Open Tubular Capillary Electrochromatography - American Chemical

Katariina O2 o1rni,‡ Petri Kovanen,‡ and Marja-Liisa Riekkola*,†. Laboratory of Analytical Chemistry, Department of Chemistry, P.O. Box 55, FIN-...
0 downloads 0 Views 268KB Size
Anal. Chem. 2006, 78, 2665-2671

Open Tubular Capillary Electrochromatography: Technique for Oxidation and Interaction Studies on Human Low-Density Lipoproteins Ruth Kuldvee,† Lucia D’ulivo,† Gebrenegus Yohannes,† Petrus W. Lindenburg,† Minna Laine,† Katariina O 2o 1 rni,‡ Petri Kovanen,‡ and Marja-Liisa Riekkola*,†

Laboratory of Analytical Chemistry, Department of Chemistry, P.O. Box 55, FIN-00014 University of Helsinki, Helsinki, Finland, and Wihuri Research Institute, Kalliolinnantie 4, FIN-00140, Helsinki, Finland

A novel, open tubular capillary electrochromatographic method was developed for the in vitro oxidation of lowdensity lipoprotein (LDL) particles. Low-density lipoprotein particles with molar mass of ∼2.5 MDa yielded a stable stationary phase at temperatures 25 and 37 °C and at pH values from 3.2 to 7.4. The quality of the coatings was not influenced by variations in the LDL concentration in the coating solutions (within the range of 2-0.015 mg/ mL) with the coating procedure used in the study. Radiolabeled LDL stationary phases and scanning electron microscopy, employed to shed light on the location and coating density of LDL particles on the inner surface of the capillary wall, confirmed the presence of an LDL monolayer and almost 100% coating efficiency (99 ( 8%). In addition, the radioactivity measurements allowed estimation of the amount of LDL present in a single capillary coating. Capillaries coated with human LDL particles were submitted to different oxidative conditions by changing the concentration of the oxidant (CuSO4), oxidation time, pH value, and temperature. The oxidation procedure was followed with electroosmotic flow mobility, which served as an indicator of the increase in total negative charges of LDL coatings, and by asymmetrical field flow fractionation, which measured the changes in size of the lipoprotein particles. The results indicated that oxidation of LDL was progressing with increasing time, temperature, and concentration of the oxidant as expected. The oxidation process was faster around neutral pH values (pH 6.57.4) and inhibited at acidic pH values (pH 5.5 and lower). Evidence from many biochemical studies supports the hypothesis that oxidative modification of low-density lipoprotein (LDL) particles is an important factor in the formation of atherosclerotic plaque. Conclusive proof for the oxidation hypothesis, as well as for factors affecting the oxidation, is still missing, however.1 The factors affecting the oxidation can be classified as intrinsic (composition of the LDL particles) and extrinsic (pH of the * To whom correspondence should be addressed. E-mail: [email protected]. Fax: +358 9 19150253. † University of Helsinki. ‡ Wihuri Research Institute. (1) Chisolm, G. M.; Steinberg, D. Free Radical Biol. Med. 2000, 28, 18151826. 10.1021/ac052006i CCC: $33.50 Published on Web 03/15/2006

