https://doi.org/10.1007/s11694-018-9904-3 ORIGINAL PAPER
Gadolinium chelate monomer based memories onto QCM electrodes
for folic acid detection in commercial follow-on baby milk
Aytaç Gültekin1 · Gamze Karanfil Celep1 · Rıdvan Say2
Received: 26 October 2017 / Accepted: 13 August 2018 / Published online: 18 August 2018 © Springer Science+Business Media, LLC, part of Springer Nature 2018
Abstract
A novel folic acid-imprinted quartz crystal microbalance (QCM) nanosensor in order to detect folic acid depending on the fabrication of folic acid-imprinted polymer film on a QCM electrode was developed with methacrylamidoantipyrine–gado-linium (III) [MAAP–Gd(III)] used as a monomer for metal coordination–chelation interaction to obtain a more selective molecularly imprinted polymers (MIPs). The developed folic acid-imprinted nanosensor on the MIP/QCM detection system showed outstanding properties such as a stronger affinity of 3.07 × 107 M−1, a short response time of 10 min, wider linear range of 0–100 µM, a selectivity of k = 8.25 and lower detection limit of 0.0080 µM. After characterizing its all features, the newly designed folic acid-imprinted QCM nanosensor was utilized to detect folic acid level in commercial follow-on baby milk in order to determine its use in real samples.
Keywords Folic acid · Molecularly imprinted polymers · Quartz crystal microbalance · Metal-chelate monomer
Introduction
Folic acid (FA; C19H29N7O6, MW = 441.4 g/mol) belong to the family of compounds with chemical structures that are similar to those of folic acid, also known as folate. FA is a water soluble vitamin with significant functions in the human body including formation of DNA and RNA, syn-thesis of red blood cells, development of the body tissues and brain in foetus, promotion of growth in children and the protection against various pediatric tumours [1, 2].
The human body is not able to synthesize FA hence should be sourced from the diets we eat (cite). Based on the importance, to determine the concentrations of FA in vari-ous food formulations such as the nutritional supplements and infant food is an interesting area of research. From the literature, various analytical methods have been used in the detection of FA, for example, liquid chromatography (LC) [3], capillary electrophoresis [4], high-performance LC [5], fluorimetric detection [6] and mass spectrometry [7]. Other
alternatives methods such as the use of electrochemical and bio-sensors have also been reported [8–14]. Despite the considerable efforts in the development of various methods in detection of FA, problems of coming up with a quicker, cheaper, and simpler method still remains unresolved.
Molecular imprinting is a method used to create tem-plate-shaped cavities in macromolecular matrices [15]. Target molecules are used as templates for imprinting the cross-linked polymers. After removing the template, the residual polymer is more selective although this selectivity is affected by factors such as the size and shape of the cavity and rebinding interactions. In order to regulate the functional monomers around the template, covalent interactions [16], non-covalent interactions [17], electrostatic interactions [18] and metal ion coordination [19] are used. In the way of specificity, strength and directionality, the interaction of metal coordination is analogous to a covalent interactions instead of electrostatic or hydrogen bonding in water [20]. The molecularly imprinted polymers (MIPs) are durable and steady synthetic receivers for the analyte of interest. The production of MIPs to perceive specific compounds via quartz crystal microbalance (QCM), which is extremely sensitive to the nanogram level of mass changing weighted down the surface of the QCM electrode, has been achieved in recent years [21–26] and the procedure is rapid, low cost, easy to use, highly specific and selective.
* Aytaç Gültekin aysari@yahoo.com
1 Department of Energy Systems Engineering, Faculty of Engineering, Karamanoğlu Mehmetbey University, 70200 Karaman, Turkey
In this work, an easy-to-use, cost-effective, and sensitive method to create FA selector memories created on selective QCM electrode surface using MIPs as the recognition mate-rial was developed with a combined molecular imprint and methacryloyl antipyrine (MAAP) to chelate the metal ions and generate an active centrum of phosphotriesterase. Cova-lent molecularly imprinted polymers were synthesised using MAAP–Gd(III) and ethylene glycol dimethacrylate (EDMA) as a metal-chelate monomer and cross-linker, respectively. These were coated on to piezoelectric crystals and the char-acteristics of the binding investigated. Selective adsorption of MIPs was used to detect the FA levels in commercial follow-on baby milk.
