Encyclopedia of Biological Chemistry - Vol_1.pdf - med - pet300

ABC Transporters

Andr ´ Goffeau, Beno ˆ t De Hertogh and Philippe V. Baret

Universit´ Catholique de Louvain, Louvain-la-Neuve, Belgium

binding receptor, the cytoplasmic NDB, and the mem-

brane TMD are believed to have arisen from a common

ancestral ABC transporter in which these three proteins

were already present. However, during evolution, the

sequence of the periplasmic solute-binding receptors

diverges more rapidly than that of the TMDs, while that

of NBDs is the least divergent. Thus, all NBDs

are homologous, but this is not true for the TMDs or

the receptors. Nevertheless, the phylogenetic clustering

patterns in bacterial ABC from different species are

generally the same for all three types of proteins, despite

their variable rate of evolution.

The topology of some eukaryotic ABC efﬂuxers can

be complex as additional TM spans occur in some

systems ( Figure 3 ) as well as extra cytoplasmic domains

of presumed regulatory function.

The ABC proteins constitute the largest family of proteins.

They are present in all living species from Archaea to Homo

sapiens . They make up to 4% of the full genome complement

of bacteria such as Escherichia coli or Bacillus subtilis . Each

eukaryote genome contains several dozens of members (over

100 in the plant Arabidopsis thaliana ). They are recognized by

a consensus ATP-binding region of approximately 100 amino

acids which include the two Walker A and B motifs

encompassing a linker or C region ( Figure 1 ). The ABC

proteins catalyze a wide variety of physiological functions,

most (but not all) of which being related to transport. This

article describes the major physiological and biochemical

functions as well as the structural properties of some of the

best-known ABC transporters using examples from the yeast

Saccharomyces cerevisiae and Homo sapiens .

Topology

Phylogeny

Most, but not all, ABC proteins are ABC transporters.

Each of those molecules contains, or is associated to, one

or two cytoplasmic ATP-binding domains named nucleo-

tide binding domains (NBDs) ( Figure 1 ) and one or two

transmembrane domains (TMDs) ( Figure 2 ). Each TMD

comprises usually six a-helix spans. Association of one

TMD to one NBD results in a half-size ABC transporter;

however, they are believed to function as homo- or

heterodimers so that the minimal functional organization

of an ABC transporter is considered to be TMD – NBD –

TMD – NBD or NBD – TMD – NBD – TMD. In eukar-

yotes, two TMDs and two NBDs are often associated in

one single molecule called full-sized ABC transporter.

The topological relation between NBD(s) and TMDs is

variable ( Figure 2 ). In bacteria two NBDs often associate

with two TMDs either as four single subunits encoded by

the same operon or in various combinations of fused

subunits. Association of other proteins may occur. The

most prominent associated bacterial protein is the

periplasmic solute-binding receptor, which in gram-

negative bacteria is found in the periplasm, and in

gram-positive bacteria is present often as a lipoprotein,

bound to the external membrane surface via electrostatic

interactions ( Figure 3 ). The three domains of the bacterial

ABC uptake transporters: namely the periplasmic

The different families of ABC proteins transport a wide

variety of substrates against their concentration gradient

using the energy of ATP hydrolysis carried out by NBD.

In bacteria, the transported substrates are either

imported in or exported out of the cell. In eukaryotes,

only extracytoplasmic exporters (transporting sub-

strates either out of the cell or into organelles) are

known up to now. Within the ABC superfamily, 61

phylogenetic families have been identiﬁed so far. These

families generally correlate with substrate speciﬁcity.