© 2006 American Chemical Society

environment, presence or absence of oxidants or antioxidants in plasma compartment, etc). Before we can confirm the oxidation hypothesis (and, more importantly, find a cure for atherosclerosis), we need to clarify the factors that contribute to or inhibit the oxidation of LDL particles.2 Various methods have been used to monitor the oxidation processes of LDL and to assess the effects of antioxidants in blood and of environmental factors. One of these methods is slab gel electrophoresis, which measures the electrophoretic mobility of LDL. A rise in the electrophoretic mobility reflects an increase in the negative charge of LDL such as during an oxidative process. The oxidative modification of LDL has also been studied by capillary electromigration techniques such as high-performance frontal analysis combined with capillary electrophoresis (CE)3-5 and capillary isotachophoresis.6 LDL particles have been reported to interact strongly with the capillary wall in the studies where LDL solutions have been used as samples and introduced into uncoated fused-silica capillaries.7 Due to the adsorption of the LDL particles, successful analysis of LDL has been achieved only after the permanent8 or dynamic9-12 coating of the capillary wall. Liu et al. have eliminated this problem by a special sample preparation technique that allowed the separation of two LDL particle species in uncoated fused silica capillary.13 In our recent paper, we have taken advantage of LDL interactions with capillary wall by a successful and stable LDL (2) Paloma¨ki, A. Doctoral thesis, Tampere University Press: Tampere, Finland, 2002, pp 1-88. (3) Mohamed, N. A. L.; Kuroda, Y.; Shibukawa, A.; Nakagawa, T.; Gizawy, S. E.; Askal, H. F.; El Kommos, M. E. J. Chromatogr. 2000, 875, 447-453. (4) Kuroda, Y.; Cao, B.; Shibukawa, A.; Nakagawa, T. Electrophoresis 2001, 22, 3401-3407. (5) Kuroda, Y.; Watanabe, Y.; Shibukawa, A.; Nakagawa, T. J. Pharm. Biomed. Anal. 2003, 30, 1869-1877. (6) Zorn, U.; Haug, C.; Celik, E.; Wennauer, R.; Schmid-Kotsas, A.; Bachem, M. G.; Gru ¨ nert, A. Electrophoresis 2001, 22, 1143-1149. (7) Ceriotti, L.; Shibata, T.; Folmer, B.; Weiller, B. H.; Roberts, M. A.; de Rooij, N. F.; Verpoorte, E. Electrophoresis 2002, 23, 3615-3622. (8) Schmitz, G.; Mollers, C.; Richter, V. Electrophoresis 1997, 18, 1807-1813. (9) Stocks, J.; Miller, N. E. J. Lipid Res. 1998, 39, 1305-1309. (10) Weiller, B. H.; Ceriotti, L.; Shibata, T.; Rein, D.; Roberts, M. A.; Lichtenberg, J.; German, J. B.; de Rooij, N. F.; Verpoorte, E. Anal. Chem. 2002, 74, 17021711. (11) Gambino, R.; Uberti, B.; Alemanno, N.; Pisu, E.; Pagano, G.; Cassader, M. Atherosclerosis 2004, 173, 103-107. (12) Zinellu, A.; Sotgia, S.; Galistu, F.; Lumbau, F.; Pasciu, V.; Pes, G. M.; Tadolini, B.; Deiana, L.; Carru, C. Talanta 2004, 64, 428-434. (13) Liu, M.-Y.; McNeal, C. J.; Macfarlane, R. D. Electrophoresis 2004, 25, 29852995.

Analytical Chemistry, Vol. 78, No. 8, April 15, 2006 2665

coating approach in capillary electrochromatography (CEC).14 This method offers a novel way to monitor changes in LDL particles. Our goal in the present study was to develop an open tubular capillary electrochromatographic technique with LDL particles as stationary phase for the in vitro oxidation of low-density lipoproteins under different conditions. As the LDL particles were immobilized onto the capillary wall, the parameters affecting conditions around LDL particles, such as pH, temperature, and oxidation time, could be easily changed. The reliability of the technique was compared with that of the classical in vitro technique carried out in a vial. EXPERIMENTAL SECTION Materials. Dimethyl sulfoxide (DMSO) was from Oy FFChemicals Ab (Yli Ii, Finland). EDTA disodium salt, aldosteroned, and acetic acid (99.7%, d 1.049 kg/L) were from Sigma Chemical Co. (St. Louis MO). Testosterone, progesterone, and copper sulfate were from Merck (Darmstadt, Germany). 17-a-Progesterone and β-estradiol were from Sigma Aldrich, and tert-butoxycarbonyl-L-[S]-methionine N-hydroxysuccinimidyl ester (NHS) was from Amersham Biosciences (Uppsala, Sweden). OptiPhase HiSafe liquid scintillation cocktail was from Wallac Oy (Turku, Finland), and Dulbecco’s phosphate-buffered saline was from Bio Whittaker (Verviers, Belgium). Instrumentation. Capillary Electrophoresis. Electrophoretic measurements were made with a Hewlett-Packard 3DCE system (Agilent, Waldbronn, Germany) equipped with a diode array detector (detection at 200 and 245 nm) and air thermostating of the capillary. The temperature of the sample carousel was maintained by water bath (Lauda Ecoline, Lauda-Ko¨nigshofen, Germany). The data acquisition rate was 100 Hz, and the response time of the detector 0.1 s. Uncoated fused-silica capillaries were from Composite Metal Services Ltd. (Worcestershire, UK). Dimensions were 50-µm i.d. and 375-µm o.d. The length of the capillary to the detector was 30.0 cm with a total length of 38.5 cm. Radioactivity Measurements. PD-10 desalting columns and the 3H labeling reagent were from Amersham Biosciences (Uppsala, Sweden). The liquid scintillation counter used for radioactivity measurements was 1219 Rackbeta, LKB from Wallac Oy (Turku, Finland). Field Flow Fractionation. Miniaturized and conventional asymmetrical flow field-flow fractionation (mAsFlFFF and AsFlFFF, respectively) channels constructed in-house were connected to a UV/visible detector. The geometrical area of the accumulation wall and the volume were 6 cm2 and 0.25 mL for the miniaturized scale and 76 cm2 and 3.7 mL for the conventional scale. The dimensions of the mAsFlFFF channel were 11-cm length, 0.7 and 0.35 cm at inlet and outlet, and 500-µm spacer thickness. Dimensions for the AsFlFFF were 38 cm × 2 cm × 500 µm. In both systems, a regenerated cellulose acetate ultrafiltration membrane with a molar mass cutoff of 10 kDa (DSS-RC70PP, Nakskov, Denmark) was laid on top of a porous frit. An HPLC pump (model PU-980, Jasco International Co., Ltd., Tokyo, Japan) was used to deliver carrier solution during the injectionrelaxation-focusing period from both the front and backside of the channel. The outlet flow from the channel was monitored with a UV/visible detector (HP1050 model 79853C, Tokyo, Japan) at

