Topological analysis of DctQ, the small integral membrane protein of the C4-dicarboxylate TRAP transporter of Rhodobacter capsulatus

Abstract

Tripartite ATP-independent periplasmic (‘TRAP’) transporters are a novel group of bacterial and archaeal secondary solute uptake systems which possess a periplasmic binding protein, but which are unrelated to ATP-binding cassette (ABC) systems. In addition to the binding protein, TRAP transporters contain two integral membrane proteins or domains, one of which is 40–50 kDa with 12 predicted transmembrane (TM) helices, thought to be the solute import protein, while the other is 20–30 kDa and of unknown function. Using a series of plasmid-encoded β-lactamase fusions, we have determined the topology of DctQ, the smaller integral membrane protein from the high-affinity C4-dicarboxylate transporter of Rhodobacter capsulatus, which to date is the most extensively characterised TRAP transporter. DctQ was predicted by several topology prediction programmes to have four TM helices with the N- and C-termini located in the cytoplasm. The levels of ampicillin resistance conferred by the fusions when expressed in Escherichia coli were found to correlate with this predicted topology. The data have provided a topological model which can be used to test hypotheses concerning the function of the different regions of DctQ and which can be applied to other members of the DctQ family.

TRAP transporter, C4-dicarboxylate, DctQ, Rhodobacter, Membrane protein topology

1 Introduction

The TRAP transporters are novel secondary solute transport systems that have been found to be widespread in both eubacteria and archaea [1,2]. Their major distinguishing feature is the involvement of a periplasmic binding protein in the transport process, in the absence of an ABC protein, and two additional membrane proteins or domains of unequal size. TRAP transporters are driven by an electrochemical ion gradient rather than ATP hydrolysis, and are thus completely distinct from ABC systems [1]. The best characterised TRAP transporter is the dctPQM-encoded high-affinity C4-dicarboxylate transport (Dct) system from the photosynthetic bacterium Rhodobacter capsulatus [1,3–5]. DctP is a 33 kDa periplasmic binding protein [5] with submicromolar K_D values for the binding of L-malate, fumarate and succinate [6]. DctM is a 45 kDa integral membrane protein predicted to contain 12 TM helices, in common with many membrane transport proteins, and is thought to catalyse substrate translocation across the cytoplasmic membrane [1]. The DctM family has weak sequence similarity to the ion transporter superfamily [2].

DctQ is a smaller integral membrane protein (26 kDa) which is essential for C4-dicarboxylate transport [1]. DctQ homologues are invariably present in TRAP transporters [2], either as discrete subunits or as domains fused to either of the other TRAP proteins, but so far have not been found in any other type of transport system. The DctQ family [2] is the most divergent of the TRAP transporter proteins in terms of sequence similarity. The function of DctQ in the transport process is unknown, but possible roles include (i) mediation of interactions between the DctP and DctM components, perhaps by acting as an ‘anchor’ for DctP, (ii) a chaperonin or assembly factor to stabilise DctM in the membrane, or (iii) participation in energy coupling to solute transport. Elucidation of the function of DctQ homologues in TRAP transporters will require a combination of biochemical and structural information. Knowledge of the topology of DctQ would be useful in testing hypotheses concerning the function of this subunit. This paper reports an analysis of the topological organisation of DctQ from R. capsulatus, using the β-lactamase translational fusion technique [7].

2 Materials and methods

2.1 Media, strains, growth conditions, standard molecular biology methods and computer analyses

Escherichia coli strain TOP10F′ (Invitrogen) was grown aerobically in Luria Broth or on agar plates at 37°C containing antibiotics where appropriate (ampicillin (Ap) at 100 μg ml⁻¹ and kanamycin (Km) at 30 μg ml⁻¹). Standard molecular biology protocols were followed for the isolation, PCR amplification, cloning and manipulation of DNA [8]. DNA sequencing was carried out by the automated DyeDeoxy chain termination method (ABIPrism/PE Biosystems). The R. capsulatus DctQ sequence [1] was analysed using the TMPRED, TOPRED2, TMHMM and HMMTOP programmes available through the ExPaSy server (http://www.expasy.ch) with the programme default parameters employed in each case.

