EP0595993A1 - Isolation of charged particles from fluids - Google Patents

Abstract

Process for the separation, from fluids such as biological fluids, and using an ion exchange medium placed on a porous membrane, of charged molecules such as proteins. Fat cells and particulate matter are prevented from crossing the membrane to prevent the ion exchange medium from clogging.

Description

ISOLATION OF CHARGED PARTICLES FROM FLUIDS

This invention relates to a process for the separation of charged molecules from a fluid using ion exchange media. The invention is particularly suitable for the extraction of protein components from biological fluids such as milk and milk products. It will be convenient to hereinafter describe the invention with particular reference to the extraction of protein components from milk and milk products but it is to be understood that the invention is not limited thereto.

Milk is a fluid secreted by all species of mammals to supply nutrition, and immune and non-immune protection to the young. Milk consists of water, proteins, fat, carbohydrates, salts, vitamins and a variety of miscellaneous components. Both young and mature humans consume large amounts of bovine milk, and the fluid thus has both nutritional and commercial significance. In broad terms, bovine milk proteins (30 - 35 g/L) consist of caseins (approximately 80%), whey proteins (approximately 20%) and a number of minor protein/enzyme constituents. Continued and expanded use of milk protein fractions has been hindered by the absence of quick and efficient isolation techniques which may be used on a commercial scale.

Whey, the yellow-green liquid that separates from the curd during the manufacture of cheese and acid casein, has long been considered a waste by-product in the dairy industry. The protein in whey accounts for about 20% of total milk protein. The primary proteinaceous constituents of whey are β-lactoglobulin and α-lactalbumin, two small globular proteins that account for some 70 - 80% of total whey protein. Minor protein components include the glycomacropeptide, serum albumin, lactoferrin, immunoglobulins, phospholipo- proteins, and a number of enzymes (including lactoperoxidase). Manufacture of spray dried whey powder and whey protein concentrate (WPC) has realised only a small portion of the potential of these proteins. It would be advantageous to isolate the individual whey proteins as it is believed that the individual whey proteins will yield products with increased nutritional, functional and biological value to the food and other industries, and increased commercial value to the dairy industry. For example, minor components such as lactoferrin and lactoperoxidase have the potential to serve as natural antibacterial agents, and thus enter areas of both human and veterinary medicine. Moreover, since lactoferrin is found in high concentration in human milk (approx. 2 g/L), a supply of purified bovine lactoferrin will facilitate the preparation of new generation infant formulae.

Because of their low concentration, isolation of minor whey protein constituents invariably involves the processing of large quantities of whey or milk. In the past this usually involved employing column or batch-wise chromatographic techniques. The isoelectric point (pi) of both lactoferrin and lactoperoxidase is greater than 9.0 while the majority of whey proteins have isoelectric points around 5.1 to 5.4. Casein has an isoelectric point of 4.6. Cation exchange chromatographic procedures have been described for the isolation of minor protein/peptide components (primarily lactoferrin and lactoperoxidase) from cheese whey. These procedures rely upon adsorption of the protein components by an appropriate cation exchange resin, usually contained in a traditional packed-bed column.

Belgian Patent Specification 901672 describes an alternative ion exchange technique based on a calcium alginate medium in which ion exchange functionality has been obtained by admixture of oxides of zirconium, titanium, silicon (quartz) or aluminium. The milk or whey is mixed with the ion exchange medium in a stirred tank whereby proteins having an isoelectric point above 7.5 are adsorbed to the ion exchange medium. After equilibration, the ion exchange gel is separated mechanically from the milk or whey, washed and diluted with calcium chloride. The Belgian process adopts this unusual methodology because conventional ion exchange columns tend to become quickly fouled using a milk product feedstock. Fat, casein fines, and other particulates, components normally found in pasteurised/separated whey, pose a problem in traditional column based chromatographic procedures as they act as column foulants, both reducing the resin effectiveness and the recovery of the product.

The problem of clogging of traditional ion exchange columns is addressed in International Patent Application PCT/SE88/00643 (WO89/04608), in which a process is described for extracting pure fractions of lactoperoxidase and lactoferrin from milk serum using a strong cation exchange bed. To avoid the clogging problem this application requires a cross-flow microfiltration of the milk by-product before contacting it with the ion exchange bed. The cross-flow microfilter used has a pore size of 1.4 microns.

Although the microfiltration step assists in avoiding clogging of the cation exchange bed, the process described in PCT/SE88/00643 still suffers from the shortcomings of column ion exchange chromatography, i.e. the requirement for very expensive column hardware and ion exchange resins, the lengthy procedures associated with resin preparation and clean-up, and the necessity to avoid the column running dry. The additional processing step of cross-flow microfiltration also involves additional equipment and processing time.

It has now been found that the benefits of ion exchange chromatography can be achieved in a single process step wherein a fluid is passed through a porous membrane which serves as a support for an ion exchange medium.

In accordance with the present invention there is provided a process for the separation of charged molecules from a fluid comprising providing an ion exchange medium disposed on a porous membrane, passing the fluid through the membrane, wherein said charged molecules are preferentially adsorbed on the medium, and eluting the adsorbed molecules from the medium.

The above process may be used for the separation of any charged molecules from fluids. The term "molecule" when used herein also encompasses aggregates of such molecules. The process is particularly suitable for treatment of biological fluids such as milk or milk products, blood or blood plasma, or other types of fluids such as fermentation fluids, fluids from cell culture, etc.

