Determination of Mercury (II) Ion Concentrations in Human Serum Albumin Using Fluorescence Spectroscopy and Multivariate Regression Analysis

Abstract

Human serum albumin (HSA) is a protein responsible for the transportation and delivery of drug molecules, fatty acids, and metabolites to various targets and the removal of waste products from the body. However, the binding of heavy metal ions, including mercury (II) (Hg(II)), can adversely affect HSA physiological properties with deleterious health consequences. The volatility of Hg at room temperature precludes the use of conventional flame, graphite furnace, or inductively couple atomic absorption spectroscopy for routine analysis of Hg in most research and medical laboratories. This study reports the first potential utility of fluorescence spectroscopy and multivariate regression analysis for the determination of Hg(II) concentration in HSA samples. The results of Fourier transform infrared (FT-IR) spectroscopy indicated the binding of Hg(II) at HSA amide I and amide II sites. Also, the binding of Hg(II) with HSA resulted in a decreased HSA ultraviolet (UV)-visible absorption and dramatic quenching of HSA fluorescence emission. The developed multivariate partial least squares regression (PLSR) model from Hg(II)–HSA fluorescence emission data was able to predict the Hg(II) concentration in HSA samples, with a root mean square percent relative error of prediction of 6.59%. The simplicity, low cost, and robustness of this method makes it a promising alternative method for rapid determination of Hg(II) concentration in biological specimens.

Keywords

Human serum albumin HSA Mercury (II) ion concentration Fourier transform infrared spectroscopy FT-IR spectroscopy Fluorescence spectroscopy Multivariate analysis Partial least squares regression PLSR

INTRODUCTION

Human serum albumin (HSA) is an extracellular protein in the blood plasma responsible for the transportation and delivery of fatty acids, drug molecules, and metabolites to different targets in the body.¹ HSA protein has a molecular weight of 66 000 Dalton, contains 585 amino acids with 17 disulfide bridges, and is approximately 60% α-helix with no β-strand.^2,3 As a transportation medium, HSA can potentially bind with small molecules, drugs, metabolites, dyes, and heavy metal ions.^4,–8 The binding of small molecules and heavy metals with HSA is well known to have a considerable influence on HSA physiological characteristics and may also induce HSA conformational changes.^4,–8 Most importantly, elevated heavy metal concentrations in humans have been implicated in various health hazards, including reduction of child intelligence quotients, lung cancer, headaches, hypertension, depression, mental disorders, cancer diseases, and neurological disorders. ^9,–12

Mercury (Hg) is a known toxic heavy metal contaminant, causing severe health problems including autism and autism spectrum disorder, diabetes, lung cancer, kidney and renal failure, oxidative DNA damage, hypertension, cardiovascular disease, stroke, and neurological disorders at chronic and/or acute concentrations.^13,–19 Humans can potentially be exposed to Hg through various routes including occupational, environmental (air, water, and soil), foods, or industrial activity.^20,–28 The harmful health consequences of elevated Hg concentration in humans necessitate the need for the development of analytical techniques capable of Hg analysis in complex biological specimens such as HSA. Conventional flame atomic absorption spectroscopy (FAAS), graphite furnace atomic absorption spectroscopy (GF-AAS), and inductively couple plasma atomic absorption spectroscopy (ICP-AAS) have been well developed for heavy-metal analysis. However, the high volatility of Hg at room temperature precludes the use of conventional FAAS, GF-AAS, or ICP-AAS for Hg analysis. Analysis of Hg is often achieved using a more specialized and relatively expensive cold vapor atomic absorption and/or cold vapor atomic fluorescence technique.^29,–33 Other analytical methods such as high-performance liquid chromatography (HPLC), capillary electrophoresis (CE), chromatographic separation with fluorescence detection, electroanalytical techniques, and nanoparticle chemical sensing have also been strategically employed for Hg analysis.^34,–40 However, the high cost of cold vapor instruments hinder routine Hg analysis in most research and medical laboratories. Chromatographic and CE separation techniques suffer from high instrumental cost, poor resolution, and long analysis times. Other methods may involve tedious and lengthy organic synthesis.

