Final Report: Design and Synthesis of CO2-Soluble Affinity Ligands for Use in CO2 Extraction of Proteins

EPA Grant Number: R824730
Title: Design and Synthesis of CO2-Soluble Affinity Ligands for Use in CO2 Extraction of Proteins
Investigators: Beckman, Eric J. , Russell, Alan J.
Institution: University of Pittsburgh
EPA Project Officer: Hahn, Intaek
Project Period: October 1, 1995 through September 30, 1998
Project Amount: $150,000
RFA: Technology for a Sustainable Environment (1995) Recipients Lists
Research Category: Nanotechnology , Sustainability , Pollution Prevention/Sustainable Development


Our ultimate objective was to design highly CO2-soluble surfactants and affinity ligands that would allow us to extract proteins from aqueous solution into CO2. The protein would be recovered without substantial activity loss via depressurization.

Summary/Accomplishments (Outputs/Outcomes):

We have succeeded in designing a series of surfactants and affinity ligands that exhibit high solubility in liquid CO2 at room temperature. These surfactants were employed to extract subtilisin Carlsberg from aqueous solution; conditions were identified where the protein could be extracted and recovered with retention of 80 percent of the original activity. In addition, we showed that use of an amphiphilic affinity ligand (here biotin-functional) allowed extraction of nearly 80 percent of the avidin from aqueous solution using very low ratios of ligand: protein (less than 10). Finally, we have extended this work to separation of enantiomers in CO2 using CO2-soluble chiral mobile phase agents, and also to formation of ligand protein complexes that are CO2-soluble, thus allowing homogeneous enzymatic chemistry in CO2.

Carbon dioxide has elicited significant interest between the academic and industrial community over the previous decade in that it exhibits properties that render it a relatively "green" solvent. CO2 is naturally abundant, is nonflammable and inexpensive, and exhibits relatively low toxicity compared with other organic solvents. Despite CO2's advantageous properties, it is still a low dielectric material, and thus incapable of solubilizing significant quantities of polar or ionic compounds. This was not deemed to be a serious problem initially, as typical organic liquids (alkanes, etc.) exhibit the same characteristics, yet these could be employed to extract polar solutes via the addition of affinity ligands such as surfactants and chelating agents. Unfortunately, early attempts to employ such affinity ligands in CO2-based extraction were stymied by the apparent very poor solubility of the ligands themselves, something that was not expected. However, while CO2 is not a particularly strong solvent, it can solubilize large quantities of material from selected classes of compounds, including certain fluorinated compounds and silicones. Beginning in 1990, we have employed such CO2-philic materials in the design of affinity ligands that are both highly CO2-soluble and can bind strongly to various target solutes, allowing extraction into CO2 of metals, proteins, and other hydrophilic materials. We have shown that use of highly CO2-soluble ligands allows one to employ CO2 in extractions of polar materials from water that were previously thought to be untenable. For the case of proteins, for example, we have found that one can solubilize proteins in CO2 using a CO2-soluble affinity ligand without the need to also transport significant quantities of water. Because one can solubilize a protein in CO2 using these ligands, one can now consider homogeneous enzymatic catalysis in CO2.

Protein Extraction into CO2

Results From Prior Years: Our initial work focused on the extraction of the protein subtilisin Carlsberg (a 27, 000 MW protease) from buffer to CO2, with recovery of the protein achieved via slow depressurization. The CO2-soluble surfactants comprised somewhat typical polar head groups, here sulfonates (anionic) and ethylene oxide oligomers (nonionic), while the CO2-philic tails were generated from perfluorinated polyethers purchased from DuPont and Ausimont. As expected, use of fluoroether rather than typical alkyl groups permitted greater than millimolar solubility of the materials in CO2 at pressures below 2,000 psi. In initial experiments, the CO2-soluble surfactants were employed to set up inverse emulsions in CO2, which were then used to extract protein from the aqueous phase. Results (Ghenciu, et al., 1998) showed that protein could indeed be transferred from the aqueous to the CO2 phase, and that the material recovered from CO2 by depressurization retained over 80 percent of its initial activity. However, one downside to these results was that a significant amount of water also was solubilized by the inverse emulsion, and thus it was difficult to concentrate the protein using this technique. Fortunately, we have made observations that eliminated this problem entirely.

We had observed early on that the presence of protein in the aqueous phase produced a significant effect on the phase behavior of the emulsion system, where the protein appeared to help stabilize the emulsion. This suggested the presence of protein-surfactant interactions, interactions that we might be able to use to aid in protein extraction. Subsequently, we employed an anionic surfactant (which should interact strongly with the positively charged [at pHs below the IP] subtilisin) in an attempt to induce the protein to migrate into CO2 without an inverse emulsion present. In our case, we added the anionic fluoroether to the system under conditions where a middle phase emulsion formed, and then used pure CO2 to extract the middle phase. Indeed, we were able to extract nearly 40 percent of the protein from the aqueous phase using this technique, while transporting almost no water.

