Adenosine transporters

ISSN 0306-3623/96 $15.00 + .00 SSDI 0306-3623(95)02053-5 All rights reserved

Gen. Pharmac. Vol. 27, No. 4, pp. 613-620, 1996 Copyright © 1996 Elsevier Science Inc. Printed in the USA. ELSEVIER

REVIEW

Adenosine Transporters James A. Thorn and Simon M. Jarvis* RESEARCH SCHOOL OF BIOSCIENCES, UNIVERSITY OF KENT, CANTERBURY, KENT, C T 2 7NJ, U.K.

TEL: 01227-827581; FAX: 01227-763912; E-MAlLS.M.JARvIS~UKC.AC.UK A B S T R A C T . 1. I n m a m m a l s , nucleoside transport is an important determinant ofthe pharmacokinetics, plasma and tissue concentration, disposition and in v i v o biological activity of adenosine as well as nucleoside analogues used in antiviral and anticancer therapies. 2. T w o broad types of adenosine transporter exist, facilitated-diffusion carriers and active processes driven by the transmembrane sodium gradient. 3. Facilitated-diffusion adenosine carriers may be sensitive (es) or insensitive (el) to nanomolar concentrations of the transport inhibitor nitrobenzylthioinosine (NBMPR). Dipyridamole, dilazep and lidoflaz. ine analogues are also more potent inhibitors of the es carrier than the ei transporter in cells other than those derived from rat tissues. 4. T h e es transporter has a broad substrate specificity (apparent Km for adenosine - 25 ~tM in many ceils at 25°C), is a glycoprotein with an average apparent M, of 57,000 in h u m a n erythrocytes that has been purified to near homogeneity and may exist i n situ as a dimer. However, there is increasing evidence to suggest the presence of isoforms of the es transporter in different cells and species, based on kinetic and molecular properties. 5. T h e ei transporter also has a broad substrate specificity with a lower affinity for some nucleoside permeants t h a n the es carrier, is genetically distinct from es but little information exists as to the molecular properties of the protein. 6. Sodium-dependent adenosine transport is present in many cell types and catalysed by four distinct systems, N1-N4, distinguished by substrate specificity, sodium coupling and tissue distribution. 7. T w o genes have been identified which encode sodium-dependent adenosine transport proteins, SNST 1 from the sodium/glucose cotransporter (SGLT 1) gene family and the rat intestinal N 2 transporter (cNT 1) from a novel gene family including a bacterial nucleoside carrier (NupC). Transcripts of cNT 1, which encodes a 648-residue protein, are found in intestine and kidney only. 8. Success in cloning the remaining adenosine transporter genes will improve our understanding of the diversity of nucleoside transport processes, with a view to better targeting of therapeutic nucleoside analogues a n d protective use of transport inhibitors. GEN PHARMAC27;4:613-620, 1996. KEY WORDS. Nucleoside transport, adenosine, nitrobenzylthioinosine, dipyridamole, sodium cotransporters, cNT1

INTRODUCTION Adenosine exerts profound effects in many tissues, organs and species and most of these actions are mediated via specific receptors (for review see Stiles, 1992; Jacobson et al., 1992). As such, adenosine has been a target for the development of new drugs. For example, adenosine is used in the treatment of paroxysmal supraventricular tachycardia and adenosine agonists and antagonists have been proposed for the treatment of a wide range of diseases including hypertension, renal failure, cardiovascular disorders and epilepsy (Jacobson et al., 1992; Liang, 1992). The termination of the actions of adenosine with its receptor involves its transport across the plasma membrane and subsequent metabolism. Adenosine is hydrophilic and specialized transport systems are required for its movement across the cell membrane. Inhibition of these transporters potentiates the actions of adenosine (Jarvis, 1988). Adenosine transport also plays a role in determining the plasma and tissue levels of *To whom correspondence should be addressed. Received 12 July 1995.

adenosine resulting from hepatic synthesis and release from, or salvage into adjacent tissues, erythrocytes and vascular endothelium. Transport is also responsible for the uptake of dietary nucleosides from the intestinal lumen, their salvage in the kidney and their distribution throughout the organism via circulating erytbrocytes. In contrast to adenosine receptors, our knowledge of the properties of adenosine transporters is more limited. Nevertheless, mammalian nucleoside transport (NT) may be divided into two broad categories-facilitated-diffusion nucleoside transport and active transport driven by an inwardly directed transmembrane sodium gradient. The purpose of this minireview is to examine the pertinent nucleoside transport literature with particular emphasis on studies during the past 8 years and to resolve areas of controversy, if possible. The review will focus on the properties of nudeoside transport systems in mammalian cells with particular emphasis on the evidence for possible isoforms of the equilibrative carriers, the inhibitor susceptibilities of the facilitated-diffusioncarriers and their molecular properties, the heterogeneity of Na+/adenosine cotransporters and the recent cloning studies on the active transporters. Previous reviews should be consulted for early references not included in this review,

614

J . A . Thorn and S. M. Jarvis

areas not touched and for different interpretations (Plagemann et al., 1988; Jarvis and Young, 1987; Jarvis, 1988; Gati and Paterson, 1989). FACILITATIVE-DIFFUSION

