Journal of Chromatography B, 981–982 (2015) 57–64

Contents lists available at ScienceDirect

Journal of Chromatography B journal homepage: www.elsevier.com/locate/chromb

Simultaneous determination of albumin and low-molecular-mass thiols in plasma by HPLC with UV detection Kamila Borowczyk, Monika Wyszczelska-Rokiel, Paweł Kubalczyk, Rafał Głowacki ∗ University of Łód´z, Faculty of Chemistry, Department of Environmental Chemistry, 163 Pomorska Str., 90-236 Łód´z, Poland

a r t i c l e

i n f o

Article history: Received 24 October 2014 Accepted 29 December 2014 Available online 12 January 2015 Keywords: Albumin Endogenic plasma thiols High performance liquid chromatography Derivatization Ultraviolet detection

a b s t r a c t In this paper, we describe a simple and robust HPLC based method for determination of total lowand high-molecular-mass thiols, protein S-linked thiols and reduced albumin in plasma. The method is based on derivatization of analytes with 2-chloro-1-methylquinolinium tetrafluoroborate, separation and quantification by reversed-phase liquid chromatography followed by UV detection. Disulfides were converted to their thiol counterparts by reductive cleavage with tris(2-carboxyethyl)phosphine. Linearity in detector response for total thiols was observed over the range of 1–40 ␮mol L−1 for Hcy and glutathione (GSH), 5–100 ␮mol L−1 for Cys–Gly, 20–300 ␮mol L−1 for Cys and 3.1–37.5 ␮mol L−1 (0.2–2.4 g L−1 ) for human serum albumin (HSA). For the protein S-bound forms these values were as follows: 0.5–30 ␮mol L−1 for Hcy and GSH, 2.5–60 ␮mol L−1 for Cys–Gly and 5–200 ␮mol L−1 for Cys. The LOQs for total HSA, Cys, Hcy, Cys–Gly and GSH were 0.5, 0.2, 0.4, 0.3 and 0.4 ␮mol L−1 , respectively. The estimated validation parameters for all analytes are more than sufficient to allow the analytical method to be used for monitoring of the total and protein bound thiols as well as redox status of HSA in plasma. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Human serum albumin (HSA) is the most abundant plasma protein and accounts for 50% of the total plasma proteins [1]. It plays an important role in regulation of osmotic pressure, buffering of plasma pH, distribution of fluid between different compartments, transporting of long chain fatty acids, bilirubin, carbon dioxide and acts as an efficient extracellular antioxidant [2]. Moreover, HSA is believed to be linked to interactions with a broad spectrum of inorganic compounds, such as calcium, magnesium, zinc and copper [3]. Antioxidant properties of HSA arise from the presence of the thiol group derived from cysteine 34 (Cys34), the only one uninvolved in intrachain disulfide bonding. Although SH groups of HSA, in chemical terms, behave like low-molecular-mass thiols, the biochemical reactivity of HSA is much more complicated due to steric hindrance, charge distribution and affinity of nucleophilic groups in biomolecule to the solvent. In contrast to most of the low-molecular-mass plasma thiols, the pKa of the thiol group of Cys34 in HSA is surprisingly low (∼5) [4], thus, at physiological

Abbreviations: CMQT, 2-chloro-l-methylquinolinium tetrafluoroborate; Cys, cysteine; Cys–Gly, cysteinylglycine; GSH, glutathione; Hcy, homocysteine; HSA, human serum albumin; TCEP, tris(2-carboxyethyl)phosphine. ∗ Corresponding author. Fax: +48 42 6355832. E-mail address: [email protected] (R. Głowacki). http://dx.doi.org/10.1016/j.jchromb.2014.12.032 1570-0232/© 2015 Elsevier B.V. All rights reserved.

conditions, HSA exists in reduced (HSA-SH) and oxidized forms (HSA-S-R) [2]. Typical plasma concentrations of HSA range from 0.6 to 0.75 mmol L−1 , therefore Cys34 provides a source of 80% of the total thiols in plasma [5]. It has been shown that SH group of Cys34 plays the crucial role in the defense against oxidative damage [2]. Oxidative stress, which is widely believed to be an important factor in the pathogenesis of several diseases such as liver and renal failure [6], uremia [7], diabetes mellitus [8], alcaptonuria [9], and obstructive sleep apnea [10], was proven to be associated with a decrease of the HSA thiol group content. Additionally, the increasing impact of reactive dicarbonyl compounds which occur during carbonyl stress leads to the cysteine (Cys) side chain carbonylation and, therefore, to the decrease of the HSA-SH content [11]. It has been established that protein carbonyls accumulate on tissue proteins during aging and disease development [12–16]. Sulfur-containing amino acids play an important role in the human metabolism. Cysteine, metabolically related to homocysteine (Hcy) and glutathione (GSH), is involved in a variety of important cellular functions, among others protein synthesis, detoxification and metabolism. However, due to its high toxicity in reduced form Cys may cause neurodegenerative changes such as excitotoxic brain damage, stroke or certain neurodegenerative disorders [17]. GSH functions as a major endogenous antioxidant and redox buffer and plays an important role in cellular defence including detoxification of xenobiotics and peroxides and the maintenance of immune function [18]. Many clinical studies have indicated that plasma total Hcy

