Electrical Separator Apparatus and Method of Counterflow Gradient Focusing

The present invention relates to an apparatus for conducting separation of biological materials, such as peptides, proteins, cells, cell organelles, cell membranes, viruses, and organic and inorganic compounds or all metal ions from a mixture. More particularly, the present invention relates to an apparatus and method for batch or continuous-flow electrophoretic separation of ionic molecules by zone electrophoresis, isoelectric focusing, counterflow gradient focusing or by electrodialysis.

Electrophoresis is an analytical and preparative tool which can separate components of a mixture on the basis of their ionic charge and mass. A mixture of ionic species is exposed to an applied voltage field, which causes the ions to migrate toward the oppositely charged electrode at a rate which depends on their electrophoretic mobility, which in turn depends on charge, mass and symmetry as well as other parameters. Different modalities of electrophoresis, e.g., zone, isoelectric focusing, isotachophoresis, moving boundary can be utilized in various instruments that are classified by the method selected to minimize convective mixing: 1) free fluid methods which use thin film, shear, capillaries, or compartmentation; or 2) supported fluid methods which use density gradient gels or other matrices. The present method is free fluid compartmentation.

Zone electrophoresis is the separation carried out in the presence of a homogenous buffer, in which sample components separate according to their mobility. No steady state is achieved, and migration continues with gradual broadening of the sample zones due to diffusion and other effects.

Isoelectric focusing (IEF) is a variant based upon the fact that most biomaterials are amphoteric in nature, i.e., are positively charges in an acidic environment and negatively charged in a basic environment. At a particular pH value, called the isoelectric point (pI), the biomaterials acquire a zero net charge due to the balance of positive and negative charges. When such amphoteric materials are exposed to an applied electrical field, in a medium exhibiting a pH gradient, they will migrate toward the pH region of their pI and become immobilized or focused in a steady state at that pH region.

Electrophoretic separation in multi-compartment electrolyzers has been known since 1912 (Ikeda et al, U.S. Pat. No. 1,015,891) when it was introduced for preparative scale protein fractionation. To the present, the basic apparatus has undergone multiple changes and has been improved in many ways.

Current designs feature membrane-separated, multicompartment electrophoretic cells with cylindrical, annular or rectangular cross-sections, and membrane-separated electrode compartments. The separation process is usually carried out in the free solution under batch process conditions. The entire process volume of the solution may be contained within the electrophoretic cell during the separation process, or the process volume of the solution may be recirculated through the electrophoretic cell until completion of the separation process.

Several electrophoretic devices for isoelectric focusing have been designed by Rilbe, with the latest improved version published in 1980. This latest device consists of a rotating cylinder with 46 subcompartments, each 1.5 cm wide and 12.5 cm in diameter, and has a total volume of 7.6 liters. The subcompartments are separated by polyvinylchloride (PVC) membranes, while, at the ends, one of the electrode compartments is separated by cellophane and the other by PVC. The cooling of the solution is achieved by a combination of inner and outer cooled surfaces. The cylinder is submerged in 360 liters of refrigerated water and rotated about the separation axis, thereby cooling the cylinder walls. Also, four glass cooling tubes pass through each subcompartment and, when the cylinder is rotated about its longitudinal axis, water is scooped into one side of the cooling tube and forced out the other.

While the internal cooling system is a beneficial feature of the apparatus, the choice of materials for the membranes and the excessive width of the subcompartments are disadvantageous. PVC exhibits a 5-10 fold greater electroosmotic effect than a woven monofilament nylon screen. Electroosmotic mixing is also increased by the small porosity of the PVC membranes, which trap colloidal and larger protein particles. PVC also binds proteins with considerable affinity. Protein precipitations at a position in the pH gradient, deviating from the respective pI, produce a net charge on the immobilized protein resulting in electroosmotic mixing. By using the chosen membrane materials, i.e., cellophane and PVC, electroosmotic transport across the electrode membrane occurs, which induces mixing. Electroosmotic transport counteracts and degrades the separation process. The large width of each subcompartment also retards the separation process. Also, the low ionic strength electrolytes, which are employed in Rilbe's apparatus, cause a voltage drop in the electrode chambers, which decreases the effective field strength in the separation compartments.

Another typical device for electrophoretic separation processes is disclosed in U.S. Pat. No. 4,588,492 issued May 13, 1986 to Bier. This apparatus, known by the trademark ROTOFOR, is essentially a smaller version of the apparatus by Rilbe. It consists of twenty annular-shaped compartments, which are arranged into a cylindrical electrophoretic cell. The individual compartments are separated by membranes, and an internal cooling tube is arranged along the longitudinal axis of the cell. The cell is usually operated in a horizontal position while rotated along its longitudinal axis. It has recently been demonstrated by Egen, N. B., et al., "Fluid Stabilization Studies of Free IEF in Cylindrical and Annular Columns" Separation Science & Technology, Vol. 22, p. 1383 (1987), however, that equal separation results could be achieved in a vertical stationary mode while minimal separation was observed in a horizontal stationary operation mode. The ROTOFOR eliminates some of the drawbacks of Rilbe's device by using woven nylon screen, small width subcompartments and ion exchange membranes to separate the electrode compartments. The disadvantage of the ROTOFOR is that, due to the annular cross section and the central inner cooling tube, scale-up to significantly larger capacities is impossible, since an increase of the cylinder diameter would result in an insufficient cooling of the solution close to the outer walls of the cylinder. The ROTOFOR is strictly a batch mode processor and cannot be operated in a continuous flow mode.

Multi-compartment electrolyzers which separate proteins by the principles of IEF are also known. For example, U.S. Pat. Nos. 4,362,613 and 4,204,929, both issued to M. Bier on Dec. 7, 1982 and May 27, 1980, respectively, disclose such apparatus. However, this is designed to operate exclusively by IEF and in the batch mode, whereby the solution is recycled between the electrolytic cell and an external heat exchanger to provide external cooling of the sample solution.

