High-performance liquid chromatographic determination of biotin in biological materials after crown ether-catalyzed fluorescence derivatization with panacyl bromide

ANALYTICAL

BIOCHEMISTRY

200,89-94

(1992)

High-Performance Liquid Chromatographic Determination of Biotin in Biological Materials after Crown Ether-Catalyzed Fluorescence Derivatization with Panacyl Bromide Jiirgen Stein,l Andreas Hahn, Bernhard

Lembcke,’ and Gertrud Rehner

Institute of Nutrition, Justus-Liebig University, Wilhelmstrasse20, D-6300 Giessen,Germany

Received

March

21, 1991

This paper reports a high-performance liquid chromatographic (HPLC) technique to determine biotin in biological samples. Biotin and the internal standard dethiobiotin are converted into fluorescent derivatives by using panacyl bromide [p-(9-anthroyloxy)phenacyl bromide] as a fluorescence label. Biotin is extracted from biological tissue with trichloroacetic acid and the extract is purified by a combination of solid-phase extraction on C,, cartridges, ion-exchange chromatography on DOWEX formate resin, and thin-layer chromatography. The purified sample extract is derivatized in the presence of a crown ether. The resulting panacyl esters can be separated on reversed-phase as well as on normal-phase HPLC. Normal phase HPLC is preferable because it provides higher sensitivity and demands less sample pretreatment. Analysis of rat intestinal tissue revealed that only about 13% of the biotin is present in free form whereas 87% is bound in proteins from which it can be released by hydrolysis. Biotin values determined by this method are comparable to those obtained by other techniques. o 1892 Academic press, IN.

Biotin (hexahydro-2-oxo-lH-thieno(3,4-d)imidazole4-pentanoic acid) is a water-soluble vitamin that is a coenzyme of some carboxylating enzymes such as pyruvate carboxylase (EC 6.4.1.1) and acetyl-CoA carboxylase (EC 6.4.1.2) (1). Its determination presents a difficult analytical problem because the molecule does not possess any suitable uv absorbance or other character’ Present address: Department of Gastroenterology, ternal Medicine, Johann-Wolfgang-Goethe University, Stern-Kai 7, D-6000 Frankfurt 70, Germany.

0003-2697192 $3.00 Copyright 0 1992 by Academic Press, All rights of reproduction in any form

Inc. reserved.

Center of InTheodor-

istic that would be suitable for its selective and sensitive detection. Up to now the most common analytical procedures have been based on microbiological assays (24). These methods are sensitive and specific but less reliable and therefore yield varying results (1,5). Within the last few years another type of method has been proposed that is based on the reaction between biotin and avidin, a basic glycoprotein found in raw eggs. This reaction is highly specific and can be used in radiometric tests which were applied to different biological materials (6-9). Different attempts were made to detect biotin after conversion to uv absorbing (10-12) or fluorescing compounds (12). Although these techniques were combined with thin-layer chromatography (11) or highperformance liquid chromatography (HPLC) (12), their suitability for analyzing biological samples has not yet been shown. Gas chromatography of trimethylsilyl derivatives of biotin was described (13,14) but is less sensitive and therefore not suited for the analysis of biological materials. Other HPLC techniques to separate biotin using uv detection of the carbonyl function of the biotin molecule (15), mass spectrometry of biotin methyl ester (16), electrochemical detection (17), or fluorimetric detection of biotin-9-anthryldiazomethane esters (18) were developed, but except for two cases (16,18) they have not been used to determine the biotin content of complex biological materials. The aim of the present study was to establish a new HPLC technique to determine biotin after precolumn derivatization of the carboxylic group with fluorescent panacylbromide [p-(9-anthroyloxy)phenacyl bromide] in the presence of a crown ether. It is shown that the analytical procedure is applicable to biological materials. 89

