Two Nucleoside Uptake Systems in Lactococcus lactis: Competition between Purine Nucleosides and Cytidine Allows for Modulation of Intracellular Nucleotide Pools

Abstract

A method for measuring internal nucleoside triphosphate pools of lactococci was optimized and validated. This method is based on extraction of ³³P-labeled nucleotides with formic acid and evaluation by two-dimensional chromatography with a phosphate buffer system for the first dimension and with an H₃BO₃-LiOH buffer for separation in the second dimension. We report here the sizes of the ribo- and deoxyribonucleotide pools in laboratory strain MG1363 during growth in a defined medium. We found that purine- and pyrimidine-requiring strains may be used to establish physiological conditions in batch fermentations with altered nucleotide pools and growth rates by addition of nucleosides in different combinations. Addition of cytidine together with inosine to a purine-requiring strain leads to a reduction in the internal purine nucleotide pools and a decreased growth rate. This effect was not seen if cytidine was replaced by uridine. A similar effect was observed if cytidine and inosine were added to a pyrimidine-requiring strain; the UTP pool size was significantly decreased, and the growth rate was reduced. To explain the observed inhibition, the nucleoside transport systems in Lactococcus lactis were investigated by measuring the uptake of radioactively labeled nucleosides. The K_m for for inosine, cytidine, and uridine was determined to be in the micromolar range. Furthermore, it was found that cytidine and inosine are competitive inhibitors of each other, whereas no competition was found between uridine and either cytidine or inosine. These findings suggest that there are two different high-affinity nucleoside transporters, one system responsible for uridine uptake and another system responsible for the uptake of all purine nucleosides and cytidine.

Lactic acid bacteria are used in the fermentation of different food products, like milk and cereals. The most important organism industrially is Lactococcus lactis, which is also used extensively as a model organism for studying the cell physiology of lactic acid bacteria. In order to optimize the performance of organisms used for starter cultures by genetic engineering or classical isolation of desired mutants, knowledge of the central biochemical pathways is crucially important. Some of the most important biochemical pathways are those responsible for the synthesis and salvage of nucleobases, nucleosides and nucleotides, since any growing organism needs nucleotides for the synthesis of DNA, RNA, and several coenzymes. Moreover, nucleotides act as allosteric effectors for a large number of enzymatic reactions. Therefore, alteration of nucleotide pool sizes may have significant effects on a large number of biochemical pathways and may lead to changes in cell physiology. The ability to control the nucleotide pools in L. lactis is dependent on detailed knowledge of nucleotide metabolism, which has been a major research area for the last 10 years in our laboratory (1, 2, 20, 21, 23-27, 33, 40, 41)

However, information concerning transport systems for nucleotide precursors in lactococci is scarce. It is known that hypoxanthine, adenine, and guanine facilitate growth of a purine-requiring mutant (34); thus, L. lactis is able to transport purine nucleobases across the membrane. It has been shown that all pyrimidine derivatives except cytosine can be metabolized (23), demonstrating permeability for these compounds. The only direct evidence for transport systems that has been established is evidence for the presence of a uracil permease (27). In the genome sequence of L. lactis IL-1403 an open reading frame encoded by pbuX has been identified as a purine nucleobase transporter (7). No open reading frames have been identified as nucleoside transporters (7). In Bacillus subtilis three different nucleoside transporters have been identified; nupC encodes a general pyrimidine nucleoside transporter (36), whereas yutK and yxjA are devoted to purine nucleoside uptake (H. H. Saxild, personal communication). The results presented in this paper suggest that L. lactis has at least two different nucleoside transporters; one is responsible for the uptake of uridine, whereas the other facilitates the uptake of the different purine nucleosides and cytidine.

In principle, two methods for determination of nucleotide pools are widely used. One method is based on high-performance liquid chromatography of extracted nucleotides, whereas the other is based on incorporation of radioactively labeled phosphate and subsequent thin-layer chromatography (TLC) of nucleotides extracted from the labeled cells. In this paper we describe a method for lactococci based on instantaneous extraction of ³³P-labeled nucleotides from lactococcal cells with formic acid, followed by separation of the nucleoside triphosphates by TLC.

MATERIALS AND METHODS

Growth conditions.

