Characterization of ConE, the VirB4 Homolog of the Integrative and Conjugative Element ICEBs1 of Bacillus subtilis

ABSTRACT

Conjugation is a major form of horizontal gene transfer, contributing to bacterial evolution and the acquisition of new traits. During conjugation, a donor cell transfers DNA to a recipient through a specialized DNA translocation channel classified as a type IV secretion system (T4SS). Here, we focused on the T4SS of ICEBs1, an integrative and conjugative element in Bacillus subtilis. ConE, encoded by ICEBs1, is a member of the VirB4 family of ATPases, the most conserved component of T4SSs. ConE is required for conjugation and localizes to the cell membrane, predominantly at the cell poles. In addition to Walker A and B boxes, VirB4 homologs have conserved ATPase motifs C, D, and E. Here, we created alanine substitutions in five conserved residues within or near ATPase motifs in ConE. Mutations in all five residues drastically decreased conjugation frequency but did not affect ConE protein levels or localization, indicating that an intact ATPase domain is critical for DNA transfer. Purified ConE is largely monomeric with some oligomers and lacks enzymatic activity, suggesting that ATP hydrolysis may be regulated or require special solution conditions. Finally, we investigated which ICEBs1 T4SS components interact with ConE using a bacterial two-hybrid assay. ConE interacts with itself, ConB, and ConQ, but these interactions are not required to stabilize ConE protein levels and largely do not depend on conserved residues within the ATPase motifs of ConE. The structure-function characterization of ConE provides more insight into this conserved component shared by all T4SSs.

IMPORTANCE Conjugation is a major form of horizontal gene transfer and involves the transfer of DNA from one bacterium to another through the conjugation machinery. Conjugation contributes to bacterial evolution by disseminating genes involved in antibiotic resistance, metabolism, and virulence. Here, we characterized ConE, a protein component of the conjugation machinery of the conjugative element ICEBs1 of the bacterium Bacillus subtilis. We found that mutations in the conserved ATPase motifs of ConE disrupt mating but do not alter ConE localization, self-interaction, or levels. We also explored which conjugation proteins ConE interacts with and whether these interactions contribute to stabilizing ConE. Our work contributes to the understanding of the conjugative machinery of Gram-positive bacteria.

INTRODUCTION

Conjugation is the transfer of DNA from a donor cell to a recipient cell through a specialized DNA translocation channel. Found in all major clades of bacteria, conjugative DNA elements are self-transmissible and contain the genes required for transfer. Conjugative DNA elements include both conjugative plasmids and integrative and conjugative elements (ICEs; also called conjugative transposons), which reside on the host chromosome and can excise to form a conjugative DNA circle (1–4). While not as well studied as conjugative plasmids, ICEs are the most abundant conjugative elements found in prokaryotes (5). The conjugation machinery encoded by a conjugative DNA element can mobilize resident plasmids as well as transferring entire chromosomes (2, 6, 7). Thus, conjugative elements are fundamental to bacterial evolution, contributing to the spread of genes conferring antibiotic resistance, new metabolic traits, and pathogenic function (8–10).

During conjugation, DNA is secreted across the cell envelope using a specialized DNA translocation channel, classified as a type IV secretion system (T4SS). The T4SSs of Gram-negative bacteria have been very well characterized and are composed of at least 11 conserved proteins (VirB1 to VirB11, using the nomenclature of the well-studied pTi plasmid found in Agrobacterium tumefaciens) (11–17). The coupling protein ATPase (VirD4) delivers the conjugative DNA or secretion substrate to the channel. Less is known about the T4SSs of Gram-positive bacteria, which are composed of a subset of the Gram-negative T4SS components, often with one or more Gram-positive-specific proteins (11, 17–20). The T4SSs found in Gram-positive bacteria generally have homologs of the VirB1 cell wall hydrolase, VirB4 ATPase, and VirD4 coupling protein and analogs of the transmembrane proteins VirB3, VirB6, and VirB8.

The VirB4 ATPase is the largest and most conserved subunit of all T4SS proteins (21–23), playing a role in energizing DNA transfer and/or assembly/stability of the T4SS (11, 12, 24, 25). Several VirB4 homologs have been shown to oligomerize, localize to the membrane, interact with other T4SS components, and bind to DNA (26–35). VirB4 homologs contain conserved Walker A and B boxes that function in ATP binding and hydrolysis, respectively, as well as conserved motif C (DX[D/E]X_1–3E), motif D (RK), and motif E ([S/T]Q) (29). Some VirB4 homologs display ATPase activity in vitro, but activity tends to be highly dependent on solution conditions and/or the oligomerization state of the protein (26, 27, 33–36).

We are interested in studying the T4SS of ICEBs1, a conjugative DNA element found integrated on the chromosome of the Gram-positive bacterium Bacillus subtilis (reviewed in reference 18) (see Fig. S1A in the supplemental material). ICEBs1 belongs to the mating pair formation (MPF) FA clade, found only in monoderms; other members include Tn916 of Enterococcus faecalis and pCW3 of Clostridium perfringens (22). Many of the genes required for regulation, exclusion, chromosomal excision, replication, transfer, and integration have been identified and characterized (37–47). The ICEBs1 T4SS includes the putative ATPase ConE (VirB4 homolog) (48), the cell wall hydrolase CwlT (VirB1 homolog) (49), the coupling protein ConQ (VirD4 homolog) (40), and the cell membrane T4SS components ConB (VirB8-like), ConC, ConD (VirB3-like), and ConG (VirB6-like) (50).

