Production and characterization of microbial biosurfactants for potential use in oil-spill remediation.

Two biosurfactants, surfactin and fatty acyl-glutamate, were produced from genetically-modified strains of Bacillus subtilis on 2% glucose and mineral salts media in shake-flasks and bioreactors. Biosurfactant synthesis ceased when the main carbohydrate source was completely depleted. Surfactin titers were ∼30-fold higher than fatty acyl-glutamate in the same medium. When bacteria were grown in large aerated bioreactors, biosurfactants mostly partitioned to the foam fraction, which was recovered. Dispersion effectiveness of surfactin and fatty acyl-glutamate was evaluated by measuring the critical micelle concentration (CMC) and dispersant-to-oil ratio (DOR). The CMC values for surfactin and fatty acyl-glutamate in double deionized distilled water were 0.015 and 0.10 g/L, respectively. However, CMC values were higher, 0.02 and 0.4 g/L for surfactin and fatty acyl-glutamate, respectively, in 12 parts per thousand Instant Ocean®[corrected].sea salt, which has been partly attributed to saline-induced conformational changes in the solvated ionic species of the biosurfactants. The DORs for surfactin and fatty acyl-glutamate were 1:96 and 1:12, respectively, in water. In Instant Ocean® solutions containing 12 ppt sea salt, these decreased to 1:30 and 1:4, respectively, suggesting reduction in oil dispersing efficiency of both surfactants in saline. Surfactant toxicities were assessed using the Gulf killifish, Fundulus grandis, which is common in estuarine habitats of the Gulf of Mexico. Surfactin was 10-fold more toxic than fatty acyl-glutamate. A commercial surfactant, sodium laurel sulfate, had intermediate toxicity. Raising the salinity from 5 to 25 ppt increased the toxicity of all three surfactants; however, the increase was the lowest for fatty acyl-glutamate.

Mechanisms of aqueous extraction of soybean oil.

Aqueous extraction processing (AEP) of soy is a promising green alternative to hexane extraction processing. To improve AEP oil yields, experiments were conducted to probe the mechanisms of oil release. Microscopy of extruded soy before and after extraction with and without protease indicated that unextracted oil is sequestered in an insoluble matrix of denatured protein and is released by proteolytic digestion of this matrix. In flour from flake, unextracted oil is contained as intact oil bodies in undisrupted cells, or as coalesced oil droplets too large to pass out of the disrupted cellular matrix. Our results suggest that emulsification is an important extraction mechanism that reduces the size of these droplets and increases yield. Protease and SDS were both successful in increasing extraction yields. We propose that this is because they disrupt a viscoelastic protein film at the droplet interface, facilitating droplet disruption. An extraction model based on oil droplet coalescence and the formation of a viscoelastic film was able to fit kinetic extraction data well.

Genetic engineering strategies for purification of recombinant proteins from canola by anion exchange chromatography: an example of beta-glucuronidase.

The elution behavior of native canola proteins from different anion-exchange resins was determined. The elution profiles showed the potential for simplified recovery of acidic recombinant proteins from canola. When Q-sepharose fast flow was used, there were three optimal salt elution points at which a recombinant protein would have minimal contamination with native proteins. The feasibility of exploiting this advantage was examined for recovery of the acidic protein beta-glucuronidase (GUS/GUSD0 from the Escherichia coli gene) along with three polyaspartate fusions to the wild-type GUS. The fusions contained 5 (GUSD5), 10 (GUSD10), or 15 (GUSD15) aspartic acids fused to the C-terminus and were chosen to extend the elution time. The three fusions and the wild-type enzyme were produced in E. coli, purified, and added to canola extracts before chromatography. The equivalence of this spiking experiment to that of extracting a recombinant protein from transgenic canola was determined in a control experiment using transgenic canola expressing the wild-type enzyme. Behavior in the transgenic and spiked experiments was equivalent. GUSD0 eluted at the earliest optimal elution point; the addition of polyaspartate tails resulted in longer retention times and better selective recovery. If one assumes binding through a single fusion (the protein is a tetramer), there is a nearly linear shift in elution within the salt gradient of 17 mM per added charge up to 10, with a reduced increment from 10 to 15. The fusions and their enzymatic activity proved very stable in the canola extracts through 7 days in cold storage, providing flexibility in process scheduling.

Suitability of immobilized metal affinity chromatography for protein purification from canola.

