III. HIGH MOLECULAR WEIGHT SYSTEMS THE PHYSICQCHEMICAL EXAMINATION OF THE CONARACHIN FRACTION OF THE GROUNDNUT GLOBULINS (ARACHZS HYPUGAEA) BY P. JOHNSON AND W. E. F. NAISMITH* Department of Colloid Science, The University, Cambridge Received 17th April, 1952 Preliminary results are reported on novel association-dissociation reactions occurring in the conarachin fraction of the groundnut globulins. It has been shown by ultra- centrifugation and light scattering that the dissociated state, corresponding to a molecular weight of ca. 190,000, is favoured by high ionic strength and high pH. With lowered ionic strength at constant pH, or lowered pH at constant ionic strength, the degree of association increases, at least four new species, with molecular weights rising as high as 2 x 106, appearing.No indication of the occurrence of these changes was obtained from electrophoresis, so that electrophoretic mobility seems to be independent of degree of association. Light-scattering measurements demonstrated that equilibrium was attained, after change in ionic strength, within 1 min. The globulins of the groundnut (Arachis hypugaea) were first investigated by Ritthausen 1 who extracted the proteins from the oil-free groundnut meal with aqueous sodium chloride and weakly basic solutions, and precipitated them by acidification. He considered the solids so obtained to be identical. Johns and Jones 2 showed that the globulins extracted with aqueous sodium chloride could be separated into two fractions, which they named arachin and conarachin, by an ammonium sulphate fractionation of the saline extract.Jones and Horn 3 later showed that arachin could be obtained from a 10 % sodium chloride extract by dilution or dilution to permanent cloudiness followed by saturation by carbon dioxide, or by 40 % saturation with ammonium sulphate. Conarachin, which was the more soluble fraction, could not be separated by dilution but could be prepared from the filtrate of the arachin precipitation by dialysis against water or complete saturation with ammonium sulphate. Irving, Fontaine and Warner,4 who carried out an electrophoretic investigation of the groundnut proteins in ammonia buffer (ionic strength I = 0.1, pH = 9.26), showed that at least 3 and probably 4 protein components were present, and that arachin and conarachin each consisted of at least 2 of these components.Because the major components of arachin and conarachin had the same electrophoretic mobilities they were regarded by these authors as being identical. Johnson 5 showed, from an ultracentrifugal examination of the proteins, that conarachin, prepared by dialysis of the supernatant liquid after precipitation of the arachin with ammonium sulphate, gave in the ultracentrifuge an ill-defined sedi- mentation diagram for which the sedimentation constant was approximately 10.927. However, the insolubility of the material and its polydisperse character indicated that during dialysis the protein had become altered and partly denatured. Conarachin, prepared by saturation with ammonium sulphate gave small quantities of slower sedimenting materials of s values 4.8 and 6.2s.Johnson, however, did * Present address-I.C.I. Ltd., Nobel Division, Dumfries Factory, Drungans, Dumfries. 98P. JOHNSON AND W. E. F . NAISMITH 99 not definitely apply the name " conarachjn " to any of the above sedimenting species. The present work is concerned with the physicochemical investigation of con- arachin, ultracentrifuge, light scattering, and electrophoretic techniques being employed. It is shown that conarachin contains at least 2 different molecular species ; one of these is very susceptible in its state of aggregation to differences in external environment. In this connection and in its solubility behaviour conara- chin is significantly different from arachin though in other properties surprising similarities are revealed.The various environmental factors which influence the state of aggregation of the conarachin proteins have been investigated and are discussed together with other association-dissociation systems. EXPERIMENTAL APPARATUS.