INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

There is an 11-step pathway for the de novo synthesis of IMP in Bacillus subtilis (22) and Escherichia coli (23). IMP is a branch point for the synthesis of AMP in two steps and the synthesis of GMP in two steps. A 12-gene pur operon encodes the enzymes required for the synthesis of IMP in B. subtilis (1). Two genes, purA and purB, are required to convert IMP into AMP. The purA gene is unlinked to the pur operon, while purB is in the operon. The purA gene encodes adenylosuccinate synthetase, and adenylosuccinate lyase is the product of purB. Purine biosynthesis is feedback regulated by end products of the pathway, and production of adenine and guanine nucleotides is balanced by regulation of the AMP and GMP branches. Expression of the pur operon is subject to dual regulation of transcription initiation and termination (1). The addition of adenine to cells results in the repression of transcription initiation, and the addition of guanosine promotes premature transcription termination in an mRNA leader region preceding the first structural gene. A purine repressor (PurR) mediates the regulation of transcription initiation. purR was cloned and overexpressed (20), and the protein was purified (19). The present studies were undertaken to assess the regulation of the branch from IMP to AMP. Earlier it was reported that adenine decreased the adenylosuccinate synthetase activity in B. subtilis and that guanosine increased the enzyme activity (17). Here we present evidence that adenine mediates a PurR-dependent repression of purA transcription and that guanosine leads to an upregulation of transcription. The repression by PurR is dependent upon a second protein, YabJ, which is homologous to a group of proteins of unknown function. On the other hand, the activation by guanosine is dependent upon an interaction of PRPP with PurR but not upon YabJ.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and vectors. The bacterial strains and vectors used in this study are listed in Table 1. Since the purA promoter is lethal to E. coli in high-copy-number plasmids (12), strain KE94 was used for the propagation of plasmids pPAL1 and pPAL3. Plasmid pPAL4 contained an inactivated purA promoter and could therefore be amplified in DH5 F'. Strain XL2 Blue was used to propagate vectors containing the purR gene. The B. subtilis transformants were selected on either Penassay broth agar (Difco) or Luria-Bertani agar plates containing either chloroamphenicol or neomycin at 5 µg/ml.

Construction and integration of a purA'-lucGR fusion. A purA'-lucGR fusion in plasmid pPAL1 was constructed in two steps. First, lacZ in pCATZ1 (2) was excised and replaced with the click beetle lucGR gene (21) from plasmid pCSS962 (9). Second, a fragment containing the 5' end of purA from nucleotides 419 to +66 (relative to the start of transcription at +1) (12) was amplified by PCR from B. subtilis DE1 chromosomal DNA and inserted into the polylinker sequence immediately upstream of the lucGR gene. In this construction, codon 7 of purA is followed by 15 nucleotides of plasmid polylinker. Translation of purA terminates at a TGA in the polylinker. Translation of lucGR is expected to start at the initiator ATG which overlaps the purA TGA stop as follows: AA. The nucleotide sequence of the purA'-lucGR junction was verified. The resulting plasmid with the purA'-lucGR transcriptional fusion was named pPAL1. Plasmid pPAL1 was integrated into the chromosome of B. subtilis DE1 by homologous recombination as shown in Fig. 1.

View larger version (13K): [in this window] [in a new window]   FIG. 1.   Integration of pPAL1 into the B. subtilis chromosome. Vector sequences are shown by a dashed line, and purA upstream sequence is shown by a solid line.

Construction and integration of mutations in the purA control site and in the purR locus. The purA gene contains an EcoRV site at nucleotide 174 (12). An MluI site at 55 was introduced into the purA control region in pPAL1 by site-directed mutagenesis. The resulting plasmid was digested with EcoRV and MluI, the 5' cohesive end was filled by Klenow fragment, and a SmaI Nm^r cartridge from pBEST 501 (6) was added to produce pPAL3. Plasmid pPAL3 thus contains a 174 to 55 cis-control-site deletion upstream of the purA'-lucGR fusion. pPAL3 was integrated into the chromosome of B. subtilis PAL1 by homologous recombination.

