The Clostridium perfringens TetA(P) efflux protein contains a functional variant of the Motif A region found in major facilitator superfamily transport proteins

In the anaerobic pathogen Clostridium perfringens, tetracycline resistance is generally encoded by the Tet(P) determinant (Abraham et al., 1988; Lyras & Rood, 1996; Sloan et al., 1994). This determinant is unusual as it consists of two overlapping genes, tetA(P) and tetB(P), which confer tetracycline resistance by two different mechanisms (Sloan et al., 1994). The tetA(P) gene encodes a tetracycline efflux protein and tetB(P) encodes a protein with a high degree of similarity to Tet(M), which confers resistance by a ribosomal protection mechanism. The previously proposed transmembrane model of the TetA(P) protein predicts the presence of three negatively charged residues within membrane-spanning regions (Fig. 1) (Bannam & Rood, 1999). Site-directed mutagenesis showed that two of these negatively charged residues, Glu-52 and Glu-59, are essential for function, whereas at Glu-89 only a negatively charged residue is required (Kennan et al., 1997). In addition, random mutagenesis of the tetA(P) gene resulted in the isolation of 31 mutants that were altered in their ability to confer tetracycline resistance (Bannam & Rood, 1999). Although not restricted to one domain, many of these residues were in the N-terminal half of the protein. Three of the mutants were located in the putative loop 23 region, suggesting that this region is important for the structure or function of TetA(P).

(29K): Fig. 1. Schematic representation of the proposed membrane structure of the TetA(P) protein. This model was generated with the TOPPRED II algorithm (Claros & von Heijne, 1994) and then modified to place the C-terminal region of the Motif A-like region wholly within loop 23 and to extend the external short loops so that they contain at least three residues. The 12 TMDs are contained within rectangles. The numbers located at the top and bottom of each TMD correspond to the first and the last amino acid within that TMD. The conserved residues that form the E59xPxGxxxDxxxRK72 motif are shown in large bold text and the important glutamate residues, E52, E59, E89, are in large text and E117 is encircled.

Major facilitator superfamily (MFS) proteins are single polypeptides that are capable of transporting small solutes across chemiosmotic gradients (Pao et al., 1998; Saier et al., 1999). Typically, these proteins utilize the proton motive force to drive the transport process. Based on their phylogeny and function, MFS proteins have been divided into 29 families consisting of over 500 proteins (Saier et al., 1999). These proteins have a wide variety of functions including uptake of essential nutrients and ions and the efflux of metabolic end products or toxic compounds. Surprisingly, with this diversity of function, hydropathy analysis predicts that the structure of these proteins is similar, with the majority containing 12 or 14 transmembrane domains (TMDs). A conserved motif, designated Motif A, is found in most members of the MFS superfamily. The consensus sequence for this motif is R/K-X-G-R/K-R/K, where X represents any amino acid and R can be replaced by K (Henderson, 1990; Marger & Saier, 1993; Paulsen & Skurray, 1993). Within the predicted structural models of MFS proteins this motif is located within the cytoplasmic loop connecting TMD2 and TMD3, with the first residue predicted to be near the cytoplasmic edge of TMD2. This physical location has been shown by extensive cysteine-scanning mutagenesis on two well studied Escherichia coli efflux proteins, lactose permease and the prototype tetracycline efflux protein, TetA(B) (Frillingos et al., 1998; Tamura et al., 2001), and further confirmed by the derivation of the first atomic structure of a member of the MFS family, the oxalate transporter, OxlT (Hirai et al., 2002, 2003). Motif A is known to be functionally important for both E. coli proteins (Jessen-Marshall et al., 1995; Yamaguchi et al., 1992). The presence of this motif within many families of the MFS has led to its designation as an MFS-specific motif. Examination of the primary sequence of TetA(P) revealed an extended variant of Motif A within the putative loop 23 region. In this study, we have analysed the role of each of the conserved residues present in this variant Motif A and have shown that this region is functionally important. In addition, we report the results of mutagenesis studies on other significant TetA(P) residues. Bacterial strains and plasmids.
All manipulations were performed with the E. coli host strain DH5α (Invitrogen Life Technologies). Strains were grown in 2xYT medium (Miller, 1972) supplemented with 100 µg ampicillin ml^-1 and various concentrations of tetracycline, where appropriate. Site-directed mutagenesis was performed on pJIR71 (Abraham et al., 1988), which is a pUC18 derivative that carries the tetA(P) gene on a 2·9 kb PstI fragment from the C. perfringens conjugative plasmid pCW3.

