Journal of Chromatography B (v.955-956, #C)

Saponaria officinalis L. (Caryophyllaceae), also known as fuller's herb or soapwort is a medicinal plant, which grows from Europe to Central Asia. Medicinal properties attributed to this plant include its antitussive and galactogogue properties. Recently, bisdesmodic saponins with very specific structural features from S. officinalis have been shown to strongly enhance the efficacy of specific targeted toxins (anti-tumor antibodies connected to protein toxins) in-vitro and in-vivo in a synergistic manner. In the presently reported novel approach we used preparative all-liquid high-speed countercurrent chromatography (HSCCC) to recover a total of 22 fractions using biphasic solvent system tert-butylmethylether/n-butanol/acetonitrile/water 1:3:1:5 (v/v/v/v) from a complex precipitated crude saponin mixture. Out of these 22 fractions, 3 fractions had the enhancer effect on anti-tumor toxins out of which one fraction (F7) was further tested elaborately in different cell lines. The molecular weight distribution and compound profiles of separated saponins were monitored by off-line injections of the sequentially collected fractions to an electrospray ion-trap mass-spectrometry system (ESI-IT-MS). The functional saponin fractions were mainly bisdesmosidc and contained saponin m/z 1861 amongst other. Using the bio-assay guided monitoring, the highly active fractions containing 2 to 3 bisdesmodic saponins (5 μg/mL) were screened for their effectiveness in enhancing the anti-tumor activity of targeted toxin Sap3-EGF, which was determined using the impedance based real-time cell cytotoxicity evaluation. This novel combination of HSCCC fractionation, MS-target-guided profiling procedure and bio-assay guided fractionation yielded 100 mg of functional saponins from a 60 g crude drug powder in a rapid and convenient manner.
Keywords: Synergistic toxicity enhancers; Bisdesmodic saponins; Preparative high-speed countercurrent chromatography; Off-line ESI-MS continuous injections; Saponaria officinalis; Caryophyllaceae;

Simultaneous determination of 12 illicit drugs in whole blood and urine by solid phase extraction and UPLC–MS/MS by Lin Zhang; Zhao-Hong Wang; Hong Li; Yong Liu; Meng Zhao; Ye Jiang; Wen-Song Zhao (10-19).
A rapid and sensitive method based on solid phase extraction and ultra performance liquid chromatography–electrospray ionization tandem mass spectrometry (UPLC–ESI-MS/MS) for the simultaneous determination of amphetamine, methamphetamine, 3,4-methylenedioxyamphetamine, 3,4-methylene-dioxymethamphetamine, N-methyl-1-(3,4-methyl-enedioxyphenyl)-2-butanamine, 3,4-methylenedioxyethylamphetamine, p-methoxymethamphetamine, ephedrine, N-methylephedrine, cathinone, methcathinone, and ketamine in whole blood and urine was developed and validated. Following solid phase extraction, the analytes were separated on ACQUITY UPLC BEH Phenyl column (100 mm × 2.1 mm, 1.7 μm) under gradient elution using a mobile phase containing of acetonitrile and 0.3% formic acid in water at a flow rate of 0.4 mL min−1 and analyzed by a triplequadrupole mass spectrometer in the multiple reaction monitoring (MRM) mode. The proposed method was linear for each analyte with correlation coefficients over 0.99. Recovery validation studies showed accuracy bias below 4.4%. Acceptable precision was also obtained with a relative standard deviation below 8.9%. The sensitivity of the assay was found to be adequate for the quantitation of the illicit drugs in whole blood and urine sample and was higher than reported methods. The present method was proved to be reliable and robust for drug screening in forensic toxicological analysis.
Keywords: Illicit drugs; Solid phase extraction; Urine; Whole blood; UPLC–MS/MS;

