J S S
ISSN 1615-9306 · JSSCCJ 38 (8) 1263–1440 (2015) · Vol. 38 · No. 8 · April 2015 · D 10609
JOURNAL OF
SEPARATION SCIENCE
Methods Chromatography · Electroseparation Applications Biomedicine · Foods · Environment
8 15
www.jss-journal.com
1433
J. Sep. Sci. 2015, 38, 1433–1439
¨ Huma Yılmaz Hasan Basan Department of Analytical Chemistry, Faculty of Pharmacy, Gazi University, Ankara, Turkey Received November 28, 2014 Revised January 20, 2015 Accepted January 20, 2015
Research Article
Development of a molecularly imprinted solid-phase extraction sorbent for the selective extraction of telmisartan from human urine A novel molecularly imprinted solid-phase extraction with spectrofluorimetry method has been developed for the selective extraction of telmisartan from human urine. Molecularly imprinted polymers were prepared by a noncovalent imprinting approach through UVradical polymerization using telmisartan as a template molecule, 2-dimethylamino ethyl methacrylate as a functional monomer, ethylene glycol dimethacrylate as a cross-linker, N,N-azobisisobutyronitrile as an initiator, chloroform as a porogen. Molecularly imprinted polymers and nonimprinted control polymer sorbents were dry-packed into solid-phase extraction cartridges, and eluates from cartridges were analyzed using a spectrofluorimeter. Limit of detection and limit of quantitation values were 11.0 and 36.0 ng/mL, respectively. A very high imprinting factor (16.1) was achieved and recovery values for the telmisartan spiked in human urine were in the range of 76.1–79.1%. In addition, relatively low withinday (0.14–1.6%) and between-day (0.11–1.31%) precision values were obtained. Valsartan was used to evaluate the selectivity of sorbent as well. As a result, a sensitive, selective, and simple molecularly imprinted solid-phase extraction with spectrofluorimetry method has been developed and successfully applied to the direct determination telmisartan in human urine. Keywords: Human urine / Molecularly imprinted polymer / Solid-phase extraction / Telmisartan DOI 10.1002/jssc.201401349
Additional supporting information may be found in the online version of this article at the publisher’s web-site
1 Introduction Telmisartan (TEL), 4׳-[(1,4׳-dimethyl-2׳-propyl[2,6׳-bi-1Hbenzimidazol]-1׳-yl)methyl] [1,1׳-biphenyl]-2-carboxylic acid, is an angiotensin II receptor antagonist and is used for the treatment of hypertension and heart failure [1]. Urinary excretion of TEL is in the range of 0.1–4% [2]. Furthermore, due to placebo-like safety and tolerability in hypertensive patients, it provides significant advantages over other members of sartan
Correspondence: Professor Hasan Basan, Gazi University, Faculty of Pharmacy, Department of Analytical Chemistry, 06330, Etiler/Yenimahalle, Ankara, Turkey E-mail: [email protected] Fax: +90312235018
Abbreviations: AA, acetic acid; ACN, acetonitrile; AIBN, N,N-azobisisobutyronitrile; CHL, chloroform; DCE, dichloroethane; DMAEMA, 2-dimethylamino ethyl methacrylate; EGDMA, ethylene glycol dimethacrylate; IF, imprinting factor; MIP, molecularly imprinted polymer; MISPE, molecularly imprinted solid-phase extraction; NIP, nonimprinted polymer; TEL, telmisartan; TOL, toluene; VAL, Valsartan
C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
family, resulting in a significant alternative for the hypertension treatment. Several analytical methods have been reported in the literature for the determination of TEL, including LC–MS/MS [3, 4], HPLC [5, 6], SPME–HPLC [7, 8], ELISA [9], LC–MS [10], CZE [11], TLC with densitometry [12], spectrophotometry [13], and spectrofluorimetry [14,15]. However, some of these analytical methods, such as LC–MS or LC– MS/MS, require high-cost instrumentation, highly skilledoperators and labor-intensive sample preparation steps, especially for biological samples such as urine, plasma, and blood. In addition, there is a sensitivity problem in CZE, especially in the trace analysis. TLC with densitometry and spectrophotometric methods lack both sensitivity and selectivity. Furthermore, unlike other methods, spectrofluorimetry provides much better sensitivity, and selectivity. However, fluorescent compounds present especially in biological samples may interfere in the determination of analyte. For that reason, an efficient clean-up step is required before spectrofluorimetric determination. Molecular imprinting is a promising technique for the selective extraction and clean-up of various types of analytes in complicated matrices such as environmental and biological
www.jss-journal.com
1434
J. Sep. Sci. 2015, 38, 1433–1439
H. Yılmaz and H. Basan
samples [16]. Molecularly imprinted polymers (MIPs) are prepared with the copolymerization of functional and crosslinking monomers in the presence of a template molecule, resulting in highly crosslinked three dimensional network. After the removal of template by extraction, three dimensional cavities whose shape, size, functionality, and spatial arrangement of functional groups complimentary to the template molecule are generated [17]. Due to exhibition of more selective affinity toward target molecules, molecularly imprinted solid-phase extraction (MISPE) has been increasingly used for the clean-up and selective extraction of analytes from complex biological and environmental fluids, compared to conventional C18 SPE sorbents [18]. In the present study, a novel MISPE sorbent was prepared using 2-(dimethylamino)ethyl methacrylate (DMAEMA) as a functional monomer, ethylene glycol dimethacrylate (EGDMA) as a cross-linker, TEL as a template, N,Nazobisisobutyronitrile (AIBN), as an initiator, and chloroform (CHL) as a porogen by photochemical UV polymerization mechanism. Before the spectrofluorimetric detection, this sorbent was successfully applied to the selective extraction of spiked TEL from complex human urine. A detailed literature survey revealed that there was only one study reporting use of MISPE for the extraction of TEL from rat urine, plasma, and pharmaceutical formulation before HPLC analysis [19]. In this study, methacrylic acid was used as a functional monomer and MIPs were prepared by thermal polymerization mechanism. Therefore, to the best of our knowledge, no MISPE with spectrofluorimetry method has been reported for the selective recognition and determination of TEL in human urine to date. Due to selective sample clean-up achieved using MISPE, most of the interfering compounds in human urine were eliminated and relatively lower LOD and LOQ values were obtained compared to the ones obtained by above mentioned study. In addition, it should be emphasized that this study is the first attempt in using DMAEMA as a functional monomer in the preparation of TEL-imprinted MISPE sorbent. Thus, a highly sensitive and selective MISPE with spectrofluorimetry method has been established for selective extraction and direct determination of TEL in human urine.
2.2 Synthesis of MISPE sorbent For the synthesis of MIP, 0.063 mmol of TEL, and 0.186 mmol of DMAEMA were dissolved in 1.0 mL of CHL in a borosilicate tube and incubated for 15 min at room temperature. To this mixture, 1.27 mmol of EGDMA and 5.0 mg of AIBN were added. The tubes were sealed and then cooled in ice. Thereafter, photopolymerization reaction was initiated using a UV lamp (6 W, 365 nm; Upland, CA, USA) and proceeded 5 h. Then, tubes containing MIP sorbent were crushed, and bulk polymer was ground using A11 basic IKA analytical mill (Staufen, Germany), and then wet-sieved to between 20 and 63 m in size using methanol. Later on, sorbent was subjected to the extraction process by using an I˙ ldam Soxhlet extraction apparatus (Ankara, Turkey) using MeOH/AA (9:1, v/v) to remove template molecules for 48 h. Finally, MISPE sorbent was dried in an incubator (Nuve, Ankara, Turkey) at 50⬚C for 6 h. The nonimprinted polymer (NIP) was also prepared as a control polymer using same procedure but in the absence of template molecule.
2.3 Swelling studies Swelling behavior of MIP/NIP sorbents was investigated through incubation studies in aqueous (pH 3.5–8.5) and organic (CHL, ACN, TOL, DCE) conditions. Typically, 50 mg MIP/NIP particles was weighed into an empty SPE cartridge (Alltech 1.5 mL, Deerfield, IL, USA). A certain amount of a solvent was added into the cartridge and sorbent was allowed to swell for 9 h until it reaches a constant weight. After removing the excess solvent, cartridge containing the sorbent was weighed. Using Eq. (1), swelling was calculated: Swelling (%) = [(Hswollen − Hdry )/Hdry ] × 100
(1)
where Hswollen is the weight of swollen sorbent and Hdry is the weight of the dry sorbent.
