News Physiol Sci 18: 191-195, 2003;
doi:10.1152/nips.01445.2003
1548-9213/03 $5.00
News in Physiological Sciences, Vol. 18, No. 5, 191-195,
October 2003
© 2003 Int. Union Physiol. Sci./Am. Physiol. Soc.
The Engine of ABC Proteins
Guillermo A. Altenberg
Membrane Protein Laboratory, Sealy Center for Structural Biology, and Department of Physiology and Biophysics, The University of Texas Medical Branch, Galveston, Texas 77555-0437
 |
Abstract
|
|---|
Proteins that belong to the ATP-binding cassette superfamily span from bacteria to humans and comprise one of the largest protein families. These proteins are characterized by the presence of two nucleotide-binding domains, and recent studies suggest that association and dissociation of these domains is a common basic molecular mechanism of operation that couples ATP binding/hydrolysis to substrate transport across membranes.
|
Introduction
|
|---|
ATP-binding cassette (ABC) proteins comprise one of the largest protein families, including 48 human transport proteins (see http://www.gene.ucl.ac.uk/nomenclature/genefamily/abc.html). The importance of ABC proteins in such diverse fields as agriculture and medicine cannot be understated. For example, ABC proteins play a major role in the resistance of microorganisms to antibiotics and antifungal agents, resistance of Plasmodium to antimalarial compounds, and herbicide resistance in plants. In addition to their role in causing microorganism and parasite resistance, ABC proteins have medical importance because they can cause cancer multidrug resistance (by transporting chemotherapeutic agents out of the cells), they are the targets of commonly used pharmacological agents (e.g., oral hypoglycemic agents used to treat type II diabetes), and mutations in the genes that encode ABC proteins result in diseases (e.g., cystic fibrosis). ABC proteins also have important physiological functions. For example, prokaryotic ABC proteins (bacterial permeases) mediate the influx of essential nutrients from the environment, where their concentration is frequently very low. Mammalian ABC proteins also participate in physiologically important processes such as maintenance of the blood-brain barrier (P-glycoprotein), vision (ABCR), and epithelial fluid secretion [the cystic fibrosis transmembrane conductance regulator (CFTR)].
Most ABC proteins are pumps that transport substrates across membranes by using the energy of ATP hydrolysis (class 1 and class 3 ABC proteins), but there is another set of these proteins (class 2), formed by two nucleotide-binding domains (NBDs) without membrane-associated domains, that are not implicated in transport but rather in cellular processes such as DNA repair (4). Although the main points brought up in this review probably apply to all ABC proteins, the discussion is focused on the ABC membrane proteins. These proteins are involved in transport across membranes, and although most of them are pumps, at least one, CFTR, is a Cl- channel (17, 18), and others regulate ion channels. One example is the sulfonylurea receptor (SUR), which is not a transport protein per se but regulates ATP-dependent K+ channels (2). ABC proteins can have two or more separate and independent functions. For instance, the multidrug resistance protein P-glycoprotein pumps a variety of compounds out of cells, but it also regulates Cl- channels (19). In CFTR, the energy of ATP hydrolysis is used for channel gating, but CFTR also regulates other ion channels and may transport organic compounds across cell membranes (17, 20).
Although the overall sequence identity among ABC proteins is low, identity among the NBDs is considerably higher. The presence of conserved NBDs is consistent with the view that the various functions of ABC proteins (e.g., ion channels, lipid transporters, peptide transporters) depend mostly on the divergent transmembrane domains (TMDs), whereas nucleotide binding and hydrolysis, common to all ABC proteins, require the conserved NBDs. Because of the large number of ABC proteins and the fact that their sequences differ significantly, it is not surprising that the substrates of ABC proteins include small inorganic ions, amino acids, large polypeptides, lipids, small sugars, and complex polysaccharides (4). Most ABC proteins are highly specific, but others, such as the multidrug resistance proteins P-glycoprotein and MRP1, have broad substrate specificity (1, 6). It is generally believed that ABC proteins are structurally related, but since our knowledge of the structure/function relationship of ABC proteins is very limited, it is unknown how structurally similar ABC proteins can accommodate such different substrates.