L. lactis cultures were grown on a synthetic medium (SA medium) based on MOPS (morpholinepropanesulfonic acid) containing seven vitamins and 19 amino acids (18) and supplemented with 1% glucose. L. lactis was cultured at 30°C in filled and capped culture flasks without aeration but with gentle agitation. For all plates, agar was added to a concentration of 15 g/liter.

Strains used.

Only L. lactis subsp cremoris MG1363 (13) and derivatives of this strain were used in this study. Strain LKH280 carrying a mutation in pyrG and cdd has been described previously (41). The purine-requiring strain MK185 and the pyrimidine-requiring strain MB420 carry an ISS1 insert in purD and pyrB, respectively. MK185 is a derivative of MK133 (21), whereas MB420 is a derivative of MB400 (27). In both of these strains plasmid pG⁺host9 (22) was cured from the chromosome by cultivating the strains at 28°C in order to select for plasmid excision. The temperature was then raised to 37°C, and the strains were grown in the absence of erythromycin, which resulted in plasmid loss. The strains were screened for the loss of erythromycin resistance and the presence of the purine or pyrimidine requirement. Strain MB22 carries an ISS1 insertion in the carB gene and was obtained by plasmid curing as described above. Isolation of the parental strain has been described previously (26).

Nucleotide pool measurements.

For labeling, 25 μCi of ³³P_i was added to a 0.5-ml culture at an optical density at 450 nm (OD₄₅₀) of 0.2 growing exponentially in SA medium, which resulted in a specific activity of 38 Ci/mol. Growth was continued until the OD₄₅₀ was 0.8. Fifty microliters of the culture was withdrawn for determination of the total radioactivity. In the standard procedure, the nucleotides were extracted by transferring 0.4 ml of the culture to 0.1 ml of 10 M prechilled formic acid. After three freeze-thaw cycles followed by incubation for 30 min in an ice-water bath, the cell debris was removed by centrifugation. For analysis 20 μl of extract was mixed with 2 μl of marker solution (nucleoside triphosphates and deoxynucleoside triphosphates [5 mM each]) and applied to polyethyleneimine (PEI) plates by 11 rounds consisting of application of 2 μl and drying under cold air. The plates were washed for 10 min in 96% ethanol, dried, and developed at room temperature in a 0.85 M KH₂PO₄ solution adjusted to pH 3.42 by addition of an equimolar H₃PO₄ solution. Following chromatography the plates were washed once in 10% (wt/vol) citrate and twice in distilled water (16). The plates were dried and chromatographed in 0.75 M LiCl supplemented with 7.5% H₃BO₃. The pH was adjusted to 6.8 by addition of solid LiOH. For determination of the specific activity of ³³P_i in individual experiments, 2 μl of a 10-fold dilution of the medium sample was spotted on a PEI plate. The plates were evaluated with an Instant Imager. Most of the spots on the chromatogram were identified by addition of markers; however, the locations of 5-phosphoribosyl 1-pyrophosphate (PRPP) and pyrophosphate were based on the spots identified by Jensen and coworkers (17). The radioactivity in an individual spot was determined by the number of total counts in an overnight exposure in the Instant Imager. By using the specific activity expressed as total counts per nanomole in the assays, the concentrations of the individual nucleotides were then determined as a function of the radioactivities in the individual spots. In order to express the concentrations of nucleoside triphosphates in nanomoles per milligram (dry weight), we used the following correlation: 0.2 mg (dry weight) ml⁻¹ corresponded to an OD₄₅₀ of 1.

ATP pool measurements.

ATP pools of exponentially growing cells of L. lactis MG1363 were measured by using a luciferin-luciferase ATP monitoring kit from Bio Orbit (catalog no. 1243-102) according to the protocols of the manufacturer. Cells of MG1363 were grown exponentially in defined media (18), and 0.9-ml portions of a culture were withdrawn at several different cell concentrations, as determined by OD₄₅₀, and processed essentially as described by Jensen et al. (19), except that 0.6 g of glass beads was added to the hot phenol.

Nucleoside uptake.