2.4 Batch rebinding studies
2 Materials and methods 2.1 Reagents Telmisartan (TEL) was kindly donated by NUVOMED (Istanbul, Turkey) and Mustafa Nevzat (Istanbul, Turkey). Valsartan (VAL) was provided by Actavis Incorporation (Malta). DMAEMA, EGDMA, and AIBN were purchased from Sigma–Aldrich (Steinheim, Germany). Dichloroethane (DCE), methanol (MeOH), CHL, toluene (TOL), acetonitrile (ACN), and acetic acid (AA) were obtained from Merck (Darmstadt, Germany). All other chemicals were of analytical reagent grade. C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
20 mg of the dry sorbent (MIP or NIP) was placed into a glass vial containing 4.0 mL of 5.0 g/mL TEL solution. The mixture was stirred for 1.5 h and centrifuged at 4000 rpm for 30 min. Then, supernatant was analyzed using a spectrofluorimeter (Varian Cary Eclipse, Mulgrave, Victoria, Australia; ex : 302 nm, em : 362 nm, slit widths: 5 nm) for the determination of free TEL in the binding medium. Amount of TEL adsorbed by MIP and NIP sorbents was found by subtracting amount of free TEL in the solution from the initial amount in the binding medium. For the determination of appropriate batch rebinding conditions, type of incubation solvent (DCE, ACN, CHL, and TOL), incubation time (30–180 min), amount of sorbent (5–40 mg), and initial concentration of TEL (1.0–15.0 g/mL) were optimized. www.jss-journal.com
Other Techniques
J. Sep. Sci. 2015, 38, 1433–1439
The equilibrium-binding capacities of MIP and NIP sorbents were determined using Eq. (2): Qe = (Ci − Cf ) × V /m
(2)
where Ci (g/mL) and Cf (g/mL) are the initial and final concentrations of TEL, V (mL) is the volume of incubation solvent, and m (g) is the mass of the sorbent. Langmuir and Freundlich [20, 21] isotherm models were applied to the MIP sorbent using Eqs. (3) and (4), respectively: Ce /Qe = (1/b Qo ) + (Ce /Qo ) LogQe = LogK F + (1/n) LogCe
(3) (4)
where Qe and Qo (mg/g) are the amounts of sorbate bound to per unit mass of sorbent at equilibrium and saturation, b (L/mg) is the Langmuir constant, Ce (mg/L) is the equilibrium concentration of sorbate, KF (L/mg) is the Freundlich constant, n is Freundlich exponent. KF is correlated to the maximum adsorption capacity. During the selectivity study, 20 mg of MIP/NIP sorbents was incubated in 4.0 mL of 5.0 g/mL TEL and VAL solutions, separately.
2.5 MISPE protocol Fifty milligrams of MIP/NIP particles was dry-packed into 1.5 mL SPE cartridge between two polyethylene frits and connected to a vacuum manifold (Alltech vacuum manifold, Deerfield, IL, USA). The cartridge was first conditioned with 2 × 0.5 mL of loading solvent (ACN or pH 7.4 distilled water). Then, 4 × 0.5 mL of 10 g/mL TEL solution prepared in ACN or distilled water (pH = 7.4) was loaded into the cartridge. Thereafter, cartridge was washed with 6 × 0.5 mL of CAN and TEL was eluted from the cartridge using 8 × 0.5 mL of MeOH/AA (9:1, v/v). Amount of TEL in the eluate was determined spectrofluorimetrically by referring to calibration curve ranging from 0.04 to 1.40 g/mL. Before the new extraction process, cartridges were regenerated using 6 × 0.6 mL of MeOH/AA (9:1, v/v).
2.6 Analysis of human urine Blank urine samples were collected from a healthy volunteer and stored in a refrigerator at –20⬚C till use. Initially, human urine samples were centrifuged at 3000 rpm for 30 min and then, supernatant was filtered through 0.45 m millipore nylon membranes. pH of the filtrate was adjusted to 7.4 with 1.0 M NaOH. Stock standard solution of TEL, 1.0 mg/mL, was prepared in ACN. Working standard solutions were obtained by adding appropriate volumes of stock solution into 10.0 mL volumetric flasks and then were diluted to the mark with the blank human urine (pH 7.4). A total of 4 × 0.5 mL of each standard solution was then passed through C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1435
MIP and NIP cartridges. After applying washing and elution steps as mentioned in Section 2.5, matrix-matched calibration curve in the range of 0.04 to 1.40 g/mL was generated. For the recovery and precision studies, urine samples (2.0 mL, pH 7.4) spiked with 0.10–0.80 g/mL TEL solutions were employed.
3 Results and discussion 3.1 Preparation of MISPE sorbent In this study, a novel MISPE sorbent was synthesized by noncovalent molecular imprinting approach for the selective extraction of TEL from human urine. Since an optimal functional monomer and template interaction is of prime importance in the formation of high affinity recognition sites for the target analyte, MIP formulation was carefully optimized and ideal molar ratio of TEL, DMAEMA, and EGDMA was found to be 1:3:20 (Supporting Information Table S1). To increase number of prepolymerization complex between DMAEMA and TEL, they were allowed to stay together for 15 min before adding cross-linker and initiator into the reaction medium. Due to presence of an acidic carboxyl group in TEL and a basic tertiary amine group in DMAEMA, there was a highly efficient and effective interaction between functional monomer and template molecules during molecular imprinting process, thus resulting in a very high imprinting factor (IF = 16.1). It was concluded that molecular interaction during molecular imprinting was mostly due to hydrogen bonding between carboxyl group of TEL and tertiary amine group of DMAEMA. Furthermore, high IF value clearly indicated that highly selective TEL imprinted recognition sites were successfully generated in MIP particles. In noncovalent imprinting technique, MIPs should be prepared in low-polarity organic solvents to avoid the disruption of weak intermolecular interactions, such as hydrogen bonding, electrostatic, and van der Waals interactions, between the functional monomer and template molecule in the prepolymerization mixture. Hence, CHL, a low-polarity and aprotic solvent, was selected as a porogen to enhance the strength of hydrogen-bonding interactions present between TEL and DMAEMA. The effect of porogen volume (0.5– 1.5 mL) was also tested and optimum value was determined to be 1.0 mL (Supporting Information Table S1). During the MIP synthesis, photochemical free-radical polymerization at ambient and cooled media were applied because it is wellknown that polymers prepared at higher temperatures (e.g. thermal polymerization) exhibit reduced molecular recognition properties. For that reason, polymerization mixture in borosilicate glass tube was cooled on ice to raise the degree of interaction between TEL and DMAEMA by reducing the molecular vibrations. Results of MISPE studies indicated that higher recovery value (81.1% with cooling and 46.6% without cooling, loading solvent was ACN) was obtained when polymerization mixture was cooled with ice. www.jss-journal.com
1436
J. Sep. Sci. 2015, 38, 1433–1439
H. Yılmaz and H. Basan
Table 1. Batch rebinding studies for the MIP and NIP in different incubation solvents
Binding % Solvent
QMIP
QNIP
Imprinting factor (IF)
Dichloroethane Acetonitrile Chloroform Toluen
83.2 94.9 82.1 80.1
9.0 5.9 10.2 9.7
9.2 16.1 8.0 8.3
During the incubation studies, 4.0 mL of 5.0 g/mL TEL solution was used. QMIP , amount adsorbed by MIP; QNIP , amount adsorbed by NIP.
