A rapid, enzymatic assay for measurement of inorganic pyrophosphate in biological samples

241

Clinica Chimica Acta, 66 (1976) 241-249 @ Elsevier Scientific Publishing Company,

Amsterdam

-

Printed

in The Netherlands

CCA 7537

A RAPID, ENZYMATIC ASSAY FOR MEASUREMENT PYROPHOSPHATE IN BIOLOGICAL SAMPLES

GEORGE

LUST*

and J.E. SEEGMILLER

Department of Medicine, School La Jolla, Calif., 92037 (U.S.A.) (Received

OF INORGANIC

of Medicine,

liniversily

of California,

San Diego,

July 22, 1975)

Summary An enzymatic assay for measurement of inorganic pyrophosphate (PPi) in biological samples is described using the UDPG pyrophosphorylase reaction and coupling it with three other enzymes in a system of phosphorylation and double reduction to form NADPH, which is measured fluorometrically. The method is specific for PP, in the presence of inorganic orthophosphate and can be used on small volumes of deproteinized ultrafiltrates of fluids or cell extracts. The assay was used here to determine the concentration of PPi in human plasma and serum, and the intracellular content of PPi in human skin fibroblasts, chondrocytes, and red blood cells. The mean PP, concentration of plasma was found to be 2.72 + 0.14 /IM (pmol/l + S.E.M.); for serum it was 6.09 * 0.36 PM. The range (95% confidence limits) was 2.14~-3.30 PM for plasma and 4.62-7.51 PM for serum. The average intracellular PP, content (pmoles PP, per lo6 cells 2 S.E.M.) was as follows: (a) skin fibroblasts, 332 ? 66; (b) articular chondrocytes, 655 _+46; red blood cells, 1.74 + 0.28.

Introduction A number of techniques are available for determination of inorganic pyrophosphate (PPi) in biological samples. Several of these use calorimetry for measuring pyrophosphate after hydrolysis to orthophosphate (Pi). This requires preliminary separation of Pi either by column chromatography or by precipitating as phosphomolybdenate with triethylamine [l-m -31. An isotopic dilution method for measurement of PP, in biologic fluids and tissues also was devel* Present 14850.

address: U.S.A.

New

York

State

College

of

Veterinarv

Medicine.

Cornell

Uniwrsity,

Ithaca.

N.Y.

242

oped [3,4], but generally these methods are complex, time consuming, and require large amounts of starting materials. The UDPG pyrophosphorylase method of Johnson et al. [ 51 is specific for PP,; however, the sensitivity of the spectrophotometric measurement of NADPH formed is limited. Adaptations for increasing the sensitivity of this method by isotopic labelling of UDPG with subsequent analysis of UTP formed [6], and spectrophotometric determination of NADPH in a special capillary microcuvette [7] have been reported. Recently Cartier and Thuillier [S] adapted the enzymatic method of Johnson et al. [ 51 by measuring the NADPH fluorimetrically, thereby increasing its sensitivity; nevertheless, a purification step to remove certain interfering substances from analytical samples was required in their procedure. A rapid, direct method with a sensitivity to allow measurement of PP, in extracts of mammalian cells cultured in vitro, as well as in biologic fluids, has been needed. In the procedure described here, PP, is determined using yeast UDPG pyrophosphorylase and fluorimetric quantification of the NADPH produced in a coupled system of phosphorylation and double reduction. The analysis of PP, may be done in either phosphate buffer or Tris buffer on ultrafiltrates of biological fluids or neutralized perchloric acid extracts of mammalian cells. Materials

