Partitioning Behavior Of Proteins Research Articles

What Can Be Learned from the Partitioning Behavior of Proteins in Aqueous Two-Phase Systems?

This review covers the analytical applications of protein partitioning in aqueous two-phase systems (ATPSs). We review the advancements in the analytical application of protein partitioning in ATPSs that have been achieved over the last two decades. Multiple examples of different applications, such as the quality control of recombinant proteins, analysis of protein misfolding, characterization of structural changes as small as a single-point mutation, conformational changes upon binding of different ligands, detection of protein-protein interactions, and analysis of structurally different isoforms of a protein are presented. The new approach to discovering new drugs for a known target (e.g., a receptor) is described when one or more previous drugs are already available with well-characterized biological efficacy profiles.

Environmentally Benign and Recyclable Aqueous Two-Phase System Composed of Distillable CO2-Based Alkyl Carbamate Ionic Liquids

Ionic liquid-based aqueous two-phase systems (IL-ATPSs) have been studied comprehensively as a potential alternative method for protein purification. One of the distinct features of IL-ATPS is that the polarity of the IL can be adjusted by different combinations of cation and anion, providing flexibility in governing the partition of proteins between phases. However, the conventional IL-ATPS usually has poor environmental footprint and low recyclability, thereby hampering the systems for practical use in protein separation. To address these shortcomings, here we explored the distillable CO2-based alkyl carbamate ILs as phase-forming components. N,N-dimethylammonium N′,N′-dimethylcarbamate (DIMCARB) was able to form ATPS with polypropylene glycol (PPG). The liquid–liquid equilibrium data of ATPSs composed of PPG with different molecular mass and DIMCARB were determined. As both phase-forming components are thermosensitive, the effect of temperature on the phase diagram of the PPG 400 + DIMCARB system was also evaluated. The partition behavior of model proteins (i.e., bovine serum albumin and lysozyme) in the PPG + DIMCARB systems were assessed. Both model proteins tend to migrate to the more hydrophilic IL-rich phase, when a higher temperature or a higher molecular mass of PPG was used. The DIMCARB was recovered via rotary evaporation, while the PPG 400 was recovered from a secondary two-phase system thermally induced from the primary ATPS. The recovered components were subsequently used to prepare a new batch of ATPS. The phase composition of the new system as well as the partition coefficients of both model proteins in the recycled system were maintained, thereby proving the feasibility of reusing both phase-forming components. Overall, the recyclable and biodegradable phase-forming components used in the newly developed PPG + DIMCARB system make the ATPS a sustainable and eco-friendly system for protein purification.

Effect of salt-based adjuvant on partition behaviour of protein in aqueous two-phase systems composed of polypropylene glycol and cholinium glycinate

Abstract Aqueous two-phase systems (ATPSs) composed of ionic liquids (ILs) have been widely explored over the past decade for application in protein purification. The interaction between the phase components of ATPSs and the target protein is one of the factors governing the partition behaviour of the protein. The addition of adjuvants like inorganic salts at a low concentration in an ATPS could manipulate the partitioning of the protein between phases. In this study, the partitioning behaviour of model proteins in ATPSs composed of cholinium glycinate, poly(propylene) glycol 400 and a salt-based adjuvant (i.e., sodium chloride, magnesium chloride and sodium sulfate) was investigated systematically. The dissociated ions with a greater hydration capacity preferentially partitioned to the phase having a higher water content. Also, the partition coefficient of the dissociated ions was independent of the concentration of salt adjuvant. The affinity of adjuvant ions for the IL-rich bottom phase was in the decreasing order of: sulfate ion > chloride ion > magnesium ion > sodium ion. This non-uniform distribution of cations and anions of ionic adjuvants in the ATPS had an influence of the partitioning of proteins in ATPS. Bovine serum albumin and lysozyme, which differ in isoelectric point and surface charge density, were used as the model proteins. The partitioning behaviour of the proteins was affected by the distribution of the dissociated ions of the ionic adjuvant in both phases of the ATPS. The electrostatic interaction between proteins and the dissociated ions of the ionic adjuvant was governed by factors such as types and charge density of the dissociated ions of the salt-based adjuvants, isoelectric point of the protein, and pH of the ATPS. The results proved the feasibility of manipulating the partitioning of target proteins in ATPSs by using inorganic salts as the adjuvant, which provides a means to improve the protein separation in IL-based ATPSs.

