permanganate footprinting. As a consequence of binding, relaxases distort the DNA around nic, exposing a stretch of ssDNA that can be cleaved (Zhang & Meyer, 1995; Guasch et al., 2003; Mihajlovic et al., 2009). Binding does not require divalent metal ions, suggesting that stable complexes need not be covalent complexes. In vivo, the MOBF complexes appear to pre-exist (Zechner et al., 1997; Moncalián et al., 1999) so that conjugation can be triggered immediately upon contact with a suitable recipient cell.

Phosphodiesterase activity on double-stranded DNA (dsDNA)

While F TraI can cleave linear dsDNA in the presence of auxiliary proteins, R388 TrwC can only cleave supercoiled DNA. This reaction is the basis for the name ‘relaxase’ applied to these proteins. The product of the reaction is, again, a nucleoprotein adduct that can only be resolved in vitro by the addition of proteases and/or SDS. In the case of R388 TrwC, only Y18 can catalyze the cleavage reaction on SC DNA (Grandoso et al., 2000). Although nicking-auxiliary proteins help in this reaction, they are not absolutely required for MOBF relaxase activity, whereas they seem to be more crucial for MOBP relaxases (Garcillán-Barcia et al., 2009). Cleavage of supercoiled DNA probably mimics the initiation reaction in vivo. The minimal nic site is composed of the proximal arm of an inverted repeat and the spacer DNA to the cleavage site. Thus, recognition and cleavage of nic during initiation does not require the extrusion of a cruciform structure. It is sufficient that the relaxase recognizes and binds to the proximal arm in dsDNA and in doing so, distorts the helix enough so that it can now cleave the nic site (Lucas and de la Cruz, unpublished results).

MOBF relaxosomes

  1. Top of page
  2. Abstract
  3. Introduction
  4. Comparison of F- and P-like conjugative systems
  5. Stages in the conjugative DNA process
  6. Relaxases, an overview
  7. MOBF relaxosomes
  8. MOBP relaxosomes
  9. MOBQ relaxosomes
  10. A model for conjugative DNA processing
  11. Concluding remarks
  12. Acknowledgements
  13. References

Organization of MOBForiT regions

In F-like conjugative systems, the organization of the oriT region reveals much about DNA processing during transfer (Fig. 1b). The nic site is located within an approximately 350-bp oriT region and is bordered by divergent promoters for orf169 or geneX in F, encoding a lytic transglycosylase homolog, and traM, a relaxosomal protein (Frost et al., 1994; Koraimann, 2003). Orf169 is encoded by the first gene that enters the recipient cell; thus, it is the start of the leading region. MOBForiTs contain a number of repeats that define the binding sites for the relaxase and auxiliary proteins TraM, TraY as well as an AT-rich region that is important for strand opening. The DNA also contains two intrinsic bends and sites for IHF (integration host factor) binding (Tsai et al., 1990) that could contribute to the conformation of the oriT region as well as three predicted sites for H–NS binding and repression of the two traM promoters (Will et al., 2004; Fig. 2).

image

Figure 2.  A detailed view of the F oriT region (Fig. 1b) based on results from Tsai et al. (1990); Stern & Schildbach (2001); Di Laurenzio et al. (1992); Lum et al. (2002); and Fekete & Frost (2002). The numbering is according to Frost et al. (1994). Solid and dashed brackets indicate DNAse I protection footprints on the top and bottom strand for primary and secondary binding sites, respectively. Boxes containing sequences on both strands indicate recognition sequences whereas double lines above and below the sequence indicate protected residues identified by hydroxy radical footprinting (sbyA and sbmA). Boxes within a single line of sequence indicate protection from hydroxyl radicals in sbmC or the start codon for traM. Hooked arrows indicate the start site for the two traM transcripts. Direct and inverted repeats within sby and sbm sites are indicated with arrows between the lines or below the line for sbmA. See text for details.

Download figure to PowerPoint

The precise nic site in F was identified by in vivo (Thompson et al., 1989) and in vitro assays (Matson & Morton, 1991) and attributed to TraI or DNA helicase I (DHI), as it was then known (Traxler & Minkley, 1988). Other F-like plasmid oriT regions have also been characterized in detail including R100 Inamoto et al., 1991), R1 (Schwab et al., 1991), pED208 (Di Laurenzio et al., 1991) and summarized in Frost et al. (1994). Whereas the nic site is conserved, small sequence changes in the adjacent 9 bp and inverted repeat define the specificity of TraI recognition and cleavage (Harley & Schildbach, 2003). The AT-rich region is present in all these plasmids as are binding sites for TraM, TraY and IHF although the order and number of binding sites vary among these plasmids (Frost et al., 1994). In common with F, all three plasmids have predicted bends immediately adjacent to nic followed by a predicted IHF site that has been demonstrated for R100 (Dempsey & Fee, 1990) and pED208 (Di Laurenzio et al., 1995). Further upstream of the IHF-binding site is a TraY-binding site in F (Nelson et al., 1993) and R100 (Inamoto & Ohtsubo, 1990). The combination of the intrinsic bend and the binding of IHF, which bends DNA at an angle of 123–140° or more (Rice et al., 1996; Seong et al., 2002), could bring TraY into close proximity with the nic site.

The nic site of F lies between nucleotides 140 and 141 on the T-strand according to the numbering of Frost et al. (1994) (Fig. 2). Immediately 5′ to nic is a 9-bp sequence (5′-G1TGGGGTGT9|GG-3′ where | is the nic site) that determines the specificity of cleavage by the relaxase. Mutation to 5′-A1TGGGATGT9-3′ resulted in no cleavage (Thompson et al., 1984), whereas mutation to 5′-G1TAGTGTGT9-3′ switched the specificity of cleavage from the F TraI to the R100 TraI relaxase (Harley & Schildbach, 2003) (mutations are italicized). These findings were confirmed using a truncated F TraI derivative TraI36 (36 kDa), corresponding to the relaxase domain (Stern & Schildbach, 2001). Using a series of overlapping 22-base ssDNA oligonucleotides, these authors found that TraI36 required at least 13 bases 5′ and two bases 3′ to nic and bound with a dissociation constant (KD) of 0.5 nM. A G to C mutation (G144′C or G6C in the sequence above and italicized below) reduced the KD over 7000-fold. When this mutation was introduced into an oriT-containing plasmid, mobilization was similarly affected. In a subsequent study, the oligonucleotide was extended to 30 bases in length to include the underlined inverted repeat that is adjacent to the nic region (5′-GCAAAAAC-TTGTTTTTGCGTGGGGTGT|GGT-3′; KD=7 nM) and is associated with cleavage and termination (Gao et al., 1994; Williams & Schildbach, 2006). The sequence of the oligonucleotides rather than formation of the hairpin appeared to be important for binding, cleavage and mobilization (Williams & Schildbach, 2006). Thus, the G144′C/G6C mutation within the 30-base oligonucleotide did not affect binding but did affect mobilization. The authors postulate that TraI36 binds the nic region in two different modes and reason that this might result in different affinities and therefore outcomes during cleavage and termination that agrees with the observations outlined in Phosphodiesterase activity on double-stranded DNA (dsDNA) above.