58

K. Borowczyk et al. / J. Chromatogr. B 981–982 (2015) 57–64

is a risk factor for coronary vascular diseases and stroke and can act as a predictor of mortality in cardiovascular patients [19,20]. Genetic and environmental determinants of total Hcy are wellknown as well as the possibility of reduction of plasma total Hcy by B-vitamin supplementation [21,22]. Increased protein S-thiolation is considered an in vivo marker of oxidative injury [23]. The dynamic interactions of sulfhydryl compounds result in a variety of disulfide forms in vivo. Thus, about 60% of plasma Cys is protein bound while 8–10% exist in reduced form [24]. In the case of Hcy the redox species percentages are greatly different, being usually more than 95% and less than 1%, respectively [24]. In normal human plasma reduced and oxidized forms of albumin amounts to ∼25% and ∼75%, respectively [25]. Reduced, free oxidized and protein-bound forms of Cys, cysteinylglycine (Cys–Gly), GSH, Hcy as well as HSA-SH and HSA-S-R, comprise the plasma redox thiol status [26]. Altered redox status of aminothiols has been observed in a number of diseases [27], thus the knowledge of concentration of particular redox forms is desirable. It has been proposed that the levels of S-glutathionylated plasma proteins as well as other forms of GSH may serve as a useful marker for oxidative stress [28]. Ratio of cysteinyl albumin was also considered as a possible biomarker of oxidative stress [29]. Several methods have been developed for the determination of HSA [25,30,31] or protein sulfhydryls [32]. Ellman’s method is the most frequently used procedure for plasma total sulfhydryls estimation [32], but determination of plasma albumin sulfhydryls requires its isolation prior to measurement, which is usually time consuming and technically demanding [33,34]. Many assays have been also proposed for low-molecular-mass thiols determination in plasma or urine [35,36]. To the best of our knowledge the problem concerning simultaneous determination of HSA and main plasma aminothiols has received no attention as yet. Here, we describe a new HPLC-UV based method for simultaneous determination of HSA, total and protein bound thiols in human plasma that relies on derivatization of analytes with CMQT and direct injection of not deproteinized sample into chromatographic column. 2. Materials and methods 2.1. Chemicals and reagents All chemicals used throughout this study were of analyticalreagent grade except for the derivatization reagent – 2chloro-1-methylquinolinium tetrafluoroborate (CMQT), that was synthesized in this laboratory [37]. Human serum albumin (HSA) as well as reduced thiols cysteine (Cys), homocysteine (Hcy), glutathione (GSH), cysteinylglycine (Cys–Gly) and its oxidized forms were received from Sigma (St. Louis, MO, USA). Lyophilized plasma for calibration was obtained from Siemens (Siemens Healthcare, Marburg, Germany). Hydrochloric acid (HCl), sodium hydroxide (NaOH), sodium hydrogen phosphate heptahydrate (Na2 HPO4 ·7H2 O), sodium dihydrogen phosphate dihydrate (NaH2 PO4 ·2H2 O) and HPLC-grade acetonitrile were from J.T. Baker (Deventer, The Netherlands). Trichloroacetic acid (TCA) and tris(2-carboxyethyl)phosphine (TCEP) were from Merck (Darmstadt, Germany). 2.2. Instrumentation The analyses were performed on a 1220 Infinity LC system from Agilent equipped with binary pump integrated with two-channel degasser, autosampler, column oven and DAD detector. A 2 ␮L of the samples were injected with the aid of an autosampler and chromatographic separation was achieved on an Aeris WIDEPORE

XB-C18 (150 × 4.6 mm) column from Phenomenex, packed with 3.6 ␮m particles. For instrument control, data acquisition and analysis, OpenLAB CDS ChemStation Edition was used. Water was purified using Milli-QRG system (Millipore, Vienna, Austria). 2.3. Human plasma samples Blood was collected into evacuated tubes containing EDTA by venipuncture, immediately placed on ice, and centrifuged at 800 × g for 15 min at room temperature. Plasma was then used for the determination of low- and high-molecular-mass thiols without delay or stored at −80 ◦ C. The investigation was performed after ´ approval by the Ethical Committee of the University of Łódz. 2.4. Stock solutions of TCEP and CMQT Stock solutions of 0.25 mol L−1 TCEP and 0.1 mol L−1 CMQT were prepared by dissolving appropriate amount of the compound in 1 mL of 0.2 mol L−1 pH 7.4 phosphate buffer. TCEP was prepared freshly each day. 2.5. Stock solutions of HSA and low-molecular-mass thiols Stock solution of albumin was prepared by dissolving 50 mg of the protein in 1 mL of 0.1 mol L−1 pH 7.4 phosphate buffer. Stock solutions of 10 mmol L−1 Cys, Hcy, GSH, Cys–Gly or their disulfides needed in the method development procedure were prepared by dissolving appropriate amount of the compound in 1 mL of 0.05 mol L−1 HCl and diluting to the volume of 2 mL. These solutions were kept at 4 ◦ C for several days without noticeable change of the analyte content. The working solutions were prepared by appropriate dilutions with water as needed, and processed without delay. 2.6. Preparation of HSA-SH and HSA-S-Cys HSA-SH as well as HSA-S-Cys were prepared according to the modified previously published procedure [38]. Briefly, HSA (50 g L−1 ) was converted to HSA-SH by treatment with 2 mmol L−1 DTT in 0.1 mol L−1 potassium phosphate buffer, pH 7.4, 0.2 mmol L−1 EDTA for 25 min at room temperature and diluted 10-fold with 0.01 mol L−1 potassium phosphate buffer, pH 5.8. Lowmolecular-mass thiols were removed from reaction mixture by ultrafiltration through a Vivaspin 500 30-kDa cut-off membrane (Sartorius) at 4 ◦ C. Under these conditions none of intrachain disulfide bonds of HSA were reduced. HSA (50 g L−1 ) was treated with 2-fold molar excess of Cys overnight at 37 ◦ C. Excess Cys/Cystine was removed from mixture by ultrafiltration through a Vivaspin 500 30-kDa cut-off membrane (Sartorius) at 4 ◦ C. The conversion of HSA to HSA-SH and HSA-S-Cys was monitored by anion exchange chromatography [38]. Both forms of HSA were then used as the calibrators for total and reduced plasma albumin determination. 2.6.1. Total thiols assay (procedure 1) The assay is based on the procedure used previously for the determination of plasma total aminothiols [39]. 25 ␮L of plasma was diluted with 90 ␮L of 0.2 mol L−1 pH 7.4 phosphate buffer. Disulfide bonds were reduced by treatment with 10 ␮L of TCEP solution for 10 min and derivatized with 10 ␮L of CMQT for 3 min at room temperature. Next, the reaction mixture was treated with 15 ␮L of 3 mol L−1 HCl, and 2 ␮L of the sample was injected onto an HPLC column.

