ANALYTICAL

BIOCHEMISTRY

199,

75-80

(19%)

Mass Assay for lnositol 1 -Phosphate in Rat Brain by High-Performance Liquid Chromatography and Pulsed Amperometric Detection Robert

J. Barnabyl

Department of Medicinal Terlings Park, Eastwick

Received

July

Chemistry, Merck, Sharp and Dohme Research Laboratories, Road, Harlow, Essex, CM20 2QR, United Kingdom

It is well established (l-3) that stimulation of cell surface receptors by agonists causes hydrolysis of phosphatidylinositol 4,5-bisphosphate by phospholipase C with production of the secondary messenger inositol 1,4,5-trisphosphate (I(1,4,5)P3).2 Removal of 1(1,4,5)P, is performed via two pathways (4-7) and is eventually

’ Correspondence address: Department of Pharmacokinetics and Metabolism, Glaxo S.p.A. Via Fleming, 2, 37100 Verona, Italy. ’ Abbreviations used: I(l)P, myo-inositol l-phosphate; 1(4)P, myoinositol 4-phosphate; 1(1,4,5)P,, myo-inositol 1,4,5 trisphosphate; PAD, pulsed amperometric detection; PI, phosphatidylinositol; RSD, relative standard deviation; SPE, solid-phase extraction; TCA, trichloroacetic acid.

All rights

$3.00

by Academic Press, of reproduction in any form 1991

Research Centre,

19, 1991

A high-performance liquid chromatographic method for direct mass measurement of inositol l-phosphate (I(l)P) in rat brain is described. Separation of I(l)P from its isomers and from endogenous components is achieved by polymeric anion-exchange chromatography with a sodium hydroxide/sodium acetate mobile phase. Detection is performed at high pH by pulsed amperometric detection at a gold electrode. Sample preparation involves liquid-liquid extraction and ion-exchange solid-phase extraction, prior to HPLC. The method is sufficiently sensitive and selective to enable facile determination of basal levels of I(l)P in small amounts of brain tissue. The applicability of the method is demonstrated by the in uiuo monitoring of I(l)P levels in rat brain after administration of the inositol monophosphatase inhibitor lithium and the cholinergic agonist pilocarpine. The method is a significant improvement over existing published mass assays for I( l)P by virtue of its simplicity, speed, sensitivity, and ruggedness. 0 1991 Academic Press, Inc.

0003-2697191 Copyright 0

Neuroscience

dephosphorylated to inositol for resynthesis of phosphatidylinositol (PI). The metabolism of inositol phosphates is further complicated by the possibility of direct production of I(l)P and inositol 1,4-bisphosphate (1(1,4)P,) from PI and phosphatidylinositol 4-monophosphate, respectively. Lithium has been used extensively in the study of inositol phosphate metabolism because of its inhibition of several enzymes involved in the dephosphorylation of 1(1,4,5)P, (8-9). The dominant effect of lithium is the noncompetitive inhibition of inositol monophosphatase, causing a dramatic elevation of inositol monophosphates, predominantly I( l)P, and a subsequent depletion of inositol. In order to study the in uiuo inhibition of inositol monophosphatase by lithium, an assay sufficiently sensitive to determine basal I(l)P levels in brain tissue is required. Most methods used to monitor agonist-induced changes in inositol phosphate levels have relied upon radiolabeling techniques, utilizing intracerebral administration of [3H]inositol to radiolabel the inositol-lipid pool. The assumption that changes in radioactivity reflect changes in inositol phosphate concentrations is only valid if the specific activity remains constant throughout the experiment; i.e., each inositol lipid pool has been radiolabeled to equilibrium. This, in practice, is often difficult to achieve and to monitor. Therefore, a direct mass assay for I(1)P in brain tissue is preferred. Recently, a number of mass assays for inositol phosphates have been developed (lo-16), most of them attempting a complete assay of all inositol phosphate isomers up to tetraphosphate, but with particular interest in I(1,4,5)P,. Typically the methods are difficult to set up, require long analysis times, and are not particularly robust. This novel, rapid, and relatively simple mass assay for I(l)P in brain tissue involves liquidliquid and solid-phase extraction followed by HPLC with pulsed amperometric detection (PAD). All simple 75

Inc. reserved.