How to manipulate partition behavior of proteins in aqueous two-phase systems: Effect of trimethylamine N-oxide (TMAO)

It is shown that the partition behavior of solutes in a given aqueous two-phase systems can be manipulated using a nonionic additive, the trimethylamine N-oxide (TMAO). To this end, partitioning of a set of proteins and small organic compounds was analyzed in aqueous Dextran-75-Polyethylene glycol (PEG-600)-0.15 M sodium sulfate-0.01 M sodium/potassium phosphate buffer, pH 7.4 two-phase systems with TMAO additive at different concentrations from 0 up to 1.95 M. Partition behavior of several proteins was also examined in aqueous Ficoll-70-Polyethylene glycol (PEG-8000)-0.01 M sodium/potassium phosphate buffer, pH 7.4 two-phase systems with TMAO additive at different concentrations in a range from 0 up to 1.5 M. It has been found that protein partitioning changes with the increasing TMAO concentration in the protein specific manner. The differences between the solvent properties of water in the phases of the aqueous two-phase systems, such as hydrophobic character, electrostatic properties, solvent dipolarity/polarizability, hydrogen bonding donor acidity, and hydrogen bonding acceptor basicity were characterized with (a) analysis of partitioning of homologous series of sodium salts of dinitrophenylated amino acids with aliphatic alkyl side-chains, and (b) solvatochromic comparison method. The results obtained show that TMAO affects partitioning of solutes via its effect on the solvent properties of the phases, in agreement with the view that partition behavior of proteins and small organic compounds is governed by the solute-water interactions in the coexisting phases.

An affinity-based aqueous two-phase mixed micellar system and its purification of Yeast 3',5'-bisphosphate nucleotidase.

Aqueous two-phase micellar system (ATPMS), as an alternative liquid-liquid extraction technique, has been extensively exploited for the precise separation or large-scale concentration of biomolecules. In this article, a novel affinity-based ATPMS composed of mixed micelles was constructed by introducing a Copper-chelated Triton X-114 (TX-Cu(II)) into an aqueous solution of hydrophobically modified ethylene oxide polymer (HM-EO). Phase diagram of the HM-EO/TX-Cu(II) system was measured, and the partitioning behavior of model proteins (YND, BSA, lysozyme) were studied by using this new system. The addition of HM-EO can result in formation of the micellar network in the micelle-rich phase, making the phase separation easier and stabler. In addition, the extractive performance of ATPMS was enhanced due to the existence of the mixed micelles composed by HM-EO and Cu(II)-chelated TX. It was found in the partitioning experiments that the hexahistidine-tagged Yeast 3',5'-bisphosphate nucleotidase (YND) was selectively extracted into the micelle-rich phase, while the histidine-poor proteins (BSA and lysozyme) remained in the micelle-poor phase. Finally, HM-EO/TX-Cu(II) was used directly to process the fermentation broth. The target protein, YND could be recovered from the cell lysate with a recovery yield of 49.23% and purification factor of 2.63. The results indicated that the new affinity-based HM-EO/TX-Cu(II) system had high partitioning performance which is promising for effectively separation of the histidine-tagged proteins.

Predicting protein partition coefficients in aqueous two phase system

The present work aims to achieve an additional insight into the protein partitioning behavior in aqueous two phase systems (ATPSs), together with a study on the viability of a semi-empirical model based on continuum electrostatics to predict the protein partition characteristics. The partitioning behaviors of 14 globular proteins, with different properties, were explored in three polymer/polymer ATPSs. By the Collander equation, a linear correlation between protein partitioning coefficients in all systems was observed. Using the semi-empirical model it was possible to predict the partitioning behavior of proteins. The electrostatic energy depends on the protein size and ATPSs characteristics and varies in agreement with the difference in phase dielectric constants. Linear correlation of nonpolar energy, and the solvent accessible surface area was observed. Polymer structure and concentration have a significant influence on model viability. A good qualitative prediction of preferred phase for studied proteins was observed.

What are the structural features that drive partitioning of proteins in aqueous two-phase systems?