Their classiﬁcation based both on functional and

phylogenetic criteria has been carried out within the

transporter classiﬁcation (TC) system developed by

Milton Saier in San Diego. The TC system has recently

been endorsed by The International Union of the

Biochemical and Molecular Biology Societies. In the

TC system, prokaryotic ABC inﬂux porters comprise 22

phylogenetic families including histidine permease, the

ﬁrst ABC transporter to be cloned and sequenced in the

laboratory of Giovanna Ames in 1982. Another famous

example is the MalEFGK operon classiﬁed in TC as a

maltooligosaccharide porter within “the carbohydrate

uptake transporter-1 (CUT1) family,” and given the TC

digit 3.A.1.1. In this operon, MalE is the receptor, MalF

ABC TRANSPORTERS

nontransporter proteins, have been classiﬁed in seven

families named ABCA to ABCG according to topologi-

cal and phylogenetical criteria that are less stringent

than those used by TC. In its present form, the HUGO

nomenclature and classiﬁcation are difﬁcult to use for

identiﬁcation of novel ABC from non-mammalian

species. For instance, the Saccharomyces cerevisiae

genome contains 32 ABC genes among which 22 (16

full-size and six small-size) are associated to transmem-

brane domains in four different topologies. Its largest

family is the full-sized Pdr5p-like family identiﬁed in

1996 by Anabelle Decottignies and Andr´ Goffeau and

shown later to be present in all fungi and plants. This

family is not detected in the animal kingdom. Con-

versely, the large human and mouse ABCA subfamily is

not represented in yeast genomes. There is a necessity to

adopt a consistent classiﬁcation system, which combines

the TC and HUGO nomenclatures.

FIGURE 1 The consensus ATP-binding region of a typical ABC

protein is made of approximately 100 amino acids (aa), including both

Walker A and B motifs and the linker C region.

and MalG are distinct TMD subunits, and MalK is a

double NBD.

The prokaryotic efﬂuxers comprise 27 families

including the multidrug exporter LmrA from the

gram-positive Lactococcus lactis well studied by Will

Koning and belonging to “the drug exporter-2 family”

(3.A.1.117).

The eukaryotic ABCs can be grouped in only 12 efﬂux

families including the famous MDR1 also named Pgp

(permease-glycoprotein), discovered in 1986 by Ira

Pastan, Michael Gottesman and colleagues, and shown

to be involved in MDR of chemiotreated tumor cells.

In the TC system, this ABC exporter is classiﬁed in “the

multidrug resistance exporter family” (3.A.1.201).

The TC system is redundant with the Human

Genome Organization (HUGO) classiﬁcation adopted

by the scientiﬁc community working on mammalian

objects (mouse or man). The 45 human or mouse

ABC proteins, comprising efﬂux transporters and

Function and Diseases

The immense variety of substrates transported in

bacteria is reﬂected by the identiﬁcation of 49 phyloge-

netic ABC families including 22 inﬂux protein com-

plexes and 27 efﬂux transporter systems. As they belong

to Archaea, gram-negative and gram-positive plasma

membranes that are widely different in organization and

composition, the number and nature of proteins

associated to given ABC transporters are variable and

their transport mechanisms may be partly different. In

bacterial and Archaea ABC, the variety of substrates:

sugars, amino acids, lipids, ions, polysaccharides,

peptides, proteins, toxins, drugs, antibiotics, xenobiotics

and other metabolites is reﬂected by the divergence of

the periplasmic sensor and that of the TMD, which must

control both speciﬁcity of substrate and part of the

coupling mechanism.

Even if all eukaryotic ABC transporters are efﬂuxers

that comprise subunits in which each TMD is fused to a

NBD, some of them are not directly involved in moving

substrates. For instance, in the cystic ﬁbrosis transmem-

brane regulator CFTR, and in the sulfonylurea receptor

SUR, the hydrolysis of ATP appears to be linked to the

regulation of opening and closing of ion channels carried

by the ABC protein itself or other proteins ( Figure 3B ).

The conservation of NBD in all ABC transporters,

however, suggests that a basic coupling mechanism

exists for efﬂux and inﬂux whatever the transported

substrate. Moreover, distantly related proteins exist

which utilize an NBD to drive diverse nontransport

processes such as DNA repair or protein-elongation or

regulation of RNAse activities.

The 32 yeast ABC proteins are in principle easy to

study, as sensitive genetic tools are available. However,

only a few successful cases of overexpressions and

Out

NBD

NBD1

NBD2

Out

NBD1

NBD2

NBD

FIGURE 2 Example of topological relations between NBDs and

transmembrane spans in full-sized and half-sized ABC transporters.

ABC TRANSPORTERS

Out

Periplasmic-binding

receptor

Out

Substrate

Membrane

Transmembrane domain

N-terminal

extension

NBD1

NBD2

Cytoplasmic nucleotide-

binding domain

NBD1

NBD2

Pore

subunit

FIGURE 3

Example of proteins associated to bacterial and mammalian ABC transporters.

in vitro UTPase or ATPase measurements of puriﬁed

ABCs have been reported. Only one puriﬁed yeast ABC

transporter, the pleiotropic drug resistance efﬂuxer

Pdr5p, has been submitted to structural studies.

