From conjugation to T4S systems in Gram-negative bacteria: a mechanistic biology perspective

Abstract

Conjugation is the process by which bacteria exchange genetic materials in a unidirectional manner from a donor cell to a recipient cell. The discovery of conjugation signalled the dawn of genetics and molecular biology. In Gram-negative bacteria, the process of conjugation is mediated by a large membrane-embedded machinery termed "conjugative type IV secretion (T4S) system", a large injection nanomachine, which together with a DNA-processing machinery termed "the relaxosome" and a large extracellular tube termed "pilus" orchestrates directional DNA transfer. Here, the focus is on past and latest research in the field of conjugation and T4S systems in Gram-negative bacteria, with an emphasis on the various questions and debates that permeate the field from a mechanistic perspective.

Keywords: DNA and Protein Secretion; bacterial conjugation; pilus biogenesis; relaxosome; type IV secretion system.

Figure 1. The various processes in which T4S systems are involved

T4S systems are involved in DNA transport during conjugation, transformation and A. tumefaciens infection, and in effector transport by a number of bacterial pathogens. This figure was modified from Grohmann et al 108.

Figure 2. The relaxosome

(A) Schematic diagram of relaxosome composition and assembly. Upper panel: Composition of the F plasmid family relaxosome. The F‐family relaxosome is composed of the relaxase TraI and three accessory proteins: TraY and TraM encoded by the plasmid and IHF encoded by the bacterial genome. All proteins assemble at the plasmid's origin of transfer (OriT) in a process that affects DNA topology around the OriT region. OriT also contains the nick site (nic). This site is flanked by regions (in orange and yellow for the regions 5′ or 3′ to nic, respectively) that, when single‐stranded, would each bind a TraI molecule. Lower panel: Schematic representation of the OriT region of the F plasmid. The binding sites for each relaxosome components are depicted by boxes coloured according to the protein to which they bind using the same protein colour‐coding shown in the upper panel. Under each box, the protein which binds to the depicted site and the name of the site are indicated. “TraI A” and “TraI B” indicate the region 5′ and 3′ to nic to which the trans‐esterase and helicase domains of two individual TraI molecules bind, respectively (depicted in Fig 2B). (B) Domain structure of TraI and binding to OriT. Upper panel: Domain structure of TraI. TraI is composed of a trans‐esterase domain (orange), a vestigial (green) and an active helicase (blue) domain, and a C‐terminal domain (grey). Boundary residues are indicated for R1 plasmid TraI. Lower panel: TraI binding to OriT. TraI is indicated schematically with each volume indicating the various TraI domains using the same colour‐coding as described above. Two TraI molecules bind OriT, one on each side of the nic site. TraI bound to sequence 5′ (indicated by the orange strip below) to the nic site is bound through its trans‐esterase domain and its overall conformation is open (not shown here). TraI bound to sequence 3′ (indicated by the yellow strip above) to the nic site is bound through its helicase domains and its overall conformation is closed (not shown here). (C) Structure of TraI in its helicase‐loaded mode. The TraI–ssDNA complex is shown with TraI and the ssDNA in ribbon and stick representation, respectively. The domains are coloured coded as in upper panel (B). The linker between the trans‐esterase and vestigial helicase domain is shown in grey.

Figure 3. The architecture of the T4S system from X. citri, H. pylori and L. pneumophila

