Evolutionary mix-and-match with MFS transporters

Abstract

Major facilitator superfamily (MFS) transport proteins are ubiquitous in the membranes of all living cells, and ∼25% of prokaryotic membrane transport proteins belong to this superfamily. The MFS represents the largest and most diverse group of transporters and includes members that are clinically important. A wide range of substrates is transported in many instances actively by transduction of the energy stored in an H⁺ electrochemical gradient into a concentration gradient of substrate. MFS transporters are characterized by a deep central hydrophilic cavity surrounded by 12 mostly irregular transmembrane helices. An alternating inverted triple-helix structural symmetry within the N- and C-terminal six-helix bundles suggests that the proteins arose by intragenic multiplication. However, despite similar features, MFS transporters share only weak sequence homology. Here, we show that rearrangement of the structural symmetry motifs in the Escherichia coli fucose permease (FucP) results in remarkable homology to lactose permease (LacY). The finding is supported by comparing the location of 34 point mutations in FucP to the location of mutants in LacY. Furthermore, in contrast to the conventional, linear sequence alignment, homologies between sugar- and H⁺-binding sites in the two proteins are observed. Thus, LacY and FucP likely evolved from primordial helix-triplets that formed functional transporters; however, the functional segments assembled in a different consecutive order. The idea suggests a simple, parsimonious chain of events that may have led to the enormous sequence diversity within the MFS.

The major facilitator superfamily (MFS) of membrane proteins represents the largest family of secondary transporters, with members from Archaea to Homo sapiens (1, 2). MFS proteins catalyze transport of a wide range of substrates in both directions across the membrane, in many instances catalyzing active transport by transducing the energy stored in an H⁺ electrochemical gradient into a concentration gradient of substrate. The lactose permease of Escherichia coli (LacY), the most extensively studied member of the MFS (3, 4), catalyzes the coupled stoichiometric transport of a galactopyranoside and an H⁺ (galactoside/H⁺ symport) (5–9). The direction of transport is dependent upon the polarity of the sugar concentration gradient (downhill sugar/uphill H⁺) or the H⁺ electrochemical gradient (downhill H⁺/uphill sugar) (3, 10, 11). Each of the 417 amino acyl side chains in LacY has been mutated (12), and functional analyses of the mutants reveals that only 92 side chains have any significant influence on sugar accumulation, lowering the rate below 50% of WT. However, fewer than 10 side chains are irreplaceable in the symport mechanism: Glu126 (helix IV), Arg144 (helix V), and Trp151 (helix V) are directly involved in galactoside recognition and binding; Tyr236 (helix VII), Glu269 (helix VIII), and His322 (helix X) are involved in both H⁺ translocation and affinity for sugar; and Arg302 (helix IX) and Glu325 (helix X) play important roles in H⁺ translocation (3, 11, 12). As shown in the inward-facing crystal structures of LacY (13–16), these residues are located at the apex of a deep central hydrophilic cavity that is open to the cytoplasm only (Fig. 1B). An important aspect of secondary transport is the alternating access mechanism. Accordingly, the catalytic cycle of a transporter does not involve significant movement of sugar- and H⁺-binding sites relative to the membrane. Rather, the protein essentially moves around the sugar, alternatively exposing both sites to either side of the membrane (3). A wealth of biochemical and biophysical data is available demonstrating that the alternating access model is operative in LacY [reviewed in Smirnova et al. (4)], and initial experiments indicate that fucose permease (FucP), a distantly related MFS member, likely functions in similar fashion (17).

Fig. 1.

Structure overview. (A) FucP [Protein Data Bank (PDB) ID 3O7Q] shown in the outward-open conformation. Helices are color-coded according to the symmetry motifs: (A boxes in C) helix I–III, blue; (B boxes) helix IV–VI, green; (C boxes) helix VII–IX, orange; (D boxes) helix X–XII, yellow. White and gray flags indicate cytoplasmic and periplasmic loops within the symmetry motifs, respectively. (A and B) Upper shows the structure in the side-view along the membrane; Lower shows the periplasmic view. The position of Asp46 is indicated as a violet sphere. (B) LacY (PDB ID 2V8N) shown in the inward-open conformation. Helices are colored according to the symmetry motifs as in A. Upper shows the structure in the side-view along the membrane in the same orientation as FucP in A; Lower shows the cytoplasmic view. The loops within the symmetry motifs are highlighted as in A. The position of Glu325 is indicated as a red sphere. (C) Schematic representation of the helix-triplets in consecutive order in the sequence. Boxes represent the helices and are colored the same as in A. The flags indicate the loops within symmetry motifs according to A and B. The positions of Asp46 and Glu325 are indicated by violet and red arrows, respectively.

