Regulation of purine biosynthesis by a eukaryotic-type kinase in Streptococcus agalactiae

Summary

Group B streptococci (GBS) are the principal causal agents of human neonatal pneumonia, sepsis and meningitis. We had previously described the existence of a eukaryotic-type serine/threonine kinase (Stk1) and phosphatase (Stp1) in GBS that regulate growth and virulence of the pathogen. Our previous results also demonstrated that these enzymes reversibly phosphorylated an inorganic pyrophosphatase. To understand the role of these eukaryotic-type enzymes on growth of GBS, we assessed the stk1-mutants for auxotrophic requirements. In this report, we describe that in the absence of the kinase (Stk1), GBS are attenuated for de novo purine biosynthesis and are consequently growth arrested. During growth in media lacking purines, the intracellular G nucleotide pools (GTP, GDP and GMP) are significantly reduced in the Stk1-deficient strains, while levels of A nucleotides (ATP, ADP and AMP) are marginally increased when compared with the isogenic wild-type strain. We provide evidence that the reduced pools of G nucleotides result from altered activity of the IMP utilizing enzymes, adenylosuccinate synthetase (PurA) and IMP dehydrogenase (GuaB) in these strains. We also demonstrate that Stk1 and Stp1 reversibly phosphorylate and consequently regulate PurA activity in GBS. Collectively, these data indicate the novel role of eukaryotic-type kinases in regulation of metabolic processes such as purine biosynthesis.

Introduction

In both prokaryotes and eukaryotes, regulation of essential cellular processes such as cell growth, metabolism, development and differentiation is accomplished by signal transduction systems. Signal transduction through phosphorylation of intracellular proteins is a crucial mechanism by which signals are transmitted within the cell. Serine/threonine and tyrosine kinases and phosphatases control reversible phosphorylation of target proteins in eukaryotes and are essential for cell cycle control and differentiation (Hanks et al., 1988; Hunter, 1995). A number of reports have described the existence of eukaryotic-type serine/threonine kinases (STK) and phosphatases (STP) in bacteria as well as other prokaryotes (for review, see Bakal and Davies, 2000). Prokaryotic STK have been shown to be necessary for regulation of various cellular functions such as growth, differentiation, pathogenicity and secondary metabolism (for review, see Zhang, 1996). STK that regulate virulence of bacterial pathogens have been described in streptococci (Rajagopal et al., 2003; Echenique et al., 2004), Yersinia pseudotuberculosis (Galyov et al., 1993) and Pseudomonas aeruginosa (Wang et al., 1998).

Research on the identification of the physiological substrates of the prokaryotic family of STK and STP has been steadily gaining interest. 6-Phosphofructokinase (PFK) is phosphorylated by an STK (Pkn4) in Myxococcus xanthus. Kinase-dependent phosphorylation increased PFK activity by 2.7-fold; this increase in activity is important for regulation of glycogen metabolism in M. xanthus (Nariya and Inouye, 2002; 2003). An STK (PknG) was described to regulate glutamine and glutamate levels in Mycobacterium tuberculosis; however, the physiological substrate for PknG is yet to be identified (Cowley et al., 2004). Elongation factor (EF-G) is reversibly phosphorylated by an STK (PrkC) and an STP (PrpC) in Bacillus subtilis (Gaidenko et al., 2002). The role of kinase-dependent phosphorylation of EF-G is not understood. YopO, an STK from yersinia, was shown to be secreted by the type III secretion apparatus and phosphorylate and consequently depolymerize host actin (Juris et al., 2000). Despite recent progress in the field, the role of the prokaryotic STK and STP particularly in the regulation of growth and virulence remains largely uncharacterized.

Group B streptococci or Streptococcus agalactiae (GBS) are β-haemolytic, Gram-positive cocci that are the principal causal agents of invasive infections in human neonates and immunocompromised adults (Baker and Edwards, 2001; Doran and Nizet, 2004). We have previously described the existence of a eukaryotic-type serine/threonine kinase (Stk1) and phosphatase (Stp1) in GBS (Rajagopal et al., 2003). In contrast to certain bacterial sp. such as M. xanthus, B. subtilis, M. tuberculosis (Pahor, 1998; Inouye et al., 2000; Kennelly, 2002; Cowley et al., 2004) that harbour multiple STK and STP homologues on their genome, GBS and other Gram-positive pathogens [e.g. Streptococcus pneumoniae (Tettelin et al., 2001), Streptococcus pyogenes (Nakagawa et al., 2003), Staphylococcus aureus (Ohta, 2001), Listeria monocytogenes (Loessner et al., 2000)] contain a single pair of genes that encode an STP and an STK respectively. We demonstrated that mutations in the GBS homologues affect growth, cell segregation and virulence. In addition, we showed that these enzymes reversibly phosphorylate an inorganic pyrophosphatase in GBS (Rajagopal et al., 2003). As inorganic pyrophosphatases (Ppases) have been implied to affect general metabolism (Lahti, 1983), we anticipated that the GBS Stk1-deficient strains may have metabolic defects due to altered pyrophosphatase activity.

The role of basic metabolic pathways on GBS infection and disease is poorly understood. GBS are auxotrophic for several amino acids indicating that the pathogen has to acquire these nutrients from the host environment (Willett and Morse, 1966; Shelver et al., 2003; Samen et al., 2004). Genome analysis of GBS has revealed the existence of a number of extracellular proteases, oligopeptide permeases and peptidases (Glaser et al., 2002; Tettelin et al., 2002), which may enable the bacteria to satisfy its amino acid requirement during growth in vivo (within the host) and in vitro (laboratory media) (Shelver et al., 2003; Samen et al., 2004). Previous studies from our laboratory demonstrated that mutations in mtaR, a regulator of methionine transport in GBS, dramatically decreased virulence (Shelver et al., 2003). Also, other GBS auxotrophs have been identified in screens of mutants that are attenuated for virulence (Jones et al., 2000). Despite the identification of auxotrophs in virulence screens, the primary focus of GBS research has been to understand the role of pathogenic factors. To elucidate the role of the two signalling enzymes, i.e. Stk1 and Stp1 on the growth of GBS, we analysed these mutants for auxotrophic phenotypes.

In this report, we describe a purine auxotrophic requirement of the Stk1-deficient strains of GBS. Inspection of the GBS genome (Tettelin et al., 2002) reveals the presence of all genes that encode enzymes essential for de novo purine and pyrimidine biosynthesis. In the absence of exogenous purines, the GBS Stk1-deficient strains are severely attenuated for growth. The intracellular nucleotide pools in these strains showed a substantial reduction in the level of G (GMP, GDP, GTP) nucleotides. We demonstrate that kinase-dependent phosphorylation decreases PurA activity and is important for the regulation of de novo purine biosynthesis in GBS.

Results

Identification of auxotrophic requirements

Our earlier results indicated that the stk1 and stp1 genes in GBS encoded a eukaryotic-type STK and STP respectively (Rajagopal et al., 2003). The operon encoding Stp1 and Stk1 is shown in Fig. 1. Sequence analysis revealed that the ATG translation start codon for stk1 has a 1 bp overlap with the TAA translation termination codon of stp1; consistent with this observation, we have previously demonstrated that both stp1 and stk1 genes were co-transcribed (Rajagopal et al., 2003). We also constructed mutants of GBS that were deficient in Stk1 (LR113, Stk1⁻) and both Stp1 and Stk1 (LR114, Stp1⁻Stk1⁻). Both these mutants had similar phenotypes (referred as Stk1-deficient strains in this report, see Experimental procedures and Table 1) and displayed altered cell growth, segregation and virulence of GBS. Furthermore, we demonstrated that these enzymes reversibly phosphorylate an inorganic pyrophosphatase in GBS (Rajagopal et al., 2003). Inorganic pyrophosphatases (Ppases) are essential for the regulation of various metabolic processes such as carbohydrate metabolism, amino acid and nucleotide biosynthesis (Lahti, 1983). Our preliminary results indicated that Stk1-dependent phosphorylation increased Ppase activity (Rajagopal et al., 2004). Therefore, we hypothesized that if kinase-dependent phosphorylation regulates pyrophosphatase activity in GBS, then the Stk1-deficient strains would be affected in some or all of the metabolic processes mentioned above and that these might manifest as auxotrophic requirements.

Fig. 1.

Physical map of the serine/threonine kinase and phosphatase operon in GBS. Thick arrows denote ORFs. Promoter and terminator regions are represented as ‘P’ and ‘T’ respectively. The genes priA, fmt, sunL, stp1 and stk1 encode primosomal protein N′ (796 aa), methionyl tRNA formyltransferase (311 aa), RNA methyltransferase (426 aa), serine/threonine phosphatase (245 aa) and serine/threonine kinase (636 aa) respectively. The gene ‘?’ is an ORF(166 aa) with no homology to known ORFs in the database. The location of the kanamycin insertion cassette (Ωkm-2) at aa13 of stk1 in LR113 and LR114 is shown. The allelic exchange replacement of stp1 with chloramphenicol acetyltransferase (cat) in LR114 is also indicated.

Table 1.