PRPP inhibition of PurR binding to DNA. Plasmid pR6H was used for overexpression and hyperproduction of PurR containing a C-terminal His tag (19). The two PurR PRPP binding site mutants, D203A and D204A, were hyperproduced from plasmids pR6H3A and pR6H4A, respectively, by the same procedure as for the wild type. The proteins were purified by using an Ni²⁺ affinity resin as described previously (19). Binding of the wild type and the two mutants to a DNA fragment containing the purA control region (163 to +47) was determined as described previously (19). The 20-µl binding mixture contained 10 mM HEPES (pH 7.6), 100 mM KCl, 1 mM EDTA, 5 mM MgCl₂, 10 fmol of ³²P-labelled DNA fragment, 5 ng of purified protein, and varied concentrations of PRPP. Free DNA and PurR-DNA complexes were separated by electrophoresis on agarose gels and quantitated by counting radioactivity with a Packard instant imager.

Primer extension mapping. The B. subtilis DE1 cells were grown in 200 ml of minimal medium (1) supplemented with 1 mM guanosine to an optical density of 650 nm (OD₆₅₀) of 0.7. The cells were poured onto ice chilled to 20°C and collected by centrifugation for 10 min at 3,000 × g at 2°C. The total RNA was isolated by the chromosomal DNA isolation method for gram-positive bacteria (15) with slight modifications and additions as necessary. The cells were lysed with lysozyme (0.5 mg/ml) in the presence of 30 U of RNasin/ml (Promega) for 30 min, followed by proteinase K treatment in 1% sodium dodecyl sulfate at 55°C for 2 h. RNA was extracted once with phenol-chloroform and once with chloroform and precipitated with isopropanol. RNA pellet was dissolved in water containing 30 U of RNasin/ml. The RNA solution was treated with RNase-free DNase and extracted with phenol-chloroform. RNA was precipitated with isopropanol and dissolved in water containing 30 U of RNasin/ml. When possible, all solutions were treated with diethylpyrocarbonate. The yield of RNA was approximately 5 mg, with an A₂₆₀/A₂₈₀ ratio of 1.7.

Luciferase assay. Expression of purA was determined by measuring luciferase activity in strains with purA-lucGR integrated into the chromosomal purA gene. The cultures for luciferase assay were grown in minimal medium (1) supplemented either with 1 mM adenine, 1 mM guanosine, or both adenine and guanosine (1 mM each) or with no added purine compounds. The medium contained either chloramphenicol (strains PAL1 and PAL4) or neomycin (other integrants) at 5 µg/ml. The 5-ml overnight cultures were centrifuged, and the pellets were suspended in 50 ml of minimal medium containing the respective supplements. The cultures were grown to an OD₆₅₀ of 0.5. From each culture, a 1-ml sample was taken, centrifuged, and resuspended in 5 ml of minimal medium containing the respective supplements. The 5-ml cultures were grown for 2 h. Samples of 1 ml were taken, and 100 µl of solution A (1 M K₂HPO₄ and 20 mM EDTA [pH 7.8]) was added to each sample. The final samples were frozen at 70°C. For the measurement of luciferase activity, the samples were thawed and centrifuged, and the supernatant was carefully removed. The cells were resuspended in 50 µl of the supernatant solution. One volume of cell lysis buffer 1 (Bio-Orbit) containing 2 mg of lysozyme/ml and 2 mg of bovine serum albumin (BSA)/ml was added, and the mixture was incubated at room temperature for 5 min. The lysate was centrifuged, and a 50-µl aliquot of the supernatant was sampled. One hundred microliters of Luciferin reagent and 100 µl of ATP reagent of GenGlow-100 kit (Bio-Orbit) were added to the extract, and the maximum light output was measured by using an LKB 1250 luminometer. The values were compared with a standard curve made by assaying known amounts of firefly luciferase standard in dilution buffer (1 mg of BSA/ml, 0.5× cell lysis buffer 1, 0.45× minimal medium, and 0.05× solution A). When necessary, the samples were diluted with dilution buffer. The results are expressed as attomoles of luciferase per 10⁸ CFU.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Integration of a purA'-lucGR fusion and purR locus mutations into the B. subtilis chromosome. A gene encoding luciferase was used as a reporter to monitor purA expression. To obtain a single-copy purA'-lucGR integrant, plasmid pPAL1 was recombined into the B. subtilis chromosome in strain DE1, as diagrammed in Fig. 1, to give strain PAL1. Strain PAL1 is purA⁺. Two purA'-lucGR derivatives, one with a promoter mutation (PAL4) and another with a cis-control-site deletion (PAL3), were constructed in a like manner. The integrations were verified by PCR amplification and DNA sequencing. Like strain PAL1, the two purA'-lucGR derivatives are purA⁺.

