The All-Alpha Domains of Coupling Proteins from the Agrobacterium tumefaciens VirB/VirD4 and Enterococcus faecalis pCF10-Encoded Type IV Secretion Systems Confer Specificity to Binding of Cognate DNA Substrates

ABSTRACT

Bacterial type IV coupling proteins (T4CPs) bind and mediate the delivery of DNA substrates through associated type IV secretion systems (T4SSs). T4CPs consist of a transmembrane domain, a conserved nucleotide-binding domain (NBD), and a sequence-variable helical bundle called the all-alpha domain (AAD). In the T4CP structural prototype, plasmid R388-encoded TrwB, the NBD assembles as a homohexamer resembling RecA and DNA ring helicases, and the AAD, which sits at the channel entrance of the homohexamer, is structurally similar to N-terminal domain 1 of recombinase XerD. Here, we defined the contributions of AADs from the Agrobacterium tumefaciens VirD4 and Enterococcus faecalis PcfC T4CPs to DNA substrate binding. AAD deletions abolished DNA transfer, whereas production of the AAD in otherwise wild-type donor strains diminished the transfer of cognate but not heterologous substrates. Reciprocal swaps of AADs between PcfC and VirD4 abolished the transfer of cognate DNA substrates, although strikingly, the VirD4-AAD_PcfC chimera (VirD4 with the PcfC AAD) supported the transfer of a mobilizable plasmid. Purified AADs from both T4CPs bound DNA substrates without sequence preference but specifically bound cognate processing proteins required for cleavage at origin-of-transfer sequences. The soluble domains of VirD4 and PcfC lacking their AADs neither exerted negative dominance in vivo nor specifically bound cognate processing proteins in vitro. Our findings support a model in which the T4CP AADs contribute to DNA substrate selection through binding of associated processing proteins. Furthermore, MOBQ plasmids have evolved a docking mechanism that bypasses the AAD substrate discrimination checkpoint, which might account for their capacity to promiscuously transfer through many different T4SSs.

IMPORTANCE For conjugative transfer of mobile DNA elements, members of the VirD4/TraG/TrwB receptor superfamily bind cognate DNA substrates through mechanisms that are largely undefined. Here, we supply genetic and biochemical evidence that a helical bundle, designated the all-alpha domain (AAD), of T4SS receptors functions as a substrate specificity determinant. We show that AADs from two substrate receptors, Agrobacterium tumefaciens VirD4 and Enterococcus faecalis PcfC, bind DNA without sequence or strand preference but specifically bind the cognate relaxases responsible for nicking and piloting the transferred strand through the T4SS. We propose that interactions of receptor AADs with DNA-processing factors constitute a basis for selective coupling of mobile DNA elements with type IV secretion channels.

INTRODUCTION

Bacterial type IV secretion systems (T4SSs) translocate DNA and protein substrates to target cells, generally by a mechanism requiring direct cell-to-cell contact (1). These systems are phylogenetically widely distributed among many Gram-negative and -positive bacterial species (2, 3). In both cell types, the T4SSs function as conjugation machines by delivering DNA substrates to bacterial recipient cells (3–5). Many Gram-negative pathogens additionally have adapted T4SSs for translocation of effector proteins to eukaryotic target cells during the course of infection (6, 7). Nearly all T4SSs are composed of a transenvelope channel and a cytoplasmic membrane ATPase, termed a type IV coupling protein (T4CP), that functions as the receptor for DNA or protein substrates (8, 9). Structural studies have identified striking architectural features of conjugation channels elaborated by model Gram-negative bacterial systems. The plasmid pKM101 conjugation system, for example, encodes 3 subunits which together assemble in 14 copies each as a large ∼1-MDa barrel-shaped substructure termed the core complex (10, 11). This complex spans the cell envelope and has been postulated to function as a structural scaffold for the translocation channel (12). Very recently, a massive 3.2-MDa substructure, comprised of nearly the entire plasmid R388 conjugation system, presented as two distinct subassemblies, the outer-membrane-associated core complex (also termed the outer membrane complex [OMC]) and an even larger inner membrane complex (IMC) (13). The T4CP of this system, TrwB, was not part of this large structure, but several lines of genetic and biochemical evidence suggest it associates with the IMC through a combination of N-terminal transmembrane domain (NTD) and extramembranous contacts (14–17).

Despite these elegant new findings, fundamental questions remain about early-stage substrate recruitment and docking reactions, as well as the route of substrate passage across the cell envelope. A crystal structure exists for the soluble domain of one member of the T4CP receptor superfamily, R388-encoded TrwB. The protomer consists of two domains, a nucleotide-binding domain (NBD) similar to RecA and DNA ring helicases and a 7-helix bundle termed the all-alpha domain (AAD). When assembled as the homohexamer, the soluble domain presents as a sphere with overall dimensions of 110 Å in diameter and 90 Å high; a central channel 20 Å wide extends from the cytoplasmic base to the NTD (18). The NTD was shown by electron microscopy to form a 25-Å-wide appendix that, when modeled with the soluble domain's X-ray structure, gives rise to an overall F₁F₀-ATPase-like ball-stem structure with a central channel traversing its entire length (19, 20). The T4CPs are phylogenetically closely related, and therefore, the TrwB hexamer is considered a structural prototype for the receptor superfamily (8). T4CPs are also related to the FtsK and SpoIIIE DNA translocases, leading to a proposal that they might pump DNA substrates through their central channels across the cytoplasmic membrane (21).

The AAD is of special interest with respect to a possible role in substrate engagement. The AAD of TrwB bears structural similarity to N-terminal domain 1 (the N domain) of the site-specific recombinase XerD (18). In XerD, the N domain is partly responsible for binding DNA and coordinating the DNA cleavage reaction by the C-terminal domain (CTD) (22, 23). In TrwB, the AAD is situated at the entrance of the channel at the cytoplasmic pole of the hexamer and forms a constriction of 8 Å at the opening of the channel (21). At this position, the AAD might bind DNA substrates as a prerequisite for passage directly through the T4CP lumen or delivery to one of the other VirB4- or VirB11-like T4SS ATPases for further substrate processing prior to translocation (24–26). NBDs are highly conserved among the T4CP superfamily members, but the AADs vary considerably in size, sequence composition, and numbers of predicted α-helical folds (Fig. 1) (27).

FIG 1.

Sequence alignments and structures of AADs from the TrwB_R388, PcfC_pCF10, and VirD4_At T4CPs. (A) Schematic showing the domain architectures of VirD4 and PcfC, together with the structural prototype TrwB, with boundaries derived from sequence alignments and Phyre 2 structural modeling (27, 51). Domain boundaries are denoted by residue numbers, and the sizes of the AADs in residues are indicated. (B) Sequence alignments of the three AADs generated by ClustalW, with α-helices shown by colored shading corresponding to the colors in panel A, based on secondary-structure predictions using the Phyre 2 Web tool (51). Weakly (periods) or strongly (colons) conserved or identical (asterisks) residues are indicated. (C) Structures of the AADs solved by X-ray crystallography for TrwB (18) or predicted by Phyre 2 structural modeling (51) for PcfC and VirD4.

Our laboratory is investigating the mechanisms of action of paradigmatic T4SSs, including the Agrobacterium tumefaciens VirB/VirD4 system, which delivers transfer DNA (T-DNA) and effector proteins to susceptible plant cells, and the Enterococcus faecalis Prg/Pcf system, which mediates conjugative transfer of the pheromone-responsive plasmid pCF10 to recipient enterococci (28). Like other conjugation machines, these systems translocate their cargo substrates through formation of mating junctions with target cells. However, they exhibit important differences both in early-stage substrate-processing/docking reactions and in overall T4SS machine architectures, the latter due in large part to the diderm (two-membrane) versus monoderm (one-membrane) nature of the Gram-negative versus Gram-positive cell envelopes (2, 29, 30). Given that type IV secretion is a major underlying mechanism for both widespread dissemination of antibiotic resistance and virulence genes among bacterial populations and interkingdom translocation of effector proteins during infection, it is imperative to define key substeps of the translocation pathway. To this end, we focused here on the substrate receptor with a goal of testing the hypothesis that the AAD imparts specificity to the DNA substrate docking reaction.

VirD4 and PcfC are the receptors of the A. tumefaciens and E. faecalis systems (Fig. 1). Previously, we showed that VirD4 and PcfC form formaldehyde (FA)-cross-linkable complexes with DNA substrates in vivo independently of other channel subunits, confirming their roles as substrate receptors (25, 26, 30). Detection of substrate-receptor cross-linking in both A. tumefaciens and E. faecalis requires the DNA transfer and replication (Dtr) proteins, which assemble as the relaxosome at the origin-of-transfer (oriT) sequences and catalyze nicking of the DNA strand (T-strand) destined for transfer. In A. tumefaciens, these Dtr proteins include the VirD2 relaxase and accessory factor VirD1, which together bind and cleave the oriT-like sequences located within the T-DNA border repeats (31, 32). Two other accessory proteins, VirC1 and VirC2, bind near the nic site and stimulate the nicking reaction (33). VirC1, a member of the ParA/MinD ATPase superfamily, and another protein termed the VirD2-binding protein (Vbp) additionally function as spatial determinants to enhance the efficiency of the T-DNA substrate–VirD4 docking reaction (29, 34, 35). In E. faecalis, the PcfG relaxase and accessory factor PcfF suffice for nicking at the pCF10 oriT sequence in vitro (36) and FA cross-linking of the pCF10 substrate with PcfC in vivo (30). A soluble form of PcfC lacking its NTD (PcfCΔN103) binds single-stranded DNA (ssDNA) and double-stranded DNA (dsDNA) substrates without sequence or strand preference and also binds PcfG and PcfF in vitro (30).