Our work here focused on ConE, the VirB4 homolog of ICEBs1. Previously, we showed that ConE is required for transfer of ICEBs1 and for ICEBs1-mediated mobilization of resident plasmids (48, 50). A fusion of green fluorescent protein (GFP) to ConE localizes to the cell membrane, with a large concentration located at the cell poles (48). ConE lacks predicted transmembrane segments, and the ICEBs1 T4SS transmembrane proteins ConB and ConD are required for localization of ConE to the membrane (50). ConB directly interacts with ConE, and this interaction likely helps recruit and/or maintain ConE at the membrane (50).

ConE requires intact ATPase motifs for function. Previously, we found that the Walker A variant ConE(K476E) and the Walker B variant ConE(D703A/E704A) were unable to support mating, indicating that conserved residues in the ATPase motifs are required for function (48). ConE(K476E)-GFP, predicted to be defective in nucleotide binding, localizes to the membrane at the cell poles, similarly to the wild-type ConE-GFP. Neither variant could be purified due to expression and/or solubility problems, so it is unknown how these mutations affect ConE activity in vitro (data not shown).

Here, we investigated the effects of alanine substitutions in five additional conserved residues found in or near the ATPase motifs in ConE on conjugation. Our rationale for characterizing additional residues in the ATPase domain was 2-fold. First, while motifs C, D, and E were defined more than a decade ago (29), little research has been done to explore their importance. Second, we created additional ATPase variants of ConE, because we wished to characterize them in vitro, which we were unable to do with our prior two variants. Here, we found that all five mutations disrupt mating but do not affect ConE localization or levels, suggesting that the ATPase activity of ConE might be required for mating. We purified and characterized wild-type ConE and two variants. We also present evidence that ConE oligomerizes and interacts with ConQ as well as ConB and explore what role the ATPase motifs play in these interactions. Finally, we explore whether the presence or absence of other T4SS proteins affects total ConE levels. Altogether, our results provide new insights into a VirB4 homolog found in a Gram-positive bacterium.

RESULTS

Identification of conserved residues in ATPase motifs of ConE.

We aligned ConE with the sequences of five different VirB4 homologs found in Gram-positive bacteria to identify conserved residues in its ATPase domain (Fig. 1). Upon alignment, the Walker A and B boxes (34), along with the previously defined VirB4-specific motifs C, D, and E, were evident (Fig. 1A to C) (29). Motifs C and E lie in the catalytic pocket (29). Motif D (RK) contains a conserved arginine that likely functions as an arginine finger (51). Arginine fingers typically are required for ATPase function but are located far from the main catalytic residues, acting in trans within an oligomeric protein. Here, we focused on five conserved residues within or near the ATPase motifs of ConE (21, 29): D498 and E502 in motif C, W706 located immediately after the Walker B box, the putative arginine finger R726 in motif D, and Q737 in motif E (Fig. 1B and C). We chose W706 specifically, as two other Walker B box variants could not be purified due to expression and/or solubility problems (data not shown).

FIG 1.

Conserved residues in ATPase motifs of ConE and its homologs. The Walker A box (red), Walker B box (purple), and motifs C, D, and E (black) are indicated (29). ConE residues D498 (pink), E502 (orange), W706 (teal), arginine finger R726 (R-finger; blue), and Q737 (green) are also indicated. (A) Schematic of ConE protein (831 residues long) with conserved motifs indicated. (B and C) Alignment of ConE (GenBank accession no. BAA19331.1) with homologs OrfD of ICESt1 in Streptococcus thermophilus (accession no. CAC67549.1), ORF SAPIG0077 of ICESa1 in livestock-associated methicillin-resistant Staphylococcus aureus subsp. aureus ST398 (abbreviated as 0077) (accession no. CAQ48515.1), Orf16 of Tn916 in Enterococcus faecalis (accession no. AAB60017.1), TcpF of pCW3 in Clostridium perfringens (accession no. ABF47331.1), and TraE/Orf5 of the broad-host-range streptococcal conjugative plasmid pIP501 (accession no. AAA99470.1). Alignment was performed by Clustal Omega using the default settings (90). An asterisk indicates a position that has a single, fully conserved residue. A colon denotes a position in which there is conservation between residues with highly similar properties. A period denotes a position in which there is conservation between residues with weakly similar properties. All numbering refers to residue numbers of the ConE sequence. Note that there is a gap in sequence between panels B and C. (D and E) Predicted structure of ConE rendered as space-filling models from the AlphaFold Monomer v2.0 database (UniProt no. P96642; structure prediction updated 9 December 2021). Residues found in the catalytic site (D) and the opposite face of ConE (E) are shown. The vast majority of the 831-amino-acid protein was predicted by AlphaFold with confidence or very high confidence (>70 pLDDT, an estimate of the confidence for each residue). Three short regions (residues 1 and 2, 125 to 157, and 741 to 749) predicted with lower confidence (50 < pLDDT < 70) are omitted from the model.

In the predicted three-dimensional (3D) structure of ConE in the Alphafold Monomer v2.0 database (52, 53), the Walker boxes, motif C, and motif E cluster together on one face of the subunit (Fig. 1D) as expected, forming the catalytic pocket. In contrast, R726 in motif D is located on the opposite face of the subunit (Fig. 1E), as would be predicted for an arginine finger. While the model of ConE is monomeric, ConE may be dimeric or hexameric, similar to some characterized homologs (27, 28, 33, 54–57). In crystal structures of hexameric homologs, the arginine residue aligning with R726 of ConE lies in close proximity to the catalytic pocket of the adjacent subunit, consistent with R726 being the putative arginine finger of ConE.

Conserved residues in ConE ATPase motifs are critical for conjugation.