and methods

Reagen ts and solutions A stock solution of sodium pyrophosphate (10 pmol/ml) (Allied Chemical Co., Morristown, N.J., reagent grade) was prepared in 0.05 M Tris/acetate buffer, pH 8.0 and stored at 4°C. It was stable for 2-3 weeks. This solution was diluted with 0.05 M Tris/acetate buffer, pH 8.0, when needed for the standard curve (O-500 pmoles Of PPi). Other solutions used were: Tris/acetate buffer, 0.1 M, pH 8.0; magnesium acetate, 0.1 M in 0.05 M Tris/acetate buffer, pH 8.0; potassium phosphate buffer, 0.1 M, pH 8.0; NADP’ (Calbiochem, San Diego, Calif.), 8.0 mg (10 pmoles) dissolved in 1.0 ml of 0.05 M Tris/acetate buffer, pH 8.0; UDPG (Calbiochem) 8.0 mg (10 ~moles) in 1.0 ml Tris buffer, pH 8.0; UDPG pyrophosphorylase (UTP:o-D-glucose-l-phosphate uridyltransferase, EC 2.7.7.9) from yeast (Sigma Chemical Company, St. Louis, MO.); phosphoglucomutase (PGM, EC 2.7.5.1, rabbit skeletal muscle) (Calbiochem); glucose-6-phosphate dehydrogenase (GGPDH), EC 1.1.1.49) from yeast (Calbiochem); yeast 6-phosphogluconatc dehydrogenase (GPGDH, EC 1.1.1.43, Calbiochem); yeast inorganic pyrophosphatase type III (EC 3.6.1.4, Sigma Chemical Co.). The following mixtures were prepared before each series of measurements: (1) Substrate solution (S) = 500 ~1 of 0.1 M phosphate buffer, pH 8.0; 100 1-11magnesium acetate; 100 ,ul UDPG; 100 ~1 NADP; 200 ~1 water. (2) Enzyme solution (E) = PGM, 20 ~1; GGPDH, 20 ~1; and GPGDH, 20 ~1; in 300 ~1 of Tris/acetate buffer, pH 8.0. (3) UDPG pyrophosphorylase (100 units) was dissolved in 4.0 ml deionized water. At 4°C this was stable for 334 weeks.

243

A diluent solution to measure fluorescence consisted of 1.1 g sodium bicarbonate, 2 g sodium carbonate and 0.3 g EDTA in 1 liter of water, pH adjusted to 10. Standard test for PPi After ultrafiltration of plasma, serum, or synovial of cells, PP, was measured in the samples as follows: pp, +

