Polyphosphate Chain Length Determination in the Range of Two to

May 13, 2019 - ... E. N. Dual role of inorganic polyphosphate in cardiac myocytes: The ..... activity of DNA and polyphosphate isolated using commerci...
2 downloads 0 Views 1MB Size
Download PDF
Article Cite This: Anal. Chem. 2019, 91, 7654−7661

pubs.acs.org/ac

Polyphosphate Chain Length Determination in the Range of Two to Several Hundred P‑Subunits with a New Enzyme Assay and 31P NMR Jonas Johannes Christ,† Sabine Willbold,‡ and Lars Mathias Blank*,† †

Downloaded via UNIV OF SOUTHERN INDIANA on July 29, 2019 at 10:02:43 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

Institute of Applied Microbiology − iAMB, Aachen Biology and Biotechnology − ABBt, Worringer Weg 1, RWTH Aachen University, Aachen D-52074, Germany ‡ Central Institute for Engineering, Electronics and Analytics, Analytics (ZEA-3), Wilhelm-Johnen-Straße, Jülich D-52428, Germany ABSTRACT: Currently, 31P NMR is the only analytical method that quantitatively determines the average chain length of long inorganic polyphosphate (>80 P-subunits). In this study, an enzyme assay is presented that determines the average chain length of polyphosphate in the range of two to several hundred P-subunits. In the enzyme assay, the average polyP chain length is calculated by dividing the total polyphosphate concentration by the concentration of the polyphosphate chains. The total polyphosphate is determined by enzymatic polyphosphate hydrolysis with Saccharomyces cerevisiae exopolyphosphatase 1 and S. cerevisiae inorganic pyrophosphatase 1, followed by colorimetric orthophosphate detection. Because the exopolyphosphatase leaves one pyrophosphate per polyphosphate chain, the polyphosphate chain concentration is assayed by coupling the enzymes exopolyphosphatase (polyP into pyrophosphate), ATP sulfurylase (pyrophosphate into ATP), hexokinase (ATP into glucose 6phosphate), and glucose 6-phosphate dehydrogenase (glucose 6-phosphate into NADPH), followed by fluorometric NADPH detection. The ability of 31P NMR and the enzyme assay to size polyP was demonstrated with polyP lengths in the range from 2 to ca. 280 P-subunits (no polyP with a longer chain length was available). The small deviation between methods (−4 ± 4%) indicated that the new enzyme assay performed accurately. The limitations of 31P NMR (i.e., low throughput, high sample concentration, expensive instrument) are overcome by the enzyme assay that is presented here, which allows for high sample throughput and requires only a commonly available plate reader and micromole per liter concentrations of polyphosphate.

I

norganic polyphosphate (polyPn, where n is the average chain length) is the linear polymer of orthophosphate (Pi) with a chain length of at least 2 and up to 1000 P-subunits, and is found in nearly all living organisms (refs 1−3; an exception is, for example, Enterococcus faecalis1). Cyclic polyP, with a chain length of 3−12 P-subunits, can be synthesized chemically and is formed spontaneously during hydrolysis of long-chain linear polyP.2 In comparison with linear polyP, cyclic polyP has no hydroxyl group with neutral pKa value at the end groups, and does not possess the (biotechnologically valuable) higher valent cation-complexing ability. In biological systems, no function of cyclic polyP has been found yet.2 Because the physicochemical properties of polyP depend on its chain length,4,5 polyPs of varying chain lengths have specific functions in vivo (ref 6, p 125ff in ref 2). Furthermore, different industrial applications of polyP require defined average polyP chain lengths.4 Thus, both the total polyP concentration and the average polyP chain length have garnered interest in studies on polyP.7 It should be noted that other important properties of polyP, such as the chain length distribution and counterion composition, are not covered in this study. An enzymatic method for specific total polyP determination is available (abbreviated as the PPX/IPP assay8). For polyP chain length determination, seven methods exist: end group titration, ion chromatography, capillary electrophoresis, the © 2019 American Chemical Society

