- Top of page
- Computer methods
Previous work defined several families of secondary active transporters, including the prokaryotic small multidrug resistance (SMR) and rhamnose transporter (RhaT) families as well as the eukaryotic organellar triose phosphate transporter (TPT) and nucleotide-sugar transporter (NST) families. We show that these families as well as several other previously unrecognized families of established or putative secondary active transporters comprise a large ubiquitous superfamily found in bacteria, archaea and eukaryotes. We have designated it the drug/metabolite transporter (DMT) superfamily (transporter classification number 2.A.7) and have shown that it consists of 14 phylogenetic families, five of which include no functionally well-characterized members. The largest family in the DMT superfamily, the drug/metabolite exporter (DME) family, consists of over 100 sequenced members, several of which have been implicated in metabolite export. Each DMT family consists of proteins with a distinctive topology: four, five, nine or 10 putative transmembrane α helical spanners (TMSs) per polypeptide chain. The five TMS proteins include an N-terminal TMS lacking the four TMS proteins. The full-length proteins of 10 putative TMSs apparently arose by intragenic duplication of an element encoding a primordial five-TMS polypeptide. Sequenced members of the 14 families are tabulated and phylogenetic trees for all the families are presented. Sequence and topological analyses allow structural and functional predictions.
Multidrug and drug-specific efflux systems are responsible for clinically significant resistance to chemotherapeutic agents in pathogenic bacteria, fungi, parasites and in human cancer cells (reviewed in [1,2]). Over 90% of the effort aimed at understanding multidrug resistance (MDR) has dealt with members of the major facilitator (MF) and ATP-binding cassette (ABC) superfamilies. The ABC superfamily includes the ATP-dependent multidrug efflux proteins such as P-glycoprotein and MRP, responsible for resistance of human tumor cells to anticancer chemotherapeutic agents, as well as PfMDR1 associated with chloroquine resistance in the malaria parasite Plasmodium falciparum. Twenty-nine such ABC transporters are found in the completely sequenced genome of Saccharomyces cerevisiae[4–6]. A few functional and structural homologs of ABC-type MDR pumps have been identified in bacteria . The MF superfamily includes protonmotive force-dependent secondary-drug efflux pumps in pathogenic micro-organisms such as QacA and NorA of Staphylococcus aureus and CaMDR1 of Candida albicans . More than two dozen members of each of the MF and ABC superfamilies have been functionally characterized as MDR pumps (see our web site http://www-biology.ucsd.edu/~msaier/transport/ for details about these and other superfamilies).
Three remaining families of recognized bacterial MDR efflux pumps are the resistance-nodulation-division (RND), the small multidrug resistance (SMR) and the multi-antimicrobial extrusion (MATE) families. Whereas some of these MDR pumps may have evolved specifically for the purpose of expelling endogenously synthesized or exogenously derived toxic substances (e.g. transporters associated with Streptomyces antibiotic-biosynthetic operons) , others may transport drugs opportunistically (e.g. the Bacillus subtilis Blt drug pump which also expels polyamines) . Thus, there may be overlap between drug-efflux and metabolite-efflux pumps. Nevertheless, each of the aforementioned superfamilies consists of members that have common structural and mechanistic features not shared by members of the other superfamilies. In this review, we define a novel ubiquitous superfamily that includes the SMR family but is unrelated to the other recognized drug-efflux superfamilies (ABC, MF, RND and MATE).
Nutrient-uptake systems in bacteria are usually energy coupled and therefore essentially unidirectional. Although this is also true of many eukaryotic uptake systems, higher organisms including animals often take up nutrients via energy-independent facilitated diffusion systems that catalyze fully reversible transport reactions. Consequently, whereas some eukaryotes use the same pathways for uptake and efflux, this is seldom true for bacteria [10,11]. Bacterial efflux systems for sugars [12,13], amino acids [14,15], and end products of metabolism [16–18] have been described. In animal cells, no such energized systems for efflux of sugars or amino acids have as yet been detected, and the few efflux systems for end products of metabolism that have been identified are very similar to, or even identical with, those catalyzing uptake (see, for example [19,20]).
