Comprehensive Two-Dimensional Field-Flow Fractionation-Liquid

Mar 10, 2007 - A novel, comprehensive two-dimensional asymmetric field-flow fractionation-liquid chromatographic system is described (AsFlFFF-RPLC)...
0 downloads 0 Views 357KB Size
Download PDF
Anal. Chem. 2007, 79, 3091-3098

Comprehensive Two-Dimensional Field-Flow Fractionation-Liquid Chromatography in the Analysis of Large Molecules Gebrenegus Yohannes,† Susanne K. Wiedmer,† Jaakko Hiidenhovi,‡ Ari Hietanen,‡ and Tuulia Hyo 1 tyla 1 inen*,†

Laboratory of Analytical Chemistry, Department of Chemistry, University of Helsinki, P.O. Box 55, FIN-00014 Helsinki, Finland, and MTT Agrifood Research, 31600 Jokioinen, Finland

A novel, comprehensive two-dimensional asymmetric field-flow fractionation-liquid chromatographic system is described (AsFlFFF-RPLC). The interface is based on a switching valve, and the whole sample is analyzed in both dimensions. The system proved to be repeatable and quantitative in the characterization of egg white proteins. Four peaks at 4, 5.5-6.0, 7.5-8.0, and 10.0-11.0 nm, and corresponding to lysozyme, ovalbumin, transferrin, and a dimer of transferrin, were obtained in the AsFlFFF first-dimension system. Lysozyme also produced an additional peak, which overlapped with ovalbumin. Twelve compounds were separated in the LC seconddimension system. Identifications were made with the help of standards (ovalbumin, ovotransferrin, lysozyme) and by comparison of the peak areas, particle sizes, and retention data with values given in the literature. The effect of heat on egg white denaturation was studied, and the unfolding of peptide bonds of the protein was found to be pronounced when the sample was heated in phosphate solution. In comprehensive two-dimensional (2D) liquid-phase separations, the entire sample is subjected to two different separation mechanisms. The separation efficiency of comprehensive 2D techniques is superior to traditional, one-dimensional methods, and more information can be obtained from the chemical and physical characteristics of the sample components. Various comprehensive 2D liquid chromatographic (LC) systems have been developed, most notably ion-exchange chromatography × reversed-phase liquid chromatography (RPLC), size-exclusion chromatography (SEC)×RPLC, and normal-phase LC×RPLC.1-5 The best separation is accomplished in orthogonal separation * Corresponding author. E-mail: [email protected], Fax: +3589-19150253. † University of Helsinki. ‡ MTT Agrifood Research. (1) Kok, S. J.; Hankenmeier, T.; Schoenmakers, P. J. J. Chromatogr. A 2005, 1098, 104-110. (2) Holland, H. A.; Jorgenson, J. W. Anal. Chem. 1995, 67, 3275-3283. (3) Opiteck, G. J.; Lewis, K. C.; Jorgenson, J. W.; Anderegg, R. J. Anal. Chem. 1997, 69, 1518-1524. (4) Po´l, J.; Hohnova´, B.; Jussila, M.; Hyo¨tyla¨inen, T. J. Chromatogr. A, in press; available on-line. (5) Dugo, P.; Favoino, O.; Luppino, R.; Dugo, G.; Mondello, L. Anal. Chem. 2004, 76, 2525-2530. 10.1021/ac062169m CCC: $37.00 Published on Web 03/10/2007