MOBF relaxases and helicases

The first relaxase investigated in detail was DHI, which was characterized by the Hoffmann–Berling laboratory as one of three Escherichia coli helicases (reviewed in Geider & Hoffmann-Berling, 1981). They also reported that DHI was a fibrous, asymmetric protein that required about 200 bases of ssDNA bound to 70–80 molecules of DHI to initiate unwinding. The high ratio of DHI : DNA probably reflects the lack of auxiliary proteins in the reaction mixture that we now know to be so important in strand unwinding, nicking and initiating DNA transfer. They later identified DHI as TraI encoded by the F plasmid and described its ability to nick the DNA at oriT and remain covalently bound to the 5′ end. This reaction required a supercoiled template and Mg2+ (Reygers et al., 1991). The location of the nick was confirmed by mapping the cleavage site on oriT-containing plasmids isolated from cells expressing the F transfer (tra) operon (Thompson et al., 1989). DHI was then assigned the additional activity of a relaxase, and the bifunctional nature of this protein elicited renewed interest in this enzyme. The purified DHI exhibits properties in vitro consistent with its essential function in conjugative DNA-strand transfer (Matson et al., 2001) including a very rapid 1100 bp s−1 rate of duplex unwinding, high processivity supporting ≥850 bp of unwinding and a 5′–3′ directional bias (relative to the strand to which it is bound) (Lahue & Matson, 1988; Sikora et al., 2006). Together these features should readily support the observed rate of translocation of the E. coli chromosome (4.6 Mbp in 100 min) originating from an integrated copy of the F plasmid, as well as concomitant replacement synthesis of the mobilized T-strand from the 3′ OH product of nic cleavage.

The structure of the relaxase domain of F TraI, TraI36, in complex with a 22-base ssDNA oligonucleotide containing nic, has been solved (Datta et al., 2003). TraI contains an HUH motif (two histidines bracketing a hydrophobic residue) that is shared by most relaxases involved in conjugative transfer or rolling circle replication (Garcillán-Barcia et al., 2009). The structure revealed a cleft lined with neutral rather than the more usual charged or aromatic amino acids (aa) and suggested a ‘knob-into-hole’ mechanism for base recognition. Datta et al. (2003) hypothesized that the ssDNA forms a U-shaped loop or ‘knob’ with two nucleotides G4 and T6 of the sequence 5′-T1GGGGTGT|GG10-3′ fitting into the pockets or ‘holes’ within the deep binding cleft. A Mg2+ ion, which is essential for cleavage, was liganded by three histidines within the hole: H146, which is not important for binding or cleavage, and H157 and H159 within the HUH motif, which are essential for cleavage. The loop is stabilized by unusual intramolecular interactions among three nucleotides G5, G7 and T8, which define the specificity of TraI for its nic site and position the scissile phosphate on G9 for cleavage and ligation to Y16 of TraI (Hekman et al., 2008).

F TraI (1756 aa; 182 kDa) can be separated into three domains with 1–306 aa defining the relaxase, 309–1504 the helicase (Byrd et al., 2002) and a third domain at the C-terminus (defined by a deletion of the last 252 aa) that binds to ssDNA (Guogas et al., 2009) and to TraM (Ragonese et al., 2007). Portions of the protein required for ssDNA translocation (L. Dostal and J. Schildbach, pers. commun.) and for its selective recognition by the TraD coupling protein (S. Lang and E. Zechner, unpublished data) have recently been mapped. All of the TraI domains are required for efficient transfer (Traxler & Minkley, 1988; Matson et al., 2001; Matson & Ragonese, 2005; Guogas et al., 2009) and the relaxase domain has been the target for inhibition by phosphonates, a possible route for preventing conjugation (Lujan et al., 2007).

In ColE1, the relaxase, MbeA (60 kDa), catalyzes DNA cleavage at the ColE1 nic site, which was mapped to 5′-CTGG/CTTA-3′ (position 1469–1462). MbeA is unusual in that it lacks a three-histidine motif that is a hallmark of conjugative relaxases even though it bears sequence similarity to the relaxase family. Residue Y19 (motif I) is involved in the nicking reaction whereas H97, E104 and N106 are equivalent to the histidine triad in Motif III. Thus, MbeA varies from most relaxases and carries a HEN motif, which might help identify related relaxases in other systems (Varsaki et al., 2003).

MOBF auxiliary proteins

The auxiliary proteins TraY and TraM of the IncF plasmids are the best studied of the proteins required for DNA processing in preparation for conjugative transfer. Interestingly, both proteins also play a role in regulating gene expression, with TraY regulating the Py promoter at the head of the F tra operon either positively (Silverman & Sholl, 1996) or negatively as in R100 (Taki et al., 1998) and stimulates the two F traM promoters, Pm1 and Pm2 (Penfold et al., 1996). TraM autorepresses both of its promoters in F (Penfold et al., 1996), R1 (Schwab et al., 1993) and R100 (Abo et al., 1991) (Fig. 2).

One or more TraY-binding sites (discussed below) are followed by two to four binding sites for TraM in combination with one (F, R100) or more (pED208) additional IHF-binding sites in R100 (Abo & Ohtsubo, 1995), R1 (Schwab et al., 1991) or pED208 (Di Laurenzio et al., 1995). The TraM-binding sites are usually represented as direct or inverted repeats (Frost et al., 1994). However, closer examination of the sequences reveals multiple short repeat sequences that are present on both strands. The most obvious examples are R1 (Schwab et al., 1991, 1993) and pED208 (Di Laurenzio et al., 1991), which have variations on the sequence GATTC repeated 15–16 times spaced 11–12 bp apart between nic and the traM promoter. F (Di Laurenzio et al., 1992; Fekete & Frost, 2002) and R100 (Abo & Ohtsubo, 1995) have fewer short repeats but they also occur 11–12 bp apart.