When used herein the term "milk or milk product" includes milk products such as skim milk, whey, colostrum etc. The process may be used to isolate cationic protein components such as lactoferrin, lactoperoxidase, growth promoting agents and lysozyme from milk or milk products. The process may also be used to isolate other charged molecules such as α-lactalbumin, glycomacro- peptide, serum albumin, immunoglobulins and enzymes. The process of the present invention may also be used for isolating charged molecules from other biological fluids for instance blood and blood products such as plasma. For example clotting factors, serum albumin and immunoglobulins may be isolated from blood or plasma.

The present invention may also be used for processing other fluids such as fermentation fluids or fluids from cell culture wherein pharmaceuticals, vitamins, hormones or other therapeutic proteins may be isolated from the fluid.

By the use of the process of the present invention, large quantities of fluid such as milk or milk products can be effectively filtered and minor protein species which may be difficult to isolate using traditional chromatographic processes, may be extracted in a single operation. Fat globules and proteinaceous aggregates which would clog traditional ion exchange media are prevented from passing through the porous membrane and are retained on the retentate side of the membrane thus allowing the ion exchange medium to trap the useful protein species, such as lactoferrin and lactoperoxidase.

It has been found that ion exchange media disposed on membranes with a thickness of the order of microns or millimetres can extract an effective yield of certain protein species from a fluid such as milk. In the past, it has been assumed that long ion exchange columns and long residence times were required to achieve an effective yield of such species.

Suitable pore sizes for the ion exchange membrane range from about 0.1 to about 1.2 microns, preferably about 0.2 to about 0.6 microns with about 0.4 microns being most preferred. As the present process does not use an easily fouled ion exchange column the pore size of the membranes used in the process of the present invention may be a little larger than those used in the cross-flow microfiltration process adopted in PCT/SE88/00643 as it is not critical to exclude all particulate matter with the membrane. Larger pores may assist in achieving higher flow-through rates in the present invention. If desired the fluid may be subjected to a microfiltration step prior to being passed through the membrane.

The fluid may be passed through the membrane either by means of a dead-end filtration technique or a cross- flow filtration technique. Conventional dead-end or static pressure filtration involves forcing a feed material against and through a vertical filter, whereas cross-flow filtration involves the passage of a feed material through a narrow gap between two parallel filters, the material passing across the filter surface at a high linear flux.

As with other ion exchange media, the membranes useable in the present invention can be quantitatively eluted using solutions of successively higher salt concentrations, and/or by successive changes in the pH of an eluting solution to shift the pH of the medium above or below the isoelectric point of the desired charged molecule such as a protein. Thus differences in pi or other binding parameters between the various charged molecules can be exploited to produce relatively pure fractions of each charged molecule as described, for instance in PCT/SE88/00643.

Preferential binding of particular proteins may also be exploited to isolate one particular protein from a fluid in preference to other proteins. For example lactoferrin binds more tightly to the membrane than lactoperoxidase and displaces lactoperoxidase from the membrane. Accordingly the membrane may be saturated with lactoferrin and a relatively pure fraction of lactoferrin may be isolated from a milk product containing both lactoferrin and lactoperoxidase.

Elution of the charged molecules can be commenced immediately after passing the fluid through the membrane i.e. before cleaning the upstream surface of the membrane of matter such as fat globules and proteinaceous debris. Alternatively the membrane can be cleaned of such matter for instance with a wash prior to introducing the eluting medium.

Unlike traditional ion exchange columns the membranes useable in the present invention may avoid problems with swelling or packing of the exchange matrix and maintenance is easier. For example, if an ion exchange membrane runs dry, it does not have to be re-packed. Furthermore sanitisation or sterilisation methods are more diverse and adaptable to existing procedures in the dairy industry such as steaming, dairy detergents or alkaline solutions such as sodium hydroxide.

Either a strong or weak cation or anion exchange medium may be used. The terms "strong" and "weak" in the context of ion exchange functional groups refer to the extent of ionisation of the group with pH of the medium. Strong ion exchange functional groups are totally ionised over a wide range of pH values. A strong cation exchange functional group (e.g. sulphopropyl) is totally ionised ( eprotonated) at pH values above 2. The exchange medium may be selected to ensure binding of the desired charged molecules. For example because the pi of lactoferrin and lactoperoxidase is around 9.5 whereas the majority of whey proteins have a pi below 5.4, a strong cation exchange medium should be selected to ensure binding of protein species with high pi values. In this case, a suitable strong cation exchange media comprises sulfonic acid functional groups disposed on a symmetrical polya ide support.

Such membranes may be prepared in a two step procedure by grafting a polymer onto an inert microporous membrane followed by the introduction of the desired cationic or anionic functional groups. It will thus be apparent that wide variations in pore size and amount of grafted polymer are feasible. The membrane useable in the present invention can be provided in the form of convenient cross-flow cartridges (modules) or dead-end f lters.

The effluent which is produced as a by-product of the process of the present invention may also have advantageous properties. For example extraction of the cationic proteins lactoferrin and lactoperoxidase from cheese whey results in a whey product stream essentially unaltered from that used as the feed material, because these cationic proteins represent less than 3% of the total whey protein. In addition, the process also ensures that this whey product stream is free of particulates and very low in microorganisms and fat. Thus, this whey product stream may be further processed into a low-fat whey protein concentrate powder with advantageous properties including high solubility.

