Thermophilic phosphoribosyltransferases Thermus thermophilus HB27 in nucleotide synthesis

Phosphoribosyltransferases are the tools that allow the synthesis of nucleotide analogues using multi-enzymatic cascades. The recombinant adenine phosphoribosyltransferase (TthAPRT) and hypoxanthine phosphoribosyltransferase (TthHPRT) from Thermus thermophilus HB27 were expressed in E.coli strains and purified by chromatographic methods with yields of 10–13 mg per liter of culture. The activity dependence of TthAPRT and TthHPRT on different factors was investigated along with the substrate specificity towards different heterocyclic bases. The kinetic parameters for TthHPRT with natural substrates were determined. Two nucleotides were synthesized: 9-(β-D-ribofuranosyl)-2-chloroadenine 5'-monophosphate (2-Сl-AMP) using TthAPRT and 1-(β-D-ribofuranosyl)pyrazolo[3,4-d]pyrimidine-4-one 5'-monophosphate (Allop-MP) using TthНPRT.


Introduction
Bacterial phosphoribosyltransferases are used in multi-enzymatic cascades that perform nucleotide synthesis de novo [1,2]. Recently, we reported on the possibility of cascade synthesis, where enzymes of thermophilic microorganisms Thermus thermophilus HB27 (phosphoribosylpyrophosphate synthetase -PRPPS and adenine phosphoribosyltransferase -APRT) and Thermus sp. 2.9 (ribokinase -RK) carry out successive transformations of ribose and adenine heterocyclic bases into the corresponding nucleotides ( Figure 1). The use of thermophilic phosphoribosyltransferases allows carrying out reactions at a higher temperature, so the concentrations of heterocyclic bases can be increased [1][2][3].
There is great interest in the development of multi-enzymatic cascades [4][5][6][7][8][9] for the preparation of nucleosides and nucleotides due to the regio-and stereospecificity of enzymes [4, 10,11], performing metabolic transformations of substrates. Phosphoribosyltransferases are increasingly being widely used as key enzymes in multi-enzymatic systems [2]. The substrate specificity of APRT limits the number of possible nucleotides that can be synthesized. Thus, for Thermus thermophilus HB27 APRT (TthAPRT), nucleotide synthesis is limited to the closest structural homologs of adenine (Table 1)  Unfortunately, 1,2,4-triazole-3-carboxamide, its analogues, guanine, hypoxanthine, and 7-deazapurins are not substrates for TthAPRT. This severely limits the usability of multi-enzymatic cascades in the synthesis of nucleotides, including the modified ones.
To expand the possible repertoire of nucleotides that could be synthesized, we obtained the recombinant form of hypoxathine phosphoribosyltransferase Thermus thermophilus (TthHPRT), investigated its substrate specificity and optimal conditions for catalytic activity, and determined the kinetic parameters of the enzyme. A comparative study of the substrate specificity of TthAPRT and TthHPRT was performed to determine the usability of thermophilic transferases in nucleotide synthesis. A scheme of purine nucleotide synthesis using TthAPRT and TthHPRT is shown in Figure 2.

Results and Discussion
Genes TT_RS08985 and TT_RS06315 from T. thermophilus HB27, coding TthHPRT and TthAPRT, were cloned into expression plasmid vectors pET 23a+ and pET 23d+, respectively. The resulting recombinant plasmid pER-TthHPRT contained fusion gene HPRT-HisTag coding TthHPRT with a C-terminal His-Tag. The resulting recombinant plasmid pER-TthAPRT contained the gene APRT coding TthAPRT without any additional sequences. Nucleotide sequences of the cloned genes were verified by sequencing. The codone GGG→AGG substitution corresponding to amino-acid Arg27Gly replacement was found in the gene encoding the TthHPRT.
The screening of available producer strains was performed to find strains, which produce target enzymes in soluble form. The resulting strains E. coli BL21(DE3)/pER-TthAPRT and E. coli C3030/pER-TthHPRT produced enzymes mainly in soluble form (>80%).
The established procedure for isolation and purification of TthHPRT includes heat treatment, immobilized metal affinity chromatography, final size-exclusion chromatography, and concentration. For TthAPRT, the protocol include heat treatment, anion exchange chromatography, hydrophobic chromatography, final size-exclusion chromatography, and concentration. Yields of both transferases were no less than 10-13 mg per liter of culture, with a purity of about 96% (as determined by SDS-PAGE).
The influence of temperature and Mg 2+ concentration on the activity of TthHPRT was investigated. The results were compared with data for adenine phosphoribosyltransferase Thermus thermophilus, obtained earlier [1].
The TthAPRT is active over a wide temperature range ( Figure 3). A maximal activity of TthHPRT (1.1 unit/mg) is observed at 60 °C. The activity at 36 °C is 5% from the maximal one and at 90 °C it is 3% from the maximal one. It is interesting, that TthAPRT shows its maximal activity at 75 °C.

