Expression of the pyrG gene determines the pool sizes of CTP and dCTP in Lactococcus lactis Casper M. Jørgensen, Karin Hammer, Peter R. Jensen and Jan Martinussen Bacterial Physiology and Genetics, BioCentrum-DTU, Technical University of Denmark, Kgs. Lyngby, Denmark The pyrG gene from Lactococcus lactis encodes CTP syn- thase (EC 6.4.3.2), an enzyme converting UTP to CTP. A series of strains were constructed with different levels of pyrG expression by insertion of synthetic constitutive pro- moters with different strengths in front of pyrG.These strains expressed pyrG levels in a range from 3 to 665% relative to the wild-type expression level. Decreasing the level of CTP synthase to 43% had no effect on the growth rate, showing that the capacity of CTP synthase in the cell is in excess in a wild-type strain. We then studied how pyrG expression affected the intracellular pool sizes of nucleotides and the correlation between pyrG expression and nucleotide pool sizes was quantified using metabolic control analysis in terms of inherent control coefficients. At the wild-type expression level, CTP synthase had full control of the CTP concentration with a concentration control coefficient close to one and a negative concentration control coefficient of )0.28 for the UTP concentration. Additionally, a concen- tration control coefficient of 0.49 was calculated for the dCTP concentration. Implications for the homeostasis of nucleotide pools are discussed. Keywords: pyrG; CTP synthase; metabolic control analysis; metabolism; EC 6.4.3.2. 1 Synthesis of ribonucleotides and deoxyribonucleotides is an essential part of cellular metabolism, as synthesis of RNA requires ribonucleotides and DNA replication is dependent on deoxyribonucleotides. The involvement of nucleotides in these central cellular pathways suggests that it is important for the cell to control the synthesis of nucleotides and to be able to maintain a steady supply of these essential precursors either by de novo biosynthesis or by uptake of precursors from the growth medium. In addition, the involvement of nucleotides in regulatory processes such as regulation of gene expression and modulation of kinetic properties of enzymes emphasizes the need for tight regulation of the level of nucleotides in the cell. Indeed, expression of genes responsible for the de novo biosynthesis of ribonucleotides in Gram-positive bacteria such as Lactococcus lactis and Bacillus subtilis are regulated by the availability of purines and pyrimidines. The pyrimidine biosynthetic genes are regulated by the RNA-binding regulatory protein PyrR that regulates gene expression by an attenuation mechanism through sensing of the UMP concentration in the cell [1–5]. However, PyrR is not involved in the regulation of expression of the pyrG gene encoding CTP synthase (EC 6.4.3.2) in L. lactis and B. subtilis,aspyrG expression is probably regulated by an attenuation mechanism responding to the CTP concentration in the cell [6,7]. The reaction catalyzed by CTPsynthase(UTP+glutamine+ATP fi CTP + glutamate + ADP + P i ) involves all four ribonucleo- tides; UTP and CTP are substrate and product, respect- ively, ATP is used as an energy source and GTP is an allosteric activator of the reaction [8]. CTP synthase has a central role in pyrimidine metabolism, as the enzyme catalyses the only reaction resulting in the amination of the pyrimidine ring into a cytosine derivative (Fig. 1). It is therefore of interest to examine to what extent this enzyme controls the fluxes and metabolite concentrations in the pathway. Here, we have determined the importance of CTP synthase for growth rate, the concentration of ribonucleotides and for the concentration of the deoxy- ribonucleotide dCTP using the methods developed for metabolic control analysis [9,10]. We show that CTP synthase has a strong inherent control on the CTP and dCTP concentrations and a negative control on the UTP concentration. Materials and methods Bacterial strains and plasmids The strains and plasmids used in this study are listed in Table 1. Plasmid pCJ31B contains the L. lactis pyrG gene, and was made from a PCR-product made with prim- ers pyrG11a (5¢-GTAGAAGCTAAAATCTGG-3¢)and SLLH7 (5¢-TACAAAAGATTTTGGGC-3¢) cloned in the TOPO TA cloning kit from Invitrogen. Chromosomal DNA purified from MG1363 was used as a template for the PCR amplification. Growth medium and growth conditions L. lactis strains were grown either in M17 broth supplied with 1% (w/v) glucose or in defined SA medium [11] with Correspondence to J. Martinussen, BioCentrum-DTU, Bacterial Physiology and Genetics, Technical University of Denmark, Building 301, DK-2800 Kgs. Lyngby, Denmark. Fax: + 45 45932809, Tel.: + 45 45252498, E-mail: jma@biocentrum.dtu.dk Enzyme: CTP synthase (EC 6.4.3.2). (Received 9 January 2004, revised 14 April 2004, accepted 16 April 2004) Eur. J. Biochem. 