Biochimica et Biophysica Acta 1828 (2013) 2015–2025

Contents lists available at SciVerse ScienceDirect

Biochimica et Biophysica Acta

j ourna l homepage: www.e lsev ie r .com/ locate /bbamem

The transmembrane domain of the T4SS coupling protein TrwBand its role in protein–protein interactions

Rosa L. Segura a, Sandra Águila-Arcos a, Begoña Ugarte-Uribe a, Ana J. Vecino a, Fernando de la Cruz b,Félix M. Goñi a, Itziar Alkorta a,⁎a Unidad de Biofísica (CSIC, UPV/EHU), and Departamento de Bioquímica y Biología Molecular, Universidad del País Vasco, Apdo. 644, 48080 Bilbao, Spainb Departamento de Biología Molecular e Instituto de Biomedicina y Biotecnología de Cantabria (IBBTEC), Universidad de Cantabria-CSIC-SODERCAN, C. Herrera Oria s/n,39011 Santander, Spain

Abbreviations: T4SS, type IV secretion system; T4CP,transmembrane domain; IR, infrared spectroscopy; DDBTH, bacterial two-hybrid; PRR, proline rich region;IPTG, β-D-thiogalactopyranoside; KDO, 2-keto-3-deo3,3′-dithiobispropionimidate; TrwEBt, Bartonella tribocor⁎ Corresponding author. Tel.: +34 94 6012568; fax: +

E-mail address: itzi.alkorta@ehu.es (I. Alkorta).

0005-2736/$ – see front matter © 2013 Elsevier B.V. Allhttp://dx.doi.org/10.1016/j.bbamem.2013.05.022

a b s t r a c t

a r t i c l e i n f o

Article history:Received 19 March 2013Received in revised form 14 May 2013Accepted 22 May 2013Available online 2 June 2013

Keywords:Bacterial conjugationType IV secretion systemCoupling proteinVirB10-like proteinsConjugative plasmid

Bacteria use type IV secretion systems to transfer genetic material and proteins from donor to recipient cells,using proteins encoded by conjugative plasmids. Among those proteins the so-called Type IV CouplingProtein plays a central role in the process. One of the best studied members of this family is TrwB, theconjugative coupling protein of R388 plasmid. Previous studies indicated that the transmembrane domainof TrwB plays a role beyond the mere anchoring of the protein to the membrane. TrwB has also beenshown to interact with other conjugative proteins, such as the VirB10-like protein of R388 TrwE. The goalof this study is to elucidate the role of the different domains of TrwB and TrwE in their biological function,and in both self- and TrwB–TrwE interactions. To this aim, a series of TrwB and TrwE deletion mutant pro-teins were constructed. Conjugation and interaction studies revealed that the transmembrane domain ofTrwB, and particularly its second transmembrane helix, is needed for TrwB self-interaction and for R388conjugative transfer and that there are contacts between TrwB and TrwE in the membrane. On the contrary,the lack of the TMD of TrwE does not completely abolish R388 conjugation although the interaction betweenTrwE–TrwB is lost. These results identify protein–protein interactions inside the membrane needed for T4SSfunction.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

Type IV secretion systems (T4SS) represent a group of proteinmachines used to translocate proteins or protein–DNA complexesfrom donor to recipient cells [1,2]. They are classified into conjugative(cT4SS), pathogenic (pT4SS) and DNA release/uptake systems [1,3].Despite the wide diversity of their substrates and functions thereis a conserved set of proteins that are found in most T4SSs [4].The paradigm of the T4SS is the VirB/VirD4 system encoded by theAgrobacterium tumefaciens Ti plasmid [3]. The majority of the Gram-negative T4SSs are constituted by an extracellular pilus (VirB2 andVirB5 proteins), three ATPases (VirB4, VirB11 and VirD4) that provideenergy and motor force for macromolecular secretion, architectureassembly and substrate pumping, and a membrane channel that en-compasses bothmembranes, constituted by VirB3-10 proteins [1,5–7].

type IV coupling protein; TMD,M, n-dodecyl-β-D-maltoside;HRP, horseradish peroxidase;xyoctonate; DTBP, dimethylum TrwE homolog34 94 6013500.

rights reserved.

Detailed knowledge of the specific protein–protein interactions inthis highly coordinated mechanism is essential to understand theoverall process. In this respect, the mechanistic knowledge of T4SSmachines has advanced in the recent years due to the availabilityof structural information [5,6,8–10]. However, functional data on invivo and in vitro interactions and information at the cell level are lag-ging behind, although they are also needed for a better understandingof the conjugation mechanism. In particular, the so-called couplingprotein (T4CP) is an essential member of T4SSs because it is thoughtnot only to connect the substrate to be transported to the transmem-brane channel but also to provide part of the energy necessary in theprocess [11]. In particular, the T4CP of R388 plasmid, TrwB, is an innermembrane protein comprising 507 residues. The protein is divided intwo domains: i) the N-terminal domain is referred as transmembranedomain (TMD) [12] and ii) a large cytosolic domain that containsthe characteristic NTP-binding motifs which are responsible for thepotential ATPase activity [13]. TrwB has been actively studied in thelast decades in different forms, namely a soluble deletion mutant(i.e., TrwBΔN70) [14–17], native TrwB purified as protein–lipid-detergent mixed micelles [12,15,18,19], and native TrwB reconstitutedin membranes [20,21]. Structural [10,15,22], biochemical [12,16,19,20],and biophysical data [18,21,22] have shed light on its biological activity.The protein self-organizes into a hexamer [10,12]. Its ATPase activity has

2016 R.L. Segura et al. / Biochimica et Biophysica Acta 1828 (2013) 2015–2025

been detected in the soluble form [16] but not in the full-length protein,neither in detergent suspension nor in reconstituted protein–lipid com-plexes [20]. Even more important, differences in nucleotide bindingaffinity and specificity were observed between TrwB and TrwBΔN70,suggesting that the TMD could regulate its nucleotide binding activity[12,19,20]. Also, IR spectroscopy studies recently published by ourgroup [18,21,23] underline the structural and functional importance ofTMD in TrwB.

Regarding T4CP and its interactions with conjugative proteins,previous investigations [24–27] indicated that TrwE, the VirB10-likeprotein of R388 plasmid, interacts with TrwB and with other couplingproteins from different conjugation systems. VirB10 is a bitopic innermembrane protein that possesses a transmembrane helix connectinga short cytoplasmatic element with a large C-terminal periplasmicdomain [28]. VirB10 from Agrobacterium is an ATP energy sensor[29] and recent structural studies have shown that TraF of thepKM101 plasmid, a VirB10 analogue, forms homo-oligomers thatinteract with VirB7 and VirB9 to form the periplasmic core complex[1,5,6,9].

Studies in different systems indicate that various proteins wouldbe involved in either a coordinated or individual manner in theassembly of T4SSs [30–35]. Nevertheless, the protein–protein interac-tions needed for such a task at the membrane level have not yet beenelucidated.

In this study, we performed a mutational dissection of TrwB andTrwE to analyze the role of their different domains in conjugativetransfer and protein–protein interactions, especially focusing on thefunctions of their N-terminal domains which link both proteins tothe membrane. We obtained results using bacterial two-hybrid(BTH), cross-linking, blue native gels and co-immunoprecipitationtechniques that indicate that some important TrwB and TrwE self-and TrwB–TrwE interactions occur in the membrane.

2. Materials and methods

2.1. Bacterial strains

The strains used in this work are shown in Table 1. For conjugationexperiments, DH5α strain was used as donor and UB1637 cells werethe recipients. The adenylate cyclase-deficient strain DHM1 was usedas a host in bacterial two-hybrid (BTH) assays [39–41]. StrainBL21C41 (DE3) was used for protein overproduction.

2.2. Plasmids

Table 2 shows the plasmids used in this work. Table S1 of thesupplementary material contains the set of plasmids constructed inthis work by using standard DNA recombination technology [44]. pETplasmids (Novagen) were used for over-expression and protein purifica-tion and for complementation, cross-linking and co-immunoprecipitationassays. Vectors pT25, pUT18, and pUT18C were used for BTH assay [41].

Table 1E. coli strains used in this work.

