Nucleoside Transport in Cells and Membrane Vesicles from Escherichia coli Kl2*

Osmotic shock treatment of cells of Escherichiu coli K12 caused a reduction in the transport of nucleosides into the cells. The strains used carried mutations in the nucleoside catabolizing enzymes. This indicated that the decrease in transport capacity was not due to loss of these enzymes during the shock treatment. Membrane vesicles, prepared from the same strains, showed a limited transport of cytidine, deoxycytidine, and uridine. Transport of purine nucleosides and of thymidine was very low in vesicles lacking the appropriate nucleoside phosphorylases and no significant stimulation was observed if the nucleoside phosphorylases were present in the membrane vesicles. These results all indicate that components outside the cytoplasmic membrane are important for nucleoside transport. Selection for resistance to fluorodeoxycytidine yielded mutants which were unable to transport any nucleoside, even when the nucleoside phosphorylases were present in high amounts. This finding is consistent with a requirement for a specific transport process prior to the initial enzymatic attack on the incoming nucleoside.

Selection for resistance to fluorodeoxycytidine yielded mutants which were unable to transport any nucleoside, even when the nucleoside phosphorylases were present in high amounts.
This finding is consistent with a requirement for a specific transport process prior to the initial enzymatic attack on the incoming nucleoside.
Nucleosides are readily taken up and catabolized by cells of Escherichia coli. It was shown previously that an energyrequiring nucleoside transport may take place independent of subsequent catabolism (1). This was demonstrated in mutant cells, unable to metabolize the nucleosides, and concentrating effects of 300-to 500-fold were estimated for uridine and cytidine (1).
Several lines of evidence have indicated the existence of at least two different nucleoside transport systems in E. coli (2)(3)(4)(5). For further reference see Ref. 5. The two systems are termed the nupC and the nupG systems',' and they may be * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
' These genes were previously termed cru and gru, respectively (4). '  by their different resistance to the uridine analogue showdomycin, since the nupC system is strongly inhibited by this antibiotic while the nupG system is not.
Gene(s) coding for the nupC system have been located at 50 min on the revised E. coli chromosome, 65% co-transducible with pt.~Z.~ This system transports nucleosides except guanosine (and deoxyguanosine) and it is regulated (5) by the cytR gene product, which is also a repressor in the control of the nucleoside-catabolizing enzymes, listed in Table I (6). Gene(s) coding for the nupG system map at 64 min on the revised E. coli chromosome, 25% co-transducible with metC3 This system is subject to a double control, since it is regulated by the cytR and the deoR gene products (5). The deoR repressor also participates in the control of the nucleosidecatabolizing enzymes (6). Komatsu and Tanaka (4,7) demonstrated the transport of adenosine and deoxycytidine in membrane vesicles of E. coli K12, wild type, or a showdomycin-resistant mutant. Similarly, Hochstadt and co-workers (8,9) prepared vesicles from Salmonella typhimurium, which were able to transport uridine, inosine, and adenosine. In this work, it was suggested that the transport consisted in a group translocation of the nucleoside, catalyzed by the appropriate nucleoside phosphorylase, i.e. uridine phosphorylase or purine nucleoside phosphorylase. On the other hand, there is evidence that extensive transport of nucleosides takes place in strains lacking the nucleoside phosphorylases (1,5). Moreover, a mutant was described which was unable to transport nucleosides although it contained normal amounts of the nucleoside-catabolizing enzymes (5). This indicated separate systems for transport and catabolism, and further evidence for such separate systems is presented below. There is, however, an intimate coupling between transport and cat@olism of nucleosides, and for reasons mentioned below this coupling has been a serious complication in the detailed measurements of nucleoside transport.
The main feature in the catabolism of nucleosides is a phosphorolytic cleavage of the N-glycosidic bond between the base and pentose moieties of the nucleosides. In these reactions, which are catalyzed by specific phosphorylases, a pentose phosphate is formed, which is then channelled into the pentose metabolism, while the bases may be used for synthesis of nucleoside monophosphates. This synthesis requires 5'phosphoribosyl-1-pyrophosphate and is catalyzed by specific phosphoribosyltransferases.