We defined the contributions of the AADs of both T4CPs to DNA substrate transfer through phenotypic studies of AAD deletion and domain-swapped variants in vivo and by assaying for AAD binding of DNA and Dtr proteins in vitro. Our findings strongly support a model in which the AAD contributes to DNA substrate recognition through selective binding of cognate Dtr proteins. We also present evidence that the non-self-transmissible MOBQ plasmid RSF1010 has evolved a mechanism to bypass the AAD substrate discrimination checkpoint, which might account for the capacity of the plasmid to translocate through many different T4SSs of Gram-negative bacteria.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

A. tumefaciens strain A348 corresponds to strain C58 carrying the octopine-type pTiA6NC plasmid (37, 38). Strain KA2000 contains an in-frame deletion of virD4 from strain A348 obtained by marker exchange eviction mutagenesis (26). C58C1 Str^r Rif^r (39) served as the recipient in mating experiments. Strain LBA4404 has the oncogenic T-DNA deleted (40), and At12516 is a virE2 mutant (41). A. tumefaciens strains were grown at 28°C in MG/L medium, consisting of LB broth supplemented with mannitol and glutamate (42). The A. tumefaciens vir genes were induced at 22°C in AB inducing medium (ABIM) (glucose-containing minimal medium [pH 5.5], 1 mM phosphate, 200 μM acetosyringone [AS]), as previously described (42). When necessary, the medium was supplemented with the following antibiotics: gentamicin (100 μg/ml), kanamycin (100 μg/ml), tetracycline (5 μg/ml), carbenicillin (100 μg/ml), spectinomycin (500 μg/ml), streptomycin (500 μg/ml), and rifampin (300 μg/ml). All antibiotics were obtained from Sigma Chemical Co. E. faecalis OG1RF carrying plasmids of interest served as the donor strain and OG1SSp or OG1ES served as the recipient in mating experiments (43, 44). E. faecalis strains were cultured in brain heart infusion (BHI) at 37°C without shaking. E. faecalis was grown in the presence of the following antibiotics as needed: erythromycin (100 μg/ml final concentration for plasmid markers; 10 μg/ml for chromosomal markers), fusidic acid (25 μg/ml), rifampin (200 μg/ml), spectinomycin (1,000 μg/ml for plasmid markers; 250 μg/ml for chromosomal markers), streptomycin (1,000 μg/ml), and tetracycline (10 μg/ml). Escherichia coli strain DH5α served as the host for plasmid construction (Gibco-BRL), and E. coli strain BL21(DE3) was used for protein production (Novagen). E. coli strains were grown at 37°C in LB broth unless otherwise indicated.

Plasmid construction.

The plasmids and oligonucleotides used in this study are listed in Tables 1 and 2, respectively. When necessary for clarification, proteins are identified by their source in subscript, e.g., TrwB_R388, and domains by their protein origin, e.g., AAD_TrwB.

TABLE 1.

Plasmids used in this study

Plasmid	Description and/or relevant featuresa	Source or reference
Vectors
pPC914KS	Crbr; PvirB expression vector	48
pPC914KS.NcoI	Crbr; PvirB expression vector	74
pGEX6P-1	Crbr; expression vector for GST tagging	GE Healthcare
pMALc2x	Crbr; expression vector for MBP tagging	NEB
pET28b(+)	Kanr; expression vector for His tagging	Novagen
pAP1	Kanr; pET28a(+) derivative	75
pML122	Genr Kanr; broad-host-range, mobilizable RSF1010 (MOBQ)	76
pAJ1	Spcr; mobA1 derivative of RSF1010 (MOBQ)	76
pAJ6	Spcr; oriT1 derivative of RSF1010 (MOBQ)	76
pXZ153	Kanr; broad-host-range IncP cloning vector	77
pBBRMCS2	Genr; broad-host-range cloning vector	46
pDL278p23	Spcr; pDL278 shuttle vector with L. lactis P23 promoter	36
pCJK21	Ermr Spcr; shuttle vector with inducible Pnis promoter	78
pCIp23	Camr; pCI372 shuttle vector with P23 promoter	30
E. coli expression plasmids
pPC2004	pGEX6P-1 with Ptac::GST-AADVirD4; VirD4 residues R201 to R346	This work
pPC2013	pET28b(+) with PT7::His6-AADVirD4	This work
pPC2009	pMALc2x with Ptac::mbp-NBD/CTDVirD4 (ΔN87/ΔAADVirD4)	This work
pPC2039	pET28b(+) with PT7::His6-NBD/CTDVirD4	This work
pPC2044	pGEX6P-1 with Ptac::GST-NBD/CTDVirD4	This work
pPC2038	pET28b(+) with PT7::His6-VirD4ΔAADVirD4	This work
pPC2071	pAPI with PT7::His6-virD2	This work
pPC2072	pAPI with PT7::His6-virD2ΔC35	This work
pPC2214	pET28b(+) with PT7::His6-virD1	This work
pPC2101	pGEX6P-1 with Ptac::GST-AADPcfC; PcfC residues L204 to N360	This work
pPC2100	pET28b(+) with PT7::His6-AADPcfC	This work
pPC2102	pMALc2x with Ptac::mbp-NBD/CTDPcfC (ΔN103/ΔAADPcfC)	This work
pPC2016	pET28b(+) with PT7::His6-NBD/CTDPcfC	This work
pPC2027	pET28b(+) with PT7::His6-PcfCΔAADPcfC	This work
pCY33	pET28b(+) with PT7::His6-pcfF	30
pCY36	pET28b(+) with PT7::pcfG-His6	30
pPC67	pET28b(+) with PT7::prgJ-His6	50
A. tumefaciens expression plasmids
pKA9	pPC914KS with PvirB::virD4	64
pPC2011	pPC914KS with PvirB::AADVirD4	This work
pPC2078	pPC914KS with PvirB::AADPcfC	This work
pPC2020	pPC914KS with PvirB::VirD4ΔAAD	This work
pPC2022	pPC914KS with PvirB::virD4-AADPcfC chimera	This work
pPC2024	pPC914KS with PvirB::NBD/CTDVirD4 (ΔN87/ΔAADVirD4)	This work
E. faecalis expression plasmids
pCY25	pCIp23 with P23::pcfC	30
pPC2104	pDL278p23 with P23::AADPcfC; PcfC residues L204 to N360	This work
pPC2105	pDL278p23 with P23::pcfC-AADVirD4 chimera	This work
pPC2106	pDL278p23 with P23::AADVirD4; VirD4 residues R201 to R346	This work
pPC2107	pDL278p23 with P23::NBD/CTDPcfC (ΔN103/ΔAADPcfC)	This work
pPC2108	pDL278p23 with P23::PcfCΔAAD	This work
pPC2109	pCJK21 with Pnis::AADPcfC	This work

^aVector plasmids conferred resistance to the following antibiotics: carbenicillin (Crb), kanamycin (Kan), gentamicin (Gen), spectinomycin (Spc), chloramphenicol (Cam), and erythromycin (Erm).

TABLE 2.

Oligonucleotides used in this study

Purpose and oligonucleotide	Sequence	Reference
Cloning
D4F	5′ CGGTGAACATATGAATTCCAGCAA 3′
D4-SD-F	5′ CTCATCATACATATGCGCAATC 3′
D4R	5′ TCACTCGAGGCATCAGCCTG 3′
AAD_Nde	5′ AATACATATGGAGCGGAAGACTCATTGT 3′
D4AAD_For_Bam	5′ AATAGGATCCGGAGCGGAAGACTCATTGT 3′
D4AAD_For_Sac	5′ AATAGAGCTCGGAGCGGAAGACTCATTGT 3′
D4AAD_Rev	5′ AAACTCGAGTCATAGAGCTCCCCTCCGGAGATCGTAAAC 3′
D4AAD_Rev_Sac	5′ AATAGAGCTCCGCCTCCGGAGATCGTAAAC 3′
ΔAAD_Rev_Sac	5′ TGAAGAGCTCTCAGAATCTAGG 3′
ΔAAD_For	5′ AATAGAGCTCCAAGAAGACCTGCATTTAT 3′
CAAD-5	5′ GGAATTCCATATGTTGTCAAATAGTGATACG 3′
CAAD-3	5′ CCGCTCGAGTTAATTCCAGGAGTCAATATC 3′
CAAD-D1	5′ CGGAGCTCCGCGTGTTTAAGTCTAAGACCTT 3′
CAAD-D2	5′ GCGAGCTCCGAAGAAAAAACAGCAGTG 3′
pcfC_F_5	5′ AAACTGTCATATGCAGACGTAC 3′
pcfCΔN103-F	5′ GCTCGTTTTCATATGGCTACACCTGC 3′
pcfC-R-XhoI	5′ AACGCACGTCCTCGAGCTATTAGGCTAGAACGTGTGTTTC 3′
ΔN103R	5′ ACATGCATGCCTGCAGCTCGAGTTAAACGTGTGTTTCTTCTT 3′
GST-pcfC-Δ103-F	5′ CATGGTCTCGGATCCGCTCGTTTTGCTACACCTGC 3′
CAAD_For_Sac	5′ CGGAGCTCGTTGTCAAATAGTGATACG 3′
CAAD_Rev_Sac	5′ GCGAGCTCGAATTCCAGGAGTCAATATC 3′
D4AAD_For_Sac	5′ AATAGAGCTCGGAGCGGAAGACTCATTGT 3′
D4AAD_Rev_Sac	5′ AAACTCGAGTCATAGAGCTCCCCTCCGGAGATCGTAAAC 3′
EMSAs
pCF10-S	5′ AAATTCGCAACATGCTAGCATGTTGCTCCGCTTGCAAAAAGAAAGCCG3′	30
pCF10-AS	5′GATCCGGCTTTCTTTTTGCAAGCGGAGCAACATGCTAGCATGTTGCG 3′	30
RB-S	5′ AGCTCAAATTACAACGGTATATATCCTGCCAGTCAGCATCATCACACCAA 3′
RB-AS	5′TTGGTGTGATGATGCTGACTGGCAGGATATATACCGTTGTAATTTGAGCT 3′

(i) E. coli expression plasmids.