Strain (E. coli) Genotype Reference

DH5α F− endA1 glnV44 thi-1 recA1 relA1 gyrA96 deoRnupG Φ80dlacZΔM15 Δ(lacZYAargF)U169, hsdR17(rK−mK+), λ–

[36], Gibco

BL21C41 (DE3) F− hsds gal (DE3) [37]UB1637 F− lys his trp rpsL recA56 [38]DHM1 F−, cya-854, recA1, endA1, gyrA96 (Nal r), thi1,

hsdR17, spoT1, rfbD1, glnV44(AS)[39], Euromedex

2.3. Purification of TrwB mutant proteins

Protein TrwB encoded by pUB3 and protein TrwBΔN50 encodedby pUB6 were expressed in BL21C41 (DE3) cells. TrwB was purifiedas previously reported [20] except that the final elution buffercontained 0.05% (w/v) DDM. TrwBΔN50 was purified as previouslyreported [19] except for some modifications. Cells that expressedTrwBΔN50 protein were harvested 20 h after induction with IPTG(1 mM) at 25 °C and the membrane fraction obtained by detergentsolubilization was applied to a P-11 column (HealthCare) and elutedwith buffer [50 mM Tris–HCl (pH 7.8), 0.1 mM EDTA, 500 mM NaCl,0.05% (w/v) DDM]. The resulting fractions were supplemented with20 mM imidazole and 0.1 mM PMSF. Subsequently they were loadedonto 5 ml HiTrap-Chelating columns (GE Healthcare) connected to aPharmacia FPLC system equilibrated with buffer A [50 mM Tris–HCl(pH 7.8), 0.1 mM EDTA, 200 mM NaCl, 0.05% (w/v) DDM], sup-plemented with 20 mM imidazole. Proteins were eluted from the col-umns with a 20–168 mM imidazole gradient in buffer A. The proteinwas concentrated using an ultrafiltration cell with YM-30 ultrafiltra-tion membrane of regenerated cellulose (Amicon) to a final volumeof 1 ml and loaded onto a 24 ml Superose 6 column (GE Healthcare)by using Pharmacia FPLC equipment. Gel filtration was performedin buffer A. The peak fractions corresponding to TrwBΔN50 werepooled, glycerol was added to a 20% (v/v) final concentration, andthe protein was stored at −80 °C. TrwB and TrwBΔN50 protein sam-ples were separated on Blue Native PAGE Novex 4–16% Bis–Tris gels(Invitrogen) at 150 V for 30 min and 250 V for 2 h at 4 °C.

2.4. Complementation assays

Conjugative mating experiments were performed with modificationsof a previously reportedmethod [27]. Briefly, 500 μl of overnight culturesof donor and recipient strains were mixed, centrifuged and the pelletplaced onto a GS Millipore filter (0.22-μm pore size) on a pre-warmedLB-agar plate for 1 h at 37 °C. Then, bacteria were washed from the filter,diluted in 2 ml LB, and suitable dilutions plated on selectivemedia (platescontained streptomycin and trimethoprim for the transconjugants andonly trimethoprim for donors). Conjugation frequencieswere normalizedfor the number of transconjugants per donor cell.

2.5. Immunoblot of TrwB and TrwB mutant proteins

BL21C41 (DE3) cells transformed with vectors pUB3, pUB4, pUB5,pUB6 pUB7, pUB11, pSU4637, pUB12, pUB13, pUB14, pUB15, pUB16or pUB17 were grown in 10 ml LB supplemented with ampicillin at37 °C. Cultures were resuspended in 1 ml 50 mM Tris–HCl (pH 7.8),0.1 mM EDTA, and 200 mM NaCl buffer supplemented with 1 mg/mllysozyme. After 30 min in ice, cells were disrupted by sonication andcentrifuged at 7000 ×g for 15 min in order to sediment non-lysed cells.

The supernatant was centrifuged at 100,000 ×g for 1 h at 4 °C, andthe new supernatant and pellet, that constitute the soluble and mem-brane fractions respectively, were analyzed by Western blotting.Immunoblotting was performed using as primary antibodies mouseanti-His (C-term) monoclonal antibody (Invitrogen) and rabbitserum anti-TrwB [12]; and as secondary antibodies goat anti-mouseand goat anti-rabbit antibodies conjugated with horseradish peroxi-dase (HRP) (Invitrogen).

2.6. Bacterial two-hybrid (BTH) assays

Derivatives of strain DHM1 co-transformed with plasmids bearingT25 and T18 fusions (Table 2) were grown at 30 °C for 48 h. As a pos-itive control, competent cells were co-transformed with the controlplasmids pKT25-zip and pUT18C-zip. Quantification of the functionalcomplementation mediated by interaction between two proteins, Xand Y, was carried out by measuring β-galactosidase activities in

Table 2Plasmids used in this work. TRAW, transfer region of the conjugative plasmid R388; AmpR, ampicillin resistance; CmR, chloramphenicol resistance; SuR, sulfonamide resistance; andTpR

, trimethoprim resistance.

Plasmid Description Phenotype Reference

pET22b(+) Expression vector AmpR, C-Terminal−His tag NovagenpET3a′ Expression vector AmpR, Rep (pMB8) [42]pSU4637 pET3a′::trwBΔN70 AmpR, Rep (pMB8) [15]pUB1 pET3a′::trwB AmpR, Rep (pMB8) [12]pUB3 pET22b(+)::trwB AmpR, His-tag [19]R388 Natural plasmid SuR, TpR, TRAW, IncW [43]pSU1443 R388 except trwB gene SuR, TpR, TRAW, IncW [13]pSU4134 R388 except trwE gene SuR, TpR, TRAW, IncW [27]pUB4 pET22b(+)::trwBΔN9 AmpR, TrwBΔN9 C-Terminal−His tag This workpUB5 pET22b(+)::trwBΔN27 AmpR, TrwBΔN27 C-Terminal−His tag This workpUB6 pET22b(+)::trwBΔN50 AmpR, TrwBΔN50 C-Terminal−His tag This workpUB7 pET22b(+)::trwBTM AmpR, TrwBN77 C-Terminal−His tag This workpUB11 pET22b(+)::trwBΔN6 AmpR, TrwBΔN6 C-Terminal−His tag This workpETDuet1 Expression vector AmpR, His tag and S tag NovagenpUB12 pETDuet1::trwE AmpR, TrwE C-Terminal−His tag This workpUB13 pETDuet1::trwEΔN42 AmpR, TrwEΔN42 C-Terminal−His tag This workpUB14 pETDuet1::trwEΔN64 AmpR, TrwEΔN64 C-Terminal−His tag This workpUB15 pETDuet1::trwEN91 AmpR, TrwEN91 C-Terminal−His tag This workpUB16 pETDuet1::trwEN201 AmpR, TrwEN201 C-Terminal−His tag This workpUB17 pETDuet1::trwEΔN192 AmpR, TrwEΔN192 C-Terminal−His tag This workpT25 Two-hybrid expression vector CmR, T25 [39]pT25 zip Two-hybrid positive control CmR, T25-leuzipper [39]pUT18 Two-hybrid expression vector AmpR, T18 [39]pUT18C Two-hybrid expression vector AmpR, T18 [39]pUT18 zip Two-hybrid negative control AmpR, T18-leuzipper [39]pUB19 pT25::trwB CmR, T25: TrwB This workpUB20 pT25::trwBΔN6 CmR, T25: TrwBΔN6 This workpUB21 pT25::trwBΔN27 CmR, T25: TrwBΔN27 This workpUB22 pT25::trwBΔN50 CmR, T25: TrwBΔN50 This workpUB23 pT25::trwE CmR, T25: TrwE This workpUB24 pT25::trwEΔN192 CmR, T25: TrwEΔN192 This workpUB25 pUT18::trwB AmpR, T18: TrwB This workpUB26 pUT18::trwBΔN9 AmpR, T18: TrwBΔN9 This workpUB27 pUT18::trwBΔN27 AmpR, T18: TrwBΔN27 This workpUB28 pUT18::trwBΔN50 AmpR, T18: TrwBΔN50 This workpUB29 pUT18::trwBTM AmpR, T18: TrwBTM This workpUB30 pUT18::trwEΔN192 AmpR, T18: TrwEΔN192 This workpUB31 pUT18C::trwB AmpR, TrwB: T18 This workpUB32 pUT18C::trwBΔN9 AmpR, TrwBΔN9: T18 This workpUB33 pUT18C::trwBΔN27 AmpR, TrwBΔN27: T18 This workpUB34 pUT18C::trwBΔN50 AmpR, TrwBΔN50: T18 This workpUB35 pUT18C::trwE AmpR, TrwE: T18 This workpUB36 pUT18C::trwEΔN42 AmpR, TrwEΔN42: T18 This workpUB37 pUT18C::trwEΔN64 AmpR, TrwEΔN64: T18 This workpUB38 pUT18C::trwEN91 AmpR, TrwEN91: T18 This workpUB39 pUT18C::trwEN201 AmpR, TrwEN201: T18 This workpUB40 pUT18C::trwEΔN192 AmpR, TrwEΔN192: T18 This workpUB41 pETDuet1::trwB AmpR, TrwB C-Terminal−S tag This workpUB42 pETDuet1::trwE::trwB AmpR, TrwE C-Terminal−His tag