(A) The X. citri OMCC. Left panel: Superposition of the heterotrimeric unit of the X. citri OMCC structure (composed of full‐length VirB7, VirB9, and VirB10) with the heterotrimeric unit of the O‐layer of the pKM101 OMCC (made of the full‐length VirB7, and the C‐terminal domains of VirB9 and VirB10). Structures are shown in ribbon representation. X. citri proteins are color‐coded in green, yellow and cyan blue for VirB10, VirB9, and VirB7, respectively. The pKM101 O‐layer heterotrimer is shown in grey. Domains of X. citri as well as some of the N‐ and C‐termini are labeled. A striking feature of all high resolution OMCC structures is the presence of long inter‐domain linkers, which project each domain of VirB9 and VirB10 a long distance away. Middle panel: Surface of the full‐length X. citri OMCC. Color‐coding is as in left panel. For the VirB10 N‐terminal domain, only density for an α‐helix (labeled “αH”) was observed and a model corresponding to residues 150 to 161 was derived. Seven of the 14 heterotrimeric units are shown so as to access a view of the interior of the OMCC. This interior is lined with VirB10 as also observed for the O‐layer of the pKM101 OMCC [65]. The I‐ and O‐layers are indicated. Right panel: Schematic diagram of the OMCC. αH (residues 150‐161 of the N‐terminal domain of VirB10) is the only secondary structure in the N‐terminal domain of VirB10 that is observed in the electron density of the X. citri OMCC. However, there are 149 residues N‐terminal to this region that are not observed, including a trans‐membrane helix (TMH) that inserts into the IM. 14 of those were hypothesized to form a channel in Legionella. (B) The H. pylori cag T4S system. Left upper panel: A slice through the side view of the composite sub‐tomogram average. Averages aligned on the periplasmic and cytoplasmic parts are stitched together using the IM as the boundary. Some regions of the density mentioned in the text are indicated and labelled. Left lower two panels: Two duplicated side views as in left upper panel are shown so as to compare the size and location of various other structures. The orange outline indicates comparison of the R388 OMCC to the cag T4S system structure; the blue outline indicates the position of the purified cag T4S system OMCC within the cag tomography structure; the magenta outline indicates the predicted location of the coupling protein Cag5/VirD4 based on the structure of the DotL/VirD4 homologue; the green outline indicates the Legionella dot/icm T4S system structure superimposed on the cag T4S system structure. Right panel: Schematic diagram of the IMC ATPases. The four “tubes” of density observed in the sub‐tomogram average together with the central density observed might correspond to either two side‐by‐side CagE/VirB4 hexamers flanking one Cag5/VirD4 hexamer (upper panel) or four CagE/VirB4 hexamers surrounding one Cag5/VirD4 hexamer (lower panel). VirB4 and VirD4 subunits are represented as cylinders colour‐coded blue and orange, respectively. (C) The Legionella dot/icm T4S system. Upper left and right panels: The Legionella T4S system observed by cryo‐ET by Ghosal et al (2017, 2018) and interpretation of the IMC ATPase organization 69, 73. Upper left panel: A slice through the side view of the composite sub‐tomogram average of the Legionella dot/icm T4S system. Some regions of the density mentioned in the text are indicated and labelled. Upper right panel: Schematic diagram of IMC ATPases. The four “tubes” of density are interpreted as projections of two side‐by‐side DotO/VirB4 ATPases. Lower left and right panels: The Legionella T4S system observed by cryo‐ET by Chetrit et al 15. Lower left panels: The top panel shows a slice of the sub‐tomogram average of the entire T4S system while the bottom panel focuses on the IMC. In the IMC, four “tubes” of density are clearly visible as for all T4S systems visualized by cryo‐ET, but, in this study, the two central tubes are bound to two additional “tubes” of density corresponding to the DotB/VirB11 ATPase. Lower right panels: Schematic diagram of the IMC ATPases: it is suggested that the four “tubes” of density are projections of a hexamer of DotO/VirB4 dimers with the DotO/VirB4 subunits involved in hexamerization stacked against the DotB/VirB11 hexamers (two views are shown here: one side view and the other 90° away).

Figure 4. Architecture of the R388 conjugative T4S system

Left panel: Schematic diagram of the R388 VirB3‐10/VirD4 structure by Redzej et al 56. The VirB4/TrwK ATPases are shown as trimers of dimers in blue while the VirD4/TrwB ATPases are shown as dimers, consistent with the observation of the electron density for that complex. Inset: side view of the density map of the negative‐stain EM structure of the VirB3‐10/VirD4 by Redzej et al 56. Right panels: Schematics of IMC ATPases. The structure shows two VirB4/TrwK trimers of dimers forming the two barrel‐like densities and two VirD4/TrwB dimers linking them. These are shown as blue and orange yellow circles, respectively. This apparatus can operate in two modes: in a pilus biogenesis mode upon binding of VirB11/TrwD to the two VirB4/TrwK ATPases, or in a substrate‐transfer mode upon binding of the relaxosome to the VirD4/TrwB. I hypothesize that VirB11 acts as a “hexamer organizer” remodelling each VirB4/TrwK trimers of dimers into active hexamers in order to execute pilus biogenesis, while the relaxosome induces remodelling of VirD4 dimers into hexamers (option 1) or mixed VirD4 and VirB4 dimers into hexamers (option 2). In option 1, the resulting VirD4 hexamer is positioned sideways; in option 2, the hexamer is central, just underneath a potential VirB10 channel and the pilus. The relaxosome is represented as in Fig 2A.

Figure 5. The F‐family pilus and mechanism of T4S

(A) The structure of the F‐family pilus. Upper left panel: One array of VirB2 pilus subunits with VirB2 shown in light blue surface representation and the phospholipid shown in sphere representation colour‐coded in white and red for carbon and oxygen atoms, respectively. Upper right panel: Five arrays of VirB2 subunits shown as in the upper left panel except that the five arrays are colour‐coded in a different colour. Lower panel: The pentameric base of the F pilus. Representation and colour‐coding are the same as in upper panels. (B) Mechanism of pilus biogenesis and substrate transfer by conjugative T4S systems. Conjugative T4S systems can operate in two modes: a pilus biogenesis mode (left) and a DNA‐transfer mode (right). VirB11 hexamer binding reshapes VirB4 to switch the T4S system to its pilus biogenesis mode. In that mode, VirB2 pilus subunits are extracted by VirB4, perhaps using a “lateral gate” mechanism to capture pilin subunits. The lateral gate mechanism was first described to account for the mechanism of the SecYEG transport apparatus 107. In the DNA transfer mode, the relaxosome is hypothesized to induce either hexamerization of VirD4 ATPase dimers to form a VirD4 homo‐hexamer situated on the side (option 1) or the formation of mixed VirB4/VirD4 hexamers located centrally, just under the VirB10 channel and the VirB2 pilus (option 2). The white arrows indicate the transfer route for each option. The T4S system and the relaxosome are as in Fig 4.