Plasmid pPC2100 produces N-terminally His₆-tagged AAD_PcfC (residues Leu204 to Asn360 of PcfC) from the P_T7 promoter. It was constructed by PCR amplification of bp 610 to 1080 of pcfC using primers CAAD-5 and CAAD-3, digestion of the PCR fragment with NdeI and XhoI, and introduction of the digested fragment into similarly digested vector pET28b(+). Plasmid pPC2101, producing glutathione S-transferase (GST)–AAD_PcfC, was constructed by excision of the AAD_PcfC fragment from pPC2100 by NdeI/XhoI digestion and introduction into similarly digested pCY34 (pGEX6p-pcfF [36]), substituting for the pcfF fragment. Plasmid pPC2016, producing an N-terminally His₆-tagged variant of PcfC lacking its NTD and AAD (here designated NBD/CTD_PcfC [Fig. 1]), was constructed by PCR amplification of bp 310 to 609 with primers pcfCΔN103-F and CAAD-D1 to generate the ΔAAD-up fragment and amplification of bp 1081 to 1830 using primers CAAD-D2 and pcfC-R-XhoI to generate the ΔAAD-down fragment. The ΔAAD-up fragment was digested with NdeI and SacI, the ΔAAD-down fragment was digested with SacI and XhoI, and both fragments were introduced into pET28b(+) digested with NdeI and XhoI. Plasmid pPC2027, producing His₆-PcfCΔAAD, was constructed by replacement of the ΔAAD-up fragment in PC16 with the full-length upstream region of pcfC (from nucleotide positions 1 to 609) generated by PCR amplification with primers pcfC-F-5 and CAAD-D1 and subsequent digestion with NdeI and SacI. Plasmid pPC2102, producing maltose-binding protein (MBP) fused to NBD/CTD_PcfC, was constructed by PCR amplification of NBD/CTD_PcfC from pPC2016 using primers GST-pcfC-Δ103-F and ΔN103R. The resulting PCR product was digested with BsaI and PstI and then cloned into vector pMALc-2x digested with BamHI and PstI. All PCR amplifications of virD4 fragments were performed using pMY1153 as the template (31). Plasmid pPC4, producing GST-AAD_VirD4, was constructed by amplification of the sequence encoding AAD_VirD4 (residues Arg201 to Arg346 of VirD4) using primers AAD_Nde and D4AAD_Rev, digestion with NdeI and XhoI, and introduction into similarly digested pPC2101, substituting for AAD_PcfC. Plasmid pPC2013, producing His₆-AAD_VirD4, was constructed by amplification of AAD_VirD4 using primers D4AAD_For_Bam and D4AAD_Rev, digestion with BamHI and XhoI, and introduction into similarly digested pET28b. Plasmid pPC2039, producing His₆-NBD/CTD_VirD4 (VirD4 with its 87-residue NTD and AAD deleted) was constructed by amplifying virD4 sequences located upstream and downstream of the AAD sequence with primers D4-SD-F and ΔAAD_Rev_Sac and ΔAAD_For and D4R, respectively. The upstream and downstream fragments were digested with NdeI/SacI and SacI/XhoI, respectively, and together introduced into pET28b(+). Plasmid pPC2038, producing His₆-VirD4ΔAAD, was constructed as described for pPC2039 except that the upstream fragment was amplified with primers D4F and ΔAAD_Rev_Sac. Plasmid pPC2044, producing GST-NBD/CTD_VirD4, was constructed by substituting the NdeI/XhoI fragment carrying NBD/CTD_VirD4 from plasmid pPC2039 for the AAD_VirD4 fragment of pPC4. Plasmid pPC2009, producing MBP-NBD/CTD_VirD4, was constructed by substituting a BamHI/XhoI fragment carrying NBD/CTD_VirD4 from pPC2044 for the AAD_VirD4 fragment of pMALc2x-AAD_VirD4 (pPC43). Plasmids pPC2071 and pPC2072, producing His₆-VirD2 and His₆-VirD2ΔC35, respectively, were constructed by introducing NdeI/XhoI fragments from pKA205 (29) and pKA172 (15) into similarly digested pAP1. Plasmid pPC2214, producing His₆-VirD1, was produced by introducing an NdeI/XhoI fragment from pKA204 (29) into similarly digested pET15b(+).

(ii) A. tumefaciens expression plasmids.

Plasmid pPC2011, producing AAD_VirD4 from the P_virB promoter, was obtained by substituting an NdeI/XhoI fragment carrying AAD_VirD4 from pPC2013 for an NdeI/XhoI fragment carrying GST in pZZ11 (45). Plasmids pPC2020 and pPC2024, producing VirD4ΔAAD and NBD/CTD_VirD4 from the P_virB promoter, respectively, were constructed by introducing NdeI/XhoI fragments from pPC2038 and pPC2039 into similarly digested pPC914KS. Plasmid pPC2078, producing AAD_PcfC from the P_virB promoter, was constructed by introducing the AAD_PcfC fragment from pPC2100 into similarly digested pPC914KS.NdeI. Plasmid pPC2022, producing VirD4-AAD_PcfC from the P_virB promoter, was constructed by amplification of the AAD_PcfC fragment with primers CAAD_For_Sac and CAAD_Rev_Sac, digestion with SacI, and introduction into SacI-digested pPC2038. From the resulting plasmid, an NdeI/XhoI fragment carrying the virD4-AAD_PcfC chimera was isolated and introduced into similarly digested pPC914KS. ColE1 plasmids expressing alleles of interest were introduced into A. tumefaciens by ligation to broad-host-range (BHR) plasmids pXZ153 (Kan^r) or pBBR1MCS2 (Gen^r) (46). Such cointegrants retained the ColE1 plasmid name plus “B” for BHR. For introduction of two plasmids into the same cell, the plasmids were first ligated to compatible BHR vectors, and the resulting cointegrant plasmids were introduced into A. tumefaciens with appropriate antibiotic selection.

(iii) E. faecalis expression plasmids.

Plasmid pPC2104, producing AAD_PcfC from the P₂₃ promoter, was obtained by introducing an NdeI/XhoI fragment carrying AAD_PcfC from plasmid pPC2100 into similarly digested pCY31 (30) to substitute for pcfCΔN103. Similarly, plasmids pPC2108, pPC2107, and pPC2106, producing PcfCΔAAD, NBD/CTD_PcfC, and AAD_VirD4 from the P₂₃ promoter, respectively, were constructed by introducing the corresponding fragments from plasmids pPC2027, pPC2016, and pPC13, respectively, into pPC31. Plasmid pPC2105, producing PcfC-AAD_VirD4 from the P₂₃ promoter, was constructed by amplification of the AAD_VirD4 sequence with primers D4AAD_For_Sac and D4AAD_Rev_Sac, digestion with SacI, and introduction into SacI-digested pPC2027. From the resulting plasmid, an NdeI/XhoI fragment carrying the pcfC-AAD_VirD4 chimera was isolated and introduced into similarly digested pCY31. Plasmid pPC2109, producing His₆-AAD_PcfC from the P_nis promoter, was obtained by introducing an NcoI-XhoI fragment carrying AAD_PcfC from pPC2100 into similarly digested pMSP3545 (47). The integrity of all cloned fragments was confirmed by DNA sequence analysis.

Conjugation assays.

A. tumefaciens donor strains carrying plasmid pML122Kan^r or pML122Gen^r were mated with the recipient strain C58C1RS as described previously (26). Briefly, exponential-phase cells (optical density at 600 nm [OD₆₀₀], ∼0.5) were harvested and diluted in induction medium (ABIM) for expression of vir genes for 8 h at 22°C. After induction, the cells were adjusted to an OD₆₀₀ of ∼0.5 before mixing donors and recipient in a ratio of 1:5. The mating mixtures were spotted onto nitrocellulose filters placed on an ABIM agar plate. After 4 days of growth at 18°C, cells were harvested from the filters, and serial dilutions were plated onto MG/L medium with appropriate antibiotic selection for transconjugant and donor cells. E. faecalis 2-h filter and broth matings were carried out as previously described (30). Briefly, donors (OG1RF background) and recipients (OG1ES or OG1SSp) were inoculated in BHI and grown overnight without shaking. Cultures were diluted 1:10 in BHI and grown for 1 h at 37°C without shaking. The donor and recipient strains were mixed in a ratio of 1:10 and allowed to mate in liquid without shaking, or 10 μl of the mating mixture was spotted onto nitrocellulose on BHI plates. After 2 h of mating, the filters were resuspended in 1 ml of 1× PBS. Mating mixtures from the liquid and solid-surface matings were serially diluted in 1× PBS, and the numbers of donors and transconjugants were obtained by plating on selective BHI agar plates. For both A. tumefaciens and E. faecalis matings, transfer efficiencies were determined by the transconjugant/donor (Tc/D) ratio. Conjugation experiments were repeated in triplicate at least three times, and the results are presented as averages of three experiments.

Plant virulence assays.

A. tumefaciens strains were tested for virulence by inoculating wound sites of Kalanchoe daigremontiana leaves as previously described (48). A348 and avirulent strain KA2000 (ΔvirD4) served as controls for the tumorigenesis assay. The virulence of strains or mixed infections was assayed at least three times on three separate leaves each time. Mixed infections were performed by coinoculation of two strains of interest on the same wound (49). Tumor formation was analyzed 6 weeks after inoculation.

