A novel structural specific creatinine sensing scheme for the determination of the urine creatinine

Biosensors and Bioelectronics 31 (2012) 90–94

Contents lists available at SciVerse ScienceDirect

Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios

A novel structural specific creatinine sensing scheme for the determination of the urine creatinine Chi-Hao Chen, Meng Shan Lin ∗ Department of Chemistry, Tamkang University, Tamsui 25137, Taiwan

a r t i c l e

i n f o

Article history: Received 1 August 2011 Received in revised form 26 September 2011 Accepted 29 September 2011 Available online 6 October 2011 Keywords: Creatinine Enzyme-free Structural specific amperometric scheme

a b s t r a c t In this work, a highly structural dependent amperometric scheme was proposed for the determination of creatinine without enzymatic assistance. The principle of this novel method is based upon the formation of a soluble copper–creatinine complex on the copper electrode surface. Subsequently, an oxidative current from the regeneration of the surface oxide layer is monitored and it is proportional to the concentration of the creatinine. This scheme can be conducted at potential of −0.1 V (vs. Ag/AgCl, 3 M) in phosphate buffer (pH 7). A typical calibration plot from 25 ␮g/dL to 1.5 mg/dL (R2 = 0.997) with a detection limit of 6.8 ␮g/dL (S/N = 3) is achieved. The relative standard deviation of 21 successive injections of 0.2 mg/dL creatinine is 0.018. Under the optimal conditions, the frequently encountered biological interferences at physiological or higher concentration were investigated. Only uric acid revealed an obvious interference (298.1%). However, a Nafion® coated copper plating electrode shows a successful decrement of the interference of the uric acid with slightly decreased sensitivity of creatinine. The feasibility of this scheme for further clinical application is demonstrated by both HPLC and FIA to evaluate the creatinine concentration in a urine sample. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Creatinine is a useful indicator for evaluating one’s kidney function in clinical laboratory. Generally, it maintains a range of 0.6–1.2 mg/dL in the serum independent of dietetic habit. However, the excretive rate in the urine might change from 1 to 2.5 g/day depending on one’s weight and diet (Cha et al., 2001). A ratio of creatinine between urine collected from 24 h period and serum is an important value to estimate the glomerular filtration rate (GFR), an indicator of one’s kidney function (Jacobi et al., 2008). Typically, a GFR level less than 60 mL/min/1.73 m2 represents the function of kidney might be less than 50% (Herget-Rosenthal et al., 2007; Yaturu et al., 2007). Besides, creatinine is also a useful biomedical marker for the diagnosis of acute myocardial infarction as well as for quantitative description of hemodialysis (Wyss and KaddurahDaouk, 2000). Jaffe reaction has been proposed to determination of the creatinine in the past; this method monitored the complex of the creatinine and picrate via a spectrophotometer. However, the problem of low specificity limits its clinical application. Although several chromatographic schemes have been proposed to improve the selectivity (Patel and George, 1981; Weber and Zanten, 1991), the tedious time-consuming operating procedure makes it unsuitable

∗ Corresponding author. Tel.: +886 2 26215656x2542; fax: +886 2 26299996. E-mail address: [email protected] (M.S. Lin). 0956-5663/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.bios.2011.09.043

for routine analysis. Biosensor seems to be a better scheme of higher selectivity and easy to use in clinical laboratory. Determination of creatinine via enzymatic products such as ammonium, urea, and hydrogen peroxide by using potentiometric or amperometric method are still a popular research field (Kale et al., 2008). For multi-enzyme, tri-enzyme based sensing scheme had been reported in the previous literatures (Killard and Smyth, 2000; Khan and Wernet, 1997). After sequential reacting with creatininase, creatinase, and sarcosine oxidase, an electrochemical active product, hydrogen peroxide, was monitored and which reflected the level of the creatinine. Bi-enzyme scheme through the creatininase and creatinase has been also proposed to monitor the urea with a polyaniline modified electrode (Pandey and Mishra, 2004). However, the aforementioned multiple enzyme based protocols are suffered from the intermediate subtracts, i.e. creatine, and sarcosine. On the other hand, determination of the creatinine by using creatinine deiminase has been reported through the ammonia ion level by using an ammonia ion selective electrode (Koncki et al., 2004; Martelet et al., 2002) and pH-sensitive device (Osaka et al., 2000). Besides, a polypyrrole (Trojanowicz et al., 1996) and polyaniline (Huang and Shih, 1999) based amperometric methods have been also reported to determinate the ammonia ion. However, poor sensitivity and limited linear range are still crucial challenges for clinic application. Several redox-active metal or metal oxide materials have been reported to promote the electrochemical sensing in pH, oxygen and hydrogen peroxide detections (Greenblatt et al., 1996; Yao