Considering the toxicity and severity of the health implications of Hg contaminants, there is an urgent need for the development of a simple, accurate, and low-cost protocol for rapid Hg analysis in biological specimens. This study explored the first possible use of fluorescence spectroscopy, guest-host HSA chemistry, and multivariate regression analysis for the determination of Hg(II) concentrations in human serum albumin. The simplicity and low cost of spectrometers and multivariate regression analysis of spectral data has recently made it increasingly attractive for rapid determination of various analytes in agricultural products, pharmaceuticals, and environmental samples.^41,–45 The practical applications of multivariate regression analysis of spectral data for the determination of diverse analytes, metabolites, and for fast screening of cancer cells with little or no sample preparation have also been well demonstrated in biomedical studies and clinical diagnoses.^46,–51 Besides, the multivariate regression approach to chemical analysis is not only rapid, sensitive, accurate, and inexpensive, but it is also robust. Once the calibration model is developed and carefully optimized, it can be used for accurate determination of analytes in samples for six months without recalibration,⁵² considerably reducing the time and the cost of chemical analysis.

EXPERIMENTAL

Chemicals and Supplies. MOPs buffer and lyophilized human serum albumin (HSA) were purchased from Sigma-Aldrich (St. Louis, MO, USA) and used as received. HPLC water and sodium hydroxide were obtained from Fisher (NJ, USA). Mercury (II) nitrate standard was purchased from Fluka (St. Louis, MO, USA). Quartz fluorocells were purchased from Starna Cells Inc. (Atascadero, CA, USA).

Sample Preparation. All the glassware was thoroughly washed in nitric acid and rinsed with a copious amount of deionized water before use. The pH meter was properly calibrated with standard buffers before use. A 30 mM MOPs buffer was made in HPLC water and adjusted to pH 7.4 using 0.1 M NaOH solution. Stock HSA and Hg(II) solutions were made in 30 mM MOPs, pH 7.4. HSA was first reconstituted in 30 mM MOPs buffer and refrigerated for 48 hours before use. Sample solution containing a fixed concentration (2 × 10⁻⁵ M) HSA and varying Hg(II) concentrations ranging from 1 × 10⁻⁴ M to 6.5 × 10⁻⁴ M were prepared in MOPs buffer. Accurate HSA concentrations were determined from the ultraviolet (UV)-visible absorption spectra of HSA solution. The Hg(II)–HSA samples were equilibrated for an hour before the Fourier transform infrared (FT-IR), UV-visible, or fluorescence spectra of samples were recorded to ensure the binding of Hg(II) with HSA. Necessary biohazard and laboratory safety precautions, including the wearing of gloves, protective laboratory coats, and eye goggles, were strictly enforced during sample preparation and chemical analysis.

Instrumental and Multivariate Data Analysis. All FT-IR spectral measurements were recorded in the absorbance mode using an FT-IR spectrometer (Shimadzu 8000S). The spectrometer was equipped with a MiRacle ZnSe 3B crystal plate attenuated total reflection (ATR) device mounted on a Shimadzu MiRacle platform. The MiRacle ZnSe 3B crystal plate permits the rapid and accurate measurement of samples with high sensitivity. The FT-IR spectrum of each sample was scanned 100 times with a resolution of 4 cm⁻¹ over a 400 cm⁻¹ to 4000 cm⁻¹ wavenumber range. The average of the 100 scans was used for spectral data analysis. Background FT-IR spectra were always measured before FT-IR measurement of the samples. The FT-IR spectral analysis was performed using IR-resolution spectral analysis software (IR-resolution, Shimadzu Inc.) available on the FT-IR instrument. The FT-IR spectrometer was always calibrated with a polystyrene standard before each use to ensure wavelength accuracy. The UV-visible spectra of the samples were recorded using a diode array UV-spectrometer (Agilent, 8453). The fluorescence emission of samples was measured using a double excitation and emission monochromator spectrofluorometer (Spex Florolog-3, model FL3-22TAU3; Jobin Yvon, Edison, NJ) equipped with a photomultiplier tube (PMT) detector. All fluorescence measurements were recorded using an excitation monochromator and an emission monochromator bandwidth of 5 nm, with excitation wavelength of 280 nm. The absorbance and fluorescence emission of the samples were recorded in a 1 cm path length quartz cell. The solution was always capped to prevent possible quenching of the fluorescence signal from atmospheric oxygen. Routine spectrometer checks and calibrations were performed before spectral measurements and all samples were blank subtracted. Multivariate regression analyses of Hg(II)-HSA fluorescence emission data were performed using chemometrics software (The Unscramble, 9.8, CAMO Inc., NY, USA).

