Báo cáo khoa học: DtpB (YhiP) and DtpA (TppB, YdgR) are prototypical proton-dependent peptide transporters of Escherichia coli

DtpB (YhiP) and DtpA (TppB, YdgR) are prototypical proton-dependent peptide transporters of Escherichia coli Daniel Harder1, Ju¨ rgen Stolz1, Fabio Casagrande2, Petr Obrdlik3, Dietmar Weitz1, Dimitrios Fotiadis2,* and Hannelore Daniel1

1 Molecular Nutrition Unit, Technical University of Munich, Freising, Germany 2 M.E. Mu¨ ller Institute for Structural Biology, Biozentrum, University of Basel, Switzerland 3 IonGate Biosciences GmbH, Frankfurt, Germany

Keywords E. coli; peptide; PTR; TppB; transport

Correspondence H. Daniel, Molecular Nutrition Unit, Technical University of Munich, Am Forum 5, D-85350 Freising, Germany Fax: +49 8161713999 Tel: +49 8161713400 E-mail: daniel@wzw.tum.de

*Present address Institute of Biochemistry and Molecular Medicine, University of Berne, Switzerland

(Received 18 February 2008, revised 20 March 2008, accepted 23 April 2008)

doi:10.1111/j.1742-4658.2008.06477.x

The genome of Escherichia coli contains four genes assigned to the peptide transporter (PTR) family. Of these, only tppB (ydgR) has been character- ized, and named tripeptide permease, whereas protein functions encoded by the yhiP, ybgH and yjdL genes have remained unknown. Here we describe the overexpression of yhiP as a His-tagged fusion protein in E. coli and show saturable transport of glycyl-sarcosine (Gly-Sar) with an apparent affinity constant of 6.5 mm. Overexpression of the gene also increased the susceptibility of cells to the toxic dipeptide alafosfalin. Transport was strongly decreased in the presence of a protonophore but unaffected by sodium depletion, suggesting H+-dependence. This was confirmed by puri- fication of YhiP and TppB by nickel affinity chromatography and reconsti- tution into liposomes. Both transporters showed Gly-Sar influx in the presence of an artificial proton gradient and generated transport currents on a chip-based sensor. Competition experiments established that YhiP transported dipeptides and tripeptides. Western blot analysis revealed an apparent mass of YhiP of 40 kDa. Taken together, these findings show that yhiP encodes a protein that mediates proton-dependent electrogenic transport of dipeptides and tripeptides with similarities to mammalian PEPT1. On the basis of our results, we propose to rename YhiP as DtpB (dipeptide and tripeptide permease B), by analogy with the nomenclature in other bacteria. We also propose to rename TppB as DtpA, to better describe its function as the first protein of the PTR family characterized in E. coli.

cells

into Escherichia coli

Uptake of peptides is thought to be mediated by three different transport systems represented by the dipeptide permease Dpp, the tripeptide permease TppB, and the oligopeptide permease Opp [1–3]. Despite the fact that these pro- teins seem to discriminate according to the backbone length of peptide substrates, they do show some over- lapping specificity. Although their prime physiological role is in the uptake of peptide-bound amino acids as

an economic process to provide energy substrates and building blocks for cellular metabolism, peptide uptake seems also to be involved in signaling processes and metabolic adaptation [4]. Whereas Opp and Dpp rep- resent ATP-binding cassette transporters with periplas- matic binding proteins [5–7], TppB belongs to the family of proton-dependent peptide symporters that utilize the proton gradient as driving force and lack cognate binding proteins [8]. After identification of the

Abbreviations AMCA, b-Ala-Lys-Ne-7-amino-4-methylcoumarin-3-acetic acid; CCCP, carbonyl cyanide m-chlorophenylhydrazone; DDM, n-dodecyl-b-D- maltoside; Gly-Sar, glycyl-sarcosine; IPTG, isopropyl-thio-b-D-galactoside; TMPD, N,N,N¢,N¢-tetramethyl-p-phenylenediamine.

FEBS Journal 275 (2008) 3290–3298 ª 2008 The Authors Journal compilation ª 2008 FEBS

3290

D. Harder et al.

Proton-dependent peptide transporters of E. coli

of

expression

(also

called

proton-dependent

carrying

gene-specific primers. The gene was cloned into the pET-21 vector fused with a C-terminal hexahistidine- dtpB in E. coli tag. The level BL21(DE3)pLysS was assayed by western blot analysis with an antibody against the hexahistidine-tag. This antibody detected a protein with an apparent molecu- lar mass of about 40 kDa in membrane preparations of dtpB-overexpressing cells (Fig. 1A, upper panel, lane 2). As no signal was obtained with isopropyl-thio- b-d-galactoside an (IPTG)-induced cells empty vector (Fig. 1A, lane 1), this band clearly repre- sented DtpB, which was also confirmed by the signal of the nickel affinity chromatography-purified DtpB (Fig. 1A, lane 3). The expected molecular mass based on the amino acid sequence of DtpB is 54.6 kDa. A similar increased mobility in SDS ⁄ PAGE was observed for DtpA, where we provided evidence that the protein was complete by probing it with a second antibody that recognized an added N-terminal tag [17].

