Organic compounds with amino functional groups are very important biomolecules,and the quantitative analysis of these compounds plays a critical role in many fields including clinical diagnostics,food analysis,and biochemical research [1, 2]. In particular,amino acid and biogenic amine analyses have become one focus of international research; and several analytical approaches such as HPLC  or CE [4, 5] coupled to a variety of detectors have been developed for this purpose. CE,because of its electrophoretic separation mechanism,is particularly suited to analyzing charged amines or amino acids. However,direct detection of these compounds is rather difficult due to their low volatility and lack of chromophores. In practice,analyte derivatization (pre-,on- or post-column) has played an important role to overcome these challenges,and the LIF detection offers especially outstanding sensitivity in a variety of detectors . A number of fluorescence reagents have been proposed for amino compounds derivatization. Several reviews describing the derivatization reagents and procedures for CE-LIF detection have been published [4, 7, 8]. Generally,naphtalene-2,3-dicarboxyaldehyde, o-phthaldialdehyde,9-fluorenylmethyl chloroformate,3- (4-carboxybenzoyl)quinoline-2- carboxaldehyde,7-fluoro- 4-nitrobenzo-2-oxa-1,3-diazole,and fluorescein isothiocyanate (FITC),etc. have been reported. Among these,FITC is one of the most suitable and widely used reagents for derivatization of primary and secondary amine groups,and the resultant products are highly fluorescent.
Saliva is a readily accessible and informative biofluid,making it ideal for the early detection of a wide range of diseases including cardiovascular,renal,and autoimmune diseases,viral and bacterial infections and,more importantly,cancer [9, 10, 11]. In recent years,several research groups have tried to analyze some amino compounds in human saliva based on CE coupled with LIF [1, 12, 13, 14, 15],mass spectrum ,or contactless conductivity detection . Because of its easy sample purification,lack of interference,and low organic solvent consumption,hollow-fiber liquid-phase microextraction (HF-LPME)  has gained attention in the field of sample preparation. HF-LPME is a microextraction process based on mass transfer through the supported liquid membrane formed by an organic solvent within the fiber holes of thewall,with extraction taking place in the porous hollow fiber lumen. The acceptor phase is not in direct contact with the sample solution,thus avoiding the solvent loss that happens easily in single-dropmicroextraction and dispersive liquid-phase microextraction. Moreover,as macromolecules,granular impurities etc. are not transferable across the fiber wall holes,HF-LPME also has amore prominent sample purification function compared with solid-phase microextraction and liquid-phase microextraction methods mentioned above,hence expanding the scope of analytical substrates. As a result,it can be used for direct analyses of samples with complex matrices.
In the present work,a novel method of HF-LPME/CE-LIF combining with FITC derivatization,has developed for the first time to determine three biogenic polyamines (putrescine (Put), cadaverine (Cad),and spermidine (Spe)) and their amino acid precursors (ornithine (Orn),lysine (Lys),and arginine (Arg)) in non-invasive saliva samples. The parameters involved in derivatization, electrophoretic separation and enrichment factors (EFs) of the targeted analytes were optimized. This method took advantage of the efficiency,reproducibility,ultra-small sample volume of CE, the high sensitivity of LIF and the enrichment capability of HFLPME. So this detection method that we developed has the superiority of low LOD (nmol) and less injection volume. The proposed method has been applied to analyze real-world saliva samples from healthy volunteers,and patients suffering from different oral diseases including halitosis,gingivitis,dental plaque, tongue cancer or gingival cancer. 2. Experimental
The laboratory-built CE-LIF system was employed and described previously . A fused-silica capillary (25 mm i.d. × 360 mm o.d.,Polymicro Technologies,Phoenix,AZ,USA) was used for the separation,and the effective capillary length was 90 cm for LIF. LIF detector (TriSepTM-2100LIF,Unimicro (Shanghai) Technologies Co.,Ltd.,Shanghai,China) was used in this work. The excitation wavelength of LIF detector was 483 nm,and the emission central wavelength was 520 nm.