New metabolite profiles of Danshensu in rats by ultraperformance liquid chromatography/quadrupole-time-of-flight mass spectrometry by Jun-fei Gu; Liang Feng; Ming-hua Zhang; Dong Qin; Jun Jiang; Xu-dong Cheng; Shu-min Ding; Shi-lin Yang; Xiao-bin Jia (20-25).
In our research, ultraperformance liquid chromatography/quadrupole-time-of-flight mass spectrometry (U-HPLC/Q-TOF-MS) was established for analyzing the metabolite profiles of Danshensu (DSS) in rat feces, bile, urine, plasma and the possible metabolic pathways were subsequently proposed after the oral dose of 80 mg/kg; rat biological samples were collected and pretreated by protein precipitation. Then, the samples were injected into an Acquity ultraperformance liquid chromatography BEHC column with mobile phase consisted of acetonitrile (solvent A)–0.1% formic acid-water (solvent B) with a linear gradient elution program. Totally, 17 metabolites of DSS were identified, including 4, 5, 4 and 4 metabolites in the feces, urine, blood, and bile samples respectively. Most of them were to our knowledge reported for the first time. The results indicated that DSS was metabolized via dehydrogenation, deoxygenation, methylation, glucuronidation, and sulfation pathways in vivo. Among these, methylation was considered as the main physiologic processes of it. This study revealed that U-HPLC/Q-TOF-MS was more accurate and sensitive to detect and identify the possible metabolites and to better understand the metabolism of DSS in vivo.
Keywords: Danshensu; U-HPLC/Q-TOF-MS; Metabolic pathways; Metabolites;

Isomerization of aspartic acid residues is one of the major causes of chemical degradation during the shelf life of biological pharmaceuticals. Monoclonal antibody biopharmaceuticals are typically stored at mildly acidic pH conditions, which can lead to the isomerization reaction. The mechanism of this non-enzymatic chemical reaction has been studied in great detail. However, the identification and quantification of the isomerization sites in a given protein still remains a challenge. We developed an ion-pair reversed-phase HPLC method for the separation of an intact monoclonal antibody variant containing a single isoaspartic acid residue from its native counterpart. We identified and characterized the isomerization site using ion-pair reversed-phase HPLC mass spectrometry methods of the reduced and alkylated antibody and the enzymatically cleaved antibody. Lys-C followed by Asp-N digestion of the antibody was used for the identification of the isomerization site. Electron transfer dissociation (ETD) mass spectrometry was used to confirm the isomerization site at a DY motif at an aspartic acid residue in the CDR-H3 region of the antibody. Tyrosine at the C-terminus of an aspartic acid residue is typically not regarded as a hot spot for isomerization. Our findings suggest that it is not possible to predict isomerization sites in proteins with confidence and all aspartic acid residues located in the CDR regions of antibodies must be considered as potential isomerization site due to the solvent exposure or the flexibility of these regions of the molecule. Additionally, the effect of the pH on the isomerization rate was evaluated using the ion-pair reversed-phase HPLC method, showing that at a lower pH the isomerization rate is faster. Storage at 25 °C for 6 months resulted in an increase of the amount of isoaspartic acid to 6.6% at pH 5.4, 6.0% at pH 5.8, and 5.6% at pH 6.2.
Keywords: Isomerization; Monoclonal antibody; Mass spectrometry; Aspartic acid; Reversed-phase high performance liquid chromatography; Critical quality attributes; Quality by design; Developability;

Considering the importance of derivatization in LC/ESI/MS analysis, the objective of this work was to develop a method for evaluation of matrix effect that would discriminate between matrix effect due to the derivatization reaction yield and from the ESI. Four derivatization reagents (TAHS, DEEMM, DNS, FMOC-Cl) were studied with respect to matrix effects using two selenoamino acids and onion matrix as model system. A novel method for assessing matrix effects of LC/ESI/MS analyses involving derivatization is proposed, named herein post-derivatization spiking, that allows evaluating effect of matrix on ESI ionization without derivatization reaction yield contribution. The proposed post-derivatization spiking method allowed to demonstrate that the reason of reduced analytical signal can be signal suppression in ESI (as in case of DNS derivatives with matrix effects 38–99%), alteration of derivatization reaction yield (TAHS, matrix effects 92–113%, but reaction yields 20–50%) or both (FMOC-Cl, matrix effects 28–88% and reaction yields 50–70%). In case of DEEMM derivatives, matrix reduces reaction yield but enhances ESI/MS signal.A method for matrix effect evaluation was developed. It was also confirmed that matrix effects can be reduced by dilution.
Keywords: Amino acids; Derivatization; Matrix effects; LC/ESI/MS;