C.-H. Chen, M.S. Lin / Biosensors and Bioelectronics 31 (2012) 90–94

et al., 2002; Lin and Shih, 2009). Among them, copper electrode has been shown a particular catalytic property in amino acids (Frei et al., 1983), sugars (Ju et al., 2007) and organic acids (Frei et al., 1984) determination. These reports purposed a surface oxide regeneration process to explain the catalytic property of the copper electrode (Hibbert et al., 1997; Baldwin et al., 1991). Recently, we also have reported a novel structural specific method for selective determination of polyamine by using a copper electrode (Lin et al., 2011). In that previous study, an amperometric method was proposed to monitor an oxidative current induced from regeneration of the oxide layer on an electrode surface after a soluble cupric–polyamine complex formation. In current study, we further extend this concept to develop an enzyme-free creatinine sensing method on a flow injection system. Owing to this special electrochemical sensing strategy, several enzymatic restrictions including low converting rate, crucial operating conditions, and suffered from the immediate products can be avoided. This work reports the detail experimental designs and analytical performances of the proposed scheme and its real application for urine creatinine determination. 2. Materials and methods 2.1. Apparatus Cyclic voltammetric experiments, electrochemical deposition, and amperometric detection in the flow injection analysis (FIA) system were carried out with a CHI 832B electrochemical workstation from CH Instruments (Austin, TX, USA). The flow rate of the FIA system was controlled with a syringe pump (74900 series, ColeParmer Instrument Company, IL, USA), and the prepared sample solutions were loaded into a 50 ␮L loop and injected by an automatic valve (ETMA-CE, VICI AG International). A GBC HPLC system (Gibco, Dandenong, Vic, Australia) with LC 1150 pump couple with Waters ␮Porasil 10 ␮m silica packing normal phase column with 10 ␮L sample loop was utilized in the HPLC analysis. 2.2. Reagents Creatinine (anhydrous form) was obtained from Sigma (St. Louis, USA). All other chemicals and solvents were of analytical grade and purchased from Riedel-deHaën (Seelze, Germany) without further purification. The conductive carbon ink (C10903D14) was obtained from Gwent Electronic Materials Ltd. (Pontypool, UK). 2.3. Fabrication and conservation of the copper plating electrode The copper layer was deposited electrochemically onto the platinum electrode surface (with ˚ = 3 mm) in the flow injection system. Before deposition, the polished platinum electrode was immersed in a 0.1 M H2 SO4 solution and electrochemically cleaned by a scanning potential between +1.0 V and −0.2 V for 20 cycles with a scan rate of 0.1 V/s. The platinum electrode was then transferred into a 0.1 M CuSO4 /0.1 M H2 SO4 solution, and a constant potential of −0.2 V was applied for 60 s. Before each experiment run, the electrode was activated by applying a −1 V in the buffer solution for 60 s (Frei et al., 1982).