interest

corresponding gene ydgR [8,9], it was genetically classi- fied as a member of the peptide transport (PTR) fam- ily oligopeptide transporter or POT family). These transporters are found essentially in all living organisms from bacteria to humans, with examples such as DtpT from Lacto- coccus lactis, Ptr2 from Saccharomyces cerevisiae, and the mammalian transporters PEPT1 and PEPT2 [10,11]. The first attempts to functionally characterize TppB in Salmonella typhimurium [12,13] and in E. coli, employing deletion mutants [14,15], provided some information on substrate preferences, and demon- strated upregulation of transport upon anaerobiosis like Dpp, was and by Leu [13,16]. Although TppB, shown to transport dipeptides and tripeptides, it seemed to prefer tripeptides, and in particular those of a hydrophobic nature. The transport mode of TppB, however, was poorly characterized. We recently cloned ydgR encoding TppB, and overexpressed the gene for a detailed functional analysis of the encoded trans- porter and its initial structural characterization. We observed a striking functional similarity between YdgR and the mammalian PEPT1 protein [17], and therefore the bacterial PTR transporters may serve as models for elucidating the structure. They are easily purified in large quantities for crystallization approaches to derive appropriate structural models. These might then be applied also to the human proteins, which are also of great for pharmacalogical applications, e.g. delivery of peptidomimetic drugs in the intestine.

that

(200 lgÆmL)1,

Three of the four E. coli members of the PTR family (yhiP, ybgH and yjdL) have not been studied yet. We here report the cloning of yhiP and the purification and characterization of the corresponding transport protein. Furthermore, we compared the functional fea- tures of YhiP with that of TppB to improve our understanding of peptide transport processes in bacte- ria. We established that tppB and yhiP code for proto- typical H+-coupled symporter proteins are specific for dipeptides and tripeptides. By analogy to similar transporters in other prokaryotes, we propose to name them Dtp (dipeptide and tripeptide permeas- es), with the first one identified being DtpA (former TppB, YdgR), and DtpB (YhiP) being described here as the second member of the PTR family in E. coli.

Results

Functional overexpression of dtpB in E. coli

Using a similar approach as previously reported for dtpA (tppB, ydgR) [17], we amplified the dtpB (yhiP) gene of E. coli from genomic DNA by PCR using

To assess the function of DtpB, we determined the uptake of radiolabeled [14C]glycyl-sarcosine (Gly-Sar), a commonly used reporter substrate for mammalian peptide transporters (Fig. 1B). Cells overexpressing dtpB showed a higher uptake of Gly-Sar than control cells. This was also the case for DtpA, indicating that both proteins mediate efficient Gly-Sar uptake. As a second approach to assess the functionality of the transport protein, we performed growth experiments (Fig. 1C) with the toxic phosphonopeptide alafosfalin, a known substrate of DtpA. E. coli cells carrying the expression vector for dtpB or the empty vector were pregrown in the presence of a nonlethal concentration of alafosfalin (500 lgÆmL)1, 2.55 mm), and gene expression was induced by addition of IPTG. Cells overexpressing dtpB showed strong growth inhibition 1 h after induction. The fact that cells carrying the empty vector, noninduced cells or induced cells with- out alafosfalin showed no growth inhibition demon- strated the specificity of DtpB-mediated uptake of alafosfalin. In similar experiments with DtpA, lower 1.02 mm) alafosfalin concentrations were needed for growth inhibition [17]. To directly compare the substrate affinities of DtpB and DtpA, we determined the uptake rates in cells overexpressing the corresponding genes as a function of Gly-Sar concen- tration (Fig. 1D). Transport was found to be saturable with an apparent Kt of 6.5 mm for DtpB and 1 mm for DtpA. We next investigated whether transport requires Na+ or H+ (Fig. 1E). When we replaced Na+ in the buffer by choline, this had only a minor effect on uptake of Gly-Sar, whereas the presence of the proton ionophore carbonyl cyanide m-chloro- phenylhydrazone (CCCP) caused complete inhibition

FEBS Journal 275 (2008) 3290–3298 ª 2008 The Authors Journal compilation ª 2008 FEBS

3291

D. Harder et al.

Proton-dependent peptide transporters of E. coli

A

B

C

D

E

Fig. 1. Overexpression of dtpB in E. coli. (A) Anti-His-tag western blot (upper panel) and Coomassie-stained (lower panel) SDS gel (12.5% acrylamide) of E. coli membranes of control cells (lane 1, 50 lg of total protein) or membranes containing overproduced DtpB (lane 2, 50 lg of total protein) or purified DtpB (0.5 lg upper panel, 1.5 lg lower panel). (B) Uptake of [14C]Gly-Sar (1 mM) by E. coli cells carrying the expression plasmid for dtpA or dtpB or the empty vector 3 h after induction with IPTG. Overexpression of either gene causes increased Gly-Sar transport relative to the control. n = 2, ± SD. (C) Growth inhibition by alafosfalin. Induced (+ IPTG) or not induced cells, carrying either the expression plasmid for dtpB or an empty plasmid, were grown in the presence of 500 lgÆmL)1 (2.55 mM) alafosfalin (+ a) or with- out alafosfalin. n = 2, ± SD. (D) Saturation kinetics for DtpA and DtpB. Uptake assays were performed with E. coli cells overexpressing dtpA or dtpB, in the presence of various concentrations of Gly-Sar. The uptake velocities were determined using four time points (10, 20, 40 and 60 s). Kt values were determined by nonlinear regression analysis of the presented data. n ‡ 2, ± SD. (E) Sodium and proton dependence of Gly-Sar uptake. For E. coli cells overexpressing dtpB, Gly-Sar (50 lM) uptake rates were determined from a 10 min time course. The rates were determined in buffers containing 150 mM NaCl (+ Na+), or cholinechloride (150 mM) or in the presence of NaCl and the protonophore CCCP (10 lM); 100% corresponds to 3.3 pmolÆs)1ÆD)1. n = 2, ± SD.

of transport. This suggests that uptake via DtpB depends on the proton-motive force, and that DtpB may function as a proton–peptide cotransporter.