Affinity selection of histidine-containing peptides using metal chelate methacrylate monolithic disk for targeted LC–MS/MS approach in high-throughput proteomics by Rajasekar R. Prasanna; Sinash Sidhik; Agamudi S. Kamalanathan; Krishna Bhagavatula; Mookambeswaran A. Vijayalakshmi (42-49).
In recent years, bottom-up approach has become the popular method of choice for large scale analysis of complex proteome samples. Peptide fractionation determines the efficiency of the bottom-up method and often the resolving power of reverse phase liquid chromatography (RPLC) is insufficient for efficient protein identification in case of complex biological samples. To overcome the inherent limitation of proteomics associated with sample complexity, we evaluated fast flow metal chelate methacrylate monolithic system – CIM (Convective Interaction Media) disk chelated with Cu(II) for targeted affinity selection of histidine-containing peptides. Initially the Cu(II)-IMAC using CIM disk was evaluated using tryptic digest of protein mixtures of 8 model proteins and was found to be highly efficient in capturing His-containing peptides with high degree of specificity and selectivity. Further the efficiency of His-peptide enrichment using CIM-IMAC was also demonstrated using complex biological samples like total Escherichia coli cell lysate. The analysis of the Cu(II)-IMAC retained peptides from tryptic digests of model protein mixture and E. coli not only demonstrated a significant reduction in sample complexity but also subsequently enabled the identification of additional peptides. His-peptide enrichment also enabled the identification of low abundant proteins that were not detected in the analysis of total E. coli digest.
Keywords: Immobilized metal affinity chromatography (IMAC); Histidine-peptides; Targeted affinity selection; Enrichment; Bottom-up approach; CIM monolithic disk;

A new, validated, sensitive and cheap method for preliminary quantitative evaluation of acetylcholine esterase inhibitory activity is presented. The proposed method combines HPTLC with data analysis by means of image processing software. An in-situ TLC autobiographic method was employed in which regions of the TLC plate which contain acetylcholinesterase inhibitors show up as white spots against the yellow background. Bleaching of the yellow color, caused by substances with acetylcholinesterase inhibitory activity was observed and recorded using a digital camera. ImageJ, JustTLC and Sorbfil, three image processing programs were evaluated for quantitative measurements. For evaluation of the assay efficiency, acetylcholinesterase inhibitory activity of different Amaryllidaceae plant extracts was expressed as Standard Activity Coefficients (SACs), which are relative measures of the activity to the well known acetylcholinesterase inhibitor eserine. We attempted to validate the method according to the ICH guideline. Different statistical data revealed that all image analysis software are able to detect the acetylcholine esterase inhibitory activity at very low concentration levels with the ImageJ program being the best of all three tested software regarding sensitivity, linearity and precision.
Keywords: HPTLC; Autobiography; Acetylcholinesterase inhibitory activity; Image processing software; Quantitative.;

A high-performance liquid chromatography (HPLC) assay with fluorescence detection (FLD) for quantification of cisatracurium in human plasma was developed and fully validated. Liquid–liquid extraction was employed for sample preparation. The separation was carried out on a C18 column with ternary mobile phase composed of 30 mmol L−1 phosphate buffer (pH 3.0), acetonitrile and methanol (60:35:5, v/v/v). Verapamil was used as the internal standard. The isocratic elution with programmed flow rate was employed by setting at 0.8 mL min−1 from 0 to 3.5 min, 0.5 mL min−1 from 3.5 to 6 min, and 1.0 mL min−1 from 6 to 10 min. The fluorescence detection was performed at 236 nm for excitation and 324 nm for emission. The assay was linear from 50 to 2800 ng mL−1, with a detection limit of 12 ng mL−1. The correlation coefficient (r) for linear regression was 0.9997. The intra-day coefficients of variation (CVs) were less than 2.0%, and the inter-day CVs were less than 4.0%. The mean recoveries were in the range of 92.1–100.4%. The total HPLC run time was less than 10 min. The developed HPLC method was fast, simple, sensitive, accurate and suitable for studying the pharmacokinetics of cisatracurium in infants and children after intravenous administration.
Keywords: Cisatracurium; HPLC-FLD; Programmed flow elution; Plasma pharmacokinetics; Infants and children;