271, 2438–2445 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04168.x 1% (w/v) glucose at 30 °C in 50-mL plastic tubes without aeration. Erythromycin was added to a concentration of 2 lgÆmL )1 , chloramphenicol to 5 lgÆmL )1 . When needed, cytidine was added to 20 lgÆmL )1 in SA medium and to 1000 lgÆmL )1 in M17 medium. Transformation of DNA to L. lactis L. lactis cells were transformed by electroporation as described by [12]. Following transformation, the cells were incubated for 2 h in M17 medium with glucose and cytidine at 1000 lgÆmL )1 for phenotypic expression. Isolation of chromosomal DNA from L. lactis Chromosomal DNA was isolated as described in [13]. Isolation of strains with altered pyrG expression A PCR product was made with the primers pyrGCP2 (5¢-ACGCTCGAGATNNNNNAGTTTATTCTTGACA NNNNNNNNNNNNNNNNNTATAATNNNNCCTC TGGGGAGCTGTTTTTG-3¢) and pyrG13b (5¢-GCTGA ACTGCAGAACTCCTGAGTTAAGGAGAG-3¢)using pCJ31B as template and the Elongase enzyme mix (Life Technologies). The pyrGCP2 primer was designed in such a way that the 3¢ end would anneal to the template upstream of the pyrG open reading frame but after the terminator in the attenuator. The downstream primer is located before the pyrG terminator, but after the stop codon. The PCR product was digested with XhoIandPstI and ligated in plasmidpLB86alsodigestedwithXhoIandPstI. After ligation, the plasmids were transformed to CJ327 (a pyrG::ISS1, cdd strain with plasmid pLB65) and plated on Table 1. Strains and plasmids used in this study. Strain or plasmid Relevant description Reference or source L. lactis cells MG1363 L. lactis ssp. cremoris strain [33] MB109 MG1363 cdd [34] CJ233 MB109/pAK80 [6] CJ295 MB109 pyrG::ISS1 [6] CJ327 CJ295/pLB65 This study CJ340 CJ327 with pLB86 integrated in attB site This study CJ381 CJ327 with pyrG gene transcribed from synthetic promoter integrated in attB site This study CJ382 CJ327 with pyrG gene transcribed from synthetic promoter integrated in attB site This study CJ383 CJ327 with pyrG gene transcribed from synthetic promoter integrated in attB site This study CJ388 CJ327 with pyrG gene transcribed from synthetic promoter integrated in attB site This study CJ405 CJ327 with pyrG gene transcribed from synthetic promoter integrated in attB site This study CJ406 CJ327 with pyrG gene transcribed from synthetic promoter integrated in attB site This study CJ407 CJ327 with pyrG gene transcribed from synthetic promoter integrated in attB site This study CJ410 CJ327 with pyrG gene transcribed from synthetic promoter integrated in attB site This study CJ411 CJ327 with pyrG gene transcribed from synthetic promoter integrated in attB site This study CJ413 CJ327 with pyrG gene transcribed from synthetic promoter integrated in attB site This study CJ418 CJ327 with pyrG gene transcribed from synthetic promoter integrated in attB site This study CJ420 CJ327 with pyrG gene transcribed from synthetic promoter integrated in attB site This study Plasmids pCJ31B Contains pyrG gene and promoter in pCR2.1-TOPO vector This study pLB65 Expresses the bacteriophage TP901-1 integrase protein; carries chloramphenicol resistance gene [17] pLB86 Integration vector with promoter-less reporter genes lacLM, erythromycin resistance gene and attP site [17] pCR2.1-TOPO Plasmid vector Invitrogen pAK80 Contains promoterless lacLM genes [14] Fig. 1. Simplified representation of pyrimidine nucleotide metabolism in L. lactis. Only reactions relevant for this study are included in the figure. The central part of the figure shows the conversion of UTP to CTP catalyzed by CTP synthase encoded by pyrG. Involvement of pyrimidine nucleotides in synthesis of DNA, RNA, and phospholipids are indicated. Breakdown of mRNA is indicated by broken arrows. For more details on nucleotide metabolism in Gram-positive bacteria, see previously published review [32]. Gene symbols refer to the fol- lowing proteins: cdd, cytidine deaminase; nrdEF, aerobic ribonucleo- tide reductase; nrdDG, anaerobic ribonucleotide reductase. Ó FEBS 2004 pyrG controls [CTP] and [dCTP] in L. lactis (Eur. J. Biochem. 271) 2439 M17 plates with glucose, erythromycin, chloramphenicol, cytidine at 1000 lgÆmL )1 and 5-bromo-4-chloro-3-indolyl- b- D -galactoside (X-gal) at 90 lgÆmL )1 . Growth experiments The strains were grown over night at 30 °ConSAplates containing glucose, erythromycin and cytidine. Several colonies from the plates were inoculated in 1 mL SA medium with glucose, erythromycin and cytidine and grown for 5–6 h at 30 °C. This growing culture was diluted and used to inoculate 10 mL of SA medium with glucose, erythromycin and cytidine so that the next day the growth experiment could start with an exponentially growing culture (D 436 <0.8) 2 . In order to remove the cytidine present in the overnight culture, the cells were centrifuged for 15 min at 3000 g 3 , washed twice with 0.9% (w/v) NaCl and resuspended in 2 mL of 0.9% (w/v) NaCl. Fifty milliliters of SA medium with glucose and erythromycin was inoculated to D 436 of 0.025 and the growth monitored by measuring D 436 .AtD 436 of 0.8, 35 mL of the culture was harvested, washed once with 0.9% (w/v) NaCl and resus- pended in 1 mL of Z buffer. b-Galactosidase measurement b-Galactosidase activity in exponentially grown cells was determined at 30 °C as previously described [14], except that cell density was measured at 436 nm and the specific activity was therefore determined as A 420 /D 436 per minute per mL