Cells were grown exponentially in SA medium, and at an OD₄₅₀ of 0.8 500 μl of cells was transferred to an Eppendorf tube containing a ¹⁴C-labeled nucleoside at a specific activity of 10 mCi/mmol. In order to determine an adequate incubation time, samples were taken after 15, 30, 45, and 60 s by using 1, 5, and 10 μM nucleoside. The medium was removed by filtration through a 0.22-μm-pore-size filter and immediately washed twice with 5 ml of water at room temperature. After drying, the radioactivity on the filters was determined with an Instant Imager. A filter with a medium sample was included in order to determine the specific activity in an individual experiment. The uptake rate was determined in nanomoles per second per microgram (dry weight). The uptake was linear up to 45 s. Consequently, an incubation time of 30 s was used for the rest of the experiments.

The classical Michaelis-Menten constants K_m and V_max were calculated by plotting the data in a double-reciprocal plot. The K_i values were determined by assuming that there was competitive inhibition; the velocity of nucleoside uptake at a substrate concentration corresponding to K_m was measured, and the concentration of the inhibitor (I) was varied. The slope of a plot of V₀/(V_i − 1) versus [I] gives 1/K_i(app), where V₀ and V_i are the velocities of uptake with inhibitor concentrations of zero and i, respectively, and K_i(app) is the apparent K_i. The true K_i value can be calculated from the following relationship: K_i = K_i(app)/(1 + S/K_m), where S is the substrate concentration. Since the substrate concentration was equal to the K_m, the true K_i was K_i(app)/2.

RESULTS

Optimization of a method for measuring nucleoside triphosphate pools in L. lactis.

Incorporation of radioactively labeled phosphate has been used successfully previously to determine the intracellular nucleotide pools of gram-negative bacteria (4, 5, 10, 12, 29, 31, 32), as well as those of a few gram-positive bacteria (28, 37). Hence, we decided to attempt to determine the intracellular nucleotide pools of L. lactis by using incorporation of ³³P_i and extraction of nucleotides with formic acid. The concentrations of formic acid used for extraction of nucleotides from Salmonella enterica serovar Typhimurium (6) and Streptococcus equisimilis (28) are very different (0.3 and 13 M, respectively). Therefore, several different concentrations were tested in order to find the concentration that resulted in the most quantitative extraction. A final concentration of formic acid of 2 M appeared to give the most quantitative extraction of nucleotides from L. lactis (data not shown). The nucleotide pools in L. lactis MG1363 were determined, and the results of a TLC experiment are shown in Fig. 1. The actual pool sizes were determined, and the results are presented in Table 1.

FIG. 1.

Autoradiogram of a TLC plate used for pool determination with the reference strain L. lactis MG1363. The identities of known spots are indicated. PPi, pyrophosphate.

TABLE 1.

Nucleotide pool sizes in MG1363^a

Nucleotide	Pool size (nmol [mg of dry wt]−1)
GTP	0.8 ± 0.2
ATP	4.0 ± 0.5
CTP	0.8 ± 0.2
UTP	1.6 ± 0.6
dGTP	0.05 ± 0.05
dATP	0.1 ± 0.1
dCTP	0.1 ± 0.1
dTTP	0.4 ± 0.2

^aMG1363 was grown at 30°C in SA medium supplemented with glucose.

In order to validate the method which we developed, we decided to measure the ATP pool of strain MG1363 using a different extraction method (hot phenol and glass beads), as well as a different enumeration assay (a luciferin-luciferase ATP monitoring kit). The ATP concentration was found to be 4.8 ± 0.4 nmol mg (dry weight)⁻¹, which corresponds well with the value found by using incorporation of ³³P_i in exponentially growing cells (4.0 ± 0.5 nmol mg [dry weight]⁻¹) and extraction with 2 M formic acid. Hence, we concluded that the method developed for labeling and extraction was efficient.

Growth inhibition of purine- and pyrimidine-requiring strains.