To determine their role in conjugation, we created alanine substitution mutations in the five conserved residues (D498A, E502A, W706A, R726A, and Q737A). The mutant conE alleles were placed unmarked on the B. subtilis chromosome within ICEBs1, replacing the wild-type conE gene. Donor strains containing either wild-type conE or the missense mutations were mated with recipient B. subtilis cells lacking the conjugative element. We found that a donor strain with wild-type conE transferred ICEBs1 with an average mating frequency of 5% (transconjugant CFU per donor CFU × 100) (Fig. 2, row A), as reported previously (37). In contrast, the W706A and motif E (Q737A) mutations had average mating frequencies of ~0.03% and ~0.04%, both more than 100-fold down from wild-type values (Fig. 2, compare rows F and J to row A). Few or no transconjugants were observed for the motif C (D498A and E502A) and motif D (R726A) mutants, with percent mating less than 1 × 10⁻⁴%, more than 70,000-fold down from wild-type values (Fig. 2, compare rows B, D, and H to row A). For all variants, mating was restored to wild-type levels by introduction of a copy of wild-type conE at an ectopic locus on the chromosome (thrC) (Fig. 2, rows C, E, G, I, and K). This indicates that the defects in mating were due to loss of ConE function and not due to other polar or downstream effects. Furthermore, quantitative Western blots showed that the levels of the ConE variants were at most 2.5-fold lower than the levels of wild-type ConE (Fig. S2), much less than the several-log-fold defects in mating (Fig. 2). Therefore, the defects in mating observed were due to the mutations likely causing significant alterations in ConE activity, rather than total protein levels.

FIG 2.

Conserved residues in or near ATPase motifs of ConE are critical for mating. Percent mating is calculated as the number of transconjugant CFU per donor CFU × 100. Data are averages from at least three experiments, and error bars indicate one standard deviation. Asterisks indicate that few or no transconjugants were observed, with percent mating for these strains being <1 × 10⁻⁴%. The allele of conE (wild-type or containing a missense mutation) at ICEBs1 is indicated in the first column. The truncated ICEBs1 derivatives integrated at thrC used for complementation of the mutations are indicated in the second column. In rows B to K, strains either have a truncated ICEBs1 containing a wild-type copy of conE at thrC (thrC::conE⁺; shown in Fig. S1B) or have a truncated ICEBs1 lacking conE at thrC (thrC::ΔconE; shown in Fig. S1C). Donor strains used were JMA168 (A), MMB1757 (B), MMB1746 (C), MMB1758 (D), MMB1747 (E), MMB1759 (F), MMB1748 (G), MMB1768 (H), MMB1767 (I), MMB1770 (J), and MMB1769 (K).

Conserved residues in ConE ATPase motifs are not required for localization.

To analyze the role of the five conserved residues on the subcellular localization of ConE, we constructed strains of B. subtilis that express fusions of the ConE variants to GFP, expressed from the presumed native promoter at a heterologous site. Previously, we reported that ConE-GFP partially complements a Walker A mutant of ConE (K476E) in trans, restoring mating ~250-fold (48). The localization of the ConE-GFP variants were observed in live cells using fluorescence microscopy. As seen previously (48), ConE-GFP localizes to the cell membrane, with a large concentration of the protein located at the cell poles (Fig. 3A). All five ConE variants (D498A, E502A, W706A, R726A, and Q737A) localized similarly to the wild type (Fig. 3B to F). Thus, mutations in the conserved ATPase motifs have a large detrimental effect on mating but do not significantly impact ConE levels or localization.

FIG 3.

Mutations in conserved residues in ConE ATPase motifs do not disrupt localization to the membrane. (A to F) The wild type or the ConE-GFP variant (shown in green) was visualized in live cells by fluorescence microscopy. Wild-type ConE-GFP localizes to the membrane, predominantly at the cell poles, in strain MMB968 (A), as do the variants ConE(D498A)-GFP in strain MMB1726 (B), ConE(E502A)-GFP in strain MMB1724 (C), ConE(W706A)-GFP in strain MMB1725 (D), ConE(R726A)-GFP in strain MMB1619 (E), and ConE(Q737A)-GFP in strain MMB1658 (F). (G) ConE-SNAP (shown in yellow) was visualized in live cells of strain MMB1678 by fluorescence microscopy.

Three observations indicate that aggregation of GFP does not drive this localization pattern. First, we previously showed that a ConE deletion variant (Δ88–808) fused to GFP mislocalizes throughout the cytoplasm (48). Second, we previously showed that ConE-GFP localization depends specifically on the ICEBs1 transmembrane protein ConB; ConE-GFP mislocalizes to the cytoplasm in a strain with conB deleted (48). Finally, we replaced conE at its native locus with a fusion of conE to SNAP, which encodes a small (20-kDa) monomeric self-labeling protein tag. Previously, a SNAP tag was reported to be the least intrusive tag for protein localization compared to a variety of tags commonly used in fluorescence microscopy (58). We found that ConE-SNAP localizes similarly to ConE-GFP (Fig. 3G) and is partially functional to support mating (>500-fold higher mating than a conE-null mutant). We conclude that ConE localizes to the membrane, predominantly at the cell poles.

Purified ConE is enzymatically inactive and largely monomeric.

To biochemically characterize the effects of the mutations, we purified wild-type and variant histidine-tagged ConE. Wild-type ConE, along with R726A and W706A mutants, eluted with decent yield, purity, and a size consistent with their calculated molecular weights (97 kDa) (Fig. 4A; also data not shown) (59). Notably, the proteins purified from the soluble fraction in the absence of detergent, consistent with ConE being a peripheral membrane protein (48, 50). Despite exploring a variety of protein expression conditions, the other three variants (D498A, E502A, and Q737A) could not be purified due to poor expression and/or poor solubility (data not shown). Using a coupled-enzyme ATPase assay, we did not detect any significant ATPase activity (<1 ATP min⁻¹ monomer⁻¹) of wild-type ConE, above that of the R726A and W706A variants, under a large variety of conditions (see Materials and Methods). With notable exceptions (29, 35, 60), the ATPase activity of VirB4 proteins has been historically difficult to detect and/or unusually sensitive to solution conditions (26, 27, 29, 33, 36, 61).