If the strains lack the appropriate phosphoribosyltransferase, the base is quantitatively excreted, but even in the presence of the enzyme a substantial amount is lost to the medium (1).
As an example, the catabolism of cytidine and uridine is shown in Fig. 1  A deamination of adenosine may also occur initially, but is not required for the phosphorolysis to take place. Table I schematizes the primary enzymic attacks exerted by the cells on exogenously added nucleosides. The enzymes are all inducible and, with the exception of adenosine deaminase, they are regulated together in two complex control systems, governed by the cytR and deoR genes (6).
The phosphorolytic attack on exogenously added nucleosides takes place with a high velocity, probably due to a localization of the enzymes at or near the cell surface (10-13). The above mentioned partial or complete excretion of the bases occurs with a similar high velocity (l), and this ter, s to disturb the standard measurements of transport rates, since in the commercially available '%-labeled nucleosides the bases always carry a radioactive label. A variable amount of counts is therefore lost from the cells into the medium when the nucleosides are cleaved phosphorolytically.
For this reason, strains with point mutations or deletions in the nucleoside phosphorylases have been used in our previous experiments whenever possible. Likewise, the strains contained mutations in the @dine-uridine kinase (1,5). In the present study, such mutants have been used to show that osmotic shock treatment of the cells reduces the ability of the cells to transport nucleosides. Membrane vesicles, prepared from the same strains, are shown to be capable of a limited transport of nucleosides. This transport is only slightly stimulated, if the vesicles are prepared from strains containing the appropriate nucleoside phosphorylases. Results with a transport negative mutant containing the enzymes support the contention that for significant utilization of the nucleosides by the cells a transport process must occur prior to catabolism.  Table II. Mutants containing the nupC system or the nupG system were obtained by mating and recombination in a transport negative strain Se 587 as reported previously (5). In the isolation of the mutant strains, used for uptake measurements, advantage was taken of the map position of the two regulatory genes, cytR and deoR, close to metB and clm4, respectively (6). Strain Se 1130 was isolated from strain S+ 430 cdd, metB, udk (14) in Assays-The procedures used for enzyme assays on cell extracts have been described previously (21), and these procedures were also applied when the enzymes were determined at the different stages during preparation of the vesicles. The suspensions of spheroplasts and of membrane vesicles were subject to sonication before the assays in order to release the enzymes, while the centrifuged cell lysates and the supernatants from spheroplast formation were used directly in the assay mixture. One enzyme unit is defined as the amount of enzyme which converts 1 nmol of substrate per min at 37°C.

Transport and Accumulation of Free Nucleosides
The strains used in the present work contain various mutations in the enzymes which catalyze nucleoside metabolism (cf. Table II). The chemical state of the intracellular nucleosides after transport was identified for each nucleoside in strains carrying the appropriate mutation. Uridine was measured in S+ 1142, and deoxycytidine, cytidine, guanosine, and thymidine in strain S+ 1130 cdd, udk, Adeo. S+ 1142 cannot metabolize uridine, while S+ 1130 is inert towards deoxycytidine and cytidine, besides being unable to phosphorolyze guanosine and thymidine.
No mutations in the guanosine kinase or the thymidine kinase genes have been available, but, as shown in Table I, very little of these nucleosides is converted to phosphorylated compounds during the time interval of the transport assay. Cytidine, deoxycytidine, and uridine appear nearly intact.
Parallel experiments were performed with cells pretreated with dinitrophenol (1 ITIM). These cells accumulated less than 10% of the activity found in the control cells and chromatography of the cell extracts did not yield discrete radioactive spots. Thus, the nucleosides may be transported and accumulated in a chemically intact state, provided the cells carry mutations in the appropriate catabolizing enzymes.