Site-directed mutagenesis.
This was performed using the USE mutagenesis kit (Amersham Biosciences) following the manufacturer's instructions, as described previously (Kennan et al., 1997). Oligonucleotide primers were synthesized on an Applied Biosystems 392 DNA/RNA synthesizer and sequences of the mutagenic oligonucleotides are available upon request. Since the selection primer eliminated the unique AatII site within the pUC18 portion of pJIR71, mutant plasmids were initially screened by their resistance to digestion with AatII. Subsequently, the entire tetA(P) gene region was sequenced to confirm that only the required mutation was present. Sequence analysis was carried out using the PRISM Ready Reaction Dye Deoxy Terminator Cycle Sequencing Kit (Applied Biosystems) and an ABI373A automated fluorescent sequencing apparatus (Applied Biosystems).

Analysis of mutants.
The tetracycline efflux protein analysed in this study originated from C. perfringens but is active in E. coli. Our previous attempts to introduce the tetA(P) gene onto shuttle plasmids that can replicate in C. perfringens and E. coli have resulted in instability (J. Sloan & J. Rood, unpublished results). Therefore, we are currently unable to critically analyse the function of the TetA(P) protein in C. perfringens. In this study, all of the TetA(P) functional analysis was performed in E. coli DH5α. For each of the tetA(P) mutants, the ability to confer tetracycline resistance was tested by determining tetracycline minimal inhibitory concentrations (MICs) as described previously (Kennan et al., 1997). Each E. coli DH5α derivative harbouring a mutated tetA(P) gene was grown to a turbidity at 550 nm of 0·8, diluted 1 in 100 and spot-tested (10 µl) in triplicate onto 2xYT agar plates containing tetracycline at concentrations ranging from 1 to 30 µg ml^-1. After incubation for 1820 h at 37 °C, the MIC was determined as the lowest concentration of tetracycline that completely inhibited growth. Immunoblot analysis of the TetA(P) protein and its derivatives was performed using inner-membrane preparations as described previously (Kennan et al., 1997). The protein concentration of these samples was measured using the BCA Protein Assay Kit (Pierce). For each sample, 10 µg was subjected to SDS-PAGE and the separated proteins were electrotransferred onto a nitrocellulose filter. The TetA(P) protein and its derivatives were detected by probing the filter with anti-TetA(P) antibodies raised against the C-terminal 17 aa of TetA(P) conjugated to diphtheria toxoid (Kennan et al., 1997), following the instructions of the ECL Western blotting system (Amersham Biosciences). Inner-membrane preparations of each TetA(P) derivative were isolated and immunoblotted at least three times.

Comparative analysis of the TetA(P) protein
Functional analysis of both the E. coli TetA(B) and lactose permease proteins has extended the consensus MFS Motif A region to G-x-x-x-D/E-R/K-x-G-R/K-R/K (Jessen-Marshall et al., 1995; Yamaguchi et al., 1992). The amino acid sequence located within the same region of the C. perfringens TetA(P) protein is A-x-x-x-D-V-x-S-R-K (Fig. 2a). The major differences between these regions are the absence of the two conserved glycine residues of Motif A, which are replaced by alanine and serine, and the replacement of the first conserved basic residue by valine. To examine whether the variant motif found in TetA(P) was present in other proteins, the TetA(P) amino acid sequence was compared to the databases. With the exception of TetA(P) proteins that have been identified from various clostridial species, the most closely related proteins (about 20 % identity) were uncharacterized MFS transport proteins from genome sequences. Alignment of the putative loop 23 regions of these proteins with the corresponding region of TetA(P) revealed a consensus sequence, ExPxxxxxDxxxRK (Fig. 2a). The last three conserved residues of this motif corresponded to the last three conserved residues found in Motif A; however, the first three conserved residues appear to be specific to this group of proteins.

