OUP user menu

Functional effects of intramembranous proline substitutions in the staphylococcal multidrug transporter QacA

Karl A. Hassan, Melanie Galea, Jingqin Wu, Bernadette A. Mitchell, Ronald A. Skurray, Melissa H. Brown
DOI: http://dx.doi.org/10.1111/j.1574-6968.2006.00411.x 76-85 First published online: 1 October 2006


The QacA multidrug transporter is encoded on Staphylococcus aureus multidrug resistance plasmids and confers broad-range antimicrobial resistance to more than 30 monovalent and bivalent lipophilic, cationic compounds from at least 12 different chemical classes. QacA contains 10 proline residues predicted to be within transmembrane regions, several of which are conserved in related export proteins. Proline residues are classically known as helix-breakers and are highly represented within the transmembrane helices of membrane transport proteins, where they can mediate the formation of structures essential for protein stability and transport function. The importance of these 10 intramembranous proline residues for QacA-mediated transport function was determined by examining the functional effect of substituting these residues with glycine, alanine or serine. Several proline-substituted QacA mutants failed to confer high-level resistance to selected QacA substrates. However, no single proline mutation, including those at conserved positions, significantly disrupted QacA protein expression or QacA-mediated resistance to all representative substrates, suggesting that these residues are not essential for the formation of structures requisite to the QacA substrate transport mechanism.

  • QacA
  • multidrug transport
  • antimicrobial resistance
  • Staphylococcus aureus
  • major facilitator superfamily
  • proline mutagenesis


The phenomenon of multidrug export, whereby a single transmembrane protein can facilitate the cellular removal of multiple structurally dissimilar compounds, was established in bacteria with the discovery and characterization of the staphylococcal qacA resistance determinant (Tennent et al., 1985). The qacA gene encodes a 514 amino acid protein, QacA, which confers resistance to at least 30 structurally dissimilar monovalent and bivalent cationic lipophilic compounds (Tennent et al., 1989; Mitchell et al., 1998; Brown & Skurray, 2001). Transport mediated by QacA is driven by an electrogenic drug:H+ antiport mechanism and adheres to classical Michaelis–Menten kinetics (Mitchell et al., 1999). Fluorometric transport assays have demonstrated that QacA transports monovalent and bivalent substrates using distinct subsets of amino acid residues, and mutagenesis studies have identified the requirement of an acidic residue at amino acid position 323 of QacA for high-level bivalent substrate transport (Mitchell et al., 1999; Xu et al., 2006).

QacA is classified within the drug:H+ antiporter 2 (DHA2) family (formerly known as the DHA14 family) of the major facilitator superfamily (MFS) of transport proteins, which includes both multidrug exporters, such as QacA, and substrate-specific exporters, such as the Gram-positive tetracycline transporters TetA(K) (Guay et al., 1993) and TetA(L) (McMurry et al., 1987; Paulsen et al., 1996b; Saier et al., 1999). The majority of MFS proteins are comprised of 12 transmembrane α-helical segments (TMS) and are thought to display a helical organization similar to that observed in the high-resolution structures of the Escherichia coli lactose permease (LacY) (Abramson et al., 2003), glycerol-3-phosphate transporter (GlpT) (Huang et al., 2003) and the recently solved EmrD multidrug exporter (Yin et al., 2006), which are members of the oligosaccharide:H+ symporter, organophosphate:Pi antiporter and drug:H+ antiporter 1 (DHA1) families of transport proteins, respectively. In these structures, the 12 TMS are organized into pseudo-symmetrical N- and C-terminal domains, which surround a putative substrate translocation region (Abramson et al., 2003; Huang et al., 2003). MFS members comprised of 14 TMS, such as QacA and other DHA2 family exporters, are likely to have evolved from a 12-TMS precursor with the procurement of two additional central TMS and may retain structural and functional similarities to 12-TMS MFS members (Jin et al., 2001; Saier, 2003).

