Extractions of pyrroloquinoline quinone from crude biological samples

Life Sciences, Vol. 47, pp. 2135-2141 Printed in the U.S.A.

Pergamon Press

EXTRACTIONS OF PYRROLOQUINOLINE QUINONE FROM CRUDE BIOLOGICAL SAMPLES

O. Suzuki , T. Kumazawa , H. Seno*, T. Urakami** and T. Matsumoto***

Department of Legal Medicine, Hamamatsu U n i v e r s i t y School of Medicine, Hamamatsu

431-31;

Mitsubishi Gas Chemical Company, Inc.,

Tokyo 105; and ***First Department of Surgery, Nagoya University School of Medicine, Nagoya 466, Japan (Received in final form October I, 1990) Summary The best conditions for extractions of free pyrroloquinoline quinone (PQQ) from crude biological samples were investigated with various organic solvents and Sep-Pak C18 cartridges. PQQ was measured with use of its native fluorescence in aqueous solution. PQQ was well extracted into n-butanol under acid conditions, and addition of NaC1 did not improve the solvent extraction. PQQ, which had been extracted into n--butanol, could be re-extracted into an aqueous phase by addition of either ~-heptane or pyridine, or combination of them. PQQ, which had been adsorbed to Sep-Pak C18 cartridges, could be eluted with a mixture of pyridine and water with very excellent recovery. The recovery of I ~g PQQ, which had been added to I g human liver, brain and I ml plasma and had undergone the n_-butanol and the Sep-Pak extractions, was 50, 75 and 105 %, respectively. From the blank fluorescence, endogenous levels of free PQQ in human liver, brain and plasma were found not greater than 0.41, 0.08 and 0.13 ~ug/g or ml, respectively, if present.

Pyrroloquinoline quinone (PQQ) was identified as a novel co-factor in various dehydrogenases purified from bacteria in early studies (1,2). Recently, a few research groups have claimed that covalently bound PQQ is located in various mammalian enzymes, such as amine oxidases (3-6) and lysyl oxidase (7,8), suggesting ubiquitous presence of PQQ in mammalian tissues. More recently, it has been reported that PQQ is nutritionally important as a vitamin or growth factor ( 9 ) . However, no one has succeeded in detecting free PQQ in mammalian tissues. This seems partly because no extraction methods have been established for PQQ. In this report, we present simple methods for extracting PQQ from crude tissues, which are applicable to various PQQ studies.

Materials and Methods MATERIALS. PQQ was obtained from Mitsubishi Gas Chemical Company, Inc. (Niigata, Japan); Sep-Pak C18 cartridges from Waters Associates (Milford, MA). Other common chemicals used were of the highest purity commercially 0024-3205/90 $3.00 +.00 Copyright (c) 1990 Pergamon Press plc

2136

PQQ Extraction From Crude Samples

Vol. 47, No. 23, 1990

available. Human liver and brain were obtained at forensic plasma from Hamamatsu Red Cross Center (Hamamatsu, Japan).

autopsies;