90 MATERIALS

STEIN

AND

METHODS

Chemicals Biotin and dethiobiotin were obtained from Sigma (Deisenhofen, Germany) and diluted in distilled water to final concentrations of 100 PM. [carboxy-14C]Biotin (specific activity 22.8 mCi/mmol) was a gift from Dr. Weber (Hoffmann-La Roche, Basle, Switzerland). It was checked by thin-layer chromatography on silica 60 layers (Merck, Darmstadt, Germany) using chloroform/methanol (95/5, v/v) as an eluent and found to be 98% pure. Dibenzo-18-crown-62 (2,3,11,12-dibenzo-1,4,7,10,13,16-hexaoxacyclooctadeca-2,lldien) and panacylbromide were obtained from Aldrich (Steinheim, Germany). The other chemicals and solvents were obtained from Merck (Darmstadt, Germany). All substances were of the highest purity available. Water was purified with a Millipore Q system (Waters, Eschborn, Germany).

Development

of the Precolumn

Derivatization

Procedure

Standards as well as sample extracts were dried under a nitrogen stream and dissolved in 100 ~1 acetone containing 250 nmol panacylbromide and 50 nmol dibenzo18-crown-6. In all cases 2.5 nmol dethiobiotin was added as an internal standard. After adding 20-30 mg K&O, the reaction tubes were closed and transferred to a water bath at either 37 or 57°C. At certain intervals 20 ~1 was withdrawn from the tube and analyzed by thinlayer chromatography as mentioned above.

Nuclear Magnetic Products

Resonance Spectra of Reaction

To characterize the panacyl esters of biotin and dethiobiotin, we recorded ‘H nuclear magnetic resonance (NMR) spectra of the derivatization product. The substances were dried under a nitrogen stream and subsequently dissolved in [2H]chloroform. Readings were taken with a Varian HA-100 spectrometer at 400 MHz. Chemical shifts were expressed in parts per million relating to tetramethylsilane as an internal standard.

Chromatography Separation of the panacyl esters of biotin and dethiobiotin was carried out on a Merck-Hitachi HPLC system consisting of a gradient former L-5000, a solvent metering pump 655A-11, a fluorescence detector F-1000 and loop injector (Rheodyne, Model 7125) with a loo-/11 syringe. All analyses were performed at room temperature. Two chromatographic systems were tested based * Abbreviations used: dibenzo-lo-crown-g, 10,13,16-hexaoxacyclooctadeca-2,11-dien; p-DACA, p-dimethylaminocinnamaldehyde; methane.

2,3,11,12-dibenzo-1,4,7,TCA, trichloroacetic acid; ADAM, g-anthryldiazo-

ET

AL.

either on normal-phase or on reversed-phase chromatography, respectively. In both cases we used a 4.6 X 150-mm column which had been filled by the upward slurry technique (19) using 2-propanol for preparing the slurry. For reversed-phase analyses Hypersil ODS 3 pm (Shandon, Frankfurt, Germany) was used as a stationary phase. Gradient elution was performed with water and methanol. Composition of the eluent was linearly changed from 60:40 (water:methanol) to 30:70 within 15 min. Flow rate was 1 ml/min. Normal-phase chromatography was done on Shandon Hypersil 3 pm. Elution was obtained isocratically with 5% methanol/95% dichloromethane at a flow rate of 1.4 ml/min.

Detection Fluorescence maxima of the biotin panacyl esters in different mobile phases were determined using a Hitachi F-2000 fluorescence spectrometer. Peaks were recorded at excitation and emission wavelengths of 380 and 470 nm, respectively. Calibration curves were calculated on the basis of peak area using least-squares regression analysis.

