FEMS Microbiology Letters 179 (1999) 187^202
MiniReview
Subunit interactions in ABC transporters: towards a functional architecture Peter M. Jones, Anthony M. George * Department of Cell and Molecular Biology, Faculty of Science, University of Technology Sydney, P.O. Box 123, Broadway, Sydney, N.S.W. 2007, Australia Received 17 June 1999 ; received in revised form 9 August 1999 ; accepted 10 August 1999
Abstract The ABC superfamily is a diverse group of integral membrane proteins involved in the ATP-dependent transport of solutes across biological membranes in both prokaryotes and eukaryotes. Although ABC transporters have been studied for over 30 years, very little is known about the mechanism by which the energy of ATP hydrolysis is used to transport substrate across the membrane. The recent report of the high resolution crystal structure of HisP, the nucleotide-binding subunit of the histidine permease complex of Salmonella typhimurium, represents a significant breakthrough toward the elucidation of the mechanism of solute translocation by ABC transporters. In this review, we use data from the crystallographic structures of HisP and other nucleotide-binding proteins, combined with sequence analysis of a subset of atypical ABC transporters, to argue a new model for the dimerisation of the nucleotide-binding domains that embraces the notion that the C motif from one subunit forms part of the ATP-binding site in the opposite subunit. We incorporate this dimerisation of the ATP-binding domains into our recently reported L-barrel model for P-glycoprotein and present a general model for the cooperative interaction of the two nucleotide-binding domains and the translocation of mechanical energy to the transmembrane domains in ABC transporters. ß 1999 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. Keywords : ABC transporter; ATP-binding domain; C motif ; Histidine P dimer; P-glycoprotein
1. Overview: ABC transporters as importers and exporters The ATP-binding cassette (ABC) transporter superfamily is a large and diverse group of integral membrane proteins involved in the ATP-dependent
* Corresponding author. Tel.: +61 (2) 9514 4158; Fax: +61 (2) 9514 4003; E-mail:
[email protected]
transport of solutes across intracellular or cell surface biological membranes [1]. Because of this role, these transporters are also known as tra¤c ATPases [2]. ABC transporters have been identi¢ed in a vast number of prokaryotes and eukaryotes and, in addition to processes such as the extrusion of noxious compounds and toxins, the uptake of nutrients, the transport of ions and peptides and cell signalling, they have been implicated in many phenomena of biomedical importance including cystic ¢brosis, adrenoleukodystrophy, multidrug resistance to chemo-
0378-1097 / 99 / $20.00 ß 1999 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 1 0 9 7 ( 9 9 ) 0 0 4 1 1 - 5
FEMSLE 8995 30-9-99 Cyaan
Magenta
Geel
Zwart
188
P.M. Jones, A.M. George / FEMS Microbiology Letters 179 (1999) 187^202
therapy and multidrug resistance in bacteria, yeasts and parasites. At the time of writing of this review, a Swiss-Prot database search for ABC transporters at the National Centre for Biotechnology Information site recovered 1111 ¢les. While many of these are still to be characterised, the number and species diversity of this superfamily is manifestly apparent. In the report of the complete genome sequence of Escherichia coli K-12, at least 80 ABC proteins were identi¢ed [3]; and ABC transporters make up one of the four major gene families in humans [4]. The ABC transporter superfamily has been featured as the subject of many excellent reviews in the past decade. Some of these reviews are concerned mostly with the evolutionary connection between ABC proteins from microorganisms to humans [1,2,5^12], while others have dealt mainly with the multidrug resistance properties of ABC proteins [13^17]. This paper will review primarily the structural and functional features of the ATP-binding components of ABC transporters, especially the recent report of the crystal structure of the HisP ATPbinding cassette subunit, and will present a model for the binding and hydrolysis of ATP and the coupling of energy to the substrate translocation process. ABC transporters are con¢gured from two ancestral structural units, hydrophobic membrane-spanning and hydrophilic ATP-binding components. In contrast to the membrane-integral domains, the nucleotide-binding domains (NBDs) show a high degree of sequence similarity and identity across the ABC superfamily [1], implying a conserved structure and function for these domains. Most ABC transporters comprise four domains: two membranespanning domains that form the transmembrane (TM) channel and two ATP-binding domains located at the cytosolic surface of the membrane, but there are variations. The structural organisation of these domains can be as single proteins, homo- or heterodimers or multicomponent systems [1,10]. In eukaryotes, transmembrane domains (TMDs) and NBDs are mostly fused and duplicated, with the TMDs N-terminal to the NBDs (examples are: Pglycoprotein; STE6; pfmdr1; cystic ¢brosis conductance regulator (CFTR); SUR; MRP). In some Saccharomyces cerevisiae ABC proteins, the NBDs are N-terminal to the TMDs. Finally, there are half mol-
FEMSLE 8995 30-9-99 Cyaan
ecules with an N-terminal TMD (TAP1 and TAP2 heterodimers; Adl) or C-terminal TMD (Drosophila melanogaster white, brown and scarlet proteins). In prokaryotes, the components are expressed generally as discrete polypeptide subunits, but variations occur in which there are fused homodimers of either the TMDs or NBDs. ABC transporters can be divided into a number of subfamilies based on the type and direction of substrate transported. In microorganisms, the most widely studied have been the bacterial periplasmic permeases that are ABC importers of amino acids, sugars, peptides and ions. These systems typically include a substrate-binding periplasmic receptor that delivers the substrate to the membrane-integral subunits. The histidine permease complex of Salmonella typhimurium, for example, comprises the periplasmic binding subunit, HisJ, the membrane channel subunits, HisM and HisQ, and two ATP-binding HisP subunits [6]. Another well-studied system is the maltose transporter complex [18], MalEFGK2 , in which MalE is the periplasmic binding subunit, MalF and MalG are the membrane channel subunits and MalK is the ATP-binding subunit. Among the ABC exporters are those involved in the export of hydrophobic drugs, toxins (colicin, haemolysin), capsule polysaccharide, proteases, peptides, ions and heavy metals (see reviews [1,7,12]). In addition to the many drug export ABC proteins (Mdrs) in bacteria, yeast, parasites and animals, each of the antibiotic-producing actinomycetes possesses at least one ABC transporter that is involved in the antibiotic biosynthetic pathway [19]. Also of interest is LmrA from Lactococcus lactis [20], a close homologue of the human MDR1 P-glycoprotein that has been shown to functionally complement the MDR1 protein in human lung ¢broblast cells [21]. Although most of the identity of LmrA to MDR1 lies in the NBDs, it shows an unusual and remarkable 23% and 27% identity to the N- and C-terminal TMDs of MDR1. The evolutionary relatedness of ABC transporters across all species makes them ideal candidates for the study of the conservation of structure and function in membrane proteins. Although this review is concerned primarily with the interaction and cooperativity of the NBDs, the TMDs of ABC transporters that form the membrane channels are equally important in any consideration
Magenta
Geel
Zwart
P.M. Jones, A.M. George / FEMS Microbiology Letters 179 (1999) 187^202
of the structure-function of these proteins. This is due to the coupling of the energy of ATP hydrolysis to the process of solute translocation through the TM channels and also to the notion that the topology of the TMDs is important for substrate recognition and binding. Most recent research suggests that the TM segments and/or the loops that connect them are functionally as well as structurally important in the translocation process. It is therefore of paramount interest to recognise that there are at least ¢ve di¡erent membrane topological models all supported to varying degrees by secondary structural predictions and site-directed scanning mutagenesis or epitope mapping. The human MDR1 P-glycoprotein is most often used as the benchmark. Brie£y, the original model deploys six TM K-helices in each duplicate half of P-glycoprotein [22^24]. Three variations of this model have one [25], two [26], or four [27] of the K-helices on one or both sides of the membrane. In all four of these models, the helices are presumed to form a single membrane channel. The ¢fth model, a more radical departure from the others, has 16 L-strands forming a L-barrel in each half of P-glycoprotein, providing the molecule with two membrane channels [28].