76

ROBERT J. BARNABY

alcohols, glycols, polyalcohols, and carbohydrates can be detected by PAD using gold electrodes at pHs > 12 (17), the response being proportional to the number and relative position of hydroxyl groups present. Hence PAD is particularly well suited to the selective and sensitive detection of inositol monophosphates.

thoroughly degas. Mobile phases were maintained in a degassed state by applying a helium pressure of 10 psi during use. All runs were carried out at ambient temperature and at a flow rate of 1 ml/min. Detection was performed in the pulsed amperometric mode (19) on a gold electrode using the following pulse sequence:

METHODS

Pulse period (s)

Applied potential (V)

o-o.5 0.51-0.59 0.60-0.65

+0.10 +0.60 -0.60

Materials myo-Inositol (1) phosphate and myo-inositol(4) phosphate were synthesized in house (18). The internal standard used was myo-inositol2-o-cyclopropyl l-phosphate and was synthesized in house by alkylation of the intermediate l-o-ally1 3,4,5,6-tetra o-benzyl myo-inositol. All compounds were prepared as their bis-cyclohexylamine salts. All solvents were of HPLC grade and obtained from Fisons (Loughborough, UK). All inorganic materials were of AnalaR grade and obtained from BDH (Poole, UK). Water was of MilliQ (>18 Mohms) quality; 50% sodium hydroxide solution (necessary to eliminate carbonate contamination) was also obtained from BDH. Diethylaminopropyl (DEA) Bond Elut solid-phase extraction cartridges (100 mg, 1 ml) were obtained from Jones Chromatography (Hengoed, UK). HPLC The instrumentation was a Dionex BioLC comprising a stainless-steel-free quaternary gradient pump module, a manual rotary microinject valve with a 25-111loop, and a pulsed electrochemical detector (consisting of a combined pulsed amperometric detector and chemically suppressed conductivity detector). All liquid pathways were inert (stainless-steel-free) and all connections made by PEEK tubing and fittings (Upchurch, Oak Harbor, U.S.A.). Autoinjection was made by a Gilson 231-401 autoinjector equipped with a titanium Rheodyne 7010 valve with a lOO-~1loop. Data acquisition and processing were either performed by a Shimadzu (Kyoto, Japan) C-R5A reporting integrator or a Hewlett-Packard (Avondale, U.S.A.) HP 3365 PC-based data system. A Dionex AS-7 polymeric strong anion exchange column (25 X 0.5 cm, 10 pm) with a Dionex AG-7 guard column (5 x 0.5 cm, 10 Km) was used. The mobile phase consisted of 100 mM sodium hydroxide and 100 mM sodium acetate. The chromatographic run was isocratic until 10 min whereupon the column was washed with 100 mM sodium hydroxide and 500 mM sodium acetate for 7 min and reequilibrated with the starting mobile phase for 7 min. The mobile phase was prepared by adding 8.0 ml of 50% (w/v) sodium hydroxide to 1 liter of filtered (0.22 pm) 100 or 500 mM sodium acetate and immediately purged with helium for 5 min to mix and

Signal integration was carried out between 0.3 and 0.5 s. Sample Preparation Sprague-Dawley rats (weight range 200-250 g) were injected ip with 10 mmol/kg lithium or saline 17 h before sacrifice and either 50 mg/kg pilocarpine or saline 90 min before sacrifice. Animals were decapitated and brains removed rapidly and suspended in liquid nitrogen for 5 min. If immediate assay was not possible samples were stored at -2O’C. Frozen brains were weighed, 50 pg internal standard (L-670,168) was added, and then the sample was homogenized with 2 ml/g ice-cold Mill;& water. To 0.3 ml of the homogenate in a l-ml Eppendorf tube was added 0.1 ml 10% TCA and 0.4 ml Mill;& water, then this mixture was vortex mixed for 5 s. After centrifugation at 4,000 rpm for 5 min, the supernatant (ca. 0.7 ml) was transferred to a glass vial and 6.5 ml (9 vol) chloroform:methanol (2:l) added. After vortex mixing for 30 set and centrifugation at 2,500 rpm for 5 min, the whole of the upper aqueous phase was transferred to a Bond Elut DEA cartridge, previously conditioned with 1 X 1 ml 12 M formic acid:acetonitrile (1:l) and 2 X 1 ml MilliQ water. The sample was slowly drawn through using slight vacuum, followed by successive washes with 1 X 1 ml MilliQ water and 1 X 1 ml 10 mM sodium acetate. The cartridge was completely dried after each of the washing steps under high vacuum. Inositol monophosphates were recovered by slow elution with 0.5 ml 100 mM sodium acetate directly into autosampler vials; 25 ~1 of the eluate was injected directly onto the HPLC. Calibration Calibration standards, usually four, were prepared by spiking l-20 pg I(l)P and 5 pg L-670,178 into 0.3 ml control rat brain homogenate. The control brain homogenates were prepared in exactly the same manner (2 ml/g in water) and at the same time as the samples to be assayed. A calibration graph was constructed by plotting the peak area ratio of I(l)P/L-670,168 against the amount of I(l)P spiked and the calibration line generated by unweighted linear regression fitting of the data.