NUCLEOSIDE TRANSPORTERS Kinetic

properties

In mammalian cells there are two facilitated systems of broad substrate specificity for purine and pyrimidine nucleosides with a low or moderate affinity ( K ~ - 20-400 LtM), differing in their sensitivity to inhibition by the synthetic nucleoside analogue, nitrobenzylthioinosine (NBMPR). The es (equilibrative, sensitive) transporter has a Ki value of 0.1-1 nM for NBMPR inhibition of nucleoside influx (Plagemann et al., 1988; Jarvis, 1988), whereas the ei (equilibrative, insensitive) carrier retains full transport activity until exposed to NBMPR concentrations in excess of 1 ~M (Belt and Noel, 1985; Lee and Jarvis, 1988a, 1988b). Of the naturally occurring nucleosides, adenosine exhibits the highest affinity for the facilitated-diffusion carriers with a Km for influx of 20-50 lxM at 25°C (Plagemann e t al., 1988; Jarvis, 1988). However, a Km value of 1-2 FaM has been reported for the es transporter in bovine chromaffin cells (Casillas et al., 1993). Similarly, the affinity of other nucleosides for es appears to differ between cell types (Aran and Plagemann, 1992; Griffith et al., 1992). In addition, the enantiometric configuration of the nucleoside is a determinant of permeant interaction with the es carrier in some cells but not others. For example, the transport rates for D-adenosine are -- 20-fold higher than the corresponding transport rates for L-adenosine in mouse erythrocytes and L1210 cells, whereas no difference is observed in rat synaptoneurosomes and bovine chromaffin cells (Gati et al., 1989; Casillas et al., 1993; Gu and Geiger, 1992). The presence of a polar hydroxyl group at the Y-position appears to be an essential structural determinant for efficient transport via the es transporter (Plagemann et al., 1990). A comparison of the affinities of nucleosides for the es and ei systems suggests that in general the ei nucleoside transporter has a lower affinity for substrates, up to 16-fold in some cells, than the es carrier (Hammond, 1992; Jarvis and Young, 1986; Lee and Jarvis, 1988a). The es and ei carriers in mammalian cells exhibit directional symmetry, with the kinetic parameters for z e r o - t r a n s influx equal to those for z e r o - t r a n s efflux (Plagemann et al., 1988; Jarvis, 1988). However, the mobility of the carrier, a term used to indicate conformational change, is either dependent on whether the transporter is in the loaded or empty state, as is the case for mammalian erythrocytes, or is independent of permeant (Jarvis, 1986, 1988; Plagemann et al., 1988, 1990). In addition, the structure of the substrate can also affect the mobility of the transporter. For example, the mobilities of the adenosine-loaded and empty es carrier of human erythrocytes are the same, whereas loading the carrier with 2-chloroadenosine decreases mobility (Jarvis, 1986). In contrast, the mobility of the pyrimidine-loaded es human erythrocyte transporter is 5-10 times greater than that of the empty carrier (Jarvis, 1986; Plagemann et al., 1990). The turnover number for the es transporter, that is transport rate divided by the number of transporters estimated from the number of high-affinity NBMPR binding sites, also varies widely by up to 20-fold (Plagemann and Wohlhueter, 1985; Griffith et al., 1992; Sobrevia et al., 1994). Thus, there is considerable heterogeneity in the permeant affinities, turnover numbers and carrier mobility of the es transporters.

Inhibitors An area of considerable confusion in the field of facilitated-diffusion nucleoside transport is the question of whether the es and ei trans-

porters can also be distinguished by their sensitivity to inhibition by other transport blockers. This confusion has resulted from three main causes. First, many of the experimental conditions used have resulted in depletion of the free inhibitor concentration leading to erroneous K~values. Second, the use of inappropriate data analysis procedures for cells possessing both es and ei where many of the stated IC50 values are based on shallow dose response curves with pseudo-Hill coefficients of < 1 resulting in an average potency for inhibiting both transporters and, finally, the failure of some workers to appreciate species differences in the potency of some compounds to act as transport inhibitors. The es transporter in rat tissues is 100 to 1000-fold less sensitive to inhibition by dipyridamole and dilazep than the corresponding es carrier in guinea-pig, human and pig cells and tissues (IJzerman et al., 1992; Jarvis and Young, 1986, 1987; Jones and Hammond, 1993; Lee, 1994; Lee and Jarvis, 1988a, 1988b; Plagemann and Woffendin, 1988). As such, es and ei transporters from rat tissues and cells exhibit equal sensitivity to inhibition by dilazep and dipyridamole. In contrast, in human, guinea-pig and mouse cells where both the es and ei transporters are present, dilazep inhibits nucleoside influx in a biphasic manner with up to a 1000-fold difference in the sensitivity of the es system compared to the ei transporter (Hammond, 1992; Lee, 1994; Lee and Jarvis, 1988a; Plagemann and Woffendin, 1988). Similarly, dipyridamole inhibits nucleoside influx in guinea-pig synaptosomes, HL-60 cells and HeLa cells with a shallow dose response curve consistent with two components differing in affinity by 1000-fold (Lee, 1994; Lee and Jarvis, 1988a; Plagemann and Woffendin, 1988). In contrast, dipyridamole is only poorly able to distinguish between es and ei in Ehrlich cells due to the low affinity of the murine es transporter for dipyridamole (Hammond, 1991; Plagemann and Woffendin, 1988). In conclusion, the ei carrier is less susceptible to inhibition by dilazep and dipyridamole than the es carrier in cells other than those derived from rat tissues and to a lesser extent those of murine origin. Nevertheless, high concentrations (100 ~tM) ofNBMPR, dilazep and dipyridamole will inhibit both es and ei. Another group of compounds that are potential inhibitors of facilitated diffusion nucleoside transport and have received much interest over the past 5 years are analogues oflidoflazine. Like dilazep and dipyridamole, the potency of lidoflazine and its analogues for inhibition of es are species related, with rat, mouse, and hamster cells up to 1000-fold less sensitive to inhibition than human, rabbit, and pig (Griffith et al., 1990; Hammond, 1991; IJzerman et al., 1992; Jones and Hammond 1993; Plagemann and Woffendin, 1988). In general, lidoflazine and its analogues are either more potent as inhibitors of es than ei or are unable to distinguish between the two equilibrative transporters. One exception is soluflazine, which is 100-fold more potent as an inhibitor of rat erythrocyte ei (IC50- 0.1 IxM) than the corresponding es system in the same cells (ICs0 I0 ~tM) (Griffith et al., 1990). Although it is beyond the scope of this review to discuss possible models in detail of the interaction of inhibitors and substrates with the es transporter, there is increasing evidence to support a model of two distinct inhibitor binding sites that are allosterically linked (Hammond, 1991). At the first site, it is suggested that NBMPR binds, together with many other transport inhibitors and nucleoside permeants, as evident from competitive inhibition profiles for the displacement of [3H]NBMPR bound (Jarvis, 1988). We propose that this site incorporates the permeation site with additional structural elements that allow the inhibitors to bind tightly. In mammalian erythrocytes, the site has also been suggested to correspond to the outward facing conformation of the es permeation site (Jarvis, 1988; Jarvis and Young, 1987). At the second site, a range of compounds

Adenosine Transporters including nucleosides and transport inhibitors are proposed to interact with a much reduced affinity that can modulate binding at the permeation site. This modulation can be observed experimentally in the finding that transport inhibitors and permeants either inhibit or enhance the rate of dissociation of bound [3H]NBMPR (Hammond, 1991; Jarvis et al., 1983; Koren et al., 1983). Further support for this model comes from binding studies with [3H]dilazep and the mioflazine analogue, R75231, that have demonstrated multiple binding sites and positive cooperativity, respectively (Gati and Paterson, 1989; IJzerman et al., 1992; Jones and Hammond, 1993). Positive cooperativity has also been proposed for the transport of L-adenosine on the es transporter in bovine chromaffin ceils (Casillas et al., 1993). The possible presence of allosteric sites on the es transporter have implications for regulation of transport by natural occurring compounds. For example, adenine nucleotides modulate nudeoside transport in adrenal chromaffin cell membrane preparations (Delicado et al., 1994).

