A Salsolinol Synthase from Rat Brain

Biologia 66/6: 1183—1188, 2011 Section Cellular and Molecular Biology DOI: 10.2478/s11756-011-0134-y

Enzymatic condensation of dopamine and acetaldehyde: a salsolinol synthase from rat brain

Xuechai Chen, Abida Arshad,HongQing,RuiWang, Jianqing Lu & Yulin Deng*

School of Life Science & Technology, Beijing Institute of Technology, 5 South Zhongguancun Street, Haidian District, Beijing 100081, People’s Republic of China; e-mail: [email protected]

Abstract: Salsolinol (1-methyl-6,7-dihydroxy-1,2,3,4-tetrahydroisoquinoline; Sal) is structurally similar to 1-methyl-4- phenyl-1,2,3,6-tetrahydropyridine, which is supposed to have a role in the development of Parkinson-like syndrome in both human and non-human subjects. In the human brain, the amount of (R)-enantiomer of Sal is much higher than (S)-enantiomer, suggesting that a putative enzyme may participate in the synthesis of (R)-salsolinol, called (R)-salsolinol synthase. In this study, the (R)-salsolinol synthase activity in the condensation of dopamine and acetaldehyde was investigated in the crude extracts from the brains of Sprague Dawley rats. Identiﬁcation of the enzymatic reaction products and enzyme activity detection were achieved by HPLC-electrochemical detection. The discovery of this enzyme activity in rat’s brain indicates the natural existence of (R)-salsolinol synthase in the brains of humans and rats, and it is distributed in most brain regions of rat with higher activity in soluble proteins extracted from striatum and substantia nigra. Key words: salsolinol; dopamine; salsolinol synthase; enzyme activity; protein puriﬁcation. Abbreviations: AcH, acetaldehyde; DA, dopamine hydrochloride; ECD, electrochemical detection; EDTA, ethylene diamine tetraacetic acid; PD, Parkinson’s disease; Sal, Salsolinol – 1-methyl-6,7-dihydroxy-1,2,3,4-tetrahydroisoquinoline; (R)-Sal, (R)-enantiomer salsolinol; (S)-Sal, (S)-enantiomer salsolinol; SD, Sprague-Dawley; THP, tetrahydropapaveroline.

Introduction However, the transport of Sal into brain through the blood-brain barrier is excluded (Origitato et al. 1981; The pathogenesis of idiopathic Parkinson’s disease Song et al. 2006), indicating the in situ synthesis of Sal (PD) remains to be elucidated. Many reports have in the brain. Sal has an asymmetric centre at C1 and revealed the involvement of neurotoxins in the de- exists as (R)- and (S)-enantiomers (Deng et al. 1997). terioration of nigrostriatal dopamine system in PD In humans, reports have shown that (R)-enantiomer of (Maruyama et al. 1997). Three endogenous alkaloids Sal can only be detected (Deng et al. 1995), whereas were proposed as potential candidates of dopaminergic others indicated that the enantiomer ratio of (R)-/(S)- neurotoxins named salsolinol (l-methyl-6,7-dihydroxy- Sal was approximately 2 (Musshoff et al. 2000). In pre- l,2,3,4-tetrahydroisoquinoline; Sal), tetrahydropapave- vious studies, the concentration of (R)-Sal was found roline (THP) derivative (1BnTIQ) and β-carbolines. to be higher than the (S)-Sal, depicting the presence Sal, in particular, is toxic to dopaminergic neurons of such biosynthetic enzyme in etiology of disease. Fig- and has recently gained much attention due to its ure 1 illustrates two proposed synthetic pathways of structural similarity with 1-methyl-4-phenyl-1,2,3,6- Sal from dopamine and acetaldehyde (AcH). The first tetrahydropyridine, a potent neurotoxin that specifi- one produces the racemic mixture of both enantiomers cally degenerates the nigrostriatal dopaminergic neu- (Robbins et al. 1968; Zhu et al. 2008), the latter pro- rons, resulting in dopamine deficiency in the striatum, duces (R)-Sal only (Naoi et al. 1996; Rojkovicova et al. and hence the symptoms of Parkinsonism (Quan et al. 2008). Accordingly, the surplus of (R)-Sal detected in 2005). Therefore, Sal may have an influential role di- the human brain may be biosynthesized by this putative rectly or indirectly in causing PD. enzyme. Numerous publications have proposed the exis- Sal is a kind of catechol isoquinoline that can be tence of (R)-salsolinol synthase, however, to our knowl- found in brain, cerebrospinal fluid, blood, and urine edge, only one paper has reported the primary purifi- (Haber et al. 1996). It can also be detected in some cation of the enzyme from human brains, published by foods and beverages, such as dried banana and port Naoi et al. (1996) from Japan. From then on, salsolinol wine, suggesting that consumption of such daily food synthase has not been extensively purified and charac- items can affect Sal level in tissues (Smythe et al. 1985). terized, and it has also not been reported whether this

