Article pubs.acs.org/ac
Method for Estimation of Protein Isoelectric Point Sari Pihlasalo,* Laura Auranen, Pekka Han̈ ninen, and Harri Har̈ ma ̈ Laboratory of Biophysics and Medicity Research Laboratory, University of Turku, Tykistökatu 6A, FI-20520 Turku, Finland S Supporting Information *
ABSTRACT: Adsorption of sample protein to Eu3+ chelate-labeled nanoparticles is the basis of the developed noncompetitive and homogeneous method for the estimation of the protein isoelectric point (pI). The lanthanide ion of the nanoparticle surfaceconjugated Eu3+ chelate is dissociated at a low pH, therefore decreasing the luminescence signal. A nanoparticle-adsorbed sample protein prevents the dissociation of the chelate, leading to a high luminescence signal. The adsorption efficiency of the sample protein is reduced above the isoelectric point due to the decreased electrostatic attraction between the negatively charged protein and the negatively charged particle. Four proteins with isoelectric points ranging from ∼5 to 9 were tested to show the performance of the method. These pI values measured with the developed method were close to the theoretical and experimental literature values. The method is sensitive and requires a low analyte concentration of submilligrams per liter, which is nearly 10000 times lower than the concentration required for the traditional isoelectric focusing. Moreover, the method is significantly faster and simpler than the existing methods, as a ready-to-go assay was prepared for the microtiter plate format. This mix-and-measure concept is a highly attractive alternative for routine laboratory work.
T
analyzer and many accessories such as capillary columns and pI marker proteins. cIEF has a higher degree of automation and quantification, high reproducibility, better resolution, higher sensitivity, and faster focusing and analysis than gel-based IEF.2−4 cIEF cannot be applied in the presence of a high salt concentration in a sample as it produces changes to the pH gradient and reduces peak capacity and reproducibility. cIEF suffers also from precipitation problems as the protein peak is highly concentrated and the solubility of the protein is low, close to the pI.5 Chemicals may be added to maintain the solubility, but the measurement accuracy may then be sacrificed.6,7 cIEF typically requires 1 μg of sample protein at a protein concentration of 0.1 g/L.2,3 The theoretical calculation of pI exploits the pKa values for the amino acid residues.8 The estimation is fast, but the amino acid sequence must be known; thus, the pI for unknown structures cannot be computed. Moreover, pKa values of the amino acid residues depend on the environmental factors, such as hydrophobicity.1 Therefore, the theoretical pI may differ from the experimental value as the theoretical calculation assumes a random coil structure, and in the experimental determination the protein possesses a native structure. The development of simple methods for the estimation of protein pI has been reported in the literature.9,10 Jaffé9 has proposed that precipitation of negatively charged proteins occurring in the presence of cationic detergents is applicable in the estimation of pI. The method is simple to perform without
he isoelectric point (pI) is an intrinsic property of proteins and defined as the value at which a protein carries a zero net charge. Thus, it determines the sign of the net charge for different pH ranges. pI values are generally determined with either gel-based isoelectric focusing (IEF) or capillary isoelectric focusing (cIEF) methods.1−3 Other less common methods are chromatofocusing, discontinuous electrophoresis, ion-exchange chromatography, and isoelectric precipitation. IEF and cIEF methods are relatively precise and can be applied to protein mixtures as the different proteins are separated from each other by their pI. However, the determination requires high concentrations of sample protein, special instrumentation, multiple assay components, solutions, and accessories, and it is both laborious and time-consuming.3 In isoelectric focusing (IEF),1,3 proteins are separated according to their pI values in a gel with a continuous pH gradient and under the influence of an electric field. The determination of pI relies on the use of pI marker proteins, and thus, high quality and precise properties should be fulfilled. The experimental procedure comprises several steps: application of the samples, focusing, fixing, equilibration, staining in a separate chamber, destaining, and preservation. Part of the solutions should be prepared fresh for each experiment and may contain toxic or corrosive reagents. The experimental procedure is long and may take a whole working day, including the preparation of the solutions. IEF suffers from the limitation of low solubility of relatively hydrophobic proteins and requires a high protein concentration, typically 1 g/L.2,3 With capillary isoelectric focusing (cIEF),1 proteins are fractionated similarly to gel-based IEF, but the fractionation is performed in a capillary column. The determination requires an © 2012 American Chemical Society