Protein partitioning in aqueous two-phase systems (ATPSs) represents a convenient, inexpensive, and easy to scale-up protein separation technique. Since partition behavior of a protein dramatically depends on an ATPS composition, it would be highly beneficial to have reliable means for (even qualitative) prediction of partitioning of a target protein under different conditions. Our aim was to understand which structural features of proteins contribute to partitioning of a query protein in a given ATPS. We undertook a systematic empirical analysis of relations between 57 numerical structural descriptors derived from the corresponding amino acid sequences and crystal structures of 10 well-characterized proteins and the partition behavior of these proteins in 29 different ATPSs. This analysis revealed that just a few structural characteristics of proteins can accurately determine behavior of these proteins in a given ATPS. However, partition behavior of proteins in different ATPSs relies on different structural features. In other words, we could not find a unique set of protein structural features derived from their crystal structures that could be used for the description of the protein partition behavior of all proteins in all ATPSs analyzed in this study. We likely need to gain better insight into relationships between protein-solvent interactions and protein structure peculiarities, in particular given limitations of the used here crystal structures, to be able to construct a model that accurately predicts protein partition behavior across all ATPSs.

Characterization of partitioning behaviors of immunoglobulin G in polymer-salt aqueous two-phase systems

The partitioning behavior of immunoglobulin G (IgG) in the aqueous two-phase system (ATPS) composed of poly(ethylene glycol) (PEG) and phosphate was studied. The parameters of ATPS exhibiting the pronounced effects on the partitioning behavior of IgG include phase composition, PEG molecular weight, and the addition of sodium chloride (NaCl). The accumulation of IgG at the interface of the ATPS increased drastically as the tie-line length (TLL) was increased. This trend was correlated with a linear relationship relating the natural logarithm of interfacial partition coefficient (ln G) to the difference of PEG concentration between the top phase and the bottom phase (Δ[PEG]), and a good fit was obtained. An attempt was made to correlate the natural logarithm of partition coefficient (ln K) to the presence of NaCl with the proposed linear relationship, ln K=α″ ln [Cl-]+β″. The proposed relationship, which serves as a better description of the underlying mechanics of the protein partitioning behavior in the polymer-salt ATPS, provides a good fit (r2>0.95) for the data of IgG partitioning. An optimum recovery of 99.97% was achieved in an ATPS (pH 7.5) composed of 14.0% (w/w) PEG 1450, 12.5% (w/w) phosphate and 5.0% (w/w) NaCl.

Interrelationship between partition behavior of organic compounds and proteins in aqueous dextran-polyethylene glycol and polyethylene glycol-sodium sulfate two-phase systems

Partition behavior of adenosine and guanine mononucleotides was examined in aqueous dextran-polyethylene glycol (PEG) and PEG-sodium sulfate two-phase systems. The partition coefficients for each series of mononucleotides were analyzed as a functions of the number of phosphate groups and found to be dependent on the nature of nucleic base and on the type of ATPS utilized. It was concluded that an average contribution of a phosphate group into logarithm of partition coefficient of a mononucleotide cannot be used to estimate the difference between the electrostatic properties of the coexisting phases of ATPS.The data obtained in this study were considered together with those for other organic compounds and proteins reported previously, and the linear interrelationship between logarithms of partition coefficients in dextran-PEG, PEG-Na2SO4 and PEG-Na2SO4-0.215M NaCl (all in 0.01M Na- or K/Na-phosphate buffer, pH 7.4 or 6.8) was established. Similar relationship was found for the previously reported data for proteins in Dex-PEG, PEG-600-Na2SO4, and PEG-8000-Na2SO4 ATPS. It is suggested that the linear relationships of the kind established in ATPS may be observed for biological properties of compounds as well.

Effects of low urea concentrations on protein-water interactions

Solvent properties of aqueous media (dipolarity/polarizability, hydrogen bond donor acidity, and hydrogen bond acceptor basicity) were measured in the coexisting phases of Dextran–PEG aqueous two-phase systems (ATPSs) containing .5 and 2.0 M urea. The differences between the electrostatic and hydrophobic properties of the phases in the ATPSs were quantified by analysis of partitioning of the homologous series of sodium salts of dinitrophenylated amino acids with aliphatic alkyl side chains. Furthermore, partitioning of eleven different proteins in the ATPSs was studied. The analysis of protein partition behavior in a set of ATPSs with protective osmolytes (sorbitol, sucrose, trehalose, and TMAO) at the concentration of .5 M, in osmolyte-free ATPS, and in ATPSs with .5 or 2.0 M urea in terms of the solvent properties of the phases was performed. The results show unambiguously that even at the urea concentration of .5 M, this denaturant affects partitioning of all proteins (except concanavalin A) through direct urea–protein interactions and via its effect on the solvent properties of the media. The direct urea–protein interactions seem to prevail over the urea effects on the solvent properties of water at the concentration of .5 M urea and appear to be completely dominant at 2.0 M urea concentration.