human

FLUOROMETRIC MEASUREMENTS OF PQQ. PQQ in 0.1 N HCI solution was quantitated with its fluorescence (10), with excitation at 375 nm and emission at 465 nm (uncorrected), on a Hitachi 650-IOS fluorescence spectrophotometer. ORGANIC SOLVENT EXTRACTION. Three milliliters of each organic solvent listed in Table I and 3 ml 0 . 1 N HCI solution were shaken vigorously in the presence of 1 jug PQQ, and c e n t r i f u g e d at 500 g for 5 min. The aqueous phase was subjected to the fluorometric measurements. The aqueous phase after shaking with an organic solvent in the absence of PQQ served as a blank test. The standard sample was made by adding l ~ g PQQ, which had been dissolved in I ~ i water, to the final aqueous phase of the blank test. The same line of experiments was also carried out in the presence of 0.8 g NaCI. From the d e c r e a s e in PQQ in the aqueous phase after organic solvent extraction, partition coefficients of it were calculated. EXTRACTION OF PQQ FROM n-BUTANOL BACK INTO AN AQUEOUS PHASE. To test reextraction of PQQ, which had been extracted into ~-butanol, back into an aqueous phase, 3 ml n-butanol saturated with 0 . 1 N HCI was shaken with 3 ml water and I ~ g PQQ in the presence of NaCI, n - h e p t a n e or pyridine, or combination of them as shown in Table II. The mixtures were centrifuged at 500 g for 5 min, and each aqueous phase was subjected to the fluorometric measurements. In these experiments, each blank test was carefully prepared by shaking the mixtures without addition of PQQ. As a standard sample, I ~ g PQQ was added to the final aqueous phase of each blank test. ISOLATION WITH SEP-PAK C18 CARTRIDGES. For pretreatment of a Sep-Pak C cartridge, 10 ml 5 % p y r i d i n e in water, 10 ml a c e t o n i t r i l e and 10 m~ distilled water were passed through it. Recoveries of PQQ during its elution from S e p - P a k C18 cartridges were tested with 3 d i f f e r e n t eluents. One microgram of PQQ dissolved in 10 ml O . 1 N HCI was loaded on the cartridges. After washing the cartridges with 20 ml 0.001 N HCI, the adsorbed PQQ was eluted with 5 % pyridine (in water), acetonitrile or methanol, 3 ml of each. They were evaporated to dryness in vacuo and the residues dissolved in 3 ml O . 1 N HCI for the fluorometric measurements. Blank and standard tests were also prepared for each elution solvent. E X T R A C T I O N OF PQQ F R O M CRUDE B I O L O G I C A L SAMPLES. To 1 g tissue or I ml plasma including I ~ g PQQ, were added 4 ml I N HCI, 40 ~ i mercaptoethanel and 0.1 ml 10 % p o t a s s i u m f e r r i c y a n i d e . The tissue was minced with surgery scissors. After addition of 10 ml n-butanol to the above mixture, it was homogenized with a Polytoron homogenizer for 5 min being cooled with ice and c e n t r i f u g e d at 500 g for 5 min. After the organic layer was c a r e f u l l y transferred to another centrifuge tube, I ml distilled water, 1 ml pyridine, O.1 g NaCI and 20 ml ~ - h e p t a n e were added to it, shaken v i g o r o u s l y and c e n t r i f u g e d as above. The aqueous layer (lower phase) was c a r e f u l l y transferred to a 20 ml-test tube and evaporated in vacuo only for 15 min to remove the residual organic solvents from the aqueous solution completely. It was then mixed with 10 ml 0.1 N HCI solution and passed through the pretreated Sep-Pak C18 cartridge at a flow rate not greater than 5 ml/min. The c a r t r i d g e was washed with 20 ml 0.001 N HCI, and finally 3 ml 5 % pyridine in water was passed t h r o u g h it to elute PQQ. The eluates were evaporated to dryness in vacuo and their residues dissolved in 3 ml 0.1 N HCI solution for their fluorometric measurements. An extract, without addition of PQQ at the initial step, was taken as a blank test for each tissue. As a standard, 1 ~g PQQ was added to the final residue of each blank sample.

Vol. 47, No. 23, 1990

PQQ Extraction From Crude Samples

2137

TABLE I Partition coefficients of PQQ with various organic the presence and absence of NaCl. Solvent

- NaCI

n-Hexane

solvents in

+ NaCI a

0.06

0.04

<0.01

0.08

Benzene

0.06

0.02

Chloroform

0.18

O.04

Ethyl acetate

1.38

0.91

~-Heptane

n-Butanol

29.3

2-Butanol

5.10

28.4 14.6

2-Methyl-S-propanol

14.9

12.3

n-Pentanol

21.7

27.6

n-Hexanol

19.4

25.3"

The values are presented as the ratio of the concentration in organic phase to the concentration in aqueous phase (0.1 N HC1). They are the means of duplicate determinations. a The amount of NaC1 added to 3 ml aqueous phase was 0.8 g.

Results Table I shows partition coefficients of PQQ between each organic solvent and 0.1 N HC1 solution in the presence and absence of NaC1. n-Butanol showed the best extraction of PQQ, followed by ~-pentanol and n-hexanol. The addition of NaC1 caused slight changes in the coefficients according to each solvent, but did not appear to improve the extraction significantly. Thus we adopted n-butanol for PQQ extraction without addition of NaC1. Table II shows recoveries of PQQ in the aqueous phase, after PQQ was shaken with water and acidified ~-butanol in the presence of NaC1, n-heptane or pyridine, or combination of them. All these additives showed positive effects on its recovery, showing that PQQ, which had been extracted into nbutanol, could be easily re-extracted into an aqueous phase with their addition. The recovery of I ~ g PQQ, which had been adsorbed to Sep-Pak C18 cartridges under acid conditions, were 103 ± 2, 47 ~ O.S and 59 ~ 4 % (means ± SD of 3 experiments) with 5 % pyridine (in water), acetonitrile and methanol as eluents, respectively. On the basis of the above data, we have established a procedure for PQQ extraction from crude biological samples by combining the shaking with n_butanol and isolation with a Sep-Pak C18 cartridge. The recovery of 1 ~g,

12 29 89 tO0 IO0

3 ml n_-Butanol b + 3 ml water + I ~ g PQQ

3 ml n-Butanol + 3 ml water + I ~ g PQQ + 0.8 g NaCI

3 ml n-Butanol + 3 ml water + I ~ g PQQ + 6 ml n-heptane

3 ml ~-Butanol + 3 ml water + I # g PQQ + I ml pyridine

3 ml n-Butanol + 3 ml water + I ~ g PQQ + 6 ml B-heptane + I ml pyridine

a The values are the means of duplicate determinations. b n_-Butanol had been saturated with 0 . 1 H C 1 solution prior to the experiments.

