by ultrafiltration prepare 50 g of a 1 g/L protein solution and ultrafiltrate with 5 volumes of a 50 mmol Tris buffer, pH 7.0.

Stock solutions of affinity ligands, e.g., PEG-bound ligand, can be prepared at a concentration of hundredfold or more, as the final concentration of these species in the 2-phase system is quite low. The cost of the PEG derivatives limits the amount of stock solution to be prepared.

3.2. Selection and Preparation of Aqueous Two-Phase Systems

The selection of the appropriate phase system depends on the final application. PEG/Dx is by far the most common pair of polymers used. Other incompatible polymers were already mentioned and summarized in Table 2. High molecular weight Dx (Dx-500 000) is highly recommended, since it can be used with low molecular weight PEGs reducing the viscosity of the phases. Also popular are PEG/K2HPO4-KH2PO4 systems. After the phase forming species have been selected, the next step is to select a particular tie-line. Good sources of tie-lines are the monograph by P. A. Albertsson (2), the book by Zaslavsky (56), which contains about 150 phase diagrams for Dx/PEG, Dx/ Polyvinylpyrrolidine (PVP), Dx/Polyvinyl alcohol (PVA), Dx/Ficoll, PEG/Polyvinyl methyl ether, and PEG-salt, and several articles (16,43,44). Some convenient 2-phase systems are shown in Table 4. As a rule of thumb, those who are using aqueous 2-

phase systems for the first time should choose equal volumes of top and bottom phases to facilitate sampling and protein partition coefficient determination. For analytical purposes, 5- or 10-g systems are very convenient. Although any buffer can be used, for acidic and neutral pHs, phosphate buffers are recommended, whereas for basic pHs, Tris buffers can be used. Buffer concentration should be kept between 20 to 50 mM. It is less cumbersome to work at room temperature since the mixing, equilibration, and sampling has to be done at the same temperature. Because protein partitioning is only marginally affected by temperature changes, low temperatures may be desirable to maintain protein stability. The final preparation of an aqueous 2-phase system is quite straightforward, and a detailed recipe can be found in Reference 11. An example is given in Procedure 2.

❖ Procedure 2. Preparation of a Dx-500 000/PEG-4000 System at Room Temperature

1. Shake the stock solutions well so that there are no density gradients.

2. Place a graduated centrifuge tube of 15 mL total volume on a weighing balance.

3. Weigh out the stock solutions into the tube in order of their increasing densities, and layer them carefully over each other. This facilitates the removal of portions of one stock solution in case of error during weighing. Because of the problem of accurately pipetting the polymer stock solutions due to their high viscosity, they are best measured by weight and are easily transferred using a Pasteur pipet with a broken tip.

4. Mix the contents of the test tube thoroughly, first by hand, and then in a rotary shaker (20 min is enough) at the equilibration temperature.

5. Let the systems settle for a period of 30 minutes to 24 hours depending on the system composition, or centrifuge them for 2 to 15 minutes at 1500x g. Poor temperature control in centrifuges makes it more convenient to sediment the systems in a water bath or in a chromatographic chamber when working at a temperature other than ambient. In general, the time of phase separation depends on the distance of the working tie-line from the critical point. Close to the critical point, the phase separation time is long. At intermediate tie-lines the phase separation time is shorter. If the more viscous phase volume is larger than the volume of the less viscous phase, the phase separation time increases. If the system is to be used in a liquid—liquid partition chro-matography system, one must chose a total concentration of polymers such that the PEG-rich phase constitutes most of the volume. This phase system must be allowed to settle for 24 to 48 hours before the PEG-rich phase is used as the mobile phase.