TraY is structurally related to the ribbon–helix–helix (RHH) family of DNA-binding proteins that dimerize and allow the β-strands or ‘ribbons’ to interact in an antiparallel fashion and contact the major groove of the DNA (Bowie & Sauer, 1990; Lum & Schildbach, 1999; Lum et al., 2002). F TraY binds homopurine–homopyridine [d(GA)n : d(CT)n] (Krasilnikova et al., 2001) and bends the DNA at an angle of ∼50–55° (Luo et al., 1994). TraY is a small (75 aa or 131 aa for F TraY) protein that is required for relaxase binding and nicking in preparation for conjugative DNA transfer in F (Nelson et al., 1995), R1 (Karl et al., 2001) and R100 (Inamoto et al., 1994). R100 TraY binds to one site in oriT, sbyA, as well as two sites near the Py promoter, sbyB and sbyC, probably as a dimer (Taki et al., 1998). It recognizes a TAA(A/T)T motif that occurs five times within sbyA (Inamoto & Ohtsubo, 1990). Inamoto et al. (1988) suggested that the F TraY (131 aa) sequence contained a tandem duplication resulting in a pseudo-dimer with two RHH motifs. Indeed, F TraY is a monomer in solution that binds to two sites at oriT sbyA and sbyC and one site at Py, sbyB (Nelson et al., 1993; Luo et al., 1994). SbyC, which is closer to nic than sbyA, overlaps the IHF site A. IHF was found to inhibit TraY binding to sbyC when present in in vitro binding assays. IHF and TraY bindings are not cooperative although sbyA and IHF site A must be correctly spaced and oriented with respect to each other for efficient nicking and mobilization (Williams & Schildbach, 2007). The requirement for a supercoiled substrate for nicking was relaxed in the presence of IHF and TraY, suggesting that these proteins increase the single-stranded character of the nic region allowing access to the relaxase (Nelson et al., 1995); however, KMnO4 mapping of the oriT region in the presence of an assembled relaxosome of the R1 plasmid suggests that localized denaturation in the sequences surrounding nic are subject to complex regulation at transfer initiation (Mihajlovic et al., 2009).

Using HR footprinting and methylation protection to more finely map F TraY-DNA interactions, Lum et al. (2002) showed that the 19-bp imperfect repeats thought to define sbyA and sbyC (Frost et al., 1994) were composed of smaller subsites that were recognized by TraY (Fig. 2). Using oligonucleotides representing sbyA, these authors showed that TraY bound to three subsites with the sequence 5′-GA(G/T)A-3′ with 5–6-bp intervening. A fourth site was occupied at high TraY concentrations (not shown). The spacing of these subsites was important, as was their orientation (two direct and one inverted repeat in sites 1, 2 and 3; Fig. 2). The stoichiometry of binding suggested that TraY binds cooperatively to subsites 2 and 3 (inverted repeat) and independently to subsite 1 at a ratio of one TraY monomer (pseudodimer) per subsite. This suggests that each half of the TraY pseudodimer supplies a β-strand that interacts with each other and binds the major groove in the DNA.

F TraM is a 127 aa protein that is conserved in sequence in other IncF transfer systems (Frost et al., 1994). TraM performs crucial functions in that it not only promotes relaxosome formation and autoregulates its expression but it is likely to interact physically with TraI and was shown to stimulate both the nic-cleavage and the oriT DNA unwinding reactions catalyzed by the bifunctional enzyme TraI (Ragonese et al., 2007; Sut et al., 2009). Moreover, recognition features of the TraM protein determine productive interactions between the T4CP, TraD and the relaxosome, which are essential to efficient plasmid transfer (Disque-Kochem & Dreiseikelmann, 1997; Sastre et al., 1998; Beranek et al., 2004; Lu et al., 2008). This evidence for strong functional integration of TraM in the relaxosome and our recent advances in understanding regulation of conjugative DNA processing and nucleoprotein uptake at the T4CP interface substantiates early predictions of TraM's involvement in signaling competent Mpf in the transfer initiation pathway (Willetts & Wilkins, 1984).

TraM binds to relatively GC-rich regions bracketed by intrinsic bends and IHF-binding sites within oriT (Frost et al., 1994). There are two to four binding sites (sbmA-D), as determined by DNAse I protection assays, that can be described as direct or inverted repeats containing small palindromes in F (Di Laurenzio et al., 1992), R1 (Schwab et al., 1993), R100 (Abo et al., 1991) and pED208 (Di Laurenzio et al., 1991). The binding site sbmC, which is nearest to nic, appears to be important for F plasmid transfer whereas sbmA and sbmB are more involved in autoregulation of the two traM promoters (Fu et al., 1991). TraM promotes nicking by the TraI relaxase both in vivo and in vitro in R1 (Kupelwieser et al., 1998; Mihajlovic et al., 2009).In the F system TraM stimulation of the in vitro reaction was observed in the presence of F TraY, IHF and DNA containing IHF sites A and B, bordering sbyA, as well as sbmC (Ragonese et al., 2007). F TraM binds sbmA and sbmB with high affinity and sbmC with lower affinity using linear DNA fragments. TraM binding is cooperative with the affinity for sbmC increasing 10-fold in the presence of sbmAB. Like TraY, F TraM bends the DNA by ≤50° (Fekete & Frost, 2002). Occupation of the TraM-binding sites substantially enhances negative supercoiling of the R1 oriT in vivo and in vitro (Mihajlovic et al., 2009). HR footprinting revealed that F TraM protected the CT bases within a 5′-CTAG-3′ subsite, which occurs twice 12 bp apart within sbmA and, by inference, sbmB, which is almost identical in sequence (Fig. 2). The HR footprint at sbmC revealed that TraM protected the center of the inverted repeat and radiated out from that point at 11–12-bp intervals toward the 5′ ends of the DNA fragment as concentrations of TraM increased (Fekete & Frost, 2002). The CTAG motif is absent in sbmC; however, all three F TraM-binding sites contain GC-rich subsites loosely described as 5′-GGC/TG-3′ that occur within the arms of the sbmC inverted repeat and 3′ to the two CTAG motifs in sbmA and sbmB (Fig. 2).

TraM has been shown to interact with the F TraD T4CP (Disque-Kochem & Dreiseikelmann, 1997) and with the last 38 aa of R1 TraD (Beranek et al., 2004). Mutational analysis of F TraM revealed that the K99E mutation prevented TraM–TraD interactions as assayed by genetic analysis and immunoprecipitation (Lu et al., 2004; Lu & Frost, 2005). F and R1 TraMs are tetramers according to analytical ultracentrifugation, spectroscopic and genetic methods (Verdino et al., 1999; Miller & Schildbach, 2003; Lu et al., 2004, 2006). The structure of the C-terminal tetramerization domain revealed an eight helical bundle with four of these helices, containing the K99 residues, protruding into the medium (Lu et al., 2006). The TraM tetramerization domain was successfully cocrystallized with a 13 aa peptide derived from the C-terminus of TraD (Lu et al., 2008). F-like coupling proteins (TraD) have a ‘specificity domain’ that extends from the C-terminus of the structure of P-type coupling proteins such as TrwB, the structure of which has been solved (Gomis-Rüth et al., 2001) and blocks RSF1010 mobilization (Sastre et al., 1998). Although the TraM–TraD C-terminal peptide interaction might only determine the first step in the binding pathway that leads to a stable complex between F-type relaxosomes and the T4CP, this result is particularly satisfying in that it addresses a longstanding observation that F TraD recognizes only closely related relaxosomes from the IncF group. One interesting outcome of the structure determination of the TraM tetramer is an explanation for the prediction by Miller & Schildbach (2003) that the equilibrium between the monomeric and tetrameric C-terminal domain, measured using circular dichroism, would be important in TraM function. Lu et al. (2006) found that four glutamic acids (E88), contributed by the four central helices, protrude into the lumen of the tetramer. These E88 residues were sensitive to deprotonation caused by temperature shift or pH changes, with the charged side chains repulsing each other, causing the tetramer to reversibly dissociate. This finding correlates with the observation that mating ability decreases rapidly as the temperature rises above 37 °C. The finding that TraM could also interact with the TraI C-terminal domain (Ragonese et al., 2007) opens up the possibility that TraM also directs the relaxase to the conjugation pore in MOBF systems leading to TraI–TraD interactions (Traxler & Minkley, 1987). However, the TraM–TraI interaction could not be verified by others (Guogas et al., 2009).