AmpR, TrwB C-Terminal−S tagThis work

pUB43 pETDuet1::trwEΔN42::trwB AmpR, TrwE ΔN42 C-Terminal−His tagAmpR, TrwB C-Terminal−S tag

This work

pUB44 pETDuet1::trwEΔN64::trwB AmpR, TrwEΔN64 C-Terminal−His tagAmpR, TrwB C-Terminal−S tag

This work

pUB45 pETDuet1::trwEN91::trwB AmpR, TrwEN91 C-Terminal−His tagAmpR, TrwB C-Terminal−S tag

This work

pUB46 pETDuet1::trwEN201::trwB AmpR, TrwEN201 C-Terminal−His tagAmpR, TrwB C-Terminal−S tag

This work

pUB47 pETDuet1::trwEN192::trwB AmpR, TrwEN192 C-Terminal−His tagAmpR, TrwB C-Terminal−S tag

This work

2017R.L. Segura et al. / Biochimica et Biophysica Acta 1828 (2013) 2015–2025

liquid cultures [41]. These data were analyzed taking into accountthat when no interaction occurs, the DHM1 strain expresses about150 units of β-galactosidase/mg bacterial dry weight. Table S2 of sup-plementary material contains the set of plasmids used in this work toanalyze TrwB or TrwE self-interactions and TrwB–TrwE interactions.

2.7. Cellular location of TrwE protein

BL21C41 (DE3) cells transformed with vector pUB12 were grownin 500 ml LB at 37 °C. When A600 reached 0.8, 1 mM IPTG was addedand cells were incubated for an additional 4 h at 30 °C with constant

shaking. The cell pellet was resuspended in 10 ml cold 10 mM Trisbuffer (pH 7.8) supplemented with 0.75 M sucrose and 0.1 mg/mllysozyme and incubated for 5 min in ice. Spheroplasts were formedby slowly diluting the suspension with 2 volumes of cold 1.5 mMEDTA at pH 7.0. Samples were sonicated and the resulting lysatewas clarified by centrifugation for 20 min at 8000 ×g at 4 °C. Thepellet containing the membrane fraction was sub-fractionated in asucrose gradient as described [45,46]. Three distinct membranefractions were obtained following their ultracentrifugation. NADHoxidase and KDO (2-keto-3-deoxyoctonate) were used as inner andouter membrane markers, respectively [47]. NADH oxidase and

2018 R.L. Segura et al. / Biochimica et Biophysica Acta 1828 (2013) 2015–2025

KDO assays confirmed the identification and purity of the low densi-ty fraction (top band) as cytoplasmic membrane material, whereasthe high density fraction (bottom band) was purified as outermembrane components. The middle density fraction (middle band)included a mixture of both membranes. The amount of TrwE wasmeasured by immunoblotting analysis, quantifying the bands in aBioRad GS-800 calibrated densitometer using Quantity One program.

2.8. Protein interactions by chemical cross-linking assay

Cross-linking experiments were performed using dimethyl-3,3′-dithiobispropionimidate 2HCl (DTBP) from Pierce. Derivatives ofEscherichia coli BL21C41 (DE3) carrying the appropriate plasmid(pUB3, pUB5, pUB6, pUB7 or pUB12) were grown in 50 ml culture at25 °C for 4 h after induction with 1 mM IPTG. Next, cells wereharvested and resuspended in 1 ml 50 mM HEPES (pH 8). Cell debrisremoval was carried out by centrifugation at 19,000 ×g for 50 min at4 °C, and the resultant clear lysates were separated into four 100 μlaliquots. DTBPwas added immediately to each aliquot to a final concen-tration of 0, 0.5, 5, or 50 mM, respectively. The samples were invertedgently to mix the suspension, and incubated at room temperature for1 h to allow cross-linking to occur. The cross-linking reaction wasquenched by adding 1 M Tris–HCl to a 100 mM final concentration,followed by incubation at 23 °C for 10 min. Cross-linked sampleswere either immediately analyzed by immunoblotting or stored at−20 °C until required.

Samples were incubated at 30 °C for 30 min in NuPAGE LDS sam-ple buffer (Invitrogen). For reducing conditions, 100 mM DTT wasadded and samples were boiled for 10 min. Cross-linked complexeswere separated by SDS-NuPAGE 4%–12% polyacrylamide gradientgel (Invitrogen). Protein transfer onto nitrocellulose membranesusing a Bio-Rad Trans-Blot semi-Dry Transfer Cell was carried out at15 V for 15 min, and immunoblotting was performed using mouseanti-His (C-term) monoclonal antibody (Invitrogen) and rabbitserum anti-TrwB as primary antibodies and HRP-conjugated goatanti-mouse or goat anti-rabbit as secondary antibodies.

2.9. TrwB–TrwE interactions by co-immunoprecipitation assays

E. coli BL21C41 (DE3) strain was separately transformed withpUB41, pUB42, pUB43, pUB44, pUB45, pUB46 or pUB47 plasmids, ex-pressing S•Tag fusion TrwB (Novagen) and His-tagged TrwE mutantproteins. Cell cultures (50 ml) were grown after induction with1 mM IPTG for 4 h at 25 °C. Subsequently, they were centrifugedand the pelleted cells were incubated in 1 ml 50 mM Tris–HCl(pH 7.8), 100 mM NaCl solution with a protease inhibitor cocktail(Roche) supplemented with 1% (w/v) DDM with constant shakingfor 1 h at 4 °C. The solubilized material of each culture was separated

Fig. 1. Mutant protein design. Schematic representation of TrwB (A) and TrwE (B) mutant ptransmembrane regions are illustrated as shaded boxes. H1 and H2 correspond to transmemthe periplasm is represented as PRR.

by centrifugation for 30 min at 70,000 ×g. To identify interactionsbetween TrwB and TrwE mutant proteins, 450 μl of supernatant wasincubated for 1 h at 4 °C with 50 μl Dynabeads-Protein A (Invitrogen)previously pre-incubated and cross-linked with anti-His commercialantibody in order to bind His-tagged TrwE mutant proteins. Beadswere washed three times with 50 mM Tris–HCl (pH 7.8), 100 mMNaCl and 1% (w/v) DDM. TrwB–TrwE interactions were analyzed byWestern blotting and immunostaining with rabbit serum anti-TrwBand HRP-conjugated goat anti-rabbit antibody, as primary and sec-ondary antibodies, respectively.

3. Results

3.1. The transmembrane domain of TrwB is essential for its biologicalactivity

TrwB is a membrane protein consisting of a small N-terminal, TMDand a bulky cytosolic domain. In particular, in the TMD a small cyto-solic stretch of 9 amino acids and two hydrophobic helixes (H1 andH2) linked by a periplasmic loop can be distinguished. To study therole of the different structural elements of TrwB on its biological func-tion, different deletion mutant proteins were constructed (Fig. 1):i) TrwBΔN6 protein lacks the first six N-terminal amino acids ofTrwB, ii) TrwBΔN9 lacks the polar residues Arg7-Lys8, the residueVal9 and the abovementioned six N-terminal amino acids, iii) TrwBΔN27 lacks the small N-terminal cytosolic domain plus the first trans-membrane helix (H1), iv) TrwBΔN50 contains only the second trans-membrane helix (H2) and the voluminous C-terminal cytosolicdomain, v) TrwBΔN70 represents the C-terminal cytosolic domainand vi) TrwBTM comprises the first 77 amino acids of TrwB and repre-sents its TMD.

Melting temperatures and secondary structure components ofsome TrwB deletion mutant proteins measured by IR spectroscopyindicate that the overall tertiary structure of those mutant proteinswas maintained [23]. TrwBΔN27 (unpublished data) and TrwBΔN50[23] mutant proteins maintained their nucleotide binding activity.In addition, both proteins were isolated from the membrane fraction,and detergent was needed to extract them from the membrane, indi-cating that both TrwBΔN27 (unpublished data) and TrwBΔN50 [23]are membrane proteins. For in vivo complementation analysis, E. colidonor cells harboring trwB-deficient R388 plasmid pSU1443 [13]were transformed with plasmids containing either trwB or trwBN-terminal deletion mutants. Plasmid pUB3 was constructed to ex-press TrwB wild-type protein and plasmids pUB4, pUB5, pUB6,pUB7 and pUB11 were designed to express TrwBΔN9, TrwBΔN27,TrwBΔN50, TrwBTM and TrwBΔN6 proteins, respectively (for moreinformation see Table 2). pSU4637was used to express TrwBΔN70 pro-tein [15]. The ability of the trwB mutant genes to restore R388ΔtrwB

roteins. Open boxes: cytosolic domain. Hatched boxes: periplasmic domain. Predictedbrane helix 1 and 2 of TrwB, respectively. The proline-rich region that is also located in

NA

DH

oxi

dase

(m

mol

.min

-1.m

L-1)

and

KD

O (

mg.

mL-1

)

0

1

2

3

4

5

6

7

Trw

E (

a.u.