Identification of metabolites of propyrisulfuron in rats by Tomoyuki Takaku; Kazuki Mikata; Hirohisa Nagahori; Yoshihisa Sogame (64-71).
The metabolites found in the urine, feces and bile of male and female rats administered with 14C-labeled herbicide, propyrisulfuron [1-(2-chloro-6-propylimidazo[1,2-b]pyridazin-3-ylsulfonyl)-3- (4,6-dimethoxypyrimidin-2-yl)urea] were identified by high-performance liquid chromatography (HPLC) with the ultraviolet (UV) and radioisotope (RI) detectors, tandem mass spectrometry and nuclear magnetic resonance (NMR). Administered 14C was excreted into the urine (5.7–29.8%) and feces (64.6–97.4%). Urine and bile samples were concentrated and purified using a solid-phase extraction cartridge, and fecal homogenates were extracted using acetonitrile. Conjugates were hydrolyzed with enzyme or hydrochloric acid solution for identification. The proposed major metabolic reactions of propyrisulfuron are as follows: (1) hydroxylation of the pyrimidine ring, propyl group, and imidazopyridazine ring, (2) O-demethylation, (3) cleavage of the pyrimidine ring, and (4) glucuronic acid and sulfate conjugation. The metabolic patterns found are not different among sulfonylurea herbicides.
Keywords: Metabolism; Propyrisulfuron; Rat; LC–MS/MS; NMR;

The enantiomeric separation of a series of racemic functionalized ethano-bridged Tröger base compounds was examined by high performance liquid chromatography (HPLC) and capillary electrophoresis (CE). Using HPLC and CE the entire set of 14 derivatives was separated by chiral stationary phases (CSPs) and chiral additives composed of cyclodextrin (native and derivatized) and cyclofructan (derivatized). Baseline separations (R s  ≥ 1.5) in HPLC were achieved for 13 of the 14 compounds with resolution values as high as 5.0. CE produced 2 baseline separations. The separations on the cyclodextrin CSPs showed optimum results in the reversed phase mode, and the LARIHC™ cyclofructan CSPs separations showed optimum results in the normal phase mode. HPLC separation data of the compounds was analyzed using principal component analysis (PCA). The PCA biplot analysis showed that retention is governed by the size of the R1 substituent in the case of derivatized cyclofructan and cyclodextrin CSPs, and enantiomeric resolution closely correlated with the size of the R2 group in the case of non-derivatized γ-cyclodextrin CSP. It is clearly shown that chromatographic retention is necessary but not sufficient for the enantiomeric separations of these compounds.
Keywords: Cyclofructan-based stationary phase; Cyclodextrin-based stationary phase; Enantiomer; Functionalized ethano-Tröger base; Principal component analysis;

Proteins play a variety of functional roles in tissues that underlie tissue health. The measurement of protein function is important to both understand normal and dysfunctional tissue states. Low-flow push–pull perfusion sampling (LFPS) has been used to collect submicroliter volumes of extracellular fluid which are well suited to capillary electrophoresis for compositional quantitative analysis. In this study, LFPS is used to deliver pharmacological agents to the in vivo retinal tissues at the probe sampling tip during sampling to measure protein function. Two native nitric oxide synthase enzymes were pharmacologically inhibited and the enzyme product NO metabolite, nitrate, was determined with capillary electrophoresis from the perfusates. LFPS delivered inhibitors including the non-selective N(G)-nitro-Larginine methyl ester (L-NAME), the nNOS selective 7-nitroindazole (7-NI), and eNOS N5-(1-imioethyl)-L-ornithine, dihydrochloride (L-NIO) were perfused to the sampling region either directly over a rat retina optic nerve head or 1-mm peripheral to the ONH. At the PONH, 65, 55 and 60% of baseline nitrate levels, respectively, were observed with inhibitor infusion. These are statistically significant (P  < 0.05) compared to saline drug infusion. However, infusion of the inhibitors to the ONH did lead to significant nitrate concentration decreases. This data suggests that the endogenous enzymes, nNOS and eNOS, are both spatially and functionally localized to the PONH at the in vivo rat retina.
Keywords: Low-flow push–pull perfusion; Capillary electrophoresis; Optic nerve head; Nitrate;

Determination of Celecoxib in human plasma using liquid chromatography with high resolution time of flight-mass spectrometry by Nagaraju Dongari; Edward R. Sauter; Brian M. Tande; Alena Kubátová (86-92).
A sensitive method for the determination of Celecoxib (CXB) in human plasma samples was developed using liquid chromatography coupled to electrospray ionization and time of flight mass spectrometry (LC–ESI–TOF–MS). A full factorial design of experiments (FF-DOE) methodology was applied to optimize the ESI conditions for CXB determination and also to predict the effects of interactions of multiple parameters affecting ionization (i.e., capillary voltage, fragmentor voltage, electrolyte and electrolyte concentration). The optimum ionization voltages were 4500 V and 220 V for capillary and fragmentor, respectively. Even though the highest ESI efficiency was obtained without electrolytes, the addition of 1.0 mM ammonium acetate was shown to be essential to buffer the matrix effect and ensure a consistent response. In contrast to previous studies, deuterated CXB was used as a recovery (surrogate) standard, which enabled the correction of CXB loss during sample preparation. The extraction recovery using solid phase extraction was 87–98%. The instrumental limit of detection of CXB (LOD), 0.33 ng/mL, and matrix affected LOD, 0.55 ng/mL, were similar and comparable to the previously reported LC–MS/MS LODs. This method was employed to determine CXB concentrations in human plasma samples. Upon administration of 400 mg CXB to the healthy women, the concentrations found in the plasma were 440–3300 ng/mL. The inter-day repeatability was less than 4% RSD.
Keywords: Celecoxib; HPLC; ESI; TOF; Human plasma; Full factorial design;

