Determination of the molecular weight of proteins in heterogeneous mixtures: use of an air-driven ultracentrifuge for the analysis of protein--protein interactions.

A high-speed air-driven ultracentrifuge (Airfuge) has been used to study the molecular weights of proteins in heterogeneous mixtures. The method is based on previous studies (M. A. Bothwell, G. J. Howlett, and H. K. Schachman, 1978, J. Biol. Chem., 253, 2073–2077) which showed that at sedimentation equilibrium in the Airfuge the fraction of a protein remaining in an upper fraction of the Airfuge tube is almost linearly related to the exponential of the reduced molecular weight of the protein. In this study the total fraction of each particular protein remaining in an upper fraction of the Airfuge tube is determined by quantitative sodium dodecyl sulfate-gel electrophoresis. This procedure allows a wide range of proteins to be analyzed in a single Airfuge experiment. The method yields the “native” molecular weights of the protein components and is independent of the shape of the macromolecules being studied. Interactions occurring between the components in solution can be detected from the Airfuge data, and procedures are described which allow the experimental data for such interactions to be analyzed in terms of an equilibrium constant for the interaction. Results obtained for the electrostatic interaction at neutral pH between lysozyme and ovalbumin (K = 1.1 × 10⁵, m^?1) and lysozyme and bovine serum albumin (K = 1.0 × 10⁵, m^?1) agree well with literature values.     

Caldesmon. Molecular weight and subunit composition by analytical ultracentrifugation

A wide range of values has been reported for the subunit and molecular weights of smooth muscle caldesmon. There have also been conflicting reports concerning whether caldesmon is a monomer or dimer. We attempted to resolve these uncertainties by determining the molecular weight of chicken gizzard smooth muscle caldesmon using the technique of sedimentation equilibrium in the analytical ultracentrifuge. Unlike previous methods that have been used to estimate the molecular weight of caldesmon, the molecular weight determined by equilibrium sedimentation does not depend upon assumptions about the shape of the molecule. We concluded that caldesmon in solution is monomeric with a molecular mass of 93 +/- 4 kDa, a value that is much less than those previously reported in the literature. This new value, in conjunction with sedimentation velocity experiments, led to the conclusion that caldesmon is a highly asymmetric molecule with an apparent length of 740 A in solution. The mass of a cyanogen bromide fragment, with an apparent mass of 37 kDa from sodium dodecyl sulfate-polyacrylamide gel electrophoresis, was determined to be 25.1 +/- 0.6 kDa using sedimentation equilibrium. These results imply that the reported molecular weights of other fragment(s) of caldesmon have also been overestimated. We have determined an optical extinction coefficient for caldesmon (E1%(280 nm) = 3.3) by determining its concentration from its refractive index which was measured in the analytical ultracentrifuge. From the above values of the molecular weight and the extinction coefficient, we redetermined that the caldesmon molecule has two cysteines and recalculated the stoichiometric molar ratio of actin/tropomyosin/caldesmon in the smooth muscle thin filament to be 28:4:1.     

Study of abnormal plasma low-density lipoprotein in rhesus monkeys with diet-induced hyperlipidemia.

Male rhesus monkeys were divided into three groups: five were fed a regular primate chow diet and were used as controls; four received an "average" American diet; and five a special low-fat primate chow diet supplemented with 25% coconut oil and 2% cholesterol. In all of these animals, the plasma low-density lipoproteins (LDL) were isolated by ultracentrifugal flotation between densities of 1.019 and 1.050 g/ml. The LDL of the five control monkeys had variable molecular weights, with a mean value of 3.12 +/- 0.21 X 10(6) (range: 2.92 X 10(6) to 3.45 X 10(6)), and an average partial specific volume of 0.969 +/- 0.003 ml/g; both were assessed by flotation equilibrium analysis in the analytical ultracentrifuge. In the individual animals, however, the physical properties of LDL were invariant with time. The administration of either an "average" American diet or a coconut oil-cholesterol diet was accompanied by hypercholesterolemia associated with changes in LDL which were characterized by increases in molecular weight to 3.52 +/- 0.21 X 10(6) (average of nine monkeys) and in partial specific volume to 0.973 +/- 0.002 ml/g. These changes were particularly evident when the molecular weight of LDL from monkeys in the normolipidemic state was compared with that obtained from the same monkeys during the hyperlipidemic state. Chemical analyses revealed that the particles from the hyperlipidemic animals had a relatively higher cholesteryl ester content, a slight increase in phospholipids, and a marked decrease to nearly complete absence of triglycerides. The other lipoprotein components, protein, carbohydrate, free cholesterol, and fatty acids, did not vary significantly from those of control LDL. It is concluded that the administration of atherogenic diets causes structural changes in LDL which appear to be accounted for, at least in part, by changes in the composition of the lipid moiety. The changes in physical and chemical properties noted in the LDL of rhesus monkeys with experimentally induced hypercholesterolemia contrast with the apparent structurally normal LDL from rhesus monkeys with spontaneous hypercholesterolemia reported previously.     