Structure ores nucleoside transport system Much of our current knowledge of the molecular properties of es transporters has come from the exploitation of NBMPR's ability to become covalently bound to the transporter upon exposure to UV light (Jarvis and Young, 1987). Utilizing NBMPR as a probe for es relies on NBMPR binding to the same polypeptide that is responsible for es type transport activity. Evidence in support of this includes: a) a correlation between NBMPR binding activity and transport capacity (Jarvis et aL, 1982), b) mutant cell lines where a loss of es transport activity is accompanied by a loss of high affinity NBMPR binding (Cass et al., 1981), c) nucleoside permeants acting as competitive inhibitors of NBMPR binding (jarvis, 1988) and d) reconstitution of both NBMPR binding activities and transport activities into phospholiposomes (Hammond, 1994; Tse et al., 1985a). Photoaffinity labelling of the es transporter with [3H]NBMPR from different species reveals considerable differences in the size of the carrier in terms of electrophoretic mobility on SDS-polyacrylamide gels. For example, in human erythrocytes and guinea pig liver and lung the es carrier has been identified as broad band of average M, 57,000 with a deglycosylated apparent M~ of 45,000 (Kwong et al., 1986, 1993; Young et al., 1983). The es transporter of pig erythrocytes exhibits a higher apparent Mr of 64,000 and 57,000 before and after deglycosylation,respectively(Kwonget al., 1986). Similar high values are observed for the native es transporter of rat tissues, but the deglycosylatedform migrates with an apparent M~ of 47,000 (Kwong et al., 1993). Despite these species differences, polyclonal antibodies against the purified es transporter of human erythrocytes cross-react with the nucleoside transporters of rabbit and pig erythrocytes and from rat liver (Kwong et al., 1992). The antibodies strongly label the deglycosylated transporter and thus these results suggest that these mammalian es transporters share common structural features. Additional evidence to support this conclusion comes from the finding that the sites of pH]NBMPR photolabelling, carbohydrate attachment and trypsin cleavage are broadly equivalent in four different species (human, pig, rat and guinea pig) and three tissues (erythrocyte, lung and liver) (Kwong et al., 1993). Furthermore, the low dipyridamole sensitivity of rat es appears to be a consequence of relatively minor differences in molecular structure and glycosylation is unlikely to contribute to this low sensitivity as dipyridamole's ability to displace NBMPR bound is not affected by endo-[3galactosidase treatment (Kwong et al., 1993). In the case of the human erythrocyte es transporter, trypsin cleavage occurs on the external surface and the site of N-linked glycosylation has been

615 located close to one end of the protein with the site of NBMPR photolabelling within 16 kDa of that site (Kwong et al., 1993). Although the human erythrocyte es transporter has been identified a s a band 4.5 polypeptide and purified to near homogeneity (Kwong et al., 1988), radiation inactivation studies suggest the carrier exists in situ as a dimer (Jarvis et al., 1980). lsofornls

O [ es n u c l e o s i d e

transporter

On the basis of results discussed above, there appears to be increasing evidence to suggest that isoforms of the es transporters from different cells and species exist with respect to both kinetic and molecular properties. Recently, antibodies to the human erythrocyte es transporter were shown to only recognise proteins in the apical surface of the human placental syncytiotrophoblast despite the presence of NBMPR binding sites in both apical and basal surfaces (Barros et aL, 1995). Thus, clearly in this tissue there are two distinct es transporters. Confirmation of the existence and the number of distinct isoforms of the es transporter will come with the identification of the genes encoding the proteins.

The NBMPR.insensitlve

nucleoside

t r a n s p o r t s y s t e m (el)

In contrast to es, our knowledge of the properties of the ei transporter are limited. As noted above, ei exhibits a broad substrate specificity, may exhibit a lower affinity for some nucleoside substrates than the es carrier and is relatively resistant to inhibition by a series of nucleoside transport inhibitors including dipyridamole. Thus, previous suggestions that pH]dipyridamole could be used as a probe of both es and ei must thus be viewed with considerable caution (Deckert et al., 1987; Marangos and Deckert, 1987). Indeed, recent results have failed to confirm earlier studies and found no evidence that the NBMPR-resistant PH]dipyridamole binding sites in CNS membranes are associated with the ei transporter (Jones and Hammond, 1992). The lack of specific molecular probes for ei transporters has resulted in virtually no information on the molecular structure of el. However, an outward facing thiol group within the permeation site has been suggested to be essential for ei transport function (Jarvis and Young, 1986). This is in contrast with the es system where a reactive thiol group sensitive to p-chloromecuriphenyl sulphonate is located within the inward facing conformation of the carrier (Tse et al., 1985b; Jarvis and Young, 1986). Recently, NBMPR-insensitive nucleoside transport was reconstituted into liposomal membranes and this may help in further studies aimed at the purification of the ei transporter(s) (Hammond, 1994).

