Protein Separation Via Ion-Exchange Chromatography And

Protein separation via ion-exchange chromatography and

identification via SDS-PAGE.

Introduction

Proteins are a sequence of aminoacids that have functional groups such as carboxylic acid, amines, sulfhydryl, guanidiino, etc. These functional groups can either donate or attract protons at varying pH conditions, leading to a negative, positive, or a neutral charge. It is the sum of the charges in the functional groups that will determine the net charge of the protein.

The net charge determines physical properties of the protein such as its attraction to a resin with charged functional groups (known as ion-exchange resins). By manipulating the charge in proteins, attractive forces between the protein and the ion-exchange resin can be either reversibly strengthened or weakened. Since proteins tend to have different functional groups, the rate at which different protein will elute in an ion-exchange column will differ from protein to protein. The greater the electrostatic potential is, the greater the attractive forces will be, and the slower the elution will take place. A salt or pH is then used to desorb those proteins with the highest electrostatic attraction between the protein and the ion-exchange resin. On one hand, the salt will competitively bind to the resin and remove the protein; pH, on the other hand, ionizes functional groups to enhance or weaken electrostatic forces.

Another physical property of proteins useful for their analysis is the molecular weight. Actually, it is possible to separate proteins based on their molecular weight. First, heat is used to disrupt the attractive forces in proteins to partially linearize the molecule. Second, B-mercaptoethanol is used as a reducing agent to disrupt disulfide bonds to linearize the protein. Finally, sodium dodecyl sulfate (SDS), an anionic detergent, binds to proteins in a ratio of 1.4:1 (mass of SDS:mass of protein). By doing so, the proteins will be rod-like structures surrounded by the negatively charged SDS molecules. Since the ratio of SDS to protein is constant, the negative charges in the SDS-surrounded protein has a positive linear dependence to the molecular weight and negative linear dependence to the displacement of protein in an electric field; therefore, the molecular weight and displacement’s negative dependence can be found.

Based on aforementioned principles, the purpose of this experiment is to separate three proteins based on their overall isoelectric points and estimate their recovery with the use of mass balances. This will be done via ion-exchange chromatography using a step-change in NaCl concentration. Additionally, identification of proteins will be done using SDS-Polyacrylamide Gel Electrophoresis (SDS-PAGE).

Theory

Electrostatic interactions

Electrostatic interactions govern ion-exchange chromatography. The attractive forces between two oppositely point-charges are described by Coulomb’s Law:

F = -q1·q2/d·r2 Eq. 1

Where q1 and q2 are the charge of point charges, d is the dielectric constant of the medium and r is the distance between q1 and q2. For this reason, the attractive force will repel if the charges are the same, and vice versa.

Protein Charge

Handerson-Hasselbach equation describes ionization behavior upon changes pH changes.

pH = pKa + log ([A-]/[HA])  Eq. 2

where pH is a measure of hydrogen activity, pKa is the negative log of the acid dissociation constant, A- is the negatively charged species, and HA is the conjugate acid. Since pKa is constant, the log ([A-]/[HA]) will increase if the pH of the solution increases; thus, the negative charge (A-) will increase. It follows from this simple analysis that proteins will have negative charges if they are above their isoelectric point (or pH at which they have no charge), and vice versa.

Mass balances

Beer-Lambert is an empirical relationship that provides a positive linear dependence of absorption of light and concentration. Typically, Beer-Lambert law is expressed as

A=lec Eq. 3

where A is absorbance, l is the path length traveled by the light, e is the extinction coefficient and c is the protein concentration.

On the other hand, the mass of the protein in column can be expressed as:

m = ∫ C· F·dt Eq. 4

where m is the mass of protein, c is the concentration of protein, f is flow rate and dt is differential time.

By combining equation equation 3 and 4, a new relationship of the mass of protein flowing out of the column with absorbance is obtained:

m = kR Eq. 5

where m is the mass of the protein that flows out of the column, k is a constant dependant on flow rate (F), extinction coefficient (e), and light path length (l), and R is the area under the curve in absorbance (A) versus time (t) plot.

Since the initial mass of protein can be determined, and the mass of the exit stream can be determined, the governing equation is.

m0-m=0 Eq. 6

where m0 is the initial mass of protein.

Langmuir isotherm

A system in which there the following equilibrium exists:

Keq,1 = [SP] / ([P][S]) –Eq. 7

Keq,2 = [SCl] / ([Cl][S]) –Eq. 8

where P- is the protein anion, S+ is the unoccupied binding site, SP is the protein-binding site complex, Cl is chloride ion, and SCl is the chloride-binding site complex, Keq are equilibrium constants. Mass balances on the sites in the resin, indicates that:

Stot = [S] + [SP] + [SCl] –Eq. 9

where Stot is the total amount of binding sites. The following Langmuir isotherm can be obtained from combining equations 7-9.