Bacterial strains, plasmids and PCR primers

Strain or plasmid	Genotype\phenotype	Reference
E. coli
MC1061	F′araD139 Δ(ara-leu)7696 Δ(lac)X74 galU galK hadR2 (rk− mk+) mcrB1 rpsL (Str+)	Wertman et al. (1986)
BL21DE3	F− ompT hsdSB (rB− mB−) gal dcm (DE3)	Novagen
S. agalactiae
A909	Wild-type (WT) Serotype Ia	Madoff et al. (1991)
LR113	A909, stk1::Ωkm-2\KmR	Rajagopal et al. (2003)
LR114	A909, Δstp1 stk1::Ωkm-2\KmR	Rajagopal et al. (2003)
AJ20C4	A909, guaA::Tn917stm\EmR	Jones et al. (2000)
Plasmids
pYJ335	ColEI ori, pE194ori, tetR/PR/Pxyl/tetO-cat\EmR ApR CmIR	Ji et al. (1999)
pLZ12spec	3.7 kb, SpR	Husmann et al. (1995)
pLR16	5.5 kb, pLZ12spec, tetR/PR/Pxyl/tetO-cat\SpR CmIR	This study
pLR16ATG	4.8 kb, pLR16, tetR/PR/Pxyl/tetO-ATG\SpR	This study
pLR16T	4.5 kb, pLR16ATG, tetR/PR/Pxyl/tetO-ATG – MCS-ΩR\SpR	This study
pAV51	5.2 kb, pLR16T, stp1a\SpR	This study
pAV52	6.4 kb, pLR16T, stk1a\SpR	This study
pAV53	7.1 kb, pLR16T, stp1stk1a\SpR	This study
pET32a	5.4 kb, His6 expression vector\ApR	Novagen
pET32CK	5.4 kb, His6 expression vector derived from pet32a, \ApR	R. Seepersaud et al. (unpublished)
pAV55	6.8 kb, pET32CK, guaBa\ApR	This study
pAV56	6.6 kb, pET32CK, purAa\ApR	This study
Primer	Sequence
pLR16F	5′-TTGGCGCGCCGTTCATTTGATATGCCTCCGAAT-3′
pLR16R	5′-TCGGCGCGCCAGATCTTCCTTCAGGTTATGACC-3′
omttbgl2	5′-GAAGATCTGTCGACATCGATCTGATCCGGTGGATGAAATT-3′
omttsph1	5′-ACATGCATGCATCCGGTGATTGATTGAGCAAGCT-3′
StpF	5′-TTGGCGCGCCATGGAAATATCTCTTTTAAC-3′
StpR	5′-GGAAGATCTTTAAACCGCCTCACTTTC-3′
Stk1F	5′-TTGGCGCGCCATGATTCAGATTGGCAAATTA-3′
Stk1R	5′-ACGCGTCGACTCACTTGGTAGTGTTGAAC-3′
GuaBF	5′-CATGCCATGGCATCAAATTGGGATAC-3′
GuaBR	5′-ACGCGTCGACGTGTACTGAATAATTTG-3′
PurAF	5′-CATGCCATGGCTTCAGTTGTTGTAGTAGG-3′
PurAR	5′-ACGCGTCGACAATATTTGACCAAACAGAC-3′

^aInsert amplified from A909 chromosomal DNA.

Em^R, erythromycin resistant; Ap^R, ampicillin resistant; Sp^R, spectinomycin resistant; Km^R, kanamycin resistant; Cm^IR, inducible chloramphenicol resistance.

To test this hypothesis, we compared nutritional requirements of the Stk1-deficient strains LR113 and LR114 to the wild-type (WT) strain A909, as described in Experimental procedures. As GBS are auxotrophic for several amino acids, a subminimal chemically defined medium (CDM; Willett and Morse, 1966) was used. To assess whether LR113 and LR114 had amino acid requirements different from that of WT, these bacteria were grown in various pools of CDM (Samen et al., 2004) that were missing one of the 20 amino acids in each case respectively. We observed that the GBS WT strain (A909) was auxotrophic for 12 amino acids (arginine, glycine, histidine, tryptophan, threonine, valine, tyrosine, phenylalanine, leucine, isoleucine, methionine and serine). The mutant strains LR113 and LR114 had amino acid requirements that were similar to the WT (data not shown).

We next assessed these strains for deficiencies in nucleotide biosynthesis. As CDM [herein called complete CDM (cCDM)] contained the purine and pyrimidine salvage compounds, i.e. adenine, guanine, xanthine and uracil, we examined nucleotide biosynthesis using deleted CDM (dCDM, lacking purine and pyrimidine compounds). Interestingly, both LR113 and LR114 were attenuated for growth in the absence of exogenous purines. As shown in Fig. 2, the WT GBS strain, A909 is prototrophic for both purine and pyrimidine biosynthesis and reach an optical density at 600 nm (OD₆₀₀) of 0.8–0.9 (3–5 × 10⁹ cfu ml⁻¹) in dCDM, similar to growth in cCDM (Fig. 2A–C). However, both LR113 and LR114 demonstrated attenuated growth in dCDM (compare Fig. 2B and C with Fig. 2A). The final OD₆₀₀ for these strains ranged between 0.3 and 0.4 [0.5–1 × 10⁸ colony-forming units (cfu) ml⁻¹]. No further increase in OD or growth was observed even on prolonged incubation (> 48 h). As the addition of exogenous pyrimidines to the growth media (Fig. 2C) did not relieve the growth defect of LR113 and LR114, this indicated that the stk1 mutants were attenuated primarily for purine biosynthesis.

Fig. 2.

Purine auxotrophy of the Stk1-deficient strains, LR113 and LR114. Nucleotide auxotrophies were assessed as described in Experimental procedures. Cell growth was monitored at OD₆₀₀ after overnight incubation at 5% CO₂ at 37°C. Media composition:

A. cCDM (contains exogenous purines and pyrimidines).

B. dCDM (lacking purines or pyrimidines).

C. dCDM (supplemented with uracil).

D. dCDM (supplemented with adenine and uracil).

E. dCDM (supplemented with xanthine and uracil).

F. dCDM (supplemented with guanine and uracil).

G. dCDM (supplemented with inosine and uracil).

H. dCDM (supplemented with adenine).

I. dCDM (supplemented with xanthine).

J. dCDM (supplemented with guanine).

K. dCDM (supplemented with adenine, guanine and xanthine).

L. THB.

The predicted purine biosynthetic pathway based on homologues present in the GBS genome (Tettelin et al., 2002) is shown in Fig. 3. To determine whether LR113 and LR114 were able to satisfy their purine requirement by utilization of the purine salvage pathway, we supplemented dCDM with exogenous purine compounds such as adenine, inosine, xanthine or guanine, either in the presence or in the absence of a pyrimidine source, i.e. uracil. We observed that xanthine and guanine were able to satisfy the purine requirements of LR113 and LR114 (Fig. 2E, F, I and J) with xanthine being the best purine source in the absence of pyrimidines. Adenine and inosine were able to satisfy the purine auxotrophic requirements of the stk1 mutants only in the presence of uracil (Fig. 2D and G). Addition of either adenine or all purines (adenine, xanthine and guanine) to dCDM, in the absence of pyrimidines (Fig. 2H and K), caused severe repression in growth of both WT and mutant strains. Adenine, inosine and guanine have been previously described to promote binding of the purine repressor PurR to the operator sites and regulate purine and pyrimidine biosynthesis in Escherichia coli and Salmonella typhimurium (Houlberg and Jensen, 1983; Zalkin and Nygaard, 1996). Further, the genome sequence of GBS reveals the existence of a purR homologue; therefore, we speculate that the severe repression of growth observed in the presence of purines and absence of pyrimidines might result from PurR-mediated repression of nucleotide biosynthesis. Interestingly, the number of cfu of LR113 and LR114 was 10-fold lower when adenine, adenosine or inosine were used as a purine source in medium containing uracil. This difference in cfu was not observed when guanine, guanosine or xanthine was added to the growth medium (data not shown).

Fig. 3.

Predicted purine biosynthetic pathway in GBS. The de novo purine biosynthetic and salvage pathways, based on homologues present in the GBS genome, are shown. Dotted lines represent the de novo pathway, filled lines represent the salvage pathway and the double lines indicate steps common to both pathways. Dashed line indicates salvage of AICA (see below) in E. coli and S. typhimurium. Precursors and products are indicated in bold letters and enzymes catalysing the reactions are represented by their genes. SAG denotes ORF numbers from the S. agalactiae genome (Tettelin et al., 2002). R5P, ribose 5-phosphate; PRPP, phosphoribosylpyrophosphate; PRA, 5-phospho-β-d-ribosylamine; GAR, 5′-phosphoribosylglycinamide; FGAR, 5′-phosphoribosyl-N-formylglycinamide; FGAM, 5′-phosphoribosyl-N-formylglycinamidine; AIR, 5′-phosphoribosyl-5-aminoimidazole; NCAIR, 5′phosphoribosyl-5-carboxyaminoimidazole; CAIR, 5′-phosphoribosyl-5-aminoimidazole-4-carboxylate; SAICAR, 5′-phosphoribosyl-4-(N-succino-carboxamide)-5-aminoimidazole; AICAR, 5′phosphoribosyl-4-carboximide-5-aminoimidazole ribonucleotide; FAICAR, 5′-phosphoribosyl-4-carboximide-5-formaminoimidazole; IMP, inosine 5′-monophosphate; XMP, xanthosine monophosphate; GMP, guanosine monophosphate; GTP, guanosine triphosphate; sAMP, adenylosuccinate; AMP, adenosine monophosphate; ATP, adenosine triphosphate; PP_i, inorganic pyrophosphate; prsA, ribose-phosphate pyrophosphokinase; purF, glutamine PRPP amidotransferase; purD, GAR synthetase; purN, GAR trans-formylase-N; purL, FGAM synthetase; purM, AIR synthetase; purK, NCAIR synthetase; purE, NCAIR mutase; purC, SAICAR synthetase; purH, AICAR trans-formylase/IMP cyclohydrolase; guaB, IMP dehydrogeanse (IMPDH); guaA, GMP synthetase; guaC, GMP reductase; purA, sAMP synthetase; purB, sAMP lyase; deoD-1/deoD-2, purine nucleoside phosphorylase; apt, adenine phosphoribosyltransferase; xpt, xanthine phosphoribosyltransferase; guaF, guanine-hypoxanthine phosphoribosyltransferase.