A method for the continuous-flow electrophoretic separation of proteins, antibodies, nucleic acids etc. in a matrix system is known and is shown in U.S. Pat. No. 4,323,439 issued to P. O'Farrell on Apr. 6, 1982. Matrices of the size-exclusion type, of the ionic type, of the adsorption-type or of a special affinity to the particles to be separated are used. The separation chamber comprises at least one matrix of continuously varying properties or two and more matrices of continuously or intermittently varying properties in the longitudinal direction. The mixture to be separated is introduced into the separation chamber in a carrier liquid which flows through the length of the chamber. An electric field is applied such that the electrophoretic movement occurs in a direction opposite to the carrier liquid flow. By countering the electrophoretic movement with counter-current flow, an equilibrium zone within the chamber is created, where the separated and concentrated ions may be withdrawn. However, matrices are expensive and the random withdrawal of the separated and concentrated ionic species will cause flow discontinuities. An electrophoretic separator in a continuous flow, free solution mode is highly desirable.

It is, therefore, an object of the present invention to provide a preparative scale, free-fluid electrical separator, which is matrix-free, which may be operated in both a batch mode and a continuous, flow-through mode and which may be scaled according to the solution quantities which are to be processed. It is another object of the present invention to provide the electrical separator with an internal cooling system to efficiently dissipate Joule heat during the separation processes, which eliminates the need for horizontal rotation or recirculation to an external cooler, and permits application of higher voltage gradients and yields shorter separation times. Another broad aspect of the invention is to provide an apparatus for electrical separation of biological materials, such as proteins, peptides, cells, cell organelles, cell membranes, viruses, and organic and inorganic compounds, and metal ions, by a single apparatus adaptable for use in zone electrophoresis, isoelectric focusing, counter-flow gradient electrophoresis and electrodialysis.

The continuous flow electrical separator of the present invention is primarily characterized by narrow and long separation chambers defined within planar spacers having internal heat exchange means, which form two parallel side walls of each separation chamber. A plurality of spacers may be arranged next to each other to form an array of parallel spacers having horizontally aligned separation chambers. Membranes, selected according to desired separation process, are provided between the planar spacers to separate the separation chambers, act as containing walls for fluid within the separation chambers and minimize convective mixing of the separated sample zones defined within each separation chamber. An electrode compartment is provided at each end of the spacer array, containing an anode and a cathode and appropriate anolyte and catholyte, respectively. A multichannel pump introduces the unprocessed sample into the individual separation chambers and pumps processed fluid from each separation chamber to a collection chamber.

The electrical separator of the present invention allows for a modular design which may be adjusted to the respective needs of various applications. It may be used for separations based on electrophoretic methods such as, without limitation, zone, counter-flow gradient, IEF, continuous flow electrodialysis, electrodecantation and for water purification based upon any of these methods.

These objects and other features and advantages of the present invention, will appear more clearly from the following more detailed description of the preferred embodiments thereof taken with reference to the accompanying figures.

FIG. 1 is a diagrammatic view of the separation apparatus according to the present invention.

FIG. 2 is a perspective view of a separation cell according to the present invention.

FIG. 3 is a cross-sectional view taken along line 3--3 of FIG. 1.

FIG. 4 is a cross-sectional view illustrating multiple adjacent separation cells according to the present invention.

FIG. 5 diagrammatically illustrates fractionation of human serum albumin and human hemoglobin by IEF using the separation apparatus of the present invention.

FIG. 6 diagrammatically illustrates the process of counter-flow gradient focusing (CFGF) according to the present invention with a deep or wide range gradient.

FIG. 6a diagrammatically illustrates the CFGF process with a shallow or narrow range gradient for high resolution focusing.

FIG. 6b diagrammatically illustrates the addition of a centering pump having a flow rate F.sub.c in the CFGF process.

FIG. 6c diagrammatically illustrates the addition of a throughput pump having a throughput flow F.sub.T in the CFGF process.

FIG. 6d diagrammatically illustrates a first set of experimental conditions used to illustrate the CFGF process of the present invention.

FIG. 6e diagrammatically illustrates a second set of experimental conditions used to illustrate the CFGF process of the present invention.

FIG. 7 diagrammatically illustrates the process of electrodialysis according to present invention.

FIG. 8 diagrammatically illustrates a scaled-up cell configuration for performing a parallel set of separations in multiple separators, but in one electrophoretic cell.

FIG. 9 diagrammatically illustrates the fluid flow for counterflow gradient mechanical pumping and the resulting pressure gradient produced thereby during the CFGF process.

FIG. 10 diagrammatically illustrates the fluid flow for electroosmotic internal pumping and the resulting pressure gradient produced thereby during the CFGF process.

FIG. 11 is a diagrammatic representation of the electrophoretic separator of the present invention set for CFGF processing.

Turning to the accompanying figures, and with particular reference to FIGS. 1-4, there is shown a multi-compartment electrophoretic separator 10 according a preferred embodiment of the present invention. The electrophoretic separator is comprised of an array of subcompartments 12 with two electrode compartments 30, 32 on either end of the array. All subcompartments 12 for this electrophoretic separation cell are identical. Each subcompartment 12 consists generally of a rectilinear planar body which has a narrow channel 16 formed within the planar body. Channel 16 acts as the separation chamber for processing the sample. The subcompartments 12 are adjacently arrayed in a co-planar fashion. Membranes 17 are interdisposed between the individual subcompartments 12. Membranes 17 which serve as delimiting inter-subcompartmental boundaries for the narrow channel 16 and effectively separate each adjacent channel 16 into discrete separation chambers.

