Development of an on-line SPR-digestion-nanoLC-MS/MS system for the quantification and identification of interferon-gamma in plasma.

An automated, on-line system for protein quantification and identification, employing Surface Plasmon Resonance (SPR), enzymatic protein digestion, nanoLC and tandem-MS (MS/MS), has been developed. For the experiments recombinant human interferon-gamma (rhIFN-gamma) in buffer or diluted bovine plasma was used as a model protein. Upon injecting 90muL of a 1mugmL(-1) solution of rhIFN-gamma in diluted plasma at a flow rate of 10muLmin(-1), 320fmol of protein was reproducibly bound to the sensor surface. After desorption of the isolated protein from the SPR surface using 10mM glycine pH 1.3, on-line digestion, nanoLC and MS/MS analysis, rhIFN-gamma could be identified on basis of peptide masses and MS/MS fragmentation data. A sequence recovery of 66% was found when a pepsin micro reactor was used. For a trypsin micro reactor the sequence recovery was 50%. In the latter case, the desorbed protein solution was pH-tuned with a TRIS buffer for optimal enzyme activity. With the identified trypsin- and pepsin-produced peptides and because parts of their amino acid sequences overlap, the protein sequence can be largely elucidated showing the potential for the analysis of unknown proteins. The SPR-digestion-nanoLC-MS/MS platform provides unattended analysis of a sample within 60min.

Recent developments in capillary isoelectric focusing.

The developments in capillary isoelectric focusing (cIEF) over the period 2003-2007 are reviewed. With the focus on technological aspects, cIEF papers published in the fields of methodology, new techniques, detection, multidimensional systems, miniaturization and applications are summarized. The methodology section covers recent research in ampholytes composition, detergents and other additives, carrier ampholyte free cIEF, coatings and other capillary modifications. In the section on new systems adjustments to the technique (e.g. dynamic IEF), different applications of cIEF (e.g. as injection system) and new devices are reported. Systems focusing on whole column imaging, fluorescence and chemiluminescence detection and coupling to mass spectrometers are discussed in the section on detection. Interfacing cIEF with MS via RPLC systems and hyphenation of cIEF with capillary electrochromatography and other capillary electrophoresis modes are also summarized. Papers focusing on miniaturization are reviewed in the section on microfluidic devices. The section on applications will show analysis of biopharmaceutical compounds and isolated proteins for metabolomic studies. For the analysis of complex biological matrices, generally multidimensional systems are needed, which are mentioned throughout this review.

Pepsin immobilized in dextran-modified fused-silica capillaries for on-line protein digestion and peptide mapping.

On-line digestion of proteins under acidic conditions was studied using micro-reactors consisting of dextran-modified fused-silica capillaries with covalently immobilized pepsin. The proteins used in this study differed in molecular weight, isoelectric point and sample composition. The injected protein samples were completely digested in 3 min and the digest was analyzed with micro-high performance liquid chromatography (HPLC) and tandem mass spectrometry (MS/MS). The different proteins present in the samples could be identified with a Mascot database search on the basis of auto-MS/MS data. It proved also to be possible to digest and analyze protein mixtures with a sequence coverage of 55% and 97% for the haemoglobin beta- and alpha-chain, respectively, and 35-55% for the various casein variants. Protease auto-digestion, sample carry-over and loss of signal due to adsorption of the injected proteins were not observed. The backpressure of the reactor is low which makes coupling to systems such as Surface Plasmon Resonance biosensors, which do not tolerate too high pressure, possible. The reactor was stable for at least 40 days when used continuously.

Development of an open-tubular trypsin reactor for on-line digestion of proteins.

A study was initiated to construct a micro-reactor for protein digestion based on trypsin-coated fused-silica capillaries. Initially, surface plasmon resonance was used both for optimization of the surface chemistry applied in the preparation and for monitoring the amount of enzyme that was immobilized. The highest amount of trypsin was immobilized on dextran-coated SPR surfaces which allowed the covalent coupling of 11 ng mm(-2) trypsin. Fused-silica capillaries were modified in a similar manner and the resulting open-tubular trypsin-reactors having a pH optimum of pH 8.5, display a high activity when operated at 37 degrees C and are stable for at least two weeks when used continuously. Trypsin auto-digestion fragments, sample carry-over, and loss of signal due to adsorption of the protein were not observed. On-line digestion without prior protein denaturation, followed by micro-LC separation and photodiode array detection, was tested with horse-heart cytochrome C and horse skeletal-muscle myoglobin. The complete digestion of 20 pmol microL(-1) horse cytochrome C was observed when the average residence time of the protein sample in a 140 cm x 50 microm capillary immobilized enzyme reactor (IMER) was 165 s. Mass spectrometric identification of the injected protein on the basis of the tryptic peptides proved possible. Protein digestion was favorable with respect to reaction time and fragments formed when compared with other on-line and off-line procedures. These results and the easy preparation of this micro-reactor provide possibilities for miniaturized enzyme-reactors for on-line peptide mapping and inhibitor screening.