Preferred embodiments of a process in accordance with the invention will now be described by way of example only with reference to the accompanying drawings in which: Figure 1: Schematic representation of the experimental setup for cross-flow membrane ion exchange processing using both purified proteins and milk-derived products. Data parameters reported herein are defined as follows: Binding capacity (mg/cm²) = {(c_p x v_p) + (c_r x v_r)}_e^_uate/module surface area (cm²); where c = concentration of solute (mg/L), and v = volume (L). Recovery (%) = [{(c_p x v_p) + (c_r x v_r)}_load + {(c_p x v_p) + (c_r x v_r)}_eluate] x 100/(Ci x v_±); where c = concentration of solute (based on activity for lactoperoxidase and on HPLC analysis for lactoferrin), v = volume (L), p = permeate, r = retentate, and i = initial feed material.

Figure 2: Schematic representation of the experimental setup for dead-end membrane ion exchange processing using both purified proteins and milk-derived products. Data parameters reported herein are defined as follows: Binding capacity (mg/cm²) = (c_e x v_e)/filter surface area (cm²); where c = concentration of solute (mg/mL), and v = volume (mL). Elution (%) = (c_e x v_e) x 100/{(Ci x v^) - (c_f x v_f) - (c_w x v_w)}, and Recovery (%) = {(c_f x v_f) + (c_w x v_w) + (c_e x v_e)} x 100/(c_± x v_±); where c = concentration of solute (based on activity for lactoperoxidase and on HPLC analysis for lactoferri ), v = volume (mL), f = filtrate, w = wash, e = eluate, and i = initial feed material.

Figure 3: Permeate flux (* ), lactoperoxidase activity (•), and lactoferrin concentration (■ ) during cross-flow membrane ion exchange filtration of pure proteins in sodium phosphate buffer, pH 6.7. Starting levels in feed material: lactoperoxidase, 5.3 IU/mL; lactoferrin, 100.0 mg/L. Figure 4: HPLC chromatogram of samples from cross- flow membrane ion exchange filtration of pure proteins in sodium phosphate buffer, pH 6.7. a, feed material; b, permeate after processing 107 L; c, 0.2 M NaCl eluate (1:24 dilution). Figure 5: SDS-PAGE electrophoretogram of samples from cross-flow membrane ion exchange filtration of pure proteins in sodium phosphate buffer, pH 6.7. Lane 1, initial feed material; lane 2, retentate during loading; lane 3, permeate during loading; lane 4, 0.2 M NaCl eluate; lane 5, 0.4 M NaCl eluate; lane 6, 1.0 M NaCl eluate; lane 1 , low molecular weight protein markers. Figure 6: Permeate flux ( -), lactoperoxidase activity (•), and lactoferrin concentration (■) during cross-flow membrane ion exchange filtration of microfiltered Cheddar cheese whey, pH 6.0. Starting levels in whey feed material: lactoperoxidase, 7.8 IU/mL; lactoferrin, 81.4 mg/L.

Figure 7: HPLC chromatogram of samples from cross- flow membrane ion exchange filtration of microfiltered Cheddar cheese whey, pH 6.0. a, feed material; b, permeate after processing 37 L; c, 1 M NaCl eluate (1:9 dilution).

Figure 8: SDS-PAGE electrophoretogram of samples from cross-flow membrane ion exchange filtration of microfiltered Cheddar cheese whey, pH 6.0. Lane 1, initial feed material; lane 2, retentate during loading; lane 3, permeate during loading; lanes 4 and 5, 1 M NaCl eluate; lane 10, low molecular weight protein markers.

Figure 9: Permeate flux (A), lactoperoxidase activity (•), and lactoferrin concentration (■) during cross-flow membrane ion exchange filtration of microfiltered Cheddar cheese whey, pH 6.9. Starting levels in feed material: lactoperoxidase, 3.4 IU/mL; lactoferrin, 40.0 mg/L.

Figure 10: HPLC chromatogram of samples from cross- flow membrane ion exchange filtration of microfiltered Cheddar cheese whey, pH 6.9. a, feed material; b, permeate after processing 53 L; c, 1 M NaCl eluate (1:12 dilution).

Figure 11: SDS-PAGE electrophoretogram of samples from cross-flow membrane ion exchange filtration of microfiltered Cheddar cheese whey, pH 6.9. Lane 6, initial feed material; lane 7, permeate during loading; lane 8, retentate during loading; lane 9, 1.0 M NaCl eluate; lane 10, low molecular weight protein markers.

Figure 12: Permeate flux (A ), lactoperoxidase activity (•), and lactoferrin concentration ( ■ ) during cross-flow membrane ion exchange filtration of non- microfiltered Cheddar cheese whey, pH 6.2. A cross-flow module with a surface area of 0.6 m² was used. Starting levels in feed material: lactoperoxidase, 11.0 IU/mL; lactoferrin, 116.0 mg/L. Figure 13: HPLC chromatogram of samples from cross- flow membrane ion exchange filtration of non- microfiltered Cheddar cheese whey, pH 6.2. a, feed material; b, permeate after processing 25 L; c, 1 M NaCl eluate (1:15 dilution). Figure 14: SDS-PAGE electrophoretogram of samples from cross-flow membrane ion exchange filtration of non-microfiltered Cheddar cheese whey, pH 6.2. Lane 1, initial feed material; lane 2, retentate during loading; lane 3, permeate during loading; lanes 4 and 5, 1.0 M NaCl eluate; lane 10, low molecular weight protein markers.