)

TrwE content NADH oxidase activityKDO content

Solubleprotein

Topband

Middleband

Bottomband

Fig. 3. TrwE cell location. Cell fractions were obtained as described in Materials andmethods section. The TrwE content in each sucrose gradient band (top, middle andbottom band) was determined by densitometry analysis of Western blots. Top bandcorresponds to low density fraction, middle band corresponds to middle density fractionand bottom band corresponds to high density fraction. TrwE content (black bars), NADHoxidase activity (white bars) and KDO content (grey bars). In the soluble fraction, none ofthe proteins was detected.

Table 3Transfer frequencies of plasmid pSU1443 (trwB−) or plasmid pSU4134 (trwE−) whencomplemented by TrwB, TrwE or their deletion mutant proteins. Transfer efficienciesof plasmids bearing different deletions in trwB or trwE genes are shown. Their comple-mentation ability was measured in cells harbouring R388 plasmids deficient in trwB(pSU1443) or trwE (pSU4134) genes. Transfer frequencies are indicated as the numberof transconjugants per donor in standard matings.

Plasmids in donors Complementary protein Conjugation frequenciesa

R388 None 0.21 ± 0.05pSU1443 None b10−8

pSU1443 + pUB3 TrwB 0.10 ± 0.09pSU1443 + pUB11 TrwBΔN6 1.00 10−4 ± 2.7 × 10−4

pSU1443 + pUB4 TrwBΔN9 8.82 10−6 ± 2.1 × 10−6

pSU1443 + pUB5 TrwBΔN27 1.26 10−5 ± 2.5 × 10−6

pSU1443 + pUB6 TrwBΔN50 2.54 10−6 ± 1.5 × 10−6

pSU1443 + pSU4637 TrwBΔN70 b10−8

pSU1443 + pUB7 TrwBTM b10−8

pSU4134 None b10−8

pSU4134 + pUB12 TrwE 0.74 ± 4.9 × 10−2

pSU4134 + pUB13 TrwEΔN42 1.69 10−4 ± 1.7 × 10−5

pSU4134 + pUB14 TrwEΔN64 1.22 10−6 ± 1.9 × 10−6

pSU4134 + pUB17 TrwEΔN192 b10−8

pSU4134 + pUB15 TrwEN91 b10−8

pSU4134 + pUB16 TrwEN201 b10−8

a Frequencies were calculated with data from at least five independent rep-licas + standard deviations.

2019R.L. Segura et al. / Biochimica et Biophysica Acta 1828 (2013) 2015–2025

conjugation was monitored by mating experiments. None of themutated trwB genes was able to restore completely the conjugationactivity of the native gene (Table 3). A significant loss of conjugationactivity was observed when the 6 first N-terminal amino acids wereremoved (TrwBΔN6 mutant protein). In fact, a progressive loss of con-jugation activity was observed upon gradual removal of the structuralelements of the TMD (i.e., TrwBΔN9, TrwBΔN27 and TrwBΔN50mutantproteins). Conjugation activity was completely abolished when theprotein lacked the whole TMD (TrwBΔN70).

To prove that the differences found in complementation assayswere not due to deficient expression levels of the mutant proteins,their expression level was analyzed (Fig. 2A). It was observed thatall mutant proteins presented sufficient expression levels and that,except for TrwBΔN70, they were located in the membrane fractionas expected. Besides, according to de Paz and co-workers [25] TrwBis in excess under standard conjugation conditions, indicating thatthe frequency differences observed are not caused by differences inthe expression levels.

3.2. The C-terminal domain of TrwE is essential for R388 conjugation

TrwE is the VirB10-like protein of R388. It is a bitopic membraneprotein that contains a small cytosolic domain comprising theN-terminal 41 amino acids, and a transmembrane helix that connectsthe cytosolic domain with a large periplasmic domain. As predicted byTMPred (http://www.ch.embnet.org/software/TMPRED_form.html),the protein has two putative TMDs, one of them including amino

Fig. 2. Western blots to assay wild-type TrwB and TrwE proteins and mutant proteinexpression levels. (A) Expression of TrwB and its deletionmutant proteins. No degradationbands were observed and only the portion of the nitrocellulose membrane containing theband is shown for convenience. (B) Expression of TrwE and its deletion mutant proteins.TrwE, TrwEΔN42, TrwEN91 and TrwEN201 were detected in the membrane fractionwhile TrwEΔN64 and TrwEΔN192 were detected in the soluble fraction.

acids 43–63 that connects the cytosolic domain and a large periplasmicdomain, and a second one, with a lower score, constituted by aminoacids 301–324 that is likely connected to the outer cell membrane.Similarly, the crystal structure of TraF confirmed that the VirB10-likeprotein of plasmid pKM101 spans both membranes of Gram-negativebacteria [5]. In the large periplasmic domain, as other VirB10-like pro-teins, TrwE also shows a Proline Rich Region (PRR) located betweenamino acids 70 and 169 and the C-terminal domain that comprisesresidues 193–395. This latter domain could be in contact with theouter membrane through the aforementioned second transmembranedomain.

First, to study the role of the cytosolic domain of TrwE in its biolog-ical activity, the TrwEΔN42mutant proteinwas constructed. It lacks thesmall cytosolic N-terminal fragment (amino acids 1 to 42) (Fig. 1). Adecrease in conjugation frequency of more than three orders of magni-tude in relation to R388 was observed (Table 3). Then, the effect ofremoving the small cytosolic domain plus the transmembrane helixon the transfer capacity of TrwE was studied (TrwEΔN64 mutantprotein) (Fig. 1). A decrease of two orders of magnitude of the transfercapacity in comparison to the previously described protein was ob-served (Table 3).When the small cytosolic domain, the transmembranehelix and the PRR were removed (TrwEΔN192 mutant protein) and acomplete loss of conjugation activity was observed (Table 3). Finally,mutant proteins that lacked either the PRR and the C-terminal regionor the C-terminal region (Fig. 1, TrwEN91 and TrwEN201 mutant pro-teins, respectively) completely lost their conjugation capacity (Table 3).

To exclude that the results described above were due to deficientexpression of the mutant proteins, expression levels of TrwE and itsdeletion mutant proteins were analyzed by Western blotting(Fig. 2B). The Fig. 2B shows sufficient expression level of the mutantproteins that, except for TrwEΔN64 and TrwEΔN192 mutant proteins,were localized in the membrane fraction.

3.3. TrwE is mainly anchored to the inner membrane

To study the intracellular localization of TrwE, total protein lysate ofE. coli BL21C41 (DE3) containing plasmid pUB12 (Table 2)was separatedinto soluble, inner membrane and outer membrane fractions. As shownin Fig. 3, most of TrwE co-localized with the NADH oxidase activity inthe upper band of the gradient, indicating that the majority of TrwEwas anchored to the bacterial inner membrane.

Fig. 4. In vivo protein–protein interactions measured by bacterial two-hybrid(BTH) assay. The appropriate plasmid pairs were introduced in the DHM1 strainin the absence of other conjugation proteins, and β-galactosidase activity wasassayed. (A) TrwB-TrwB interactions. (B) TrwB–TrwE interactions. (C) TrwE–TrwEinteractions.

2020 R.L. Segura et al. / Biochimica et Biophysica Acta 1828 (2013) 2015–2025

3.4. Protein–protein interactions between TrwB, TrwE and their deletionmutant proteins

Complementation studies (Table 3) demonstrated that the TMD ofTrwB is essential for R388 conjugation. This result led us to studyTrwB self-interactions and interactions with other conjugative pro-teins in search of their role in the biological activity of TrwB. In fact,previous studies using random mutagenesis of TrwB indicated thatTrwB interacts with cognate and heterologous VirB10-like proteinsin the membrane [25,27].

To shed further light on the role of the different domains of TrwBand TrwE in protein–protein interactions, the following studies wereperformed. BTH experiments based on the interaction of the T25 andT18 domains of adenylate cyclase of Bordetella pertussis weredesigned and the results of TrwB self-interactions, interactionsbetween TrwB and TrwE or its deletion mutant proteins and TrwEself-interactions are shown in Fig. 4 (panels A, B and C, respectively)(additional information is given in Table S2). Even more, self-interactions between TrwB and its deletion mutant proteins wereanalyzed by DTBP chemical cross-linking in the membrane (Fig. 5).Additionally, blue native gel electrophoresis was also performed toanalyze TrwB and TrwBΔN50 self-interactions and results areshown in Fig. 6. Finally, interactions between TrwB and TrwE or itsdeletion mutant proteins were studied by co-immunoprecipitationassays (Fig. 7).