Recently, a crystallographic model of FucP was obtained with the central cavity open to the periplasm, an outward-facing conformation opposite to that of LacY (18). It has been proposed that the transformation between the inward- and outward-facing conformations is due to interconversion between conformations of neighboring inverted triple-helix repeats in the N- and C-terminal six-helix bundles (19). Thus, MFS transporters may have arisen by intragenic multiplication of the triple-helix motif to two pseudosymmetrical six-helix bundles (20, 21), the most common topological feature of MFS transporters. Although FucP and LacY are structurally related MFS sugar/H⁺ symporters, the two share only weak sequence similarities (∼10% sequence identity). Therefore, it is difficult to perform functional annotations of FucP using available structural and biochemical data from LacY and vice versa.

Mutagenesis of FucP reveals that Asp46 mutants exhibit l-fucose-induced Trp fluorescence quenching, and the mutants bind with essentially WT affinity (17), a conclusion consistent with isothermal calorimetry measurements (18). Moreover, although Asp46 mutants do not catalyze l-fucose/H⁺ symport, the D46A mutant catalyzes counterflow (18), a reaction that does not involve H⁺ translocation. In this respect, Asp46 mutants exhibit the main characteristics of neutral replacement mutants of Glu325 in LacY, which are also unable to catalyze lactose/H⁺ symport, but catalyze counterflow, as well as equilibrium exchange. This and additional evidence (reviewed in refs. 3, 10, 11, and 22) indicate that Glu325 is an essential player in the symport mechanism of LacY and directly involved in deprotonation. However, Glu325 in the C-terminal six-helix bundle of LacY is located at a different position from Asp46 in FucP, which is in the N-terminal six-helix bundle (Fig. 1). This observation prompted an analysis of alternative arrangements of the triple-helix motifs that might shed light on the location of functionally significant residues in LacY and FucP with respect to each other.

Results and Discussion

The crystallographic structure models of LacY (15) and FucP (18) were separated into four subdomains corresponding to the triple-helix inverted structural symmetry motifs (19) (Fig. 1C; repeat-A, H1–H3; repeat-B, H4–H6; repeat-C, H7–H9; and repeat-D, H10–H12), and the domains were then superimposed with regard to structure and sequence homology. The conservation of residues in FucP was analyzed with respect to mutations in LacY that cause greater than 50% inhibition of the lactose transport rate (12). As a result of the strong symmetry within each transporter, all superimpositions display similar qualities with root mean square deviations (rmsd) ranging from ∼1.8 Å to ∼2.6 Å (Fig. 2). However, only superimposition 4 (Figs. 2 and 3: FucP-A/LacY-D; FucP-B/LacY-C; FucP-C/LacY-B; FucP-D/LacY-A,) exhibits remarkable conservation of residues in FucP relative to functionally significant residues in LacY (12) (Fig. 2 and Table S1; 51% similarity and 27% identity). Importantly, the orientation of the superimposed triple-helix motifs in LacY and FucP are inverted with respect to each other (Fig. 3), and the cytoplasmic loops superpose on the periplasmic loops. This superposition was tested based on the phenotype of 34 mutants in FucP (Fig. 4). Additionally, most of the interactions between subdomains within LacY are also conserved in FucP (Figs. 5 and 6). When subdomain FucP-D is superimposed on LacY-A, Asp406 in FucP aligns with Asp68 in LacY (Fig. 5A). This residue interacts with Lys131 in LacY, forming an essential charge pair between helices II and IV. Replacement of Asp68 blocks a conformational transition involving the two helices, which inactivates transport (23). In FucP, Asp406 is charge-paired with Arg283 in subdomain FucP-C (Fig. 5B), and transport is severely inhibited in FucP mutant D406A (Fig. 4A). Thus, the Arg283/Asp406 pair corresponds structurally and probably functionally to the charge pair between subdomains A and B in LacY; however, in FucP, it is subdomains C and D that interact.