(23K): Fig. 2. (a) Alignment of regions containing the extended Motif A. The name of each protein is listed, followed by the name of the host organism. The putative transporters that have similarity to TetA(P) (GenPept accession no. AAA20116) include HP1165 from H. pylori (NP_223809) (Tomb et al., 1997), BBI26 from Borrelia burgdorferi (AAC66191) (Fraser et al., 1997), ProP from Thermoanaerobacter tengcongensis (NP_622076) (Bao et al., 2002), CAP0086 from Clostridium acetobutylicum (NP_149250) (Nolling et al., 2001), VNG1687C from Halobacterium sp. (NP_280454) (Ng et al., 2000) and YxaM from Bacillus subtilis (P42112) (Yoshida et al., 1995). For comparison, the same regions of the prototype E. coli tetracycline-resistance protein TetA(B) (P02980) (Nguyen et al., 1983), the lactose permease LacY (GREC) (Buchel et al., 1980) and the sequence of the extended Motif A (Yamaguchi et al., 1992) are included. The variant Motif A found in TetA(P) is shown at the top of the diagram and similar regions from each of the proteins are shown, with the numbers of the first and last amino acid within the aligned region indicated. The last column shows the total length of each protein. In the extended Motif A, R may be replaced by K. Conserved residues are shown in bold. (b) Analysis of individual residues within the variant Motif A of TetA(P). The single amino acid code of each independent substitution at that position is listed and the resulting tetracycline MIC encoded by that TetA(P) derivative is shown by the subscript. Substitutions presented in white text on black background are from this study and those in normal print are from previous studies (Bannam & Rood, 1999; Kennan et al., 1997). No immunoreactive TetA(P) protein was observed for the following derivatives: P61S, A66V, D67K, D67E, D67N, R71C, K72E.

Mutagenesis of the TetA(P) protein
Site-directed mutagenesis was used to analyse the role of each of the conserved residues. Analysis of derivatives of TetA(P) in which Phe-58 was replaced independently with alanine or tyrosine exhibited tetracycline resistance similar to wild-type (Table 1), showing that this residue was not required for TetA(P) function. The next conserved residue, Glu-59, was previously found to be essential for tetracycline resistance, as changes to aspartate, glutamine or lysine all abolished function (Kennan et al., 1997). Similarly, the importance of Pro-61 was shown previously by the detection of a random mutant, P61S, which no longer encoded tetracycline resistance (Bannam & Rood, 1999). To further analyse Pro-61, changes to glycine, alanine or tryptophan were introduced. None of these mutants was able to confer tetracycline resistance (Table 1). To determine whether the various mutations had an effect on the expression levels of the TetA(P) protein, immunoblots of inner-membrane protein preparations of each mutant derivative were probed with anti-TetA(P) antiserum (Fig. 3). Less immunoreactive protein was consistently observed for P61W, although normal levels of the P61A and P61G derivatives were found. Together, these results confirm that a proline at position 61 is required for TetA(P) function.

Table 1. Tetracycline resistance levels conferred by TetA(P) mutants

(44K): Fig. 3. Immunoblots of inner-membrane proteins prepared from E. coli DH5α cells harbouring wild-type or mutant TetA(P) derivatives. Each lane contained approximately 10 µg of total protein and is indicated by the mutation carried by the corresponding TetA(P) derivative. Exceptions are the derivatives harbouring wild-type TetA(P) (pJIR71; WT), or the pUC18 vector control (C). Only the relevant protein of each blot is shown. The fourth panel is more heavily exposed, revealing the presence of a non-specific band under TetA(P).

Motif A from the E. coli TetA(B) protein has two functionally important glycine residues (Yamaguchi et al., 1992). At the same positions in the clostridial TetA(P) protein are the residues Ala-63 and Ser-70. It was possible that the activity of TetA(P) would be optimized if these residues were changed to glycine to more closely resemble Motif A. The A63G derivative conferred wild-type tetracycline resistance whereas S70G had an intermediate level of resistance (Table 1), and immunoreactive protein was observed for both proteins (Fig. 3). Clearly, converting these residues to the consensus glycine residue did not increase the level of TetA(P)-mediated tetracycline resistance. Interestingly, sequencing projects have identified four additional TetA(P) proteins from various clostridial sources (GenPept accession nos T45464, BAB71965, BAB71967 and NP_348075), each of which is conserved within this Motif A region. Unlike our TetA(P) protein, all four have a glycine residue at position 63. In contrast, these proteins all have a serine at position 70.