of culture (units/D 436 ). Nucleotide pool determinations The concentration of nucleotides in the cell was determined by thin layer chromatography on PEI-plates of 33 PO 43 - labeled nucleotides extracted from exponentially growing cells at a cell density of D 436 ¼ 0.8 as described previously [15]. Determination of concentration control coefficients The experimental data of reporter gene activity from the strains with altered pyrG expression and corresponding growth rate or concentrations of (deoxy)ribonucleo- tides were plotted and fitted to functions using the software program GRAFIF (Erithacus Software Ltd, Harley, Surrey, UK). Wild-type b-galactosidase expres- sion level was obtained by plotting the relative CTP, UTP, and dCTP concentrations against the specific b-galactosidase activity (z) in units/D 436 4 and fitting to the functions f(z) ¼ 2.49–2.49 · exp() 0.0237 · z 1.362 )for CTP, f(z) ¼ 1.55–0.268 · ln(z) for UTP, and f(z) ¼ 1.27– 1.27 · exp()0.131 · z) 5 for dCTP. The average of the b-galactosidase activities obtained for the three nucleo- tides at the wild-type pool level was taken as wild-type b-galactosidase expression level of 9.5 units/D 436 .The CTP concentration data were then fitted to the func- tion f(x) ¼ 3.81–3.81 · exp()0.00101 · x 1.352 ), where x is expression of the reporter genes relative to the wild-type level of 9.5 units/D 436 . When the CTP concentration data were fitted to the function f(x) ¼ 0.0068 · 0.998x · x 1.225 , a similar concentration control coefficient was obtained at the wild-type level. To determine control of pyrG expression on the UTP concentration, the data were fitted to two functions: f(x) ¼ 9.20–1.132 · ln(x) and f(x) ¼ 10.06 · 0.999x · x-0.191. Both functions gave sim- ilar concentration control coefficients at the wild-type level. The ATP and GTP concentrations were fitted to the linear functions f(x) ¼ 7.81–0.00342 · xandf(x)¼ 1.98–0.00176 · x. The change in dCTP concentration was found to fit the equation f(x) ¼ 0.614 · [1 – exp()0.0128 · x)]; fitting the data to a quadratic equa- tion [f(x) ¼ 0.198 + 0.00236 · x ) 0.00000266 · x2) did not affect the calculated concentration control coefficient at the wild-type level. Control coefficients for CTP synthase (PyrG) on the concentration of (deoxy)ribonu- cleotides ([NTP]) were calculated from the equation ¼ fd([NTP])/[NTP]g/fd(b-gal)/b-galg. The growth rate data were fitted to the function f(x) ¼ 0.650 · [1 – exp() 0.2197 · x)], although fitting the data to the function f(x) ¼ 0.664 · 0.1061/x gave almost identical control coefficients except at very low x values. Results Isolation of strains with altered pyrG expression Strains with different constitutive expression levels of the 6 pyrG gene encoding CTP synthase were isolated using a PCR strategy where the pyrG gene under control of synthetic constitutive promoters was integrated on the chromosome of L. lactis [16]. A degenerate primer con- taining a promoter with consensus )10 and )35 promoter regions separated by a randomized spacer of 17 nucleo- tidesplusa3¢-end with homology to the upstream sequence from pyrG and a primer with homology to a region downstream of pyrG were used to generate PCR products covering the entire pyrG gene downstream of synthetic promoters. Expression from pyrG in the wild type is regulated by the concentration of CTP in the cell by an attenuation mechanism in the 5¢-end of the pyrG mRNA [6]. The primer was constructed so the PCR product did not contain these regulatory signals to obtain constitutive expression. The downstream primer was designed in such a way that, after the pyrG open-reading frame, the terminator could not be located on the PCR product. The obtained DNA fragment was inserted in the integration vector pLB86, which contains an erythromycin resistance marker, the lacLM reporter genes as well as the attachment site attP from the bacteriophage TP901-1. In the presence of the TP901-1 integrase, pLB86 will insert with high frequency at the attB site on the chromosome; the promoter library was therefore transformed to a strain carrying plasmid pLB65 expressing the integrase [17]. The promoter fusions were integrated on the chromosome of the CTP synthase deficient strain CJ295 (pyrG::ISS1), which also carries a mutation in the cdd gene encoding cytidine deaminase (Fig. 1) in order to prevent degrada- tion of cytidine added to the growth medium. The transformants were selected on rich media with a cytidine concentration of 1000 lgÆmL )1 even though we have previously isolated pyrG mutants on defined media with only20or50lg of cytidine per milliliter [6,8]. However, 2440 C. M. Jørgensen et al. (Eur. J. Biochem. 