2. Structural elements of the nucleotide-binding domains Until recently, high resolution structural data have not been available for any member of the ABC transporter superfamily. Although ABC transporters have been studied for over 30 years, very little is known about the mechanism by which the energy of ATP hydrolysis is used to transport substrate across the membrane. The recent report of the high resolution crystal structure of HisP, the NBD subunit of the histidine permease complex of S. typhimurium [29], represents a signi¢cant breakthrough toward the elucidation of the mechanism of solute translocation by ABC transporters. The crystal structure of HisP correlates well with a large body of biochemical, genetic and biophysical data. HisP subunits were crystallised as two di¡erent symmetry-related dimers, only one of which was illustrated and discussed [29]. Dimerisation of the HisP subunits is consistent with experimental data from HisP and
FEMSLE 8995 30-9-99 Cyaan
189
other ABC transporters and suggests close cooperativity between the two NBDs for ATP hydrolysis to occur. The local architecture of the ATP-binding site in HisP is also consistent with the known structures of sequence-related nucleotide-binding proteins such as ras P21 (Ras) [30], RecA [31], adenylate kinase [32], myosin [33] and the F1 -ATPase [34]. All of these proteins contain two short, highly conserved consensus sequences, the Walker A and B motifs [35] for nucleotide-binding (Fig. 1). The phosphate-binding loop (P-loop) or Walker A motif (G-X-S/T-G-X-GK-S/T-S/T) is a glycine-rich loop followed by an uncapped K-helix. Residues within this structure interact with the phosphate groups and the magnesium ion of the bound Mg2 -nucleotide complex [33,36]. The Walker B motif is h-h-h-h-D, where `h' is a hydrophobic residue. In the known structures of purine nucleotide-binding proteins, this sequence constitutes a buried L-strand within the core of the nucleotide-binding fold [33]. The highly conserved aspartate residue is involved in the coordination of the catalytic Mg2 ion. Mutational analyses of ABC transporters have established that the conserved lysine and aspartate within the Walker A and Walker B motifs, respectively, are essential for ATP hydrolysis and substitution of these and other residues within these motifs are not well tolerated with respect to ATP hydrolysis (reviewed in [9]). All ABC transporters contain within each NBD a highly conserved signature sequence (L-S-G-G-Q-Q/ R/K-Q-R), also known as the C motif [37] or peptide linker [5]. This motif is located immediately N-terminal to the Walker B motif (Fig. 1) and is the site of mutations which severely impair function in many ABC transporters. It has been suggested that this region is involved in the transduction of the energy of ATP hydrolysis to the conformational changes in the membrane-integral domains required for translocation of the substrate [13]. However, mutations in this region are known to abolish ATPase activity and it has also been suggested that the C motif forms part of the nucleotide-binding pocket [38]. In this review, we use data from the crystallographic structures of HisP and other nucleotidebinding proteins, combined with sequence analysis of a subset of atypical ABC transporters, to argue a model for the dimerisation of the NBDs that em-
Magenta
Geel
Zwart
190
P.M. Jones, A.M. George / FEMS Microbiology Letters 179 (1999) 187^202
braces the notion that the C motif from the opposite subunit forms part of the ATP-binding site. We incorporate this dimerisation of the ATP-binding domains into our recently reported L-barrel model for P-glycoprotein (Pgp) [28] and present a general model for the cooperative interaction of the two NBDs and the translocation of mechanical energy to the TMDs in ABC transporters.