-The ultracentrifugal measurements were made using a Phywe air-driven instrument equipped with a Philpot 6 diagonal Schlieren optical system. In view of con- siderable temperature rise during sedimentation velocity determinations, rotor tempera- tures were recorded throughout and a correction factor evaluated to reduce sedimentation constants to the viscosity and density of water at 20" C . For peaks of reasonable area and resolution, such S& values (given in Svedberg units) are accurate to f 2-3%. Since, in this work, the different components observed are in the main the different aggregated states of a single protein, the relative areas quoted give a useful indication of relative concentrations. In view of other approximations, no correction for dilution in the sectoral ultracentrifuge cell has been applied to area measurements.Peaks are numbered 1, 2, 3, . . . beginning with the most rapid. It should be noted that in the single sedimentation diagrams reproduced here, it is not always possible, owing to the large range of s values, to include all observed components. Diagrams are chosen at a stage of sedimentation which gives resolution of the maximum number of components. The apparatus and methods of Goring and Johnson 7 were used for the measurements of light scattered at 90" to the incident beam and over a GO" arc around this direction.The temperature was 18 f 2" C. The solutions were prepared dust free by ultrafiltration through nitrocellulose membranes as described by Goring and Johnson.8 Using this technique, dissymmetry values as low as 1.04 were reproducibly obtained. Electrophoretic measurements were made in a Tiselius electrophoresis apparatus manufactured by L.K.B. Produkter, Fabriksaktiebolag, Stockholm, and provided with the Svensson modification of the diagonal Schlieren optical system. Changes of ionic strength were routinely performed by dialysis of the protein solution in a slowly rotating Cellophane bag against a large volume of the required buffer solution. Osmotic equilibrium was achieved in ca.1 h and 3 or 4 changes of solvent were always carried out. MATERwLs.-Buffer salts were A.R. or of equivalent purity. The groundnut meal from blanched Spanish Bunch nuts, defatted by extraction with petrol, was supplied by T.C.I. Ltd. (Nobel Division). PREPARATION OF PROTEINS.-The groundnut meal was extracted at room temperature for 4 to 5 h with a 40 % saturated solution of ammonium sulphate. According to previous work3 this should leave the arachin and extract only the conarachin. The latter was pre- cipitated from the extract which had been filtered through paper pulp, by saturating to 85 % with solid ammonium sulphate. No more protein was precipitated by further addition of ammonium sulphate and tests for protein by trichloracetic acid and biuret were negative.The precipitate after redissolving in 40 % saturated ammonium sulphate and reprecipitating at 85 % saturation (40/85 fraction) was stored under 85 % saturated ammonium sulphate. In most of the experiments described, the fractions precipitated over the ranges 40-65 % (40/65) and 65-85 % saturation with ammonium sulphate (65/85) were used. According to the above statements, conarachin is precipitated between 40 and 85 "/, saturation with ammonium sulphate. However, Jones and Horn 3 defined conarachin as the fraction precipitating from 10 % NaCl solution on adding ammonium sulphate to between 40 and 85 % saturation. It has, however, been confirmed that the presence of 10 % NaCl does not appreciably affect the fractionation limits so that the two definitions of conarachin are equivalent.100 GROUNDNUT GLOBULINS Jones and Horn3 state that after the arachin was removed at 40 % saturation with ammonium sulphate from a 10 % aqueous NaCl extract of the meal, there was no further precipitation on the addition of ammonium sulphate until a fraction separated at between 70 % and 80 % saturation.We found that a fraction precipitated (from either a 10 % saline extract after removal of the arachin, or from a 40 % saturated ammonium sulphate extract) at between 40 % and 65 % saturation with ammonium sulphate, and another fraction between 65 % and 85 % saturation. The fractionation, however, was by no means sharp, as will later be seen. Arachin was prepared by precipitation from a 10 % saline extract of the groundnut meal by 40 % saturation with ammonium sulphate and purified by several similar repre- cipitations.drying to constant weight at 120" C for 4 h, conarachin was found by the micro Kjeldahl method to contain 18.2 % nitrogen (cf. 18.3 % quoted by previous workers 3). Using the value 18.2 % the results of fig. 1 were obtained by interferometer measurements and DETERMINATION OF PROTEIN CONCENTRATION AND PARTIAL SPECIFIC VOLUME.