3.4.1. The TMD of TrwB is the minimal element necessary for TrwB–TrwBinteractions

Due to the major role that the TMD of TrwB has shown in differentproperties of TrwB [18–21,23], it was interesting to analyze the role ofthis TMD in the oligomerization of the protein. Fig. 4A shows theinteractions between TrwB and its deletion derivatives using theBTH assay. The following interactions: TrwB–TrwB, TrwB–TrwBΔN6,TrwB–TrwBΔN9, TrwB–TrwBΔN27, TrwB–TrwBΔN50 and TrwB–TrwBTM were detected. They indicate that the N-terminal domainof TrwB mediates TrwB–TrwB interactions. It is important to notethat BTH negative results found here (TrwBΔN27–TrwBΔN27,TrwBΔN27–TrwBΔN50, TrwBΔN50–TrwBΔN50) could be false nega-tive results due to steric effects that impede to restore the adenylatecyclase activity [40]. Therefore, to complete the data obtained inBTH assays, DTBP cross-linking (Fig. 5) and blue native gels (Fig. 6)were accomplished. Interactions between TrwB–TrwB, TrwBΔN27–TrwBΔN27 and TrwBΔN50–TrwBΔN50 were observed by DTBPcross-linking (Fig. 5A). In all cases homo-oligomers (dimers, tetra-mers and higher molecular weight complexes) were detected. Fur-thermore, when TrwB, TrwBΔN27 or TrwBΔN50 was not denaturedby heating prior to loading onto the gel they formed stable dimerseven in the absence of cross-linker (Fig. 5A, lanes 0). Similarly, blue-native gel electrophoresis (Fig. 6) showed that TrwB and TrwBΔN50were able to self-interact. In particular, the results obtained withTrwBΔN50 (Figs. 5A and 6) highlighted the importance of the secondtransmembrane helix (H2) in TrwB oligomerization.

Interestingly, the TrwB–TrwBTM interaction observed in BTHassays (Fig. 4A) underlines that the first 77 amino acids of the proteinare required to maintain the interactions between TrwB monomers.Furthermore, cross-linking showed that this domain alone (TrwBTMmutant protein) forms dimers (Fig. 5B), suggesting that the TMD ofTrwB could be the minimal element necessary for TrwB self-interactions. On the contrary, TrwBΔN70 self-interaction by DTBPcross-linking was not detected (Fig. 5C).

3.4.2. The transmembrane domains of TrwB and TrwE are involved intheir interactions

T4CPs are supposed to interact with other proteins of the T4SS. Inparticular, interactions between T4CP and VirB10-like proteins havebeen already described [24–27] and more specifically, between TrwE,

Fig. 5. TrwB–TrwB interactions analyzed by chemical cross-linking assay. (A) TrwB, TrwBΔN50, TrwBΔN27, (B) TrwBTM and (C) TrwBΔN70 in presence of the chemical cross-linkerDTBP. After treatment with DTBP the resultant products were separated on a Bis–Tris 4%–12% SDS-polyacrylamide gradient gel. Lanes 0 mean that no cross-linker was added.(D) Control samples (TrwBTM, TrwB, TrwBΔN27 and TrwBΔN50) were boiled in the presence of 100 mM DTT without previous treatment with the cross-linker. All sampleswere then transferred to a nitrocellulose membrane for Western blotting and proteins were detected using specific anti-TrwB or commercial anti-His antibodies as primaryantibodies and were recognized by HRP-conjugated secondary antibodies.

2021R.L. Segura et al. / Biochimica et Biophysica Acta 1828 (2013) 2015–2025

the VirB10-like protein of R388, and TrwB [27]. Nevertheless, more in-formation about the role of the different domains of TrwB and TrwE intheir interaction is needed. In order to do so, interactions betweenTrwB–TrwE and their deletion mutant proteins were analyzed byBTH and co-immunoprecipitation assays (Figs. 4B, 7, respectively).

Fig. 6. TrwB (A) and TrwBΔN50 (B) self-interaction confirmed by blue native gel electrophopolyacrylamide gradient gel. M termed lanes correspond to molecular weight markers.

Interactions between TrwB and TrwE were observed by BTH assay,even when TrwE lacked either the cytosolic (TrwEΔN42) or the peri-plasmic domain (TrwEN91) (Fig. 4B). Similarly, when TrwB–TrwE in-teractions were analyzed by co-immunoprecipitation, mutant proteinslacking the TMD of TrwE lost their interaction capacity with TrwB

resis. TrwB and TrwBΔN50 were purified and proteins were loaded onto 4%–16% native

Fig. 7. TrwB interaction with TrwE and its deletion mutant proteins was analyzed byco-immunoprecipitation. Co-immunoprecipitation was performed with anti-His-Dynabeads.E. coli BL21C41 (DE3) strain was separately transformed with pUB41, pUB42, pUB43,pUB44, pUB45, pUB46 or pUB47 plasmids (see supplementary material Table S1).Co-expression of S•Tag fusion TrwB and His-tagged TrwE or its mutant proteins is describedin Material and methods section. Total cell lysate was incubated with Dynabeads-anti-Hisand possible TrwB interaction with either the native form of TrwE or TrwE mutant proteinswas analyzed using anti-TrwB specific antibodies. pUB41 which only expresses TrwB wasused as negative control (lane named TrwB control-) to prove that TrwB did not show anyunspecific interaction with the resin.

2022 R.L. Segura et al. / Biochimica et Biophysica Acta 1828 (2013) 2015–2025

(TrwEΔN64 and TrwEΔN192 mutant proteins). Interactions weredetected between TrwB and TrwEN91 or TrwEN201 mutant proteins,indicating that there are interactions that occur in the first 91 aminoacids of TrwE where the transmembrane helix is located (Fig. 7). Inthe same way, the TMD of TrwB alone (TrwBTM mutant protein) wasable to interact with TrwE, as shown by BTH assays (Fig. 4B).

Finally, BTH and co-immunoprecipitation showed that TrwB–TrwE interactions were lost when the transmembrane helix of TrwEwas removed (TrwEΔN64) (Figs. 4B and 7). Nevertheless, the lackof this interaction does not abolish completely the in vivo activity ofR388 as shown by TrwEΔN64 that maintained part of the conjugationcapability (Table 3).

3.4.3. The TMD of TrwE contributes to TrwE self-interactionsWhen self-interactions between TrwE and its deletion mutant

proteins were analyzed by BTH technique it was observed thatTrwEΔN42 and TrwEN91 mutant proteins interacted with TrwE(Fig. 4C). From these results we infer that there are interactionsbetween TrwE monomers that occur in the membrane or in theperiphery of the membrane, both in the absence of the cytoplasmicor periplasmic domain of TrwE.

4. Discussion

The coupling protein of plasmid R388, TrwB, is a membraneprotein that consists of a bulky cytosolic domain and a small TMD.Recently, it has been reported that both TMD and the membrane itselfhad an effect on structural and functional aspects of TrwB. IR experi-ments showed that when the TMD was completely or partiallyremoved (TrwBΔN70 and TrwBΔN50, respectively) the percentageof periodically ordered elements of secondary structure decreased[18,23]. Moreover, when TrwB and TrwBΔN50 mutant protein wereinserted into the membrane (proteoliposomes) a more ordered struc-ture and a higher denaturation temperature was observed [21,23].Regarding the nucleotide binding activity, it has been described thatthe TMD confers specificity for purine nucleotides [19]. Even more,

when TrwB was inserted into the membrane a strong increment ofthe affinity and specificity for ATP was observed [20]. Nevertheless,up to date DNA-dependent ATPase activity has been only reportedfor the soluble mutant protein TrwBΔN70 [14,16,17]. The fact thatthis ATPase activity has not been found in TrwB neither in detergentnor when inserted into liposomes, and the observed differences inthe affinity and specificity for ATP between the wild type and thesoluble mutant protein of TrwB, led us to think that the TMD couldregulate the ATPase activity of TrwB by interacting with other R388conjugative proteins.

There are several studies that report protein–protein interactionsbetween T4CPs and relaxosomal proteins. In particular, studies onR388 plasmid [15–17,27], RP4 [48], and R1 [49,50], support theexistence of such contacts. In particular, the abovementioned andother authors [24–27,31] indicate that T4CP–VirB10 interactions area key matter in the conjugative process. However, these studiesdescribe interactions in the cytosol, while interactions in the mem-brane remained unexplored.