* Corresponding author

c 2011 Institute of Molecular Biology, Slovak Academy of Sciences 1184 X. Chen et al.

Fig. 1. Synthetic pathways of Sal from dopamine and acetaldehyde. Pathway A is non-enzymatic Pictet-Spengler reaction; it produces racemic mixture of both (R)- and (S)-Sal. Pathway B is catalyzed by (R)-salsolinol synthase, producing only (R)-Sal. enzyme exists in the other animal brains or not. Hence adjusted to pH 7.4 with 1 M HCl. Following centrifugation ◦ further research in this ﬁeld is necessary to explore the at 100,000×g for 1 h at 4 C, the supernatant was collected mechanism of action and possible role of (R)-salsolinol as crude enzyme. Crude enzyme fraction was incubated in synthase in the pathogenesis of PD. 50 mM Tris-HCl (pH 7.4) with 1 mM DA and 1 mM AcH in a total volume of 1 mL (Naoi et al. 1996; De-Eknamkul In this study we conducted a series of experiments ◦ et al. 1997). After incubation at 37 C for 30 min, 20 µLof to examine the existence of salsolinol synthase in rat 1 M perchloric acid containing 1 mM EDTA and sodium brain. Chiral analysis was done to demonstrate the fact metabisulphite was added to stop the reaction. After mix- that this enzyme has an enantio-selectivity. Moreover, ing and centrifugation at 17,000×g for 15 min, the sample the distribution of this enzyme in various regions of was diluted 25 fold and ﬁltered through a 0.22 µm cellulose brain was thoroughly investigated. membranes before storing at −80 ◦C for further experimen- tations.

Material and methods HPLC-electrochemical detection apparatus and chromato- Chemicals and substrates graphic conditions Dopamine hydrochloride (DA), AcH, Sal, sodium 1-hepta- Enzyme activity was assayed using HPLC-electrochemical nesulphonate and HPLC-grade methanol were purchased detection (ECD) (ESA, Chelmsford, MA, USA) as described from Fisher Scientific Canada (Edmonton, Canada). Tris- previously (Maruyama et al. 1996). Briefly, the oxidation HCl, perchloric acid, ethylene diamine tetraacetic acid voltage of a model 5020 guard cell was 400 mV, 350 mV, (EDTA), sodium metabisulphite, citric acid and pentobar- 100 mV and −50 mV. Samples were separated on a reverse- bital sodium were purchased from Sigma. Disodium EDTA phased Inertosil ODS-C18 column (4.6 (i.d.) × 150 mm), was purchased from Sino-American Biotechnology Com- with a mobile phase (pH 4.0) of citric acid (40 mM), pany. All other chemicals were obtained from the Beijing Na2HPO4 (20 mM), Na2EDTA (0.3 mM), 5% methanol, Chemical reagent Company in China and all of the chemi- and 1.0 mM heptane sulphonic acid at a flow rate of 1.0 cals used were of analytical grade and organic solvents were mL/min. The quantity of Sal was calculated using a stan- of HPLC grade. A Milli-Q water purifying system was uti- dard curve generated by standard substances. The salsolinol lized to generate 18.2 MΩ deionized water. synthase activity was demonstrated by the increment of Sal during per min per mg of protein at 37 ◦C, and is represented −1 −1 Brain dissection as nmol min mg . The protein was quantified using the Sprague-Dawley (SD) rats, weighing 200 ± 20 g, were ob- Bradford method (Bradford 1976). tained from Central Animal House of Chinese Academy of Medical Sciences. Rats were kept under standard laboratory Preliminary purification of salsolinol synthase conditions, at room temperatures and were supplied with The crude enzyme fraction was used for preliminary purifi- food and water. The rats were anesthetized with sodium cation. The sample was first presented on a PA-DEAE-02 pentobarbital (64.8 mg/mL, 50 mg/kg) and decapitated (ac- column (20 (i.d.) × 100 mm) at the rate of 3.0 mL/min. A cording to institutional protocols). The brains were removed non-linear gradient elution of eluents A (25 mM Tris-HCl by standard autopsy procedures and cortex, striantum, sub- buffer solution and 1 M NaCl solution) and B (25 mM Tris- stantia nigra, and cerebellum were quickly dissected on a HCl buffer solution, pH 7.4) was used for the separation. ◦ cold plate at 4 C for subsequent investigations (Musshoff The elution programmer was optimized and conducted as ◦ et al. 1999). All specimens were frozen at −80 Cuntilbio- follows: over the range from 0.01 min to 5.00 min, 0% A chemical analysis. was used, from 5 min to 25 min, 0–60% A was used, 60– 90% A was used in the range of 25 min to 45 min, and then Sample preparation 90–100% A was used from 45min to 48min, finally, sepa- Tissue samples (0.2–0.5 g) were weighed and transferred to ration was done for period from 48 min to 53.01 min with centrifugation tubes. The sample were homogenized with 100% A. Detection was accomplished via UV at 280 nm. In an Ultra Turrax in 1 mL of a solution of 50 mM Tris-HCl, total, 53 fractions were collected after every min. The same Salsolinol synthase from rat brain 1185