Substrate specificity of DtpB

for DtpA,

determined with

To assess the substrate specificity of DtpB (and DtpA, for comparison), we performed competition experi- ments with 1 mm Gly-Sar as a substrate and 10 mm competitors in E. coli cells overexpressing either dtpB or dtpA. The uptake rates in the presence of the com- petitor are presented in Table 1. The substrate specific- ity data presented here for DtpA essentially match b-Ala-Lys[b-Ala-Lys-Ne-7- those amino-4-methylcoumarin-3-acetic acid(AMCA)], a fluo- rescent dipeptide analog used as a substrate in our previous study [17]. Similar to DtpA, neither the amino acid Ala nor the tetrapeptide Ala-Ala-Ala-Ala caused significant inhibition of DtpB-mediated trans-

port. In the case of DtpB, essentially all tested compet- itors reduced transport to a lower extent than with DtpA, which reflects its generally lower affinity. Despite this, there were differences between DtpB and DtpA when inhibitory effects of individual peptides were compared. In the case of DtpB, peptides contain- ing d-stereoisomers of l-Ala essentially failed to inhibit uptake, whereas strong inhibition was observed with d-Ala when it was placed in an N-termi- nal position. Virtually no competition was seen in both transporters with peptides containing only d-Ala. We also assessed the effects of charged amino acid residues and their spatial position on the uptake of Gly-Sar. In the case of DtpA, a positively charged amino acid at the N-terminus caused stronger inhibition than one at the C-terminus. A negatively charged residue was the C-terminus. This was also clearly preferred at observed with DtpB, but here inhibition was higher with Gly-Asp than in the case of DtpA. Selected b-lactam

FEBS Journal 275 (2008) 3290–3298 ª 2008 The Authors Journal compilation ª 2008 FEBS

3292

D. Harder et al.

Proton-dependent peptide transporters of E. coli

For DtpB,

100% = 50 pmolÆs)1ÆD)1. ND,

antibiotics (peptidomimetics) are known substrates of mammalian peptide transporters. DtpA showed efficient inhibition when an excess of cefadroxil, cefalexin or cephradine was added, whereas only cephradine and cefalexin significantly reduced Gly-Sar uptake via DtpB.

Table 1. Substrate specificity of DtpA and DtpB. Uptake of Gly- Sar in % ± standard deviation (n ‡ 2). Competitor was present at a concentration of 10 mM (10-fold excess) except otherwise indicated. The uptake time was 30 s. For DtpA, 100% = 200 pmolÆs)1ÆD)1. not determined.

[14C]Gly-Sar uptake (%)

Reconstitution of DtpB and DtpA

DtpB

DtpA

Competitor

100 ± 11 71 ± 12a 32 ± 6a 124 ± 20 17 ± 3a 17 ± 3a 110 ± 20 102 ± 3 88 ± 3 102 ± 3 99 ± 2 57 ± 12a 17 ± 1a 97 ± 5 101 ± 18 99 ± 19 57 ± 4a 78 ± 2a 90 ± 23

None Gly-Sar Gly-Sar (100 mM) Ala Ala-Ala Ala-Ala-Ala Ala-Ala-Ala-Ala L-Ala-D-Ala D-Ala-L-Ala D-Ala-D-Ala Gly-Lys Lys-Gly Gly-Asp Asp-Gly Cefamandole Cefuroxime Cephradine Cefalexin Cefadroxil

100 ± 19 27 ± 11a ND 76 ± 2 1 ± 0a 1 ± 0a 78 ± 22 63 ± 11a 5 ± 5a 84 ± 32 90 ± 40 2 ± 1a 26 ± 6a 54 ± 16a 145 ± 50 100 ± 2 29 ± 5a 35 ± 7a 19 ± 1a

a Significant (P < 0.05) reduction relative to assay containing no competitor.

inhibition by the protonophore Although transport CCCP (Fig. 1E) already suggested H+-dependence, we characterized the transport mode in more detail by reconstitution in a cell-free system. Membranes of E. coli containing the overproduced transport proteins were fused with proteoliposomes containing bovine cytochrome c oxidase. A transmembrane proton gradi- ent (inside negative) is generated when cytochrome c oxidase is supplied with electrons, which are trans- ferred from reduced ascorbate via N,N,N¢,N¢-tetram- ethyl-p-phenylenediamine (TMPD) and cytochrome c. Before addition of ascorbate, we observed a slow influx of Gly-Sar into all types of vesicles. Addition of ascorbate, however, caused a marked increase in Gly- Sar influx into vesicles containing either DtpA or DtpB, but not into control vesicles (Fig. 2A,B). Efflux of Gly-Sar was observed when the proton gradient was dissipated by addition of CCCP, indicating accumula- tion of substrate above the extravesicular concentra- tion in the presence of the proton gradient.