Selective protein removal and desalting using microchip CE.

This paper describes the on-line sample pretreatment and analysis of proteins and peptides with a poly(methylmethacrylate) (PMMA) microfluidic device (IonChip). This chip consists of two hyphenated electrophoresis channels with integrated conductivity detectors. The first channel can be used for sample preconcentration and sample clean-up, while in the second channel the selected compounds are separated. Isotachophoresis (ITP) combined with zone electrophoresis (CZE) was used to preconcentrate a myoglobin sample by a factor of about 65 before injection into the second dimension and to desalt a mixture of six proteins with 100 mM NaCl. However, ITP-CZE could not be used for the removal of two proteins from a protein/peptide sample since the protein zone in the ITP step was too small to remove certain compounds. Therefore, we used CZE-CZE for the removal of proteins from a protein/peptide mixture, thereby injecting only the peptides into the second CZE separation channel.

An improved coating for the isolation and quantitation of interferon-gamma in spiked plasma using surface plasmon resonance (SPR).

A study was initiated to investigate the use of surface plasmon resonance (SPR) for the detection in plasma of a high pI model protein, recombinant human interferon-gamma (IFN-gamma). Initially a number of self-assembled monolayers (SAMs) and hydrogel-derivatised SAM-coatings were characterised for the adsorptive and desorptive properties of plasma components. Next a monoclonal anti-IFN-gamma antibody, MD-2, was covalently attached to dextran-modified mercaptoundecanoic acid surfaces that performed best. On coatings consisting of carboxyl-modified dextran (CMD) a difference in interaction behaviour was observed when IFN-gamma was injected in either buffer or diluted plasma. During the injection of IFN-gamma in buffer, an acceleration of the interaction process was observed and the signal continued to increase after the injection plug had passed. Upon injection of diluted plasma spiked with IFN-gamma, the response increased without acceleration of the binding process. After the injection was finished, some of the bound material desorbed as expected, resulting in a signal decrease. On non-charged dextrans, the interaction between the antibody-modified surface and IFN-gamma in either plasma or buffer was similar. During sample injection the response increased with a binding rate depending on the concentration of IFN-gamma present in solution. When the injection was finished, some of the bound material was washed away from the surface and only a minor contribution of non-specific adsorbed plasma components was noticeable. From the coatings tested, the non-modified dextran-coated SPR sensor disks prove to be best suited for the detection of IFN-gamma in complex matrices like plasma. The interaction of IFN-gamma in both diluted plasma and buffer is comparable and concentrations of IFN-gamma of 250 ng ml-1 and higher can be detected in both buffer and 100x-diluted plasma. The non-specific adsorption of plasma components is low, whereas the specific IFN-gamma response is relatively high.

On the response of a label-free interferon-gamma immunosensor utilizing electrochemical impedance spectroscopy.

The research on our flow-injection, label-free, non-faradaic impedimetric immunosensor for interferon-gamma (IFN-gamma) has been extended. The sensor is prepared by immobilization of anti-IFN-gamma antibodies on a self-assembled monolayer (SAM) of acetylcysteine, deposited on polycrystalline gold. A multi-frequency impedance method is described, which allows time-resolved measurement of Nyquist plots. To these plots, an equivalent circuit was fitted, which is discussed in terms of a two-layer structure (inner and outer layer) of the interfacial region. Because binding of IFN-gamma mainly causes a decrease of Q (a constant-phase element), this element is considered as the outer layer. Several aspects of the impedimetric sensor response are studied, including the dependence on detection frequency, target concentration and applied dc potential. For quantitative detection of IFN-gamma, an optimum of the signal-to-noise (S/N) ratio of the out-of-phase impedance component (Z'') was found at about 100 Hz. At a dc-potential of +0.2 V versus a saturated calomel reference electrode, the sensor response is higher than at 0.0 V. Logarithmic dose-response curves of IFN-gamma in the concentration range of 10(-18) to 10(-9) M were obtained using two procedures: by successive injections over a single electrode, and by using freshly prepared electrodes for each measurement. Using the latter method, the repeatability is impaired. The need for in situ complementary techniques for a correct interpretation of the studied parameters is discussed.