Each subcompartment 12 is provided with a linear array of openings 20 which receive a bolt or other fastening means to clamp together the array of aligned subcompartments 12. Each subcompartment also is provided with a linear array of alignment openings 22, which receive alignment pins to maintain a uniform width of the channel 16 along its longitudinal aspect. The array is bolted or clamped together, whereby the electrode compartments function as end-plates which provide a rigid structural support. With the electrode compartments 30, 32 provided on either end of the array, an electrical field applied to the electrophoretic separator runs perpendicular to the planar aspect of subcompartments 23 and the membranes 17. Appropriate electrolytes are pumped into their respective electrode compartments. Continuous recycling of the electrolytes removes electrolytic gases.

The membranes 17 reduce inter-subcompartmental fluid convection and mixing of the separated sample zones. By providing a parallel array of membranes 17 perpendicular to the separation axis confines fluid convection to the individual subcompartments 12, which is beneficial to the separation process while preventing detrimental fluid convection between subcompartments 12.

Each subcompartment 12 is provided with a heat exchanger 14 to dissipate Joule heat generated during the electrophoretic separation. The heat exchanger 14 is disposed in the channel 16 of each subcompartment 12 such that at least a substantial portion of the longitudinal inner walls of channel 16 formed by the planar body 12, are associated with heat exchanger 14. According to the preferred embodiment of the invention, heat exchanger 14 comprises either a quadrilateral or circular cross-sectional tubing having a depth or diameter corresponding to the thickness of the planar body 12. The heat exchanger 14 connects to inlet and outlet ports 18 at the ends of the length of the cavity. Inlet and outlet ports 18 are preferably formed by providing co-axially aligned inlet and outlet ports 18 in adjacent arrayed subcompartments 12, each port 18 having fluid conduit 19 between the port 18 and the heat exchanger 14. Fluid conduit 19 may be an integral with heat exchanger 14 or a discrete component.

Sample inlet 11 and collection 13 ports are also provided at each end of channel 16 and are in fluid flow communication therewith. Both inlet port 11 and collection port 13 are connected to a multichannel pump 60 which pumps the sample solution into channels 16 of individual subcompartments 12 and concurrently removes separated solution from the collection ports 13.

The volume of the electrophoretic separator according to the present invention is determined by the number of subcompartments 12 and the length of channel 16 in each subcompartment 12.

According to the best mode contemplated for the invention, each subcompartment 12 is preferably 0.1 to 0.4 cm in thickness, and may be die cut, extruded, molded or otherwise formed as may be known in the art. Regardless of the length of channel 16 or number of subcompartments 12, the thickness of the each subcompartment 12, and, hence, channel 16, i.e., the distance between two membranes 17, should be about 0.1 to about 0.4 cm. A thickness of excessive width is detrimental to the electrophoretic separation process due to the greater time required for a particle to cross each subcompartment. A thickness which is too small makes sample input and output difficult and may allow adjacent membranes to touch, thereby occluding flow through the channel 16. The length of the planar body 12 and cavity 16 may vary to accommodate different volumes. A greater distance between the sample inlet and exit ports requires unnecessary fabrication costs. A much shorter path between the sample inlet and exit ports, on the other hand, does not provide enough travel distance for an effective separation. Further, an effective dimension of the channel 16 which forms the separation chamber, should be about 50 cm in length, 0.2 cm in width and 0.3 cm in thickness, which yields an effective total volume of 3 ml within channel 16.

Because of its dielectric properties, a silicone rubber material is well suited for sample containment in an electrophoretic separator, however, other electrically insulating materials are also effective. Because of its elastic properties, silicone rubber, together with the membranes 17, acts as a gasket between adjacent subcompartments 12 when the subcompartment array (FIG. 1) is bolted or clamped together.

The membranes used in the embodiment of the present invention are preferably woven monofilament nylon screens. Compared to PVC screens, frequently employed in electrophoretic separators of the prior art such as the Rilbe machine, nylon screens lower inter-compartmental electroosmosis by a factor of 5 to 10.

The electrode compartments 30, 32, provided at each end of the subcompartment array, provide a cavity for a platinum electrode. The electrode compartments 30, 32 are constructed from a suitable plastic material, preferably acrylic, and serve as a rigid end plate allowing the subcompartment array to be bolted together. In order to remove electrolytic gases from the electrolytic cavity, they are equipped with inlet and exit ports for the flow-through operation of the electrolytes. The electrolytes may be recirculated within each compartment 30 or 32, may be flowed through each compartment 30, 32 in a single pass or the anolyte may be mixed with the catholyte.

To facilitate assembly of the apparatus 10, the subcompartments 12, and the electrode compartments 30, 32, have a plurality of openings perpendicular to their flat faces. Bolts are inserted into the openings, which form one long bore hole upon the co-axial alignment of the components, which tightened to seal the subcompartments and electrode compartments against leakage. Additionally, a plurality of alignment openings 22 are provided in the planar bodies 12 and adjacent to or in close proximity to the channel 16. The alignment openings 22 serve to maintain uniform longitudinal alignment of channel 16, by engagement upon alignment pins (not shown).

Sample inlet 11 and collection 13 ports may be disposed within the silicone rubber material and pass from each end of the channel 16 and extend external to the planar body of the subcompartment 12. Each of the sample inlet 11 and collection 13 ports are connected to the multichannel pump 60.

Continuous-flow processing requires the provision of means for dissipating the Joule heat generated by the applied electric field. The apparatus of the present invention has the heat exchange means 14 for the recirculation of a coolant either attached to, or made a part of the planar body of each subcompartment 12 such that the heat exchange means 14 forms side walls of channel 16. The heat exchange means 14 according to the preferred embodiment of the invention, consists of cooling tubes 14 having the same cross-sectional diameter or depth as that of the channel 16. Each subcompartment 12 is equipped with a parallel pair of cooling tubes 14, each of which has inlet and exit ports 18 provided at the ends of the longitudinal axis of the subcompartment 12. The inlet and outlet ports 18 of all subcompartments 12 open into the same feed lines running perpendicular to the planar aspect of each subcompartment 12 through the array. The feed lines on either end of the silicone rubber spacer are formed by openings perpendicular to the planar aspect of the planar body of each subcompartment 12 and are in fluid flow communication upon bolting or clamping of the subcompartment array. The coolant is supplied from a coolant supply 50 and is continuously recirculated by a pump 52.