Put (≥98.0%),Cad (≥99.0%),Spe (≥98.0%),Orn (>98.0%),Lys (99.0%),Arg (99.0%),and FITC (99.0%) were purchased from Sigma- Aldrich (St. Louis,MO,USA); poly(sodium-p-styrenesulfonate) (PSS) and cetrimonium bromide (CTAB) were purchased from China National Pharmaceutical Group Corporation (Shanghai, China),and they were all of analytical grade and used as received. The stock solution of each analyte (1.0 mmol/L) was prepared with doubly distilled water,and that of FITC (1.0 mmol/L) was prepared with acetone; and all stock solutions were stored in a refrigerator at 4 ℃. A fresh,mixed standard solution was prepared daily by diluting the stock solution with a NaOH solution (300 mmol/L) to the desired concentrations.
Saliva samples of healthy volunteers were collected randomly from students in our laboratory,and those of patients suffered from halitosis,gingivitis,dental plaque,tongue cancer or gingival cancer were collected in Shanghai Ninth People’s Hospital (Shanghai,China). All samples were stored at -20 ℃. To a 1.5 μL micro-tube with 500 μL thawed saliva sample containing analytes of interest was added 500 μL of acetonitrile. The mixture was mixed thoroughly,centrifuged (High-speed desktop centrifuger, Flying Pigeon TGL-16C,Suzhou Bozhao Electronics Co.,Ltd., Suzhou,China) for 15 min at 10,000 rpm,and then filtered through 0.22 mm nylon filters. An appropriate amount of filtrate was derivatized by adding 180 μL of FITC (1.0 mmol/L) as a derivatization reagent,360 μL of borate buffer (pH 9.23,200 mmol/L) and ultra-pure water to a total volume of 1 μL in a 1.5 μL centrifuge tube. After gentle mixing,the reaction mixture was incubated in the dark at room temperature for 12 h. The derivatization ratio of FITC/analyte of 3:1 was chosen as a good compromise considering the peak responses of all analytes and peak interference generated from FITC. (The effect of the derivatization ratio on the peak areas of the analytes was shown in Fig. S1 in Supporting information.) All the samples were repeatedly injected three times for regression analysis and error analysis. Before extraction,the reactants were stored at 4 ℃ in darkness.
Q3/2 Accurel PP polypropylene microporous hollow-fiber membrane (200 μm wall thickness,600 mm inner diameter, 0.2 μm pore size,75% porosity) were obtained from Membrana (Wuppertal,Germany). The hollow fiber was cut into segments with a length of 5 cm,and the detailed preparation can be found in our previous work . The optimization of dynamic three-phase HF-LPME procedure was conducted using a unified standard solution (10.0 nmol/L). An 8 μL of mixed standard solution (the pH value of the donor phase was adjusted with HCl to the final concentration of 5 mmol/L) was placed in a 10 μL beaker,and a small stirring bar was placed in the solution to ensure efficient stirring during the extraction,which was covered with aluminum foil to prevent targeted analytes from photodecomposition and evaporation. At the same time,the magnetic stirrer was switched on to start the dynamic extraction at 500 rpm. After a prescribed time,the magnetic stirrer was switched off and the hollow fiber was removed from the sample solution. One end of the hollow fiber was cut carefully with a sharp blade,and the needle tip of a microsyringe was carefully inserted into the hollow fiber. The acceptor solution in the hollow fiber was withdrawn into the syringe,and was ready for the CE-LIF analysis. Each piece of hollow fiber was used only for a single extraction. 3. Results and discussion
To identify the optimum electrophoretic conditions,several factors including the pH and concentration of the running buffer, buffer additives,separation voltage and injection time were investigated,respectively. Under the optimum conditions,six analytes could be well separated with the main coexisting interference compounds in saliva samples in 12 g/L PSS (3 mmol/L) CTAB (80 mmol/L) Na2B4O7-NaOH buffer (pH 12.35) at the separation voltage of 22 kV within 30 min; and the injection time was 8 s (at 22 kV).