3. Conserved features of the ATP-binding pocket of HisP and other P-loop proteins Comparisons of the active site structures of P-loop proteins Ras, RecA, adenylate kinase, myosin and the F1 -ATPase, reveal a high degree of conservation in the nature and distribution of the ligands that coordinate the triphosphate moiety. The three-dimensional structure of the P-loop in these proteins is remarkably similar, with measured rms deviations in CK main chain positions being of the magnitude of the error in the crystallographic coordinates. Backbone amino groups of the P-loop form extensive hydrogen bonds to the phosphate oxygen atoms of the bound nucleotide. The O-amino group of the highly conserved lysine is coordinated with oxygen atoms of the L- and Q-phosphoryl groups, assisting in stabilising and orienting the Q-phosphoryl group during transfer. A divalent cation, usually Mg2 , is absolutely required for catalysis and is bound to oxygen atoms of the L- and Q-phosphoryl groups. In the majority of purine nucleotide-binding proteins (including G-proteins, Ras, EF-Tu and myosin), the ¢rst coordination sphere of the catalytic Mg2 is completed by the hydroxyl oxygen of the conserved serine/threonine of P-loop residue 8, two water molecules and a second hydroxyl side chain. Notably, the second protein ligand to the divalent cation generally comes from a region quite removed from the P-loop, which in many cases undergoes functionally important conformational changes during the cata-
lytic cycle [33]. In G-proteins, Ras, EF-Tu and myosin, a conserved glycine occurs three residues downstream of the Walker B aspartate residue. This glycine has been shown to be involved in the binding of the Q-phosphate [30,39,40] and is pivotal to the generation of conformational change resulting from ATP hydrolysis [41,42]. For the most part, the structure of the ATP-binding pocket in HisP conforms to the general pattern evaluated in the studies outlined above. The overall fold of the HisP molecule was found to be related most closely to that of RecA and the K- and L-subunits of the F1 -ATPase [29]. This fold di¡ers from that of transducin, GiK1, Ras, EF-Tu, adenylate kinase and myosin in a number of respects. Most signi¢cantly, in ras-like proteins, adenylate kinase and myosin, the buried Walker B L-strand is located adjacent to the L-strand immediately preceding the Ploop, while in HisP, RecA and F1 -ATPase, it is situated one strand further along the L-sheet which forms the backbone of the nucleotide-binding fold. This places the Walker B conserved aspartate at a greater distance from the Mg2 -binding site and in a di¡erent orientation in HisP, RecA and F1 -ATPase. The exact coordination of the catalytic Mg2 in these three proteins is uncertain. A scheme supported by mutational analysis has been put forward for the F1 ATPase in which the position occupied by the second hydroxyl ligand in Ras-like domains and myosin is occupied by a water molecule in F1 -ATPase [43]. However, an earlier electron spin resonance study indicated that in one conformation of the chloroplast F1 -ATPase, a second hydroxyl side chain is a direct ligand to the catalytic Mg2 [44]. There is another signi¢cant di¡erence between these two groups of proteins. The binding of a conserved glycine to the Q-phosphate suggests a mechanism for direct coupling of ATP hydrolysis to conformational changes in Ras-like domains and myosin but no equivalent mechanism has been suggested by the crystal structures of HisP, RecA and F1 -ATPase. C
Fig. 1. Sequence alignment of NBDs from a set of ABC transporters. The conserved motifs A, B and C are indicated by thick overbars and the non-conserved loop by a striped overbar. Residues involved in the binding of ATP by HisP are indicated by asterisks. Designations for this and the next two ¢gures are as follows: Residue numbers are bracketed after the names at the beginning of each row. Residue identity and similarity are indicated by heavy and light shading, respectively. Sequences were obtained from the Swiss-Prot Database. The alignment was produced using MacVector (Oxford Molecular) and improved manually by introducing gaps (represented as dashes).
FEMSLE 8995 30-9-99 Cyaan
Magenta
Geel
Zwart
P.M. Jones, A.M. George / FEMS Microbiology Letters 179 (1999) 187^202
FEMSLE 8995 30-9-99 Cyaan
Magenta
Geel
Zwart
191
192
P.M. Jones, A.M. George / FEMS Microbiology Letters 179 (1999) 187^202
Fig. 2. Sequence alignment of NBDs from a set of yeast and fungi ABC transporters. N- and C-terminal NBDs are grouped together in the top and bottom halves of each block of residues. The non-conserved loop region of each NBD has been omitted from between the ¢rst (A) and second (B) blocks of the alignment. The numbered, arrowed residues are referred to in the text.
Recent experimental evidence supports a model of RecA function in which DNA rotation is directly coupled to an ATP hydrolytic motor [45]. In RecA, the phosphates of the bound ADP are close to, and point toward, the DNA-binding site. A disordered loop, implicated in DNA-binding, lies immediately downstream from the L-strand adjacent and parallel to the L-strand N-terminal to the P-loop. This highly conserved, disordered loop thus lies directly above
FEMSLE 8995 30-9-99 Cyaan
the phosphates of the bound nucleotide. Interestingly, residues in the equivalent region of the F1 ATPase form the major speci¢c interactions between the K3 L3 sub-assembly and two long helices of the Qsubunit, which is proposed to rotate. Thus, in all Ploop proteins of known structure, residues immediately downstream from the L-strand adjacent to the P-loop L-strand have been implicated in conformational changes central to the function of these pro-
Magenta
Geel
Zwart
P.M. Jones, A.M. George / FEMS Microbiology Letters 179 (1999) 187^202
teins. The crystal structure of the HisP dimer reveals that the equivalent region in HisP forms a loop and a short helix (K7). However, it is not clear yet how this region might be involved in conformational changes in the HisP dimer.
4. The C motif or signature sequence The NBDs of ABC transporters contain a large loop (approx. 100 residues, known as the non-conserved loop) between the Walker A (P-loop) and Walker B motifs. This region is generally not well conserved between ABC transporters and has been proposed to transmit the conformational changes which couple ATP hydrolysis to the transport function [13,46]. The C motif joins this non-conserved loop to the Walker B motif, and thus to the core of the catalytic site. Domain replacement mutagenesis and sequence analysis have supported the notion that regions immediately N-terminal to the C motif are involved in a mechanistic aspect of the transport function [47,48]. Biochemical and mutagenic experiments with the bacterial maltose transporter have shown that residues within the C-terminal half of the non-conserved loop undergo conformational changes and interact with regions of the membrane-spanning domains [49]. The C motif itself is a hot-spot for mutations which cause loss of function in ABC transporters, particularly mutations altering the very highly conserved serine at position two and glycine at position four (Fig. 1) [50]. In a recent study [51], it was noted that, of 283 NBDs in the Swiss-Prot database, only seven of them lacked the glycine at position four of the C motif. The crystal structure of the HisP monomer is consistent with the idea that regions of the non-conserved loop and the C motif are involved in interactions with the membrane-spanning domains. The Cterminal half of the non-conserved loop and the C motif form a helical domain at the end of arm II of the `L' shaped HisP subunit [29]. Mutational data available for HisP also implicate the arm II region in interactions with the membrane-spanning domains. These mutations are characterised by a loss of regulatory control by the membrane-spanning domains on HisP, resulting in constitutive ATPase activity and by a looser attachment of HisP to the
FEMSLE 8995 30-9-99 Cyaan
193
transporter complex [29]. Further, it should be noted that the HisP arm II-mutations which result in constitutive ATP hydrolysis are of residues not highly conserved across ABC transporters. In contrast to these data, most mutations in the C motif abolish ATP hydrolysis. In particular, mutation S154F in the C motif of HisP reduced ATP-binding to 40^ 60% ([50]; and references therein). The crystal structure of HisP reveals that most residues that are as highly conserved as several of the C motif residues in ABC transporters, are those involved in ATP-binding and/or catalysis (Fig. 1). Indeed, in a di¡erent symmetry-related HisP dimer [29], two residues immediately upstream of the C motif in one monomer make contact with Q-phosphate oxygens of the bound nucleotide in the opposite NBD. These two residues (R145 and K149) are illustrated on the ¢gure that delineates the ATP-HisP hydrogen bonding coordination sphere [29] but are not discussed. We propose that this alternative con¢guration is closer to the native conformation of the dimer, and that residues within each C motif are involved in ATPbinding and catalysis in the opposite monomer (see below), thereby directly coupling ATP hydrolysis in one monomer to conformational changes in the arm II region of the opposite monomer.