.!!!!%?yr’??hosph

UDPG

~~~~~~~~~ _p Glucose-6-P 6-P-gluconate

UT’

fluid, or after extraction

+ glucose_l_jJ

J’~sehoglwomutase _~~~~~~ .._ glucose-6-P

+ NADP

06?ZDE 6-P-gluconate

+ NADP !!J?!%! ribulose-5-P

+ NADPH + CO, + NADPH

Under the conditions of the standard test, the fluorescence of the NADPH formed was measured in borosilicate tubes (12 mm X 75mm, 10 mm light path). The standard test included in a final volume of 250 ~1: 50 ~1 of solution S (substrates); 10 ~1 of solution E (enzymes); O-20 ~1 of PPi standards; 5 ~1 of UDPG pyrophosphorylase; and deionized water to make the final volume 250 ~1. Incubation was at room temperature (22°C) for 20 min. The reaction was stopped by addition of 2 ml of the bicarbonate/carbonate/EDTA diluent. Tubes were then mixed by swirling on a shaker, dipped into water and the outsides dried with lens paper to remove fingerprints and other dirt. Fluorescence was read in a Turner fluorometer model 111 (Palo Alto, Calif., 94303), adapted with an excitation filter at 366 nm (Corning filter No. 7760) and emission filters above 450 nm (Kodak light filters No. 48, 2A and 8). For analytical samples the standard PP, was omitted and 50 1-11of unknown sample was used. The relative fluoresence of this was compared to a blank sample from which the UDPG pyrophosphorylase was omitted. Preparation

of samples

Fluids Human blood was obtained by venipuncture from 15 normal adult individuals, seven females and eight males. To obtain plasma, the blood was immediately placed into heparinized tubes and inverted several times and placed in crushed ice. The remaining blood was placed into glass tubes and allowed to clot at room temperature for 30 min to obtain serum. Both types of fluids then were centrifuged at 2400 rpm (HL-8 rotor) in a RC-3 Sorvall refrigerated centrifuge (1700 X g) for 15 min to obtain clear plasma and serum, respectively. Proteins were removed by ultracentrifugation using an Amicon centriflo membrane (CF50 or CF25) and filter apparatus (Amicon Corporation, Lexington, Mass.) and centrifuged at 1600 rpm in an RC3 Sorvall refrigerated centrifuge (650 X g) until enough filtrate was collected (about 30 min). The ultrafiltrate (usually 0.5 ml) was used for PP, analysis.

244

Cell extracts Human skin fibroblasts and human chondrocytes from knee and hip joint articular cartilage were cultured in monolayer for seven days in plastic flasks [ 10, 131. Both types of cells were grown at 37°C in Coon’s F, 2 medium supplemented with 10% fetal calf serum and in the presence of 10% CO,/90% air. Cells were scraped from the surface with a rubber policeman in ice-cold phosphate buffered saline (PBS). Cell pellets were obtained in a refrigerated centrifuge (Sorvall RC3) (1700 X g). This pellet was resuspended thoroughly in 0.5 ml PBS and immediately made 1 M in perchloric acid (PCA) by addition of concentrated PCA. The macromolecular precipitate was removed by centrifugating (1700 X g) and saved for DNA analysis, and the supernatant solution was neutralized to pH 7.0-8.0 with KOH. The precipitate of KClO, was removed by centrifugation and the supernatant (about 0.5 ml) was collected and either analyzed for PP, directly or frozen at -20°C until used. DNA content was determined on the macromolecular pellet by the method of Levya and Kelley [9] . DNA content was converted to cell number by means of a standard curve, with cell number being determined by photoelectric scanning using a Coulter counter (Coulter Electronics, Inc., Hialeah, Fla.) (amplification setting 16). Red blood cells were obtained from heparinized blood by centrifuging at 2400 rpm (1700 X g) for 15 min. RBC’s were washed three times with ice-cold PBS and finally re-suspended in ice-cold PBS. Cell numbers per ml were quantified with a Coulter counter at an amplification setting of l/4. Experimental

results