Fig. 2. In vitro transport studies of DtpB and DtpA. (A) Uptake of Gly-Sar (175 lM) in membrane vesicles generated by fusion of membranes from dtpB-overexpressing or control cells with cytochrome c oxidase-con- taining proteoliposomes. After 20 min (arrow), a proton gradient was established by addition of a mix of ascor- bate ⁄ TMPD ⁄ cytochrome c and destroyed by addition of 10 lM CCCP as indicated. (B) As (A), but with membranes of dtpA-over- expressing cells (one of two similar experi- ments). (C) Uptake of Gly-Sar (175 lM) in liposomes containing purified DtpB and cytochrome c oxidase. After 20 min (arrow), a proton gradient was established by addi- tion of a mix of ascorbate ⁄ TMPD ⁄ cyto- chrome c and destroyed by addition of 10 lM CCCP as indicated. n = 2, ± SD. (D) As (C), with purified DtpA instead of DtpB. n = 2, ± SD.

FEBS Journal 275 (2008) 3290–3298 ª 2008 The Authors Journal compilation ª 2008 FEBS

3293

D. Harder et al.

Proton-dependent peptide transporters of E. coli

(12.5% (1) (3) (4) Ni2+–nitrilotriacetic acid flow-

Fig. 3. Purification of DtpB. Coomassie-stained SDS gel acrylamide) of Ni2+–nitrilotriacetic acid purified DtpB. Lanes: (2) pellet after solubilization; DDM-solubilized membranes; supernatant after solubilization; through; (5) wash; (6) elution at about 150 mM imidazole.

Fig. 4. Electrophysiological experiment on proteoliposomes. The electrical response of liposome-reconstituted DtpB (dashed line) or DtpA (dotted line) to a change from a solution without substrate (Gly) to a solution with substrate (Gly-Gly). The solid line shows the recording from a sensor loaded with protein-free liposomes. The current peaks indicate proton cotransport, and the negative peak is back-flow of charges out of the liposomes (one of > 10 similar experiments using ‡ 2 sensors).

but not in control liposomes lacking the transport proteins. The current spikes at the changing of the buffer are artefacts of the working valves. The negative current peaks observed after changing back to the Gly- containing washing buffer most likely represent the backflow of charges (Gly-Gly with protons) out of the liposomes, as the substrate gradient is now reversed. As the transport studies were performed with Na+-free buffers at pH 7.0, the observed currents must originate from proton movement coupled to dipeptide transloca- tion in a symport mechanism. As in the cytochrome c oxidase-energized vesicles, DtpA caused higher signals than DtpB, suggesting that DtpA has a higher turnover rate or higher stability in vitro.

Discussion

Despite the fact that peptide uptake in E. coli has been studied for more than 20 years [1–3], essentially no biochemical characterization has been performed on any of the proteins. Another prokaryotic member of the PTR family, DtpT from L. lactis, has been studied in more detail, and some features of its membrane topology are known [20]. With the focus on the four E. coli genes that carry the typical PTR family motif [11], we first cloned and overexpressed dtpA, and then purified and characterized the corresponding protein (DtpA ⁄ TppB ⁄ YdgR) [17]. Here we provide a detailed functional analysis of DtpB as the second as yet

Next, we purified DtpB and DtpA from E. coli membranes. The membranes were solubilized with the detergent n-dodecyl-b-d-maltoside (DDM), and the proteins were purified by metal-affinity chromato- graphy, making use of the C-terminal His-tags. This procedure yielded about 2–3 mg of relatively pure pro- tein per liter of cell culture (Fig. 3). For reconstitution into liposomes, the detergent was removed by adsorp- tion to Bio-Beads. Subsequently, the liposomes contain- ing the peptide transporters were fused with vesicles containing cytochrome c oxidase, using a freeze–thaw– sonication cycle. Gly-Sar uptake after initiation of the proton gradient increased almost 10-fold in the case of DtpB and 60-fold in the case of DtpA (Fig. 2C,D), and substrate efflux was observed after addition of CCCP. To unequivocally demonstrate the electrogenic nature of the transport process, we next used a chip-based assay system, in which proteoliposomes containing DtpB or DtpA were adsorbed onto a gold surface of the SURFE2Rone setup (IonGate Biosciences GmbH, Frankfurt, Germany). This system is based on a solid- support membrane technology, and allows detection of capacity-coupled currents induced by movement of charged molecules across a lipid bilayer [18,19]. After adsorption of the proteoliposomes on the gold surface, the SURFE2Rone setup was used to let different buffers flow for a certain time over this surface. Transport was initiated by the exchange of a buffer containing Gly (Gly included for charge and osmotic equilibration only) against a buffer containing the dipeptide Gly-Gly. Figure 4 shows that Gly-Gly induced significant cur- rents in proteoliposomes containing DtpA or DtpB,

FEBS Journal 275 (2008) 3290–3298 ª 2008 The Authors Journal compilation ª 2008 FEBS

3294

D. Harder et al.

Proton-dependent peptide transporters of E. coli

uncharacterized PTR member in E. coli. To assess the differences between DtpA and DtpB and better under- stand why there are four putative peptide transporters in E. coli, we compared the transport features of DtpA and DtpB under identical experimental conditions.