© 2007 American Chemical Society

systems consisting of complementary techniques, where each dimension provides maximum separation selectivity for one type of distribution but as far as possible is insensitive to the differences in the other type of distribution. Thus, in the ideal case, the separation mechanisms in the two separation dimensions are completely independent. In practice, unfortunately, the selectivity in the two systems is often more or less correlated. In SEC, for example, currently the most widely used method for the separation of large molecules, it is difficult to avoid chemical interactions with the stationary phase and to obtain a separation based purely on size. Thus, in a SEC×RPLC combination, the separation mechanisms tend to be correlated. Also liquid chromatography at critical conditions has been used in comprehensive twodimensional combinations with SEC. Although this technique is not very straightforward with regard to, for example, analysis time and optimization, it has been applied successfully to the characterization of various polymers.6-8 Also, other multidimensional techniques, such as combination of capillary isoelectric focusing and hollow fiber flow field-flow fractionation, have been used in the fractionation of large molecules, such as proteins.9 Field-flow fractionation (FFF) is an alternative technique that has been successfully applied for the separation and fractionation of large molecules. The advantage of the FFF methods is that the separation of analytes is achieved solely through the interaction of the sample with an external, perpendicular physical field, rather than by the interaction with a stationary phase. It is a straightforward task to adjust size fractionation range, resolution, and run time by manipulating the channel flow and the strength of the field. In addition, due to the low pressures used in flow FFF (FlFFF), the shear effects between analytes and the channel surfaces are minimized and the shear-induced degradation is greatly reduced. Thus, proteins and other molecules, sensitive to these effects, can be fractionated without structural alteration. FlFFF, which is one of the four subclasses of FFF, is a versatile method applicable to many scientific and technological studies of synthetic and biological polymers and environmental colloids of (6) Jiang, X.; Van der Horst, A.; Lima, V.; Schoenmakers, P. J. J. Chromatogr. A 2005, 1076, 51-61. (7) Im, K.; Kim, Y.; Chang, T.; Lee, K.; Choi, K. J. Chromatogr. A 2006, 1103, 235-242. (8) Coulier, L.; Kaal, E. R.; Hankemeier, T. J. Chromatogr. A 2005, 1070, 7987. (9) Kang, D.; Moon, M. H. Anal. Chem. 2006, 78, 5789-5798.

Analytical Chemistry, Vol. 79, No. 8, April 15, 2007 3091

Figure 1. AsFlFFF×RPLC system.

diameters ranging from 2 nm (molar masses of a few thousand daltons) to ∼50 µm.10 Sizes of macromolecules and particles separated by FlFFF, and in particular by asymmetrical flow FFF (AsFlFFF), can be calculated and the molar masses are similar to those obtained by SEC.11 However, instead of separation being achieved with a stationary phase as in SEC, the separation depends on the application of a cross-flow force field and a laminar flow. FlFFF works well beyond the exclusion limits of chromatography,12 and AsFlFFF has successfully been applied to fractionation and determination of particle sizes and molar masses for a number of biopolymers, such as proteins.13-16 As well, it offers applications related to the formation of protein aggregates.17-19 Eluent compatibility is a critical requirement in 2D liquid-phase techniques because the eluent from the first dimension becomes the sample solvent in the second dimension. Since FFF allows considerable scope for choice of the eluent, it is relatively easy to combine FFF with a suitable LC mode. In a comprehensive FFF×LC setup, the flow from the FFF is collected in a loop of the transfer valve and then injected for separation on the LC column while the second fraction is filling the other sampling loop. From the instrumental point of view, then, the interface is the same as used in LC×LC. A novel, comprehensive two-dimensional AsF1FF×RPLC system, capable of giving information on the properties of high-molar mass molecules, was constructed in this work. The comprehensive system combines separation based on two different physical properties of a compound, namely, molecular size and hydrophobicity. The two separation mechanisms can be considered orthogonal, as the hydrophobicity does not have any significant effect on the retention in AsFlFFF, while in RPLC, the retention is mainly based on hydrophobicity. The applicability of the system is demonstrated in the separation of egg white proteins. In addition to their nutritional importance, egg white proteins have multiple functional properties, such as foaming, emulsification, and heatsetting, making them essential ingredients for the food industry. They are also of increasing interest for incorporation in healthpromoting products. Egg white has a protein content of 10-12%, comprising mainly ovalbumin, ovotransferrin, ovomucoid, globulins, and lysozyme.20 Typically, egg white proteins have been separated by chromatographic and electromigration techniques, including RPLC, SEC, ion-exchange chromatography, countercurrent chromatography, sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and isoelectric focusing.20-26 However, none of these methods allows simultaneous separation of all the proteins present in egg white. The two-dimensional AsFlFFF×RPLC system demonstrated 3092