FIG 4.

Purified His6-ConE and His6-ConE(R726A) are mixtures of oligomers in vitro. Gels were stained using Coomassie blue G-250. Protein molecular mass standards (M) are indicated. (A) SDS-PAGE analysis of purified His6-ConE and His6-ConE(R726A). (B) BN-PAGE analysis of His6-ConE and His6-ConE(R726A), with the quantities of protein loaded indicated. (C) BN-PAGE analysis of His6-ConE and His6-ConE(R726A) (20 μg), with various molecules added. Lanes oligo, single-stranded DNA oligonucleotide; lanes M13, M13 phage circular single-stranded DNA; lanes dsDNA, plasmid circular supercoiled DNA.

The oligomeric state of purified VirB4 homologs varies from monomers to mixtures of oligomers and is often highly dependent on the solution conditions (27–29, 33, 55, 56). We used blue native polyacrylamide gel electrophoresis (BN-PAGE) to assess the oligomerization state of purified his-tagged ConE and the R726A variant (Fig. 4B). While monomeric ConE is 97 kDa, we found that the majority of purified ConE runs with an apparent molecular weight of ~165 ± 5 kDa, which likely corresponds to an irregularly shaped monomer. The second most prominent band is likely dimeric or trimeric ConE, with an apparent molecular weight of 313 ± 35 kDa. A small fraction of the protein ran as larger oligomers. The R726A protein variant has an oligomerization profile similar to that of the wild-type protein (Fig. 4B), indicating that the arginine finger residue is not essential for self-interaction under these conditions. The addition of ATP, single-stranded DNA, or double-stranded DNA did not significantly alter the oligomerization states of either His6-ConE or His6-ConE(R726A) (Fig. 4C). Similar results were obtained via analysis of purified ConE using analytical gel filtration (see Materials and Methods) (Fig. S3). Altogether, our data indicate that purified ConE is largely monomeric in solution, with some evidence of oligomerization.

ConE interacts with itself, ConB, and ConQ in vivo.

The ICEBs1 T4SS is composed of several proteins (ConB, ConC, ConD, ConE, ConG, and CwlT), along with the coupling protein ConQ, but it is unclear how these proteins interact within the DNA translocation channel. To evaluate which of these proteins interact with ConE in vivo, we used a bacterial two-hybrid (BACTH) assay based on the interaction between the T18 and T25 domains of the adenylate cyclase enzyme (62). As we previously reported (50), we observed that ConE interacts with the bitopic membrane protein ConB (Fig. 5A), as indicated by blue colonies on indicator plates. Furthermore, we found that ConE interacts with itself, consistent with our observation that ConE forms some oligomers in vitro (Fig. 4B). Finally, we found evidence that ConE may interact with the coupling protein ConQ, although this interaction appeared weaker (lighter blue) than the interaction of ConE with itself or ConB (Fig. 5A). In contrast, no interaction was observed between ConE and ConC, ConD, ConG, or CwlT, as well as the negative controls, in which one fusion partner was missing (Fig. 5A; Fig. S4). We conclude that ConE interacts with itself, ConB, and ConQ in vivo.

FIG 5.

ConE interactions with other ICEBs1 proteins detected by BACTH assays. BACTH assays were carried out with E. coli transformed with plasmids containing the T18 and T25 domains of adenylate cyclase (91, 92) alone or fused to ICEBs1 genes, as indicated. Interacting partners reconstitute adenylate cyclase by bringing together the T18 and T25 domains. Strains were plated onto LB agar plates containing X-Gal, ampicillin, kanamycin, and IPTG. (A) ConE interacts with itself, ConB, and ConQ in vivo. Plates were photographed 24 h after placement at room temperature. (B) Summary of effects of mutations in the ATPase motifs of ConE on interaction with itself, ConB, and ConQ in vivo. Plates were inspected 24 h after placement at room temperature for color development. ++, medium blue; +, light blue; −, beige similar to negative controls. Photos of each interaction as well as all negative controls are shown in Fig. S4 to S7.

Using the BACTH assay, we examined whether conserved residues in the ATPase motifs of ConE were required for interaction with itself, ConB, and ConQ. We were particularly interested in how these mutations might affect self-interaction, as the variants are likely on the surfaces that mediate oligomerization. We found that the five alanine mutations (D498A, E502A, W706A, R726A, and Q737A) in ConE did not disrupt self-interaction or interaction with ConB (Fig. 5B; Fig. S5 and S6). This result is consistent with our in vitro result showing that the arginine finger variant had an oligomerization profile similar to that of the wild-type ConE. In contrast, we found that the ConE D498A and W706A variants appeared defective for interaction with ConQ in this qualitative assay (Fig. 5B; Fig. S7). We conclude that mutations in conserved residues in the ATPase motifs do not disrupt ConE self-interaction or interaction with ConB, but some mutations disrupt interaction with ConQ.

ConE protein levels in strains lacking various T4SS components.