Experimental
ReagentsFA, azobisisobutyronitrile (AIBN), allyl mercaptan, tri-ethyamine (TEA), and acetonitrile (ACN) were pro-vided by Sigma Aldrich (Milwaukee, WI, USA). 3,4-Eth-ylenedioxy-N-methylamphetamine (EDMA) was purchased from Fluka (Buchs, Switzerland). Whole other chemicals were obtained from Merck (Darmstadt, Germany). The ultra pure water used in the tests was recieved from a Zeneer Power II water purification system.
Equipment
The topographic characteristics of non-imprinted polymer (NIP) and MIP films on the QCM electrode surface were examined by an atomic-force microscope (AFM) (hpAFM, Nanomagnetics Instruments, Oxford, UK).
In order to practise microgravimetric measurements, an AT-cut, Ti/Au polished, quartz crystals of 5 MHz and a quartz crystal analyser (Standford Research Systems, Model QCM200 Quartz Crystal Microbalance Digital Controller) were run. For the AT-cut shear mode QCM, Sauerbrey’s equation has been designated:
where ∆F is the measured frequency shift by the reason of the added mass in Hertz, F0 is the basic oscillation fre-quency of the dry crystal, ∆m is the surface mass loading in grams, pq is the density of quartz (2.65 g cm− 3), 𝜇q is the shear modulus (2.95 × 1011 dyn cm−2), and A is the elec-trode area (0.19 ± 0.01 cm2). Equation (1) guesses that a frequency change of 1 Hz corresponds to a mass increase of (1) ΔF = −2 F
2 o Δm A(pqμq)1∕2
1.03 ng cm−2 on the electrode for the 5 MHz quartz crystals expended in the present work [26].
Sample preparation
The MAAP monomer, which has a π electron-rich aromatic ring, was synthesised to use in our previous study and was characterised according to previously published procedures [27]. Next, we merged molecular imprinting with the tal-ent of MAAP to chelate metal ions [Gd (III)], which gener-ate decent complexes with FA [28] to ingender an active centre on the polymer. Likewise, a metal-chelate mono-mer, MAAP–Gd(III), was synthesised and characterised according to published procedures [29]. A ligand exchange monomer, methacrylamidoantipyrine–gadolinium (III)/FA (MAAP–Gd(III)–FA) was synthesised using MAAP–Gd(III) and template molecule FA. MAAP–Gd(III) (0.1 mmol) and FA (0.1 mmol) were lyzed in ethanol and the two solutions were intermingled and the ultimate solution was lastly stirred for 24 h. When FA was inserted into the monomer system, a broad OH band (3300–3600 cm−1), which became apparent due to the –OH group of the –COOH functional group, and a peak at 1590 cm−1 due to a C=N stretching vibration in the chemical structure of FA were observed from the Fourier transform infrared (FTIR) spectra. In addition, a carbonyl peak of carboxylic acid in a chemical structure of FA at 1700 cm−1 shifted to 1662 cm−1, which indicated an interac-tion between Gd3+ of MAAP–Gd(III) and FA.
Before covering, a piranha solution (1:3, 30% H2O2/ concentrated H2SO4) was used to purify the gold surfaces. After inserting thiol groups upon the QCM gold surface, solution of allyl mercaptan (2-propane-1-thiol) (0.30 mmol) was dripped into the cleaned surfaces. The allyl mercaptan thiol groups provided the interaction between the gold area of the QCM electrode and imprinted polymer and the allyl group of allyl mercaptan achives the metal–chelate monomer polymerisation from the current site. The QCM electrode was then cleaned with ethyl alcohol and ultra pure water to move away the surplus thiols.