An interesting feature of membrane transport proteins, including transporters of the MFS, is that the TMS of these proteins contain a high proportion of proline residues relative to the TMS of integral membrane proteins that do not function in transport (Deber et al., 1986). Proline is unique among the common amino acids in that the backbone nitrogen is covalently bound to its side chain, forming a pyrrolidine ring. Consequently, these residues disrupt α-helical structures; they lack a backbone amide proton for α-helical hydrogen bonding and the bulky proline side chain can be in steric conflict with surrounding residues (von Heijne, 1991). The severity of distortions to TMS generated by proline residues can vary. Studies of protein crystal structures have shown that at the extremes of the transmembrane region, proline residues may induce helical termination or initiation, whereas proline residues located centrally within TMS could introduce a kink or bend, whose degree is influenced by the surrounding tertiary protein structure (Ceruso & Weinstein, 2002; Cordes et al., 2002). The disruption to stable α-helical H-bonding induced by proline residues may also increase conformational flexibility within the helix and could generate a potential proton or cation binding site through the nonhelically satisfied backbone carbonyl H-bond acceptor (Sansom, 1992; Bright & Sansom, 2003).

Transmembrane proline residues have been shown to be significant for transport function in a number of MFS transporters including DHA family transporters. For example, two of 13 amino acid residues essential for tetracycline resistance in the TMS of the E. coli DHA1 family tetracycline transporter TetA(B) are proline residues (Tamura et al., 2001). Additionally, replacement of TMS-bound proline residues within the TetA(K) and TetA(L) export proteins can reduce or abolish tetracycline transport and/or produce an apparent leakage of K+, an alternative coupling ion for TetA(K)- and TetA(L)-mediated transport (Ginn et al., 2000; De Jesus et al., 2005). Furthermore, the crystal structures of the LacY and GlpT transporters show that proline residues are associated with the formation of kinks and bends within their TMS (Abramson et al., 2003; Huang et al., 2003). These distorted TMS may function in the tight closure of the central translocation region, or could introduce conformational plasticity to membrane-bound regions required for large molecular motions implied in the alternating access transport mechanism, thought to be a general feature of all MFS transporters (Abramson et al., 2003; Huang et al., 2003). The QacA multidrug transporter contains 10 intramembranous proline residues, several of which are conserved among related DHA family transporters (Fig. 1). In this study, these residues were investigated to determine whether they impart structural or functional properties required for the overall QacA substrate transport mechanism, or for the accommodation, recognition and transport of specific QacA substrates.

Figure 1

Amino acid sequence alignment of QacA and related bacterial MFS export proteins. The 10 intramembranous proline residues of QacA targeted in this study, and those conserved in corresponding positions of related transporters, are labelled and highlighted by grey shading. The sequence names are shown on the left and their grouping into MFS DHA1 and DHA2 families. Numbered grey bars above the alignment indicate the predicted positions of the TMS (1–14) in QacA; the broken line of TMS 10 identifies the boundaries predicted from solvent accessibility studies in this region (Xu et al., 2006). The approximate locations of TMS in other DHA family exporters aligned close to those of QacA, however, in DHA1 12-TMS family members, TMS 7–12 are aligned with TMS 9–14 of the 14-TMS QacA transporter. Sequence numbers on the right refer to the position of the rightmost residue on each line. The conserved amino acid sequence motifs, motif C and motif H (gxxxGPxiGGxl and WxwxFlINvPIg, respectively; where residues in uppercase and lowercase are conserved in at least 90% and 40% of exporters, respectively, and x represents any amino acid) which contain proline residues targeted in QacA, are identified in the alignment and labelled.

Materials and methods

Bacterial strains, plasmids and media

The E. coli strain DH5α [supE44 ΔlacU169 (φ80lacZΔM15) hsdR17 recA1 endA1 gyrA96 thi-1relA1] (Hanahan, 1983) was used for routine plasmid cloning procedures and for Western blot analyses of QacA protein expression. Escherichia coli BHB2600 (supE, supF, hsdR, met) (Hohn, 1979) was used in minimum inhibitory concentration (MIC) analysis and for ethidium (Et) transport assays. The plasmids used were: pBluescript II SK (Stratagene); the pBluescript-based qacA clone, pSK4322 (Xu et al., 2006); pTTQ18 (Stark, 1987); and the pTTQ18-based qacA clone, pSK7291 (Saidijam et al., 2006). Escherichia coli strains were grown in Luria–Bertani media. Ampicillin was used at a concentration of 100 µg mL−1 for plasmid selection where appropriate and other selective compounds were used at the concentrations described.