Molecular weights of nitrogenase components from Azotobacter vinelandii.

Meniscus depletion sedimentation equilibrium ultracentrifuge experiments were performed on purified MoFe and Fe proteins of $\text{Azotobacter vinelandii}$. The MoFe protein was found to have a molecular weight of 245,000, using an experimentally confirmed partial specific volume of 0.73. The MoFe protein formed one band on sodium dodecyl sulfate gel electrophoresis and had a subunit molecular weight of 56,000. The subunit molecular weight from ultracentrifuge experiments in 8 M urea was 61,000. The molecular weight of the Fe protein was calculated to be 60,500 in meniscus depletion experiments. Similar experiments in 8 M urea solvent indicated a subunit molecular weight of 30,000. A subunit molecular weight of 33,000 was obtained from sodium dodecyl sulfate gel electrophoresis experiments.     

Sizes and mass distributions of clathrin-coated vesicles from bovine brain

Clathrin-coated vesicles obtained from bovine brain have been studied by ultracentrifugation and dynamic light scattering techniques to provide information on their sedimentation and mass distributions and their average diffusion coefficients. "Uncoated" vesicles, obtained by removing the protein coat from coated vesicles, have been similarly characterized. For typical preparations, maximal values of approximately 210 and 95 S are observed for the sedimentation coefficients of coated and uncoated vesicles, respectively. Corresponding values for the average molecular weights, determined from values of average sedimentation and diffusion coefficients, are 49 X 10(6) and 13 X 10(6); values obtained by equilibrium sedimentation are 37.2 X 10(6) and 10.6 X 10(6). In order to obtain these results, some minor modifications of sedimentation and light-scattering techniques have been devised which may have application to other studies of size distributions of large particles.     

Preparation of a coated vesicle-enriched fraction from plant cells

A fraction rich in coated vesicles has been prepared from suspension-cultured cells of tobacco (Nicotiana tabacum L.) by sucrose gradient centrifugation. Isolated, negatively-stained plant coated vesicles are approx. 100 nm in diameter, and show the characteristic basket-like structure of the clathrin coat previously reported for both plant [2–5] and animal [1, 6–9] coated vesicles. Analysis of the various plant subcellular fractions by SDS polyacrylamide gel electrophoresis demonstrates that a polypeptide of 190 000 D is enriched in parallel with the morphologically identifiable coated vesicles. It is postulated that this polypeptide is plant clathrin with a molecular weight about 10 000 D greater than that previously reported for animal clathrin [1, 6].     

Improved methods for the structural analysis of the amylose-rich fraction from rice flour

Cooking and sensory properties of rice are largely determined by the amylose content and structure. For relationships between functional and structural properties, a more accurate method to determine the structure of amylose is required. Here we calibrate size exclusion chromatography (SEC) columns, using Mark-Houwink parameters for linear starch and pullulan standards, to obtain the true molecular weight distribution of linear starch. When the molecular weight distribution is reported relative to pullulan, rather than the actual molecular weight which is readily obtained from universal calibration, it is seen that the molecular weights of longer amylose chains are greatly underestimated. We validate the SEC method to enable the measurement of the hydrodynamic volume distribution of the starch by examining reproducibility and recovery. Analysis of the starch in the sample pre- and post-SEC shows that 20% of the carbohydrate is not recovered. Comparison of the weight-average degree of polymerization, X(w), of (undebranched) starch of pre- and post-SEC is made using iodine binding as well as Berry plots of data from multi-angle laser light scattering (MALLS). These both show that current SEC techniques for starch analysis lead to significant loss of high molecular weight material. Indeed, for the systems studied here, the values for X(w) after SEC are about three times lower than those before SEC. Iodine-starch complexes of pre- and post-SEC samples reveals that the SEC techniques give reliable data for the amylose fraction but not for amylopectin. We address reports in the literature suggesting that the conventional isoamylase method for debranching starch would lead to incomplete debranching and thus incorrect molecular weight distributions. However, it is shown using (1)H NMR that isoamylase can completely debranch the amylose (to within the detection limit of 0.5%), and by SEC that successive incubation with isoamylase, alpha-amylase, and beta-amylase can degrade the amylose-rich fraction completely to maltose. We develop a method to obtain a hot water soluble fraction (HWSF), rich in undamaged amylose molecules, directly from rice flour, avoiding the structural degradation of previous techniques. With appropriate sample handling, the formation of associations between starch chains is minimized. With the combination of calibrated and validated SEC methods, and an improved extraction of amylose from rice, the X(w) for both HWSF and debranched HWSF are found to be much larger than has previously been reported.     