In order to achieve the polymerisation, the reaction solu-tion including the metal–chelate (MAAP–Gd(III)–FA) pre-organised monomer, the crosslinking agent of EDMA, and precursor of AIBN in ethanol was prepared in ethyl alcohol. After that, a driblet of the prepared mixture was drained on the all activated electrode of QCM. Pure nitrogen gas was pumped into the cell for 10 min to deaerate the space com-pletely because the presence of oxygen would have avoided polymerisation. Since FA is a large biomolecule that cannot sustain ultraviolet radiation [30], polymerisation was real-ised at 60 °C and 4 h in an oven (Fig. 1). As a reference, the NIP-coated QCM electrodes (NIPs) were also designed in the same way as MAAP–Gd(III).
Sensor measurements
The FA-imprinted QCM nanosensor was practised for the real-time determining of FA. Solutions containing known amounts of FA were prepared in water/acetonitrile (ACN) (9:1, v/v) [31] solution and the frequency of the nanosensor was watchedup to it stabilised. The frequency shift for all concentration (0–100 µM) of FA was calculated using the equation ΔF = F0 − F1 and the evaluation was realised in trip-licate. Laterevery analysis, ACN/TEA (4:1, v/v) was used to removethe template [31]. The FA-imprinted QCM nanosensor was leached with deionised water after the desorption step was realised. This laundering procedure was iterated until the nanosensor requency equalized to the F0 value.
After the calibration line was created, the commercially available follow-on babymilk was used to determine the unknown amount of FA. The follow-on babymilk in two dif-ferent trade Marks were purchased, diluted 1/100 in water/ace-tonitrile (ACN) (9:1,v/v) solution, and dropped of 128 µL onto QCM electrode. After performing frequency measurements, the amount of FA was determine by using the calibration line and then calculated by taking into account the amount of drop-ping and the ratio of dilution for real level.
Results and discussion
AFM topography studiesThe two- and three-dimensional AFM images, as shown in Fig. 2, depict the surface topography of pure Ti/Au elec-trode and NIP- and MIP-covered gold elecelec-trodes. The sur-face roughness values of pure Ti/Au electrode and NIP- and MIP-coated QCM electrodes were specified to be 0.82 nm, 32 nm, and 126 nm, respectively. The roughness of the sur-face was well dispersed over the entire the electrode sursur-face. This outcome indicated that homogenous FA imprinting on the QCM electrode has been ably executed. This property is one of the important parameters that checks the selectivity and identification speed of the nanosensor [26].
Optimization of analytical parameters
In order to reach optimum conditions for FA determination, the major factors that should be considered are solvents, initiators, crosslinkers and the temperature and time of the polymerisation. One of the most important factors in prepar-ing MIPs is the selection of the crosslinker. Crosslinkers are Fig. 1 Schematic illustration of folic acid molecular imprinting on allyl mercaptan modified Ti/Au QCM electrode
Fig. 2 Two-dimensional (left) and three-dimensional (right) AFM image of a pure Ti/Au electrode(s), b MIP-covered QCM electrodes, and c NIP-covered QCM electrodes. Scanning mode: dynamic; scanning area: 5.0 × 5.0 µm
used to form polymer networks that remember the analyte after removing the template from the polymer. In order to obtain effective imprinting, the concentration ratio of the crosslinker/functional monomer is of capital importance. Different crosslinker/functional monomer concentration ratios cause the formation of various binding sites in MIPs, affecting selectivity. If the molar ratios are too small, the binding sites of the template molecules converge with each other and an effective result for imprinting cannot be obtained. On the other hand, if the molar ratios are too large, the activity of imprinting decrease again as a result of the crosslinkers showing non-covalent interactions with the functional monomers and template molecules. When viewed from this regard, optimising the concentration ratio of the crosslinker/functional monomer becomes significant for effective imprinting. From this point of view, we pre-pared polymerisation mixtures with different crosslinker/ functional monomer ratios (0.1, 0.15, 0.20 and 0.25). Since no polymerisation was observed in the other solutions, the selected crosslinker/functional monomer ratio was 0.25.