Chemicals and reagents

Ethidium, rhodamine 6G (R6G), pyronin Y (PyY), benzalkonium (Bc), dequalinium (Dq), 4′,6-diamidino-2-phenylindole (DAPI), chlorhexidine (Ch) and carbonyl cyanide m-chlorophenylhydrazone (CCCP) were obtained through Sigma, and diamidinodiphenylamine (Dd) was provided by Rhone-Poulenc Rorer (Dagenham, UK). Other reagents and materials were reagent grade and purchased from commercial sources.

DNA manipulations

Plasmid DNA was isolated via an alkali lysis method, or using the Quantum Prep plasmid miniprep kit (Bio-Rad). Restriction endonucleases were obtained from New England Biolabs and were used in accordance with the manufacturer's instructions. Bacterial plasmid transformations were conducted using standard molecular biological techniques. Oligonucleotides were obtained from GeneWorks (Australia). Nucleotide sequencing was conducted at the Australian Genomic Research Facility (Brisbane, Australia) using BDT V2 chemistry, and DNA sequences were assembled and stored using SEQUENCHER V3.1.1 (Gene Codes Corp.).

Construction of qacA mutants

Site-directed qacA mutants were prepared via the QuikChange™ (Strategene) method using Pfu DNA polymerase (Strategene). Pairs of complementary oligonucleotide primers were designed to incorporate the desired codon change(s) and a silent mutation that either introduced or removed an endonuclease restriction site to aid in screening. In a number of cases, degenerate primers were designed to enable the incorporation of alanine, glycine or serine codon changes into a single PCR reaction. The plasmid pSK4322 (Xu et al., 2006) was used as the qacA template for mutagenesis. This pBluescript-based plasmid facilitates low-level expression of QacA when carried in E. coli DH5α or BHB2600 cells without induction and is ideal for the functional analysis described here. In attempting to construct potentially toxic alanine and serine substitutions at amino acid position 161 of QacA, a tightly controlled pTTQ18-based (Stark, 1987) QacA expression vector, pSK7291, was used (Saidijam et al., 2006). For heightened suppression of the Ptac promoter in pSK7291 when constructing these mutant clones, 0.5–1% glucose was included in all media. The qacA coding region of this plasmid, as well as all mutated plasmids, was fully sequenced to ensure the absence of spurious secondary mutations.

Western blot analysis of QacA mutant proteins

The level of expression of mutant QacA proteins was established via Western blot analysis. Escherichia coli DH5α cells expressing mutant QacA proteins from pSK4322-based plasmids were cultured to the mid-exponential phase (OD600 nm=0.6), and equal volumes of cells, equating to c. 100 µg of total protein, were collected and subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis. After separation via electrophoresis, cellular proteins were transferred onto the polyvinylidine difluoride membrane, blotted using a rabbit anti-QacA antibody raised against a C-terminal QacA peptide and detected using a colorimetric detection system. Blots were scanned using a GS-710 calibrated imaging densitometer (BioRad) and analysed using the Quantity One® software package (BioRad). Mutant protein expression levels were determined as a percentage of wild-type QacA protein expression.

MIC analysis of QacA mutants

The MIC for each mutant qacA clone was determined using standard agar dilution methods as described previously (Xu et al., 2006). Substrates were chosen as representatives from different chemical classes and were used in the following concentration ranges: the monovalent dyes ethidium (50–2000 µg mL−1), rhodamine 6G (50–2000 µg mL−1), and pyronin Y (25–500 µg mL−1); the monovalent quaternary ammonium compound (Qac) benzalkonium (20–70 µg mL−1), and the bivalent Qac dequalinium (20–400 µg mL−1); the bivalent diamidines DAPI (5–60 µg mL−1), and diamidinodiphenylamine (25–750 µg mL−1); and the bivalent biguanidine chlorhexidine (1–12 µg mL−1). Cells were grown at 37°C for 24–48 h and the MIC determined as the lowest substrate concentration required to fully inhibit bacterial growth.

Fluorimetric ethidium transport assays of QacA mutants

Fluorimetric transport assays of E. coli BHB2600 cells expressing QacA derivatives were conducted as outlined elsewhere (Mitchell et al., 1999). The initial transport velocity for each substrate concentration was determined and Michaelis–Menten curves were fitted to these data points using KaleidaGraph version 4.0 (Synergy Software). As a negative control, the ethidium transport capacity of E. coli BHB2600 cells carrying the pBluescript vector was also tested using the same conditions.