PPX/IPP assay, polyacrylamide gel electrophoresis (PAGE), size exclusion chromatography, and 31P nuclear magnetic resonance (31P NMR). End group titration is limited to pure polyP and gram amounts thereof, rendering this method unsuitable for the analysis of biological samples.9 Ion chromatography and capillary electrophoresis are restricted to polyP with a maximum chain length of 30 and 45 Psubunits, respectively.9−12 The PPX/IPP assay is based on the differential hydrolysis of a polyP sample with the enzyme Saccharomyces cerevisiae exopolyphosphatase 1 (scPpx1p; EC 3.6.1.11; reaction 1 in Figure 1), and a mixture of scPpx1p and S. cerevisiae inorganic pyrophosphatase 1 (scIpp1p; EC 3.6.1.1; reaction 2 in Figure 1). Subsequently, the released Pi is determined colorimetrically. The PPX/IPP assay, however, is limited to the determination of a polyP length in the range of 2−80 Psubunits.8 In PAGE, the polyP is separated on a polyacrylamide gel by an electric current and subsequently stained.13 PAGE provides information on the chain length distribution. However, PAGE requires a sequencing apparatus, is time-consuming, allows for Received: January 31, 2019 Accepted: May 13, 2019 Published: May 13, 2019 7654

DOI: 10.1021/acs.analchem.9b00567 Anal. Chem. 2019, 91, 7654−7661

Article

Analytical Chemistry

Figure 1. Mechanism of the enzyme assay. The present method for the determination of the average polyP chain length is divided into assays for the quantification of the total polyP concentration (A) and the polyP chain concentration (B). The combination of both values can be used to calculate the average polyP chain length. Individual reactions (reactions 1−3, 5−8) and the overall reactions (reactions 4 and 9) are depicted. Red arrows indicate cascade reactions.

Figure 2. Equations used in this study.

PAGE, size exclusion chromatography provides information on the chain length distribution, and is dependent on polyP size standards.15−18

the analysis of only ca. 20 samples at a time, and gives only nonquantitative results. With calibrated standards, the average polyP chain length can be estimated with PAGE.13,14 Similar to 7655

DOI: 10.1021/acs.analchem.9b00567 Anal. Chem. 2019, 91, 7654−7661

Article

Analytical Chemistry In 31P NMR, polyP gives three resonance peaks: terminal P (PP1) at ca. −7 ppm, P at positions two (PP2) and three (PP3) in the polyP chain at ca. −20.17 to −21.7 ppm, and P at position four or higher in the polyP chain (PP4) at ca. −22.5 ppm (ref 19; p 27 in ref 2). 31P NMR differentiates between linear polyP, cyclic polyP, and Pi.20,21 Lindner et al. proposed a formula to calculate the average polyP chain length from the peak areas of PP1−PP4 (eq 1 in Figure 220). To overcome the analytical challenge of peak overlap, the formula uses PP1, PP2, or PP3 in the denominator, depending on which has the best resolution. With NMR, the average polyP chain length is quantitatively determined. However, NMR requires a special instrument, is limited to the analysis of one sample at a time, and is completed in several minutes to an hour per sample. Sometimes, PP1 cannot be resolved due to the overlap of β-P and γ-P with nucleotide diphosphate and nucleotide triphosphate, respectively. Further, 31P NMR lacks sensitivity for determinations down to micromoles per liter.19 Nevertheless, the ability of 31P NMR to quantitatively size polyP up to hundreds of P-subunits has been, until now, unrivaled. In conclusion, there is a need for an assay that is quantitative, allows for high throughput processing, can quantify in the range of micromoles per liter, is insensitive to biological impurities, and requires only commonly available laboratory equipment. The average polyP chain length can be calculated by dividing the total polyP concentration by the concentration of the polyP chains (eq 2 in Figure 2). These two variables were quantified in this study by two enzymatic assays (Figure 1). Total polyP was quantified according to a recent study (Figure 1A8). Every polyP was hydrolyzed irreversibly by scPpx1p into one polyP2 and (n − 2) Pi (reaction 1 in Figure 1). The molecular reasons for the inability of scPpx1p to cleave polyP2 are discussed in ref 22 (see also ref 23 for a biochemical characterization of scPpx1p). At the same time, polyP2 was irreversibly hydrolyzed by scIpp1p into two Pi’s (reaction 2 in Figure 1). The hydrolysis of a polyP sample with scPpx1p and scIpp1p, and the subsequent colorimetric Pi detection (reaction 3 in Figure 1), allowed for the quantification of the total polyP concentration. Overall, polyPn was quantified as n blue complexes (reaction 4 in Figure 1). The polyP chain concentration was detected fluorometrically (Figure 1B). From every polyP chain, scPpx1p produced one polyP2 (reaction 5 in Figure 1). PolyP2 reacted with adenosine 5′-phosphosulfate to produce ATP and sulfate (catalyzed by ATP sulfurylase; EC 2.7.7.4; reaction 6 in Figure 1). Although the reaction of ATP sulfurylase is reversible, in vitro the enzyme favors the reaction toward ATP (equilibrium constant 10−8, see ref 24). Hexokinase (EC 2.7.1.1) irreversibly transferred one Pi from ATP to glucose (reaction 7 in Figure 1). Glucose 6-phosphate dehydrogenase (EC 1.1.1.49) reversibly produced one fluorescent NADPH per glucose 6phosphate from one nonfluorescent NADP+ (reaction 8 in Figure 125). The NADPH concentration was quantified on the basis of its fluorescence. The overall reaction is shown in reaction 9 in Figure 1. The quantification of polyP2 by coupling ATP sulfurylase, hexokinase, and glucose 6-phosphate dehydrogenase was inspired by Drake et al.26 The first goal of this study was to develop an enzyme assay that allows for the determination of the average polyP chain length in the range of two to several hundred P-subunits. In the second aim, both the performance of the enzyme assay and 31P