To achieve an optimized cooling effect, the cooling tubes 14 are thin walled, achieving high thermal transfer rates, and have high dielectric strength to avoid electrical conduction to the coolant. Suitable materials are thin plastic materials, such as tetrafluoroethylene or fluorinated ethylpropylene resins marketed under the trademark TEFLON or polypropylene co-polymers, polyethelyene or silicone. The cooling tubes 14 may have any suitable cross-sectional shape, but are preferably circular or quadrilateral. Cooling tubes 14 have a cross-sectional dimension which corresponds to the thickness of the planar body forming the subcompartment 12, and, therefore, have substantially the same width or depth as that of the channel 16. The cooling tubes 14 may be affixed to the side walls of the channel 16 by gluing, welding, or other suitable method of affixation as may be known in the art, the cooling tubes may be molded directly into the material forming the planar body of the subcompartment 12, or the cooling tubes 14 may be extruded or otherwise formed as an integral part of the planar body of the subcompartment 12.

Sufficient cooling of the sample volume in the cavity is achieved only when the entire volume of the sample fluid is in close proximity to the surface area of the heat exchange means 14. It has been found preferable to have the sample fluid within a range of about 0.05 cm to about 0.15 cm away from any heat exchange means 14 surface to provide effective cooling of the sample fluid. Thus, the desirable width of channel 16, as measured by the lateral distance between the two parallel cooling tubes 14 is from about 0.1 cm to about 0.3 cm.

Providing internal cooling adjacent to the separation chamber permits use of a higher applied potential for the separation process which increases both the resolution and the speed of the separation. An advantage of the apparatus according to the present invention over designs of the prior art is the configuration of a long narrow separation chamber having internal cooling. As long as the ratio of cooling surface area to process volume remains constant or is increased, the device can be scaled to any sample cavity volume desired, simply by increasing the length and the number of subcompartments 12, without loss of resolution.

An alternative system design for the IEF mode, which increases sample volume processing, features the alignment of two or more subcompartments 12 in series. Instead of constructing an electrophoretic cell having longer subcompartments 12 and, therefore longer channels 16, two or more subcompartments 12 may be employed in series. In this manner, an awkwardly elongated cell design is avoided. A serial arrangement of subcompartments 12 permits application of a higher voltage to the downstream array. As the sample solution is pumped through the separation chamber channel 16 while exposed to a voltage gradient perpendicular to the sample flow and membranes 17, there is a decrease of conductivity in the direction of the sample flow due to the formation of the pH gradient where buffer components become less changed. The loss in conductivity permits application of greater voltages across the cell. In standard electrophoretic cell designs, it is impossible to provide a compensating increased voltage for only that part of the separation chamber which experiences the loss of conductivity of the fluid sample. However, by providing two or more subcompartments 12, in series, according to the present invention, a different voltage may be applied to each serial subcompartment 12 array. Therefore, the effect of conductivity loss because of the ongoing separation, i.e., depletion of charged particles, permits higher voltages, resulting in better and faster separation results.

According to the preferred embodiment of the invention, and the best mode contemplated for the invention, for zone electrophoresis, counter-flow gradient focusing or isoelectric focusing, it is desirable to employ about twelve subcompartments 12 in the array. Fewer than 10 subcompartments compromises separation, while more than 20 subcompartments complicates construction of the apparatus without further improving the separation to a significant extent. For electrodialysis, a single subcompartment may be used in conjunction with an anode and cathode, wherein an appropriate anion and cation exchange membrane, is interdisposed between the anode and cathode, respectively, and the subcompartment.

The capacity can be scaled by providing an array of a plurality of separator subcompartments 12 extending in the direction perpendicular to both the length of the subcompartment and to the electrophoresis axis. This scale-up is illustrated in FIG. 8. The scaled up cell configuration consists of a series of parallel subcompartments 12, each having a plurality of linearly adjacent separation channels 16, with each separation channel bounded by a body of a cooling fluid 15. This scaled-up configuration is equivalent to performing a parallel set of separations in multiple separators, but in one electrophoretic cell.

The above described electrical separator is a versatile instrument for the separation of proteins, cells and salts. It may operate according to various electrophoretic separation principles, for which different operational modes are required. The electrical separator of the present invention is easily adapted to various modes simply by altering the number of subcompartments, the selection and arrangement of membranes, buffers and anion or cation exchange resins. The present invention is, therefore, a unique versatile apparatus.

When the apparatus of the present invention is used for IEF, a pH gradient is created in the array. The pH gradient and, therefore, the resolution, may be easily adjusted by manipulating the buffers inherently required for the isoelectric focusing process. Such adjustment may occur by, for example, judiciously tailoring the gradient either with isoelectric focusing buffers having appropriate pH ranges selected from commercial sources, by prefractionation of commercial pH gradient mixtures in order to produce a more narrow pH gradient than commercially available, or by using mixtures of simple buffers. Most conveniently applicable are ampholyte buffers which are commercially available in the pH range of 3.0 to 10.0 in intervals of 0.01 pH units. By using these buffers a virtually continuous pH gradient may be created in the direction of the electric field in the separation cell.

The method is applied such that unprocessed fluid containing a desalted sample is mixed with the appropriate carrier ampholyte buffer and then pumped directly into each subcompartment of the separation cell. The sample and buffer mixture is continuously pumped into the cell using a multichannel pump 60. While it moves through the cell, it is continuously processed. The processed fluid is then pumped from each of the focusing subcompartments 12 into individual collection vessels 40 by the same pump 60. Thus, an apparatus with 12 subcompartments will require a 24 channel pump.