Previously, it has been shown that the levels of some T4SS proteins depend on the presence of other T4SS proteins in the cell, possibly due to protein-protein interactions stabilizing a protein and/or protecting it from degradation (63–66). We were interested in determining whether ConE levels were altered in strains missing either of its two interaction partners, ConB and ConQ, or other T4SS components. To test this, we grew cells that contained deletions of various T4SS genes and then measured ConE protein levels in B. subtilis cells using quantitative Western blots. We found that ConE levels were approximately the same in the absence of ConB, ConQ, or ConG (Fig. 6). In contrast, ConE levels were reduced approximately 2- and 5-fold in the absence of ConC and ConD, respectively (P < 0.01) (Fig. 6). We conclude that the stability and/or expression of ConE relies on the presence of ConC and ConD. The finding with ConD is consistent with our prior complementation studies that showed that conE is not translated efficiently in the absence of conD (48). The two genes may be translationally coupled, as the coding region of the end of conD overlaps the coding of the first 37 codons of conE.

FIG 6.

Cellular ConE protein levels were decreased 2-fold and 5-fold in strains with deletions of conC and conD, respectively. Total protein extract from cells (~5 μg loaded) was analyzed by quantitative Western blot probing with rabbit polyclonal anti-ConE antibodies. The averages and standard deviations are reported for the wild-type (JMA168), ΔconB (MMB1275), ΔconC (MMB1271), ΔconD (MMB1274), ΔconE (MMB951), ΔconG (MMB1283), ΔconQ (CAL848), and ΔcwlT (TD20) strains. An asterisk denotes statistical significance (P < 0.01) compared to the wild-type value, as determined by a two-tailed heteroscedastic t test. At least three independent biological replicates were used to obtain averages.

DISCUSSION

While VirB4 proteins are the signature subunits of all T4SSs, the individual family members differ in behavior in various important ways, including enzymatic properties and oligomerization state (11, 19, 21–23). Here, we contribute to our understanding of the VirB4 family of ATPases found in Gram-positive bacteria. While residues in motifs C, D, and E have been described as conserved and located close to or in the catalytic pocket (29), few studies have analyzed whether they are functionally important. We found that five residues within and around conserved ATPase motifs in ConE are critical for mating of B. subtilis (Fig. 2). The residues are located in or near the active site (Fig. 1D and E), but their alteration does not dramatically affect steady-state ConE levels (Fig. S2) or localization (Fig. 3). Given their location and conservation in VirB4 ATPases, we hypothesize that these residues may be critical for mating because they contribute to ATP binding and/or hydrolysis. Arginine fingers have been shown to be critical for conjugation and/or ATPase activity for several homologs, including VirB4 from Thermoanaerobacter pseudethanolicus (29), TrbE of RP4 (34), HerA in archaea (57), and the structurally similar VirD4-like TrwB coupling protein (67). Similar to our results (Fig. 2), mutation of TrbE’s motif C completely abolished mating, while mutation of the Q residue in motif E reduced mating moderately (34).

Mutation of conserved residues in the ATPase motifs of ConE disrupted mating but did not disrupt localization to the membrane (Fig. 3). Thus, proper localization of ConE at the membrane is not sufficient to support its function in mating. Since these residues are hypothesized to be critical for ATP binding and hydrolysis, we hypothesize that ATPase activity may not be essential for its subcellular localization. Similarly, ATPase mutants of VirB4 of A. tumefaciens and PrgJ of pCF10 are able to correctly partition to the membrane (27, 68). In contrast, the localization of another VirB4 homolog, TrhC of conjugative plasmid R27, is dependent on intact ATPase motifs for proper subcellular localization (69).

Purified ConE did not exhibit ATPase activity in vitro under the many conditions explored. With some exceptions (29, 35, 60), the ATPase activity of VirB4 proteins has been historically difficult to detect and/or unusually sensitive to solution conditions (27, 29, 33, 36, 61). The VirB4 homolog TrwK of plasmid R388 was initially reported as inactive in vitro (34). Later, TrwK was shown to hydrolyze ATP in the presence of acetate ions and absence of sodium chloride, solution conditions that also promoted hexamerization (33). Additionally, the VirB4 homolog TraB of plasmid pKM101 is dimeric and enzymatically inactive if purified from membrane fractions but is hexameric and active if purified from soluble cytoplasmic fractions (26). PrgJ of pCF10 is enzymatically active and dimeric but is inhibited by glycerol or HEPES buffer (27). VirB4 of CTn4 of Clostridioides difficile is also sensitive to solution conditions; it has much lower ATPase activity in NaCl than in KCl (36). The ATPase activity of ConE may require special solution conditions or interaction with another T4SS component(s) or membrane and/or may be regulated. One interesting model is that ConE enzymatic activity is autoinhibited to prevent futile ATP hydrolysis in the absence of conjugation (70–73). As a precedent, the VirB4 homolog TrwK possesses an autoinhibitory domain that inhibits its ATPase activity in vitro (74).

Purified VirB4 proteins vary in oligomerization state from monomers to dimers to mixtures of oligomers, including hexamers (27, 29, 33, 36, 75, 76). Structural analysis of various assembled T4SSs favors the idea that VirB4 proteins associate with the conjugation machinery as a hexamer of dimers (24, 25, 77–79), or possibly as two side-by-side hexamers (56, 80). Similar to characterized homologs, ConE interacts with itself in vivo (Fig. 5). Purified ConE is primarily monomeric, with traces of dimers and higher-order oligomers (Fig. 4). Although it is possible that the arginine finger plays some role in self-interaction, we did not detect any changes in the ability of the arginine finger variant to oligomerize in vivo or in vitro (Fig. 4B and 5B; Fig. S6). Similarly, mutation of the arginine finger in the coupling protein TrwB decreased DNA transfer but did not cause a defect in self-interaction (67). We were able to complement the mating defects of the individual variants by addition of the wild-type conE gene in trans at an ectopic locus (Fig. 2). Since no dominant negative effects were observed, ConE oligomers may be functional in vivo even when composed of a mixture of functional and nonfunctional subunits.