View larger version (19K): [in this window] [in a new window]   FIG. 2.   Gene arrangement of wild-type (PAL1) and various purR integrants. purR6H is a purR+ derivative with six His codons at the 3' end, followed by two translation stop codons. Truncation of the yabJ or the purR yabJ operon is marked by an apostrophe. The site of mutations D203A and D204A is marked by an arrow.

Regulation of purA. The level of purA expression was determined by using the lucGR gene as a reporter. In the purA'-lucGR transcriptional fusion, a DNA fragment containing nucleotides 419 to +66 of the purA operon was ligated immediately upstream from the lucGR gene. A basal luciferase activity of 498 amol per 10⁸ CFU was obtained from strain PAL1 (purA'-lucGR) grown in medium without added purines (Table 2). For a luciferase control, a 10 promoter mutation was incorporated into purA'-lucGR in strain PAL4. The 10 promoter mutation abolished luciferase activity (data not shown), indicating that all of the activity was derived from purA'-lucGR expression. To examine regulation by PurR, purR and O_purA operator mutations were incorporated into the purA'-lucGR reporter strain. The levels of luciferase activity in these strains are given in Table 2. In the purR⁺O⁺_purA wild type, there was 10-fold repression of purA expression by adenine and a 4.5-fold activation by guanosine. When adenine and guanosine were combined, repression by adenine overrode the activation by guanosine.

Primer extension mapping. To rule out the possibility that activation by guanosine in the wild-type cells is due to a shift of the transcription initiation site, the 5' end of the purA transcript in DE1 cells grown with excess guanosine was determined by primer extension mapping by using the same primer previously used for mapping the 5' end of purA mRNA (12). The length of primer extension product was 117 nucleotides (data not shown), the same length as that previously determined with the cells grown in the absence of purines.

Role of yabJ in the regulation of purA. To test the significance of the yabJ gene downstream from purR for regulation of purA, yabJ was disrupted to obtain strain N6H. The data in Table 3 show that regulation by purR6H in strain N6H was perturbed and is not similar to that by purR in PAL1 (Table 2). Repression of basal expression by adenine was abolished in N6H (purR6H 'yabJ), although the upregulation by guanosine was retained.

PRPP inhibition of PurR binding to DNA. A K_d of 7.9 µM was determined previously for the interaction of PurR with purA-control-site DNA (19). Given that PRPP inhibits the binding of PurR to the pur operon control site (20), a similar inhibition by PRPP was expected for PurR binding to purA. The data in Fig. 3 show PRPP inhibition of the PurR-purA operator DNA interaction. For 50% inhibition of binding, approximately 30 µM PRPP was required, but complete inhibition was not attained even at 2.5 mM PRPP (data not shown).

View larger version (17K): [in this window] [in a new window]   FIG. 3.   PRPP inhibition of PurR binding to the purA control region DNA. The effect of PRPP on binding is shown for wild-type PurR () and two mutants, PurR D203A () and D204A (). The fraction of DNA bound to PurR in the absence of PRPP (0.82 for wild type, 0.81 for D203A, and 0.89 for D204A) was normalized to 1.0. The apparent Kd values for binding to operator DNA for the wild type and the mutants were similar.