Protein production and purification.

For detection of protein production in A. tumefaciens, cultures were induced in ABIM for 12 h and normalized to an OD₆₀₀ of 0.5, and cells from equivalent culture volumes were harvested by centrifugation at 5,000 × g for 10 min at 4°C. Cell pellets were resuspended in lysis buffer A (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1 mM EDTA, 20% sucrose, 10 mg/ml lysozyme, 1× complete protease inhibitor cocktail [Roche]) and incubated for 20 min at room temperature, and the cells were lysed by sonication (Branson Sonifier 250). The cell lysates were recovered by two rounds of centrifugation at 15,700 × g for 30 min to remove unbroken cells and cell debris. For corresponding analyses of protein abundance in E. faecalis, strains of interest were induced with 20 ng/ml of cCF10 peptide or nisin for 1.5 h, and cells were harvested by centrifugation. The cell pellet was frozen at −20°C, thawed to facilitate cell breakage, and then resuspended in lysis buffer A supplemented with 10 μg/ml mutanolysin and incubated for 20 min at 37°C. The suspension was sonicated and centrifuged twice at 15,700 × g for 30 min to remove unbroken cells and cell debris. Soluble lysates of A. tumefaciens and E. faecalis were mixed with 5× Laemmli buffer, boiled for 5 min, and loaded on SDS-polyacrylamide gels on a per cell (CFU) equivalent basis. The proteins were transferred onto nitrocellulose membranes, and blots were developed with anti-VirD4 or -PcfC antibodies and goat anti-rabbit secondary antibodies conjugated to alkaline phosphatase for visualization (Bio-Rad). Molecular mass markers were from Gibco-BRL (Grand Island, NY). AAD_VirD4 and AAD_PcfC were highly enriched as follows. Overnight cultures (2.5 ml) of strains carrying pPC2004 (GST-AAD_VirD4), pPC2101 (GST-AAD_PcfC), or pGEX6p (GST only) were inoculated into 250 ml fresh LB medium. The cultures were incubated with shaking at 37°C to an OD₆₀₀ of 0.4, induced with 0.8 mM (final concentration) IPTG (isopropyl-β-d-thiogalactopyranoside), and incubated with shaking at room temperature (RT) for 3 h. The cells were pelleted by centrifugation at 6,000 × g for 10 min. The pellet was resuspended in 2 ml of lysis buffer A, incubated at 37°C for 10 min, and lysed by sonication. The lysates were clarified by two rounds of centrifugation at 15,000 × g at 4°C for 30 min, and the resulting supernatant was mixed with 100 μl of glutathione-Sepharose beads and gently shaken for 1 h at 4°C. The mixture was loaded onto a 1.0- by 10-cm column, and the beads were washed 5 times with wash buffer (50 mM Tris-HCl [pH 7.4], 0.15 M NaCl, 1 mM EDTA, 1 mM dithiothreitol [DTT]) to remove unbound protein. Bound GST-AAD protein or GST alone was eluted from the column with buffer A supplemented with 10 mM glutathione. Alternatively, the GST tag was cleaved from the AADs by treatment with PreScission enzyme (2 units) at 4°C for 6 h. The protein-Sepharose mixture was centrifuged for 5 min at 0.5 × g, and the supernatant was mixed with 50 μl of glutathione-Sepharose beads for 2 h at 4°C to remove residual GST and PreScission enzyme. The purified AAD_PcfC or AAD_VirD4 fragments were recovered by centrifugation for 5 min at 500 × g. His₆-tagged forms of PcfF, PcfG, VirD1, VirD2, VirD2ΔC35, and PrgJ were highly enriched by Co²⁺ affinity chromatography as previously described (30, 50). Briefly, 500-ml cultures of BL21(DE3) strains carrying the relevant plasmids were grown at 37°C to an OD₆₀₀ of 0.4, induced with IPTG (0.1 mM final concentration), and incubated at room temperature for 3 h. Cells were harvested by centrifugation at 6,000 × g for 10 min, resuspended in 5 ml buffer A, and incubated for 10 min at 37°C. The cells were sonicated, and the lysate was centrifuged at 40,000 × g at 4°C for 30 min. The supernatant was mixed with 1 ml of Co²⁺ (Talon; Clontech) affinity resin, washed extensively with buffer A containing 10 mM imidazole, and incubated for 12 h at 4°C. The bound proteins were eluted with buffer A containing 200 mM imidazole. Soluble domains of VirD4 and PcfC lacking their NTDs and AADs (designated NBD/CTDs) were enriched as described above for enrichment of the GST-AADs, except that cell lysates were mixed with 100 μl of amylose resin and gently shaken for 1 h at 4°C. The beads were washed 5 times with wash buffer (50 mM Tris-HCl [pH 7.4], 0.15 M NaCl, 1 mM EDTA, 1 mM DTT) to remove unbound protein, and the bound MBP-NBD/CTD protein or MBP alone was eluted from the column with buffer A (minus lysozyme) supplemented with 10 mM maltose.

EMSAs.

Binding of purified AADs and NBD/CTDs to DNA substrates was assessed by electrophoretic mobility shift assays (EMSAs) using ssDNA and dsDNA substrates. The ssDNA substrates included 50-mers corresponding to the sense or antisense strands of the minimal oriT region of plasmid pCF10 (30) or the nic region of the T-DNA right-border repeat (38). The ssDNA oligonucleotides were synthesized and purified by PAGE (Sigma-Aldrich). A 39-bp fragment provided in the DIG gel shift kit (Roche) was used as a short, nonspecific dsDNA substrate. ssDNA and dsDNA substrates were labeled at the 3′ ends with digoxigenin (DIG)-11-ddUTP using the DIG gel shift kit (Roche) according to the manufacturer's instructions. EMSAs were carried out in 20-μl reaction mixtures containing the recombinant proteins, DIG-labeled ssDNA or dsDNA, 0.2 μg of poly(dA-dT), 0.1 μg of poly-l-lysine, and 1× reaction buffer (20 mM Tris-HCl [pH 8.0], 0.1 M KCl, 10% glycerol, 5 mM EDTA). The reaction mixtures were incubated at RT for 15 min and loaded onto 8% polyacrylamide gels in 0.5× Tris-borate-EDTA buffer (pH 7.9). After electrophoresis at 100 V for 1.5 h, the gels were transferred to a nylon membrane using the Genie electrophoretic blotter (Idea Scientific). DIG-labeled DNA was visualized by an enzyme immunoassay according to the manufacturer's instructions (DIG gel shift kit; Roche).

Protein interaction assays.

For GST pulldown assays, soluble fractions of E. coli cells producing GST-AAD_VirD4, GST-AAD_PcfC, or GST alone were mixed with 100 μl of glutathione-Sepharose beads for 1 h at RT. The Sepharose beads were pelleted by centrifugation at 500 × g for 5 min and washed three times with 1× physiologically buffered saline (PBS). The beads were resuspended in 500 μl of binding buffer (20 mM Tris-HCl, pH 8.0, 0.1 M NaCl, 10 μg bovine serum albumin). Enriched His₆-VirD1, His₆-VirD2, His₆-PcfF, PcfG-His₆, or PrgJ-His₆ (20 μg) protein, prepared as described above, was incubated with the beads prebound with GST or the GST-AAD fusions for 3 h at RT. The beads were pelleted, and extensively washed beads were eluted twice with elution buffer (10 mM glutathione, 0.1 M NaCl). Cell lysates containing the His₆-tagged proteins were prebound with 100 μl of Co²⁺ resin as described above, and the beads with bound proteins recovered by centrifugation were mixed with enriched forms of GST-AAD_VirD4, GST-AAD_PcfC, or GST alone. After incubation at 4°C for 3 h, the resins were washed extensively with binding buffer with 20 mM imidazole, and the bound proteins were eluted with buffer A containing 250 mM imidazole. The eluted fractions were assayed for PcfC, PcfF, and PcfG proteins by SDS-PAGE and immunostaining. For the MBP pulldowns, cell lysates containing the MBP-NBD/CTD_VirD4 and MBP-NBD/CTD_PcfC proteins, or MBP alone, were mixed with 100 μl of amylose resin, and the resin with bound proteins recovered by centrifugation was mixed with enriched forms of the His₆-tagged Dtr proteins. After incubation at 4°C for 3 h, the resins were washed extensively with binding buffer, and the bound proteins were eluted with binding buffer containing 10 mM maltose. The eluates from the pulldown assays were analyzed by SDS-PAGE and immunostaining.

RESULTS

The all-alpha-domains of VirD4 and PcfC are essential for T4CP function and exert negative dominance.

We identified domains of VirD4 and PcfC by protein sequence alignments and structure predictions based on the X-ray structure of TrwB and by using the Phyre 2 algorithm (51) (Fig. 1A). The AADs of VirD4 and PcfC share a low degree of primary sequence similarity with each other and with that of TrwB (Fig. 1B). AAD_VirD4 and AAD_PcfC also are larger and are predicted to adopt distinct α-helical bundle folds in comparison to the 7-helix bundle of AAD_TrwB (Fig. 1C).

To evaluate the functional importance of the AADs, we deleted the AADs of both receptors and assayed for effects on substrate transfer (Fig. 2). Native VirD4, as well as a VirD4ΔAAD variant, accumulated at comparable levels when produced from the AS-inducible P_virB promoter, yet only the native protein supported T-DNA transfer and tumor production on wounded K. daigremontiana leaves (Fig. 2A). Similarly, although native PcfC and the PcfCΔAAD mutant protein accumulated at high levels in E. faecalis donors, only the former supported high-frequency plasmid transfer in both liquid and solid-surface (filter) matings (Fig. 2B). We also tested whether separate production of the AAD and ΔAAD might reconstitute a functional T4CP by production of both protein pairs in A. tumefaciens and E. faecalis, but neither of these strains transferred their cognate DNA substrates at detectable levels (Fig. 2).

