| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Abstracts of Oak Ridge Posters |
(1 Department of Chemistry "G. Ciamician", Bologna, Italy; 2 Wyatt Technology Europe GmbH, Dernbach, Germany;
aaddress correspondence to this author at: Department of Chemistry "G. Ciamician", Via Selmi 2, 40126 Bologna, Italy; fax 39-0-51-2099452, e-mail pierluig.reschiglian{at}unibo.it)
Determination of total cholesterol (TC), HDL and LDL cholesterol, and triglycerides (TG) is used to assess lipoprotein abnormalities and coronary artery disease (CAD) risk. Furthermore, studies have demonstrated that small, dense LDL particles penetrate more easily into the arterial intima (1), exhibit increasing binding to arterial wall proteoglycans(2), and are more prone to oxidative stress(3). Small, dense LDL particles also have a prolonged plasma half-life because of their lower binding affinity for the LDL receptor(4). The presence of small, dense LDL particles in plasma is, therefore, considered to be proatherogenic.
Analytical methods for LDL size characterization are therefore expected to improve classification, diagnosis, and therapy of dyslipidemic patients. However, simple methods for routine LDL size measurement are not yet available. LDL size is mainly measured by nondenaturing PAGE (5), which is time-consuming and may be unsuitable for large numbers of samples. Easier methods for measuring LDL size have been proposed, including high-performance gel-filtration chromatography (HPGC)(6). Both PAGE and HPGC can give accurate size estimations only if appropriate standards are available. Moreover, recent studies have shown that LDL particles may be discoidal, with diameter and height that are not significantly correlated(7). PAGE and HPGC cannot give information on particle conformation.
Flow field-flow fractionation (FlFFF) is a separation technique in which macromolecules and particles are separated by the combined actions of an axial flow and a perpendicular cross-flow (8). No stationary phase is present, and the separation mechanism is sufficiently gentle not to alter the native structure of proteins and protein complexes(9). Proteins are fractionated according to their difference in diffusion, with retention time that is inversely proportional to the analyte diffusion coefficient. In principle, with the use of FlFFF it is possible to obtain the hydrodynamic radius (Rh) of proteins from their retention time values. A microchannel, prototype variant of FlFFF, the hollow-fiber (HF) FlFFF that uses a piece of porous HF as separation channel [(10) and references therein], has shown advantages not only because the channel is of reduced volume, but also because it is simple to construct and inexpensive, making it potentially disposable(11). Disposable FlFFF channels should be particularly advantageous in clinical analysis, for which sample carryover must be strictly avoided.
Since its early development stage, FlFFF has been used to fractionate lipoproteins (12). FlFFF was used to characterize LDL particles of patients with CAD(13), and HF FlFFF of LDL particles from CAD patients and from healthy donors was recently reported(14). However, no independent, uncorrelated measurements of the fractionated LDL size were used to fully characterize serum lipoproteins in size and shape.
Multiangle light scattering (MALS) allows for the absolute determination of molar mass (Mw) and particle size over a broad range (15). If MALS is not coupled to a size-based separation device, however, it gives only mean values. Online FlFFF-MALS makes it possible to size-sort macromolecules and particles and to online characterize the narrowly dispersed fractions without using standards. Thus, distribution of the absolute Mw values can be obtained even for broadly dispersed samples if the concentration and specific refractive index increment (dn/dc) are also measured or known. The root mean square gyration radius (Rg) can be calculated from the angular dependence of the MALS signals alone, without information on dn/dc(15).
This study was the 1st to investigate the feasibility of FlFFF-MALS as a method for size and shape characterization of lipoproteins from whole human serum samples. Because MALS was applied to HF FlFFF for the accurate estimation of Mw and size values of protein aggregates (16), we compared the performance of the method when used with either commercial FIFFF or prototype HF FlFFF instruments. The method will require further validation and verification to be applied on a routine scale.
Human blood serum samples were obtained from healthy donors who gave informed consent. Samples were stored at –20 °C until analysis. In some experiments, sera were stained before the analysis with Sudan Black B (SBB; Fluka), a dye for lipid components, which generates a specific absorption maximum at 600 nm. For these experiments, a volume of 2 µL SBB solution (1% wt/vol in ethanol) was added to 100 µL serum and 400 µL phosphate buffer (7.5 mmol/L KH2PO4, 7.5 mmol/L Na2HPO4 x 2 H2O, pH = 7.2).
The commercial FlFFF system was the model Eclipse 2 (Wyatt Technology Europe). The channel had a trapezoidal shape, with 350-µm thickness and 175-mm length, and with breath decreasing from 18 mm (inlet end) to 3 mm (outlet end). The accumulation wall was a 30 000-Da cutoff regenerated cellulose membrane.