A standard curve showing the relationship of increasing PP, concentration with an increase in relative fluorescence is illustrated in Fig. 1. The curve was linear from 50 to 500 pmoles of PP,. A novel 2-fold increase in sensitivity was effected by the addition of 6-phosphogluconate dehydrogenase (GPGDH), resulting in a second reduction of NADP’ to NADPH. The overall reaction usually was complete in 7-10 min as judged by a leveling off of the maximum fluorescence, but to ensure completion of the reactions with unknown samples, a 20-min incubation was used.

It.88

rl

CC\:ENTRATION Fig.

1. Standard

The

dotted

line

n~rvc’

4

;

5’8

PPI ip~comoles m 0 25ml) showing

represents

the relationship

the standard

tebt donr

between without

the concentration addition

of PPi and rclaticr

of 6-phosphopluconate

fluoresence.

dehydrog(,nasr.

245

The following substances were found to be without effect in the standard test at concentrations of 0.20 and 1 mM, and 5 mM: potassium phosphate (when Tris/acetate buffer was used), sodium chloride, ATP, ADP, GTP, GDP, UTP, UDP, glucose, dithiothreitol, and sodium fluoride. Addition of 5-phosphoribosyl l-pyrophosphate consistently increased the relative fluorescence, presumably because of breakdown of this compound into PPi. CaCl, and EDTA were without effect at the two lower concentrations, but inhibited the assay by 20% and 30% respectively at 5 mM. Equimolar amounts of both CaCl, and EDTA were not inhibitory. Ribose 5-phosphate inhibited the reaction sequence by 10% at 1 mM and about 25% at 5 mM. This inhibition was observed with short-term incubation times (5 min) but was less pronounced at 20 min of incubation as used in these tests. Recovery of PP, added to samples of plasma, serum, and to samples treated with 1 M perchloric acid and neutralized with 4 N KOH are presented in Table I. PP, was recovered nearly quantitatively. We have used the standard test described here to determine the concentration of PP, in the plasma and in the serum of 15 normal human adults of both sexes. Mean concentration of PP, (’ standard error of the mean) for plasma was 2.72 2 0.14 PM; for serum the PP, concentration was 6.09 + 0.36 PM. The complete set of measurements is given in Table II. The experimental values obtained correlate well with PP, concentrations reported previously in the literature. Specificity for PPi was verified in these samples by pre-incubating selected ultrafiltrates with yeast inorganic pyrophosphatase. This treatment abolished the fluorescence in those samples. The intracellular PP, contents of human skin fibroblasts, articular cartilage chondrocytes, and red blood cells are presented in Table III, part A. Results are expressed as pmoles per million cells. Two separate fibroblast lines from two normal individuals were studied. Twelve subcultures of line B and ten subcultures of line J were analyzed. Mean pyrophosphate contents (k standard error of the mean) were 266 + 22 and 398 ? 33 pmol/106 cells, respectively. Four chondrocyte lines were studied with the intracellular PP, contents per lo6 cells ranging from 787 ? 85 to 580 * 71. Individual variations, per culture, in the chondrocytes were rather large as may be seen by the large standard deviations; however, when duplicate cultures were analyzed at the same time, the

TABLE

I

RECOVF:RY

PPi added

OF

PPi E’ROM

SAMPLES

OF

PLASMA.

SERUM

AND

PERCHLORIC

Perchloric

PPi recovered*

ACID

acid

EXTRACTS

extract

(nmoles) Serum

Plasma

100

nltl0les

%

nm0les

70

IllllOllZS

%

98

99.0

99

97.0

97

98.2

50

48.8

97.5

46.5

93

46.4

93

10

9.3

9.0

90

9.1

91

* Samples

wt’re

93 analyzed

exactly

as unknowns

were

done.

See Methods.

246 TABLE

II

CONCENTRATIONS

OF

PPi IN HUMAN

PLASMA

AND

PPi

SEKUM

(~InUl/l)

Plasma

1

36

M

2.45

7.70

2

28

M

2.15

6.08

3

31

M

2.45

9.10

4

2R

M

3.46

7.35

5

20

F

3.60

7.20

6

54

M

2.80

5.00

7

37

M

4.00

6.55

8

26

M

3.64

5.74

9

24

F

2.66

6.72

10

27

M

2.52

7.56

11

26

F

2.50