Validation and clinical evaluation of a UHPLC method with fluorescence detector for plasma quantification of doxorubicin and doxorubicinol in haematological patients by Jonás Samuel Pérez-Blanco; María del Mar Fernández de Gatta; Jesús María Hernández-Rivas; María José García Sánchez; María Luisa Sayalero Marinero; Francisco González López (93-97).
A rapid and simple UHPLC–fluorescence detection method for the quantification of doxorubicin and its main metabolite, doxorubicinol, in human plasma has been developed. The method was also validated for its application in therapeutic drug monitoring, a clinical approach used in the optimization of oncologic treatments. Following a single protein precipitation step, chromatographic separation was achieved using a C18 column (50 mm × 2.10 mm, particle size 1.7 μm) at 50 °C with a mobile phase consisting of water (containing 0.4% triethylamine and 0.4% orthophosphoric acid)/acetonitrile (77:23, v/v). Flow rate was 0.50 mL/min and fluorescence detection with an excitation wavelength of 470 nm and an emission wavelength of 548 nm was used. The method met the specifications of linearity, selectivity, sensitivity, accuracy, precision and stability of the FDA and EMA guidelines for the validation of bioanalytical methods. Linearity for the drug (8–3000 ng/mL) and the metabolite (3–150 ng/mL) was observed (R 2  > 0.992) and the maximum intra-day and inter-day precision coefficients of variation were less than 14% for both. The lower limits of quantification were 8 and 3 ng/mL for doxorubicin and doxorubicinol, respectively. The method was successfully applied to the quantify plasma concentrations of doxorubicin and doxorubicinol in 33 patients diagnosed with haematological malignancies in which broad ranges for drug (8.3–2766.0 ng/mL) and metabolite (4.8–104.9 ng/mL) levels were measured adequately.
Keywords: Doxorubicin; Doxorubicinol; UHPLC–fluorescence method; Method validation; Drug plasma levels;

Comparison of solid phase- and liquid/liquid-extraction for the purification of hair extract prior to multi-class pesticides analysis by Radu-Corneliu Duca; Guillaume Salquebre; Emilie Hardy; Brice M.R. Appenzeller (98-107).
The present study focuses on the influence of a purification step – after extraction of pesticides from hair and before analysis of the extract – on the sensitivity of analytical methods including compounds from different chemical classes (both parent and metabolites). Sixty-seven pesticides and metabolites from different chemical classes were tested here: organochlorines, organophosphates, carbamates, pyrethroids, ureas, azoles, phenylpyrazoles and neonicotinoids. Two gas chromatography-negative chemical ionization–tandem mass spectrometry methods and one based on ultra-performance liquid chromatography–electrospray tandem mass spectrometry were used. Seven solid-phase extraction cartridges: C18, S-DVB, PS-DVB, GCB, GCB/PSA, SAX/PSA and Florisil/PSA were tested and compared to more classical liquid–liquid extraction procedures using ethyl acetate, hexane and dichloromethane. Although LLE allowed obtaining good results for some compounds, on the whole, SPE clearly provided better recovery for the majority of the pesticide residues tested in the present study. GCB/PSA was clearly the best suited to non-polar compounds such as organochlorines, pyrethroids and organophosphates, with recovery ranging from 45.9% (diflufenican) to 117.1% (parathion methyl). For hydrophilic metabolites (e.g. dialkyl phosphates and other organophosphate metabolites, pyrethroid metabolites, phenols and carbamate metabolites), the best results were obtained with PS-DVB, with recovery ranged from 10.3% (malathion monocarboxylic acid) to 93.1% (para-nitrophenol). For hydrophilic parent pesticides (e.g. neonicotinoids, carbamates, azoles) and metabolites without nucleophilic functions, the best recovery was obtained with SAX/PSA, with recovery ranging from 52.1% (3-hydroxycarbofuran) to 100.9% (3,4-dichloroaniline). Solid phase extraction was found to be more suitable than the liquid–liquid extraction for pesticides and their metabolites determination in terms of number of extracted compounds and their recovery. Moreover, the use of solid phase extraction cartridges has enabled the reduction of the analytical background noise, resulting in better chromatographic separations.
Keywords: Hair; Pesticides; Solid phase extraction; Liquid–liquid extraction; Multi-class analysis;