K. Borowczyk et al. / J. Chromatogr. B 981–982 (2015) 57–64

59

Fig. 2. Representative chromatograms of HSA after reduction with TCEP and derivatization with CMQT. Chromatographic conditions as described in Section 2.8. Fig. 1. Representative chromatogram for total forms of Hcy, GSH, Cys, Cys–Gly and HSA in human plasma after reduction with TCEP and derivatization with CMQT. Chromatographic conditions as described in Section 2.8; UN, unidentified peak.

2.6.2. Protein S-bound thiols assay (procedure 2) 50 ␮L of plasma was diluted with 450 ␮L of 0.1 mol L−1 pH 7.4 phosphate buffer and then low-molecular-mass thiols and disulfides were removed by ultrafiltration through a Vivaspin 500 30-kDa cut-off membrane (Sartorius) at 4 ◦ C. Devoid of free and oxidized low- molecular-mass thiols plasma was then processed according to the procedure 1. Fig. 3. The influence of TCEP quantity on reduction yield of HSA disulfides; n = 3.

2.6.3. HSA-SH assay (procedure 3) 25 ␮L of plasma was diluted with 100 ␮L of 0.2 mol L−1 pH 7.4 phosphate buffer and derivatized with 10 ␮L of CMQT for 3 min at room temperature. Next, the reaction mixture was treated with 15 ␮L of 3 mol L−1 HCl, and 2 ␮L of the sample was injected onto an HPLC column.

2.7. Derivatization of HSA-SH with CMQT For monitoring the kinetics of the reaction of HSA-SH with CMQT, HSA prepared according to the procedure described in Section 2.6 was used. To 50 ␮L of 1 g L−1 HSA solution, 75 ␮L of phosphate buffer (pH 7.4, 0.2 mol L−1 ) was added. The mixture was vortex mixed, followed by addition of 5 ␮L of CMQT. 10 ␮L of 3 mol L−1 HCl was added at intervals of 0.5 min–250 min at 25 ◦ C and 5 ␮L aliquots were injected into the chromatographic system.

2.8. Chromatographic separation For separation of the S-quinolinium derivatives of all thiols (including HSA) gradient elution was used (0–6 min 7–62% B, 6–8 min 62–7% B, 8–10 min 7% B, where: (A) was 0.2% TCA adjusted to pH 2.25 with 1 mol L−1 NaOH and (B) acetonitrile). When exclusively HSA was determined the chromatographic separation was accomplished using the same mobile phase, but the gradient profile was as follows: 0–3 min 30% B, 3–4 min 30–70% B, 4–5 min 70–30% B, 5–7.5 min 30% B. In both cases the temperature was 25 ◦ C, the flow-rate 1 mL min−1 and detector wavelength 355 nm. For peaks identification retention times and UV spectra were used.

2.9. Calibration standards 2.9.1. Total thiols Calibration standards were prepared by spiking 25 ␮L of human plasma with appropriate disulfides (HSA-S-Cys for HSA) at the following concentrations: 0, 0.3, 0.9, 1.2, 1.5, 1.8, 2.1, 2.4 and 3.0 g L−1 plasma for HSA, 0, 20, 50, 100, 150, 200 and 250 ␮mol L−1 plasma for Cys, 0, 1.0, 2.0, 5.0, 10, 20, 30 and 40 ␮mol L−1 for Hcy and GSH and 0, 5.0, 10, 20, 40, 60, 80 and 100 ␮mol L−1 for Cys–Gly. Then the samples were processed according to the procedure 1.

2.9.2. Protein bound thiols Calibration standards were prepared by spiking 25 ␮L of human plasma obtained according to the procedure 2 with appropriate disulfides of low-molecular-mass thiols at the following concentrations: 0, 5.0, 10, 20, 50, 100, 150 and 200 ␮mol L−1 plasma for Cys, 0, 0.5, 1.0, 2.5, 5.0, 10, 20 and 30 ␮mol L−1 for Hcy and GSH and 0, 2.5, 5.0, 10, 20, 40 and 60 ␮mol L−1 for Cys–Gly. Then the samples were processed according to the procedure 1.

2.9.3. HSA-SH Calibration standards were prepared by spiking 25 ␮L of human plasma with HSA-SH at the following concentrations: 0, 0.2, 0.4, 0.6, 0.8, 1.0, 1.4, 1.8, 2.0 and 2.4 g L−1 plasma. Then the samples were processed according to the procedure 3. In all cases calibration standards were prepared in four replicates.

60

K. Borowczyk et al. / J. Chromatogr. B 981–982 (2015) 57–64

Fig. 4. General chemical derivatization reaction of HSA with CMQT.