The structure of the DNA-binding domain has been more difficult to obtain. The residues that contact the DNA are known to be near the N-terminus in F and R1 TraM (Schwab et al., 1993; Miller & Schildbach, 2003; Lu et al., 2004). A nuclear magnetic resonance structure of the first 56 aa, minus the initial methionine, which is cleaved off in vivo, suggested that it contained three helices that were capable of dimerization (Stockner et al., 2001). Based on homology at the immediate N-terminus to known RHH proteins such as VirC2 and CopG, the DNA-binding domain might contain an RHH motif with the -strand within residues 2–8 (Lu et al., 2009). Thus, the tetramerization domain would be linked by an unstructured sequence to two dimeric DNA-binding domains in either a helical or an RHH conformation. Each TraM tetramer would bind to short GC-rich inverted repeats bracketing the CTAG sequences identified by Fekete & Frost (2002). In this scenario, the direct repeats in sbmA and sbmB can be viewed as inverted repeats with smaller inverted repeats within each arm (Fig. 2). The stoichiometry of TraM binding is two tetramers per binding site, that is, 6 TraM tetramers bound to sbmABC, with the possibility of loading more TraM onto the DNA at high concentrations (Lu et al., 2006). Interestingly, the MbeC protein of ColE1, which is the auxiliary protein in that system, has also been found to be an RHH DNA-binding protein (Varsaki et al., 2009).

MOBP relaxosomes

  1. Top of page
  2. Abstract
  3. Introduction
  4. Comparison of F- and P-like conjugative systems
  5. Stages in the conjugative DNA process
  6. Relaxases, an overview
  7. MOBF relaxosomes
  8. MOBP relaxosomes
  9. MOBQ relaxosomes
  10. A model for conjugative DNA processing
  11. Concluding remarks
  12. Acknowledgements
  13. References

MOBP relaxases and primases

In 1981, Lanka and Barth described an anisometric protein, Pri, encoded by the IncPα plasmid RP4 that complemented a dnaG mutation and hence showed primase activity (Lanka & Barth, 1981). This primase appeared to increase the efficiency of conjugative transfer of ssDNA by priming discontinuous DNA synthesis in the recipient cell. The pri gene encodes two overlapping products named TraC and TraC* (Lanka et al., 1984) with the larger form (120 kDa) being transported to the recipient cell (Rees & Wilkins, 1990). Similar results were also reported for the IncI plasmid ColIb that encodes the primase suppressor of DnaG (Sog) (Rees & Wilkins, 1989). This was an important finding that suggested conjugative pores could transport protein as well as DNA.

This initial work on RP4 primase set the stage for the fruitful studies on relaxosomes done by the Lanka laboratory in the 1990s. Analysis of the sequence of the RP4 and closely related R751 oriT regions (Pansegrau et al., 1988; Fürste et al., 1989) provided the basic framework for the description of oriT regions in general. They reported that the nic site is bordered by divergent promoters that transcribe the relaxase and leader operons, with traJ and traK being the first genes, respectively (Fig. 1c). The leader operon is so named because it encodes the first genes to enter the recipient cell. The oriT region is about 350 bp in size and contains a number of symmetric repeats that are bound by the transfer proteins, TraJ and TraK, encoded near oriT, which define the specificity of plasmid transfer, i.e. the variants produced by RP4 and R751 are not interchangeable (Ziegelin et al., 1991). The relaxase of RP4 was identified as being TraI (82 kDa), which is attached to the 5′ end covalently (Pansegrau et al., 1990b). The 5′ end at nic remained inaccessible, whereas the 3′ end could be extended by DNA PolI in vitro. They concluded that the nicking reaction required a supercoiled template, Mg2+, TraJ and TraI, and that there must be an ordered sequence of events that define the process to generate the ssDNA required for transfer. Thus, the stage was set for more detailed studies on the construction of relaxosomes.

RP4 auxiliary proteins

RP4 encodes a transfer region separated into two parts by IS8 (IS21) and a kanamycin resistance locus (Lessl et al., 1992) with Tra1 encoding oriT and the leader, relaxase and primase (TraC) operons as well as an operon encoding the auxiliary proteins TraN and TraO (Fig. 1c). The relaxase operon also encodes the T4CP TraG and the pilin cyclase TraF, the only gene product in Tra1 required for P-pilus assembly. The Tra2 region encodes the transferosome components for P-pilus assembly as well as the entry exclusion proteins (reviewed in Schröder & Lanka, 2005). A remarkable feature of RP4 is its independence from host-encoded proteins such as IHF that modulate DNA conformation and play a prominent role in F relaxosome biology. Instead, the RP4 relaxosome has three auxiliary proteins, TraH, -J and -K, to provide the conformation and specificity that the transfer process requires.

TraJ (13.2 kDa) binds oriT as demonstrated by gel retardation assays (Fürste et al., 1989) and specifically binds the right arm of a 19-bp inverted repeat 8 bp from the nic site by DNAse I protection and HR footprinting (Fig. 3). Examination of the sequence of the right arm of the inverted repeat reveals the presence of a shorter inverted repeat with the sequence G/AGTG/A. The central GT in each half of the repeat is located on the same face of the DNA as the nic site and coincides with the strongest protection from hydroxyl radicals. This sequence is less conserved in the left arm of the 19-bp inverted repeat, which might load TraJ in vivo in response to perturbations in the structure of the DNA that could then form a hairpin (Ziegelin et al., 1989).

image

Figure 3.  A detailed view of the RP4 oriT region (Fig. 1c) based on Fürste et'al. (1989) and Ziegelin et'al. (1992). Numbering is according to Pansegrau et'al. (1994). sri and srj refer to the sites for recognition by TraI (T-strand) and TraJ, respectively. The extensive DNAse I protection footprint is bracketed by arrows above the line and marked by sbK. Solid and dashed brackets indicate DNAse I protection footprints on the top and bottom strand for primary and secondary binding sites, respectively. Filled vertical arrowheads indicate positions of hypersensitivity to DNAse I. Bases that are protected from hydroxyl radicals are boxed on each line of sequence. Hooked arrows indicate the approximate start site for the transcripts from the PL1 and PR1 promoters. The inverted repeats, thought to be associated with TraJ recognition and binding, are indicated between the lines. The start codons for traJ and traK are boxed. See text for details.