Received: June 7, 2012 Accepted: September 3, 2012 Published: September 4, 2012 8253
dx.doi.org/10.1021/ac301569b | Anal. Chem. 2012, 84, 8253−8258
Analytical Chemistry
Article
the investment of any specialized equipment. The cationic detergent is added to protein solutions at varying pH values. The pI of the protein corresponds to the pH of the most acidic buffer solution precipitating the protein. The downside to the method is that the pH values of the test solutions must be rechecked when applied in acidic or basic solutions. The method requires a protein concentration above 10 g/L and is applicable only to highly soluble proteins. Moreover, the estimation depends on the visual observation and is thus subjective. Another relatively simple method, which is based on the changes in the adsorption of the protein to an ion-exchange material as a function of pH, has been developed by Yang et al.10 Approximately 1 h is needed to assay pI values at the desired temperatures, and the method is relatively accurate. The pI is estimated from the determined quantity of unadsorbed protein as a function of pH. The method requires more than 0.1 mg of protein for one pH, and for several pH values the total quantity can be several milligrams. To measure the unbound protein, a centrifugation step is required after the adsorption and subsequent assaying of the supernatant either by a UV absorbance of 280 nm or by enzymatic activity. A cross-partitioning11−13 method utilizes the partition coefficient of proteins in the presence of salts. The partition coefficient versus pH is measured for alkali chloride and alkali sulfate systems. The pI is determined from the cross-point of these two curves for both alkali systems. The experimental procedure contains few technical difficulties, preparation of several solutions, desalting of proteins, and centrifugation and takes hours to complete. The method requires a high protein concentration of 1 g/L and ∼1 mg of protein at each pH, and thus the total quantity is tens of milligrams. Moreover, the pI and cross-point values differ, e.g., for trypsin and αchymotrypsin.14 Previously, we have constructed and reported a mix-andmeasure fluorescence quenching, time-resolved luminescence resonance energy transfer (TR-LRET), and dissociation-based methods for the quantification of proteins using the adsorption of an analyte to nanoparticles.15−18 These microtiter-well assays are extremely sensitive with detection limits of subnanogram quantities of the protein. Here, we developed a sensitive and fast method for the estimation of the protein isoelectric point. The estimation relies on the reduced adsorption of the sample protein at a pH above its pI, due to the electrostatic repulsion between the particles and protein with charges of equal sign. The dissociation system was applied to detect the adsorption of the protein: adsorption of the sample protein at a pH below its pI and a fast decrease of pH. The adsorbed protein prevents dissociation of a europium ion from the surface-conjugated Eu3+ chelate, and a high time-resolved luminescence signal is monitored (Figure 1). On the contrary, the protein incubated with the particles at a pH above the pI leads to dissociation of the Eu3+ ion and a reduced luminescence signal. A ready-to-go high-throughput microtiter plate format containing the buffer components for different pH values was prepared. The analytical conditions of the method were optimized for the dissociation buffer, the concentration of sample protein, and the number of particles. The performance of the method was tested with four proteins possessing different pI values and for the presence of possible interfering agents.
Figure 1. Schematic illustration of the assay principle. (a) At a pH below the pI, the sample protein adsorbs efficiently onto the nanoparticle surface due to attractive electrostatic interactions and a high luminescence signal is detected. (b) At a pH above the pI, the sample protein adsorption is less efficient due to repulsive charges between the protein and the particles and the protection of the Eu3+ chelate on the particle surface is lost leading to a low luminescence signal.