Development of an electrochemical immunosensor for direct detection of interferon-gamma at the attomolar level.

An electrochemical immunosensor for direct detection of the 15.5-kDa protein interferon-gamma (IFN-gamma) at attomolar level has been developed. Self-assembled monolayers (SAMs) of cysteine or acetylcysteine are formed on electropolished polycrystalline Au electrodes. IFN-gamma adsorbs physically to each of these SAMs. With injections of 100 mM KCl, IFN-gamma can be removed in the flow without damaging the acetylcysteine SAM. However, the cysteine SAM is affected by these KCl injections. In an on-line procedure in the flow, a specific antibody (MD-2) against IFN-gamma is covalently attached following carbodiimide/succinimide activation of the SAM. The activation of the carboxylic groups, attachment of MD-2, and deactivation of the remaining succinimide groups with ethanolamine are monitored impedimetrically at a frequency of 113 Hz, a potential of +0.2 V versus SCE, and an ac modulation amplitude of 10 mV. Plots of the real (Z') and imaginary (Z") component of the impedance versus time provide the information to control these processes. In the thermostated setup (23.0 degrees C), samples of unlabeled IFN-gamma (in phosphate buffer pH 7.4) are injected and the binding with immobilized MD-2 is monitored with ac impedance or potential-step methods. While the chronoamperometric results are rather poor, the ac impedance approach provides unsurpassed detection limits, as low as 0.02 fg mL-1 (approximately 1 aM) IFN-gamma. From a calibration curve (i.e. Z" versus the amount injected), recorded by multiple 50-microL injections of 2 pg mL(-1) of IFN-gamma, a dynamic range of 0-12 pg mL(-1) could be derived. However, when nonspecific adsorption is taken into account, which has been found to be largely reduced through injections of 100 mM KCl, a much smaller dynamic range of 0-0.14 fg mL(-1) remains. The immunosensor can be regenerated by using a sequence of potential pulses in the flow by which the SAM with attached MD-2 and bound IFN-gamma is completely removed. When the developed procedures described above are repeated, the response of the immunosensor is reproducible within 10%.

Reduction of Cys36-Cys42 and Cys64-Cys74 disulfide bonds in recombinant human granulocyte colony stimulating factor.

The Cys36-Cys42 and Cys64-Cys74 disulfide bonds in recombinant methionyl human granulocyte colony-stimulating factor were reduced to sulfhydryls with dithiothreitol or mercury. Both reduction reactions are dependent on the pH. The reduction reaction with dithiothreitol increased in rate with increasing pH; between pH 7-9 and above pH 10.5 this increase was less than in other regions. These observations are explained by repulsive forces between dithiothreitol and regions in granulocyte colony-stimulating factor which intensify in these pH-regions. The hydroxyl catalysis causes the overall increase in k(obs) in the pH-region studied. The reduction of the disulfides with mercury is, as could be expected from the Nernst equation for disulfide reduction, also pH dependent: the half-wave potential decreases with increasing pH as predicted by theory.

Liposome-mediated enhancement of the sensitivity in immunoassays of proteins and peptides in surface plasmon resonance spectrometry.

A recently developed liposome sandwich immunoassay for interferon-gamma (IFN-gamma), to be applied in microtiter plates, is tailored for surface plasmon resonance (SPR) spectrometry. The assay is performed on a thin (approximately 20 nm) polystyrene layer that covers a gold surface. This way, analytical data obtained from microtiter plate technology can directly be extrapolated toward SPR. For assaying the antigen IFN-gamma, a 16-kDa cytokine, a capture monoclonal antibody is physically adsorbed onto the polystyrene surface. After addition of the sample containing IFN-gamma, a biotinylated detecting antibody is added. Avidin is used as a bridging molecule between the biotinylated antibody and the biotinylated liposomes. All solutions are prepared with PBS buffer (10 mM, pH 7.4). This avoids additional changes in index of refraction caused by the use of various buffer solutions in immunoassays on microtiter plates for coating, binding, and washing procedures. It is shown that, when liposomes are used, a substantial enhancement of the detection limit is achieved. The "liposome" strategy improves the sensitivity for the IFN-gamma assay approximately 4 x 10(4) times and the detection limit to low picomolar. The method is generally applicable to other sandwich immunoassays.