91

solution was stirred at 200 rpm. The pre-concentration was then stopped, and after 5 s, the differential pulsed voltammogram was used to monitor the stripping peaks between −0.2 V and 0.4 V. 2.5. Procedure of the real sample test Fresh urine sample was donated by a healthy volunteer. Twofold of the acetone (volume to volume) was added into this fresh urine sample to remove the protein in the sample. After centrifugation, the acetone was removed with decompression. Subsequently, the sample was diluted with buffer solution in the appropriate concentration for subsequent analysis (100 folds in FIA and 10 folds in HPLC dilution, respectively). The diluted sample was subsequently filtrated with a 0.45 ␮m nylon filter before the HPLC injection. 3. Results and discussion 3.1. Mechanism study The selectivity and sensitivity of this novel method is based upon the chelating ability between creatinine and the copper layers rather than the redox behavior of creatinine itself. In the mean time, an apparent oxidative current was obtained after a potential assisted surface oxide regeneration process (PASOR) after converting portion of the cuprous oxide into a soluble Cu–creatinine complex. Cyclic voltammetry was conducted to understand the basic surface reaction between creatinine and copper based electrode. Fig. 1 shows the voltammograms at the presence of blank (a), 5 mg/dL (b), 10 mg/dL (c), 15 mg/dL (d), and 20 mg/dL (e) of creatinine on the copper plated electrode. It is observed that the oxidative current increases proportionally to the concentration of the creatinine but the reductive peak decreases inversely. This result is similar to a typical EC reaction (Bard and Faulkner, 2001), however, creatinine is known as an electrochemical inactive species, i.e. a direct electron transfer between creatinine and electrode or catalyst is not feasible. Thus, these voltammograms might attribute to the PASOR process as our previous report (Lin et al., 2011). In order to proof this hypothesis, a mimic surface dissolving process was conducted in presence of oxygen and the oxygen acts as an oxidizer to provide an oxidative strength as the electrochemical potential scan. Here, the chelating capabilities of cupric oxide, cuprous oxide, and the copper metal to the creatinine were investigated. Briefly, a 50 mg of each powder was incubated in a 10 mL phosphate buffer solution, pH 7, containing 50 mg/dL creatinine, respectively. After well mixing for 20 min, a simple stripping analysis from each solution was adapted to investigate the chelating capability via

2.4. Procedure of stripping analysis In order to identify the chelating abilities of copper, cuprous oxide and cupric oxide, 50 mg of each powder was incubated in the 10 mL, 50 mg/dL creatinine, respectively. After well mixing for 20 min, the solution was filtrated with 0.45 ␮m nylon filter. Subsequently, 0.5 mL filtrations were spiked into 9.5 mL, 0.1 M H2 SO4 , respectively. A pre-concentration potential (−0.2 V) was applied to the glassy carbon electrode (with ˚ = 3 mm) for 120 s, while the

Fig. 1. Typical cyclic voltammograms of creatinine on a copper-deposited electrode in 50 mM phosphate, pH 6.5. Concentration of the creatinine: 0 mg/dL (a), 5 mg/dL (b), 10 mg/dL (c), 15 mg/dL (d), and 20 mg/dL (e). Scan rate: 50 mV/s.

92

C.-H. Chen, M.S. Lin / Biosensors and Bioelectronics 31 (2012) 90–94

Scheme 1. Proposed creatinine measurement mechanism on a copper based electrode.

of electrochemical regeneration of the surface oxide. Fig. 2 shows the statistical signal of three successive injections of 0.5 mg/dL creatinine at various operating potentials from 0 to −0.15 V with an interval of 0.025 V, and the best sensitivity appears at −0.1 V. In the prior section, we have proved that the copper metal and cuprous oxide has better complex affinity then cupric oxide, thus, the decreased sensitivity in higher operating potential range might attributed to cupric oxide accumulation. In contrast to higher potential range, a decreased trend in the lower operating potential might result from slower surface oxide regeneration rate. Moreover, this operating potential does not suffer from most biological antioxidants and heavy metal ions, therefore, −0.1 V was chosen as optimized potential. Subsequently, the acidity is investigated, Fig. S1 shows the influence of the pH from 6 to 8 with an interval of 0.5 U, and the best sensitivity appears at pH 7. Because the

150 120

Current (nA)

the formation rate of soluble cupric–creatinine complex. The current signal of copper powder and cuprous oxide are 2.13 ± 0.21 ␮A and 1.16 ± 0.16 ␮A, respectively, but no detectable signal can be found in either the cupric oxide or any buffer incubated solutions. Although a stable crystal structure of cupric–creatinine complex has been reported (Niclos-Gutierrez et al., 2005), our result indicates that only the copper metal and cuprous oxide powder can provide an effective cupric ion with creatinine to form the cupric–creatinine complex. Besides, there is no response between cupric oxide and creatinine might attribute to the slow kinetic rate of direct converting the cupric oxide into creatinine complex. It is well known that the coordination number of the cuprous complex is 4, but 4, 5, or 6 for cupric one (Zelewsky, 1996). This report implies higher valence state of the copper based materials require more ligands to form a stable complex. Thus, the higher formation rate on cuprous oxide and copper metal powders might attribute to the generation of another empty molecular orbital for creatinine to form the complex in the oxidative process of copper or cuprous oxide. This assumption indicates a suitable operating voltage is required to control the electrode surface at copper/cuprous oxide co-existed state. Besides, because portion of the oxide layer was dissolved in the complexation process, the decreasing reductive peak observed in Fig. 1 might result from converting portion of the surface cuprous oxide into cupric–creatinine complex. According to the prior discussion, a proposed mechanism is shown in Scheme 1.