Origins o~ es and ei

It is difficult to understand why two NT systems of almost identical physiological function have evolved. Converent evolution of two NT genes is unlikely as both transporters may be expressed in the same tissue, with no apparent selective advantage. The analogous family of facilitated glucose transporters (GLUTs) (Mueckler, 1994) would appear to be the result of gene duplication followed by evolutionary diversification to perform specialised functions in different tissues. Thus, one explanation is that ei and es are the result of a gene duplication event so recent in evolutionary history that little functional specialisation has occurred. This and other hypotheses will remain speculative until the DNA sequence(s) encoding the transporters are identified. However, the generation of mutant strains of cultured cells in which one or other

616

J . A . Thorn and S. M. Jarvis

T A B L E 1. Comparison of kinetic properties and tissue location of m a m m a l i a n sodium-dependent nucleoside transporters. Transporter Substrates Urd Thy Cyt Ado Guo Ino Formycin B AZT ddC Inhibitors Urd Thy Cyt Ado Guo Ino Formycin B AZT ddC Na + Coupling Tissue distribution of expression

References

N1 Km (gM) 9-40 No (2,15)

N2 (cNT1)

N3

N4

Km(~M) 11-13 5-8

Km(gM) 18 13

Km(gM) 4-5 27

8-9 8-11 + (2) 3-45 ICs0 (IIM) 41 840 320 8 15 12-35 (Ki=36) No (500) 1:1 Kidney Intestine Spleen Liver Macrophages Monocytes L1210 cells Vijayalakshmi and Belt, 1998 Plagemann and Aran, 1990 Dagnino et al., 1991 Williams and Jarvis, 1991 Moseley et al., 1991 Ruiz-Montasell et al., 1992

SNST1

+ (1)

30 No 490 510 ICs0 (ltM) 8-22 13 6 12-21 640 1300 900 + (1000) No (500) 1:1 Kidney Intestine

IC50 (laM) 8-10 13-16 4-13 + (100) 8-30 9-10 22-42 No (200) No (200) 2:1 Choroid Plexus

IC5o (taM) + (100) + (100) + (100) + (100) + (100) No (I00) No (100) No (100) No (100) 1:1 Human Kidney

Inhibition + (1 raM) + (1 mM) + (1 mM) + (1 mM)

+ (I mM) Not tested mRNA in: Kidney Heart

RatJounum

Vijayalakshimi and Belt, 1988 Williams and Jarvis, 1991 Huang et al., 1994

Wu et al., 1992, 1994 Huang etal., 1993

Gutierrez et al.,

1992, 1993

Pajor and Wright, 1992

+ (x) = Uptake or inhibition reported at a concentration of x. No (x) = No uptake or inhibition reported at a concentration of x.

transport activity is lost (Crawford et al., 1990; Vijayalakshmi et al., 1992) suggests distinct genetic determinants ofei and es activity. ACTIVE NUCLEOSIDE TRANSPORT About 10 years ago, it became apparent that further diversity in nucleoside transport mechanisms exist with a report of adenosine being transported across renal brush-border membrane vesicles by a Na+-cotransport system(s) (Le Hir and Dubach, 1984). Subsequent work has extended these observations and found that Na+-dependent nucleoside transport is widespread (present in macrophages, spleenocytes, intestine, kidney, liver, choroid plexus and several cultured cell lines) and catalysed by more than one system (Dagnino et al., 1991; Darnowski et al., 1987; Gutierrez et al., 1992; Huang et al., 1993; Jarvis, 1989; Moseley etal., 1991; Plagemann and Aran, 1990; Ruiz-Montasell et al., 1992; Vijayalakshmi and Belt, 1988; Williams and Jarvis, 1991; Wu et al., 1992). Much of the detailed information

on the kinetic properties of the active nucleoside carriers has come from studies on membrane vesicles and a few investigations of cultured cells. In general, major features of the active nucleoside transport systems in membrane vesicles are an overshoot in solute uptake in the presence of an inward Na ÷ gradient and a depolarization of the membrane potential (Williams et al., 1989; Williams and Jarvis, 1991). The requirement for Na ÷ is generally specific with only one report suggesting that the cation Li ÷ may be able to replace Na + (Moseley et al., 1991). The effect of anions with differing permeability and valinomycin-induced K ÷ diffusion potential on the uptake of nucleosides by membrane vesicles have indicated that Na ÷dependent nudeoside transport is an electrogenic process involving the net transfer of positive charge (Gutierrez and Giacomini, 1993; Williams and Jarvis, 1991; Williams et al., 1989). In rat renal brushborder membrane vesicles, anions may be directly involved in the transport process (Lee et al., 1990).

Adenosine Transporters

617 CARBOXYL TERMINI

AMINO TERMINI Amino acid 1 7 5 - 1 7 6

496-501 541-544

[]

R

NUDC

I

[]

B.subtilis NUPC

I

[]

E.coli UNKNOWN 1

[]

[]

E.coli UNKNOWN 2

[]

[]

B.subtihs N17G

m

!

cNT 1 I

E.coli

D.rnelanogaster

}

FIGURE 1. T h e N2 Na +/nudeoside transporter gene (cNT1) family. Scale representation of protein sequence alignments of members of the rat cNT1 gene family. A m i n o acid n u m b e r i n g refers to the cNT1 sequence. Hatched boxes represent highly conserved peptide motifs.

[] I

STS

100 AMINO ACIDS

612-614

NEFVA MOTIF

Substrate specificity Detailed studies on the substrate specificity, and hence the possible number of different Na + nucleoside cotransporters, has been mainly derived from studying the inhibitory effects of nucleosides on the influx of radioactive permeant. Four separate transporters have been characterised (Table 1). Two systems of nomenclature have been used, based on either substrate specificity (cif, cit and cib) or numerical order of characterisation (N 1, N2, N3 and N4). The former system is becoming cumbersome, for example N4 is classified as "cit-like," so we prefer the adoption of the numerical system. The NI (ci~) system which mainly transports purine nucleosides, uridine and deoxyuridine, is widely distributed, whereas N2 (cit) has only been found in kidney and intestine and transports pyrimidinenudeosides, adenosine and analogues of adenosine. Human renal brush-border membrane vesicles possess a single Na +-dependent nudeoside transporter with a substrate specificity similar to N2 but also accepting guanosine (Gutierrez and Giacomini, 1993; Gutierrez et al., 1992). This transporter has been termed N4. The final active nucleoside transporter is N3 (cib) that has a broad substrate specificity and is found in rabbit choroid plexus and rat jejunum (Wu et al., 1992, 1994; Huang et al., 1993). It is interesting to note that both adenosine and uridine appear to be substrates for all four cotransporters. However, caution should be exercised as transport inhibition can occur without the inhibitor being transported. Indeed, tubercidin inhibits Na+-dependent thymidine transport by mouse enterocytes without itself entering the intestinal cells in a Na+-dependent manner (Vijayalakshmi and Belt, 1988). Nevertheless, supporting evidence for the designated permeant specificities come from trans-stimulation experiments (Gutierrez et al., 1992; Le Hir, 1990; Ruiz-Montasell et al., 1992). The sodium:nucleoside coupling ratios of N1-N4 have been estimated from Na + activation curves at fixed nucleoside concentrations and shown to be consistent with h l for N1, N2 and N4, but 2:1 for N3 (Dagnino et al., 1991; Gutierrez and Giacomini, 1993; Moseley et al., 1991; Plagemann and Aran, 1990; Williams and Jarvis, 1991; Wu et al., 1992). In general, the affinity of N1-N4 for the nucleoside permeants is high (Kinvalues 1 to 40 pM at 22°C, 100 mM external Na +) (Table 1). Increasing the concentration of Na + increases the affinity of

FANF MOTIF

nucleoside for zero-trans influx without an effect o n Vmax (Dagnino et al., 1991; Williams and Jarvis, 1991). These data are consistent with the view that Na ÷ binds first to the carrier thereby increasing the carrier's affinity for nucleoside.