5. Sequence alignments identify atypical NBD subunits The crystal structure of HisP provides an explanation for the highly conserved nature of particular residues within the NBDs of ABC transporters. In addition to those within the Walker A and Walker B sequences (see above), the HisP structure identi¢es Q100, E179 and H211 as interacting with the bound ATP through hydrogen bonds with intermediate water molecules (underscored with asterisks in Fig. 1). Mutagenic analysis has indicated that H211 is essential to a step subsequent to ATP hydrolysis in the transport process (reviewed in [9]). Q100 and E179 appear to be coordinating a putative attacking nucleophile, with E179 being the most likely candidate for the catalytic base in the hydrolysis reaction. Even the conservative substitution E179D eliminates transport and ATP hydrolysis in HisP [29]. Supporting evidence for the proposal that residues
Magenta
Geel
Zwart
194
P.M. Jones, A.M. George / FEMS Microbiology Letters 179 (1999) 187^202
Fig. 3. Sequence alignment of NBDs from a set of bacterial sugar permeases. N- and C-terminal NBDs are grouped together in the top and bottom halves of each block of residues. The non-conserved loop region of each NBD has been omitted from between the ¢rst (A) and second (B) blocks of the alignment. The numbered, arrowed residues are referred to in the text.
FEMSLE 8995 30-9-99 Cyaan
Magenta
Geel
Zwart
P.M. Jones, A.M. George / FEMS Microbiology Letters 179 (1999) 187^202
within the C motif are involved in ATP-binding and catalysis in the opposite monomer comes from sequence alignments of a subset of ABC transporters which contain one apparently non-functional NBD subunit. Fig. 2 shows a sequence alignment of the Nand C-terminal NBDs of a group of Pgp-like ABC transporters from a small number of yeast and fungi. The N-terminal NBDs of these transporters contain substitutions for the Walker A lysine (equivalent to K45 of HisP), the glutamic acid following the Walker B aspartate (equivalent to E179 of HisP) and the conserved histidine (equivalent to His211, HisP) (Fig. 2, numbered 1 to 3). Though their activity status is unknown, these proteins would be expected to have one non-catalytic NBD [9]. Indeed the N-terminal NBD of Pdrb contains a three-residue deletion within the P-loop, and may not even bind ATP. The N-terminal NBDs of these transporters, however, contain a consensus sequence C motif. Conversely, the C-terminal NBDs, while appearing to constitute a catalytically active subunit with consensus Walker A and B motifs, contain non-conservative substitutions within the highly conserved SGG region of the C motif (Fig. 2). A similar pattern is revealed by an alignment of bacterial sugar transporters (Fig. 3), whose C-terminal NBDs have substitutions for the Walker A conserved lysine (Fig. 3, indicated by the number 1), corresponding to K45 in HisP, and for the distal histidine (Fig. 3, indicated by the number 3), corresponding to H211 in HisP. In Mgla_mycpn and Mgla_mycge, the Walker B aspartate and the following glutamic acid are also changed (Fig. 3). These ABC transporters contain consensus sequence C motifs in these presumably non-catalytic N-terminal NBDs and atypical C motifs in the apparently functional N-terminal NBDs. The nonequivalence of the NBDs in the CFTR has also been argued on the basis of biochemical and mutagenic data [52]. The N-terminal NBD1 of this ABC transporter lacks the conserved glutamic acid and histidine (E179 and H211 in HisP), while the C-terminal NBD2 of CFTR contains C motif substitutions at positions three (G1226H) and six (R1231L). Thus, while the reason for the apparent functional asymmetry of the NBD pairs in some ABC transporters remains unclear, the alignment analysis discussed above is consistent with the contention that the C motif in each NBD interacts with,
FEMSLE 8995 30-9-99 Cyaan
195
and has a role essential to, the function of the opposite NBD.
6. An alternative model of the dimerisation of the HisP subunit of the histidine permease complex We have examined the crystal structure of HisP to determine if the conformation of the monomer allows the possibility that residues within each C motif can be involved in ATP-binding and catalysis in the opposite monomer, thereby directly coupling ATP hydrolysis to conformational changes in the nonconserved loop. We have postulated that the serine at position two of the C motif constitutes a second hydroxyl ligand to the catalytic magnesium and that the backbone amide of the highly conserved glycine at position four contacts a Q-phosphate oxygen in a manner analogous to that discussed above for the Qphosphate glycine ligand in myosin and the G-related proteins. Fig. 4 depicts two views of a manual orientation of two identical HisP monomers, using the atomic coordinates of the crystal structure of the HisP monomer [29]. The serine at position two in each C motif has been positioned in approximately the same position, relative to the triphosphate moiety, as the second hydroxyl ligand to the catalytic Mg2 in myosin and Ras-like domains, with the difference that the serine in the case of HisP is being donated from an adjacent subunit in the dimeric structure (Fig. 4B). The remaining regions of each monomer were rotated to avoid obvious inter-monomer steric clashes. Fig. 4A illustrates a top view, that is from the membrane surface, of how the `knobs' of one monomer ¢t into the `holes' in the opposite monomer. In this aspect, the HisP dimer is £anked on the left and right by the two equivalent K4 helices [29], each of which are part of the C-terminal regions of the respective non-conserved loops. We have recently proposed a complete revision of the canonical 12 TM helix model for P-glycoprotein membrane topology. In our new model, the two membrane-associated domains in each half of the molecule form two membrane-spanning L-barrels [28]. The intracytoplasmic loops within each of these domains form two cytoplasmic helix bundles, one beneath each barrel. In our model, a helix from the non-conserved loop of
Magenta
Geel
Zwart
196
P.M. Jones, A.M. George / FEMS Microbiology Letters 179 (1999) 187^202
Fig. 4. K-Carbon stereo diagrams of the manually oriented dimer of the HisP NBDs. The atomic coordinates of the crystal structure of the monomer were used to generate the ¢gures on a Silicon Graphics computer using the program Setor [62] and they were plotted in Setorplot. Panels A and B depict top and side views of the dimer, respectively.