This work demonstrates that proper selection of a metal ion and chelating ligand enables recovery of a his(6)-tagged protein from canola (Brassica napus) extracts by immobilized metal affinity chromatography (IMAC). When using Co(2+) with iminodiacetate (IDA) as the chelating ligand, beta-glucuronidase-his(6) (GUSH6) can be purified from canola protein extract with almost homogeneous purity in a single chromatographic step. The discrimination with which metal ions bound native canola proteins followed the order Cu(2+) < Ni(2+) < Zn(2+) < Co(2+) in regard to elimination of proteins coeluted with the fusion protein. IDA- and nitrilotriacetate (NTA)-immobilized metal ions showed different binding patterns, whose cause is attributed to a more rigid binding orientation of the his(6) in forming a tridentate with Me(2+)-IDA than in forming a bidentate with Me(2+)-NTA. The more flexible binding allows for multisite interactions over the protein.

Propionic acid production by extractive fermentation. I. Solvent considerations.

Solvent selection for extractive fermentation for propionic acid was conducted with three systems: Alamine 304-1 (trilaurylamine) in 2-octanol, 1-dodecanol, and Witcohol 85 NF (oleyl alcohol). Among them, the solvent containing 2-octanol exhibited the highest partition coefficient in acid extraction, but it was also toxic to propionibacteria. The most solvent-resistant strain among five strains of the microorganism was selected. Solvent toxicity was eliminated via two strategies: entrapment of dissolved toxic solvent in the culture growth medium with vegetable oils such as corn, olive, or soybean oils; or replacement of the toxic 2-octanol with nontoxic Witcohol 85 NF. The complete recovery of acids from the Alamine 304-1/Witcohol 85 NF was also realized with vacuum distillation.

Contribution of protein charge to partitioning in aqueous two-phase systems.

Protein partitioning in aqueous two-phase systems based on phase-forming polymers is strongly affected by the net charge of the protein, but a thermodynamic description of the charge effects has been hindered by conflicting results. Many of the difficulties could be because of problems in isolating electrochemical effects from other interactions of phase components. We explored charge effects on protein partitioning in poly(ethylene glycol)-dextran two-phase systems by using two series of genetically engineered charge modifications of bacteriophage T4 lysozyme produced in Escherichia coli. The two series, one in the form of charged-fusion tails and the other in the form of charge-change point mutations, provided matching net charges but very different polarity. Partition coefficients of both series were obtained and interfacial potential differences of the phase systems were measured. Multi-angle laser light scattering measurements were also performed to determine second virial coefficients. A semi-empirical model accounting for the roles of both charge and non-charge effects on protein partitioning behavior is proposed, and the results predicted from the model are compared to the results from the experiments.

Fed-batch fermentation with and without on-line extraction for propionic and acetic acid production by Propionibacterium acidipropionici.

Fed-batch propionic and acetic acid fermentations were performed in semi-defined laboratory medium and in corn steep liquor with Propionibacterium acidipropionici strain P9. On average, over four experiments, 34.5 milligrams propionic acid and 12.8 milligrams acetic acid were obtained in about 146 h in laboratory medium with 79 milligrams glucose added over five feeding periods. The highest concentration of propionic acid, 45 milligrams, was obtained when the glucose concentration was not allowed to drop to zero. In corn steep liquor 35 milligrams propionic acid and 11 milligrams acetic acid were produced in 108 h from 59.4 milligrams total lactic acid provided as seven feedings of corn steep liquor. Extractive fed-batch fermentations were conducted in semi-defined medium using either flat-sheet-supported liquid membranes or hollow-fiber membrane extraction to remove organic acids from the culture medium. As operated during the course of the fermentation, these systems extracted 25% and 22% of the acetic acid and 36.5% and 44.5% of the propionic acid, respectively, produced in the fermentation. Total amounts of acids produced were about the same as in comparable nonextractive fermentations: 30-37 milligrams propionic acid and 13 milligrams acetic acid were produced in 150 h. Limitations on acid production can be attributed to limited substrate feed, not to failure of the extraction system.

Water reuse in the L-lysine fermentation process.

L-Lysine is produced commercially by fermentation. As is typical for fermentation processes, a large amount of liquid waste is generated. To minimize the waste, which is mostly the broth effluent from the cation exchange column used for l-lysine recovery, we investigated a strategy of recycling a large fraction of this broth effluent to the subsequent fermentation. This was done on a labscale process with Corynebacterium glutamicum ATCC 21253 as the l-lysine-producing organism. Broth effluent from a fermentation in a defined medium was able to replace 75% of the water for the subsequent batch; this recycle ratio was maintained for three sequential batches without affecting cell mass and l-lysine production. Broth effluent was recycled at 50% recycle ratio in a fermentation in a complex medium containing beet molasses. The first recycle batch had an 8% lower final l-lysine level, but 8% higher maximum cell mass. In addition to reducing the volume of liquid waste, this recycle strategy has the additional advantage of utilizing the ammonium desorbed from the ion-exchange column as a nitrogen source in the recycle fermentation. The major problem of recycling the effluent from the complex medium was in the cation-exchange operation, where column capacity was 17% lower for the recycle batch. The loss of column capacity probably results from the buildup of cations competing with l-lysine for binding. (c) 1996 John Wiley & Sons, Inc.