Function of Components
Outside the Cytoplasmic Membrane in Nucleoside Transport There have been reports that components outside the cytoplasmic membrane participate in the transport of nucleosides. Thus, the Ts receptor protein (the tsx gene product) stimulates the utilization of certain nucleosides (16,22). The specificity of this system is seen in Fig. 2. All nucleosides tested except cytidine and deoxycytidine are affected by the absence of the tsx product. The effect is most pronounced when nucleosides are present at low (0.1 PM) concentration. Strains lacking the tsx gene product grow well on agar plates with nucleosides as sole carbon sources (1 mM concentration), provided the necessary enzymes for nucleoside catabolism are present in the cells (22).
Osmotic shock treatment, carried out according to the procedure of Heppel (19), causes a reduction in the transport of all nucleosides. Two strains, S+ 1130 and S+ 1131, containing the nupG and the nupC system, respectively, were investigated and both systems were impaired (Table IV) of glycine before and after osmotic shock treatment was used as a control, since transport of this amino acid occurs independent of shock-releasable proteins (23). The rate of glycine transport is reduced approximately 50% after the shock treatment (Table IV). This probably reflects partial damage of cells, while the more extensive (80 to 90%) reduction of nucleoside transport suggests that the shock treatment releases factors which are essential for the transport of nucleosides.
Osmotic shock treatment causes partial release of the nucleoside-catabolizing enzymes (5), but since the strains used in the experiments in Table III carry mutations in these enzymes, the reduced transport cannot be ascribed to a loss of the enzymes during the shock procedure.

Enzymatic
Properties of Membrane Vesicles Routinely, the strains used for transport measurements harbored a deletion in the genes which code for thymidine phosphorylase and purine nucleoside phosphorylase. Some of the strains also contained mutations in cytidine deaminase or uridine phosphorylase (Table II). It has been suggested, however, that nucleoside phosphorylases perform transport func- tions in membrane vesicles through a group translocation process (8,9). To test the influence of the enzymes on transport, membrane vesicles also were prepared from a strain containing the nupG system and carrying the intact genes for the nucleoside phosphorylases.
Due to the cytR and deoR gene mutations, the enzymes, when present, were synthesized in maximally derepressed amounts. These enzymes are soluble proteins and it seemed important to know to what extent they were retained in the final vesicle preparations. Table V shows the distribution of enzyme activities as measured at each step of the procedure in the preparation of vesicles from two transporting strains and from two negative strains. The activities of the three nucleoside phosphorylases as well as of cytidine deaminase were determined, since these four enzymes catalyze the initial, specific cleavage of the transported nucleosides (Table I). Three important features emerge from the table.  1. Very little enzyme activity is released in the medium, when spheroplasts are formed by lysozyme-EDTA treatment. The bulk of activity does not appear in the medium until after lysis. This supports the notion that the enzymes are located inside the cytoplasmic membrane rather than in the periplasmic space (10).
2. Small, but significant enzyme activities are retained by the vesicles, but no enrichment of the enzyme is observed. Specific activities drop 50-to loo-fold, compared to crude extracts. Thus, there is no evidence that the nucleoside phosphorylases are integrated in the cytoplasmic membrane. 3. The residual activities in the membrane vesicles of the four enzymes are the same in vesicles derived from the transport negative strains, and from the corresponding transporting strains.

Comparison between Nucleoside Transport in Whole Cells and Membrane Vesicles
The nupG System-Cells of S$I 285, which contains the nupG system, but lacks thymidine phosphorylase and purine nucleoside phosphorylase, transport all nucleosides efficiently (Fig. 3A). The ratio between intracellular and extracellular concentrations may be calculated, assuming an intracellular water volume of 0.05% of the cell suspension volume (24) at a cell density of 2 X 10' cells/ml. At the substrate concentration employed (1.8 X 10m6 M), a ratio of 590 was found for deoxy-@dine, in agreement with values found previously for cytidine (1). For the amino acid proline, a ratio of 122 was calculated.
Membrane vesicles prepared from this strain retained some ability to transport cytidine, deoxycytidine, and uridine. Based on the assumption that the inner volume of membrane vesicles from E. coli K12 is 5 pl/mg of protein (25), the concentrating effects for membrane vesicles were calculated to be lo-fold for deoxycytidine and 28-fold for proline. Thus, the ability of the vesicles to transport and concentrate proline was retained to a higher extent than the ability to transport the nucleosides.