each ATP-binding domain (K6 helices [28]) is proposed to form part of each helix bundle. These K6 helices are homologous with the £anking K4 helices of the HisP dimer (Fig. 4A). The non-conserved loops of the NBDs have been proposed to provide communication with the membrane-integral domains [13] and experimental data for the maltose transporter has supported this notion [49]. In our model, the incorporation of a helix from the non-conserved loop of each NBD into the helix bundle formed by
FEMSLE 8995 30-9-99 Cyaan
the intracytoplasmic loops of each membrane-associated domain enables direct communication between the NBDs and the membrane-associated domains [28]. If the two membrane-spanning L-barrels are positioned above our proposed NBD dimer (Fig. 4A), extending to the left and right, then their respective cytoplasmic helix bundles would be positioned appropriately to engage the £anking K4 (HisP) helices of the NBD non-conserved loops. Thus this alternative dimerisation for HisP is consis-
Magenta
Geel
Zwart
P.M. Jones, A.M. George / FEMS Microbiology Letters 179 (1999) 187^202
tent with our L-barrel model for Pgp and allows us to propose an outline of the overall architecture of the complete ABC transporter. Fig. 4B shows a side view of our alternative conformation of the HisP dimer with the membrane at the top of the ¢gure and perpendicular to the plane of the page. This view illustrates the position with respect to the membrane of two well characterised residues. W105 has a £uorescence spectrum in vitro which is una¡ected by the dimerisation status of HisP or by ATP binding, but is altered by the presence of lipids [53,54]. Its location near the membrane surface in this conformation of the HisP dimer (Fig. 4B) is consistent with these ¢ndings. Experiments with a membrane-impermeant biotinylating agent have indicated that K204 in HisP is accessible to the periplasm [29]. According to our model for the architecture of the transporter, K204 is located near the cytoplasmic surface of the membrane and would be accessible to the periplasm through its respective L-barrel channel. The K4 helix [29] in the view presented in Figs. 4B and 5 is located near the middle of the dimer and is almost vertical. As discussed above for our Pgp model, the equivalent of this HisP NBD K4 helix is integrated into a helix bundle formed from the intracytoplasmic loops within the TMDs. We further suggested that the loops at the top of each intracytoplasmic helix bundle form the principal substratebinding and translocation sites [28]. The helix bundle is proposed to undergo a conformational shift from a membrane-embedded state to a lower or more cytoplasmic location during the transport cycle. According to this scheme, the K4 helix is in the equivalent of the membrane-embedded conformation in the HisP crystal structure. When the helix bundle retracts into the cytoplasm, the top of the helix bundle and thus the substrate-binding site would drop to a position more proximal to the K204 residue in HisP, that is the loop joining K6 to L10 [29]. Signi¢cantly, mutation T578C in the mouse Mdr3 alters the drug resistance pro¢le of this transporter [48], suggesting that the residue at this position interacts with the substrate-binding site. The position of this mutation maps to the residue T205, immediately Cterminal to K204 in HisP, and would thus be close to the top of the helix bundle, and the substrate-binding site, in our proposed architecture.
FEMSLE 8995 30-9-99 Cyaan
197
Further evidence in support of our proposed structure comes from experiments with MalK, the NBD subunit of the bacterial maltose transporter [55]. The M187 residue in MalK also maps to the loop between K6 to L10 containing K204 in HisP. Substitutions of M187 suppress deleterious mutations in the EAA region of the membrane-integral subunits [55]. In our L-barrel model for Pgp [28], the equivalent E(E)A region, positions 282^284 in MDR1 [56], is located at the top of the helix bundle. Other MalK substitutions which suppress EAA mutations (V149M, V149I and V154I) [55], map to HisP residues A169 and V174, which are located in the loop between K5 and L9 in HisP (Fig. 5). This region would also be proximal to the top of the helix bundle in our ABC transporter model. Finally, substitutions in the `ERGA' region (residues 522^525) in the NBD1 of murine Mdr3 were found to alter the drug speci¢city of this transporter [48] and these residues map to the region near E144 in HisP (Fig. 5); again a region which would be close to the top of the helix bundle or substrate-binding site in the L-barrel model of Pgp. It has been suggested [48] that these residues are involved in signal transduction of substrate-binding events from the membrane-associated domains to the NBDs. The architecture we have outlined here for the Pgp transporter is entirely consistent with this idea.
7. The C motif as a phosphate-binding helix With the placement of the C motif serine residue as indicated (Fig. 4B), the axis of the C motif helix points precisely at the centre of the Q-phosphate group of the bound ATP. The importance of the helix dipole in the binding and stabilisation of phosphate groups in enzyme catalytic sites is well documented [57,58]. A study of dinucleotide-binding proteins concluded that the helix dipole was the most important factor in pyrophosphate-binding, while glycine residues at the N-terminus of the helix permitted hydrogen bonding of phosphate oxygens to backbone NH groups, albeit at slightly greater bond distances than usual [57]. A similar study found phosphate-binding helices in 20 di¡erent enzymes [58]. These enzymes bind a wide variety of phosphate-containing ligands and, in every case, a
Magenta
Geel
Zwart
198
P.M. Jones, A.M. George / FEMS Microbiology Letters 179 (1999) 187^202
Fig. 5. Schematic diagram of manually oriented dimer of the HisP NBDs, in the same view as Fig. 4B. ATP molecules are coloured green, residues discussed in the text are coloured red.
glycine residue is present at the N-terminus of the phosphate-binding helix. The crystal structure of phosphoglycerate kinase reveals that an active site helix has an N-terminal glycine which forms a backbone hydrogen bond to a phosphate ion [59]. It has been concluded that interaction with this helix is the most important factor in the stabilisation of the catalytic transition state in phosphoglycerate kinase. Thus, in the orientation proposed in our dimerisation model of HisP, the C motif helix, with its two highly conserved N-terminal glycine residues, ful¢lls these requirements for phosphate-binding. Since the N-terminus of the C motif is connected by a short
FEMSLE 8995 30-9-99 Cyaan
segment to K4, it could, in a manner related to the proposed catalytic mechanism of myosin and Raslike proteins (discussed above), be directly involved in the generation of conformational changes in the helix bundle resulting from ATP hydrolysis.