Analytical Chemistry, Vol. 79, No. 8, April 15, 2007

in this work provides a successful separation of egg white proteins and offers great potential for the characterization of high-molar mass molecules in a variety of matrices. EXPERIMENTAL SECTION Instrumentation. The AsFlFFF×LC apparatus (Figure 1) was assembled from an AsFlFFF device, an LC system, and an interfacing 10-port, high-pressure, two-position valve (C2-1000EP, VICI Valco). The AsFlFFF channel was constructed in-house in a manner similar to that used by other groups. 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. A Mylar spacer with thickness of 500 µm, with the channel cut away, was placed between an ultrafiltration membrane and an upper glass plate. The nominal channel dimensions were 38 cm × 2 cm × 500 µm. An HPLC pump (model PU-980, Jasco International Co., Ltd., Tokyo, Japan) was used to move the carrier liquid. Samples were introduced to the channel at 1.0 mL/min for 1-5 min by a second HPLC pump. During the injectionrelaxation-focusing period, the carrier liquid was delivered from both the front and the backside of the channel. The outlet flow from the channel was monitored with a UV-vis detector (HP1050 (10) Schimpf, E. M.; Caldwell, K.; Giddings, J. C. In Field Flow Fractionation Handbook; Wiley-Interscience: New York, 2000; pp 3-30. (11) Barth, H. G.; Boyes, B. E.; Jackson, C. Anal. Chem. 1998, 70, 251R-278R. (12) Hansen, M. Am. Biotechnol. Lab. 2002, 20, 15-16. (13) Wahlund, K.-G.; Litzen, A. J. Chromatogr. 1989, 461, 73-87. (14) Litzen, A.; Wahlund, K.-G. J. Chromatogr. 1989, 476, 413-421. (15) Lee, H.; Williams, S. K. R.; Allison, S. D.; Anchordoquy, T. J. Anal. Chem. 2001, 73, 837-843. (16) Liu, M.-K.; Giddings, J. C. Macromolecules 1993, 26, 3576-3588. (17) Silveira, J. R.; Raymond, G. J.; Hughson, A. G.; Race, R. E.; Sim, V. L.; Hayes, S. F.; Caughey, B. Nature 2005, 437, 257-261. (18) Luo, J.; Leeman, M.; Ballagi, A.; Elfwing, A.; Su, Z.; Janson, J-C.; Wahlund, K.-G. J. Chromatogr., A 2006, 1134, 236-245. (19) Zhu, R.; Frankema, W.; Huo, Y.; Kok, W. Th. Anal. Chem. 2005, 77, 45814586. (20) Miguel, M.; Manso, M. A.; Lo´pez-Fandino, R.; Ramos, M. Eur. Food. Res. Technol. 2005, 221, 542-546. (21) Shibusawa, Y.; Iino, S.; Shindo, H.; Ito, Y. J. Liq. Chromatogr. Relat. Technol. 2001, 24, 2007-2016. (22) Vachier, M. C.; Piot, M.; Awade´, A. C. J. Chromatogr. B 1995, 664, 201210. (23) Rothemund, D. L.; Thomas, T. M.; Rylatt, D. B. Protein Expression Purif. 2002, 26, 149-152. (24) Desert, C.; Gue´rin-Dubiard, C.; Nau, F.; Jan, G.; Mallard, J. J. Agric. Food Chem. 2001, 49, 4553-4561. (25) Mo ¨ller, C. C.; Thomas, D.; Van Dyk, D.; Rylatt, D.; Sheehan, M. Electrophoresis 2005, 25, 35-46. (26) Nau, F.; Mallard, A.; Pages, J.; Bru´le G. J. Liq. Chromatogr. Relat. Technol. 1999, 22, 1129-1147.

