Analytical Biochemistry 269, 353–358 (1999) Article ID abio.1999.4042, available online at http://www.idealibrary.com on
Photometric Assay for Polynucleotide Phosphorylase Laura Fontanella,* Sabrina Pozzuolo,* Alessandra Costanzo,* Rebecca Favaro,† Gianni Deho`,† and Paolo Tortora* ,1 *Department of General Physiology and Biochemistry and †Department of Genetics and Biology of Microorganisms, University of Milan, Via Celoria 26, I-20133 Milano, Italy
Received November 30, 1998
Polynucleotide phosphorylase (PNPase) is a prokaryotic enzyme that catalyzes phosphorolysis of polynucleotides with release of NDPs. It is also believed to play a key role in turnover of prokaryotic transcripts, thus regulating gene expression. At the moment, only radioisotopic methods are available for assaying PNPase in crude extracts; these involve incubating [ 32 P]phosphate and poly(A) in the presence of the enzyme, separating [ 32 P]phosphate from [ 32 P]ADP, and quantifying ADP by scintillation counting. Photometric assay using pyruvate kinase and lactate dehydrogenase as auxiliary enzymes is not feasible in crude extracts because of endogenous ATPase activities, which regenerate ADP from the ATP released by pyruvate kinase. Here, we present a simple photometric assay that uses a cyclic detection system which, due to the sequential action of pyruvate kinase and hexokinase, results in an exponential increase of ADP and glucose 6-phosphate. Glucose 6-phosphate is then revealed by a glucose6-phosphate dehydrogenase reaction. Based on the theoretical model, a linear increase in absorbance is predicted as a function of the square of the reaction time, with a slope proportional to PNPase activity. Experimental data confirmed the theoretical predictions and showed that the assay was quantitative and unquestionably specific. We also devised a simple procedure for determining absolute enzyme activities (expressed in micromoles of product formed per minute) using exact amounts of pure PNPase as internal standards. © 1999 Academic Press Key Words: polynucleotide phosphorylase; enzyme assay; Escherichia coli; Pseudomonas putida.
To whom correspondence should be addressed. Fax: 139-022362451. E-mail:
[email protected]. 1
0003-2697/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
Polynucleotide phosphorylase (PNPase; EC 2.7.7.8) 2 is a prokaryotic enzyme catalyzing (a) polymerization of ribonucleoside diphosphates with release of phosphate, (b) the reverse reaction, namely the phosphorolytic cleavage of polyribonucleotides, and (c) an exchange reaction between [ 32 P]phosphate and the b-phosphate of ribonucleoside diphosphates. Since the pioneering investigations of Grunberg-Manago and of Ochoa (see, e.g., Refs. 1 and 2), considerable work has been carried out to identify the catalytic mechanism and understand its physiological role(s). The purified enzyme proved very useful in the synthesis of definedsequence oligonucleotides (3). Furthermore, recently there has been renewed interest in PNPase, as this enzyme would appear to play a key role in the turnover of prokaryotic transcripts, thus regulating gene expression (4 –10). It has also been shown that in Escherichia coli PNPase, which is encoded by the pnp gene (11, 12), is a component of a multienzyme complex, the degradosome, the other identified components being ribonuclease E, an RNA helicase, enolase, and polyphosphate kinase (4 – 8). In this complex, PNPase processively degrades mRNAs form the 39 end. Finally, PNPase was also used to solve the structure of the S1 RNA binding domain, a conserved feature among several RNA binding proteins (13). Despite this growing interest, the only currently available methods for assaying PNPase in crude extracts are cumbersome and time-consuming. Basically, they involve incubating [ 32P]phosphate and poly(A) in the presence of the enzyme, subsequent separation of [ 32P]phosphate from [ 32P]ADP by an extraction procedure, and quantification of ADP by scintillation counting (14). By contrast, when assaying a purified enzyme, a much simpler, continuous enzyme assay could be adopted: in this case, the activity may be detected in a 2 Abbreviations used: polynucleotide phosphorylase, PNPase; polyadenylic acid, poly(A); PEP, phosphoenolpyruvate; G6PD, glucose-6-phosphate dehydrogenase.