271) Ó FEBS 2004 L. lactis pyrG mutants do not grow on M17 media with cytidine at such low concentrations. Apparently, com- pounds present in M17 medium inhibits cytidine metabo- lism, most likely by inhibiting the uptake of cytidine [15]. Adding cytidine in excess at 1000 lgÆmL )1 relieves the inhibitory effect and allows for growth of L. lactis pyrG mutants on rich medium. More than 1500 colonies with potentially different pyrG expression levels were isolated. Forty strains with colors of colonies ranging from white to blue on X-gal indicator plates were purified, and after an initial screening of b-galactosidase activity, 12 strains were selected for further analysis. The pyrG gene from these strains was amplified by PCR and sequenced in order to exclude strains having mutations in the pyrG gene. Estimation of pyrG expression from b-galactosidase activity The 12 selected strains were grown in defined medium without cytidine and harvested during exponential growth for determination of the level of pyrG expression. Several attempts to measure CTP synthase activity in L. lactis crude extracts using an assay based on the conversion of UTP to N 4 -hydroxy-CTP in the presence of hydroxyl- amine [18] was unfortunately unsuccessful. This is in accordance with earlier results from B. subtilis, where no CTP synthase activity could be detected in crude extracts [7]. The estimation of pyrG expression in the constructed strains was therefore based on reporter gene activity, as b-galactosidase is expressed in an operon with pyrG.The specific b-galactosidase activity in the selected group of strains with altered pyrG expression was determined and varied over a 20-fold range from 0.3 to 63 units/D 436 .The nucleotide pool sizes in the 12 strains were determined, and it was found that the CTP, UTP and dCTP pools were affected by the pyrG expression level. Figure 2 shows the correlation between the nucleotide concentrations and the specific b-galactosidase activity in the range from 4 to 25 attenuance units. The nucleotide pool sizes are shown relative to the wild-type concentrations from an isogenic strain carrying a wild-type pyrG gene (CJ233). From the wild-type concentrations of the three nucleotides in Fig. 2, it can therefore be estimated that a specific activity of b-galactosidase of 9.5 units/attenuance results in the same nucleotide concentrations found in a wild-type strain. A specific b-galactosidase activity of 9.5 units/D 436 was therefore taken as the reference level where the CTP synthase activity is the same as in a wild-type cell. CTP synthase has no control on the growth rate of L. lactis The strains with altered expression of pyrG were grown in defined medium without cytidine and the specific growth rate was determined. The b-galactosidase activity in each strain was calculated relative to 9.5 units/D 436 ,which reflects the wild-type CTP synthase level. The strains then have b-galactosidase activities ranging from 3% of the wild-type level to 6.6-fold increased expression. Figure 3 shows the specific growth rate of the strains with altered expression of pyrG plotted against the relative b-galac- tosidase activity. No change in growth rate was observed around the wild-type level and only when the b-galactosi- dase activity dropped below 40% of the wild-type level, the growth rate was affected. From the curve fit in Fig. 3, control coefficients for pyrG expression on the specific growth rate for all lacLM levels were determined. The level of CTP synthase had zero control on the growth rate in the range from 40 to 600% relative b-galactosidase expression; the control increased to more than 0.5 at expression levels below 5% of the wild-type level, reflect- ing that CTP synthase is an essential enzyme for growth of L. lactis. (Deoxy)ribonucleotide pool sizes in mutants with different levels of pyrG expression Figure 4A shows the variation of the CTP concentration as a function of pyrG expression measured as b-galactosidase activity from the lacLM reporter genes. There is a clear correlation between expression of lacLM and the CTP pool Fig. 2. Correlation of specific b-galactosidase activity and concentra- tions of the nucleotides CTP, dCTP, and UTP. Concentrations of CTP, UTP and dCTP relative to the wild type are plotted against the specific b-galactosidase activity in strains with modulated pyrG expression. The horizontal stippled line indicates the wild-type nucleotide con- centration, which has been set to one for all three nucleotides. Fig. 3. Specific growth rate and control coefficient for CTP synthase. The specific growth rate is shown as function of b-galactosidase activity from mutants with altered pyrG expression given as percent relative to the wild-type activity. The curve fitted to the experimental data is shown as a thin line; the thick black line indicates the calculated control coefficient. Ó FEBS 2004 pyrG controls [CTP] and [dCTP] in L. lactis (Eur. J. Biochem. 