Strains carrying a mutated pyrG gene are unable to grow in the absence of cytidine. Since salvage of the pyrimidine nucleoside cytidine and salvage of the pyrimidine nucleoside uridine share at least the enzyme uridine kinase (udk) (Fig. 2), we investigated whether growth of a pyrG strain was inhibited by addition of high concentrations of uridine. The doubling time of strain LKH280 (pyrG cdd) was determined in defined medium supplemented with 20 μg of cytidine per ml and was found to be 56 min. If the same medium was supplemented with 500 μg of uridine per ml, the doubling time was twofold greater. This could be explained by uridine acting as a competitive inhibitor of cytidine uptake and/or on the phosphorylation of cytidine to CMP by the uridine kinase. We tested whether other nucleosides could act as inhibitors of growth of a cytidine-requiring strain, and we found that all common purine nucleosides suppress growth of LKH280. In contrast, no effect was seen when deoxyuridine and thymidine were used as inhibitors. Surprisingly, purines, including inosine, adenosine, and guanosine and the corresponding deoxyribonucleosides, were more efficient growth inhibitors than uridine. At concentrations of 500 μg/ml and above, complete inhibition of growth was observed, and concentrations around 100 μg/ml resulted in a twofold increase in doubling time.

FIG. 2.

Uptake and salvage of ribonucleosides in L. lactis. Abbreviations: AR, adenosine; GR, guanosine; IR, inosine; CR, cytidine; UR, uridine; A, adenine; Hx, hypoxanthine; G, guanine; U, uracil. The udk gene encodes the uridine kinase, which is responsible for the phosphorylation of both uridine and cytidine. IMP and UMP are the end products of the purine and pyrimidine biosynthetic pathways, respectively.

Since purines were found to inhibit growth of the cytidine-requiring strains, we wondered whether the reverse situation could be established. Hence, the purine-requiring strain MK185 (purD::ISS1) was used in a growth experiment in which the cells were grown in defined medium supplemented with 20 μg of inosine per ml. Different amounts of cytidine were added, and no effects were seen at concentrations below 500 μg/ml. If the cytidine concentration was higher, inhibition was observed. A cytidine concentration of 2,000 μg/ml resulted in a twofold reduction in the growth rate. The same result was found when cytidine was replaced by deoxycytidine. In similar experiments the effect of uridine on growth of the purine-requiring strain was investigated. Interestingly, no effect of uridine was observed even at a concentration of 5,000 μg/ml. Similar results were obtained when deoxyuridine and thymidine were used as inhibitors.

It is noteworthy that addition of inosine to strain MK185 led to an increased growth rate compared to the wild-type growth rate (Table 2). Hence, inosine was added to wild-type strain MG1363, and the doubling time was reduced from 56 to 46 min in the presence of inosine. Addition of pyrimidines had no effect on the growth rate of the wild-type strain. This finding suggests that MG1363 has a partial purine requirement.

TABLE 2.

Effects of the presence of different nucleotide precursors on nucleotide pool sizes in purine-, pyrimidine-, and cytidine-requiring strains

Strain	Genotype	Generation time (min)	Addition(s)a	Pool sizes (nmol [mg of dry wt]−1)
				GTP	ATP	CTP	UTP	PRPP
MG1363	WTb	56		0.8	4.0	0.8	1.6	2.9
MG1363	WT	49	A (15), Hx (15), GR (30)	2.9	7.9	0.5	1.7	0.4
MG1363	WT	49	IR (20)	3.9	9.0	0.8	2.4	0.4
LKH280	pyrG cdd	55	CR (20)	0.6	4.0	0.7	1.9	1.4
LKH280	pyrG cdd	112	CR (20), IR (100)	4.1	8.7	0.1	4.8	0.3
MK185	purD::ISS1	46	IR (20)	3.7	8.1	0.8	2.0	0.4
MK185	purD::ISS1	91	IR (20), CR (2,000)	0.9	3.1	1.3	1.7	1.4
MB420	pyrB::ISS1	174	CR (20), IR (100)	1.9	5.0	4.2	0.6	0.1
MB22	carB::ISS1	188		1.0	5.1	0.5	0.9	0.7

^aAll experiments were conducted at 30°C in SA medium with glucose, nucleosides, and nucleobases added. The values in parentheses are final concentrations (in micrograms per milliliter). Abbreviations: A, adenine; Hx, hypoxanthine; GR, guanosine; CR, cytidine; IR, inosine.

^bWT, wild type.

Finally, the growth of the pyrimidine-requiring strain MB420, in which an ISS1 element is integrated into the pyrB open reading frame, was investigated. Using either uridine or cytidine as a general pyrimidine source, we analyzed competition by purines. We found that growth on a combination of cytidine and inosine inhibited the growth of the cells, whereas no effect was seen when uridine was used in the presence of purines even at extremely high concentrations.

