Mediated Transport of Nucleosides in Human Erythrocytes

Transport of nucleosides in human erythrocytes, previously shown to occur by facilitated diffusion, was further characterized using procedures that measured efflux from cells containing radioactive uridine or thymidine. E&x of either uridine or thymidine was accelerated several fold in the presence of extracellular uridine or thymidine. Similar apparent kinetic constants for accelerative exchange diffusion of uridine and thymidine and mutual acceleration of efiux indicated that both nucleosides are transported with equal facility by the same mechanism. Specificity of nucleoside transport was investigated by assessing the ability of structurally related compounds to accelerate uridine efflux when added at different concentrations to the extracellular medium. Various pyrimidine riboand 2’-deoxyribonucleosides were accepted as permeants by the transport mechanism, which appeared to be less tolerant of modifications in the sugar than in the base portion of the permeant molecule.


SUMMARY
Transport of nucleosides in human erythrocytes, previously shown to occur by facilitated diffusion, was further characterized using procedures that measured efflux from cells containing radioactive uridine or thymidine. E&x of either uridine or thymidine was accelerated several fold in the presence of extracellular uridine or thymidine. Similar apparent kinetic constants for accelerative exchange diffusion of uridine and thymidine and mutual acceleration of efiux indicated that both nucleosides are transported with equal facility by the same mechanism.
Specificity of nucleoside transport was investigated by assessing the ability of structurally related compounds to accelerate uridine efflux when added at different concentrations to the extracellular medium. Various pyrimidine ribo-and 2'-deoxyribonucleosides were accepted as permeants by the transport mechanism, which appeared to be less tolerant of modifications in the sugar than in the base portion of the permeant molecule.
Transfer of nucleosides across the plasma membrane in several mammalian cell types is mediated and exhibits the characteristics of "facilitated diffusion" (l-7). Facilitated diffusion, a nonconcentrative form of transport, controls the rate at which particular permeants are transferred across the cell membrane down a concentration gradient (8). Criteria for recognition of transport by facilitated diffusion include saturability of rate, inhibition by compounds structurally analogous to the permeant, and demonstration of "trans" effects. In the latter, movement of permeant from the cis membrane face is influenced by the flow of the same or related permeant from the opposite or trans membrane face.
Accelerative exchange diffusion and counter transport are trans effects that have been observed in studies of facilitated diffusion of glucose (8) and nucleosides (1) in human erythrocytes. Accelerative exchange diffusion, as defined by Stein @J, occurs when the rate of outward transfer of permeant is accelerated by inward transfer of the same or a related permeant * This work was supported by the Medical Research Council of Canada and the National Cancer Institute of Canada. present at the opposite membrane face. In counter transport, a permeant, initially equally distributed across the membrane, is driven outward against its own concentration gradient by the inward flow of a related permeant down its concentration gradient.
In general, when flow of permeant in one direction is influenced by flow of a related permeant in the opposite direction, both permeants are transported by the same mechanism (8-13).
Previous studies of nucleoside transport in human erythrocytes indicted that rates of uptake of uridine and thymidine were saturable and uptake of both nucleosides was inhibited by several other nucleosides (1, 2). Accelerative exchange diffusion was demonstrated when efflux of labeled uridine or thymidine occurred more rapidly into nucleoside-containing than into nucleoside-free medium; free bases or free sugars did not accelerate efflux. Inosine-driven counter transport, of uridine was also demonstrated.
In the present work, accelerative exchange diffusion of nucleosides in human erythrocytes was examined using procedures based on measurements of outward transport of uridine or thymidine, neither of which is phosphorylated or cleaved by erythrocytes (1). Concentration-dependent acceleration of efflux of radioactive uridine by thymidine and of radioactive thymidine by uridine was demonstrated and these experiments formed the basis for use of the accelerative exchange procedure to investigate permeant specificity of nucleoside transport in human erythrocytes.

Erythrocytes
were obtained by centrifuging (1700 x g, 15 min) human blood received from the Red Cross Society Blood Transfusion Service, Edmonton, Alberta, after 21 to 28 days of storage at 4" in acid citrate-dextrose solution A (U.S.P.). After removal of supernatant and white cells, erythrocytes were washed and sedimented (1700 x g, 15 min) three times in TEVbuffered saline (which consisted of 140 mM NaCI, 1.4 mM MgS04. H20, and 18 mM TES, at pH 7.4), discarding the upper layer of cell sediment after each wash. Erythrocyte sediments prepared in this way have an extracellular inulin space of 7 to 11 y0 and a total water content of 70 to 75 y0 (I).