Using the multichannel pump, the apparatus is first primed with the unprocessed sample fluid, whereby the exit ports lead back to the unprocessed fluid container. Next, the coolant is circulated through the cooling tubes allowing sufficient time for the temperature equilibration. When the temperature equilibrium is achieved in the apparatus, the electrical power is applied. The fluid collected at the exit ports is recycled into the unprocessed sample fluid reservoir until an element of sample has had sufficient time to traverse the length of the cell while exposed to the electric field. The fluid flowing from the exit ports is not recycled and is then continuously collected.

As an example, which is not intended to limit the scope of the invention, a ten subcompartment array 10 is provided, with each separation chamber channel 16 having a volume of 3 ml, providing a total sample cavity volume of 30 ml is provided. The multi-channel pump 60 is set to introduce and withdraw at the rate of 0.5 ml/min per channel. It, therefore, will require 6 min for the sample to complete one pass through the separation chamber channel 16 in each subcompartment 12. Thus, 6 min from the initial application of the power the sample withdrawn from the collection port 13 may be fed directly into the appropriate collection vessels 40.

Complete separation time depends upon a large number of variables such as the steepness of the pH gradient, the concentration of the sample constituents, the amount of power applied, and the magnitude of the protein mobility.

In all IEF devices, the criterion for the resolution of sample components is whether two proteins having closely spaced isoelectric points may be resolved, or, at which minimum isoelectric point difference, between the two sample components, may they still be resolved.

As the resolution increases with the applied voltage gradient, maximizing the voltage is essential for a fast and clean separation. However, the voltage maximum in IEF is limited by the efficiency of heat removal from the sample solution. The more efficient cooling provided by the present invention permits application of higher applied voltages to the apparatus of the invention compared with the IEF apparatus of the prior art. For example, a comparison of the apparatus of the present invention with the ROTOFOR system of the prior art (Table 1) shows that the apparatus of the present invention is superior in its heat removal properties while permitting use of greater maximum applied voltages. The maximum applied power of the present invention is compared with the ROTOFOR having a ceramic cooling tube, 8.6 cm long and 1 cm in diameter, which effectively removes 15 watts. Maximum heat removal is calculated by multiplying the total cooling tube surface area of the apparatus of the present invention by the ratio 15/ROTOFOR cooling tube area.

Thus, the apparatus of the present invention permits a higher applied field strength with a concomitant greater heat removal capacity than ROTOFOR device. Both effects result in a greater resolution and throughput.

Another advantage of the apparatus of the present invention is that material separated in the flow-through mode may be collected in as little time as is required for the sample to traverse one length of the cell. Where, for example a flow rate of 1 ml/min is employed, the focused material may be collected after only three minutes from the onset of the applied potential. This facilitates rapid analysis of the eluate during the run, which permits adjustments of the operating parameters, if needed, during the run.

FIG. 5 presents the results of a separation of a two component mixture of human hemoglobin and human serum albumin by IEF using the electrical separator 10 of the present invention. The process was carried out in an IEF solution containing 1.0 ml of human hemoglobin (80 mg/ml) and 0.5 ml of blue stained human serum albumin (250 mg/ml) diluted to 500 ml with 1% (w/v) pentaethylenehexamine carrier ampholyte. Ion exchange membranes were configured in IEF orientation. Absorbance of the human serum albumin was measured at 602 nm and absorbance of human hemoglobin was measured at 540 nm. The experimental conditions are summarized in Table 2.

The data illustrates that a complete separation of the components was accomplished utilizing the apparatus 10 of the present invention in the continuous IEF mode under various conditions which influence the resolution and focusing of the separation.

For a batch operation, the entire sample solution containing the appropriate buffer system is introduced through the inlet ports 11 into the subcompartments 12 of the apparatus 10, and are contained within the separation chamber 16. The pump is turned off to induce the separation process on a non-flowing sample. When the separation is complete, the fractions are collected at the respective outlet ports of the subcompartments by pumping them into the individual collection vessels.

Another separation mode in which the apparatus 10 of the present invention operates has been termed by the inventors counter-flow gradient focusing mode. Counter-flow gradient focusing fundamentally operates on the principle of zone electrophoresis whereby components separate due to their electrophoretic mobility through a medium. In a fixed amount of time, species of a high electrophoretic mobility move further in an electrophoretic field than species of a low electrophoretic mobility.

Conventional zone electrophoresis processing is performed when there is no fluid flow in the direction of the electrophoretic migration. The counter flow process described in O'Farrell, U.S. Pat. No. 4,323,439 contemplates introduction of the carrier fluid flow in a direction opposite to the direction of the electrophoretic migration of the ions to be separated, thereby slowing each ion migration down by the same rate along the separation axis of the separation cell.

According to the present invention, however, there is provided a counter-flow gradient which counteracts the electrophoretic migration of the ions to be separated. The counter-flow gradient, as distinguished from the uniform counter-flow employed by O'Farrell, non-uniformly slows migration by slowing the faster species more than the slower species.

The inventive counter-flow gradient focusing (CFGF) method incorporates a system of fluid pumping which generates a buffer flow that opposes the direction of electromigration. The system is generally illustrated in FIG. 11. The counter-flow exists as a flow speed gradient which increases in the electrophoretic direction. As a result of this flow pattern, for example, a negatively charged sample component introduced into the top of the separator apparatus 10, adjacent to the cathode (-) will migrate towards the anode (+) under the influence of the electric field, while moving towards the bottom of the cell. As the sample component is migrating towards the anode it becomes increasingly slowed by the opposing counter-flow until the particle reaches the position in the separator 10 where the electrophoretic speed is balanced by the counter-flow. The particle is focused and will remain at this position until the buffer flow sweeps it from the subcompartment 12 in which it is focused and into a collector 40. This results in an equilibrium focusing process and a better separation.

Thus, if the sample component is introduced into the top of the separator apparatus 10 on the anodic side of its focused position, where the counter-flow is greater than the electrophoretic speed, the component will be forced towards the cathode by the counter-flow, until it reaches the same focused position. Regardless of the input position, the sample component will focus at the same subcompartment 12 position and is collected from the same exit position of the separator apparatus 10. As every sample species experiences the opposing forces of the electric field and fluid counterflow, band widening diffusion is minimized.