3.2 Swelling study Swelling of MIP sorbent greatly affects its recognition ability because size and shape of imprinted cavity may change significantly, resulting in loss of selectivity. For this reason, swelling studies for MIP and NIP sorbents were performed in various solvents and results were given in Supporting Information Table S2. At lower pH values (pH < 7.4), both MIP and NIP sorbents reached a higher swelling % (52.4–58.4% for MIP). However, at pH ࣙ 7.4, they have lower swelling % values (37–38.1%). This pH-sensitive behavior can be attributed to the protonation of tertiary amine functional group in the DMAEMA monomer at lower pH values, resulting in swelling of polymeric sorbent due to repulsion of positive charge bearing polymeric chains each other. Therefore, it was concluded that higher recognition ability of MIP sorbent would be at higher pH values because of the fact that less size and shape change would occur. Thus, for the aqueous samples, it was decided that TEL loading should be performed at pH 7.4 medium, probably due to expecting less swelling and in turn, higher selectivity. Swelling studies were also investigated in various organic solvents including CHL, ACN, TOL, and DCE. As can be seen from Supporting Information Table S2, the lowest and highest swelling degrees were reached by using ACN and CHL, respectively. In light of these observations, it might be concluded that geometry of imprinted cavities would change less if ACN is selected as the loading solvent.
3.3 Batch rebinding studies Before MISPE studies, MIP and NIP sorbents are generally tested to investigate their target analyte recognition and binding abilities, i.e. imprinting effect. It is apparent from Table 1 that the highest TEL-binding percent and in turn highest IF value were obtained using ACN as an incubation solvent. This is consistent with the swelling study results, stating that less swelling and in turn less change in geometry of MIP occurs in ACN solvent. Subsequently, optimum incubation time and sorbent amount were found to be C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
90 min and 20 mg (results are not shown here), respectively. Incubation time is of great importance because it determines how much analyte would be adsorbed by the sorbent. It should be given enough time for the analyte molecules to fill the imprinted cavities situated on the surface and interior parts of the sorbent. The effect of target analyte, TEL, initial concentration on the MIP/NIP-binding capacity was also investigated, and the results are illustrated in Fig. 1. The maximum-binding capacity for MIP was achieved when 5.0 mg/L initial TEL concentration was used. The binding capacities for MIP and NIP sorbents were found to be 0.97 and 0.092 mg TEL/g sorbent, respectively. Thus, high difference in binding capacities of MIP and NIP particles can be easily attributed to the molecular imprinting effect. Equilibrium adsorption results were evaluated using Langmuir and Freundlich models to estimate adsorption behavior of MIP sorbent as well. The Langmuir model states that maximum adsorption takes place when a saturated monolayer of solute molecules is present on the sorbent surface [20]. On the other hand, the Freundlich model is a multisite adsorption isotherm for heterogeneous surfaces [21]. The adsorption data fitted well in the Langmuir equation with regression coefficient value of 0.9998 compared to 0.4960 obtained using the Freundlich model (Supporting Information Table S3). The b and Qo values were determined from the slope and intercept of isotherm graph to be 0.076 L/mg and 0.9708 mg/g, respectively. Good fitting of experimental data to Langmuir isotherm proves the existence of homogeneous TEL-binding sites on MIP sorbent. Selectivity of MIP and NIP sorbents to TEL was investigated using valsartan (VAL). VAL was chosen due to its molecular structure similar to TEL to some extent and its ability to form hydrogen bonds with functional groups in MIP cavities. In addition, it contains carboxyl and nitrogen functional groups like TEL. As depicted in Supporting Information Table S4, MIP showed significantly higher selectivity or affinity toward TEL molecules compared to VAL molecules as far as IF values were concerned. Almost 10 times higher IF value (IF for TEL is 16.1 and IF for VAL is 1.51) is certainly due to imprinting effect generated by template TEL molecules during MIP synthesis. 3.4 Optimization of MISPE conditions For the determination of trace levels of analytes in biological fluids, it is mostly desirable to have a MISPE protocol providing high level of analyte recovery and extracts free from interfering compounds. For that reason, several parameters including loading solvent type, pH of aqueous loading solvent, loading time, washing solvent, and volume of elution solvent were carefully optimized. Before the loading step, MIP cavities were first activated using loading solvent to maximize the interactions between TEL molecules and functional groups in imprinted cavities. When ACN was used as loading solvent, 81.1 and 29.4% recovery values were obtained for MIP and NIP, respectively. On the other hand, when TOL was used, corresponding values for MIP and NIP were found www.jss-journal.com
Other Techniques
J. Sep. Sci. 2015, 38, 1433–1439
1437
Figure 1. Effect of initial concentration of TEL on binding capacity for MIP and NIP sorbents in batch study.
to be 65.5 and 31.7%, respectively. Thus, ACN was selected as optimum loading solvent for further experiments. Effect of sample loading pH on the recovery of TEL was investigated in the pH range of 3.5–8.5. As shown in Supporting Information Fig. S1, best recovery value was achieved at pH 7.4. This pH-dependent change in recovery can be attributed to the swelling behavior of the sorbent mentioned in Section 3.2. At lower pH values (pH < 7.4), size and shape selective abilities of the sorbent decrease due to swelling of imprinted sites, resulting in diminished recognition ability. At higher pH values (pH ࣙ 7.4), since the sorbent swells less, loss in the recognition ability is relatively low. For this reason, pH 7.4 was specified as optimum sample loading pH for the aqueous samples. It might be concluded that swelling has a great influence on the recognition ability of the MIP because swelling causes size and shape of the imprinted sites of the MIP to change, leading to loss of its memory effect. Effect of loading time was also studied because it determines strength or efficiency of the interaction between analyte and imprinted cavities and also diffusion time from bulk solution to rebinding sites. Two options were tested in that firstly MISPE study was done by directly passing to the washing step right after the sample loading step and 57.4 and 36.3% recovery values were obtained for MIP and NIP, respectively. In the second study, we waited for 3 min after loading TEL solution before passing to the washing step. Thus, TEL molecules found enough time to interact with the functional groups of the imprinted cavities. As a result of these interactions, relatively higher recovery values were attained (80.8.0% for MIP and 31.1% for NIP). To achieve a selective extraction, a clean-up step with an organic solvent is applied before the elution step and this step is more crucial in MISPE applications than in conventional SPE [22]. Thus, washing solvent should suppress the nonspecific interactions without disrupting the selective specific interactions between the imprinted cavities and the target molecule. When ACN was used as the washing solvent, C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
recovery values for MIP and NIP were 81.1 and 29.4%, respectively. On the other hand, when TOL was tried, recovery value for MIP was 72.5% and that for NIP was 37.5%. It can be easily concluded that higher recovery value for MIP and higher specific binding result were obtained when ACN was employed as the washing solvent. Thus, ACN was selected as the most appropriate washing solvent. Moreover, the volume of ACN (1.0–4.0 mL) was also tested. Since the highest recovery value was generated, 3.0 mL of ACN was determined as appropriate washing solvent volume for further studies. Supporting Information Fig. S2 displays the effect of the elution solvent, MeOH/AA (9:1, v/v), volume on the TEL recovery. Due to higher recovery and specific binding values, 4.0 mL elution solvent was determined as appropriate elution solvent volume. As the volume of the elution solvent increases, amount of TEL eluted from the cartridge increases, reaching maximum value when 4.0 mL was used. Since the strength of interactions between imprinted cavities and TEL molecules was very high, 2.0 or 3.0 mL of elution solvent was not enough to elute all of TEL molecules, thus requiring 4.0 mL. Because an acidic elution solvent is used, sorbent and, in turn, imprinted cavities swell to a certain degree, leading to disruption of intermolecular attraction between TEL molecules and functional groups in imprinted cavities. As a result, TEL molecules are loosely held inside the sorbent and easily eluted from the sorbent. The influence of the sorbent amount (25, 50, and 75 mg) packed into the SPE cartridges on TEL recoveries was also studied. The highest recovery value (81.1%) was achieved when 50 mg MISPE sorbent was used.