NMR for the determination of the polyP chain length was evaluated with polyP’s that comprised 2 to ca. 280 P-subunits.

■

MATERIALS AND METHODS Chemicals and Enzymes. PolyP3 (penta-sodium triphosphate, item number ACRO393961000) was obtained from VWR (Radnor, PA). PolyPs “Budit 7” (average chain length: 138) and “Budit 4” were a kind gift from the Chemische Fabrik Budenheim (Budenheim, Germany). The polyP2 (tetra-sodium diphosphate decahydrate, item number T883), black polystyrene 96-well microtiter plates (item number CEK8), transparent polystyrene 96-well microtiter plates (item number 9293), NADP+ (disodium salt, ≥97% purity, item number AE13), salmon sperm DNA, and RNA sodium salt were obtained from Carl Roth (Karlsruhe, Germany). PolyP’s “p100”, “mode 155”, “mode 385”, and “p700” were a kind gift from Dr. James H. Morrissey (University of Michigan Medical School). The polyP’s with a homogeneous chain length were polyP2 and polyP3. All other polyP’s were a mixture of varying chain lengths. The polyP concentrations (including polyP 2) are reported as Pi. Adenosine 5′phosphosulfate (sodium salt, 95% purity, item number 21226) was purchased from Cayman Chemicals (Ann Arbor, MI). Sterile filtration was performed with 0.2-μm-pore poly(ether sulfone) membranes (Nalgene, Thermo Fisher Scientific, Waltham, MA). ATP-free water was prepared by sterilefiltering and autoclaving Milli-Q water (Merck Millipore, Burlington, MA). ATP sulfurylase (from S. cerevisiae, item number M0394L) was purchased from NEB (Ipswich, MA). Hexokinase (from S. cerevisiae, item number H6380) and glucose 6-phosphate dehydrogenase (from Leuconostoc mesenteroides, item number G8529) were obtained from SigmaAldrich. scPpx1p and scIpp1p were prepared according to ref 8 (NCBI accession numbers L28711 and NP_009565, respectively). scPpx1p, scIpp1p, and glucose 6-phosphate dehydrogenase were dissolved and/or diluted in filter-sterilized 0.1% (w/v) bovine serum albumin (protease free), 50% (v/v) glycerol, 20 mM Tris-Cl, and 50 mM KCl at pH 7.5. ATP sulfurylase was diluted in filter-sterilized 10 mM Tris-Cl, 50 mM NaCl, 0.1 mM dithiothreitol, 0.1 mM EDTA, and 50% (v/v) glycerol at pH 7.5. All enzymes were stored at −20 °C. The specific activities of scIpp1p and scPpx1p were assayed by measuring the Pi that was released from polyP2 and Budit 7, respectively, during a 10 min period. The final assay concentrations were as follows: 5 mM magnesium acetate, 20 mM Tris-Cl, 50 mM ammonium acetate, 1 mM polyP2 or Budit 7, and pH 7.5. The released Pi was quantified colorimetrically, as described below. One unit of scIpp1p was the amount of enzyme that generated 1 μmol of Pi per minute at 37 °C. One unit of scPpx1p activity was defined as the release of 1 pmol Pi per minute at 37 °C.23 The activities of the other enzymes were taken as stated by the manufacturer. Enzymatic Determination of the Total PolyP Concentration. Total polyP was determined by enzymatic hydrolysis with scPpx1p and scIpp1p, and subsequent colorimetric Pi detection according to ref 8. Briefly, the samples for the total polyP analysis were diluted in dilution buffer (filter-sterilized 1 mM sodium 3-(N-morpholino)propanesulfonic acid (MOPS) and 0.1 mM EDTA at pH 7, prepared with ATP-free water). A total of 50 μL of enzyme reaction buffer (15 mM magnesium acetate, 60 mM Tris-Cl, 150 mM ammonium acetate, pH 7.5, filter-sterilized), which 7656