Liposomes and immunoassays.

Various aspects of the application of liposomes as a label in immunoassays are reviewed. Methods for the preparation of liposomes, from the basic film method to the more advanced dehydration-rehydration method, are discussed. Furthermore, the markers used in liposome labels, as well as the methods to conjugate liposomes to antigens or antibodies, are summarized. Liposome immunoassays are applied as homogeneous or heterogeneous assays. Homogeneous assays often rely on the lytic activity of complement on antibody-associated liposomes. Another group of homogeneous assays utilizes the inhibitory action of antibodies on the activity of conjugates of mellitin (a bee venom protein) with a hapten. Free mellitin conjugates are able to lyse liposomes effectively. Heterogeneous liposome immunoassays, performed either competitively or non-competitively, resemble more closely standard enzyme linked immunosorbent assays, with the enzyme being replaced by a liposome label. Washing steps are used to separate antigen-specifically bound liposomes from unbound liposomes. All bound liposomes are lysed with a detergent, giving an instantaneous amplification. Flow-injection liposome immunoassays and liposome immunosensors are also described as examples of other possible immunoassay formats.

Application of xanthine oxidase-catalyzed luminol chemiluminescence in a mouse interleukin-5 immunoassay.

A chemiluminescent substrate reagent for use in a sandwich immunoassay for the model antigen mouse interleukin-5 (IL-5) was developed using xanthine oxidase and luminol. Various parameters involved in this chemiluminescent reaction have been studied, including the substrate hypoxanthine, luminol and the Fe(II)-EDTA complex. Addition of the Fe(II)-EDTA complex enhances the chemiluminescence signal considerably. The xanthine oxidase-catalyzed chemiluminescent immunoassay was compared to horseradish peroxidase-linked immunoassays with luminol as chemiluminescent, and tetramethyl benzidine as colorimetric substrate. The detection limit of the xanthine oxidase-luminol assay was found to be about 0.6 pg/ml IL-5, whereas the peroxidase-catalyzed immunoassays have detection limits of about 1.3 (HRP-TMB) and 2.9 pg/ml (HRP-luminol) IL-5.

Structural aspects of antioxidant activity of flavonoids.

Flavonoids, a group of naturally occurring antioxidants and iron chelators, might be used as cardioprotective agents in doxorubicin-induced cardiotoxicity, which is believed to be caused by the formation of oxygen free radicals. To investigate the underlying molecular mechanism, we tested a large group of flavonoids from all major structural subclasses on their ability to inhibit doxorubicin (enzymatically)-induced and Fe2+/ascorbate (nonenzymatically)-induced microsomal lipid peroxidation (LPO) and to chelate Fe2+. In addition, we measured half peak oxidation potentials (Ep/2). LPO inhibition data gave a good qualitative correlation with the oxidation potentials. Most flavonoids tested chelated Fe2+, but there were large differences in the chelating capacity. For good scavenging activity, a catechol moiety on ring B is required. The 3-OH moiety can function as a chelation site and can also be oxidized. The 3-OH group in combination with a C2 C3 double bond, increases the scavenging activity. Fe2+ chelation only plays a role in the LPO inhibition by less active scavengers. Chelation can then raise the activity to the level of the most active scavengers, possibly by site-specific scavenging. It can be concluded that Ep/2 values and iron chelating activity can almost completely describe the LPO inhibiting behaviour of the flavonoids.

3,5-Disubstituted analogues of paracetamol. Synthesis, analgesic activity and cytotoxicity.