Multiple sequence alignment

Amino acid sequences of the selected MFS DHA family drug transporters were aligned with QacA using ClustalW 1.81 (Thompson et al., 1994). Accession numbers (SwissProt) are as follows: QacA (P23215); LfrA (Q50392); NorB (Q7A5M0); EmrB (P27304); SmvA (Q5PHX7); PqrB (Q9L1B4); VceB (O51919); SgcB (Q9RMF9); TetA(K) (P14512); TetA(L) (P23054); LmrA (P46104); FarB (Q9RQ29); NorA (P21191); Bmr (P33449); MdfA (Q46966); EmrD (P31442); LmrP (Q48658); and TetA(B) (P02980).


Conservation of intramembranous proline residues within QacA-related transporters

The QacA multidrug efflux protein, encoded on the Staphylococcus aureus plasmid pSK1, contains 20 proline residues, 10 of which are predicted to be membrane embedded on the basis of hydropathy, gene fusion and solvent-accessibility analyses (Fig. 1; P73, P126, P161, P182, P211, P240, P318, P328, P331, P344) (Paulsen et al., 1996a; Xu et al., 2006). Amino acid sequence alignment revealed that the intramembranous proline residues P126, P161, P182 and P318 are conserved within a range of QacA-related multi- and single drug transporters from the DHA1 and DHA2 MFS families (Fig. 1) (Paulsen et al., 1996b). Two of these residues are conserved within previously identified amino acid sequence motifs: P161 in TMS 5, within the DHA-specific amino acid sequence motif C, and P182 in TMS 6, within the DHA2-specific motif H (Fig. 1) (Paulsen et al., 1996b). Owing to the high level of structural diversity among DHA family substrates, conserved amino acid residues, such as these proline residues, may participate in functions common to the transport mechanism of these exporters, such as proton translocation or conformational transitions required during substrate flux, and were therefore of particular interest to this investigation.

Construction and expression of proline-substituted QacA mutants

Glycine, alanine or serine substitutions of the 10 intramembranous prolines in QacA were made to represent conservative (glycine) and nonconservative (alanine and serine) amino acid changes in terms of α-helix-breaking propensities relative to proline (O'Neil & DeGrado, 1990). When targeting the proline residue at amino acid position 161 of QacA for mutagenesis using degenerate oligonucleotide primers, an absolute bias for glycine substitution was noted. Repeated attempts to construct P161A (P161 to A) or P161S QacA mutants using change-specific primers failed; the changes could not be incorporated into the pSK4322 or pSK7291 qacA expression plasmids carried in E. coli DH5α without the introduction of deleterious secondary mutations, typically a frame-shift. Notably, the codon changes were easily made in a qacA template that had been rendered translationally inactive due to the incorporation of a stop sequence (TAA) at codon position 21 of qacA. As expression from the pTTQ18-based, pSK7291 clone was quite tightly controlled (Stark, 1987), this suggests that even low-level expression of QacA proteins containing either alanine or serine substitutions for P161 is toxic in E. coli.

Following construction, the relative expression levels of the mutant proteins were determined via Western blot analysis using a QacA-specific antibody. Expression levels of these QacA derivatives ranged from 57% of wild-type QacA expression, found for the P161G QacA mutant, to 125% of the wild-type level for the P182A QacA mutant (Table 1).

View this table:
Table 1

Antimicrobial resistance profiles and expression levels of QacA mutants

Monovalent substratesBivalent substrates
Wild-type QacA MIC (µg mL−1)
QacA mutationExpression (% wild-type QacA)Mutant MIC (% wild-type QacA)
No QacA253010605152520
  • Percentage wild-type QacA expression levels are the average of two Western blot analyses.

  • MIC values were determined in E. coli BHB2600 cells expressing QacA proteins as previously described (Mitchell et al., 1999).

  • Percentage wild-type QacA MIC values are the average of two separate experiments.

  • ND: MIC levels were not determined for these compounds.

  • Qac, quaternary ammonium compounds; Et, ethidium; R6G, rhodamine 6G; PyY, pyronin Y; Bc, benzalkonium; Dq, dequalinium; DAPI, 4′,6-diamidino-2-phenlindole; Dd, diamidinodiphenylamide; Ch, chlorhexidine.