Figure 15: Permeate flux (A), lactoperoxidase activity (•), and lactoferrin concentration (■) during cross-flow membrane ion exchange filtration of non- microfiltered Cheddar cheese whey, pH 6.2. A cross-flow module with a surface area of 0.5 m² was used. Starting levels in feed material: lactoperoxidase, 11.2 IU/mL; lactoferrin, 116.0 mg/L.

Figure 16: HPLC chromatogram of samples from cross- flow membrane ion exchange filtration of non-micro- filtered Cheddar cheese whey, pH 6.2. a, feed material; b, permeate after processing 26 L; c, 1 M NaCl eluate (1:5 dilution).

Figure 17: SDS-PAGE electrophoretogram of samples from cross-flow membrane ion exchange filtration of non-microfiltered Cheddar cheese whey, pH 6.2. Lane 1, initial feed material; lane 6, retentate during loading; lane 7, permeate during loading; lanes 8 and 9, 1.0 M NaCl eluate; lane 10, low molecular weight protein markers.

Figure 18: HPLC chromatogram of samples from dead-end membrane ion exchange filtration of microfiltered Cheddar cheese whey, pH 6.2. a, feed material; b, permeate after filtering 50 mL; c, 1 M NaCl eluate (1:3 dilution).

Figure 19: SDS-PAGE electrophoretogram of samples from dead-end membrane ion exchange filtration of microfiltered Cheddar cheese whey, pH 6.2. Lane 1, low molecular weight protein markers; lane 4, initial feed material; lane 5, filtrate (permeate) during loading; lane 6, 1.0 M NaCl eluate. Figure 20: HPLC chromatogram of samples from dead¬ end membrane ion exchange filtration of microfiltered Cheddar cheese whey, pH 7.0. a, feed material; b, 1 M NaCl eluate (1:3 dilution).

Figure 21: SDS-PAGE electrophoretogram of samples from dead-end membrane ion exchange filtration of microfiltered Cheddar cheese whey, pH 7.0. Lane 1, low molecular weight protein markers; lane 2, filtrate (permeate) during loading; lane 3, 1.0 M NaCl eluate; lane 4, initial feed material. Figure 22: HPLC chromatogram of samples from dead-end membrane ion exchange filtration of dialyzed and microfiltered Cheddar cheese whey, pH 7.0. a, feed material; b, permeate after filtering 30 mL; c, 1 M NaCl eluate (1:3 dilution). Figure 23: SDS-PAGE electrophoretogram of samples from dead-end membrane ion exchange filtration of dialyzed and microfiltered Cheddar cheese whey, pH 7.0. Lane 1, low molecular weight protein markers; lane 2, filtrate (permeate) during loading; lane 3, 1.0 M NaCl eluate; lane 6, initial feed material.

Figure 24: Filtrate (permeate) flux during dead-end filtration of microfiltered (A) and non-microfiltered (B) Cheddar cheese whey, pH 6.5 using a membrane ion exchange Sartobind S filter (5.4 cm², 0.45 μm pore)(β) or a conventional Minisart N filter (5.3 cm , 0.2 μm pore)(■ ). Flux rates were determined at 50^βC at a constant applied pressure of 50 kPa.

EXAMPLES

General Methodology

Raw materials

Dairy whey was a byproduct of Cheddar cheese production, and was obtained fresh either from commercial cheese manufacturers or prepared "in-house" at the CSIRO Dairy Research Laboratory. The whey was separated (40°C) and pasteurized (72°C, 15 sec) prior to use. Non-fat milk was prepared by separation (35°C - 40°C) and pasteurization (72^CC, 15 sec). For proving trials with pure proteins, cytochrome c (equine heart) was obtained from Boehringer Mannheim, and bovine lactoferrin and lactoperoxidase were isolated from cheese whey essentially as described previously (Law, B.A., and Reiter, B. (1977) The isolation and bacteriostatic properties of lactoferrin from bovine milk whey,

J. Dairy Res. 44, 595-599). Milk ultrafiltrate was prepared by collecting the permeate stream during ultrafiltration (18,000 molecular weight cutoff membrane) of non-fat milk at 50°C.

Membrane processing

Pretreatmeπt. For some examples, only where indicated, the cheese whey and non-fat milk raw materials were pretreated using microfiltration prior to membrane ion exchange. For such experiments involving membrane ion exchange in cross-flow configuration, cheese whey was first microfiltered using a Sartorius cellulose triacetate cross-flow module (0.6 m² = 6 x 10^ cm², 0.45 μm pore) at 45°C. For experiments involving membrane ion exchange in dead-end configuration, cheese whey was first microfiltered using Sartorius Minisart N (5.3 cm², 0.2 μm pore) and non-fat milk using Nalgene cellulose acetate (3.8 cm², 0.45 μm pore) at 20°C.

Membrane ion exchange (cross- flow configuration) . Experiments were carried out using a Sartorius Sartocon II plant incorporating custom-made cross-flow modules fabricated with Sartorius SCX1 polysulphone membrane material (strong cation exchanger) (0.2 μm pore) in two configurations - wide channel (0.5 m² = 5 x 10^ cm²), and narrow channel (0.6 m² = 6 x 10^ cm²). Other common operating conditions included: pressure, 200 kPa (inlet) and 100 kPa (outlet); and temperature, 45°C (loading and washing) and 20°C (elution). Cross-flow modules were pre-equilibrated prior to use and washed with 10 mM NaCl (45 L), and bound protein was eluted with 1 M NaCl (20 L), unless stated otherwise in each example.