MASS

ASSAY

FOR

INOSITOL

77

l-PHOSPHATE

C

a

2 \

i

i hh

. r 0 Retention

time (mind

Retention

FIG. 1. Representative chromatograms of samples obtained homogenate (0.3 ml) spiked with 5 pg L-670,168 and 3 pg I(1)P. sample (0.3 ml) + 5 ag L-670,168 after lithium and pilocarpine (2) I(l)P 9.0 min, (3) I(4)P 10.0 min.

time (mins)

Retention

time (mind

during I(l)P determination in rat brain homogenate samples. (a) Control brain (b) Control brain homogenate (0.3 ml) + 5 c(g L-670,168. (c) Brain homogenate administration. Peak identification and retention times: (1) L-670,168 7.5 min,

Extrapolating this line to y = 0 (standard addition method) provides the basal concentration of I( l)P. Resubstituting this concentration value into the regression equation (y = mr) does not give the measured peak area ratio of control brain homogenate but a value approximately 10% less, due to endogenous interference (see Validation of Methodology). This residual value is calculated for each calibration curve prepared and subtracted from the sample peak area ratio before calculation of sample concentration. Only the slope (mx) term of the regression equation (y = mz + c) is used in the final calculation of concentration. RESULTS

Chromatography Under the conditions described, the retention time of I(l)P was typically 9.3 min and that for the internal standard, L-670,168, was 7.8 min. No interference in L-670,168 measurement, by coextracted endogenous material, was apparent. Typical chromatograms obtained from control rat brain homogenate and standards are shown in Fig. 1. A small peak (typically 15 20% of the basal I(l)P peak area) was present and

possesseda retention time similar, but not identical, to that of inositol4-phosphate I(4)P. Addition of I(4)P to the sample gave a similar single, unsymmetrical peak at the same retention time, indicating that some of the peak area may be attributed to I(4)P. Validation of Methodology The assay was validated by reference to recovery, linearity, reproducibility, and selectivity. Recovery was determined both by comparing peak areas of control brain homogenate spiked with I( l)P before and after extraction (no internal standard), and also by scintillation counting after spiking brain homogenate samples with 3H-labeled I(l)P. Both methods gave similar recoveries of approximately 70%, the major loss arising from the TCA deproteinization and solvent extraction steps. The recovery can be improved by washing the precipitate after TCA treatment and performing a second extraction of the organic phase, but at the cost of a substantial increase in analysis time. These steps were not included as the sensitivity and reproducibility were more than sufficient for the purposes of the assay.

78

ROBERT

J. BARNABY

The linearity of I(l)P calibrations was tested in the range O-20 pg spiked in 0.3 ml control rat brain homogenate. Linear regression fitting consistently produced excellent calibrations (? > 0.995) and enabled basal levels of I(l)P to be determined by extrapolation of the regression line to zero (standard addition technique). Basal levels of I(l)P typically ranged from 50 to 150 nmol/g wet wt. The reproducibility of the method was determined as the relative standard deviation (RSD) of replicate assays of samples obtained from animals dosed with lithium and pilocarpine or saline. Six replicates of samples at high (2.5 pmol/g), intermediate (1 pmol/g), and low (0.1 pmollg) concentrations were performed on the same day. RSD values obtained were 2.5, 4.6, and 7.5% for the high, intermediate, and low concentration levels, respectively. The selectivity of the assay for I(1)P was demonstrated by incubation of control rat brain homogenate with various amounts of bovine recombinant inositol monophosphatase expressed and purified as described elsewhere (20). Figure 2 shows the effect of enzyme addition. Approximately 90% of the I(l)P peak was removed by addition of 5 ~1 pure enzyme. Further hydrolysis did not occur with longer incubation time and it was inferred that the remaining peak was attributable to a coextracted endogenous component and compensated for (see Calibration) in the determination of sample I( l)P content. The level of this endogenous component was observed to be reproducible between batches of prepared standards. Also the peak, occurring at a similar retention time to 1(4)P, was reduced in area by 75% upon incubation with enzyme. Inositol l-monophosphatase hydrolyzes I(4)P at approximately the same rate as I(l)P (8,9), suggesting that there is a 25% interference in basal I(4)P determination. This was confirmed by performing a standard addition assay for I(4)P content, (spiking 0,0.5,1, or 2 pg I(4)P into 0.3 ml control brain homogenate). Complete separation of I(l)P from possible interfering endogenous compounds such as glucose l-phosphate, glucose 6-phosphate, fructose l-phosphate, and adenosine monophosphate was observed, under the chromatographic conditions used. Characterization

of I(l)P

Response

in Vivo

I( l)P levels in rat brain samples obtained after ip administration of saline, lithium, or lithium + pilocarpine were successfully determined and the results are shown in Table 1. Four animals were used per group. An approximate lo-fold elevation in I( l)P levels was observed after lithium administration and was potentiated by pilocarpine treatment, giving I( l)P levels 17-fold those of basal. I(4)P levels were estimated, assuming little significant interference in determination of elevated levels.