280 nm. The shaking bath 58-16, serial no. 19999-6, was from Techne Inc. (Princeton NJ). Other Equipment. A Radiometer PHM 220 Lab pH meter was used in the pH measurements. Distilled water was further purified with a Millipore water purification system (Millipore S.A., Molsheim, France). Millipore filters (Bedford, MA) were used for filtering background electrolytes (BGEs). In ζ potential measurements carried out by Zetasizer 3000 HSA (Malvern Instruments, Malvern, Worcs, UK), oxidized lipoprotein solution was injected with a syringe via an injection port to the measurement cell until a clear reflection from the whole pathway of light going through the cell was seen. One run consisted of 10 individual measurements, and our results at pH 4 and 7.4 were mean values of three runs. Methods. Isolation of LDL Particles from Human Plasma. Human LDL (d ) 1.019-1.050 g/mL) was isolated from plasma of healthy volunteers by sequential ultracentrifugation in the presence of 3 mM EDTA15,16 Briefly, solid KBr was added to plasma to adjust its density to 1.019 g. Very low-density lipoprotein and intermediate density lipoprotein were removed, and the density of the bottom fractions was adjusted to 1.050 g/mL with solid KBr. After ultracentrifugation for 72 h at 35 000 rpm, LDL was recovered from the top of the centrifuge tubes, recentrifuged (d ) 1.060 g/mL) for 24 h at 35 000 rpm, and dialyzed extensively against LDL buffer (1 mM EDTA and 150 mM sodium chloride in water, pH 7.4, adjusted with sodium hydroxide). The amount of LDL is expressed in terms of its protein concentration, which was determined by the method of Lowry et al.17 with bovine serum albumin (BSA) as standard. EDTA was removed from LDL in a PD-10 desalting column equilibrated with Dulbecco’s phosphatebuffered saline. Sample and Buffer Preparation. The ionic strength of the BGE solutions (phosphate and acetate buffers) was 20 mM with pH adjusted to 3.2-7.4 with 1.0 M sodium hydroxide. Before use, the BGEs were filtered through 0.45-µm Millipore filters using a Millipore vacuum system. The BGE solution did not contain any LDL. Concentrations of the original LDL samples were 1.0-11.4 mg/ mL. The LDL solutions used for coating the capillary were diluted with the running BGE solution. The concentration of LDL in coating solutions was 0.1 mg/mL if not stated differently. The stock solutions of steroids were prepared in methanol. The steroid sample consisted of 20 µg/mL aldosterone, 20 µg/ mL testosterone, 50 µg/mL β-estradiol, and 50 µg/mL progesterone in BGE solution. All solutions were stored at +4 °C. Capillary Coating. Before coating, the fresh capillary was submitted to hydrochloric acid pretreatment as described in ref 14. LDL coating was applied to the capillary inner surface as follows: after preconditioning, the capillary was flushed with LDL solution at 50 mbar for 40 min and left to stand filled with the LDL solution for 15 min. Electroosmotic Flow (EOF) Measurements and CE Separations. For measurement of the charge of the LDL coatings, six successive runs with DMSO as EOF marker were performed both in freshly coated and in oxidatively modified capillaries. After coating, the capillary was flushed with BGE solution for 15 min. Between runs it was flushed for 2 min. CE conditions for steroid separations were as follows: voltage 20 kV, temperature of the capillary cassette 25 or 37 °C, injection of steroids for 3 s at 50 mbar.

(14) Kuldvee, R.; Wiedmer, S. K.; O ¨o ¨rni, K.; Riekkola, M.-L. Anal. Chem. 2005, 77, 3401-3405.

(15) Weisgraber, K. H.; Rall, S. C., Jr. J. Biol. Chem. 1987, 262, 11097-11103. (16) Radding, C. M.; Steinberg, D. J. Clin. Invest. 1960, 39, 1560-1569.