8. Cooperativity of the nucleotide-binding domains A ¢nal argument in support of our proposed dimerisation model of the NBDs comes from studies indicating strong cooperativity between the two NBDs in the catalytic cycle of the complete trans-
Magenta
Geel
Zwart
P.M. Jones, A.M. George / FEMS Microbiology Letters 179 (1999) 187^202
porter. Data for a number of ABC transporters have demonstrated that two functional NBDs are required for normal transport activity and normal rates of ATP hydrolysis (reviewed in [9]). Mutations or covalent labelling of residues within the Walker A consensus region of either NBD abolish steady-state ATP hydrolysis. Covalent inhibition of Pgp with NBD-Cl (7-chloro-4-nitrobenzo-2-oxa-1,3-diazole) indicated that reaction with one NBD abolished steady-state ATP hydrolysis and prevented reaction of NBD-Cl with the opposite NBD [60]. Vanadate trapping experiments showed that trapping of ATP at one site blocked catalysis at both sites [60], while mutation of the Walker A lysine residue in either NBD eliminated nucleotide trapping entirely [61]. Together, these data indicate signi¢cant communication and strong cooperativity between the two NBDs in the catalytic cycle of the transporter. The dimer as presented by Hung et al. [29] does not appear to provide insight into the nature of the cooperativity and would, presumably, require it to be coordinated almost entirely by the TMDs. In contrast, the alternative dimer proposed here (Fig. 4) exhibits an intimate contact between the core consensus regions of the two NBDs that is consistent with the observed cooperativity between the two subunits. Our proposed interaction of each C motif with the opposite subunit of the dimer immediately accounts for the requirement for two NBDs in the transport function. Examination of this alternative dimer also indicates possible avenues of communication between the catalytic sites of the two NBDs. In particular, the putative catalytic base E179 [29], which is immediately downstream from the Walker B aspartate residue, is connected by a six-residue loop to another highly conserved aspartate residue (D185) which in our dimer forms part of the ATPbinding site of the opposite NBD (Fig. 5). Interestingly, this aspartate residue is not conserved in the N-terminal NBD of the sugar transporters discussed above (Fig. 3, indicated by the number 2) and would form part of what we propose to be the non-catalytic ATP-binding site.
9. Summary In the context of the conserved features of the
FEMSLE 8995 30-9-99 Cyaan
199
active site in P-loop proteins, we have posited an alternative dimerisation of the HisP subunits in which the C motif SGG region in each monomer interacts with the bound Mg2 -ATP in the opposite monomer. The justi¢cation for this idea comes from a number of facts: (a) the report of a HisP dimer which appears to approximate this alternative conformation [29]; (b) the SGG region in ABC transporters is very highly conserved and mutation of the serine or the second glycine has a deleterious e¡ect on ATP-binding and hydrolysis. Most other highly conserved residues in the NBDs of ABC transporters are involved with ATP-binding and/or hydrolysis; (c) a second protein hydroxyl ligand to the catalytic Mg2 is found in all P-loop ATPases where the coordination of the cation is known, and this side chain always comes from regions remote from the P-loop sequence. A second hydroxyl ligand has been implicated in the coordination of the catalytic Mg2 in the F1 -ATPase; (d) highly conserved glycine residues are found to be involved in the coordination of phosphate groups in the active centres of most mononucleotide- and dinucleotide-binding proteins of known structure; (e) a subset of ABC transporters which appears to contain a degenerate active site in one NBD subunit carry a non-consensus C motif in the opposite NBD subunit; (f) while it has been postulated that the C motif may interact directly with the TMDs [29], it might be expected that a complementary highly conserved region would be found in these domains. Despite extensive studies of ABC transporters, such a region is yet to be identi¢ed. Manual alignment of the atomic coordinates of the HisP subunits to approximate this alternative con¢guration reveals aspects of the dimer which are consistent with experimental data from a number of ABC transporters and other nucleotide-binding proteins. This alternative dimer provides insights into the nature of NBD subunit interactions and cooperativity, and into the mode of interaction of the C motif with the bound Mg2 -ATP complex. Incorporation of this NBD dimer into our recently published L-barrel model for Pgp produces an outline for the architecture of the complete ABC transporter. This architecture is consistent with a mechanism for the transport process that accommodates extensive mutational analyses of the NBDs identify-
Magenta
Geel
Zwart
200
P.M. Jones, A.M. George / FEMS Microbiology Letters 179 (1999) 187^202
ing residues in HisP and other ABC transporters involved in the interaction of the membrane-integral domains with the NBDs.
Acknowledgements We are indebted to Drs Sung-Hou Kim and Giovanna Ferro-Luzzi Ames and their co-workers for generously providing the atomic coordinates of the HisP crystal structure ahead of its availability at the Brookhaven Protein Databank. This work was supported by a grant from the Clive and Vera Ramaciotti Foundation.
References [1] Higgins, C.F. (1992) ABC transporters : From microorganisms to man. Annu. Rev. Cell Biol. 8, 67^113. [2] Ames, G.F.-L., Mimura, C.S. and Shyamala, V. (1990) Bacterial periplasmic permeases belong to a family of transport proteins operating from Escherichia coli to human: Tra¤c ATPases. FEMS Microbiol. Rev. 75, 429^446. [3] Blattner, F.R., Plunket III, G., Bloch, C.A., Perna, N.T., Burland, V., Riley, M., Collado-Vides, J., Glasner, J.D., Rode, C.K., Mayhew, G.F., Gregor, J., Davis, N.W., Kirkpatrick, H.A., Goeden, M.A., Rose, D.J., Mau, B. and Shao, Y. (1997) The complete genome sequence of Escherichia coli K12. Science 277, 1453^1474. [4] Tatusov, R.L., Koonin, E.V. and Lipman, D.J. (1997) A genomic perspective on protein families. Science 278, 631^637. [5] Ames, G.F.-L., Mimura, C., Holbrook, S. and Shyamala, V. (1992) Tra¤c ATPases: a superfamily of transport proteins operating from Escherichia coli to humans. Adv. Enzymol. 65, 1^47. [6] Doige, C.A. and Ames, G.F.-L. (1993) ATP-dependent transport systems in bacteria and humans: Relevance to cystic ¢brosis and multidrug resistance. Annu. Rev. Microbiol. 47, 291^319. [7] Fath, M.J. and Kolter, R. (1993) ABC transporters: Bacterial exporters. Microbiol. Rev. 57, 995^1017. [8] Dean, M. and Allikmets, R. (1995) Evolution of ATP-binding cassette transporter genes. Curr. Biol. 5, 779^785. [9] Schneider, E. and Hunke, S. (1998) ATP-binding cassette (ABC) transport systems: Functional and structural aspects of the ATP-hydrolysing subunits/domains. FEMS Microbiol. Rev. 22, 1^20. [10] Croop, J.M. (1998) Evolutionary relationships among ABC transporters. Methods Enzymol. 292, 101^116. [11] Schwiebert, E.M. (1999) ABC transporter-facilitated ATP conductive transport. Am. J. Physiol. Cell Physiol. 276, C1^ C8.