FIG 2.

Contributions of the AADs of VirD4 and PcfC to T4CP function in vivo. (A) Virulence of A. tumefaciens strains producing variant forms of VirD4 on K. daigremontiana. (Top) Leaf wound sites were inoculated with A. tumefaciens A348 (WT) producing native VirD4 or the ΔvirD4 mutant KA2000 (ΔvirD4). Where indicated, strains carried broad-host-range plasmids producing native VirD4, VirD4ΔAAD, AAD_VirD4, or AAD_PcfC. The indicated wound sites were inoculated with the following strain (the plasmid names correspond to those of the ColE1 plasmids plus “B” for BHR): 1, A348; 2, KA2000; 3, KA2000(pPCB2020); 4, KA2000(pPCB2020, pPCB2011); 5, A348(pPCB2011); 6, A348(pPCB2078); 7, KA2000(pKA9). (Bottom) Immunoblot showing steady-state levels of VirD4 (72 kDa) and VirD4ΔAAD (56 kDa) produced in A. tumefaciens A348 (WT) or KA2000 (ΔvirD4) strains. The lane numbers of the analyzed strains correspond to those presented in the virulence assay. (B) (Top) Histogram showing transfer frequencies of pCF10 from OG1RF donors producing variant forms of PcfC in 2-h filter and liquid matings, presented as the number of transconjugants per donor cell (Tc/Donor). Experiments were repeated at least 3 times, and the histogram depicts the average values with standard deviations. (Middle) Immunoblot showing steady-state levels of PcfC (67 kDa) and PcfCΔAAD (50 kDa) in E. faecalis OG1RF strains. Lanes: 1, OG1RF(pCF10); 2, OG1RF; 3, OG1RF(pCF10ΔpcfC, pPC2108); 4, OG1RF(pCF10ΔpcfC, pPC2108, pPC2109); 5, OG1RF(pCF10, pPC2104); 6, OG1RF(pCF10, pPC2106). (Bottom) Schematic showing analyzed OG1RF strains that produced combinations of native PcfC or PcfCΔAAD with the AAD_PcfC or AAD_VirD4 variants shown.

If the AAD functions in substrate binding, its production in an otherwise wild-type (WT) background might exert negative dominant effects on transfer, for example, through substrate binding and sequestration. To examine this possibility, we assayed for inhibitory effects accompanying AAD production on DNA transfer. Strikingly, A. tumefaciens A348 expressing P_virB::AAD_VirD4 from a BHR plasmid was strongly suppressed in its capacity to induce tumor formation on susceptible plants (Fig. 2A). Similarly, E. faecalis OG1RF(pCF10) donors engineered to express P₂₃::AAD_PcfC transferred pCF10 at frequencies 3 orders of magnitude lower than WT donors in both liquid and solid-surface matings (Fig. 2B). Our antibodies to VirD4 and PcfC did not react against the respective AADs, preventing assessments of steady-state accumulation of these domains. Next, we tested whether the AADs inhibited transfer of DNA substrates in the heterologous species. We expressed P_virB::AAD_PcfC in A. tumefaciens WT A348 and, conversely, P₂₃::AAD_VirD4 in E. faecalis strain OG1RF(pCF10). Neither domain exerted negative dominance, as monitored by A. tumefaciens-mediated T-DNA transfer to plants or conjugative transfer of pCF10 to E. faecalis recipients (Fig. 2).

A VirD4-AAD_PcfC chimera supports transfer of a mobilizable MOBQ plasmid through the A. tumefaciens VirB/VirD4 T4SS.

The above-mentioned findings were consistent with a role for receptor AADs in binding of cognate DNA substrates. To further test this model, we reciprocally swapped the AADs of VirD4 and PcfC and assayed for functionality of the chimeric proteins. The chimeric proteins VirD4-AAD_PcfC and PcfC-AAD_VirD4 accumulated at detectable levels in A. tumefaciens and E. faecalis donor strains but did not support T-DNA transfer to plant cells or plasmid transfer to E. faecalis recipients (Fig. 3A and B). Strikingly, however, A. tumefaciens donors producing VirD4-AAD_PcfC transferred the mobilizable plasmid pML122 (RSF1010 derivative; MOBQ incompatibility group) through the VirB channel to A. tumefaciens recipients, establishing that this chimeric protein engages productively with the plasmid transfer intermediate and the VirB channel (Fig. 3C). As expected, VirD4-AAD_PcfC did not support the transfer of RSF1010 derivatives sustaining deletions of the oriT sequence (pAJ6) or the mobA relaxase (pAJ1) (Fig. 3C). VirD4 with its AAD deleted also did not mobilize pML122 transfer at detectable frequencies, which confirms the requirement for an AAD—from either VirD4 or PcfC—for transmissibility of the MOBQ plasmid through the VirB/VirD4 T4SS. We also tested whether production of AAD_VirD4 or AAD_PcfC disrupted pML122 transfer through the native VirB/VirD4 T4SS, but neither AAD exerted dominant-negative effects (Fig. 3). These findings suggest that the MOBQ plasmid substrate is capable of docking with VirD4 through a mechanism that bypasses an AAD substrate discrimination checkpoint.

FIG 3.

Functionality of T4CP chimeras in A. tumefaciens and E. faecalis. (A) Virulence of A. tumefaciens strains producing variant forms of VirD4 on K. daigremontiana. (Top) Leaf wound sites were inoculated with A. tumefaciens A348 (WT) producing native VirD4 or the ΔvirD4 mutant KA2000 (ΔD4). Where indicated, KA2000 was engineered to produce native VirD4 or the VirD4-AAD_PcfC (D4-AAD_C) chimera. Mixed infections were carried out using strains Mx358 (virE2 mutant; translocates T-DNA) and LBA4404 (ΔT-DNA; translocates VirE2). (Bottom) Immunoblot showing steady-state levels of VirD4 (72 kDa) and VirD4-AAD_PcfC (73 kDa), labeled - and *, respectively, detected by development with anti-VirD4 antibodies. The strains for the virulence assays and immunoblot are numbered as follows: 1, A348; 2, KA2000 (ΔvirD4); 3, KA2000(pKA9); 4, KA2000(pPCB2022); 5, LBA4404 (ΔT-DNA); 6, Mx358 (virE2 mutant); 7, LBA4404 × Mx358 coinoculation; 8, KA2000(pPCB2022) × LAB4404 coinoculation; 9, KA2000(pPCB2022) × Mx358 coinoculation. (B) (Top) Transfer frequencies of pCF10 from donors producing variant forms of PcfC in 2-h filter and liquid matings, presented as the number of transconjugants per donor cell (Tc/Donor), with average values and standard deviations shown. 1, OG1RF(pCF10); 2, OG1RF(pCF10ΔpcfC); 3, OG1RF(pCF10ΔpcfC, pPC2105). (Bottom) Immunoblot showing steady-state levels of PcfC (67 kDa) and PcfC-AAD_VirD4 (66 kDa), labeled as - and *, respectively, detected by development with anti-PcfC antibodies. The lower bands correspond to an unknown cross-reactive species. (C) (Top) Transfer of the MOBQ plasmid pML122 or, where indicated, the ΔoriT (pAJ6) or mobA mutant (pAJ1) variant through the A. tumefaciens VirB/VirD4 T4SS. 1, A348(pML122); 2, A348(pML122, pPCB2011); 3, A348(pML122, pPCB2078); 4, KA2000(pML122, pPCB2022); 5, KA2000(pAJ6, pPCB2022); 6, KA2000(pAJ1, pPCB2022); 7, KA2000(pML122, pPCB2020). (Bottom) Analyzed A. tumefaciens strains that produced combinations of VirD4, VirD4-AAD_PcfC, or VirD4ΔAAD with the AAD_VirD4 or AAD_PcfC variants shown.

In the A. tumefaciens transfer system, the efficiency of T-DNA delivery to the plant nucleus is significantly enhanced by cotranslocation of effector proteins, such as the VirE2 single-stranded DNA-binding protein (SSB). Upon cotransfer of the T-DNA transfer intermediate (VirD2–T-strand) and effector proteins into a plant cell, the latter bind and facilitate delivery of the T-DNA intermediate to the nucleus (52). Interestingly, coinoculation on the same wound site of two avirulent strains, a ΔT-DNA mutant and a virE2 mutant, incites WT tumor production as a result of VirE2 transfer by the former strain and T-DNA transfer by the latter into the same plant cell (Fig. 3A) (49, 53). To determine if VirD4-AAD_PcfC was selectively impaired in mediating T-DNA or VirE2 effector translocation, we coinoculated plant wound sites with a strain producing the chimeric receptor with either the ΔT-DNA mutant or the virE2 mutant. Neither mixed infection incited tumor production (Fig. 3A), indicating that the chimeric protein fails to engage productively with either substrate despite its capacity to support translocation of the MOBQ plasmid.

The AADs of VirD4 and PcfC bind DNA substrates without sequence specificity.

We next assayed for binding of the AADs to DNA and Dtr components of the relaxosome in vitro. Previously, we showed that a His₆-tagged soluble form of PcfC, His₆-PcfCΔN103, binds ssDNA and dsDNA substrates without sequence preference in vitro (30). We were unable to enrich His₆-tagged forms of AAD_VirD4 and AAD_PcfC due to problems of insolubility, but the addition of a GST tag followed by removal of the tag by PreScission protease cleavage yielded highly enriched soluble forms of both domains (Fig. 4A).