353
354
FONTANELLA ET AL.
mixture containing poly(A), phosphate, phosphoenolpyruvate, NADH, and suitable amounts of pyruvate kinase and lactate dehydrogenase. On reaction with phosphoenolpyruvate, ADP gives rise to ATP and pyruvate; this latter compound is then reduced to lactate with stoichiometric formation of NAD 1, allowing the reaction kinetics to be monitored by measuring the decrease in absorbance at 340 nm (15, 16). The renewed interest in PNPase demands simpler assays to be used for crude extracts to facilitate the study of this enzyme under different growth conditions, in addition to screening and isolating new mutants. This would greatly help in further defining its physiological role(s), as well as its biochemical and structural properties. To develop a simple photometric assay for crude extracts, we first addressed the issue of why the procedure described by Godefroy et al. (15, 16) was only feasible with purified enzyme. We discovered that, due to ATPase activity in E. coli extracts, the ATP produced by pyruvate kinase was reconverted into ADP with continuous recycling between the two compounds. This resulted in an exponential increase in absorbance, which made it impossible to make a reliable assessment of PNPase activity by conventional kinetic measurements. In this paper, we present a new, simple photometric assay that relies on a cyclic reaction system which, thanks to the sequential action of pyruvate kinase and hexokinase, also results in an exponential increase in ADP and glucose 6-phosphate. The latter compound is then revealed by a glucose-6phosphate dehydrogenase reaction. Based on the theoretical background, it is predicted that, under our experimental conditions, there will be a linear increase in absorbance as a function of the square of the reaction time, with the slope of this profile proportional to PNPase activity, and that interference by endogenous ATPase will be minimized. Experimental data confirmed the theoretical model. The method also allowed determination of absolute enzyme activity.
Bacterial Strains and Growth Conditions The E. coli C strains C-5612 (pnp-7;Tn10) and C-5641 (pnp::Tn5) were previously reported (10); C-5613 is the isogenic pnp 1;Tn10 strain of C-5612, and was constructed as described (10). The pnp-7 mutant is defective in PNPase activity (17) and carries a nonsense (TGA) mutation, as determined by sequencing (Zangrossi et al., manuscript in preparation); plasmid pAZ8 (10) carries the pnp 1 gene. Pseudomonas putida KT2440 was described by Bagdasarian et al. (18); its pnp-1 (INS pKRF7A) derivative PPM-103 (Favaro et al., manuscript in preparation) was constructed as follows: a DNA fragment internal to the P. putida pnp gene (coordinates 889-2473, GenBank Accession number Y18132) and containing the V-Km interposon (19) was cloned in the suicide plasmid vector pKNG101 (20). The resulting plasmid pKRF7A was transferred by conjugation into P. putida KT2440 and exconjugants with the pnp gene interrupted by the insertion of the plasmid were selected and characterized by Southern blot and PCR analysis. Bacterial cultures were grown in 2.18 % K 2HPO 4, 1.7% KH 2PO 4, 0.5% yeast extract (Difco), 0.6% glucose (14) with aeration to the late exponential phase (OD 650 1.6 –1.7) at either 37°C (E. coli) or 30°C (P. putida). Production of Cell Extracts At the end of growth, cells were harvested by centrifugation at 6000g for 15 min, washed with 50 mM Tris–HCl, pH 7.4, and resuspended in the same buffer to a density of 200 mg/ml. Aliquots (150-mg) were disrupted by vigorous shaking with 1-g glass beads (212–300 mm diameter) for five 1-min periods with idle intervals of 1 min at 0°C. The homogenates were then incubated with 6 units of bovine pancreas DNase for 10 min at 37°C and centrifuged at 12,000g for 20 min at 4°C. Supernatants were extensively dialyzed against 50 mM Tris–HCl, pH 7.4, aliquoted, and stored frozen at 220°C. Protein was assayed by the Coomassie Plus protein assay reagent (Pierce, Rockford, IL) using bovine plasma g-globulin as standard protein.