FEMSLE 8995 30-9-99 Cyaan
[12] Saurin, W., Hofnung, M. and Dassa, E. (1999) Getting in or out: Early segregation between importers and exporters in the evolution of ATP-binding cassette (ABC) transporters. J. Mol. Evol. 48, 22^41. [13] Hyde, S.C., Emsley, P., Hartshorn, M.J., Mimmack, M.M., Gileadi, U., Pearce, S.R., Gallagher, M.P., Gill, D.R., Hubbard, R.E. and Higgins, C.F. (1990) Structural model of ATPbinding proteins associated with cystic ¢brosis, multidrug resistance and bacterial transport. Nature 346, 362^365. [14] Bellamy, W. and Weinstein, R.S. (1994) Functional diversity in the multidrug resistance gene family. Lab. Invest. 71, 716^ 720. [15] Bolhuis, H., van Veen, H.W., Poolman, B., Driessen, A.J.M. and Konings, W.N. (1997) Mechanisms of multidrug transporters. FEMS Microbiol. Rev. 21, 55^84. [16] Hrycyna, C.A. and Gottesman, M.M. (1998) Multidrug ABC transporters from bacteria to man: an emerging hypothesis for the universality of molecular mechanism and function. Drug Res. Updat. 1, 81^83. [17] van Veen, H.W. and Konings, W.N. (1998) The ABC family of multidrug transporters in microorganisms. Biochim. Biophys. Acta 1365, 31^36. [18] Ehrmann, M., Ehrle, R., Hofmann, E., Boos, W. and Schlosser, A. (1998) The ABC maltose transporter. Mol. Microbiol. 29, 685^694. [19] Me¨ndez, C. and Salas, J.A. (1998) ABC transporters in antibiotic-producing actinomycetes. FEMS Microbiol. Lett. 158, 1^8. [20] van Veen, H.W., Vanema, K., Bolhuis, H., Oussenko, I., Kok, J., Poolman, B., Driessen, A.J.M. and Konings, W.N. (1996) Multidrug resistance mediated by a bacterial homolog of the human multidrug transporter MDR1. Proc. Natl. Acad. Sci. USA 93, 10668^10672. [21] van Veen, H.W., Callaghan, R., Soceneantu, L., Sardini, A., Konings, W.N. and Higgins, C.F. (1998) A bacterial antibiotic-resistance gene that complements the human multidrugresistance P-glycoprotein gene. Nature 391, 291^295. [22] Gerlach, J.H., Endicott, J.A., Juranka, P.F., Henderson, G., Sarangi, F., Deuchlars, K.L. and Ling, V. (1986) Homology between P-glycoprotein and a bacterial haemolysin transport protein suggests a model for multidrug resistance. Nature 324, 485^489. [23] Chen, C.-J., Chin, J.E., Ueda, K., Clark, D.P., Pastan, I., Gottesman, M.M. and Roninson, I.B. (1986) Internal duplication and homology with bacterial transport proteins in the mdr1 (P-glycoprotein) gene from multidrug-resistant human cells. Cell 47, 381^389. [24] Gros, P., Croop, J. and Housman, D. (1986) Mammalian multidrug resistance gene : Complete cDNA sequence indicates strong homology to bacterial transport proteins. Cell 47, 371^380. [25] Bibi, E. and Beja, O. (1994) Membrane topology of multidrug resistance protein expressed in Escherichia coli. J. Biol. Chem. 269, 19910^19915. [26] Skach, W.R., Calayag, M.C. and Lingappa, V.R. (1993) Evidence for an alternate model of human P-glycoprotein structure and biogenesis. J. Biol. Chem. 268, 6903^6908.
Magenta
Geel
Zwart
P.M. Jones, A.M. George / FEMS Microbiology Letters 179 (1999) 187^202 [27] Zhang, J.-T. and Ling, V. (1995) Involvement of cytoplasmic factors regulating the membrane orientation of P-glycoprotein sequences. Biochemistry 34, 9159^9165. [28] Jones, P.M. and George, A.M. (1998) A new structural model for P-glycoprotein. J. Membr. Biol. 166, 133^147. [29] Hung, L.-W., Wang, I.X., Nikaido, K., Liu, P.-Q., Ames, G.F.-L. and Kim, S.-H. (1998) Crystal structure of the ATP-binding subunit of an ABC transporter. Nature 396, 703^707. [30] Pai, E.F., Krengel, U., Petsko, G.A., Goody, R.S., Kabsch, W. and Wittinghofer, A. (1990) Re¢ned structure of the triî resolution: phosphate conformation of H-ras-p21 at 1.35 A implications for the mechanism of GTP hydrolysis. EMBO J. 9, 2351^2359. [31] Story, R.M., Weber, I.T. and Steitz, T.A. (1992) The structure of the E. coli recA protein monomer and polymer. Nature 355, 318^325. [32] Diederichs, K. and Schultz, G.E. (1990) Three-dimensional structure of the complex between the mitochondrial matrix adenylate kinase and its substrate AMP. Biochemistry 29, 8138^8144. [33] Smith, C.A. and Rayment, I. (1996) Active site comparisons highlight structural similarities between myosin and other Ploop proteins. Biophys. J. 70, 1590^1602. [34] Abrahams, J.P., Leslie, A.G.W., Lutter, R. and Walker, J.E. î resolution of F1 -ATPase from bo(1994) Structure at 2.8 A vine heart mitochondria. Nature 370, 621^628. [35] Walker, J.E., Saraste, M., Runswick, M.J. and Gay, N.J. (1982) Distantly related sequences in the K- and L-subunits of ATP synthase, myosin, kinases and other ATP-requiring enzymes and a common nucleotide-binding fold. EMBO J. 1, 945^951. [36] Saraste, M., Sibbald, P.R. and Wittinghofer, A. (1990) The Ploop ^ a common motif in ATP- and GTP-binding proteins. Trends Biochem. Sci. 15, 430^434. [37] Bianchet, M.A., Ko, Y.H., Amzel, L.M. and Pedersen, P.L. (1997) Modeling of nucleotide-binding domains of ABC transporter proteins based on a F1 -ATPase/recA topology: Structural model of the nucleotide-binding domains of the cystic ¢brosis transmembrane conductance regulator (CFTR). J. Bioenerg. Biomembr. 