-After FIG. 1.-a n (from interferometer) against c for conarachin. were used for all concentration determinations. The partial specific volume V was calculated from the equation,g 1 - - w dm (1 - V p ) = - - m dw' where p = density of the solution for which the solute weight fraction is w, and m = mass of a given volume of the solution. m was determined with a pyknometer for a range of protein concentrations.From fig. 2 in which m is plotted against w, the value ij = 0.72 0.004 was obtained. Absence of curvature indicates the constancy of this value over the concentration range covered. It is of interest to note that the more insoluble globulin from the groundnut, arachin, also has the value 6 = 0-72 f 0-005.5 the different conarachin preparations in phosphate buffer, I = 0.5 and pH 7-5. Three IN PHOSPHATE BUFFER (0.032 M Na2HP04 . 12H20, 0.003 M KH2P04, 0.04 M NaCI) AT ULTRACENTRIFUGE RESULTS : FRACTIONATION.-Tabk 1 summarizes the results on TABLE 1 .-SEDIMENTATION AND DERIVED DATA FOR THE DIFFERENT CONARACHIN FRACTIONS I = 0.5, pH = 7.5 ; PROTEIN CONC. = 0.8 g/100 ml wt.-average mol. wt. peak 1 peak 2 peak 3 --I___ --__--I_ __ _._.__ S s area .? a:7a (calc.) area fraction 0' % 0 A 40185 15 16 8.7 44 2 40 150,000 B 40/65 14 20 - I 2 80 90,000 C 65/85 18 4 8.7 78 2 18 175,000P . JOHNSON AND W . E . F. NAISMITH 101 main components occur in the 40185 fraction, mainly two occurring in the other fractions. The slow component of s w 2s occurs largely in the 40/65 fraction whilst the prominent peak of s = 8.7 is largely precipitated in the 65/85 fraction. Most of the work here discussed is concerned with the latter material. Although the s2 component appears to be inert compared with s8.7 (see later) it is desirable to remove it completely from the fraction but no simple method has yet been devised. On lowering the ionic strength of the 40/85 fraction to I = 0.1, pH = 7.7, a profound change occurred in the sedimentation pattern, the following sio values being observed : (1) 30, (2) 20, (3) 12.5 and (4) ca.2, with the respective percentage peak areas (1) 11 (2) 37 (3) 19 and (4) 33. Apparently the s2 component remains unaltered but the other components disappear completely, new ones appearing. Thus definite evidence of asso- ciation at low ionic strengths was obtained. On adding an equal volume of arachin solution, also at I = 0.1, pH = 7-7, of equal total protein concentration, the resulting sedimentation pattern was in good agreement with that expected from the addition of the separate patterns, arachin appearing as a separate component with sio = 14.2 (cf. sio = 14.6 for arachin alone at Z = O.l).lO Thus distinguishability from conarachin was proved as well as the lack of appreciable interaction effects between arachin and con- arachin components.In a similar experiment with the 65/85 fraction at I = 0.5, pH = 7.5, sedimentation patterns were again apparently additive (fig. 3), arachin appearing as a new w x / 0 2 FIG. 2.-m (mass of a given volume of solution of conarachin for which the solute weight fraction is w) against w. component with s;o = 13.5 to be compared with sio = 13-3s for sedimentation alone at this ionic strength.7 Thus arachin and conarachin are separate and distinguishable sys tems. Since the 65/85 fraction contains a high proportion of the labile s;., component with much smaller amounts of s2 and s15 than the other fractions, further work was concen- trated on it. The ~ 1 5 and 32 components have not, as yet, received any further direct study.Since diffusion data are not available for the different components, molecular weight calculation depends on some type of assumption concerning frictional ratios (flfo). As many seed globulins approximate to ellipsoids of revolution of relatively small asymmetry, fife has been assumed throughout to be 1.3 which allows for a contribution from hydration as well as from asymmetry (e.g. 30 % hydration, axial ratio 4). Table 2 contains observed sedimentation constants and molecular weights calculated thus, both sets of values being suitably rounded. Using these figures and relative proportions as indicated by peak areas, weight average molecular weights were calculated for the fractions and are included in table 1 and elsewhere. * A species of sedimentation constant 8.7 Svedberg units is for convenience termed the sg.7 component.102 GROUNDNUT GLOBULINS 65/85 FRACTION : EFFECT OF VARIATION OF IONIC STRENGTH.