3.5 Analytical parameters To demonstrate the suitability of the developed MISPE with spectrofluorimetry method for the determination of TEL
www.jss-journal.com
1438
J. Sep. Sci. 2015, 38, 1433–1439
H. Yılmaz and H. Basan
Table 2. Recovery and precision results for the assay of TEL in spiked human urine Recovery (%) (n = 3)
Spiked TEL (g/mL)
Precision Within-day (RSD%, n = 5)
Between-day (RSD%, n = 5)
MIP
NIP
0.10 0.40 0.80
1.6 0.31 0.14
1.31 0.44 0.41
76.1 77.3 79.1
32.1 32.1 32.1
in human urine, analytical parameters including linearity, LOD, LOQ, within-day (repeatability), and between-day (reproducibility) precisions, and accuracy were validated. For the validation of linearity, calibration curve was constructed using optimized MISPE protocol for the blank urine samples spiked with TEL over a concentration range of 0.04– 1.40 g/mL. A relatively high correlation coefficient value (r = 0.9965) indicated the linearity of the calibration curve. LOD and LOQ were calculated using the SD on the concentration of standard solution having the lowest concentration and slope of the calibration curve. It should be stated that LOD (11.0 ng/mL) and LOQ (36.0 ng/mL) values were relatively low compared to those obtained by Mudiam et al. [19]. In their study, before the HPLC analysis, a combination of MISPE and dispersive liquid–liquid microextraction for the selective preconcentration and determination of TEL in rat urine was described, resulting in relatively higher LOD (190 ng/mL) and LOQ (630 ng/mL) values compared to our proposed method. Thus, it can be concluded that the proposed method is much more sensitive and lower levels of TEL can be selectively determined as a result of better clean-up protocol provided by the newly synthesized MIP sorbent. Human urine samples spiked with low (0.10 g/mL), medium (0.40 g/mL), and high concentrations (0.80 g/mL) of TEL were prepared and analyzed for the determination of withinday and between-day precision of the proposed method. The within-day precision was calculated by performing five extractions of independently prepared urine samples spiked with TEL at different concentrations over a day. Between-day precision was assessed by studying five extractions of independently prepared urine samples spiked with TEL at different concentrations in three different days. Results of withinday and between-day precision studies were presented in Table 2 and relatively satisfactory RSD values ranging from 0.14–1.6% for the within-day and 0.41–1.3% for the betweenday studies were achieved. These results prove the precision of the proposed MISPE with spectrofluorimetry method. For recovery studies, three different concentration levels (0.10– 0.80 g/mL) of TEL spiked into urine samples were studied, Table 2 and relatively high recovery values were achieved. The difference between the recoveries of MIP and NIP sorbents clearly indicates the success of molecular imprinting. These recovery values are comparable to the ones obtained by Ferreiros et al. (74%) in that before the HPLC with fluorimetric detection, a clean-up step consisting of a SPE with C8 cartridges was employed for urine samples [23]. In addition, Zhang et al. [11] prepared a polymeric monolithic C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
microextraction using poly(methacrylic acid-ethylene glycol dimethacrylate) in conjunction with CZE and reached 81.7– 84.2% recovery values in human urine. The LOD and LOQ values they found were 15.0 and 50.0 ng/mL, respectively. It can be concluded that the LOD and LOQ values for the proposed MISPE with spectrofluorimetry method are comparable to the values obtained by Zhang et al. Since TEL in complex human urine was selectively retained and cleaned-up using the proposed method, high recovery, and low LOD values were attained without applying complicated sample pretreatment procedures. Furthermore, low LOD and high recovery values in aqueous rebinding conditions revealed that the proposed MISPE sorbent had high water compatibility, making direct application of human urine samples to MISPE cartridges possible.
4 Concluding remarks In this study, a novel MISPE sorbent was synthesized and used in combination with a very simple and cost effective spectrofluorimetry method for the selective extraction and determination of TEL in human urine. A very high imprinting factor value (16.1) confirmed the success of molecular imprinting. The developed MISPE with spectrofluorimetry method combines sensitive and selective features of spectrofluorimetry with the highly selective properties of the molecular imprinting technique and, in turn, uses their synergistic effect in the determination of TEL in relatively complex human urine sample. The spiked TEL recovery values were in the range of 76.1–79.1% for the proposed method. In addition to the high recovery values, selective and effective sample clean-up procedure was developed. Since most of the interfering compounds in human urine were removed, direct determination of TEL in complex human urine sample was achieved. Due to selectivity and sensitivity of the proposed method, trace levels of TEL in biological samples can be easily determined in clinical and routine analyses without using costly instrumentation and tedious sample preparation methods. ¨ The present study was financially supported by OYP department at Gazi University, Turkey. The authors have declared no conflict of interest.
5 References [1] Krishna, P. R. V., Kasturi, J. K., Babu, A. V., Vijay, K., Raman, G. V., Chakravarthy, I. E., Babu, B. H., Int. J. Pharm. Sci. Res. 2005, 4, 3630–3634. [2] Wienen, W, Entzeroth, M., vanMeel, J. C. A., Stangier, J., Busch, U., Ebner, T., Schmid, J., Lehmann, H., Matzek, K., Kempthorne-Rawson, J., Gladigau, V., Hauel, N. H., Cardiovasc. Drug Rev. 2000, 18, 127–154.
www.jss-journal.com
J. Sep. Sci. 2015, 38, 1433–1439
[3] Hempen, C., Schwarz, L. G., Kunz, U., Karst, U., Anal. Chim. Acta 2006, 560, 41–49. [4] Gupta, V. K., Jain, R., Lukram, O., Agarwal, S., Dwivedi, A., Talanta 2011, 83, 709–716.
Other Techniques
1439
[13] Ilango, K., Kumar, P. S. S., Pharm. Meth. 2012, 3, 112–116. [14] Belal, T. S., Mahrous, M. S., Khalek, M. M. A., Daabees, H. G., Khamis, M. M., Luminescence 2014, 29, 893– 900.
[5] Brunetto, M. R., Contreras, Y., Clavijo, S., Torres, D., Delgado, Y., Ovalles, F., Ayala, C., Gallignani, M., Estela, J. M., Martin, V. C., J. Pharm. Biomed. Anal. 2009, 5, 194– 199.
[15] Cagigal, E., Gonzalez, L., Alaso, R. M., Jimenez, R. M., J. Pharm. Biomed. Anal. 2001, 26, 477–486.
[6] Torrealday, N., Gonzalez, L., Jimenez, R. M., Lastra, E. O., J. Pharm. Biomed. Anal. 2003, 32, 847–857.
[17] Lavignac, N., Allender, C. J., Brain, K. R., Anal. Chim. Acta 2004, 510, 139–145.
[7] Nie, J., Zhang, M., Fan, Y, Wen, Y., Xiang, B., Feng, Y. Q., J. Chromatogr. B 2005, 828, 62–69.
[18] Lai, J. P., Chen, F., Sun, H., Fan, L., Liu, G. L., J. Sep. Sci. 2014, 37, 1170–1176.
[8] Nie, J., Zhao, Q., Huang, J., Xiang, B., Feng, Y. Q., J. Sep. Sci. 2006, 29, 650–655.
[19] Mudiam, M. K. R., Chauhan, A., Singh, A. K., Sharma, V. P., Saxena, P. N., Bioanalysis 2013, 5, 847–858.
[9] Hempen, C., Schwarz, L. G., Kunz, U., Karst, U., Anal. Chim. Acta 2006, 560, 35–40.
[20] Langmuir, L., J. Am. Chem. Soc. 1918, 40, 1361–1403.
[10] Chen, B., Liang, Y., Wang, Y., Deng, F. L., Zhou, P., Guo, F., Huang, L., Anal. Chim. Acta 2005, 540, 367–373.
[16] Won, J. C., Song, H. Y., Faiz, A., J. Sep. Sci. 2013, 36, 609–628.
[21] Freundlich, H. M. F., Z. Phys. Chem. 1906, 57(A), 385–470.
[11] Zhang, M., Wei, F., Zhang, Y. F., Nie, J., Feng, Y. Q., J. Chromatogr. A 2006, 1102, 294–301.
[22] Caro, E., Marce, R. M., Borrull, F., Cormack, P. A. G., Sherrington, D. C., TrAC, Trends Anal. Chem. 2006, 25, 143– 154.
[12] Tsvetkova, D. D., Obreshkova, D. P., J. Planar Chromatogr. 2012, 25, 326–330.
[23] Ferreiros, N., Iriarte, G., Alonso, R. M., Jimenez, R. M., Talanta 2007, 73, 748–786.