[S P] = (Keq,1 Stot [P])/(1+ [P] Keq,1+ [Cl] Keq,2 ) –Eq. 10

Experimental Methods

Reagents

All proteins were purchased from Sigma (St. Louis, MO, U.S.A.). Cytocrhrome C, Myoglobin and Bovine Serum Albumin (BSA) were dissolved in 0.02 M Tris buffer (pH 8.5) to 0.15 mg/ml, 0.07 mg/ml and 0.3 mg/ml respectively. Previous to loading, the dissolved protein was filtered in a 0.22 μm hydrophilic polyvinylidene (PVDF) to remove particulate matter.

Extinction coefficient was previously determined to be 0.0610±0.002 for BSA, 0.21±0.002 for Cytocrhrome C, and 0.170±0.001 for Myoglobin using a Shimadzu UV-Vis Spectrophotometer Model #UV160U.

Ion-exchange column

The resin used was Macro-Prep High Q Support (methylacrylate copolymer) from Bio-Rad (Hercules, CA, U.SA.). All solutions (sterilization, regeneration, equilibration, protein, wash, elute) were introduced into the column using a peristaltic pump at a flow rate of approximately 1/6 (1mL/min) solution volume flow per unit of liquid volume per minute.

Ionic strength in the medium was step-changed using solutions of 0.02 M Tris buffer with 0 M, 0.3 M, 0.5 M and 1.0 M NaCl (pH 8.5).

Data acquisition and validation

Conductivity and UV readings were collected using LabView from National Instruments (Austin, TX, U.S.A.). Fractions were 3 mL in volume and analyzed further in 15% Tris-HCl ready gels (SDS-PAGE).

SDS-PAGE was ran on Mini-PROTEAN II and the low range molecular weight standard was used; both manufactured by Bio-Rad.

Results

Molecular weight estimation

The low range molecular weight standard was Bio-Rad was ran along with the eluted proteins. From this standard, a rough molecular weight estimations can be made: cytochrom C is 14-21 KDa and Myoglobin about 21 KDa. BSA is difficult to determine because the standard is smeared. This compared to the theoretical values: cytochrom should have 12KDa, Myoglobin 17.6 KDa, and BSA 66.5 KDa.

A positive control of the proteins used was run in the gel. From the positive control it is clear that the fractions with each protein were identified.

Protein separation

Two different SDS-Page gels were run because the first gel did not show the stain corresponding Myoglobin.

Figure 2 shows that the fractions 9, 15, 17 and 23 contained cytochrome C; fraction 25 contained BSA. It also shows that fraction 18 contained myoglobin.

Thus, these gels suggest that the elution order was: cytochrome C, myoglobin and BSA.

Protein mass balances

In figure 3, gels identified the peak 1 and 2 to be cytochrome C (fractions 9, 15, 17), peak 3 to be myoglobin (fraction 18), and peak 4 to be BSA (fraction 25).

Figure 3 shows 4 peaks. Using polymath, the area under the peaks (of the A.U. versus time plot) were found to be, from left to right, 0.106±0.024, 0.036±0.007, 0.059±0.007, and 0.119±0.013 minutes.

By using equation 5, the mass that the peaks represent are, from left to right, 0.438±0.100, 0.149±0.029, 0.174±0.021, and 0.975±0.110 mg of protein. Which represent percent recovery of 97.8±21.5% for cytochrome C, 62.0±7.7% for myoglobin, and 81.3±9.2% for BSA.

Salt concentration effects

No step change was necessary to recover 73.0±16.7% of cytochrome C. A step change in NaCl resulted in further cytochrome C desorption from column. The first step change, corresponding to 0.1 M NaCl, desorbed 24.8±4.9% cytochrome C and myoglobin out of the column; the second step change of 0.3 M NaCl detached BSA.

Discussion

Protein elution order

Because all the proteins were buffered at a pH of 8.5, it is expected from the Hendelson-Hasselbach equation that the net charge for cytochrome C will be positive. In other words, cytochrome C should be repelled from the column. However, results show that only 73.0±16.7% is repelled; the rest elutes after addition of 0.1 M NaCl. Results suggest that part of the cytochrome C binds -as opposed to what is predicted by the net charge model used. This behavior is not strange because net charges do not take into account that the surface charges might not be symmetrically distributed. In fact, it has been suggested by retention mapping studies that the net charge model is inadequate (Kopaciewicz et al., 1983). The only bovine cytochrome NMR shows that the negatively charged residues, indicated in figure 4, are on the surface (Muskett and Whitford, 2003). An improved model, therefore, is that sometimes the negatively charged surfaces have a probability to attach to the resin despite the protein’s positive net charge. However, since these charges are weak, it is easy to desorb them –a 0.1 M increase in NaCl detaches them immediately as indicated in figure 3.

The order of elution of the other proteins followed expectations. At pH 8.5 myoglobin should elute first because its pI is 7.2 and its electrostatic force, indicated by Coulomb’s Law, is weaker –as opposed to 4.7 pI of BSA.

An approach to quantitatively assess the electrostatic attractions between a protein and the resin is by treating each ionizable functional group as a force vector. The addition of these vectors should give an idea of the total force that the protein is attracted to the resin. However, the implementation is out of the scope of this lab report.

