Use of an Escherichia coli Recombinant Producing Thermostable Polyphosphate Kinase as an ATP Regenerator To Produce Fructose 1,6-Diphosphate

Abstract

Heat-treated Escherichia coli producing Thermus polyphosphate kinase regenerated ATP by using exogenous polyphosphate. This recombinant could be used as a platform to produce valuable compounds in combination with thermostable phosphorylating or energy-requiring enzymes. In this work, we demonstrated the production of fructose 1,6-diphosphate from fructose and polyphosphate.

ATP is the most important biological phosphoryl donor and is required for many enzymatic reactions. The direct addition of ATP not only is expensive but also leads to the accumulation of inhibitory by-products, such as ADP or AMP (19). Furthermore, high concentrations of ATP inhibit enzymes such as phosphofructokinase (PFK) (19). These problems can be overcome by including ATP-regenerating systems in the reactions.

Inorganic polyphosphate (polyP), a linear chain of many tens or hundreds of phosphates, is an inexpensive phosphagen; a commercial form of polyP costing $9/lb can provide ATP equivalents that, if purchased separately, would cost over $2,000/lb, while other phosphagens able to regenerate ATP, such as phosphoenolpyruvate and phosphocreatine, cost more than ATP. ATP can be synthesized from polyP and ADP with polyP kinase (PPK) (8). This ATP regeneration has recently been applied to the practical synthesis of an oligosaccharide, N-acetyllactosamine (13). Also, ATP can be regenerated from AMP using PPK and polyP-AMP phosphotransferase (15, 18).

The genus Thermus has thermostable enzymes that have enormous potential for many industrial applications (4, 12, 14). The production of thermostable enzymes in mesophilic hosts inactivates only host enzymes by a simple heat treatment, which reduces the production of by-products (1, 2). We expected that Thermus thermophilus PPK (PPK_T), whose amino acid sequence has 30% identity with that of E. coli PPK (7), would be thermostable and that heat-treated E. coli producing PPK_T could be directly used as an ATP regeneration system. We examined the use of this system to generate fructose 1,6-diphosphate (FDP), which can be used to reduce ischemic injury in the myocardium, brain, and kidney (3, 5, 10, 11, 16).

A DNA fragment encoding PPK_T was amplified from T. thermophilus HB27 DNA using the primer pair ppk1/ppk2 (Table 1) and then inserted into pET21-b (EMD Biosciences, Darmstadt, Germany). The resultant plasmid, pETPPK_T, was introduced into E. coli Rosetta(DE3)pLysS (EMD Biosciences). His-tagged PPK_T was purified from the E. coli recombinant by HisTrap chromatography (GE Healthcare, Piscataway, NJ). The rates of ATP production by purified PPK_T (20 ng) were measured in 60 μl of 5 mM polyP65 (65 phosphate residues; Sigma, St. Louis, MO), 100 μM ADP, 40 mM (NH₄)₂SO₄, 4 mM MgCl₂, and 50 mM HEPES-KOH (pH 7.2). The generated ATP was determined using a bioluminescence assay (CLSII; Roche Diagnostics, Basel, Switzerland). PPK_T synthesized 3.6 × 10⁶ pmol of ATP per min per mg protein at 70°C (Fig. 1), which is slightly lower than the maximum velocity of E. coli PPK (8, 9). The optimal temperature for PPK_T was between 60°C and 80°C (Fig. 1), whereas E. coli PPK lost its activity over 50°C (data not shown). The optimal pH of PPK_T was between 5 and 6. Also, PPK_T required magnesium for activity, and high concentrations of ammonium sulfate and potassium chloride inhibited its activity. PPK_T preferred to use a long-chain polyP (>25 phosphate residues) (Fig. 1).

TABLE 1.

DNA primers

Primer	Sequencea
ppk1	5′-GGAATTCCATATGCACCTCCTTCCCGAAGC-3′
ppk2	5′-GAAGTCGACTAGCTCCAGGCGCTGGGCGT-3′
pfk1	5′-GGCCATATGAAACGCATCGGGGTGTT-3′
pfk2	5′-TATAAGCTTGAGGGCCAGCACCTGCGATA-3′
pfk3	5′-GAAGAATTCAATGAAACGCATCGGGGTGTT-3′
pfk4	5′-TATAAGCTTCTAGAGGGCCAGCACCTGCG-3′
fk1	5′-AAGGTTCCATATGGTGGAGAAGCCCGCTATC-3′
fk2	5′-TTTGTCGACGCCCGCAAGCCACCGGGCCC-3′
fk3	5′-TTTGATATCCTAGCCCGCAAGCCACCGGG-3′

^aThe underlined sequences represent additional restriction enzyme sites.

FIG. 1.

Effect of temperature (A), pH (B), MgCl₂ (C), (NH₄)₂SO₄ (D), KCl (E), and the length of polyP (F) on the ATP synthesis activity of PPK_T. All data shown are representative of two replicates and were almost perfectly reproduced unless otherwise noted. Bars in panel F represent standard deviations in triplicate experiments.

We expected that heat treatment would increase the membrane permeability and allow ATP and polyP to penetrate the cell membrane. A 5-μl sample of heat-treated E. coli harboring pETPPK_T (7.8 mg ml⁻¹) at 70°C for 10 min was mixed with 55 μl of 50 mM MOPS (morpholinepropanesulfonic acid)-NaOH buffer (pH 6.0) containing 0.08 mM ADP, 40 mM (NH₄)₂SO₄, 4 mM MgCl₂, and 5 mM polyP65. The heat-treated E. coli recombinant synthesized ATP from external ADP and polyP, leading to the extracellular accumulation of ATP. After heat treatment, PPK_T activity was detected in the precipitated cell fraction but not in the supernatant (Fig. 2). More than 60% of PPK_T activity was retained even after a 1-week incubation at 70°C (see the supplemental material).

FIG. 2.

ATP synthesis activity by E. coli producing PPK_T. After heat treatment, recombinant E. coli was precipitated. Heat-treated cells (filled circles) and the supernatant (open circles) were examined for the ability to synthesize ATP.

To allow the production of FDP in this system, we cloned genes encoding predicted fructokinase (FK) and PFK (20) from T. thermophilus. The FK and PFK DNAs amplified using primer pairs fk1/fk2 and pfk1/pfk2, respectively (Table 1), were inserted into pET21-b. The combined use of the purified PPK_T, FK, and PFK synthesized FDP from fructose and polyP75 (data not shown). Also, the combined use of heat-treated E. coli recombinants producing FK, PFK, and PPK_T synthesized FDP at 70°C (Fig. 3). The reaction was performed in a 500-μl mixture containing 50 mM HEPES-KOH (pH 7.0), 10 mM fructose, 10 mM KCl, 10 mM MgCl₂, 1 mM ADP, 30 mM polyP75, 4 mg of E. coli (PPK_T) cells, 0.5 mg of E. coli (PFK) cells, and 0.5 mg of E. coli (FK) cells. When we used 30 mM ATP instead of polyP, however, we did not observe FDP generation (Fig. 3). This may be ascribed to the fact that FK is perfectly inhibited by 10 mM ATP (see the supplemental material). ATP regeneration overcame this inhibitory effect of ATP.

FIG. 3.

Synthesis of FDP from fructose and polyP75 by E. coli recombinants. FDP was synthesized using 4 mg of E. coli cells producing PPK_T, 0.5 mg of E. coli cells producing PFK, and 0.5 mg of E. coli cells producing FK (filled circles). The ratio of PPK to PFK to FK in terms of protein concentration was approximately 8:1:3. FDP was measured as previously reported (6). In a parallel reaction, ADP and polyP75 were replaced by 30 mM ATP as the phosphoryl donor (open circles). FDP was also synthesized using 5.8 mg of E. coli cells producing PPK_T, PFK, and FK simultaneously (filled triangles).

We expected the rate of FDP synthesis to increase if the FK, PFK, and PPK_T reactions occurred in the same cell. We constructed an E. coli recombinant producing FK, PFK, and PPK_T simultaneously. Two DNAs encoding FK and PFK were amplified using primer pairs fk1/fk3 and pfk3/pfk4 (Table 1), respectively, and then inserted into one vector, pRSF-Duet (Takara Co., Kyoto, Japan). The resulting plasmid was introduced into E. coli harboring pETPPK_T. The heat-treated E. coli strain producing FK, PFK, and PPK_T simultaneously successfully synthesized FDP (Fig. 3). The rate of FDP synthesis increased approximately twofold compared to that of the mixture of heat-treated E. coli strains separately producing FK, PFK, and PPK_T. A nearly 100% yield from fructose was accomplished within 3 h.

Here, we demonstrated that PPK_T is a thermostable PPK and that heat-treated E. coli producing PPK_T can be used as an ATP regenerator for at least 1 week at 70°C. The biggest advantages of the present method are the stability of PPK_T and the use of an inexpensive phosphagen for ATP regeneration. During the review of our paper, Sato et al. published a study of ATP-requiring d-amino acid dipeptide synthesis using an E. coli strain producing Thermosynechococcus PPK (17), which is thermostable but less so than Thermus PPK. While further research is required to achieve synthetically useful product concentrations, the ATP regenerator described in this paper could be used as a general platform to produce valuable compounds by thermostable phosphorylating or energy-requiring enzymes.