DOI: 10.1021/acs.analchem.9b00567 Anal. Chem. 2019, 91, 7654−7661

Article

Analytical Chemistry

Figure 3. Stepwise protocol for the enzyme assay. The chain concentration is determined in (A) (measuring value) and (B) (blank). The total polyP concentration is quantified in (C) (measuring value) and (D) (blank).

read, which reflected the NADPH concentration. For each sample, a blank without scPpx1p and ATP sulfurylase was measured. Every day, the concentration of the 100 μM polyP2 standard was verified by determining the total polyP. The polyP2 standards (1, 2.5, 5, 7.5, 10 μM polyP2 in dilution buffer) were prepared daily from the 100 μM polyP2 standard, and treated as samples, except that no blank was measured. The fluorescence values were converted to polyP2 concentrations using a calibration curve. The polyP chain concentration was calculated according to eq 4 in Figure 2. As for every polyP in this study, the concentration of polyP2 is reported as the monomer (i.e., μM Pi). However, the chain concentration is equal to the polyP2 polymer concentration (i.e., μM polyP2). To convert between the monomer and the polymer concentrations, a factor of 0.5 was included in eq 4 in Figure 2. The “blank” in eq 4 (Figure 2) refers to a measurement without ATP sulfurylase and scPpx1p, which allows contaminating NADPH, ATP, and glucose 6-phosphate to be subtracted. Enzymatic Determination of the PolyP Concentration, the Average PolyP Chain Length, and the Pi Concentration. The workflow of the enzyme assay that has been developed here is shown in Figure 3. The test was divided into the determination of the polyP chain concentration (Figure 3A and B), and the determination of the total polyP concentration (Figure 3C and D). Figure 3D describes the measurement of the blank of Figure 3C and the P i concentration. Before this assay was performed, a pretest was performed in single measurements, to determine the appropriate dilutions. The sample concentration that was required for Figure 3A was estimated, if the approximate chain length was known (eq 5 in Figure 2). A more wasteful, but time-saving, alternative was to omit the pretest, and directly run the main test with several sample dilutions. In the main test, the blanks and samples were analyzed in triplicate for the total polyP assay and chain concentration assay. The KH2PO4 and polyP2 standards were analyzed in duplicate. The 100 μM polyP2 standard was analyzed in six measurements in the total polyP assay, because this was the

contained 11 000 U scPpx1p and 0.013 U scIpp1p, was added to a 100 μL sample with a concentration of 5−200 μM polyP (optimal: 100 μM polyP) in a transparent microtiter plate. After an incubation at 1 h ± 5 min at 37 °C, 50 μL of Pi detection reagent (2.4 mM ammonium heptamolybdate, 600 mM H2SO4, 0.6 mM antimony potassium tartrate, and freshly added ascorbic acid up to 88 mM) was added. After an incubation for exactly 2 min at room temperature, the absorbance then was read at 882 nm, which indicated the Pi concentration. The KH2PO4 standards (20, 65, 110, 155, and 200 μM Pi in dilution buffer) were treated as samples, except that no blank was measured. For every sample, a blank without scPpx1p and scIpp1p was tested. The blank was equivalent to the Pi concentration of the sample. The Pi concentrations were calculated with a calibration curve. The total polyP minus the blank concentration was calculated per eq 3 in Figure 2. Enzymatic Determination of the PolyP Chain Concentration. Both sodium MOPS buffer (50 mM, pH 7) and TMG buffer (450 mM Tris-acetate, 30 mM magnesium acetate, 12 mM glucose, pH 7.6, storage at −20 °C) were prepared with ATP-free water and filter-sterilized. The aliquots of adenosine 5′-phosphosulfate (2 mM in MOPS buffer), NADP+ (20 mM in MOPS-buffer), and 100 μM polyP2 were stored at −20 °C. The polyP2 detection reagent was freshly prepared by mixing the following per reaction: 25 μL of TMG buffer, 5 μL of adenosine 5′-phosphosulfate (2 mM), 5 μL of NADP+ (20 mM), 1 μL of scPpx1p (33 000 U μL−1), 1 μL of ATP sulfurylase (0.075 U μL−1), 1 μL of hexokinase (0.075 U μL−1), 1 μL of glucose 6-phosphate dehydrogenase (0.075 U μL−1), and 11 μL of ATP-free water. The samples and polyP2 standards were dissolved and diluted in dilution buffer. A total of 100 μL of sample, which contained 1−10 μM polyP2 (optimal: 5 μM), was pipetted into a black microtiter plate. After the addition of 50 μL of polyP2 detection reagent, the plate was incubated for 55 ± 5 min at room temperature and then for 5 min at 25 °C in the plate reader, to equilibrate the temperature. Next, the fluorescence (excitation at 340 nm with a bandwidth of 9 nm, emission at 460 nm with a bandwidth of 9 nm, automatic gain adjustment scaled to the high wells “10 μM polyP2” with maximum scale value) was 7657