2.2 Subcloning of dctQ into pYZ4 and generation of dctQ ′-blaM fusions

The ∼0.7 kb R. capsulatus dctQ gene was amplified from pDCT207 [1,3] by PCR with pfu polymerase, using the following primers; DCTQTOPF, 5′-_GGCCGAATTCCATG_GTGCGCATCCTCGACCGG-3′ and DCTQTOPR, 5′-_GGCCAAGCTTCCGGTCAGCATCAGCGCGATG-3′ (mismatches are subscripted, EcoRI and HindIII sites are single underlined, the dctQ start codon is double underlined and the NcoI site is italicised in bold). The EcoRI/HindIII digested PCR product was ligated into pZERO-2 (Invitrogen) and transformed into E. coli TOP10F′ to give pIQTOP1. The 0.7 kb NcoI/HindIII fragment of pIQTOP1 containing dctQ was then ligated into the Km^r plasmid pYZ4 [7] to give pQTOPA. Nested 3′ deletions of dctQ were made by digesting HindIII and KpnI cut pQTOPA with exonuclease III using a commercial deletion kit (Pharmacia) according to the manufacturers instructions. The products were rendered blunt-ended by S1 nuclease treatment, end-filled with Klenow enzyme, digested with SacI and then ligated with the 0.85 kb SacI-SmaI blaM cassette from the plasmid pLH21 [7]. The ligation products were then used to transform TOP10F′ to Km^r. Several site-specific fusions were also constructed because fusions to certain regions of DctQ were not obtained by the deletion method. Reverse primers were designed which incorporated a 5′EcoRV site for cloning purposes, and were used with the DCTQTOPF primer to PCR amplify the required dctQ segments. The resulting PCR products were cloned into pZeRO-2, the truncated dctQ genes excised as NcoI/EcoRV fragments and cloned into pYZ4. The SacI/SmaI digested BlaM cassette was then ligated into these plasmids after SacI/EcoRV digestion. All transformants were tested for their ability to grow at high and low cell densities by patching onto L-agar containing 100 μg ml⁻¹ Ap, and their MICs determined in triplicate by serial dilution onto L-agar containing 30 μg ml⁻¹ Km, 0.1 mM IPTG and 0–200 μg ml⁻¹ Ap. The position of the fusion junctions in plasmids isolated from the transformants was determined by DNA sequencing using the primer BLAM1 (5′-CTCGTGCACCCAACTGA-3′), identical to codons 14–18 of blaM.

2.3 Immunodetection of DctQ′-BlaM fusion proteins

Transformants were grown overnight in LB plus 30 μg ml⁻¹ Km and 0.1 mM IPTG. Cells from 1.5 ml culture were centrifuged (13 000×g, 5 min), resuspended in SDS sample buffer (100 mM Tris–HCl pH 6.8, 2% w/v SDS, 5% v/v β-mercaptoethanol, 7% v/v glycerol and 1% w/v bromophenol blue), boiled for 5 min, centrifuged again (13 000×g, 10 min) and 30 μl aliquots analysed by SDS–PAGE on 10% polyacrylamide gels. Proteins were transferred to a nitrocellulose filter using a Trans-Blot cell (Bio-Rad) and fusion proteins detected on the blots using a 1/500 dilution of rabbit anti-BlaM polyclonal antiserum (5 Prime to 3 Prime Inc.) as primary antibody, and a 1/30 000 dilution of alkaline phosphatase conjugated goat anti-rabbit antibody (Sigma) as secondary antibody.

3 Results

3.1 Hydropathy and predicted topology of DctQ

The sequence of DctQ was analysed using several topology prediction programmes which evaluate both hydropathy properties and topogenic characteristics in the primary sequence. The results of the TOPPRED2, TMHMM and HMMTOP analyses all strongly predicted four TM helices located at similar positions, with the N- and C-termini of the protein located in the cytoplasm. The TMHMM output is shown in Fig. 1. The predicted TM regions are; TM1, residues 11–29; TM2, residues 69–91; TM3, residues 112–134; TM4, residues 153–175. However, TMPRED predicted five TM helices, with the N-terminus outside and the C-terminus located in the cytoplasm. Four of these TM regions were in approximately the same positions as those predicted by the other programmes. The additional region (residues 32–51) has a low average hydrophobicity, and seems unlikely to be able to span the membrane. A model of DctQ topology based on the predictions from the three programmes indicating four TM regions is shown in Fig. 2, with the helix ends predicted by the TMHMM programme (Fig. 1). Examination of the distribution of positively charged (Lys+Arg) amino acids in DctQ showed that there are a total of four in the predicted periplasmic loops (all located in PL1; Fig. 2), and 12 in the regions of the protein predicted to be cytoplasmic, a charge bias of eight. This is in accordance with the ‘positive inside’ rule [9] and supports the orientation of the protein as shown in Fig. 2.