90 60 30

3.2. Optimization of the FIA

0 After investigation of the basic electrochemical behavior, this novel creatinine sensing scheme was optimized on a FIA system. Several parameters including potential, pH, buffer concentration, sample volume, and flow rate were carefully studied. The applied potential was optimized firstly since this parameter affects not only the oxide composition of the electrode surface but the rate

-150 -125 -100

-75

-50

-25

0

Potential (mV) Fig. 2. The influences of potential on the response of the copper based creatinine sensor: 50 ␮L sample solution containing 0.5 mg/dL creatinine was injected in to the FIA system. The buffer is 62.5 mM PBS pH 6.5, and the flow rate was kept at 1.0 mL/min.

C.-H. Chen, M.S. Lin / Biosensors and Bioelectronics 31 (2012) 90–94

indicates a less interferences and higher accuracy for creatinine determination.

800

3.4. Interference study

600

Current (nA)

93

200

pka s of the creatinine are 4.8 and 9.2, respectively, the creatinine might become positive charge in low pH solution and decrease the chelating behavior. Besides, hydroxyl ion is a well-known ligand for a metal ion. The sensitivity dropped in higher pH might attribute to the higher stability of cupric hydroxide. Moreover, pH 7.0 is very close to a biological fluid, thus, we choose this acidity as the optimal condition. The flow rate and the sample volume were also investigated, after evaluation of the sensitivity and peak resolution, 1.0 mL/min and 50 ␮L are chosen as the optimal flow rate and sample injection volume for the subsequent analytical performance and real sample studies (see supplementary Figs. S2 and S3).

The goal of the current paper is to provide a simple method for clinical application; not only the sensitivity but also the selectivity should be considered in the design process. The interference of several biological species were investigated and shown in Table 1. Among them, uric acid, acetaminophen and ascorbic acid are the most frequently concerned interferences for an electrochemical sensor. Glucose has been reported as an effective analyte on copper electrode (Song et al., 2009). Creatine and sarcosine are intermediates in the enzyme based biosensor. Ammonium ion, chloride ion and urea are the major constituent in the urine. The concentration of all the interferences used are equal or higher then its physiological condition. The result of Table 1 shows current scheme do not suffered from ascorbic acid as well as several amine based interferences including ammonium, acetaminophen, and sarcosine cause a slight influence (less than 10%). It is noticed that uric acid possesses a significant influence in this scheme (298.1%). However, because the charge of the uric acid and creatinine are negative and partial positive in a pH 7 buffer, respectively, thus, an anionic membrane, Nafion® , was introduced to minimize the influence of uric acid. Briefly, a 3 ␮L Nafion® solution (1%) was dropped on the electrode surface and dried in the 4 ◦ C refrigerator for 1 h. After coating with a Nafion® film on the electrode surface, the response of 1 mg/dL creatinine reduces slightly from 332.1 nA to 291.4 nA, but the interference of the uric acid decreases from 298.1% to 5.0%. This membrane also prevents the effective permeation of those amine based species, and the influences of all the interferences were decreased to less than 5%. Apparently, this approach provides a suitable selectivity for a further clinical application.

3.3. Analytic performances

3.5. Analysis of real sample

After optimization of the major operating conditions, Fig. 3 shows a typical calibration plot of this proposed scheme. A linear range starting from 25 ␮g/dL to 1.5 mg/dL or 1.8 ␮M to 108 ␮M (R2 = 0.997) is obtained. The detection limit is estimated around 6.8 ␮g/dL (S/N = 3). This range is suitable for serum analysis (0.9–1.5 mg/dL) as well as a diluted urine sample. The coefficient of variation of 21 successive detections of 0.2 mg/dL creatinine is only 1.8%, which indicates a good stability of current method. Table S1 summarizes the analytical performance of current scheme alone with other prior creatinine publications. Compared to the previous reports, our scheme possesses an equal or better linear range with simpler experimental preparation, rapid analysis, and cost effective. Besides, this scheme can be conducted in a biological closed acidity with a dramatic operating voltage (−0.1 V), which