Substrate specificity

of

as

determined

proteins [23,24]. DtpB, like DtpA, discriminates dipep- tides on the basis of the position of charges within the substrate. Peptides containing a positively charged side chain in an N-terminal position are good competitors, but the ability to compete is lost when the charge is present at the C-terminus. The reversed pattern was observed for negatively charged side chains, which were generally better tolerated at the C-terminal position. A similar charge preference is found in the mammalian peptide transporters [25,26]. Of the b-lactam antibiotics tested, only cephradine and cefalexin showed modest inhibition of Gly-Sar influx via DtpB, whereas for DtpA, the best inhibition was seen with cefadroxil, fol- lowed by cephradine and cefalexin. The mammalian peptide transporter PEPT1, which is expressed in the intestine, also clearly prefers those substrates but does not transport cefamandole or cefuroxime, which are considered to be inactive when administered orally [27]. Taken together, these findings show that DtpB has essentially the prototypical substrate characteristics of a dipeptide ⁄ tripeptide transporter of the PTR family. Despite its lower affinity for Gly-Sar and alafosfalin, it is similar to DtpA in the substrate recognition pattern, and also shows similarities to the mammalian trans- porters of the PEPT subgroup.

Transport mode

The dtpB gene has already been cloned and expressed in a global membrane proteome analysis approach in E. coli, where, by the use of green fluorescent protein and PhoA fusion constructs, the C-terminus was found to be localized in the cytoplasm [21]. We show here that DtpB overproduced in E. coli has an apparent molecular mass by 40 kDa SDS ⁄ PAGE, and can be purified in quantities suffi- cient for reconstitution experiments with conservation of function (Figs 1A and 3). Gly-Sar, a commonly used radiotracer substrate for mammalian peptide transporter proteins, showed saturation kinetics for uptake into cells overexpressing dtpB with a fairly low affinity (Kt = 6.5 mm), and transport was sensitive to a protonophore, suggesting proton coupling. A previ- ous study with E. coli mutants lacking the two major peptide transporters Opp and Dpp but containing DtpA could not detect Gly-Sar uptake [22]. As we clearly showed Gly-Sar uptake in cells overexpressing dtpA and in liposomes containing the purified protein (Figs 1B,D and 2D), this finding may result from low endogenous expression of dtpA in wild-type cells. The apparent affinity of DtpA for Gly-Sar (1 mm) is in the same range as for many other model dipeptides [17]. Alafosfalin, a known toxic substrate for bacterial pep- tide transporters, was reported to be specifically trans- ported by DtpA [12]. We show here that DtpB also confers alafosfalin toxicity in E. coli cells overpro- ducing the protein, but higher concentrations of ala- fosfalin are needed than for DtpA (2.55 mm instead of 1.02 mm) to obtain toxicity. This indicates a lower affinity for alafosfalin (as with Gly-Sar) or a lower expression of dtpB.

Inhibition of substrate influx via DtpB in E. coli cells in the presence of CCCP already suggested proton coupling. This was confirmed by reconstitution of the purified protein into proteoliposomes containing cyto- chrome c oxidase for energization, as previously shown for the L. lactis ortholog DtpT [28]. When an electro- chemical proton gradient was generated, uptake via DtpB as well as via DtpA proceeded until the gradient was collapsed by adding a protonophore. Moreover, capacity-coupled currents obtained with immobilized proteoliposomes on the SURFE2Rone chip show the electrogenic nature of the DtpB- and DtpA-mediated transport process, similar to on-chip recordings with membranes containing the mammalian PEPT trans- porters [29]. This unequivocally establishes that dipep- tide transport by the two bacterial transporters occurs by H+-symport.

Conclusions

DtpB is shown here to represent the second proton- dependent dipeptide and tripeptide transporter of E. coli, with functional features similar to those of other PTR family members, including those from mammals. Although DtpB appears to represent a sys-

The substrate specificity was addressed by competi- tion experiments with Gly-Sar as substrate (Table 1). The inhibition pattern observed for DtpA confirmed previous data obtained by using the fluorescent sub- strate b-Ala-Lys(AMCA) [17]. As DtpB failed to trans- port b-Ala-Lys(AMCA), we here employed Gly-Sar and compared the substrate specificities of both pro- teins under identical experimental conditions. DtpB, like DtpA, has a clear preference for dipeptides and tripeptides composed of l-amino acids. The chain length restriction and some other features match well with data for the mammalian PEPT1 and PEPT2

FEBS Journal 275 (2008) 3290–3298 ª 2008 The Authors Journal compilation ª 2008 FEBS

3295

D. Harder et al.

Proton-dependent peptide transporters of E. coli

Growth experiments

tem of lower affinity, there is an obvious redundancy in substrate specificity as compared with DtpA. This includes Gly-Sar, which we here clearly show is also a DtpA substrate. The question of why E. coli has two proteins with similar functional features will probably be resolved when regulation and expression of the corresponding genes under different environmental conditions are studied.