DOI: 10.1021/acs.analchem.9b00567 Anal. Chem. 2019, 91, 7654−7661

Article

Analytical Chemistry

activity, and to counteract inhibiting substances. Further, the rapid polyP2 consumption by these enzymes pulled the reaction of scPpx1p to completion. The reaction speed, starting from the intermediate products (i.e., polyP2, ATP, and glucose 6-phosphate; reactions 6−8 in Figure 1), was sufficiently fast that the reaction was completed in under a minute (data now shown). In the chain concentration assay, one polyP2 was expected to be released per polyP chain. Thus, the required polyP concentration for the assay depended on the polyP chain length (eq 5 in Figure 2; reaction 5 in Figure 1). For example, polyP2 was used at 5 μM, whereas polyP380 was used at 950 μM, to elicit the same polyP2 concentration (5 μM) after treatment with the enzymes. Thus, for scPpx1p, the highest substrate concentration was given for the longest polyP. Seven thousand U scPpx1p per reaction sufficed to hydrolyze 1 mM of the longest available polyP (polyP mode 384) to polyP2 in 1 h. To accommodate for the variance in enzyme activity and even longer polyP, a 4.7-fold excess (33 000 U scPpx1p reaction−1, which was the highest available enzyme concentration) and 1 h incubation time were chosen for the standard assay. The authors previously showed that 2200 U scPpx1p per reaction suffice to hydrolyze 200 μM polyP3 under similar reaction conditions.8 Full conversion of polyP3 (one of the shortest polyPs) by scPpx1p was thus ensured. The fluorescence intensity of pure NADPH in water decreased with rising temperature. This effect was notable when measuring the NADPH fluorescence over time in a microtiter plate, which was at room temperature and then placed in a 37 °C plate reader (data not shown). To keep the microtiter plate at a constant temperature, the assay was performed at room temperature, and the microtiter plate was preincubated for 5 min at 25 °C in the plate reader prior to the fluorescence being read. A linear relationship between the polyP2 concentration and the fluorescence was observed in the range of 1−10 μM polyP2. The signal to blank ratio was 1.7 for 1 μM polyP2 and 8.0 for 10 μM polyP2. The regression coefficient R2 was excellent (0.999). The limit of quantification (10 × standard error of regression × (slope)−1) was 1.0 μM polyP2. Two hypotheses were confirmed, to ensure proper function of the chain concentration assay. The first hypothesis was as follows: under standard assay conditions, ATP sulfurylase catalyzes ATP formation only from polyP2, but not from polyPn≥3. If ATP sulfurylase also acts on polyPn≥3, ATP would form from a polyP, which does not contain polyP2. This hypothesis was tested with polyP3, because, according to the manufacturer, it contained only polyP with a chain length of 3 P-subunits. Under standard assay conditions, but without scPpx1p, 3.7 μM polyP2 was detected from 100 μM polyP3. This small amount of polyP2 was explained by the nonenzymatic hydrolysis of polyP 3 to polyP 2 or by the contamination of the polyP3 with polyP2 . The latter speculation was later verified by 31P NMR (Table 2). Thus, hypothesis one was confirmed. The second hypothesis was as follows: under standard assay conditions, scPpx1p hydrolyzes polyPn≥3, but not polyP2, indicating that no polyP2 is hydrolyzed by scPpx1p in the relevant polyP2 range (1−10 μM polyP2). To test this hypothesis, the polyP2 standards were analyzed for their polyP2 concentration at various scPpx1p concentrations (0; 2200; 7000; 33 250; 66 500; and 133 000 U scPpx1p reaction−1). The hydrolysis of polyP2 through scPpx1p would