MATERIALS AND METHODS
Reagents The following enzymes (supplied by Sigma, St. Louis, MO) were used in the PNPase assays reported here: type II rabbit muscle pyruvate kinase, type II rabbit muscle L-lactate dehydrogenase, hexokinase from a yeast overproducing strain, recombinant glucose-6phosphate dehydrogenase from Leuconostoc mesenteroides, and PNPase from E. coli. All other biochemicals were also supplied by Sigma. Prior to the assays, poly(A) was extensively dialyzed against 5 mM Tris– HCl, pH 7.4. PNPase was also purified in our laboratory. The purification procedure will be published elsewhere.
Productions of Antibodies Pure E. coli PNPase was used to elicit polyclonal antibodies in rabbits using standard immunological methods (21). Enzyme Assay Unless otherwise stated, the assay devised in this work was performed on a mixture containing, in a volume of 1.5 ml, 50 mM Tris–HCl, pH 7.4, 0.1 M KCl, 5 mM MgCl 2, 20 mg/ml poly(A), 1.5 mM phosphoenolpyruvate, 20 mM glucose, 0.5 mM NAD 1, 0.6 U/ml pyruvate kinase, 2 U/ml hexokinase, 4 U/ml glucose-6-phosphate dehydrogenase, and suitable amounts of cell extracts (typically, 20 to 80
PHOTOMETRIC POLYNUCLEOTIDE PHOSPHORYLASE ASSAY
mg protein) and/or pure enzyme. Stock solutions of pyruvate kinase were prepared by a twofold dilution in the assay buffer of the commercially available ammonium sulfate suspension. A clear solution was thus obtained, in which pyruvate kinase was stable over several months. This made it possible to add exact amounts of enzyme to each assay mixture, an essential prerequisite for the reproducibility of the test (see “Theoretical background”). The assay mixtures were prepared immediately before use. Absorbance was continuously recorded at 340 nm and 25°C. After recording the baseline for about 10 min, the reaction was started by addition of 0.75 M phosphate, pH 7.4, to attain a 10 mM final concentration. The slight linear increase that took place before phosphate addition was extrapolated and subtracted from the exponential increase detected thereafter. Alternatively, absorbance of the blank (without phosphate added) was directly subtracted using a dual beam spectrophotometer. Absorbance was then plotted at intervals of 2 min as a function of the square of the reaction time. In such plots, after a short lag time, profiles were linear for at least 15 min, whereas later a decline in linearity was observed. The slope was determined using a suitable linear regression method of an Enzfitter Elsevier-Biosoft program. Pure PNPase was also assayed as described by Godefroy (15) with minor modifications. In a volume of 1.5 ml, the assay mixture contained 50 mM Tris–HCl, pH 7.4, 0.1 M KCl, 5 mM MgCl 2, 20 mg/ml poly(A), 1 mM phosphoenolpyruvate, 0.1 mM NADH, 1.3 U/ml pyruvate kinase, 10 U/ml lactate dehydrogenase, and suitable amounts of pure enzyme. The reaction was started by addition of phosphate to a 10 mM final concentration. Absorbance was continuously recorded at 340 nm and 25°C. In both assays, poly(A) and phosphate concentrations were largely saturating and each activity value was obtained from the mean of at least three independent determinations. Standard deviations never exceeded 10% of the respective mean values. One unit of enzyme activity is defined as the amount of enzyme which releases 1 mmol of ADP/min under the working conditions. Theoretical Background The assay relies on the following reaction scheme: PNPase
poly(A)n 1 Pi | - 0 poly(A)n21 1 ADP
355
The rate of appearance of ADP is given by the following equation: d[ADP] 5 A PN 2 A PK 1 A HK , dt
[1]
where t is the reaction time, A PN is PNPase activity, A PK is the rate of ADP subtraction by pyruvate kinase, and A HK is the rate of ADP regeneration by hexokinase. Since pyruvate kinase activity is limiting with respect to that of hexokinase, A HK equals A PK, so the Eq. [1] may be rewritten as follows: d[ADP] 5 A PN dt
[2]
whose integrated form is: [ADP] 5 A PNt 1 [ADP]0,
[3]
where [ADP] 0 is the starting ADP concentration, when t 5 0. However, [ADP] 0 5 0 under our working conditions, as we subjected cell extracts to extensive dialysis prior to assay. Thus, the Eq. [3] may be simplified as follows: [ADP] 5 A PNt