Extraction of Biotin from Biological Sample Cleanup Procedure

Materials

and

The method described was designed primarily to determine the biotin content of the gut wall. Therefore it was validated using intestinal tissue as a model matrix. Rat gut was obtained from male Wistar rats (Winkelmann, Borchen, Germany) with an average body weight of 230-250 g. The animals were killed by a blow on the neck; the whole small intestine was removed and rinsed with ice-cold saline. All further steps were carried out at 4’C. Two to three grams of gut tissue was cut by scissors and subsequently homogenized in 5 ml of a 5% trichloroacetic acid (TCA) solution after dethiobiotin had being added (5 nmol). The homogenate was centrifuged for 15 min at 10,OOOgand the pellet was reextracted with 5 ml of TCA two more times. The combined supernatants were applied to a C,, solid phase extraction column (Sep-Pak, Waters, Eschborn, Germany) that had been pretreated with 10 ml of methanol and 10 ml of water. Polar compounds were removed by washing the column with 10 ml of 2% acetonitrile and biotin was eluted with 10 ml of 15% acetonitrile. The eluted fraction was then transferred to a 0.8 X 7-cm DOWEX 1 X 8 formate column (200-400 mesh, Serva, Heidelberg, Germany) which was washed afterward with 10 ml of water and 10 ml of 0.1 N potassium formate. Biotin and dethiobiotin were eluted with an

CHROMATOGRAPHIC

additional

30 ml of 0.1 N potassium

formate.

DETERMINATION

The eluate

wasagainapplied to a preconditionedSep-Pak&cartridge, washed with 10 ml of water, and eluted with 10 ml of methylformate. The methylformate was evaporated under a nitrogen stream and the extract was dissolved in 50 ~1 of chloroform. For further purification this solution was chromatographed on silica 60 layers using chloroform/methanol/ potassium formate (90/9/l) as an eluent. To simplify the identification of the substance spots, standard solutions of biotin and dethiobiotin were chromatographed on every layer. When chromatography was terminated, the edges of the layer containing the standards were cut off with scissors and coloured with p-dimethylaminocinnamaldehyde (p-DACA, 10). The correspondingsample spots were scraped off, transferred to reaction tubes, lyophilized and derivatized as described above. During the method development the loss of substance within the whole procedure was calculated after adding 0.5 &i [14C]biotin to the homogenate. Recovery of added unlabeled biotin was determined by carrying out the whole analysis with homogenates from rat gut which were divided into two parts and by adding 0.5 nmol of biotin to one half. Analysis

of Rat Intestinal

Tissue and Rat Liver

Rat intestinal tissue (duodenum, proximal jejunum, distal jejunum) was analyzed as described above. Free biotin was analyzed directly whereas protein-bound biotin could be measured after hydrolytic treatment of the homogenate with 6 N sulfuric acid as explained by others (20). For comparison with established methods the biotin content of rat liver was determined as described above.

a

I

OF

91

BIOTIN

RESULTS

Derivatization Procedure

As can be seenfrom the radiochromatogramsin Fig. la conversion of biotin into the corresponding panacyl ester is a function of incubation time. The derivatization kinetics (Fig. lb) emphasize that about 90% of the biotin could be converted within 3 h at 57°C but efhciency varied from 73 to 98%. Nevertheless, when correcting these variations using dethiobiotin as an internal standard the technique was highly reproducible. Results of derivatizing the same sample 10 times indicated a coefficient of variation of 7.8%. NMR

Spectra of Derivatization

Products

It could be concluded from the spectra of the panacyl esters of biotin and dethiobiotin that only the carboxylic function of biotin participates in the formation of the panacyl ester whereas th.e amide functions of the molecule can still be detected at 5.8 and 6.3 ppm. Thus the reaction sequence shown in Fig. 2 is established. Chromatography

and Detection

As Figs. 3a and 3b indicate, both reversed-phase and normal-phase HPLC were well suited to achieve complete separation of the panacylesters of biotin and dethiobiotin within 7 min. The fluorescence maxima of both derivatives were found to be 380 and 470 nm for excitation and emission wavelengths, respectively, in both mobile phases (Fig. 4). As can be seen from Fig. 5, normal-phase chromatography, i.e., an aprotic eluent, provides higher sensitivity. Furthermore, additional tests revealed that reversed-phase chromatography was disturbed by side products of the derivatization reaction

b

FIG. 1. Conversion of biotin to the corresponding panacyl ester as examined of biotin (B) and biotin panacyl ester (BPE) obtained at distinct intervals Time-dependent formation of biotin panacyl ester at 37 and 57°C.

by thin layer underlining

chromatography. the conversion

(a) Typical radiochromatograms of biotin to its panacyl ester.