^aTwo-monocistronic copies of prsA (SAG0018 and SAG1097) are present on the GBS genome.

Purine biosynthetic intermediates such as AICA (4-aminoimidazole-5-carboxamide) have complemented purine auxotrophs of E. coli and S. typhimurium (Zalkin and Nygaard, 1996). AICA when phosphoribosylated by the enzyme Apt (adenine phosphoribosyltransferase) in the presence of phosphoribosylpyrophosphate (PRPP) forms AICAR (5′phosphoribosyl-4-carboximide-5-aminoimidazole ribonucleotide), which restores de novo purine biosynthesis (see Fig. 3). Therefore, we substituted dCDM with purine biosynthetic intermediates such as AICA or AICAR; however, these intermediates were unable to restore normal growth to the mutant strains LR113 and LR114 either in the presence or in the absence of uracil (data not shown). It is possible either that GBS is defective in the uptake of these intermediates of de novo purine biosynthesis or that these mutants are attenuated in subsequent enzymatic steps of the purine biosynthetic pathway. Collectively, these data indicated that the de novo purine biosynthetic pathway is severely affected in the stk1 mutants and that these strains can utilize the purine salvage pathway to satisfy these auxotrophic requirements.

Doubling time

We then compared doubling times of LR113 and LR114 to the WT strains in the presence and absence of exogenous purine sources and the results are shown in Table 2. As expected, the doubling time of the mutant strains, LR113 and LR114, in dCDM was significantly longer (115 and 120 min) than the WT strain (68 min, see Table 2). Interestingly, the doubling times of LR113 and LR114 were also longer in the presence of purine salvage compounds such as adenine or inosine in dCDM containing uracil, when compared with WT. In contrast, the doubling time of the mutant strains were similar to WT in rich media or in dCDM containing uracil and either guanine or xanthine. These results suggest that the increase in doubling time of the Stk1-deficient strains in the presence of adenine might be the cause for the 10-fold drop in cfu. The attenuation of growth observed in the stk1 mutants in dCDM may result from altered Stk1-dependent regulation of Ppase activity. Alternatively, these strains might be affected for enzyme(s) that are required for both the salvage of adenine/inosine and de novo purine biosynthesis.

Table 2.

Altered generation time of LR113 and LR114.

		Doubling time in minutes
Strain	Genotype	dCDMa	Ab	Ib	Gb	Xb	cCDMa	THB
A909	Wild type	68	53	50	55	68	50	28
LR113	stk1::Ωkm-2	115	70	64	55	68	64	28
LR114	Δstp1stk1::Ωkm-2	120	75	72	50	68	60	28

^acCDM (supplemented with exogenous purines and pyrimidines), dCDM (devoid of purines and pyrimidines).

^bdCDM supplemented with uracil and various purine sources; A, adenine; I, inosine; G, guanine; X, xanthine.

Doubling time or generation time was compared during exponential phase growth of the GBS strains.

THB, Todd Hewitt Broth.

Constitutive expression of Stk1 during growth of GBS

As the Stk1-deficient strains LR113 and LR114 demonstrated attenuated growth in dCDM, we sought to confirm that WT GBS encodes Stk1 during growth in minimal media (cCDM and dCDM). Although we had previously demonstrated that the stk1 gene in GBS encodes an STK that autophosphorylates at serine residues, these experiments were conducted using GBS that were grown in rich-media [Todd Hewitt Broth (THB)] (Rajagopal et al., 2003). Using in vitro phosphorylation, we had shown that Stk1 activity was detected in membrane fractions of GBS, consistent with the presence of a trans-membrane domain in the coding sequence (Rajagopal et al., 2003). To confirm Stk1 expression during growth of GBS in minimal media, we performed in vitro phosphorylation reactions (in the presence of [γ-³²P]-ATP) on membrane fractions isolated from A909 and the Stk1-deficient strain LR113, grown in dCDM, cCDM and THB respectively. The reaction products were analysed on a 10% SDS-PAGE followed by autoradiography. As shown in Fig. 4 (lanes 1–3, 5–7 and 9–11), the band at ∼65 kDa, corresponding to autophosphorylated Stk1 was observed in membrane fractions of the WT during various growth phases, irrespective of the media. This band was absent in LR113 (Fig. 4, lanes 4, 8 and 12) indicating the absence of Stk1 expression in the mutant, as described previously (Rajagopal et al., 2003). These data confirmed our initial hypothesis that WT GBS constitutively expressed Stk1 during growth in either rich or minimal media.

Fig. 4.

Stk1 is expressed during growth of GBS in rich and minimal media. In vitro phosphorylation reactions were performed on membrane fractions of WT and LR113 as described in Experimental procedures. Lanes 1–3 indicates membrane fractions isolated from WT GBS (A909) grown in THB to an OD₆₀₀ of 0.15, 0.3 and 0.6 respectively. Lanes 5–7 indicates membrane fractions isolated from A909 grown in dCDM to an OD₆₀₀ of 0.15, 0.3 and 0.6 respectively. Lanes 9–11 indicates membrane fractions isolated from A909 grown in cCDM to an OD₆₀₀ of 0.15, 0.3 and 0.6 respectively. Lanes 4, 8 and 12 indicate membrane fractions isolated from LR113 grown to an OD₆₀₀ of 0.3 in THB, dCDM and cCDM respectively.

Intracellular inorganic pyrophosphate (PP_i)

An arrest in purine biosynthesis has been previously described to be associated with an increase in intracellular pools of inorganic pyrophosphate (PP_i) (Kukko and Saarento, 1983). As a consequence, the enzymatic reactions that liberate PP_i as a by-product can be severely affected due to non-allosteric feedback inhibition (Lahti, 1983). As we had previously described reversible phosphorylation of PpaC (encoding an inorganic pyrophosphatase) by Stk1 and Stp1 in GBS (Rajagopal et al., 2003), we hypothesized that the growth arrest of LR113 and LR114 in dCDM might result from accumulation of intracellular PP_i. Measurements of intracellular PP_i pools was performed as described in Experimental procedures and the results are shown in Fig. 5. Surprisingly, we observed that PP_i levels were significantly higher in the WT rather than in the Stk1-deficient strains, during growth in dCDM (Fig. 5). In contrast, PP_i levels were substantially higher in LR113 and LR114 during growth in cCDM. This might result from the fact that PP_i is an essential by-product of the purine salvage pathway and also that levels of PP_i can be expected to be higher in actively growing cultures. These data imply that kinase-dependent regulation of Ppase activity is most likely to occur when intracellular PP_i pools are significantly higher. Taken together, these results indicate that the cause of growth arrest in LR113 and LR114 in dCDM does not result from accumulation of PP_i within the cell.

Fig. 5.

Intracellular inorganic pyrophosphate (PP_i) levels are not increased during growth of LR113 and LR114 in dCDM. Intracellular PP_i levels were determined as described in Experimental procedures. The amount of PP_i is expressed as micromoles per mg (dry weight). All experiments were performed in triplicate and repeated for reproducibility. cCDM represents media containing purines and pyrimidines and dCDM represents media lacking purines and pyrimidines.

Sensitivity to azaserine and purine analogues

We next tested the hypothesis that the Stk1-deficient strains may be decreased in activity of enzymes essential for de novo purine biosynthesis (see Fig. 3 for purine biosynthetic pathway). We therefore assessed sensitivity of these strains to inhibitors of purine biosynthesis. Azaserine has been previously shown to inhibit glutamine PRPP amidotransferase (purF), the first enzyme involved in de novo purine biosynthesis (Gollub and Gots, 1956). We observed that the Stk1-deficient strains were comparable to WT in sensitivity to azaserine [minimum inhibitory concentration (MIC) = 0.05 μM] in dCDM. These results indicated that the Stk1-deficient strains are possibly defective in subsequent steps of the de novo purine biosynthetic pathway and corroborates the lack of complementation for purine biosynthesis by AICA, in these strains. B. subtilis strains resistant to 2-flouroadenine (2-FA) had decreased Apt activity (Saxild and Nygaard, 1987). The Stk1-deficient strains LR113 and LR114 were similar to WT (A909) in sensitivity to 2-FA (data not shown).

We also compared the sensitivity of WT and the stk1 mutants with mycophenolic acid. Mycophenolic acid (MA) is an inhibitor of IMP dehydrogenase (IMPDH), an enzyme that catalyses the conversion of IMP to XMP (Zalkin and Nygaard, 1996). Interestingly, both LR113 and LR114 demonstrated severe sensitivity to MA, both in dCDM or when adenine was included as the sole purine source in the medium (see Table 3) suggesting that these mutants may have altered levels/activity of IMPDH. This sensitivity was abolished when these strains were grown in medium containing guanine as the sole purine source, consistent with the fact that salvage of guanine for synthesis of purines does not require IMPDH (Table 3). As a control, we included a GMP synthetase (guaA) mutant AJ20C4 (Jones et al., 2000). This strain contains a transposon (Tn917stm) insertion in guaA and requires either guanine or guanosine as a purine source (see Fig. 3 for purine biosynthetic pathway). As expected, the guaA mutant AJ20C4 was not sensitive to MA as this strain is defective in the subsequent step for GMP synthesis (see Table 3).