Fig. 2. Production of Sal detected by HPLC-ECD. (A) Chromatographic patterns: a: enzyme, b: DA+AcH, c: DA+enzyme, d: AcH+Enzyme, e: DA+AcH+enzyme. (B) Quantity of Sal from enzyme, DA+AcH and DA+AcH+enzyme. The values were obtained from eight rat brains. fractions from each run were combined, lyophilized and de- lower (Fig. 2A, a). Quantitation of the Sal is shown in tected the salsolinol synthase activity by HPLC-ECD. The Figure 2B. It should be noted that DA and Ach can active fraction was then applied on Sephacryl S-100 HR (16 produce a small amount of Sal spontaneously, without × mm (i.d.) 60 cm) which was equilibrated with 25 mM the presence of enzyme, so it should be subtracted when Tris-HCl buffer solution (pH = 7.4) and 0.15 M NaCl at the calculating the enzyme activity. We also attempted to rate of 0.5 mL/min. Detection was accomplished via UV at 280 nm. In total, 39 fractions (2 mL) were collected after use different measures to deactivate the enzyme, such every 4 min, and detected the activity by HPLC-ECD. as heating, treating with strong acid, urea, or organic solvents; the result showed that treatment with perchlo- HPLC analysis of the enantiomers of Sal ric acid was the best method to deactivate the enzyme A β-cyclodextrin-bonded column, nucleodex β-OH (4.0 (data not shown). These results clearly suggested the (i.d.) × 200 mm; Macherey-Nagel, Dtiren, Germany) was existence of enzyme activity in the crude protein frac- used to separate the enantiomers of Sal. For the determi- tions of rat brain that rapidly catalyzed the condensa- nation of dopamine and Sal enantiomers, coupled columns, tion of DA and AcH into Sal. β a -cyclodextrin-bonded column connected with a reverse- The enzyme was purified from crude enzyme frac- phased Inertosil ODS-C18 column in series was employed tion by HPLC using DEAE column and a gel filtration (Naoi et al. 1996). The mobile phase was 100 mM sodium phosphate buffer (pH 3.6) containing 5% methanol at a flow column. Crude fraction was separated first by DEAE rate of 0.4 mL/min. The voltammograms of authentic stan- column, and the chromatogram was recorded at a wave- dard of (R)-Sal and the corresponding peak in the sample length of 280 nm. The result showed that the activity were measured with an HPLC-ECD system. Quantitation peak was detected between 17 and 21 min. Then, 17 was performed by comparison of the peak area of the sam- and 21 min fractions were collected and separated by ple with that of the standard of Sal. gel filtration chromatography, and salsolinol synthase was eluted from 137 to 153 min. The specific activity of Results salsolinol synthase after purification was compared with the activity of crude fraction. Figure 3 documents sig- Crude enzyme fractions prepared from the rat brain nificant increase of the specific activity of salsolinol syn- were incubated with DA and AcH at pH 7.4, the re- thase after purification, when it was purified 8.8- and action products appeared in the incubation mixture as 115.2-fold, respectively, suggesting the existence of this detected by HPLC-ECD. As shown in Figure 2, the in- enzyme (details of the preliminary purification of this cubation of the two substrates with rat brain extracts enzyme will be published elsewhere). Further researches produced a much large amount of Sal (Fig. 2A, e), about purification and characterization are under inves- whereas no much Sal was detected in the condensa- tigations. tion of substrates by the non-enzymatic Pictet-Spengler The separation of DA and Sal enantiomers was reaction (Fig. 2A, b) or in the crude extract contain- found to be dependent on the concentration and pH ing no substrates (Fig. 2A, a). Furthermore, a series value of the buffer, as well as on the concentration and of controls were used to rule out the possibility of in- nature of the organic solvent. An increase in the po- volvement of other factors in the brain extracts that larity of the organic solvent (e.g. methanol) enhanced might affect the process. First, we analyzed two groups the separation. Na2HPO4-NaH2PO4 was most effec- of single-substrate reactions, one containing DA only, tive buffer for the separation of the enantiomers of Sal. and the other one only containing AcH. No increase of The optimal mobile phase selected was 100 mM sodium Sal was detected in either groups compared to positive phosphate buffer (pH 3.6) containing 5% methanol at reaction (Fig. 2A, c,d), and the endogenous Sal was a flow rate of 0.4 mL/min. The value of the separa- 1186 X. Chen et al.