Osmotic Virial Coefficients as Access to the Protein Partitioning in Aqueous Two-Phase Systems

A promising alternative to state of the art chromatographic separations of therapeutic proteins is the extraction of the target protein using an aqueous two-phase system (ATPS). The use of an additional salt working as a displacement agent can influence the protein partitioning behavior in ATPS and thus enable a selective purification of the target protein. The selection of a suitable ATPS for protein extraction requires information concerning the protein-protein interactions (second osmotic virial coefficient B22 ) as well as the interactions between protein and solute (displacement agent and phase-forming components) (cross virial coefficient B23 ). In this work, the partitioning behavior and the precipitation affinity of immunoglobulin G (IgG) is considered within a polyethylene glycol (PEG)-phosphate ATPS. The influence on IgG partitioning upon addition of NaCl and (NH4)2 SO4 was investigated. In order to access the IgG precipitation affinity and the IgG partitioning behavior, the B22 and B23 values were determined for several combinations of solute [PEG, phosphate buffer, NaCl, and (NH4)2 SO4 ] and IgG based on static light scattering measurements. A qualitative estimation of the IgG precipitation affinity and the suitability of a solute as potential displacement agent within the PEG-phosphate ATPS on the basis of the measured B22 and B23 values is presented.

Responses of proteins to different ionic environment are linearly interrelated

Protein partitioning in aqueous two-phase systems (ATPS) is widely used as a convenient, inexpensive, and readily scaled-up separation technique. Protein partition behavior in ATPS is known to be readily manipulated by ionic composition. However, the available data on the effects of salts and buffer concentrations on protein partitioning are very limited. To fill this gap, partitioning of 15 proteins was examined in dextran–poly(ethylene glycol) ATPSs with different salt additives (Na2SO4, NaClO4, NaSCN, CsCl) in 0.11M sodium phosphate buffer, pH 7.4. This analysis reveals that there is a linear relationship between the logarithms of the protein partition coefficients determined in the presence of different salts. This relationship suggests that the protein response to ionic environment is determined by the protein structure and type and concentrations of the ions present. Analysis of the differences between protein structures (described in terms of proteins responses to different salts) and that of cytochrome c chosen as a reference showed that the peculiarities of the protein surface structure and B-factor used as a measure of the protein flexibility are the determining parameters. Our results provide better insight into the use of different salts in manipulating protein partitioning in aqueous two-phase systems. These data also demonstrate that the protein responses to different ionic environments are interrelated and are determined by the structural peculiarities of protein surface. It is suggested that changes in ionic microenvironment of proteins may regulate protein transport and behavior in biological systems.

Structural features important for differences in protein partitioning in aqueous dextran–polyethylene glycol two-phase systems of different ionic compositions

Partitioning of 15 proteins in dextran–70-polyethylene glycol (PEG)-8000 aqueous two-phase systems (ATPSs) in the presence of 0.01M sodium phosphate buffer, pH7.4 was studied. The effect of salt additives (NaCl, CsCl, Na2SO4, NaClO4 and NaSCN) at different concentrations on the protein partition behavior was examined. The salt effects on protein partitioning were analyzed by using the Collander solvent regression relationship between the protein partition coefficients in ATPSs with and without salt additives. The results obtained show that the presence and concentration of salt additives affect the protein partition behavior. Analysis of ATPSs in terms of the differences between the relative hydrophobicity and electrostatic properties of the phases does not explain the protein partition behavior. The differences between protein partitioning could not be explained by the protein size. The structural signatures for the proteins were constructed from partition coefficient values in four ATPSs with different salt additives, and the structural distances were calculated using cytochrome c as the reference structure. The structural distances for all the examined proteins (except lysozyme) were found to be interrelated. Analysis of about 50 different descriptors of the protein structures revealed that the partition behavior of proteins is determined by the peculiarities of their surfaces (e.g., the number of water-filled cavities and the averaged hydrophobicity of the surface residues) and by the intrinsic flexibility of the protein structure measured in terms of the B-factor (or temperature factor).