Table 3.

The Stk1-deficient strains, LR113 and LR114, are more sensitive to the purine analogues mycophenolic acid and 6-mercaptopurine.

	% Survival
	Mycophenolic acid				6-Mercaptopurine
	AU	GU	cCDM	dCDM	GU	cCDM	dCDM
A909	76	98	78	81	62.5	75	0.015
LR113	0.001	95	5.2	0.008	0.041	62.5	0
LR114	0.002	100	4.4	0.009	0.032	66.6	0
AJ20C4	ND	95	95	ND	0.0	0.0	ND

Sensitivity of the GBS strains to purine analogues was performed as described in Experimental procedures and is denoted as percent survival. Mycophenolic acid was used at a final concentration of 0.3 mM and 6-mercaptopurine was used at a final concentration of 2 mM. AU, dCDM containing adenine and uracil; GU, dCDM containing guanine and uracil; ND, not determined.

The purine analogue 6-mercaptopurine (6-MP) when phosphoribosylated by guanine phosphoribosyltransferase (guaF) to thio-IMP can inhibit both glutamine PRPP amidotransferase (purF) and IMPDH (guaB) (Zalkin and Nygaard, 1996). The strains LR113 and LR114 were more sensitive to 6-MP when compared with WT in dCDM (see Table 3). This might also result from inhibition of IMPDH and possibly glutamine PRPP amidotransferase. Interestingly, LR113 and LR114 displayed increased sensitivity to 6-MP even in dCDM containing guanine as the sole purine source. This was surprising as salvage of guanine does not utilize IMPDH. However, we also observed that the guaA mutant displayed severe sensitivity to 6-MP, under these conditions. We speculate that the increased sensitivity of LR113, LR114 and the guaA mutant to 6-MP might result from decreased levels of GuaF enzyme required for salvage of guanine in the presence of 6-MP, rather than inhibition of IMPDH or PRPP amidotransferase. Alternatively, these strains might have a higher requirement of guanine nucleotides for growth.

Intracellular nucleotide pools

Intracellular nucleotide pools of WT, LR113 and LR114 were measured during growth in dCDM and also in cCDM. We hypothesized that if the attenuation in purine biosynthesis of the Stk1-deficient strains resulted from altered IMPDH activity, then the pools of guanine nucleotides (GMP, GDP and GTP) will be lower in these strains when compared with the WT. Nucleotide pool measurements were carried out as described in Experimental procedures and the results are shown in Table 4. We observed that the strains LR113 and LR114 had similar ATP and slightly increased AMP pools during growth in dCDM (Table 4). ADP pools were also marginally increased in LR113 and LR114 (data not shown). In contrast, the guanine nucleotides (GMP, GDP and GTP) were significantly lower in LR113 and LR114 when compared with WT, during growth in dCDM. The UTP and CTP pools in these strains were slightly decreased when compared with WT, under these growth conditions. The ratio of ATP:GTP was ∼40-fold higher in the Stk1-deficient strains. This increase in ATP:GTP ratio was not observed during growth of these strains in cCDM (see Table 4) or in THB (data not shown).

Table 4.

The Stk1-deficient strains demonstrate a substantial decrease in intracellular guanine nucleotide pools.

	Strain	AMP	GMP	GDP	ATP	GTP	CTP	UTP	ATP:GTP	CTP:UTP
dCDM	A909	20	15	23	10.5	1.3	10	1.2	8	8
dCDM	LR113	28	5	3	10.1	0.2	7	0.7	50	10
dCDM	LR114	32	6	4	12	0.3	7.2	0.8	40	9
cCDM	A909	41	43	26	1132	45	85	354	25	0.24
cCDM	LR113	15	31	30	1508	60	108	522	25	0.2
cCDM	LR114	15	25	50	916	45	74	340	20	0.21
cCDM	AJ20C4	BD	34	18	1138	41	68	323	27	0.21
GU	A909	9	6	8	1365	47	170	472	29	0.36
GU	LR113	57	3	7	342	14	47	160	24	0.3
GU	LR114	48	4	8	397	18	47	133	22	0.35
AU	A909	12	8	21	938	60	59	307	15	0.2
AU	LR113	103	4	23	428	36	62	157	12	0.39
AU	LR114	92	3	21	383	32	55	140	12	0.39

Intracellular nucleotide concentrations were estimated by HPLC as described in Experimental procedures. Nucleotide concentrations are expressed as picomoles per mg (dry weight). dCDM (lacking purines and pyrimidines), cCDM (contains purines and pyrimidines), GU (dCDM with guanine and uracil), AU (dCDM with adenine and uracil). BD denotes below the limit of detection. ATP:GTP and CTP:UTP indicate respective ratios.

We then compared nucleotide pools of WT, LR113 and LR114 during growth in media containing uracil and either adenine or guanine as purine sources. Under these growth conditions, the ratio of ATP:GTP in LR113 and LR114 was similar to WT (Table 4). In contrast to WT, both LR113 and LR114 had significantly higher AMP pools (>10-fold) when adenine was added to the growth media (similar results were obtained with inosine, data not shown). This increase in AMP was lowered (approximately fivefold) in the presence of guanine. It is possible that decreased IMPDH activity in these strains might lead to an increase in intracellular IMP pools. As a consequence, the accumulated IMP can be converted to AMP by the action of enzymes encoded by purA and purB (see Fig. 3 for purine biosynthetic pathway). Alternatively, the stk1 mutants may have increased PurA activity, which may lower intracellular IMP levels thereby affecting IMPDH activity and consequently, GMP synthesis.

Interestingly, the pools of G nucleotides were also lower in LR113 and LR114 when compared with WT, during growth in dCDM containing guanine as the sole purine source. As GMP can be reduced to IMP by GMP reductase (guaC), this discrepancy in GMP pools might also result from an imbalance in regulation of GuaB and PurA activities on IMP in LR113 and LR114. As the XMP and IMP nucleotide peaks did not separate under the high-performance liquid chromatography (HPLC) conditions used, we were unable to discern their concentrations within the cell. We speculate that the growth arrest of LR113 and LR114 in the absence of purines might result from significantly lower GTP levels. GTP is essential for continued synthesis of 10 formyl-tetrahydrofolate, an important cofactor in purine and pyrimidine biosynthesis (Zalkin and Nygaard, 1996; Sauer et al., 1998). Further, addition of folic acid did not alleviate the growth arrest of the Stk1-deficient strains (data not shown) suggesting that these bacteria may not take up folic acid, similar to E. coli (Zalkin and Nygaard, 1996). This indicates that folic acid must be synthesized de novo for nucleotide biosynthesis and one carbon metabolism. Alternatively, it is possible that an increase in intracellular ATP:GTP ratios that is evident in LR113 and LR114 only during growth in dCDM might be the cause for the severe attenuation in growth of the stk1 mutant strains.

It is noteworthy that a leaky guaB (IMPDH) mutant of S. typhimurium (Jensen, 1979) had increased ATP:GTP ratio (∼10-fold) and a 1.5-fold decrease in CTP:UTP ratio when compared with the WT strain, during growth in media lacking guanine as a purine source. The discrepancy in ratios of ATP:GTP and CTP:UTP was not observed in the leaky guaB strain during growth in the presence of guanine (Jensen, 1979). Although no significant differences were observed in the ratio of CTP:UTP in LR113 and LR114, the increase in ATP:GTP ratio is similar to the guaB S. typhymurium strain. These data provide further support to the hypothesis that the Stk1-deficient strains have altered IMPDH activity. It is not known whether the leaky guaB mutation in S. typhimurium conferred a growth arrest in the absence of exogenous purines.

Purine enzyme activities

To analyse whether the decrease in G nucleotide pools in LR113 and LR114 during growth in dCDM can be correlated to decreased IMPDH activity or conversely, an increase in PurA activity, we compared activities of enzymes encoded by guaB, guaA, guaC and purA in total cell extracts, as described in Experimental procedures. Interestingly, we observed that the IMPDH activity was approximately twofold decreased in LR113 and LR114 compared with WT (see Table 5). Also, these strains demonstrated a twofold increase in PurA activity (Table 5). These data confirmed that the Stk1-deficient strains had decreased IMPDH activity and that the increase in AMP pools results from the increased PurA activity observed in these strains. GMP reductase activity was not detected in cell extracts of LR113 and LR114, presumably due to lower G nucleotide levels. It has been previously reported that activity of GMP reductase was significantly higher in the presence of guanine or G nucleotides (Garber et al., 1980). These data indicate that the altered levels of G and A nucleotides during growth of the Stk1-deficient strains in dCDM compares well with the differences in activities of the enzymes that regulate these pools.

Table 5.

Altered activity of purine biosynthetic enzymes in the GBS Stk1-deficient strains LR113 and LR114.

	Specific activity
	IMPDH (GuaB)	XMP synthetase (GuaA)	GMP reductase (GuaC)	Adenylosuccinate synthetase (PurA)
A909	1.6 ± 0.04	5.3 ± 0.9	0.096 ± 0.01	2.4 ± 0.1
LR113	0.6 ± 0.05	5.7 ± 0.4	ND	4.5 ± 0.2
LR114	0.8 ± 0.05	5.5 ± 0.7	ND	4.9 ± 0.2

Isolation of bacterial cell extracts and activity assays were performed as described in Experimental procedures. One unit is defined as the amount of enzyme required to utilize or release one nanomole of substrate or product per minute (see Experimental procedures). Specific activity is denoted as UNITS per mg protein.

ND, not detected.

