Browsing by Author "Peter K. Kilpatrick, Committee Member"

Now showing 1 - 4 of 4

Fc-binding Hexamer Peptide Ligands for Immunoglobulin Purification
(2008-08-18) Yang, Haiou; Carla Mattos, Committee Member; Patrick V. Gurgel, Committee Member; Jason M. Haugh, Committee Member; Ruben G. Carbonell, Committee Chair; Peter K. Kilpatrick, Committee Member
Antibodies and their fragments have found a wide array of applications as pharmaceutical compounds, in addition to their common usage in the purification and localization of proteins. Antibody-based therapeutics accounts for roughly 20% of the therapeutic products in development in the USA, with the majority being of the immunoglobulin G (IgG) isotype. For IgG purification, affinity chromatography has been greatly used where Staphylococcus aureus Protein A and Streptococcus Protein G are the most common affinity ligands for IgG. However, the drawbacks associated with these two proteins have given rise to the searching for alternative affinity ligands for antibody purification. We therefore set our research goal as to find a small peptide that can bind hIgG through its Fc portion and can be used in purification of antibodies and Fc-fusion proteins. Small peptides are interesting for their advantages of being more stable, less immunogenic, less expensive and milder in elution than protein ligands. Peptides were searched by a radiolabeled-screening technique in a combinatorial linear hexamer peptide library built on solid phase Toyopearl AF-Amino resins. The screening identified a family of ligands with homogenous composition of His + aromatic group + positively charge group having the ability to match Protein A in binding human IgG (hIgG) through its Fc portion. The selectivity to the Fc portion is comparable to Protein A. The HWRGWV ligand of the Fc-binding peptide family has been investigated in many aspects and exhibits some interesting tributes. It has broad binding spectrum. It can bind all subclasses of hIgG, human IgD, IgE, IgM, and to a less extent human secretory IgA. It also displays the ability to retain chicken and several mammalian IgGs. Deglycosylation of hIgG has no influence on its binding to the HWRGWV ligand and the ligand does not compete with Protein A or Protein G in binding hIgG. It is suggested by the mass spectrometry data that HWRGWV binds to the pFc portion of hIgG and interacts with the amino acids SNGQPEN in the loop Ser383 — Asn389 by specific interactions. The selectivity of the HWRGWV resin to Fc over Fab is affected by its peptide density which also influences the equilibrium constant for hIgG. Increasing density increases both the association constant which is in the order of 105 M-1 and the binding of the Fab fragment. A ligand density of around 0.1 meq—g was determined to have both the specificity and appropriate affinity to Fc. HWRGWV demonstrates the ability to purify IgG. It can isolate hIgG from mammalian cell culture media containing 10% fetal calf serum (cMEM) with more than 95% of both purity and yield. The ligand was also used to isolate hIgG from Cohn II+III paste and an yield of 82% and purity of 73% were obtained in one step. The bound IgG can be recovered using phosphate buffer at pH 4 and its binding capacity for hIgG is 130 mg⁄g-dry-resin. Acetylation of the N-terminal amine does not pose any influence on either the static binding or the dynamic isolation of hIgG. Temperature has no significant influence on hIgG isolation from cMEM. Increasing peptide density improve the yield but with a compensation of the purity to a similar degree. Feed hIgG concentration in the range of 0.5 — 10 mg⁄mL affects the recovery yield where the yield is favored with higher IgG concentration. The separation of hIgG from cMEM by HWRGWV is comparable to Protein A at an initial hIgG concentration of 10 mg⁄mL and to A2P agarose gel at both 10 and 0.5 mg⁄mL, but using a milder pH 4 elution condition. HWRGWV was immobilized on Sepharose CL-4B. HWRGWV-modified Sepharose CL-4B can isolate hIgG from cMEM with similar purity to and a lower yield than on Toyopearl AF-Amino under the conditions optimized on the latter matrix. Our experiments also show that 2% AcOH is not strong enough to completely remove the bound proteins on HWRGWV. A better column regeneration and sanitation procedure is needed for longer column lifetime. This work demonstrates the possibility to use a peptide as short as six amino acids to mimic Protein A in IgG isolation by binding through the Fc portion.