2666

Analytical Chemistry, Vol. 78, No. 8, April 15, 2006

Oxidation Procedure. Oxidation of LDL was carried out either in vial (before coating onto the capillary) or in capillary (after coating). Oxidation in vial was carried out as follows. EDTA-free LDL and 5 µM CuSO4 solution were mixed in the desired ratio in a vial for a predetermined time (0.5-12 h). The oxidation was terminated by adding EDTA to the mixture (so that the concentration of EDTA in the mixture would be 1 mM). After that, the mixture was coated onto a capillary as described in the Capillary Coating section. For oxidation in the capillary, the capillary was coated, as described in Capillary Coating, with LDL solution containing EDTA. After coating, the capillary was filled with 5 µM CuSO4 solution (in BGE) for the desired time (0.5-12 h). The oxidation was terminated by replacing CuSO4 solution in capillary by 1 mM EDTA in BGE solution. Scanning Electron Microscopy (SEM) Studies. Taking SEM photos of the inside of the capillary was technically impossible as the capillaries could not be opened without damage to the inner coating. As a means of elucidating the distribution of LDL particles on the capillary surface, a fused-silica capillary was coated on the outside. First the polyimide coating of a capillary was burned off and the surface was cleaned with methanol. Then the capillary was submitted to acid pretreatment, and LDL coating was applied approximately as described in Capillary Coating. After coating, the capillary was rinsed with MilliQ water and dried in a desiccator at least overnight. The capillary was glued onto a SEM stage and coated with chromium (thickness 3-4 nm). The distribution of the particles was then examined with a Hitachi S4800 field emission scanning microscope operating with an acceleration voltage of 1.0 kV. FFF Studies. The carrier liquid used in the AsFlFFF work was phosphate buffer at pH 7.4, 6.5, and 3.2 and acetate buffer at pH 4 and 5. The ionic strength of the buffers was 20 mM. Actual channel thickness was calculated, at 20 °C, from BSA solution in 8.5 mM phosphate buffer with 150 mM NaCl solution, having a diffusion coefficient of 6.21 × 10-7 cm2/s.18 Unless otherwise specified, the main and cross-flow rates used in mAsFlFFF were 0.3 and 0.5 mL/min, respectively, and those in conventional AsFlFFF 1.8 and 1.4 mL/min, respectively. LDL Oxidation for AsFlFFF Experiments. Lipoprotein oxidation was performed by incubation of LDL (2 µL of 4.6 mg of protein/ mL in EDTA-free PBS) with 23 µL of 5 µmol/L CuSO4‚5H2O at 37 °C for up to 24 h. The reactants were placed in an Eppendorf plastic container, and incubation was undertaken in a horizontal shaking bath operating at 25 cycles/min. Oxidation was terminated by addition of 2 µL of 1 mM EDTA. Radiolabeling of LDL Particles. LDL particles were labeled with 3H following the Bolton and Hunter procedure19 as described previously,20 except that the Bolton and Hunter reagent was replaced with NHS. The ester group of NHS reacts with nucleophiles in the same manner as the Bolton and Hunter reagent, labeling, preferentially, lysine and the N-terminus, and, more slowly, nucleophiles such as the sulfhydryl group of cysteine, the guanidinium group of arginine, and the indole ring of tryptophan. The labeling reaction was conducted in a vessel precooled on ice to maximize labeling specificity and efficiency. The LDL particles in 0.1 M borate buffer (pH 8.5) at a protein concentration of 5-10 (17) Lowry, O. H.; Rosebrough, N. J.; Farr, A. L.; Randall, R. J. J. Biol. Chem. 1951, 193, 265-275. (18) Meechai, N.; Jamieson, A. M.; Blackwell, J. J. Colloid Interface Sci. 1999, 218, 167-175. (19) Bolton A. E.; Hunter W. M. Biochem. J. 1973, 133, 529-539. (20) Kovanen P. T., Kokkonen J. O., J. Biol. Chem. 1991, 266, 4430-4436.