Of the few examples of sugar/amino acid/metabolite-efflux pumps currently known, almost all have been described only within the last couple of years. Further, except for the work of Krämer [10,11], there was, until recently, little recognition of the physiological need for such systems, particularly for essential nutrient building blocks such as sugars and amino acids. Metabolite efflux therefore represents a poorly understood and little studied physiological phenomenon. Evidently, the toxic effects of high concentrations of sugars, amino acids, other metabolites, and their detrimental analogs, have provided evolutionary pressure for the acquisition of these efflux systems. To what extent they will prove to be important, and the number of exporter families that will provide these functions, remain to be determined. Our phylogenetic analyses of transport systems have led us to suggest that numerous prokaryotic permease families are concerned exclusively with solute export.
The sugar-efflux transporter (SET) family [transporter classification (TC) no. 2.1.20) is a group of recently described permeases within the MFS[12,13]. Members of this family include proteins from a wide variety of Gram-negative and Gram-positive bacteria, although functional data are available only for two closely related Escherichia coli members (see  for a review). The two characterized permeases, SetA and SetB, both mediate efflux of glucose and lactose, and SetA, but not SetB, also mediates efflux of isopropyl thio-β-d-galactoside, galactose and certain sugar-containing antibiotics. SetA-mediated efflux is inhibited by a large variety of d sugars and l sugars as well as glycosides of α-anomeric and β-anomeric configurations, suggesting broad specificity. Although not established, a proton : sugar antiport mechanism was inferred.
Bost et al.  and Caroléet al.  have both identified the E. coli protein, YdeA, a member of the drug : H+ antiporter family (DHA2) (TC no. 2.1.2) of the major facilitator superfamily (MFS), as a sugar-efflux pump that accommodates both l-arabinose and isopropyl β-d-thiogalactopyranoside. This system will undoubtedly prove to be of broad specificity, as has been documented for SetA.
Krämer, Eggeling and their coworkers described an efflux pump for l-lysine and l-arginine, the so-called LysE protein of Corynebacterium glutamicum[14,15,24]. Recently, two distantly related members of this family in E. coli, RhtB and RhtC, were suggested to mediate efflux of homoserine and threonine, respectively, based primarily on growth inhibition studies [25,26]. E. coli alone has at least six paralogs of this family, all of which may prove to be amino-acid-efflux systems . The LysE superfamily (TC families 2.75–2.77) may comprise a superfamily of permeases devoted solely to efflux.
These examples may represent just the tip of the iceberg. We predict that many members of these families, as well as many proteins in new families, will prove to function in the efflux of metabolites, macromolecules and hydrophobic substances. Most of these families are also expected to be prokaryote specific.
In this article we describe a ubiquitous superfamily that includes members known to mediate (a) export of drugs, (b) nutrient uptake, (c) efflux of nutrients and metabolites, and (d) exchange of metabolites across intraorganellar membranes of eukaryotes. Most of the members of this superfamily belong to families for which few or no functional data are available. For the prokaryotic members, we expect that many of the proteins will prove to be pumps for drug or metabolite efflux. For the eukaryotic members, organellar metabolite or drug transport have been established or seem quite probable. We summarize published evidence for the functions of some of these proteins and suggest functions for some of the others. The work reported serves to characterize the interrelationships of members of this superfamily. It should provide a guide for the elucidation of functional characteristics for the diverse group of proteins that comprise the 14 families of the drug/metabolite transporter (DMT) superfamily.
- Top of page
- Computer methods
Sequences of the proteins that comprise the 14 families within the DMT superfamily were obtained by an initial screening procedure involving recursive psiblast searches without iteration until all recognizable members had been retrieved from the NR databases (e value ≤ 10−4 as obtained with this program) . This value suggests but does not establish homology as detailed in the second paragraph of this section (see also Table 3 for rigorous proof of homology). Subsequently, phylogenetic trees were constructed based on multiple alignments developed with the clustal x 1.8  and tree programs. Average hydropathy, similarity and amphipathicity plots were generated using the alignment based on the tree program, with a sliding window of 21 residues with hydrophobicity values as described by Kyte & Doolittle . Amphipathicity plots were generated with an angle set at 100 ° as is appropriate for an α helix. The clustal and tree programs gave concurring results (see  for evaluation of these and other programs concerned with phylogenetic tree construction).