Washed erythrocytes were "loaded" by incubating 40 to 50% Issue of May 25, 1972 C. E. Cass and .A. R. P. Paterson 3315 cell suspensions in TES-buffered saline containing labeled nucleoside for 40 min at 37", conditions sufficient to achieve an equilibrium distribution of labeled nucleoside. Loading was terminated by centrifugntion at 1700 X g for 15 min. Since nucleosides are not concentrated in erythrocytes (1)) internal and external nucleoside concentrations were considered equal at the end of the incubation and the intracellular nucleoside concent'ration was obtained by measuring extracellular radioactivity. Efflux of radioactive lmcleosides from loaded cells was determined by following the time course of appearance of radioactive nucleoside in the medium using at least four reaction mixtures, one for each point of a given time course. Each reaction mixture JTas processed to the stage where a sample of cell-free medium was obtained before preparing the next mixture. Efflux was measured at 25" using solutions and reaction vessels that were kept at 25" prior to use. Packed cell sediments were dispensed in 0.25-or 0.50.ml portions (Glaspak disposable glass syringes, Becton, Dickinson and Co.) into lo-or 25-ml flasks and the assay was initiated by rapid addition of 2.5 or 5.0 ml of test medium.
Such reaction mixtures were stirred magnetically, and aft,er a timed interval, nucleoside transport in each mixture was terminated by rapid addition of 2.5 or 5.0 ml of TES-buffered saline containing 50 PM HTG.
Zero time points were obtained by first adding TES-buffered saline containing HTG, followed 10 s later by addition of test medium.
Immediately after HTG addition, l-to 2-ml portions of the assay mixture were centrifuged with 5 ml of di-1-butylphthalate for 13 min at 1500 X g to separate medium from cells (1). Because of differences in specific gravity, centrifugation at room temperature with dibutylphthalnte (14) results in a three-layered system with dibutylphthalate separating erythrocytes (lower layer) from the aqueous phase (upper layer).
Radioactivity in the aqueous phase of each sample was assayed in triplicate in 5 ml of Bray's counting solution (15) using liquid scintillation counting.
Radioactivity did not appear in the dibutylphthalate phase. The hematocrit of each assay mixture was determined using a capillary tube method.
Efflux, expressed as micromoles per min per ml of packed cells, was calculated after determination of hematocrit, specific activity, and initial rate of appearance of l*C-nucleoside in the incubation medium.2 Hematocrits from the incubation mixtllres used for each rate determination were averaged. Specific activities of labeled uridine or thymidine were calculated from measurements of radioactivity and optical density of the medium used for loading rells. Initial rates were obtained from the time course of appearance of radioactivity in the incubation medium; points were taken at 5-to 10-s intervals and straight lines fitted to the data by the method of least squares.
FIG. 1. Inhibition of efflux of uridine and thymidine by HTG. Data (expressed as radioactivity per ml of cell-free supernatant) were obtained from duplicate sets of flasks for each concentration of HTG tested. A, efflux from cells loaded with 9.4 mM [2-'4C]uridine was measured into medium containing 9.4 mM nonradioactive uridine.

Inhibition
of Nucleoside Transport by 6-[(~-Elydroxy-5-nitrobenzyl) thio]guanosine-A procedure for measurement of initial efflux of uridine or thymidine was developed using the thioguanosine derivative, HTG, to rapidly terminate nucleoside transport.4 In Fig 14C]thymidine were rapidly suspended in medium containing nonradioactive uridine or thgmidine and HTC was added (final concentration of 10 PcLRf) at the times indicated by arrows. For analysis, 0.5.ml samples were removed from the incubation mixtures at timed intervals and immediat.ely centrifuged with dibutylphthalate to obtain cell-free medium. Results in Fig. 1 indicate that 10 ELM HTG inhibited efflux of uridine and thymidine almost completely during the 60-s incubation period.
Inhibitiou of efflux was also observed at 6 and 0.6 pM HTG; partial inhibition was observed at 0.06 /.LK The use of HTG in combination with the dibutylphthalate centrifugation method, which separates cells from medium within 30 s (I), permitted sampling at 5-to 10-s intervals. Samples taken after 10 to 15 min of incubation in medium containing HTG indicated that a slow loss of radioactive nucleoside occurred, but at a rate less than 2% of that observed in medium without inhibitor.
To study kinetics of efflux of uridine or thymidiue during accelerative exchange diffusion, a unidirectional flux of radioactive permeant rnust be measured.
To determine whether the extracellular volume of medium was sufficient to prevent significant reentry of radioactivity during the incubation period, efflux of uridine was measured in cell suspensions of 5, 10, and 20% (by volume).