mg/mL were added to the precooled labeling agent. After 30 min, the reaction was terminated by addition of 100 µL of 0.2 M glycine in 0.1 M borate buffer (pH 8.5). The labeled LDL sample was purified on a PD-10 desalting column and dialyzed against LDL buffer. Coating was performed as described in Capillary Coating. Three different lengths of capillaries were coated with radiolabeled LDL particles: 38.5 (the length normally used for separations), 77 (twice the size of separation capillary), and 192.5 cm (five times the size of separation capillary). When longer capillaries were used, the preconditioning rinses, 3H LDL (0.07 mg/mL) flush, and final BGE flush were increased correspondingly with the capillary length. After coating, the capillaries of longer size were cut into pieces, and the pieces, just as the 38.5-cm capillaries, were flushed with freshly prepared hot piranha acid (3:1 sulfuric acid and hydrogen peroxide (v/v) for 60 min. The LDL coating material was collected and its radioactivity was measured. Radioactivity Measurements. Samples were dissolved in 2 mL of MilliQ and 3 mL of OptiPhase HiSafe 3 liquid scintillation cocktail in a polystyrene vial and left to stand for at least 1 h. Radioactivity was measured over a period of 10 min with a liquid scintillation counter, 3H window: channels 5-364. With a 3H standard, the counting efficiency was determined at 22.6%. Values were not corrected for decay because the half-time of 3H is sufficiently long (12.3 years) relative to the age of the sample. RESULTS AND DISCUSSION To determine the applicability of our previously described14 coating procedure for low-density lipoprotein particles to the present in vitro oxidation studies, we carried out dilution experiments in which the amount of human lipoproteins per coating was much reduced. Furthermore, we investigated the distribution and stability of the lipoprotein stationary phase in the capillary and the amount of lipoprotein per coating. Concentration Studies on LDL Coating. In our previous study, where we demonstrated that small changes in ionic strength do not markedly influence the parameters of the LDL coating, the LDL samples were diluted by a factor of 4, at which point the LDL concentration was 1 mg/mL or higher. Here, we first determined the optimal LDL concentration for the coating. The EOF mobility, which is used to characterize the coating stability, was found to be constant up to an LDL dilution of 500 times (the concentration of the original LDL sample was 7 mg/mL), corresponding to an LDL concentration of 0.015 mg/mL in coating solution. In fact, the EOF mobility appeared to become slightly slower with the dilution. At LDL concentrations lower than 0.015 mg/mL, the coating deteriorated dramatically due to the fact that the amount of LDL particles present in the capillary were most probably not sufficient anymore for coverage of the capillary wall. Although the LDL concentration of 0.015 mg/mL had proved successful, the concentration of 0.1 mg/mL was mostly used in our further studies to ensure there were no complications due to deterioration. The results of the dilution study are important in two respects. They allowed us to reduce the amount of LDL (i.e., to save time and costs), and they showed that even large changes in the original LDL sample concentration (during the present study the LDL concentration in original samples varied more than 10-fold) do not produce significant changes in the coating performance. Distribution of LDL Particles in the Capillary. The capability of scanning electron microscopy (SEM) was tested in clarificaAnalytical Chemistry, Vol. 78, No. 8, April 15, 2006

2667

Figure 2. Radioactivity measurements of 10 LDL-coated capillaries. The capillaries were coated with radiolabeled LDL as described in the Experimental Section. The LDL coatings were then removed with piranha acid, and the radioactivity of the effluent was measured.

Figure 1. SEM photos of (A) an uncoated fused-silica capillary after burning off the polyimide coating and cleaning the capillary surface and (B) an LDL-coated fused-silica capillary. In (B), the capillary has been coated on the outside, the polyimide coating of the capillary has been burned off, and the capillary has been submitted to acid preconditioning and a coating procedure similar to the procedures used to coat the capillary on the inside. For more details see the Experimental Section.

tion of the LDL coating uniformity along the capillary. Even though clear differences were found in the coated and uncoated capillaries (Figure 1), the resolution was too low to give information about the number of particle layers or the quantity of the lipoprotein particles. To obtain this information, the capillaries were coated with radiolabeled LDL according to the standard procedure described in the Experimental Section. After the coating procedure, the capillary was rinsed with piranha acid to remove the LDL coating for radioactivity measurements on the procedure described in the Experimental Section. Additionally, fused-silica capillaries of 192.5 (five times longer than our normal capillary length) and 77 cm were coated with radiolabeled LDL and cut into pieces; every piece was separately rinsed with piranha acid according to the same procedure, and the solution was collected for radioactivity measurements. As can be seen from Figure 2, the 10 capillaries yielded similar values for LDL particle content (average 2 × 10-7 µmol/capillary; and 5 × 10-9 µmol/cm). The results confirm our earlier repeatability findings for EOF mobilities in different LDL-coated capillaries.14 The similar amounts of LDL in different parts of long capillaries confirmed our expectation that the LDL particles are evenly distributed along the whole capillary. If the LDL particles preferably interact with the capillary wall rather than with each other, they should form a monolayer on the wall. Knowing the radioactivity (disintegrations per minute (dpm) per centimeter) of the capillary, the radioactivity per microgram of apoB (17 142 dpm), Avogadro’s number, and the molecular mass of the LDL particle (2.5 × 106), and taking into 2668 Analytical Chemistry, Vol. 78, No. 8, April 15, 2006