We used BACTH assays to explore which T4SS proteins ConE interacts with. We confirm that ConE interacts with ConB, as shown previously (50). This interaction is functionally important, as ConE lacks transmembrane segments, and ConB is required for ConE to localize at the membrane. We also discovered an interaction between ConE and ConQ. Both of these ATPases have been seen in close proximity near the inner membrane complex in assembled T4SS complexes (25, 80). Although we observed a negative result using BACTH assays, we suspect that ConE also directly interacts with ConD (VirB3), since ConE displays some localization defects in the absence of ConD (50). Furthermore, homologs of each have been shown to interact in a number of T4SSs (31, 79), and their genes are fused in some conjugative elements (64).

Overall, our analysis has contributed to the characterization of a VirB4 protein and its conserved motifs. Future genetic and biochemical analyses should provide more insight into the mechanism by which VirB4-like proteins interact with the conjugation machinery and how ATP hydrolysis is coupled to T4SS assembly and/or secretion.

MATERIALS AND METHODS

Media and growth conditions.

For B. subtilis and Escherichia coli cells, routine growth and strain construction were done in LB medium at 37°C. Antibiotics were used at standard conditions (81). Strains containing ICEBs1 were grown in the presence of kanamycin (5 μg/mL) to maintain selection of ICEBs1. Growth conditions for particular experiments are described below.

Strains, alleles, and plasmids.

E. coli strains used for routine cloning were AG1111 (MC1061 F′ lacI^q lacZM15 Tn10) and DH5α. B. subtilis strains used in experiments are derived from JH642 (82) and their relevant genotypes are listed in Table S1. B. subtilis strains were constructed via natural transformation (81). Strains cured of ICEBs1 (ICEBs1⁰), the spontaneous streptomycin resistance allele (str-84), the amyE::[Pspank(hy)-rapI spc] allele, and Δ(rapI phrI)342::kan were described previously (37). The lacA::[Pxis-(conD conE-mgfpmut2) tet] allele expressing conD followed by conE fused to a 23-amino-acid linker and monomeric gfpmut2 from the presumed native promoter Pxis at the ectopic locus lacA was described previously (48). The ΔcwlT19 allele was previously described (49). All newly cloned constructs were confirmed by sequencing.

(i) Construction of unmarked mutations in conE on the chromosome. Plasmids containing missense mutations of conE were constructed using site-directed mutagenesis. The parental wild-type plasmid pMMB1566 was constructed by ligating the PCR-amplified conE-yddF-conG region from ICEBs1 into the EcoRI/BamHI-digested vector pEX44 that encodes chloramphenicol resistance (83). The plasmid pMMB1673 contains the conE-SNAP allele and provides chloramphenicol resistance. It was produced by isothermal assembly (83) of the vector pEX44 cut with EcoRI/BamHI, a PCR-generated DNA fragment containing the conE gene lacking its stop codon, and a PCR-generated DNA fragment containing a five-amino-acid linker (GSGSG) followed by SNAP. Using a previously described strategy (38), the wild-type conE gene within ICEBs1 was replaced on the chromosome with each unmarked mutant conE allele or conE-SNAP. All other upstream and downstream genes in ICEBs1 were kept intact.

(ii) Construction of ICEBs1 truncation complementation alleles. Previous research has shown that full complementation of mutations in several genes in ICEBs1, including conE, cannot be achieved by addition of the single gene at an ectopic locus (48, 50). In most cases, full complementation was achieved by the addition of a second truncated copy of ICEBs1 containing the missing gene at the ectopic locus. The truncated ICEBs1 constructs for complementation of the conE alleles were constructed as described previously (40). The truncated ICEBs1 derivative integrated at thrC, thrC1699::[ICEBs1 (ΔyddF-attR::cat) mls], contains conE and all the upstream ICEBs1 genes but is missing all of the ICEBs1 genes downstream of conE (Fig. S1B). The construct thrC1753::[ICEBs1 (ΔconE-attR::cat) mls] is essentially the ICEBs1 insertion at thrC but is also missing conE (Fig. S1C). These alleles were constructed by transforming CAL1496 with products of splicing by overlap extension (SOE) PCR (84) that deleted the desired region of ICEBs1, replacing the tetracycline resistance gene with a chloramphenicol resistance gene.

(iii) Construction of conE-gfp mutations on the chromosome. Mutations of conE-gfp were constructed via site-directed mutagenesis of the parental plasmid, pMMB786, which can be integrated at the ectopic locus lacA (48). pMMB786 expresses conD followed by conE fused to monomeric gfpmut2 (mgfp-mut2) from their presumed native promoter, Pxis (83, 84). The plasmids were transformed into B. subtilis strain JMA222, selecting for tetracycline resistance. Double crossovers of the conE-gfp fusions into lacA were verified by PCR. Chromosomal DNA containing the conE-gfp fusions was transformed into either JMA168 (conE⁺) or MMB951 (ΔconE).

(iv) Construction of his6-conE and his10-conE wild-type and mutant expression plasmids. pMMB1242, a plasmid that overexpresses His6-ConE(K476E), was constructed by ligating a PCR product containing conE(K476E) in-frame with the coding sequence for an N-terminal hexahistidine tag into pBAD24 (85) downstream from the pBAD promoter. pMMB1313, encoding wild-type His6-ConE, was constructed by site-directed mutagenesis of pMMB1242. pMMB1503, encoding His6-ConE(R726A), was constructed by site-directed mutagenesis of pMMB1313. pMMB1682, a plasmid that overexpresses His10-ConE, was constructed by ligating a PCR product containing conE in frame with the coding sequence for an N-terminal 10-histidine tag into pBAD24. pMMB1700, encoding His10-ConE(W706A), was constructed by site-directed mutagenesis of pMMB1682. All His-tagged versions of conE include the 831-amino-acid-encoding open reading frame (ORF) for conE (previously named yddE), as annotated in the original B. subtilis genome sequencing (86).