Fig. 6. Kinetics of the derivatization reaction of HSA with CMQT; n = 3. Reaction conditions are described in Section 2.7.

Fig. 5. Comparison of the absorption spectra of HSA and its 2-S-quinolinium derivative.

of the separation conditions led to a very well resolution and good peak symmetry. 3.2. Sample preparation

3. Results and discussion 3.1. Chromatography Reversed-phase high performance liquid chromatography (RPHPLC) has some limitations, especially when attempting to analyze samples that contain a mix of compounds possessing different physicochemical properties including a large span of size of the particles and hydrophobicity. The choice of chromatographic conditions directly affects the quality of the separation. Chemical derivatization influences selectivity, thus a standard mixture of CMQT-derivatized analytes was used for optimization of chromatographic conditions. The method optimization involved changes to the TCA concentration, gradient profile and flow rate to ensure that the method can efficiently separate all sample components. The amount of organic modifier and pH of mobile phase were altered to affect changes in retention and selectivity of the separation, primarily by changing the hydrophobicity of the eluent and degree of ionization of the analytes. Above briefly mentioned experiments enabled us to establish optimum chromatographic conditions for all analytes, specified in Section 2.8. As can be seen from the chromatogram shown in Fig. 1 the four CMQT derivatives of small thiols elute by pairs; GSH in close proximity to Hcy and Cys in close proximity to Cys–Gly. 2-S-quinolinium derivative of HSA exhibits the highest hydrophobicity and elutes as the last. All analytes, including HSA, elute within 8 min in contrast to our earlier published methodologies concerning low-molecular-mass thiols where total separation time was 11–12 min [24,39]. If only HSA is determined, chromatographic separation can be significantly shortened and accomplished within 4 min (Fig. 2). Despite high particle size discrepancy of the analytes, in both cases optimization

Due to a large number of individual compounds in biological sample, leading to difficulty in resolving the analytes of interest, low concentrations of exogenous or endogenous compounds of interest resulting in detection difficulties, and conjugation of analytes to protein and/or low-molecular-mass components of the analyzed mixture a suitable sample preparation is required. Taking under consideration the above, it is not surprising that the majority of bioanalytical methods do not use just one simple separation step, but rather involve several sample pretreatment steps which simplify the matrix, and often preconcentrate and chemically modify the analytes [35,36]. Such approach leads to a relatively purified material being introduced to the final separation unit, making the separation simpler and reliable. There are also several disadvantages of this approach, namely multiple steps usually require more lengthy and more refined sample handling and there is more chance for errors. Analytical methods for total thiols determination usually have employed reduction of disulfide bonds, deproteinization, and chemical modification followed by HPLC separation and quantification [35,36]. Thus, any tool that helps reduce the number of experiments shortens the time and cost. In our methodology the sample preparation step is significantly simplified and consists of reduction of disulfide bonds with the use of TCEP and derivatization of thiol compounds with CMQT. Such approach allows direct injection of protein containing sample, simultaneous separation and quantitative determination of low- and high-molecular-mass thiols in plasma. The new assay uses fewer steps in sample work-up which significantly speeds up the analytical procedure. The sample preparation time is substantially shortened by elimination of deproteinization stage in contrast

K. Borowczyk et al. / J. Chromatogr. B 981–982 (2015) 57–64 Table 1 Performance of the plasma thiols assay. Thiol

3.3. Disulfide bonds cleavage

Precision [%] −1

[␮mol L

]

Accuracy [%]

Intra-day

Inter-day

Intra-day

Inter-day

Total Hcy 1 10 40

0.0 0.0 8.2

2.5 3.4 10.2

99.9 102.2 99.1

95.6 97.4 105.2

Cys 20 150 300

0.2 6.3 7.5

3.6 7.8 8.2

101.1 104.0 97.3

103.6 96.4 95.3

Cys–Gly 5 40 100

6.3 6.3 8.8

7.2 8.3 8.9

96.7 100.2 103.3

103.5 104.3 98.7

5.9 4.5 4.0

6.2 5.3 4.8

102.2 98.8 101.3

97.6 96.4 98.1

4.4 4.3 5.3

4.4 5.0 5.5

97.2 98.0 94.0

96.8 96.6 103.5

Protein S-bound Hcy 2.9 0.5 0.0 5 1.3 30

4.3 2.1 2.5

107.5 100.4 104.5

94.2 98.7 95.7

Cys 5 50 200

2.5 0.9 1.7

3.6 2.7 3.2

99.7 107.3 108.4

97.8 92.3 109.3

Cys–Gly 2.5 20 60

1.1 0.9 0.7

2.1 1.8 2.4

102.3 109.5 101.4

105.4 92.1 94.3

GSH 0.5 5 30

7.5 2.0 1.9

8.9 5.1 3.2

92.9 99.9 104.5

108.7 95.6 108.2

HSA-SH [g L−1 ] 0.2 6.4 3.4 1.0 5.3 2.4

7.6 4.5 3.2

98.6 108.2 98.6

105.7 92.3 93.8

GSH 1 10 40 HSA [g L−1 ] 0.3 1.5 3.0

61

to earlier published methods where proteins were removed from plasma before chromatographic separation in different ways [20], mostly by precipitation followed by centrifugation [24] or dialysis [39].