the distribution of Sal and its metabolites, which can be detected in many areas of the brain. It was reported that the highest concentration of Sal has been detected in the striatum and substantia nigra (Musshoff et al. 1999). Our result demonstrated that striatum and substantia nigra contained higher activity of salsolinol synthase than other regions. These results suggested that the distribution of Sal does not depend on its substrate, but on the distribution of its synthesizing enzymes. Sal- solinol synthase activity was also detected in SH-SY5Y cells (human neuroblastoma) as 2.7 nmol min−1 mg−1 which is higher than the activity in rat brain, suggesting the activity of salsolinol synthase in human brain Fig. 3. The specific activity of salsolinol synthase purified by dif- may be higher than in rat brain. ferent methods compared with crude fraction. One was purified by DEAE column chromatography and 53 fractions were collected after every min, the other was purified by gel filtration column Discussion chromatography and 39 fractions were collected after every 4 min. The enzyme activity was detected by HPLC-ECD according to We report evidences for the presence of a putative sal- the increment of Sal during per min per mg protein in the con- dition of 37 ◦C, represented as nmol min−1 mg−1. solinol synthase in rat brain, which catalyzes the condensation of DA and AcH to (R)-Sal. The activity of salsolinol synthase was detected by HPLC-ECD. Prepa- tion factor (α) corresponding to the enantiomers of Sal ration of the crude enzyme fractions with active enzyme was 1.30, and it also shows that the racemic mixture activity could be achieved by using 50 mM Tris-HCl, of Sal was well separated from each other and from its pH 7.4. The crude extracts, after being preliminary pu- catecholamine precursor dopamine under the selected rified by DEAE and gel filtration column chromatog- conditions with coupled columns (Fig. 4A). Figure 4B raphy, contained higher activity. The formation of Sal shows that the ratio of (R)-Sal and (S)-Sal produced was clearly the result of an enzymatic reaction since no by non-enzymatic Pictet-Spengler reaction was about 1. high concentration of Sal was observed in the similar However, in enzymatic reaction, (R)-Sal exhibits an ap- incubation mixture containing deactivated enzyme ex- proximately 1.49-fold increase compared to the produc- tracts, enzyme extracts containing single-substrate, or tion of (S)-Sal. These results indicated that the surplus non-enzymatic Pictet-Spengler reaction. (R)-Sal was biosynthesized by (R)-salsolinol synthase, It is proposed that there are two synthetic path- and the two pathways of Sal production co-exist in rat ways for Sal. Our results suggest that both pathways of brain. Sal synthesis co-exist. Non-enzymatic Pictet-Spengler Salsolinol synthase activity could be detected in all reaction produces both (R)- and (S)-Sal, and an addi- of these regions (Fig. 5). This result was the same with tional enzymatic activity produces (R)-Sal more than