|
Domain organization of ABC proteins
|
|---|
Figure 1A
illustrates the domain organization of ABC proteins. The core structure (shown in yellow) consists of two multispanning TMDs (TMD1 and TMD2) and two NBDs (NBD1 and NBD2). The NBDs of ABC proteins are probably positioned near the plasma membrane but are not integrated into it. Some ABC proteins contain additional domains or associated proteins. Most members of the MRP (ABCC) subfamily have an additional multispanning TMD (TMD0 in Fig. 1A) of unknown function (6). A NBD1-TMD2 linker polypeptide of variable length is present in many ABC proteins. In some cases, the linker is important for regulation of the protein function by phosphorylation (17, 18). In CFTR, this linker domain is known as regulatory domain (R domain), and its phosphorylation by protein kinase A is a critical event for channel activation (18). Finally, many bacterial permeases associate with substrate-binding periplasmic proteins that deliver the substrate to the transport complex (4).
View larger version (35K):
[in this window]
[in a new window]
|
FIGURE 1. General domain structure of ATP-binding cassette (ABC) proteins. A: basic core structure of ABC proteins is depicted in yellow. Other domains and associated proteins not common to all ABC proteins are not colored. TMD, transmembrane domain; NBD, nucleotide-binding domain; PBP, periplasmic solute-binding protein. A substrate bound to the PBP is shown as a filled circle. B: different polypeptide arrangements forming the basic core domain of ABC proteins. Separate polypeptides are depicted in different colors. Examples of each arrangement include histidine permease (a), the transporter associated with antigen processing TAP (b), E. coli YhiGHI (c), the ribose transporter RbsAC (d), the iron chelate transporter FhuCB (e), and the multidrug resistance protein P-glycoprotein (f). See Ref. 4 for additional information.
|
|
The number of polypeptides that form the basic core structure varies (see Fig. 1Bb
, where separate polypeptides are depicted in different colors). In many bacterial proteins, each of the four domains is encoded as a separate polypeptide, whereas a single polypeptide contains all of the domains in most mammalian ABC proteins (4). One consequence of the varied arrangements is that TMD1 and TMD2 and/or NBD1 and NBD2 are not necessarily different; e.g., some functional ABC membrane proteins consist of two identical TMDs and two identical NBDs.
Recently, two structures of complete ABC proteins, the Escherichia coli transporters of lipids (MsbA) and vitamin B12 (BtuCD), have been resolved by X-ray crystallography (3, 12). Although the structures of these transporters provide a framework to understand some aspects of substrate transport by ABC proteins, many issues remain open. It is not clear yet whether there are significant analogies between the proposed mechanisms of transport by MsbA and BtuCD and those of other ABC proteins such as P-glycoprotein and MRP1. In this context, although the basic functional unit of two TMD-NBD tandems may be general in the ABC family, the absence of significant homology in the TMD segments and the finding of major differences in the helix packing between MsbA and BtuCD suggest that the basic arrangements and mechanisms of transport may vary among proteins of the ABC superfamily.