(b)

STEIN ET AL.

92

P C HN’

‘NM

sr-Cl-l*COOH

FIG. ester.

2.

Reaction

sequence

for

the

formation

of biotin

panacyl 260

and by free panacyl bromide. These impurities had to be removed by an additional thin-layer chromatographic procedure before HPLC analysis was performed.

Extraction Liver

Procedure

and Analysis

of Rat Gut and Rat

As judged by the radioactivity of [14C]biotin, the final HPLC peak contained 61.2 f 4.9% (n = 8) of the substrate added to the gut homogenates. Recovery of biotin was 93.8 + 3.9% (n = 10) as calculated by means of the internal standard. Fluorescence spectra recorded from the peaks during the analysis of biological samples corresponded to those of standard solutions. Figure 6 shows chromatograms of free biotin and total biotin, i.e., after hydrolysis of protein-bound biotin underlining that only minor impurities are still present in the sample. Table 1 summarizes the biotin content of different

b

a

360

440

520

600

(nm)

FIG. 4. Fluorescence spectra of biotin panacyl mobile phases indicating that relative fluorescence presence of an aprotic eluent.

ester in different is higher in the

parts of rat small intestine indicating that the major amount of biotin, i.e., ca. 85%, exists as bound biotin. Significant differences concerning the biotin content and the biotin pattern of the different intestinal segments were not found. Biotin content in rat liver (Table 2) as determined by our method is within the same range as found by other investigators indicating the accuracy of the technique described. DISCUSSION

Problems in biotin analysis mainly arise from the lack of a suitable detection system which permits the quantification of the molecule in the picomolar range. Therefore formation of uv-absorbing or fluorescing derivatives is essential when chromatographic procedures are to be used in biotin analysis. Reaction of biotin and dethiobiotin with panacylbromide gives rise to the for-

l

.

P b Y P a

normal-phase reversed-phase

HPLC (r=0.999) HPLC (r=0.993)

1

-*----S

0

2

4

6

8

10 12

(min)

O-2

(min)

FIG. 3. Chromatographic separation of the panacyl esters of dethiobiotin (1) and biotin (2) applying normal-phase (a) and reversedphase (b) HPLC. Retention times: (a) normal phase HPLC: biotin 6.46 min, dethiobiotin 5.56 min; reversed-phase HPLC: biotin 4.67 min, dethiobiotin 5.71 min.

__C_----a---I 200

I 400

pm01

I 600

/

’

I 800

I 1000

injection

FIG. 6. Calibration curves for the HPLC determination of biotin panacyl ester as obtained after carrying out the whole sample cleanup with standard solutions of biotin.

CHROMATOGRAPHIC

DETERMINATION

OF

93

BIOTIN

TABLE

Content of Rat Liver as Determined by Various Investigators”

Biotin

Biotin content mean + SD (pg/g)

Procedure HPLC Microbiological Microbiological Microbiological Microbiological 0 As far dimension

4

6

8 10 12

in 2,

mation of highly fluorescing esters of these substances and can be used for HPLC determination of biotin in biological materials. Carboxylic function is potentially suited for derivatization as can be seen from various techniques developed for the labeling of organic acids such as fatty acids (2128) as well as from the few attempts to form uv absorbing or fluorescing derivatives of biotin (12,18). Until now biotin content of biological materials has been determined by HPLC applying fluorescence derivatization in one case only (18). However, examination of the chromatogram presented in (18) indicates that. biotin appears as a very small peak among very large impurities; the method using 9-anthryldiazomethane (ADAM) as a derivatization reagent is thus not optimal for analyzing biological samples. A derivatization reagent for carboxylic functions should have two properties: The derivatives formed should permit sensitive detection and the reaction should be relatively specific for carboxylic functions. These requirements are fulfilled by several fluorescent labels including panacyl bromide. When standard solutions of biotin were analyzed after carrying out the whole cleanup procedure the detection limit was about 10 pmol using normal-phase and about 100 pmol apply-

in Different Duodenum

Free biotin Protein bound Total biotin

biotin

0.97 + 0.21 assay assay assay assay

Present

2.38 0.30

1.23 + 0.47 0.89 f 0.12

as necessary the original rglg fresh weight.