■
MATERIALS AND METHODS Materials. Amino-modified polystyrene particles, 240 nm in diameter, were purchased from Spherotech Inc. (Libertyville, IL). 9-Dentate Eu(III) chelate, {2,2′,2″,2‴-{[4′-(4‴-isothiocyanatophenyl)-2,2′:6,6″-terpyridine-6,6″-diyl]bis(methylenenitrilo)}tetrakis(acetato)}europium(III), was from the Laboratory of Biophysics (Turku, Finland) and was synthesized and characterized according to the literature.19 The DELFIA enhancement solution was from Wallac, PerkinElmer Life and Analytical Sciences (Turku, Finland). Polystyrene low-fluorescence microtiter plates with 96 wells were purchased from Nunc GmbH & Co. KG (Langenselbold, Germany). All the proteins (albumin from bovine serum (BSA), fraction 5, >99%; carbonic anhydrase from bovine erythrocytes, 90%; γ-globulins from bovine blood, 97%; trypsinogen from bovine pancreas, 92%; pepsin from porcine stomach mucosa, 55%; and unfractionated histone from calf thymus, high purity stated by the manufacturer), concentrated hydrochloric acid, sodium tetraborate, potassium chloride, oxalic acid, Triton X-100, polyethylene glycol 3000 (PEG 3000), and sodium chloride were ordered from Sigma-Aldrich Co. (St. Louis, MO). Citric acid and sodium hydroxide were purchased from FF-Chemicals (Yli-Ii, Finland). Tris(hydroxymethyl)aminomethane was purchased from Calbiochem-Novabiochem Co. (La Jolla, CA), glycine from J. T. Baker (Deventer, The Netherlands), potassium dihydrogen phosphate from Merck KGaA (Darmstadt, Germany), and ethanol (ETAX A) from Altia Finland (Helsinki, Finland). High-purity Milli-Q water was used to prepare all aqueous solutions. Methods. Preparation of Microtiter Wells with Dried Buffer Components (Ready-to-Go Plates). A universal buffer containing 25 mM sodium tetraborate, citric acid, tris(hydroxymethyl)aminomethane, potassium dihydrogen phosphate, and potassium chloride with a pH varying from 4.3 to 9.8 was prepared by adjusting the pH with concentrated hydrochloric acid and sodium hydroxide (original pH). The buffer was diluted to half, and 2.8 μL of the dilution was added to the 8254
dx.doi.org/10.1021/ac301569b | Anal. Chem. 2012, 84, 8253−8258
Analytical Chemistry
Article
The ratio between signals measured at a pH close to the pI (5.0) and above the pI (6.7) was monitored at varying concentrations of BSA. In this test, 70 μL of BSA in water and 5.0 μL of the particles labeled with 9-dentate Eu3+ chelate (21000 chelates per particle, 15 amol) were mixed in the wells of the ready-to-go plate. After mixing, 10 μL of 0.5 M glycine (pH 0.8) was added. In the final optimized assay setup, the same concentration was used for all proteins. Estimation of the Isoelectric Point of Different Proteins with the Method. To test the performance of the developed method for the estimation of pI values for different proteins, four proteins (BSA, carbonic anhydrase, γ-globulin, and trypsinogen) with different isoelectric points were tested for the method. The signals as a function of adsorption buffer pH were measured using an optimized, typical protocol of the method. Effects of the Interfering Agents on the Method. To test the tolerance of the developed method for general interfering agents in the protein samples, four chemical agents [Triton X100, sodium chloride, ethanol, and polyethylene glycol 3000 (PEG 3000)] were studied. BSA was diluted to 210 μg/L in either water or water solutions containing varying concentrations of contaminants. The tests were run using an optimized, typical protocol for the method with 15 amol of particles, and the ratio between signals measured at a pH close to the pI (5.0) and above the pI (5.7) was monitored at varying concentrations of contaminants.
polystyrene microtiter plate wells. The plates were dried at 50 °C overnight. The final pH values in the assay were measured after dissolving the dried components in the wells into 70 μL of water. The pH values in this article are given as final values in the assay, unless otherwise stated. Labeling of Amino-Modified Polystyrene Particles with Eu3+ Chelate. Amino-modified polystyrene particles, 240 nm in diameter, were labeled with 9-dentate Eu3+ chelate, {2,2′,2″,2‴{[4′-(4‴-isothiocyanatophenyl)-2,2′:6,6″-terpyridine-6,6″-diyl]bis(methylenenitrilo)}tetrakis(acetato)}europium, as described previously.18 Before labeling, the particles were washed by centrifugation (19000g for 5 min) several times. The 9-dentate (20 nmol) Eu3+ chelate was added to 250 μL of a 30 mM carbonate buffer (pH 9.9) containing 0.11 pmol of particles. After overnight incubation, the particles were washed in water by centrifugation (19000g for 5 min) several times. The number of Eu3+ chelates per particle was determined with a DELFIA enhancement solution, as recommended by the manufacturer, except that the europium ions were first dissociated from the chelates in 1 M hydrochloric acid before mixing and enhancing with DELFIA. Zeta Potential Measurements for Amino-Modified Polystyrene Particles Unlabeled and Surface-Labeled with a 9Dentate Eu3+ Chelate. Zeta potentials were measured with a Zetasizer Nano Series Nano ZS instrument (Malvern Instruments Ltd., Malvern, U.K.) as described previously.18 The zeta potential of unlabeled amino-modified polystyrene particles (44 pM) was measured in water and a universal buffer containing 0.5 mM sodium tetraborate, citric