|
Role of NBDs in different ABC proteins
|
|---|
Many ABC proteins from lower organisms contain two identical NBDs, and the NBD1 and NBD2 of many other ABC proteins, including P-glycoprotein, have very similar sequences and functional properties (1, 4, 16). In contrast, the two NBDs of other ABC proteins have very different functional roles. Proteins of this kind include those that belong to the multidrug resistance protein subfamily, such as CFTR, SUR, and MRP1 (6, 8, 18). Both NBDs are required for proper function of mammalian ABC proteins, and a mandatory positive catalytic cooperativity between the NBDs occurs in P-glycoprotein (1, 18). In this protein, NBD1 and NBD2 are equivalent and ATP binds randomly to either NBD. Hydrolysis occurs at only one NBD per hydrolysis cycle, and NBD1 and NBD2 alternate in the hydrolysis of ATP. Although there are some asymmetries between the NBDs, both NBDs of P-glycoprotein behave fairly similarly from a functional point of view (1, 18). The current working hypothesis for CFTR channel gating also proposes cooperativity between NBD1 and NBD2 (18). The simplest interpretation of single-channel analysis studies using a number of nucleotide analogs and mutations at NBD1 and/or NBD2 is that the main effects of ATP hydrolysis by NBD1 and NBD2 are to "open" and "close" the channel, respectively. However, recent data point toward a more complex picture that involves a more stable nucleotide interaction and a slower rate of hydrolysis at NBD1 compared with NBD2. The NBDs of SUR and MRP1 also have distinct functions (2, 6, 8). It has been proposed that their ATPase activity resides at NBD2, whereas NBD1 binds but does not significantly hydrolyze ATP.
|
NBD structure
|
|---|
There is general agreement that the structure of the NBDs of different ABC proteins is very similar, independently of the level of conservation in primary sequence (3, 9–12). A representative NBD (HisP, the ATP-binding subunit of a histidine permease; see Ref. 9) is depicted in Fig. 2A. The ribbon model shows that each NBD is L-shaped with two arms or lobes. Arm I contains the basic domain homologous to the F1-ATPase (red 
-helices and ß-sheets), with the motifs A and B (conserved motifs are underlined in Fig. 2B) as well as the antiparallel ß-sheet subdomain (ABCß, shown in yellow) that interacts with the ribose and base of the nucleotide. Motifs A and B are far apart in the sequence (>90 amino acids) but close in the tertiary structure. The structural aspects of the binding of - and ß-nucleotide phosphates and Mg2+ are similar for ABC proteins and the F1-ATPase. ABC proteins and the F1-ATPase share an /ß motif (Rossman fold) that consists of a central core of ß-sheets (red in Fig. 2A) surrounded by -helices (red in Fig. 2A). The central region in the NBDs contains the conserved motifs A and B. The motif A (Gly-X4-Gly-Lys-Thr/Ser), also called P-loop, is involved in the binding of the ß- and 
-phosphates of nucleotide triphosphates (pyrophosphate binding site) and is located between a ß-sheet (red ß1 in Fig. 2B) and an -helix (red 1 in Fig. 2B) of the NBD. Interestingly, most of the hydrogen bonds are formed between the main-chain nitrogens and the phosphates, which indicates that the side chains are not important, consistent with the variability of the central X4 region. The conserved Lys is believed to contribute most of the energy of the pyrophosphate/motif A binding. Motif B is in a highly hydrophobic ß-sheet that ends in an acidic residue that interacts with Mg2+. In contrast to the pyrophosphate interactions, the interactions of the adenine and sugar with the protein are different in the F1-ATPase and monomer NBDs of ABC proteins (see Refs. 10 and 11). The ribose and adenine are buried in the F1-ATPase, whereas in ABC proteins the ABCß subdomain (yellow in Fig. 2A) moves the sugar and base away from the ATP-binding core, resulting in an active site much more exposed to solvent. The pyrophosphate-binding site is also more exposed to solvent in ABC proteins. The exposure of the ATP bound to each NBD monomer may facilitate potential interactions with the other NBD (as proposed for the Rad50 NBDs; see Association between NBDs) and TMD residues.
Arm II contains the -helical subdomain (ABC; green -helices in Fig. 2). The -helical subdomain is formed by three phylogenetically conserved -helices and contains the most conserved sequence among ABC proteins, the motif C or signature sequence. This subdomain probably participates in nucleotide binding (see Association between NBDs-phosphate linker; purple dots in Fig. 2B) and His loops, which contain conserved Gly and His residues involved in Mg2+ interaction and presentation of the attacking water, are located in the hinge region between the arms.