Fig. 4. Determination of the (R)- and (S)-enantiomers of Sal by using chiral high-performance liquid chromatographic column. They were analyzed with various types of mobile phases changing pH value of the buffer as well as the concentration and nature of the organic solvent. (A) Chromatograms of standard of DA, (R)-Sal and (S)-Sal. It represents the effect of the buffer concentration, pH and methanol concentration on the separation. (B) Quantitation of Sal of samples from the Pictet-Spengler reaction and reaction with enzyme. Production of salsolinol by DA+AcH+enzyme had subtracted the Sal produced by non-enzymatic Pictet-Spengler reaction. The production of (R)-Sal was compared to (S)-Sal (n =4;*p < 0.05). Salsolinol synthase from rat brain 1187

Fig. 5. The speciﬁc activity of salsolinol synthase in four major regions of rat brain, such as cortex, striatum, substantia nigra, and cerebellum. Activity was detected in a similar way like that in Figure 3, and the Sal produced by non-enzymatic Pictet-Spengler reaction was subtracted while calculating the enzyme activity.

(S)-Sal. This result is different from earlier research, the support provided by the professors in the School of Life which showed that only (R)-Sal could be detected in Science and Technology, Beijing Institute of Technology and human brain with PD (Naoi et al. 1996). However, it all colleagues in the laboratory. is similar to the report where the enantiomeric ratio R S of ( )-/( )-Sal was approximately 2 (Zhu et al. 2008). References The disagreements in above mentioned findings may be because salsolinol synthase in our research was ex- Bradford M.M. 1976. A rapid and sensitive method for quantita- tracted from the rat brain and the activity of the en- tion of microgram quantities of protein using the principle of 72: zyme in rat brain is lower than the activity in human protein-dye binding. Anal. Biochem. 248–254. R De-Eknamkul W., Ounaroon A., Tanahashi T., Kutchan T.M. brain with PD. However, we confirmed that ( )-Sal was & Zenk H.M. 1997. Enzymatic condensation of dopamine much more than (S)-Sal after catalyzed by crude pro- and secologanin by cell-free extracts of Alangium Lamarckôň. tein fraction, suggesting the existence of (R)-salsolinol Phytochemistry 45: 477–484. synthase. This is the first time that the activity of (R)- Deng Y., Maruyama W., Dostert P., Takahashi T., Kawai M. & Naoi M. 1995. Determination of the (R)- and (S)-enantiomers salsolinol synthase in rat brain is detected. of salsolinol and N-methylsalsolinol by use of a chiral high- We also detected the activities of salsolinol syn- performance liquid chromatographic column. J. Chromatogr. thase in several brain regions. The highest level of activ- B Biomed. Appl. 670: 47–54. ity was found in the substantia nigra which is in agree- Deng Y., Maruyama W., Kawai M., Dostert P., Yamamura H., Takahashi T. & Naoi M. 1997. Assay for the (R)- and (S)- ment with previous findings describing the distribution enantiomers of salsolinols in biological samples and foods with of Sal in human brain (Maruyama et al. 1997). There ion-pair high-performance liquid chromatography using 13– was no significant correlation between the activity of cyclodextrin as a chiral mobile phase additive. J. Chromatogr. salsolinol synthase and the levels of DA, which is only B Biomed. Sci. Appl. 689: 313–320. Haber H., Winkler A., Putscher I., Henklein P., Baeger I., Georgi present in striatum and substantia nigra. It might be M. & Melzig M.F. 1996. Plasma and urine salsolinol in hu- due to the fact that DA exists selectively in the ni- mans: effect of acute ethanol intake on the enantiomeric com- grostriatal dopaminergic neurons, while (R)-Sal is dis- position of salsolinol. Alcohol Clin. Exp. Res. 20: 87–92. tributed uniformly among brain regions. These results Luk L.Y., Bunn S., Liscombe D.K., Facchini P.J. & Tanner R M.E. 2007. Mechanistic studies on norcoclaurine synthase suggest that the concentration of ( )-Sal does not de- of benzylisoquinoline alkaloid biosynthesis. Biochemistry 46: pend on its precursor, but on the activity of the syn- 10153–10161. thesizing enzymes. Maresh J.J., Giddings L.A., Friedrich A., Loris E.A., Panjikar Further work on salsolinol synthase will include at- S., Trout B.L., Stockiqt J., Peters B. & O’Connor S.E. 2008. Strictosidine synthase: mechanism of a pictet-Spengler cat- tempts to isolate and characterize the enzyme from rat alyzing enzyme. J. Am. Chem. Soc. 130: 710–723. brains or cell cultures, and to determine its amino acid Maruyama W., Abe T., Tohgi H., Dostert P. & Naoi M. 1996. sequence. It will be of great interest to compare the A dopaminergic neurotoxin, (R)-N-methylsalsolinol, increases sequence of salsolinol synthase with other two Pictet- in parkinsonian cerebrospinal fluid. Ann. Neurol. 40: 119– Spengler-ases, such as strictosidine synthase and norco- 122. Maruyama W., Sobue G., Matsubara K., Hashizume Y., Dostert claurine synthase (Luk et al. 2007; Maresh et al. 2008). P. & Naoi M. 1997. A dopaminergic neurotoxin, 1(R), This study is currently under investigation in our lab- 2 (N)-dimethyl-6,7-dihydroxy-1,2,3,4-tetrahydroisoquinoline, oratory. N-methyl(R)salsolinol, and its oxidation product, 1,2(N)- dimethyl-6,7-dihydroxyisoquinoliniumion, accumulate in the nigrostriatal system of the human brain. Neurosci. Lett. 223: Acknowledgements 61–64. Musshoff F., Schmidt P., Dettmeyer R., Priemer F., Jachau K. & Madea B. 2000. Determination of dopamine and dopamine- The authors thank the support from the National Natural derived (R)-/(S)-salsolinol and norsalsolinol in various hu- Science Foundation of China (No. 20435020), and partial man brain areas using solid-phase extraction and gas chro- support from the Ministry of Industry and Information of matography/mass spectrometry. Forensic Sci. Int. 113: 359– China (No. A2220060002). The authors also acknowledge 366. 1188 X. Chen et al.