C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.jss-journal.com
ISSN 1615-9306 · JSSCCJ 38 (8) 1263–1440 (2015) · Vol. 38 · No. 8 · April 2015 · D 10609
JOURNAL OF
SEPARATION SCIENCE
Methods Chromatography · Electroseparation Applications Biomedicine · Foods · Environment
8 15
www.jss-journal.com
1433
J. Sep. Sci. 2015, 38, 1433–1439
¨ Huma Yılmaz Hasan Basan Department of Analytical Chemistry, Faculty of Pharmacy, Gazi University, Ankara, Turkey Received November 28, 2014 Revised January 20, 2015 Accepted January 20, 2015
Research Article
Development of a molecularly imprinted solid-phase extraction sorbent for the selective extraction of telmisartan from human urine A novel molecularly imprinted solid-phase extraction with spectrofluorimetry method has been developed for the selective extraction of telmisartan from human urine. Molecularly imprinted polymers were prepared by a noncovalent imprinting approach through UVradical polymerization using telmisartan as a template molecule, 2-dimethylamino ethyl methacrylate as a functional monomer, ethylene glycol dimethacrylate as a cross-linker, N,N-azobisisobutyronitrile as an initiator, chloroform as a porogen. Molecularly imprinted polymers and nonimprinted control polymer sorbents were dry-packed into solid-phase extraction cartridges, and eluates from cartridges were analyzed using a spectrofluorimeter. Limit of detection and limit of quantitation values were 11.0 and 36.0 ng/mL, respectively. A very high imprinting factor (16.1) was achieved and recovery values for the telmisartan spiked in human urine were in the range of 76.1–79.1%. In addition, relatively low withinday (0.14–1.6%) and between-day (0.11–1.31%) precision values were obtained. Valsartan was used to evaluate the selectivity of sorbent as well. As a result, a sensitive, selective, and simple molecularly imprinted solid-phase extraction with spectrofluorimetry method has been developed and successfully applied to the direct determination telmisartan in human urine. Keywords: Human urine / Molecularly imprinted polymer / Solid-phase extraction / Telmisartan DOI 10.1002/jssc.201401349
Additional supporting information may be found in the online version of this article at the publisher’s web-site
1 Introduction Telmisartan (TEL), 4׳-[(1,4׳-dimethyl-2׳-propyl[2,6׳-bi-1Hbenzimidazol]-1׳-yl)methyl] [1,1׳-biphenyl]-2-carboxylic acid, is an angiotensin II receptor antagonist and is used for the treatment of hypertension and heart failure [1]. Urinary excretion of TEL is in the range of 0.1–4% [2]. Furthermore, due to placebo-like safety and tolerability in hypertensive patients, it provides significant advantages over other members of sartan
Correspondence: Professor Hasan Basan, Gazi University, Faculty of Pharmacy, Department of Analytical Chemistry, 06330, Etiler/Yenimahalle, Ankara, Turkey E-mail: [email protected] Fax: +90312235018
Abbreviations: AA, acetic acid; ACN, acetonitrile; AIBN, N,N-azobisisobutyronitrile; CHL, chloroform; DCE, dichloroethane; DMAEMA, 2-dimethylamino ethyl methacrylate; EGDMA, ethylene glycol dimethacrylate; IF, imprinting factor; MIP, molecularly imprinted polymer; MISPE, molecularly imprinted solid-phase extraction; NIP, nonimprinted polymer; TEL, telmisartan; TOL, toluene; VAL, Valsartan
C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
family, resulting in a significant alternative for the hypertension treatment. Several analytical methods have been reported in the literature for the determination of TEL, including LC–MS/MS [3, 4], HPLC [5, 6], SPME–HPLC [7, 8], ELISA [9], LC–MS [10], CZE [11], TLC with densitometry [12], spectrophotometry [13], and spectrofluorimetry [14,15]. However, some of these analytical methods, such as LC–MS or LC– MS/MS, require high-cost instrumentation, highly skilledoperators and labor-intensive sample preparation steps, especially for biological samples such as urine, plasma, and blood. In addition, there is a sensitivity problem in CZE, especially in the trace analysis. TLC with densitometry and spectrophotometric methods lack both sensitivity and selectivity. Furthermore, unlike other methods, spectrofluorimetry provides much better sensitivity, and selectivity. However, fluorescent compounds present especially in biological samples may interfere in the determination of analyte. For that reason, an efficient clean-up step is required before spectrofluorimetric determination. Molecular imprinting is a promising technique for the selective extraction and clean-up of various types of analytes in complicated matrices such as environmental and biological
www.jss-journal.com
1434
J. Sep. Sci. 2015, 38, 1433–1439
H. Yılmaz and H. Basan
samples [16]. Molecularly imprinted polymers (MIPs) are prepared with the copolymerization of functional and crosslinking monomers in the presence of a template molecule, resulting in highly crosslinked three dimensional network. After the removal of template by extraction, three dimensional cavities whose shape, size, functionality, and spatial arrangement of functional groups complimentary to the template molecule are generated [17]. Due to exhibition of more selective affinity toward target molecules, molecularly imprinted solid-phase extraction (MISPE) has been increasingly used for the clean-up and selective extraction of analytes from complex biological and environmental fluids, compared to conventional C18 SPE sorbents [18]. In the present study, a novel MISPE sorbent was prepared using 2-(dimethylamino)ethyl methacrylate (DMAEMA) as a functional monomer, ethylene glycol dimethacrylate (EGDMA) as a cross-linker, TEL as a template, N,Nazobisisobutyronitrile (AIBN), as an initiator, and chloroform (CHL) as a porogen by photochemical UV polymerization mechanism. Before the spectrofluorimetric detection, this sorbent was successfully applied to the selective extraction of spiked TEL from complex human urine. A detailed literature survey revealed that there was only one study reporting use of MISPE for the extraction of TEL from rat urine, plasma, and pharmaceutical formulation before HPLC analysis [19]. In this study, methacrylic acid was used as a functional monomer and MIPs were prepared by thermal polymerization mechanism. Therefore, to the best of our knowledge, no MISPE with spectrofluorimetry method has been reported for the selective recognition and determination of TEL in human urine to date. Due to selective sample clean-up achieved using MISPE, most of the interfering compounds in human urine were eliminated and relatively lower LOD and LOQ values were obtained compared to the ones obtained by above mentioned study. In addition, it should be emphasized that this study is the first attempt in using DMAEMA as a functional monomer in the preparation of TEL-imprinted MISPE sorbent. Thus, a highly sensitive and selective MISPE with spectrofluorimetry method has been established for selective extraction and direct determination of TEL in human urine.
2.2 Synthesis of MISPE sorbent For the synthesis of MIP, 0.063 mmol of TEL, and 0.186 mmol of DMAEMA were dissolved in 1.0 mL of CHL in a borosilicate tube and incubated for 15 min at room temperature. To this mixture, 1.27 mmol of EGDMA and 5.0 mg of AIBN were added. The tubes were sealed and then cooled in ice. Thereafter, photopolymerization reaction was initiated using a UV lamp (6 W, 365 nm; Upland, CA, USA) and proceeded 5 h. Then, tubes containing MIP sorbent were crushed, and bulk polymer was ground using A11 basic IKA analytical mill (Staufen, Germany), and then wet-sieved to between 20 and 63 m in size using methanol. Later on, sorbent was subjected to the extraction process by using an I˙ ldam Soxhlet extraction apparatus (Ankara, Turkey) using MeOH/AA (9:1, v/v) to remove template molecules for 48 h. Finally, MISPE sorbent was dried in an incubator (Nuve, Ankara, Turkey) at 50⬚C for 6 h. The nonimprinted polymer (NIP) was also prepared as a control polymer using same procedure but in the absence of template molecule.
2.3 Swelling studies Swelling behavior of MIP/NIP sorbents was investigated through incubation studies in aqueous (pH 3.5–8.5) and organic (CHL, ACN, TOL, DCE) conditions. Typically, 50 mg MIP/NIP particles was weighed into an empty SPE cartridge (Alltech 1.5 mL, Deerfield, IL, USA). A certain amount of a solvent was added into the cartridge and sorbent was allowed to swell for 9 h until it reaches a constant weight. After removing the excess solvent, cartridge containing the sorbent was weighed. Using Eq. (1), swelling was calculated: Swelling (%) = [(Hswollen − Hdry )/Hdry ] × 100
(1)
where Hswollen is the weight of swollen sorbent and Hdry is the weight of the dry sorbent.