Ligands from Combinatorial Peptide Libraries for Virus Removal
(2004-09-22) Salm, Jeffrey R; Jan Genzer, Committee Member; Dennis T. Brown, Committee Co-Chair; Ruben G. Carbonell, Committee Chair; Peter K. Kilpatrick, Committee Member
Small peptides were investigated as affinity ligands for virus removal from human blood plasma. Sindbis virus (SV) was radiolabeled with ³⁵S and screened against a solid phase peptide library. The screening identified 9 hexapeptides that were synthesized on a TosoBioseparations Toyopearl 650M Amino Resin. The peptide sequences SGKPVA and IATDGG were found to remove approximately 3 logs and 2.4 logs of SV from buffer respectively and approximately 1.5 logs of SV from 50% human blood plasma (HBP). Toyopearl Amino resin was used as a control and bound only 0.6 logs of SV in both buffer and 50% HBP. Good agreement was seen between infectious quantification methods and quantification of the virus using radiation. Injections of several batches of SV showed variations between the batches. A similar screening procedure was also applied to radiolabeled canine parvovirus (CPV). Screening against a solid phase library identified 25 leads through to be specific to CPV. Tests with a portion of these leads found less than 0.5 logs of CPV clearance in both buffer and 50% HBP. Electron microscopy of the CPV used in these experiments showed that large aggregates of virus were present in the radiolabeled virus. Infectious assays for CPV were also problematic and inconsistent. Work done in collaboration with the American Red Cross (ARC) identified several leads capable of binding porcine parvovirus (PPV) and human B-19 parvovirus (B-19). Selected leads were tested for viral clearance from 1 ml samples of spiked buffer and 50% HBP. The resins were shown to bind at least 4 logs of PPV and B-19 but less than 0.5 logs of CPV. Experiments with 10 ml samples of spiked buffer and 7.5% HBP showed that the resins were capable of 6 logs of PPV clearance. 10 ml experiments performed in 50% HBP showed an initial clearance of 6 logs of virus. Subsequent fractions showed decreasing clearance with complete breakthrough occurring by the 9th 1 ml fraction. Experiments that compared three different column systems showed that each system produced identical log clearance results. This work shows that screening a solid phase library can identify peptide ligands and that the leads can be used for viral clearance from HBP.
Peptide ligands that bind to staphylococcal enterotoxin B (SEB)
(2004-05-24) Wang, Guangquan; Steven W. Peretti, Committee Member; Zhilin Li, Committee Member; Robert M. Kelly, Committee Member; Peter K. Kilpatrick, Committee Member; Ruben G. Carbonell, Committee Chair
Staphylococcal enterotoxin B (SEB) is a primary toxin that causes food poisoning. It also acts as a superantigen that interacts with the major histocompatibility complex class II molecule (MHCII) and T cell receptor (TCR) to activate large amounts of T cells leading to autoimmune diseases. Highly purified SEB is needed in research and can be used as a standard in current detection methods for SEB. SEB is also a contaminant in protein A purification from Staphylococcal aureus fermentation broth. Inexpensive, robust ligands with high affinity would be suitable replacements for antibodies in biosensors for the detection of SEB. Affinity adsorption processes using short peptides as ligands show great promise in purifying and detecting proteins in comparison with other methods due to their high stability and low cost. By screening a solid-phase combinatorial peptide library, a short peptide ligand, YYWLHH, has been discovered that binds with high affinity and selectivity to SEB, but only weakly to other staphylococcal enterotoxins that share sequence and structural homology with SEB. Using column affinity chromatography with an immobilized YYWLHH stationary phase, it was possible to separate SEB quantitatively from Staphylococcus aureus fermentation broth, a complex mixture of proteins, carbohydrates and other biomolecules. The immobilized peptide was also used to purify native SEB from a mixture containing denatured and hydrolyzed SEB, and showed little cross reactivity with other SEs. To our knowledge this is the first report of a highly specific short peptide ligand for SEB. Such a ligand is a potential candidate to replace antibodies for detection, removal and purification strategies for SEB. Modeling the transport and kinetic processes in peptide affinity chromatography allows for a direct measurement of the rate of binding of SEB to peptide ligands. It can also provide design parameters for columns that can be used to either remove or detect SEB as well as other pathogens. The mass transfer parameters of SEB were either measured using pulse experiments or determined from correlations. The adsorption isotherms of SEB on YYWLHH resins with different peptide densities were fitted to bi-Langmuir isotherms. The general rate (GR) model was used to fit experimental breakthrough curves to obtain the intrinsic rate constants of the adsorption-desorption kinetics of the protein on the peptide ligands. An analysis of the number of transfer units in the column revealed that both intraparticle mass transfer and intrinsic adsorption rates were important rate-limiting steps for