data

work 31 32 33 34

have

been

changed

to the

ing reversed-phase HPLC. A third important factor responsible for adequate derivatization is the reactivity of the labeling reagent in the picomolar range which is necessary for the analysis of biological samples. This was reported to be sufficient with panacyl bromide only (29) and could be confirmed during the method development by derivatizing very small amounts of biotin (ca. 22 pmol, Fig. 1). Formation of biotin esters such as panacyl derivatives may be disturbed by the poor reactivity of the carboxylic function. Reactivity can be enhanced by adding a crown ether as a catalyst as first described by Durst et al. (21). Crown ethers complex metal salts, i.e., they are also able to form the anionic form of biotin. Furthermore, they contribute to the dissolution of the anionic molecule into an aprotic eluent. Thus the nucleophilic properties of the biotinate anion are enhanced and nucleophilic attack of biotin on panacyl bromide is facilitated. Furthermore, in contrast to other chemicals used for initiating the reaction such as triethylamine (30), crown ether catalyzed derivatization is not disturbed by traces of water remaining in the sample. Nevertheless, to correct possible deviations arising from the derivatization procedure as well as from the sample pretreatment, we propose using dethiobiotin as an internal standard. The recovery rate of our method is much higher and deviations are lower than those observed analyzing ADAM esters of biotin (18), although these investigators did not apply any cleanup procedure. We assume that higher precision and recovery are at least partially due to the use of an internal standard.

TABLE Content

Reference

(min)

(min)

FIG. 6. Typical chromatograms of free (a) and total (b) biotin rat intestinal tissue as determined by normal-phase chromatography of biotin and dethiobiotin panacyl ester. Peaks: 1, dethiobiotin; biotin.

Biotin

1

0.43 2.66 3.09

+ 0.07 k 0.45 + 0.56

Parts

of Rat % 13.9 86.1 100

Small

2

Intestine

(nmol/g

Jejunum 0.44 2.87 3.31

+ 0.08 + 0.52 * 0.58

f

SD;

% of Total

Biotin,

% 13.3 86.7 100

N = 12) Ileum

0.41 2.77 3.18

+ 0.07 + 0.49 f 0.57

% 12.9 87.1 100

STEIN ET AL.

94

Furthermore, the sample pretreatment presented in this paper is very specific as can be seen from the fact that only minor impurities were observed in the chromatogram. ACKNOWLEDGMENTS The investigations were sponsored by the Deutsche Forschungsgemeinschaft (Re 437/5-l). The authors are grateful to Professor Dr. Sattler for giving them the opportunity to work in the laboratories of the Radiation Center (Central Department) of the Justus Liebig University, Giessen. We thank Dr. Weber from Hoffmann-La Roche, Basle, for supplying us with [“Clbiotin and Dr. Kalinowski from the Institute of Organical Chemistry of the Justus Liebig University for performing the NMR analyses.

11. Groningsson,

K., and Jansson,

L. (1979)

J. Pharm.

Sci. 66,364-

365. 12. Desbene, P. L., Coustal, them. 128,359-362.

S., and Frappier,

F. (1983)

Anal.

Bio-

13. Viswanathan, V., Mahn, F. P., Venturella, V. S., and Senkowski, B. Z. (1970) J. Pharm. Sci 69,400-406. 14. Johnson, R. N., and Boyden, G. R. (1977) J. Pharm. Sci. 66, 1212-1213. 15. Chastain, J. L., Bowers-Komro, D. M., and McCormick, D. B. (1985) J. Chromatogr. 330, 153-158. 16. Azoulay, M., Desbene, P.-L., J. Chromatogr. 303,272-276. 17. Kamata, Nakayama,

K., Hagiwara, K., and Akiyama,

Frappier,

F., and Georges,

T., Takahashi, M., K. (1986) J. Chromatogr.