-At Constant pH (7.6 & 0.2) and protein concentration (0.8 g/100 ml) the ionic strength of a solution of the 65/85 fraction was varied by stages from 0.5 to 0-025 by dialysis.Sedimentation diagrams and constants are shown in table 3 and calculated weight-average molecular weights and relative areas of the various components in table 4. Since it was not possible to use the same preparation for all experiments, some variability in peak areas is to be expected from fractionation differences. Thus the areas of ~ 1 2 . 5 and s20 components at I = 0.1 TABLE 2.4BSERVED SEDIMENTATION CONSTANTS AND MOLECULAR WEIGHTS CALCU- LATED ASSUMING f/fo = 1 *3.0 820 2 9 12.5 14 20 30 50 (Svedberg units) mol. wt. 19,000 190,000 290,000 380,000 560,000 1 - 1 x 106 2.2 X 106 TABLE 3.-sEDIMENTATION DIAGRAMS AND CONSTANTS * OF THE 65/85 CONARACHJN FRACTION AT pH = 7-6 ( f 0.2) AND PROTEIN CONCENTRATION = 0.8 g/100 ml buffer composition (M) NazHPO KH2P04 NaCl I sedimentation constants 8.7 2 - A 0032 0.003 0.40 0-50 18, B 0-032 0.003 0.25 0.35 20, 9.3 2 C 0016 04015 0.15 0-20 30, 19, 12.6, 2 - - D 0.016 0.0015 0.05 0.10 32 21, 12-5, 2 . _ E 0.008 0.0008 - 0.025 33 20, 12.8, 2 - * any component occurring > 30 % (by area) underlined. TABLE 4.-wEIGHT-AVERAGE MOLECULAR WEIGHTS AND RELATIVE AREAS (%) FOR THE DIFFERENT SEDIMENTING COMPONENTS OF THE 65/85 FRACTION AT pH = 7.6 (4: 0.2) AND PROTEIN CONCENTRATION = 0.8 g/100 ml 330 s20 s12.5 s9 sz wt.- average mol.wt. (calc.) 0.50 - 4 - 78 18 175,000 0.35 - 4 - 84 12 185,000 0.20 4 43 40 - 13 405,000 0-10 4 60 23 - 13 450,000 0.025 3 54 31 - 12 430,000 were somewhat variable even under identical conditions as far as could be judged. How- ever, tables 3 and 4 show clearly the disappearance of S8.7, with lowered ionic strength, the appearance or growth of ~12.5, s20 and to a smaller extent 830, and the relative constancy of ~ 2 . At the lowest ionic strength, sedimentation constants and patterns may be subject to charge effects ; the slight indication of smaller degree of association at I = 0-025 than at I = 0-1 is not therefore necessarily significant.At higher ionic strength than 0.5, no further change in pattern was observed. Weight-average molecular weight values indicate clearly the occurrence of increased association with lowered ionic strength. Replacement of NaCl by Na2S04 at I = 0.5 did not introduce any sedimentation changes ; nor did the addition to a similar solution of CaCl2 to M/1000 or HgC12 to 5 x 10-4 M (i.e. 1 atom of Hg/2 moles of protein) give any sign of change in degree of association (cf. Hughes 11). Since dialysis occurs over several hours, in order to estimate the speed of the processes involved, experiments were performed in which the ionic strength was raised from 0.1 to 0.5 by addition of solid NaCl and lowered from 0.5 to 0.1 by diluting a concentratedA B C FIG.3.-Sedimentation dia- grams of: ( a ) arachin, (b) 65/ 85 conxachin, (c) mixture of (a) and (b) ; phosphate buffx I=0*5, pH=7.5 ; total protein conc. * 0.8 g i l 0 0 ml. A A B B C D C E TABLE 1. TABLE 5 . [To fuce page 102 TABLE 3.A B C D E C TABLE 10. F G TABLE 6. [To face page 103P . JOHNSON AND W . E . F . NAISMITH 103 protein solution with water. In both cases, ultraceiitrifuge runs were imniediately per- formed, sedimentation patterns obtained being characteristic of the final ionic strength with no further changes occurring. Thus the changes occurring were complete within the first hour (see later). Whilst the effects of extreme change in ionic strength seemed completely reversible (i.e. 0.1 --'t 0-5 or vice versa) it was observed that if the intermediate ionic strength 0.2 was approached from higher and lower values, some differences in the sedimentation patterns persisted even after prolonged dialysis.This result is not understood and may indicate the importance of some factor as yet overlooked. EFFECT OF PROTEIN coNcmTRATIoN.