Download figure to PowerPoint

Whereas TraJ appears to allow TraI to specifically nick DNA, stable relaxosomes could not be isolated without the addition of an acidic, oligomeric protein TraH (12.9 kDa) that is encoded from a different reading frame within traI. TraH is not a DNA-binding protein but, based on its ability to complex with TraI or TraJ as assayed by glycerol gradient centrifugation or column chromatography, it is thought to bind both proteins thereby stabilizing the relaxosome. This complex can be visualized on supercoiled DNA by electron microscopy (Pansegrau et al., 1990a).

A third protein involved in relaxosome formation is TraK (14.6 kDa) that has been implicated in relaxosome formation in vivo and in vitro as well as increasing the abundance of open circular plasmid species isolated from relaxosomes. Like TraJ, TraK also appears to be a transfer specificity determinant based on mobilization assays of oriT-containing plasmids (Fürste et al., 1989; Waters et al., 1991). TraK is a basic, proline-rich protein that binds DNA within the leader region with a Kapp of 4 nM, which is on the opposite side of nic compared with the TraJ-binding site (sbj) (Fig. 3). It binds to intrinsically bent DNA and wraps the DNA around it to form a compact nucleoprotein complex that shortens a linear DNA fragment by 170–180 bp. Using DNAse I protection footprinting assays, TraK was shown to bind one side of the helix whereas HR footprinting showed that TraK bound to short sites of 8 bp separated by 1–3 bp containing hypersensitive sites, on both strands, in an alternating fashion. A 49-bp inverted repeat was implicated in initiating TraK binding but was not a substrate for binding in in vitro assays. Thus, TraK appears to aggregate along the DNA and form a nucleoprotein complex that could aid in strand separation in preparation for DNA transport with the transferred strand being unwound from the exterior surface of this complex. The authors also suggest that TraK might introduce positive supercoils within the nucleoprotein complex that would permit negative supercoiling and DNA melting near nic, thereby increasing accessibility to the TraI relaxase (Ziegelin et al., 1992).

Interestingly, the identification of large repeats might be misleading and the true recognition sequence for RP4 TraK, as well as F TraY and TraM, might be short, densely packed subsites within the DNAse I protection assay footprints. The pattern of binding of these proteins to DNA is reminiscent of the Par type I and II proteins involved in DNA partitioning after replication and during cell division (Schumacher, 2008). The ability to bind numerous times on one face of the DNA and bend it into a superhelical structure, as envisioned for the pSK41 plasmid partition system Schumacher et al., 2007), is intriguing. One can imagine TraY, TraM and IHF in the F-like systems, or TraK and perhaps TraJ in the P-like systems, forming such a structure that would facilitate strand opening in preparation for the entry of the relaxase and helicase, where appropriate, as well as the DNA replication machinery in the donor cell. The T4CPs are distant relatives of DNA partitioning proteins such as SpoIIIE and FtsK (Gomis-Rüth et al., 2004). It is entirely possible that the two processes of partitioning and conjugative DNA transfer, which is really a partitioning reaction into another cell, are evolutionarily related.

The mechanism for unwinding the DNA in promiscuous plasmids such as RP4 remains somewhat mysterious. Unlike the bifunctional TraI relaxase–helicase of F, RP4 and related plasmids do not encode a helicase nor has there been any evidence that they avail themselves of a host-encoded helicase to unwind the DNA. The TrbB ATPase of the VirB11 family (reviewed in Schröder & Lanka, 2005) has been implicated in DNA transport (Chen et al., 2005) but it appears to be more involved in assembly or disassembly of the T4SS and translocation of protein or DNA across the inner membrane than DNA unwinding. RP4 and several other promiscuous plasmids in both Gram-negative and -positive bacteria encode a homolog of topoisomerase III (TraE encoded in Tra1 of RP4) that could aid in plasmid resolution during vegetative or conjugative replication (Li et al., 1997). Similarly, KorB appears to have a dual role in regulation of RP4 gene expression and partitioning (Lukaszewicz et al., 2002) but it also has not been implicated in conjugative DNA transfer. Conjugative DNA replication is not thought to be required for DNA transfer, at least in the F system (Sarathy & Siddiqi, 1973; Kingsman & Willetts, 1978); however, it might help drive this process in P-like transfer systems.

MOBQ relaxosomes

  1. Top of page
  2. Abstract
  3. Introduction
  4. Comparison of F- and P-like conjugative systems
  5. Stages in the conjugative DNA process
  6. Relaxases, an overview
  7. MOBF relaxosomes
  8. MOBP relaxosomes
  9. MOBQ relaxosomes
  10. A model for conjugative DNA processing
  11. Concluding remarks
  12. Acknowledgements
  13. References

MOBQ relaxases

The R1162 relaxase belongs to probably the largest family of relaxases, termed the MOBQ superfamily (Francia et al., 2004; Garcillán-Barcia et al., 2009). Members of this family are found not only in the proteobacteria but in the firmicutes and cyanobacteria as well, and are encoded by many different plasmids, not just those of the Q incompatibility group. The relaxases of R1162, RSF1010 and R3200B, independently isolated but essentially identical plasmids, have been taken as prototypes of the group, but they have the unusual feature of being fused at the C-terminal end to a second active protein, the replicative primase RepB′. The two proteins are joined by translational read-through across the intervening gene mobB, which is translated in a different reading frame (Scholz et al., 1989; Fig. 1a). The primase is not strictly required for mobilization, but contributes to complementary strand synthesis in both donor and recipient cells (Henderson & Meyer, 1999; Becker & Meyer, 2002; Parker & Meyer, 2002).

The DNA-processing domain of the R1162 relaxase (minMobA) makes up the N-terminal 186 aa of the protein (Becker & Meyer, 2002). The crystallographic structure of this domain has been solved (Monzingo et al., 2007) and bears a close resemblance to the corresponding domains of the F and R388 relaxases (Datta et al., 2003; Guasch et al., 2003), despite no convincing homology at the amino acid level. The active sites are also similar, with three histidines coordinating a covalent metal ion to enable phosphodiester cleavage by a tyrosine residue (Y25). Cleavage involves a transesterification with the formation of a tyrosyl phosphodiester linkage. Thus, this nucleophile is unavailable for a second cleavage reaction until it is released following recircularization of the transferred strand. Nevertheless, the covalently linked relaxase can apparently undertake a second cleavage reaction in vivo, although a second nucleophile has not been identified. Exhaustive mutagenesis has shown that such a nucleophile is not in the active site as defined by the crystal structure (Monzingo et al., 2007).

Although minMobA can process oriT DNA, it is not competent for transfer. The signal allowing recognition of a T4CP is missing but is contained within the ∼100 aa adjacent to the DNA-processing domain (Parker & Meyer, 2007). This region is poorly conserved at the amino acid level among different members of the MobQ relaxases, even when the plasmids recognize the same coupling protein (Meyer, 2000). Interestingly, there is a second signal, also active for conjugative transfer, encoded within the primase domain of the full-length protein (Parker & Meyer, 2007).