Reversible phosphorylation of purine enzymes by Stk1 and Stp1

The difference in activity of GuaB and PurA in LR113 and LR114 prompted us to analyse reversible phosphorylation of these proteins by Stk1 and Stp1. A human STK (PKB/Akt) was recently described to phosphorylate IMPDH via a pleckstrin homology (PH) domain; a twofold increase in activity of IMPDH was observed on kinase-dependent phoshorylation (Ingley and Hemmings, 2000). Although we did not locate a PH domain in the coding sequence of Stk1, we hypothesized that Stk1 and possibly Stp1 may reversibly phosphorylate GuaB and/or PurA. To our knowledge, kinase-dependent phosphorylation of PurA has not been described. To test whether Stk1 and Stp1 reversibly phosphorylate GuaB and/or PurA, we constructed C-terminal His-tag constructs to the GBS purA and guaB genes and recombinant fusion proteins were purified as described in Experimental procedures. Subsequently, in vitro phosphorylation reactions were conducted on the GuaB–His₆ and PurA–His₆ fusion proteins either in the presence or in the absence of Stk1 (see Experimental procedures). The reaction products were analysed on a 10% SDS-PAGE, stained with coomassie to localize proteins and exposed to autoradiography (see Fig. 6A). We have previously described autophosphorylation of the GST–Stk1 fusion protein (rStk1, 96 kDa) in the presence of [γ-³²P]-ATP (Rajagopal et al., 2003) and an autophosphorylated product corresponding to rStk1 can be seen at ∼96 kDa (Fig. 6A, lanes 3–5). Neither GuaB nor PurA autophosphorylated in the presence of [γ-³²P]-ATP alone (Fig. 6A, lanes 2 and 6). Interestingly, we observed a phosphorylation product at ∼49 kDa, corresponding to PurA–His₆ in the presence of rStk1 (Fig. 6A, lane 5). As we did not observe autophosphorylation of the PurA–His₆ fusion protein, this indicates kinase-dependent phosphorylation of PurA.

Fig. 6.

Reversible phosphorylation of PurA by Stk1 and Stp1. All in vitro phosphorylation reactions were performed in the presence of 10 μCi [γ-³²P]-ATP. Phosphorylation reactions of PurA and GuaB fusion proteins were performed in the presence or absence of rStk1 as described in Experimental procedures. Recombinant Stp1 was also added subsequent to in vitro phosphorylation reaction containing rStk1 and PurA. All reaction products were analysed on a 10% SDS-PAGE gels, stained with coomassie and subjected to autoradiography. A. Lane 4 represents autophosphorylation of rStk1; lanes 2 and 6 represent addition of 10 μCi [γ-³²P]-ATP to GuaB and PurA respectively. Lanes 3 and 5 indicate addition of rStk1 to GuaB and PurA in the presence of 10 μCi [γ-³²P]-ATP respectively. Lane 1 represents addition of rStp1 subsequent to the in vitro phosphorylation reaction represented in lane 5. The position and molecular masses (kDa) of protein standards are indicated on the left. Open arrows indicate breakdown products of Stk1 upon autophosphorylation.

B. Time-course dephosphorylation of rStk1 and PurA by rStp1. Phosphorylated rStk1 and PurA were incubated with rStp1 in the presence of Mn²⁺ as described in Experimental procedures. Aliquots of the reaction were removed at various time intervals (0–60 min) and analysed on SDS-PAGE followed by autoradiography.

The open arrows represent breakdown products of rStk1 observed on autophosphorylation (seen in all lanes containing rStk1, Fig. 6A, lanes 3–5). Interestingly, kinase-dependent phosphorylation of the GuaB–His₆ fusion protein was not observed indicating that GuaB may not be a target of Stk1 (see Fig. 6A, lane 3).

We also observed that both PurA and Stk1 were dephosphorylated when Stp1 was added to purified, phosphorylated rStk1 and PurA–His₆ (see Experimental procedures and Fig. 6A, lane 1). In addition, we analysed time-course dephosphorylation of rStk1 and PurA–His₆ by rStp1. Recombinant Stp1 was added to purified, phosphorylated rStk1 and PurA–His₆ in the presence of Mn²⁺ as described in Experimental procedures. Aliquots were removed at various time points and the products were analysed on SDS-PAGE followed by autoradiography. Figure 6B shows that rStp1 dephosphorylated both rStk1 and PurA–His₆ and that dephosphorylation increased with time. Phosphoimager analysis revealed that the rate of rStp1-mediated dephosphorylation of rStk1 was approximately twofold greater than rStp1-mediated dephosphorylation of PurA–His₆ (data not shown). Taken together, these results indicate reversible phosphorylation of the GBS PurA by Stk1 and Stp1.

Effect of phosphorylation on PurA enzyme activity

We also analysed the effect of phosphorylation on PurA (adenylosuccinate synthetase) activity. In vitro phosphorylation of PurA–His₆ was performed in the presence of rStk1 and ATP. As controls, we included reactions that contained unphosphorylated PurA (i.e. lacked the phosphate donor, ATP or the rStk1 protein, see Experimental procedures). All reactions were purified using G25 sephadex spin columns respectively. Subsequently, PurA assays were conducted as described in Experimental procedures and the results are summarized in Table 6. Interestingly, we observed a 1.7- to 1.8-fold decrease in activity of phosphorylated PurA (reactions performed after phosphorylation in the presence of rStk1 and ATP) when compared with unphosphorylated PurA (control phosphorylation reactions, i.e. lacking either rStk1 or ATP), indicating that phosphorylation decreased PurA enzyme activity. These data are consistent with the approximately twofold increase in PurA enzyme activity observed in the Stk1-deficient strains compared with WT (A909).

Table 6.

Effect of phosphorylation on PurA activity.

	Specific activity (U mg−1)
PurA–His6	79 ± 4.0
PurA–His6 + rStk1	80.55 ± 2.7
PurA–His6 + rStk1 + [ATP]	44.58 ± 3.2

In vitro phosphorylation of PurA–His₆ fusion protein was performed in the presence of rStk1 and ATP as described in Experimental procedures. Controls included unphosphorylated PurA–His₆ (reactions that did not contain either ATP or the rStk1 protein). All reactions were purified and PurA (adenylosuccinate synthetase) enzyme assays were conducted in the presence and absence of aspartic acid respectively (see Experimental procedures). One unit is defined as the amount of enzyme required for formation of one nanomole of sAMP per minute. The average specific activity and standard deviation obtained from three independent experiments is reported.

As we did not observe kinase-dependent phosphorylation of GuaB (IMPDH), we speculate that the decrease in levels/activity of IMPDH in the stk1 mutants might either from result a consequence of the increase in PurA activity or from other factors that regulate IMPDH activity in GBS.

Complementation

To confirm that the attenuated growth of LR113 and LR114 in dCDM was linked to the stk1 deficiency in these strains, we performed complementation experiments. Plasmids containing complementing clones were constructed using the tetracycline (tet)-inducible expression vector as described in Experimental procedures. We confirmed that the tet-inducible expression system (Ji et al., 1999) was functional in GBS. As mentioned previously (Geissendorfer and Hillen, 1990), this inducible promoter has a low level of expression even in the absence of the inducer (anyhydrotetracycline, Atc). However, a dose-dependent increase in chloramphenicol-resistant (Cm^R) colonies was obtained in the presence of varying concentrations of the inducer Atc. The genes stk1, stp1 and both stk1 and stp1 were amplified from the GBS genome, cloned into pLR16T, and introduced into WT, LR113 and LR114 respectively (see Experimental procedures for details).

The results from the complementation experiments are shown in Fig. 7. We observed that strains LR113 and LR114 with the stk1 gene in trans were complemented for growth in dCDM [compare growth of LR113 or LR114 with pLR16T (vector) with strains containing complementing constructs LR113/pStk1 and LR114/pStk1 in Fig. 7]. This suggests that the mutation in stk1 was responsible for the attenuated purine biosynthesis in LR113 and LR114. As complementation of the growth defect was observed even in the absence of induction, this indicated that the basal level of expression from the tet-inducible promoter was sufficient for complementation. In addition, we observed that an increase in expression of Stp1 in the WT strain (caused attenuated growth in dCDM (A909/pStp1) (Fig. 7); this attenuation in growth was not observed in the Stk1-deficient strains LR113 and LR114 (LR113/pStp1 and LR114/pStp1 in Fig. 7). We have previously demonstrated that Stp1 dephosphorylates Stk1 (Rajagopal et al., 2003), therefore, we hypothesize that Stp1 might be a negative regulator of Stk1 or that an imbalance in the Stp1:Stk1 ratio might prevent phosphorylation of Stk1 targets. Collectively, the results presented here indicate that the eukaryotic-type kinase regulates PurA activity and consequently de novo purine biosynthesis. Further characterization of these enzymes and domains essential for phosphorylation is currently in progress in our laboratory and will provide clues on the nature of the sensors or activators of these signalling cascades.

Fig. 7.

Complementation restores normal growth to the Stk1-deficient strains in media lacking purines and pyrimidines (dCDM). Complementation constructs and growth in dCDM was performed as described in Experimental procedures. Increasing concentrations of the inducer, Anhydrotetracycline, were also added to dCDM (i.e. 0.5 and 1 μg ml⁻¹). The vector pLR16T was introduced into A909, LR113 and LR114. pStp1 denotes plasmid pAV51 that encodes the WT GBS stp1 allele, pStk1 denotes plasmid pAV52 that encodes the WT GBS stk1 allele and pStp1Stk1 denotes plasmid pAV53 that encodes the WT GBS stp1 and stk1 alleles.