271) 2441 size. Increased reporter gene expression to 660% of the wild- type level results in an increase in the CTP concentration of almost 2.5-fold; decreased expression results in a sevenfold decrease in the CTP pool size. With respect to the UTP pool size, the correlation appeared to be inverse (Fig. 4B), with UTP decreased more than twofold at increased CTP synthase activity and increased twofold at low activity. No significant correlation was observed for the purine nucleo- tides ATP and GTP, although there is a tendency for the concentration of these nucleotides to decrease with increas- ing pyrG expression (Fig. 4C,D). The variation of the concentration of the deoxyribonucleotide dCTP with respect to altered pyrG expression is shown in Fig. 5. The pattern of changes in dCTP pool size resembles the one for CTP. No significant changes in the concentrations of the deoxyribonucleotides dATP, dGTP and dTTP were observed with different levels of pyrG expression (data not shown). CTP synthase has a positive control on the CTP and dCTP concentrations and negative control on the UTP concentration The primary data on Figs 4 and 5 already indicates that the level of CTP synthase is important for the pool sizes of CTP, UTP and dCTP. In order to calculate how important the Fig. 5. Effects of changing pyrG expression on the concentration of dCTP. The concentration of dCTP is shown as a function of relative b-galactosidase activity in percent of the wild-type level. The concen- tration of the deoxyribonucleotide is given in nanomoles per mg (dry weight). The experimental data are fitted to the thin black line; the calculated concentration control coefficient is shown as a thick black line. The dCTP pool size in the reference strain CJ233 is shown by a stippled line. Fig. 4. Effects of changing pyr G expression on the concentrations of ribonucleotides. The concentrations of CTP, UTP, GTP and ATP (A, B, C, and D) are shown as functions of relative b-galactosidase activity in percent of the wild-type level. The concentrations of the ribo- nucleotides are given in nmolÆmg )1 (dry weight). The experimental data are fitted to the thin black lines; the calculated concentration control coefficients for CTP and UTP are shown as thick black lines. Ribonucleotide pool sizes in the reference strain CJ233 are shown by horizontal stippled lines. 2442 C. M. Jørgensen et al. (Eur. J. Biochem. 271) Ó FEBS 2004 enzyme activity is, we used metabolic control analysis to quantify the effect of a sustained modulation of CTP synthase, i.e. in terms of so-called inherent control coeffi- cients [19,20]. The curve fits shown in Fig. 4A,B were used for calculation of the concentration control coefficients for CTP synthase on the concentrations of CTP and UTP in L. lactis. CTP synthase (PyrG) has a high inherent control on the CTP concentration, as at the wild-type level the calculated concentration control coefficient C ½CTP PyrG is 1.03 (Fig. 4A). The control coefficient increased to 1.35 at very low expression of the pyrG gene and decreased to zero at high expression levels. The concentration control coefficient on the UTP concentration was calculated from two different curve fits, with almost identical results. The control coefficients in Fig. 4B is from the equation [UTP](x) ¼ 9.20–1.13 · ln(x). The concentration control coefficient decreased from )0.11 at very low expression of pyrG to )0.28 at the wild-type level and to )0.63 at high expression levels. CTP synthase was found to have control on not only the CTP concentration but also the dCTP concentration. At the wild-type level, the concentration control coefficient was calculated to ¼ 0.49 (Fig. 5). At low expression levels, this value increased to one and decreased to zero at high expression levels. Strains with decreased CTP and dCTP concentrations have reduced growth at 15 °C In Escherichia coli, deletion of the cmk gene encoding cytidine monophosphate kinase, results in reduced CTP and dCTP pool sizes, probably due to impaired reutiliza- tion of nucleotides generated from, e.g. mRNA turnover [21], and an interesting phenotype of the E. coli cmk mutant is an inability to grow at low temperatures. As some of the isolated L. lactis strains with altered pyrG expression have very low CTP and dCTP pool sizes, they were tested for growth on solid medium at 15 °Cand 30 °C. Table 2 compares the growth of seven strains with reduced pyrG expression to the growth of a pyrG::ISS1 mutant with a vector inserted in the attB site (CJ340) as well as to the growth of a pyrG + strain (CJ233). The three strains with the lowest pyrG expression level, and thus the lowest CTP and dCTP concentrations in the cell, grew significantly slower at 15 °Cthanat30°C compared to strains with normal pyrG expression. One of these strains has 43% pyrG expression relative to the wild-type andnogrowthdefectat30°C. However, the two strains with the lowest pyrG expression levels at 3–4% of the wild-type level (CJ381 and CJ388) also showed reduced growth rate at 30 °C compared to the wild-type, but these strains are growth impaired at 15 °C, as they do not grow, even after 8 days of incubation at 15 °C. Addition of cytidine to the growth medium restores growth, suggesting that the reduced growth rate is indeed related to the decreased CTP and dCTP pool sizes. Discussion In this work, we have modulated the expression of the L. lactis pyrG gene encoding CTP synthase. To our knowledge, this is the first time an enzyme in nucleotide metabolism has been the subject of metabolic control analysis. We have established that the level of CTP synthase has no control on the growth rate at the wild-type level in L. lactis. At highly reduced pyrG expression, CTP synthase controls the growth rate, thus confirming that CTP synthase is an essential enzyme for growth of L. lactis in the absence of cytidine. CTP synthase has a