-In experiments reported above the protein con- centration was kept constant since it was conceivable that variations in it would affect the equilibrium position. No observable change in the proportions of the sedimenting components was observed at I = 0.5, pH = 7.5, on varying the protein concentration from 0.4 to 2.5 g/100 ml, but at I = 0.1, pH = 7.7, it was found that increase in protein concentration tended to favour greater association (table 5). The constancy of the area of the s2 component is again noticeable, whilst increased total protein concentration causes a transfer from the ~ 1 2 .5 area to that of s20. A simiIar conclusion is reached from the increase in weight-average molecular weight with protein concentration. VARIOUS CONCENTRATIONS IN PHOSPHATE BUFFER : (Na2HP04 . 12H20,0.016 M ; KH2P04, TMLE 5.-sEDIMENTATION AND DERIVED DATA FOR THE 65/85 CONARACHIN FRACTION AT 0.0015 M ; NaCl, 0.05 M) I = 0.1, pH = 7.7 wt .-average mol. wt. peak 1 peak 2 peak 3 peak 4 .__I__ ~ _ _ _ _ protein concentra- tion area area area area (calc.) A 2-10 - * ca. 2 22 74 11-5 12 2 12 475,000 (g/100ml) (yo) (%> (Yo) (70) B 0.85 36 4 23 60 12-4 23 2 13 450,000 C 0.43 33 4 20 56 12.3 28 2 12 440,000 * not measurable because of ill-defined peak.EFFECT OF pH vARIA-rIoN.-Solutions of 65/85 conarachin at constant protein concen- tration (0.8 g/100 ml) and ionic strength (0-l), were prepared at various pH values by dialysis from I = 0.1, pH =I 7.7, and examined in the ultracentrifuge. Summarized data is presented in tables 6 and 7. At pH values below 4 irreversible changes occurred since on dialyzing back to pH7.7, considerable precipitation of protein took place and the soluble protein now gave an altered sedimentation pattern. On the other hand, re- dialysis to I = 0.1, pH = 7.7 after exposure to pH's above 6 gave the characteristic pattern for these conditions. With increasing pH above 6, the constancy of the s2 area was again shown, and there was a gradual transfer from the more rapid to the slower sediment- ing components which was confirmed by weight-average molecular weight calculations. At I = 0.1, pH = 10.3, the sedimentation pattern was very similar to that at Z = 0.5, pH = 7.5, and similar weight-average molecular weights were observed.Thus high pH and high ionic strength affect the degree of association of the protein very similarly. The main peak at pH9.4 was noticeably broader than those obtained at other pH values and the s& value, 14.7, seemed to fall between those of the two well-defined species ~ 1 2 . 5 and s20. Whilst these facts have not been further investigated, it seems possible that some rapid reversible type of reaction (e.g. 2~12.5 + s20) is involved. At I = 0.5, pH variation within the range 9-4-59 had little or no effect on the ultra- centrifugal pattern.Only when the pH was reduced to 4.7 (which is approximately the isoelectric point of conarachin at I = 0.5) was considerable aggregation observed, the pattern in table 6 being obtained. At I = 0-35, lowering of the pH from 7.6 to 6.1 resulted in only slight association. Thus apparently the solvent action of the aqueous medium at I = 0-5 is so good that only the discharge of the protein at or near the isoelectric point affects the state of association. On the other hand, at lower ionic strengths the weaker solvent action is shown by the higher state of association at a given pH, and by the effect on association of small pH changes. EFFECT OF TIME.-NO change in the ultracentrifugal pattern of a solution of the 65/85 conarachin in phosphate buffer at I = 0.5, pH = 7.5 was observed with time even after104 GROUNDNUT GLOBULINS 3-4 weeks.At I - 0.1, pH =: 7.7, little change was observed over a period of two weeks but thereafter gradual dissociation occurred until after 3-4 weeks a pattern similar to that at I = 0.5 was observed. This result, which was confirmed, was most surprising, and is difficult to reconcile with the definite reversible changes in association which occur on changing the ionic strength. Its occurrence with old and often obviously infected protein solutions suggests the effects of microbiological attack. TABLE 6.-sEDIMENTATION DIAGRAMS AND CONSTANTS * OF THE 65/85 CONARACHIN FRACTION AT DIFFERENT pH VALUES I = 0.1, protein conc. = 0.8 g/100 ml PH _______-____ ~ -__ -- buffer composition (M) sedimentation constants - -_ gl ycine 0.09 NaOH 0.0 1 A NaCl 0.0098 10.3 21, gly cine 0.094 NaOH 0.0056 B NaCl 0.014 9.4 9.4 2 14.7 2 - Na2HP04 .12H20 0.01 6 C KH2P04 0.0015 7.7 32, 21, 12.5 2 - NaCl 0.05 Na2HP04. 