2.4 Batch rebinding studies
2 Materials and methods 2.1 Reagents Telmisartan (TEL) was kindly donated by NUVOMED (Istanbul, Turkey) and Mustafa Nevzat (Istanbul, Turkey). Valsartan (VAL) was provided by Actavis Incorporation (Malta). DMAEMA, EGDMA, and AIBN were purchased from Sigma–Aldrich (Steinheim, Germany). Dichloroethane (DCE), methanol (MeOH), CHL, toluene (TOL), acetonitrile (ACN), and acetic acid (AA) were obtained from Merck (Darmstadt, Germany). All other chemicals were of analytical reagent grade. C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
20 mg of the dry sorbent (MIP or NIP) was placed into a glass vial containing 4.0 mL of 5.0 g/mL TEL solution. The mixture was stirred for 1.5 h and centrifuged at 4000 rpm for 30 min. Then, supernatant was analyzed using a spectrofluorimeter (Varian Cary Eclipse, Mulgrave, Victoria, Australia; ex : 302 nm, em : 362 nm, slit widths: 5 nm) for the determination of free TEL in the binding medium. Amount of TEL adsorbed by MIP and NIP sorbents was found by subtracting amount of free TEL in the solution from the initial amount in the binding medium. For the determination of appropriate batch rebinding conditions, type of incubation solvent (DCE, ACN, CHL, and TOL), incubation time (30–180 min), amount of sorbent (5–40 mg), and initial concentration of TEL (1.0–15.0 g/mL) were optimized. www.jss-journal.com
Other Techniques
J. Sep. Sci. 2015, 38, 1433–1439
The equilibrium-binding capacities of MIP and NIP sorbents were determined using Eq. (2): Qe = (Ci − Cf ) × V /m
(2)
where Ci (g/mL) and Cf (g/mL) are the initial and final concentrations of TEL, V (mL) is the volume of incubation solvent, and m (g) is the mass of the sorbent. Langmuir and Freundlich [20, 21] isotherm models were applied to the MIP sorbent using Eqs. (3) and (4), respectively: Ce /Qe = (1/b Qo ) + (Ce /Qo ) LogQe = LogK F + (1/n) LogCe
(3) (4)
where Qe and Qo (mg/g) are the amounts of sorbate bound to per unit mass of sorbent at equilibrium and saturation, b (L/mg) is the Langmuir constant, Ce (mg/L) is the equilibrium concentration of sorbate, KF (L/mg) is the Freundlich constant, n is Freundlich exponent. KF is correlated to the maximum adsorption capacity. During the selectivity study, 20 mg of MIP/NIP sorbents was incubated in 4.0 mL of 5.0 g/mL TEL and VAL solutions, separately.
2.5 MISPE protocol Fifty milligrams of MIP/NIP particles was dry-packed into 1.5 mL SPE cartridge between two polyethylene frits and connected to a vacuum manifold (Alltech vacuum manifold, Deerfield, IL, USA). The cartridge was first conditioned with 2 × 0.5 mL of loading solvent (ACN or pH 7.4 distilled water). Then, 4 × 0.5 mL of 10 g/mL TEL solution prepared in ACN or distilled water (pH = 7.4) was loaded into the cartridge. Thereafter, cartridge was washed with 6 × 0.5 mL of CAN and TEL was eluted from the cartridge using 8 × 0.5 mL of MeOH/AA (9:1, v/v). Amount of TEL in the eluate was determined spectrofluorimetrically by referring to calibration curve ranging from 0.04 to 1.40 g/mL. Before the new extraction process, cartridges were regenerated using 6 × 0.6 mL of MeOH/AA (9:1, v/v).
2.6 Analysis of human urine Blank urine samples were collected from a healthy volunteer and stored in a refrigerator at –20⬚C till use. Initially, human urine samples were centrifuged at 3000 rpm for 30 min and then, supernatant was filtered through 0.45 m millipore nylon membranes. pH of the filtrate was adjusted to 7.4 with 1.0 M NaOH. Stock standard solution of TEL, 1.0 mg/mL, was prepared in ACN. Working standard solutions were obtained by adding appropriate volumes of stock solution into 10.0 mL volumetric flasks and then were diluted to the mark with the blank human urine (pH 7.4). A total of 4 × 0.5 mL of each standard solution was then passed through C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1435
MIP and NIP cartridges. After applying washing and elution steps as mentioned in Section 2.5, matrix-matched calibration curve in the range of 0.04 to 1.40 g/mL was generated. For the recovery and precision studies, urine samples (2.0 mL, pH 7.4) spiked with 0.10–0.80 g/mL TEL solutions were employed.
3 Results and discussion 3.1 Preparation of MISPE sorbent In this study, a novel MISPE sorbent was synthesized by noncovalent molecular imprinting approach for the selective extraction of TEL from human urine. Since an optimal functional monomer and template interaction is of prime importance in the formation of high affinity recognition sites for the target analyte, MIP formulation was carefully optimized and ideal molar ratio of TEL, DMAEMA, and EGDMA was found to be 1:3:20 (Supporting Information Table S1). To increase number of prepolymerization complex between DMAEMA and TEL, they were allowed to stay together for 15 min before adding cross-linker and initiator into the reaction medium. Due to presence of an acidic carboxyl group in TEL and a basic tertiary amine group in DMAEMA, there was a highly efficient and effective interaction between functional monomer and template molecules during molecular imprinting process, thus resulting in a very high imprinting factor (IF = 16.1). It was concluded that molecular interaction during molecular imprinting was mostly due to hydrogen bonding between carboxyl group of TEL and tertiary amine group of DMAEMA. Furthermore, high IF value clearly indicated that highly selective TEL imprinted recognition sites were successfully generated in MIP particles. In noncovalent imprinting technique, MIPs should be prepared in low-polarity organic solvents to avoid the disruption of weak intermolecular interactions, such as hydrogen bonding, electrostatic, and van der Waals interactions, between the functional monomer and template molecule in the prepolymerization mixture. Hence, CHL, a low-polarity and aprotic solvent, was selected as a porogen to enhance the strength of hydrogen-bonding interactions present between TEL and DMAEMA. The effect of porogen volume (0.5– 1.5 mL) was also tested and optimum value was determined to be 1.0 mL (Supporting Information Table S1). During the MIP synthesis, photochemical free-radical polymerization at ambient and cooled media were applied because it is wellknown that polymers prepared at higher temperatures (e.g. thermal polymerization) exhibit reduced molecular recognition properties. For that reason, polymerization mixture in borosilicate glass tube was cooled on ice to raise the degree of interaction between TEL and DMAEMA by reducing the molecular vibrations. Results of MISPE studies indicated that higher recovery value (81.1% with cooling and 46.6% without cooling, loading solvent was ACN) was obtained when polymerization mixture was cooled with ice. www.jss-journal.com
1436
J. Sep. Sci. 2015, 38, 1433–1439
H. Yılmaz and H. Basan
Table 1. Batch rebinding studies for the MIP and NIP in different incubation solvents
Binding % Solvent
QMIP
QNIP
Imprinting factor (IF)
Dichloroethane Acetonitrile Chloroform Toluen
83.2 94.9 82.1 80.1
9.0 5.9 10.2 9.7
9.2 16.1 8.0 8.3
During the incubation studies, 4.0 mL of 5.0 g/mL TEL solution was used. QMIP , amount adsorbed by MIP; QNIP , amount adsorbed by NIP.