(v) Construction of BACTH protein fusion plasmids. Table S2 lists the BACTH plasmids constructed and used. ICEBs1 genes were cloned in-frame into vectors carrying either the T25 or T18 domains of the Bordetella pertussis cyaA gene for adenylate cyclase as previously described (62). The T18-based plasmids confer ampicillin resistance, while the T25-based plasmids confer kanamycin resistance. Based on their predicted membrane topologies (18), T18 was fused to the N termini of ConB, ConC, ConG, ConQ, and CwlT to ensure that the T18 domain was intracellular. Mutations in conserved ATPase motifs were introduced via site-directed mutagenesis.

Mating assays.

Mating assays were performed as previously described (50) with two minor changes: (i) donor and recipient cells were grown in LB medium, and (ii) cells were allowed to mate for only 2 h, not 3 h (37, 48). To allow selection and monitoring of ICEBs1 acquisition, donor cells contained ICEBs1 marked with a kanamycin resistance cassette. Recipient cells (CAL419) lacked ICEBs1 (ICEBs1°), were comK null to prevent acquisition of DNA via transformation, and possessed streptomycin resistance (str-84) (81). Percent mating was calculated as the number of transconjugant CFU per donor CFU multiplied by 100. The averages and standard deviations of at least three independent experiments are reported. The fold decrease relative to the wild type was calculated by dividing the percent mating of the wild type by the percent mating of the mutant (or in some cases, the upper limit of percent mating of the mutant when no mating was observed).

Western blot analysis.

To measure the protein levels of the five ConE variants, B. subtilis cells were grown and induced as described for the mating assays. Cell pellets were resuspended in SDS sample buffer (50 mM Tris-HCl [pH 6.8], 10% glycerol, 2% SDS, 0.02% bromophenol blue, 1% beta-mercaptoethanol), adjusting the volume of buffer to the optical density at 600 nm. Samples were sonicated for 10 s, followed by incubation at 99°C for 5 min. Proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) on 4 to 20% gradient polyacrylamide gels (Bio-Rad). A series of 2-fold dilutions of each sample were run with each replicate to verify that the signals were in the linear quantitative range. Lysate from a ΔconE (MMB951) strain (48) was used to verify the specificity of the antibody. Proteins were then wet-transferred to a polyvinylidene difluoride (PVDF) membrane. Membranes were blocked with bovine serum albumin (BSA) blocking buffer (5% BSA in Tris-buffered saline with 0.1% Tween 20 [TBST]), probed with anti-ConE rabbit polyclonal antisera (generated by Covance using purified His10-ConE protein) diluted 1:5,000 in BSA blocking buffer, and incubated with anti-rabbit horseradish peroxidase (HRP)-conjugated secondary antibody diluted 1:2,500 in BSA blocking buffer. Chemiluminescent signals were detected using Pierce enhanced chemiluminescence (ECL) Western blot substrate in a GeneGnome imager. Band intensity was measured using ImageJ. At least four independent biological replicates were used to obtain averages.

To measure the protein levels of ConE within strains carrying gene deletions for various ICEBs1 proteins, a similar procedure was followed except that a 7.5% polyacrylamide gel and nitrocellulose membrane (0.2 μm) were used. Membranes were reversibly stained for total protein using Revert total protein stain (LI-COR), blocked in milk blocking buffer (5% nonfat dried milk in Tris-buffered saline), probed with anti-ConE rabbit polyclonal antisera (Covance) diluted 1:5,000 in milk blocking buffer with 0.1% Tween 20, and incubated with anti-rabbit IRDye 800CW goat secondary antibody (LI-COR) diluted 1:20,000 in milk blocking buffer with 0.1% Tween 20. Infrared signals were detected using a LI-COR Odyssey Fc imaging system, and band intensity was measured using Image Studio software. Signal intensity was normalized to total protein. At least three independent biological replicates were used to obtain averages.

Live-cell fluorescence microscopy.

Cells were grown and induced in S7 defined minimal medium (87) and prepared for visualization of ConE-GFP as described previously (50, 87). Cells were visualized at room temperature using a Nikon H550L microscope equipped with a 100× Plan Fluor phase-contrast objective, a high-resolution monochrome cooled charge-coupled device (CCD) Andor digital camera, and Chroma filter set 96362 (for GFP). NIS Nikon Elements 4.0 software was used to process images. Each strain was examined in at least three independent experiments. Nearly the same procedure was used for visualization of ConE-SNAP except that cells were stained with tetramethylrhodamine TMR-Star dye (0.6 μM) during the last 30 min of induction with IPTG (isopropyl-β-d-thiogalactopyranoside) and washed 10 times with 10 volumes of phosphate-buffered saline prior to immobilization on agarose pads. ConE-SNAP was visualized using a Chroma filter set for TMR-Star dye with excitation filter ET545/25x (545 ± 30 nm) and emission filter ET605/70m (650 ± 30 nm).

Protein expression and purification.