It might be expected that plasmids such as R1162/RSF1010/R300B, with a very broad host range for replication, have also undergone selection for the promiscuous use of different T4CP. Indeed, IncQ plasmids are efficiently mobilized by IncP, IncIα and IncX plasmids (Willetts & Crowther, 1981), the vir system of Agrobacterium tumefaciens (Beijersbergen et al., 1992), the DOT–ICM system of Legionella pneumophila (Vogel et al., 1998) and an uncharacterized T4SS encoded by Salmonella enterica serovar Typhi (Baker et al., 2008). The F factor, on the other hand, mobilizes R1162/RSF1010/R300B very poorly. This seems to be due to active exclusion by the C-terminal end of the F coupling protein TraD: when this tail is deleted, RSF1010 is now mobilized much more efficiently (Cabezón et al., 1997; Sastre et al., 1998).

MOBQoriT and auxiliary proteins

The oriT of R1162/RSF1010/R300B is a compact locus made up of no more than 38 bp (Brasch & Meyer, 1987). One major reason for this is that, in contrast to F and R388, there are no binding sites for chromosomally encoded proteins. Independence from host proteins is not surprising for a mobilization system functional in many different cytoplasms. Nevertheless, the MobQ relaxase is most active for cleavage of oriT DNA when this DNA is single-stranded. Thus, as with the larger oriTs of the self-transmissible plasmids, in the relaxosome, a mechanism for strand separation at nic is required.

In addition to the relaxase, R1162 encodes two small proteins required for efficient transfer: MobB (MW 15,115) and MobC (MW 10,885). An active relaxosome has been assembled in vitro from purified components (Scherzinger et al., 1992). Both MobB and MobC are required for full activity, and are present at high stoichiometric amounts in the complex. There is localized disruption of the DNA duplex within the relaxosome at an AT-rich region adjacent to the nic site. Presumably, this strand separation is necessary to generate the single-strand character required for efficient cleavage. Extension of the nascent melted region through nic requires MobC (Zhang & Meyer, 1997). MobC might cause wrapping of the DNA, inducing torsional stress that is relieved by greater strand separation. However, it is important to note that, in contrast to the auxiliary proteins of the F oriT, there are no clearly delineated binding sites for MobC (or for MobB). This is another reason that the R1162 oriT is small in comparison with the F locus.

Although MobB is required for a relaxosome that is fully functional for nicking in vitro, the effect of its absence from the reaction mixture is small (Scherzinger et al., 1992). In contrast, plasmids lacking MobB are mobilized at only about 0.1% of the normal frequency. This suggested that MobB is required in some other way for transfer, in addition to increasing the efficiency of nicking in the relaxosome. Recently, a Cre reporter system, originally devised to examine the flux of Vir proteins in type IV secretion (Vergunst et al., 2000) was used to determine the requirements for secretion of the R1162 relaxase by the R751 (IncP) transfer system. It was found that MobB was required for transport of the relaxase into a recipient cell, even in the absence of plasmid DNA. MobB is associated with the membrane, and has a single, putative membrane-binding domain that is important for transfer (Parker & Meyer, 2007). The details of the requirement for MobB are not yet understood, but it is likely that by anchoring the relaxase to the membrane, it enhances the interaction of the relaxase with the T4CP of the type IV transporter.

The R1162/RSF1010/R300B oriT consists of an inverted repeat with 10-bp arms, an adjacent, AT-rich region where strand separation is initiated in the relaxosome, and a GC-rich segment that includes the nic site (Fig. 1a). This general arrangement is highly conserved among the oriTs associated with the large family of MobQ relaxases (Becker & Meyer, 2003) with the size and sequence of the inverted repeat showing the greatest variation. The arm of the inverted repeat distal from the nic site, at one end of oriT, is required for rejoining the ends of the transferred strand but not for initiating transfer (Kim & Meyer, 1989). It is likely that the inverted repeat recreates a duplex region for binding by the relaxase, covalently linked to the 5′ end of the strand, at termination of a round of transfer. In agreement with this, minMobA binds much more strongly to an oriT containing the intact inverted repeat, compared with an oligonucleotide lacking the outer arm (Bhattacharjee & Meyer, 1993). The inverted repeat is not required in vitro for correct cleavage at nic, however (Scherzinger et al., 1993).

Because the outer arm of the inverted repeat is not required for initiation of transfer, only 19 bp of oriT DNA are sufficient for this step with the R1162 Mob proteins. In addition, there is a high degree of tolerance for base changes at many positions within this 19 bp: the consensus sequence for initiating transfer is WWW NNNNTAAR TGCGC. A consequence of this relaxed specificity is that sites in the chromosome matching the consensus can serve as origins for the transfer of chromosomal DNA when R1162 is in the cell, thus providing a new source of plasticity for the bacterial genome (Meyer, 2009).

Because only one DNA strand is transferred, synthesis of the complementary strand in the recipient and replacement strand synthesis in the donor, are intrinsic steps in conjugation. However, determining the exact mechanism(s) of these syntheses, and establishing their relationship to replicative DNA synthesis, has proved difficult. Plasmid derivatives of R1162 were introduced by electroporation into donor cells lacking the plasmid-encoded proteins for vegetative replication of its DNA, and the resulting cells were mated. Transferred DNA was then rescued by site-specific recombination into the chromosome of the recipient (Parker & Meyer, 2005). These results revealed that the machinery for vegetative replication does not need to be active for successful transfer. However, when the plasmid-encoded primase RepB′ is synthesized in the donor, and there is an appropriately oriented priming site on the plasmid introduced by electroporation, then this protein efficiently initiates complementary strand synthesis in the donor. Thus, a component of the system for vegetative replication can be recruited to assist during conjugation. This result does not rule out rolling-circle replication for strand replacement in the donor, but does indicate that other mechanisms can be involved.

A model for conjugative DNA processing

  1. Top of page
  2. Abstract
  3. Introduction
  4. Comparison of F- and P-like conjugative systems
  5. Stages in the conjugative DNA process
  6. Relaxases, an overview
  7. MOBF relaxosomes
  8. MOBP relaxosomes
  9. MOBQ relaxosomes
  10. A model for conjugative DNA processing
  11. Concluding remarks
  12. Acknowledgements
  13. References