3.2 Swelling study Swelling of MIP sorbent greatly affects its recognition ability because size and shape of imprinted cavity may change significantly, resulting in loss of selectivity. For this reason, swelling studies for MIP and NIP sorbents were performed in various solvents and results were given in Supporting Information Table S2. At lower pH values (pH < 7.4), both MIP and NIP sorbents reached a higher swelling % (52.4–58.4% for MIP). However, at pH ࣙ 7.4, they have lower swelling % values (37–38.1%). This pH-sensitive behavior can be attributed to the protonation of tertiary amine functional group in the DMAEMA monomer at lower pH values, resulting in swelling of polymeric sorbent due to repulsion of positive charge bearing polymeric chains each other. Therefore, it was concluded that higher recognition ability of MIP sorbent would be at higher pH values because of the fact that less size and shape change would occur. Thus, for the aqueous samples, it was decided that TEL loading should be performed at pH 7.4 medium, probably due to expecting less swelling and in turn, higher selectivity. Swelling studies were also investigated in various organic solvents including CHL, ACN, TOL, and DCE. As can be seen from Supporting Information Table S2, the lowest and highest swelling degrees were reached by using ACN and CHL, respectively. In light of these observations, it might be concluded that geometry of imprinted cavities would change less if ACN is selected as the loading solvent.
3.3 Batch rebinding studies Before MISPE studies, MIP and NIP sorbents are generally tested to investigate their target analyte recognition and binding abilities, i.e. imprinting effect. It is apparent from Table 1 that the highest TEL-binding percent and in turn highest IF value were obtained using ACN as an incubation solvent. This is consistent with the swelling study results, stating that less swelling and in turn less change in geometry of MIP occurs in ACN solvent. Subsequently, optimum incubation time and sorbent amount were found to be C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
90 min and 20 mg (results are not shown here), respectively. Incubation time is of great importance because it determines how much analyte would be adsorbed by the sorbent. It should be given enough time for the analyte molecules to fill the imprinted cavities situated on the surface and interior parts of the sorbent. The effect of target analyte, TEL, initial concentration on the MIP/NIP-binding capacity was also investigated, and the results are illustrated in Fig. 1. The maximum-binding capacity for MIP was achieved when 5.0 mg/L initial TEL concentration was used. The binding capacities for MIP and NIP sorbents were found to be 0.97 and 0.092 mg TEL/g sorbent, respectively. Thus, high difference in binding capacities of MIP and NIP particles can be easily attributed to the molecular imprinting effect. Equilibrium adsorption results were evaluated using Langmuir and Freundlich models to estimate adsorption behavior of MIP sorbent as well. The Langmuir model states that maximum adsorption takes place when a saturated monolayer of solute molecules is present on the sorbent surface [20]. On the other hand, the Freundlich model is a multisite adsorption isotherm for heterogeneous surfaces [21]. The adsorption data fitted well in the Langmuir equation with regression coefficient value of 0.9998 compared to 0.4960 obtained using the Freundlich model (Supporting Information Table S3). The b and Qo values were determined from the slope and intercept of isotherm graph to be 0.076 L/mg and 0.9708 mg/g, respectively. Good fitting of experimental data to Langmuir isotherm proves the existence of homogeneous TEL-binding sites on MIP sorbent. Selectivity of MIP and NIP sorbents to TEL was investigated using valsartan (VAL). VAL was chosen due to its molecular structure similar to TEL to some extent and its ability to form hydrogen bonds with functional groups in MIP cavities. In addition, it contains carboxyl and nitrogen functional groups like TEL. As depicted in Supporting Information Table S4, MIP showed significantly higher selectivity or affinity toward TEL molecules compared to VAL molecules as far as IF values were concerned. Almost 10 times higher IF value (IF for TEL is 16.1 and IF for VAL is 1.51) is certainly due to imprinting effect generated by template TEL molecules during MIP synthesis. 3.4 Optimization of MISPE conditions For the determination of trace levels of analytes in biological fluids, it is mostly desirable to have a MISPE protocol providing high level of analyte recovery and extracts free from interfering compounds. For that reason, several parameters including loading solvent type, pH of aqueous loading solvent, loading time, washing solvent, and volume of elution solvent were carefully optimized. Before the loading step, MIP cavities were first activated using loading solvent to maximize the interactions between TEL molecules and functional groups in imprinted cavities. When ACN was used as loading solvent, 81.1 and 29.4% recovery values were obtained for MIP and NIP, respectively. On the other hand, when TOL was used, corresponding values for MIP and NIP were found www.jss-journal.com
Other Techniques
J. Sep. Sci. 2015, 38, 1433–1439
1437
Figure 1. Effect of initial concentration of TEL on binding capacity for MIP and NIP sorbents in batch study.
to be 65.5 and 31.7%, respectively. Thus, ACN was selected as optimum loading solvent for further experiments. Effect of sample loading pH on the recovery of TEL was investigated in the pH range of 3.5–8.5. As shown in Supporting Information Fig. S1, best recovery value was achieved at pH 7.4. This pH-dependent change in recovery can be attributed to the swelling behavior of the sorbent mentioned in Section 3.2. At lower pH values (pH < 7.4), size and shape selective abilities of the sorbent decrease due to swelling of imprinted sites, resulting in diminished recognition ability. At higher pH values (pH ࣙ 7.4), since the sorbent swells less, loss in the recognition ability is relatively low. For this reason, pH 7.4 was specified as optimum sample loading pH for the aqueous samples. It might be concluded that swelling has a great influence on the recognition ability of the MIP because swelling causes size and shape of the imprinted sites of the MIP to change, leading to loss of its memory effect. Effect of loading time was also studied because it determines strength or efficiency of the interaction between analyte and imprinted cavities and also diffusion time from bulk solution to rebinding sites. Two options were tested in that firstly MISPE study was done by directly passing to the washing step right after the sample loading step and 57.4 and 36.3% recovery values were obtained for MIP and NIP, respectively. In the second study, we waited for 3 min after loading TEL solution before passing to the washing step. Thus, TEL molecules found enough time to interact with the functional groups of the imprinted cavities. As a result of these interactions, relatively higher recovery values were attained (80.8.0% for MIP and 31.1% for NIP). To achieve a selective extraction, a clean-up step with an organic solvent is applied before the elution step and this step is more crucial in MISPE applications than in conventional SPE [22]. Thus, washing solvent should suppress the nonspecific interactions without disrupting the selective specific interactions between the imprinted cavities and the target molecule. When ACN was used as the washing solvent, C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
recovery values for MIP and NIP were 81.1 and 29.4%, respectively. On the other hand, when TOL was tried, recovery value for MIP was 72.5% and that for NIP was 37.5%. It can be easily concluded that higher recovery value for MIP and higher specific binding result were obtained when ACN was employed as the washing solvent. Thus, ACN was selected as the most appropriate washing solvent. Moreover, the volume of ACN (1.0–4.0 mL) was also tested. Since the highest recovery value was generated, 3.0 mL of ACN was determined as appropriate washing solvent volume for further studies. Supporting Information Fig. S2 displays the effect of the elution solvent, MeOH/AA (9:1, v/v), volume on the TEL recovery. Due to higher recovery and specific binding values, 4.0 mL elution solvent was determined as appropriate elution solvent volume. As the volume of the elution solvent increases, amount of TEL eluted from the cartridge increases, reaching maximum value when 4.0 mL was used. Since the strength of interactions between imprinted cavities and TEL molecules was very high, 2.0 or 3.0 mL of elution solvent was not enough to elute all of TEL molecules, thus requiring 4.0 mL. Because an acidic elution solvent is used, sorbent and, in turn, imprinted cavities swell to a certain degree, leading to disruption of intermolecular attraction between TEL molecules and functional groups in imprinted cavities. As a result, TEL molecules are loosely held inside the sorbent and easily eluted from the sorbent. The influence of the sorbent amount (25, 50, and 75 mg) packed into the SPE cartridges on TEL recoveries was also studied. The highest recovery value (81.1%) was achieved when 50 mg MISPE sorbent was used.