Extraction of charged fusion proteins in reversed micelles: comparison between different surfactant systems.

Behavior of a series of fusion proteins of varying charge in reversed micellar extraction was studied. The proteins consisted of the enzyme glucoamylase from Aspergillus awamori joined to short peptides containing from 0-10 additional aspartate residues. The fusions were partitioned into two different cationic surfactant systems, one based on the surfactant trioctylmethylammonium chloride (TOMAC) and the other on cetyltrimethylammonium bromide (CTAB). These two systems differed chiefly in micelle size as measured by the water to surfactant ratio, wo. Water numbers were determined for the TOMAC system, with values of approximately 10, and as a function of pH and ionic strength for CTAB for each of the mutant enzymes. For the CTAB system, water numbers were as low as 50 with NaCl concentrations of 500 mM and as high as 68 at 300 mM NaCl (95% confidence level of 2.4). The enzyme partitioned most strongly using CTAB, with maximal recoveries approaching 95%. However, in the CTAB system, there were no significant differences in behavior between the mutants because of the relatively large micellar size, even under high salt concentrations. Extraction of the control enzyme from clarified cell broth indicated that broth components did not significantly interfere with partitioning.

Genetically engineered charge modifications to enhance protein separation in aqueous two-phase systems: Charge directed partitioning.

This report continues or examination of the effect of genetically engineered charge modifications on the partitioning behavior of proteins in aqueous two-phase extration. The genetic modifications consisted of the fusion of charged peptide tails to beta-galactosidase and charge-change point mutations to T4 lysozyme. Our previous article examined the influence of these charge modifications on partitioning as a function of interfacial potential difference. In this study, we examined charge directed partitioning behavior in PEG/dextran systems containing small amounts of the charged polymers diethylaminoethyl-dextran (DEAE-dextran) or dextran sulfate. The best results were obtained when attractive forces between the protein and polymer were present. Nearly 100% of the beta-galactosidase, which carries a net negative charge, partitioned to the DEAE-dextran-rich phase regardless of whether the phase was dextran or PEG. In these cases, cloudiness of the protein-rich phases suggest that strong charge interactions resulted in protein/polymer aggregation, which may have contributed to the extreme partitioning. Unlike the potentialdriven partitioning reported previously, consistent partitioning trends were observed as a result of the fusion tails, with observed shifts in partition coefficient (K(p)) of up to 37-fold. However, these changes could not be solely attributed to charge-based interactions. Similarly, T4 lysozyme, carrying a net positive charge, partitioned to the dextran sulfate-containing phase, and displayed four- to sevenfold shifts in K(p) as a result of the point mutations. These shifts were two to four times stronger than those observed for potential driven partitioning. Little effect on partitioning was observed when the protein and polymer had the same charge, with the exception of beta-galactosidase with polyarginine tails. The high positive charge density of these tails provided for a localized interaction with the dextran sulfate, and resulted in 2- to 15-fold shifts in K(p). (c) 1995 John Wiley & Sons, Inc.

Broth recycle in a yeast fermentation.

Fermentation is a water-intensive process requiring treatment of large amounts of effluent broth. It is desirable to increase the ratio of product produced to the volume of effluent by minimizing the discharge of effluent from the fermentation process. A study of recycling spent fermentation process. A study of recycling spent fermentation broth for the subsequent fermentation was carried out with Apiotrichum curvatum an oleaginous yeast, as the working culture. Spent broth from a defined medium was recycled t replace as much as 75% of the water and salts for subsequent batches and this was repeated for seven sequential batches without affecting cell mass and lipid production. A 64% vlume reduction of wastewater was achieved in this manner. However, when using whey permeate as the medium, lipid production dropped after three consecutive recycle operations at 50% recycle, and after two consecutive recycle operations at 75% and 100% recycle. Accumulation of ions in the broth appeared to be responsible for the inhibition. An ion exchange step was able to eliminate the ion buildup and restore fermentation performance. (c) 1994 John Wiley & Sons, Inc.

Reversed micellar extraction of charged fusion proteins.