Cells were grown in LB to an attenuance (D600 nm) of 1, and transferred to a 12-well plate (1 mL per well); IPTG (0.1 mm) and alafosfalin (500 lgÆmL)1) were then added. Growth was continuously recorded by measuring D600 nm while shaking at 37 (cid:2)C in a plate reader (Varioscan, Thermo, Vantaa, Finnland). For conversion to standard D600 nm units, a calibration curve was used.

Experimental procedures

Purification of the His-tagged proteins

Cloning and expression of the transport genes in E. coli

The dtpB (yhiP) gene was cloned into the pET-21 vector (T7 promoter, C-terminal hexahistidine-tag, Novagen), using the primers 5¢-AAAAAGCTTATGAATACAACAA CACCCATG-3¢ and 5¢-AAACTCGAGATGGCTTTCCG GCGTCGC-3¢, as described previously for dtpA (ydgR) [17]. Expression was performed as previously described [17].

Western blot analysis

E. coli inner membranes were prepared for purification as described in [17]. From a sample corresponding to 250 lL of culture, the proteins were separated by SDS ⁄ PAGE and blotted onto poly(vinylidene difluoride) membranes (Milli- pore, Bedford, MA, USA) following detection with the Novagen His-tag antibody [17]. Cell pellets from 500 mL of culture were resuspended in 20 mL of lysis buffer (10 mm Hepes ⁄ Tris, pH 7.4, 1 mm dithiothreitol, 0.5 mm EDTA) and broken by sonication (10 cycles of 30 s). Membranes were prepared as previously in 10 mm described [17] and solubilized (60 min, 4 (cid:2)C) Hepes ⁄ Tris (pH 7.4), 150 mm NaCl, 1% DDM, 5% glyc- erol, 0.1 mm tris-2-carboxyethylphosphine and 30 mm imid- azole at protein concentrations of 1–3 mgÆmL)1. After centrifugation (40 000 g, 20 min), the supernatant was loaded onto an Ni2+–nitrilotriacetic acid column (HisTrap FF; GE) with an FPLC column (A¨ KTA; Amersham ⁄ GE Healthcare, Uppsala, Sweden) and washed with running buffer (10 mm Hepes ⁄ Tris, pH 7.4, 150 mm NaCl, 30 mm imidazole, 0.06% DDM, 5% glycerol, 0.1 mm tris-2-car- boxyethylphosphine). Protein was eluted with a gradient from 30 to 250 mm imidazole in running buffer. Elution yielded a sharp peak at about 150 mm imidazole, which was shown by SDS ⁄ PAGE and western blotting to contain the purified transporters.

Transport assays in intact cells

Reconstitution into proteoliposomes

For reconstitution, 250 lL of E. coli lipids at a concentration of 20 mgÆmL)1 in CHCl3 (total extract; Avanti Polar Lipids, Alabaster, AL, USA) were dried to a thin film under a stream of N2 and further incubated for 30 min in vacuum. The lipids were resuspended in 500 mL of buffer (50 mm KPO4, pH 6.3), sonicated until clearing (3 · 3 s, probe-type soni- fier), and destabilized with 0.5% DDM. One hundred micro- grams (about 100 lL) of purified protein as eluted from the Ni2+–nitrilotriacetic acid column was added. After 10 min on ice, the detergent was removed by adding 200 mg of Bio-Beads SM-2 (BioRad, Hercules, CA, USA), and this was followed by incubation for 4 h at 4 (cid:2)C. The detergent removal step was repeated three times with new Bio-Beads and incubation at 4 (cid:2)C for a total of 24 h.

Uptake with cytochrome c oxidase-energized vesicles

FEBS Journal 275 (2008) 3290–3298 ª 2008 The Authors Journal compilation ª 2008 FEBS

3296

Transport assays were performed 3 h after induction with IPTG with [14C]Gly-Sar (custom-synthesized by Biotrend, Cologne, Germany). Approximately 3 · 109 cells were har- vested by centrifugation (2500 g, 5 min), and resuspended in 1 mL of buffer (25 mm Hepes ⁄ Tris, pH 7.4, 150 mm NaCl, 5 mm glucose). The assay volume of 50 lL consisted of 20 lL of cells (6 · 107 cells), 5 lL of 10 mm Gly-Sar [14C]Gly-Sar, final concentration (containing 0.1 lCi of 1 mm, 2 mCiÆmmol)1), and 25 lL of buffer (control) or a 20 mm competitor solution (final concentration 10 mm). The cells were incubated with the substrate for 30 s at 37 (cid:2)C, and this was followed by filtration (0.45 lm mixed cellulose esters; ME25 Whatman, Dassel, Germany) and washing twice with ice-cold buffer. Uptake of [14C]Gly-Sar was quantified by liquid scintillation counting. Significance (P < 0.05) was determined by t-test using sigmaplot. For determination of apparent Kt values, uptake rates at differ- ent substrate concentrations were determined by linear regression of time course experiments (10, 20, 40 and 60 s). Then, Kt values and maximal uptake rate were determined by nonlinear regression of the rates versus concentration using sigmaplot. To generate a proton gradient, membrane vesicles or prote- oliposomes were fused with proteoliposomes containing cytochrome c oxidase [30,31]. Cytochrome c oxidase (1 mg,

D. Harder et al.

Proton-dependent peptide transporters of E. coli

GmbH for the possibility of using the SURFE2Rone setup. This work was supported by the Sixth Frame- work Programme of the European Union (EUGIN- DAT, European Genomics Initiative on Disorders of Plasma Membrane Amino Acid Transporters).