The bulk of plasma thiols, including HSA, occurs in the disulfide forms rendering them inaccessible to derivatization reagent. In order to determine their total contents disulfide bonds must be cleaved with suitable reducing reagent to liberate a free thiol. For this purpose TCEP was used [35,36,39]. We have found that reduction reaction yield firmly depends on TCEP concentration. The diagram showing the influence of molar excess of the reagent on reduction efficiency of oxidized HSA is shown in Fig. 3. Importantly, we observed good reproducibility of the results for all analytes including HSA. It is in a good agreement with our earlier findings concerning reduction of disulfides in plasma [39]. Then we found that TCEP reduces all dimer and mixed disulfides in plasma completely but not protein bound Cys. The latter is reduced in 80%, but analytically advantageous is a good repeatability of the reaction [39]. We have also studied the influence of TCEP concentration on the derivatization of HSA. Finally, for the determination of HSA and low-molecular-mass thiols in plasma samples we have chosen 10 ␮L of TCEP, representing final concentration in plasma sample of 23.5 mmol L−1 .

3.4. Derivatization of HSA The main challenges in the assay of biological thiols lie in their unfavorable physicochemical properties and high reactivity [35,36]. In present assay we have exploited CMQT, a highly reactive and thiol specific UV derivatization reagent, widely used for derivatization of hydrophilic aminothiols [24,35–37,39]. The use of diode array detector was advantageous to obtain spectral information including maximum absorption and peak purity of the analytes. The derivatization of HSA with CMQT is a one-step reaction shown on Fig. 4. The reaction takes advantage of great susceptibility of quinolinium molecule at 2-position to nucleophilic displacement, and the high nucleophilicity of the HSA thiolate anion at slightly alkaline conditions. Moreover, high hydrophilicity of CMQT facilitates its efficient reaction with protein sulfhydryls. The process comes to a halt after acidification and the 2-S-quinolinium derivative resulted from the reaction exhibits well defined absorption maximum at 355 nm. Importantly, significant spectral differences of 2-S-quinolinium derivative in comparison to native HSA, among other beneficial increase in absorption ability, were observed (Fig. 5). As CMQT was used in a large excess relative to the analyte a bathochromic shift also becomes advantageous. Kinetics of the derivatization reaction of HSA with CMQT is shown on Fig. 6. Since 50-fold molar excess of derivatization reagent compared to analyte was used, the shape of the curve proves that under these conditions, reaction requires a relatively short time (2 min). This is

Table 2 Validation data. Analyte

Total Hcy GSH Cys Cys–Gly HSA

Regression equation

Linear range [␮mol L−1 ]

R2

Precision (%)

Accuracy (%)

Min

Max

Min

Max

LOQ [␮mol L−1 ]

LOD [␮mol L−1 ]

y = 0.212x + 2.02 y = 0.297x + 1.75 y = 0.414x + 84.55 y = 0.262x + 15.5 y = 150.5x + 104

1.0–40 1.0–40 20–300 5.0–100 3.1–37.5

0.997 0.999 0.999 0.995 0.994

0.0 4.0 1.3 0.7 1.2

10.8 12.5 11.8 14.0 13.8

89.6 96.0 97.3 96.7 88.0

102.7 112.2 113.2 103.3 105.2

0.4 0.4 0.2 0.3 0.5

0.15 0.15 0.08 0.10 0.15

Protein S-bound y = 0.186x + 1.88 Hcy y = 0.287x + 1.33 GSH y = 0.455x + 77.4 Cys y = 408x + 13.24 Cys–Gly y = 139.3x + 45.96 HSA-SH

0.5–30 0.5–30 5–200 2.5–60 3.1–37.5

0.994 0.992 0.994 0.999 0.998

0.0 1.8 1.2 0.7 3.3

3.1 7.5 3.9 4.6 14.7

92.3 88.3 92.0 87.3 91.0

104.5 104.1 107.7 109.5 104.7

0.4 0.4 0.2 0.3 0.5

0.15 0.15 0.08 0.10 0.15

6.5 6.8 6.0 7.7 7.4 6.3 6.8 7.0 6.7 7.7 7.4 7.6 7.1 6.3 8.3 7.8 9.2 8.1 5.9 7.1 48.3 45.3 45.0 45.2 44.4 44.2 44.1 44.4 43.0 47.3 44.1 41.4 41.7 38.5 42.8 39.8 42.7 40.0 35.9 40.1 54.1 56.5 67.7 63.8 51.4 40.9 49.5 49.5 51.3 43.9 33.2 42.5 39.9 45.9 38.1 62.6 44.7 47.8 44.3 39.3 16.9 20.8 34.8 20.8 18.9 20.6 30.1 25.2 22.1 23.3 21.8 22.3 22.3 20.6 19.1 23.8 18.4 19.6 15.0 15.9 35.3 41.8 43.5 39.3 38.9 42.2 63.7 52.2 44.5 52.8 53.1 53.7 55.6 52.3 49.2 58.3 40.8 43.4 36.4 40.8 90.0 86.8 55.2 91.3 76.6 60.6 70.2 71.5 72.9 68.4 53.9 68.8 62.4 75.1 62.8 99.2 64.2 76.4 63.9 65.5 204.2 172.3 141.8 189.5 130.5 145.1 119.3 162.4 134.3 164.0 172.5 157.6 216.9 169.0 158.2 168.1 124.2 152.5 143.3 131.6 256.2 225.1 217.3 249.6 180.8 200.7 177.8 232.3 190.9 238.6 258.0 234.0 346.0 262.5 247.6 260.2 191.9 211.1 241.6 202.7 75.8 74.1 90.9 98.6 66.6 50.5 59.0 57.5 62.9 60.8 46.5 57.5 44.1 19.5 26.2 45.9 45.5 46.5 29.8 45.3 4.2 4.2 4.9 3.4 3.2 2.3 4.2 3.3 3.0 2.3 2.7 1.8 1.7 0.9 1.7 1.8 3.4 2.7 1.7 2.8 6.4 7.5 5.7 5.9 3.9 5.0 4.6 6.3 5.2 6.4 8.8 5.3 9.4 7.1 5.2 8.3 3.4 3.3 4.6 3.2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