The length of the separation chamber 16 between the input 11 and outlet 13 ports, in the present invention, is much greater than the distance between adjacent chambers 16 or subcompartments 12. This permits ample time for particles to reach equilibrium between the counter-flow gradient and their electrophoretic mobility before the particles exit from the outlet port 13.

In a further embodiment of the present invention, one or more of the subcompartments may have inlet ports for the introduction of the sample solution. The sample solution may be introduced into any one of the inlet ports of the sample subcompartments with one unit of inlet flow. When this configuration is employed, an additional unit of flow must be withdrawn from this subcompartment in order to compensate for the additional unit of inlet flow.

FIG. 6 is a schematic representation of the counter-flow gradient focusing mode of the present invention. Unlike the IEF mode, where each of the subcompartments of the separation cell is provided with an inlet and an exit port, in one method to generate the counter-flow gradient, buffer fluid is introduced only into an end subcompartment 12 on the array. The magnitude of the gradient equals the difference across adjacent membranes which, in turn, equals F.sub.w. The remaining inlets 11 are plugged or otherwise blocked. An input flow of the fluid sample in buffer, having a flow rate F.sub.i, is introduced into the end subcompartment 12 and pumped into the apparatus 10 on a continuous mode. All collection ports 13 are open. The pump 60 withdraws an even fractional portion of the flow pumped out of the collection port 13. In accordance with a first CFGF technique, withdrawal of the separated fluid from each subcompartment 12 occurs at a withdrawal flow rate of F.sub.w in which F.sub.w =F.sub.i /# of subcompartments in the array. In this manner, each adjacent membrane 17 has a decreasing counter-flow across it. Thus, for example, if an introductory counter flow consists of 10 units of flow in a ten subcompartment array, one unit of flow is withdrawn from collection port 13 of subcompartment A, and nine units of flow must cross the membrane to the adjacent subcompartment B. In the adjacent subcompartment B, one unit of flow will be withdrawn from its collection port 13, and eight units of flow will cross the boundary to the next subcompartment C. The counter-flow opposing the direction of electrophoretic velocity, from subcompartment to subcompartment, is thereby reduced by one unit each time and a linear counter-flow step gradient is created. For ions electromigrating opposite the fluid flow, the counter-flow is accordingly strongest at the final destination of the migrating particles and increases in the direction of migration of the particles. Thus, the fastest electromigrating particle will approach the fastest counter-flow.

The resolution of counter-flow gradient can be increased by providing a mechanism to cause a shallower gradient of counter-flow than 1 flow unit as shown in the example above. This second application of CFGF is accomplished by decreasing the fluid withdrawal rate. FIG. 6a illustrates an example in which the fluid withdrawal rate was reduced from 1 unit of flow to 0.1 unit of flow. A concomitant decrease in the counter-flow gradient, then, permits species with closely spaced mobilities to be separated by a greater distance. Two species with mobility differences less than 1 flow unit would be collected in the same fraction when F.sub.w =1, yet they would be collected in different fractions when F.sub.w <1.

However, as decreasing F.sub.w decreases the counter-flow gradient (increases resolution), it concomitantly decreases the sample throughput and changes the position of the gradient. Both of these secondary changes are, at first glance, potentially detrimental to the separation, but are rectified with an additional pump for each condition. In order to reposition the desired gradient to the center subcompartments of the electrophoretic cell, while maintaining the shallow gradient, a centering withdrawal pump 80 is provided on the exit port of the subcompartment adjacent to the cathode as illustrated in FIG. 6b.

In the example illustrated in FIG. 6b, the centering withdrawal pump is set to a centering flow rate F.sub.c, e.g., 5X, in order to maintain the counter-flow gradient as 0.1X and to position the gradient between 5.9 and 5.1. The value of F.sub.c does not effect the magnitude of the counter-flow gradient nor the throughput of sample components of interest. A centering flow of 8X, for example would position the gradient between 8.9 and 8.1, and so on. Thus, by choosing the values of F.sub.c and F.sub.w, any desired gradient and position within the gradient can be selected.

In order to simultaneously control the throughput rate, the magnitude of the counter-flow gradient and the position within the gradient, it is necessary to add an additional pump, a throughput pump 68 having an input flow rate F.sub.T per subcompartment, as illustrated in FIGS. 6c and 11. As used in FIG. 6c:

F.sub.i =input flow to the subcompartment adjacent to the anode;

F.sub.w =withdrawal flow per subcompartment=throughput rate;

F.sub.T =throughput pump flow per subcompartment;

F.sub.c =centering pump withdrawal flow;

F.sub.N =flow rate across the membrane adjacent to the anode;

F.sub.1 =flow rate across the membrane adjacent to the cathode;

Counter-flow gradient limits=F.sub.N to F.sub.1 ;

N=number of membranes=number or subcompartments-1; and

G=the counter-flow gradient= ##EQU1##

Before operating the electrophoretic cell in the CFGF mode, the flow limits F.sub.N and F.sub.1, and the throughput rate per subcompartment F.sub.W are selected on the basis of the desired resolution, throughput and position within gradient. The value of F.sub.W is then set on a fraction pump 70. Using F.sub.N, F.sub.1, and F.sub.W, the flows for the gradient G, the centering pump F.sub.C, the throughput pump F.sub.T, and the input flow F.sub.i are calculated according to the following equations: ##EQU2## The values for F.sub.C and F.sub.T are manually or automatically set on their respective pumps 80, 70. The sample can be input on any one or any combination of throughput pumps F.sub.T. Table 3 presents some representative calculations using flow parameters from FIGS. 6, 6a, 6b and 6d and other parameters selected to demonstrate 1) changing the throughput while maintaining the same gradient and flow limits; 2) changing the position within the gradient, while maintaining the gradient and throughput; and 3) changing the gradient, while maintaining the throughput.