~olecular properties Two approaches have been used to isolate the genes encoding putative Na + nucleoside cotransporters. The first approach was to identify cDNAs based on homology to the Na+/glucose cotransporter SGLT1. SNST1 was isolated from a rabbit kidney cDNA library and when expressed in Xenopus laevis oocytes stimulated only a 2-fold increase in Na+-dependent nucleoside transport (Pajor and Wright, 1992). The inhibitor profile appears to be possibly similar to N3 but the tissue specificity of mRNA expression (present in kidney and heart) is not consistent with the N l-N4 subtypes (Table 1). Recent studies from this laboratory have failed to detect active nucleoside transport in guinea pig cardiomyocytes and coronary endothelial cells (Conant and Jarvis, 1994, unpublished observations). Thus, it is possible that the physiological substrate for SNST1 has not yet been identified. Low levels ofnucleoside transport activity may be the result of substrate cross reaction such as seen between the Na +/glucose and Na +/myo-inositol transporters from the same gene family (Hager et al., 1995). A more convincingstrategy has been expression cloning, resulting in the isolation of a rat jejunal cDNA encoding a 648-residue Na ÷dependent nucleoside transport protein termed cNTI (Huang et al., 1994). Expression of cNTI in Xenopus oocytes stimulates Na +dependent uridine influx by 10,000-fold and the inhibition profile of uridine and thymidine influx indicates N2 type transport. Transcripts for cNT1 are detected in intestine and kidney which parallel the kinetic data that has demonstrated only N2 transport activity in these two tissues (see Table 1). Hydropathy plots are consistent with up to 14 transmembrane domains and sequence database searches show that cNTI is a member of a novel gene family with no sequence similarity to date to any other mammalian protein. Nevertheless, the family contains the NupC proton-driven nucleoside transporter from E. coli (Craig et al., 1994) and B. subtiUus (Genbank accession number X82174), with amino acid sequence

618 similarity in the order of 30%. Other family members are also bacterial and of u n k n o w n function, except for a short sequence from D. melanogaster (Genbank accession number G01280) which aligns with a highly conserved hexapeptide motif (NEFVAY) towards the carboxyl terminus of the gene. Conservation with the shorter bacterial genes is limited to the middle and carboxyl terminus of cNT1 (Fig. 1). It is, therefore, tempting to suggest that the conserved regions encode domains of similar function (nucleoside transport) whereas the nonconserved regions encode dissimilar function (sodium or proton dependency). Such proposals might be too simplistic given the observation that both Na ÷ and H ÷ can support glucose transport by SGLT1 (Hirayama et al., 1994) and will remain speculative until tested by mutagenesis and chimera studies.

J . A . Thorn and S. M. Jarvis lated in diabetes (Sobrevia et al., 1994). Finally, the possible use of nucleoside transport blockers as therapeutic agents needs to be further elucidated. One of the most promising applications is the use ofinhibitors as cardioprotective agents where inhibition of adenosine transport would attenuate adenosine catabolism during ischemia and delay its washout upon reperfusion thereby enhancing the protective role of endogenous adenosine. Draflazine, an analogue oflidoflazine, with excellent bioavailability, shows particular promise as a cardioprotective agent (Van Belie, 1993). There are, therefore, major roles to be played by both molecular biologists, pharmacologists, biochemists and clinicians towards the further understanding and exploitation of nucleoside transport. Note added in