account the spatial limitation of round particles, we can calculate the approximate number of LDL particles in the capillary (as well as their average coverage percentage in capillary). The diameters of the LDL particles ranged from 20 to 25 nm, and on the basis of our AsFFF studies, an LDL particle size of 22 nm was selected as the average size to be used in the calculations. Calculations after radioactivity measurements on 10 3H LDL-coated capillaries resulted in a coating efficiency value of ∼99 ( 8%, confirming that the LDL particles were covering the inside surface of the capillary as a monolayer. Stability of LDL Coating/Particles under Acidic pH. Our recent studies14 showed that LDL-coated capillaries could be used in pH range of 4.5-7.4 with no sign of deterioration of the coating. Here, we were able to show that an LDL coating could be used sequentially for separations at pH 7.4 and acidic pH values (pH 4.0 and 4.5); that is, the pH of the BGE could be changed back and forth without notable pH hysteresis (Supporting Information, Figure S-1). We also confirmed by AsFFF that no changes in the sizes or distribution of the LDL particles occurred at pH values 3.2-7.4 (data not shown). Oxidation of LDL. The oxidation of LDL in the body probably takes place either in the subendothelial space of the arterial wall or within the plasma compartment. The oxidation process is influenced by the composition of the LDL particle (the ratio of polyunsaturated to monounsaturated fatty acids, the antioxidant content of LDL) and by the surrounding microenvironment (the concentration of oxidants, pH, presence of local antioxidants).21,22 Even though several oxidation studies have been published worldwide, the link between oxidation of LDL and atherosclerosis and its clinical complications has not yet been proven. Many in vitro systems have been used to generate oxidized LDL: macrophage-generated oxidants,23,24 cytochrome p450,25,26 and catalytically active copper and iron (e.g., Cu2+ or Fe2+). Of these, incubation with copper ion27 has been the most common system. (21) Esterbauer, H.; Gebicki, J.; Puhl, H.; Jurgens, G. Free Radical Biol. Med. 1992, 13, 341-90. (22) Liu, M.-L. Doctoral thesis, Helsinki, Finland, 2002. (23) Witztum, J. L.; Steinberg, D. Trends Cardiovasc. Med. 2001, 11, 93-102. (24) Chisholm, G. M.; Steinberg, D. Free Radical Biol. Med. 2000, 28, 18151826. (25) Aviram, M.; Kent, U. M.; Hollenberg, P. F. Atherosclerosis 1999, 143, 253260. (26) Sevanian, A.; Ursini, F. Free Radical Biol. Med. 2000, 29, 306-311. (27) Burkitt, M. J. Arch. Biochem. Biophys. 2001, 394, 117-135.

Figure 3. Influence of oxidation on EOF mobility. Coating solution: LDL diluted 1:100 (v/v) with BGE solution. Coating conditions: 40 min rinsing and 15 min standing with coating solution. BGE solution: phosphate, I ) 20 mM, pH ) 7.4. Running conditions: fused-silica capillary with total length of 38.5 cm (30 cm to the detector) and i.d./ o.d. 50/375 µm; capillary temperature 25 °C; injection 5 s at 50 mbar; running voltage 20 kV; detection at 214 nm. Oxidation is carried out with 5 µM CuSO4 solution in BGE.

The chemical and biological properties of copper-oxidized LDL resemble those in LDL particles modified by macrophages. Copper ion was used as the oxidizing agent throughout the present study. One of the parameters used to measure the extent of LDL oxidation is the total charge of the LDL particles. When LDL particles are attached to the capillary wall in CEC, the charge of the particles can be indirectly measured as the EOF mobility inside the capillary. The more the oxidation of the particles, the more negatively charged they become and the faster is the EOF in the capillary (Figure 3). When the oxidation reaches a plateau, so does the EOF mobility. We used EOF mobility as an indirect measure of LDL oxidation throughout our studies. AsFlFFF was employed side by side to monitor changes in LDL particle sizes and possible fusion/aggregation of the particles during oxidation. The results are shown in Figure 4. Without oxidation (Figure 4A) and also after 0.5-h oxidation (data not shown), the average size of the LDL particles is 22 nm, and no additional peaks indicating fusion/aggregation of particles are seen. After 3-h oxidation with 5 µM CuSO4 at pH 7.4, the average size of the main LDL peak is still 22 nm, but an extra, vaguely defined low-intensity peak can be seen on the right side of Figure 4B, indicating an increase in the particle size through LDL aggregation or fusion or both. After 5-h oxidation (Figure 4C), the latter peak is better defined, with higher intensity; however, the peak with average size of 22 nm was still the main peak. From these results, we can conclude that, although a slight increase in aggregation occurs during the oxidation, the intact 22-nm particles are still in the majority. Comparison of LDL Oxidations in Vial and Capillary. There are two ways to study the oxidation of LDL by capillary

Figure 4. Influence of oxidation on LDL particle size, measured by AsFlFFF. The major peak corresponds to LDL particles. AsFlFFF conditions as described in the Experimental Section. Oxidation with 5 µM CuSO4, pH 7.4, 37 °C. (A) 0-, (B) 3-, and (C) 5-h oxidation. Carrier solution: 8.5 mM phosphate buffer, 0.02% NaN3, 150 mM NaCl, pH 7.4, I, 20 mM. Relaxation focusing: frontal flow rate 0.1 mL min-1; flow inward from outlet 2.6 mL min-1; injection 1.0 mL min-1for 10 min; relaxation time 30 min. Flow rates during elution: 1.8 and 1.4 mL/min for the main and cross-flow rates; UV detection at 280 nm.