The influence of the magnesium ion concentration on the
TthHPRT activity is nonlinear. The activity increases rapidly while the magnesium chloride concentration increases from 0 to 1 mM ( Figure 4). Further increasing of the concentration (up to 5 mM) does not increase the activity significantly. Since the reaction rate increases rapidly with increasing the magnesium chloride concentration to values equivalent to the concentration of 5-phosphoribosyl-α-1-pyrophosphate (1 mM), it can be assumed that the presence of magnesium ions promotes the proper spatial orientation of the substrate. The reaction also proceeds in the absence of magnesium ions in solution. A similar dependence is observed for TthAPRT. After optimization of the reaction conditions, kinetic parameters for TthHPRT were determined (Table 2).
Based on the K m values, the affinity of 5-phosphoribosyl-α-1pyrophosphate for the active site is much lower than that of heterocyclic bases. The similar situation we observed for TthAPRT [1]. Comparison of the synthesis rates of inosine-5'monophosphate and guanosine-5`-monophosphate showed that the first is synthesized 4.6 times faster. The literature data for similar enzymes (see Table 2) confirm a poor affinity of PRPP to the active site: K m for hypoxanthine is 17 fold less then for PRPP, although for the human enzyme K m is only 5 fold less. Comparing two enzymes from different strains of Thermus thermophilus, we can conclude that TthAPRT from HB8 (in contrast with HB27), synthesizes guanosine-5`-monophosphate faster. This may be due to the difference in reaction conditions. Kinetic data are displayed by double reciprocal plot ( Figure 5). Determination of substrate specificity of TthHPRT was performed in comparative experiments with TthAPRT. The process of nucleotide synthesis was monitored by a liquid chromatography-mass spectrometry analysis of the reaction mixture.
The data is presented in the Table 3. As expected, TthHPRT is specific to 6-oxopurines, while TthAPRT is specific to 6-aminopurines. Both enzymes do not recognize thymine as a substrate. This is consistent with data that pyrimidine heterocyclic bases are substrates for uracyl phosphoribosyltransferase Escherichia coli [14] Hypoxanthine 37 ---PRPP 330 --- and orotate phosphoribosyltransferase only [2]. Unfortunately, we did not find any product in reactions with compounds based on 1,2,4-triazole-3-carboxamide, which was also observed for E. сoli HPRT [15,16]. However, allopurinol and 8-azaguanine are substrates for TthHPRT, and 2-chloroadenine is a substrate for TthAPRT. For 2-chloroadenine and 8-azaguanine, reaction at a higher temperature is preferable because of their low solubility in water (less than 1 mM at 37 °C). Interestingly, allopurinol proved to be a good substrate for both TthHPRT and TthAPRT, unlike hypoxanthine, which differs only in the position of one of the nitrogen atoms. Probably, the presence of nitrogen atom at C7 position of purine heterocycles plays an important role in reactions catalyzed by phosphoribosyltransferase, and also affects the substrate properties of TthHPRT and TthAPRT.  Two nucleotides were synthesized using TthHPRT or TthAPRT (see Figure 6). Synthesis of 2-Cl-AMP was performed at 75 °C. This allowed to achieve a concentration of 0.5 mM of the initial 2-chloroadenine. The reaction progress was monitored by HPLC. After 2 days (the product content in the reaction mixture was 54%), the reaction mixture was concentrated and the desired product was isolated by column chromatography on ionexchange sorbents (anion and then cation-exchange). The yield of 2-Сl-AMP was 37%. A second nucleotide (Allop-MP) was synthesized at a lower temperature (60 °C). After 2 days, the product content in the reaction mixture was 55%. The product was isolated in the same way, with a yield of 32%.

Conclusion
The recombinant adenine phosphoribosyltransferase and hypoxanthine phosphoribosyltransferase from Thermus thermophilus HB27 were purificated with yields no less than 10-13 mg per litre of culture. A comparative study of substrate specificity of these enzymes towards different heterocyclic bases was carried out and temperature-dependence and magnesium chloride concentration-dependence of enzymes activity were determined.
The protein concentration was determined by the Bradford method [17], using BSA as a standard.
Protein purity was determined by electrophoresis in a polyacrylamide gel under denaturing conditions [18].
Cloning and creation of producer strain: Genes TT_RS08985 and TT_RS06315, encoding TthHPRT and TthAPRT, respectively, were amplified on the genomic DNA template of the T. thermophilus HB27 strain by a polymerase chain reaction (PCR) using synthetic primers. The genes were cloned into the expression vectors pET-23a+ and pET-23d+ respectively. The E. coli strains BL21(DE3)/pER-TthAPRT and C3030/pER-TthHPRT produced the target enzymes mainly in soluble form (culturing conditions: 4 h grow at 37 °С after supplementing with 0.4 mM IPTG).   14 mmol) and TthAPRT (5 units) were added, and the reaction mixture was incubated at 75 °C for 2 days; the reaction progress was monitored by HPLC. The reaction mixture was neutralized with 2 N hydrochloric acid and concentrated in vacuo to ca. 10 mL. The precipitate was filtered off, the filtrate was applied to the column with DEAE-Toyopearl 650C, bicarbonate form, 40 × 140 mm, and the product was eluted with triethylammonium bicarbonate (0.1 M). Fractions were concentrated in vacuo to ca. 10 mL, applied to the column with CM-Sephadex C-25, sodium form, 20 × 160 mm, and the product was eluted with water to give, after evaporation and drying in vacuo under P 2 O 5 , 16 mg (0.037 mmol; 37%) of 9-(β-D-ribofuranosyl)-2-chloroadenine 5'-monophosphate of 99% purity (HPLC ppm.