An airfuge centrifugation procedure for the measurement of ligand binding to membrane-associated and detergent-solubilized plasma membrane receptors

A method is described in which high-speed centrifugation of membranes through an oil phase is used to separate membrane-bound and detergent-solubilized polypeptide receptor-iodinated ligand complexes from unbound ligands. Three centrifuges, the Brinkmann Eppendorf (5412), the Beckman Microfuge B and the Beckman Airfuge were evaluated for this capability. Under the conditions described, the Beckman Airfuge surpassed the others in recovering previously 125I- and 32P-labelled cell membranes. The Airfuge method was compared with the more classically employed membrane filtration method to measure specific [125I]insulin and [125I]thrombin binding to human placental membranes and an enriched plasma membrane fraction from mouse embryo fibroblasts, respectively, are found to be 4 to 6 times more sensitive. For example, specific binding of ligand to its receptor was demonstrated with 5 micrograms of protein. With slight modifications, the polyethyleneglycol 6000 method of precipitating 125I-labelled ligand-soluble receptor complexes can be adapted to the Airfuge sedimentation through oil procedure.     

Sedimentation velocity analyses of the effect of hydrostatic pressure on the 30 S microtubule protein oligomer

Increasing hydrostatic pressure in the analytical ultracentrifuge by increasing rotor velocity and overlayering protein samples with oil caused a depolymerization of the 30 S oligomer of microtubule protein. This results indicates that the reaction of 6 S microtubule protein to form the oligomer was accompanied by a positive volume change. The effect of hydrostatic pressure on the 6 S to 30 S transition was employed to demonstrate the presence of a rapidly reversible equilibrium between these components by showing polymerization or depolymerization of the oligomer during the course of ultracentrifugation. The magnitude of the partial specific volume change accompanying this reaction was estimated from mass fraction measurements of microtubule protein solutions at a variety of hydrostatic pressures to be about 9 X 10(-4) ml g-1.     

Molecular weight of human high-molecular-weight kininogen light chain by equilibrium sedimentation in an air-driven ultracentrifuge

High-molecular-weight kininogen, a nonenzymatic glycoprotein of the intrinsic blood coagulation system, is proteolytically cleaved by kallikrein as an early event in the activation of this system. The light chain of cleaved kininogen retains the ability to form specific noncovalent complexes with prekallikrein and factor XI, other members of this system. We have determined the molecular weight of human kininogen light chain by equilibrium sedimentation in buffers of differing density, using an air-driven benchtop ultracentrifuge. The resulting molecular weight (30,500 +/- 800 g/mol) and partial specific volume (0.660 +/- 0.008 ml/g) are consistent with the idea that a sizeable fraction of the carbohydrate of high-molecular-weight kininogen is associated with the light chain. This level of precision is relatively easy to attain. The procedures are detailed, along with expressions for error propagation, to permit ready application of the technique.     

Partial molar volumes of proteins: amino acid side-chain contributions derived from the partial molar volumes of some tripeptides over the temperature range 10-90 degrees C

The partial molar volumes of tripeptides of sequence glycyl-X-glycine, where X is one of the amino acids alanine, leucine, threonine, glutamine, phenylalanine, histidine, cysteine, proline, glutamic acid, and arginine, have been determined in aqueous solution over the temperature range 10-90 degrees C using differential scanning densitometry . These data, together with those reported previously, have been used to derive the partial molar volumes of the side-chains of all 20 amino acids. The side-chain volumes are critically compared with literature values derived using partial molar volumes for alternative model compounds. The new amino acid side-chain volumes, along with that for the backbone glycyl group, were used to calculate the partial specific volumes of several proteins in aqueous solution. The results obtained are compared with those observed experimentally. The new side-chain volumes have also been used to re-determine residue volume changes upon protein folding.     