Linearity of the calibration line
Five calibration standards for FA were prepared in order to evaluate the relationship between the concentrations of FA and imprinted quartz crystal versus frequency shifts. The lin-earity of the relationship (y = 23.229 × − 27.8) was evaluated over the concentration range 0–100 µM with coefficient (R2) of 0.9911 (Fig. 3).Compared with similar studies in the literature, it is seen that the working range is quite wide [32]. All tests were executed three times and the samples were analysed in three times as well. The detection limit, described as the ana-lyte concentration providing a frequency shift tantamount to three standard deviations of the target, plus the net target fre-quency shift, was 0.0080 µM (3.5 ppb) for the MAAP–Gd(III)-based nanosensor. In the literature, MIP/QCM systems were investigated for FA sensing with 15.4 µM [30, 31], 0.08 µg/L
(0.18 µM) [32] and 30 nM (0.030 µM) [33] detection limits. Therefore, very low detection limits for a new FA-imprinted QCM nanosensor were obtained for this study, as compared with the results of previous similar works.
Measurements of the binding interaction of MIP‑based QCM nanosensors by way of ligand interactions
The ligation of FA to the designed FA-imprinted metal-chelate polymer [MAAP–Gd(III)] on the gold quartz crystal gave rise to a variance in mass, ∆m, that was observed in the crystal frequency. The QCM electrodes were cleaned down deionised water and then made dry. The frequency (F0) was followed in an open area after drying and then the FA solution was distilled into a bounded-type detector unit. The nanosensor fre-quency diminished after distilling the FA solution and follow-ing arrived a fixed value after 10 min (Fig. 4). The reaction achieved equilibrium impetuously, supposing a powerful FA molecular interaction with the imprinted polymer on the QCM electrode. When a NIP was operated, fewer binding of the FA molecules to the NIP was surveyed.
The binding interactivity and equilibrium data between the imprinted polymer and FA template molecule can be achived by Scatchard analysis. Aforesaid analysis is applied by the fol-lowing equation:
where Q is the quantity of FA bound to the polymer, as computed by the mass frequency shiftover the supplemen-tation of the analyte, and C is the concentration of free FA.
(2) Q C = Qmax KD − Q KD y = 23.229x - 27.8 R² = 0.9911 0 20 40 60 80 100 120 140 0 0.01 0.1 1 10 100 Frequency Shi (Hz )
Concentraon of folic acid (µM)
Folic acid imprinted sensorr
Non-imprinted sensor
Fig. 3 Calibration curves of the folic acid imprinted and
Qmaxdescribes the apparent maximum number of binding sites and KD is the equilibrium dispersal constant of the metal-chelate copolymer based on ligand exchange. The results obtained from the linearised form of the Scatchard isotherm by plotting Q/C as a function of Q are:
Therefore, the association constant (Ka) for the bind-ing of FA to the MAAP–Gd(III)-based MIP nanosensor is 3.07 × 107 M−1 and the maximum number of ligand exchange interaction sites, Qmax, is 2.94 nmol. The high value of Ka in-dicates that the affinity of the binding sites is very intensive.
Selectivity studies
In order to understand interactive relation between the bind-ing sites of the MIP-based QCM nanosensor and its tem-plate molecules, and the talent of the imprinted polymer to define the FA molecules, the adsorption of (+)-amethopterin (4-amino-10-methylfolic acid), which is similar in chemical structure to FA, on the imprinted quartz crystal nanosensor was investigated.
The FA-imprinted polymers adsorption capacity for the adsorption of FA plus relevant molecule (+)-amethopterin was surveyed (Fig. 5). According to experimental results, Q values of FA and (+)-amethopterin were determined to be 0.363 and 0.044 mg/g for the MIP-based nanosensor and 0.038 and 0.028 mg/g for the NIP-based nanosensor, respectively. Selectivity coefficients (k) and relative selectiv-ity coefficients (k′) values are listed in Table 1. There have been no other selectivity studies about FA in the literature using the MIP-based QCM technology. Thus, it is tough to (3)
Q
C = −30.735Q + 90.496
contrast this outcome with the literature. As shown in Fig. 4, although (+)-amethopterin has an extremely similar chemi-cal structure to FA, the prepared FA-imprinted nanosensor is eight times more selective to FA than (+)-amethopterin.
