The 1.8-kbp DNA fragment from E. coli K12 genomic DNA including ygjG and aer-ygjG intergenic region (Fig. 1A) was cloned into pUC18 vector and sequenced. The obtained plasmid was designated as pUC18-ygjG. The computer analysis of the ygjG upstream regulatory region revealed recognizable consensus sequence of σ⁵⁴-dependent promoter with score 70.7 21. The localization of this promoter is shown on Fig. 1. Likewise, according to DNA array data, the nitrogen limitation results in 3- to 5-fold increase in levels of ygjG transcripts 22. For more detail characterization of the role of the 5'-flanking region in the ygjG expression and nitrogen starvation induction we compared the expression levels of three different ygjG upstream regions fused to the lacZ reporter gene in plasmid pBRP 23 under different growth conditions. pBR-σ⁵⁴-lacZ-1 and pBR-σ⁵⁴-lacZ-2 plasmids carried 343-bp KpnI-AccIII and 56-bp BglI-AccIII fragments, respectively, containing predicted σ⁵⁴ promoter sequence, and pBR-lacZ-3 plasmid harbored 282-bp KpnI-BglI fragment upstream of σ⁵⁴ promoter. E. coli TG1 cells were transformed by each of plasmids and β-galactosidase productions in three media: LB broth and salt M9 media 24 with ammonium or supplemented with proline as a sole nitrogen source were analyzed (Table 1).

The all three recombinant strains were found to have low β-galactosidase activity (below 200 Miller unites 24) in LB broth. The strain harboring pBR-lacZ-3 also revealed low levels of expression in all media, pointing that the 282-bp fragment doesn't include noticeable promoter sequence. The strain carried plasmid pBR-σ⁵⁴-lacZ-2 containing practically σ⁵⁴ sequence alone, showed 3 – 5 times higher lacZ expression levels than pBR-lacZ-3 in both variant of minimal media. But lacZ activity was dramatically increased in the strain harboring pBR-σ⁵⁴-lacZ-1, especially under nitrogen limited conditions. The increase was about 9-fold in M9 medium with ammonium and 28-fold in M9 with proline in comparison with pBR-σ⁵⁴-lacZ-2. Thus, the 56-bp region contains promoter sequence, and we concluded that σ⁵⁴ promoter localized in the BglI-AccIII fragment is the main promoter of ygjG. The KpnI-BglI fragment is strongly essential for ygjG expression in minimal media and activation of ygjG expression (up to 6-fold) under nitrogen limitation. Because enhancer proteins are absolutely necessary for σ⁵⁴-dependent transcription 25 we reanalyzed the sequence of this fragment and found some putative binding sites for E. coli activators: two general nitrogen regulatory protein (NR_I) binding sites and binding site for integration host factor (IHF) (Fig. 1B). In addition, existence of factor of inversion stimulation (FIS) binding site was suggested http://www.ncbi.nlm.nih.gov:80/. Thus the activation of ygjG expression is complicated, depends on growth conditions and needs further investigations.

TG1 cells harboring pUC18-ygjG were grown in three media (see above), the total RNAs were isolated and used to identify the transcriptional initiation site of ygjG mRNA by primer extension experiment. The same cDNA fragment was observed as the major transcript for all three RNA preparations, and the cytidine (C) was identified as the first mRNA base (Fig. 2, also shown as "+1" in Fig. 1). The quantities of ygjG cDNA synthesized were estimated using Typhoon imager. The highest level of transcript was observed for RNA preparation from cells grown under nitrogen limitation, while only about 20% and less than 1% of it were detected if cells were grown in M9 and LB media, accordingly. Also, crude protein extracts from these cells were examined for PATase activity with putrescine as amino group donor and with 2-OG as acceptor. Enzyme activity was determined routinely by glutamate formation, which was measured by capillary zone electrophoresis (CE) (Fig. 3). The second product of the reaction, γ-aminobutyraldehyde, was unstable 16 and formed cycled Δ¹-pyrroline at alkaline pH. PATase activity was highest in extracts of cells grown under nitrogen limitation (47 nmol·min^-1·mg^-1). It was significantly reduced in cell extracts from M9 medium (nearly proportionally to decreasing of ygjG transcript level), and was practically undetectable in cell extracts from LB medium (near to the limit of accuracy for the CE method). Thus, the level of ygjG transcription and PATase activity correlated in cells grown in each of three media used.