Seven 3,5-disubstituted analogues of paracetamol were synthesised in order to compare their physicochemical, pharmacological and toxicological properties with those of paracetamol (4'-hydroxyacetanilide, acetaminophen). Oxidation of the phenolic structure is likely involved in the analgesic action of paracetamol as well as in its toxification by cytochrome P450. The effect of disubstitution adjacent to the phenolic hydroxyl group was studied in order to establish possible structure-activity relationships. 3,5-Substituents with electron-donating capacities (R = -CH3, -OCH3, -SCH3) decreased the half-wave oxidation potential substantially by 0.07 V to 0.16 V, whereas electron-withdrawing substituents (R = -F, -Cl, -Br, or -I) increased the oxidation potential by 0.04 V to 0.06 V when compared to paracetamol. Electron-donating substituents (R = -CH3, -OCH3, -SCH3) increased the mouse brain cyclooxygenase inhibiting capacity of paracetamol. Electron-withdrawing halogen substituents (R = -F, -Cl, -Br or -I) decreased this inhibiting capacity. In agreement with this, the in vivo analgesic activity of the 3,5-dihalogenated analogues was lower when compared to paracetamol. Electron-donating substituents (R = -CH3, -OCH3, -SCH3) decreased the cytotoxicity of paracetamol, when measured as leakage of lactate dehydrogenase from freshly isolated rat hepatocytes, almost completely. Most 3,5-dihalogen substituents (R = -F, -Cl or -Br) diminished it slightly. The fourth electron-withdrawing substituent (R = -I) strongly lowered the cytotoxicity of paracetamol in this test system. In conclusion, a higher cycylooxygenase inhibitory potency of 3,5-disubstituted analogues of paracetamol seemed to correlate with a lower cytotoxicity. 3,5-Disubstituted analogues with electron-donating substituents might therefore be safer analgesics than paracetamol itself. The opposite probably applies to analogues of paracetamol with electron-withdrawing substituents at the 3- and 5- positions of the aromatic nucleus.

Chemiluminescence and immunoassays.

The principles of chemiluminescence and the application of chemiluminescent labels and substrates in immunoassays are reviewed. Various immunoassay formats and all known classes of chemiluminescent molecules, including 1,2-dioxetanes, luminol and derivatives, acridinium esters, oxalate esters and firefly luciferins are described as well as the many sensitizers and fluorescent enhancers. Recent promising developments are discussed.

Mercuric chloride down-regulates T cell interferon-gamma production in brown Norway but not in Lewis rats; role of glutathione.

Injection of a low dose of mercuric chloride into Brown Norway (BN) rats caused a marked decrease in the concanavalin A (ConA)-induced generation of interferon-gamma-producing cells (IFN-gamma pc) in spleen cell cultures prepared 1 h after mercury administration. A second injection 48 h later caused a further diminution of IFN-gamma pc down to 30% of the number generated in splenocyte cultures of phosphate-buffered saline (PBS)-injected controls. Injection of Lewis rats with either one or two doses of HgCl2 revealed no inhibitory effect on splenic IFN-gamma production. The presence of the reduced form of glutathione (GSH) in the culture medium was found to be essential in these experiments. In the absence of GSH there was an overall 20-fold reduction of the number of IFN-gamma pc in splenocyte cultures of normal or PBS-injected rats, which was further reduced to a 60- to 70-fold-lower level in cultures of rats exposed to HgCl2. This mercury-mediated extra reduction could be fully reversed with an excess (2 mM) of GSH in Lewis but not in BN splenocyte cultures. Since the bivalent Hg2+ ion is known to bind to and inactivate sulfhydryl groups of proteins and low molecular weight thiols, most notably GSH, we investigated a possible role for thiols in IFN-gamma production. It was found that the generation of IFN-gamma pc in normal BN and Lewis splenocyte cultures was strongly dependent on GSH or its precursor cysteine in the culture medium. Other thiol compounds were also effective but disulfides were completely inactive. Depletion of intracellular GSH in ConA-stimulated splenocytes by buthionine sulfoximide (BSO), an inhibitor of de novo GSH biosynthesis, strongly inhibited the generation of IFN-gamma pc. The inhibitory effect of BSO was not abolished by the addition of interleukin-2 (IL-2), but was mimicked with antibodies directed to the IL-2 receptor. The data stress the importance of GSH in the enhancement of IL-2-mediated IFN-gamma production and are most consistent with a model in which mercury interferes with T cell IFN-gamma production by affecting the intracellular availability of GSH. The strain-specific susceptibility to mercury-mediated inhibition of IFN-gamma production is discussed.

On-line phase-transfer catalysed dansylation of phenolic compounds followed by normal-phase liquid chromatography with fluorescence detection.