electrochromatography: to perform the oxidation in a vial and then coat the capillary with the oxidized LDL particles or to coat a capillary and carry out the oxidation procedure inside the capillary. The conventional in vitro procedure for LDL oxidation is carried out in a vial, but since the LDL sample cannot then contain any generally used preservative, it is susceptible to oxidation during storage. In the second approach, the LDL sample contains the preservative EDTA until the coating is completed, at which time the original LDL buffer is rinsed out and replaced by the background electrolyte. An LDL sample prepared in this way is stable much longer than an “unprotected” LDL sample. The oxidation procedure in capillary is far more flexible than that in vial. It allows the introduction of the oxidant for a desired time and easy termination of the process, e.g., by addition of a preserving agent (EDTA in our case). The process can easily be repeated by replacing the preserving agent with oxidant. In addition, the environment around the LDL particle (e.g., pH, presence of oxidants or antioxidants) can effortlessly be altered. By comparison, the addition of components to the vial is easy, but their removal is most difficult. Finally, the processes in capillary are more automated and less prone to human errors. The EOF mobilities that we obtained after oxidations carried out in vial and in capillary confirm the close similarity of the oxidation processes in the two systems during 5-h oxidation (Figure 5). The results are somewhat different at longer oxidation times (12-h oxidation). However, because oxidation times over 5 h are biologically less relevant, we concentrated on the 5-h oxidation studies. Analytical Chemistry, Vol. 78, No. 8, April 15, 2006

2669

Figure 5. Comparison of average EOF mobilities after different oxidation times in vial and in capillary. Oxidations were carried out in vials and in capillary at 25 °C and pH 7.4, with 5 µM CuSO4 as oxidizing agent. Running conditions as in Figure 3.

The repeatability of the oxidation process was evaluated by performing a 3-h oxidation procedure in coated capillaries and in vials. In both cases, the RSD for the oxidation process was less than 4% at 37 °C and less than 2% at 25 °C (not shown). Bearing in mind that oxidation processes in general are complicated and multilevel processes and that their mechanisms are not yet clear, our results show the oxidation procedure in capillary to be highly repeatable. Interactions between LDL and Steroids during Oxidation. To clarify the changes in the hydrophobicity of an LDL particle surface during the oxidation uncharged aldosterone, testosterone, β-estradiol, and progesterone were selected due to their favorable extent of interactions with LDL particles, which allowed relatively short analysis times, on average 6 min at 37 °C (Figure 6). Retention factors were used to study the effect of the oxidation of LDL particles on the interactions between steroids and the stationary phase. Aldosterone, which is the least hydrophobic steroid, interacted very weakly with the LDL stationary phase, and it migrated almost together with the EOF marker. When the EOF mobility increased with oxidation, the separation window for aldosterone became narrower, revealing still weaker interactions with the LDL particles. As to the other model steroids, their interactions with LDL particles became slightly stronger as a function of oxidation time, as can be seen from Table 1, even though the total analysis time was shortened. A possible explanation for this is that extensive oxidation of LDL particles with copper leads preferentially to degradation of the hydrophobic regions of

Figure 6. Separation of steroids in LDL-coated capillary. Sample: 1, aldosterone; 2, testosterone; 3, β-estradiol; 4, progesterone. Detection wavelength 245 nm. Coating and running conditions as in Figure 3.

apoB-100 and their relocation onto the particle surface. Due to such degradation low molecular mass, fragments of apoB-100 are released from LDL loosening the surface and helping the hydrophobic core lipids to enter the particle surface.28 Then LDL particles become more hydrophobic during oxidation, and the hydrophobic interactions with steroids are enhanced.29 LDL Oxidation at 25 and 37 °C. Because the temperature of the environment is expected to affect the oxidation process of LDL, we performed oxidation procedures both at 25 °C and at physiological temperature 37 °C. A temperature of 25 °C is technically less demanding and allows us to compare our results with those of several other in vitro studies. As can be seen from Figure 7, and as was expected, the oxidation process is faster at higher temperature. In addition, at 37 °C the oxidation appears to reach a plateau, while at 25 °C, it is still progressing. Thus, the physiological temperature 37 °C should be used to obtain a reliable picture of the in vivo oxidation process (Figure 7). Influence of pH on Oxidation of LDL. The acidity of the environment is increased at inflammatory and ischemic sites.30,31 Atherosclerotic lesions are in several ways similar to chronic inflammatory sites, and evidence from many sources suggests that the extracellular pH is acidic. Altered acidity of the environment has been shown to affect the rate and extent of LDL oxidation.30 To study the effect of pH on copper-induced LDL oxidation, we carried out 5-h oxidation experiments at five different pH values. All EOF measurements were performed at pH 7.4 to facilitate the interpretation of the pH-related changes in LDL oxidation. The results presented in Figure 8A demonstrate that the influence of pH on LDL oxidation is complex. The mechanism of the oxidation

Table 1. Effect of LDL Oxidation Time on the Retention Factors (k) of Steroids. Oxidation of “Protected” LDL Inside a Capillary, pH 7.4, 37 °Ca oxidation time (h) 0 1 2 3 4 5 change of EOF mobility/k value (%) a