The predicted AUG initiation codon for the ygjG gene of 1290 bp (denoted downstream as ORF1) which codes 429-amino acid polypeptide (SWISS-PROT Database, P42588) is situated 126 nucleotides downstream of the mRNA start. Analysis of the nucleotide sequence revealed the other ORF (denoted as ORF2) encoding 459-amino acid protein with UUG initiation codon (Fig. 1B), which is 30 amino acids upstream from an in frame ORF1 start codon. Because both putative ORFs have the potential Shine-Dalgarno (SD) sequence 26, we decided to express each of them as his₆-tag fused recombinant protein to determine if these recombinant constructs actually yield a catalytically active enzyme.

The 1.29-kbp (ORF1) and 1.38-kbp (ORF2) DNA fragments of pUC18-ygjG were cloned in translation fusion with N-terminal his₆-tag sequence under the control of the T7 promoter in pET15b(+) vector. The resulting plasmids, pET-Ht-ORF1 and pET-Ht-ORF2, were expressed in E. coli BL21(DE3) cells. SDS-PAGE of cellular extracts demonstrated a high level of expression of these proteins. To elucidate which of polypeptides encodes an active enzyme, crude extracts were examined for PATase activity. The essential level of glutamate formation activity was observed in crude cell extracts of the BL21(DE3) harboring pET-Ht-ORF2. It was equal to 2.16 μmol·min^-1·mg^-1. The BL21(DE3)(pET-Ht-ORF1) strain exhibited only traces of PATase activity. The BL21(DE3) strain transformed with pET15b(+) as a control revealed no activity. Obviously, the UUG is the real initiation codon of the ygjG encoding 459-amino acid protein with a calculated molecular mass (M_r) of 49,644 Da.

In order to confirm the identity of ORF2 as the gene encoding PATase, we purified recombinant protein from the BL21(DE3)(pET-Ht-ORF2) strain by immobilized-metal-affinity chromatography (IMAC) to give a final preparation having a specific activity of 11.68 μmol·min^-1·mg^-1 with yield of 81%. The purity of Ht-YgjG recombinant protein was about 95%. SDS-PAGE showed one distinctive band with a relative M_r of about 52 kDa. This value well corresponds to the sum of mass of 49.6 kDa for ORF2 coding protein and 2.1 kDa for his₆-tag leader peptide.

Some parameters of PATase reaction were determined. The enzyme was active at the alkaline pH, with the maximum activity at pH 9.0 in Tris-HCl buffer and displayed a broad temperature optimum between 20°C and 80°C with maximum activity at 60°C. The K_m values for putrescine and 2-OG were determined to be 9.2 mM and 19.0 mM, respectively.

Substrate specificity of purified Ht-YgjG was measured with different donors and acceptors of amino group (Table 2). Several compounds (putrescine, cadaverine, spermidine, agmatine, and ornithine) were tested for their ability to function as amino donors for 2-OG. Among them putrescine was found to be the best amino group donor for the aminotransferase activity of Ht-YgjG. Slightly lower activity with cadaverine (97%) and significantly lower with spermidine (32%) was detected. Also, three keto acids: 2-OG, α-ketobutyrate and pyruvate were used as amino acceptors for putrescine. The enzyme did exhibit significant activity toward α-ketobutyrate (38%) and, to less extent, pyruvate (12%) with putrescine. Almost no activity was detected with ornithine as amino acceptor, while ornithine aminotransferase activity was annotated for ygjG in GeneBank Database. GABA, diaminopimelic acid, acetylornithine were completely inactive as amino group donors.