Data (not shown) indicated that initial rates of outward flow of radioactivity were independent of extracellular rolume when cell suspensions of 10% or less were used.
Throughout this work the sampling procedure described in Fig. 1 was used; the time courses of appearance of radioactivity in the medium were consistently linear for 15 to 30 s, indicating unidirectional movement of radioactivity during the period of measurement.
Apparent initial rates were calculated from time courses as described in "Materials and Methods." In Fig. I efflux of uridine occurred at an apparent initial rate of 5.5 pmoles per min per ml of packed cells and efflux of thymidine at 6.2 pmoles per min per ml of packed cells.
Accelerative Exchange Diffusion of Uridine and Thymidine-In several transport systems with a facilitated diffusion mechanism, the kinetics of unidirectional flux of permeant are described by the Michaelis-Menten equatiou (8,19,20). li,,, represents the maximum flux and K, represents the permeant concentration at which flus is half-maximal; the meaning of these "constants" in descriptions of transport kinetics is unknown.
In this de scription of kinetics of accelerative exchange diffusion, sV~,, represents maximum efflux when the concentration of intracellular permeant is above saturation for transport from the internal membrane face and efflux is a function of the concentration of extracellular nucleoside. SK, is defined as that extracellular concentration of added nucleoside at which efflux is one-half SV*'ll,,,.
To study the dependence of efflus on the presence of nucleoside at t'he trans membrane face, it was necessary to measure fluxes from cells containing radioactive permeant at coilcentrations sufficient to saturate the transport mechanism at the cis membrane face. Studies of efflur of radioactive uridine or thymidine at different intracellular concentrations were conducted with extracellular concentration held constant (10 mr\ll or with extracellular and intracellular concentrations equal.5 In these experiments efflux was maximal at intracellular concentrations above 4 mhf, indicating saturation of transport at the cis membrane face. In the experiments of Fig. 2 erythrocytes were loaded with excess radioactive uridine or thymidine (5 to 10 mM) and the ability of either nucleoside to accelerate outflow of the other was tested at several extracellular concentrations.
In preparing incubation mixtures, 10 volumes of medium were added to each volume of packed, loaded cells, thereby diluting 125.fold the radioactive permeant present in the extracellular volume of packed cell sediments.
The initial extracellular concentrations of nucleoside after addition of nucleoside-free medium to cells loaded with 5 to 10 mM uridine or thymidine were about 0.04 to 0.08 mM, and uridine efflux occurred at a rate of 0.7 to 1.2 pmoles per min per ml of packed cells. Efflus of either nucleoside increased as extracellular concentrxtious of uridine or thymidine were increased and reached limiting values at estracellular concentrations of about 4 mn~ The acceleration of uridine effluz by thymidine and of thymidine efflux by uridine indicates that transport of both nuc~leosides is mediated by a single system. Plots of the reciprocals of efflux and of the concentration of added extracellular nucleoside for each of the four possible combinations of uridine and thgmidine (Fig. 2) indicate that efflux was a function of the concen tratiou of extracellular nucleoside and was saturable. Values derived from data in Fig. 2, A to D, for the apparent half-saturation constants, SK,, were 0.1 rnM; values derived for ST;,,,, w-ere 4 3 to 7.1 pmoles per min per ml of packed cells. Because the apparent kinetic constants for accelerative exchange diffusion are of the same magnitude for all combinations, both uridine and thymidine appear to be transported equally well.
Issue of ?\Iny 25,1972 C. E. Cuss md A. R. I'. Paterson 3317 lose of Accelerative I<xc/uznge DiJksion to Study Transport Specificity--ii procedure using uridine as the radioactive permeant was devised to e?tamine the ability of the uridine-thymidine transport mechanism to accept other nucleosides as permeants. Cells were loaded with excess [2-Wjuridine (6 mar)6 and efflus of radioactivity \vas measured in medium eonta,ining different concentrations of nonradioactive test nucleoside. As controls, efflus was measured into (n) TES-buffered saline and (b) TES-buffered saline caontnining the equilibrium concent'ration of nonradioactive uridine.6 Efflur into medium that contained test nucleoside was related to efflus into nonradioactive uridine by expressing t,he former as a percentage of uridine-accelerat.ed efflus, which O~YWS at maximum rat.es when the equilibrium concelltmtion is 6 mu. The procedure is illustrated in Fig. 3 where deosycytidine \T-as teated as a l)ossible permeant.
Comparison of efflus into medium containin? different concent,rations of deoxycytidine with efflux into 'lW-buffered saline indicates that uridine outflow was accelerated by estracellular deosycytidine. These data demonstrate that the rate of uridine outflow was n function of the est,racellular concaentrstion of deosycytidine and was saturable.