In order to start the separation process, the separation cell is first primed with the buffer fluid to expel all air. The inlet tubes 11 (one for each sample subcompartment) may be joined together to fill the separation cell with the buffer solution using a pump. After filling the separation cell, the sample to be separated may then either be introduced into the buffer reservoir or pumped directly into the subcompartment 12 of choice.

In order to begin separation, the cooling system must be switched on and, after a temperature equilibrium is reached between the coolant and the buffer fluid, the power supply is turned on. Depending on the power supply means, the limits of the applied power, the applied voltage and the applied current must be previously determined. Typical limits may be 100-200 V, 100-200 W, and 100-200 mA. These limits depend on many factors, including the buffer conductivity, the magnitude of the counter-flow gradient and the cooling rate.

Each of the exit tubes is connected to its respective collection vessel. The sample collection may thus be performed in a continuous and automatic manner.

Table 4 presents the results of a CFGF separation of two colored proteins, human hemoglobin (Hb) human serum albumin (A), introduced in three runs at different subcompartment input positions. Human Hb was isolated from packed, saline washed red blood cells. The cells were mixed with three volumes of deionized water, and centrifuged to removed lysed membranes. The Hb was spectrophotometrically determined at 80 mg/ml. Human serum albumin, 250 mg/ml from Baxter Healthcare, mixed 1:11 with saturated bromophenol blue in 0.1M NaOH. Blue stained Alb was then diluted to 80 mg/ml with deionized water. Equal volumes of Hb and albumin were mixed and diluted to 2 mg/ml total protein for separation. The running buffer used consisted of 10 mM Tris borate, pH 8.5 as a 1:10 dilution of electrolyte buffer. In the first run, the protein solution was introduced at subcompartment B, in the second run the sample mixture was pumped into the subcompartment E and in the third run the sample mixture was input into subcompartment G. For each experimental run, the protein input was 2 mg/ml, the sample input rate (F.sub.s) was 0.55 ml/min, the sample process rate was 1.1 ml/min, the fraction pump speed (F.sub.W) was 0.55 ml/min, no centering pump (F.sub.c) or throughput pump (F.sub.T) were employed. The voltage was 230, with a 0.55 gradient in ml/min*fraction, at a minimum counter-flow of 0.55 ml/min and maximum counter-flow of 5.0 ml/min.

The results clearly show that successful separation for all three experiments was achieved. The two proteins were collected in separate fractions from different subcompartments without any overlap.

All three experiments were run at the same counter-flow input and the same voltage. The results indicate that the albumin exited the separator in subcompartments C and D, while the hemoglobin exited in subcompartments F and G with a clear subcompartment E therebetween. These results also demonstrated that the position of the introduction of the proteins does not influence focusing position. A protein is focused by attaining its balance between electrophoretic velocity and the counter-flow and always migrates into its respective fraction defined by its electrophoretic properties. That the final position of the protein focusing is independent of its input position confirms that the counter-flow gradient focusing mode of the present invention is not a zone electrophoresis methodology, since the final position in zone electrophoresis will depend on the input position.

As an example of the application of counter-flow gradient focusing for biological cell separation, a mixture of sheep and rabbit red blood cells (Srbcs and Rrbcs, respectively) were separated in a 10-subcompartment electrophoretic cell. The Srbcs and Rrbcs were prepared from packed cells which were washed three times with running buffer (triethanolamine, 15 mM KAc; 4 mM; glycine, 0.24M; glucose, 11 mM; sucrose, to 0.31 Osm; HAc, to pH 7.2; 90 .mu.S) and stored at 4.degree. C. in running buffer for several days before use. A mixture of equal volumes of washed and packed Srbcs and Rrbcs, diluted 1:50 with running buffer was introduced into subcompartment 5 in these experiments. The differences in experimental conditions were the fluid withdrawal rate F.sub.W (equal to the gradient G) and the centering pump flow rate F.sub.c. Table 5, below, presents the results of the separations for these three experiments, clearly showing that a cell mixture could be resolved by this technique. By calculating the flows through the membranes according to the above equations, the flow limits which bound the Srbcs and Rrbcs shown in Table 5 were calculated: Srbcs were 1.1-1.6, 1.1-1.5 and 1.1-1.3, respectively for the three experiments, while the Rrbcs were 0.5-0.9, 0.7-1.1, and eluted from F.sub.c, respectively for the same three experiments. By manipulating F.sub.c and F.sub.W it was demonstrated that the gradient could be decreased while maintaining the red blood cells positioned in the center of the electrophoretic cell.

In another example of cell separations, a sample of human red blood cells were initially injected in vitro with Plasmodium falciparum malaria. The experimental protocol included washing packed human red blood cells three times with RPMI 1646 medium and mixing 0.2 ml of washed packed cells with 10 ml of RPMI 1640 culture medium containing 5% human serum, then culturing at 37.degree. C. for 3 days. Prior to separation, the cells were washed three times with running buffer. Packed, washed cells were diluted 1:100 with running buffer for use. Before staining, the infected cells were distinguished from the healthy ones by observing the presence of massive dark substructures in the infected cells. Thus, it was determined by counting the cells under the microscope that about 3% of the cell culture was infected. When this mixture was separated by counter-flow gradient focusing, fractions 5-8 contained the rbcs with 2,1, 19 and 25% infection, respectively. This indicates that the normal and infected rbcs had sufficient mobility differences such that an 8-fold enrichment in infection was achieved in fraction 8.

In the present configuration, the membranes separating the subcompartments were obtained from a commercially available source. They are composed of a woven monofilament nylon or other materials such as Teflon, polyethylene, polypropylene, polyester or urethane. Several porosities were used and a relation between resolution and pore size was observed. When the membrane pore size was too small for cell passage, the electrophoretic and counter-flow movement of the cells was prevented. A pore size larger than that needed for cell passage did not prevent convective fluid mixing between subcompartments and poor resolution was observed.