1

DctQ topology prediction from the TMHMM programme. The probability scores for transmembrane helical regions are shown as vertical lines with the amino acid position of the ends of the predicted helices given above the bars at the top of the figure. Predicted cytoplasmic (IN) and periplasmic (OUT) regions are shown as solid and dashed lines respectively.

2

Model of the topology of DctQ derived from secondary structure prediction and DctQ′-BlaM fusion analysis. The amino acid sequence of DctQ is shown with the ends of the TM regions predicted by the TMHMM programme (Fig. 1). Positively charged residues (arginine+lysine) are shown in black to illustrate their asymmetric distribution. The BlaM fusions are labelled alphabetically in rectangular boxes along with the Ap MIC value. Fusions A, B, D, E, F, I, J and O were derived from exonuclease digestion, while fusions C, G, H, K, L, M and N were constructed by PCR as described in Section 2. PL1,2, periplasmic loops; TM1–4, transmembrane helices.

3.2 Generation of dctQ ′-blaM fusions and analysis of the Ap resistance of DctQ′-BlaM fusions

In order to test the topology of DctQ predicted from the above analyses, fusions were constructed between a series of progressively truncated dctQ derivatives (generated using exonuclease III or derived by PCR) and a blaM gene encoding a leaderless β-lactamase. The fusion junctions were determined by DNA sequencing and a total of 15 in-frame fusions were characterised in detail. These fusions were distributed throughout the length of DctQ, except for the very N-terminal region including most of the first predicted TM helix. However, fusions in this region of a protein are often difficult to interpret as there is usually insufficient truncated polypeptide to allow proper membrane insertion [10]. The MIC values for Ap were determined in transformants harbouring the in-frame plasmid-borne fusions. Those transformants which exhibited MIC values of >20 μg ml⁻¹ Ap were presumed to contain fusions in which the BlaM region is exported to the periplasm, while MIC values of <20 μg ml⁻¹ indicated fusions where the BlaM region remains in the cytoplasm [7]. The MIC values are shown in Fig. 2.

There are two groups of fusions which strongly support an orientation of the protein where both the N- and C-termini are located in the cytoplasm. The first group are the six N-terminal fusions (A–F) which all had high MIC values of >200 μg ml⁻¹ Ap. Fusions A and F are located near the end and the beginning respectively of the first two predicted TM regions, while fusions B–E are located in the hydrophilic region between TM1 and TM2. The high MIC values of fusions B–E strongly indicate that this region is located in the periplasm. The second group are the C-terminal fusions M–O, which had MIC values of <10 μg ml⁻¹ Ap. Fusion M is located at the very end of predicted TM4, while fusions N and O are in the hydrophilic region at the extended C-terminus of the protein. The very low MIC values of these fusions show unequivocally that the C-terminus of DctQ is located in the cytoplasm. Fusion L is located within the hydrophilic loop between TM3 and TM4, and had an Ap MIC of 75 μg ml⁻¹. This is consistent with a periplasmic location for this region, again supporting the orientation of the protein as shown in Fig. 2. Two of the fusions isolated (H and I) were located in the hydrophilic region between TM2 and TM3 which is predicted to be located in the cytoplasm. The MIC of fusion I was within the range that would be consistent with such a cytoplasmic location, but fusion H had a higher than expected MIC. Interestingly, fusion G, which is located towards the C-terminal end of TM2 also had a high MIC. In contrast, fusion J, towards the N-terminal end of TM3 had a very low MIC, while the MIC of fusion K (at the C-terminal end of TM3) was much higher. A fusion junction formed only a third of the way through TM3 may well prevent correct membrane insertion, thus the β-lactamase moiety would be expected to be cytoplasmic (consistent with the low MIC of fusion J). However, the greater amount of TM3 sequence contained in fusion K appears to be sufficient to specify a periplasmic location for the β-lactamase moiety, giving a high MIC value.