A study of the on-line phase-transfer catalysed dansylation of phenolic compounds is presented. The extraction-dansylation is performed in the extraction coil of a home-made flow-injection extraction unit. After phase separation, the organic phase is fed to a normal-phase liquid chromatographic system with fluorescence detection. Ethynyloestradiol, oestradiol and paracetamol are used as test compounds. The influence of temperature on the reaction is examined. Calibration graphs showed good linearity (r greater than 0.996) and limits of detection are satisfactory (8 x 10(-7) M for ethynyloestradiol, 2 x 10(-6) M for oestradiol and 5 x 10(-7) M for paracetamol). The method is not applicable for the assay of oxychinoline, phenylephrine and morphine.

Rapid determination of benzalkonium chloride in pharmaceutical preparations with flow injection liquid-liquid extraction.

Benzalkonium chloride was assayed by on-line extraction of the benzalkonium ion with picrate to chloroform. The absorbance of picrate was measured. The extractions were performed with a home-made flow injection extraction unit. Calibration curves (1.5-180 x 10(-4)% w/v) were straight lines (r = 0.9993) and the relative standard deviation of a series of injections was less than or equal to 2%. Pharmaceutical benzalkonium preparations, containing xylometazoline, timolol, phenylephrine or carbachol could also be assayed. The method was compared with a modified HPLC assay.

Determination of penicillin in pharmaceutical formulations by flow injection analysis using an optimised immobilised penicillinase reactor and iodometric detection.

An automated assay for the determination of penicillin in formulations suitable for use in pharmaceutical quality control is presented. The method is based on the classical iodometric penicillin assay which is incorporated in a flow injection analysis (FIA) system. The required hydrolysis is performed on-line by using an immobilised penicillinase reactor. Packed-bed and single-bead-string enzyme reactors are compared. It turns out that a packed-bed penicillinase reactor (10 cm x 1.5 mm i.d.) provides complete hydrolysis within short residence time, while only little back-pressure is generated. This enzyme reactor is stable for at least 9 months. Enzymatic hydrolysis of the beta-lactam ring results in the formation of the corresponding penicilloic acid, which consumes iodine. The iodine consumption is determined colorimetrically by measuring the decrease of the absorbance of the blue coloured iodine/starch complex. The optimum reactor length and flow rate for the colourimetrical detection reaction are determined. The optimised method is applied to the assay of penicillin in formulations and the results are compared with the "true" results obtained with a reference method: a mercurimetric titration. The reliability of the flow injection method is evaluated quantitatively by determining the maximum total error (MTE). The reliability is shown to be highest when measuring at a 0.3-mM level. Eight formulations including capsules, tablets and injectables containing penicillin G, amoxicillin or flucloxacillin are assayed. The MTE does not exceed the 6% level and the most probable MTE is between 1.5 and 3.5%.

Automated reversed-phase chromatographic analysis of etoposide and teniposide in plasma by using on-line surfactant-mediated sample clean-up and column-switching.

An automated high-performance liquid chromatographic method for the plasma assay of two neutral drugs, etoposide and teniposide, involving direct plasma injection is presented. The problematic nature of protein precipitation has been circumvented by adding the anionic surfactant sodium dodecyl sulphate to the plasma at a final concentration of 38 mM. Plasma samples are loaded on to a clean-up column with an aqueous mobile phase with which the analyte(s) is (are) retained, whereas the solubilized plasma proteins are flushed to waste. Next, the retained compounds are eluted from the clean-up column on to the analytical column by using the chromatographic mobile phase with a higher elution capacity. The column-switching technique is used to achieve an automated assay. At least 10 ml of plasma, representing 100 repeated injections of 100 microliters or five repeated injections of 2 ml, can pass through the clean-up column without increasing the back-pressure. The recovery increased considerably from 10-30% to 90-95% on adding surfactant to the plasma samples prior to the analysis. The relative standard deviation of the proposed clean-up procedure is 3.5% (n = 6) for both drugs measured at the 2 micrograms/ml level without using an internal standard. The limit of determination with 100-microliters injections is 0.10-0.15 microgram/ml for ultraviolet detection and is seven times lower with electrochemical detection. Teniposide was determined in patients' plasma and the results agreed well with those obtained by the conventional procedure involving manual liquid-liquid extraction prior to chromatographic analysis.