[4]
Furthermore, the overall reaction rate, detected by measuring the appearance of NADH, equals the rate of the step catalyzed by pyruvate kinase, due to the fact that the activity of this enzyme is limiting with respect to both hexokinase and glucose-6-phosphate dehydrogenase. If one also assumes that the Michaelis constant of pyruvate kinase for ADP (0.3 mM) (22) is much higher than the concentration of the nucleotide under the assay conditions, then the reaction rate will be given by: d[NADH] V PK [ADP], 5 dt K PK
[5]
where V PK and K PK are the apparent maximal velocity and Michaelis constant, respectively. If the Eq. [4] is replaced into the Eq. [5], it follows that:
pyruvate kinase
ADP 1 PEP O ¡ ATP 1 pyruvate hexokinase
glucose 1 ATP O ¡ glucose 6-phosphate 1 ADP
d[NADH] V PK A t 5 dt K PK PN
[6]
Integrating this equation gives:
G6PD
glucose 6-phosphate 1 NAD 1 O ¡ 6-phosphogluconate 1 NADH
[NADH] 5
V PK A PN 2 t 1 [NADH]0 2K PK
[7]
356
FONTANELLA ET AL.
Since at zero time NADH is absent, the Eq. [7] may be rewritten as follows: [NADH] 5
V PK A PN 2 t 2K PK
[8]
This relationship states that, when plotting [NADH] versus t 2, a linear increase of [NADH] should be observed, with a slope proportional to both PNPase and pyruvate kinase activities. All these predictions were confirmed by our results, which support the theoretical model. In addition, endogenous ATPase should not cause any appreciable interference because its activity was substantially lower than that of hexokinase added to the assay mixture. RESULTS AND DISCUSSION
Initial attempts aimed at determining PNPase activity in E. coli crude extracts were performed using the pyruvate kinase/lactate dehydrogenase detection system. However, using this method an exponential decrease in absorbance as a function of time was seen. In addition, initial reaction rates showed a strong variability, which prevented us from reliably determine enzyme activity (data not shown). This finding also justifies why only radioisotopic assays were used to date in crude extracts. This interference might be due to ATPase activities, which results in the reconversion of ATP released by pyruvate kinase into ADP. As expected, we detected indeed significant ATPase activity in crude extracts (data not shown). The cyclic reaction system that we subsequently developed to assay PNPase was devised to avoid this interference by taking advantage of continuous recycling between ADP and ATP, which also should result in an exponential increase in absorbance. Based on the theoretical background (see Materials and Methods), a plot of absorbance as a function of the square of the reaction time should produce linear profiles, with slopes proportional to both PNPase and pyruvate kinase activities. Moreover, removal of ATP by endogenous ATPase should be negligible due to the much higher activity of hexokinase added as auxiliary enzyme. Our experimental results confirm indeed the theoretical forecasts because the increase in absorbance was exponential as a function of time (Fig. 1A) and linear as a function of the square of time (Fig. 1B). Furthermore, as expected the quantity [NADH] z min 22 was directly proportional to the amount of crude extract, up to about 80 mg of total protein (Fig. 1C). In keeping with the theoretical predictions, pyruvate kinase activity up to 1 U/ml under our working conditions was also directly proportional (data not shown). This implies that the sensitivity of the assay might be suitably increased by increasing the amount of this auxiliary enzyme.
FIG. 1. Determination of PNPase activity by the cyclic reaction system. The absorbance was monitored continuously at 340 nm as a function of the reaction time (A); values recorded at intervals of 2 min were also plotted as a function of the square of the reaction time (B). The indicated amounts of crude extract (expressed as micrograms of protein) of the wild-type E. coli C-5613 strain were added to the reaction mixtures. In C, the slopes of the profiles shown in B are given as a function of micrograms of protein.