Analysis of FA in commercial follow‑on baby milk
The FA-imprinted QCM nanosensor was examined against follow-on milk for babies. Milk specimens was dropped on the prepared FA-imprinted QCM nanosensor. The FA level in the givenspecimens was confirmed by the calibration curve using the frequency values. Table 2 summarises the results obtained for the determination of FA in two follow-up on baby milkpurchased from local supermarkets. The levels of FA detected by the prepared FA-imprinted QCM nanosen-sor were 13.78 and 16.92 (µg FA/100 mL) for different sam-ples of A and B, respectively. Our ability to determine FA level in baby foods using the FA-imprinted QCM nanosensor was determined by comparing our findings with the amounts labelled on the packages of commercial baby foods.
Besides, the high pressure liquid chromatography (HPLC, Shimadzu UV–Vis detector) was used in order to compare the analytical performances of the FA imprinted QCM nanosensor. Chromatographic separation was exe-cuted on a Perkin Elmer C18 (10 µm, 300 × 3.9 mm) col-umn. The amount of folic acid in baby foods using both MAAP–Gd(III) based on the QCM nanosensor and HPLC were given in Table 2. The nanosensor outcomes were vali-dated by relative chromatographic measurements.
All measurements were realised three times at a 95% confidence level. The accuracy of the measurement was controlled with the standard addition method. The standard
Fig. 5 QCM responses of the FA-imprinted nanosensors for FA and (+)-amethopterin (all concentrations are 100 µM)
Table 1 Selectivity of the FA-imprinted QCM nanosensor Q(mg/g)
(imprinted) Q (mg/g) (non-imprinted)
k
(imprinted) k (non-imprinted) k′
FA 0.363 0.038 8.25 1.35 6.11
(+)-Ame- thop-terin
0.044 0.028
Table 2 FA levels in some commercial follow-on baby milk (µg FA/100 mL)
a The standard values listed in Table 2 are the values given on the packages of commercial follow-on baby milk
A B
Standarda 15.00 18.00
FA-imprinted QCM
sensor 13.78 ± 0.026 16.92 ± 0.030
addition method is a quantitative analysis method used to minimize matrix effects that interfere with analyte measur-ing signals. This method is especiallypractical when stand-ard and samples show an alternation in features like ion strength, salinity, viscosity or other types of impurity and interferences. In this method, different amount of standard are directly added to some amount of the sample and the concentration of the analytes is then subtracted via the inter-section of the derived curve with the negative X-axis [34]. It can be said that although the same nanosensor was used 21 times, leaching with ACN/TEA (4:1, v/v), only an insub-stantial (only 5%) measurement effect of the nanosensor was reported (Fig. 6).
Conclusions
The FA-imprinted QCM nanosensor with good reusabil-ity, a wide linear range (0–100 µM), a short response time (10 min), a low detection limit (0.0080 µM), a strong affinity for binding sites (3.07 × 107 M−1), and high selectivity (8.25 for MAAP–Gd (III)–FA complex with respect to (+)-amethop-terin having similar structure with FA) using the molecular imprinting technique with polymerization of metal-chelating monomer [MAAP–Gd(III)] in the presence of a template molecule (FA) was successfully investigated. Besides, in order to find out the congruity of the investigated nanosen-sor in real specimens, the QCM nanosennanosen-sor has been used to determine FA levels in commercial follow-on baby milks. The FA levels were found to be 13.78 ± 0.026 and 16.92 ± 0.030 (µg FA/100 mL) with the FA-imprinted QCM nanosensor and; 15.47 ± 0.008 and 18.54 ± 0.007 (µg FA/100 mL) with HPLC for different follow-on baby milks samples of A and B, respectively.The designed FA-imprinted QCM nanosensor
had a good holding stability and the sensing unit could be developed to be reutilized for long periods of time.
References
1. S. Cheraghi, M.A. Taher, H. Karimi-Maleh, J. Food Comput. Anal.
62, 254 (2017)
2. A.G. Herzog, D.B. MacEachern, H.B. Mandle et al., Epilepsy Behav.
72, 156 (2017)
3. F. Wang, M. Cao, N. Wang et al., Food Chem. 239, 62 (2018) 4. R.H.F. Cheung, J.G. Hughes, P.J. Marriott, D.M. Small, Food Chem.