model 79853C, Tokyo, Japan) at 280 nm. Capillary Teflon tubes (i.d. 0.5 mm), restrictors (from a local electrical shop), and threeway valves (V101T; Upchurch Scientific, Oak Harbor, WA) were used to control the carrier liquid flows. An Agilent ChemStation for LC and LC/MS (Palo Alto, CA) was used for data acquisition. The carrier liquid used in the AsFlFFF was composed of 8.5 mM phosphate (I ) 20 mM), 150 mM NaCl, and 0.02% NaN3 at pH 7. 4. Relaxation/focusing was as follows: frontal flow rate 0.1 mL min-1; flow inward from outlet 3.3 mL min-1; injection 1.0 mL min-1 for 2-5 min; relaxation time 12 min. Flow rates during elution were 0.4 mL min-1 for the outflow rate and 2.6 mL min-1 for the cross-flow rate. However, for the larger aggregates (after heating), the flow rates during elution were 2.6 mL min-1 for the outflow rate and 0.4 mL min-1 for the cross-flow rate. The actual channel thickness (w) calculated from the retention time of bovine serum albumin in 8.5 mM phosphate-150 mM NaCl solution, at pH 7.4 (D ) 6.21 × 10-7 cm2/s)27 was 498 µm using an equation R ) t0/tr ) 6DV0/Vcw2. The LC was an Agilent 1100 LC (degasser, pump, injector, column compartment, and UV-vis detector), and the column was a 4.6-mm-i.d., 50-mm-long Chromolith C18 (Merck). The LC system was connected to a PC and controlled by Agilent Chemstation software that allowed identical method initiation and termination. UV detection was done at wavelengths 214 and 280 nm. The LC eluent was linear and consisted of 0.025% trifluoroacetic acid (TFA) in water (eluent A) and 0.025% TFA in acetonitrile (eluent B), used in gradient elution mode. The gradient program, which was repeated every 5 min, was as follows: 0 min, eluent A 88% + 12% eluent B; 4.5 min, eluent A 20% and eluent B 80%; 5 min, eluent A 88% and eluent B 12%. The flow rate was 2.2 mL/min. 2D Contour Plot Processing. A conversion program was used to transfer raw chromatographic data into a 2D array (program by Philipp Marriott, RMIT, Melbourne, Australia). Visualization was performed in Transform (Noesys Research Systems International, Crowthorne, UK). Reagents. Chemicals and solvents were of HPLC or analytical grade unless noted otherwise. Acetonitrile (Labscan, Dublin, Ireland) and water (laboratory-made with a Milli-Q device, Millipore, Molsheim, France) with addition of MS-grade TFA (Fluka, Buchs, Switzerland) were used in the mobile phase. Ovalbumin (MW 45 000 amu), ovotransferrin (MW 76 000 amu), and lysozyme (MW 14 400 amu) for use as standards were isolated from egg white by preparative ion-exchange chromatography. In addition, lysozyme standard was purchased from Sigma (St. Louis, MO). Hen eggs were purchased from a local store. Isolation of Proteins from Egg White. Five liters of egg white was diluted with 15 L of distilled water, and the mixture was adjusted to pH 6 with 3 M HCl. The solution was left overnight at 4 °C to enable ovomucin precipitation. The mixture developed a white, gelatinous precipitate. The supernatant was separated from the precipitate by decanting. Before anion-exchange chromatography, the mixture was adjusted to pH 8 with 3 M NaOH. Pilot-scale chromatography was performed with 3200 mL (12.5 cm × 18 cm i.d.) of Q Sepharose Fast Flow anion exchanger (Amersham Biosciences, Uppsala, Sweden). All steps were carried (27) Meechai, N.; Jamieson, A. M.; Blackwell, J. J. Colloid Interface Sci. 1999, 218, 167-175.