The type of organic solvent,compositions of both donor and acceptor phases,stirring rate and extraction time were also investigated in order to obtain good extraction efficiency for the three-phase HF-LPME procedure. In three-phase HF-LPME,the type of organic solvent plays an important role in the extraction efficiency and the analyte preconcentration. So,the commonlyused organic solvents such as 1-octanol and toluene have been examined in this work. The experimental results showed that 1- octanol as the extraction solvent can offer the best enrichment characteristics for the targeted analytes and exhibited good reproducibility. The compositions of both donor and acceptor phases are also very important parameters affecting the extraction efficiency in HF-LPME. The pH value of the donor phase is adjusted to deionize the analytes,while the acceptor phase is adjusted to ionize them. Therefore,the concentrations of HCl as the donor phase and NaOH as the acceptor phase were studied,respectively. As shown in Fig. 1a and b,when the 5 mmol/L HCl and 300 mmol/L NaOH were used as donor media and acceptor media,respectively, the extraction process could provide the highest EFs. Some research has shown that addition of salt may improve the extraction efficiency by decreasing the aqueous solubility of the organic analytes. So,the effect of salt concentration was investigated in the range of 50-250 g/L as shown in Fig. 1c,and 200 g/L was selected as the concentration of sodium chloride in this experiment. Besides,the appropriate enhancement of the stirring rate can promote the diffusion of donor phase and the transfer of targeted analytes,and then shorten the extraction time; when the stirring rate was higher than 500 rpm,vortex phenomenon disrupted the organic-phase membrane and resulted in the loss of EFs. So,500 rpm was selected as the optimum stirring rate in this work. Finally,the effect of extraction time on EFs of the targeted analytes was also investigated as shown in Fig. 1d. In order to balance the extraction efficiency and analysis time,4 h was chosen as the optimum time in this work.
|Fig. 1.Effects of (a) HCl concentration, (b) NaOH concentration, (c) NaCl concentration and (d) extraction time on the EFs of the analytes. Optimum extraction conditions: supported liquid membrane: 1-octanol; donor phase: 5 mmol/L HCl; acceptor phase: 300 mmol/L NaOH; NaCl concentration: 200 g/L; stirring rate, 500 rpm; extraction time, 4 h. (a), (b) and (c) were conducted under the extraction time of 1 h.|
To determine the linearity of the six analytes of interest,a series of the mixed standard solutions containing 0.01-5.0 μmol/L of each analyte were tested. The correlation between peak area and concentration of each analyte was subjected to regression analysis to obtain the calibration equations and correlation coefficients,as listed in Table 1. The results showed an excellent linearity (r2 ≥ 0.998) between peak area and analyte concentration at three orders of magnitude. The LODs of six analytes ranged from 0.0072 nmol/L to 0.26 nmol/L (S/N = 3). And six analyte-FITC adducts could be pre-concentrated up to 218-fold.
The reproducibility of the method was evaluated by intraday precision and interday precision analyses,respectively. The RSD was used as a measure of precision. The intraday precision was assessed by making seven repetitive injections of the mixed standard solution of six analyte-FITC adducts at three different concentrations (5.0,0.50 and 0.050 μmol/L) under the selected optimum conditions. The assay results showed that the RSDs of peak area and migration time were within 0.3%,4.6% for Put,0.3%, 4.8% for Cad,1.3%,4.7% for Arg,0.6%,4.6% for Lys,0.6%,4.6% for Orn, and 0.7%,4.8% for Spe,respectively. Besides,the interday precision was also estimated by making repetitive injections of a mixed standard solution of 0.50 μmol/L for five consecutive days for three replicates,and the RSDs of peak areas and migration time were within 1.3% and 3.0%,respectively. The repeatability data indicated that it was feasible to determine the amino compounds based on the proposed CE-LIF method.
To further evaluate the reliability of the method,recovery experimentswere performed by a standard addition method. Under the optimum conditions,recovery data were determined with the real salivasamples of No.7 healthy volunteer,No.4 gingivitis patient, and gingival cancer patient at two concentration levels,respectively. The average recovery was in the range of 75.4%-120% with corresponding RSDs of 0.1%-3.0%,which indicated that the CE-LIF method was sufficiently accurate for the simultaneous determination of the targeted compounds. (The detailed results of recovery with the real samples were listed in Table S1 in Supporting information.)