7.98

12

30

1’

2.35

3.95

13

33

F

2.19

3.85

14

22

F

1.90

3.40

15

26

I;

2.20

5.35

Mea11

2.72

6.09

Ran@?

1.90.-4.00

x40--9.10

’ 0.29

S.D.

to.14

S.E.M. 95q1

‘0.73

Confidence

Literature

Limits

2.14

IO.36

’

3.30

4.67

Values

11.81

1.10-5.90 (mran

121

3.53)

1.70.-2.90 (mean

1.80)

141

3.90-~9.20

TABLE: PAKT

+ 7.51

III A.

CYTES

PPi

AND

CONTENT RED

OF

BLOOD

HUMAN

SKIN

~IBKORLASTS,

AKTICULAK

CAKTILAGE

CELLS

__ Cril

,,;:

tsyr

11ltrd.ccIlular (Prn<,l,‘l

Skin

fibrocvtrs Overall

Articular

c11011dr0cytrs

01 rrall Red

blood

cells)

I3

I2

266

10

398

102.7

332

93

C2

11

787

213

C7

8

649

2 219

Cc,

9

6 04

187

!62

Cl1

6

580

1 172

,71

655

+

1.74

1

averagr cells

0”

.J average cartilage

PI+

11

+

76.9

i22 ’ 3 3.5 166 ‘85 ’ 76

92 0.91

46

. 0.28

CHK’,NDKO-

247

PART

Cell

B.

PPi

CONTENT

type

OF

FIBROBLASTS

Cell (X

number IO”)

AND

CHONDROCYTES

PP, content

AT

TWO

pmol/lO”

CELL

DENSITIES

cells

(pmoles)

Fibroblasts

B

3.65

910

Fibroblasts

B

7.30

7690

250 230

Chondrorvtes

C:

1.80

1100

620

Chonclrocvtes

C>

3.60

2250

640

results usually were within 5% of each other. Different culture conditions from one time to another (e.g., cell density) and differences in cell size may be responsible for the variability in intracellular PPi content. Duplicate analyses of eleven individual, washed red blood cell extracts were more consistent. The intracellular content of PP, in RBC’s was only about 1.74 pmoles per million cells, but it must be noted that RBC’s are small cells compared to fibroblasts and chondrocytes. In Table III, part B, results are given which indicate that the extraction procedure and analytical test is linear with increasing cell density. Doubling the cell numbers of fibroblasts and chondrocytes in each assay doubled the total intracellular amount of PP,; content per lo6 cells remained constant. Again, chondrocytes contained about twice the intracellular amount of PPi as fibroblasts. Discussion The role of PPi in metabolism is poorly understood, largely because the methods of determination available until recently were complex and lacked the specificity and the sensitivity required to determine quantities in the nanomole and picomole range. Many biosynthetic reactions result in the formation of PP,, and this compound may have a fundamental part in regulation of metabolic sequences in biology. PPi has been found in liver, blood, synovial fluid, and in bone and has been implicated in the control of calcium and phosphate metabolism in bone [ 1,2,4,6,8,11,12]. The UDPG pyrophosphorylase method of Johnson et al. is specific for PP, and uses commercially available enzymes and substrates. The sensitivity of this method [5] was limited by the spectrophotometric measurement of the NADPH at 340 nm. In a cell volume of 0.5 ml, one nmole of NADPH (and PP,) results in an increase of the absorbance of only 0.0125 at 340 nm. Alteman et al. [ 73 increased the sensitivity of this method by using a 5-cm optical pat,h in a micro cell. Floodgard [6] used a tritiated UDPG of high specific activity and analyzed for [‘HI UTP. Fluorimetric measurement of NADPH can increase the sensitivity 100-fold over the spectrophotometric method. This was done to advantage by Cartier and Thuillier [8] who reported on the PPi content of plasma, urine and bone. These authors used the UDPG pyrophosphorylase from beef liver which was sensitive to the inhibitory action of Cl ions and phosphate. Their system also was affected by other interfering substances which required the use of a puri-