References

1 Payne JW & Smith MW (1994) Peptide transport by micro-organisms. Adv Microb Physiol 36, 1–80. sonifier. For 2 Higgins CF & Gibson MM (1986) Peptide transport in (2.5 lCi, 175 lm final bacteria. Methods Enzymol 125, 365–377.

3 Payne JW & Marshall NJ (2001) Peptide transport. In Microbial Transport Systems (Winkelmann G ed.), pp. 139–164. Wiley-VCH, Weinheim, Germany.

4 Detmers FJ, Lanfermeijer FC & Poolman B (2001) Pep- tides and ATP binding cassette peptide transporters. Res Microbiol 152, 245–258.

5 Abouhamad WN, Manson M, Gibson MM & Higgins CF (1991) Peptide transport and chemotaxis in Escheri- chia coli and Salmonella typhimurium: characterization of the dipeptide permease (Dpp) and the dipeptide- binding protein. Mol Microbiol 5, 1035–1047. [32]) was reconstituted with purified from bovine heart 40 mg of E. coli lipids and 12 mg of octyl-glucoside in 2 mL of buffer (50 mm KPO4, pH 6.3), and detergent was removed by dialysis. These proteoliposomes were mixed with E. coli membranes (usually 2 mg of lipid plus 100 lg of membrane protein) or with liposomes containing purified transporter (usually 27 lg of protein) in a final volume of 250 lL. After addition of 1 mm MgSO4, a cycle of freezing (liquid nitrogen) and thawing (room temperature) was per- formed, which was followed by sonication for 3 s with a the uptake assay, 2.5 lL of probe-type [14C]Gly-Sar concentration, 57 mCiÆmmol)1) was added to the proteoliposomes, which were stirred at 37 (cid:2)C. At the indicated times, 25 lL aliquots were withdrawn, diluted with 2 mL of 100 mm LiCl, and filtered (0.45 lm, mixed cellulose esters: ME25, Whatman). The fil- ters were washed with 2 mL of 100 mm LiCl, and this was followed by scintillation counting. Proton gradient genera- tion was started by the addition of 20 mm ascorbate, 200 lm TMPD and 20 lm cytochrome c (Sigma, St Louis, MO, USA). Fresh ascorbate was added (1 : 50) when the assay solution had turned blue, due to oxidized TMPD. CCCP (Sigma) was used at a final concentration of 10 lm to destroy the proton gradient.

6 Higgins CF (2001) ABC transporters: physiology, struc- ture and mechanism – an overview. Res Microbiol 152, 205–210.

Electrical measurements of reconstituted transporters with the SURFE2Rone setup

7 Hiles ID, Gallagher MP, Jamieson DJ & Higgins CF (1987) Molecular characterization of the oligopeptide permease of Salmonella typhimurium. J Mol Biol 195, 125–142. 8 Smith MW, Tyreman DR, Payne GM, Marshall NJ &

Payne JW (1999) Substrate specificity of the periplasmic dipeptide-binding protein from Escherichia coli: experi- mental basis for the design of peptide prodrugs. Micro- biology 145, 2891–2901. 9 Goh EB, Siino DF & Igo MM (2004) The Escherichia

coli tppB (ydgR) gene represents a new class of OmpR- regulated genes. J Bacteriol 186, 4019–4024.

10 Daniel H, Spanier B, Kottra G & Weitz D (2006) From bacteria to man: archaic proton-dependent peptide transporters at work. Physiology (Bethesda) 21, 93–102. 11 Steiner HY, Naider F & Becker JM (1995) The PTR

family: a new group of peptide transporters. Mol Micro- biol 16, 825–834. 12 Gibson MM, Price M & Higgins CF (1984) Genetic

characterization and molecular cloning of the tripeptide permease (tpp) genes of Salmonella typhimurium. J Bac- teriol 160, 122–130. The solid-supported membrane technology allows detection of capacitively coupled currents [18]. The currents are caused by the shift of electrical charges as the transporters go through the transport cycle, and originate from the movement of charged substrates, cotransported ions or pro- tein moieties carrying (partial) charges. Adsorption of the proteoliposomes to the gold surface sensors was performed as described previously [19]. Using the SURFE2Rone setup (IonGate Biosciences GmbH, Germany), the solution flow- ing over the adsorbed liposomes could be changed rapidly and the charging of the gold chip measured on-line. DtpB- and DtpA-mediated transport was activated when a buffer without substrate (30 mm Gly, 140 mm KCl, 25 mm Hepes, 25 mm Mes, 2 mm MgCl2, pH 7.0) was exchanged for the same buffer containing the substrate Gly-Gly (30 mm) instead of Gly. After a rapid fluid exchange to the Gly- Gly-containing solution, the charging of the proteolipo- somes on the sensor driven by the H+ ⁄ peptide symport was measured. The wash-out of dipeptide reversed the cur- rent flow, representing an electrogenic back-flux of peptide and proton along the now reversed substrate gradient.