acid, tris(hydroxymethyl)aminomethane, potassium dihydrogen phosphate, and potassium chloride with a pH varying from 2.5 to 11.5. The zeta potential of particles that were surface-labeled with a 9-dentate Eu3+ chelate (75000 chelates per particle, 44 pM) was measured in water. Assay for Estimation of Protein Isoelectric Point. Stock solutions of all proteins were prepared in water and stored at 4 °C. In a typical microtiter plate assay, 70 μL of the sample protein solution at a concentration of 210 μg/L diluted in water and 5.0 μL of the particles labeled with 9-dentate Eu3+ chelate (21000 chelates per particle, 15 amol) in water were mixed in the well of the ready-to-go plate. After mixing, 10 μL of 0.5 M glycine buffer (pH 0.8) was added. The final concentration of the particles was 0.17 pM. After 30 min of incubation, luminescence emission intensities were measured in a 400 μs window after a 400 μs delay time using a Victor2 multilabel counter (Wallac, Perkin-Elmer Life and Analytical Sciences) with 340 nm excitation and 615 nm emission wavelengths. Assay Optimizations. The assay contained two buffers: a universal buffer for obtaining varying pH values for adsorption (adsorption buffer as dried components) and an optimal buffer for dissociation of the Eu3+ ion from the Eu3+ chelate (dissociation buffer). The dissociation buffer was optimized for pH and the components. Two buffers, 0.5 M glycine and oxalic acid, and a hydrochloric acid solution were tested at different pH values. In this test, 70 μL of 210 μg/L BSA in water and 5.0 μL of the particles labeled with 9-dentate Eu3+ chelate (21000 chelates per particle, 7.3 amol) were mixed in the well of the ready-to-go plate (final pH values 5.0 and 6.7). After mixing, 10 μL of hydrochloric acid solution, 0.5 M glycine, or 0.5 M oxalic acid at varying pH values was added to obtain different final pH values. The concentration of the sample protein was optimized with BSA for estimation of the protein pI as precisely as possible.
■
RESULTS AND DISCUSSION A nanoparticle-based noncompetitive method was developed to estimate isoelectric points of proteins. The sample protein adsorbs to the nanoparticles at a pH below its pI due to the attractive electrostatic interactions between negatively charged particles and the positively charged protein (Figure 1). The adsorption efficiency decreases at increasing pH as the particles and protein repel each other, both having a negative charge. The adsorption of the sample protein to a Eu3+ chelate-labeled particle at a pH below the pI protects the chelate from dissociation at decreased pH values, while the lanthanide ion from the unprotected chelate is prone to dissociation. Thus, a high luminescence signal is detected at a pH below the pI and a low signal at a pH above the pI. The assay was optimized for analytical conditions, such as the concentration of a sample protein, number of particles, and the dissociation buffer and its pH. The performance of the assay was tested for four different proteins and for the presence of various potential interfering agents. A ready-to-go high-throughput microtiter plate format containing the buffer components for different pH values was prepared to develop a simple system for the end users. Universal buffers (see Materials and Methods for the composition) having pH values between 4.3 and 9.8 were dried in the microtiter wells. Because the pH changed after drying and subsequent addition of water, the final pH values were measured after dissolving the dried components into water (see Figure S1 of the Supporting Information for the relation between the final pH in the assay and the original pH of the universal buffer added to the wells). The final pH values were obtained at intervals of approximately 0.1−0.4 on the pH scale and are given in this article unless otherwise stated. The developed assay relies on the charge of the particles and protein. Therefore, zeta potentials were measured for the particles. The zeta potentials of the unlabeled and labeled 8255
dx.doi.org/10.1021/ac301569b | Anal. Chem. 2012, 84, 8253−8258
Analytical Chemistry
Article
particles were −55 and −43 mV in water (assuming Smoluchowski’s20 approximation), respectively. As this difference was not large, the zeta potential was measured as a function of pH in the universal buffer only for the unlabeled particles (see Figure S2 of the Supporting Information). The zeta potential of the amino-modified nanoparticles was clearly negative (−59 to −48 mV) and did not change significantly at a pH range from 2.5 to 11.5. The negative zeta potential originates from the surface groups on the polystyrene particle, dominated by the sulfite groups [pKa 1.9 (−SO3H)] and influenced by the amino, carboxylic, and sulfate groups [pKa 11 (−NH2), pKa 4.7 (−COOH), and pKa −3 (−SO4H)]. This negative charge provides the basis for the method, as the method is viable only if the charge of the surface of the particles was not a function of pH. Consequently, the attractive electrostatic interaction and adsorption are optimal at a pH below the pI of the protein.21,22 In contrast, the adsorption efficiency decreases at pH values above the pI of the protein due to the repulsion between the protein and the particle, with charges of equal signs. The protein sample is added in water to the prepared readyto-go