To check the specificity of our assay, we determined PNPase activity both in E. coli and in P. putida wildtype and pnp mutant strains. Our assay did not reveal any measurable activity in the E. coli strains C-5641 and C-5612, which carry a nonsense and a Tn10 insertion mutation in the pnp gene, respectively, nor in the P. putida strain PPM-103 in which pnp is interrupted by the insertion of a plasmid; in contrast, PNPase activity was detected in the wild-type P. putida extract (Table 1). We also assayed enzyme activity in a crude extract from the wild-type E. coli strain C-5613 after incubation either with an antiserum anti-PNPase or with the preimmune serum. After pretreatment with the antiserum, the assay did not reveal any appreciable activity, which in contrast was not significantly reduced by pretreatment with the preimmune serum (Table 1). Western blots previously performed on crude extracts of the E. coli strain C-5613/pAZ8 using antiPNPase antibodies showed a level of immunologically reactive protein about 3- to 4-fold higher than in wild-
357
PHOTOMETRIC POLYNUCLEOTIDE PHOSPHORYLASE ASSAY
type strains (data not shown); thus, we also assayed PNPase activity in this strain. In line with the immunological determinations, it showed activity that was about 3-fold higher activity than the wild type (Table 1). On the whole, these results clearly indicate that our assay is strictly specific for PNPase activity. One possible drawback of the assay is that it does not allow a direct determination of the absolute activity (expressed as micromoles of product formed per minute). To overcome this problem we first determined enzyme activity of samples of pure PNPase by the pyruvate kinase/lactate dehydrogenase assay; then we employed such samples as internal standards in crude extracts using the cyclic reaction system. In this way, activity expressed as slope (Fig. 1C; Table 1) could be converted in absolute values. Under the standard assay conditions, our measurements produced a conversion factor of 0.97 mU z ml 21/mM NADH z min 22 (Table 2). This corresponds to an overall change in absorbance, after 10 min of reaction, 10- to 20-fold higher in the cyclic reaction system than in the pyruvate kinase/ lactate dehydrogenase assay using equal amounts of PNPase. This highlights the remarkable sensitivity of the cyclic assay method. These experiments also showed that the activity of pure PNPase was slightly underestimated in crude extracts (Table 2). This is probably due to endogenous adenylate kinase: in fact, this enzyme converts ADP into ATP and AMP, and
TABLE 1
PNPase Assay in E. coli and P. putida Strains Using the Cyclic Reaction System
E. coli C-5613 (wild type) E. coli C-5613 plus antiserum a E. coli C-5613 plus preimmune serum a E. coli C-5641 (pnp::Tn5) E. coli C-5612 (pnp-7) E. coli C-5613/pAZ8 P. putida KT2440 (wild type) P. putida PPM-103 (pnp-1)
Sample assayed Pure PNPase (0.097 mU/ml) E. coli strain C-5613 (22 mg protein) E. coli strain C-5613 (20 mg protein) 1 pure PNPase (0.097 mU/ml) E. coli strain C-5613 (44 mg protein) E. coli strain C-5613 (44 mg protein) 1 pure PNPase (0.097 mU/ml)
[NADH] z time 22 (mM z min 22) 0.122 0.108 0.210 (0.102) 0.210 0.310 (0.100)
Note. Activities, expressed as increase of mM NADH concentration per min 2, were measured using different amounts of crude extract from the wild-type strain C-5613 before and after addition of an amount of pure PNPase equivalent to 0.097 mU/ml, as assessed using the pyruvate kinase/lactate dehydrogenase detection system. The increase in activity resulting from addition of pure enzyme is given in parentheses.