112, 507 (2009)
5. A. Chaudhary, J. Wang, S. Prabhu, Biomed. Chromatogr. 24, 919 (2009)
6. R.A.S. Lapa, J.L.F.C. Lima, B.F. Reis, J.L.M. Santos, E.A.G. Zagatto, Anal.Chim.Acta 351, 223 (1997)
7. P. Koufopantelis, S. Georgakakou, M. Kazanis et al., J. Chromatogr. B 877, 3850 (2009)
8. S. Cakır, I. Atayman, O. Cakır, Microchim. Acta 126, 237 (1997) 9. S. Wei, F. Zhao, Z. Xu, B. Zeng, Microchim. Acta 152, 285 (2005) 10. C. Wang, C. Li, L. Ting, X. Xu, C. Wang, Microchim. Acta 152, 233
(2006)
11. V.D. Vazeand, A.K. Srivastava, Electrochim. Acta 53, 1713 (2007) 12. M. BostroÈmCaselungheand, J. Lindeberg, Food Chem 70, 523
(2000)
13. F. Gudarzy, A.B. Moghaddam, S. Mozaffari et al., Microchim. Acta
180, 1257 (2013)
14. H. Karimi-Maleh, M. Hatami, R. Moradi et al., Microchim. Acta
183, 2957 (2016)
15. O. Ramstroand, K. Mosbach, Curr. Opin. Chem. Biol. 3, 759 (1999) 16. M.J. Whitcombe, M.E. Rodriquez, P. Villar, E.N. Vulfson, J. Am.
Chem. Soc. 17, 7105 (1995)
17. M.A. Markowitz, G. Deng, B.P. Gaber, Langmuir 16, 6148 (2000) 18. B. Sellergrenand, K.J. Shea, J. Chromatogr. 654, 17 (1993) 19. R. Say, E. Birlik, A. Ersöz, F. Yılmaz, T. Gedikbey, A. Denizli, Anal.
Chim. Acta 480, 251 (2003)
20. R. Say, A. Gültekin, A. Özcan, A. Denizli, A. Ersöz, Anal. Chim. Acta. 640, 82 (2009)
21. T.Y. Lin, C.H. Hu, T.C. Chou, Biosens. Bioelectron. 20, 75 (2004) 22. S. Diltemiz Emir, D. Hür, A. Ersöz, A. Denizli, R. Say, Biosens.
Bioelectron. 25, 599 (2009)
23. A. Gültekin, A. Ersöz, A. Denizli, R. Say, Talanta 93, 364 (2012) 24. A. Gültekin, A. Ersöz, A. Denizli, R. Say, Sens. Actuators B 162,
153 (2012)
25. E. Özkütük Birlik, S. Diltemiz Emir, E. Özalp, T. Gedikbey, A. Ersöz, Mater. Chem. Phys. 139, 107 (2013)
26. A. Gültekin, G. Karanfil, M. Kuş, S. Sönmezoğlu, R. Say, Talanta
119, 533 (2014)
27. A. Ersöz, A. Denizli, I. Sener, A. Atilir, S. Diltemiz, R. Say, Sep. Purif. Technol. 38, 173 (2004)
28. R. Say, Anal. Chim. Acta 579, 74 (2006)
29. B.B. Prasad, R. Madhuri, M.P. Tiwari, P.S. Sharma, Biosens. Bio-electron. 25, 2140 (2010)
30. D.C. Apodaca, R.B. Pernites, R.R. Ponnapati, F.R. Del Mundo, R.C. Advincula, ACS Appl. Mater. Interface 3, 191 (2011)
31. B.B. Prasad, R. Madhuri, M.P. Tiwari, P.S. Sharma, Sens. Actuators. B 146, 321 (2010)
32. R. Madhuri, M.P. Tiwari, D. Kumar, A. Mukharji, B.B. Prasad, Adv. Mater. Lett. 2, 264 (2011)
33. W.A. Henne, D.D. Doorneweerd, J. Lee, P.S. Low, C. Savran, Anal. Chem. 78, 4880 (2006)
34. E. Bertozzini, L. Galluzzi, A. Penna, M. Magnani, J. Microbiol. Met. 87, 17 (2011)