MusshoffF.,SchmidtP.,DettmeyerR.,PriemerF.,WittigH.& Smythe G.A. & Duncan M.W. 1985. Precise GC/MS assays for Madea B. 1999. A systematic regional study of dopamine and salsolinol and tetrahydropapaveroline: the question of arti- dopamine-derived salsolinol and norsalsolinol levels in human facts and dietary sources and the influence of alcohol. Prog. brain areas. Forensic Sci. Int. 105: 1–11. Clin. Biol. Res. 183: 77–78. Naoi M., Maruyama W., Dostert P., Kohda K. & Kaiya Song Y., Xu J., Hamme A. & Liu Y.M. 2006. Capillary liquid T. 1996. A novel enzyme enantio-selectively synthesizes chromatography-tandem mass spectrometry of tetrahydroiso- (R)salsolinol, a precursor of a dopaminergic neurotoxin, N- quinoline derived neurotoxins: a study on the blood-brain methyl(R)salsolinol. Neurosci. Lett. 212: 183–186. barrier of rat brain. J. Chromatogr. A. 1103: 229–234. Origitato O., Hanningen J. & Collins M.A. 1981. Rat brain sal- Zhu W., Wang D., Zheng J., An Y., Wang Q., Zhang W., Jin L., solinol and blood-brain barrier. Brain Res. 224: 446–451. Gao H. & Lin L. 2008. Effect of (R)-Salsolinol and N-Methyl- Quan Z., Song Y., Peters G., Shenwu M., Sheng Y., Hwang H. (R)-salsolinol on the balance impairment between dopamine & Liu Y.M. 2005. Chiral CE separation of dopamine-derived and acetylcholine in rat brain: involvement in pathogenesis of neurotoxins. Anal. Sci. 21: 115–119. Parkinson disease. Clin. Chem. 54: 705–712. Robbins J.H. 1968. Alkaloid formation by condensation of bio- genic amines with acetaldehyde. Clin. Res. 16: 350. Received March 15, 2011 Rojkovicova T., Mechref Y., Starkey J.A., Wu G., Bell R.L., Accepted August 3, 2011 McBride W.J. & Novotny M.V. 2008. Quantitative chiral analysis of salsolinol in different brain regions of rats genet- ically predisposed to alcoholism. J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 863: 206–214.