Cleaning and sanitation of the membrane was effected with 1.5% (w/v) P3-Ultrasil 53 (10 L) at 37°C, and the membranes were stored in the presence of 1 M NaCl in 20% (v/v) ethanol at 4°C. A schematic representation of the cross-flow configuration experiments, including a definition of terminology and data parameters, is given in Figure 1.

Membrane ion exchange (dead-end configuration) . Experiments were carried out using Sartorius Sartobind S filter units in Minisart configuration (5.4 cm² =

5.4 x lO""* m², 0.45 μm pore). Binding and recovery data were collected at ambient temperature and at 50"C using a controlled flow-rate of 10 mL/min. Flux data were collected at 50^βC using a constant pressure of 50 kPa. Filters were pre-equilibrated prior to use and washed with 10 mM sodium phosphate, pH 7.0 (10 mL), and bound protein was eluted with 1 M NaCl in 10 mM sodium phosphate, pH 7.0 (10 mL). Cleaning and sanitation of the filter was effected with 1 M NaOH (10 mL), and the Sartobind filters were stored in the presence of 1 M NaCl in 20% (v/v) ethanol at 4°C. A schematic representation of the dead-end configuration experiments, including a definition of terminology and data parameters is given in Figure 2.

Analytical Procedures High performance liquid chromatography . Qualitative and quantitative determination of protein components in the raw materials and in samples following membrane ion exchange processing was carried out by high performance liquid chromatography (HPLC) using a Waters system. Samples (100 μL) were loaded onto a Mono S HR5/5 column (Pharmacia) equilibrated with 50 mM sodium phosphate, pH 7.5, and elution was effected with a linear salt gradient (0 - 1.0 M NaCl) in the same buffer over 14 min at a flow rate of 1.0 mL/min. Column effluent was monitored continuously at 220 nm. The areas of peaks, representing proteins of interest (e.g., lactoferrin, lactoperoxi¬ dase), were determined by electronic integration using the Delta Junior data analysis software package (Digital Solutions Pty. Ltd., Australia). Standard lactoperoxi- dase and lactoferrin eluted from the column, under the stated conditions, with retention times of 9.7 min and 18.9 min, respectively.

Polyacrylami e gel electrophoresis . Determination of the identity, purity and molecular size of protein components in the raw materials and in samples following membrane ion exchange processing was carried out by polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulphate (denaturing conditions) (SDS- PAGE) using a vertical slab gel apparatus as described previously (Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of the bacteriophage T4, Nature 277, 680-685). Proteins in samples (50 μL) were separated in a linear 10 - 15% gradient gel, and protein bands were stained with Coomassie Brilliant Blue R. Low molecular weight marker proteins (Pharmacia) were used to calibrate the gel. Markers included: phosphorylase b (94 kDa), bovine serum albumin (67 kDa), ovalbumin (43 kDa), carbonic anhydrase (30 kDa), soybean trypsin inhibitor (20.1 kDa) and α-lactalbumin (14.4 kDa). Under the conditions used, lactoferrin and lactoperoxidase, appear on the gel at an equivalent molecular weight of 80 kDa.

Spec tropho tome try. Activity of the enzyme lactoperoxidase was determined spectrophotometrically in a continuous assay using the artificial substrate 2,2'- azino-Jbis(3-ethyl-benzthiazoline-6-sulphonic acid)(ABTS), essentially as described previously (Putter, J. , and Becker, R. (1983) Peroxidases. In Methods of Enzymatic Analysis (Bergmeyer, H.U., ed.), Vol. 3 (3rd Edition), pp. 286-293, Verlag Chemie, Weinheim). Assay mixtures (2.38 mL) contained 1.67 mM ABTS and 0.18 mM H₂θ2 in 100 mM citrate buffer, pH 5.5. The reaction was initiated by the addition of enzyme containing solution (20 μL), and the rate of change of absorbance at 405 nm was measured on a recording spectrophotometer at 25°C. One unit of enzyme activity catalyzes the oxidation of 1 μmole of ABTS per min under the stated conditions (IU). For proving trials with pure proteins, estimates of protein content in the feed, wash and eluate fractions were calculated from absorbance at 280 nm.

Examples of Membrane Ion Exchange

Cross-Flow Filtration using Pure Proteins

Example 1. In a trial to determine the binding capacity of the membrane ion exchanger (cross-flow configuration) for pure proteins, 200 L of 10 mM sodium phosphate (pH 6.7) containing 30 mg/L lactoperoxidase and 100 mg/L lactoferrin was used as feed material. The Sartocon II plant was configured to recycle the retentate stream. Samples of permeate and retentate were collected during loading for later analysis by HPLC, SDS-PAGE and enzyme assay. After loading, the membrane was washed with 10 L of 10 mM sodium phosphate pH 6.7, and elution of the proteins was carried out at 20°C with three 10-L batches of the sodium phosphate buffer containing increasing concentrations of NaCl, viz. 0.2 M, 0.4 M and 1.0 M. During elution the retentate stream was not recycled. Samples were collected from permeate and retentate fractions during elution for subsequent analysis. Results of permeate flux measurements, and analysis of the permeate for lactoferrin and lactoperoxidase during membrane loading are shown in Figure 3. Results of HPLC and SDS-PAGE analysis of samples from the trial are depicted in Figures 4 and 5, respectively. Data describing binding capacity of the membrane for lactoferrin and lactoperoxidase, and recovery of these proteins following elution, are presented in Table 1.