RESULTS AND DISCUSSION

Fourier transform infrared spectroscopy was initially used to evaluate the HSA binding sites with Hg(II). As expected, the FT-IR spectrum of Hg(II) was transparent and featureless. Figure 1 shows the FT-IR spectrum of HSA and the resulting Hg(II)–HSA complexes. HSA shows the characteristic amide I (1653 cm⁻¹) and amide II (1541 cm⁻¹) peaks corresponding to carbonyl (C=O) and amine (C–N) of amino acid residues of HSA.^8,53 The binding of Hg(II) with HSA resulted in a significant reduction of the amide I and amide II peaks, indicating that Hg(II) binds with HSA at these two binding sites. Mercury has strong affinity for thiol (sulfur) containing groups. Therefore, Hg(II)–HSA complexation can potentially occur through metal–sulfur linkage with involvement of C=O and/or C–N groups of amide I and amide II residues of HSA molecules. Other factors, including a charge transfer, may also influence Hg(II)–HSA complexation.

Fig. 1.

FT-IR spectra of HSA and Hg(II)-HSA complexes.

Figure 2 shows the UV-visible spectra of samples containing a fixed concentration (2 × 10⁻⁵ M) of HSA and varying Hg(II) concentration. HSA showed the characteristic HSA absorption λ_max at 280 nm, with a molar absorption coefficient of 35 219 M⁻¹cm⁻¹.² The absorption at 280 nm is a result of an π→π* absorption transition of the lone pair electrons of the carbonyl and amine group of HSA tryptophan residue.^2,54 The absorbance of HSA progressively decreased with an increase in Hg(II) concentrations in HSA samples.

Fig. 2.

UV-visible spectra of 2 × 10⁻⁵ M HSA and HSA solutions of varying Hg(II) concentrations ranging between 1 × 10⁻⁴ M Hg and 6.5 × 10⁻⁴ M Hg.

Figure 3 shows the emission spectra of samples containing a 2 × 0⁻⁵ M HSA concentration and varying Hg(II) concentrations. HSA is a highly fluorescent molecule because of its tryptophan residue, with an emission λ_max at 350 nm. As expected, Hg(II) has no chromophore or fluorophore; therefore, Hg(II) showed no UV-visible absorption or fluorescent emission. The emission of HSA was notably quenched with an increase in Hg(II) concentration. It is of significant interest to note that the binding of Hg(II) with HSA resulted in a gradual blue shift of HSA emission λ_max towards shorter wavelengths. The blue shift became more pronounced with increasing Hg(II) concentrations. The binding of Hg(II) with HSA obviously resulted in significant quenching of HSA emission. For instance, 2 × 10⁻⁴ M Hg(II) quenched HSA emission by approximately 20%. At 6 × 10⁻⁴ M Hg(II), nearly 45% quenching of HSA emission was observed for Hg(II). The binding of Hg(II) with HSA may result in protein agglomeration and induce HSA conformational or stereochemical changes. Human serum albumin protein agglomeration results in poor HSA solubility and/or precipitation in solution. Insolubility of HSA in solution will hinder effective transportation and delivery of necessary materials to the target. Also, changes in HSA conformation may influence HSA secondary structure and HSA stereochemistry. Changes in HSA stereochemistry may possibly hinder HSA chiral recognition ability and negate effective HSA binding of chiral metabolites or chiral drugs, ultimately affecting drugs utilization and drug therapeutic efficacy in humans.

Fig. 3.

Emission spectra of HSA and HSA solutions of varying Hg(II) concentrations: 1, 2 × 10⁻⁵ M HSA; 2, 1 × 10⁻⁴ M Hg; 3, 1.25 × 10⁻¹ M Hg; 4, 1.5 × 10⁻⁴ M Hg, 5, 1.75 × 10⁻⁴ M Hg; 6, 2 × 10⁻⁴ M Hg; 7, 2.5 × 10⁻⁴ M Hg; 8, 3 × 10⁻⁴ M Hg; 9, 3.5 × 10⁻⁴ M Hg; 10, 4 × 10⁻⁴ M Hg; 11, 4.5 × 10⁻⁴ M Hg; 12, 5 × 10⁻⁴ M Hg; 13, 5.5 × 10⁻⁴ M Hg; 14, 6 × 10⁻⁴ M Hg; 15, 6.5 × 10⁻⁴ M Hg (excitation λ, 280 nm).