Table 3. Interfamilial binary comparisons for the 14 families of the DMT superfamily. The gap program was used to align the sequences with a gap opening penalty of 8, a gap extension penalty of 2 and 300 random shuffles.
| Family|| Name|| gi number|| Residues||Total residues|| vs|| Family|| Name|| gi number|| Residues||Total Residues||Gap score SD|
|SMR||Orf1 Sty||2921424||39–107||111|| ||DME||Orf1 Ssp||1001762||91–159||337||16.3|
|SMR||Y4nH Rsp||2496716||48–110||117|| ||DME||Orf1 Hin||1573895||76–137||306||14.6|
|SMR||Orf1 Sco||5420027||40–102||106|| ||BAT||Orf1 Mja||2826397||62–135||137||10.2|
|BAT||Orf1 Mja||2826397||40–135||137|| ||DME||Orf1 Spy|| 596092||178–274||277||10.3|
|GRP||GlcU Sxy||2226001||1–288||288|| ||DME||Orf6 Pho||3257019||1–285||285||12.1|
|CEO||Orf1 Cel||3877364||1–330||330|| ||DME||Orf1 Lhe||3850047||1–280||280||9.4|
|DME||Orf1 Ssp||1001762||90–159||337|| ||BAT||Orf1Aae||2984057||66–133||143||13.6|
|DME||YcxE Bsu||6648095||53–180||287|| ||RhaT||RhaT Eco|| 132528||36–160||344||10.1|
|DME||Pago Sty||6093644||1–304||304|| ||P-DME||Orf1 ath||6957732||120–432||432||10.1|
|TPT||CptP Zma||1352200||101–409||409|| ||DME||LicB1 Hin|| 97168||1–305||305||18|
|RarD||Y680 Hin||2833492||1–298||298|| ||DME||Orf1 Pab||5457510||1–295||295||10.3|
|UGA||CGI-19 Hsa||4680677||1–382||382|| ||UAA||Orf1 Cel||7497861||1–318||318||18.5|
|TPT||Orf3 Ath||3367515||51–410||410|| ||UGA||Orf1 Rno||2136348||1–322||322||17.6|
|DME||YcaY Ckl|| 731339||1–311||311|| ||POP||Orf21 Ath||7487848||751–1128||1130||9|
|DME||LicB1 Hin|| 97168||1–305||305|| ||TPT||Orf1 Ath||3983125||100–410||410||13.2|
|TPT||Cpt2 Bol||1706110||101–402||402|| ||GMA||Orf4 Ath||7486328||1–296||296||16|
|DME||YoaV Bsu||6137261||1–292||292|| ||CSA||CmsT Cgr||2499225||1–336||336||11.3|
Family assignments were based on the phylogenetic results and on the statistical analyses obtained with the gap program . Thus, our standard criterion for establishing homology between two proteins is 9SDs, when two comparable regions of more than 60 residues (the size of a typical protein domain) are compared using the gap program, with 500 random shuffles, a gap opening penalty of 8, and a gap extension penalty of 2 , as outlined and rationalized by Saier .
The main features of the 14 families are presented in Table 1, the protein members of the 14 families are presented in Table 2, and the gap scores for representative interfamilial sequence comparisons are presented in Table 3. Comparison scores for members of a single family within the DMT superfamily always exceeded 15 SDs. Phylogenetic analyses of all members of these 14 families are presented in the text, and inclusion of outliers (members of the other 13 families) always revealed that they were more distantly related to the member of a family than the members of that family were to each other. Thus, family assignment was always based on phylogeny, and comparison scores using the gap program established distant phylogenetic relationships.
Table 1. The 14 established families in the DMT superfamily.