The molecular steps involved in nucleotide hydrolysis by ABC proteins are not well understood, but hydrolysis of ATP by ABC proteins most likely involves approximation of substrates and cofactors (ATP and Mg2+) and base activation of water. The nucleophilic side chain of an amino acid in the NBD would act as a base by donating an electron pair to a water molecule, which increases the nucleophilicity of water. This "activated" water attacks the bond between the ß- and -phosphates.
|
Association between NBDs
|
|---|
It is well established that there is functional interaction between the two NBDs, and each NBD has to be "normal" for the ABC proteins to display proper function. Nucleotide binding to the NBD monomer produces conformational changes via an induced-fit mechanism (10). The structural changes induced by ATP on each NBD, although important, are not likely responsible for the power stroke that couples the use of ATP energy to substrate transport. It has been proposed that changes in the interactions between the two NBDs during the transport cycle can provide this power stroke (12, 15).
Details on how the NBDs interact physically have only recently started to emerge. The comparison of the high-resolution crystal structures of isolated NBDs of a number of ABC proteins (bacterial and human TAP1), the two available full-length bacterial ABC proteins, and a few ABC-related proteins have created controversy (see Ref. 11). Dimers have been identified or proposed for all of the NBDs crystallized, but the proposed structural relationships between the NBDs vary significantly (see Fig. 3). For instance, HisP dimers are proposed to be arranged "back-to-back" in such a way that the two arm II structures form the common surface that interacts with the TMDs and the motifs A and C of each monomer face away from those of the other monomer. As a result, the two ATP molecules bound would be 
40 Å apart, and the motif C would not participate in nucleotide binding. The proposed dimeric structure of MalK, the ATP-binding subunit of a bacterial maltose transporter, shows a slightly asymmetric interlocking association with the motifs A of each monomer facing each other but still far away (40 Å apart), whereas the motifs C are closer to each other (18 Å).
A different picture has emerged from the analysis of the structure of the NBDs of Rad50 and the full-length ABC protein BtuCD (7, 12). Rad50 is a DNA repair enzyme that does not belong to the ABC superfamily, but its catalytic domains are related to the NBDs of ABC proteins. The NH2- and COOH-terminal NBDs of Rad50 coexpressed and copurified [Rad50 catalytic domain (Rad50CD)] dimerize in the presence of ATP (7). In the dimeric crystal structure of Rad50CD obtained in the presence of a nonhydrolyzable ATP analog, residues in motifs A and B of one monomer interact with the pyrophosphate moiety of ATP, whereas motif C of the other monomer contributes additional interactions with the phosphate, adenine, and ribose.
The structure of BtuCD obtained in the absence of ATP resembles that of Rad50CD and suggests that motif A of one NBD and motif C of the other NBD move away from each other in the absence of nucleotide (7, 12). Recently, structural studies of NBDs from Archaebacteria support a Rad50-like mechanism (15). According to these data, ATP binding induces dimerization of the NBDs, forming a nucleotide sandwich dimer. Mutagenesis studies of HisP, MalK, and Rad50CD support or are compatible with each of the proposed dimer structures (see Ref. 11). Recently, however, most attention has been focused on the relevance of the Rad50CD model because of important supporting data that include: 1) the structure of BtuCD (12), 2) vanadate photocleavage experiments on the maltose transporter complex that suggest proximity of motif A of one NBD and motif C of the other NBD during ATP hydrolysis (5), and 3) cross-linking experiments on P-glycoprotein that suggest that motif A of NBD1 and motif C of NBD2 of P-glycoprotein are very close to each other at same stage during the ATP binding/hydrolysis cycle (13). The fact that in the Rad50CD model ATP binding and hydrolysis require two NBDs helps explain the cooperativity of ATP hydrolysis in ABC proteins. It is possible that isolated ABC NBDs and MsbA crystallized as monomers and that the proposed dimeric structures are artifactual (i.e., intermolecular forces present during crystal formation produce contacts between monomers that do not represent biologically relevant association interfaces between monomers).