P126, P161, P182 and P318 of QacA, proline residues conserved within DHA family transporters

Despite the conservation of QacA P126, P161, P182 and P318 at corresponding positions within DHA family proteins (Fig. 1), substitution of these residues for either glycine or alanine did not result in dramatic changes in the catalytic activity of QacA. Significantly, none of the mutations made to these residues abolished QacA-mediated resistance to all representative substrates. QacA P161, which could not be substituted for alanine or serine, was substituted for glycine with a loss of resistance to the monovalent dyes ethidium, rhodamine 6G and pyronin Y (Table 1). Consistent with the poor ethidium resistance, the P161G QacA mutant demonstrated the lowest ethidium transport capacity of the mutants analysed (Fig. 2). However, the level of resistance conferred by the P161G QacA mutant to all other representative QacA substrates was close to, or in the case of benzalkonium, DAPI and diamidinodiphenylamine, greater than the wild-type level (Table 1). These results suggest that glycine constitutes a suitable replacement for P161 of QacA, but a proline at this position is optimal and plays a role in the recognition or transport of a limited subset of QacA substrates.

Figure 2

Michaelis–Menten curves depicting ethidium (Et) efflux from Escherichia coli BHB2600 cells expressing QacA derivatives, wild-type QacA (black triangle), and no QacA (upside-down black triangle). The initial rate of Et efflux for each Et concentration is expressed as the change in fluorescent units per second (Δ FU S−1).

Mutations to P126 of QacA resulted in only small reductions to QacA-mediated resistance to most of the substrates tested. However, the P126A mutant displayed a 70% reduction in resistance to dequalinium and the P126G mutant a 55% reduction in resistance to rhodamine 6G from the wild-type levels (Table 1). Interestingly, the low initial ethidium transport rates mediated by the P126A and P126G QacA mutants were sufficient for the maintenance of ethidium resistance at 70% of the wild-type QacA level, which could reflect a capacity for sustained, long-term transport by these mutants (Fig. 2a; Table 1). Similar to the trend seen for the P126 QacA mutants, the P318A QacA mutant demonstrated only small reductions in resistance to the majority of the QacA substrates and a 65% reduction in resistance to rhodamine 6G from the wild-type level (Table 1). The initial ethidium transport capacity of the P318A mutant was close to the wild-type level (Fig. 2d). Substitutions of P182 in QacA had little impact on QacA-mediated resistance (Table 1) or transport function (Fig. 2b).

P73, P211, P240, P328, P331 and P344, nonconserved proline residues in QacA

In general, the effect of substituting membrane-embedded proline residues that are not conserved between QacA-related transporters was not profound, nor could the results be discerned from those obtained with the conserved proline residue replacements. However, notable reductions in resistance to the four monovalent substrates and the biguanidine chlorhexidine were observed for the P328A QacA mutant (Table 1). The ethidium transport rate for this mutant was also significantly reduced from that of the wild-type protein (Fig. 2d). Modest reductions of the resistance capacity to a number of QacA substrates were observed for mutants of P73, P240, P331 and P344, whereas mutations at P211 of QacA reduced resistance levels to rhodamine 6G only (Table 1). As seen for the P126 mutant proteins, the P73S/G and P344A QacA mutants demonstrated surprisingly low initial ethidium transport relative to the conferred ethidium resistance levels (Fig. 2a and d; Table 1). The P211A/G, P240A/G, P331A and P344G QacA mutants maintained moderate ethidium transport capacities, reflecting the ethidium resistance levels of these mutants (Fig. 2c and d).


Proline residues can be a strong influence on the structure of TMS, possessing unique structural properties, which may promote helical termination, allow the introduction of a kink or bend or introduce flexibility or a potential cation-binding site (Sansom, 1992; Cordes et al., 2002; Bright & Sansom, 2003). Interestingly, the TMS of polytopic membrane transport proteins contain a high proportion of proline residues, relative to the TMS of membrane proteins not involved in transport (Deber et al., 1986). Given the potential importance of TMS-bound proline residues for the structure and function of transport proteins and the significant functional roles of proline residues in DHA family transporters (Ginn et al., 2000; Tamura et al., 2001; De Jesus et al., 2005), it was the goal of this study to determine the influence of the 10 QacA TMS-bound proline residues for the drug transport activity of the QacA transporter.