Effect of salt additives on protein partition in polyethylene glycol–sodium sulfate aqueous two-phase systems

Partitioning of 15 proteins in polyethylene glycol (PEG)–sodium sulfate aqueous two-phase systems (ATPS) formed by PEG of two different molecular weights, PEG-600 and PEG-8000 in the presence of different buffers at pH7.4 was studied. The effect of two salt additives (NaCl and NaSCN) on the protein partition behavior was examined. The salt effects on protein partitioning were analyzed by using the Collander solvent regression relationship between the proteins partition coefficients in ATPS with and without salt additives. The results obtained show that the concentration of buffer as well as the presence and concentration of salt additives affects the protein partition behavior. Analysis of ATPS in terms of the differences between the relative hydrophobicity and electrostatic properties of the phases does not explain the protein partition behavior. The differences between protein partitioning in PEG-600–salt and PEG-8000–salt ATPS cannot be explained by the protein size or polymer excluded volume effect. It is suggested that the protein–ion and protein–solvent interactions in the phases of ATPS are primarily important for protein partitioning.

Effects of chemical modifications in the partition behavior of proteins in aqueous two-phase systems: A case study with RNase A

Chemical modification of proteins is gaining importance due to the improvement in properties and the broader range of applications that these protein conjugates have. Once modified, several purification strategies need to be applied to isolate the conjugates of interest. Aqueous two-phase systems (ATPS) are an attractive alternative for the primary recovery of proteins and their conjugates. However, to better understand which biochemical parameters affect in greater degree the partition behavior of these modified proteins in ATPS, it becomes necessary to characterize the partition behavior of different species. In this work, ribonuclease A (RNase A) was selected as a model protein to address the partition behavior of chemically modified proteins in ATPS. Native, mono-PEGylated, Uniblue A, Dabsyl Chloride, and Direct Red 83 chemically modified RNase A's were partitioned in 16 different polyethylene glycol (PEG)-potassium phosphate ATPS. Results suggest that while the effects of system design parameters govern the partition of native RNase A, the behavior of the chemically modified species is more influenced by the physicochemical characteristics of the modifying molecules, that in most cases promote partition toward the top polymer-rich phase with recovery percentages as high as 86%. It has been found that both, the hydrophobicity and molecular weight of the modifying species play a preponderant role in conjugate partition behavior since as hydrophobicity increases partition is promoted towards the PEG-rich phase balancing the effect of the molecular weight of the modifying molecules that tends to shift partition towards the salt rich phase.

Partitioning Behavior of Wastewater Proteins in Some Ionic Liquids-Based Aqueous Two-Phase Systems

The recovery of wastewater proteins is gaining more attention in environmental protection. In the present work, phase diagram of the [Cnmim]X (n = 4, 6, 8; X = Cl and Br) + K2HPO4 aqueous two-phase systems have been determined, and the aqueous two-phase systems have been used to remove bovine serum albumin (BSA), lysozyme, and hemoglobin from the simulated wastewater. The influence of pH value, the composition of the aqueous two-phase systems, and the temperature of the systems on the partitioning of the proteins has been studied. It has been found that 76–100% of the proteins could be extracted into the ionic liquid-rich phase in a single step partition. The partition coefficients of lysozyme slightly increases but those of BSA and hemoglobin decrease with increasing tie line length (TLL). The higher extraction efficiency of the proteins is achieved at the pH value close to their isoelectric point and under higher temperatures. The results indicate that the IL-based aqueous two-phase systems have the potential for the separation of wastewater proteins.

An aqueous anonic/nonionic surfactant two-phase system in the presence of salt. 2. Partitioning of ice structuring proteins

Three ice structuring proteins (ISPs) with different molecular weights and net charges at pH 7.4 were partitioned in an aqueous anionic/nonionic surfactant two-phase system of SDS and Trition X-114 in the presence of 100 mM sodium chloride. The effects of temperature and SDS concentration on the protein partition coefficients were investigated. It is found that the protein partition coefficients increase with the increase of temperature and SDS concentration. Compared to the reported net dependence of protein partitioning in aqueous mixed surfactant two-phase systems, the observed protein partitioning behavior is driven primarily by excluded volume interactions between the proteins and the surfactant aggregates, rather than electrostatic interactions. Specifically, the ISPs partition preferentially to the micelle-poor phase, where they experience less excluded-volume interactions with the surfactant aggregates. The improvements in protein partitioning to the micelle-poor phase upon the SDS addition appear to result from the growth of the surfactant aggregates by forming mixed micelles, which enhances the excluded-volume interactions between the proteins and the surfactant aggregates.