TABLE 1: Binding capacity and recovery of purified bovine lactoferrin and lactoperoxidase when applied together to a strong cation exchange cross-flow membrane in buffered solution at pH 6.7.

Protein" Binding capacity Recovery (%)' (mg/cm²) lactoferrin 2.37 107 lactoperoxidase 0.22 80

"Purified proteins were dissolved in 10 mM sodium phosphate, pH 6.7 and presented to the narrow channel membrane (0.6 m²) at 45 "C . ^bDefinition is provided in Figure 1. Cross-Flow Filtration using Cheddar Cheese Whey

Example 2. In this trial, Cheddar cheese whey at pH 6.0 was microfiltered prior to membrane ion exchange. The microfiltered whey (120 L) was passed over the narrow channel cross-flow module (0.6 m²) using the Sartocon II plant configured to recycle the retentate stream. Samples of permeate and retentate were collected during the trial to enable subsequent determination of the lactoferrin and lactoperoxidase content by HPLC, SDS-PAGE, and enzyme assay. Following loading of the whey feed material, the membrane was washed with 45 L of 10 mM NaCl. Elution was subsequently effected with 20 L of 1 M NaCl at 20°C, with the retentate stream in non-recycle mode. Following completion of the elution process, samples were collected from the pooled permeate and retentate streams for subsequent analysis. Data describing permeate flux and "breakthrough" of the proteins of interest in the permeate stream during the trial are shown in Figure 6. Results of HPLC and SDS-PAGE analysis of samples from the trial are depicted in Figures 7 and 8, respectively. Data describing binding capacity of the membrane for lactoferrin and lactoperoxidase, and recovery of these proteins following elution, are presented in Table 2.

TABLE 2: Binding capacity and recovery of bovine lactoferrin and lactoperoxidase from separated/past¬ eurized Cheddar cheese whey (pre-microfiltered) at pH 6.0 when applied to a strong cation exchange cross-flow membrane.

Protein" Binding capacity Recovery (%)' (mg/cm²)^b lactoferrin 0.38 83 lactoperoxidase 0.03 101

"Whey (pH 6.0), containing lactoferrin and lacto¬ peroxidase, was presented to the narrow channel membrane (0.6 m²) at 45°C. ^bDefinition is provided in Figure 1.

Example 3. In this trial, experimental conditions were as described above for Example 2 with the exception that the whey was adjusted to pH 6.9 with a concentrated solution of NaOH prior to membrane ion exchange. Following pH adjustment, the whey (97 L) was micro- filtered and then subjected to cross-flow membrane ion exchange as described for Example 2 , with the exception that elution was carried out with 10 L of IM NaCl at 20°C. Data describing permeate flux and "breakthrough" of the proteins of interest in the permeate stream during the trial are shown in Figure 9. Results of HPLC and

SDS-PAGE analysis of samples from the trial are depicted in Figures 10 and 11, respectively. Data describing binding capacity of the membrane for lactoferrin and lactoperoxidase, and recovery of these proteins following elution, are presented in Table 3. TABLE 3: Binding capacity and recovery of bovine lactoferrin and lactoperoxidase from separated/pasteurized Cheddar cheese whey (pre-microfiltered) at pH 6.9 when applied to a strong cation exchange cross-flow membrane.

"Whey (pH 6.9), containing lactoferrin and lacto¬ peroxidase, was presented to the narrow channel membrane (0.6 m²) at 45^βC. ^bDefinition is provided in Figure 1.

Example 4. In this trial, Cheddar cheese whey at pH 6.2 was used, as described above for Example 2, with the exception that the whey was not microfiltered prior to membrane ion exchange. The whey (80 L) was passed over the narrow channel cross-flow module (0.6 m²) using the Sartocon II plant configured to recycle the retentate stream. Other processing conditions were as described for Example 2 , with the exception that elution was carried out with 10 L of IM NaCl at 20^βC. Data describing permeate flux and "breakthrough" of the proteins of interest in the permeate stream during the trial are shown in Figure 12. Results of HPLC and SDS-PAGE analysis of samples from the trial are depicted in Figures 13 and 14, respectively. Data describing binding capacity of the membrane for lactoferrin and lactoperoxidase, and recovery of these proteins following elution, are presented in Table 4. TABLE 4: Binding capacity and recovery of bovine lactoferrin and lactoperoxidase from separated/pasteurized Cheddar cheese whey (not microfiltered) at pH 6.2 when applied to a strong cation exchange cross-flow membrane.

Protein" Binding capacity Recovery (%)¹ (mg/cm²)^b lactoferrin 0.25 86 lactoperoxidase 0.01 102

"Whey, containing lactoferrin and lactoperoxidase, was presented to the narrow channel membrane (0.6 m²) at 45°C. definition is provided in Figure 1.

Example 5. This trial was a repeat of that described for Example 4 , with the exception that the whey (pH 6.2, 76 L, same batch as that used in Example 4 ) was passed over the wide channel cross-flow module (0.5 m²). Other processing conditions were as described for Example 2, with the exception that elution was carried out with 10 L of IM NaCl at 20^βC. Data describing permeate flux and "breakthrough" of the proteins of interest in the permeate stream during the trial are shown in Figure 15. Results of HPLC and SDS-PAGE analysis of samples from the trial are depicted in Figures 16 and 17, respectively. Data describing binding capacity of the membrane for lactoferrin and lactoperoxidase, and recovery of these proteins following elution, are presented in Table 5. TABLE 5: Binding capacity and recovery of bovine lactoferrin and lactoperoxidase from separated/pasteurized Cheddar cheese whey (not microfiltered) at pH 6.2 when applied to a strong cation exchange cross-flow membrane.