Multivariate Analysis Determination of Hg(II) Ion Concentrations from HSA Samples. The high volatility of Hg at room temperature precludes the use of conventional AAS for Hg analysis. The matrices of biological specimens are often complex. The utility of a traditional univariate regression analysis approach (spectra analysis at a single wavelength, typically at λ_max) for chemical analysis may be severely hindered in such biological specimen matrices without prior and often tedious sample extraction. Other analytical challenges including overlapping of analyte spectra with other components in biological sample matrices may further preclude the effective use of a univariate regression analysis without tedious analyte spectra deconvolution procedures. As noted earlier, the binding of Hg(II) with HSA resulted in a significant blue shift of HSA emission λ_max. Such shift in HSA emission λ_max prevents the possible use of univariate regression analysis of the fluorescent data for sample calibration. These challenges can be eliminated by the use of multivariate regression analysis.

Therefore, the determination of Hg(II) concentration in human serum albumin by multivariate regression analysis of fluorescence emission of Hg(II)–HSA complexes was explored. To achieve this goal, multivariate partial least squares (PLS-1) regression analysis was used to correlate changes in fluorescence emission of Hg(II)–HSA complexes in Fig. 3 with Hg(II) concentration in HSA samples. Detailed mathematical descriptions of multivariate regression calibrations in analytical spectroscopy for chemical analysis have been reported elsewhere.^52,58,–61 In brief, a multivariate regression equation can be simplified and represented by Eq. 1:

y = b_{0} + x_{1} b_{1} + x_{2} b_{2} + x_{3} b_{3} + ……… x_{n} b_{n}

(1)

where, y is the dependent variable [Hg(II) concentration in this study], x₁, x₂,…, x_n are the independent variables (emission intensity at various wavelengths in this study), b0 is the intercept of the regression equation, and b₁, b₂,…, b_n are the regression coefficients of x-variables. Equation 1 can be expressed in matrix notation as shown in Eq. 2.

[Y] = [X] b

(2)

where, [Y] contains the matrix values of the dependent variables for all samples, [X] is a matrix composed of values of the independent variables of all samples, and b contains the regression vector. In other words, the regression vector relates the independent and dependent variables. The goal in developing any regression model is to first determine the values of the regression vector using a data set of known independent and dependent variables, a process known as regression or model calibration. Once the value of the regression vector is established from the regression equation in the calibration phase, the value of the regression vector can then be combined with the independent variables of future unknown samples to predict the dependent variables of the future samples. This process is known as regression or model validation.

The first critical step in any multivariate regression analysis involves the removal of colinearity in the spectral data. Spectral data colinearity is typically eliminated using modern principal component analysis (PCA), where the original spectral data set is transformed and transposed to a new orthogonal variance scaled data set.^55,–57 The principal components (PCs) in the new variance scale data set are orthogonal to each other, eliminating colinearity in the data set. In addition to the removal of colinearity, the dimensions of a data set are often reduced, allowing the use of fewer PCs to represent the data, ultimately eliminating the inherent noise in a data set. The next critical phase in a multivariate regression analysis is careful optimization of the regression models by selecting optimum wavelength regions where changes in spectra correlate most with analyte concentrations of interest.

The wavelength region between 288 nm and 500 nm was found to be the optimum wavelength region where Hg(II) concentrations correlated most with changes in the fluorescence emission of Hg(II)-HSA complexes. Using five principal components, the developed PLS1 model for Hg(II) from Hg(II)-HSA emission data resulted in a slope of 0.99464, a square correlation coefficient of 0.99464, and offset of 1.716 × 10*⁶. A perfect model would have a slope of 1, a square correlation coefficient of 1, and offset of 0. While the merit of the PLS1 model was significant, the practical application of any regression model is the ability of the model to predict analyte concentration. Figure 4 shows the plot of the actual Hg(II) concentration in the HSA sample versus the predicted Hg(II) concentration by the PLS1 model. Clearly, the predicted Hg(II) concentrations were in close agreement with the actual Hg(II) concentrations in HSA samples.

Fig. 4.

Plot of actual versus predicted Hg(II) concentrations in human serum albumin samples.

Table I shows the result of the predicted and actual Hg(II) concentrations in HSA of the independent test validation samples. The ability of the regression model to correctly predict the Hg(II) concentration in the HSA sample was evaluated by the root mean square percent relative error (RMS%RE). An RMS%RE of 6.59% was obtained for the prediction of Hg(II) concentration in HSA of the validation samples.

TABLE I.

The predicted and actual Hg(II) concentration from human serum albumin.