When the analytical procedure had been validated it was applied to the analysis of plasma samples donated by 20 apparently healthy volunteers. Detailed data are inserted in Table 3. Plasma was collected, treated and analyzed as described in procedures 1–3. Usually, content of protein bound thiols is calculated by subtraction of free thiols from a total bulk [24]. Our present methodology allows direct determination of total endogenic lowand high-molecular-mass thiols, protein S-linked thiols as well as HSA-SH. Direct comparison of chromatographic signals of HSA in plasma samples before and after ultrafiltration enables precise estimation of protein loss. Such approach allows facile determination of real protein S-thiolation. Average concentrations of total and protein S-bound Hcy, GSH, Cys and Cys–Gly in plasma were

Table 3 Content of different forms of thiols in plasma.

3.6. Application of the method to real samples

Hcy (␮mol L−1 ) protein-S-Hcy

% of total

Total

GSH (␮mol L−1 ) protein-S-GSH

% of total

Three concentrations representing the entire range of the calibration curves were studied: one near the lower limit of quantitation, one near the center and one near the upper boundary of the standard curve. Measured concentrations were assessed by the application of calibration curves obtained on that occasion. Detailed data concerning intra- and interday precision and accuracy are given in Table 1. The linearity of the method was tested using seven- to ten-point calibration plots, and at each concentration four replicates were assayed, independently for total and protein bound thiols as well as reduced HSA as described in procedures 1–3 and Section 2.9. Validation parameters of all fitted calibration curves were satisfactory as can be seen in Table 2.

6.4 6.3 4.4 4.5 5.1 3.7 8.1 5.7 5.1 3.4 4.7 3.2 4.4 5.4 6.3 6.2 7.8 6.4 5.4 6.1

Total % of total Cys (␮mol L−1 ) protein-S-Cys Total

measured amount − endogenous content × 100% added amount

Total

Accuracy (%) =

Cys–Gly (␮mol L−1 ) protein-SCysGly

Standard addition method was used for calibration of the method. Peaks of analytes were identified by comparison of their retention times with those of an authentic standards. The limit of detection (LOD) was determined experimentally, and was taken as the concentration that produced a detector signal that could be clearly distinguished from the baseline (larger than three times the baseline noise). The limit of quantification (LOQ) was taken as the concentration that produced a detector signal 9 times larger than the baseline noise. The LOQ, was determined by spiking a proxy matrix (0.9% NaCl in pH 7.4 10 mmol L−1 phosphate buffer) with decreasing concentrations of analyte until signal to noise ratio of 9:1. Via this simple yet efficient method very good precision as well as accuracy were obtained. Precision and accuracy were calculated using the results of the analysis of samples spiked with known amounts of analytes and analyzed in triplicate following the guidelines for biological sample analysis [40]. Precision is expressed in terms of relative standard deviation, whereas accuracy as the percentage of analyte recovery calculated by expressing the mean measured amount as percentage of added amount. Accuracy was calculated with the use of a formula:

48.6 65.4 47.5 75.7 48.7 43.0 60.5 66.5 67.6 48.1 39.9 62.7 50.5 70.9 50.9 88.1 99.0 95.5 66.7 77.9

% of total

3.5. Validation of the method

2.7 4.4 3.2 3.7 1.8 2.6 2.7 4.1 3.4 3.1 4.4 3.3 4.8 4.3 2.7 4.8 3.4 3.0 2.8 2.5

Total

HSA (g L−1 ) HSA-SH

% of total

in agreement with earlier findings concerning low molecular mass thiols [24,37,39]. Importantly, quinolinium derivative of HSA similarly to CMQT derivatized thiols, is stable under acidic conditions for at least 24 h, so prepared sample does not have to be directly measured. Since other plasma thiols are derivatized during the same sample preparation step, the method allows simultaneous determination of HSA, Cys, Hcy as well as GSH and Cys–Gly in plasma at the same analytical wavelength. We have also found that under these conditions CMQT do not reacts with HSA-S-R as well as fibrinogen, hemoglobin and myoglobin (data not shown).

13.4 15.0 13.2 17.0 16.7 14.1 15.4 15.8 15.5 16.3 16.8 18.3 17.0 16.3 19.3 19.5 21.6 20.3 16.4 17.7

K. Borowczyk et al. / J. Chromatogr. B 981–982 (2015) 57–64

Sample no.