3.3 Immunodetection of DctQ′-BlaM fusions

The use of BlaM fusions for determining topology relies on the assumption that the MIC values measured solely reflect the cellular location of the BlaM domain, rather than the abundance or stability of the fusion. To compare the amounts of the 15 DctQ′-BlaM fusions expressed, equal quantities of total cell protein from each of the transformants was subjected to SDS–PAGE, followed by Western blotting and immunodetection of BlaM using an anti-β-lactamase antibody (Fig. 3). Multiple immunoreactive bands were present on the blot, probably representing breakdown products of the full-length fusions. Such instability has been reported previously in topology studies (e.g. [11]). The full-length fusions were readily apparent as bands with the greatest molecular mass in each lane, the size of which increased with the position of the fusion junction from the N- to the C-terminus of DctQ. It is clear from Fig. 3 that all except one of the fusions (fusion K) were present in similar amounts in the various transformants. In particular, fusions M–O, conferring Ap MIC values of <10 μg ml⁻¹ were as abundant as fusions A–F, conferring Ap MIC values of >200 μg ml⁻¹. The exception was fusion K, which only produced faintly immunoreactive bands, yet conferred an Ap MIC of 50 μg ml⁻¹. This is a TM3-BlaM fusion, which may be particularly unstable.

3

Immunoblot analysis of DctQ′-BlaM fusions. Lane 1, molecular mass markers. Lanes A–O, equal quantities of cell protein from transformants expressing fusions A–O were separated by SDS–PAGE and transferred to nitrocellulose. After incubation with anti-β-lactamase antibody, fusion proteins on the blot were visualised using an alkaline phosphatase conjugated secondary antibody as described in Section 2.

4 Discussion

DctQ is an unusual protein which only shares significant sequence similarity with orthologous proteins in the TRAP transporter family. Simple hydropathy analyses of the DctQ family [2] suggest each protein contains four hydrophobic regions capable of spanning the membrane. Analyses of the R. capsulatus DctQ sequence with a variety of topology prediction programmes support this conclusion, with the TMPRED programme alone predicting five TM regions. Four TM regions with an N- and C-terminus inside topology were particularly strongly predicted by the TMHMM and the HMTOP programmes, which use recently developed hidden Markov models for more accurate topology prediction.

The actual topology of DctQ was tested experimentally using the β-lactamase fusion approach, in which the cellular location of the fusion junction is reported by the level of Ap resistance conferred by the BlaM moiety. This technique is robust, and has been used with success for both eukaryotic and prokaryotic membrane proteins [7] but, like any method based on the analysis of a translational fusion, suffers from the potential disadvantage that the topology of a truncated, not native, version of the target protein is being determined. Particular fusions may be unstable, not inserted correctly in the membrane or subject to degradation. Usually, however, such individual problems are readily apparent when a collection of fusions is analysed. In this work, we isolated 15 in-frame fusions which were located in all of the predicted loops, the C-terminal region and some of the TM regions of DctQ. From the immunoblot analysis, only one of these fusions (K) appeared to be significantly more unstable or under-expressed than the others. In interpreting the fusion data, it should be noted that fusions to TM regions may be less reliable than those to loop regions, which are less likely to disturb the native topology of the protein. The groups of fusions in the first hydrophilic loop and in the C-terminal region were found to be particularly informative, as they all had Ap MIC values that were fully consistent with predicted periplasmic and cytoplasmic locations of these regions respectively, but which would be incompatible with the opposite orientation.

The topological data reported in this study will be particularly useful in defining the function of DctQ and related proteins, in particular their interaction with the DctP and DctM homologues. For example, two of the hydrophilic loops in DctQ have been shown to be periplasmic and thus provide potential surfaces for interaction with DctP, while the TM regions and the cytoplasmic C-terminal extension are possible DctM interaction sites. Residues within these regions can now be targeted for mutagenesis and functional studies. DctQ also differs from the other orthologues so far identified, in having a significant hydrophilic C-terminal extension and a larger periplasmic loop connecting TM1 and TM2. These regions may thus have a role specific to the Dct system of R. capsulatus.

Acknowledgements

This work was supported by a grant from the UK Biotechnology and Biological Sciences Research Council to D.J.K. and S.C.A. We thank Dr Paul Golby for discussion and strains.

Author notes

1

Present address: John Innes Centre, Colney Lane, Norwich, NR4 7UH, UK.