adsorption to the resin particles. The substitution level of peptide on the resin's surface has a significant impact on the resin's performance. The effects of peptide density on both equilibrium (batch format) and dynamic adsorption (column format) were investigated. The results revealed that the binding mechanism might change from univalent to multivalent adsorption with an increase in peptide density, thus explaining the variation of dissociation constants, maximum capacity, and rate constants with increases in peptide density. Adsorption processes can play an important role in helping to remove infectious pathogens such as SEB from solution without affecting other desirable components in a product stream. However, little work has been done on the design of adsorption columns specifically for the removal of several logs of infectious components. This requirement is much more stringent than the normal design of affinity separation columns where at most two logs of removal are sufficient. A column design strategy was developed, aimed at the removal of several logs of an infectious agent from a known volume of a process stream in a fixed amount of time. Two design options are analyzed with the general rate (GR) model, one fixing the column length and varying the fluid velocity, the other fixing the fluid velocity and varying the column length. The results indicate that the reduction in pathogen concentration is highly dependent on the residence time in the column, which is in turn dependent on the flow rate and column geometry. The theory, with no adjustable parameters, is shown to predict with great accuracy the effect of residence time on the log removal of SEB from an aqueous stream using an affinity resin with the peptide YYWLHH. Using a conventional Enzyme-Linked Immunosorbent Assay (ELISA) method, up to 5 log removal of SEB was detected, and the experimental log removal results agreed well with the theoretical prediction. The detection of minute amounts of SEB using High Performance Liquid Chromatography (HPLC), ELISA or Mass Spectrometry (MS) requires a concentration step prior to analysis. This can be accomplished by solid-phase affinity extraction on a peptide ligand column. It was shown that the YYWLHH resin has a great potential in sample preparation for SEB detection due to its high affinity and capacity. The peptide column was shown to capture all the SEB in highly diluted samples and it was possible to release SEB in a small volume of elution buffer. The enrichment of SEB from the YYWLHH column enabled an ELISA method to be able to work on diluted samples that cannot be analyzed without the peptide-affinity solid-phase extraction. The detection sensitivity of the ELISA method was improved from ng/ml to pg/ml using the YYWLHH column for SEB concentration.
Phase Equilibria of Diatomic Lennard-Jones Molecules Using Monte Carlo Simulation
(2006-08-06) Galbraith, Aysa Lamia; Orlin Velev, Committee Member; Keith E. Gubbins, Committee Member; Carol K. Hall, Committee Chair; Peter K. Kilpatrick, Committee Member
The overall aim of this research is to use computer simulations to study vapor-liquid and solid-liquid phase behavior of simple nonspherical molecules with a special focus on determining how the differences in the components' molecular size and intermolecular interactions affect the type of phase diagrams observed. We first calculate vapor-liquid phase diagrams for binary mixtures of diatomic Lennard-Jones molecules using Monte Carlo simulations and the Gibbs-Duhem integration method. We plot pressure versus composition vapor-liquid phase diagrams for the binary mixtures O₂-N₂, CO₂-C₂H₆ and N₂-C₂H₆ at different temperatures. We then add a quadrupole term to the two-center Lennard-Jones potential model and we observe that this further improves the agreement with experimental data. We also investigate the dependence of Henry's constant on the temperature, pressure and binary interaction parameter. We explore the effect of varying the molecular size ratio from σ₁₁/σ₂₂=1.0 to 1.40, intermolecular attraction ratio from ε₁₁/ε₂₂=0.70 to 1.20 and binary interaction parameter from δ₁₂=0.70 to 1.10 on the dumbbell mixture's phase behavior. We then examine the solid-liquid phase equilibria for systems containing pure Lennard-Jones dumbbell molecules and their mixtures. We begin by calculating the equations of state for systems containing C₂H₆, CO₂ and F₂, all of which are modeled by two-center Lennard-Jones potential for linear diatomic molecules. We then use the Frenkel-Ladd thermodynamic integration method to calculate the free energies. The equations of state and the free energies are used to obtain solid-liquid coexistence points which are needed to start the Gibbs-Duhem integration. The solid-liquid phase equilibria for pure and binary mixtures of Lennard-Jones dumbbells are predicted using the Gibbs-Duhem integration method. We use three model Lennard-Jones binary mixtures with varying bondlengths, σ₁₁/σ₂₂, ε₁₁/ε₂₂ ratios to calculate the solid-liquid phase diagram. Mixtures I, II and III show solid solution, azeotrope and eutectic phase diagrams, respectively. We then investigate the effects of molecular size and intermolecular attractions on the solid-liquid phase diagrams of binary Lennard-Jones mixtures using our three model mixtures.