Sample	Predicted Hg(II) concentration (M)	Actual Hg(II) concentration (M)	% Relative error
1	1.34E-04	1.35E-04	−1.11
2	2.93E-04	2.75E-04	6.44
3	3.84E-04	4.25E-04	−9.67
4	4.38E-04	4.75E-04	−7.83
5	4.85E-04	5.25E-04	−7.71
6	5.40E-04	5.75E-04	−6.03
7	6.68E-04	6.25E-04	6.91
8	6.29E-04	6.75E-04	−6.89
9	7.14E-04	7.25E-04	−1.53
RMS %RE			6.59

The use of simple, rapid, and relatively inexpensive spectroscopy in conjunction with the multivariate regression analysis strategy reported in this study is promising. The techniques may be a viable alternative technique for rapid determination of Hg(II) concentration, particularly in biological samples. The practical utility of multivariate regression spectral analysis for direct determination of protein and water content in agricultural products with little or no sample preparation has been widely demonstrated.^44,45 Additionally, multivariate regression analysis of spectral data is increasingly being used in biomedical, medical, and clinical studies for effective diagnosis of various cancer cells in complex biological matrices and for the determination of venous blood and muscle pH.^46,50 The use of multivariate regression analysis of spectral data for analysis of muscle oxygen saturation with no sample pretreatment without influence from skin and fat has also been reported.⁴⁷ Another advantage of multivariate regression modeling is its robustness. The robustness of multivariate regression models for accurate prediction of analyte concentrations up to six months without recalibration has been demonstrated.⁵² With further development and more studies, the reported multivariate regression technique for Hg(II) analysis in HSA samples used in this study can be adapted for the analysis of other toxic and volatile heavy metals such as arsenic. The method can also be employed to analyze other metals including Pb, Cd, Cr, Ni, or Co in HSA or other biological specimens.

CONCLUSION

Fourier transform infrared, ultraviolet-visible, and fluorescence spectroscopy were utilized to investigate the binding of Hg(II) with human serum albumin. In addition, this study reported the first use of fluorescence spectroscopy, guest-host HSA chemistry, and multivariate regression analysis in combination for the prediction of Hg(II) concentration in human serum albumin samples. The results of FT-IR analysis indicated the binding of Hg(II) at HSA amide I and amide II sites. In general, the binding of Hg(II) with HSA resulted in a decrease of HSA UV-visible absorption and dramatic quenching of HSA fluorescence emission. The developed PLS1 regression model from Hg(II)-HSA fluorescence emission data was able to correctly predict Hg(II) concentrations of independent validated HSA samples, with low root mean square percent relative error of prediction. The simplicity, rapidity, and low cost of the reported method in this study makes it a promising alternative method for a rapid determination of Hg(II) concentration in biological samples. The protocol reported in this study for Hg(II) analysis can also be adapted for the determination of other toxic heavy metals including As, Pb, Cd, Cr, Ni, or Co in HSA or other biological specimens.

Footnotes

ACKNOWLEDGMENT

The study was supported by NSF HBCU-UP: Award # 0927905.

References

Bendedouch

Chen

S.H.

“Structure and interparticle interactions of bovine serum albumin in solution studied by small angle neuron scattering”. J. Phys. Chem. 1983. 87(9): 1473–1477.

Pace

C.N.

Vajdos

Fee

Grimsley

Gray

“How to measure and predict the molar absorption coefficient of a protein”. Protein Sci. 1995. 4: 2411–2423.

J.X.

Carter

D.C.

“Atomic structure and chemistry of human serum albumin”. Nature. 1992. 358: 209–215.

Peters

T.J.

“All about albumin”. London: Academic Press, 1996.

Waheed

A.A.

Sridhar

R.K.

Gupta

P.D.

“Mechanism of dye binding in the protein assay using eosin dyes”. Anal. Biochem. 2000. 287: 73–79.

Hushcha

T.O.

Luik

A.L.

Naboka

Y.N.

“Conformation changes of albumin in its interaction with physiologically active compounds as studied by quasi-elastic light scattering spectroscopy and ultrasonic method”. Talanta. 2000. 53: 29–34.

Cui

Kong

Qin

Zhang

Liu

Lei

Yao

“Specific interaction of 40-O-(a-L-Cladinosyl) daunorubicin with human serum albumin: The binding site II on HSA molecular using spectroscopy and modeling”. J. Photochem. Photobiol. B. 2009. 95: 162–169.

Yue

Chen

Qin

Yao

“A study of the binding of C.I. Direct Yellow 9 to human serum albumin using optical spectroscopy and molecular modeling”. Dyes Pigments. 2008. 79:176–182.

Agency for Toxic Substances and Diseases Registry (ASTDR) (2005). http://www.atsdr.cdc.gov/toxprofiles/tp13.pdf. Accessed November 16, 2009.