Fig. 2.

Alternative aliments of FucP structure symmetry motifs on LacY structure symmetry motifs (A, helix-triplet H1–H3; B, helix-triplet H4–H6; C, helix-triplet H7–H9; D, helix-triplet H10–H12). Respective rmsds of the Cα atoms are indicated in blue for each superposition. The overall similarity/identity of residues is taken from Table S1 (side-chain similarity matrix: AVLI, FYW, CM, ST, KRH, DENQ, PG). The sequence similarity/identity for the combination of segment superimpositions generating the best sequence alignment (superimposition 4) is highlighted in red (compare Fig. S1 for conventional alignment of all residues).

Fig. 3.

Structure alignment of helix-triplets from LacY and FucP (helices 1–3, blue; helices 4–6, green; helices 7–9, orange; helices 10–12, yellow). The alignments are oriented with the LacY cytoplasmic side to the top. The flags indicate the loops within symmetry motifs according to Fig. 1 (white, cytoplasmic loop; gray, periplasmic loop). The numbers on the flags indicate the two helices that are connected by the respective loop.

Fig. 4.

Transport activity of FucP mutants. (A) In vivo transport: cell-based active transport assay. Transport activities of the FucP mutants are means of four measurements and are normalized to WT. Control refers to transport by fucP-deficient E. coli cells transformed with empty vector. Error bars represent SDs. The residues are colored according to residual activity of WT: 0–25%, red; 25–50%, orange; 50–100%, green; >100%, blue. On top of given histograms, the activities of the corresponding mutated residues in LacY are indicated for comparison. (B) in vitro transport, counterflow assays with proteoliposomes reconstituted with purified FucP. Counterflow activities of given FucP mutants are the means of four measurements and are normalized to the WT. Control refers to the counterflow activity of liposomes without reconstituted FucP. Color-coding is the same as in A. See Materials and Methods for details.

Fig. 5.

Superimposition of the triple-helix motifs of LacY and FucP. Significantly equivalent residue pairs are shown as sticks. For orientation, flags colored according to Figs. 1 and 2 indicate the position of loops connecting the helices (upper flag, FucP; lower flag, LacY; gray, periplasmic loop; white, cytoplasmic loop); the stoke of the flag is color-coded according to the symmetry motif. Glycines at tightly packed locations of helices II, V, VIII, and XI in LacY and the corresponding Gly-residues in FucP are indicated as pink segments. (A) Superimposition of FucP-D (yellow) on LacY-A (blue). Examples of residue pairs potentially important for transporter flexibility are shown in dark green and violet. For a complete alignment see Table S1. (B) Superimposition of FucP-C (orange) on LacY-B (green). A dashed red line indicates interactions between residues Arg144, Glu126 in LacY and Arg312, Gln274 in FucP. (C) Superimposition of FucP-B (green) on LacY-C (orange). (D) Superimposition of FucP-A (blue) on LacY-D (yellow).

Fig. 6.

Schematic representation of the superimposition of the triple-helix motifs. Helix-triplets are numbered from A through D according to the sequence order, and are colored as in Fig. 1. Overlapping side-chain positions are shown in the same color for corresponding helices in LacY and FucP. Positions of side-chains shown in Fig. 5 are boxed. Contacts between side-chains are indicated as broken lines: blue, conserved charged pairs; red, conserved intratriplet interaction of Arg in the sugar binding-site with a polar side-chain; green, conserved coordination of an acidic side-chain with two alkaline residues; gray, other prominent interactions (−⋅−, putative interactions between Gly residues in FucP). This alignment does not alter the structural integrity of LacY and FucP (Fig. S2).