If instructions are followed carefully, the preparation of aqueous 2-phase systems should be routine laboratory work. Still, a common source of frustration for those using aqueous 2-phase systems for first time is their apparent lack of reproducibility. As indicated before, systems are normally prepared according to some published binodal. Often, the prepared system differs from the one published. Specifically, the ratio between top and bottom phase volumes of both published and prepared systems may not be the same. The appearance of only one phase after following a published recipe step-by-step is equally frustrating. These apparent inconsistencies cause people to believe that the lack of reproducibility is an inherent property of aqueous 2-phase systems. Fortunately, this is not true. The most common reasons for these inconsistencies are:

• The selected tie-line is too close to the critical point. So, small differences in the molecular weight or molecular weight distribution of the polymers, presence of additives, or differences in temperature moved the system into the one phase region. Addition of small amounts of one of the two polymers will move the system into the 2-phase region.

• The selected tie-line is too far from the critical point. When working at longer tie-lines poor mixing is normally the cause for the lack of production of 2 phases. Since the denser stock solution is added to the centrifuge tube first, it is quite difficult to mix the residue of stock solution that is trapped in the tip of the tube with the rest of the solution. So, the 2-phase system is actually prepared using a considerable smaller amount of one of the two polymers. To assure good mixing, mix the content of the tube in a vortex mixer and inspect the tip and walls of the tubes for stock solution residues. Continue mixing after no deposits are present, and place the tube in a rotary shaker.

3.3. Preparation of Liquid—Liquid Partitioning Chromatography Systems

Specific guidelines for the use of this method can be found in the various articles by Muller (35-37). The main steps are outlined in Procedure 3.

❖ Procedure 3. Preparation of

Partitioning Chromatography Systems

1. Measure the partition coefficient of the raw materials in batch systems before attempting to run a liquid-liquid partition chromatography (LLPC) column. The partition coefficient of the target protein must be adjusted to be between 0.3 and 0.1 (the data shown in Figure 2 can be used for this purpose), and the salt concentration should be high enough to shield elec

Figure 4. Purification of hemoglobin from E. coli. Cell homogenate is mixed with Dx and PEG solutions. The Dx-rich phase is discarded, and Na2SO4 is added to the PEG-rich phase. The Na2SO4-rich phase is discarded, and MgSO4 is added to the PEG-rich phase. The protein is recovered from the salt-rich phase.

Figure 4. Purification of hemoglobin from E. coli. Cell homogenate is mixed with Dx and PEG solutions. The Dx-rich phase is discarded, and Na2SO4 is added to the PEG-rich phase. The Na2SO4-rich phase is discarded, and MgSO4 is added to the PEG-rich phase. The protein is recovered from the salt-rich phase.

trostatic interactions between the proteins and the gel (ionic strength of 0.05 or higher).

2. After the optimum conditions to obtain an appropriate partition coefficient have been identified, prepare enough top phase at the right pH and at the right ionic strength to elute the column. The top (PEG-rich) phase is allowed to equilibrate for several days in the presence of small amount of bottom phase.

3. The column is packed according to Muller (36) and equilibrated until the UV noise of the effluent has dropped below 0.005 OD units at 280 nm. This ensures that all the Dx-rich phase that is not bound to the beads has been washed out.

4. The sample is dissolved in the mobile phase (it should not exceed 2%—3% of the bed volume for analytical runs and as twice as much for preparative runs) and injected into the column. The elu-tion is started immediately.

As in any chromatographic separation, sample preparation is quite important. If the starting material is a cell homogenate, solids must be sedimented out by centrifu-gation for about 15 minutes at 2000x g. The clear supernatant is mixed with the appropriate amount of PEG that is going to be used as a mobile phase. If aggregates are observed, they must be eliminated by centrifugation. If no precipitation is observed, more PEG is added (to reach 30%), the liquid is cooled in an ice bath for 10 minutes, and the protein precipitate is removed and resuspended in buffer.

Some features of these type of systems are: (i) the partition coefficients must be sufficiently different from 1 to make an impact on the retention times. That is, they must be in the right range to make a multicomponent chromatographic separation possible; (ii) the elution volumes cor relate quite well with the partition coefficients of the proteins obtained in batch experiments, so scale-up is straightforward; (iii) very low amounts of Dx are needed, which is of direct benefit as far as costs go. In addition, if the losses are proportional to the total Dx, then the losses are expected to be low as well; (iv) PEG precipitates large proteins (above 200 000 Da) at the stationary phase—mobile phase interface. This is avoided at all costs, as the precipitate clogs the column; and (v) the eluant contains significant amounts (around 10% wt/wt) PEG. Depending on the final application, this can be removed as described in the following section.