Dynamics of MOBF relaxosomes

The premise that the relaxosome in the traditional sense is insufficient for in vivo transfer initiation has long been discussed, but it has been challenging to move beyond negative results and gain insights into later steps in the initiation pathway. The studies described below provide evidence for distinctly regulated stages of initiation and introduce an experimental framework to address the underlying mechanisms. The transition from preinitiation, i.e. high-affinity interactions of TraI relaxase, to initiation of unwinding at a melted duplex by the TraI helicase domain requires some sort of activation. Because recruitment and activation of the conjugative helicase does not occur on naked DNA but instead on a loading platform provided by the relaxosome, recent efforts have focused on understanding the regulation of helicase operation within the context of the relaxosome. TraI requires ssDNA 5′ to the duplex junction for helicase activity (Kuhn et al., 1979) and definition of the length requirements for TraI loading outside of oriT demonstrated that >30 nt are necessary (Csitkovits & Zechner, 2003; Sikora et al., 2006). Reconstitution of MOBF relaxosomes on supercoiled DNA in vitro has not led to localized melting of oriT or to initiation of helicase activity (Mihajlovic et al., 2009). Consequently, little is known about sequence-specific loading of a conjugative helicase and its control. To gain insights into helicase regulation, linear heteroduplex substrates containing dsDNA-binding sites for plasmid R1 relaxosome proteins and varying regions of open duplex for TraI helicase loading were constructed to model intermediate structures in the initiation pathway (Csitkovits et al., 2004; Sut et al., 2009). Variation in the sequence and length of DNA exposed in single-stranded form around nic, as well as the additional presence of auxiliary relaxosome proteins, revealed novel insights into control of a conjugative helicase. Firstly, efficient helicase activity initiating from a melted oriT duplex, required twice the length of ssDNA (60 nt) than in other in vitro systems (Csitkovits et al., 2004). Remarkably, the study also revealed a form of TraI helicase inhibition that acts on one or more steps of the multicomponent activation cycle. The effect is apparently autoregulatory and involves sequence specificity at oriT (Sut et al., 2009). Repression was alleviated by removing either the phosphodiesterase domain from TraI or the specific relaxase recognition elements from the DNA substrates. Thus, the effect could involve the distinct conformations of high-affinity binding observed previously for the TraI phosphodiesterase domain with sequences 5′ to nic (Williams & Schildbach, 2006). These data are also consistent with results of a general mutagenesis of F plasmid traI (Haft et al., 2006). Hyperconjugation phenotypes of traI insertion mutants implied that in the wild-type situation, a repression mechanism involving traI imposes a rate limitation upon the conjugation process. Thus, the current model proposes that autoinhibition indeed reflects the inherent contradiction of combining a high-affinity site-specific DNA-binding activity with a helical motor in the same protein. Physical tethering of the TraI functional domains has persisted in evolution; thus, it would appear to be advantageous to localize the conjugative helicase to oriT via sequence recognition properties of the phosphodiesterase domain yet simultaneously repress a concerted unwinding event. It follows that a commitment to unwinding from nic requires activation of the molecular control switch linking the two domains.

The autoinhibition of TraI was apparent in vitro despite sufficient ssDNA to effectively load the helicase; thus, the controlling mechanism operating in conjugation is very likely to involve additional proteins. At present, the nature of the helicase repression remains unknown. Application of the heteroduplex system described above provided the first evidence that auxiliary proteins could enhance TraI-catalyzed oriT unwinding (Sut et al., 2009). Auxiliary proteins TraM and IHF stimulated the helicase activity of the cloned and purified TraI helicase domain as well as the full-length protein in a manner directly correlated with the proximity of the proteins' specific binding sites to the position of duplex opening. By contrast TraY had no effect. The cytoplasmic domain of the R1 T4CP, TraDΔN130, increased helicase efficiency on all substrates in a manner consistent with cooperative interaction and sequence-independent DNA binding.

The above study thus revealed the potential importance of the physical interface between the relaxosome and the T4CP for the regulation of conjugative DNA processing and strand transfer initiation. The R388 T4CP, TrwB, has been characterized in detail; thus, we know that proper multimerization to active hexamers (in the membrane) is driven by cooperative interactions among monomers, as well as by binding to DNA and NTP. In turn, a latent capacity for NTP hydrolysis is induced in TrwB by the resulting flux in protein conformation (Tato et al., 2007). Early models of coupling protein function invoked a physical bridge between plasmid and T4SS transporter, without implications on the DNA side except the necessity to distinguish one plasmid from another (Cabezón et al., 1997). The sum of our current understanding implies a more substantial role for T4CPs.

DNA processing of MOBF relaxosomes

From the analysis of the MOBF relaxase crystal structures, together with the biochemical properties manifested in the in vitro reactions (MOBF relaxosomes) plus the biochemical details of the DNA-processing reactions (Dynamics of MOBF relaxosomes), we have inferred a model for DNA processing at the initiation of plasmid strand transfer, which has the following steps (Fig. 4):

image

Figure 4.  Model for regulation of T-strand production and delivery to the translocation machinery of the MOBF T4SS (indicated by ‘uptake of DNA’ in a box). Steps in the initiation reaction cascade (I–VII) are described in DNA processing of MOBF relaxosomes. Protein components are illustrated as follows: the bifunctional TraI phosphodiesterase (blue) with active site tyrosine (Tyr) and TraI helicase domains (light blue), IHF (orange), TraY (purple), TraM (green) and the T4CP TraD (yellow). See DNA processing of MOBF relaxosomes for details. © Journal of Bacteriology 19: 6888-6899, doi:10.1128/JB. 00920-09, reproduced with permission from the American Society for Microbiology.

Download figure to PowerPoint

Preinitiation (relaxosome formation and nic cleavage)

The relaxosome assembles on supercoiled oriT DNA and assumes a complex (but yet undefined) higher order structure determined by the sum of topological distortions induced by each effector interacting with the DNA (Fig. 4, step I). Within the relaxosome, F TraI is proficient for cleaving-joining activity at nic and is stimulated by IHF, TraY and TraM (Fig. 4, step II). T4CP TraD also stimulates nic cleavage by TraI revealing the capacity for regulatory interactions at the physical interface of relaxosome and the conjugative pore (Fig. 4, step III). Localized melting of oriT at this stage may not be sufficient to load the helicase and extensive duplex unwinding does not occur. Thus, the relaxosome complex is preformed and in a potentially competent state for elongation, but no irreversible reaction has occurred.

Activation (formation of the T-strand)

Translocation on ssDNA and entry into the duplex require modulation of a molecular switch in the full-length bifunctional relaxase and, possibly, an extension of duplex melting at nic. Conditions that would induce these changes presumably emerge from intercellular contacts between bacteria and are communicated over conjugative pili to the T4SS (Fig. 4, step IV). In that case, docking of the preinitiation complex to the T4SS via TraD would confer competence for T-strand production. Interactions between TraD and relaxosome components thus complete the signal transduction pathway to protein and nucleoprotein targets. TraI switches to translocation mode altering the N-terminal domain's high-affinity interactions with the inverted repeat and surrounding sequences. Both 3′ and 5′ ends of nic remain sequestered by TraI. Protein translocation or increased exposure of ssDNA alters TraD conformation and ligand interaction (Fig. 4, step V). The latent NTPase of the T4CP is activated (Fig. 4, step VI). Uptake of the adduct of TraI and the T-strand into the TraD translocation channel is mediated by recognition of translocation signals on the TraI protein. Simultaneous handoff of the 3′ end of nic to DNA polymerase III and commitment of the 5′ pilot protein to the translocation pathway marks the end of the initiation cycle (Fig. 4, step VII).