Recovery of PQQ in aqueous phase (%)a

Composition of the mixture for shaking

Recovery of PQQ in aqueous phase after shaking PQQ with acidified n-butanol and water in the presence of NaCI, n-heptane or pyridine

TABLE II

Z O

O

C~

O

o

m

~D

co

Vol. 47, No. 23, 1990

PQQ Extraction From Crude Samples

2139

which had been added to 1 g human liver, 1 g human brain and 1 ml human plasma, were 50 ± 2, 75 ± I and 105 + I % (means ± SD of 3 experiments), respectively. If free endogenous PQQ is present in human liver, brain and plasma, it should be included in the blank fluorescence of an extract without addition of PQQ in this study. The blank fluorescence intensity for the liver, brain and plasma extracts was equal to fluorescence intensity of 0.41, 0.08 and 0.13 ~ g PQQ/g or ml, respectively (the mean values of 3 experiments). Therefore, actual levels of free PQQ in the human samples are not greater than the above values, even if present.

Discussion There are some reports dealing with determination of free PQQ (11-14), which can be grouped into high-performance liquid chromatography (HPLC) and enzymatic methods. The HPLC methods suffer from persistent adsorption of PQQ to HPLC columns, which causes long bleeding of PQQ during its repeated running (11,13). The enzymatic assays require tedious preparation of apoenzyme of glucose dehydrogenase from bacteria (11-14). Both groups of methods have never been applied to crude biological samples with high protein contents, but applied only to relatively clean aqueous solutions of PQQ. This seems partly because no extraction methods have been established for PQQ as mentioned earlier. We have used fluorometric detection of PQQ in the present study because of its simplicity (10). Although it does not seem highly specific to PQQ, it gave no problems for the present addition tests when blank and standard samples were carefully taken. From the blank fluorescence for the extracts of human liver, brain and plasma, the levels of free PQQ were found not greater than 0.41, 0.08 and 0.13 ~g/g or ml, if present. The latter value for human plasma is in contrast with that of Fl~ckiger et al. (15), who reported that PQQ level in human plasma is as high as 1 0 ~ g T m l when measured by their spectrophotometric method with nitroblue tetrazolium reduction. To test whether the above values are actual levels of free PQQ in human samples, studies by gas chromatography/mass spectrometry (GC/MS) are now in progress in our laboratory. We have added potassium ferricyanide and mercaptoethanol to tissues at the initial step of extraction. The ferricyanide was added to prevent PQQ from its binding to hemoglobin (16). Mercaptoethanol was added to reduce PQQ and thus to avoid reaction of PQQ with amino acids and amines. We have adopted n_-butanol for liquid-liquid extraction of PQQ, because it showed the highest partition coefficient (Table I). PQQ probably forms a reversible adduct with n-butanol (10), which can be easily backed to free PQQ in aqueous solution (Table II). Very recently, Duine's group has proposed 2hexanol extraction, coupled with HCl-hydrolysis of enzyme proteins, to identify PQQ in'dopa decarboxylase (17) and galactose oxidase (18), but has not applied it to any crude tissue. We preferred n_-butanol to n-hexanol because the former requires a shorter time for evaporation in addition to the above better extraction of PQQ (Table I). The tissues should be homogenized together with ~-butanol, because it seems very useful to detach PQQ from its adsorption to membranes and proteins. When a tissue was homogenized only with acid aqueous solution, a major amount of PQQ precipitated with proteins during centrifugation and its

2140

PQQ Extraction From Crude Samples

r e c o v e r y in the observation).