The biochemical study of human ABC transporters is

often more advanced than that of the yeasts. The Pgp

protein responsible for multiple drug resistance (MDR)

in human cells is especially well studied. One strong

impetus for the study of mammalian ABC transporters is

their involvement in diseases. Many mendelian diseases

and complex genetic disorders are caused by ABC

transporters including cystic ﬁbrosis, adrenoleuko-

dystrophy, Stargardt disease, Tangier disease, immune

deﬁciencies, progressive familial intrahepatic cholesta-

sis, Dublin – Johnson syndrome, Pseudoxanthoma elas-

ticum, persistent hyperinsulinemic hypoglycemia of

infancy due to focal adenomatous hyperplasia, X-linked

sideroblastosis and anemia, age-related macular

degeneration, familial hypoapoproteinemia, Fundus

ﬂavimaculatis, Retinitis pigmentosum, cone rod dystro-

phy etc. Cell lines isolated from diseased tissues allow

molecular study of the involved ABC transporter.

Moreover, a variety of drug-resistant cell lines is

available from MDR or MDR-related protein (MRP)

tissues. Basic studies of human ABC transporters would

greatly beneﬁt from heterologous expression of human

ABC transporter genes in yeast or other cells, but this

technology is far from being satisfactory yet. Mean-

while, knockout technology in the mouse may be needed

to begin to understand the molecular and physiological

functions of the mammalian transporters.

bound at their interface. Each nucleotide-binding site

comprises a Walker A motif from monomer 1 and the C

motif from monomer 2. This results from a “head-to-

tail” arrangement of the two interacting monomers.

This is supported by biochemical arguments and is

coherent with the cooperative hydrolysis for ATP

hydrolysis observed with MalK.

More recently, three structures of complete dimeric

ABC transporters comprising both NDB and TMD were

obtained: that of the presumed phospholipid ﬂippase

MsbA from E. coli and Vibrio cholera and that of the

vitamin B12 importer BtuCD from E. coli .The

structures obtained were dissimilar, which may not be

too surprising taking into account the different con-

ditions used, the different numbers of TM spans and the

different functions (import or export) of the proteins

analyzed. No generalization can be made, for instance,

on the angle between the TM spans and the membrane

plane or on the identiﬁcation of the interaction domains

between the TM spans. The nature of communication

between the NBD and TMD is variable and carried out

either through the long and complex so-called intra-

cellular domain named ICD in MsbA, or through a short

L -shaped linker between the transmembrane spans 6 and

7 in BtuCD. The nature, the size, the orientation, and the

location of the so-called chamber (or water channel, or

pore, or cone) presumed to be involved in substrate

binding are also variable. No consensus interaction

points between the NBDs (open or closed conformation)

were observed. Obviously, more structures are needed

on several transporters carrying out similar functions,

such as drug efﬂux, for instance, to clarify these issues

and to reach a consensus interpretation of the basic

structural elements involved in the transport and in the

coupling mechanism.

In contrast, recent analyses at the electron

microscopy level of a bacteria (BmrA or YccV from

Bacillus subtilis ) and a yeast (Pdr5p) drug efﬂux ABC

transporter came to a remarkably coherent set of

conclusions. In both cases, the basic structural unit

seems to comprise four joining NBDs (that corresponds

to two full-size Pdr5p or four half-size BmrA), which are

Structure and Biochemical

Mechanism

In 1998, the ﬁrst high-resolution structure of a NBD,

that from the histidine ABC importer HisP, was

reported. Five years later, about six related structures

were available and a consensus view emerged. NBDs are

organized as dimers and two molecules of ATP are

ABC TRANSPORTERS

transports hundreds of different chemicals, apparently

contradicting the famous key/slot concept described in

all enzymology textbooks. It must be recognized that the

identiﬁcation of presumed ABC substrates are often

based on indirect data such as resistance or sensitivity of

mutants or drug-induced transcription proﬁles. These

measurements should in principle be corroborated

by direct transport measurements, which are often

difﬁcult to obtain.