FUTURE PROSPECTIVES

proof

Since submission of this manuscript, a rat liver cDNA (SPNT) encodThe marked heterogeneity in nucleoside transporters that exists ing a sodium/nucleoside transport protein with the transport characwith respect to substrate specificity, inhibitor sensitivity, mechanism teristics of N1 has been isolated (Che, M., Oritz, D. F. and Arias, and tissue distribution is an important determinant of the pharmaco- I. A. (1995). J. Biol. Chem. 270, 13596-13599). The nucleotide sekinetics, disposition and biological activity in vivo of physiological quence is predictive of a 659 amino acid protein with 14 transmemnucleosides and drugs. Moreover, such diversity in transporters brane domains and a molecular weight of approximately 72kDa. suggests that it may be possible in the future to target more precisely The SPNT amino acid sequence is 64% identical to cNT 1, confirming drugs to certain tissues and to use specific inhibitors to modulate it as a member ofthi s nudeoside transporter gene family. Recently, an H. influenzae NUPC nucleoside transporter sequence was identified the activity of defined nucleoside transporters. Such an approach will require a detailed knowledge of transporter structure and mecha- (Fleischmann, R. D., et al. (1995). Science 269, 496-512). This is nism. With the recent application of recombinant D N A technology also a member of the gene family, with a protein sequence similarity to nucleoside transporters, progress towards these goals is underway to cNT1 of the same order as the E. coli NUPC sequence. and will continue in this and other laboratories engaged in the Work in this laboratory during the past 5 years, part of which has been summarized cloning of each nudeoside transporter type. Identification of the in this review, was supported by grants #ore the Medical Research Council, the cDNA sequences of the transporters will allow secondary structural National Kidney Research Fund, the Nuffield Foundation, the British Council and the Royal Society. models to be generated and tested by chemical modification and site-directed mutagenesis. High level expression and protein purification would be the first step towards crystallographic determination References of protein conformation. Aran J. M. and Plagemann P. G. W. (1992)High affinity, equilibrative nudeOften more than one transporter type is expressed in a cell or oside transporter of pig kidney cell lines (PK-15). Biochim. Biophys. Acta. tissue, making it difficult to characterise one system in isolation 1108, 67-74. without major biochemical and/or pharmacological manipulation. Barros L. F., Yudelevich D. L., Jarvis S. M., Beaumont N., Young J. D. and Baldwin S. A. (1995) Immunolocalisation of nucleoside transporters in Thus, the cloning and expression of each nucleoside transporter human placental trophoblast and endothelial cells: evidence for multiple type should help in defining the precise substrate profile for the transporter isoforms. Pflugers Archiv.-Eur. J. Physiol. 429, 394-399. carrier. For example, in contrast to previous studies with bovine . Belt J. A. and Noel L. D. (1985) Nudeoside transport in Walker 256 rat carcinoma and $49 mouse lymphoma cells. Differences in sensitivity to renal membrane vesicles expressing both N 1 and N2 (Williams and nitrobenzylthioinosine and thiol reagents. Biochem. J. 232, 681-688. Jarvis, 1991), expression ofcNT1 in oocytes revealed that Y-azido-YBoumah C. E., Harvey C. M., Paterson A. R. P., Baldwin S. A., Young deoxythymidine (AZT) and 2',Y-dideoxycytidine(ddC) are low affinJ. D. and Cass C. E. (1994)Functionalexpression of the nitrobenzylthioiity (Kin 0.5 raM) permeants for intestinal N2 (Huang et al., 1994). nosine-sensitive nudeoside transporter of human cl3oriocarcinoma (BeWo) cells in isolated oocytes of Xenopus laevis. Biochem. J. 299, 769Injection of Xenopus oocytes with poly(A) + RNA from a number 773. of sources has resulted in the expression of N1 (Huang et al., 1993), Casillas T., Delicado E. G., Carmona F. G. and Miras-Portugal T. (1993) N3 (Wu and Giacomini, 1994) and N4 (Gutierrez and Giacomini, Kinetic and allosteric cooperativity in L-adenosinetransport in chromaffin 1994) suggesting that expression cloning of these active nucleoside cells: a menmonial transporter. Biochemistry 32, 14203-14209. transporters should be possible. Alternatively, if N 1-N4 are closely Cass C. E., Kolassa N., Uehara Y., Dahlig-Harley E., Harley E. R. and Paterson A. R. P. (1981)Absence of bindingsites for the transport inhibirelated, hybridisation screening ofcDNA libraries with cNT 1 probes tor nitrobenzylthioinosine on nucleoside transport deficient mouse lymmay yield related gene sequences. By analogy with active and passive phoma cells. Biochim. Biophys. Acta. 649, 769-777. glucose transporters, which are not related, it seems unlikely that Conant A. R. and Jarvis S. M. (1994) Nucleoside influx and efflux in guinea es and ei transporters will be members of the N2 family. Nevertheless, pig ventricular myocytes: inhibitionby analogues of lidoflazine. Biochem. Pharmacol. 48, 873-880. work is in progress using antibodies to purified h u m a n erythrocyte Craig J. E., Zhang Y., and Gallagher M. P. (1994)Cloning of the nupC gene es (Kwong et al., 1992) to screen expression libraries coupled with of Escherichia coli encoding a nucleoside transport system and identificaoocyte expression of es (Boumah et al., 1994) to attempt to clone tion of an adjacent insertion element, IS 186. M0l. Microbiol. 11, 1159this equilibrative carrier. 1168. Despite the exciting developments that may lead from the molecu- Crawford C. R., Ng C. Y. C., Noel L. D. and Belt J. A. (1990) Nucleoside transport in L 1210murineleukaemia cells:evidence for three transporters. lar biology studies, additional studies on the biochemical characterJ. Biol. Chem. 265, 9732-9736. ization and regulation ofnucleoside transport are required, especially Dagnino L., Bennet L. L., Jr., and Paterson A. R. P. (1991)Substrate specificin the context of neoplastic transformation and disease states. For ity, kinetics and stoichiometry of sodium dependent adenosine transport in L1210/AM mouse leukaemia cells. J. Biol. Chem. 266, 6312-6317. example, es nucleoside transport activity appears to be down regu-