supernatant

fraction

became

Vol. 47, No. 23, 1990

very

low

(unpublished

Sep-Pak CIR cartridges were sometimes used for clean-up of PQQ or its adducts (19). ~ n all cases, methanol or a methanol-water mixture was used for elution of PQQ from the cartridges. During our experiments to follow this procedure, we noticed that the recovery of PQQ in the methanol eluate was low, e s p e c i a l l y when small amounts of PQQ (less than I ~g) had been adsorped to the cartridges. We have tested many solvent systems as eluents and have found that pyridine in water (5-50 %) is very suitable for elution of PQQ from the cartridges and gives 100 % recovery as shown in the text. PQQ was almost freely soluble, and stable for a long period in the pyridine solution (pH 8.5). The pyridine can be completely removed by evaporation when it is necessary. To identify covalently bound PQQ in purified mammalian enzymes, the enzymes was reacted with hydrazines in situ, such as 2,4-dinitrophenylhydrazine and phenylhydrazine, before digestion with proteolytic enzymes; the PQQ-hydrazone liberated from proteins was finally detected by HPLC (3,4,7) and by resonance Raman spectroscopy (5,6,8). The in situ derivatization was aimed to avoid reaction of PQQ with amino acids or amines to form various unidentified products during the digestion. However, the hydrazine methods were far from being quantitative, because the yields of PQQ-hydrazones were very low and also variable under various conditions (4)In addition, phenylhydrazine in solution rapidly decomposed by autoxidation, which made the analyses much less reliable (20). In the present study, we have d e m o n s t r a t e d that free PQQ is well extracted into ~-butanol and can be easily re-extracted into an aqueous phase by addition of NaCI, pyridine and/or ~-heptane. To isolate PQQ with Sep-Pak C18 cartridges, a pyridine-water mixture has been found of the best choice as an eluent. By combining the ~-butanol extraction with Sep-Pak C18 cartridge isolation, I ~ g PQQ, which had been added to human brain and plasma, could be detected with satisfactory recovery, though it was 50 % with human liver. The present i n f o r m a t i o n s on PQQ e x t r a c t i o n s seem useful for any type of studies on PQQ analyses, such as GC, HPLC and GC/MS, with crude tissues, enzymes, foods and bacteria.

References 1.

S. A. SALISBURY, 280

2. 3. 4. 5. 6. 7. 8.

9. 10.

H. S. FORREST, W. B. T. CRUSE and

O. KENNARD,

Nature

843-844 (1979).

J. A. DUINE, J. FRANK, JZN. and J. A. JONGEJAN, FEMS Microbiol. Rev. 32 165-178 (1986). C.L. LOBENSTEIN-VERBEEK, J. A. JONGEJAN, J. FRANK and J. A. DUINE, FEBS Lett. 170 305-309 (1984). R. A. VAN DER MEER, J. A. JONGEJAN, J. FRANK, JZN. and J. A. DUINE, FEBS Lett. 206 111-114 (1986). R. S. MOOG, M. A. MCGUIRL, C. E. COTE and D. M. D00LEY, Proc. Natl. Acad. Sci. USA 83 8435-8439 (1986). G. F. KNOWLES, K. B. PANDEYA, F. X. RIUS, C . M . SPENCER, R. S. MOOG, M. A. MCGUIRL and D. M. DOOLEY, Biochem. J. 241 603-608 (1987). R. A. VAN DER MEER and J. A. DUINE, Bioehem. J. 239 789-791 (1986). P. R. WILLIAMSON, R. S. MOOG, D. M. DOOLEY and H. M. KAGAN, J. Biol. Chem. 261 16302-16305 (1986). J. KILLGORE, C. SMIDT, L. DUICH, N. ROMERO-CHAPMAN, D. TINKER, K. REISER, M. MELKO, D. HYDE and R. B. RUCKER, Science 245 850-852 (1989). R. H. DEKKER, J. A. DUINE, J. FRANK, JZN., P. E. J. VERWIEL and J.

Vol. 47, No. 23, 1990

11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

PQQ Extraction From Crude Samples

2141

WESTERLING, Eur. J. Biochem. 125 69-73 (1982). J. A. DUINE, J. FRANK, JZN. and J. A. JONGEJAN, Anal. Biochem. 133 239-243 (1983). M. AMEYAMA, M. NONOBE, E. SHINAGAWA, K. MATSUSHITA and O. ADACHI, Anal. Biochem. 151 263-267 (1985). M. A. G. VAN KLEEF, P. DOKTER, A. C. MULDER and J. A. DUINE, Anal. Biochem. 162 143-149 (1987). 0. GEIGER and H. GORISCH, Anal Biochem. 164 418-423 (1987). R. FLUCKIGER, T. W00DTLI and P. M. GALLOP, Biochem. Biophys. Res. Commun. 153 353-358 (1988). K. SATO, Y. KATSUMATA, M. AOKI, M. 0YA, S. YADA and 0. SUZUKI, Forensic Sci. Int. 17 177-184 (1981). B. W. GROEN, R. A. VAN DER MEER and J. A. DUINE, FEBS Lett. 237 98102 (1988). R. A. VAN DER MEER, J. A. JONGEJAN and J. A. DUINE, J. Biol. Chem. 264 7792-7794 (1989). J. A. DUINE and J. FRANK, JR., Biochem. J. 187 221-226 (1980). J. G. ROBERTSON, A. KUMAR, J. A. MANCEWICZ and J. J. VILLAFRANCA, J. Biol. Chem. 264 19916-19921 (1989).