The prototype HF FlFFF channel was built up as described elsewhere (11)(17). A piece of polyacrylonitrile HF membrane was sheathed by 2 pieces of one-eighth inch outer diameter Teflon tube. A T connection was positioned between the 2 tubes to make the radial flow outlet. Hand-tight male fittings were positioned at the channel inlet and outlets. The HF membrane had a 30 000 Mr cutoff, nominal inner radius of 0.40 mm, and a length of 240 mm (dried conditions).
The mobile phase was phosphate buffer (7.5 mmol/L KH2PO4, 7.5 mmol/L Na2HPO4 x 2 H2O, pH = 7.2) degassed with a 1100 Series vacuum degasser (Agilent Technologies) and delivered by a 1100 HPLC iso-pump (Agilent). The injected sample volume was 10 µL (with the commercial FlFFF channel) or 4 µL (with the prototype HF FlFFF channel). With either the commercial FlFFF or the prototype HF FlFFF channel, flow rates and patterns were controlled using the Eclipse 2 separation system. The FlFFF and HF FlFFF operation conditions are reported in the Fig. 1
legend.
|
Fractionated components were online detected as a function of retention time, and a serum profile (fractogram) was recorded. Detection was performed by an HP1100 UV and visible diode array detector (Agilent), operated at 280 and 600 nm wavelengths, a refractive index detector model Optilab rEX (Wyatt Technology), and a MALS detector model DAWN HELEOS (Wyatt Technology). ASTRA software version 5.3.1 (Wyatt Technology) was used to handle signals from the detectors and to compute the Rg values.
A serum profile obtained using the commercial FlFFF channel is shown in Fig. 1A
. The concentrations were known for TC (5.05 mmol/L, 195 mg/dL) and TG (5.36 mmol/L, 473 mg/dL). Total analysis time was approximately 1 h. The scattering intensity signal shows bands not only for the main lipoprotein classes (HDL, LDL, VLDL) but also for the most abundant serum proteins (albumin, IgG). Assignment of the lipoprotein bands was then necessary. The MALS trace was compared with the UV trace at 280 nm, in which the lipoprotein bands were hardly detectable owing to their low concentration. Using SBB-stained samples, we then compared the MALS trace with the UV trace at 600 nm, in which the stained lipoproteins were the only detected species. One band ascribed to IgG (band 2) and 3 bands ascribed to the HDL coretained with albumin (band 1), to LDL (band 3), and to VLDL (band 4) were found. According to FlFFF theory, lipoproteins are eluted at retention times that increase with increasing Rh values (12). The Rh values we obtained (top x axis) were in reasonable agreement with those previously obtained by other method: 2–7 nm for HDL, 8–17 nm for LDL, and 15–40 nm for VLDL(18). From the MALS signal we also were able to evaluate the corresponding Rg values as plotted in Fig. 1A
. Teerlink et al.(7) calculated the height values for model disk-like LDL particles. By applying a hydrodynamic model for disk-like particles(19), we obtained for LDL a predicted range for Rg/Rh of 0.787–0.830. The experimental LDL Rg/Rh values we obtained for a set of serum samples were in the range 0.38–1.38. These values are higher than those expected for either a homogeneous sphere (Rg/Rh = 0.774) or for the above disk-like model (Rg/Rh = 0.787–0.830). However, LDL particles are core shell with nonhomogeneous density(20), with the shell denser than the core. Because Rg/Rh increases with increasing contribution of the shell to the LDL mass distribution, our results are compliant with a core-shell, discoidal conformation. Finally, from band intensities the relative abundance of the different lipoprotein classes can be determined: high intensity of band 4, for instance, reflects the known correlation existing between TG concentration and VLDL content.
Compared with the profile in Fig. 1A
, the profile obtained with the same serum sample by use of the prototype HF FlFFF channel (Fig. 1B
) shows a lower signal-to-noise ratio, likely because of the lower sample load. The lipoprotein classes (bands 1, 3, 4), however, are fractionated with comparable resolution, and the calculated Rh values are also comparable. Because of the different flow conditions used in HF FlFFF, the VLDL band (band 4) is broader, but, as also plotted in Fig. 1B
, the increase in Rg values across the VLDL fraction indicates that band broadening is not due to a separation efficiency lower than in commercial FlFFF but to the high size-based selectivity of HF FlFFF.
Acknowledgments
Grant/funding support: This work was partially supported by the Italian Ministry of the University and Research (Progetti di Ricerca di Interesse Nazionale 2006, 2006033944). The contribution of Wyatt Technology Europe was partly supported by a research grant of the regional government of Rheinland-Pfalz under the Innovation program.
Financial disclosures: C.J. and D.R. are a managing director and application scientist, respectively, of Wyatt Technology Europe.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
