626 BIOCHEMICAL AND SYNTHETIC POLYMER SEPARATIONS 80% ACN 75% ACN 70% ACN 12 3 4 5 6 7 12 3 4 5 6 7 8 9 5 13 12 11 10 9 8 7 6 1-4 0 5 10 15 (min)20 Figure 13.34 Isocratic HILIC separations of a homologous mixture of 3-hydroxy-2-nitropyridinyl-β- D-maltooligoglycosides. Conditions: 200 × 4.6-mm Poly- Hydroxyethyl A column; acetonitrile-water mobile phases; 2.0 mL/min. Numbers indicate degree of polymerization of each compound. Adapted from [75]. Polymer-based strong cation-exchange columns [122] and hydrophilic size-exclusion columns [123] have also been used for the separation of carbohy- drates, with the usual HILIC mobile phases (acetonitrile-water mixtures). These separations are not based on either ion exchange or size exclusion, as evidenced by increased retention with increasing concentration of organic solvent. 13.6.2 Ion-Moderated Partition Chromatography Sulfonated polystyrene-divinylbenzene resins are used for the separation of a wide variety of mono- and oligosaccharides, as well as mixtures of carbohydrates with alcohols and other small molecules [124, 125]. Separation is based on a combination of size exclusion and ligand exchange. Ligand exchange involves transition-metal ions that are tightly held by the resin sulfonic-acid groups; the metal ion then provides a positive charge that interacts with a very slight negative charge on the sugar molecule (the ‘‘ligand’’). For oligosaccharide separations, the primary mechanism is size exclusion. Resins with a low percentage of cross-linking are preferred, in order to allow penetration of the oligosaccharides into the packing; Figure 13.35 shows the separation of oligosaccharides in a corn syrup. For monosaccharides, ligand exchange of the sugar hydroxyls with the fixed counter-ion on the resin is the primary mechanism. 13.6 SEPARATION OF CARBOHYDRATES 627 (min) 1 2 3 4 5 6 7 8-11 30 20 10 Figure 13.35 Analysis of oligosaccharides in corn syrup by ion-moderated partition chro- matography on a 4% cross-linked strong cation exchange resin in the silver form. Sample: corn syrup (glucose, 1; 2–11, Dp 2–11). Conditions: 300 × 7.8-mm Aminex HPX 42-A column; water as mobile phase; 85 ◦ C; 0.4 mL/min. Adapted from [124]. A mechanism for ligand exchange has been proposed [126]: carbohydrate hydroxyls exchange with water molecules held in the hydration sphere of the fixed cation. The stability of the solute-cation complex increases with increased availabil- ity for coordination, and carbohydrate retention increases with the stability of the complex. Ligand-exchange selectivity is dependent on the nature of the counter-ion and the size, and on the structure and stereochemistry of the carbohydrate. Sep- arations are isocratic with water or dilute solutions of sulfuric acid (5 − 10 mM) as mobile phase, usually at elevated temperatures (40–85 ◦ C). Recommended resin cross-linkages and ionic forms for particular applications are listed in Table 13.7. Table 13.7 Strong Cation-Exchange columns for Ion-Moderated Partition Chromatography of Carbohydrates Cross-Linkage Ionic Application Percentage Form 8 Calcium Monosaccharides and class separation of di-, tri-, and tetrasaccharides 8 Lead Pentoses and hexoses in wood products 8 Hydrogen Carbohydrates in solution with fatty acids, alcohols, and ketones 8 Sodium Sugars in samples with high salt 8 Potassium Mono-, di-, and trisaccharides in corn syrup and brewing wort 4 Silver Oligosaccharides 4 Calcium mono- and disaccharides in starch hydrolysates 628 BIOCHEMICAL AND SYNTHETIC POLYMER SEPARATIONS Note that these are fixed-ion resins; the column is converted to a specific form by the manufacturer before packing and is maintained in that form for the life of the column. In-column conversion from one form to another is not recommended, since resins can shrink and swell with changes in ion form, leading to likely column failure. Sodium-form columns are useful for samples containing high salt concentra- tions (e.g., molasses), potassium-form columns are useful for analysis of corn syrup, silver-form columns provide good selectivity for oligosaccharides, and calcium-form columns are used for analysis of starch hydrolysis products. 