In conclusion, cytidine competes with both uridine and purine nucleosides, whereas no competition between uridine and purine nucleosides was observed. The most plausible explanation for this is that uridine and cytidine compete at the level of phosphorylation for the uridine kinase, while cytidine and purine nucleosides compete at the level of transport. Since no competition was observed between uridine and purine nucleosides, it could be predicted that these molecules are transported by different uptake systems. The inhibitory effect of purine nucleosides on cytidine uptake cannot be at the level of nucleoside phosphorylation, since the same enzyme, uridine kinase, is responsible for phosphorylation of uridine; hence, the inhibition most likely occurs at the level of transport into the cell. Consequently, the uptake of cytidine and the uptake of purine nucleosides are probably mediated by the same system.

Characterization of nucleoside transport in L. lactis.

In order to verify the assumption that cytidine and purine nucleosides have a common transporter while uridine uses a separate system, transport assays were performed. A prerequisite for measuring the true kinetic parameters for uptake is that no metabolism of the substrate can occur inside cell. This is not the case in the available strains. What we actually measured was a combination of transport and internal metabolism. Therefore, the derived kinetic constants cannot be regarded as true kinetic parameters for nucleoside uptake but rather are apparent kinetic parameters for nucleoside utilization. Consequently, the purpose of these experiments was to support the hypothesis for the growth inhibitory observations presented above.

Using different concentrations of the nucleosides in the assays allowed calculation of K_m and V_max as described in Materials and Methods. To test whether the three nucleosides might function as inhibitors, the K_i values were determined as described in Materials and Methods. The derived kinetic constants are shown in Table 3. The K_m values for inosine and cytidine are in the same range (approximately 2 μM), whereas the K_m for uridine is 1 order of magnitude higher. Furthermore, uridine showed no inhibitory effect on cytidine and inosine transport and visa versa. On the other hand, competition between cytidine and inosine was demonstrated. These findings explain the observation made in the growth experiments. Inosine seems to be a much better inhibitor of cytidine transport than cytidine is of inosine transport. Again, this finding accounts for the observation made in the growth experiments in which purine nucleosides were found to be much more efficient inhibitors than cytidine. Intuitively, it should be expected that the K_i for a compound acting as an inhibitor is identical to the K_m for the compound serving as a substrate. As stated above, we could not measure true kinetic parameters for uptake in this experiment, since metabolism of the substrate was occurring inside the cells. The data do, however, fit very well with the data obtained in the growth experiments. No effects were seen when uridine and inosine were used together, whereas inosine was a very potent inhibitor of cytidine uptake. Moreover, the finding that the K_i for cytidine with inosine was higher than the K_i for the opposite reaction correlates with the fact that a very high concentration of cytidine had to be used before any inhibition of the growth of a purine-requiring strain was observed.

TABLE 3.

Kinetic constants for nucleoside utilization

Substrate	Km (μM)	Specificity coefficient (U/μM)a	Vmax (U)b	Ki (μM) with the following inhibitors:
				Uridine	Cytidine	Inosine
Uridine	16 ± 3	4.0 ± 0.8	64 ± 3		NIc	NI
Cytidine	2.1 ± 0.2	5.7 ± 2.9	12 ± 6	NI		0.6 ± 0.2
Inosine	1.7 ± 0.2	17 ± 4	29 ± 7	NI	17 ± 4

^aV_max/K_m.

^bOne unit was defined as transport of 10³ molecules per s per cell across the membrane, assuming that one OD₄₅₀ unit corresponds to 8 × 10⁸ cells per ml.

^cNI, no inhibition observed.

Genes encoding potential nucleoside transporters cannot be identified on the chromosome of L. lactis.