EOF mobility × 104 (cm2/s‚V)

testosterone

retention factor k β-estradiol

progesterone

3.11 (1.7) 3.41 (0.6) 3.67 (1.6) 3.89 (0.2) 4.00 (1.5) 4.07 (0.5)

0.09 (0.8) 0.09 (0.6) 0.11 (1.1) 0.13 (0.3) 0.14 (2.2) 0.14 (1.5)

0.22 (0.9) 0.22 (0.3) 0.27 (1.3) 0.29 (7.3) 0.34 (2.6) 0.36 (1.7)

0.67 (1.3) 0.72 (0.9) 0.89 (1.6) 0.98 (1.3) 1.04 (3.9) 1.04 (3.6)

26

The RSD values of six parallel experiments are given in parentheses.

2670 Analytical Chemistry, Vol. 78, No. 8, April 15, 2006

34

50

42

Figure 7. Oxidation of LDL in capillary at 25 and 37 °C, pH 7.4. Other experimental conditions as in Figure 3.

process appears to be similar at pH values 7.4 and 6.5, with the oxidation reaching a plateau at the end of 5 h. The more acidic environment (pH values 4, 4.5, and 5.0) clearly retards the oxidation. The relative change of EOF mobility increases progressively with the pH, at the end of 5 h being 4, 5, 8, 22, and 26% for pH values 4.0, 4.5, 5.0, 6.5, and 7.4, respectively. The supporting ζ-potential measurements at pH 4.0 revealed an insignificant increase in LDL particle charges during the 5-h oxidation, while at pH 7.4, the change was notably greater. This is in keeping with our CE measurements indicating that the oxidation process (while Cu2+ is used as the oxidant) is more pronounced at higher pH values. These results are also in accordance with the literature where the studies have indicated that copper-mediated oxidation is slower at acidic pH compared to physiological pH.32 It has been suggested that the reason may be due to the ability of copper to bind to the protein part of LDL particles.33 Repetitions after the 5-h oxidation procedure at pH 7.4 gave repeatable results as demonstrated in Figure 8B and in Table S-1 in Supporting Information. CONCLUSIONS An open tubular capillary electrochromatographic method was developed for study of LDL and its oxidative modifications.The results of the CEC and AsFlFFF methods demonstrated that the LDL particle stationary phase is stable over pH range 3.2-7.4. Radioactivity measurements confirmed that LDL coatings form a (28) O ¨o ¨rni, K.; Pentika¨inen, M. O.; Ala-Korpela, M.; Kovanen, P. T. J. Lipid Res. 2000, 41, 1073-1714. (29) Wang, Z.; Yu, H.; Hong, J.; Li, X. Wuhan Daxue Xuebao, Yixueban 2004, 25, 236-238. (30) Leake, D. S. Atherosclerosis 1997, 129, 149-157. (31) Naghavi, M.; John, R.; Naguib, S.; Siadaty, M. S.; Grasu, R.; Kurian, K. C.; van Winkle, W. B.; Soller, B.; Litovsky, S.; Madjid, M.; Willerson, J. T.; Casscells, W. Atherosclerosis 2002, 164, 27-35. (32) Rodrı´guez-Malaver, A. J.; Leake, D. S.; Rice-Evans, C. A. FEBS Lett. 1997, 406, 37-41. (33) Kalyanamaran, B.; Artholine, W. E.; Parthasarathy, S. Biochim. Biophys. Acta 1990, 1035, 286-292.

Figure 8. (A) Oxidation of LDL in capillaries at different pH values. The EOF mobilities have been normalized to remove the effect of slight differences in the coatings. (B) Comparison of three parallel LDL oxidation procedures in capillary. Capillary temperature 37 °C. Other experimental conditions as in Figure 3.

monolayer inside the capillary and revealed the amount of LDLs in the capillary. The radioactivity experiments also indicated that the inside wall of the capillary was evenly coated with LDL particles. The in vitro oxidation of LDL particles can be reliably carried out in capillary, and the oxidation mechanism is comparable to that obtained with the classical method carried out in a vial. Open tubular capillary electrochromatography is thus an excellent new and automated tool for LDL oxidation studies carried out under different environmental conditions. ACKNOWLEDGMENT Marja Siitari-Kauppi is thanked for her help in radioactivity analysis and Marianna Kemell for her help with scanning electron microscopy. We are grateful to Jari Hautala for his help in the ζ-potential measurements of LDL particles. Financial support was provided by the Academy of Finland under Grants 210194 and 20914 (M.-L.R.), 206296 (M.-L.R. and R.K.) and 80631 (K.O ¨ .). SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review November 11, 2005. Accepted February 22, 2006. AC052006I

Analytical Chemistry, Vol. 78, No. 8, April 15, 2006

2671