a

i ’ iI

i

I ii l-2I”\ L

I

0

I

I

I

I

5

10

15

20

Retention

time (mind

FIG. 2. Validation of the I(l)P assay. (a) Control rat brain homogenate (0.3 ml). (b) Control brain homogenate (0.3 ml) incubated with 5 ~1 pure bovine inositol monophosphatase for 16 h at room temperature. Peak identification and retention times: (1) L-670,168 7.8 min, (2) I(l)P 9.3 min, (3) I(4)P 10.2 min.

Basal levels were determined by standard addition. I(4)P levels also rose to about the same degree as I(l)P after lithium treatment but did not rise again significantly after pilocarpine administration. DISCUSSION Triple-pulse amperometric detection of carbohydrates was first described by Hughes and Johnson (21) and further developed by Rocklin and Pohl(22) using a gold electrode. Products of the oxidation of polyols rapidly fouled nobel metal electrodes so that conventional constant potential (dc) amperometric detection is impossible. Pulsed amperometry, involving application of high and low potential pulses after the integration of the signal at the working potential, overcomes this problem by repeatedly cleaning and replenishing the active electrode surface. Detection at a gold electrode is only ac-

MASS TABLE

ASSAY

FOR

INOSITOL

1

Effect of Lithium Administration on I(l)P and I(4)P Levels in Whole Rat Brain with and without Pilocarpine Potentiation Concentration Treatment (hr before sacrifice)

(pmol/g

wet wt)

I(l)P

I(W’

2-h saline

0.14 (0.01)

0.053" (0.007)

17-h lithium 1.5-h saline

1.50 (0.39)

0.37

(0.19)

17-h lithium 1.5-h pilocarpine

2.43 (0.52)

0.42

(0.18)

Note. Numbers in parentheses indicate a Determined by standard addition.

standard

deviation.

complished at high (> 12) pHs. Matching the column mobile phase to detector requirements is preferred for convenience and to avoid problems with noise generated from postcolumn mixing. This is possible with the recent availability of polymeric anion-exchange columns which are highly efficient when used with sodium hydroxide mobile phases. Pulsed amperometry of inositol monophosphates is a highly selective detection technique; however, some sample clean-up is required to avoid interferences from the sample matrix and to maintain the column and electrode reproducibility. The sample preparation procedure uses well-established methods (11,12) for inositol phosphate extraction. Solid-phase extraction (SPE) on disposable silica-based ion-exchange cartridges was used instead of commonly used Dowex resins because of the improved speed of batch processing and the lack of contamination from the packing itself. Diethyl aminopropylbonded silica was chosen as a medium strength anion-exchange material so that the ionic strength of the eluent required to quantitatively recover inositol monophosphates was similar to that of the mobile phase. The use of formate or chloride as SPE eluent caused some decrease in detector response and injection of solutions containing an appreciably higher concentration of salt than the mobile phase caused peak broadening and splitting. Most mass assays for inositol phosphates have required a desalting step before chromatography, hence direct injection of the eluate imparts an immediate advantage in terms of speed and recovery. The sensitivity of I(l)P detection is excellent: ca 5 pmol can be detected (S/N = 3) on column. In fact, basal levels of I( l)P in rat brain are such (20-60 pg/g wet wt) that only a small fraction of the final sample volume needs to be injected, enabling I(l)P levels to be monitored in discrete regions of the brain. The selectivity for inositol monophosphate determination is good, although significant interference in basal I(4)P measurement is observed. Alteration of the column and/or mo-