high positive control on the concentrations of CTP and dCTP and a negative control on the UTP concentration. dCTP is synthesized from CTP by a reaction catalyzed by ribonucleotide reductase, which converts the ribose part to 2¢-deoxyribose. In L. lactis,two ribonucleotide reductases have been identified: NrdDG, required for strict anaerobic growth and NrdEF that does not function in the absence of oxygen [22,23] (Fig. 1). The two lactococcal ribonucleotide reductases have different substrate specificity as NrdDG is an NTP reductase and NrdEF is an NDP reductase. Although the data in Figs 4A and 5 show that the dCTP concentration varies in a similar way as the CTP concentration when pyrG expression is altered, it is not possible from the available data to conclude whether reduction occurs at the tri- or diphosphate level in our experiments. Table 2. Growth of strains with modulated pyrG expression at 15 °C and 30 °C. The relative pyrG expression is given as per cent of wild-type b-galactosidase activity. The activity in the pyrG strain CJ340 is defined as 0%; activity in the pyrG + strain CJ233 is defined as 100%. Growth was on solid defined medium with erythromycin at 2 lgÆmL )1 and 1% glucose as carbon source in the presence or absence of cytidine (20 lgÆmL )1 ). 3/4, no growth; +++, good growth after 4 days at 15 °C or two days at 30 °C. Strain Genotype Relative pyrG expression (%) Growth at 15 °C Growth at 30 °C None Cytidine None Cytidine CJ406 cdd pyrG::ISS1 attB::pyrG + 99 +++ +++ +++ +++ CJ418 cdd pyrG::ISS1 attB::pyrG + 89 +++ +++ +++ +++ CJ410 cdd pyrG::ISS1 attB::pyrG + 88 +++ +++ +++ +++ CJ405 cdd pyrG::ISS1 attB::pyrG + 77 +++ +++ +++ +++ CJ413 cdd pyrG::ISS1 attB::pyrG + 43 + +++ +++ +++ CJ381 cdd pyrG::ISS1 attB::pyrG + 4.3 3/4 +++ + +++ CJ388 cdd pyrG::ISS1 attB::pyrG + 3.3 3/4 +++ + +++ CJ340 cdd pyrG::ISS1 attB::pLB86 0 3/4 +++ 3/4 +++ CJ233 cdd 100 +++ +++ +++ +++ Ó FEBS 2004 pyrG controls [CTP] and [dCTP] in L. lactis (Eur. J. Biochem. 271) 2443 The finding that CTP synthase has no control on the growth rate but a strong control on the CTP and dCTP concentrations is perfectly in line with the theory of metabolic supply and demand analysis [24]. This theory predicts that for biosynthetic pathways, such as those leading to the biosynthesis of amino acids or nucleotides, the control of the flux should reside in the demand for the end-product, whereas the supply determines the degree of concentration control. Recently it was found that DNA supercoiling in E. coli is under tight homeostatic control with 87% of imposed changes being counteracted by homeostatic mechanisms [19,20]. The homeostasis was found to take place at the metabolic and genetic level with 72 and 28%, respectively. Here it is important to remember that in the analysis of CTP synthase, the feedback mechanism that may have acted on the level of pyrG expression has been removed. The strength of the feedback regulation of CTP on pyrG expression, i.e. the elasticity of pyrG expression for the CTP concentration [25], has not been quantified, but the regulation appears to be quite strong: when a pyrG mutant was starved for cytidine the CTP pool dropped more than 10-fold and at the same time the expression of a reporter gene fusion to the pyrG promoter increased 37-fold [6]. Therefore, the control exerted by CTP synthase in a ÔnormalÕ cell, i.e. where the regulation of pyrG expression is operative, could very well differ significantly from the control measured in the current study, where the elasticity of pyrG expression is zero. The control coefficients we have obtained by modulating the level of transcription are called inherent control coeffi- cients [19,20]. In the experiments, product inhibition by CTP on CTP synthase was intact, and this feedback inhibition [8] was shown to be unable to fully counteract an increase in the CTP concentration in strains with increased pyrG expression of up to 250% of the wild-type level (Fig. 4A). The results show that the feedback inhibition of the CTP synthase enzyme is incomplete in vivo. In conclusion, the homeostasis of the CTP pool in the wild-type cell is primarily a matter of regulation of pyrG expression exerted by the attenuator found immedi- ately in front of the pyrG open reading frame. Strains with pyrG expression from 43% to 665% of the wild-type level have growth rates at the wild-type level at 30 °C, implying that the need for CTP in these strains is similar to the wild-type. This suggests that for strains with pyrG expression of 43%, the average flow of substrate through each CTP synthase enzyme is increased approxi- mately 2.5-fold and that the average in vivo activity of CTP synthase is correspondingly increased. Increased in vivo activity of CTP synthase may be due to the reduced CTP concentration, as the L. lactis CTP synthase enzyme is feedback inhibited by CTP [8], as well as due to an increase in the substrate concentration. It was not possible to detect CTP synthase activity in L. lactis cell extracts, and determination of pyrG expression in the constructed strains with altered pyrG expression was therefore dependent on b-galactosidase activity measure- ments. As