reference for the calibration of the chain concentration assay. The appropriate dilutions for Figure 3A and B were not the same, because the latter was the blank of the former and because the linear range of polyP2 detection was narrow (i.e., 1 log). For the total polyP and its blank determination, 100 μM total polyP, which was equal to the middle of the linear range of the Pi detection, was chosen (Figure 3C and D). For the detection of Pi in the presence of polyP, the total polyP concentration should not exceed 200 μM, to avoid nonenzymatic polyP hydrolysis during Pi detection. The average polyP chain length was calculated per eq 2 in Figure 2. The concentrations in this equation refer to the undiluted sample. 31 P NMR for the Determination of the Average PolyP Chain Length. The solution 31P NMR spectra were obtained using a Bruker 600-MHz spectrometer that was operated at a 31 P frequency of 242.95 MHz, and equipped with a prodigyprobe (a broadband CryoProbe that uses N-cooled RF coils and preamplifiers to deliver sensitivity enhancement over room-temperature probes by a factor of 2−3 for nuclei from 15 N to 31P). The samples were measured with a D2O-field lock at room temperature, and the chemical shifts were referenced to 85% orthophosphoric acid (0 ppm). The NMR parameters that were generally used were as follows: 32 K data points, 60 s repetition delay, 0.7 s acquisition time, 30° pulse width, and 64 scans. The peak areas were calculated by integration of the spectra that were processed with the 1 Hz line-broadening, using Bruker Topspin software, and in cases in which the peak areas could not be integrated unambiguously, MestReNova was used for curve fitting. The individual P compounds were identified on the basis of their chemical shifts and splits. These compounds included Pi (1−2 ppm), polyP2 (−7 to −8 ppm), polyP (−6 to −8, −20 to −23 ppm), and cyclic polyP from approximately −23 to −25 ppm. The accurate values of the chemical shifts of these signals depended on the pH and the residual concentration of the divalent cations in the solution. The chain length of the linear polyP was calculated from the signal intensities of PP1, PP2, PP3, and PP4. In cases in which the integration of the signal of the core phosphate group (PP4) was not possible due to overlays, the signals were fitted. This step also applied to the signal for the polyP2. The average polyP chain length for polyPn≥3 was calculated according to eq 1 in Figure 2. The overall average polyP chain length was calculated as the weighted mean of polyPn≥3 and polyP2 (eq 6 in Figure 2). The relative error of the method was ca. 2%.

■

RESULTS The total polyP and Pi determinations (Figure 3C and D; Figure 1A) were already established in a previous study.8 The present study developed a method to measure the polyP chain concentration (Figure 3A and B; Figure 1B). The combination of the total polyP and polyP chain determinations allowed the quantification of the average polyP chain length. Because the determination of the polyP chain concentration was an end point assay, it was possible to use enzymes and substrates in excess. With regard to the upper limit of detection (10 μM polyP2), excesses of 10-fold for adenosine 5′-phosphosulfate, 300-fold for glucose, and 100-fold for NADP+ were chosen, to drive the reactions to completion. ATP sulfurylase, hexokinase, and glucose 6-phosphate dehydrogenase were used at a 4500fold excess, to accommodate for the variance in enzyme 7658

DOI: 10.1021/acs.analchem.9b00567 Anal. Chem. 2019, 91, 7654−7661

Article

Analytical Chemistry Table 1. Interference of Varying Substances on the Enzyme Assaya interfering substance concentration [μM]

Table 2. Chain Length Determination of Seven PolyP’s with the Enzyme Assay and 31P NMRb mean chain length ± standard error of the mean

interference [%]

mass [g] to 1 g of analyte

total polyP concentration

polyP chain concentration

polyP chain length

100000 400000 750000

378.3 1513.0 2836.9

10 100 1000

0.1 0.9 9.2

30 300 3000

0.8 8.3 83.0

20 200 2000

0.4 4.2 42.3

−2 2 2

0 1 −10

−2 1 13

1 5 4

0 5 2

0 0 2

1 3 −96

−1 b.t.h. b.t.h.