this latter compound is not revealed by the detection system. In conclusion, our new assay satisfies the two basic requirements which are essential for any enzyme assay, i.e., specificity and linearity. In addition, this method is remarkably sensitive and far less cumbersome and time-consuming than the radioisotopic tests available at present; it therefore represents a valuable tool for investigations into the genetic and physiological role of PNPase and on its involvement in turnover and stability of prokaryotic transcripts. ACKNOWLEDGMENTS
Activity
Strain
TABLE 2
Determination of Absolute PNPase Activity in E. coli Extracts
[NADH] z time 22 (mM z min 22)
% of wild type
0.217 0.004
100 2
0.212
98
bd b bd b 0.694 0.049 bd b
bd b bd b 320 100 bd b
Note. Activities are expressed either as mM increase of NADH concentration per min 2, or as percentage of the activity found in the respective wild-type strains. 45 mg of crude extract protein was used for each determination. a Crude extracts (230 mg protein in 50 ml) were preincubated for 2 h at room temperature with 6 ml of either antiserum or preimmune serum. The immunoprecipitate was removed by centrifugation and the activity determined in an amount of supernatant equivalent to 45 mg of crude extract protein. b Below detection limit.
This work was supported by grants from the Consiglio Nazionale delle Ricerche, Rome, Grants 97.01028.49 (Progetto Finalizzato Biotecnologie Ambientali) and 97.04298.CT04. Thanks are due to Dr. C. Portier, who provided anti-PNPase antibodies in the early stages of this work.
REFERENCES 1. Singer, M. F., Hilmoe, R. J., and Grunberg-Manago, M. (1960) J. Biol. Chem. 235, 2705–2712. 2. Ochoa S., and Mii S. (1961) J. Biol. Chem. 236, 3303–3311. 3. Gillam, S., and Smith, M. (1980) Methods Enzymol. 65, 687– 701. 4. Carpousis, A. J., Van Houwe, G., Ehretsmann, C. and Krisch, H. M. (1994) Cell 76, 889 –900. 5. Hajnsdorf, E., Steier, O., Coscoy, L., Teysset, L., and Re´gnier, P. (1994) EMBO J. 13, 3368 –3377. 6. Py, B., Causton, H., Mudd, E. A., and Higgins, C. F. (1994) Mol. Microbiol. 14, 717–729. 7. Zilha˜o, R., Cairra˜o, F., Re´gnier, P., and Arraiano, C. M. (1996) Mol. Microbiol. 20, 1033–1042. 8. Blum, E., Py, B., Carpousis, A. J., and Higgins C. F. (1997) Mol. Microbiol. 26, 387–398.
358
FONTANELLA ET AL.
9. Luttinger, A., Hahn, J., and Dubnau, D. (1996) Mol. Microbiol. 19, 343–356. 10. Piazza, F., Zappone, M., Sana, M., Briani, F., and Deho`, G. (1996) J. Bacteriol. 178, 5513–5521. 11. Reiner, A. M. (1969) J. Bacteriol. 97, 1431–1436. 12. Regnier, P., Grunberg-Manago, M., and Portier, C (1987) J. Biol. Chem. 262, 63– 68. 13. Bycroft, M., Hubbard, T. J., Proctor, M., Freund, S. M., and Murzin, A. G. (1997) Cell 88, 235–242. 14. Soreq, H., and Littauer, U. Z. (1977) J. Biol. Chem. 252, 6885– 6888. 15. Godefroy, T., (1970) Eur. J. Biochem. 14, 222–231.
16. Bon, S., Godefroy, T., and Grunberg-Manago, M. (1970) Eur. J. Biochem. 16, 363–372. 17. Reiner, A. M. (1969) J. Bacteriol. 97, 1437–1443. 18. Bagdasarian, M., Lurz, R., Ruckert, B., Franklin, F. C., Bagdasarian, M. M., Frey, J., and Timmis, K. N. (1981) Gene 16, 237–247. 19. Fellay, R., Frey, J., and Krisch, H. (1987) Gene 52, 147–154. 20. Kaniga, K., Delor, I., and Cornelis, G. R. (1991) Gene 109, 137–141. 21. Harlow, E., and Lane, D. (1988) Antibodies: A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. 22. Bergmeyer, H. U. (1974) Methods of Enzymatic Analysis, VCH, Weinheim; Academic Press, New York/London.