Phospholipid-binding properties of bovine factor V and factor Va.

Factor V and factor Va binding to single bilayer phospholipid vesicles was investigated by light-scattering intensity measurements. This technique allows the measurement of free and phospholipid-bound protein concentrations from which equilibrium constants can be obtained. As controls, the Ca2+-dependent phospholipid binding of prothrombin and factor X were also studied. The average values obtained for the dissociation constants (Kd) and lipid to protein ratio at saturation, moles/mole (n), for prothrombin (Kd = 2.3 X 10(-6) M, n = 104) and factor X (Kd = 2.5 X 10(-6) M, n = 46) binding to vesicles containing 25% Folch fraction III and 75% phosphatidylcholine in the presence of 2 mM Ca2+ were in agreement with those reported in the literature. The average factor V and factor Va values for the dissociation constants and lipid to protein ratio at saturation (moles/mole) were Kd = 7.2 X 10(-8) M and n = 270 for factor V and Kd = 4.4 X 10(-7) M and n = 76 for factor Va. In contrast to prothrombin and factor X, factor V and factor Va demonstrated Ca2+-independent lipid binding. In addition, the number of factor V and factor Va molecules bound per vesicle was found to be dependent both on the phosphatidylserine content of the vesicle and the ionic strength of the buffer.     

The interaction of calmodulin with erythrocyte membrane proteins

The method of sedimentation equilibrium in an air-driven ultracentrifuge (Airfuge) has been employed to investigate the interaction of 125I-calmodulin with the cytoskeletal components of the human red cell membrane. The results indicate significant calcium-dependent calmodulin binding activity in the low and high ionic strength extracts of the human erythrocyte membrane. The interaction of 125I-calmodulin with the low ionic strength extract proteins is analysed quantitatively. Further purification of the high ionic strength extract comprising mainly band 2.1 and band 4.1 results in the elution of calmodulin binding activity in a purified fraction of band 4.1.     

Physical characteristics of human transferrin from small angle neutron scattering.

The technique of small angle neutron scattering has been used to determine the molecular shape, the volume, and the molecular weight of pooled human transferrin in an aqueous solution isotonic with blood. Analysis of the measurements assuming a spheroidal molecular shape indicates that an oblate spheroid with semi-axes of length 46.6 +/- 1.4, 46.6 +/- 1.4 and 15.8 +/- 3.8 A, and a molecular volume of (144 +/- 45) X 10(3) A3 is the best simple approximation to the shape of the transferrin molecule. The radius of gyration, Rg, determined from a Guinier plot is 30.25 +/- 0.49 A, in agreement with Rg calculated for the oblate spheroidal shape. The molecular weight is determined to be (75 +/- 5) X 10(3). The shape-independent molecular volume is found to be (98 +/- 10) X 10(3) A3. The difference in the two volumes suggests that transferrin is not a uniform spheroid but may have a more complex shape.     

The amino acid sequence of plastocyanin from spinach. (Spinacia oleracea L.).

The amino acid sequence of spinach (Spinacia oleracea L.) plastocyanin was determined. It consists of a single polypeptide chain of 99 residues and has a sequence molecular weight of 10415. The sequence was determined by using a Beckman 890C automatic sequencer and by the dansyl--phenyl isothiocyanate analysis of peptides obtained by the enzymic digestion of purified CNBr fragments. Overlap through the two methionine residues was not shown. Sedimentation equilibrium in the ultracentrifuge gave a molecular weight for spinach plastocyanin of about 9000, in contrast with the value of 21000 reported previously by Katoh et al. (1962).     