The genome sequence of L. lactis subsp. lactis strain IL-1403 has been published previously (7). Strain MG1363 used in this study has very high sequence similarity to IL-1403. At the amino acid level the two strains diverge very little (around 1%). Therefore, the genome sequence of IL-1403 was searched for the presence of potential nucleoside transporters. As queries the three nucleoside transporters from B. subtilis NupC, YutK, and YxjA (Saxild, personal communication) were used. The best hit was with the putative multidrug efflux transporter BMR2, encoded by blt, with less than 25% identity in only a part of the peptides. When other known nucleoside transporters from Escherichia coli (30), Plasmodium falciparum (8), and Leishmannia (9, 39) were used as queries, no candidate sequences with levels of identity of more than 25% were found on the lactococcal chromosome.

Strains with altered nucleotide pools.

In order to determine whether the inhibition of growth by combined addition of nucleosides is a consequence of reduced nucleotide pools, growth experiments in the presence of ³³P_i were conducted. Subsequently, the nucleotide pools were determined as described above. Table 2 shows the sizes of nucleotide pools in different purine- and pyrimidine-requiring strains grown in the presence of combinations of different nucleosides that led to reduced growth rates. As mentioned above, addition of purines to the medium stimulates growth of L. lactis wild-type strain MG1363, implying that cells growing in the absence of exogenous purines suffer from purine starvation. This was verified by the pool measurements; in the presence of purines the GTP pool size increased fourfold, whereas the ATP pool size doubled in wild-type strain MG1363 when either a mixture of purine nucleobases or inosine was used as an exogenous purine source (Table 2). The same effect was seen in purine-requiring strain MK185; addition of excess cytidine together with inosine resulted in decreases in the purine pool sizes. The amount of ATP was even lower than the amount found in the wild type.

The size of the CTP pool alone can be significantly decreased by using a pyrG strain. Restricting cytidine uptake in strain LKH280 (pyrG cdd) by addition of inosine led to a 1-order-of-magnitude decrease in the CTP pool size compared to the wild-type MG1363 pool size. Simultaneously, the sizes of the purine pools increased since the strain no longer experienced purine starvation like wild-type strain MG1363 in the absence of purines. The size of the pool in strain LKH280 supplemented with only 20 μg of cytidine per ml was identical to the size of the wild-type MG1363 pool.

Strain MB22 has a partial pyrimidine requirement caused by insertion of an ISS1 element into the carB gene (26). In this strain, the carbamoyl phosphate required for de novo synthesis of pyrimidines is obtained by degradation of arginine through the arginine deiminase pathway. Previously, it was shown that this pathway could produce sufficient carbamoyl phosphate only at lower growth rates (26). The sizes of the nucleotide pools were determined in this strain, and it was found that purines are present at wild-type levels, whereas the pyrimidine pools are significantly smaller.

In order to obtain pyrimidine starvation in an alternative way, strain MB420 (pyrB::ISS1) was grown on cytidine as a pyrimidine source in the presence and absence of inosine. As described above, inosine inhibits cytidine uptake, resulting in a reduced growth rate. The pool sizes in the repressed situation are shown in Table 2. The sizes of the purine pools were found to be slightly elevated, whereas the sizes of the pyrimidine pools were changed. The UTP pool was small, as expected, whereas surprisingly, the size of the CTP pool increased fivefold. This observation suggests that under the experimental conditions used, which inhibited cytidine transport activity, the pathway through phosphorylation of cytidine all the way to CTP was more efficient than the deamination of cytidine to uridine and the subsequent conversion to UTP.

DISCUSSION

Substrate specificity of nucleoside uptake systems.

In this study we demonstrated that nucleoside transport in L. lactis has the following specificities: one system is responsible for the uptake of uridine and probably deoxyuridine, whereas another system has a broader specificity and is responsible for the uptake of cytidine and deoxycytidine and all purine ribonucleosides and deoxyribonucleosides. This is an unprecedented pattern. In E. coli two different uptake systems located in the inner membrane have been detected. One system, encoded by nupG, is capable of transporting all nucleosides, whereas the C system, encoded by nupC, is responsible for the uptake of adenosine, deoxyadenosine, cytidine, deoxycytidine, uridine, deoxyuridine, and thymidine (30). In addition, the tsx gene encodes an outer membrane protein which facilitates the transfer of all nucleosides except cytidine and deoxycytidine across the membrane (15). The nucleoside transporters from Leishmannia have been cloned and characterized, and it has been determined that they have yet another substrate specificity. Two genes have been identified which encode two different transporters, which are responsible for the transport of adenosine and pyrimidine nucleosides (LdNT1) and for the transport of inosine and guanosine (LdNT2) (9, 39)