FIG 4.

DNA-binding activities of AAD_VirD4 and AAD_PcfC. (A) Enrichment of GST-tagged and untagged AAD_VirD4 and AAD_PcfC by glutathione affinity chromatography, as described in the text. Lanes: 1 and 2, AAD_VirD4 and AAD_PcfC following removal of GST tags; 3 and 4, GST-tagged AAD_VirD4 and AAD_PcfC; 5, GST alone; M, molecular size markers (in kilodaltons). (B) EMSAs testing for binding of AAD_VirD4 to the 3′-end-DIG-labeled, dsDNA and ssDNA substrates listed above each gel. Lanes for the dsDNA 39-mer: 1, DNA probe only; 2 to 5, DNA probe plus AAD_VirD4 at concentrations of 72 nM (2), 144 nM (3), and 504 nM (3 to 5). In the last two lanes, AAD_VirD4 was preincubated with unlabeled dsDNA competitor at concentrations 10- and 100-fold higher than that of the DIG-labeled substrate. Lanes for the ssDNA probes: 1, DNA probe only; 2, DNA probe plus AAD_VirD4 (576 nM); 3, same as lane 2 except that AAD_VirD4 was preincubated with unlabeled competitor DNA at a 100-fold-higher concentration than the DIG-labeled substrate. (C) EMSAs testing for binding of AAD_PcfC to the 3′-end-DIG-labeled dsDNA and ssDNA substrates listed above each gel. The lanes are as described for panel B using the same AAD_PcfC protein concentrations.

By EMSAs (30, 50), we determined that both AAD_VirD4 and AAD_PcfC bound ssDNA and dsDNA substrates without sequence preference (Fig. 4B and C). Both AADs formed complexes with a 39-bp nonspecific dsDNA substrate and with ssDNA 50-mers corresponding to the sense or antisense strands of the pTiA6NC oriT-like right-border repeat (for AAD_VirD4) or the pCF10 oriT sequence (for AAD_PcfC). Although we could detect binding of the AADs to longer 186- to 200-bp dsDNA substrates spanning the pCF10 oriT or A. tumefaciens T-DNA right-border repeat, these results were not reproducible, possibly because of weak-affinity interactions (data not shown). For all tested substrates, DNA binding was inhibited, at least to some extent, by preincubation with unlabeled competitor DNAs (Fig. 4B and C).

The AADs of VirD4 and PcfC bind only cognate Dtr processing factors.

VirD4 and PcfC form FA-cross-linkable complexes with DNA substrates in vivo, but only in strains competent for assembly of the relaxosome at the T-DNA border or oriT sequences (25, 30). Furthermore, purified His₆-PcfCΔN103 binds the relaxosome components PcfF and PcfG, which together catalyze nicking within the pCF10 oriT sequence (30). To test whether the AADs of VirD4 and PcfC bind one or more of the Dtr processing proteins in vitro, we obtained highly enriched GST-tagged forms of the AADs (Fig. 4A) and His₆-tagged forms of PcfF and PcfG, as well as the corresponding His₆-tagged forms of A. tumefaciens VirD1 and VirD2 required for nicking at T-DNA border sequences (Fig. 5). For pulldown assays, GST-AAD_VirD4 or GST alone was prebound to glutathione affinity resin, the protein-bound resin was mixed with the His-tagged Dtr factors, and complexes were eluted with glutathione. As shown in Fig. 5A, His₆-VirD2 coeluted with GST-AAD_VirD4 but not GST alone. In reciprocal pulldown assays, His₆-VirD2 prebound to Co²⁺ resin formed a complex with GST-AAD_VirD4 but not GST alone. Despite repeated efforts, we were unable to detect an interaction between His₆-VirD1 and GST-AAD_VirD4 (Fig. 5A). In contrast, both His₆-PcfF and PcfG-His₆ from E. faecalis formed complexes with GST-AAD_PcfC but not GST alone (Fig. 5C).

FIG 5.

Interactions between the AADs and Dtr proteins. (A) (Top) Cobalt affinity pulldown assays testing for GST-AAD_VirD4 interactions with Dtr proteins. His₆-VirD1, His₆-VirD2, or His₆-VirD2ΔC35 was highly enriched by cobalt affinity chromatography (lanes labeled “load”) and bound to Co²⁺ beads, and the protein-bead complexes were mixed with GST-AAD_VirD4 (second lane) or GST alone (third lane). Following extensive washes, the bound proteins were eluted from the beads with imidazole, and proteins recovered in the eluates were identified by immunostaining with anti-GST and anti-His antibodies (shown on the left of each gel). (Bottom) Reciprocal GST affinity pulldowns. Enriched GST-AAD_VirD4 or GST alone was bound to glutathione beads, and the protein-bead complexes were mixed with the His-tagged proteins shown at the top. Following extensive washes, the bound proteins were eluted from the beads with glutathione and identified by immunostaining with anti-GST and anti-His antibodies (shown on the left of each gel). (B) Cobalt and GST affinity pulldown assays testing for GST-AAD_PcfC interaction with His₆-VirD2, as described for panel A. (C) Cobalt and GST affinity pulldown assays testing for GST-AAD_PcfC interaction with His₆-PcfF and PcfG-His₆, as described for panel A. (D) Cobalt and GST affinity pulldown assays testing for GST-AAD_VirD4 interaction with PcfG-His₆, as described for panel A.

Full-length VirD2, and just a C-terminal fragment of the relaxase, mediate intercellular transfer of Cre when fused to the recombinase, as monitored by excision of a lox cassette in engineered plant cells using the Cre recombinase assay for translocation (CRAfT) (54, 55). Conversely, C-terminal-truncation mutations block relaxase-mediated T-DNA transfer to plant cells (53, 55). Recently, we further showed that, although VirD2 with its C-terminal 35 residues deleted (VirD2ΔC35) exhibits wild-type relaxase activity, the processed T-DNA substrate fails to interact productively with VirD4 (15). To examine whether VirD2's C terminus contributes to the VirD2-AAD_VirD4 interaction, we repeated the pulldown assays using a His₆-VirD2ΔC35 variant. We were unable to detect an interaction between the ΔC35 relaxase mutant and GST-AAD_VirD4 (Fig. 5A), suggesting that VirD2's C terminus participates directly or indirectly in mediating the AAD_VirD4 interaction.

Taken together, the above-mentioned findings suggest that the AADs of PcfC and VirD4 participate in a network of interactions with DNA and relaxosome components and also might contribute to substrate discrimination through binding of cognate relaxases. To further test for a possible discrimination function, we assayed for binding of relaxases with the heterologous GST-AAD_PcfC or GST-AAD_VirD4 variant. Although His₆-VirD2 bound GST-AAD_VirD4 (Fig. 5A), we were unable to detect binding with GST-AAD_PcfC (Fig. 5B). Similarly, His₆-PcfF and PcfG-His₆ bound GST-AAD_PcfC (Fig. 5C), but PcfG-His₆ failed to bind GST-AAD_PcfC (Fig. 5D). The AADs of VirD4 and PcfC thus selectively bound only their cognate relaxases, and AAD_PcfC additionally bound the accessory factor PcfF.

VirD4 and PcfC soluble domains lacking the AADs exhibit nonspecific substrate binding.

Our data indicated that receptor AADs confer specificity for binding of DNA substrates through Dtr protein interactions but did not exclude the possibility that other T4CP domains, including the conserved NBDs, as well as the sequence-variable CTDs present in a large number of T4CPs (8), also contribute to the substrate docking reaction (Fig. 1). We tested the potential for soluble domains of VirD4 and PcfC lacking their AADs, designated NBD/CTD_VirD4 and NBD/CTD_PcfC, respectively, to bind substrates. In the in vivo genetic tests for negative dominance, neither NBD/CTD_VirD4 nor NBD/CTD_PcfC suppressed substrate transfer when produced in otherwise wild-type A. tumefaciens A348 or E. faecalis OG1RF(pCF10) strains, respectively (Fig. 6). These truncation mutants thus failed to sequester substrate or otherwise impair T4SS machine function, despite accumulating at levels comparable to or greater than those of the native T4CP in the merodiploid strains (Fig. 6).

FIG 6.

Effects of producing NBD/CTD truncation mutants on native VirD4 and PcfC function in vivo. (A) Virulence of A. tumefaciens strains producing VirD4 variants on K. daigremontiana. (Top) Leaf wound sites were inoculated with A. tumefaciens A348 producing native VirD4 (WT) or the ΔvirD4 mutant KA2000 (ΔvirD4) without or with the NBD/CTD_VirD4 expression plasmid. (Bottom) Immunoblot showing steady-state levels of VirD4 (72 kDa) and NBD/CTD_VirD4 (46 kDa) produced in the strains listed and detected by development with anti-VirD4 antibodies. Strains for virulence assays and immunoblot: 1, A348; 2, KA2000 (ΔvirD4); 3, A348(pPCB2024); 4, KA2000(pPCB2024). (B) (Top) Transfer frequencies of pCF10 from donors producing variant forms of PcfC in 2-h filter and liquid matings, presented as the number of transconjugants per donor cell (Tc/Donor), with average values and standard deviations shown. 1, OG1RF(pCF10); 2, OG1RF(pCF10ΔpcfC); 3, OG1RF(pCF10, pPC2107). (Bottom) Immunoblot showing steady-state levels of PcfC (67 kDa) and NBD/CTD_PcfC (39 kDa) detected by development with anti-PcfC antibodies.