"Whey, containing lactoferrin and lactoperoxidase, was presented to the wide channel membrane (0.5 m²) at 45 "C. definition is provided in Figure 1.

Dead-End Filtration using Pure Proteins

Example 6. In a trial to determine the binding capacity of the membrane ion exchanger (dead-end configuration, Sartobind S, 5.4 cm²) for pure proteins, 20 mL of 10 mM sodium phosphate (pH 7.0) containing 0.6 mg/mL cytochrome c, or 0.6 mg/mL lactoperoxidase, or 0.6 mg/mL lactoferrin were used as feed materials. The experiment was carried out at 20°C. Samples of the feed, filtrate, wash and eluate were collected for later analysis by HPLC and spectrophotometry. Data describing binding capacity of the membrane for cytochrome c, lactoferrin and lactoperoxidase, and recovery of these proteins following elution, are presented in Table 6. TABLE 6: Binding capacity, elution and recovery of purified equine cytochrome c, and bovine lactoferrin and lactoperoxidase when applied in isolation to a strong cation exchange dead-end membrane in buffered solution at pH 7.0.

"Purified proteins were dissolved in 10 mM sodium phosphate, pH 7.0 and presented to the Sartobind S filter (5.4 cm²) at 20°C. ^bDefinition is provided in Figure 2. ^cData represents mean and standard deviation of 4 determinations. ^dData represents mean and standard deviation of 8 determinations.

Example 7. In this trial, the binding capacity of the membrane ion exchanger (dead-end configuration,

Sartobind S, 5.4 cm²) for the pure proteins lactoferrin and lactoperoxidase dissolved in milk ultrafiltrate (pH 6.7) was determined at both 20°C and 50°C. The experiment was carried out as described for Example 6 with the exception that the feed material constituted non-fat milk ultrafiltrate containing 0.6 mg/mL lactoperoxidase or 0.6 mg/mL lactoferrin. Data describing binding capacity of the membrane for lactoferrin and lactoperoxidase, and recovery of these proteins following elution, are presented in Table 7. TABLE 7: Binding capacity, elution and recovery of purified bovine lactoferrin and lactoperoxidase when applied in isolation to a strong cation exchange dead-end membrane in milk ultrafiltrate at pH 6.7.

Protein" Temp. Binding Elution Recovery

(°C) capacity (%)^b (%)^b

"Purified proteins were dissolved in milk ultrafiltrate, pH 6.7 and presented to the Sartobind S filter (5.4 cm²) at 20°C or at 50^βC. "Definition is provided in Figure 2. ^cData represents mean and standard deviation of 4 (20^βC) or 2 (50°C) determinations.

Example 8. In this trial, the binding capacity of the membrane ion exchanger (dead-end configuration, Sartobind S, 5.4 cm²) for lactoferrin and lactoperoxidase dissolved in 10 mM sodium phosphate (pH 7.0), when presented as a solution containing both proteins, was determined at both 20°C and 50"C. The experiment was carried out as described for Example 6 with the exception that the feed material (60 mL) contained 0.03 mg/mL lactoperoxidase and 0.15 mg/mL lactoferrin, concent¬ rations that mimic those found in milk and whey. Data describing binding capacity of the membrane for lactoferrin and lactoperoxidase, and recovery of these proteins following elution, are presented in Table 8. TABLE 8: Binding capacity, elution and recovery of purified bovine lactoferrin and lactoperoxidase when applied together to a strong cation exchange dead-end membrane in buffered solution at pH 7.0.

"Purified proteins were dissolved in sodium phosphate buffer, pH 7 and presented to the Sartobind S filter (5.4 cm²) together at 20°C or at 50°C. ^bDefinition is provided in Figure 2.

Dead-End Filtration using Cheddar Cheese Whey

Example 9. In this trial, microfiltered Cheddar cheese whey at pH 6.2 (150 mL) was presented to the membrane ion exchange filter (Sartobind S, 5.4 cm²) at 20°C. Samples of the feed, filtrate, wash and eluate were collected for later analysis by HPLC, SDS-PAGE and spectrophotometry. Results of HPLC and SDS-PAGE analysis of samples from the trial are shown in Figures 18 and 19, respectively. Data describing binding capacity of the membrane for lactoferrin and lactoperoxidase, and recovery of these proteins following elution, are presented in Table 9. TABLE 9: Binding capacity, elution and recovery of bovine lactoferrin and lactoperoxidase from separated/pasteurized Cheddar cheese whey (microfiltered) at pH 6.2 when applied to a strong cation exchange dead- end membrane.

"Whey, containing lactoferrin (63.1 mg/L) and lactoperoxidase (2.4 IU/mL), was presented to the Sartobind S dead-end filter (5.4 cm²) at 20°C. ^bDefinition is provided in Figure 2.

Example 10. This trial was a repeat of that described in Example 9, with the exception that the whey was adjusted to pH 7.0 with 1 M NaOH prior to membrane ion exchange. Results of HPLC and SDS-PAGE analysis of samples from the trial are shown in Figures 20 and 21, respectively. Data describing binding capacity of the membrane for lactoferrin and lactoperoxidase, and recovery of these proteins following elution, are presented in Table 10.