62

K. Borowczyk et al. / J. Chromatogr. B 981–982 (2015) 57–64

5.8 and 3.4, 5.4 and 2.8, 231.2 and 157.9, 46.9 and 21.6 ␮mol L−1 , respectively. These values constitute 63.7, 55.2, 71.8 and 48.3% of total thiols, respectively. The average concentration of total HSA was 42.9 g L−1 (0.67 ␮mol L−1 ) while reduced form amounted to 7.2 g L−1 (0.11 ␮mol L−1 ), representing 16.8%. These results clearly demonstrate that the predominating form of thiols in tested samples is the protein S-bound form. Obtained results are in agreement with earlier findings concerning levels of different forms of plasma thiols [24,26,27,30,34]. 4. Conclusions The precise and accurate determination of plasma thiols requires a comprehensive approach thus constitutes an important and sophisticated analytical challenge. The developed and validated method proved to be rapid, selective and convenient for the simultaneous determination of different forms of HSA and lowmolecular-mass thiols in human plasma. The biggest benefit of presented methodology is that it allows the retention, separation and detection at the same analytical wavelength of compounds with widely different physic–chemical properties in a single chromatographic run. Streamlined sample work-up, involving only reduction of disulfide bonds (for total and protein bound thiols) and derivatization allows direct injection of protein containing sample onto chromatographic column. The method is sensitive, precise and shows very good linearity at both low and high concentration levels. It can be used effectively to quantitate the redox status of plasma albumin and should be helpful in explanation of mechanisms regulating S-thiolation of proteins. The simplicity in sample preparation, the quickness in the analytical times and the low costs of our proposed method make it a reliable tool for analytical laboratories or research groups when an elevated number of samples must be analyzed daily. The use of UV-detector known for its stability and low demand in terms of maintenance is also advantageous. Conflict of interest The authors have declared no conflict of interest. Acknowledgements This work was supported in part by grants from the University of Łódz´ and National Science Center (Grant No. 2012/07/B/ST5/00765). References [1] S. Curry, H. Mandelkow, P. Brick, N. Franks, Crystal structure of human serum albumin complexed with fatty acid reveals an asymmetric distribution of binding sites, Nat. Struct. Biol. 5 (1998) 827–835. [2] T. Peters, All about albumin: biochemistry, in: Genetic and Medical Applications, Academic Press, San Diego, 1996. [3] M. Kleinova, O. Belgacem, K. Pock, A. Rizzi, A. Buchacher, G. Allmaier, Characterization of cysteinylation of pharmaceutical-grade human serum albumin by electrospray ionization mass spectrometry and low-energy collision-induced dissociation tandem mass spectrometry, Rapid Commun. Mass Spectrom. 19 (2005) 2965–2973. [4] R. Narazaki, M. Hamada, K. Harada, M. Otagiri, Covalent binding between bucillamine derivatives and human serum albumin, Pharm. Res. (NY) 13 (1996) 1317–1321. [5] G.J. Quinlan, G.S. Martin, T.W. Evans, Albumin: biochemical properties and therapeutic potential, Hepatology 41 (2005) 1211–1219. [6] K. Oettl, V. Stadlbauer, F. Petter, J. Greilberger, C. Putz-Bankuti, S. Hallstrom, C. Lackner, R.E. Stauber, Oxidative damage of albumin in advanced liver disease, Biochim. Biophys. Acta 1782 (2008) 469–473. [7] M.L. Wratten, L. Sereni, C. Tetta, Oxidation of albumin is enhanced in the presence of uremic toxins, Ren. Fail. 23 (2001) 563–571. [8] Z. Rasheed, R. Ali, Reactive oxygen species damaged human serum albumin in patients with type 1 diabetes mellitus: biochemical and immunological studies, Life Sci. 79 (2006) 2320–2328.