79

l-PHOSPHATE

bile phase conditions may be useful in separating the endogenous interferences from I(4)P. Perhaps the major advantages of this method over existing mass methods for I(l)P are speed, simplicity, and ruggedness. Typically 30-40 samples, including standards, can be assayed in a 24-h period. The column and detector require very little maintenance. The column is washed for 2 h with 1 M sodium hydroxide, typically after 3 days of continuous use. To date, no attempt has been made to investigate the applicability of the method to determine higher inositol phosphates. It is anticipated that detector sensitivity will decrease as hydroxyls are replaced by phosphate groups, but to what degree is uncertain. The opposite trend is observed for conductivity detection (X,16), whereby sensitivity is increased for the higher phosphates. Both amperometric and conductivity detectors, contained in the Dionex pulsed electrochemical detector, can operate simultaneously in series, so that if required the system could measure all inositol phosphate isomers in a single run, using PAD for the lower phosphates and conductivity for the higher phosphates. The constraints would then be made on the column and mobile phase to optimize isomer separation by stepwise gradient programs. In summary, a simple, rapid, and robust mass assay for I(l)P in rat brain has been developed. The assay is sufficiently sensitive, reproducible, and specific to monitor basal levels and small changes in I(l)P concentrations following lithium administration. I(4)P can also be measured but with some interference in determination of basal levels. This procedure overcomes many of the problems associated with the use of radiolabeled precursors and is a significant improvement over previously published methods for the mass assay of I( l)P in tissue. ACKNOWLEDGMENTS I am indebted to D. Booth, M. Russell, and I. Mawer for the synthesis of the I(l)P, L-670,168, and I(4)P standards used in the assay and also thank S. Burton and E. Brawn for the preparation of this manuscript.

REFERENCES 1. Berridge,

M. J. (1984)

Biochem.

J. 220,

345-360.

2. Downes,

C. P., and Michell, R. H. (1985) in Molecular Mechanisms of Transmembrane Signalling (Cohen, P., and Housley, M. D., Eds.), pp 3-56, Elsevier, Amsterdam.

3. Abdel-Latif, A. A. (1986) Pharmacol. Rev. 38, 227-272. 4. Downes, C. P., and Michell, R. H. (1981) Biochem. J. 198,133140.

5. Downes, 6.

C. P., Mussatt, them. J. 203, 169-177. Hawkins, P. T., Michell,

M. C., and Michell, R. H., and Kirk,

R. H. (1982)

C. J. (1984)

Biochem.

BioJ.

218,785-793. 7. Storey, 8.

Nature Ragan,

D. J., Sears, S. B., Kirk, C. J., and Michell, R. H. (1984) (London) 312,374-376. I. C., Watling, K. J., Gee, N. S., Aspley, S., Jackson, R. G.,

80

ROBERT Reid, G. G., Baker, R., Billington, D. C., Barnaby, son, P. D. (1988) Btichem. J. 249,143-148.

9. Ackermann, K., Gish, B. G., Honchan, (1987) Biochem. J. 242,517-524. 10. Sherman, W. R., Ackermann, and Zinbo, M. (1986) Biomed. 341. 11. Dean,

N. M., and Beaven,

R. J., and Lee-

M. P., and Sherman,

K. E., Berger, Environ. Mass

M. A. (1989)

J. BARNABY

W. R.

R. A., Gish, B. G., Spectrum 13, 333-

Anal.

Biochem.

183,199-

209. 12. Palmer,

S., and

Wakelam,

M.

(1989)

Biochim.

Biophys.

Acta

1014,239-246. 13. Meek,

J. L. (1986)

14. Mayr,

G. W. (1988)

Proc.

15. Smith, matogr.

R. E., MacQuarrie, Sci. 27,491-495.

N&Z.

Biochem.

Acad.

Sci. USA

83,

4162-4166.

J. 254,585-591. R. A., and Jope,

R. S. (1989)

J. Chro-

16. Lin, T-N., Sun, G. Y., Premkumar, N., MacQuarrie, R. A., and Carter, S. R. (1990) B&hem. Biophys. Res. Commun. 167(3), 1294-1301. 17. Johnson, D. C., and LaCourse, W. R. (1990) Anal. Chem. f32(10), 589-597. 18. Billington, D. C., Baker, R., Kulagowski, J. J., and Mawer, (1987) J. Chem. Sot. Chem. Commun. 4. 314-316. 19. Rocklin, 49-64.

R. D., Henshall,

20. Diehl, R. E., Whiting, meyer, D., Schoepfer, Biol. Chem. 265(11),

A., and Rubin, P., Potter, R., Bennett, 5946-5949.

R. B. (1990)

I. M.

Znt. Lab. 20,

J., Gee, N., Ragan, C. I., LineC., and Dixon, R. A. (1990) J.

21. Hughes, S., and Johnson, D. C. (1981) Anal. Chim. Acta. 22. Rocklin, R. D., and Pohl, C. A. (1983) J. Liq. Chromatogr. 1577-1590.

132,ll. 6(g),