3.5 Analytical parameters To demonstrate the suitability of the developed MISPE with spectrofluorimetry method for the determination of TEL
www.jss-journal.com
1438
J. Sep. Sci. 2015, 38, 1433–1439
H. Yılmaz and H. Basan
Table 2. Recovery and precision results for the assay of TEL in spiked human urine Recovery (%) (n = 3)
Spiked TEL (g/mL)
Precision Within-day (RSD%, n = 5)
Between-day (RSD%, n = 5)
MIP
NIP
0.10 0.40 0.80
1.6 0.31 0.14
1.31 0.44 0.41
76.1 77.3 79.1
32.1 32.1 32.1
in human urine, analytical parameters including linearity, LOD, LOQ, within-day (repeatability), and between-day (reproducibility) precisions, and accuracy were validated. For the validation of linearity, calibration curve was constructed using optimized MISPE protocol for the blank urine samples spiked with TEL over a concentration range of 0.04– 1.40 g/mL. A relatively high correlation coefficient value (r = 0.9965) indicated the linearity of the calibration curve. LOD and LOQ were calculated using the SD on the concentration of standard solution having the lowest concentration and slope of the calibration curve. It should be stated that LOD (11.0 ng/mL) and LOQ (36.0 ng/mL) values were relatively low compared to those obtained by Mudiam et al. [19]. In their study, before the HPLC analysis, a combination of MISPE and dispersive liquid–liquid microextraction for the selective preconcentration and determination of TEL in rat urine was described, resulting in relatively higher LOD (190 ng/mL) and LOQ (630 ng/mL) values compared to our proposed method. Thus, it can be concluded that the proposed method is much more sensitive and lower levels of TEL can be selectively determined as a result of better clean-up protocol provided by the newly synthesized MIP sorbent. Human urine samples spiked with low (0.10 g/mL), medium (0.40 g/mL), and high concentrations (0.80 g/mL) of TEL were prepared and analyzed for the determination of withinday and between-day precision of the proposed method. The within-day precision was calculated by performing five extractions of independently prepared urine samples spiked with TEL at different concentrations over a day. Between-day precision was assessed by studying five extractions of independently prepared urine samples spiked with TEL at different concentrations in three different days. Results of withinday and between-day precision studies were presented in Table 2 and relatively satisfactory RSD values ranging from 0.14–1.6% for the within-day and 0.41–1.3% for the betweenday studies were achieved. These results prove the precision of the proposed MISPE with spectrofluorimetry method. For recovery studies, three different concentration levels (0.10– 0.80 g/mL) of TEL spiked into urine samples were studied, Table 2 and relatively high recovery values were achieved. The difference between the recoveries of MIP and NIP sorbents clearly indicates the success of molecular imprinting. These recovery values are comparable to the ones obtained by Ferreiros et al. (74%) in that before the HPLC with fluorimetric detection, a clean-up step consisting of a SPE with C8 cartridges was employed for urine samples [23]. In addition, Zhang et al. [11] prepared a polymeric monolithic C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
microextraction using poly(methacrylic acid-ethylene glycol dimethacrylate) in conjunction with CZE and reached 81.7– 84.2% recovery values in human urine. The LOD and LOQ values they found were 15.0 and 50.0 ng/mL, respectively. It can be concluded that the LOD and LOQ values for the proposed MISPE with spectrofluorimetry method are comparable to the values obtained by Zhang et al. Since TEL in complex human urine was selectively retained and cleaned-up using the proposed method, high recovery, and low LOD values were attained without applying complicated sample pretreatment procedures. Furthermore, low LOD and high recovery values in aqueous rebinding conditions revealed that the proposed MISPE sorbent had high water compatibility, making direct application of human urine samples to MISPE cartridges possible.
4 Concluding remarks In this study, a novel MISPE sorbent was synthesized and used in combination with a very simple and cost effective spectrofluorimetry method for the selective extraction and determination of TEL in human urine. A very high imprinting factor value (16.1) confirmed the success of molecular imprinting. The developed MISPE with spectrofluorimetry method combines sensitive and selective features of spectrofluorimetry with the highly selective properties of the molecular imprinting technique and, in turn, uses their synergistic effect in the determination of TEL in relatively complex human urine sample. The spiked TEL recovery values were in the range of 76.1–79.1% for the proposed method. In addition to the high recovery values, selective and effective sample clean-up procedure was developed. Since most of the interfering compounds in human urine were removed, direct determination of TEL in complex human urine sample was achieved. Due to selectivity and sensitivity of the proposed method, trace levels of TEL in biological samples can be easily determined in clinical and routine analyses without using costly instrumentation and tedious sample preparation methods. ¨ The present study was financially supported by OYP department at Gazi University, Turkey. The authors have declared no conflict of interest.
5 References [1] Krishna, P. R. V., Kasturi, J. K., Babu, A. V., Vijay, K., Raman, G. V., Chakravarthy, I. E., Babu, B. H., Int. J. Pharm. Sci. Res. 2005, 4, 3630–3634. [2] Wienen, W, Entzeroth, M., vanMeel, J. C. A., Stangier, J., Busch, U., Ebner, T., Schmid, J., Lehmann, H., Matzek, K., Kempthorne-Rawson, J., Gladigau, V., Hauel, N. H., Cardiovasc. Drug Rev. 2000, 18, 127–154.
www.jss-journal.com
J. Sep. Sci. 2015, 38, 1433–1439
[3] Hempen, C., Schwarz, L. G., Kunz, U., Karst, U., Anal. Chim. Acta 2006, 560, 41–49. [4] Gupta, V. K., Jain, R., Lukram, O., Agarwal, S., Dwivedi, A., Talanta 2011, 83, 709–716.
Other Techniques
1439
[13] Ilango, K., Kumar, P. S. S., Pharm. Meth. 2012, 3, 112–116. [14] Belal, T. S., Mahrous, M. S., Khalek, M. M. A., Daabees, H. G., Khamis, M. M., Luminescence 2014, 29, 893– 900.
[5] Brunetto, M. R., Contreras, Y., Clavijo, S., Torres, D., Delgado, Y., Ovalles, F., Ayala, C., Gallignani, M., Estela, J. M., Martin, V. C., J. Pharm. Biomed. Anal. 2009, 5, 194– 199.
[15] Cagigal, E., Gonzalez, L., Alaso, R. M., Jimenez, R. M., J. Pharm. Biomed. Anal. 2001, 26, 477–486.
[6] Torrealday, N., Gonzalez, L., Jimenez, R. M., Lastra, E. O., J. Pharm. Biomed. Anal. 2003, 32, 847–857.
[17] Lavignac, N., Allender, C. J., Brain, K. R., Anal. Chim. Acta 2004, 510, 139–145.
[7] Nie, J., Zhang, M., Fan, Y, Wen, Y., Xiang, B., Feng, Y. Q., J. Chromatogr. B 2005, 828, 62–69.
[18] Lai, J. P., Chen, F., Sun, H., Fan, L., Liu, G. L., J. Sep. Sci. 2014, 37, 1170–1176.
[8] Nie, J., Zhao, Q., Huang, J., Xiang, B., Feng, Y. Q., J. Sep. Sci. 2006, 29, 650–655.
[19] Mudiam, M. K. R., Chauhan, A., Singh, A. K., Sharma, V. P., Saxena, P. N., Bioanalysis 2013, 5, 847–858.
[9] Hempen, C., Schwarz, L. G., Kunz, U., Karst, U., Anal. Chim. Acta 2006, 560, 35–40.
[20] Langmuir, L., J. Am. Chem. Soc. 1918, 40, 1361–1403.
[10] Chen, B., Liang, Y., Wang, Y., Deng, F. L., Zhou, P., Guo, F., Huang, L., Anal. Chim. Acta 2005, 540, 367–373.
[16] Won, J. C., Song, H. Y., Faiz, A., J. Sep. Sci. 2013, 36, 609–628.
[21] Freundlich, H. M. F., Z. Phys. Chem. 1906, 57(A), 385–470.
[11] Zhang, M., Wei, F., Zhang, Y. F., Nie, J., Feng, Y. Q., J. Chromatogr. A 2006, 1102, 294–301.
[22] Caro, E., Marce, R. M., Borrull, F., Cormack, P. A. G., Sherrington, D. C., TrAC, Trends Anal. Chem. 2006, 25, 143– 154.
[12] Tsvetkova, D. D., Obreshkova, D. P., J. Planar Chromatogr. 2012, 25, 326–330.
[23] Ferreiros, N., Iriarte, G., Alonso, R. M., Jimenez, R. M., Talanta 2007, 73, 748–786.
C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.jss-journal.com