Transport of guanosine, adenosine, and thymidine was negligible in membrane vesicles from strain S+ 285, which lacks the appropriate nucleoside phosphorylases, i.e. purine nucleoside phosphorylase and thymidine phosphorylase. When membrane vesicles were prepared from strain S+ 1135, containing the enzymes in high amounts, some enzyme activity was retained (Table V) but this activity was not able to stimulate the transport of purine nucleosides or thymidine significantly (Fig. 3B). A transport negative mutant, S4 861, was derived from S+ 285 by selection for resistance to FdCyd (10 ag/ml). The mutant strain was unable to transport any nucleoside in whole cells as well as in membrane vesicles (Fig. 3, A and B). A derivative of this mutant, L!Q 1135, containing all nucleosidecatabolizing enzymes in high amounts (Table V) did not show significant transport, neither in whole cells nor in membrane vesicles. Transport of the ammo acid proline was unaltered in cells and vesicles of this mutant strain, compared to the parent strain (Fig. 3, A and B).
The nupC System-Cells of strain S+ 304, containing the same deletion mutation in pm-me nucleoside phosphorylase and thymidine phosphorylase as strain S+ 285, transport all nucleosides except guanosine (5). The transport pattern in membrane vesicles from this strain was similar to that found for the nupG system. Some transport of @dine, deoxycytidine, and uridine was observed, but the rate was reduced compared to whole cells (Fig. 4). Only insignificant transport of purine nucleosides or thymidine could be detected in the membrane vesicles (not shown).
Also here, a transport negative derivative, S+ 1145, resistant whole ceils to FdCyd, was isolated and tested. None of the nucleosides were transported by this mutant strain, neither in whole cells nor in vesicles (shown for pyrimidine nucleosides only (Fig.  4)). The levels of cytidine deaminase and uridine phosphorylase were unaltered in the mutant strain.

Characteristics of the Two Transport Systems
Transport of deoxycytidine was examined in more detail in membrane vesicles containing either of the two transport systems. To ensure that catabolism did not interfere with the results, these vesicles were prepared from two strains containing a mutation in cytidine deaminase in addition to the previously mentioned enzyme mutations (Table II). Only the results for deoxycytidine are shown, but transport of cytidine and uridine responds to the various inhibitors in a manner similar to that of deoxycy'tidine transport. As shown in Fig. 5, the two transport systems may be distinguished from each other in whole cells by their behavior towards guanosine or showdomycin as inhibitors of deoxycytidine (or cytidine) transport. Thus, in cells containing the nupC system, the transport is not inhibited by guanosine (20 PM), but strongly inhibited by showdomycin (0.1 mM). Conversely, deoxycytidine transport in the nupG system is strongly inhibited by guanosine, but much less affected by showdomycin (Fig. 5). The difference in behavior towards showdomycin and guanosine is very pronounced in whole cells (Fig. 5, A and C) , but it is not very evident in vesicles (Fig. 5, B and D). Here, both systems are inhibited by showdomycin as well as by guanosine and the difference observed is only quantitative.
The inhibition by guanosine and showdomycin in the vesicles is specific for nucleosides. Neither of the two compounds affects proline transport significantly (not shown). In other respects, the two transport systems showed the same properties in the membrane vesicles as found in whole cells. In both systems, the transport was energy-requiring as shown by the stimulation by the artificial electron donor, PMS + ascorbate. It was inhibited by dinitrophenol (0.1 mM) and-to a lesser extent-by N-ethylmaleimide (4 mM) (Fig.  5). Adenosine, uridine, and thymidine, added in lo-fold excess, were strong inhibitors (not shown). If unlabeled substrate is added in excess 3 min after addition of the labeled substrate, the major part of the radioactivity is lost from the vesicles (Fig. 5). This indicates that the nucleoside is transported and accumulated without appreciable degradation.