The half-saturation constant, sK 711, was derived from reciprocal plots of uridine rf?lus uerahs deosycytidine concentration.
Comparison of SK, for efflus into medium containing deosycyt,idine wit,h that for efflux into uridine (Table I) indicates that deoxycytidine is almost as effective an accelerator of efflus as uridine.
,4t saturation, deosycytidine appears to accelerate efflus of radioact,ive uridine to a great,er extent than estracellular uridine--i.e. ST'm,, for efflus in t.he presence of deosyctytidine is greater than sV,,,~~ for efflus in the presence of uridine.
Other pyrimidine nucleosides \vere first t.ested for abi1it.y to accelerate uridine efflus at 0.5 and 6.0 mnl. If acceleration of efflus was observed at these concentrations, the compound was then tested at several additional concentrations in order to derive kinetic constants.
For each compound listed in Table I, determinations of efflus were performed using erythrocytes obt'ained from a single individual.
Graphic results for two of the nucleosides listed in Table I are presented in Fig. 4. The data in Table  I and Fig. 4 have been rela.ted to efflus into nonradioa.ctive uridine by expressing efflus in the presence of test nucleoside as Dhe percentage of efflus in the presence of 6 mu nonradioactive uridine.
When values for SK, are compwed (Ta.ble I), extracellular uridine and thymidine appear to accelerate efflux of radioactive uridine at the lowest estracellular concentrations.
Similar SK, values were obtained for both members of several pairs of ribosyl and 2'-deoxyribosyl nucleosides (compare uridine and 2'-deoxyuridine; cytidine and 2'-deoxycytidine). The compounds of Ta.ble I least able t.o accelerate uridine efflus are dihydrouridine, in which the base moiety dihydrouracil differs from uracil in having a puckered non-planar conformation (21), and pseudouridine, in which the ribose moiety is linked to Cg of the pyrimidine ring (22). The ability of dihydrouridine and pseudouridine t.o accelerate efflus of uridine indi-0 When cells were loaded by incubation with [2-14C]uridine, internctl uridinc concentrations at equilibrium were 5.5 to 6.5 mM. Media containing the equilibrium concentration of nonradioactive uridine were prepared to within 570 of the internal concentration. 3. Acceleration of uridiae efflux by 2'-deoxycytidine. Cells loaded with 6.1 rnbl [W%]uridine were incubated in TESbuffered saline or in TES-buffered saline cont,aining nonradioactive nucleoside.
The appearance of radi0activit.y in t,hc medium versus t.ime is pIot.ted for data obtained during 2 days using a single unit of blood. Panel 3: A, 6.1 mM uridine; 0, TKSbuffered saline; and q , 0.5 IIIM deoxycytidine.
Apparent kinet.ic constants were obtained from lines fitted by the method of least squares to reciprocals of efflux (expressed as percentage of maximum uridine-accelerated efflux of uridine) and the extracellular concentration of added test nucleoside. catea t.hat interaction of permeant with the nucleoside t.ransport system does uot require an aromatic haze moiety or the N-glycosidic linkage.
The pattern of these results was confirmed in other experiments in which uridine efflus was measured at extracellular concentrations of test nucleoside well above saturation for accelerative exchange diffusion. Initial rates of appearance of radioact,ivity in the medium were determined when cells loaded with 5.5 to 6.0 mM [2-14Cluridine were incubated in media containing different concentrations of nonradioactive nucleoside. insets contain plots Other nucleosides :dso accelerated uridine efflus but were tested only at n single concentration (Table  II)  derived from reciprocal plots of data in Fig. 5; however, assuming a maximum rate of loo%, visual inspection suggests a halfsaturation value for 2'Wmethyluridine of approximately 6.0 mr\l.
Failure of a particular compound to accelerate uridine efflux indicates either that it is not a permeant for the uridine transport system or that, if it, is a permeant, it is transported without Issue of May 25, 1972 C. 3. Cass ad A. R. P. Paterson 3319 Efflnx from cells loaded with 5.5 to 6.1 mu uridine into medium containing nonradioactive test compound is compared with efflux into TES-buffered saline and into 5.5 to 6.1 rnkc nonradioactive uridine.
For each preparation the concentration of external uridine xv-as within 57; of t,he internal uridine concentration; the concentrations given below are those of the external test nucleoside.   Table III. These data indicate that (a) nucleotides, free uracil, and free sugars do not accelerate uridine efflux, (b) nucleosides that are present in an ionized form do not. accelerate uridine effluq7 and (c) modification of the ribosyl or 2'-deosyribosyl moiety elimillntes or greatly reduces ability to accelerate uridine efflus.