For BN-PAGE and some ATPase assays, His6-ConE and His6-ConE (R726A) were purified from E. coli DH5α-derived strains MMB1313 and MMB1503, respectively. Cultures were grown in LB containing ampicillin (100 μg/mL) to an optical density at 600 nm (OD₆₀₀) of ~0.6 before induction with arabinose (0.4% [wt/vol]) for 4 h at 37°C. Cell pellets were lysed on ice for 30 min in lysis buffer (30 mM Tris-HCl [pH 8.0], 250 mM NaCl, 5 mM β-mercaptoethanol, 0.5 mM phenylmethylsulfonyl fluoride [PMSF]) containing 0.5 mg/mL lysozyme. The lysate was sonicated with a Misonix ultrasonic liquid processor XL-2000 series sonicator and cleared by centrifugation at 6,000 × g for 10 min. The soluble portion was added to Ni²⁺ resin (Sigma His-Select nickel affinity gel) and placed on ice for 15 min. The resin was washed with wash buffer 3I (lysis buffer containing 3 mM imidazole) and wash buffer 5I (lysis buffer containing 5 mM imidazole) twice each. The protein was eluted with elution buffer (30 mM Tris-HCl, 250 mM NaCl, 500 mM imidazole, 5 mM β-mercaptoethanol). The eluted protein was dialyzed at 4°C twice against dialysis buffer (20 mM Tris-HCl [pH 8], 150 mM potassium acetate, 2 mM magnesium acetate, 0.1 mM EDTA, 1 mM dithiothreitol [DTT], 20% glycerol). Glycerol was added to the purified protein to 50% prior to storage at −20°C.

For analytical gel filtration and some ATPase assays, His6-ConE and His6-ConE (R726A) were purified from E. coli NEB Express strains MMB1667 and MMB1670, respectively. His10-ConE and His10-ConE(W706A) were purified from NEB Express strains MMB1682 and MMB1700, respectively. Cultures were grown in LB containing ampicillin (100 μg/mL) to an OD₆₀₀ of ~0.6 before induction with arabinose (0.4% [wt/vol]) for 4 h at 37°C. Cell pellets were resuspended in lysis/wash buffer (50 mM potassium phosphate [pH 8.0], 500 mM NaCl, 30 mM imidazole, 5 mM β-mercaptoethanol, 0.5 mM PMSF) and cells were lysed with an LM20 Microfluidizer (Siemens). The lysate was treated with Turbo nuclease (4 U/mL). Soluble fractions were collected after centrifugation at 14,000 rpm for 10 min and incubated with Ni Sepharose 6 Fast Flow (GE Healthcare) for 30 min at 4°C. After three washes with lysis/wash buffer, the proteins were eluted with elution buffer K (10 mM Tris [pH 8.0], 75 mM potassium acetate, 10 mM magnesium acetate, 250 mM imidazole [pH 8.0], 1 mM DTT) or elution buffer N (10 mM Tris [pH 8.0], 150 mM NaCl, 250 mM imidazole [pH 8.0], 1 mM DTT). The eluates were dialyzed against dialysis buffer K (10 mM Tris [pH 8.0], 75 mM potassium acetate, 10 mM magnesium acetate, 1 mM DTT, 10% glycerol) or dialysis buffer N (10 mM Tris [pH 8.0], 150 mM NaCl, 1 mM DTT, 10% glycerol). Some protein at this stage was concentrated using a Vivaspin 6 concentrator (Sartorius; 30,000 molecular weight cutoff [MWCO]). In some cases, glycerol was added to the purified protein to 50%.

Coupled enzyme ATPase assays.

ATPase activity was measured using a coupled enzyme ATPase assay (88). Purified His-tagged ConE, wild-type or mutant (varied from 0.2 to 10 μM, in monomer equivalent), was analyzed at room temperature in dialysis buffer N with 2 or 10 mM magnesium chloride or in dialysis buffer K with 10 mM magnesium acetate. The reactions also contained 2 mM ATP, 0.6 mM NADH, 7.5 mM phosphoenolpyruvate, and 20 U/mL of each pyruvate kinase and lactate dehydrogenase. The decrease in absorbance at 340 nm was monitored. A large variety of solution conditions were tried, including conditions that were previously shown to promote the ATPase activity for other VirB4 homologs (26, 27, 33, 60), such as addition of single-stranded or double-stranded linear DNA, supercoiled plasmid DNA, more or no glycerol, and higher concentrations of ATP. Additionally, some reactions were performed at 37°C or with ConE purified on the same day or with gel filtration fractions enriched for either oligomeric or monomeric ConE. Control experiments showed that addition of the His tag did not alter the ability of ConE to support mating in vivo (data not shown).

Analysis of oligomerization state of purified ConE.

Proteins were analyzed by BN-PAGE according to the manufacturer’s protocols (Invitrogen), as described previously (89). In some cases, protein was preincubated for 30 min with one of the following prior to addition of sample buffer: circular double-stranded plasmid DNA pMMB488 (36.5 nM), single-stranded M13 phage DNA (27 nM), single-stranded oligonucleotide primer oMMB11 (51 μM), or ATP (1 mM). BN-PAGE was performed at least three times.

The oligomerization state of ConE was also analyzed by analytical gel filtration. Analytical gel filtration was performed two to five times with purified ConE using a Superdex 200 10/300 GL column (AKTA; GE Healthcare) equilibrated in dialysis buffer N or K as described previously (89). The mobility of protein standards, run separately, was used to estimate the molecular weight of His6-ConE in solution.

BACTH interaction assays.

The BACTH assay is based on reconstitution of the adenylate cyclase protein from its T18 and T25 domains (62). T18 fusion and T25 fusion plasmids were transformed into E. coli strain BTH101 [F⁻ cya-99 araD139 gal15 galK16 rpsL1 (Str^R) hsdR2 mcrA1 mcrB1], with selection on LB agar with ampicillin (100 μg/mL) and kanamycin (50 μg/mL). Colonies on the transformation plates were resuspended in LB, normalized to optical density, and spotted on LB plates containing X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside; 80 μg/mL), IPTG (0.5 mM), ampicillin (100 μg/mL), and kanamycin (25 μg/mL). After overnight incubation at 30°C, the plates were placed at room temperature and then photographed 24 h later. BACTH assays were performed three or more times.