Adenosine Transporters Darnowski J. W., Holdridge C. and Hanschumacher R. E. (1987) Concentrative uridine transport by murine spleenocytes: kinetics, substr ate specificity and sodium dependency. Cancer Res. 47, 2614-2619. Deckert J., Bisserbe J. C. and Marangos P. J. (1987) Quantitative [3H]dipyridamole autoradiography: evidence for adenosine transporter heterogeneity in guinea-pig brain. Naunyn-Schmiedeberg Arch. Pharmacol. 335,660-666. Delicado E. G., Casillas T., Sen R. P. and Miras-Portugal M. T. (1994) Evidence that adenine nucleotides modulate nucleoside-transporter function: characterisation of uridine transport in chromaffin cells and plasma membrane vesicles. Eur. J. Biochem. 225, 355-362. Gati W. P. and Paterson A. R. P. (1989) Nucleoside transport. In: (Edited by Agre P. and Parker J. C.) Red Blood Cell Membranes: Structure, Function, Clinical Implications pp. 635-661. Marcel Dekker, New York. Gati W. P., Dagnino L. and Paterson A. R. P. (1989) Enantiomeric selectivity of adenosine transport systems in mouse erythrocytes and L1210 cells. Biochem. J. 263, 957-960. Griffith D. A., Conant A. R. and Jarvis S. M. (1990) Differential inhibition of nucleoside transport systems in mammalian cells by a new series of compounds related to lidoflazine and mioflazine. Biochem. Pharrnacol. 40, 2297-2303. Griffith D. A., Doherty A. J. and Jarvis S. M. (1992) Nucleoside transport in cultured LLC-PK1 epithelia. Biochim. Biophys. Acta. 1106, 303-310. Gu J.G. and Geiger J.D. (1992) Transport and metabolism of D-[SH]adenosine and L-[3H]adenosine in rat cerebral cortical synaptoneurosomes. J. Neurochem. 58, 1699-1705. Gutierrez M. M., Brett C. M., Ott R. J., Hui A. C. and Giacomini K. M. (1992) Nucleoside transport in brush border membrane vesicles from human kidney. Biochim. Biophys. Acta. 1105, 1-9. Gutierrez M. M. and Giacomini K. M. (1993) Substrate selectivity, potential sensitivity and stoichiometry ofNa +-nucleoside transport in brush border membrane vesicles from human kidney. Biochim. Biophys. Acta. 1149, 202-208. Gutierrez M. M. and Giacomini K. M. (1994) Expression of a human renal sodium nucleoside cotransporter in Xenopus laevis oocytes. Biochem. Pharmacol. 48, 2251-2253. Hager K., Hazama A., Kwon H. M., Loo D. D. F., Handler J. S. and Wright E. M. (1995) Kinetics and specificity of the renal Na+/myo-inositol cotransporter expressed in Xenopus oocytes. J. Membr. Biol. 143, 103-113. Hammond J. R. (1991) Kinetic analysis of ligand binding to Ehrlich cell nucleoside transporter: pharmacological charcterisation of allosteric interactions with the [SH]nitrobenzylthioinosinebinding site. Mol. Pharmacol. 39, 771-779. Hammond J. R. (1992) Differential uptake of [3H] guanosine by nucleoside transporter subtypes in Ehrlich ascites tumour cells. Biochem. J. 287, 431-436. Hammond J. R. (1994) Functional reconstitution of pharmacologically distinct subtypes of nucleoside transporters in liposomal membranes. J. Pharamacol. Exp. Ther. 271, 906-917. Hirayama B. A., Loo D. D. F. and Wright E. M. (1994) Protons drive sugartransport through the Na + glucose cotransporter (SGLT1). J. Biol. Chem. 269, 21407-21410. Huang Q.-Q., Harvey C. M., Paterson A. R. P., Cass C. E. and Young J. D. (1993) Functional expression ofN a +-dependent nucleoside transport systems of rat intestine in isolated oocytes of Xenopus laevis. J. Biol. Chem. 268, 20613-20619. Huang Q.-Q., Yao S. Y. M., Ritzel M. W. L., Paterson A. R. P., Cass C. E. and Young J. D. (1994) Cloning and expression of a complementary DNA encoding a mammalian nucleoside transport protein. J. Biol. Chem. 269, 17757-17760. IJzerman A. P., Kruidering H., van Weert A., van Belle H. and Janssen C. (1992) [SH]R75231-a new radioligand for the nitrobenzylthioinosinesensitive and-resistant nucleoside transport proteins. Naunyn-Schmiedeberg's Arch. Pharmacol. 435, 558-563. Jacobson K. A., van Galen P. J. M. and Williams M. (1992) Adenosine receptors: pharmacology, structure-activity relationships and therapeutic potential. J. Med. Chem. 35, 407-422. Jarvis S. M. (1986) Trans-stimulation and trans-inhibition of uridine efflux from human erythrocytes by permeant nucleosides. Biochem. J. 233, 295297. Jarvis S. M. (1988) Adenosine transporters. In: (Edited by Cooper D. M. F. and Londos C.) Adenosine Receptors, pp. 113-123. Alan R. Liss Inc., New York. Jarvis S. M. (1989) Characterisation of sodium-dependent nucleoside transport in rabbit intestinal brush border membrane vesicles. Biochim. Biophys. Acta. 979, 132-138. Jarvis S. M. and Young J. D. (1986) Nucleoside transport in rat erythrocytes:

619 two components with differences in sensitivity to inhibition by nitrobenzylthioinosine and p-chloromecuriphenylsulphonate../. Membr. Biol. 93, 1-10. Jarvis S. M. and Young J. D. (1987) Photoafflnity labelling of nucleoside transporter polypeptides. Pharmacol. Ther. 32, 339-359. Jarvis S. M., Young J. D. and EUory J. C. (1980) Nucleoside transport in human erythrocytes: apparent molecular weight of the nitrobenzykhioinosine-binding complex estimated by radiation inactivation analysis. Biochem..l. 190, 373-376. Jarvis S. M., Hammond J. R., Paterson A. R. P. and Clanachan A. S. (1982) Species differences in nucleoside transport: a study of uridine transport and nitrobenzylthioinosine binding by mammalian erythrocytes. Biochem. J. 208, 83-88. Jarvis S. M., Janmohamed S. N. and YoungJ. D. (1983) Kinetics of nitrobenzylthioinosine binding to the human erythrocyte nucleoside transporter. Biochem. J. 216, 661-667. Jones K. W. and Hammond J. R. (1992) Heterogeneity of [3H]dipyridamole binding to CNS membranes: correlation with [SH]nitrobenzylthioinosine binding and [SH]uridine influx studies. J. Neurochem. 59, 1363-1371. Jones K. W. and Hammond J. R. (1993) Interaction of the mioflazine derivative R75231 with the nucleoside transporter: evidence for positive cooperativity. Fur. J. Pharmacol. 246, 97-104. KorenR.,CassC. E. andPatersonA. R. P.(1983)Thekineticsofdissociation of the inhibitor of nucleoside transport, nitrobenzykhioinosine, from the high affinity binding sites of cultured hamster cells. Biochem. J. 216, 299308. Kwong F. Y. P., Baldwin S. A., Scudder P. R., Jarvis S. M., Choy M. Y. M. and YoungJ. D. (1986)Erythrocyte nucleoside and sugar transport. EndolS-galactosidase and endoglycosidase-F digestion of partially purified human and pig transporter proteins. Biochem. J. 240, 349-356. Kwong F. Y. P., Davies A., Tse C. M., Young J. D., Henderson P. J. F. and Baldwin S. A. (1988) Purification of the human eyrthrocyte nucleoside transporter by immunoaffinity chromatography. Biochem. J. 255, 243249. KwongF. Y. P.,FinchamH. E.,DaviesA.,BeaumontN.,HendersonP. J. F., Young J. D. and Baldwin S. A. (1992) Mammalian nitrobenzykhioinosine-sensitive nucleoside transport proteins: immunological evidence that transporters differing in size and inhibitor specificity share sequence homology. J. Biol. Chem. 267, 21954-21960. Kwong F. Y. P., Wu J. S. R., Shi M. M., Fincham H. E., Davies A., Henderson P. J. F., Baldwin S. A. and Young J. D. (1993) Enzymatic cleavage as a probe of the molecular structures of mammalian equilibrative nucleoside transporters. J. Biol. Chem. 268, 22127-22134. Le Hir M. (1990) Evidence for separate carriers for purine nucleosides and for pyrimidine nucleosides in the renal brush border membrane. Renal Physiol. Biochem. 13, 154-161. Le Hir M. and Dubach U. C. (1984) Sodium gradient energised concentrative transport of adenosine in renal brush border vesicles. Pfluger Archly. Eur. J. Physiol. 401, 58-63. Lee C. W. (1994) Decrease in equilibrative uridine transport during monocytic differentiation of HL-60 leukaemia: involvement of protein kinase C. Biochem. J. 300, 407-412. Lee C. W. and Jarvis S. M. (1988a) Nucleoside transport in rat cerebralcortical synaptosomes. Biochem. J. 249, 557-564. Lee C. W. and Jarvis S. M. (1988b) Kinetics and inhibitor specificity of adenosine transport in guinea pig cerebral cortical synaptosomes: evidence for two nucleoside transporters. Neurochem. Int. 12, 483-492. Lee C. W., Chesseman C. I. and Jarvis S. M. (1990) Transport characteristics of renal brush border Na ÷- and K+-dependent uridine carriers. Am. J. Physiol. 258, F1203-1210. Liang B. T. (1992) Adenosine receptors and cardiovascular function. Trends Cardiovasc. Med. 2, 100-108. Marangos P. J. and Deckert J. (1987) [3H]Dipyridamole binding to guinea pig brain membranes: possible heterogeneity of central adenosine uptake sites. J. Neurochem. 48, 1231-1236. Moseley R. H., Jarose S. and Permoad P. (1991) Adenosine transport in rat liver plasma vesicles. Am. J. Physiol. 261, G716-722. Meuckler M. (1994) Facilitative glucose transporters. Eur. J. Biochem. 219, 713-725. Pajor A. M. and Wright E. M. (1992) Cloning and functional expression of a mammalian Na + nucleoside cotransporter: a member of the SGLT family. J. Biol. Chem. 267, 3557-3560. Plagemann P. G. W. and Aran J.M. (1990) Characterisation of Na +dependent, active nucleoside transport in rat and mouse peritoneal macrophages, a mouse macrophage cell line and normal rat kidney cells. Biochim. Biophys. Acta. 1028, 289-298.