In this study we have investigated the role of the differentdomains of TrwB and TrwE in the R388 transfer capacity aswell as protein–protein interactions between these two proteins.In order to do so, we constructed deletion mutant proteins ofTrwB and TrwE. Fig. 1 shows deletion mutant proteins used inthis study.

In vivo complementation analysis indicated that the N-terminal 6amino acids of TrwB are functionally important since the conjugationcapacity of the mutant protein TrwBΔN6 decreases significantly,suggesting that this small fragment that is predicted to be located inthe cytosol could be an important element for the transfer of R388plasmid probably by its interaction with the relaxosome as previouslydescribed [25]. However, the interaction observed by BTH experi-ments between TrwB and TrwBΔN6 suggests that those firstN-terminal 6 amino acids of TrwB are not essential for TrwB oligo-merization. Moreover, when the transfer activity of TrwBΔN6 andTrwBΔN9 was compared, a significant decrease in conjugation fre-quency of TrwBΔN9 mutant protein was observed. The latter proteinlacks the polar residues Arg7–Lys8. In this regard, the importance ofpolar residues in the proper insertion of a model transmembranehelix in the membrane has been demonstrated by von Heijne andco-workers [51,52]. Therefore, this result may suggest that whenthe abovementioned polar residues are removed (TrwBΔN9 mutantprotein) helix H1 could not be properly oriented/inserted in the mem-brane, affecting thereby the biological activity of the protein. Comple-mentation assays of TrwBΔN27 and TrwBΔN50 mutant proteinsshowed that deletion of the first transmembrane helix (H1) preservedsome in vivo activity that was totally abolished when the proteinlacked the whole TMD, losing the membrane anchorage (TrwBΔN70mutant protein). Previous work has reported the role of the TMD onbiochemical and structural aspects of TrwB [10,12,18–21,23], butthis is the first time that the relationship between the differentelements of the TMD of TrwB and the R388 transfer capacity (invivo activity) is shown.

When the first helix was removed it was observed that TrwBΔN50mutant protein formed homo-oligomers in the membrane as shownby cross-linking and when TrwBΔN50 was purified and analyzed byblue native gel experiments (Table 4), suggesting that the secondtransmembrane helix (H2) could play a key role in TrwB self-interactions. BTH experiments confirmed the interaction betweenTrwB and TrwBTM (Fig. 4A) and cross-linking analysis showed thatnot only TrwB forms different oligomers but also that the TMD(TrwBTM) alone forms dimers (Fig. 5A and B, respectively). Onthe contrary, the soluble mutant protein TrwBΔN70 that lacks theTMD did not form oligomers when studied by cross-linking(Fig. 5C). These results are in agreement with previous BTH studies,gel filtration and analytical ultracentrifugation experiments carriedout on soluble mutant proteins of T4CPs that revealed that mutations

Table 4Summary of conjugation frequencies and protein–protein interactions of TrwB and TrwE proteins. Schematic representation of TrwB and TrwE mutant proteins. Open boxes: cytosolicdomain. Hatched boxes: periplasmic domain. Shaded boxes: predicted transmembrane regions. PRR: Proline Rich Region. Conjugation frequencies. Conjugation frequencies werenormalized to the number of transconjugants per donor cell. Symbols + and – account for positive or negative protein–protein interactions. BTH: bacterial two hybrid assay.CL: chemical cross-linking. BNG: blue native gel electrophoresis. IP: co-immunoprecipitation. GF: gel filtration. AUC: analytical ultracentrifugation.

Schematic representationComplementary

proteinConjugationFrequencies

In vivo activity Protein-Protein interactions

TrwB TrwB∆N70

TrwB∆N50

TrwB∆N70

TrwE∆N42

TrwE∆N64

TrwE∆N192

TrwB∆N27

TrwB∆N9

TrwB∆N6

TrwBTM TrwE

PRR

BTH +BTH +

CL +

CL +

BNG +

BTH +CL +

BNG +

BTH +

BTH +

BTH +

CL +BTH +

BTH +

BTH +

BTH +

N t

N t

N t N t

N t N t

N t

N t

N t

N t

N t

N t

N t

N t

N t

N t

N t

N t

N t

N t

N t

N t

N t

N t

N t

N t

N t

IP +

BTH +

BTH +

BTH +

BTH +

BTH +

BTH -

BTH -

BTH +

BTH +

IP +

BTH +

IP +

BTH -

IP -

BTH -

IP -

BTH -

IP +

BTH -

IP +

AUC - (*)

G F - (*)CL -

TrwEN201

TrwEN91

<10-8

<10-8

<10-8

<10-6

<10-8

1.22 10-6

2.54 10-6

1.26 10-5

8.82 10-6

1.00 10-4

1.69 10-4

0.74

0.10TrwB

TrwBTM

TrwE

1 201PRR

1 91

192 407

407

407

407

64PRR

PRR

PRR

42

1

1 77

70 507

507

507

507

507

507

50

27

9

6

3

Nt: not tested.

(*) Results obtained from Matilla et. al. [14].

2023R.L. Segura et al. / Biochimica et Biophysica Acta 1828 (2013) 2015–2025

in the TMD of the T4CPs prevent self-interaction [14,27,53]. In partic-ular, data from TrwBΔN70 showed that in vitro oligomerization de-pends on its binding to G4-DNA quadruplex structure [14]. However,when in vivo oligomerization of full length TraD from F plasmid wasstudied it was concluded that DNA binding was not necessary [53].Previously reported structural work suggests that TMD of TrwBcould be the main responsible for oligomerization of TrwB [10,12]but this is the first time that interaction between the TMD alone hasbeen described. These results led us to propose that the initial andmain interaction for TrwB oligomerization could occur in the mem-brane between the helices present in the TMD, although interactionsthrough the cytosolic domain also occur. In fact, mutations in the nu-cleotide binding domain of TrwB, TraD of F plasmid or TcpA of pCW3plasmid did hinder monomer–monomer interactions [25,53,54],

suggesting that subsequent interactions in the cytosol would stabilizethe oligomers.

Previous studies on wild type proteins have reported TrwB–TrwEinteractions [26–28,31], as well as interactions between TrwBand the VirB10 homologs of A. tumefaciens, Bartonella tribocorumor Brucella suis T4SSs [24,55]. Our experiments using BTH andco-immunoprecipitation techniques provided new informationabout the role of the different domains of TrwB and TrwE in theirinteractions (Figs. 4B and 7). In particular, it is the first time thatinteractions between TrwBΔN6-TrwE, TrwB–TrwEN91, TrwB–TrwEΔN42 and especially, TrwBTM–TrwE have been reported. They sug-gest that there are interactions in the membrane or in its periphery.In agreement with our findings, in previous studies no interactionwas detected when a TrwE mutant protein lacking the first 64

2024 R.L. Segura et al. / Biochimica et Biophysica Acta 1828 (2013) 2015–2025

amino acid residues (i.e., the small cytosolic portion and the TMD)was used [27]. Same behavior was observed for AgrobacteriumVirB10 [28], suggesting that the transmembrane portion is neededfor these interactions. Interestingly, mutant proteins TrwEΔN42and TrwEN91 tested in our work further narrow the interactionregion between TrwB and TrwE, indicating that there are interac-tions in the membrane in the absence of either cytosolic or periplas-mic domains of TrwE, including its PRR motif, and also when thecytosolic domain of TrwB was removed (TrwBTM–TrwE) (Figs. 4Band 7). In this regard, interactions between T4CPs and VirB10-likeproteins from different T4SSs appear to be slightly different. Studieswith R27 plasmid [26,31] and A. tumefaciens [28] reported that theperiplasmic segment and the PRR of VirB10-like protein were essen-tial for TraG–TrhB and VirB10–VirD4 interactions. Similarly, interac-tion studies between different TrwB cytosolic mutant proteins andTrwE or VirB10-homologs from Bartonella tribocorum suggestedthat the soluble domain of TrwB was not involved in TrwB–TrwEand TrwB–TrwEBt interactions [25].