Regarding speciﬁcity of transport, one of the best-

studied yeast transporters is Pdr5p through the capacity

of its deleted mutants to gain growth sensitivity to an

amazing variety of xenobiotics. From a large screen of

several hundreds of toxic compounds, no chemical

determinants for transported substrate speciﬁcity could

be identiﬁed among a wide range of compounds

including the fungicides anilinopyrimidines, benzimida-

zoles, benzenedicarbonitriles, dithiocarbamates, guani-

dines, imidothiazoles, polyenes, pyrimidynyl carbinols,

and strobilurine analogues, the urea derivative and

anilide herbicides, a wide collection of ﬂavonoids and

steroids, several membrane lipids resembling detergents,

and newly synthesized lysosomotropic aminoesters.

However, it could be concluded that Pdr5p shows

considerable substrate overlap with members of the

same ABC phylogenetic family or even with yeast drug

efﬂuxers from other families such as Snq2p or Yor1p.

This was demonstrated by the numerous cases showing

full sensitivity only in double or triple mutants. The most

promiscuous substrates were: itraconazole, miconazole,

nystatin, antimycin, nigericidin, and tetradodecylam-

monium bromide, which are transported by at least

three distinct yeast ABC transporters. In contrast many

substrates show relative speciﬁcity for a given pleio-

tropic drug resistance exporter. Prominent examples

include cycloheximide, benomyl, ﬂuxilazole, nuarimal,

and soraphen for Pdr5p. From a ﬁrst large scale screen it

was concluded that Pdr5p’s most efﬁcient substrates

were valinomycin, the antifungal azoles and rhodamine

6G. These compounds are inhibiting the growth of a

sensitive PDR5 deletant at concentrations below the

micromolar range. Up to now, the most potent

competitive inhibitor of binding of rhodamine 6G to

the yeast Pdr5p is the oestradiol derivative RU 49953,

which exhibits a K i of 23 nM.

Other recent systematic screens have provided more

precise chemical information on the determinants of

Pdr5p speciﬁcity. From such studies it was concluded

that Pdr5p is capable of transporting substrates that

neither ionize nor have electron pair donors and that are

much simpler in structure than those handled by the

human Pgp. The substrate optimum surface volume is

about 200 ˚ 3 . Analysis of the interactions between

imidazole derivatives, organotin and other compounds

argues, that, as also established for Pgp, the Pdr5p

comprises at least two substrate-binding sites. One site

Out

ATP

ADP

ATP

Drug

ADP

ATP

ADP

FIGURE 4 The alternating catalytic sites hypothesis for P-glyco-

protein, according to Alan Senior. The NBDs have three ligands and

four states: free, ATP bound, ADP þ Pi bound, and ADP bound. They

alternate in such a way that two ATPs never bind simultaneously. The

binding of one ATP to one NBD induces hydrolysis at the other ATP.

The drug is transported out during Pi release.

related to the TMDs through four distinct stalks.

Each NBD is oriented at a ﬁxed 908 angle relative to

its neighbor NBDs. This raises the possibility of

concerted rotation movements of the NBDs implying a

certain ﬂexibility of the stalks. No intramolecular or no

intramembrane pores were observed even though there

is room (or chamber) between the four stalks that join

together at their NBDs tips.

These latter observations are difﬁcult to reconcile

with the mechanism of alternating sites established for

the drug exporters, the human Pgp and the bacterial

LmrA ( Figure 4 ). In these cases the monomeric NBDs are

similar and both are able to hydrolyze ATP. Hydrolysis

of ATP at one NBD is believed to be responsible for drug

release outside. Upon release of ADP and Pi, the other

NBDs bind and then hydrolyze ATP while the drug is

taken up inside. However, the simplest version of this

“two-cylinder” mechanism cannot function when the

two NBDs partners are not equivalent in a dimeric

arrangement as it is the cases of yeast Pdr5p and human

CFTR, where one of the two NBDs from a full-sized

ABC molecule might be unable to hydrolyze ATP.

The Substrate Speciﬁcity of the

ABC Multidrug Exporters

One of the most intriguing contemporary biochemical

problems is the characterization of the interactions of

the TMDs with their transported substrate. The most

extraordinary feature in this context is the apparent lack

of speciﬁcity of the yeast Pdr5p and human Pgp that

ABC TRANSPORTERS

might use only hydrophobic binding interactions. Some

substrates may bind to two sites, others associate more

speciﬁcally to only one site. However, the Pdr5p

substrates-binding sites, behave differently from those

of Pgp. This concept of overlapping substrate-binding

sites may reconcile many previous observations con-

cerning the substrate broad speciﬁcity of Pdr5p and Pgp,

which up to now appeared contradictory.