fication step. In their test, PP, was eluted from an ion-exchange column and analyzed. In the method described here, we used the UDPG pyrophosphorylase of yeast which was not sensitive to chloride ions or to phosphate. The commercially available substrates and enzymes used in the coupled system were free of interfering activities (e.g., ATP, hexokinase) so that a precipitation step was not needed to eliminate interfering compounds (e.g., glucose, Cl-, phosphate) from samples. A new and useful observation was that the addition of 6-P-gluconate dehyrogenase to the coupled assay system increased the sensitivity by a factor of 2-fold. Under the conditions of the standard test, a sample of 50 ~1 of ultrafiltrate of plasma or serum gave reproducible results with a sensitivity down to 50 pmoles of PP, . A further useful and time-saving device was the use of Centriflo (Amicon Corp.) membrane filter cones to obtain ultrafiltrates from biological fluids. In this way, ultrafiltrates of plasma were obtainable within 30 min. The total time needed to do PPi analysis on 10 samples of ultrafiltrates or cell extracts was about 30 min. The validity of this test was proved by the fact that the PP, concentrations of plasma measured with it corresponded very well with values reported in the literature. We report here a plasma PP, of 2.72 + 0.14 PM (2 S.E.M.); Russell et al. [l] reported 3.53 + 0.19 @‘I; Cartier and Thuillier [8] 3.50 ? 0.11 PM, McCarty et al. [ 21 and Silcox and McCarty [ 41 1.80 + 0.06 PM. The PP, concentration of serum was greater than twice that of plasma. Our results on this confirm the findings of Silcox and McCarty [4], who previously observed this phenomenon. They presented evidence that the additional PP, of serum originated from platelets during the “release reaction” 141. In order to measure the intracellular PP, content of mammalian cells grown in culture, a sensitivity in the picomole range is needed to keep the cell numbers manageable. The content of PP, for articular chondrocytes was found to be about twice that of skin fibroblasts. The reasons for this are unknown but may be due partly to variations in cell sizes or differences in activities and compartmentation of degradative enzymes of PP,, such as pyrophosphatase and alkaline phosphatase. Chondrocytes are known to synthesize more sulfated proteoglycans [13], a process which releases PP,. Additional work is in progress to delineate the metabolic origin of the PP, in articular chondrocytes. These studies eventually may contribute to an understanding of arthritic diseases such as “pseudo-gout” (articular chondrocalcinosis), a condition where large human numbers of microcrystals of calcium pyrophosphate dihydrate appear in diarthroses.

Acknowledgements

This work was supported in part by grant 5-ROI-AM-13622 from the National Institutes of Health, PHS feliowship grant IF32 AM 05005-01, and a grant from the Kroc Foundation for the Advancement of Medical Science. The technical assistance of J. James and L. Radell is acknowledged.

References 1

Russell.

R.G.G..

Bisaz,

S.,

Donath,

A..

D.B.

Morgan.

and

Fleisch.

H.

(1971)

J.

Clin.

Invwt.

50.

961-967 2

McCarty.

3

Cartier,

4

Silcox.

5

Johnson.

D..J.,

Solomon,

P. (1957) D.L.

S.D..

Bull.

and

Sac.

McCarty.

J.C.,

Warnock,

Chem. D.J.

Shanoff,

M.L.

Biol.

(1973)

M..

Bass. Biochem.

39, J.

Clin.

ST..

and

Polyan,

Invest.

52,

E.

(1971)J.

Lab.

Clin.

Med.

78,

216-229

169 .J.A.

Boeri,

1863-m1870

and

Hansen.

R.G.

(1968)

Anal.

Hiochem.

137-144 6

Floodgard,

7

Altmm,

8

Cartier,

9

Levya,

A. and

10

Green,

W.T.

11

A. R.D.. P.H.

Fleisch,

12

Silcox, Sokoloff,

Eur.

J.

Muniz.

0..

Pita,

Thuillier.

L.

and

Kclley, (1971)

H. and

(Rasmussen, 13

(1970)

H..

D.C. L..

and

W.N. Clin.

Russell. ed.).

R.G.G.

Anal.

D.J. C.J.

Rel. (1970)

51. and

PP.

273-280

Howell,

Anal.

(1974)

Sect.

15.

and

(1974)

Orthop.

McCarty.

Malemud,

J.C.

Biochrm.

Res.

Green

(1973)

75.

61. 62,

J.

Pergamon. Clin.

*Jr., W.T.

Arthritis

Rheum.

16,

171.-178

of Pharmacology

and

416-428 173-179

248-260

in International

61m-100.

(1973)

D.S.

Biochem.

Invest. (1970)

Encyclopedia Oxford 52.

1590-1600 Arthritis

Rheum.

13.

118-124

Therapeutics

26,