Discussion

Over the last two decades, a number of reports have described the presence of eukaryotic-type signalling enzymes in prokaryotic systems. A few reports also describe the identification of physiological substrates for these enzymes. Despite the recent advances, much remains to be learned about the upstream activating signals and signal transduction mechanisms of eukaryotic-type kinases and phosphatases in prokaryotic species. A comparison of the phenotypes of the various prokaryotic STK and STP mutants in conjunction with the nature of their targets indicates that these enzymes regulate diverse functions in these organisms.

We had previously described reversible phosphorylation of an inorganic pyrophosphatase (PpaC) by Stk1 and Stp1 in GBS (Rajagopal et al., 2003). As inorganic pyrophosphatases (Ppases) have been implied to influence general metabolism (Lahti, 1983), we speculated that the Stk1-deficient strains might be defective in metabolic processes that are regulated by Ppases. Therefore, we characterized the mutants for auxotrophic phenotypes. In this study, we provide evidence that Stk1-deficient strains of GBS are unable to sustain de novo purine biosynthesis and are consequently growth arrested. These strains are able to utilize exogenous purines to compensate their auxotrophic requirement. It was previously demonstrated that purine or pyrimidine starvation in E. coli induced an accumulation of intracellular inorganic pyrophosphatase (PP_i) and an arrest in growth (Kukko and Saarento, 1983). In contrast, we observed that the GBS Stk1-deficient strains did not accumulate PP_i during their attenuated growth in dCDM. As starvation for purines or pyrimidines in E. coli was achieved using analogues that inhibit nucleotide biosynthesis (Kukko and Saarento, 1983), the growth arrest observed presumably resulted from the lack of the nucleotide(s) rather than the cell's inability to tolerate higher PP_i levels or PP_i-mediated inhibition of the biosynthetic pathway. Thus, it remains to be determined whether intracellular accumulation of PP_i due to altered Ppase activity can trigger a growth arrest in GBS. It is also noteworthy that while kinase-dependent phosphorylation of Ppases has been described in other systems besides GBS (Rudd and Franklin-Tong, 2003), the effect of kinase-dependent phosphorylation on Ppase activity is yet to be understood.

We propose that negative regulation of PurA activity by a eukaryotic-type STK is important for the accurate balance of purine nucleotide pools and regulation of de novo purine biosynthesis in GBS. Our results indicate that Stk1 and Stp1 reversibly phosphorylate PurA (adenylosuccinate synthetase), an enzyme involved in synthesis of AMP and that phosphorylation negatively regulates activity of the enzyme. In the absence of Stk1, a twofold increase in PurA activity was observed during growth of the stk1 mutants in dCDM. Although this increase in PurA activity per se seems modest, it becomes significant due to the concomitant decrease in IMPDH activity, G nucleotide pools and more importantly, the growth arrest observed in the stk1 mutants during growth in dCDM. Our proposed model is that Stk1 phosphorylates and inhibits PurA when intracellular A nucleotide pools or ATP concentrations increase and consequently suppresses further AMP and ATP synthesis. As a result of PurA inhibition, the net intracellular IMP pools for IMPDH activity is increased, which promotes XMP and consequently G nucleotide synthesis. In the absence of Stk1, PurA activity is not regulated in response to an increase in intracellular ATP concentrations; therefore, PurA activities are increased leading to a continual synthesis of A nucleotides and this eventually decreases the net IMP pools for IMPDH activity (also see below) and G nucleotide biosynthesis.

We speculate that Stp1 may primarily be a negative regulator of Stk1 in vivo, although it can dephosphorylate both Stk1 and PurA in vitro. We observed that an increase in Stp1 expression decreased growth of the WT GBS strain in dCDM (see Fig. 7) indicating that excess Stp1 expression may inhibit kinase (Stk1) function. Interestingly, however, the double mutant LR114 that is defective in expression of both Stk1 and Stp1 was complemented for normal growth in dCDM in the presence of Stk1 alone (i.e. absence of Stp1; see Fig. 7). This suggests that Stp1-mediated dephosphorylation of PurA might be dispensable for de novo purine biosynthesis/growth of GBS in dCDM.

It is noteworthy that while Stk1-dependent phosphorylation of GuaB (IMPDH) was not observed, an approximately twofold decrease in specific activity of the IMPDH enzyme was seen in the Stk1-deficient strains during growth in dCDM. These data indicate that guaB expression or activity maybe regulated by a mechanism independent of Stk1-dependent phosphorylation in GBS. As changes in intracellular nucleotide concentrations have been previously described to regulate expression of purine biosynthetic genes and particularly IMPDH activity (Lopez et al., 1981), it is a formal possibility that transcription and/or translation of GuaB may be decreased due to an imbalance in intracellular nucleotide pools.

The critical role of intracellular nucleotide concentrations and their balance on growth of organisms has been previously described (Saxild and Nygaard, 1991). It is interesting that the growth arrest observed in LR113 and LR114 occurred only in the absence of purines, despite the fact that G nucleotide pools were lower in these strains when guanine was the sole purine source. Although the guanine nucleotide pools were lower in LR113 and LR114 during growth in the presence of guanine, the ratio of intracellular nucleotide pools was comparable to the WT. Further, as we did not observe any increase in stress-related factors such as ppGpp (guanosine 3′-diphosphate 5′-triphosphate) or the folate stress response nucleotide (ZTP, 5-amino 4-imidazole carboxamide riboside 5′-triphosphate; Bochner and Ames, 1982) during growth of these strains in dCDM (data not shown), the growth arrest most probably results from the imbalance in intracellular nucleotide concentrations. One possibility is that in the absence of purines, the net intracellular concentrations of GTP and other precursor G nucleotides drop below threshold values. Of note, PurA is itself a GTP-utilizing enzyme; therefore, the increase in PurA activity combined with the decrease in G nucleotide biosynthesis might eventually cause all GTP requiring enzymes (including PurA) to be non-functional, leading to the growth arrest of the stk1 mutants in the absence of purines. Consistent with this hypothesis, we observe that intracellular nucleotide levels, even in the WT strain, is at least 10-fold lower in the absence of exogenous purines and pyrimidines. This indicates that de novo purine biosynthesis is less efficient than the salvage pathway in GBS and suggests that any perturbation might significantly affect the growth of the bacterium. Another possibility is that the Stk1-deficient strains may require higher GTP levels for growth. The increased sensitivity of the Stk1-deficient strains and the guaA mutant to 6-MP (in the presence of guanine) supports this hypothesis.

While it was previously thought that nucleotide biosynthesis was regulated primarily at the level of the PurR repressor, recent reports indicate additional levels of control. A new direction in the regulation of metabolic pathways is the existence of riboswitches. The 5′ untranslated mRNA of the xpt-pbuX operon in B. subtilis was shown to contain a conserved G-box that binds guanine and prematurely terminates expression of the operon and consequently regulates purine biosynthesis (Mandal et al., 2003). An attenuation in growth of B. subtilis, due to the G-box riboswitch was observed only in the presence of excess guanine in the media. Although the G-box riboswitch is present in the xpt-pbuX operon of GBS (Mandal et al., 2003), the cessation of growth in the Stk1-deficient strains was observed only in the absence of purines. Whether expression/activity of IMPDH (GuaB) is regulated by riboswitches during growth of GBS in the absence of exogenous purines and/or due to changes in adenine nucleotide pools remains to be determined. Collectively, these results indicate that multiple levels of control (i.e. transcriptional and post-translational) exist in the regulation of purine biosynthesis in bacterial systems.

To our knowledge, reports on biosynthesis or acquisition of purines and pyrimidines in GBS have been scarce, if not non-existent. These pathways have been extensively studied in Gram-negative organisms such as E. coli and S. typhimurium (Zalkin and Nygaard, 1996). Mutations in de novo purine biosynthetic genes of Salmonella or E. coli result in histidine or thiamine auxotrophies due to interlinking pathways between AICAR and ATP for thiamine and histidine biosynthesis respectively (Zalkin and Nygaard, 1996). In contrast, we observed that WT GBS are auxotrophic for thiamine and histidine. Consistent with this observation, the GBS genome indicates the absence of the genes/enzymes in the biosynthetic pathway leading from AICAR and ATP to thiamine and histidine respectively.

It is intriguing that while the GBS genome indicates the absence of multiple amino acid biosynthetic pathways, these bacteria have retained the genes essential for de novo purine and pyrimidine biosynthesis. Recently, a number of reports describe purine acquisition in Gram-positive organisms such as B. subtilis (Beaman et al., 1983; Saxild and Nygaard, 1987; 1991; Mandal et al., 2003), Lactococcus lactis (Martinussen et al., 2003). However, unlike GBS, the genome of the oral streptococcal pathogen Streptococcus mutans (Ajdic et al., 2002) shows the existence of genes/enzymes essential for histidine biosynthesis from nucleotide precursors. This disparity among related species might result from the fact that pathogens have evolved to either retain or eliminate essential and non-essential genes based on selection pressure and ecological niches. This paradigm is supported by the observation that parasitic organisms such as leishmania and trypanosomes lack the genes essential for de novo purine and pyrimidine biosynthesis; these organisms solely rely on salvage of purine and pyrimidine sources for growth (Fairlamb, 1989).