TABLE 10: Binding capacity, elution and recovery of bovine lactoferrin and lactoperoxidase from separated/pasteurized Cheddar cheese whey (microfiltered) at pH 7.0 when applied to a strong cation exchange dead-end membrane.

"Whey, containing lactoferrin (61.4 mg/mL) and lactoperoxidase (4.6 IU/mL), was presented to the Sartobind S dead-end filter (5.4 cm²) at 20°C. ^bDefinition is provided in Figure 2.

Example 11. This trial was a repeat of that described in Example 9 , with the exception that the whey (50 mL) was dialyzed against 10 mM sodium phosphate, pH 7.0 prior to membrane ion exchange. Results of HPLC and SDS-PAGE analysis of samples from the trial are shown in Figures 22 and 23, respectively. Data describing binding capacity of the membrane for lactoferrin and lacto¬ peroxidase, and recovery of these proteins following elution, are presented in Table 11.

TABLE 11: Binding capacity, elution and recovery of bovine lactoferrin and lactoperoxidase from separated/pasteurized Cheddar cheese whey (dialyzed, microfiltered) at pH 7.0 when applied to a strong cation exchange dead-end membrane.

"Whey, containing lactoferrin (55.8 mg/L) and lactoperoxidase (5.6 IU/mL), was presented to the Sartobind S dead-end filter (5.4 cm²) at 20°C. ^bDefinition is provided in Figure 2.

Example 12. In this trial, filtrate (permeate) flux rates for dead-end membrane ion exchange filtration (Sartobind S, 5.4 cm², 0.45 μm pore) were compared with flux rates for conventional filtration (Minisart N, 5.3 cm², 0.2 μm pore) using both microfiltered and non-microfiltered Cheddar cheese whey, pH 6.5. The experiments were conducted at 50°C using a constant applied pressure of 50 kPa. Results are reported in Figure 24.

Dead-End Filtration using Non-Fat Milk Example 13. In this trial, the binding capacity of the membrane ion exchanger for lactoferrin and lactoperoxidase in milk was determined. Microfiltered non-fat milk at pH 6.7 (60 mL) was presented to the dead-end membrane ion exchange filter (Sartobind S, 5.4 cm², 0.45 μm pore) at 20"C. Other conditions were as described above for Example 6. Data describing binding capacity of the membrane for lactoferrin and lactoperoxidase, and recovery of these proteins following elution, are presented in Table 12.

TABLE 12: Binding capacity and recovery of bovine lactoferrin and lactoperoxidase from separated/past¬ eurized and microfiltered milk at pH 6.7 when applied to a strong cation exchange dead-end membrane.

Protein" Binding capacity Recovery (%)'

(mg/cm²)^b lactoferrin 0.47 122 lactoperoxidase <0.01 94

"Non-fat milk, containing lactoferrin (75.5 mg/L) and lactoperoxidase (9.7 IU/mL), was presented to the Sartobind S dead-end filter (5.4 cm²) at 20°C. ^bDefinition is provided in Figure 2.

Although the invention has been illustrated by reference to the particular Sartorius membranes described above it will be readily apparent that the use of other membrane structures, functional groups and porosities falls within the spirit and scope of the invention.

Those skilled in the art will appreciate that the invention described herein is susceptible to variations and modifications other than those specifically described. It is to be understood that the invention includes all such variations and modifications which fall within its spirit and scope.

Claims

CLAIMS : -

1. A process for the separation of charged molecules from a fluid comprising providing an ion exchange medium disposed on a porous membrane, passing the fluid through the membrane, wherein said charged molecules are preferentially adsorbed on the medium, and eluting the adsorbed molecules from the medium.

2. A process according to claim 1 wherein the fluid is a biological fluid.

3. A process according to claim 1 wherein the charged molecules are one or more proteins.

4. A process according to claim 1 wherein the pore size of the membrane is sufficiently small to prevent fat globules and particulates in the fluid from passing through the membrane.

5. A process according to claim 1 wherein the fluid is milk or a milk product.

6. A process according to claim 5 wherein the milk product comprises whey.

7. A process according to claim 1 wherein the fluid is blood or a blood product.

8. A process according to claim 1 wherein said charged molecules are selected from hormones, vitamins, enzymes, clotting factors, immunoglobulins, peptides. lysozyme and antibodies.

9. A process according to claim 1 wherein the membrane has a pore size in the range about 0.1 μm to about 1.2 μm.

10. A process according to claim 1 wherein the membrane has a pore size in the range about 0.2 μm to about 0.6 μm.

11. A process according to claim 1 wherein the adsorbed charged molecules are eluted from the membrane prior to removing fat globules and particulates from the retentate side of the membrane.

12. A process according to claim 1 further comprising the step of removing fat globules and particulates from the retentate side of the membrane prior to elution of the charged molecules.

13. A process according to claim 1 wherein the ion exchange medium is a strong ion exchange medium.

14. A process according to claim 13 wherein the ion exchange medium is a strong cation exchange medium.

15. A process according to claim 5 wherein the charged molecules comprise lactoferrin and/or lactoperoxidase.

16. A process according to claim 5 wherein lactoferrin is separated from milk or a milk product containing lactoferrin and lactoperoxidase.

17. A charged molecule isolated from a fluid using the process of claim 1.

18. A process according to claim 1 substantially as hereinbefore described with reference to any one of the Examples.