out at a flow rate of 500 mL min-1. Nonbound material was removed by washing the column with water until the absorbance reached the baseline. Bound material was recovered by applying a step gradient of 0.2 M NaCl until the absorbance reached the baseline, and then a step gradient of 0.5 M NaCl to remove the rest of the proteins. The purities of ovotransferrin, lysozyme, and ovalbumin were 76.4, 87.0, and 91.4%, respectively, determined by RP-HPLC using a known method.28 Preparation of Egg White Samples. Egg white was diluted 3-fold with 0.05 M Tris-HCl buffer solution at pH 9.0, containing 0.4 M NaCl and 10 mM β-mercaptoethanol. β-Mercaptoethanol was used to reduce and separate ovomucin subunits. The egg white stock solution (buffered) was gently stirred overnight at 4 °C and then filtered through cellulose-acetate 0.45-µm filters prior to analysis.29 The amount of egg white used in the AsFlFFF analysis was 0.2 mL, diluted to 50 mL with 8.5 mM phosphate buffer containing 150 mM NaCl, 0.02% NaN3, at pH 7.4. In the study of the effect of temperature on the egg white proteins, 3 mL of egg white stock solution was diluted to 50 mL with 8.5 mM phosphate-saline buffer containing 150 mM NaCl, 0.02% NaN3, at pH 7.4. The solutions were heated for 120 min at 55, 63, 68, 73, or 80 °C. After cooling down, the samples were filtered through a 0.45-µm ashless filter. Before injection to the FFF channel, 10 mL of filtrate was further diluted to 50 mL. For determination of the effect of sodium dodecyl sulfate (SDS) on the egg white proteins, 3 mL of egg white stock was diluted to 50 mL with 0.5% SDS solution and the resultant mixture heated for 120 min at 80 °C. After cooling down, the sample was filtered through a 0.45-µm ashless filter. Before injection to the FFF channel, 10 mL of filtrate was further diluted to 50 mL. Calculations. The diffusion coefficient (D) of a spherical protein in a dilute aqueous solution can be approximated by the Stokes-Einstein equation, which considers a rigid sphere solute diffusing in a solvent continuum

D ) kT/3πηdH

(1)

where k is the Boltzmann constant, T is the absolute temperature, η is the viscosity of the suspending liquid, and dH is the hydrodynamic particle diameter. In AsFlFFF, this diffusion coefficient can be calculated from the retention times using equation

D ) t0w2V˙ c/6tRV 0

(2)

where w is the channel thickness, V˙ c is the volumetric cross-flow rate (channel inlet flow minus channel outlet flow), V0 the void volume of the channel, t0 the void time, and tR the retention time.13 The particle size (dH) of comparable spherical particles is obtained by combining the diffusion coefficient of eqs 1 and 2 to give

dH ) 2kTV 0tR/πηw2t0V˙ c

(3)

RESULTS AND DISCUSSION In the development of the comprehensive 2D AsFlFFF×RPLC, conditions were optimized for the two methods off-line and then (28) Croguennec, T.; Nau, F., Pezennec, S.; Brule, G. J. Agric. Food Chem. 2000, 48, 4883-4889. (29) Awade´, A. C.; Efstathiou, T. J. Chromatogr. B 1999, 723, 69-74.

Analytical Chemistry, Vol. 79, No. 8, April 15, 2007

3093

fine-tuned for the on-line setup. The repeatability was studied, and the system was applied to study the composition of egg white and the changes in protein composition occurring in heated and SDS-denatured egg samples. Ion-Exchange Chromatography of Egg White Proteins. IEC fractionation in pilot scale was used for the preparation of egg white protein standards. The supernatant, or “mucin-free” egg white, was used for the fractionation. The fractionation was done at pH 8 to ensure that all major egg proteins except lysozyme would be negatively charged and bind to the anionic exchanger. The volume of the anionic exchanger used to fractionate the acidic proteins was first determined experimentally so that the exchangerbinding sites would be saturated with ovalbumin and other proteins in the mucin-free egg white with high affinity for the anionic exchanger. As a result, during loading, all proteins except lysozyme were bound to the exchanger until the dynamic resin capacity in the conditions of fractionation was reached. As soon as resin capacity was exceeded, negative proteins were selectively displaced according to their pI. The first fractions collected were the “ovotransferrin fraction” and “lysozyme fraction”. A NaCl gradient was applied to elute the tightly bound proteins from the column. The “ovalbumin fraction” was obtained with an isocratic elution with 0.2 M NaCl, and the fraction was collected until the absorbance at 280 nm reached baseline. Then 0.5 M NaCl was applied to the column to remove the rest of the proteins. The ovalbumin, ovotransferrin, and lysozyme fractions were characterized by RPLC.28 Optimization of AsFlFFF and RPLC. In the first part of the study, egg white samples were fractionated by AsFlFFF off-line before RPLC separation of the fractions. An automated system for collecting fractions from AsFlFFF was constructed and utilized. It was soon noticed, however, that the RPLC profiles changed, the ovalbumin peak was decreased and new peaks appeared in the chromatogram. This was an indication that the egg white proteins changed after collection of the fractions, and thus, the off-line system was abandoned. The change was investigated by analyzing egg samples in series by RPLC. In RPLC, the chromatographic profile of the egg sample began to change after 30min exposure of the sample to ambient conditions. However, storage at +4 °C overnight had no influence on the peak profiles. Although, in principle, it would be possible, therefore, to perform the AsFlFFF analysis and collection of fractions in a cooled environment, in practice, the on-line combination is more feasible, as will be demonstrated below. Figure 2A and B shows the fractograms of a crude egg white sample and the standards by off-line AsFlFFF, respectively. In Figure 2A, a relatively large amount of egg sample was introduced to AsFlFFF to see the minor components of the egg white, causing some tailing of the peaks. Four peaks, corresponding to proteins with diameters of 4, 5.5-6.0, 7.5-8.0, and 10.0-11.0 nm were identified as lysozyme, ovalbumin, ovotransferrin, and a dimer of ovotransferrin (peaks a-d in Figure 2A). Lysozyme showed an additional peak that overlapped with ovalbumin, which is probably due to the impurity of the isolated standard. Commercial lysozyme standard gave only one peak in the fractionation. Off-line RPLC separation of the FFF fractions was carried out with a monolithic column of porous structure having both mezoand macropores and, thus, suitable for the separation of molecules 3094 Analytical Chemistry, Vol. 79, No. 8, April 15, 2007