The feasibility of this method for real biological application was demonstrated by determination of the creatinine level in a urine sample. In order to obtain a credible value, a bared copper electrode is used as the HPLC detector to define the actual concentration of creatinine in the urine sample by using a standard addition method. The inset graph in Fig. 4 shows the actual chromatograms obtained from addition of 0 mg/dL (a), 25 mg/dL (b), 50 mg/dL (c), and 75 mg/dL (d) (diluted factor had been calculated) creatinine in the ten fold diluted urine sample, respectively. After calculation of the dilution factor, the intercept of line B in Fig. 4 shows a 98.6 mg/dL creatinine level in this urine sample. Good reproducibility of the retention time and suitable peak resolution indicates the accuracy of this simple assay is acceptable. The line A of Fig. 4 shows the same real sample analysis by a Nafion® /Cu electrode on the FIA

Current (nA)

400

200

(a)

(b)

(c)

(d)

100 0 -100 -200 1000

0

1200

1400

1600

1800

Time (s)

0

1000

2000

3000

4000

Concentration (µg/dL) Fig. 3. Typical calibration curve of the Cu based creatinine sensor: the inset diagram is the actual responses of sequential injection of 0.3 mg/dL (a), 0.4 mg/dL (b), 0.5 mg/dL (c), and 0.75 mg/dL (d) creatine.

Table 1 The interference studies of copper modified electrode in the presence and absence of Nafion® membrane. This data are calculated from three successive injections. Analyte

Concentration (mg/dL)

Creatinine Ascorbic acid Acetaminophen Creatine Ammonium Sarcosine Uric acid Uria Chloride Glucose

1 1 1 1 1 1 1 10 10 100

Nafion® coated electrode

Bare electrode Signal (nA) 332.1 11.9 21.7 11.4 30.3 18.6 990.4 20.0 20.3 15.4

± ± ± ± ± ± ± ± ± ±

1.6 1.3 2.1 2.5 3.4 2.6 15.6 1.5 1.8 1.3

Ratio %

Signal (nA)

100 3.6 6.5 3.4 9.1 5.4 298.1 6.0 6.1 4.6

291.4 −12.6 5.3 −10.2 11.2 11.0 14.8 6.3 −5.2 −6.8

± ± ± ± ± ± ± ± ± ±

1.4 2.2 1.4 1.2 3.3 2.3 3.8 1.6 1.8 1.2

Ratio % 100 −4.3 1.8 −3.5 3.8 3.7 5.0 2.1 −1.8 −2.3

94

C.-H. Chen, M.S. Lin / Biosensors and Bioelectronics 31 (2012) 90–94

(A) (B)

500

180

Acknowledgment

150

We are grateful for the financial support from the National Science Council in Taiwan (Grant No. NSC 97-2113-M-032-004-MY3).

90 (a)

100 s

200

60

Current (nA)

120 300

100 nA

Current (nA)

400

Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bios.2011.09.043.

(d)

100

30

0

0 0

Appendix A. Supplementary data

25

50

75

Creatinine addtion (mg/dL) Fig. 4. Standard addition analysis of the real sample by using FIA integrated with Nafion® /Cu electrode (A) and HPLC integrated with Cu electrode (B). Inset: actual chromatograms after addition of 0 mg/dL (a), 25 mg/dL (b), 50 mg/dL (c), and 75 mg/dL (d) (dilution factor has been calculated) creatinine in a 10-fold diluted urine sample, respectively.

system by using similar standard addition method with a diluted factor of 100, and a 96.2 mg/dL creatinine level is obtained. Compared to the analytical result found in HPLC, a suitable correlation (R = 0.980) indicates the Nafion® /Cu electrode on the FIA system possesses a feasibility in the clinical analysis. Besides, this approach also shows this scheme possess a potential to integration with the other commercial separating tools. 4. Conclusions In this work, an indirect oxidative signal from PASOR process is proposed to measure the concentration of creatinine without enzymatic assistance. A suitable operating potential (−0.1 V) provides a significant advantage in avoiding the influence of biological antioxidants as well as the intermediate species including urea, creatine, and sarcosine. Both of the Nafion® modified copper electrode on the FIA system or bare copper electrode coupled with HPLC can provide an accurate, convenient and rapid method in real urine sample determination. Besides, creatinine is a useful diluting factor for clinical diagnosis to evaluate other biological compounds in a urine sample. The relative applications in clinical diagnosis for urine and blood samples are currently developed in our laboratory.