Acknowledgements

13 Jamieson DJ & Higgins CF (1984) Anaerobic and leu- cine-dependent expression of a peptide transport gene in Salmonella typhimurium. J Bacteriol 160, 131–136. 14 Payne JW, Grail BM & Marshall NJ (2000) Molecular recognition templates of peptides: driving force for

We would like to thank Daniela Kolmeder for excel- technical assistance and IonGate BioSciences lent

FEBS Journal 275 (2008) 3290–3298 ª 2008 The Authors Journal compilation ª 2008 FEBS

3297

D. Harder et al.

Proton-dependent peptide transporters of E. coli

molecular evolution of peptide transporters. Biochem Biophys Res Commun 267, 283–289. 15 Payne JW, Grail BM, Gupta S, Ladbury JE, Marshall Thondorf I et al. (2006) The renal type H+ ⁄ peptide symporter PEPT2: structure–affinity relationships. Amino Acids 31, 137–156. 25 Daniel H, Morse EL & Adibi SA (1992) Determinants

NJ, O’Brien R & Payne GM (2000) Structural basis for recognition of dipeptides by peptide transporters. Arch Biochem Biophys 384, 9–23. of substrate affinity for the oligopeptide ⁄ H+ symporter in the renal brush border membrane. J Biol Chem 267, 9565–9573.

16 Jamieson DJ & Higgins CF (1986) Two genetically distinct pathways for transcriptional regulation of anaerobic gene expression in Salmonella typhimurium. J Bacteriol 168, 389–397.

26 Kottra G, Stamfort A & Daniel H (2002) PEPT1 as a paradigm for membrane carriers that mediate elec- trogenic bidirectional transport of anionic, cationic, and neutral substrates. J Biol Chem 277, 32683– 32691. 27 Bretschneider B, Brandsch M & Neubert R (1999)

17 Weitz D, Harder D, Casagrande F, Fotiadis D, Obrdlik P, Kelety B & Daniel H (2007) Functional and struc- tural characterization of a prokaryotic peptide trans- porter with features similar to mammalian PEPT1. J Biol Chem 282, 2832–2839. 18 Meyer-Lipp K, Ganea C, Pourcher T, Leblanc G & Intestinal transport of beta-lactam antibiotics: analysis of the affinity at the H+ ⁄ peptide symporter (PEPT1), the uptake into Caco-2 cell monolayers and the trans- epithelial flux. Pharm Res 16, 55–61. 28 Hagting A, Knol J, Hasemeier B, Streutker MR, Fang Fendler K (2004) Sugar binding induced charge translo- cation in the melibiose permease from Escherichia coli. Biochemistry 43, 12606–12613.

G, Poolman B & Konings WN (1997) Amplified expres- sion, purification and functional reconstitution of the dipeptide and tripeptide transport protein of Lactococ- cus lactis. Eur J Biochem 247, 581–587.

19 Zuber D, Krause R, Venturi M, Padan E, Bamberg E & Fendler K (2005) Kinetics of charge translocation in the passive downhill uptake mode of the Na+ ⁄ H+ antiporter NhaA of Escherichia coli. Biochim Biophys Acta 1709, 240–250.

29 Kelety B, Diekert K, Tobien J, Watzke N, Dorner W, Obrdlik P & Fendler K (2006) Transporter assays using solid supported membranes: a novel screening platform for drug discovery. Assay Drug Dev Technol 4, 575–582. 20 Hagting A, vd Velde J, Poolman B & Konings WN (1997) Membrane topology of the di- and tripeptide transport protein of Lactococcus lactis. Biochemistry 36, 6777–6785.

30 Driessen AJ, de Vrij W & Konings WN (1985) Incor- poration of beef heart cytochrome c oxidase as a pro- ton-motive force-generating mechanism in bacterial membrane vesicles. Proc Natl Acad Sci USA 82, 7555– 7559. 21 Daley DO, Rapp M, Granseth E, Melen K, Drew D & von Heijne G (2005) Global topology analysis of the Escherichia coli inner membrane proteome. Science 308, 1321–1323. 31 Stolz J, Stadler R, Opekarova M & Sauer N (1994)

22 Payne JW, Payne GM, Gupta S, Marshall NJ & Grail BM (2001) Conformational limitations of glycylsarco- sine as a prototypic substrate for peptide transporters. Biochim Biophys Acta 1514, 65–75.

Functional reconstitution of the solubilized Arabidopsis thaliana STP1 monosaccharide-H+ symporter in lipid vesicles and purification of the histidine tagged protein from transgenic Saccharomyces cerevisiae. Plant J 6, 225–233.

FEBS Journal 275 (2008) 3290–3298 ª 2008 The Authors Journal compilation ª 2008 FEBS

3298

23 Terada T, Sawada K, Irie M, Saito H, Hashimoto Y & Inui K (2000) Structural requirements for determining the substrate affinity of peptide transporters PEPT1 and PEPT2. Pflugers Arch 440, 679–684. 24 Biegel A, Knutter I, Hartrodt B, Gebauer S, Theis S, 32 Yu C, Yu L & King TE (1975) Studies on cytochrome oxidase. Interactions of the cytochrome oxidase protein with phospholipids and cytochrome c. J Biol Chem 250, 1383–1392. Luckner P, Kottra G, Rastetter M, Zebisch K,