Y. (1984)

Uehara, S., 356,326-

330. 18. Hayakawa,

K., and Oizumi,

J. (1987)

413, 247-

J. Chromutogr.

250. 19. Bristow, P. A., Brittain, (1977) J. Chromatogr.

REFERENCES 1. Bonjour, J.-P. (1984) in Handbook of Vitamins (Machlin, L. J., Ed.), pp. 403-435, Dekker, New York/Basel. 2. Baker, H., Frank, O., Matovitch, V. B., Pasher, I., Aaronson, A., Hutner, S. H., and Sobotka, H. (1962) Anal. Biochem. 3,31-39.

3. Baker,

H., Frank, O., Ziffer, H., Goldfarb, S., Leevy, Sobotka, H. (1964) Am. J. Clin. Nutr. 14, l-6.

C. M.,

and

4. Baker,

H., Frank, O., Thomson, A. D., Langer, A., Munves, E. D., De Angelis, B., and Kaminetzky, H. A. (1975) Am. J. Clin. Nutr. Z&59-65. 5. Bonjour, J.-P. (1977) Znt. J. Vit. Nutr. Rex 47, 107-118.

6. Horsburgh, 7.

T., and Gompertz, D. (1978) Clin. Chim. Aeta 82, 215-223. Dakshinamurti, K., and Allan, L. (1979) in Methods in Enzymology (McCormick, D. B., and Wright, L. D., Eds.), Vol. 62, pp. 284-287, Academic Press, San Diego.

8. Sanghvi, R. S., Lemons, (1982) Clin. Chim. Acta 9. Mock, D. M., and DuBois,

R. M.,

Baker,

H.,

and

Thoene,

J. G.

D. B. (1986)

Anal.

Biochem.

153,272-

278. 10. McCormick,

226-236.

D. B., and

Roth,

20. Schreiner, J. M. (1985) 21. Durst, H. D., Milano, 22. 23. 24. 25. 26.

J. A. (1970)

Anal.

Biochem.

34,

E. (1975)

Diinges,

W. (1977)

Diinges,

W., and Seiler,

Distler, Korte,

(1984)

Ann.

N.Y.

Anal.

Chem.

H., Hayashi, T., Naruse, 234, 121-130.

H., Hayashi, J. Chromatogr.

Ad.

Sci. 447,

and

243,153-157. H., and N. Takagi

T., Sato, M., Tatsumi, 309, 43-52.

M. (1979)

(1982)

M., and Takagi,

J. N.

W. J. M., and Offermann 215-222.

Cox, J. W., and Pullen, R. H. (1984) And. Chem. W. D., and Peterson, M. B. (1982) And.

30. Watkins, 30-40.

M., and Suschetet,

S. A.,

192, 240-246.

H., Hulshoff, A., Underberg, F. B. J. M. (1984) J. Chromatogr. 290,

Lhuissier, 312-316.

420-422.

145,483-488.

J. Chromatogr.

28. Lingeman, 29.

B. F.

49,442-444.

N. (1978)

W. (1980) J. Chromatogr. W. D. (1982) J. Chromatogr.

Tsuchiya, Chromatogr.

C. M., and Williamson,

M., Kikta, E. J., Connelly, an&. Chem. 47, 1797-1801.

Grushka,

27. Tsuchiya,

31.

124,85-90.

P. N., Riley,

131, 57-64.

56,1866-1870. Biochem.

Znt. J. Vit. Nutr.

125,

Res. 49,

32. Baker, H. (1985) Ann. N.Y. Acad. Sci. 447,129-132. 33. Misir, R. (1986) Znt. J. Vit. Nutr. Res. 57, 65-70. 34. Watanabe, T., and Endo, A. (1989) J. Nutr. 119. 255-261.