E. coli strains TG1(K12 supE hsdΔ5 thi Δ (lac-proAB) F'[traD36 proAB⁺ lacI^qlacZΔM15]) and BL21(DE3)(B F^- dcm ompT hsdS(r_B- m_B-) gal λ (DE3)) were used as a recipients, MG1655 (K12 λ^-, F^-) 31 was used for gene amplification. Plasmids pUC18 (Fermentas), pBRP 23 and pET15b(+) (Novagen) were used as a vectors for cloning.

The E. coli K12 ygjG gene with upstream aer-ygjG intergenic region was amplified from MG1655 genomic DNA using two oligonucleotides 5'-TTGGATCCGATTAATTTGATTTAGATCGCA-3' and 5'-TTCTGCAGCCTGCGGGCGTACGCGTCG-3' as primers. The 1,8 kb PCR product was digested by BamHI and PstI and cloned into pUC18 vector, yielding pUC18-ygjG plasmid. The insertion in pUC18-ygjG was sequenced for frame verification.

The 338-bp KpnI-AccI, 56-bp BglI-AccIII and 282-bp KpnI-BglI fragments were excised from pUC18-ygjG, fused with lacZ reporter gene and cloned into pBRP vector (pBR322-based plasmid containing pUC18 polylinker region), yielding pBR-σ⁵⁴-lacZ-1, pBR-σ⁵⁴-lacZ-2 and pBR-lacZ-3 plasmids, respectively.

The ygjG ORF1 and ORF2 sequences were amplified by PCR from pUC18-ygjG using phosphorilated downstream primers 5'-CCGATATCATGAAAGCACTTAACCGAGAG-3' and 5'-CCGATATCTTGAACAGGTTACCTTCGAG-3', respectively, designed so as to provide translation fusion between his₆-tag leader sequence and the gene of interest, and upstream primer 5'-TTCTGCAGCCTGCGGGCGTACGCGTCG-3'. PCR products were cloned into blunted vector pET15b(+)/NdeI-BamHI to construct plasmids pET-Ht-ORF1 and pET-Ht-ORF2, respectively, and insertions were sequenced.

Strains were routinely grown in LB broth or M9 medium 24. Nitrogen limited conditions were provided using M9 medium lacking ammonium and supplemented with 2 mM of a proline as nitrogen source. All media contained 100 μg/ml ampiciline. Strains were grown aerobically at 37°C.

For expression of lacZ fusions, E. coli TG1 harboring pBR-σ⁵⁴-lacZ-1, pBR-σ⁵⁴-lacZ-2 or pBR-lacZ-3 plasmids were grown overnight in M9 medium, washed by 15 mM NaCl, diluted to optical density at 600 nm (OD₆₀₀) of 0.07 by LB, M9 or M9 nitrogen limited medium. The resulting cultures were grown until reaching on OD₆₀₀ of about 3.0 (stationary phase).

The synthesis of his₆-tag fused recombinant proteins was induced in BL21(DE3) harboring pET-Ht-ORF1 or pET-Ht-ORF2 plasmid. When the cell density in LB medium had reached OD₆₀₀ of 1.0, 1 mM isopropyl-β-D-thiogalactopyranoside was added followed by 2 h incubation.

For preparation of cell-free extracts bacteria were harvested by centrifugation and washed twice with TE buffer. The cell pellets were stored frozen for several days at -70°C without significant loss of enzyme activity. Frozen cells were thawed, suspended to 0,025 g (wet weight) per ml in buffer A (20 mM potassium phosphate, pH 7.4, 1 mM EDTA, 1 mM dithiothreitol (DTT), 1 mM phenylmethylsulfonyl fluoride (PMSF), 10 μM PLP), and disrupted by sonication followed by centrifugation to remove debris.

The TG1 harboring pUC18-ygjG has been grown in LB, M9 media and under nitrogen limited conditions, and the total RNAs were isolated using RNAsy MiniKit (Sigma) according to the manufacturer's recommendations.

Oligonucleotide 5'-CGCTTCTCAATGAGATTCAGGGCGTGG-3' which is complementary to the region from +44 to +71 relatively to the ygjG ORF2 initiation codon was radiolabeled by T4 polynucleotide kinase (Pharmacia) with γ-[³³P]ATP (6000 Ci/mmol).