Figure 2. Fractograms obtained by AsFlFFF (off-line) for (A) crude egg white and (B) lysozyme, ovalbumin, and ovotransferrin standards and crude egg white. Flow rates during the elution period were in (A) V˙ out 2.62 mL min-1 and V˙ cout 0.4 mL min-1 and in (B) V˙ out 0.40 mL min-1 and V˙ cout 2.64 mL min-1. Carrier, 8.5 mM phosphate buffer, 0.02% NaN3, 150 mM NaCl, pH 7.4; relaxation focusing, flow rate at inlet 0.1 mL min-1; flow inward from outlet 3.3 mL min-1; injection, 1 mL min-1 for 2-5 min; relaxation time, 12 min. UV detection at 280 nm. Peaks a-d in (A) correspond to diameters of 4, 5.5-6.0, 7.58.0, and 10.0-11.0 nm.

of sizes ranging from micro- to macromolecules. In addition, the porous structure means that the monolithic column has much lower flow resistance than a particle packed column. Thus, very high flow rates can be applied to accelerate the analysis. This is a particular benefit of the comprehensive setup.

Figure 3. Color plot of AsFlFFF×RPLC analysis of egg white sample. The area marked with an asterisk (/) has been zoomed to show the minor components. (a) is a slice of the RPLC separation at retention time 50-55 min showing the separation of compounds with sizes corresponding to 6.1-6.9 nm. (b) is a slice of the AsFlFFF separation at RPLC retention time of 2 min, showing the size separation of ovomucoid (omcd) and G3 ovoglobulin (og 3). Flow rates during the elution period were V˙ out 0.08 mL min-1 and V˙ cout 2.91 mL min-1. See the Experimental Section for further running conditions. Other abbreviations: lz, lysozyme; og 2, G2 ovoglobulin; ot, ovotranferrin; ova, ovalbumin; and ui, unidentified compound.