Regulation of purA by PurR PRPP binding mutants. The PurR D203A and D204A mutations were incorporated into the chromosome of strain PAL1 (purA'-lucGR) in order to determine the in vivo role of PRPP as a PurR effector. The two Asp mutants purR6H/D203A and purR6H/D204A, encoding repressors having a C-terminal His tag, were integrated into PAL1 to give yabJ mutant strains N3A (purR6H/D203A purA'-lucGR) and N4A (purR6H/D204A purA'-lucGR) (Fig. 2). The purR6H gene encodes a repressor with a His tag identical to that used for the in vitro PurR-purA operator DNA binding experiments (19) (Fig. 3). The results in Table 3 show that in the two PurR 'yabJ Asp mutants, strains N3A and N4A, repression by adenine was lost, as in the yabJ mutants with His-tagged or wild-type purR. In addition, the upregulation of guanosine was completely lost in both of the PurR PRPP binding mutants. It thus appears that the two regulatory events, repression by adenine and upregulation by guanosine, were separated in strain N6H. YabJ was required for repression by adenine but not for upregulation by guanosine. The PRPP effector site was necessary for upregulation by guanosine.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

It is not surprising that regulation of purA expression should contribute to controlling de novo AMP synthesis and to balancing the production of AMP with GMP. The data in Table 2 establish that purA expression, monitored by a luciferase reporter gene, is repressed by the addition of adenine to cells and is upregulated by added guanosine. Mutant analysis indicates that both repression and upregulation are dependent upon the interaction of PurR with the purA control region. Direct in vitro evidence for this binding has been reported previously (19). The data are consistent with the view that transcriptional regulation of purA by PurR contributes to controlling the production of AMP and to maintaining the balanced synthesis of adenine and guanine nucleotides.

The data reported in Table 3 have brought to light an important new aspect of PurR function. Repression of purA by PurR depends upon yabJ, an overlapping downstream gene. This gene encodes a member of a protein family with unknown function. Data in Tables 2 and 3 show that yabJ⁺ is needed for PurR-mediated purA repression by adenine although not for upregulation by guanosine. The yabJ mutation thus appears to uncouple the repression and activation functions of PurR.

How can PurR mediate repression by adenine and upregulation by guanosine, and what role might yabJ have? A working model shown in Fig. 4 explains these results. First, we consider three states of the purA control region as follows: (i) little or no PurR bound, giving maximal purA expression; (ii) partial occupation by PurR, giving basal expression; and (iii) saturation by PurR, resulting in full repression. purA expression in strains NMW (purR::neo purA'-lucGR) and PAL3 (O_purA1 purA'-lucGR) (Table 2) reflects the unoccupied state I control region that leads to high constitutive expression. In these mutants, there is no binding of PurR to the control region. State II, partial occupancy by PurR, giving basal expression, is seen in the wild-type grown without purines (Table 2). Partial occupancy of the purA control region results from the interplay between the endogenous purine compounds and the PRPP pool. PRPP, which inhibits PurR binding to the purA control region (Fig. 3), has been proposed to be the key regulatory molecule for de novo purine nucleotide synthesis in B. subtilis based on its exclusive ability to influence PurR-DNA binding in vitro (20). Adenine lowers the PRPP concentration in B. subtilis (17), thus allowing PurR to bind to the control region and shift the III equilibrium toward II. Guanosine, on the other hand, increases the cellular PRPP pool (17), promoting a shift toward state I. According to this model, YabJ is needed for the state II-to-state III conversion, in which additional PurR is bound and the control site is saturated with PurR. This YabJ-dependent step may correspond to the high PurR/control site binding stoichiometry detected in vitro at an elevated PurR concentration (19). Alternatively, YabJ could inhibit the ability of endogenous adenine to lower the intracellular PRPP level. In contrast to the state II-state III equilibrium, YabJ is not required for the state II-to-state I shift promoted by guanosine.

View larger version (12K): [in this window] [in a new window]   FIG. 4.   Hypothetical model explaining roles of YabJ and PRPP. PRPP is required to inhibit binding of PurR, leading to state I. The PRPP concentration in the cell is increased by Guo and decreased by Ade. Basal activity results from partial occupancy of the control region with PurR. Repression requires the binding of additional molecules of PurR (state III). YabJ is essential for this additional PurR binding, perhaps as a PurR-YabJ complex. The possible association of YabJ with PurR in state III is not addressed.

This model explains the loss of repression by adenine in the yabJ mutants N6H and NPR (Table 3). These strains exhibit basal expression when grown with excess adenine. Guanosine, however, increases the PRPP pool and shifts the equilibrium toward high state I activity. A mutation in the PurR PRPP site abolishes this PRPP-mediated shift in the yabJ mutant strains N3A and N4A (Table 3).