We were unable to enrich His₆- or GST-tagged forms of NBD/CTD_VirD4 or NBD/CTD_PcfC due to aggregation, but MBP fusions were soluble in E. coli and amenable to enrichment (Fig. 7A). Interestingly, both MBP-tagged NBD/CTDs shifted all DNA substrates tested in the EMSAs, including the 39-mer nonspecific dsDNA and short ssDNA substrates corresponding to both strands of the pCF10 oriT or A. tumefaciens T-DNA region (Fig. 7B). In control experiments, enriched MBP (Fig. 7A) did not detectably bind these DNA substrates (Fig. 7B). Thus, the NBDs or CTDs of both T4CPs bind DNA without sequence or strand preference.

FIG 7.

In vitro binding activities of NBD/CTD_VirD4 and NBD/CTD_PcfC. (A) Coomassie-stained protein gel showing enrichments of MPB-NBD/CTD_VirD4 (82-kDa), MBP-NBD/CTD_PcfC (88-kDa), and MBP alone (migrating at ∼48-kb) by amylose affinity chromatography. M, molecular size markers. (B) EMSAs showing binding of MBP-NBD/CTD_PcfC (top) and MBP-NBD/CTD_VirD4 (bottom) to the 3′-end-DIG-labeled nonspecific dsDNA 39-mer and the ssDNA 50-mer substrates encompassing the nicked strands of pCF10 oriT (for MBP-NBD/CTD_PcfC) or T-DNA right-border repeat (for MBP-NBD/CTD_VirD4). NBD/CTD_PcfC EMSA lanes: DNA probe plus MBP or NBD/CTD_PcfC (1) at 84 nM (2), 168 nM (3), 336 nM (4), or 504 nM (5 and 6). In lanes 5 and 6, the protein was preincubated with unlabeled 39-mer dsDNA competitor or 50-mer ssDNAs at concentrations 10- and 100-fold higher than that of the DIG-labeled substrate. NBD/CTD_VirD4 EMSA lanes: (1) DNA probe alone or (2 to 5) plus NBD/CTD_VirD4 at the same concentrations as the corresponding lanes of NBD/CTD_PcfC EMSAs or (6) plus MBP. (C) MBP affinity pulldowns assaying for interactions of NBD/CTD_PcfC with Dtr proteins. Enriched MBP-NBD/CTD_PcfC or MBP alone (lanes labeled “load”) was bound to amylose beads, and the protein-bead complexes were mixed with enriched His₆-PcfF, PcfG-His₆, or His₆-VirD2. Following extensive washes, the bound proteins were eluted from the beads with maltose, and proteins recovered in the eluates were identified by immunostaining with anti-MBP or anti-His antibodies (shown on the left of each gel). (D) MBP or GST affinity pulldowns assaying for interactions of PrgJ-His₆ with MBP-NTD/CTD_PcfC or GST-AAD_PcfC, as described for panel C and Fig. 5A, respectively.

We were unable to detect binding of the MBP-NBD/CTD_VirD4 variant with cognate VirD1 or VirD2 or heterologous PcfF and PcfG (data not shown). In contrast, MBP-NBD/CTD_PcfC bound not only the cognate Dtr proteins His₆-PcfF and PcfG-His₆ but also heterologous His₆-VirD2 (Fig. 7C). MBP alone did not detectably bind these His-tagged Dtr proteins (Fig. 7C). Although MBP-NBD/CTD_VirD4 bound DNA substrates, we cannot exclude the possibility that this variant fails to bind Dtr proteins because of a folding defect. Even so, our in vivo and in vitro findings together suggest that the AADs and not the NBDs or CTDs of VirD4 and PcfC confer DNA substrate recognition through selective binding of cognate Dtr proteins.

Finally, once PcfC binds the pCF10 substrate, the receptor delivers its cargo to the VirB4-like PrgJ ATPase, possibly for further processing prior to translocation (50). Soluble PcfCΔN103 binds PrgJ in vitro (50), prompting us to test whether AAD_PcfC or NBD/CTD_PcfC participates in PrgJ binding. As shown in Fig. 7D, PrgJ-His₆ coeluted with MBP-NBD/CTD_PcfC but not MBP alone. In E. coli, PrgJ accumulates as both a full-length protein and a presumptive degradation product (50), and the NBD/CTD_PcfC variant bound both forms of this ATPase. In contrast, in the corresponding GST pulldown assays, PrgJ-His₆ did not form a detectable complex with GST-AAD_PcfC (Fig. 7D). These findings suggest that PcfC's NBD or CTD, or both domains, mediates receptor-PrgJ contacts necessary for the first step of substrate transfer through the Prg/Pcf T4SS.

DISCUSSION

In this study, we defined the contributions of the AADs of T4CPs from the A. tumefaciens VirB/VirD4 and E. faecalis Prg/Pcf systems to binding of DNA substrates. We presented evidence that the AADs of A. tumefaciens VirD4 (VirD4_At) and E. faecalis pCF10-encoded PcfC (PcfC_pCF10) participate in selection of T-DNA and pCF10 substrates, respectively, through binding of the associated Dtr processing proteins. Other domains of these T4CPs, including the NBDs and CTDs, also might participate in substrate binding but lack the discriminating function of the AAD for cognate DNA substrates. Furthermore, we identified differences between T4CP domain-Dtr protein contacts in the A. tumefaciens and E. faecalis transfer systems that might be explained by variations in early-stage substrate processing and recruitment reactions evolved by these two systems. Finally, our analyses of chimeric T4CPs suggested that MOBQ plasmids have evolved the capacity for promiscuous transfer through different T4SSs, at least in part by evolving a mechanism to bypass the T4CP AAD specificity checkpoint. We discuss our findings in the context of a new model (Fig. 8) depicting the AAD as a key regulatory domain required not only for docking of cognate substrates but also for substrate-driven channel activation.

FIG 8.

Model depicting proposed contributions of the AADs of VirD4 and PcfC to the substrate docking reaction. (Left) Schematic of the VirB/VirD4 T4SS spanning the Gram-negative A. tumefaciens cell envelope. The relaxosome assembles at the oriT-like nic site within the right T-DNA border sequence, and VirD2 relaxase nicks and remains covalently bound to the 5′ end of the T-strand. The spatial adaptors VirC1 and Vbp significantly enhance the efficiency of VirD2–T-strand docking with VirD4 (brown hexamer); adaptor interactions with VirD2 and VirD4 are not yet specified. VirD2 binds the AAD_VirD4 substrate specificity determinant by a mechanism dependent on the relaxase's C-terminal translocation signal. The docked substrate then transfers through the lumen of the VirD4 hexamer or, more likely, to one or more of the VirB4 or VirB11 ATPases for further processing and substrate transfer through a channel formed by the IMC. The promiscuous MOBQ plasmid RSF1010 bypasses the AAD_VirD4 substrate discrimination checkpoint by MobB-mediated spatial colocalization of the RSF1010 relaxosome with the VirB/VirD4 T4SS and one or more unspecified contacts between MobA's internal or C-terminal translocation signals and VirD4. (Right) Schematic of the Prg/Pcf T4SS spanning the Gram-positive E. faecalis cell envelope. The accessory factor PcfF recruits PcfG to the pCF10 oriT sequence, and the two Dtr factors mediate docking of the relaxosome through direct contacts with the AAD of PcfC. PcfG nicks the pCF10 substrate, remains covalently bound to the 5′ end of the transferred strand, and pilots the DNA substrate across the cytoplasmic membrane either via the lumen of the PcfC hexamer or through a translocation channel formed by the IMC.

The T4CP AAD is a substrate specificity checkpoint.

The notion that the AAD participates in DNA substrate binding originated from X-ray crystallography studies showing that the AAD of TrwB_R388 is positioned at the cytoplasmic entrance of the lumen of the NBD hexamer and exhibits structural similarity to a DNA-binding domain of XerD recombinase (18). Mutational studies then supplied evidence for the functional importance of AADs associated with two T4CPs, TraG_RP4 and TrwB_R388. Specifically, insertions of a 31-residue peptide (i31) in the AAD of TraG abrogated RP4 plasmid transfer (56), and substitutions of AAD residues predicted by the X-ray structure to be surface exposed or to line the cytoplasmic entrance of the hexamer lumen similarly abolished TrwB_R388 function (57).

We extended these earlier findings by showing that VirD4_At and PcfC_pCF10 with their AADs deleted are nonfunctional and production of these domains in trans in otherwise WT strains inhibited substrate transfer (Fig. 2). Importantly, negative dominance observed in the merodiploid strains was substrate specific, as evidenced by selective inhibition of pCF10 transfer by AAD_pCF10 in E. faecalis and of T-DNA transfer by AAD_VirD4 in A. tumefaciens. We confirmed the prediction (18) that T4CP AADs bind DNA (Fig. 4); however, binding lacked detectable oriT/T-DNA sequence or strand preference, arguing against the notion that these domains contribute to substrate selection solely through their DNA-binding activities (Fig. 4). Interestingly, an earlier study showed that the TrwB_R388 receptor preferentially binds potential G-quadruplex (G4) DNA sequences and also that G4 DNA stimulates TrwB_R388 ATPase activity to a greater extent than other DNA sequences (58, 59). These findings, coupled with the results of the present studies and additional observations discussed below, support a proposal that AAD binding of DNA is not required for substrate docking per se but nevertheless plays an important role in activation of the T4CP as a prerequisite for subsequent substrate transfer through the T4SS channel.