13.6.3 High-Performance Anion-Exchange Chromatography Neutral carbohydrates are not retained on IEC columns under usual conditions, but they are weak acids that can partially ionize at pH > 12 (Table 13.8). Their separation can be achieved under alkaline conditions by means of polymer-based anion-exchange columns [127, 128]. Commercially available columns for high-performance anion exchange (HPAE) are based on polystyrene/divinylbenzene (for monosaccharide separations) or ethylvinylbenzene/divinylbenzene (for oligosaccharide separations). Both supports consist of nonporous particles covered with a fine layer of sulfonated latex microbeads (Fig. 13.36). Sugar alcohols are weaker acids than nonreduced sugars and require a high-capacity ion exchanger for their separation by HPAE. For the latter application a macroporous vinylbenzene-chloride/divinylbenzene functionalized with alkyl quaternary groups is used. Monosaccharides (including neutral and amino sugars) can be separated isocratically using a mobile phase of dilute aqueous sodium hydroxide. Separation of acidic carbohydrates (sialic acid, sialyated and phosphorylated oligosaccharides) can be achieved using sodium hydroxide/sodium acetate mobile phases, either isocratically or with a sodium acetate gradients. Oligo- and polysaccharides (including high-mannose, hybrid, and complex oligosaccharides) can be separated using sodium hydroxide or sodium hydroxide/sodium acetate gradients (Fig. 13.37). See also the carbohydrate separation of Figure 7.21. HPAE is typically coupled with pulsed-amperometric detection (PAD). At high pH, carbohydrates are electrochemically oxidized at the surface of a gold electrode Table 13.8 Dissociation Constants for Common Carbohydrates Sugar pK a Fructose 12.03 Mannose 12.08 Xylose 12.15 Glucose 12.28 Galactose 12.39 Dulcitol 13.43 Sorbitol 13.60 α-Methyl glucoside 13.71 13.6 SEPARATION OF CARBOHYDRATES 629 5 μm SO 3 − SO 3 − SO 3 − SO 3 − SO 3 − SO 3 − SO 3 − NR 3 + NR 3 + NR 3 + R 3 N + R 3 N + NR 3 + NR 3 + NR 3 + R 3 N + R 3 N + NR 3 + NR 3 + NR 3 + R 3 N + R 3 N + micro-particle surface anion-exchange nano- p articles Figure 13.36 Pellicular anion-exchange-resin bead (Dionex). The bead structure consists of a nonporous sulfonated microparticle with surface-associated latex nanobeads that are func- tionalized with a strong anion-exchanger. The microparticles range in diameter from 5.5 to 10 μm, depending on the column type. The nanobeads range in diameter from 43 to 275 nm, also depending on column type. The nanobeads are immobilized on the microparticle surface by electrostatic interactions of the ion-exchange groups. Adapted from [127]. 1. disialylated, triantennary 2. disialylated, triantennary 3. trisialylated, triantennary 4. trisialylated, triantennary 5. tetrasialylated, triantennary 6. tetrasialylated, triantennary 7. trisialylated, triantennary 1 3 2 4 7 01020304050 (min) 5 6 Figure 13.37 Separation of bovine fetuin N-linked oligosaccharide alditols by high-performance anion-exchange chromatography, using a gradient of 0.0–0.5M sodium acetate in 100-mM sodium hydroxide. Reprinted from [128] with permission from Dionex Corp. by application of a positive potential [129]. The measured current from this reaction is proportional to carbohydrate concentration, with detection limits in the low picomole range. The reaction produces oxidation products that poison the electrode surface, so the detector is pulsed to a high voltage to clean the surface, then pulsed again to a low voltage to reduce gold oxide back to gold. 630 BIOCHEMICAL AND SYNTHETIC POLYMER SEPARATIONS 13.7 SEPARATION OF VIRUSES The large size of the virus particle represents an important consideration in their chromatographic separation, one that has several consequences. First, viruses can be expected to display large values of S (reversed-phase) or m (ion exchange), such that small changes in mobile phase concentration will have large effects on their isocratic retention [3]. Therefore the isocratic separation of viruses is impractical for all modes of chromatography, except size-exclusion chromatography. In gradient elution, a virus will elute at a discreet point in the gradient, in most cases with k ∗ ≈ 0. The second consequence of viral size is slow diffusion. For example, D m for adenovirus is 5 × 10 −8 cm 2 /s, about 10-fold slower than for a large protein. Slow diffusion should result in low values of N ∗ and broad peaks. However, peak broadening due to slow diffusion is offset by the small value of k ∗ (Eq. 9.5, assuming a very large value of S), and observed peak widths for viruses are similar to those for proteins. The third consequence of viral size is their large hydrodynamic volume, which restricts their entry into the pore system of the packing. For this reason viral particles interact with only a small fraction of the total column surface area, and column capacities for viruses are 20- to 