❖ Procedure 4. Determination of Partition Coefficients

1. Approximately 1 mL of the protein solution to be purified is added to the phase-forming species mixture replacing an equal amount of buffer. Mixing and phase separation are done as described above for systems that do not contain any protein.

2. Mixing has to be done carefully. It has to be vigorous enough to allow distribution of the proteins between the 2 phases but gentle enough to prevent protein denaturation (a rotary shaker is highly recommended, whereas the use of a vortex mixer is discouraged).

3. The phase systems are centrifuged at 1500x g for 20 minutes to speed phase settling.

4. Sampling is done by pipetting carefully 1 mL of top phase and 1 mL of bottom phase from each partitioning tube (the amount pipetted should be controlled by weighing for more precise sampling). Impurities may accumulate at the liquid—liquid or liquid—air interfaces. They do not constitute a problem unless they are pipetted during sampling, so a positive pressure on the pipet as it enters the phases is always recommended. Blank phase systems are sampled in the same manner.

5. The samples are diluted with buffer. The actual dilution depends on the particular protein and on its partition coefficient. Since the viscosity of the phases is very high, improper mixing of the sample and the dilution buffer may result. Uncontrollable scattering from regions of different densities within the sample produces erroneous absorbance readings. As a general rule, mix the sample of the phase with the dilution buffer and stir in a vortex mixer.

6. Leave the solution resting and stir again. Inspect the solution to detect density differences along the axial direction of the test tube. Continue stirring until the solution is completely transparent. Because of the relatively high absorbance of the blanks at 280 nm, an absorbance reading of protein containing samples between 0.5 and 1 is recommended. Hemoproteins are conveniently measured at 540 nm. Standard protein tests like Bradford's test (8) can be also used.

7. The partition coefficient is calculated from K = [Absorbancesamp[e - Absor-

banceblank]top/[Absorbancesample - Ab-sorbanceblank]bottom.

8. For preparative applications a precise mass balance is not necessary. For analytical purposes, a protein mass balance can easily be performed, since the volumes of the phases are very easy to measure, and the density of each phase is well correlated with polymer concentration. If the mass balance is not close (within 5%), check for the formation of a precipitate at the liquid—liquid interface. If a precipitate is present, one should use more diluted protein solutions (a decrease of 50% in protein concentration is usually enough). If no precipitate is present, poor sampling is probably the source of error.

3.4. Removal of PEG

Depending on the final application of the protein purified by aqueous 2-phase extraction, it may be desirable to eliminate all or most of the polymer that contaminates the protein of interest. Although the overall yield of the separation may be reduced, one of the easiest ways of eliminating the polymer (usually PEG) from the protein solution is to repartition the PEG-rich phase against a salt (phosphate or sulfate) rich phase. This is accomplished rather easily by first separating the top (PEG-rich) phase from the bottom (Dx-rich) phase and by adding either solid sodium phosphate or sulfate directly into the PEG-rich phase. By driving the protein into the salt-rich phase most of the PEG is eliminated. If the size of the protein is sufficiently different from the size of PEG, PEG-protein mixtures can be separated by ultrafiltration and by gel permeation chro-matography. For example, we have used Ultra free-20 (Sigma) centrifuge tubes with a nominal molecular weight cut-off of 10 000 to separate lysozyme from PEG-4000. The samples were centrifuged at 12 000x g for 30 minutes. Up to 85% of the PEG present is eliminated in this way.

3.5. Methods for Characterization Experiments

3.5.1. Cross-Point Determination

Prepare 2 sets of Dx/PEG systems (Set A which contains alkali chloride and Set B which contains alkali sulfate) spanning a pH range from 3.5 to 11.5. Two runs are highly recommended for precision work. In the first run, 4 or 5 different pH values are enough. In the second run, 5 or 6

points should be obtained in the neighborhood of the cross-point.