Our location studies of TrwE confirmed that this protein is mainlyanchored to the inner membrane of E. coli in absence of other R388conjugative proteins as predicted for other VirB10 homologs [28,56].Conjugation assays demonstrated that the periplasmic domain wasessential for TrwE biological activity, in agreement with the resultsobserved for VirB10 [28]. Nevertheless, when the N-terminal portion(the cytosolic segment and the transmembrane helix) of TrwE wasremoved, the conjugation capacity diminished but it was notcompletely abolished. These results suggest that the membrane an-chorage of TrwE is important for the correct formation of the T4SSand for the interaction with the coupling protein as described byother authors [5,27,28]. However, despite the lack of membrane an-chorage its transfer capacity is partially maintained, suggesting thatthe membrane anchorage could be overcome by interactions of theperiplasmic domain of TrwE with other proteins that form the T4SS.Recently, structural studies on plasmid pKM101 divided the T4SScore complex in two layers: the inner (I) layer that is composed ofthe N-terminal of TraF and TraO and the outer (O) layer that is mainlycomposed of the C-terminal domain of TraF and TraO and the wholeprotein TraN [6]. This structure could be similar to the T4SS corecomplex of the R388 plasmid as shown by the fact that when theN-terminal of TrwE is removed (TrwEΔN64) the plasmid did notcompletely lose the conjugation capacity (Table 3), indicating thatthe periplasmic domain of TrwE could maintain the necessary inter-actions with TrwH and TrwF proteins to build the T4SS core complex,as it was observed for TraF from pKM101 [6]. Similarly, studies withVirB10 from A. tumefaciens, demonstrated that a mutant proteinwithout cytoplasmic and TMDs formed a precipitable complex withVirB7–VirB9 although it failed to interact with VirD4 [28]. Studieson VirB10-like proteins point to the periplasmic domain as respon-sible for self-interactions [26,55]. Our results showed that therealso exist important contacts that involve the N-terminal domainof TrwE in absence of the periplasmic domain (TrwE–TrwEN91interactions).

Our previous studies reported the effect of the TMD of TrwB on itsin vitro activity and stability [18,19,21,23]. In this work the importantrole of this domain in transfer capacity, self-interaction, and interac-tion with other proteins of the secretion apparatus (i.e., TrwE) hasbeen described. This study has determined that the TMD of TrwB isnecessary for its biological activity and for TrwB self-interactions.Moreover, we found that the TMDs of both TrwB and TrwE proteinsare involved in TrwB–TrwE interactions. Our data suggest that,although the conjugation capacity of R388 is not completely lostwhen TMD of TrwE is removed, TMD does take part in TrwEself-interactions. In conclusion, this work underscores the role ofthe membrane and the TMD of TrwB and TrwE in the T4SS function.

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.bbamem.2013.05.022.

Acknowledgements

This workwas supported with funds from the SpanishMinisterio deEducación y Ciencia (grant no. BFU2007-62062), from MICINN (grantno. BFU2010-22103) and from LSHM-CT-2005_019023 (European VIFramework Program). R.L.S. was a postdoctoral scientist supportedby CSIC I3P and Basque Government postdoctoral fellowships. A.J.V.was a postdoctoral scientist supported by the University of the BasqueCountry. S.A-A. was a pre-doctoral student supported by the BasqueGovernment. B.U-U.was supported by aUniversity of the BasqueCountrypre-doctoral fellowship. Work in the F.C. laboratory was supportedby grant 282004/FP7-HEALTH.2011.2.3.1-2 (European VII FrameworkProgram).

References

[1] R. Fronzes, P.J. Christie, G. Waksman, The structural biology of type IV secretionsystems, Nat. Rev. Microbiol. 7 (2009) 703–714.

[2] F. de la Cruz, L.S. Frost, R.J. Meyer, E.L. Zechner, Conjugative DNA metabolism inGram-negative bacteria, FEMS Microbiol. Rev. 34 (2010) 18–40.

[3] C.E. Alvarez-Martinez, P.J. Christie, Biological diversity of prokaryotic type IVsecretion systems, Microbiol. Mol. Biol. Rev. 73 (2009) 775–808.

[4] C. Smillie, M.P. Garcillan-Barcia, M.V. Francia, E.P. Rocha, F. de la Cruz, Mobility ofplasmids, Microbiol. Mol. Biol. Rev. 74 (2010) 434–452.

[5] V. Chandran, R. Fronzes, S. Duquerroy, N. Cronin, J. Navaza, G.Waksman, Structure ofthe outer membrane complex of a type IV secretion system, Nature 462 (2009)1011–1015.

[6] R. Fronzes, E. Schäfer, L. Wang, H.R. Saibil, E.V. Orlova, G. Waksman, Structure of atype IV secretion system core complex, Science 323 (2009) 266–268.

[7] G. Schröder, E. Lanka, The mating pair formation system of conjugative plasmids—a versatile secretion machinery for transfer of proteins and DNA, Plasmid 54(2005) 1–25.

[8] R. Bayliss, R. Harris, L. Coutte, A. Monier, R. Fronzes, P.J. Christie, P.C. Driscoll, G.Waksman, NMR structure of a complex between the VirB9/VirB7 interactiondomains of the pKM101 type IV secretion system, Proc. Natl. Acad. Sci. U. S. A.104 (2007) 1673–1678.

[9] L. Terradot, R. Bayliss, C. Oomen, G.A. Leonard, C. Baron, G. Waksman, Structuresof two core subunits of the bacterial type IV secretion system, VirB8 from Brucellasuis and ComB10 from Helicobacter pylori, Proc. Natl. Acad. Sci. U. S. A. 102 (2005)4596–4601.

[10] F.X. Gomis-Rüth, G. Moncalián, R. Pérez-Luque, A. González, E. Cabezón, F. de laCruz, M. Coll, The bacterial conjugation protein TrwB resembles ring helicasesand F1-ATPase, Nature 409 (2001) 637–641.

[11] E. Cabezon, F. de la Cruz, TrwB: an F(1)-ATPase-like molecular motor involved inDNA transport during bacterial conjugation, Res. Microbiol. 157 (2006) 299–305.

[12] I. Hormaeche, I. Alkorta, F. Moro, J.M. Valpuesta, F.M. Goñi, F. de la Cruz, Purificationand properties of TrwB, a hexameric, ATP-binding integralmembrane protein essen-tial for R388 plasmid conjugation, J. Biol. Chem. 277 (2002) 46456–46462.

[13] M. Llosa, S. Bolland, F. de la Cruz, Genetic organization of the conjugal DNAprocessing region of the IncW plasmid R388, J. Mol. Biol. 235 (1994) 448–464.

[14] I. Matilla, C. Alfonso, G. Rivas, E.L. Bolt, F. de la Cruz, E. Cabezon, The conjugativeDNA translocase TrwB is a structure-specific DNA-binding protein, J. Biol. Chem.285 (2010) 17537–17544.

[15] G. Moncalián, E. Cabezón, I. Alkorta, M. Valle, F. Moro, J.M. Valpuesta, F.M. Goñi, F.de la Cruz, Characterization of ATP and DNA binding activities of TrwB, thecoupling protein essential in plasmid R388 conjugation, J. Biol. Chem. 274(1999) 36117–36124.

[16] I. Tato, S. Zunzunegui, F. de la Cruz, E. Cabezon, TrwB, the coupling proteininvolved in DNA transport during bacterial conjugation, is a DNA-dependentATPase, Proc. Natl. Acad. Sci. U. S. A. 102 (2005) 8156–8161.

[17] I. Tato, I. Matilla, I. Arechaga, S. Zunzunegui, F. de la Cruz, E. Cabezon, The ATPaseactivity of the DNA transporter TrwB is modulated by protein TrwA: implicationsfor a common assembly mechanism of DNA translocating motors, J. Biol. Chem.282 (2007) 25569–25576.

[18] I. Hormaeche, I. Iloro, J.L. Arrondo, F.M. Goñi, F. de la Cruz, I. Alkorta, Role of thetransmembrane domain in the stability of TrwB, an integral protein involved inbacterial conjugation, J. Biol. Chem. 279 (2004) 10955–10961.

[19] I. Hormaeche, R.L. Segura, A.J. Vecino, F.M. Goñi, F. de la Cruz, I. Alkorta, The trans-membrane domain provides nucleotide binding specificity to the bacterial conju-gation protein TrwB, FEBS Lett. 580 (2006) 3075–3082.

[20] A.J. Vecino, R.L. Segura, B. Ugarte-Uribe, S. Aguila, I. Hormaeche, F. de la Cruz, F.M.Goñi, I. Alkorta, Reconstitution in liposome bilayers enhances nucleotide bindingaffinity and ATP-specificity of TrwB conjugative coupling protein, Biochim.Biophys. Acta 1798 (2010) 2160–2169.

[21] A.J. Vecino, I. de la Arada, R.L. Segura, F. de la Cruz, F.M. Goñi, I. Alkorta, Membraneinsertion stabilizes the structure of TrwB the R388 conjugative plasmid couplingprotein, Biochim. Biophys. Acta 1808 (2011) 1032–1039.