50-fold lower than for proteins. Ion exchange, hydrophobic interaction, and metal-chelate chromatography have all been used for viral purification [130]. However, ion-exchange chromatog- raphy is most often employed as the capture step in virus purification because of its satisfactory yield and purity [3]. In the case of adenoviruses the isoelectric point of the surface is IP ≈ 6, so anion-exchange chromatography is the preferred procedure for separating viruses. Additives such as sucrose, magnesium, and glycerol that are used to stabilize the virus are compatible with anion-exchange conditions. Both strong and weak anion-exchange columns have been employed for ade- novirus purification, based on cross-linked polystyrene-divinyl benzene, hydrophilic polymer, polyacrylamide, or cross-linked agarose and dextran supports [5]. Mono- liths have been used for viral separations [3], although these columns are not commercially available in the larger sizes that are required for large-scale purifica- tion. The mobile-phase pH should be ≈2 units above the isoelectric point of the virus, in order to avoid viral aggregation; pH values between 7.5 and 9 have been used for adenovirus purification. Sodium chloride gradients are generally used, with salt concentration of 0–0.3M for sample loading and ≥ 0.6M for elution. In a study of static binding capacity as a function of ionic strength, it was found that capacity passes through a maximum at about 0.3M. Loading of the virus at or slightly above 0.3M is recommended, in order to allow contaminating proteins to be washed from the column during the loading step [3]. Elevated column temperatures should be avoided, in order to minimize any loss of viral activity. The advantage of IEC for adenovirus purification is its ability to distinguish aggregated and disrupted forms from the intact virus [130]. However, the removal of empty capsids from intact virus is less certain [5, 130]. The resolution of the early-eluting p53 adenovirus and late-eluting DNA is shown in Figure 13.38. For further details on virus separations, see [3]. 13.8 SIZE-EXCLUSION CHROMATOGRAPHY (SEC) 631 0 102030min 280 nm 260 nm virus (ACN53) DNA Figure 13.38 Separation of adenovirus by anion-exchange chromatography. Separation con- ditions: column, 50 × 6.6-mm Fractogel DEAE-650 M column; gradient, 300–600-mM NaCl (50-mM Tris, pH-8.0 plus 2-mM MgCl and 2% sucrose) in 10 min. Adapted from [130]. 13.8 SIZE-EXCLUSION CHROMATOGRAPHY (SEC) Size-exclusion chromatography (SEC) separates compounds according to their molecular size in solution, as a result of the exclusion of larger molecules from smaller pores in the column packing [131]. When applied to synthetic polymers with organic solvents as mobile phase and polymeric column packings (Section 13.10.3.1), the technique is referred to as gel permeation chromatography (GPC). When sep- arating biopolymers such as proteins with aqueous buffers as mobile phases and hydrophilic column packings, the technique is termed gel filtration.Gelfiltration can be used either as a preparative tool to isolate biologically active species (often in concert with other chromatographic techniques in a multi-stage purification pro- cess), or as an analytical tool to obtain information about solute molecular size or shape, aggregation state, or the kinetics of ligand-biopolymer binding. Historically gel filtration has employed soft gels such as dextrans, agarose, or polyacrylamide [132–134]). These packings are compressible and therefore are only compatible with mobile-phase flow by means of gravity or low-pressure pumps. Soft gels may be stabilized by cross-linking, in which case they can be used with higher flow rates and pressures of a few hundred psi. Analytical gel filtration is most often carried out with rigid supports: a silica matrix modified with a hydrophilic stationary phase or cross-linked organic polymers. These materials are mechanically stable at pressures of a few thousand psi or higher, and can be used with HPLC systems. 