a clear correlation was observed between nucleotide pool sizes and b-galactosidase activity, the wild-type b-galactosidase activity was established using the in vivo concentrations of nucleotides (Fig. 2). This deter- mination is important, as calculations of control coefficients are dependent on knowledge of enzyme activities. However, even changing the estimate of the wild-type b-galactosidase activity with 25% does not result in significant changes in the concentration control coefficients. A variation of ± 25% in the wild-type b-galactosidase activity results in changes in the concentration control coefficients for CTP and UTP with less than 8% and less than 18% for dCTP (data not shown). A strain with pyrG expression reduced to 43% of the wild-type level has decreased CTP and dCTP pool sizes and show reduced growth at 15 °C, whereas no effect on growth was observed at 30 °C (Table 2). The growth defect is relieved by cytidine addition, suggesting that the observed slow growth at 15 °C is a result of reduced CTP and dCTP pool sizes. Bacteria grown at low temperatures have several metabolic problems compared to growth at the optimal growth temperature, including reduced enzyme activities, low membrane fluidity, and decreased initiation of transla- tion [26,27]. These factors may all be related to the impaired growth at 15 °C of strains with decreased CTP and dCTP pool sizes at 30 °C. Reduced synthesis or activity of CTP synthase at 15 °C may affect the growth of cells with low expression of pyrG, whereas cells with expression levels close to the wild-type have excess CTP synthase capacity and are thus not affected by a decrease in CTP synthase activity. To maintain membrane structure and function, B. subtilis change the fatty acid composition in the membrane during cold shock [28,29]. Altered CTP and dCTP pool sizes may result in perturbations of membrane synthesis, as both nucleotides are used in the biosynthesis of phospholipids through the synthesis of CDP-diacylglycerol from phos- phatidic acid [30]. Dramatically changed CTP and dCTP pool sizes in L. lactis may therefore inhibit the adaptation of thecellmembranetogrowthat15°C resulting in slow growth at this temperature. The importance of CTP synthase in E. coli phospholipid biosynthesis was investi- gatedinapyrG mutant starved for CTP by resuspending cells in medium lacking cytidine thereby reducing the CTP pool size. This results in accumulation of phosphatidic acid, the substrate for CDP-diacylglycerol synthetase [31] and increased resistance to the antibiotic erythromycin. How- ever, the mechanisms for CTP and dCTP involvement in lipid biosynthesis and adaptation to growth at low temper- atures remain unclear. Acknowledgements We thank Martin Willemoe ¨ s for many helpful discussions. The FØTEK Program and the DFFE supported this work through the Centre for Advanced Food Studies. References 1. Turner, R.J., Lu, Y. & Switzer, R.L. (1994) Regulation of the Bacillus subtilis pyrimidine biosynthetic (pyr) gene cluster by an autogenous transcriptional attenuation mechanism. J. Bacteriol. 176, 3708–3722. 2. Bonner, E.R., D’Elia, J.N., Billips, B.K. & Switzer, R.L. (2001) Molecular recognition of pyr mRNA by the Bacillus subtilis attenuation regulatory protein PyrR. Nucleic Acids Res. 29, 4851–4865. 2444 C. M. Jørgensen et al. (Eur. J. Biochem. 271) Ó FEBS 2004 3. Andersen, P.S., Martinussen, J. & Hammer, K. (1996) Sequence analysis and identification of the pyrKDbF operon from Lacto- coccus lactis including a novel gene, pyrK, involved in pyrimidine biosynthesis. J. Bacteriol. 178, 5005–5012. 4. Martinussen, J. & Hammer, K. (1998) The carB gene encoding the large subunit of carbamoylphosphate synthetase from Lacto- coccus lactis is transcribed monocistronically. J. Bacteriol. 180, 4380–4386. 5. Martinussen, J., Schallert, J., Andersen, B. & Hammer, K. (2001) The pyrimidine operon pyrRPB-carA from Lactococcus lactis. J. Bacteriol. 183, 2785–2794. 6. Jørgensen, C.M., Hammer, K. & Martinussen, J. (2003) CTP limitation increases expression of CTP synthase in Lactococcus lactis. J. Bacteriol. 185, 6562–6574. 7. Meng, Q. & Switzer, R.L. (2001) Regulation of transcription of the Bacillus subtilis pyrG gene, encoding cytidine triphosphate synthetase. J. Bacteriol. 183, 5513–5522. 8. Wadskov-Hansen, S.L., Willemoe ¨ s, M., Martinussen, J., Hammer, K., Neuhard, J. & Larsen, S. (2001) Cloning and verification of the Lactococcus lactis pyrG gene and characteri- zation of the gene product, CTP synthase. J. Biol. Chem. 276, 38002–38009. 9. Heinrich, R., Rapoport, S.M. & Rapoport, T.A. (1977) Metabolic regulation and mathematical models. Prog. Biophys. Mol. Biol. 32, 1–82. 10. Kacser, H. & Burns, J.A. (1973) The control of flux. In Rate Control of Biological Processes (Davies, D.D., ed.), pp. 65–104. Cambridge University Press, London. 11.Jensen,P.R.&Hammer,K.(1993)Minimalrequirementsfor exponential growth of Lactococcus lactis. Appl. Environ. Micro- biol. 59, 4363–4366. 12. Holo, H. & Nes, I.F. (1989) High-frequency transformation, by electroporation, of Lactococcus lactis subsp. cremoris grown with glycine in osmotically stabilized media. Appl. Environ. Microbiol. 