29, 503^524. [38] Manavalan, P., Dearborn, D.G., McPherson, J.M. and Smith, A.E. (1995) Sequence homologies between nucleotide-bindingregions of CFTR and G-proteins suggest structural and functional similarities. FEBS Lett. 366, 87^91. î [39] Noel, J.P., Hamm, H.E. and Sigler, P.B. (1993) The 2.2 A crystal structure of transducin-K complexed with GTPQS. Nature 366, 654^663. [40] Kjeldgaard, M., Nissen, P., Thirup, S. and Nyborg, J. (1993) The crystal structure of elongation factor EF-Tu from Thermus aquaticus in the GTP conformation. Structure 1, 35^ 50. [41] Mixon, M.B., Lee, E., Coleman, D.E., Berghuis, A.M., Gilman, A.G. and Sprang, S.R. (1995) Tertiary and quaternary structural changes in GiK1 induced by GTP hydrolysis. Science 270, 954^960. [42] Park, S., Ajtai, K. and Burghardt, T.P. (1997) Mechanism for
FEMSLE 8995 30-9-99 Cyaan
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
[52]
[53]
[54]
[55]
[56]
201
coupling free energy in ATPase to the myosin active site. Biochemistry 36, 3368^3372. Weber, J., Hammond, S.T., Wilke-Mounts, S. and Senior, A.E. (1998) Mg2 coordination in catalytic sites of F1 -ATPase. Biochemistry 37, 608^614. Houseman, A.L.P., LoBrutto, R. and Frasch, W.D. (1995) E¡ects of nucleotides on the protein ligands to metals at the M2 and M3 metal-binding sites of the spinach chloroplast F1 ATPase. Biochemistry 34, 3277^3285. MacFarland, K.J., Shan, Q., Inman, R.B. and Cox, M.M. (1997) RecA as a motor protein. J. Biol. Chem. 272, 17675^ 17685. Mimura, C.A., Holbrook, S.R. and Ames, G.F-L. (1991) Structural model of the nucleotide-binding conserved component of periplasmic permeases. Proc. Natl. Acad. Sci. USA 88, 84^88. Beaudet, L. and Gros, P. (1995) Functional dissection of Pglycoprotein nucleotide-binding domains in chimeric and mutant proteins. J. Biol. Chem. 270, 17159^17170. Beaudet, L., Urbatsch, I.L. and Gros, P. (1998) Mutations in the nucleotide-binding sites of P-glycoprotein that a¡ect substrate speci¢city modulate substrate-induced adenosine triphosphatase activity. Biochemistry 37, 9073^9082. Mourez, M., Je¨hanno, M., Schneider, E. and Dassa, E. (1998) In vitro interaction between components of the inner membrane complex of the maltose ABC transporter of Escherichia coli: modulation by ATP. Mol. Microbiol. 30, 353^363. Hoof, T., Demmer, A., Hadam, M.R., Riordan, J.R. and Tummler, B. (1994) Cystic ¢brosis-type mutational analysis in the ATP-binding cassette transporter signature of human P-glycoprotein MDR1. J. Biol. Chem. 269, 20575^20583. Bakos, Eè., Klein, I., Welker, E., Szabo¨, K., Mu«ller, M., Sarkadi, B. and Va¨radi, A. (1997) Characterisation of the human multidrug resistance protein containing mutations in the ATPbinding cassette signature region. Biochem. J. 323, 777^ 783. Szabo¨, K., Szaka¨cs, G., Hegedu«s, T. and Sarkadi, B. (1999) Nucleotide occlusion in the human cystic ¢brosis transmembrane regulator. J. Biol. Chem. 274, 12209^12212. Nikaido, K., Liu, P.-Q. and Ames, G.F.-L. (1997) Puri¢cation and characterisation of HisP, the ATP-binding subunit of a tra¤c ATPase (ABC transporter), the histidine permease of Salmonella typhimurium. Solubilisation, dimerisation and ATPase activity. J. Biol. Chem. 272, 27745^27752. Liu, C.E., Liu, P.-Q., Wolf, A., Lin, E. and Ames, G.F.-L. (1999) Both lobes of the soluble receptor of the periplasmic histidine permease, an ABC transporter (tra¤c ATPase), interact with the membrane-bound complex. J. Biol. Chem. 274, 739^747. Mourez, M., Hofnung, M. and Dassa, E. (1997) Subunit interactions in ABC transporters: a conserved sequence in hydrophobic membrane proteins of periplasmic permeases de¢nes an important site of interaction with the ATPase subunits. EMBO J. 16, 3066^3077. Shani, N., Sapag, A. and Valle, D. (1996) Characterisation and analysis of conserved motifs in a peroxisomal ATP-binding cassette transporter. J. Biol. Chem. 271, 8725^8730.
Magenta
Geel
Zwart
202
P.M. Jones, A.M. George / FEMS Microbiology Letters 179 (1999) 187^202
[57] Wierenga, R.K., De Maeyer, M.C.H. and Hol, W.G.J. (1985) Interaction of pyrophosphate moieties with K-helixes in dinucleotide-binding proteins. Biochemistry 24, 1346^1357. [58] Hol, W.G.J. (1985) The role of the K-helix dipole in protein function and structure. Prog. Biophys. Mol. Biol. 45, 149^195. [59] Bernstein, B.E., Michels, P.A.M. and Hols, W.G.J. (1997) Synergistic e¡ects of substrate-induced conformational changes in phosphoglycerate kinase activation. Nature 385, 275^278.
FEMSLE 8995 30-9-99 Cyaan
[60] Senior, A.E. (1998) Catalytic mechanism of P-glycoprotein. Acta Physiol. Scand. Suppl. 163, 213^218. è ., Mu«ller, M., Roninson, I., [61] Szabo¨, K., Welker, E., Bakos, E Va¨radi, A. and Sarkadi, B. (1998) Drug-stimulated nucleotide trapping in the human multidrug transporter MDR1. J. Biol. Chem. 273, 10132^10238. [62] Evans, S.V. (1993) Setor ^ Hardware-lighted three-dimensional solid model representations of macromolecules. J. Mol. Graph. 11, 134^138.
Magenta
Geel
Zwart