632 BIOCHEMICAL AND SYNTHETIC POLYMER SEPARATIONS 13.8.1 SEC Retention Process SEC is the simplest form of chromatography, in which retention depends only on the relative penetration (or ‘‘permeation’’) of solute molecules into and out of the pores of the stationary phase; molecules are separated on the basis of their size in solution (for polymers of the same chemistry and architecture, this size correlates with molecular weight). In contrast to other modes of chromatography such as RPC or IEC, in which solutes are retained by interacting with the stationary phase, SEC (under ideal conditions) involves no interaction of solute and stationary phase. Molecules that are too large to enter any of the pores elute in a volume of mobile phase that is equal to the interstitial volume between the stationary-phase particles (V 0 ). Molecules that are small enough to freely enter all of the pores elute in a volume equal to the interstitial volume plus the volume of the pore system (V i ). Molecules of intermediate sizes enter some fraction of the pore system, depending on their size or shape, and elute between V 0 and V 0 + V i . The total mobile-phase volume or dead-volume V m can be expressed as the sum of the interstitial volume and the pore volume: V m = V 0 + V i (13.6) The extent to which a solute can penetrate the pore system is governed by its distribution coefficient K D , which is related to its elution volume V R by K D = V R − V 0 V i (13.7) Equations (13.6) and (13.7) can be combined: V R = V 0 + K D V i (13.8) From the expression above it can be seen that molecules too large to enter the pores will all have K D = 0 and will co-elute at V 0 . Similarly all molecules small enough to freely penetrate the entire pore system will have K D = 1 and co-elute at V m . Molecules of intermediate size will have K D values between zero and one and will be separated according to size, with larger molecules eluting before smaller molecules. The relationship between molecular size and retention volume can be used to estimate the molecular weight M of an analyte. A calibration plot of log M versus retention volume (or K D ) will exhibit an approximately linear segment between V 0 and V i , as in Figure 13.39a (solid portion of curve). The range in sample molecular weights corresponding to this linear segment is referred to as the fractionation or separation range. If the plot is constructed using standard proteins whose shapes are similar to that of an analyte protein, the retention time of the analyte (or retention volume as in Fig. 13.39b) will correspond to a specific molecular weight on the plot, allowing an estimate of its molecular weight. The relationship between log M and K D is nearly linear for K D values between about 0.2 and 0.8, but with curvature at the ends of the plot (as in the SEC calibration curve of Fig. 13.39a). In theory, SEC column packings with a single, absolute pore size will provide a separation that spans 1.5 decades of molecular weight. In practice, however, this separation range 13.8 SIZE-EXCLUSION CHROMATOGRAPHY (SEC) 633 0 KD = 0.0 0.4 0.8 1.0 (a) (b) Total exclusion Fractionation range V m V m V 0 log M Total permeation V 0 V i Volume Volume Figure 13.39 Hypothetical SEC calibration curve (a) and chromatogram (b). Adapted from [135]. will cover roughly two decades or a little more, since the size of the pores in a given packing material will vary somewhat. Although SEC is often used to estimate protein molecular weight, it should be understood that retention is actually determined by the hydrodynamic diameter of the solute, which is only indirectly related to molecular weight. The hydrodynamic diameter of a protein (or other molecule) is related to its radius of gyration or Stokes radius, and this can vary with solute hydration and molecular shape. Two proteins with similar molecular weights but different shapes (e.g., spherical vs. oblate vs. rod-like, or native vs. denatured) will have different hydrodynamic diameters and therefore display significantly different retention volumes (Fig. 13.40). To obtain accurate molecular-weight estimates using gel filtration, it is necessary that the proteins used to construct the calibration plot and the analytes all have similar shapes. An alternative approach is to perform calibration and analysis under denaturing conditions, so that both calibrant and analyte proteins are converted to linear random-coil conformations, with retention times that better correlate with molecular weight. 