63

[9] D. Giustarini, I. Dalle-Donne, S. Lorenzini, E. Selvi, G. Colombo, A. Milzani, P. Fanti, R. Rossi, Protein thiolation index (PTI) as a biomarker of oxidative stress, Free Radic. Biol. Med. 53 (2012) 907–915. [10] P. Faure, R. Tamisier, J.P. Baguet, A. Favier, S. Halimi, P. Levy, J.L. Pepin, Impairment of serum albumin antioxidant properties in obstructive sleep apnoea syndrome, Eur. Respir. J. 31 (2008) 1046–1053. [11] J.M. Acimovic, B.D. Stanimirovic, N. Todorovic, V.B. Jovanovic, L.M. Mandic, Influence of the microenvironment of thiol groups in low molecular mass thiols and serum albumin on the reaction with methylglyoxal, Chem. Biol. Interact. 188 (2010) 21–30. [12] E.R. Stadtman, Protein oxidation and aging, Science 257 (1992) 1220–1224. [13] J.A. Thomas, R.J. Mallis, Aging and oxidation of reactive protein sulfhydryls, Exp. Gerontol. 36 (2001) 1519–1526. [14] C.D. Smith, J.M. Carney, P.E. Starke-Reed, C.N. Oliver, E.R. Stadtman, R.A. Floyd, W.R. Markesbery, Excess brain protein oxidation and enzyme dysfunction in normal aging and in Alzheimer disease, PNAS 88 (1991) 10540–10543. [15] C.N. Oliver, B.W. Ahn, E.J. Moerman, S. Goldstein, E.R. Stadtman, Age-related changes in oxidized proteins, J. Biol. Chem. 262 (1987) 5488–5491. [16] I.M. Gladstone Jr., R.L. Levine, Oxidation of proteins in neonatal lungs, Pediatrics 93 (1994) 764–768. [17] A.K. Abbas, W. Xia, M. Tranberg, H. Wigström, S.G. Weber, M. Sandberg, S-sulfocysteine is an endogenous amino acid in neonatal rat brain but an unlikely mediator of cysteine neurotoxicity, Neurochem. Res. 33 (2008) 301–307. [18] A. Meister, M.E. Anderson, Glutathione, Annu. Rev. Biochem. 52 (1983) 711–760. [19] S.H. Mudd, J.D. Finkelstein, H. Refsum, P.M. Ueland, M.R. Malinow, S.R. Lentz, D.W. Jacobsen, L. Brattström, B. Wilcken, D.E.L. Wilcken, H.J. Blom, S.P. Stabler, R.H. Allen, J. Selhub, I.H. Rosenberg, Homocysteine and its disulfide derivatives: a suggested consensus terminology, Arteriosc. Thromb. Vasc. 20 (2000) 1704–1706. [20] H. Refsum, A.D. Smith, P.M. Ueland, E. Nexo, R. Clarke, J. McPartlin, C. Johnston, F. Engbaek, J. Schneede, C. McPartlin, J.M. Scott, Facts and recommendations about total homocysteine determinations: an expert opinion, Clin. Chem. 50 (2004) 3–32. [21] G. Saposnik, J.G. Ray, P. Sheridan, M. McQueen, E. Lonn, Homocysteine-lowering therapy and stroke risk, severity, and disability: additional findings from the HOPE 2 trial, Stroke 40 (2009) 1365–1372. [22] J.D. Spence, H. Bang, L.E. Chambless, M.J. Stampfer, Vitamin intervention for stroke prevention trial: an efficacy analysis, Stroke 36 (2005) 2404–2409. [23] P. Di Simplicio, F. Franconi, S. Frosalí, D. Di Giuseppe, Thiolation and nitrosation of cysteines in biological fluids and cells, Amino Acids 25 (2003) 323–339. [24] E. Bald, G. Chwatko, R. Głowacki, K. Ku´smierek, Analysis of plasma thiols by high-performance liquid chromatography with ultraviolet detection, J. Chromatogr. A 1032 (2004) 109–115. [25] K. Ohyamaa, N. Kishikawaa, A. Matsuo, T. Imazato, Y. Ueki, M. Wada, K. Nakashima, N. Kuroda, Determination of the ratio between mercaptalbumin and nonmercaptalbumin by HPLC with fluorescence probe specifically binding to albumin, J. Pharm. Biomed. Anal. 88 (2014) 170–173. [26] P.M. Ueland, M.A. Mansoor, A.B. Guttormsen, F. Muller, P. Aukrust, H. Refsum, A.M. Svardal, Reduced, oxidized and protein-bound forms of homocysteine and other aminothiols in plasma comprise the redox thiol status - a possible element of the extracellular antioxidant defense system, J. Nutr. 126 (1996) 1281–1284. [27] T. Apeland, O. Kristensen, M.A. Mansoor, The aminothiol redox status in haemodialysis patients does not improve with folate therapy, Scand. J. Clin. Lab. Investig. 69 (2009) 265–271. [28] W.A. Kleinman, D. Komninou, Y. Leutzinger, S. Colosimo, J. Cox, C.A. Lang, J.P. Richie Jr., Protein glutathiolation in human blood, Biochem. Pharmacol. 65 (2003) 741–746. [29] Y. Ogasawara, Y. Mukai, T. Togawa, T. Suzuki, S. Tanabe, K. Ishii, Determination of plasma thiol bound to albumin using affinity chromatography and high-performance liquid chromatography with fluorescence detection: ratio of cysteinyl albumin as a possible biomarker of oxidative stress, J. Chromatogr. B 845 (2007) 157–163. [30] V.B. Jovanovic, I.D. Pavicevic, M.M. Takic, A.Z. Penezic-Romanjuk, J.M. Acimovic, L.M. Mandic, The influence of fatty acids on determination of human serum albumin thiol group, Anal. Biochem. 448 (2014) 50–57. [31] T. Hayashi, K. Suda, H. Imai, S. Era, Simple and sensitive high-performance liquid chromatographic method for the investigation of dynamic changes in the redox state of rat serum albumin, J. Chromatogr. B 772 (2002) 139–146. [32] G.L. Ellman, Tissue sulfhydryl groups, Arch. Biochem. Biophys. 82 (1959) 70–77. [33] H.M. van Eijk, D.R. Rooyakkers, B.A. van Acker, P.B. Soeters, N.E. Deutz, Automated isolation of high-purity plasma albumin for isotope ratio measurements, J. Chromatogr. B 731 (1999) 199–205. [34] V.B. Jovanovic, A.Z. Penezic-Romanjuk, I.D. Pavicevic, J.M. Acimovic, L.M. Mandic, Improving the reliability of human serum albumin-thiol group determination, Anal. Biochem. 439 (2013) 17–22. [35] K. Ku´smierek, G. Chwatko, R. Głowacki, P. Kubalczyk, E. Bald, Ultraviolet derivatization of low-molecular-mass thiols for high performance liquid chromatography and capillary electrophoresis analysis, J. Chromatogr. B 879 (2011) 1290–1307. [36] K. Ku´smierek, G. Chwatko, R. Głowacki, E. Bald, Determination of endogenous thiols and thiol drugs in urine by HPLC with ultraviolet detection, J. Chromatogr. B 877 (2009) 3300–3308.

64

K. Borowczyk et al. / J. Chromatogr. B 981–982 (2015) 57–64

[37] E. Bald, R. Głowacki, 2-Chloro-1-methylquinolinium tetrafluoroborate as an effective and thiol specific UV-tagging reagent for liquid chromatography, J. Liq. Chromatogr. Relat. Technol. 24 (2001) 1323–1339. [38] R. Głowacki, H. Jakubowski, Cross-talk between Cys34 and lysine residues in human serum albumin revealed by N-homocysteinylation, J. Biol. Chem. 279 (2004) 10864–10871.

[39] R. Głowacki, E. Bald, Fully automated method for simultaneous determination of total cysteine, cysteinylglycine, glutathione and homocysteine in plasma by HPLC with UV absorbance detection, J. Chromatogr. B 877 (2009) 3400–3404. [40] FDA Guidance for Industry Bioanalytical Method Validation, 2001, http://www. fda.gov/cder/guidance/4252fnl.pdf