L In a previous paper, a K, of 0.3 PM for cytidine transport in whole cells was determined (7). K, for transport of cytidine and deoxycytidine in membrane vesicles was estimated to 0.5 to 0.8 PM, i.e. the same order of magnitude as that found in whole cells.

Uridine Transport
In both transport systems, uridine transport exhibits the same properties as deoxycytidine transport, but the measurements require special consideration, since the membrane vesicles contain some uridine phosphorylase activity (Table V), which is sufficient for a rapid phosphorolytic splitting of uridine. This is concluded from Fig. 6, where transport is measured of uridine, labeled either uniformly or in the base moiety alone. In the transport assay, only [ U-%]uridine gives rise to accumulation of radioactivity in the membrane vesicles. In both cases, the uridine must be transported and phosphorylized and the ribose moiety retained by the vesicles, probably as ribose l-phosphate (9). Since the strains contain upp mutations (Table II), no uracil phosphoribosyltransferase is available to form uridine monophosphate, and the free uracil is excreted into the medium. When uniformly labeled uridine is used as a substrate, the radioactivity from the labeled ribose is retained by the vesicles (Fig. 6A). If [2-'%]uridine is employed, the pentose is unlabeled and no radioactivity is accumulated (Fig. 6B). Only in membrane vesicles prepared from a strain lacking uridine phosphorylase is it possible to demonstrate a transport of [2-%]uridine which equals that of [ U-"C]uridine.
This uridine transport is energy-requiring and shows saturation kinetics (Fig. 6, A and B).
If the same measurements are performed in vesicles prepared from the transport negative strain S+ 861, neither of the labeled uridines causes any accumulation of radioactivity (Fig. 6) vesicles in the same amounts as in the vesicles prepared from transporting cells (Table V). DISCUSSION

Transport Systems in Whole Cells and Membrane
Vesicles-The two nucleoside transport systems in E. coli, the nupC and the nupG systems, are strongly affected by osmotic shock treatment, indicating that in both cases extramembraneous components are important for the transport. One outer membrane protein, the Ts receptor, has been shown to stimulate nucleoside transport, but the loss of this protein after osmotic shock cannot explain the general reduction in nucleoside transport, since the receptor was shown to have no influence on the transport of cytidine and deoxycytidine (Fig.   2).
The present results further demonstrate that membrane vesicles from cells containing either of the two transport systems are able to carry out, some nucleoside transport. The transport rates are low compared to whole cells, but in this low efficiency range the nucleosides seem to fall in two categories. Cytidine, deoxycytidine, and uridine are transported in significant amounts, while purine nucleosides and thymidine are poorly transported under the standard assay conditions employed in this investigation. The presence of residual amounts of nucleoside phosphorylases does not stimulate this transport significantly.
Transport of Cytidine, Deoxycytidine, and Uridine-These nucleosides are transported by both transport systems, in cells as well as in vesicles, but in vesicles the two systems exhibit properties that are in some respects different from those found in whole cell experiments.
The transport data for deoxycytidine (Fig. 5) show that the vesicles containing the nupC system have lost partially the resistance to guanosine, which is characteristic for this system, when deoxycytidine transport is measured in whole cells. Similarly, the resistance to showdomycin is partially lost in vesicles, containing the nupG system.
The change in response to inhibitors in membrane vesicles, as well as the strongly reduced transport in osmotically shocked cells, indicates that components of the two transport systems are lost during the preparation of vesicles or during the osmotic shock treatment.
These components are important, but not essential for nucleoside transport, since membrane vesicles retain the ability to carry out some transport. The stimulation of this transport by PMS + ascorbate (Fig.  5) implies that the nucleoside transport systems are able to derive their energy directly from the energized membrane.
The data available at present do not allow a decision between the following two possibilities: 1) membrane vesicles, prepared from strains containing either the nupG or the nupC system, each carry a separate transport system or 2) one common membraneous translocation system serves two different components or systems, located outside the cytoplasmic membrane.