Analysis of the basic residues of Motif A
In TetA(P), the first basic residue of the extended Motif A is replaced by valine. To examine the functional significance of Val-68, it was replaced independently by leucine, threonine, alanine and, most importantly, the consensus arginine residue. These mutations had varying effects on tetracycline resistance (Table 1), although all yielded immunoreactive membrane proteins (Fig. 3). The conservative V68L change produced wild-type tetracycline resistance, whereas the V68A and V68T mutants had an intermediate resistance. Unexpectedly, replacement of Val-68 with arginine, as in Motif A, abolished tetracycline resistance. Unlike other tetracycline efflux proteins, it is clear that at this position in TetA(P) a positively charged residue significantly disrupts efflux function.

At position 71 and 72 in TetA(P) there are the positively charged amino acids arginine and lysine, respectively. These residues form part of the Motif A consensus sequence and are highly conserved. To examine the role of Arg-71 in TetA(P), it was independently changed to lysine, glutamate and glutamine. Immunoreactive proteins were observed for all three derivatives although there was less R71E compared to wild-type (Fig. 3). The R71K, R71E and R71Q mutants were unable to confer tetracycline resistance (Table 1), which indicated that there was a requirement for arginine at this position. Similar analysis was performed on the positively charged Lys-72 residue. Immunoreactive protein was observed for K72R, whereas K72E and K72Q had reduced levels (Fig. 3). The K72E and K72Q derivatives were unable to confer tetracycline resistance but K72R gave an intermediate resistance level (Table 1). These results suggest that there is a requirement for a basic amino acid at position 72 within TetA(P).

Functional significance of Glu-117
Our previous computer-based topology models of TetA(P) predicted that the variant Motif A was located within TMD3 (Bannam & Rood, 1999). After examining the well studied topology models of TetA(B) and lactose permease (Frillingos et al., 1998; Tamura et al., 2001), and taking into account the mutagenesis results, we considered that this region was more likely to be located within the intergenic loop region 23 (Fig. 1). After modifying the TetA(P) model to compensate for this prediction, we observed that we had now introduced a negatively charged residue, Glu-117, into the putative TMD4 (Fig. 1). To examine the significance of this residue, it was independently changed to glutamine, lysine and aspartate. The resultant E117Q and E117K derivatives no longer conferred tetracycline resistance (Table 1) and no immunoreactive protein was observed for the E117K derivative (Fig. 3). In contrast, the E117D derivative was fully functional, indicating that an acidic amino acid side chain is required at this position.

Glutamate residues in putative loop 45
To date, four functionally important glutamate residues have been identified within TMDs of the TetA(P) protein, Glu-52, Glu-59, Glu-89 and now Glu-117. Within the putative loop 45 region of TetA(P) there are five glutamate residues (Fig. 1). To examine the functional importance of this region, Glu-122, Glu-123 and Glu-131 were changed independently to glutamine. For each of these derivatives a TetA(P) protein was detected (Fig. 3) and they still conferred resistance to tetracycline (Table 1), indicating that these glutamate residues are not essential for TetA(P) function.

The conserved Motif A has been identified in most members of the MFS family of transport proteins, indicating that it has an important functional and/or structural role (Jessen-Marshall et al., 1997; Paulsen et al., 1996). A variant of this motif (E₅₉xPxGxxxDxxxRK₇₂) was identified in the tetracycline efflux protein, TetA(P), from C. perfringens. Extensive site-directed mutagenesis has now been performed on all of the conserved residues within this region of TetA(P) (Fig. 2b). Our results have shown that the integrity of each of these conserved residues is required for the function of this tetracycline-resistance protein.

A proline residue occupies the third conserved position in the TetA(P) variant Motif A region. Proline is known to have a high propensity to induce turns within proteins and to place conformational constraints on the surrounding region. Mutation of Pro-61 to residues with a smaller side-chain volume, such as alanine and glycine, abolished TetA(P) function. In addition, the non-polar bulky side chain of tryptophan at this position was also not acceptable for TetA(P) function. These results imply that side-chain spatial volume is critical at this position and that the structural constraint introduced by the presence of Pro-61 is essential for biological activity. Requirement for a proline within a similar region has also been noted for the E. coli TetA(B) protein, although a different effect was observed. When Pro-59 of TetA(B) was replaced with cysteine, resistance to tetracycline was not affected; however, this mutant demonstrated a reduced ability to transport tetracycline within everted vesicles (Kimura-Someya et al., 1998).