As the nylon was specified to have a neutral charge, i.e., no surface charge, a small degree of natural oxidation results in carboxylate ions which, in turn, caused the nylon to have a negative surface charge. Because the negative charge was immobilized on the nylon membranes, there was a resulting rise in electroosmotic fluid pumping in the same direction as the counter-flow (cathodic) and a hydrostatic pressure gradient in the opposite direction. FIG. 6e illustrates this principle.

To determine the fluid withdrawal rate which would produce a counter-flow gradient equivalent to the electroosmotic pressure gradient, the fluid withdrawal rate was slowly increased, thereby increasing the magnitude of the counter-flow gradient until which point the mechanical pumping pressure gradient cancelled the electroosmotic gradient pressure. The magnitude of the mechanical pumping at this null point was within the limits of the pumping range used to obtain good resolution in cell separations. This result suggests that either the counter-flow gradient focusing technique could be improved by reducing the surface charge of the membrane to zero and rely solely on the mechanical pump to generate the flow gradient, or a choice of nylon surface charges could be selected to aid the separation process.

The surface properties of the membrane can be altered by covalently binding appropriately charged groups to the membrane surface. In addition to the negative surface charges (from carboxylate or sulfonate groups, positive surface charges (from substituted ammonium ions such as quaternary ammonium ions), hydrophilic groups (from aromatic and aliphatic polyols), hydrophobic groups (from aromatic and aliphatic hydrocarbons) and specific affinity groups (from antibodies, substrates or ligands) are contemplated.

Specific affinity groups allow the electrophoretic separation to take on a new dimension using, for example, antibodies to immobilize the sample of choice on the membrane while eluting the unwanted contaminants. Subsequently, the cells of choice can be eluted by the use of dissociating reagents to break the antigen-antibody bond, such as polyethylene glycol, pH change or ionic strength change, or by using other techniques such as voltage pulsing, electric field reversal, sonication or mechanical vibration. In another mode, the impurities are immobilized on the membrane surface using antibodies specific to the impurities while the material of choice elutes and is collected in the void volume. Subsequent dissociation conditions could then recharge the electrophoretic cell for an additional separation.

Conventional zone electrophoresis, of course, may be performed in the apparatus of the present invention by simply omitting the counter flow gradient principle.

FIG. 7 is a schematic representation of yet another embodiment of the present invention relating to the application of the apparatus of the present invention for electrodialysis, whereby the principle of electrophoresis and a continuous flow operation are employed.

The configuration of a preferred embodiment comprises one subcompartment 12 with two electrode compartments 30, 32 on either side. The compartments are separated by ion exchange membranes 19. The sample to be treated is introduced through the inlet port 11 of the subcompartment 12 and exits through the collection port 13 at the opposite end of the subcompartment 12.

The electrodialysis device according to this embodiment of the invention is operated such that the sample solution is pumped into the subcompartment 12 in such a manner to first expel all the air in the separation chamber channel 16. When the channel 16 is filled with the sample solution, electrical current is applied at preset voltage, current and power limits. The ionic species, which are continuously pumped into the subcompartment 12, migrate under the electrophoretic force through the ion exchange membranes into the electrode compartments, leaving the depleted or cleaned solution behind, which may be continuously pumped off at the exit ports.

The effectiveness of the electrodialysis method of the present invention was demonstrated by desalting water which contained 0.15M NaCl.

Finally, according to another embodiment of the present invention, the present invention contemplates a water purification process, in which pathogenic organisms, toxins and/or heavy metals may be separated, concentrated and removed from a water sample. Waste water may contain pathogenic organisms, toxins and heavy metals. The water purification process of the present invention is useful in generating potable water from waste streams, gray water, or industrial or agricultural waste streams containing one or more of pathogenic organisms, toxins, toxic metals, and colloidal soil particles. The water purification process is particularly useful in bioregenerative closed systems required in confined environments such as in spacecraft or submarines where recycling of water is a necessity.

According to the present invention, there is provided a method of water purification which employs the counter-flow gradient focusing methodology and continuous flow electrodialysis previously disclosed. Standard buffer systems may be employed as a counter-flow buffer solution. Alternatively, an acid and a base pair may be introduced into the anode and cathode electrodes, respectively, whereby no flow of the respective ions across the separation cell occurs since the respective ions are already at their destination. If, for example, NaOH is introduced into the cathode compartment and H.sub.2 SO.sub.4 is introduced at the anode, the Na.sup.+ ion remains at the cathode and the Cl.sup.- ion stays at the anode; no electrolyte transport is observed. Alternatively, standard buffers such as tris acetate, pH 7.2, can be used as electrolytes to furnish the buffering ions to the separation chamber to facilitate the removal of pathogens and toxins containing ionizable functional groups. The buffers used in this process are subsequently removed in tandem by continuous flow electrodialysis in the present invention from the water in those subcompartments that do not contain pathogenic organisms or pathogens.

Water samples containing heavy metal ions and pathogens have been purified successfully with the above described counter-flow gradient focusing method. As an example, which is not intended to be limiting of the scope of the invention, lead has been separated from a waste stream containing E. coli and Rhizobia using the counter-flow focusing methodology. A twelve subcompartment array was employed as previously discussed. As summarized in Table 6, there was a sharp focusing of lead ions, with minimal or no presence of lead ions noted in other subcompartments of the apparatus.

The sharp focusing which can be achieved with the counter flow gradient leads to the concentration of undesired materials into very small volumes in certain subcompartments, while maximizing the volume of purified water in other subcompartments. The contaminants may thus be removed from those specific subcompartments, while purified water may be pumped off others in a continuous mode.

The effectiveness of the separation process for biological materials may be monitored visually, using bioluminescent assay and/or the polymerase chain reaction employing specific gene probes.

The present invention is, of course, in no way restricted to the specific disclosure of the specification, examples and drawings, but also encompasses any modifications within the scope of the appended claims.