References Baldwin, R.P., Luo, P., Zhang, F., 1991. Anal. Chem. 63, 1702–1707. Bard, A.J., Faulkner, L.R., 2001. Electrochemical Methods: Fundamentals and Applications, 2nd ed. John Wiley & Sons Inc., New York. Cha, G.S., Shin, J.H., Choi, Y.S., Lee, H.J., Choi, S.H., Ha, J., Yoon, I.J., Nam, H., 2001. Anal. Chem. 73, 5965–5971. Frei, R.W., Kok, W.T., Brinkman, U.A.T., 1983. J. Chromatogr. 256, 17–26. Frei, R.W., Kok, W.T., Groenedijk, G., Brinkman, U.A.T., 1984. J. Chromatogr. 315, 271–278. Frei, R.W., Kok, W.T., Hanekamp, H.B., Bos, P., 1982. Anal. Chim. Acta. 142, 31–45. Greenblatt, M., Shuk, P., Ramanujachary, K.V., 1996. Electrochim. Acta 41, 2055–2058. Herget-Rosenthal, S., Bökenkamp, A., Hofmann, W., 2007. Clin. Biochem. 40, 153–161. Hibbert, D.B., Hidayat, A., Alexander, P.W., 1997. Talanta 44, 239–248. Huang, H.J., Shih, Y.T., 1999. Anal. Chim. Acta 392, 143–150. Jacobi, D., Lavigne, C., Halimi, J., Fierrard, H., Andres, C., Couet, C., Maillot, F., 2008. Diabetes Res. Clin. Pract. 80, 102–107. Ju, H., Zhai, C., Li, C., Qiang, W., Lei, J., Yu, X., 2007. Anal. Chem. 79, 9427–9432. Kale, G.M., Lad, U., Khokhar, S., 2008. Anal. Chem. 80, 7910–7917. Khan, G.F., Wernet, W., 1997. Anal. Chim. Acta 351, 151–158. Killard, A.J., Smyth, M.R., 2000. Trends Biotechnol. 18, 433–437. ˛ S., 2004. Talanta 64, 603–608. Koncki, R., Radomska, A., Bodenszac, E., Glab, Lin, M.S., Chen, C.H., Chen, Z., 2011. Electrochim. Acta 56, 1069–1075. Lin, M.S., Shih, W.C., 2009. Biosens. Bioelectron. 24, 1679–1684. Martelet, C., Soldatkin, A.P., Montoriol, J., Sant, W., Jaffrezic-Renault, N., 2002. Mater. Sci. Eng. C 21, 75–79. Niclos-Gutierrez, J., Tribet, M., Covelo, B.A., Sicilia-Zafra, G., Navarrete-Casas, R., Choquesillo-Lazarte, D., Gonza’lez-Pe’rez, J.M., Castineiras, A., 2005. J. Inorg. Biochem. 99, 1424–1432. Osaka, T., Komaba, S., Amano, A., Fujino, Y., Mori, H., 2000. Sens. Actuators B: Chem. 65, 58–63. Pandey, P.C., Mishra, A.P., 2004. Sens. Actuators B: Chem. 99, 230–235. Patel, C.P., George, R.C., 1981. Anal. Chem. 53, 734–735. Song, W., Wang, W., Zhang, L., Tong, S., Li, X., 2009. Biosens. Bioelectron. 25, 708–714. Trojanowicz, M., Lewenstam, A., Krawczynski vel Krawczyk, T., Lahdesmaki, F., Szczepek, W., 1996. Electroanalysis 8, 233–243. Weber, J.A., Zanten, A.P., 1991. Clin. Chem. 37, 695–700. Wyss, M., Kaddurah-Daouk, R., 2000. Phys. Rev. 80, 1107–1213. Yao, S., Wang, M., Madou, M., 2002. Sens. Actuators B: Chem. 81, 313–315. Yaturu, S., Reddy, R.D., Rains, J., Jain, S.K., 2007. Cytokine 37, 1–5. Zelewsky, A.V., 1996. Stereochemistry of Coordination Compounds. John Wiley & Sons Inc., New York.