620 Plagemann P. G. W. and Woffendin C. (1988) Species differences in sensitivity of nucleoside transport in erythrocytes and cultured cells to inhibition by nitrobenzylthioinosine,dipyridamole, dilazep and lidoflazine. Biochim. Biophys. Acta. 969, 1-8. Plagemann P. G. W. and Wohlhueter R. M. (1985) Nitrobenzylthioinosinesensitive and resistant nucleoside transport in normal and transformed rat cells. Biochim. Biophys. Acta. 816, 381-395. Plagemann P. G. W., Wohlhueter R. M. and Woffendin C. (1988) Nucleoside and nucleobase transport in animal cells. Biochim. Biophys. Acta. 947, 405-443. Plagemann P. G. W., Aran J. M., Wohlhueter R. M. and Woffendin C. (1990) Mobility of nucleoside transporter of human erythrocytes differs greatly when loaded with different nucleosides. Biochim. Biophys. Acta. 1022, 103-109. Ruiz-Montasell B., Casado J. F., Felipe A. and Pastor-Anglada M. (1992) Uridine transport in basolateral plasma membrane vesicles from rat liver. J. Membr. Biol. 128, 227-233. Sobrevia L., Jarvis S. M. and Yudilevich D. L. (1994) Adenosine transport in cultured human umbilical vein endothelial cells is reduced in diabetes. Am. J. Physiol. 267, C39-47. Stiles G. L. (1992) Adenosine receptors. J. Biol. Chem. 267, 6451-6454. Tse C. M., Belt J. A., Jarvis S. M., Paterson A. R. P., Wu J. S. and Young J. D. (1985a) Reconstitution studies of the human erythrocyte nucleoside transporter. J. Biol. Chem. 260, 3506-3511. Tse C. M., Wu J. S. R. and Young J. D. (1985b) Evidence for the asymmetric binding of p-chloromercuriphenylsulfonate to the human-erythrocyte nucleoside transporter. Biochim. Biophys. Acta. 818, 316-324. Van Belle H. (1993) Nucleoside transport inhibition- a therapeutic approach to cardioprotection via adenosine. Cardiovasc. Res. 27, 68-76.

J . A . T h o r n and S. M. Jarvis Vijayalakshmi D. and Belt J. A. (1988) Sodium dependent nucleoside transport in mouse intestinal epithelial cells. J. Biol. Chem. 263, 19419-19423. VijayalakshmiD., DagninoL., Belt J. A.,GatiW. P.,CassC. E. andPaterson A. R. P.(1992) L1210/B23.1 cellsexpressequilibrative, inhibitor-sensitive nucleoside transport activity and lack two parental nucleoside transport activities. J. Biol. Chem. 267, 16951-16956. Williams T. C., Doherty A. J., Griffith D. A. and Jarvis S. M. (1989) Characterisation of sodium dependent and sodium independent nucleoside transport systems in rabbit brush-border and basolateral plasma-membrane vesicles from the renal outer cortex. Biochem. J. 264, 223-231. WilliamsT. C. andJarvisS. M.(1991)Multiplesodium-dependentnucleoside transport system in bovine renal brush border membrane vesicles. Biochem. J. 274, 27-33. Wu X., Gutierrez M. M. and Giacomini K. M. (1994) Further characterisation of the sodium-dependent nucleoside transport (N3) in choroid plexus from rabbit. Biochim. Biophys. Acta. 1191, 190-196. Wu X. and Giacomini K. M. (1994) Expression of the choroid plexus sodiumnucleoside cotransporter (N3) in Xenopus laevis oocytes. Biochem.Pharmacol. 48, 432-434. Wu X., Yuan G., Brett C. M., Hui A. C. and Giacomini K. M. (1992) Sodium dependent nucleoside transport in choroid plexus from rabbit: evidence for a single transporter for purine and pyrimidine nucleosides. J. Biol. Chem. 267, 8813-8818. YoungJ. D.,Jarvis S. M., Robbins M. J. and Paterson A. R. P. (1983)Photoaffinity labelling of the human erythrocyte nudeoside transporter by Ne-(pAzidobenzyl)adenosine and nitrobenzylthioinosine. J. Biol. Chem. 258, 2202-2208.