The self-assembly of collagen molecules

The aggregation of native acid-soluble collagen (N-ASC) and of pronase-treated acid soluble collagen (P-ASC) was examined in solution under conditions which varied from those of minimum collagen-collagen interaction to those leading to incipient fiber formation. Molecular weights and weight distributions were determined in the analytical ultracentrifuge using the Yphantis high speed sedimentation equilibrium and Aarchiblad approach-to-equilibrim techniques. The aggregation was pH and ionic strength dependent in each case. Under conditions of minimum aggregation (low pH, low ionic strength), N-ASC showed the presence of permant aggregates. At higher pH and ionic strength, a higher fraction of aggregate was formed but these were of the same charcter and molecular weight as the permanent aggregates. The aggregates were of a single molecular size, with a weight of 1.5 × 10⁶ daltons, compared with a monomer collagen weight of 3.1 × 10⁵ daltons. The P-ASC formed aggregates also but to a much lower extent and the maximum aggregate size corresponded to dimers in molecular weight. These data show the major importance of molecular end-regions in collagen aggregation to form native type fibers and, by virtue of the discrete size of the N-ASC aggregates, support the microfibrillar hypothesis for the assembly of collagen fibrills.     

Purification and properties of beta-galactosidase from fungus, Curvularia inaequalis]

Highly purfied beta-galactosidase from fungus Curvularia inaequalis cultural fluid with a specific activity of 50 units per mg of protein was obtained by 2-fold purification of the enzyme, using chromatography on DEAE-cellulose and on hydroxylapatite. The enzyme was found to hydrolyze o-nitrophenyl-beta-D-galactopyranoside (pH optimum of 3.7--4.5) and lactose (pH optimum 3.9--5.3). The isoelectric point was observed at pH 4.4 the temperature optimum was 60 degrees C. The molecular weight (115 000--126 000) and the amino acid composition of the enzyme were determined. Km values for o-nitrophenyl-beta-D-galactopyranoside and lactose were 0.55-10(-3) M and 4.5-10(-3) M respectively. Disc-electrophoresis in polyacrylamide gel revealed a single band with a specific activity. The homogeneity of the enzyme was found in ultracentrifuge.     

The thermal depolymerization of porcine submaxillary mucin

The time dependence of the molecular weight, radius of gyration, and hydrodynamic size distribution for porcine submaxillary mucin (PSM) in solution have been studied using static and dynamic light scattering. The weight average molecular weight (Mw) of PSM in 6 M guanidine HCl, pH 7, is initially 3 X 10(6) and decreases with time in three phases: rapidly from 3-2 X 10(6), less rapidly from 2-0.9 X 10(6), and slowly below 0.9 X 10(6). The rates of decrease are much greater at pH 2. The energy of activation associated with each phase is 20 kcal/mol, which is similar to that reported for peptide bond cleavage at an aspartic acid residue. Addition of mercaptoethanol to PSM in 6 M guanidine HCl leads to a rapid decrease in Mw to 0.9 X 10(6), followed by a very slow further decrease. These results suggest that native PSM consists of subunits (Mw = 0.9 X 10(6] that are linked by disulfide bonds to form dimers (Mw = 2 X 10(6] and then higher aggregates. This cross-linking appears to occur at unglycosylated regions of the protein core, which are believed to be richer in aspartic acid than the rest of the molecule.     

Molecular weight and hydrodynamic properties of a proteinase A inhibitor from baker's yeast.

The molecular weight of the proteinase A inhibitor IA3 from baker's yeast was determined by different methods. From gel-filtration experiments, a molecular weight of 19 000 was calculated for the native inhibitor, while under denaturing conditions a molecular weight of 7400 was found. From electrophoretic experiments with the native protein, a molecular weight of 9000 was calculated. A similar value was obtained from the analytical ultracentrifuge, even at a protein concentration of 12 mg/ml. The diffusion coefficient and the partial specific volume were measured and from these data the frictional ratio and the Stokes radius were calculated. These parameters indicate that the relatively high apparent molecular weight calculated from the gel-filtration experiments is caused by the assymetric shape of the inhibitor molecule rather than by an aggregation of subunits.     

Multiple forms of human gamma-glutamyltransferase: preparation and characterization of different molecular weight fractions

Eight fractions of human gamma-glutamyltransferase were prepared from liver tissue, serum and bile by gel filtration. Bile, pooled serum from patients with high gamma-glutamyltransferase activities and serum in which liver tissue had been incubated, each contained an enzyme fraction with molecular weight greater than 10(6). A fraction of about 80,000 molecular weight was obtained from bile, and by incubation of liver tissue in serum or sodium chloride solution, but not from the serum pool. The main enzyme fraction in native serum had a molecular weight of about 300,000, and the molecular weight of gamma-glutamyltransferase partially purified from liver was initially 160,000. The fractions had similar Km and Ki values, and differences in heat stability and binding to concanavalin A were not marked.