microtiter plate wells containing the dried universal buffer components to obtain different pH values for the adsorption. The particles labeled with Eu3+ chelate are added to the wells. Finally, the pH is decreased to dissociate the Eu3+ ion from the Eu3+ chelate with a dissociation buffer. Two buffers, 0.5 M glycine and 0.5 M oxalic acid at different pH values, were tested and compared to the hydrochloric acid solutions for the dissociation. In the test, the potency to detect the difference in the efficiency of the adsorption to the particles was monitored for BSA adsorbed at pH values below and above its pI (pH values of 5.0 and 6.7, respectively). After the adsorption, the dissociation solution was added to the wells and the luminescence signal was measured. The ratio between the signals at these two adsorption pH values was the highest for glycine at pH 0.8, which was selected for further studies (see Figure S3 of the Supporting Information for the signal ratio as a function of the dissociation buffer pH). Rapid dissociation of the Eu3+ ion is required to prevent the adsorption of the nonbound positively charged sample protein at the dissociation step. The number of particles was tested using 0.5 M glycine (pH 0.8) as the dissociation buffer. The number of particles did not affect the performance significantly. The signal ratio between pH values of 5.0 and 6.7 for BSA adsorption increased approximately 20% when the number of particles was doubled. On the basis of these tests, a quantity of 15 amol of particles was chosen for further studies. The concentration of sample protein was optimized with BSA for the estimation of pI by taking the signal ratio at pH values of 5.0 and 6.7 (Figure 2). The ratio was the highest at 210 μg/L BSA corresponding to 3.2 nM protein. The ratio decreased at lower concentration due to lower coverage of the particle surface and at higher concentration as the tendency for protein adsorption increased. This high sensitivity method can be applied at 500−5000 times lower concentrations compared to the conventional IEF and cIEF methods.2,3 Four different proteins having pI values between 5 and 9 were selected to demonstrate the functionality of the assay concept (Figure 3). The adsorption at a pH below 4 could not be exploited due to the instability of the chelate, and thus, the current method is limited to pI values above 4. Lanthanide chelates having higher stability may potentially increase the dynamic range to low pH values, although proteins with very
Figure 2. Optimization of the sample protein concentration. The ratio between signals with adsorption pH close to the pI (5.0) and above the pI (6.7) as a function of the BSA concentration.
Figure 3. Estimation of the pI. Normalized signal as a function of the adsorption buffer pH for different proteins.
low pI values are relatively rare.23−26 The performance of the method is weaker at high pH values, as efficient dissociation is not achieved for a pH higher than 10.5 with the dissociation buffer used. It might be possible to measure higher pI values if the buffer solutions are changed. However, the range is limited with the selected buffers. In Figure 3, the normalized luminescence signals for each protein are presented as a function of the adsorption pH showing the pH-dependent adsorption profiles. The average coefficient of variation was 6% for the measured signals. The decrease in the luminescence signal is observed at a pH close to the pI of each protein. In the pH range of the higher plateau on the curves, the protein carries a net positive charge and adsorbs efficiently to the particles. In contrast, both the protein and the particles carry a total negative charge at the lower plateau. The charge of the protein changes during the decreasing curvature, and a linear fitting to this curvature, by minimizing the sum of the squared differences between an observed value and a fitted value, was used to estimate the pI. The midpoint of the transition (i.e., the normalized luminescence signal at 0.5) gave an estimate for the protein pI due to the zero net charge. The horizontal line in Figure 3 indicates the estimated pI values.10 Generally, it is relatively difficult to define a correct pI, as it depends on the structure of a protein. If the protein is assumed to have a random coil or a native structure, the pI may vary due to the changes in the pKa values of the amino acids having 8256
dx.doi.org/10.1021/ac301569b | Anal. Chem. 2012, 84, 8253−8258
Analytical Chemistry
Article
The performance of the developed assay was examined in the presence of four contaminants commonly found in biological samples. Due to the high tolerance shown for the corresponding protein quantification assay based on the same detection principle, we expected the developed method to have a good tolerance for contaminants.18 Detergent Triton X-100, NaCl salt, organic solvent ethanol, and polymer PEG 3000 were tested for their highest tolerated concentration. Nonionic detergents, such as Triton X-100, are suitable in the extraction of proteins from cell preparations. Proteins are stored in buffers with physiological salt concentrations, and organic solvents and polyethylene glycol (PEG) are used as precipitants in protein crystallization. The highest tolerated concentration was defined as the concentration resulting in a