In TetA(B), a positively charged amino acid is found at position 67, whereas in TetA(P) it is a valine residue. When a positive charge was introduced into TetA(P) at this position, to construct the V68R derivative, resistance to tetracycline was abolished. Only non-polar residues were found to be functionally acceptable at this position, indicating that the role of this residue within TetA(P) is different to that within TetA(B). In contrast, TetA(P) does contain the two highly conserved basic residues at the end of Motif A. Mutation of Arg-71, to a negatively charged, polar or even to the basic amino acid lysine, led to loss of tetracycline resistance, suggesting a specific requirement for arginine at this position. A similar loss of function was observed with the R71C mutant (Bannam & Rood, 1999). Clearly, charge and side-chain length are important at this position in TetA(P). This result differs from that observed with TetA(B), where the corresponding residue, Arg-70, can be changed to lysine and still retain function (Yamaguchi et al., 1992; Kimura et al., 1998).

Mutation of the second basic residue, Lys-72, showed that proteins with either lysine or arginine at this position were able to confer tetracycline resistance. Replacement of this residue with other amino acids led to a loss of resistance. This result differed from that observed with TetA(B) as when the corresponding residue, Arg-71, was replaced with a neutral amino acid biological function was not affected (Yamaguchi et al., 1992). As this region of the TetA(P) protein has only two basic residues, rather than the usual three found in MFS transporters, it is possible that removal of one basic side chain is detrimental to the overall structure and therefore the correct steric positioning of this region. Independent replacement of each of the three positive charges within lactose permease was found not to affect lactose transport (Jessen-Marshall et al., 1995); similarly, double mutants still retained function. However, a triple mutant had negligible levels of permease in the membrane and consequently was completely defective for transport (Pazdernik et al., 2000). It has been suggested that for lactose permease the topology of this region is important for the ability of the protein to undergo the conformational changes required for substrate transport (Cain et al., 2000). Similar results were observed for the eukaryotic GlutI glucose transporter (Sato & Mueckler, 1999). The positively charged residues were required for correct structure of this protein, specifically for the correct positioning of this region with respect to the membrane (Sato & Mueckler, 1999). It would be of interest to examine whether the mutation of the positively charged regions of TetA(P) are causing similar disruptions to the protein structure. This would require further information regarding the native membrane topology of TetA(P), which could be obtained using either alkaline phosphatase fusions or more specifically by cysteine scanning. Attempts to carry out the former have not proven to be productive (T. Bannam & J. Rood, unpublished results) and mutation of all four cysteine residues in TetA(P) significantly reduces biological activity, limiting the value of performing cysteine-scanning mutagenesis (P. Johanesen & J. Rood, unpublished results).

Previous studies have shown that Glu-52 and Glu-59 are essential for TetA(P) function (Kennan et al., 1997). In addition, this study has shown that at position 117 of TetA(P) there is a requirement for a negatively charged residue, as previously observed at position 89 (Kennan et al., 1997). As it is known that the E. coli TetA(B) protein transports tetracycline in the form of a divalent cationtetracycline complex (Yamaguchi et al., 1990), it would be of interest to examine whether any of these glutamate side chains are involved intimately with the transport of tetracycline. By contrast, we showed that several of the glutamate residues, which are proposed to be in loop 45, are not involved in tetracycline transport as changing these residues to the neutral glutamine had little effect on tetracycline resistance.

Finally, the conserved variant Motif A, ExPxxxxxDxxxRK, was found in putative transport proteins that have limited similarity to the TetA(P) protein. Although the solute that these proteins are capable of transporting has yet to be identified, they are unlikely to encode resistance to tetracycline. None of the host strains from which these sequences were derived is known to be tetracycline resistant. The HP1165 protein from Helicobacter pylori does not encode tetracycline resistance (Gerrits et al., 2002). In addition, the only conserved functional glutamate residue that they contain is that found within the variant Motif A region. Conservation of this motif amongst proteins that may be capable of transporting different solutes suggests that its role in transport is of a general nature. It is not likely that this region is involved in interaction with the specific target molecules to be transported. Instead, its role may be in maintaining the structural conformation required for transmembrane stability or in providing a common functional role such as energy translocation via the proton motive force.

This work was supported by a research grant from the Australian Research Council. K. A. F. was the recipient of an Australian Postgraduate Award.

Footnotes

^†, Chelsea L. Salvado †Present address: Department of Microbiology and Molecular Genetics, Medical College of Wisconsin, Milwaukee, WI 53226, USA. ‡Present address: Monash Institute of Reproduction and Development, Monash University, Victoria 3800, Australia.