In an earlier report from one of our laboratories (20), Weng et al. stated that a yabJ disruption had no effect on the expression or regulation of purR or the pur operon. Unfortunately, this mutant was lost, and the result cannot be replicated. We assume that the earlier result was incorrect and that yabJ has a similar role in the regulation of purA, the pur operon, and purR.

What is YabJ and how does it work? The deduced amino acid sequence of YabJ (125 residues) is homologous to a group of 35 other proteins of unknown function from archaea, procaryotes, and eucaryotes. Some organisms (e.g., E. coli and yeast) encode several YabJ paralogs, whereas some have only one YabJ ortholog. In pairwise comparisons of these 35 sequences with YabJ, there is an identity of 21 to 53% over a span of 117 to 125 amino acids. A multiple alignment of all the sequences does not reveal any invariant residues, although approximately 10 conserved amino acids can be identified. YabJ belongs to the YER057c/YjgF protein family of unknown function (PROSITE accession no. PS01094). The family is defined by a conserved signature motif located at the C terminus of these proteins, consisting of the following amino acids: P-[AT]-R-[SA]-X-[LIVMY]-X₂-[AK]-X-L-P-X₄-[LIVM]-E. The consensus pattern is between amino acids 100 and 117 in YabJ. The YER057c/YjgF motif is conserved in all 36 YabJ homologs presently in sequence databases, although the degree of conservation varies to some extent.

The homologs from rat and human have 45 and 44% identity with YabJ, respectively. These proteins have been shown to inhibit cell-free protein synthesis at a high concentration (14, 18). Samuel et al. (16) found that the YabJ homolog Hrp12 from mouse had some similarity to heat shock proteins Hsp70 and Hsp90. They also noticed that the purified mouse Hrp12 could be phosphorylated in vitro with protein kinase C. Melloni et al. (13) reported that the bovine and goat YabJ homologs had the capacity to activate calpains. An aldR gene product from Lactococcus lactis (4), which is 52% identical to B. subtilis YabJ, has been suggested to interfere with branched-chain amino acid synthesis, although aldR does not encode an isoleucine biosynthetic enzyme. Mutations that allow thiamine synthesis in the absence of both PurF (glutamine phosphoribosylpyrophosphate amidotransferase) and the pentose phosphate pathway in Salmonella typhimurium have been localized in a gene encoding a YabJ homolog, YjgF. Moreover, in these mutants the isoleucine biosynthetic pathway seems to be affected as in L. lactis (3).

A common molecular function for YabJ and its 35 homologs is implied by their high degree of sequence identity and similar sizes. However, no certain biological function or common molecular function has emerged from studies of YabJ homologs. In some cases, as in L. lactis aldR and in the present study, the target seems to be related to the operon containing the gene for the YabJ homolog. Our work adds a new dimension to the question of the function of a YabJ homolog, which for the first time, can be located to a specific target. The effect of the disruption of YabJ on the regulation of transcription has been established by using a promoter-reporter system. Given the complex PurR-DNA interaction that has been studied in vitro (19), several possibilities exist for YabJ function, all of which require adenine dependence and guanosine independence in vivo. One possibility is that YabJ associates with PurR and promotes an interaction with DNA that is more stable than the interaction that is possible with PurR alone. A stoichiometry of two or six PurR dimers per pur operon was reported in studies with a DNA control-site fragment (19). Perhaps interaction of PurR and YabJ favors the higher binding stoichiometry and a more stable PurR-DNA interaction. Another possibility is that YabJ may stabilize PurR and increase its half-life. It has been speculated that some YabJ homologs may have a chaperone-like function (13). To test this hypothesis, it will be necessary to determine the intracellular level of PurR in yabJ⁺ and yabJ strains.

The results presented here provide important new information on the YER057c/YjgF protein family of unknown function. We have shown by using a promoter-reporter system that B. subtilis YabJ, a member of this family, affects the purine repressor-mediated regulation of purA. The established promoter-reporter system will be valuable for future in vivo studies of YabJ. In addition, a collaborative study of the three-dimensional structure of YabJ (19a) provides important clues about the function of YER057c/YjgF family members.