None of the substitutions made to TMS-bound proline residues abolished, or significantly reduced, QacA-mediated resistance to all representative QacA substrates. This finding is particularly significant for the P126, P161, P182 and P318 QacA residues, which are well conserved within the TMS of DHA family proteins and in some cases positioned within conserved motifs, and were thus hypothesized to be of general importance for H+-driven transport in these proteins (Fig. 2) (Paulsen et al., 1996b). This dispensability of proline residues located within TMS suggests that these residues do not form functionally essential structures within QacA. Alternatively, after proline substitution, helices in important regions of QacA may be forced to assume functionally competent tertiary conformations, similar to those of the wild-type protein, due to structural constraints imposed by surrounding helices (Ceruso & Weinstein, 2002). Similarly, substitutions of intramembranous proline residues in LacY did not prevent formation of structures essential for substrate transport, although one residue was identified as important in substrate binding and recognition (Consler et al., 1991).

Despite maintaining the capacity for significant catalytic transport activity, some QacA mutants displayed substrate-specific reductions in resistance. For example, the P126A/G and P318A QacA mutants conferred less than 50% of wild-type resistance to a single representative QacA substrate, the P161G mutant conferred poor resistance to monovalent dyes and the P328A mutant displayed limited resistance to all representative monovalent substrates and chlorhexidine (Table 1). Multidrug transporters are likely to use different subsets of amino acid residues within flexible substrate binding and translocation regions for the binding and transport of compounds with unique structural properties (Neyfakh, 2002). This has been observed in the crystal structures of the E. coli multidrug exporter AcrB in complex with several substrates (Yu et al., 2003), and demonstrated by the differential impact of D323 mutations in QacA for monovalent and bivalent substrate transport (Mitchell et al., 1999; Xu et al., 2006). Additionally, crystal structures of the qacA transcriptional repressor protein QacR, in complex with a number of structurally distinct substrates, show that these compounds, many also being substrates of QacA, are bound by different subsets of amino acid residues within one large flexible substrate-binding region (Schumacher et al., 2001; Murray et al., 2004). Therefore, substitutions of transmembrane proline residues within QacA that incur substrate-specific functional consequences could alter the localized tertiary structure around the binding or translocation regions of QacA, such that amino acid residues ideally recruited for the transport of selected substrates are no longer positioned appropriately for transport. A similar proposal was made for the human P-glycoprotein after systematic mutational analyses of the transmembrane proline residues in this multidrug transporter revealed observable changes in resistance against some substrates, but not others (Loo & Clarke, 1993). Alternatively, these proline residues may be specifically required for the formation of binding or translocation regions for some substrates. For example, the nonhelically satisfied backbone carbonyl H-bond acceptor, formed after the introduction of a proline residue within a TMS, may directly participate in the binding or translocation of specific substrates.

It was noted in this study that QacA proteins containing non-conservative alanine and serine substitutions for P161 could not be expressed in E. coli, possibly due to an inherent toxicity of these mutant QacA proteins. Additionally, the P161G QacA mutant, while tolerated in E. coli, failed to confer significant resistance to monovalent dyes (Table 1). P161 of QacA is central to amino acid sequence motif C, which is conserved within TMS 5 of DHA family members (Paulsen et al., 1996b; Ginn et al., 2000). Owing to its conservation among H+-driven exporters, motif C has been implicated in the H+-antiport mechanism of these transporters (Paulsen et al., 1996b). Molecular modelling has suggested that the GP dipeptide of this motif introduces a bend into TMS 5 (Varela et al., 1995), which could be important for the formation of a permeability barrier within these proteins (Tamura et al., 2001; De Jesus et al., 2005). The substitution of P161 for alanine or serine could perturb the natural structure of TMS 5 and result in complications to proton coupling or the permeability barrier within QacA that render the protein toxic to E. coli cells. As glycine confers flexibility to an α-helix, the P161G QacA mutant may still accommodate a bend in TMS 5, which prevents the potential lethality seen from P161A and P161S QacA mutants; however, this mutant may not allow the specific molecular motions or interactions necessary for the translocation of monovalent dyes. It is likely that future studies, including high-resolution structural analysis, will further confirm the structural significance of P161 in QacA and the corresponding residue in other proton-coupled antiporters.


This work was supported by Project Grant 301938 from the National Health and Medical Research Council (Australia).


  • Editor: Anthony George


View Abstract