13.8.2 Columns for Gel Filtration The column packings used for SEC must be as inert as possible, so as to minimize any interactions with analytes (which would negate a relationship between solute 634 BIOCHEMICAL AND SYNTHETIC POLYMER SEPARATIONS The hydrodynamic diameter of a molecule is defined by a sphere with a diameter equal to the length of the molecule. hydrodynamic diameter Sphere MW = 10,000 Da hydrodynamic diameter Rod MW = 10,000 Da Figure 13.40 Molecular ‘‘size’’ and molecular shape compared. Adapted from [135]. molecular weight and SEC retention). For gel filtration, this is achieved by the use of hydrophilic packings whose interactions with the aqueous mobile phase are stronger than with protein solutes. The pore-volume of the packing should be as large as possible, and the support material should be mechanically stable for the required flow rates and pressures. Packings of different pore sizes may be needed for proteins of different molecular weight. Smaller pores are used for small molecules, larger pores for larger molecules. If a sample contains proteins of widely different molecular weights, two columns of different pore size can be connected in series (the order is not important). To obtain the greatest fractionation range, the connected columns should have packings with pores about 10-fold difference in size, and the two columns should be closely matched in terms of efficiency or values of H (Section 2.4.1). 13.8.2.1 Support Materials Two types of materials are used for SEC columns: hydrophilic bonded silicas and hydrophilic organic polymers. Silica is the most widely used material for HPLC packings because of its mechanical stability, acceptable porosity, and availability in a range of pore sizes. However, the silica surface interacts strongly with proteins, so it must be derivatized. The most common approach is to react surface silanols with an organosilane reagent to form a diol-type or carbohydrate-like coating that is covalently attached to the silica. A limitation of silica-based SEC packings is their instability under alkaline conditions. Silica dissolves at pH values above 8, leading to reduced column lifetime. One manufacturer (the Zorbax GF columns from Agilent Technologies) uses a zirconyl cladding to stabilize the silica support for operation above pH-8. Because of silica’s limitations, several manufacturers offer gel-filtration columns based on hydrophilic organic polymers. These include polymethacrylate supports, 13.8 SIZE-EXCLUSION CHROMATOGRAPHY (SEC) 635 proprietary hydrophilic polymers, and semi-rigid cross-linked agaroses and dextrans. These columns are more stable under high-pH operation but are less efficient and less able to tolerate high pressures. 13.8.2.2 Pore Size and Porosity High-performance SEC packings are available in pore sizes ranging from 10 to 400 nm. A column should be selected with a pore size such that solutes of interest elute within 0.2 ≤ K D ≤ 0.8. Manufacturers provide calibration plots in their product literature for this purpose; see the example of Fig. 13.41 for two different gel-filtration columns (the x-axis in Fig. 13.41 is in units of retention time, corresponding to a specified flow rate). In this case the approximate fractionation range of the GF-250 column is 4000 ≤ M ≤ 200, 000 Da, while that of the GF-450 column is 10,000 ≤ M ≤ 1,000,000 Da. By connecting the two columns in series, the fractionation range would be much wider: 4000 ≤ M ≤ 1,000,000 Da. Columns (of the same dimensions) with a narrow pore-size distribution will be characterized by higher resolution over a narrow range of analyte molecular weights, for example, 1000 ≤ M ≤ 30, 000. That is, such columns will exhibit a calibration plot with a shallow slope and a reduced fractionation range. Columns with a wide pore-size distribution will be characterized by a wider fractionation range, that is, a calibration plot with a steep slope. Sample resolution or the ability to separate two solutes of different molecular weight increases for a shallow slope of log M versus retention time or volume. Thus the choice of a particular column for a given sample represents a compromise between resolution and retention range. M 6 8 10 12 14 Retention time ( min ) GF-250 GF-450 IgM Thy β-Gal IgG BSA Oval Myo NaN 3 10 7 10 6 10 5 10 4 10 3 10 2 Figure 13.41 Calibration curves of proteins for GF-250 and GF-450 columns; 0.2M sodium phosphate (pH-7.5) mobile phase (nondenaturing conditions). Reprinted with permission of Agilent Technologies, Inc. . preparative tool to isolate biologically active species (often in concert with other chromatographic techniques in a multi-stage purification pro- cess), or as an analytical tool to obtain information. too large to enter any of the pores elute in a volume of mobile phase that is equal to the interstitial volume between the stationary-phase particles (V 0 ). Molecules that are small enough to. entry into the pore system of the packing. For this reason viral particles interact with only a small fraction of the total column surface area, and column capacities for viruses are 20- to 50-fold