In the on-line combination, the carrier in the AsFlFFF needs to be suitable for RPLC separation as well, and fully aqueous phosphate-saline buffer (pH 7.4) proved to be appropriate. As this is a weak eluent for the RPLC separation, efficient reconcentration of the transferred fractions took place during the transfer from AsFlFFF to RPLC, and relatively large fractions could be handled without significant peak broadening. The flow rate of the carrier is critical in the on-line connection, because the size of the transferred fraction and the time required for the AsFlFFF separation are both dependent on it. Too high flow rate will result in large fraction volumes, which can cause severe band broadening in the RPLC separation. On the other hand, very low flow rates of the carrier increase the time required for the AsFlFFF fractionation. Here, the inlet flow was optimized to 0.08 mL/min, which still resulted in a reasonable fraction volume (400 µL), and the time of AsFlFFF was not excessive. The cross-flow rate was set to 2.91 mL/min. In contrast to FFF, where no stationary phase is used, adsorption onto the LC column can be a problem and recoveries less than 70% are often encountered in RPLC analysis of egg white proteins.29 Such poor recoveries would cause memory effects to be seen in the next analysis. In the comprehensive AsFlFFF×RPLC setup, memory effects in RPLC could cause serious problems, since the next fraction is transferred to the LC directly after the previous one. To avoid memory effects, the eluent composition must be chosen carefully. It has been shown that the concentration of buffer may be critical for the recovery,26,29 and the recovery can be increased by using a low concentration of TFA and acetonitrile as the organic solvent. With a fast gradient and high flow rate, the recovery of egg white proteins was good, and only a small fraction of proteins were left in the column (5-9%) even

when a high-concentration sample (20 µL of 1:3 diluted egg white, v/v) was injected into the column. With the on-line setup, the recoveries were even better. The RPLC separation was optimized to obtain as fast and efficient a separation of the FFF fractions as possible. A fast linear gradient program was applied, and the total analysis time was ∼4 min. The FFF peaks are wide, so a cycle time of 5 min was chosen. It would be possible to increase the modulation time, i.e., shorten the RPLC separation time by increasing the flow rate in LC further. In a short monolithic column, the flow rate could be increased up to 3-4 mL/min if required. However, this would also increase the eluent consumption substantially. Analysis of Egg White by Comprehensive AsFlFFF×RPLC. An example of an AsFlFFF×RPLC separation of crude egg white is shown in Figure 3. Figure 3a shows the color plot of the whole analysis with AsFlFFF retention in the x-axis and RPLC separation in the y-axis, with intensities of the peaks presented by color. Areas marked with an asterisk have been zoomed separately to show the minor components better. As an example, Figure 3a shows the RPLC separation from a AsFlFFF fraction at retention time of 50-55 min, corresponding to a size range of 6.1-6.9 nm. In Figure 3b, a part of the AsFlFFF separation at RPLC retention time of 2 min is shown, presenting the size separation of ovomucoid (omcd) and G3 ovoglobulin (og 3). Altogether, 12 compounds were separated. The peaks were identified by reference to standards (ovalbumin, ovotransferrin, lysozyme) and by comparison of peak areas, particle sizes, and retention data with values given in the literature. The two main peaks corresponded to ovalbumin and ovotransferrin with RPLC retention times of 3.18 and 2.58 min, respectively. Their calculated diameters were 5.5-6.0 and 7.5-8.0 nm, respecAnalytical Chemistry, Vol. 79, No. 8, April 15, 2007

3095

Figure 4. AsFlFFF determination of egg white proteins after heating at 25, 55, 63, 68, 73, and 80 °C for 120 min using phosphate-saline buffer. Flow rates during the elution period: (A) V˙ out 0.4 mL min-1 and V˙ cout 2.63 mL min-1; (B) V˙ out 2.62 mL min-1 and V˙ cout 0.4 mL min-1. (A) shows nonaggregated particles and (B) aggregated particles. Carrier, 8.5 mM phosphate buffer, 0.02% NaN3, 150 mM NaCl, pH 7.4; relaxation focusing, flow rate at inlet 0.1 mL min-1; flow inward from outlet 3.3 mL min-1; injection, 1 mL min-1 for 3-5 min; relaxation time, 12 min. UV detection at 280 nm.

Table 1. Relative Amounts and Particle Sizes of Proteins in Egg Whitea compound

amount (%)

diameter (nm)

ovalbumin ovotransferrin ovomucoid lysozyme G2 ovoglobulin G3 ovoglobulin unidentified 1 (ui1) unidentified 2 (ui2) unidentified 3 (ui3) unidentified 4 (ui4) unidentified 5(ui5) unidentified 6 (ui6)

57.8 14.5 2.9 0.5 0.5 1.1 0.3 0.8 0.1 0.3 0.2