Both AAD_PcfC and AAD_VirD4 selectively bound the cognate Dtr proteins required for processing of pCF10 and T-DNA (Fig. 5). Structural resolution of the AAD-Dtr protein interface(s) requires further work; however, an accumulating body of evidence points to the importance of VirD2's C-terminal translocation signal for contact with VirD4 in the A. tumefaciens system. We (15, 25, 29) and others (55) have supplied evidence that VirD2's C terminus is dispensable for nicking at T-DNA border sequences but is required for docking of the T-DNA transfer intermediate with native VirD4 and for successful transfer of T-DNA to plants. Furthermore, we recently reported that VirD2-mediated docking of the T-DNA substrate with the VirD4 T4CP corresponds to one of two important intracellular signals required for activation of the VirB/VirD4 channel (15). The second is represented by the ATP hydrolysis activities of the VirD4 and VirB11 ATPases (60). Together, these intracellular signals induce a conformational change in the channel subunit VirB10, which is necessary for substrate passage through the VirB channel to the cell surface (15, 60). Importantly, a VirD2ΔC35 truncation mutant catalyzes formation of a VirD2ΔC35–T-strand complex, but this is a dead-end transfer intermediate, as shown by its failure to induce the VirB10 conformational switch (15). These observations, combined with our present finding that the VirD2ΔC35 mutant fails to bind AAD_VirD4 in vitro (Fig. 5), suggest that VirD2's C-terminal domain is essential not only for VirD2-AAD_VirD4 binding but also for coupling a substrate-binding signal to conformationally driven changes in the VirB/VirD4 T4SS required for channel activation and substrate passage (Fig. 8).

Distinguishing features of the DNA substrate docking reactions in the A. tumefaciens VirB/VirD4 and E. faecalis Pcf/Prg transfer systems.

Although the VirD2-AAD_VirD4 interaction is a proposed early step of T-DNA transfer, the A. tumefaciens VirB/VirD4 system has evolved additional mechanisms to optimize the efficiency of substrate engagement and delivery to plant cells. VirD2, for example, carries a second internal translocation signal between residues 205 and 264 that is important for efficient T-DNA transfer to plants (55). Further studies are needed to determine if or how this signal coordinates with the C-terminal domain to mediate substrate-VirD4 docking. Additionally, two accessory factors that enhance the efficiency of T-DNA transfer have been identified. One, the ParA/MinD-like ATPase VirC1, assembles as a component of the relaxosome and stimulates VirD2 nicking at T-DNA borders (33). VirC1 also spatially couples the VirD2–T-strand transfer intermediate with the VirB/VirD4 channel through a combination of partner interactions with VirD2 and VirD4 and an intrinsic affinity for the membrane at the cell poles (29). The second factor, termed VirD2-binding protein (Vbp), also promotes coupling of the VirD2–T-strand complex and VirD4 through binary contacts (34, 35).

The importance of these accessory factors for efficient T-DNA substrate docking potentially derives from an unusual feature of this “conjugation” system. In other well-characterized conjugation systems, such as those encoded by the E. coli F, RP4, and R388 plasmids, the early-stage processing reactions, including relaxosome assembly and oriT cleavage, occur concomitantly with binding of the relaxosome with the T4CP (4, 61). In striking contrast, in the A. tumefaciens system, processing of the T-DNA substrate is physically uncoupled from the docking and translocation reactions. Indeed, upon induction of the vir regulon encoding this transfer system with plant inducing signals, an individual A. tumefaciens cell can accumulate as many as 50 copies of the processed VirD2–T-strand complex over the ensuing postinduction period (15, 29). In nature, preloading of the T-DNA substrate in induced A. tumefaciens cells might have evolved as a mechanism to enhance the efficiency or amount of T-DNA substrate delivered once bacterial donors are able to form productive contacts with plant target cells. Accordingly, such a system would necessitate an efficient substrate recruitment system, and VirC1 and Vbp might have been coopted for this purpose (29, 34, 35).

At this time, the network of contacts formed between these adaptors and the VirD proteins for substrate-receptor docking is not completely defined. We were unable to detect VirD1 contacts with any of the VirD4 domains (Fig. 5 and 7), suggesting that this accessory factor does not contribute to the docking reaction or, alternatively, is dissociated from the transfer intermediate prior to substrate docking. We also have not been able to detect a direct interaction between VirD2 and VirD4 in A. tumefaciens cell extracts through standard pulldown assays (29, 34), in spite of our ability here to detect VirD2-AAD_VirD4 complex formation in vitro (Fig. 5). The VirC1 and Vbp adaptors might contribute not only to substrate recruitment but also to stabilization of the T-DNA substrate–VirD4 interaction in vivo.

In the E. faecalis pCF10 system, in contrast, several lines of evidence suggest that the relaxosome binds directly to the PcfC receptor for coupled substrate processing and translocation. This system transfers pCF10 to neighboring enterococci at extremely high frequencies in less than 15 to 30 min following exposure of pheromone sensing by donor cells. Indeed, nearly identical kinetics profiles have been reported for accumulation of prg and pcf transcripts and Prg and Pcf proteins as well for plasmid transfer, highly suggestive of spatiotemporal coupling of these reactions (30, 50, 62). Studies also have established that the Dtr proteins PcfF and PcfG both suffice for and impart specificity to the pCF10 processing and receptor-binding reactions (36). We extended these findings, first, by showing that the pCF10 DNA substrate binds the PcfC receptor only in cells competent for assembly of the catalytically competent relaxosome, that is, cells harboring a plasmid with an intact oriT sequence and producing both PcfF and PcfG (30). Furthermore, we showed that PcfF and PcfG bind a purified soluble form of PcfC (30). Finally, here we demonstrated binding of these Dtr proteins with the AAD_PcfC in vitro (Fig. 5). Although we also detected binding of both Dtr factors to the NBD/CTD_pCF10 fragment, this fragment also bound nonspecifically to the VirD2 relaxase (Fig. 7), implicating the AAD and not the NBD or CTD of PcfC in substrate discrimination. Thus, we propose that upon binding of PcfF and PcfG at pCF10's oriT sequence, both Dtr factors form a ternary complex with the AAD_pCF10, thereby physically coupling the pCF10 relaxosome to the PcfC receptor. Further definition of these Dtr-receptor contacts is of special interest in view of the fact that the C termini of PcfF and PcfG lack clusters of positively charged or hydrophobic residues characteristically found in the C-terminal translocation signals of VirD2 and other T4SS substrates (54, 63).

Contributions of other T4CP domains to substrate binding.

As noted above, the NBD/CTD truncation mutants of both VirD4 and PcfC bound DNA nonspecifically, and NBD/CTD_pCF10 also nonselectively bound Dtr proteins (Fig. 7). The biological relevance of these findings remains to be evaluated, but it should be noted that several studies have identified functions of the various T4CP domains that could contribute—directly or indirectly—to substrate binding. The NTDs, for example, are unlikely to directly bind substrates, but studies have identified several NTD functions, including binding interactions with other channel subunits, stimulatory effects on T4CP oligomerization and catalytic activity, and contributions to T4CP spatial positioning in the cell (17, 64–67), that could indirectly affect substrate docking. By virtue of their possible accessibility to secretion substrates in the cytoplasm, the NBDs and CTDs might coordinate with the AAD to recruit substrates through direct substrate engagement. In fact, the best-characterized substrate-T4CP contact to date is that between the CTD of TraD encoded by F-like plasmids and TraM, a component of the F relaxosome. Structural studies have determined that the last 7 amino acids (V₇₁₂EPGDDF₇₁₇) of TraD's CTD fit into a positively charged groove of TraM. Further mutational analyses of this partner interaction confirmed that the TraD CTD-TraM interaction is required for highly efficient F plasmid transfer and, conversely, is strongly inhibitory for F plasmid mobilization of MOBQ plasmids (68, 69). These findings warrant further studies exploring the importance of the CTDs of PcfC_pCF10, VirD4_At, and other T4CPs that carry such domains for possible DNA or effector protein contacts.

MOBQ plasmid mobilization bypasses the AAD specificity checkpoint.

Finally, our finding that the VirD4-AAD_PcfC chimera supported transfer of the MOBQ plasmid pML122 raises the intriguing question of how the MOBQ plasmid transfer intermediate bypasses the proposed AAD substrate specificity checkpoint. We envision two scenarios that are not mutually exclusive. First, MobA relaxases of MOBQ plasmids might carry distinct translocation signals capable of mediating relaxosome docking with heterologous AADs, including that of AAD_PcfC. Indeed, MobA has been shown to carry a negatively charged translocation motif reminiscent of that of VirD2 at its C terminus (54), as well as two large internal translocation signals, each capable of mediating substrate transfer (70). The internal signals are unrelated to each other, with the exception of a shared 7-residue motif (G[E/D]R[L/M]R[V/F]T) that, interestingly, is also present in internal translocation signals recently identified in the TraI_R16 and TrwC_R388 relaxases (71–73). Such signals might conform to a general motif that for MobA enables binding to heterologous AADs.

The second docking scenario derives from earlier work establishing that another MOBQ-encoded factor termed MobB plays an important role in recruitment of the MOBQ relaxosome to different T4SSs. Interestingly, MobB was also shown to be important for the functionality of both of MobA's internal translocation signals (70). MobB has an intrinsic affinity for the cytoplasmic membrane, which led to a proposal that MobB functions to colocalize the MOBQ plasmid relaxosome at the membrane in spatial juxtaposition with T4SSs. In this capacity, MobB would contribute to the stabilization of inherently weak MobA relaxase contacts with heterologous T4CPs (70). In the A. tumefaciens chimeric system, MobB might stabilize weak contacts between MobA_pML122 and the AAD_PcfC or mediate MOBQ relaxosome docking with the chimeric receptor completely independently of an AAD_PcfC contact (Fig. 8). Our finding that the VirD4ΔAAD mutant did not support pML122 transfer (Fig. 2) does not preclude the latter bypass mechanism, since, as noted above, the AAD likely contributes in ways other than binding of cognate Dtr factors to overall T4CP function.