We have investigated the use of charged fusion tails with the enzyme glucoamylase in reversed micellar extraction. The addition of the charged tails increased the fraction of enzymatically active protein recovered at a given pH, with the tails containing the largest number of charges being recovered at the highest level. The series of mutations also allows for investigation of the charge-dependent behavior of reversed micellar extraction. However, in this case, the change in protein charge via fusions had a lesser impact than did the change in charge via a pH change. The difference may be due to the difficulty of partitioning the hydrodynamically larger fusion protein.

Ion exchange immobilization of changed beta-galactosidase fusions for lactose hydrolysis.

The use of charged peptides fused to enzymes for immobilization onto ion-exchange membranes was explored for the enzyme x-galactosidase. The additional charged peptides, containing 1, 5, 11, and 16 aspartates, fused to x-galactosidase, for the most part did not interfere with the kinetic behavior for lactose hydrolysis. There was a 2-fold decline in V(m) for the 16-aspartate fusion, but the others were quite similar to the wild type enzyme (BGWT). BGWT and the fusions all retained approximately 50% of their activities when adsorbed onto ion-exchange membranes. In contrast to BGWT, the enhanced binding strength of the 11 aspartate fusion provided the ability to hydrolyze whey permeate at 0.3 M ionic strength without enzyme leakage, and to immobilize the enzyme directly from diluted cell extract with 83% purity.

Genetically engineered charge modifications to enhance protein separation in aqueous two-phase systems: Electrochemical partitioning.

We have examined the effect of genetically engineered charge modifications on the partitioning behavior of proteins in dextran/polyethylene glycol two-phase systems containing potassium phosphate. By genetically altering a protein's charge, the role of charge on partitioning can be assessed directly without the need to modify the phase system. The charge modifications used are of two types: Charged tails of polyaspartic acid fused to beta-galactosidase and charge-change point mutations of T4 lysozyme which replace positive lysine residues with negative glutamic acids. The partition coefficient K(p) for these proteins was related to measured interfacial potential differences Deltaphi using the simple thermodynamic model, In K(p) = In K(o) + (F/RT)Z(p) deltaphi. The protein net charge Z(p) was determined using the Henderson-Hasselbalch relationship with modifications based on experimentally determined titration and isoelectric point data. It was found that when the electropartitioning term Z(p) deltaphi was varied by changing the pH, the partitioning of T4 lysozyme was quantitatively described by the thermodynamic model. The beta-galactosidase fusions displayed qualitative agreement, and although less than predicted, the partitioning increased more than two orders of magnitude for the pH range examined. Changes in the partitioning of lysozyme due to the various mutations agreed qualitatively with the thermodynamic model, but with a smaller than expected dependence on the estimated charge differences. The beta-galactosidase fusions, on the other hand, did not display a consistent charge based trend, which is likely due either to the enzyme's large size and complexity or to nonelectrostatic contributions from the tails. The lack of quantitative fit with the model described above suggests that the assumptions made in developing this model are oversimplified. (c) 1994 John Wiley & Sons, Inc.

Characterization and polyelectrolyte precipitation of beta-galactosidase containing genetic fusions of charged polypeptides.

Genetically engineered versions of beta-galactosidase were constructed through the addition of charged polypeptide fusion tails for the purpose of enhancing polyelectrolyte precipitation. Negatively charged aspartic acid tails and positively charged poly(arginine) tails were added to beta-galactosidase from Escherichia coli. These fusion proteins were all shown to possess specific activity equal to that of the native enzyme. Gel permeation and ion-exchange chromatography provided evidence concerning the integrity of the tails as well as their altered charge characteristics. All enzymes containing charged tails displayed enhanced polyelectrolyte precipitation over the native enzyme. An optimal number of charged residues, beyond which no further enhancement of precipitation was observed, was found to be approximately 10 residues for each type of tail. No interference from nucleic acids was observed in the precipitation of positively tailed beta-galactosidase.

Charged fusions for selective recovery of beta-galactosidase from cell extract using hollow fiber ion-exchange membrane adsorption.

We explored the use of charged fusions for selective recovery of beta-galactosidase from cell extract using a low-cost, easily scaled, fast, charge-based separation technique-ion exchange on hollow fiber ion-exchange membranes (HFIEMs). The additional charges carried by a series of anionic fusion tails allowed selective binding and release of beta-galactosidase from Escherichia coli cell extract using the HFIEM cartridge. The purification factors increased with fusion length. The beta-galactosidase was recovered in active form. For the longest fusion studied, more than sixfold enrichment in specific activity was attained. The specific activity of the recovered fraction is comparable with that of commercial wild-type beta-galactosidase and affinity-purified fusion protein.