Residues involved exclusively in sugar recognition and binding in LacY are also located in the subdomain LacY-B. The indole-ring of Trp151 in subdomain B and the galactopyranosyl ring provide the primary hydrophobic interaction between LacY and substrate (24). Superimposition of FucP-C and LacY-B reveals that Phe308 in FucP is at precisely the same position as Trp151 in LacY (Fig. 5B). An aromatic side chain at position 151 in LacY, preferably Trp, is an absolute requirement for binding and transport (24). FucP mutants F308A and F308D transport very poorly, consistent with the possible functional equivalence of Phe308 to Trp151 in LacY (Fig. 4A). Another essential sugar-binding residue in LacY is Arg144 (25), which is positioned two helix turns from Trp151 toward the cytoplasm (Fig. 5B). In FucP, Arg312 is on the cytoplasmic side of Phe308, and, as in LacY, mutation of Arg312 drastically impairs transport (Fig. 4). Neutral replacement of Glu126 in LacY, another irreplaceable residue in close proximity to Arg144, completely abolishes binding and transport of substrate (25). In FucP, mutation of Gln274 (Fig. 5B), which is close to Arg312, markedly impairs transport (Fig. 4). Thus, it appears that subdomains LacY-B and FucP-C contain many residues that are essential for sugar recognition and binding. However, the putative sugar-binding site in FucP seems to be organized upside-down within the subdomain, but with the same orientation with respect to the membrane relative to LacY. Arg144 likely forms a salt bridge with Glu269 in subdomain LacY-C (14, 15). An interaction is observed in FucP between Arg312 and Gln159 in subdomain FucP-B (Fig. 5C). However, FucP mutant Q159A exhibits reasonably good transport activity (Fig. 4). Superimposition of subdomains LacY-C and FucP-B exhibits remarkable conservation with respect to the location of functionally important residues. Four side-chains important for transport in LacY (12)—Asp237, Asp240, Glu269, and Asn272—align with Glu135, Asn139, Gln159, and Asn162, respectively, in FucP (Fig. 5C), and mutation of these residues in FucP also inhibits transport (Fig. 4). Although the residues in FucP-B are shifted by two helix-turns with respect to LacY-C, the spatial arrangement of the side-chains is remarkably well conserved. However, a residue corresponding to Arg302 (26, 27) in LacY is not found in FucP. Superimposition of subdomain LacY-D, where most of the residues involved in H⁺ translocation and affinity for sugar are located, on subdomain FucP-A, where Asp46 is located, also reveals conservation of functional residues. In addition to Asp46 in FucP, which is located in the same position as Glu325 in LacY, Asn42 is in the same position as His322 and Trp38 aligns with Lys319 (Fig. 5D). As in LacY, mutation of these residues in FucP severely cripples transport (Fig. 4), thereby implicating the residues as participants in substrate binding and/or H⁺ translocation (11).

Conservation of FucP residues involved in subdomain contacts is also detected in LacY-D. However, the residues in FucP are coupled with different partners. For example, in LacY, Lys358 forms a charged pair with Asp237 (28–30). Positions Lys358 and Asp237 approximate the positions of Tyr74 and Glu135, respectively, although shifted by two helix-turns in FucP (Fig. 5D). The distance between the hydroxyl group of Tyr74 and the carboxyl group of Glu135 is 5.3 Å. However, Glu135 interacts with Tyr365 with a distance of 2.4 Å. A residue corresponding to Tyr365 that might interact with Asp237 or Asp240 is not conserved in LacY.

It seems apparent that a meaningful sequence alignment between LacY and FucP can be obtained only if the order of the triple-helix motifs is altered (Fig. 6). The evolutionary relationship suggested by the structural similarity of the domains (19) is consistent with the presence of evolutionarily conserved sequence motifs. It appears that the evolutionary origins of these transporters are triple-helix bundles, which correspond to the inverted structural symmetry motifs (19), fused in tandem by intragenic duplication. The observations are consistent with the notion that the MFS evolved by intragenic duplication of helix-triplets (20). However, fusion of the triple-helix bundles likely occurred multiple times during MFS evolution, and, simultaneously, mutations may have occurred with selective effects on transport function that led to a mix-and-match type of process (Fig. 6). Consequently, MFS transporters may share only weak sequence similarities in conventional, linear sequence aliments despite similar functions and a likely alternating access mechanism of action. With a growing number of crystallographic structures of MFS transporters, the location of functionally important residues within the MFS becomes increasingly important in a structural context. This concept of evolution in the MFS may allow a more useful annotation of sequence motifs than a conventional, linear sequence aliment. It may also reveal interesting and important mechanistic differences between transporters.

Materials and Methods

Structure and Sequence Alignments.