|Family TC no.|| Family name|| Abbr.||No. of membersa||Source organismsb||Size range (no. residues)||No. putative TMSsc||Well-characterized examplesd|
|2.A.7.1||4 TMS small multidrug resistance||SMR||30||B||103–121|| 4||EmrE Eco|
|2.A.7.2||5 TMS bacterial/archaeal transporter||BAT||6||B, Ar||137–143|| 5||–|
|2.A.7.3||10 TMS drug/metabolite exporter||DME||102||B, Ar||246–353||10||YdeD Eco|
|2.A.7.4||Plant drug/metabolite exporter||P-DME||28||Pl||251–432||10||–|
|2.A.7.5||Glucose/ribose porter||GRP||6||B||280–294||10||RbsU Lsa|
|2.A.7.6||l-rhamnose transporter||RhaT||2||B||344||10||RhaT Eco|
|2.A.7.7||Chloramphenicol-sensitivity protein||RarD||9||B||258–300|| 9 or 10||–|
|2.A.7.8||C. elegans ORF||CEO||6||An||202–839|| 9 or 10||–|
|2.A.7.9||Triose phosphate transporter||TPT||53||An, Pl, Y||246–520|| 6–9||CptR Zma|
|2.A.7.10||UDP-N-acetylglucosamine : UMP antiporter||UAA||7||An, Y||316–352||10||Mnn2 Kla|
|2.A.7.11||UDP-galactose : UMP antiporter||UGA||15||An, Pl, Y||322–465|| 9 or 10||Orf1 Hsa|
|2.A.7.12||CMP-sialate : CMP antiporter||CSA||24||An, Y||263–424||10||Orf2 Mmu|
|2.A.7.13||GDP-mannose : GMP antiporter||GMA||11||An, Pl, Y||249–372||10||Gog5 Sce|
|2.A.7.14||Plant organocation permease||POP||12||Pl||315–1128||10, 30||Pup1 Ath|
- Top of page
- Computer methods
In this paper, we have described a superfamily of topologically diverse transporters and have proposed a pathway for their origin. The SMR family proteins exhibit four established TMSs; the BAT family exhibits five putative TMSs, and the DME family proteins exhibit 10 putative TMSs. The putative 10-TMS topology for one member of the DME family (PecM of E. chrysanthemi) has recently been established . These data therefore establish that the two halves of PecM are of opposite orientation in the membrane. The pathway proposed for the evolution of these proteins is:
It is impossible to say whether four gave rise to five, or five gave rise to four, but it seems highly likely that five gave rise to 10 (Fig. 16). Other families within the DMT superfamily include proteins that exhibit as many as 10 TMSs, but possibly as few as six. We have not investigated the pathways that may have led to the loss of TMSs from the putative primordial permease protein, presumably of 10 TMSs, in some of the families within the DMT superfamily. However, equivalent evolutionary events have been documented for other families [78–81]. Detailed topological analyses and further sequence comparisons of DMT superfamily members may shed further light on this.
The functions and sources of the proteins included in the DMT superfamily are worthy of note. Six of the 14 families are exclusively prokaryotic and eight are exclusively eukaryotic. Thus, none of the families cross the prokaryotic/eukaryotic barrier. This fact argues against frequent horizontal transmission of genetic material between these two groups of organisms. However, members of the DME and P-DME families are clearly more closely related to each other than they are to other DMT superfamily members. Although no functional data are available for P-DME family proteins, we suggest that these proteins are chloroplast/plastid transporters derived from transporters in cyanobacteria, the chloroplast precursor bacteria. This is supported indirectly by the observation that TPT family members are found in chloroplasts and plant plastids, whereas other eukaryotic DMT superfamily proteins are apparently localized to intracellular organelles such as the endoplasmic reticulum and Golgi apparatus rather than to the cytoplasmic membrane of the eukaryotic cell. In eukaryotes, subcellular localization is sometimes (but not always) an evolutionarily conserved trait [37,82]. Thus, none of the mitochondrial carrier family proteins have so far been found in eukaryotic cell plasma membranes; they are apparently restricted to organelles such as mitochondria and peroxisomes . In contrast, both human ENT1 and ENT2 transporters of the equilibrative nucleoside transporter family (TC no. 2.A.57) have been shown to be present in both cell surface membranes and organellar membranes including mitochondria, nuclear envelopes and lysosomes (see  for a summary of the evidence). Why eukaryotic subcellular localization should be a well-conserved evolutionary trait in some families but not others is not at present apparent.
This paper brings together a large amount of functional data for transporters characterized from diverse sources. Phylogenetic relationships between these proteins were not previously recognized. Because these proteins are evolutionarily related, they are not only expected to exhibit basic structural similarities, they should also exhibit fundamental mechanistic similarities. The phylogenetic data presented should therefore facilitate interpretation and consolidation of data obtained from studies of many DMT superfamily porters. We hope that biochemists and molecular biologists with primary interests in this diverse group of transporters will take advantage of the reported relationships, and use them to provide guidance for the design of experiments and the interpretation of results.