10.

Tong

von Schirnding

Y.E.

Prapamontol

“Environmental lead exposure: a public health problem of global dimensions”. Bull. W.H.O. 2000. 78(9): 1068–1077.

11.

Schwartz

“Low-level lead exposure and children's IQ: a meta-analysis and search for a threshold”. Environ. Res. 1994. 65(1): 42–55.

12.

Vivoli

Bergomi

Borella

Fantuzzi

Caselgrandi

“Cadmium in blood urine and hair related to human hypertension”. J. Trace Elem. Electrol. Health. Dis. 1989. 3(3): 139–145.

13.

Houston

M.C.

“Role of Mercury Toxicity in Hypertension, Cardiovascular Disease, and Stroke”. J. Clin. Hypertens. 2011. 13(8): 621–627.

14.

Curtis

J.T.

Chen

Buck

D.J.

Davis

R.L.

“Chronic inorganic mercury exposure induces sex-specific changes in central TNFα expression: Importance in autism?”. Neurosci. Lett. 2011. 504: 40–44.

15.

Monteiro

D.A.

Rantin

F.T.K.A.L.

“Inorganic mercury exposure: toxicological effects, oxidative stress biomarkers and bioaccumulation in the tropical freshwater fish matrinxa*, Brycon amazonicus (Spix and Agassiz, 1829”. Ecotoxicology. 2010. 19(1): 105–123.

16.

Alattia

J.R.

Kuraishi

Dimitrov

Chang

Lemaitre

Fraering

P.C.

“Mercury is a direct and potent γ-secretase inhibitor affecting Notch processing and development in Drosophila”. FASEB J. 2011. 25(7): 2287–2295.

17.

Toimela

T.A.

Tähti

“Effects of mercuric chloride exposure on the glutamate uptake by cultured Retinal Pigment Epithelial Cells”. Toxicol. Vitro. 2001. 15: 7–12.

18.

Lund

B.O.

Miller

D.M.

Woods

J.S.

“Mercury-induced H₂O₂ production and lipid-peroxidation in-vitro in rat-kidney mitochondria”. Biochem. Pharmacol. 1991. 11(Suppl. 42): S181–S187.

19.

Rabenstein

D.L.

Isab

A.A.

“A proton nuclear magnetic resonance study of the interaction of mercury with intact human erythrocytes”. Biochim. Biophys. Acta. 1982. 721: 374–384.

20.

Bose-O'Reilly

Lettmeier

Roider

Siebert

Drasch

“Mercury in breast milk—a health hazard for infants in gold mining areas?”. Int. J. Hyg. Environ. Health. 2008. 211(5-6): 615–623.

21.

Castilhos

Z.C.

Rodrigues-Filho

Rodrigues

A.P.

Villas-Boas

R.C.

Siegel

Veiga

M.M.

Beinho

V.C.

“Mercury contamination in fish from gold mining areas in Indonesia and human health risk assessment”. Sci. Total Environ. 2006. 368(1): 320–325.

22.

Corbett

C.E.

K.M.

Costa

A.N.

Gyuricza

J.V.

Corbett

J.F.

Frizzarini

de Araujo Andrade

D.C.

Cordeiro

Stravogiannis

Chassot

C.A.

Vieira

J.L.

Pinheiro

M.C.

“Health evaluation of gold miners living in a mercury-contaminated village in Serra Pelada, Para, Brazil”. Arch. Environ. Occup. Health. 2007. 62(3): 121–128.

23.

Cortes-Maramba

Reyes

J.P.

Francisco-Rivera

A.T.

Akagi

Sunio

Panganiban

L.C.

“Health and environmental assessment of mercury exposure in a gold mining community in Western Mindanao, Philippines”. J. Environ. Manage. 2006. 81(2): 126–134.

24.

Tomicic

Vernez

Belem

Berode

“Human mercury exposure associated with small-scale gold mining in Burkina Faso”. Int. Arch. Occup. Environ. Health. 2011. 84: 539–546.

25.

EPA. “Fate and transport of mercury in the environment”. Mercury study report to Congress” (Vol III). EPA-452/R-97-005. U.S. Environmental Protection Agency, 1997.

26.

Evans

M.S.

Muir

Lockhart

W.L.

Stern

Ryan

Roach

“Persistent organic pollutants and metals in the freshwater biota of the Canadian Subarctic and Arctic: an overview”. Sci. Total Environ. 2005. (351-352): 94–147.

27.