Under the optimum conditions,the proposed method has been applied for simultaneously determining above three biogenic polyamines and their precursor amino acids in different saliva samples based on HF-LPME/CE-LIF. The saliva samples were collected from healthy volunteers and patients suffered from different oral diseases including halitosis,gingivitis,dental plaque, tongue cancer or gingival cancer,and the saliva electropherogramof tongue cancer patient was shown in Fig. 2b. By a standard addition method and comparing the migration time of targeted analyteswith those of the standard mixture solution electropherogram (Fig. 2a), six analytes were determined in different real-world samples. As seen from the sample electropherograms,a baseline separation (R ≥ 1.0) could be obtained for the analytes fromthemain coexisting substances as well as FITC under the optimized experimental conditions,despite the fact that the matrix of the saliva samples is rather complicated therefore will certainly challenge the separation of the analytes. (The resolution data of the targeted analytes in the real samples with this HF-LPME/CE-LIF method were listed in Table S2 in Supporting information.)
|Fig. 2.Typical electropherograms of (a) a mixed standard solution and (b) saliva sample of tongue cancer patient. CE-LIF conditions: fused-silica capillary: 25 μm i.d. × 90 cm (effective length); running buffer: 12 g/L PSS/3 mmol/L CTAB/80 mmol/L Na2B4O7-NaOH buffer (pH 12.35); separation voltage: 22 kV; injection time: 8 s (at 22 kV); excitation wavelength of LIF detector: 483 nm, emission central wavelength was 520 nm; peak identifications: (1) Put, (2) Cad, (3) Arg, (4) Lys, (5) Orn, (6) Spe, and (7) FITC. Extraction time: 4 h and other extraction conditions were the same as in Fig. 1.|
The detailed assay data for the tested samples were listed in Table 2. Generally speaking,the contents of Put,Cat and Spe in the saliva samples of patients were visibly higher than those of healthy volunteers,particularly for gingivitis,gingival cancer and halitosis patients; while their precursor amino acids in some patients’ salivawere also higher than those of healthy volunteers. Besides,in the matter of healthy volunteers,the contents of Put,Cat and Spe before brushing were clearly higher than those after brushing,indicating that brushing teeth can effectively decrease the levels of biogenic polyamines in saliva. Therefore, the levels of biogenic polyamines or the ratio of biogenic polyamines and precursor amino acidsmay have some correlation with certain oral diseases; and the above experimental conclusions were concordant with those of the above reported literature. For example,Ta´ bi et al.  have found that in the absence of oral hygiene,the mol fraction of Cad to Cad plus Lys in saliva increased significantly,indicating the presence of higher amount of bacterial Lys decarboxylase,which may contribute to periodontal diseases; Ye et al.  have also thought that Cad level in saliva exhibited some correlation with halitosis. However,further investigations based on larger sample size were suggested in order to collect statistically significant biological data.
In this work,a novel analyticalmethod has been developed and applied for the simultaneous determination of three biogenic polyamines and their precursor amino acids in human saliva samples based on HF-LPME-CE-LIF after FITC derivatization,and at the present stage the results showed that the levels of biogenic polyamines or the ratio of biogenic polyamines and precursor amino acids might have some correlation with certain oral diseases,and that in a way brushing teeth could decrease the levels of these polyamines in saliva. Therefore,the resultant electropherograms could provide more information for clinic diagnosis,providing a potential method for early non-invasive diagnosis of some oral diseases. The LODs of the proposedmethod are comparable to or significantly better (0.0072-0.26 nmol/L vs. 0.1-1.7 × 103 nmol/L) than those of reported CE-LIF methods [1, 12, 13, 14, 15]. Compared with other pretreatment methods reported, the proposed HF-LPME method has advantages in extraction efficiency,extraction time or organic solvent consumption. In order to collect statistically significant biological data,further investigations were suggested to discriminate between the pathological state of oral disease patients and the physiological conditions of healthy subjects.Acknowledgments
This work was supported by the Natural Science Foundation of China (No.21205042),the Special Funds for theDevelopmentofMajor Scientific Instruments and Equipment (No. 2011YQ15007205),and the Daxia Foundation of ECNU (No. 2012DX-185).Appendix A. Supplementary data
Supplementary data associated with this article can be found,in the online version,at http://dx.doi.org/10.1016/j.cclet.2014.01.037.