The outward flux of radioactive uCdine was inhibited by four nucleosides (Table IV).
Inhibit'ion is indicated by a lower value for efflux into medium containing test compound than for efflux into TICS-buffered saline. To further examine their in-  to inhibit the efflux of uridine, this interaction may be prevented by t,he 1,resence of extracellular uridine.

DISCUSSIOX
Kinetic models proposed for facilitated diffusion are based primarily on data obtained from studies of monosaccharide t,ransport in human erythrocytes. The classical carrier model assumes mediation of transport by a carrier capable of moving or rotating within the membrane (g-13), (25)(26)(27). Lieb and Stein (28) have proposed a noncarrier model in \vhich transport is accomplished by a membrane-bound oligomeric protein that is simultaneously exposed to intra-and extracellular solutions and capable of undergoing substrate-induced conformational changes resulting in movement of permeant within the oligomeric protein and thence across the membrane.
Studies of exchange diffusion of monosaccharides in erythrocytes (11-13) indicate that when structurally related permeants are transported by the same system of facilitated diffusion, accelerative exchange diffusion occurs when the two permeants are present on opposite sides of the membrane.
During acce1eratiT-e exchange diffusion t,he transport mechanism, after releasing outgoing permeant molecules at the external membrane face, reorients or returns to the internal membrane face more ral'idly Vol. 247, n-o. 10 values for sV,,,, for efflus and results from studies of uptake of uridine and thymidine (I) and from the kinetics of equilibrium exchange diffusion of uridine and thymidine* suggest similar maximal rates of inflow for both nucleosides.
A procedure based on the dependence of uridine efflux on the concentration of permeant present at the external membrane face was used to eramine the chemical specificity of the nucleoside transport system in erythrocytes. Among pyrimidine nucleosides that act as permeants in the uridine-thymidine transport system are such diverse structural analogs of uridine as 2'.deosyuridine, 5-bromouridine, 5.aminouridine, 6-methyluridine, 3.deazauridine, dihydrouridine, pseudouridine, and 4-thiouridine.
The C-glycosides, pseudouridine and formycin B (I), are permeants indicating that Dhe N-glycosidic linkage is not crucial for interaction of permeant with the transport mechanism.
Earlier studies (I) demonstrated increased efflux of radioactivity when cells loaded with uridine were incubated in media containing 1.0 to 2.0 1llM adenosine, inosine, or guanosine, indicating that purine nucleosides are also permeants.
Failure of 6-azauridine and orotidine to accelerate uridine efflux suggests that the presence of charged groups on the base is not accommodated by the transport mechanism.
The transport mechanism is less tolerant of changes in the sugar portion of the permeant molecule than in the base. There appears to be little tolerance for substitution on the 2'-or 3'.hydrosyl groups since methylation at either position greatly reduces the ability of uridine or cytidine to accelerate efflux and the 2') 3'.substituted uridine derivatives inhibit efflux of uridine when assayed in the absence of external uridine.
However, the 2'-hydroxyl group is not essential for interaction of permeant with t,he transport mechanism since several pairs of pyrimidine ribo-and deosyribonucleosides that differ only at the 2' position have comparable SK, values; furthermore, arabinosylcytosine, in which the 2'Qydroxyl group is cis to the glycosidic linkage, promotes efflus as effectively as 2'-deoxyuridine.
These results and structure-act'ivity studies with purine nucleoside permeantsg support the conclusion that the sugar moiety is more important than the base in recognition of permeant by the transport mechanism.
HTG and 6-(methylmercapto)purine ribonucleoside inhibit' uridine efflus in the presence and absence of excess extracellular uridine, and HTG is inhibitory at concentrations as low as 0.06 PX. That HTG and other related S-substituted 6-thiopurine ribonucleosides (2) inhibit nucleoside transport suggests that addition of bulky hydrophobic groups to the sulfur atom of 6-thiopurine ribonucleoaide, s imparts a high affinity for the nucleoside transport mechanism. Inhibition of uridine efflux by the uridine derivative, hydroxynitrobenzylthiouridine, was much less effective than inhibition by HTG, suggesting greater 8 C. E. Cass aud A. 1:. I'. Paterson, unpublished results. 9 3'.l>eoxy-3'.methyladenosine, 9.p-1,.psicofuranosyladenille, 9-B-L-ribofuranosrl:lderline, alld 9~B-o-xylofuranos$ladelline inhibit or have no effect on uridine eifiux when assayed-according to the procedure described in this report (C. E. Cass and A. R. P. Paterson, unpublished results).
affinity of the inhibitory site or sites on the transport mechanism for purine ribonucleosides than for pyrimidine ribonucleosides.