[22] F.X. Gomis-Rüth, G. Moncalián, F. de la Cruz, M. Coll, Conjugative plasmid proteinTrwB, an integral membrane type IV secretion system coupling protein. Detailedstructural features and mapping of the active site cleft, J. Biol. Chem. 277 (2002)7556–7566.

2025R.L. Segura et al. / Biochimica et Biophysica Acta 1828 (2013) 2015–2025

[23] A.J. Vecino, R.L. Segura, I. de la Arada, F. de la Cruz, F.M. Goñi, J.L.R. Arrondo, I.Alkorta, Deletion of a single helix from the transmembrane domain causes largechanges in membrane insertion properties and secondary structure of the bacterialconjugation protein TrwB, Biochim. Biophys. Acta 1818 (2012) 3158–3166.

[24] H.D. de Paz, F.J. Sangari, S. Bolland, J.M. García-Lobo, C. Dehio, F. de la Cruz, M.Llosa, Functional interactions between type IV secretion systems involved inDNA transfer and virulence, Microbiology 151 (2005) 3505–3516.

[25] H.D. de Paz, D. Larrea, S. Zunzunegui, C. Dehio, F. de la Cruz, M. Llosa, Functionaldissection of the conjugative coupling protein TrwB, J. Bacteriol. 192 (2010)2655–2669.

[26] M.W. Gilmour, J.E. Gunton, T.D. Lawley, D.E. Taylor, Interaction between the IncHI1plasmid R27 coupling protein and type IV secretion system: TraG associates with thecoiled-coil mating pair formation protein TrhB, Mol. Microbiol. 49 (2003) 105–116.

[27] M. Llosa, S. Zunzunegui, F. de la Cruz, Conjugative coupling proteins interact withcognate and heterologous VirB10-like proteins while exhibiting specificity forcognate relaxosomes, Proc. Natl. Acad. Sci. U. S. A. 100 (2003) 10465–10470.

[28] S.J. Jakubowski, J.E. Kerr, I. Garza, V. Krishnamoorthy, R. Bayliss, G. Waksman, P.J.Christie, Agrobacterium VirB10 domain requirements for type IV secretion and Tpilus biogenesis, Mol. Microbiol. 71 (2009) 779–794.

[29] E. Cascales, P.J. Christie, Agrobacterium VirB10, an ATP energy sensor required fortype IV secretion, Proc. Natl. Acad. Sci. U. S. A. 101 (2004) 17228–17233.

[30] P. Chiang, M. Habash, L.L. Burrows, Disparate subcellular localization patternsof Pseudomonas aeruginosa Type IV pilus ATPases involved in twitching motility,J. Bacteriol. 187 (2005) 829–839.

[31] J.E. Gunton, M.W. Gilmour, G. Alonso, D.E. Taylor, Subcellular localization andfunctional domains of the coupling protein, TraG, from IncHI1 plasmid R27,Microbiol. 151 (2005) 3549–3561.

[32] C. Guynet, A. Cuevas, G. Moncalián, F. de la Cruz, The stb Operon balances therequirements for vegetative stability and conjugative transfer of plasmid R388,PLoS Genet. 7 (2011) e1002073.

[33] P.K. Judd, R.B. Kumar, A. Das, The type IV secretion apparatus protein VirB6 ofAgrobacterium tumefaciens localizes to a cell pole, Mol. Microbiol. 55 (2005) 115–124.

[34] R.B. Kumar, Y.H. Xie, A. Das, Subcellular localization of the Agrobacteriumtumefaciens T-DNA transport pore proteins: VirB8 is essential for the assemblyof the transport pore, Mol. Microbiol. 36 (2000) 608–617.

[35] R.B. Kumar, A. Das, Polar location and functional domains of the Agrobacteriumtumefaciens DNA transfer protein VirD4, Mol. Microbiol. 43 (2002) 1523–1532.

[36] S.G. Grant, J. Jessee, F.R. Bloom, D. Hanahan, Differential plasmid rescue fromtransgenic mouse DNAs into Escherichia coli methylation–restriction mutants,Proc. Natl. Acad. Sci. U. S. A. 87 (1990) 4645–4649.

[37] B. Miroux, J.E. Walker, Over-production of proteins in Escherichia coli: mutanthosts that allow synthesis of some membrane proteins and globular proteins athigh levels, J. Mol. Biol. 260 (1996) 289–298.

[38] F. de la Cruz, J. Grinsted, Genetic and molecular characterization of Tn21, a multipleresistance transposon from R100.1, J. Bacteriol. 151 (1982) 222–228.

[39] G. Karimova, J. Pidoux, A. Ullmann, D.A. Ladant, Bacterial two-hybrid systembased on a reconstituted signal transduction pathway, Proc. Natl. Acad. Sci. U. S. A.95 (1998) 5752–5756.

[40] G. Karimova, A. Ullmann, D. Ladant, A bacterial two-hybrid system that exploitsa cAMP signaling cascade in Escherichia coli, Methods Enzymol. 328 (2000)59–73.

[41] G. Karimova, N. Dautin, D. Ladant, Interaction network among Escherichia colimembrane proteins involved in cell division as revealed by bacterial two-hybridanalysis, J. Bacteriol. 187 (2005) 2233–2243.

[42] F.W. Studier, B.A. Moffatt, Use of bacteriophage T7 RNA polymerase to directselective high-level expression of cloned genes, J. Mol. Biol. 189 (1986) 113–130.

[43] N. Datta, R.W. Hedges, Trimethoprim resistance conferred by W plasmids inEnterobacteriaceae, J. Gen. Microbiol. 72 (1972) 349–355.

[44] J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular Cloning: a Laboratory Manual,Cold Spring Harbor Lab. Press, Plainview, New York, 1989.

[45] R.A. De Maagd, B. Lugtenberg, Fractionation of Rhizobium leguminosarum cellsinto outer membrane, cytoplasmic membrane, periplasmic, and cytoplasmiccomponents, J. Bacteriol. 167 (1986) 1083–1085.

[46] M.J. Osborn, J.E. Gander, E. Parisi, J. Carson, Mechanism of assembly of the outermembrane of Salmonella typhimurium, J. Biol. Chem. 247 (1972) 3962–3972.

[47] A. Weissbach, J. Hurwitz, The formation of 2% keto-3 deoxyheptonic acid inextracts of Escherichia coli B, J. Biol. Chem. 234 (1959) 705–709.

[48] G. Schröder, S. Krause, E.L. Zechner, B. Traxler, H.J. Yeo, R. Lurz, G. Waksman, E.Lanka, TraG-like proteins of DNA transfer systems and of the Helicobacter pyloritype IV secretion system: inner membrane gate for exported substrates?J. Bacteriol. 184 (2002) 2767–2779.

[49] S. Mihajlovic, S. Lang, M.V. Sut, H. Strohmaier, C.J. Grube, G. Koraimann, E.Cabezón, G. Moncalián, F. de la Cruz, E.L. Zechner, Plasmid r1 conjugative DNAprocessing is regulated at the coupling protein interface, J. Bacteriol. 191 (2009)6877–6887.

[50] M.V. Sut, S. Mihajlovic, S. Lang, C.J. Gruber, E.L. Zechner, Protein and DNA effec-tors control the TraI conjugative helicase of plasmid R1, J. Bacteriol. 191 (2009)6888–6899.

[51] I. Mingarro, G. von Heijne, P. Whitley, Membrane–protein engineering, TrendsBiotechnol. 10 (1997) 432–437.

[52] G. von Heijne, Membrane proteins: from sequence to structure, Annu. Rev.Biophys. Biomol. Struct. 23 (1994) 167–192.

[53] R.J. Haft, E.G. Gachelet, T. Nguyen, L. Toussaint, D. Chivian, B. Traxler, In vivooligomerization of the F conjugative coupling protein TraD, J. Bacteriol. 189 (2007)6626–6634.

[54] J.A. Steen, T.L. Bannam, W.L. Teng, R.J. Devenish, J.I. Rood, The putative couplingprotein TcpA interacts with other pCW3-encoded proteins to form an essentialpart of the conjugation complex, J. Bacteriol. 191 (2009) 2926–2933.

[55] Z. Ding, Z. Zhao, S.J. Jakubowski, A. Krishnamohan, W. Margolin, P.J. Christie, Anovel cytology-based, two-hybrid screen for bacteria applied to protein–proteininteraction studies of a type IV secretion system, J. Bacteriol. 184 (2002)5572–5582.

[56] Z. Liu, A.N. Binns, Functional subsets of the virB type IV transport complexproteins involved in the capacity of Agrobacterium tumefaciens to serve as a recipientin virB-mediated conjugal transfer of plasmid RSF1010, J. Bacteriol. 185 (2003)3259–3269.