55, 3119–3123. 13. Johansen, E. & Kibenich, A. (1992) Characterization of Leuco- nostoc isolates from commercial mixed strain mesophilic starter cultures. J. Dairy Sci. 75, 1186–1191. 14. Israelsen, H., Madsen, S.M., Vrang, A., Hansen, E.B. & Johansen, E. (1995) Cloning and partial characterization of regulated pro- moters from Lactococcus lactis Tn917-lacZ. integrants with the new promoter probe vector, pAK80. Appl. Environ. Microbiol. 61, 2540–2547. 15. Martinussen, J., Wadskov-Hansen, S.L. & Hammer, K. (2003) Two nucleoside uptake systems in Lactococcus lactis: competition between purine nucleosides and cytidine allows for modulation of intracellular nucleotide pools. J. Bacteriol. 185, 1503–1508. 16. Solem, C. & Jensen, P.R. (2002) Modulation of gene expression made easy. Appl. Environ. Microbiol. 68, 2397–2403. 17. Brøndsted, L. & Hammer, K. (1999) Use of the integration ele- ments encoded by the temperate lactococcal bacteriophage TP901- 1 to obtain chromosomal single-copy transcriptional fusions in Lactococcus lactis. Appl. Environ. Microbiol. 65, 752–758. 18. Willemoe ¨ s, M. & Larsen, S. (2003) Substrate inhibition of Lac- tococcus lactis cytidine 5¢-triphosphate synthase by ammonium chloride is enhanced by salt-dependent tetramer dissociation. Arch. Biochem. Biophys. 413, 17–22. 19. Jensen, P.R., van der Weijden, C.C., Jensen, L.B., Westerhoff, H.V. & Snoep, J.L. (1999) Extensive regulation compromises the extent to which DNA gyrase controls DNA supercoiling and growth rate of Escherichia coli. Eur. J. Biochem. 266, 865–877. 20. Snoep, J.L., van der Weijden, C.C., Andersen, H.W., Westerhoff, H.V. & Jensen, P.R. (2002) DNA supercoiling in Escherichia coli is under tight and subtle homeostatic control, involving gene- expression and metabolic regulation of both topoisomerase I and DNA gyrase. Eur. J. Biochem. 269, 1662–1669. 21. Fricke,J.,Neuhard,J.,Kelln,R.A.&Pedersen,S.(1995)Thecmk gene encoding cytidine monophosphate kinase is located in the rpsA operon and is required for normal replication rate in Escherichia coli. J. Bacteriol. 177, 517–523. 22. Jordan, A., Pontis, E., Aslund, F., Hellman, U., Gibert, I. & Reichard, P. (1996) The ribonucleotide reductase system of Lac- tococcus lactis. Characterization of an NrdEF enzyme and a new electron transport protein. J. Biol. Chem. 271, 8779–8785. 23. Torrents, E., Buist, G., Liu, A., Eliasson, R., Kok, J., Gibert, I., Graslund, A. & Reichard, P. (2000) The anaerobic (class III) ribonucleotide reductase from Lactococcus lactis: catalytic prop- erties and allosteric regulation of the pure enzyme system. J. Biol. Chem. 275, 2463–2471. 24. Hofmeyr, J.S. & Cornish-Bowden, A. (2000) Regulating the cel- lular economy of supply and demand. FEBS Lett. 476, 47–51. 25. Burns, J.A., Cornish-Bowden, A., Groen, A.K., Heinrich, R., Kacser, H., Porteous, J.W., Rapoport, S.M., Rapoport, T.A., Stucki, J., Tager, J.M., Wanders, R.J.A. & Westerhoff, H.V. (1985) Control analysis of metabolic systems. Trends Biochem. Sci. 10,16. 26. Thieringer, H.A., Jones, P.G. & Inouye, M. (1998) Cold shock adaptation. Bioessays 20, 49–57. 27. Weber, M.H. & Marahiel, M.A. (2002) Coping with the cold: the cold shock response in the Gram-positive soil bacterium Bacillus subtilis. Phil. Trans. R. Soc. Lond. B 357, 895–907. 28. Aguilar, P.S., Cronan, J.E. & de Mendoza, D. 8 (1998) A Bacillus subtilis gene induced by cold shock encodes a membrane phos- pholipid desaturase. J. Bacteriol. 180, 2194–2200. 29. Klein,W.,Weber,M.H.W.&Marahiel,M.A.(1999)Coldshock response of Bacillus subtilis: isoleucine-dependent switch in the fatty acid branching pattern for membrane adaptation to low temperature. J. Bacteriol. 181, 5341–5349. 30. Cronan, J.E. & Rock, C.O. (1996) Biosynthesis of membrane lipids. In Escherichia coli and Salmonella Typhimurium: Cellular and Molecular Biology (Neidhardt, F.C., ed.), pp. 612–636. ASM Press, Washington DC. 31. Ganong, B.R. & Raetz, C.R. (1982) Massive accumulation of phosphatidic acid in conditionally lethal CDP-diglyceride syn- thetase mutants and cytidine auxotrophs of Escherichia coli. J. Biol. Chem. 257, 389–394. 32. Switzer, R.L., Zalkin, H. & Saxild, H.H. (2002) Purine, pyri- midine, and pyridine nucleotide metabolism. In Bacillus subtilis and its Closest Relatives: from Genes to Cells (Sonnenshein, A.L., Hoch,J.A.&Losick,R.,eds),pp.255–269.ASMPress,Wash- ington DC. 33. Gasson, M.J. (1983) Plasmid complements of Streptococcus lactis NCDO 712 and other lactic streptococci after protoplast-induced curing. J. Bacteriol. 154, 1–9. 34. Martinussen, J. & Hammer, K. (1995) Powerful methods to establish chromosomal markers in Lactococcus lactis:ananalysis of pyrimidine salvage pathway mutants obtained by positive selections. Microbiology 141, 1883–1890. Ó FEBS 2004 pyrG controls [CTP] and [dCTP] in L. lactis (Eur. J. Biochem. 271) 2445 . Expression of the pyrG gene determines the pool sizes of CTP and dCTP in Lactococcus lactis Casper M. Jørgensen, Karin Hammer, Peter R. Jensen and. the pool sizes of CTP, UTP and dCTP. In order to calculate how important the Fig. 5. Effects of changing pyrG expression on the concentration of dCTP. The