Mutations in purine biosynthetic genes have been previously described to severely attenuate virulence of both Gram-negative and Gram-positive pathogens. Purine biosynthesis is essential for salmonella virulence; its purine requirement has been extensively utilized as a tool to identify promoters that are regulated in vivo (Mahan et al., 1993). A guaBA deletion mutant of Shigella fexineri was described to be significantly attenuated in host cell invasion and intracellular growth (Noriega et al., 1996; Cersini et al., 2003). Mutations in de novo purine biosynthetic genes in Brucella sp. was previously described to affect the organism's ability survive and replicate within the phagosome (Alcantara et al., 2004). Interestingly, purine auxotrophic strains of GBS have been previously isolated in signature-tagged mutagenesis (STM) screens for decreased virulence (Jones et al., 2000). Given that the pathogen has to adapt and survive in a variety of host compartments (i.e. amniotic fluid, lung, blood and brain) during its infection cycle, identification of purine auxotrophs that are attenuated for virulence indicates that GBS encounters purine deficiencies in vivo. GBS has the ability to overcome these purine limitations by its ability to synthesize purines de novo. We have previously observed that the Stk1-deficient strains LR113 and LR114 were 50- to 100-fold attenuated for sepsis infections (Rajagopal et al., 2003). This suggests that the decreased virulence of these strains might in part result from their inability to sustain purine biosynthesis. Also, a twofold increase in expression of stk1 and purine biosynthetic genes was observed during growth of a related streptococci, S. pneumoniae in human blood (Orihuela et al., 2004). These data provide further evidence that eukaryotic-type kinases in prokaryotic pathogens regulate virulence, either by their influence on metabolic pathways and/or by factors essential for virulence.

Experimental procedures

Bacterial strains and media

Bacterial strains, plasmids and primers used in this study are listed in Table 1. All chemicals were purchased from Sigma-Aldrich, USA, unless mentioned otherwise. Recombinant DNA and other techniques were performed as described (Sambrook et al., 1989). Open reading frame (ORF) and blast homology searches were performed using the NCBI Internet server (http://www.ncbi.nlm.nih.gov). Molecular biology reagents were purchased from New England Biolabs (NEB), USA.

The bacterial strain A909 (WT) is a Type Ia capsular polysaccharide clinical isolate of GBS (Madoff et al., 1991). LR113 is a derivative of A909 and has a ΩKm-2 insertion cassette within the coding region [amino acid 13 (aa13)] of stk1; this strain is Stk1-deficient, as demonstrated previously (Rajagopal et al., 2003). LR114 is an isogenic A909 strain that also has the ΩKm-2 cassette at aa13 of stk1, identical to LR113. In addition, the coding sequence of the gene encoding stp1 was allelically replaced with a gene conferring resistance to chloramphenicol in LR114. Our previous results have confirmed that the LR113 and LR114 strains are deficient for Stk1 and both Stp1 and Stk1 respectively (Rajagopal et al., 2003). Given that the phenotypes of LR113 and LR114 are similar if not identical from our previous observations (Rajagopal et al., 2003) and in this report (see Results), we have often referred to LR113 and LR114 as Stk1-deficient strains.

Escherichia coli was cultured in lysogeny broth (LB; Bertani, 2004). Routine cultures of S. agalactiae (group B streptococci, GBS) strains were performed in THB (Difco Laboratories) in 5% CO₂ at 37°C. Chemically defined medium [CDM (Willett and Morse, 1966), herein called complete CDM (cCDM), contains adenine, xanthine, guanine and uracil at a final concentration of 0.1 mM] or dCDM (deleted CDM, i.e. devoid of purines or pyrimidines) were used to assess auxotrophic requirements. Amino acid deficiencies of GBS strains were examined using pools of cCDM not containing any one of the 20 amino acids, as described (Samen et al., 2004). To evaluate nucleotide biosynthesis, cell growth was monitored in dCDM as well as dCDM containing various purine sources either in the presence or in the absence of pyrimidines. Evaluation of growth in cCDM, dCDM or respective pools was performed as described previously (Shelver et al., 2003). Briefly, GBS was grown to log phase (OD₆₀₀ 0.3) in THB. Subsequently, the cells were harvested by centrifugation, washed twice with an equal volume of sdH20 and diluted 1–50 in dCDM, cCDM or respective pools. Cell growth was monitored at 600 nm after overnight incubation in 5% CO₂ at 37°C. For consistency in all assays (see below), GBS were grown in various media to an OD₆₀₀ of 0.3.

Antibiotics were added at the following concentrations when necessary: spectinomycin 300 μg ml⁻¹ for GBS and 50 μg ml⁻¹ for E. coli; kanamycin 1000 μg ml⁻¹ for GBS and 50 μg ml⁻¹ for E. coli, chloramphenicol 5 μg ml⁻¹ for GBS and 10 μg ml⁻¹ for E. coli. AICA and AICAR were included in dCDM up to a concentration of 1 mM.

Inorganic pyrophosphate (PP_i) levels

To quantify intracellular levels of PP_i, GBS strains were grown in either dCDM or cCDM to an OD₆₀₀ of 0.3. PP_i was extracted from these samples as described for extraction of nucleotides (see below). Subsequent to freeze-drying, the samples were resuspended in 400 μl of sdH₂O and equal amounts were used for quantification of PP_i, using the pyrophosphate reagent kit from Sigma-Aldrich, USA. PP_i assays were performed as described by the manufacturer. All experiments were performed in triplicate and repeated for reproducibility.

Sensitivity to analogues

Purine analogues were used at the following final concentration: 6-MP, 2 mM; MA, 0.3 mM and 2-FA, 0.25 mM. To estimate percentage survival, GBS were grown in THB to an OD₆₀₀ of 0.3; cells were harvested and washed twice and resuspended to the original volume using sdH₂O. Serial dilutions of each strain were plated on cCDM, dCDM or dCDM containing various purine sources (see Results), in the presence and absence of the analogue. The plates were incubated at 37°C for 24–48 h. Per cent (%) survival, for any given media and inhibitor, was calculated as follows:

(Total cfu in the presence of the analogue)/(total cfu in the absence of the analogue) × 100.

Intracellular nucleotide pools

Intracellular nucleotide pools were extracted as described below. Briefly, GBS were grown in 100 ml of cCDM, dCDM or dCDM containing uracil and various purine sources to an OD₆₀₀ of 0.3. The cells were washed in an equal volume of sdH₂O and the cell pellet was resuspended in 10 ml of ice-cold 0.2 N HPLC grade formic acid (Fisher). Nucleotides were extracted at 4°C for 2 h and centrifuged for 10 min at 6000 g. Supernatants were filtered using a formic acid compatible 0.22 μm pore-size filter (Millipore, USA) and freeze-dried using a FreeZone® 4.5 system (Labconco, MO, USA) that had a dry ice-ethanol solvent trap to compensate for the low eutectic temperature of formic acid containing samples. The free-dried sample was then resuspended in 400 μl of buffer A (5 mM NH₄H₂PO₄ pH 2.8). Intracellular nucleotide concentrations were determined by HPLC using a Hypersil SAX column (4.6 × 250 mm, Thermo Hypersil-Keystone, Bellefonte, PA) equipped with a guard column. A Shimadzu LC-10 HPLC system that contained a SIL-10AD_VP autoinjector and a SIL-10AD_VP UV-Vis detector was used. Twenty microlitres of sample were injected into the column, which was pre-equilibrated with buffer A. Nucleotides were eluted using a discontinuous gradient of 5 mM NH₄H₂PO₄ pH 2.8 (buffer A) and 750 mM NH₄H₂PO₄ pH 3.7 (buffer B) as per manufacturer's instructions. The gradient employed was an increase in buffer B, from 0% to 30% over 25 min and 30% to 100% buffer B, over the next 10 min at a constant flow rate of 1.0 ml min⁻¹. The absorbance was recorded at 254 nm and the peaks were integrated using the software supplied with the system, i.e. Shimadzu EZ Start Version 7.2.

Standard HPLC nucleotide markers for adenine, cytosine, guanine and uracil (mono-, di- and triphosphates) were purchased from Sigma-Aldrich, USA. Nucleotide standards were analysed and standard curves were obtained for individual experiments. The nucleotide concentrations were determined from the peak areas and standard curves, as described previously (Sigoillot et al., 2003). The corelation ‘OD₆₀₀ of 0.3’ corresponding to 0.06 mg (dry weight) ml⁻¹ was used to express concentration of the nucleotides in picomoles per milligram (dry weight).

Purine enzyme activity assays

For measurements of purine enzyme assays, bacterial extracts were prepared as previously described (Jensen, 1979) with minor modifications. GBS strains were grown in 500 ml of dCDM to an OD₆₀₀ of 0.3, cells were harvested and washed in 0.9% NaCl, and resuspended in 10 ml of 0.1 M Tris HCl, 2 mM EDTA, pH 8.0, as previously described (Jensen, 1979). The cells were then disrupted by sonication and centrifuged at 6000 g for 10 min, to remove unlysed cells and cell debris. Total protein concentration in the cell extracts was estimated using the Bradford method and equal amounts of protein was used in individual experiments. All experiments were performed in triplicate and repeated for reproducibility.

A DU® 800 UV-Vis spectrophotometer (Beckman Coulter, USA) was used for all measurements. All assays were carried out at 37°C for 30 min unless mentioned otherwise. In all cases, 100 μg of total protein was used for activity measurements. Also, to confirm the dose dependency of the different assays, varying concentrations of total protein was used (data not shown). One unit is defined as the amount of enzyme that utilizes or produces 1 nmol of substrate or product per min respectively (Jensen, 1979). Specific activities are denoted as units per milligram total protein.