To investigate this effect further, we prepared a fluoroether-functional affinity ligand for avidin (which has four binding sites for biotin), namely a fluoroether-PEG-biotin material (Ghenciu and Beckman, 1997). As before, we added this ligand to a CO2-buffer biphasic mixture under conditions where a middle phase emulsion forms, and then pumped pure CO2 through the system to see if we could solubilize the avidin-ligand adducts. Indeed, at ligand-to-protein ratios less than 20, we could extract nearly 80 percent of the avidin present in the buffer, without the transport of any significant amounts of water. We also found that extractions using a ligand without the PEG spacer are only half as effective, possibly because the presence of the PEG induces greater surface activity in the ligand. Extractions performed with a ligand without the biotin produce essentially no transport of protein into the CO2.

Results From 1998: The work on protein extraction was successful to the point that Genencor International, our industrial partner throughout the work, decided to file for both domestic and foreign patent protection, and to set up facilities in their own laboratory to further explore the area. We decided to explore other applications for the use of CO2-philic amphiphiles in separations of compounds of interest to the biotechnology industry. Two areas were examined: (1) the use of CO2-soluble amphiphiles to allow homogeneous enzymatic reactions in CO2; and (2) the separation of enantiomers using CO2 and CO2-soluble resolving agents.

In the first area, we are endeavoring to modify both an oxoreductase and NAD(H) to allow homogenous reaction in CO2. There are a number of situations where an oxoreductase is employed to convert a hydrophobic material to a hydrophilic product, and thus conducting such a reaction in a biphasic CO2/water mixture would allow for continuous processing and collection or product without producing a contaminated aqueous stream (as would be the case when a conventional organic solvent is used). However, neither the enzyme nor the NAD(H) cofactor is soluble in CO2, and thus would obviously partition to the aqueous phase, leaving the substrate stranded in the organic. Consequently, we have devised a strategy by which both the enzyme and the cofactor can be dissolved/dispersed in CO2. First, NAD(H) was derivitized via reaction with (1) bromine, followed by (2) a short chain aliphatic diamine. The pendant amine group is then reacted with a functional CO2-phile, leading to the CO2-soluble NAD(H). With regard to the enzyme, it has been shown that proteins will form complexes with certain sugar-terminal surfactants, allowing them (the enzymes) to be dispersed in an organic solvent. Consequently, we have prepared a fluorinated 9, and thus CO2-philic) version of the surfactant. Here, a fluorinated alcohol is reacted with N-protected glutamic acid to form the di-ester. Following deprotection, the amino diester is then reacted with a sugar-lactone, forming the surfactant. Future work will explore the solubility of the derivitized NAD(H) in CO2, as well as that of the enzyme-surfactant complex. Finally, we propose to examine the rate of a model reaction in CO2 using this system.

In addition to this work on use of enzymes in CO2, we also explored the design of highly CO2-soluble resolving agents to be used to separate the enantiomers of ibuprofen. Typically, ibuprofen is produced as a racemic mixture, after which it is separated by crystallization using an appropriate resolving agent. The extent of the separation is based on the kinetics of precipitation of each diastereomeric salt. We have explored the extension of the use of resolving agents to CO2 where ibuprofen is used as the model racemic mixture. We generated a series of CO2-soluble resolving agents from both lysine and quinine. The type of CO2-philic tail as well as the location of the tail on either lysine or quinine was varied, and the extent to which the derivitized agents bind to each enantiomer of ibuprofen was measured in CO2. At this point, we attempted to separate the enantiomers of ibuprofen using pressure alone, as well as a mixture of agents and pressure. Enantiomeric excesses near 60 percent were achieved with a pressure-based separation using a fluorinated lysine-resolving agent.

Future Activities:

This is the final report for the grant, which was completed in September 1998. In the future, we will continue working on designing systems to allow homogeneous enzymatic chemistry in CO2.

Journal Articles on this Report : 2 Displayed | Download in RIS Format

Other project views: All 11 publications 3 publications in selected types All 2 journal articles
Type Citation Project Document Sources
Journal Article Ghenciu EG, Beckman EJ. Affinity extraction into carbon dioxide .1. Extraction of avidin using a biotin-functional fluoroether surfactant. Industrial & Engineering Chemistry Research 1997;36(12):5366-5370 R824730 (Final)
not available
Journal Article Ghenciu EG, Russell AJ, Beckman EJ, Steele L, Becker NT. Solubilization of subtilisin in CO2 using fluoroether-functional amphiphiles. Biotechnology and Bioengineering 1998;58(6):572-580. R824730 (Final)
not available

Supplemental Keywords:

carbon-dioxide-soluble affinity ligands, biological extraction, separation of enantiomer, racemic target molecules, chiral mobile phase ligands, carbon-dioxide-philic tails, fluoroalkyl, fluoroether., RFA, Scientific Discipline, Sustainable Industry/Business, cleaner production/pollution prevention, Environmental Chemistry, Sustainable Environment, Technology for Sustainable Environment, Ecological Risk Assessment, bilogical extraction, cleaner production, environmentally benign solvents, depressurization, carbon dioxide extraction, proteins, ligands, chiral mobile phase agents

Progress and Final Reports:

Original Abstract
  • 1996
  • 1997