Fusion tails for the recovery and purification of recombinant proteins.

Several fusion tail systems have been developed to promote efficient recovery and purification of recombinant proteins from crude cell extracts or culture media. In these systems, a target protein is genetically engineered to contain a C- or N-terminal polypeptide tail, which provides the biochemical basis for specificity in recovery and purification. Tails with a variety of characteristics have been used: (1) entire enzymes with affinity for immobilized substrates or inhibitors; (2) peptide-binding proteins with affinity to immunoglobulin G or albumin; (3) carbohydrate-binding proteins or domains; (4) a biotin-binding domain for in vivo biotination promoting affinity of the fusion protein to avidin or streptavidin; (5) antigenic epitopes with affinity to immobilized monoclonal antibodies; (6) charged amino acids for use in charge-based recovery methods; (7) poly(His) residues for recovery by immobilized metal affinity chromatography; and (8) other poly(amino acid)s, with binding specificities based on properties of the amino acid side chain. Fusion tails are useful at the lab scale and have potential for enhancing recovery using economical recovery methods that are easily scaled up for industrial downstream processing. Fusion tails can be used to promote secretion of target proteins and can also provide useful assay tags based on enzymatic activity or antibody binding. Many fusion tails do not interfere with the biological activity of the target protein and in some cases have been shown to stabilize it. Nevertheless, for the purification of authentic proteins a site for specific cleavage is often included, allowing removal of the tail after recovery.

Recovery of a charged-fusion protein from cell extracts by polyelectrolyte precipitation.

Beta-galactosidase served as a model system to explore the feasibility of enhancing the selectivity of a low-cost, easily scaled separation method-precipitation. Enhanced selectivity was sought by fusing the enzyme with polypeptide tails including 5 and 11 aspartates. The unfused protein could not be selectively removed from the Escherichia coli cell extract by precipitation with polyethylenimine (PEI), but the longest fusion could be selectively removed. The presence of nucleic acids limited the purification attainable. Pretreatment with nuclease followed by diafiltration resulted in an extract from which the same fusion could be precipitated with greater than fivefold enrichment, while the untailed enzyme remained unenriched by the same precipitation step. Selectivity is attributed to the binding strength of the polyanionic tails to the polycationic PEI.

Precipitation of nucleic acids with poly(ethyleneimine).

Removal of nucleic acids from cell extracts is a common early step in downstream processing for protein recovery. We report on the precipitation of nucleic acids from a homogenate of Saccharomyces cerevisiae by addition of the cationic polyelectrolyte poly(ethyleneimine) (PEI), focusing on the effect of PEI dosage on particle size, protein loss, and extent of nucleic acid removal in both batch and continuous mode. Better than 95% removal of nucleic acids from yeast homogenates was achieved by means of precipitation with PEI with protein losses of approximately 15% with or without previous removal of cell debris. The coprecipitated protein is predominately large molecular weight material and exhibits both low and high isoelectric points. Such treatment does not aggregate the cell debris; size distribution of the precipitated particles from a continuous precipitator is very similar to that for protein precipitation.

Polyelectrolyte precipitation of beta-galactosidase fusions containing poly-aspartic acid tails.

Protein recovery from industrial microbial processes can be very expensive, often exceeding the cost of protein production. We have genetically engineered 3 beta-galactosidase (beta-gal) fusion proteins containing poly-aspartic acid tails to test the effect of the tails on recovery by the relatively inexpensive method of polyelectrolyte precipitation. The fusion proteins, designated T1, T2, and T3, were constructed with C-terminal tails of 5, 11, and 16 aspartic acid residues, respectively. The fusion proteins were expressed in Escherichia coli, and purified by affinity chromatography. T1 and T2 had specific activities similar to that of wildtype beta-gal, whereas the specific activity of T3 was about half that of T1 and T2. The increased net charge of the fusion proteins compared to wildtype beta-gal was indicated both by ion-exchange chromatography and their migration pattern in non-denaturing polyacrylamide gel electrophoresis. All three tails enhanced polyethyleneimine (PEI) precipitation of the fusion proteins compared to wildtype beta-gal. At a low PEI/protein ratio (0.01, g g-1), recovery by precipitation of T2 and T3 was more than 2 X that of the beta-gal control, whereas that of T1 was only slightly greater than that of the control. At a higher PEI/protein ratio (0.03, g g-1) the amount of precipitation of all three fusion proteins was nearly the same, about 1.5 X that of the control.