Che'telat

Amyot

Cloutier

Poulain

“Metamorphosis in chironomids, more than mercury supply, controls methylmercury transfer to fish in high arctic lakes”. Environ. Sci. Technol. 2008. 42: 9110–9115.

28.

Boening

D.W.

“Ecological effects, transport, and fate of mercury: a general review”. Chemosphere. 2000. 40: 1335–1351.

29.

Vereda

A.E.

Siles

C.M.T.

Garcia de

T.A.

Canada

R.P.

Cano

P.J.M.

“Mercury speciation in sea food by flow injection cold vapor atomic absorption spectrometry using selective solid phase extraction”. Talanta. 2008. 77(1): 53–59.

30.

Dabeka

R.W.

Bradley

McKenzie

A.D.

“Routine, high-sensitivity, cold vapor atomic absorption spectrometric determination of total mercury in foods after low-temperature digestion”. J. AOAC Int. 2002. 85(5): 1136–1143.

31.

Lalor

DeSilva

Rattray

Wright

P.R.D.

“Design and characterization of a vapour generation system for determination of mercury by atomic absorption spectrometry”. J. Anal. Atom. Spectrom. 2005. 20(7): 659–661.

32.

da Silva

M.J.

Paim

A.P.S.

Pimentel

M.F.

Cervera

M.L.

de la Guardia

“Determination of mercury in rice by cold vapor atomic fluorescence spectrometry after microwave-assisted digestion”. Anal. Chim. Acta. 2010. 667(1-2): 43–48.

33.

Gil

Costas

Pena

De La Calle

Cabaleiro

Lavilla

Bendicho

“On-line UV photoreduction in a flow-injection/stopped-flow manifold for determination of mercury by cold vapour-atomic absorption spectrometry”. Anal. Methods. 2010. 2(11): 1798–1802.

34.

Meers

J.D.

Jahromi

E.Z.

Heyne

Gailer

“Improved RP-HPLC separation of Hg²⁺ and CH₃Hg⁺ using a mixture of thiol-based mobile phase additives”. J. Environ. Sci. Health. A. 2012. 47(1): 149–154.

35.

Yan

Chen

Xia

Jiang

“Human Serum Albumin-Mercurial Species Interactions”. J. Proteome Res. 2007. 6: 2277–2286.

36.

Zhou

Wang

Chen

Mao

“Sensitive and selective voltammetric measurement of Hg²⁺ by rational covalent functionalization of graphene oxide with cysteamine”. Analyst. 2012. 137(2): 305–308.

37.

Kodamatani

Kanzaki

Tomiyasu

Saito

Kono

“Determination of Organic and Inorganic Mercury Species as Emetine Dithiocarbamate Complexes by High-Performance Liquid Chromatography with Electrogenerated Tris(2,2′-bipyridine)ruthenium(III) Chemiluminescence Detection”. Anal. Lett. 2011. 44(17): 2769–2779.

38.

Hezard

Fajerwerg

Evrard

Colliere

Behra

Gros

“Gold nanoparticles electrodeposited on glassy carbon using cyclic voltammetry: Application to Hg(II) trace analysis”. J. Electroanal. Chem. 2012. 664: 46–52.

39.

Liu

Chen

Zhang

Wang

“Reversible “off-on” fluorescent chemosensor for Hg²⁺ based on rhodamine derivative”. Spectrochim Acta, Part A. 2012. 85(1): 38–42.

40.

Sabri

Y.M.

Kojima

Ippolito

S.J.

Wlodarski

Kalantarzadeh

Kaner

R.B.

Bhargava

S.K.

“QCM based mercury vapor sensor modified with polypyrrole supported palladium”. Sensors Actuators, B Chem. 2011. 160(1): 616–622.

41.

Tran

C.D.

Oliveira

, “Chiral ionic liquid that function as both solvent and chiral selectors for the determination of enantiomeric composition of pharmaceutical products”. Anal. Chem. 2006. 78: 1349–1356.

42.

Zhou

Lin

Welch

C.J.

Ellison

“Determination of the enantiomeric excess of an M3 antagonist drug substance by chemometric analysis of the IR spectra of different guest-host complexes”. Chirality. 2006. 18(5): 306–313.

43.

Zhu

Anslyn

E.V.

“Facile Quantification of Enantiomeric Excess and Concentration with Displacement Assays: An example in the Analyses of α-Hydroxyacids”. J. Am. Chem. Soc. 2004. 126: 3676–3677.

44.

Lee

J. H.

Choung

“Determination of Protein Content in Pea by Near infrared Spectroscopy”. Food Sci. Biotechnol. 2009. 18(1): 60–65.