|||E. Pobozy, W. Czarkowska, M. Trojanowicz, Determination of amino acids in saliva using capillary electrophoresis with fluorimetric detection, J. Biochem. Biophys. Methods 67 (2006) 37.47.|
|||S.S. Li, H.L. Wu, Y.J. Liu, H.W. Gu, R.Q. Yu, Simultaneous determination of tyrosine and dopamine in urine samples using excitation-emission matrix fluorescence coupled with second-order calibration, Chin. Chem. Lett. 24 (2013) 239.242.|
|||A. Onal, S.E.K. Tekkeli, C. Onal, A review of the liquid chromatographic methods for the determination of biogenic amines in foods, Food Chem. 138 (2013) 509.515.|
|||V. Poinsot, M.A. Carpéné, J. Bouajila, et al., Recent advances in amino acid analysis by capillary electrophoresis, Electrophoresis 33 (2012) 14.35.|
|||M. Castro-Puyana, V. García-Can. as, C. Simó, A. Cifuentes, Recent advances in the application of capillary electromigration methods for food analysis and foodomics, Electrophoresis 33 (2012) 147.167.|
|||J.C.M. Waterval, H. Lingeman, A. Bult, W.J.M. Underberg, Derivatization trends in capillary electrophoresis, Electrophoresis 21 (2000) 4029.4045.|
|||W.J.M. Underberg, J.C.M. Waterval, Derivatization trends in capillary electrophoresis: an update, Electrophoresis 23 (2002) 3922.3933.|
|||H.A. Bardelmeijer, H. Lingeman, C. De Ruiter, W.J.M. Underberg, Derivatization in capillary electrophoresis, J. Chromatogr. A 807 (1998) 3.26.|
|||M. Sugimoto, D.T. Wong, A. Hirayama, T. Soga, M. Tomita, Capillary electrophoresis mass spectrometry-based saliva metabolomics identified oral, breast and pancreatic cancer-specific profiles, Metabolomics 6 (2010) 78.95.|
|||M. Mori, W.H. Hu, J.S.H. Fritz, T. Tsue, S. Kaneta, Tanaka, Determination of inorganic anions in human saliva by zwitterionic micellar capillary electrophoresis, Fresenius J. Anal. Chem. 370 (2001) 429.433.|
|||C. Streckfus, L. Bigler, M. Tucci, J.T. Thigpen, Preliminary study of CA15-3, c-erbB- 2, epidermal growth factor receptor, cathepsin-D, and p53 in saliva among women with breast carcinoma, Cancer Invest. 18 (2000) 101.109.|
|||A. Zinellu, S. Sotgia, E. Pisanu, et al., Quantification of neurotransmitter amino acids by capillary electrophoresis laser-induced fluorescence detection in fluids, Anal. Bioanal. Chem. 98 (2010) 1973.1978.|
|||Y.H. Deng, H. Wang, H.S. Zhang, Determination of amino acid neurotransmitters in human cerebrospinal fluid and saliva by capillary electrophoresis with laserinduced fluorescence detection, J. Sep. Sci. 31 (2008) 3088.3097.|
|||T. Tá bi, Z. Lohinai, M. Pálfi, M. Levine, É. Szöko, CE-LIF determination of salivary cadaverine and lysine concentration ratio as an indicator of lysine decarboxylase enzyme activity, Anal. Bioanal. Chem. 391 (2008) 647.651.|
|||H.M. Tseng, Y. Li, D.A. Barrett, Profiling of amine metabolites in human biofluids by micellar electrokinetic chromatography with laser-induced fluorescence detection, Anal. Bioanal. Chem. 388 (2007) 433.439.|
|||P. Tuůma, K. Málková , E. Samcová, K. Štulík, Rapid monitoring of arrays of amino acids in clinical samples using capillary electrophoresis with contactless conductivity detection, J. Sep. Sci. 33 (2010) 2394.2401.|
|||S. Pedersen-Bjergaard, K.E. Rasmussen, Liquid.liquid.liquid microextraction for sample preparation of biological fluids prior to capillary electrophoresis, Anal. Chem. 71 (1999) 2650.2656.|
|||J.B. Zhang, M.J. Li, Z. Li, et al., Study on urinary profile of inborn errors of metabolism by 18-crown-6 modified capillary electrophoresis with laser-induced fluorescence detection, J. Chromatogr. B 929 (2013) 102.106.|
|||Y.L. Pan, F. Chen, M.Y. Zhang, et al., Sensitive determination of chloroanilines in water samples by hollow fiber-based liquid-phase microextraction prior to capillary electrophoresis with amperometric detection, Electrophoresis 34 (2013) 1241.1248.|
|||W. Ye, Y.W. Li, X.P. Feng, Correlation study between cadaverine level in saliva and halitosis, Shanghai J. Stomatol. 16 (2007) 347.350.|