Modulation of Membrane Domain Sizes and Properties by Hybrid Lipids

At present, there is a fundamental discontinuity between model membranes and real cellular membranes: The macroscopic Lo domains in model membranes and the nanoscopic lipid rafts found in cell membranes not only differ in size, but also in lipid order and protein partition behavior. Recently, new experimental evidence shows that nanometer-size Lo domains can exist in model membranes; however, the lipid mixtures must have at least 4-component and one of the components needs to be a “hybrid lipid”, which is a lipid having one saturated acyl chain and one unsaturated chain.

Flowsheet simulation of aqueous two‐phase extraction systems for protein purification

AbstractBACKGROUND: Aqueous two‐phase extraction (ATPE) has many advantages as an efficient, inexpensive large‐scale liquid–liquid extraction technique for protein separation. However, the realization of ATPE as a protein separation technology at industrial scales is rather limited due to the large, multidimensional design space and the paucity of design approaches to predict phase and product behavior in an integrated fashion with overall system performance. This paper describes a framework designed to calculate suitable flowsheets for the extraction of a target protein from a complex protein feed using ATPE. The framework incorporated a routine to set up flowsheets according to target protein partitioning behavior in specific ATPE systems and a calculation of the amounts of phase‐forming components needed to extract the target protein. The thermodynamics of phase formation and partitioning were modeled using Flory‐Huggins theory and calculated using a Gibbs energy difference minimization approach.RESULTS: As a case study, suitable flowsheets to recover phosphofructokinase from a simple model feedstock using poly(ethylene glycol)‐dextran (PEG6000‐DxT500) and poly(ethylene glycol)‐salt (PEG6000‐Na3PO4) two‐phase systems were designed and the existence of feasible solutions was demonstrated. The flowsheets were compared in terms of product yield, product purity, phase settling rate and scaled process cost. The effect of the mass flowrates of phase‐forming components on product yield and purity was also determined.CONCLUSION: This framework is proposed as a basis for flowsheet optimization for protein purification using ATPE systems. Copyright © 2010 Society of Chemical Industry

Prediction of the partitioning behaviour of proteins in aqueous two-phase systems using only their amino acid composition

The prediction of the partition behaviour of proteins in aqueous two-phase systems (ATPS) using mathematical models based on their amino acid composition was investigated. The predictive models are based on the average surface hydrophobicity (ASH). The ASH was estimated by means of models that use the three-dimensional structure of proteins and by models that use only the amino acid composition of proteins. These models were evaluated for a set of 11 proteins with known experimental partition coefficient in four-phase systems: polyethylene glycol (PEG) 4000/phospate, sulfate, citrate and dextran and considering three levels of NaCl concentration (0.0% w/w, 0.6% w/w and 8.8% w/w). The results indicate that such prediction is feasible even though the quality of the prediction depends strongly on the ATPS and its operational conditions such as the NaCl concentration. The ATPS 0 model which use the three-dimensional structure obtains similar results to those given by previous models based on variables measured in the laboratory. In addition it maintains the main characteristics of the hydrophobic resolution and intrinsic hydrophobicity reported before. Three mathematical models, ATPS I–III, based only on the amino acid composition were evaluated. The best results were obtained by the ATPS I model which assumes that all of the amino acids are completely exposed. The performance of the ATPS I model follows the behaviour reported previously, i.e. its correlation coefficients improve as the NaCl concentration increases in the system and, therefore, the effect of the protein hydrophobicity prevails over other effects such as charge or size. Its best predictive performance was obtained for the PEG/dextran system at high NaCl concentration. An increase in the predictive capacity of at least 54.4% with respect to the models which use the three-dimensional structure of the protein was obtained for that system. In addition, the ATPS I model exhibits high correlation coefficients in that system being higher than 0.88 on average. The ATPS I model exhibited correlation coefficients higher than 0.67 for the rest of the ATPS at high NaCl concentration. Finally, we tested our best model, the ATPS I model, on the prediction of the partition coefficient of the protein invertase. We found that the predictive capacities of the ATPS I model are better in PEG/dextran systems, where the relative error of the prediction with respect to the experimental value is 15.6%.