T-strand transport

Conjugative mobilization of the E. coli chromosome starting at oriT of an integrated copy of the F plasmid (Hfr) requires ∼100 min. That observation implies a rate of unwinding of at least 800 bp s−1. Concomitant replacement strand synthesis is expected as this affords maximum protection for the plasmid genome. In the absence of a helicase to provide unwound template, DNA Pol III achieves relatively low rates of DNA synthesis. Thus, if replacement strand synthesis occurs at an equivalent rate to T-strand mobilization, TraI-catalyzed activity at the junction of plasmid strands must work in concert with DNA Pol III (Mok & Marians, 1987; Kim et al., 1996). It follows that if the monomer of TraI covalently linked to the 5′ end of the T-strand is actively secreted in the initial step of translocation, then a second monomer positioned at the duplex junction is by necessity distinct from that piloting the T-strand uptake and translocation through the T4 machinery. The relaxase–T-strand complex is transported to the recipient cell by the pilus-associated T4SS (reviewed by Christie et al., 2005). The details of the transport process are not known, but it is supposed that the relaxase acts as a pilot protein, being transported to the recipient cell and thus threading the DNA into the transport channel. The T-strand appears in the recipient as ssDNA and the leading region is expressed, possibly from ssDNA-specific promoters or by the activity of primases.

T4CP activation

T-strand transport, which can be as extensive as the complete E. coli chromosome, is thought to be energized by the T4CP that connects the relaxosome to the T4SS via interactions with the relaxosome components, including the relaxase. T4CPs are integral membrane proteins, containing two transmembrane α-helices in their N-terminal region. The T4CP, R388 TrwB, has been shown to be a DNA-dependent ATPase. Its activity is greatly enhanced by interaction with the nicking-auxiliary protein TrwA. TrwB, which is isolated mainly as a monomer, probably assembles onto the DNA as it is threaded into the T4SS. In vitro, R388 TrwB assembles on ssDNA nonspecifically, forming most probably hexamers, which have enhanced ATPase activity (Tato et al., 2005; Haft et al., 2006). Transport of the T-DNA probably involves both pushing from the donor via the TrwB motor, and pulling from the recipient via the TrwC helicase motor.

Termination (recircularization)

Once in the recipient, MOBF-type relaxases can track on the incoming DNA in search for the nic-site that they will recognize as a termination site. A strand-transfer reaction – catalyzed this time on purely ssDNA – reforms an ssDNA circle representing the donor DNA. The termination site can also be transferred precleaved from the donor by a second relaxase molecule. The involvement of the TrwC relaxase in the termination reaction within the recipient cell has been proven using specific antibodies (Garcillán-Barcia et al., 2007).

Lagging-strand synthesis

This is still a largely unresolved question although progress has been made using the IncQ plasmid R1162. No nonessential gene appears to be involved in replacement strand synthesis in the recipient cell in E. coli-to-E. coli matings (Pérez-Mendoza & de la Cruz, 2009).

Concluding remarks

  1. Top of page
  2. Abstract
  3. Introduction
  4. Comparison of F- and P-like conjugative systems
  5. Stages in the conjugative DNA process
  6. Relaxases, an overview
  7. MOBF relaxosomes
  8. MOBP relaxosomes
  9. MOBQ relaxosomes
  10. A model for conjugative DNA processing
  11. Concluding remarks
  12. Acknowledgements
  13. References

Unlike many conjugation systems, the MOBFPQ relaxosomes are remarkable in that conjugative DNA transfer occurs almost immediately after contact is established between the donor and recipient cells. This is in contrast to many systems that require induction by small molecules (e.g. plant wound compounds, homoserine lactone derivatives or pheromones; Wilkins & Frost, 2001) that activate transfer region transcription and assembly of the conjugative pore before transfer can begin. Another remarkable feature of these systems is the apparently random distribution of potential conjugative pores on the cell surface (Gilmour et al., 2001; Lawley et al., 2002b). Although few systems have been studied in detail, other systems localize the newly expressed transfer apparatus at the poles of the cell (e.g. Kumar & Das, 2002), which simplifies directing the T-strand to the pore. Thus, there must be mechanisms in MOBFPQ systems that allow the relaxosome to sense that a T4CP–T4SS has formed a competent mating pore somewhere in the cell envelope and direct the plasmid to that site. Currently, this is another premier challenge in the field.

Although the auxiliary proteins are important in both in vitro and in vivo assays at all stages of Dtr, the location and binding characteristics of these proteins during the cell and growth cycles are unknown. For instance, nicking and ligation are in equilibrium in growing cells with the supercoiled, closed circular form predominating. During vegetative replication, the DNA is presumably ligated and the relaxosomal complex makes way for the replication machinery in the cell. How these two processes are coordinated remains unknown. Plasmids in the IncF and P groups are replicated at the mid-point of the cell and partitioned to the quarter positions in step with cell division (Gordon et al., 1997; Gilmour et al., 2001). As the cells enter stationary phase, the levels of the relaxase, nicking at oriT, and the auxiliary proteins decrease to undetectable levels (Frost & Manchak, 1998), whereas the components of the transferosome and the coupling protein are relatively stable. Thus, the auxiliary proteins must connect cell growth and plasmid replication to relaxase activity as well as monitor conjugative pore formation. They could also be involved in the transition from preinitiation (nicking and religation) to initiation (nicking and unwinding) that is key to the process. Considerable progress has been made in understanding the stages in readying a plasmid for transfer since the review of Lanka & Wilkins (1995). However, there is much to be done if we wish to link the biochemistry of the relaxosome to the events that trigger transfer and establishment of the T-strand in the recipient cell.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Comparison of F- and P-like conjugative systems
  5. Stages in the conjugative DNA process
  6. Relaxases, an overview
  7. MOBF relaxosomes
  8. MOBP relaxosomes
  9. MOBQ relaxosomes
  10. A model for conjugative DNA processing
  11. Concluding remarks
  12. Acknowledgements
  13. References

The authors wish to thank Dr Joel Schildbach for unpublished results and Silvia Lang for providing Fig. 4. This work was supported by the Canadian Institutes for Health Research, grant MT 11249 (L.S.F.); Austrian FWF grants P18607, W901-B05 (DK: Molecular Enzymology) and the EU grant FP6 ESR 019023 (E.L.Z.). Work in the FdlC laboratory was supported by grants BFU2005-03477/BMC (Spanish Ministry of Education), RD06/0008/1012 (RETICS research network, Instituto de Salud Carlos III, Spanish Ministry of Health) and LSHM-CT-2005_019023 (European VI Framework Program). Work in the Meyer laboratory was supported in part by the US Public Health Service (NIH GM-37462).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Comparison of F- and P-like conjugative systems
  5. Stages in the conjugative DNA process
  6. Relaxases, an overview
  7. MOBF relaxosomes
  8. MOBP relaxosomes
  9. MOBQ relaxosomes
  10. A model for conjugative DNA processing
  11. Concluding remarks
  12. Acknowledgements
  13. References