In summary, although the balance of the data may have recently shifted toward a Rad50CD-like NBD dimer structure, it is presently unclear how NBDs interact in functional ABC proteins with the complete core domain structure. Crystallization artifacts can explain some of the apparently conflicting results, but different NBD arrangements in different ABC proteins and/or the dynamics of the NBD associations during the transport cycle can be important. In the ABC protein SUR (MRP subfamily), recent mutagenesis data suggest that motif C is not directly involved in ATP binding (14), as proposed for Rad50CD and ABC proteins with identical NBDs or NBDs of similar function (e.g., P-glycoprotein).
|
Toward a unified mechanism of transport by ABC proteins
|
|---|
If dimerization of the NBDs does indeed occur in ABC proteins, it could not only explain the cooperativity between the NBDs but could also provide a unified molecular mechanism of coupling between ATP hydrolysis and substrate transport. A scheme of possible interactions between the NBDs during the ATP binding/hydrolysis cycle is depicted in Fig. 4. Since each ATP binding site is conformed by residues in the two NBDs, ATP binding promotes dimerization of the NBDs. After ATP hydrolysis, the electrostatic repulsion between the inorganic phosphate bound to motif C and the ADP bound to motif A could lead to the dissociation of the NBD dimer. Phosphate is rapidly released, and on release of ADP the cycle can restart. The power stroke for substrate transport can be the formation or the dissociation of the dimers. Because of the tight association between the NBDs and the TMDs, the pulling out of the NBDs after ATP hydrolysis may open the "gate" of the "channel" formed by intracellular loops in BtuCD (12). Whether the tightening of the packing of some helices on the cytoplasmic side is a key step for ABC proteins that mediate substrate efflux (e.g., P-glycoprotein), as opposed to influx (BtuCD), is unknown. In this context, it is interesting that low-resolution images of P-glycoprotein suggest that the major conformational changes are associated with ATP binding and not hydrolysis (16). Since cross-linking experiments suggest that the drug-binding pocket of P-glycoprotein has a funnel-like shape, with the larger opening facing the extracellular side, similar to that of the vitamin B12 importer BtuCD (12), it cannot be ruled out that the power stroke is ATP binding for the efflux pumps (closing access of substrates to the cytoplasmic side or inner leaflet of the plasma membrane) and ATP hydrolysis (opening access of substrates to the cytoplasm) for the influx pumps.
Undoubtedly, more high-resolution structures of ABC proteins will soon be available and will provide essential missing information, but more detailed and sophisticated biochemical and biophysical experiments will be needed to understand how ABC proteins work and to identify basic common steps of the molecular mechanism of transport by such diverse proteins.
|
Acknowledgments
|
|---|
I thank Dr. A. Ritchie, Dr. L. Reuss, and Mr. Sung H. Lee for critical reading of the manuscript.
I apologize to all colleagues whose work was not cited due to space limitations.
|
References
|
|---|
- Ambudkar SV, Dey S, Hrycyna CA, Ramachandra M, Pastan I, and Gottesman MM. Biochemical, cellular, and pharmacological aspects of the multidrug transporter. Annu Rev Pharmacol Toxicol 39: 361–398, 1999.[ISI][Medline]
- Bryan J and Aguilar-Bryan L. Sulfonylurea receptors: ABC transporters that regulate ATP-sensitive K(+) channels. Biochim Biophys Acta 1461:285–303, 1999.[Medline]
- Chang G and Roth CB. Structure of MsbA from E-coli: a homolog of the multidrug resistance ATP binding cassette (ABC) transporters. Science 293: 1793–1800, 2001.[Abstract/Free Full Text]
- Dassa E and Bouige P. The ABC of ABCs: a phylogenetic and functional classification of ABC systems in living organisms. Microbiol Res 152: 211–229, 2001.