45.

Fujiwara

Aoki

Ishikawa

Atsumi

Nishikawa

Kamiya

Komaru

Aquaculture Sci. 2009. 57: 469–472.

46.

Yang

Y.E.

Soyemi

O.O.

Landry

M.R.

Soller

B.R.

“Noninvasive in Vivo Measurement of Venous Blood pH During Exercise Using Near-Infrared Reflectance Spectroscopy”. Appl. Spectrosc. 2007. 61(2): 223–229.

47.

Yang

Soyemi

O.O.

Scott

P.J.

Landry

M.R.

Lee

S.M.C.

Stroud

Soller

B. R.

“Quantitative measurement of muscle oxygen saturation without influence from skin and fat using continuous wave near infrared spectroscopy”. Opt. Exp. 2007. 15(21): 13715–13730.

48.

Khanmohammadi

Garmarudi

A.B.

Ghasemi

“Back-propagation artificial neural network and attenuated total reflectance-Fourier transform infrared spectroscopy for diagnosis of basal cell carcinoma by blood sample analysis”. J. Chemometrics. 2009. 23(10): 538–544.

49.

Harvey

T.J.

Gazi

Henderson

Snook

R.D.

Clarke

N.W.

Brown

Gardner

“Factors influencing the discrimination and classification of prostate cancer cell lines by FTIR microscopy”. Analyst. 2009. 134: 1083–1091.

50.

Khanmohammadi

Nasiri

G.K.

Samani

Bagheri

G.A.

“Diagnosis of basal cell carcinoma by infrared spectroscopy of whole blood samples applying soft independent modeling class analogy”. J. Cancer Res. Clin. Oncol. 2007. 133: 1001–1010.

51.

Taleb

Diamond

McGarvey

J.J.

Beattie

J.R.

Toland

Hamilton

P.W.

“Raman microscopy for the chemometric analysis of tumor cells”. J. Phys. Chem. B. 2006. 110: 19625–19631.

52.

Busch

K.W.

Swamidoss

I.M.

Fakayode

S.O.

Busch

M.A.

“Determination of the Enantiomeric Composition of Guest Molecules by Chemometric Analysis of the UV-visible Spectra of Cyclodextrin Guest-Host Complexes”. J. Am. Chem. Soc. 2003. 125: 1690–1691.

53.

Tang

J.H.

S.D.

Chen

X.G.

“Spectroscopic studies of the interaction of anti-coagulant rodenticide diphacinone with human serum albumin”. J. Mol. Struct. 2005. 779(1-3): 87–95.

54.

Pavia

D.L.

Lampman

G.M.

Kriz

G.S.

“Introduction to Spectroscopy”. Fort Worth, TX: Harcourt Brace College, 1996.

55.

Adams

M.J.

“Chemometrics in Analytical Spectroscopy”. Cambridge, UK: Royal Society of Chemistry, 1995.

56.

Malinowski

E.R.

“Factor Analysis in Chemistry”. New York: Wiley, 1991.

57.

Martens

Naes

“Multivariate Calibration”. New York: Wiley, 1998.

58.

Fakayode

S.O.

Busch

M.A.

Busch

K.W.

“Determination of the Enantiomeric Composition of Samples by Multivariate Regression Modeling of Spectral data obtained with Cyclodextrin Guest-Host Complexes—Effect of an Achiral Surfactant and Use of Mixed Cyclodextrin”. Talanta. 2006. 68: 1574–1583.

59.

Fakayode

S.O.

Busch

M.A.

Bellert

D.J.

Busch

K.W.

“Determination of the Enantiomeric Composition of Phenylalanine by Chemometric Analysis of the Fluorescence Spectra of Cyclodextrin Guest-host Complexes”. Analyst. 2005. 130: 233–241.

60.

Fakayode

S.O.

Williams

A.A.

Busch

M.A.

Busch

K.W.

Warner

I.M.

“The Use of Polymeric Surfactant as a Chiral Selector for Chiral Analysis by Chemometric Analysis of the Fluorescence Spectra of Guest-Host Complexes”. J. Fluoresc. 2006. 16: 659–670.

61.

Williams

A.A.

Fakayode

S.O.

Lowry

Warner

I.M.

“A Versatile Chiral Selector for Determination of Enantiomeric Composition of Fluorescent and non-Fluorescent Chiral Molecules Using Steady-State Fluorescence Spectroscopy”. Chirality. 2009. 21: 305–315.