Protein recovery

Protein recovery was not constant. Percent recovery of 97.8±21.5 was observed for cytochrome C, 62.0±7.7 for myoglobin, and 81.3±9.2 for BSA. It is interesting to observe that the recovery of repelled proteins, such as cytochrome C, is higher than attracted proteins, such myoglobin and BSA. Although this seems to indicate that proteins that are most strongly attached to the resin will have less protein recovery, it is not always the case. BSA, the protein that is most strongly bounded, has a recovery of 81.3±9.2 compared to the 62.0±7.7 for myoglobin. Thus, even if the errors are considered, results indicate that some protein did not exit.

To address these considerations, we can use cytochrome C as a working example. As seen earlier, the behavior of proteins binding to the column is a probabilistic phenomenon (recall that some cytochrome C was even found in fraction 23). Proteins that bind to the column from patches of negative charge in the surface will tend to attach stronger. Thus, a reasonable speculation will be that myoglobin’s charges are distributed asymmetrically -as seen in figure 5. BSA on the other hand, has charged distributed throughout its tertiary structure.

Another factor that might have an effect on attractive forces is the size. Whereas it is difficult for BSA to attach different sizes simultaneously to the ion-exchange resin, it is easy for myoglobin to orient its negative charges towards the ion-exchange matrix.

Error analysis

The raw A.U. data from the UV unit was noisy. Even when buffer was running and the apparatus was recalibrated to zero; the A.U. values fluctuated intermittently from -0.06 to 0.05. Since the fluctuations were symmetrical about 0 A.U, it is reasonable to assume that the noise is a random error; the average of all the values of the noise over time should be the set point. The set-point was calculated to be -0.00027, the standard deviation due to noise was determined to be 0.00140.

The standard deviation was added to each data-point that corresponded to each peak in figure 3. A new set of data-points with standard deviation was obtained. These data was then inputted into Polymath to calculate the area under the curve using the Simpson’s area approximation algorithm.

The area under the peak was subtracted from the area under the peak with error. The difference should be the error of the area due to noise.

Molecular weight determination

Although the molecular weight estimations based on the low range molecular weight standard were off by 2 to 4 KDa. It did follow the prediction that the proteins with lower molecular weight would cover a longer distance. Despite these differences, it does not posses significant relevance for the experiment. Positive controls were used to determine the identity of the proteins. Since the eluted protein aligned with the positive controls, it can be concluded that the eluted proteins are indeed the proteins of interest –and not contamination.

Salt concentration effects

Decisions on steps of NaCl concentrations were made arbitrarily. However, the resolution was good since no fraction contained more than one protein. A better approach to this is by constructing Langmuir isotherms (see Figure 5). By doing so, it is possible to determine the precision at which the protein of interest will detach from the ion-exchange resin. This is useful in proteins with close isoelectric points require to be separated using ion-exchange chromatography.

Another important aspect from figure 5 is that the detachment of proteins becomes subsequently difficult. An increase of 0.5 mg/ml decreases the protein concentration on resin from ~0.08 to 0.06, whereas an increase of 90 mg/ml decreased the protein concentration on resin from ~0.01 to 0. This is because the Langmuir isotherm is represented as:

[S P] = (Keq,1 Stot [P])/(1+ [P] Keq,1+ [Cl] Keq,2 )

Considering that figure 5 shows the effect of equilibrium binding of a monovalent protein, it is even probable for a plurivalent (such as myoglobin) protein to attach strongly to the ion exchange resin. However, as the Langmuir isotherm shows, it is difficult to detach the last portion to obtain a 100% recovery.

Conclusions

Although the study of ion-exchange chromatography based on the net charge model is inadequate, as it does not describe the reason why two peaks for the same protein was observed or the reason why the recovery of different proteins was different, it is generally a reasonable description of protein elution order.

The study of protein structure can shed light on how proteins will elute. Asymmetrical distribution of charges will typically increase the ways in which proteins can attach; high molecular weight, on the other hand, will decrease the ways in which protein can be absorbed into the ion-exchange matrix. Since the protein structure is complex and variable, it is difficult to develop a model that takes into account the structure of the protein that desires to be separated on a case-per-case basis.

Despite the limitations of the existing net charge model, the resolution was good. None of the fractions had more than one protein and the resolution is 5.26±0.1. Recovery of the protein was fairly decent: 97.8±21.5% for cytochrome C, 62.0±7.7% for myoglobin, and 81.3±9.2% for BSA. Despite a 100% protein recovery is desired, Langmuir isotherm shows us that doing so is inefficient –since it will require high quantities of competitors for the binding site.

Nomenclature

References

Bhattacharya, A.A., Curry, S., Franks, N.P. 2000. Binding of the general anesthetics propofol and halothane to human serum albumin. High resolution crystal structures. J.Biol.Chem. 275,38731-38738.

Harrison, R.G., Todd P., Rudge, S.R., Petrides D.P. 2003. Bioseparations science and engineering. New York, 2003.

Maurus, R., Overall, C.M., Bogumil, R., Luo, Y., Mauk, A.G., Smith, M., Brayer, G.D. 1997. A myoglobin variant with a polar substitution in a conserved hydrophobic cluster in the heme binding pocket. Biochim.Biophys.Acta.1341, 1-13.