17 amino acids. In ABC transporters, the transmembrane domains

usually comprises six continuous transmembrane spans.

transporter classiﬁcation (TC) Classiﬁcation of over 800 transporters

families, developed by Milton Saier (UCSD), based on a combina-

tion of mechanistic and phylogenetic criteria.

Walker A and B Small consensus of amino acid sequences involved

in ATP binding.

F URTHER R EADING

Dean, M. (2002). The Human ATP-Binding Cassette (ABC) Trans-

porter Superfamily . Monograph Bethesda (MD), NCBI, National

Library of Medicine (US).

Decottignies, A., and Goffeau, A. (1997). Complete inventory of the

yeast ABC proteins. Nat. Genet. 15, 137 – 145.

Gottesman, M. M., and Ambudkar, S. V. (2001). Overview: ABC trans-

porters and human disease. J. Bioenerg. Biomembr. 33, 438 – 453.

Hipfner, D. R., Deeley, R. G., and Cole, S. P. C. (1999). Structural

mechanistic and clinical aspects of MRP1. Biochim. Biophys. Acta

1461, 359 – 376.

Martinoia, E., Klein, M., Geisler, M., Bovet, L., Forestier, C.,

Kolukisaoglu, U., Muller-Rober, B., and Schulz, B. (2002). Plant

ABC transporters: more than just detoxiﬁers. Planta 214,

345 – 355.

Schmitt, L., and Tamp´, R. (2002). Structure and mechanism of ABC

transporters. Curr. Opin. Struct. Biol. 12, 754 – 760.

Senior, A. E., al-Shawi, M. K., and Urbatsch, I. L. (1995). The catalytic

cycle of P-glycoprotein. FEBS Lett. 27, 285 – 289.

Conclusion

The present frontiers in the study of ABC transporters

are challenging. The evolutionary history of this large

ubiquitous family has to be unraveled. Additional

atomic structures have to be produced. Better heter-

ologous overexpression systems have to be developed to

allow further biochemical studies. Speciﬁc inhibitors of

drug (and other) ABC exporters have to be screened for.

The physiological mechanisms of ABC-linked diseases

have to be further studied in mouse knockouts. Systems

prone to speciﬁc inhibition of ABC transporters

expression by interfering RNA have to be explored.

Genetic therapy has to be developed.

S EE A LSO THE F OLLOWING A RTICLE

MDR Membrane Proteins

B IOGRAPHY

Andr´ Goffeau is Emeritus Professor at the Universit´ Catholique de

Louvain (Belgium), Department of Physiological Chemistry. He

investigates the P-type ATPases and the ABC transporters in

S. cerevisiae . He has initiated and organized the sequencing of the

yeast genome. He is interested in the phylogenetic classiﬁcation of yeast

membrane proteins.

G LOSSARY

human genome organization (HUGO) An international association

that comprises scientists involved in all aspects of the sequence of

the human genome and its analysis.

linker C Small amino acids sequence signature involved in ATP

binding of ABC proteins.

Pdr5p The major yeast ABC transporter, involved in pleiotropic drug

resistance.

Pgp The ﬁrst ABC transporter (glycoprotein) shown to be involved in

mammalian multidrug resistance.

transmembrane span and transmembrane domain (TMS and

TMD) The existence of the transmembrane span is predicted by

frequency of hydrophobic residues in a reading window of about

Benoˆt De Hertogh is Ing´nieur Agronome at the Universit´ Catholique

de Louvain (Belgium), Department of Quantitative Genetics. He is

completing a Ph.D. thesis on population genetics on the genetical basis

of fertility in livestock. He is interested in comparative genomics of

membrane proteins.

Philippe V. Baret is Associate Professor at the Universit ´ Catholique de

Louvain (Belgium), Department of Quantitative Genetics. He is model-

ing the relationship between genes and phenotypes, with a peculiar

interest in quantitative trait loci mapping and cladistic approaches.

Encyclopedia of Biological Chemistry - Vol_1.pdf

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: