JOUR~At OF BIOSClEIqCE AND BIOElqOII~ERIHG Vol. 91, No. 6, 557-563. 2001

A Novel ATP Regeneration System Using Polyphosphate-AMP Phosphotransferase and Polyphosphate Kinase

ATSUSHI KAMEDA, 1 TOSHIKAZU SHIBA, 1. YUMI KAWAZOE, 1 YASUHARU SATOH, 1 YOSHIHARU IHARA, 1§ MASANOBU MUNEKATA, 1 KAZUYA ISHIGE, 2 AND TOSHITADA NOGUCHI 2

Division o f Molecular Chemistry, Graduate School o f Engineering, Hokkaido University, Sapporo 060-86281 and Biochemicals Division, Yamasa Corporation, Choshi, Chiba 288-0056, 2 Japan

Received 15 January 2001/Accepted 9 March 2001

Polyphosphate-AMP phosphotransferase (PAP) and polyphosphate kinase (PPK) were used for designing a novel ATP regeneration system, named the PAP-PPK ATP regeneration system. PAP is an enzyme that catalyzes the phospho-conversion of AMP to ADP, and PPK catalyzes ATP formation from ADP. Both en- zymes use inorganic polyphosphate [poly(P)] as a phosphate donor. In the PAP-PPK ATP regeneration system, ATP was continuously synthesized from AMP by the coupling reaction of PAP and PPK using poly(P). Poly(P) is a cheap material compared to acetyl phosphate, phosphoenol pyruvate and crcatine phosphate, which are phosphate donors used for conventional ATP regeneration systems. To achieve efficient synthesis of ATP from AMP, an excessive amount of poly(P) should be added to the reaction solution because both PAP and PPK consume poly(P) as a phosphate donor. Using this ATP generation reaction, we constructed the PAP-PPK ATP regeneration system with acetyI-CoA synthase and succeeded in synthesizing acetyl-CoA from CoA, acetate and AMP. Since too much poly(P) may chelate Mg 2+ and inhibit enzyme activity, the Mg 2+ concentration was optimized to 24 mM in the presence of 30 mM poly(P) in the reaction. In this reaction, ATP was regenerated 39.8 times from AMP, and 9 9 . 5 ~ of CoA was converted to acetyI-CoA. In addition, since the PAP-PPK ATP regeneration system can regenerate GTP from GMP, it could also be used as a GTP regener- ation system.

[Key words: acetyl-CoA synthesis, ATP regeneration system, inorganic polyphosphate, polyphosphate-AMP phosphotransferase, polyphosphate kinase]

Enzymatic systems for ATP regeneration have been constructed using several combinations of phosphate donors and enzymes, such as acetyl phosphate (AcP) and acetate kinase, phosphoenol pyruvate (PEP) and pyruvate kinase (1-3), and creatine phosphate (PC) and creatine kinase (4). ATP regeneration has been applied to enzymatic synthesis of GMP, CDP-choline and other materials in the industry (5-7). Although these systems work well in certain applications, there are problems regarding the high costs of chemicals and lack of availa- bility of a method for regenerating ATP from AMP. Since some ATP-dependent enzymes produce AMP, an efficient system for regenerating ATP from AMP is also needed (8). For these reasons, a new economical system for enzymatic regeneration of ATP from both ADP and AMP would be useful, particularly for industrial use.

One of the promising candidates as an economical phosphate donor for an ATP regeneration system is an inorganic polyphosphate [poly(P)] (9, 10). Poly(P) inclu- des linear polymers of orthophosphate with high-energy phospho-anhydrate bonds (11, 12). A large amount of poly(P) is routinely produced as sodium hexametaphos- phate (about 13 to 18 phosphate residues) for food ad- ditives and other industrial uses. It is inexpensive compar- ed to AcP, PEP and PC.

Several enzymes involved in poly(P) metabolism have been identified in many bacteria, yeast and some eukary- otes (13, 14). Polyphosphate kinase (PPK) is an enzyme

* Corresponding author, e-mail: shiba@dove~mc.eng.hokudai.ac.jp phone/fax: + 81-(0)11-706-7816 § Present address: Hokkaido Food Processing Research Center,

Bunkyodaimidori-cho, Ebetsu, Hokkaido 069-0836, Japan.

catalyzing the synthesis of poly(P) polymerizing the ter- minal phosphate of ATP to poly(P) (15). The Esch- erichia coli PPK can also catalyze the kination of not only ADP but also other nucleoside diphosphates using poly(P) as a phosphate donor, yielding nucleoside tri- phosphates (NTP) (16). Using E. coli PPK and poly(P), we have designed an NTP regeneration system and suc- ceeded in synthesizing sugar nucleotides in a cyclic syn- thesis system for oligosaccharides (17).

Resnick and Zehnder constructed an ATP regenera- tion system from poly(P) and AMP by PAP and adeny- late kinase (AdK) of Acinetobactor johnsonii strain 210A (18). PAP catalyzes the phospho-conversion of AMP to ADP using poly(P) as a phosphate donor. This activity has been found in Acinetobacter (19) and Myx- ococcus xanthus (13). Although PAP was purified from an Acinetobacter strain (20, 21), a gene encoding PAP has not yet been identified. Resnick and Zehnder cou- pled partially purified PAP with adenylate kinase (AdK) for complete regeneration of ATP from AMP via ADP.

Instead of the combinations of PAP from A. john- sonii and AdK, we employed PAP from M. xanthus and PPK to construct a novel ATP regeneration system. While PAP activity is mainly detected in the supernatant of A. johnsonii (20), it is also found in the membrane fraction of M. xanthus. Repeated washing of this mem- brane fraction completely removes other poly(P)-degrad- ing activity and facilitates preparation of an active PAP fraction. The PAP activity in the membrane fraction was also stable in solution at 4°C for at least 2 months. In addition, since PPK catalyzes the NTP regeneration reaction using poly(P) only while AdK requires ATP to catalyze NTP synthesis from nucleoside diphosphates

557

558 KAMEDA ET AL. J. Blosci. BIo~'~o.,

(NDP) (22), it might be possible in the future to design an NTP regeneration system from nucleoside monophos- phates (NMP) by the coupling reactions of PAP and PPK using poly(P) as the sole phosphate donor. Actually, the active PAP fraction of M. xanthus has GDP synthetic activity from GMP and could generate GTP by coupling with PPK.

Using these advantages of PPK, PAP and poly(P), we designed a novel ATP regeneration system, named the PAP-PPK ATP regeneration system. Combinations of PPK and PAP can catalyze ATP formation from AMP using poly(P) as a phosphate donor as follows.

AMP + poly(P)n --, ADP + poly(P)n- 1

ADP + poly(P)n --~ ATP + poly(P)n- 1

Using the PAP-PPK ATP regeneration system, we suc- ceeded in realizing continuous synthesis of acetyl-CoA from CoA catalyzed by acetyl-CoA synthase (ACS).

MATERIALS AND METHODS

Bacterial strains and plasmids M. xanthus DK101 was employed for preparation of partially purified PAP. E. coil JM109 was the host strain for preparation of plus- raids and overproduction of His-tagged proteins. Plas- mid pQE30 for expression of His-tagged proteins was purchased from QIAGEN (Germany).

Materials The Ni-NTA agarose for purification of His-tagged protein was purchased from QIAGEN. Sodium phosphate glasses of type 75 + [Poly(P)75] whose average chain length is 75 phosphate residues was purchased from Sigma. Polyethyleneimine-cellulose thin-layer chro- matography (PEI-TLC) plates were purchased from Merck. 5-5'-Dithiobis-2-nitrobenzoic acid (DTNB) was purchased from Wako-Junyaku (Osaka). N-Dodecyl-N,N-dimethyl- 3-ammonio-l-propanesulfonate (SB-12) was purchased from Boehringer Mannheim. Restriction and DNA modi- fication enzymes were purchased from Takara-Shuzo (Kyoto).

Partial purification of PAP M. xanthus DK101 was grown at 30°C in 1 l of CTT medium (1% Bacto casitone, 10mM Tris-HC1 (pH 8.0) and 8 mM MgSO4) with shaking until the stationary growth phase was reached. Then the culture was harvested by centrifuga- tion, and the cell pellet was suspended in 40ml of 20mM Tris-HC1 (pH8.0) and sonicated (SONIFIE 450D, BRANSON, USA) for 4 rain at output 5 to obtain a cell-free extract. DNase (20 pg/ml), RNase (20 pg/ml) and MgC12 (4mM) were added, and the mixture was incubated at 4°C for 16h. After incubation, the cell extract was centrifuged (19,500 x g, 60 min, 4°C) and the supernatant was discarded. The pellet fraction was resuspended in 15 ml of washing buffer [20 mM Tris-HCl (pHS.0), 1% Triton X-100 and 1% SB-12], and the resuspended fraction was recentrifuged to wash out the proteins that do not have PAP activity. This washing step was repeated three times, and the pellet fraction thus obtained was resuspended in 5 ml of 20 mM Tris- HCI (pH 8.0). This fraction was used as partially puri- fied PAP (3.62× 10 -3 units/mg protein) for each reac- tion in this study.

Measurement of PAP activity The standard reac- tion mixture (20 pl) contained 50 mM Tris-HCl (pH 8.0), 40 mM (NH4)2SO4, 4 mM MgC12 (buffer A), 1 mM AMP and 0.238mM (in terms of phosphate residues)

[32p]poly(P), whose average chain length is approximate- ly 750 phosphate residues. The reaction was initiated by the addition of PAP and performed at 37°C. Initial con- centrations of AMP, poly(P) and PAP fractions differed from experiment to experiment. [32p]ADP generated in the reaction was separated on a PEI-TLC plate devel- oped using 0.75 M KH2PO4 (pH 3.5), and the radioac- tivity was quantitated using a radio-image analyzer (BAS2000, Fujix). One unit of PAP activity was defined as the amount of PAP producing one pmol of ADP per minute.

Overproduction and purification of His-tagged PPK The E. coli ppk gene was amplified by polymerase chain reaction (PCR) using an upstream primer (5'-AAGGTA CCACACAGAATTCATTAAAGAGGA-3') containing a BamHI site and a downstream primer (5'-AAACTGCA GGCGGCAACCGAGCGTTCTG-3') containing a HindIII site (23). A 2.2-kb DNA fragment containing the entire open reading frame of the ppk gene was cloned into the pQE30 expression vector (QIAGEN) to obtain pQEPPK. The ppk gene was inserted downstream of the T5 pro- moter that is inducible by isopropyl beta-D-thiogalac- topyranoside (IPTG). E. coli JM109 transformed by pQEPPK was cultured in LB medium with 100pg/ml ampicillin at 30°C until the mid-logarithmic growth phase (OD600=0.5). Then IPTG was added to a final con- centration of I mM and the bacterium was continuously cultured for 4 h. Cells were harvested by centrifugation, and the cell pellet was suspended in 10 ml of lysis buffer [50ram NaH2PO4 buffer (pH 8.0), 300raM NaC1 and 10mM imidazole]. Cells were disrupted by sonication (SONIFIE 450D, BRANSON) for 3min at output 3, and the His-tagged PPK was purified using a Ni-NTA column according to the manufacturer's instructions (QIAGEN). One unit of PPK was defined as the amount of PPK producing one pmol of ATP per minute. The specific activity of the purified His-tagged PPK was 1.407 units/rag protein.

Overproduction and purification of ACS The ACS-coding region of the E. coil acs gene was amplified by PCR using an upstream primer (5'-AAGGATCCAG CCAAATTCACAAACACACCATTCCT-3') containing a BamHI site and a downstream primer (5'-AAAAGC TTACGATGGCATCGCGATAGCCTG-3') containing a HindIII site (24). A 1.9-kb DNA fragment containing the entire open reading frame of the acs gene was cloned into the pQE30 vector to yield pQEACS. The acs gene was inserted downstream of the T5 promoter. E. coil JM109 transformed by pQEACS was cultured in LB medium with 100pg/ml ampicillin at 37°C until an OD~0o of 0.3 was reached. Then IPTG was added to a final concentration of 0.4 mM and the bacterium was continuously cultured for 6 h. Extraction and purifica- tion of His-tagged ACS was carried out in the same way as that for His-tagged PPK. One unit of ACS was defined as the amount of ACS producing one pmol of acetyl-CoA from CoA and acetate per minute. The specific activity of the purified His-tagged ACS was 3.58 units/rag protein.

Evaluation of the PAP-PPK ATP regeneration system in terms of acetyI-CoA synthesis Reaction was per- formed using an appropriate buffer (400 pl) containing appropriate amounts of acetate, CoA, Poly(P)75, AMP, PAP, PPK and ACS as described in the legends of Figs. 7, 8 and 9. The progress of the reaction was evaluated by two independent methods: measurement of the

VOL. 91, 2001 PAP-PPK ATP REGENERATION SYSTEM 559

TABLE 1. Partial purification of PAP from M. xanthus

Protein Activity Specific activity Purification Yield Fraction Step (mg) (U) (U/mg) (fold) (%)

I Crude extraction 53 3.17 0.06 1.0 100 II Pellet fractionation 19 3.12 0.16 2.7 95 III Detergent washing 0.25 2.99 12 200 94

amount of CoA consumed in the reaction as described previously (25) and direct measurement of the amount of acetyl-CoA synthesized by HPLC (Shimadzu, Kyoto, SCL-10A). For HPLC analysis, the reaction was stopped by adding 5 pl of IN CHaCOOH to 25 pl of the reaction mixture, and boiling the reaction mixture for 2min. Then 100mM KH2PO4 (120pl) was added, and the mixture was centrifuged to remove insoluble materials. The supernatant was applied to a Mightysil RP-18 GP column (150-3 mm, Kanto Chemical Co. Inc., Tokyo) pre-equilibrated with 50mM KH2PO, in methanol (10:90, v/v) (A) and analyzed using 10% of 50mM KH2PO, in methanol (30 : 70, v/v) (B). CoA and acetyl- CoA were separated by a linear gradient of 10-100% of

A

ADP

poly(P)---~ O O ~

Time(rain) 0 5 10 20 30 40 60 80

B

0.25<

g

13 ~ 0.15- e~ ~ 0.1-

~ 0.05-

o 0 20 40 60 80

Time (min)

FIG. 1. PAP activity of fraction III in M. xanthus. (A) Time course of PAP-dependent [32p]ADP synthesis from [32p]poly(P) and AMP. The reactions were performed at 37°C in buffer A containing 1 mM AMP, 0.238 mM [32p]poly(P) and 1.66 × 10 -4 units of PAP (fraction III). Reaction mixtures were separated on a PEI-TLC plate and visualized using a radio-image analyzer (BAS2000, Fujix). (B) Determination of amounts of poly(P) and ADP in PAP reaction. Poly(P) and ADP concentrations were calculated by measuring the intensity of radioactive spots visualized using a radio-image analyzer as shown in panel A. Poly(P) is shown by diamonds, and ADP by squares.

B (25min) with a flow rate of 0.7ml/min at 37°C. Under these conditions, the retention times of CoA and acetyl-CoA were 9.0 and 19.9 min, respectively.

Other procedures [32p]poly(P) was synthesized in vitro using the purified E. coil PPK and purified as described by Akiyama et al. (26). Concentrations of syn- thesized [32p]poly(P) were determined by the method of Wurst and Kornberg (27). PPK activity was measured using [32p]poly(P) as described by Ahn and Kornberg (15). Concentrations of poly(P) are given in terms of phos- phate residues. All DNA manipulations were carried out as described by Sambrook et al. (28).

RESULTS

Preparation of PAP in M. xanthus In order to con- struct an ATP regeneration system using PAP and PPK, we prepared PAP from M. xanthus. M. xanthus accumu- lates a large amount of cellular poly(P), and PAP is the major poly(P)-degrading activity in this bacterium (13). Unlike that of the PAP of Acinetobacter (19-21), the activity of the enzyme of M. xanthus appeared in the membrane fraction in the stationary growth phase. We partially purified the PAP activity from the membrane fraction. The PAP activity was 200-fold purified with successive washings of the membrane fraction using de- tergents as shown in Table 1. The purified fraction (frac- tion III) had no other poly(P)-degrading activities such as exopolyphosphatase (Fig. 1A). All of the [32p]poly(P) was converted to ADP by PAP activity in fraction III depending on incubation time (Fig. 1A, B). For this rea- son, we employed fraction III as a PAP enzyme for con- struction of the PAP-PPK ATP regeneration system.

25

~ 4) 2sO4

15

50 100 150 Concentration (raM)

FIG. 2. Enhancement of PAP activity by NH4 + and SO42-. The PAP activity was assayed in 10 pl of the reaction mixture containing 50 mM Tris-HC1, 4 mM MgCI2, 1 mM AMP, 0.1 mM [32p]poly(P) and appropriate amounts of indicated salts at 37°C. After 10rain of incubation, the assay mixture was applied to PEI-TLC plates and developed in 0.4 M LiCI and 1 M formic acid. The p2p]ADP spot was quantitated using a radio-image analyzer (BAS2000, Fujix). KCI is shown by diamonds, NH4CI by triangles, K2SO4 by squares and 0NI-I4)2SO4 by circles.

560 KAMEDA ET AL. J. BIO$CI. BIOEN¢3.,

AMP

poly(P)n -"N

PAP

poly(p)n.1 "~ ' /

ADP

poly(P). ~

po ly (p)n . l~ / PPK

ATP

FIG. 3. Scheme of the ATP generation reaction by PAP and PPK. PAP catalyzes the poly0a)-dependent ADP formation from AMP, and PPK catalyzes the poly(P)-dependent ATP formation from ADP. Both enzymes commonly use poly(P) as a phosphate donor for kination reaction.

Basic properties of PAP Similar to the P A P in Acinetobacter (20), the M. xanthus enzyme had an op- timal pH o f 8.5 and required Mg 2+ (data not shown). The P A P activity was strongly stimulated by SO42- and NH4 + salts at K2SO4 and NH4CI concentrations o f up to 80 mM and 150 mM, respectively; there was no stimulato- ry effect of KCI (Fig. 2). Thus, (NH4)2SO4 strongly stimu- lated the activity, and the optimal concentration was approximately 40 mM (Fig. 2).

Coupling of PAP and PPK for ATP generation from AMP Using the P A P enzyme prepared as described above, we first attempted to generate AT P from A M P by coupling the enzyme with PPK. His-tagged PPK was purified as described in Materials and Methods, and used as a PPK enzyme in this AT P generation reaction. The scheme for the AT P generation reaction is shown in Fig. 3. Both P A P and PPK use poly(P) as a phosphate donor for A D P and AT P syntheses, respectively. Figure 4 shows time courses o f ATP and ADP formation in

1 .~ o.1

~ o.1

-~ 0.0

~2 0 ao 60 90

Time (rain)

FIG. 4. Time course of ATP generation reaction. The reaction was performed in buffer A (20 ILl) containing 1 mM AMP, 0.238 mM [32p]poly(P), 2.87 x 10 -4 units of PPK and 1.66 x 10 -4 units of PAP. Two-microliter reaction mixtures were spotted and separated on a PEI-TLC plate and analyzed using a radio-image analyzer. ATP is shown by triangles, ADP by squares, poly(P) by circles, and Pi by diamonds.

this reaction. Along with the decrease in poly(P) concen- tration, both A D P and ATP concentrations increased until poly(P) was exhausted in the reaction. Two phosphate molecules were supplied to synthesize both A D P and ATP, and more than 92% of poly(P) was converted to A D P or ATP after 90m in o f reaction. Only 8% o f poly(P) was converted to orthophosphate (Pi) because the purified PPK enzyme has little exopolyphosphatase, ATPase and /o r A D P hydrolase activity. The reason why the efficiency of ATP formation was much less than that o f A D P formation will be discussed in next section.

Effect of initial poly0P) concentration on ATP forma- tion Since both P A P and PPK use poly0 a) to phos- phorylate A M P or ADP, the poly(P) concentration in the ATP generation reaction should be crucial f o r effi- cient synthesis o f ATP. Thus, we evaluated the efficiency of ATP formation at various poly(P) concentrations (Fig. 5). The total amount o f synthesized ATP increased de- pending on the poly(P) concentration that was added to the reaction. In the case o f an initial poly(P) concentra- tion o f 0.238 mM (Fig. 5, squares), 23.5% of poly(P) was converted to ATP. When more poly(P) was added to the reaction (0.768mM, Fig. 5, diamonds), more ATP was synthesized, and 57.2% of poly(P) was convert- ed to ATP. This clearly shows that poly(P) is a limiting factor for the ATP generation reaction. Since both P A P and PPK require poly(P) as a phosphate donor, the reac- tion rates o f both enzymes are therefore limited by poly(P). To achieve an efficient synthesis o f ATP, an excessive amount o f poly(P) should be added to the reac- tion mixture. Theoretically, the initial poly(P) concentra- tion must be more than twice as much as the initial AMP concentration in the reaction mixture to equally distribute poly(P) to P A P and PPK. Since the initial con- centration of poly(P) is much less than that o f AMP, the efficiency of ATP formation was also much less than that of A D P formation, as shown in Fig. 4.

The initial reaction rate was not simply dependent on the poly(P) concentration, and the maximum rate was observed when 0.768 mM poly(P) was added to the reac- tion mixture. When the poly(P) concentration was more than 1.83 mM, reaction rate decreased with increasing poly(P) concentration, because too much poly(P) may chelate Mg 2+ and inhibit enzyme activity. These results

0.4

0.3-

I 0.2-

0.1-

0[ 0 30 60 90

Time (rain)

FIG. 5. Effect of initial poly(P) concentration on ATP genera- tion. Reactions were performed in buffer A (20 pl each) containing I mM AMP, 2.87 x 10 -4 units of PPK and 1.66 x 10 -4 units of PAP. Initial [~2P]poly(P) concentrations in the reactions were varied as follows: 0.238 mM (squares), 0.768 mM (diamonds), 1.83 mM (circles) and 2.89 mM (triangles).

VOL. 91, 2001 PAP-PPK ATP REGENERATION SYSTEM 561

poly(P) n poly(P) n.1

t m l y ( p ) n . l ~ poly(P) n

A'IT AMP

Acetate + CoA - Acetyl-CoA ACS

FIG. 6. Scheme of the acetyl-CoA synthetic reaction with the PAP-PPK ATP regeneration system. Acetyl-CoA was synthesized from acetate and CoA using ATP as an energy source and released AMP and PPi. ATP was regenerated from AMP using poly(P) as a phosphate donor by the coupling reaction of PAP and PPK, catalyz- ing ADP formation from AMP and ATP formation from ADP.

indicate that a sufficient m o u n t of poly(P) for regenerat- ing both ADP and ATP should be added to the reaction mixture, but optimization of the Mg 2+ concentrat ion is also important .

Construction of a P A P - P P K A T P regeneration system for aeetyi-CoA synthesis To construct the P A P - P P K ATP regeneration system, the reaction of acetyl-CoA synthesis as catalyzed by ACS was used. Acetyl-CoA was synthesized from acetate and CoA using ATP as an energy source generating AMP and pyrophosphate (PPi). The scheme of the reaction of acetyl-CoA synthe- sis using this A T P regeneration system is shown in Fig. 6. ATP should be regenerated from AMP, which is produced by ACS, by the coupling reaction of P A P and PPK using poly(P) as a phosphate donor. The reaction of acetyl-CoA synthesis was monitored by measuring the amount of CoA in the reaction.

As shown in Fig. 7, the amoun t of CoA continuously decreased with the reaction time only when the reaction mixture contained complete components of the PAP- PPK ATP regeneration system. No CoA decrease was

1.5

e~

0 50 100 150 Time (min)

FIG. 7. Essential requirements for acetyl-CoA synthetic reaction using the PAP-PPK ATP regeneration system. Reactions were per- formed in buffer A (400 pl) containing 9.4 mM acetate, 1 mM CoA, 2raM AMP and 10raM poly(P)75 in the presence of ACS only (squares), ACS+PAP (diamonds), ACS+PPK (circles) or ACS~- PAP÷PPK (triangles) at 37°C. The amounts of the enzymes added to the reactions were 2.48 × 10 -2 units in the case of ACS, 6.64 × 10-4units in the case of PAP, and 1.15)< 10-3 units in the case of PPK. The decrease in the amount of CoA in the reaction mixture was measured as described in Materials and Methods.

2.5

0.5

0 I i a m i ~ - -

0 50 100 150 200 Time (rain)

FIG. 8. Optimization of Mg 2÷ concentration in the PAP-PPK ATP regeneration system at ahigh poly(P) concentration. Reactions were performed in reaction mixtures (400 pl each) containing 50 mM Tris-HC1 CuH 8.0), 40mM (NI-I4)2SO+, 11.3 mM acetate, 2mM CoA, I mM AMP, 30ram pOly(P)75, 1.24)< 10-2units of ACS, 3.05 × 10 -4 units of PAP and 2.90)< 10 -3 units of PPK at 37°C. MgCI2 concen- trations were varied as follows: 0 mM (open squares), 4 mM (open diamonds), 9 mM (open circles), 14 mM (open triangles), 24 mM (solid squares), 34 mM (solid diamonds) and 44 mM (solid circles). The decrease in the amount of CoA in the reaction mixture was measured as described in Materials and Methods.

MgZ + concentration

0mM 4mM 9mM 14 mM

---- 24 mM 34 mM

~, 44 mM

observed when the reaction mixture did not contain both P A P and P P K or either P A P or PPK. This indicates that both P A P and P P K are required for A TP regenera- t ion and acetyl-CoA synthesis.

Optimization of Mg 2+ concentration in the PAP-PPK A TP regeneration system In order to achieve high- efficiency regeneration of A TP from A MP by the coupling reaction of P A P and PPK, optimization of Mg 2+ concentrat ion was carried out at high poly(P) con- centration in the reaction mixture. In the presence of 30raM poly(P), the initial reaction rate increased with MgC12 concentrat ion up to 24 raM, and the reaction rate decreased when MgCI2 concentrat ion was more than

12.5

1o[

~ 7.5

8

0 5 10 15 20

Time (h)

FIG. 9. Application of the PAP-PPK ATP regeneration system to the acetyl-CoA synthetic reaction. The reaction was performed in a reaction mixture (400 pl) containing 50 mM Tris-HC1 (pH 8.0), 40 mM (NH4)2SO4, 24 mM MgCl2, 11.3 mM acetate, 10 mM CoA, 0.25 mM AMP, 30 mM poly(P)Ts, 1.24 x 10 -2 units of ACS, 1.22 x 10 -3 units of PAP, 1.16 x10-2 units of PPK and 0.1unit of inorganic pyro- phosphatase at 30°C. The progression of the reaction was monitored and is represented as means of the decreased level of CoA in the reac- tion. CoA is shown by squares.

562 KAMEDA ET AL. J. BIoscI. BIOENG.,

TABLE 2. Formation of acetyl-CoA using the PAP-PPK ATP regeneration system

CoA concentration (raM) Acetyl-CoA formed initial final used ( m M ) Regeneration

10.00 0.05 9.95 9.95 39.8 times

24 mM (Fig. 8). Based on these results, acetyl-CoA syn- thesis together with ATP regeneration reaction was per- formed using 30 mM poly(P) and 24 mM MgC12.

Large-scale synthes is o f a e e t y l - C o A under op t ima l con- d i t ions To perform the reaction on a larger scale, acetyl-CoA was synthesized using 10 mM CoA and 11.3 mM acetate as initial substrates. The reaction proceeded continuously for up to 16 h, since the amount of CoA in the reaction continuously decreased during the incuba- tion period (Fig. 9). After 18 h of incubation, 9.95 mM CoA was consumed and 9.95 mM acetyl-CoA was synthe- sized (Table 2). This indicates that all the CoA added to the reaction mixture was converted to acetyl-CoA without any loss. Since only 0.25 mM AMP was initially added to the reaction mixture, ATP was regenerated ap- proximately 39.8times during the 18h incubation period. To avoid both Mg 2+ chelation due to a high poly(P) concentration (30 mM) and the inhibitory effect of PPi caused by ATP hydrolysis, relatively high con- centrations of Mg 2+ (24 mM) and inorganic pyrophos- phatase (0.1 unit) were added to the reaction mixture. The reaction was performed at 30°C since ACS showed decreased activity during prolonged incubation at 37°C.

GTP can also be regenerated f r o m G M P us ing the P A P - P P K A T P regenerat ion sys tem To determine whether the PAP-PPK ATP regeneration system can catalyze generation of other nucleoside triphosphates from nucleoside monophosphates, this system was ap- plied to regeneration of GTP from GMP. As shown in Fig. 10, GTP was formed from GMP in the presence of

P i ~

GDP---4~

GTP---4~

Poly(P)---4~

lanes 1 2 3 4 5

PAP + + +

PPK + + +

GMP + + + +

FIG. 10. Application of the PAP-PPK ATP regeneration system to GTP generation from GMP. Reactions were performed in buffer A (20 pl each) containing 1.44 mM [32p]poly(P) in the presence or ab- sence of 0.5 mM GMP, 1.22 × l0 -4 units of PAP and/or 1.15 × 10 -4 units of PPK at 37°C for 18 h. Reaction mixtures were separated on a PEI-TLC plate and visualized using a radio-image analyzer (BAS2000, Fujix). Reactions were performed with PAP+PPK+GMP (lane 1), P A P + P P K (lane 2), PPK+GMP (lane 3), PAP+GMP (lane 4) or GMP only (lane 5).

both PAP and PPK in the reaction mixture (Fig. 10, lane 1), whereas no GTP formation was observed in the presence of only PAP with GMP (Fig. 10, lane 4) or PPK with GMP (Fig. 10, lane 3). Since the purified PPK enzyme has low exopolyphosphatase activity, Pi forma- tion was observed (Fig. 10, lanes 1 to 3). These results indicate that the PAP-PPK ATP regeneration system can also be used for GTP regeneration.

DISCUSSION

We constructed a novel ATP regeneration system, named the PAP-PPK ATP regeneration system, using a coupling reaction of PAP and PPK. The initial poly(P) concentration is critical for this ATP generation reaction because both PAP and PPK require poly0 a) as a phos- phate donor, and the reaction rates of both enzymes are therefore limited by poly(P). The initial poly(P) concen- tration should be much higher than the AMP concentra- tion for efficient production of ATP and a high turnover rate of regeneration. For example, a minimum of 20 mM poly(P) would be required for 10times regeneration of 1 mM ATP from AMP. However, unless PAP and PPK can share equal amounts of poly(P) for their reaction, this 10 times regeneration would not be achieved. There- fore, a balance of poly(P) utilization is critical for efficient ATP regeneration. If too much AMP is added to the reaction mixture, PAP activity rapidly consumes poly(P) for ADP synthesis, and it is impossible to sup- ply sufficient poly(P) for ATP synthetic reaction by PPK, resulting in generation of ADP rather than ATP.

To achieve a high poly(P) concentration in the reac- tion mixture, it is necessary to avoid the cation chelation effect of poly(P). Since many enzymes, including P A P and PPK, require cations such as Mg 2+, the addition of a high MgC12 concentration would be needed to over- come the chelation effect induced by a high poly(P) con- centration. In the case of acetyl-CoA synthetic reaction, we added 30 mM poly(P) in the reaction mixture with 24mM MgCI2. However, the excess amount of MgC12 formed a precipitate with poly(P), and the reaction rate decreased (Fig. 8). Thus, optimization of the balance of poly(P) and Mg 2+ concentrations is necessary to achieve an efficient regeneration reaction.

Taking together these observations, we tried to con- struct a PAP-PPK ATP regeneration system using ACS. In acetyl-CoA synthesis, 9.95 mM acetyl-CoA was synthe- sized from 10ram CoA, 11.3 mM acetate and 0.25mM AMP after 18 h of incubation (Table 2). This means that 99.5% of the substrate was converted to the product and that ATP was approximately 39.8times regenerated from AMP. These results indicate that the PAP-PPK ATP regeneration system works well in the acetyl-CoA synthetic reaction. Since acetyl-CoA is about five times more expensive than CoA, this reaction can be used for inexpensive in vitro synthesis of acetyl-CoA.

A possible problem with this ATP regeneration system is the direct inhibitory effect of poly(P) on enzymes. Since poly(P) is highly negatively charged, it may bind to enzymes and inhibit their activities. This inhibition could be dependent on the characteristics of the enzymes used in reactions. In the acetyl-CoA synthetic reaction, no inhibition was observed up to 30 mM poly(P) (data not shown).

In the industry, poly(P) is a promising candidate as a new phosphate donor because of its stability and low

VOL. 91, 2001 PAP-PPK ATP REGENERATION SYSTEM 563

cost. Recently, three methods for ATP regeneration from AMP using poly(P) have been presented. Ishige and Noguchi demonstrated that PPK and AdK form a complex in the presence of polyphosphate, resulting in PAP activity in E. cog (29). Although they did not show an experimental result of the ATP regeneration system using PPK and AdK, it would also be possible to re- generate ATP from AMP. The use of a combination of PAP and AdK for ATP regeneration from AMP has been reported by Resnick and Zehnder (18). They applied their regeneration system to glucose-6-phosphate (G-6-P) production with hexokinase (HK) using 100 mM glucose. According to their results, approximately 16mM ATP was produced from 5 mM AMP for 16 mM G-6-P synthesis. In their system, ATP seems to be much less efficiently re- generated (around 3 to 4 times) that it is in the PAP-PPK regeneration system (39.8 times regenerated). In addition, the reaction appeared to be stopped when only 16% of glucose was converted to G-6-P. Although further ana- lyses are needed to evaluate the efficiency of both ATP regeneration systems, it is considered that the PAP-PPK ATP regeneration system is not inferior to other ATP regeneration systems.

A comparison of these poly(P)-dependent ATP re- generation systems showed the possible advantage of the PAP-PPK ATP regeneration system (Fig. I0). Since PAP can catalyze GDP formation from GMP, it was possible to regenerate GTP from GMP using this system. If a new enzyme that catalyzes the conversion of pyrimidine mono-phosphate to pyrimidine diphosphate can be found or constructed by mutagenesis; NTP regeneration from NMP could be achieved by coupling with PPK.

REFERENCES

1. Kim, D.M. and Swarm, J. R.: Prolonging cell-free protein syn- thesis with a novel ATP regeneration system. Biotechnol. Bioeng., 66, 180-188 (1999).

2. Craas, D.C., Kazlanskas, R.J., Hirschbein, B.L., Wong, C.H., Abril, O., and Whitesides, G.M.: Enzymatic regenera- tion of adenosine 5"triphospate: acetylphosphate, phosphoenol- pyruvate, methoxycarbonylphosphate, dihydroxyacetone phos- phate, 5-phospho-alpha-D-ribosyl pyrophosphate, uridine-5'- diphosphoglucose. Methods Enzymol., 136, 236-280 (1987).

3. Kondo, H., Tomioka, l., Nakajima, H., and Imahori, K.: Con- struction of a system for the regeneration of adenosine 5'-tri- phosphate, which supplies energy to bioreactor. J. Appl. Bio- chem., 6, 29-38 (1984).

4. Shin, Y.H. and Whitesides, G.M.: Large-scale ATP-requiring enzymatic phosphorylation of creatine can be driven by en- zymatic ATP regeneration. J. Org. Chem., 42, 4165-4166 (1977).

5. Fujio, T., Nishi, T., Ito, S., and Maruyama, A.: High level expression of XMP aminase in F~cherichia coil and its applica- tion for the industrial production 5"guanylic acid. Biosci. Bio- tech. Biochem., 61, 840-845 (1997).

6. Fujio, T. and Maruyama, A.: Enzymatic production of pyrimi- dine nucleotides using Corynebacterium ammoniagenes cells and recombinant Escherichia coli cells: enzymatic production of CDP-chofine from orotic acid and choline chloride (Part I). Biosci. Biotech. Biochem., 61, 956-959 (1997).

7. Fujio, T., Teshiba, S., and Maruyama, A.: Construction of a plasmid carrying both CTP synthetase and a fused gene formed from cholinephosphate cytidylyltransferase and choline kinase genes and its application to industrial CDP-choline production: enzymatic production of CDP-chofine from orotic acid (Prat II). Biosci. Bintech. Binchem., 61, 960-964 (1997).

8. Kitabatake, S., Dombou, M., Tomioka, I., and Nakajima, H.: Synthesis of P1, P4-di(adenosine 5") tetraphosphate by leucyl- tRNA synthetase, coupled with ATP regeneration. Biochem.

Biophys. Res. Commun., 146, 173-178 (1987). 9. Murata, K., Uehida, T., Kato, J., and Chibata, I.: Polyphosphate

kinase: distribution, some properties and its application as an ATP regeneration system. Agric. Biol. Chem., 52, 1471-1477 (1988).

10. Hoffman, R. C., Jr., Wyman, P.L., Smith, L.E., Nolt, C.L., Conley, J. L., Hevel, J. M., Warren, J. P., Reiner, G.A., and Moe, O.A., Jr.: Immobilized polyphosphate kinase: prepara- tion, properties, and potential for use in adenosine 5'-triphos- phate regeneration. Biotechnol. Appl. Biochem., 10, 107-117 (1988).

11. Kulaev, I.S.: The biochemistry of inorganic polyphosphates. John Wiley and Sons Inc., New York (1979).

12. Shiba, T., Tsutsumi, K., Ishige, K., and Noguehi, T.: Inorganic polyphosphate and polyphosphate kinase: their novel biological functions and applications. Biochemistry (Moscow), 65, 315- 323 (2000).

13. Komberg, A.: Inorganic polyphosphate: toward making a forgotten polymer unforgettable. J. Bacteriol., 177, 491-496 (1995).

14. Kulaev, I., Vagabov, V., and Kulakovskaya, T.: New aspects of inorganic polyphosphate metabolism and function. J. Bio- sci. Bineng., 88, 111-129 (1999)

15. Ahn, K. and Kornberg, A.: Polyphosphate kinase from Esch- erichia coli. Purification and demonstration of a phosphoen- zyme intermediate. J. Biol. Chem., 265, 11734-11739 (1990).

16. Kuroda, A. and Komberg, A.: Polyphosphate kinase as a nucleoside diphosphate kinase in Escherichia coli and Pseudom- onas aeruginosa. Proc. Nail. Acad. Sci., 94, 439--442 (1997).

17. Noguchi, T. and Shiba, T.: Use of Escherichia coli poly- phosphate kinase for oligosaccharide synthesis. Biosci. Biotech. Binchem., 62, 1594-1596 (1998).

18. Resniek, S.M. and Zehnder, A. J.: In Vitro ATP regeneration from polyphosphate and AMP by polyphosphate: AMP phos- photransferase and adenylate kinase from Acinetobacter john- sonii 2IOA. Appl. Environ. Microbiol., 66, 2045-2051 (2000).

19. van Groenesti~n, J.M., Bentvelsen, M.M., Deinema, M.H., and Zehnder, A.J.: Polyphosphate-degrading enzymes in Aci- netobacter spp. and activated sludge. Appl. Environ. Micro- biol., 55, 219-223 (1989).

20. Bonting, C.F., Kortstee, G.J., and Zelmder, A. J.: Properties of polyphosphate: AMP phosphotransferase of Acinetobacter strain 210A. J. Bacteriol., 173, 6484--6488 (1991).

21. Bonting, C.F., Kortstee, G. $., and Zehnder, A.J.: Properties of polyphosphatase of Acinetobacter johnsonii 210A. Antorde. Van. Leeuwenhoek, 64, 75-81 (1993).

22. Lu, Q. and Inouye, M.: Adenylate kinase complements nucleo- side diphosphate kinase deficiency in nucleotide metabolism. Proc. Natl. Acad. Sci. USA, 93, 5720-5725 (1996).

23. Akiyama, M., Crooke, E., and Koraberg, A.: The poly- phosphate kinase gene of Escherichia coil Isolation and se- quence of the ppk gene and membrane location of the protein. J. Biol. Chem., 267, 22556-22561 (1992).

24. Kumari, S., Tiehiel, R., Eisenbach, M., and Wolfe, A. J.: Clon- ing, characterization, and functional expression of acs, the gene which encodes acetyl coenzyme A synthetase in Esch- erichia coli. J. Bacteriol., 177, 2878-2886 (1995).

25. Ellman, G.L.: Tissue sulfhydryl groups. Arch. Biochem. Bio- phys., 82, 70-77 (1959).

26. Akiyama, M., Crooke, E., and Kornberg, A.: An Exopoly- phosphatase of Escherichia coll. J. Biol. Chem., 268, 633- 639 (1993).

27. Wurst, H. and Komberg, A.: A soluble exopolyphosphatase of Saccharomyces cerevisiae purification and characterization. J. Biol. Chem., 269, 10996-11001 (1994).

28. Sambrook, J., Fritseh, E. F., and Maniatis, T.: Molecular clon- ing: a laboratory manual, 2nd ed. Cold Spring Harbor Labora- tory Press, Cold Spring Harbor, New York (1989).

29. lshige, K. and Noguehi, T.: Inorganic polyphosphate kinase and adenylate kinase participate in the polyphosphate: AMP phosphotransferase activity of Escherichia coli. Proc. Natl. Acad. Sci. USA, 97, 14168-14171 (2000).