2 n.c. n.c.

−3 1 −21

−2 3 b.t.h.

−1 −2 n.c.

−2 2 −11

4 6 12

−5 −4 −21

polyP polyP2 polyP3 m. 155 Budit 4 p700 p100 m. 385

NaSO4

ADP

cyclic polyP [relative to linear polyphosphate]a

enzyme assay 1.9 2.9 6.0 20.3 53.1 54.2 274.3

± ± ± ± ± ± ±

0.0 0.0 0.2 0.1 1.1 0.8 5.3

2.0 2.9 5.6 17.0 50.4 44.1 297.3

4 3 −6 −16 −5 −19 8

0.00 0.02 0.83 6.81 1.53 6.35 0.29

Determined with 31P NMR. bFor the enzyme assay, the experimental design was 2 × 7 × 3 (number of independent experiments done on separate days × number of tested polyP’s × number of replicate measurements). For 31P NMR, the experimental design was 1 × 7 × 1. Abbreviation: m., mode.

reducing the salt concentration. Concluding, the analysis of polyP in seawater should be possible. Sulfate and ADP were tested as potential inhibitors of the enzymatic reactions that are depicted in reactions six and seven in Figure 1, respectively. Sulfate concentrations at 10-fold higher and lower concentrations than the effective sample concentration of adenosine 5′phosphosulfate (100 μM) were tested, and had no effect on the enzyme assay. ADP concentrations at 1-, 0.1-, and 0.01-fold the effective sample concentration of glucose (3000 μM) were tested. Thirty μM ADP showed no interference. As measured in the blank of the chain concentration assay, ADP was contaminated with 2.1 molar percent ATP, preventing analysis of the interfering properties of 300 and 3000 μM ADP. In theory, the influence of ADP should be minimal, because the hexokinase reaction is irreversible. DNA and RNA were tested as interfering substances, because they are often copurified with polyP from biological samples. The 20 and 200 μM DNA and RNA had no effect. The 2000 μM DNA or RNA interfered strongly with the enzyme assay. For the analysis of biological samples that contain high ratios of nucleic acid to polyP (i.e., >4 g of nucleic acid to 1 g of polyP), this effect should be considered (in this case, the use of RNase and DNase during polyP extraction is recommended). It was uncertain why there was a measurable chain concentration blank for DNA (0.4 molar percent NADPH equivalent with regard to the DNA monomer). None of the chain concentration blank-quantifiable substances (i.e., ATP, glucose 6-phosphate, and NADPH) were expected from the salmon sperm DNA. In summary, the enzyme assay was sufficiently robust for the analysis of most biological samples. Seven polyP standards were analyzed for their mean polyP chain length with the enzyme assay (column 2 in Table 2). For the enzyme assay, coefficients of variance (CV; standard deviation divided by the mean) ± standard error of the CV were 2.1 ± 0.3% for interassay, 3.1 ± 0.5% for intra chain concentration assay, and 1.5 ± 0.2% for intra total polyP assay. The low coefficients of variance indicated that the enzyme assay was reproducible. No significant correlation between the interassay CV and the chain length was found (slope: 0.003 CV [%] per P-subunit, multiple correlation coefficient r =

RNA 0.4 4.4 43.9

deviation between methods [%]

a

DNA

20 200 2000

P NMR

31

NaCl

a

PolyP m100 (average polyP chain length by enzyme assay: 54 Psubunits) was spiked with the interfering substances, whose concentration is displayed in column one (DNA and RNA as monomer). The chain length of polyP m100 made it possible that the same total polyP concentration (150 μM) was used for both the determination of the total polyP concentration and the chain concentration. To illustrate the degree of contamination, column two shows the concentration of the interfering substances as the mass ratio “gram interfering substance to 1 g of polyP”. The total polyP concentration and the polyP chain concentration were determined with the standard enzyme assay. The average polyP chain length was calculated from these two concentrations. The mean interference from one experiment with three replicate measurements per substance concentration combination is reported in comparison to the control without the interfering substance. The following monomer molecular weights were assumed: 103 g mol−1 for polyP m100, 327 g mol−1 for DNA, and 339 g mol−1 for RNA. Abbreviations: b.t.h., blank too high; n.c., not calculable, because the chain concentration was not measured.