Aliquots of mRNA (5 μg) were denaturated at 65°C for 10 min, incubated in reaction buffer (50 mM Tris-HCl, pH 8.3, 75 mM KCl, 3 mM MgCl₂, 20 mM DTT) containing 0.5 ng of the labeled oligonucleotide, rTth reverse transcriptase (5 u) (Perkin Elmer) and 1 mM deoxyribonucleotide triphosphates at 37°C for 1 h. The synthesized cDNAs were denaturated at 65°C for 10 min and analyzed by electrophoresis on 6% acrylamide gel. The same primer was used for the DNA sequence ladder run on the same gel. The quantities of synthesized cDNAs were estimated using Typhoon 9210 imager and ImageQuant version 5.2 software (Amersham, Molecular Dynamics).

The Ht-YgjG was purified from 150 ml of BL21(DE3)(pET-Ht-ORF2) induced culture. Frozen cells (60 mg) were thawed, suspended in 75 ml of buffer B (20 mM Tris-HCl, pH 8.0, 1 mM PMSF) and disrupted by sonication followed by centrifugation to remove debris (10 000 g, 20 min, 4°C). The imidazol and NaCl were added to supernatant up to final concentrations of 50 mM and 500 mM, respectively. The supernatant was applied to a His-Trap column (1 ml) (Pharmacia) equilibrated with the same buffer. The protein was eluted with 10 ml of buffer C (20 mM Tris-HCl, pH 8.0, 400 mM imidazol, 500 mM NaCl). The Ht-YgjG containing fractions were applied on Sephadex G-25 column (Pharmacia) equilibrated with buffer D (20 mM potassium phosphate, pH 7.4, 15% glycerol, 1 mM EDTA, 1 mM DTT, 1 mM PMSF, 10 μM PLP) and eluted in the same buffer. The Ht-YgjG containing fractions were combined, aliquoted and stored at -70°C until required.

The protein concentration was estimated following the method described by Bradford with bovine serum albumin as the standard 32. Proteins were separated by SDS-PAGE (10%) 33.

β-Galactosidase assay was performed in cell-free extracts of TG1 cells harboring pBR-σ⁵⁴-lacZ-1, pBR-σ⁵⁴-lacZ-2 and pBR-lacZ-3 plasmids according to Miller 24.

PATase activity was routinely assayed as the formation of L-glutamic acid from putrescine and 2-OG at 37°C. The assay mixture contained (in 0.1 ml of total volume) 100 mM Tris-HCl, pH 9.0, 25 μM PLP, 10 mM putrescine, 10 mM 2-OG. The reaction mixture was preincubated at 37°C for 5 min, and the reaction was started by the addition of cell extract or enzyme solution (5 μl). The reaction was stopped by adding 15 μl of HCl (10%). The analysis of glutamic acid formation was carried out using a Quanta 4000E Capillary Electrophoresis System (Waters) with an uncoated fused-silica capillary (75 μm inner diam. × 60 cm) at 25 kV potential. The injection was performed hydrostatically for 25 s. The separation buffer consisted of 50 mM Tris-base, 25 mM benzoic acid, pH 8.5, 0.25 mM tetradecyl-trimethyl-ammonium bromide (for indirect UV detection at 254 nm). The calibration at three concentrations of the L-glutamic acid (0.025 mM, 0.05 mM, 0.1 mM) and three concentrations of the 2-OG (0.05 mM, 0.1 mM, 0.2 mM) was carried out.

The kinetic constants of the Ht-YgjG were evaluated by varying the concentration of putrescine at a 2-OG concentration of 20 mM or by varying the concentration of 2-OG at a putrescine concentration of 15 mM. The enzyme kinetics data were fitted to Michaelis-Menten kinetics and K_m values were calculated. The substrate specificity of Ht-YgjG was determined using assay as described above with different polyamines and amino acids as amino group donors and 2-OG as acceptor, or with different acceptors and putrescine as amino group donor.