❖ Procedure 5. Cross-Point Determination

1. Set A. Add to a 10-mL centrifuge tube (37): (i) 2.5 g of Dx stock solution; (ii) 1.0 g of PEG stock solution; (iii) 3 g of sodium chloride; (iv) 2.5 g of buffer; and (v) 1 g of the protein stock solution.

2. Set B. Add to a 10-mL centrifuge tube: (i) 2.5 g of Dx stock solution; (ii) 1.0 g of PEG stock solution; (iii) 3 g of or sodium sulfate solution; (iv) 2.5 g of buffer; and (v) 1 g of the protein stock solution.

3. The final phase system composition is 7.5% (wt/wt) Dx, 5.0 (wt/wt) PEG, 0.1 M alkali chloride or 0.05 M alkali sulfate, and 0.04 M glycine or sodium phosphate buffer.

4. Prepare blanks of the phases without added protein.

5. Mix, equilibrate, and sample the phases as explained in Section 3.4.6.

6. The pH in each phase is measured with a microelectrode directly on the undiluted phases. Because of the high viscosity of the phases, the pH measurements must be done over a relatively long period of time.

7. The partition coefficients of Sets A and B are plotted versus the pH. The pH and the partition coefficient values at which one Kp versus pH line (Set A) crosses the other one (Set B) and are read from the axes.

The lines of Kp versus pH may not cross each other because of errors in pH or in the values of the partition coefficients. One must be sure that the pH has been measured long enough to reach equilibrium and that the pH of both phases agrees within the experimental uncertainty (approximately 0.05 pH units). Erroneous values of Kp are generally due to poor sampling, and a protein mass balance should be done to assure that sampling has been done correctly.

The sensitivity of cross-partitioning depends upon the angle at which the 2 lines intersect. If the lines are perpendicular, the sensitivity is at a maximum, while parallel or nearly parallel lines yield no cross-point or a "cross-point range". The slope of the lines depends upon the type of salt, the change in net charge of the protein with pH, the specific interactions between the ions and the proteins, and the salt-induced changes in the interactions between polymer and protein. The sensitivity can be manipulated by varying the molecular weight of the polymers, the temperature, the concentration of the polymers, and the type of salt. So, cross-partitioning should be done using the lowest possible PEG molecular weight to minimize problems associated with the high viscosity of the phases. If the sensitivity is not good enough, the experimentalist needs to explore different conditions until a good sensitivity is found.

The pH and the partition coefficient at the cross-point are only marginally dependent on the combination of salts used and on their concentration. For example, NaCl can be replaced by potassium chloride and/or sodium sulfate by lithium sulfate without affecting the results. Still, some small differences in pH values at the cross-point with different salts have been observed. These differences are similar to those encountered in the electrophoretic determination of isoelectric points, which can also be slightly affected by the salt used.

This independence of cross-partitioning on the type and concentration of salt makes cross-partitioning a viable option for determining the isoelectric point of proteins that are stable only at high salt concentrations. In contrast, the type and concentration of salts have a strong influence on the shape of the ln Kp versus pH curves.

3.5.2. Surface Hydrophobicity

Dx/Ficoll-400 systems are prepared by weighting stock solutions of the polymers to a final concentration of 12.5% Ficoll and 10.8% Dx-70. The systems are prepared in sodium phosphate buffer at pH 7.4 at a concentration of (56):

where the concentration of the buffer is varied from 0.01 to 0.11 M and the concentration of NaCl is varied from 0 to 0.15 M. The protein(s) of interest is partitioned in this set of systems as indicated above. The logarithm of the partition coefficient is plotted versus the ionic strength. The zero intercept yields a parameter that represents the strength of all the interactions of the protein with an aqueous environment relative to that of a methylene group and the slope yields a parameter that reflects the strength of the hydration interactions of all the iono-genic groups of the protein relative to that of the a-carboxyl group of DNP-amino acid.

0 0

Post a comment