Drug detection by tandem mass spectrometry on the basis of adduct formation by Marek Dziadosz; Michael Klintschar; Jörg Teske (108-109).
Keywords: Adduct formation; Electrospray ionisation; LC–MS/MS; Ion fragmentation;

fRMSDchiral: A novel algorithm to represent differences between positions of stereoisomers in complex with dissymmetric binding site by Rafal D. Urniaz; Ewelina E. Rutkowska; Anita Plazinska; Krzysztof Jozwiak (110-115).
The ability of molecules to distinguish between optical isomers is crucial for living systems. The change of position of one enantiomer in respect to the position of the second enantiomer within an asymmetric binding site may be analyzed on different levels. Root Mean Square Deviation (RMSD) may be used for such analyses with low precision. Additional fragment level variants of RMSD allow for more precise definition of differences in location of the main molecular features responsible for recognition of stereoisomers by a selector. Three fRMSDchiral parameters appear to be very useful to precisely quantify the change in orientations of stereoisomers. Proposed calculation emerges as interesting assistance in interpretation of consequences of formation differential interaction(s) responsible for a chiral recognition process.
Keywords: Chiral recognition; Molecular modeling; Docking; Molecular dynamics; Fragment RMSD;

In dispersive liquid–liquid microextraction, a few hundred microliters to a few milliliters of water-miscible dispersive solvent are commonly used to assist emulsification in aqueous samples. In the present study, a consistent and automatic up-and-down-shaker-assisted dispersive liquid–liquid microextraction (UDSA-DLLME) that does not require a dispersive solvent was developed. The enrichment factors (EFs) of the targets obtained using the automatic shaker were 361–1391 for UDSA-DLLME, 51–77 for ultrasonication, and 298–922 for vortexing. The linearity of the method was in the range 0.2–200 μg L−1, and its limit of detections was within 0.02–0.04 μg L−1. The intraday and interday relative standard deviations ranged from 5.7 to 10.0% and 5.5 to 10.3%, respectively. The relative recoveries of river and lake samples spiked with 2.0 μg L−1 of triazines were 94.2–102.2% and 98.5–104.1%, respectively. The technique provided high repeatability and recovery. No matrix interference from river and lake water was observed. The method also achieved high EFs compared with those obtained through other emulsification methods such as vortexing and ultrasonication. UDSA-DLLME is an alternative sample preparation technique with good performance.
Keywords: Dispersive liquid–liquid microextraction (DLLME); Sample preparation; Water sample analysis; Gas chromatography–mass spectrometry (GC–MS); Triazine herbicides;

In this study, we have proposed and developed a novel, environmental-friendly and simple method for separation of nine hydrophilic and hydrophobic components in Danshen using microemulsion liquid chromatography. The proposed method was optimized via the preliminary screening experiment and the experimental design. The following factors were investigated in preliminary screening experiment: pH of mobile phase, column type, the nature of surfactant, the nature of oil phase and additives. In order to simultaneously optimize resolution and analysis time, the chromatographic optimization function (COF) was adopted to evaluate chromatograms. The central composite design (CCD) was used to create the matrix of experiments for mapping the chromatographic response surface. Finally, the COF values were fitted into a second order polynomial model and the response surface methodology (RSM) was employed to find the optimal eluent constituents. The reliability of the established model was confirmed by the good agreement obtained between experimental data and predictive values. Based on the results from the preliminary screening experiment and the CCD optimization, the optimal mobile phase was identified as a solution consisting of 6.68% (w/w) polyoxyethylene lauryl ether (Brij35), 0.84% (w/w) cyclohexane, 6.92% (w/w) n-butanol, 85.56% (w/w) phosphate buffer (pH 6.60) and 8 mM cetyltrimethyl ammonium bromide (CTAB).
Keywords: Microemulsion liquid chromatography; Salvia miltiorrhiza; Phenolic acids; Tanshinone; Central composite design; Chromatographic optimization function;