have been visible in a decreased slope of the polyP 2 concentration−fluorescence−calibration curve in comparison with a calibration curve, from which scPpx1p was omitted. The slope of the calibration curve fell by −0.028% per 1000 U PPX per reaction. This indicated that the slope was essentially the same for all scPpx1p concentrations. Thus, hypothesis two was confirmed. At this point, the assay development was complete. The interference of certain substances on the enzyme assay was investigated (Table 1). An interference of ≤5% was considered to be acceptable. NaCl was tested, to demonstrate that seawater is a possible sample matrix. The salt concentration of seawater is ca. 600 mM NaCl. One hundred and 400 mM NaCl had no effect on the enzyme assay. At 750 mM NaCl the chain concentration assay was inhibited by −10%, which resulted in a +13% increase in the measured chain length. The enzyme assay requires micromole per liter polyP concentrations (e.g., 150 μM polyP for polyP54). Thus, polyP samples in seawater probably must be diluted, effectively 7659

DOI: 10.1021/acs.analchem.9b00567 Anal. Chem. 2019, 91, 7654−7661

Article

Analytical Chemistry 0.289, p = 0.529). Concluding, the enzyme assay showed reproducible performance for short polyP as well as for very long polyP. The chain length of the same seven polyPs was also quantified by 31P NMR (Table 2). A highly significant positive correlation between the chain lengths that were determined by both methods was noted (slope: 1.0897 P-subunits determined with 31P NMR per P-subunit determined with the enzyme assay, r = 0.998, p = 1.3 × 10−7), confirming that these methods generated similar results. The deviation between methods was −4 ± 4% (mean ± standard error of the CV). No significant correlation between the deviation between methods and the polyP chain length was observed (slope: −0.067 deviation between methods [%] per P-subunit, r = 0.64, p = 0.119). 31P NMR was used to quantify the content of cyclic polyP relative to the linear polyP (column 5 in Table 2). A highly significant negative correlation between the content of cyclic polyP and the deviation (negative deviation = chain length31P NMR < chain lengthenzyme assay) between methods was found (slope: −3.1% deviation between both methods per 1fold excess of cyclic to linear polyP, r = 0.914, p = 0.003). In other words, a high cyclic polyP content gave a lower reading for the chain length by 31P NMR as compared to the enzyme assay.

only a low sample concentration (μM range) and common laboratory equipment (multichannel pipet, plate reader, 37 °C incubator). Advantages and Disadvantages of the Enzyme Assay in Comparison with the PPX/IPP Assay. The recently published PPX/IPP assay sizes polyP in the range of 2−80 Psubunits.8 However, with greater chain lengths, the PPX/IPP assay becomes increasingly inaccurate, because the difference in Pi concentration between the scPpx1p and scPpx1p/scIpp1p wells becomes hard to resolve analytically. In the PPX/IPP assay, a significant increase in interassay CV was noted (an increase of 0.67 CV [%] per P-subunit, r = 0.964, p = 0.00188). In other words, for polyP3 a CV of 0.5% and for polyP28 a CV of 17% was measured.8 In the enzyme assay from this study, the CV remained unchanged at ca. 2% over the entire range of polyP lengths (increase of 0.003 CV [%] per P-subunit), which is a major advantage over the PPX/IPP assay. The enzyme assay from the presented study is based on two independent assays (i.e., total polyP and chain concentration assay). The PPX/IPP assay, in contrast, is “one” assay, which makes it easier to perform and less prone to interassay calibration errors. Further, the PPX/IPP assay does not coquantify ATP, glucose 6-phosphate, and NADPH. In conclusion, for short polyP (2 to ca. 30 P-subunits), the PPX/IPP assay is a viable alternative to the enzyme assay from this study. PolyP Chain Length Standards for PAGE. The polyP’s “p100”, “mode 155”, “mode 385”, and “p700” are used as chain length standards for PAGE (see, for example, refs 6 and 14 for the application of similar polyP standards in PAGE). The numbers in the name indicate the average size that was determined with PAGE. The average polyP chain lengths of these standards were determined in this study to be 44, 5, 297, and 50 P-subunits by 31P NMR, respectively. Certain polyacrylamide gel concentrations allow for the resolution of only a certain range of polyP lengths (the longer is the polyP, the lower is the required polyacrylamide concentration). The separation of, for example, a polyP with an average chain length of 500 P-subunits on a low-strength (2%) polyacrylamide gel would not show contaminating polyP with a chain length of