The structural symmetry motifs according to Radestock et al. (19) were generated from the crystallographic coordinates of LacY (A, Thr7-Asn102; B, Leu104-Phe187; C, Lys220-Ser309; D, Ala311-Leu400) (15) and FucP (A, Arg22-Met115; B, Asn116-Thr229; C, Arg258-Ala345; D, Gly347-Phe431) (18), and the initial superimposition of the structural symmetry motifs from LacY and FucP was carried out by analyzing one point per residue (Cα atoms) using the program COOT version 0.7 (31). Residue types were not used; only their spatial proximities were used. The structural superposition was further improved using procedure “MatchMaker” in UCSF-Chimera (32). First, pair-wise sequence alignments of the protein fragments were created using the blocks substitution matrix 62 (BLOSUM-62) by the Needleman–Wunsch algorithm (33), and then the coordinates of the protein fragments were aligned using residue pairs from the sequence alignments. A secondary structure term, analogous to residue similarity, but the values depend on what type of secondary structure the residues are in (helix, strand, or other) (34) was included in the score as a weight relative to residue similarity term [total score = 0.30 (residue similarity score) + 0.70 (secondary structure score) − gap penalties]. No positional restrains were applied to functionally significant residues in LacY or in FucP in the sequence or structure alignments. Identity of the sequences was calculated by counting identical side chains in FucP fragments aligned with the functionally significant residues in LacY (Table S1) and dividing by total number of significant residues [n = 92; side-chain similarity matrix (amino acid single letter code): AVLI, FYW, CM, ST, KRH, DENQ, PG].

In Vivo Active Transport Assay.

E. coli BL21 (DE3), which does not contain fucp in the genome, was transformed with plasmids encoding FucP and grown in Luria Bertani broth at 37 °C with brisk aeration. Cells were harvested at an OD₆₀₀ of ∼1.0. After washing three times with 5 mM Mes (pH 6.5)/150 mM KCl (MK buffer), the cells were resuspended in the same buffer to an OD₆₀₀ of 2.0. Glycerol (20 mM final concentration) was added to 100-μL aliquots as an energy scource, and the samples were incubated at 25 °C for 3 min. l-[5,6-³H]fucose (60 Ci/mmol) was then added to a final concentration of 0.17 μM to initiate transport. Reactions were terminated by diluting the suspensions with 2.0 mL of ice-cold MK buffer at 30 s. Cells were immediately collected by filtration on 0.45-μm cellulose acetate filters. The filters were washed with another 4.0 mL of ice-cold MK buffer, dried, and assayed for radioactivity by liquid scintillation spectrometry. All experiments were repeated at least three times.

Preparation of Proteoliposomes.

Proteoliposomes were prepared as described (18) in 50 mM potassium phosphate [(KP_i; pH 7.0)/2 mM MgSO₄ (KPM 7.0 buffer)/20 mg/mL pre-extruded phospholipids/1.25% octyl-β-d-glucopyranoside (β-OG)/20 mM l-fucose/WT or a given FucP mutant (10 μg protein/mg lipid)]. β-OG was removed by incubation with 400 mg/mL Bio-Beads SM2 (Bio-Rad) overnight. After the removal of β-OG, the proteoliposomes were frozen in liquid nitrogen and thawed at room temperature for 5–10 cycles. After extrusion through a 400-nm membrane filter, the proteoliposomes were harvested by ultracentrifugation at 100,000 × g for 1 h and washed twice with ice-cold KPM 7.0 buffer to remove the external l-fucose. The proteoliposomes were resuspended in ice-cold KPM 7.0 buffer to a final concentration of 100 mg phospholipids/mL immediately before the counterflow assay.

Counterflow Assay.

Counterflow assays were carried out at 25 °C. An aliquot (2 μL) of concentrated proteoliposomes was diluted into 100 μL of KPM 7.0 buffer containing 0.17 μM l-[5,6-³H]fucose (1 μCi; 60 Ci/mmol). Reactions were terminated at 30 s by diluting the samples with 2 mL of ice-cold KPM 7.0 buffer and rapidly filtering through 0.22-μm filters (Millipore). The filters were washed with another 4.0 mL of ice-cold KPM 7.0 buffer, dried, and assayed for radioactivity by liquid scintillation spectrometry. All experiments were repeated at least three times.