Nucleoside transporters with broad substrate specificity exist. In the malarial parasite P. falciparum the nucleoside transporter PfNT1 has been identified, and this transporter has the ability to transport a large number of purines and pyrimidines and their analogs across the membrane of the parasite (8). A nucleoside transporter with broad specificity is also known from humans; hENT1 has been identified in erythrocytes and has been shown to facilitate the uptake of all purine and pyrimidine nucleosides (14). Other protozoans, like Tryanosoma brucei and Toxoplasma gondii, have been shown to possess multiple nucleoside transporters (11, 35). None of these transporters has specificity like that of L. lactis. When nucleobase uptake systems are considered, a transporter has been found in Saccharomyces cerevisiae that has specificity for adenine, guanine, hypoxanthine, and cytosine.

It is not unprecedented that enzymes that are involved in the metabolism of nucleosides have specificity for adenosine, guanosine, and cytidine and the corresponding deoxyribonucleosides. In Mycoplasma a deoxynucleoside kinase that phosphorylates deoxyadenosine, deoxyguanosine, and deoxycytidine has been identified (42). Moreover, a ribonucleoside hydrolase from Lactobacillus delbrueckii was found to have specificity for all ribonucleosides (38).

Control of internal nucleotide pool sizes by addition of nucleosides to the growth medium.

In addition to being precursors in the synthesis of RNA and DNA, nucleotides are allosteric effectors for a large number of biochemical reactions in the cell. Therefore, altered nucleotide pool sizes must result in intracellular flux changes in many biochemical pathways. Reduced growth rates are obtained by growing mutants with mutations in nucleotide biosynthesis genes under conditions in which the availability of nucleotide precursors is limited. Normally, this is accomplished by performing laborious experiments with fermentors, which requires large quantities of synthetic media. If nucleotide pool sizes were measured by radioactive labeling, fermentor experiments would be extremely expensive due to the large volumes of media. Here we demonstrate that it is possible to obtain partial nucleotide starvation by performing simple batch experiments on the milliliter scale without expensive equipment. We have shown that it is possible to change both the purine and pyrimidine pools. The most dramatic effect was seen during specific CTP starvation. Here the size of the CTP pool was reduced 1 order of magnitude, and the UTP/CTP ratio increased 24-fold. The opposite effect (reduced UTP pool size and a 14-fold increase in the CTP/UTP ratio) was observed with the pyrB strain.

PRPP pools versus purine pools: implications for regulation of expression of the purine biosynthetic genes.

Addition of exogenous purines to the growth medium leads to repression of the purine biosynthetic genes in L. lactis. It has previously been shown that this control is exerted by the transcriptional activator PurR, which is very similar to the PurR repressor from B. subtilis (51% identity at the amino acid level) (21). In B. subtilis it was shown that PRPP is the true inducer (43). The effect of purine addition on the PRPP concentration in B. subtilis was proposed to be a result of two events. Conversion of purine bases to the corresponding nucleoside monophosphates catalyzed by phosphoribosyl transferases would deplete the PRPP pool. Furthermore, the increases in the concentrations of all purine nucleotides would include an increase in the concentration of ADP, which is known to be a very potent inhibitor of the PRPP synthase in B. subtilis (3). Since a PRPP binding site was found in the L. lactis PurR protein, it was speculated that PRPP also is the effector in L. lactis, in this case converting PurR to its activating conformation. Table 2 shows that a small PRPP pool is correlated with the presence of exogenous purine, thus supporting the hypothesis that PRPP can function as a reporter of purine nucleotide status within the cell. From the data presented in Table 2 it was concluded that reference strain MG1363 has a partial purine requirement, since addition of exogenous purines led to increases in both the growth rate and the sizes of the purine nucleoside triphosphate pools. During severe purine starvation when purine uptake is inhibited by cytidine in a purine-requiring strain, the purine pools intuitively must be at the lowest levels. Since the GTP concentration is the same as that found in the wild type and the level of ATP is only slightly reduced, the data suggest that the sizes of the purine pools in the wild-type strain growing in a medium without exogenous purines are at the lowest values possible for a growing strain.