Transport of Purine Nucleosides and Thymidine-The inability of the membrane vesicles to support significant transport of these nucleosides confirms the participation in nucleoside transport of protein factors localized outside the cytoplasmic membrane. It also indicates that transport of purine nucleosides and thymidine requires one or more components different from those functioning in the transport of cytidine nucleosides and uridine. Such components might be lost during the vesicle preparation procedure due either to instability or to a localization outside the cytoplasmic membrane. So far, all attempts to isolate binding protein(s) for nucleosides have failed.
Separate Systems for Transport and Catabolism-There is abundant evidence that transport of nucleosides into whole cells may take place independent of the nucleoside-catabolizing enzymes. Thus, in strains devoid of thymidine phosphorylase and purine nucleoside phosphorylase, an efficient transport and accumulation of thymidine, guanosine, and adenosine are observed (5) (Fig. 3A and Table III). Furthermore, strains deficient in cytidine deaminase or uridine phosphorylase, the fast enzymes to attack cytidine and uridine, respectively, retain the ability to transport these nucleosides although no catabolism takes place (1, 2) (Table III). Also, the regulation of the transport process by the cytR and deoR repressors occurs independent of the presence of enzymes (5). In the present paper, it is shown that the presence of the nucleoside-catabolizing enzymes is not sufficient for transport to take place. Selection for resistance to FdCyd in a transporting strain, containing the nupG system, yielded a mutant, S+ 861, which is unable to transport any nucleoside ( Fig. 2A). A derivative of this mutant, containing all nucleoside-catabolizing enzymes in high amounts did not show significant nucleoside transport either (Fig. 3A).
Experiments with membrane vesicles corroborate the results obtained with whole cells. Cytidine and deoxycytidine are transported in significant amounts by membrane vesicles containing either of the two transport systems, but lacking cytidine deaminase, while no transport can be measured in membrane vesicles prepared from the transport negative strain (Fig. 3B). Similarly, uridine is transported by vesicles, lacking uridine phosphorylase, while vesicles from the transport negative strain, which contain uridine phosphorylase, are unable to support transport (Fig. 6). In the case of purine nucleosides and thymidine, the presence of the corresponding nucleoside phosphorylases does not stimulate transport in vesicles from the transport negative strain (Fig. 3B).
The experiments on uridine transport in membrane vesicles demonstrate the intimate coupling between transport and phosphorolytic cleavage of the nucleoside (Fig. 6). Here the transported uridine is quantitatively phosphorylized by tl .e uridine phosphorylase even at the fust sampling at 10 s. .4 similar rapid phosphorolytic cleavage of uridine by membre ,re vesicles from 5'. typhimurium has been interpreted as a group translocation of the nucleoside through the cytoplasmic membrane, catalyzed by uridine phosphorylase (8,9). In the present study, the lack of transport in whole cells and vesicles from S+ 861 and S+ 1135 (Fig. 3, A and B) speaks against such a mechanism as the major way of nucleoside transport, since these strains contain uridine phosphorylase in the same amounts as the parent strain. Moreover, if the uridine phosphorylase was spanning the cytoplasmic membrane as suggested (9), an enrichment of this enzyme would be expected in the membrane vesicles as compared to the whole cells. Instead, a substantial loss in specific activity occurs (Table V).
The nature of the mutation to transport negativity is being investigated.
The simultaneous loss of transport function for all nucleosides but not for proline indicates that the mutation is in a component essential and specific for nucleoside transport. Since the transport negativity is expressed in both whole cells and membrane vesicles, the mutated component probably resides in the cytoplasmic membrane.
In summary, the present results provide strong evidence that cells of E. coli contain energy-requiring transport systems which will transport nucleosides into the cells independent of the catabolic process. This does not exclude the possibility that in wild type cells the nucleoside phosphorylases may mediate some transport.
The failure of these enzymes to support transport in strain S+ 1135, nupG, nupC, is consistent, although, with the hypothesis that a transport is required prior to the initial enzymatic attack on the incoming nucleo-side. This also places the nucleoside catabolizing enzymes inside the permeability barrier of the cell, a location which agrees with the failure of EDTA-lysozyme treatment (Ref. 10 and this paper) to cause their release.