- Fetsch EE and Davidson AL. Vanadate-catalyzed photocleavage of the signature motif of an ATP-binding cassette (ABC) transporter. Proc Natl Acad Sci USA 99: 9685–9690, 2002.[Abstract/Free Full Text]
- Hipfner DR, Deeley RG, and Cole SPC. Structural, mechanistic and clinical aspects of MRP1. Biochim Biophys Acta 1461: 359–376, 1999.[Medline]
- Hopfner KP, Karcher A, Shin DS, Craig L, Arthur LM, Carney JP, and Tainer JA. Structural biology of Rad50 ATPase: ATP-driven conformational control in DNA double-strand break repair and the ABC-ATPase superfamily. Cell 101: 789–800, 2000.[ISI][Medline]
- Hou YX, Cui L, Riordan JR, and Chang XB. ATP binding to the first nucleotide-binding domain of multidrug resistance protein MRP1 increases binding and hydrolysis of ATP and trapping of ADP at the second domain. J Biol Chem 277: 5110–5119, 2002.[Abstract/Free Full Text]
- Hung LW, Wang IX, Nikaido K, Liu PQ, Ames GFL, and Kim SH. Crystal structure of the ATP-binding subunit of an ABC transporter. Nature 396: 703–707, 1999.
- Karpowich N, Martsinkevich O, Millen L, Yuan YR, Dai PL, MacVey K, Thomas PJ, and Hunt JF. Crystal structures of the MJ1267 ATP binding cassette reveal an induced-fit effect at the ATPase active site of an ABC transporter. Structure 9: 571–586, 2001.[Medline]
- Kerr ID. Structure and association of ATP-binding cassette transporter nucleotide-binding domains. Biochim Biophys Acta 1561: 47–64, 2002.[Medline]
- Locher KP, Lee AT, and Rees DC. The E. coli BtuCD structure: a framework for ABC transporter architecture and mechanism. Science 296: 1091–1098, 2002.[Abstract/Free Full Text]
- Loo TW, Bartlett MC, and Clarke DM. The LSGGQ motif in each nucleotide-binding domain of human P-glycoprotein is adjacent to the opposing Walker a sequence. J Biol Chem 277: 41303–41306, 2002.[Abstract/Free Full Text]
- Matsuo M, Dabrowski M, Ueda K, and Ashcroft FM. Mutations in the linker domain of NBD2 of SUR inhibit transduction but not nucleotide binding. EMBO J 21: 4250–4258, 2002.[Medline]
- Moody JE, Millen L, Binns D, Hunt JF, and Thomas PJ. Cooperative, ATP-dependent association of the nucleotide binding cassettes during the catalytic cycle of ATP-binding cassette transporters. J Biol Chem 277: 21111–21114, 2002.[Abstract/Free Full Text]
- Rosenberg MF, Velarde G, Ford RC, Martin C, Berridge G, Kerr ID, Callaghan R, Schmidlin A, Wooding C, Linton KJ, and Higgins CF. Repacking of the transmembrane domains of P-glycoprotein during the transport ATPase cycle. EMBO J 20: 5615–5625, 2001.[ISI][Medline]
- Schwiebert EM, Benos DJ, Egan ME, Stutts J, and Guggino WB. CFTR is a conductance regulator as well as a chloride channel. Physiol Rev 79: 145–166, 1999.[Abstract/Free Full Text]
- Senior AE and Gadsby DC. ATP hydrolysis cycles and mechanism in P-glycoprotein and CFTR. Sem Cancer Biol 8: 143–150, 1997.[ISI][Medline]
- Vanoye CG, Castro AF, Pourcher T, Reuss L, and Altenberg GA. Phosphorylation of P-glycoprotein by PKA and PKC modulates swelling-activated Cl- currents. Am J Physiol Cell Physiol 276: C370–C378, 1999.[Abstract/Free Full Text]
- Wersto RP, Rosenthal ER, Crystal RG, and Spring KR. Uptake of fluorescent dyes associated with the functional expression of the cystic fibrosis transmembrane conductance regulator in epithelial cells. Proc Natl Acad Sci USA 93: 1167–1172, 1996.[Abstract/Free Full Text]
Copyright © 2003 by the Int. Union Physiol. Sci./Am. Physiol. Soc.
Top
Introduction
Domain organization of ABC...