IMPDH (GuaB) was measured as previously described (Mangasanik, 1963). This assay measures the rate of formation of NADH in the presence of IMP, which is monitored as an increase in absorbance at 340 nm (A₃₄₀). Simultaneous control reactions lacking IMP were always run for each strain and the rate of IMP-independent NADH formation from NAD⁺ was subtracted from the total rate to obtain IMP-dependent reaction rate as mentioned (Jensen, 1979). Further, controls that contained all components of the reaction except the bacterial extract was included and differences in A₃₄₀, if observed, was also subtracted from the total rate. A standard curve for different concentrations of NADH was used to corelate A₃₄₀ with NADH concentration.

Adenylosuccinate synthetase (PurA) was assayed as described (Juang et al., 1993). This assay measures the rate of formation of sAMP in the presence of aspartate. Activity was measured as an increase in absorbance at 280 nm (A₂₈₀). Control reactions that lacked either bacterial extract or aspartate were simultaneously run for each strain. Specific activity was calculated using the molar extinction coefficient of 11 700 M⁻¹ cm⁻¹ as described (Juang et al., 1993).

GMP reductase (GuaC) measurements were performed as described (Garber et al., 1980). Appropriate controls and blanks were also included and the rate of NADPH oxidation to NADP was continuously monitored at 340 nm. An extinction coefficient of 6.2 × 10³ was used to calculate specific activity also as previously described (Garber et al., 1980).

GMP synthetase (GuaA) was measured as described (Jensen, 1979). The change in absorbance at 290 (A₂₉₀), reflective of GMP produced per minute in the reaction, was recorded. Specific activities were determined as described previously (Jensen, 1979). Controls included blank reactions, i.e. no bacterial extract or reactions lacking XMP were included and the rate of change observed, if any, were subtracted from the total rate, as described (Jensen, 1979).

Construction of C-terminal His-tag constructs and purification of recombinant proteins

A derivative of pET32a called pET32CK that lacks all the N-terminal fusion tags (i.e. the thioredoxin tag, S-tag and N-terminal his-tag) was used (kind gift from Drs R. Seepersaud and A. Jones, unpublished data). The pET32CK vector contains a C-terminal his-tag and was used to clone and purify C-terminal his-tag GuaB and PurA fusion proteins as follows. DNA fragments containing guaB and purA were amplified from WT A909 chromosomal DNA using high-fidelity polymerase chain reaction (PCR). The following primer pairs guaBF and guaBR and purAF and purAR were used for amplification of guaB and purA respectively. The PCR products were digested with the enzymes for which restriction sites were engineered in the primers, cloned in frame into the multiple cloning site (MCS) of pET32CK to obtain C-terminal His₆ fusion proteins respectively. The plasmid pAV55 encodes GuaB–His₆ and plasmid pAV56 encodes PurA–His₆ fusion proteins on IPTG induction respectively. His-tag fusion proteins were purified using a Ni²⁺ column according to manufacturer's instructions (Novagen, USA).

Reversible phosphorylation

In vitro reversible phosphorylation assays were carried out in the presence of GST–Stk1 (recombinant Stk1, rStk1) fusion protein and recombinant Stp1 (rStp1, also a GST fusion) as described previously (Rajagopal et al., 2003). His-Tag fusion proteins were dialysed to remove imidazole and protein concentrations were estimated using the Bradford's method. [γ-³²P]-ATP (10 μCi) was added to 1 μg of purified his-tag fusion protein (i.e. GuaB or PurA), in the presence or absence of 1 μg of rStk1 in kinase reaction buffer (Rajagopal et al., 2003). The reaction was incubated at 37°C for 15 min, stopped by the addition of SDS-PAGE sample buffer and analysed on SDS-PAGE gels followed by autoradiography. To test dephosphorylation of the his-tag fusion protein by rStp1, the reactions containing autophosphorylated rStk1 and PurA–His₆ fusion protein was purified using a G25 sephadex spin column to remove unincorporated nucleotides. MnCl₂ was added at the optimal concentration for rStp1 activity (3 mM; see Rajagopal et al., 2003) in the presence of 1 μg of rStp1. Dephosphorylation reaction was carried out for 30 min and analysed on a 10% SDS-PAGE, following which the gels were stained with coomassie and subjected to autoradiography.

Time-course dephosphorylation

For time-course dephosphorylation, in vitro phosphorylation was performed with reactions containing 5 μg of Stk1 and 5 μg of PurA–His₆ in the presence of kinase buffer and 50 μCi [γ-³²P]-ATP for 15 min. Subsequently, unincorporated [γ-³²P]-ATP was removed and the reaction products were purified as described above. To monitor time-course dephosphorylation of rStk1 and PurA–His₆ by rStp1, MnCl₂ was added at a final concentration of 3 mM and 2 μg of rStp1 was added to the reaction. Aliquots were removed at various time intervals including time zero (that is reflective of phosphorylated proteins present in the reaction) and the reactions were terminated by the immediate addition of SDS-PAGE sample buffer. All samples were analysed on an SDS-PAGE followed by autoradiography.

Effect of phosphorylation on PurA activity

For analysis of the role of kinase-dependent phosphorylation on PurA activity, we first performed in vitro phosphorylation (as described above) of PurA in the presence of rStk1 (1:5 molar ratio) and 1 μM ATP. A molar excess of rStk1 was used to ensure that PurA–His₆ was in the phosphorylated state, as described for analysis of IMPDH activity (Ingley and Hemmings, 2000). Controls included reactions that did not contain either ATP or the rStk1 protein. Subsequently, all reactions were individually purified on G25 sephadex spin columns. Adenylosuccinate synthetase (PurA) assays were performed as described above and in Juang et al. (1993). Activity was measured as an increase in absorbance at 280 nm (A₂₈₀) over the linear period of the assay (60 min). Control reactions that lacked aspartate were simultaneously run for each of the reactions mentioned above, i.e. phosphorylated and unphosphorylated PurA. Specific activity was calculated using the molar extinction coefficient of 11 700 M⁻¹ cm⁻¹ as described (Juang et al., 1993).

Construction of the P_xyltet-inducible vector, pLR16T

The plasmid vector pLR16T was derived from a GBS complementation vector, pLZ12spec (Husmann et al., 1995). To construct an inducible expression system in GBS, we adapted the tetracycline-inducible expression system in pYJ335 for use in GBS. Plasmid pYJ335 has been extensively used as an inducible expression vector in the Gram-positive S. aureus (Ji et al., 1999). A 1.8 kb SalI fragment of pYJ335 that included the tetR gene (tet repressor), its promoter (P_R), the strong xyl/tet promoter–operator fusion and a promoterless chloramphenicol acetyltransferase gene (cat) present downstream to the xyl/tet promoter and a transcriptional terminator (Ji et al., 1999) was gel-eluted, blunt-ended and ligated to EcoRI-digested and blunt-ended pLZ12spec to give pLR16. Subsequently, we sequenced the DNA in the 1.8 kb region cloned into pLR16 to enable us to engineer unique restriction sites.

To test whether the inducible xyl/tet promoter was functional in GBS, pLR16 was electroporated into GBS. We observed a dose-dependent increase in Cm^R colonies in the presence of increasing concentrations of the inducer, Atc. In the absence of the inducer, a few Cm^R colonies (i.e. 20 cfu per 100 ng of plasmid DNA) were obtained. This indicated a low level of expression from the P_xyl/tetO promoter even in the absence of the inducer, as mentioned previously (Geissendorfer and Hillen, 1990). However, a 50- and 100-fold increase in Cm^R colonies was obtained in the presence of 0.5 μg ml⁻¹ Atc and 1.0 μg ml⁻¹ Atc respectively. These data confirmed that the tet-inducible system was functional in GBS. As mentioned previously, Atc is not bactericidal and likewise we observed no adverse effect of Atc on growth of GBS.

The cat gene was eliminated from pLR16 by inverse PCR using the primers pLR16F and pLR16R, such that only the ATG start codon of cat, downstream to P_xyl/tetO, was retained. This plasmid was called pLR16ATG and contains a functional ribosome binding site (RBS) but lacked both the coding sequence for cat and the transcriptional terminator. Subsequently, we inserted a transcriptional terminator (Ω_R) that is functional in GBS (Chaffin and Rubens, 1998); Ω_R was PCR amplified from pCIV2 (Okada et al., 1993) using the primers omttbgl2 and omttsphI and was cloned downstream to the tetR/P_R/P_xyl/tetO-ATG start codon in pLR16ATG. The forward primer, omttbgl2, was engineered to provide an MCS, downstream to the tetR/P_R/P_xyl/tetO -ATG start codon and upstream to Ω_R and the resulting plasmid was called pLR16T. The region in pLR16T can be represented as: tetR/P_R/P_xyl/tetO-ATG-MCS-Ω_R and was sequenced to confirm that there were no errors during the various cloning steps.

Construction of complementation constructs

DNA fragments containing either stp1, stk1 and both stp1 and stk1 was amplified from WT A909 chromosomal DNA using high-fidelity PCR. The following primer pairs StpF and StpR, Stk1F and Stk1R, and Stp1F and Stk1R were used for amplification of stp1, stk1 and both stp1 and stk1 respectively. The PCR products were digested with AscI and Bgl2 or AscI and SalI for the stp1–stk1 fusion and cloned in frame into the MCS of pLR16T that was previously digested with the respective enzymes. The plasmid pAV51 encodes the WT stp1 allele, pAV52 encodes WT stk1 allele and plasmid pAV53 encodes WT stp1 and stk1 alleles. These plasmids were electroporated into A909, LR113 and LR114, as described previously (Framson et al., 1997). The vector pLR16T was also introduced into each of the mutant strains as controls. Cell growth was monitored in dCDM or cCDM containing spectinomycin, as described above. Growth of all strains was compared relative to growth of the WT derivative, i.e. A909/pLR16T (OD₆₀₀ = 1.0).