12H20 0.0086 NaCl 0.05 D KH2P04 0.024 6.3 50, 32, - CH3COONa 0.05 NaCl 0.05 29, 17, - E CH3COOH 0.23 3-8 glycine 0.01 6 HCI 0.004 F NaCl 0.096 2.9 I = 10-5, protein conc. = 0.74 g/100 ml 12.3 2 2 2 2 CH3COONa 0.1 NaCl 0.4 0.0048 4.7 46, 27, - G CH3COOH 11.7 2 * any component occurring > 30 % (by area) underlined. DIFFERENT SEDIMENTING COMPONENTS OF THE 65/85 FRACTION AT DIFFERENT pH VALUES I = 0.1, protein conc. = 0.8 g/100 ml TABLE 7.-wEIGHT-AVERAGE MOLECULAR WEIGHTS AND RELATIVE AREAS (%) FOR THE s2 wt.-average mol. PH s50 s30 S20 s l 4 s12.5 $9 wt.(calc.) 10.3 - - 5 - - 83 12 190,000 9.4 - - - 87 - - 13 330,000 7.7 - 4 60 - 23 - 13 450,000 6.3 5 70 - - 13 - 12 900,000 3.8 - 3 42 - - 39 16 345,000 2.9 - - - 30 - - 70 130,000 I = 0.5, protein conc. = 0 . 7 4 g/100 ml 4.7 8 58 - - 21 - 13 880,000P. JOHNSON AND W . E . F . NAISMITH 65/85 and 40/65 fractions of conarachin in phosphate buffer, Z = 0.5, pH = 7-5, has given weight-average molecular weights of 190,000 and 70,000 respectively to be compared with 175,000 and 90,000 from ultracentrifuge calculations. In view of the approximations made and the fact that light scattering and ultracentrifuge work was carried out on different protein preparations, the agreement in molecular weight values is as good as could be expected. C/T against c (c = concentration in g/cm3) for the 65/85 fraction in phosphate + NaCl buffers at pH 7.6 0-2 over a range of ionic strengths.Each plot gives a horizontal 105 LIGHT SCATTERING EXAMINATION.-EXam~atiOn by light Scattering Of SOlUtiOnS Of the 65/85 FRACTION : EFFECT OF VARIATION IN IONIC STRENGTH.-Fig. 4 Contains plots Of i c x 10' (p +') 2 2 4 6 8 I 1 /o FIG. 4 . 4 7 against c for the 65/85 conarachin fraction at pH 7.6 ( + 0.2) and different ionic strengths. 0 I = 0.5. 0 I = 0.35. A Z = 0.2. I = 0.1. straight line and the intercepts on the C/T axis were used to calculate the weight-average molecular weights of table 8, column 2. Satisfactory agreement with calculated values from ultracentrifuge data is obtained. It was further shown that changes produced were readily and quantitatively reversible at the extremes of the ionic strength range, but at TABLE 8 .-CHANGE IN WEIGHT-AVERAGE MOLECULAR WEIGHT OF THE 65/85 FRACTION WITH IONIC STRENGTH AT pH = 7.6 (& 0.2) weight-average molecular weight - calculated from ultracentrifugal data Z light scattering 0.50 190,000 175,000 0.35 210,000 185,OOO 0.20 395,000 405,000 0.10 550,000 450,000 I = 0-2 the molecular weight value again depended to a small extent upon whether the solution had been previously at a higher or lower ionic strength.To estimate the rate at which association or dissociation occurred, rapid changes in ionic strength were caused by addition of water or concentrated sodium chloride solutions (carefully filtered) to the protein solutions. Light scattering measurements taken within 1 min of such additions corresponded with the final jonic strengths and no further change was observed. D106 GROUNDNUT GLOBULINS EFFECT OF pH VARIATION.-Fig.5 contains plots of C/T against c for the 65/85 fraction in buffers at I = 0.1 and various pH values, horizontal straight lines again being obtained. Table 9 contains derived weight-average molecular weights compared with calculated ultracentrifugal values. Reasonable agreement is again obtained, and as previously suggested on the basis of ultracentrifuge patterns, the weight-average molecular weight at I = 0.1, pH = 10.3, agrees well with that for I = 0.5, pH = 7-5. Within the pH range covered, the changes observed were readily reversible. It was shown by ultracentrifuge examination that increased protein concentration at I = 0.1, pH = 7.7, caused greater association and higher weight-average molecular weights, and this effect might be expected to modify the horizontal C / T against c plots expected for a moderately charged protein in solution at appreciable salt concentration.However, the concentration effects observed occurred largely at above 1 % protein con- centration, whilst light scattering data was obtained at lower values where any curvature FIG. 5.+/7 against c for the 65/85 conarachin fraction at I = 0.1 and different pH values. 0 pH = 10-3. A pH = 9.4. pH = 7.7. El pH = 6-3. due to altered degree of association would be very small. No such curvature is detectable in the light scattering results reported here. TABLE g.-CHANGE IN WEIGHT-AVERAGE MOLECULAR WEIGHT OF THE 65/85 FRACTION WITH pH AT IONIC STRENGTH = 0.1 _____ weight-average molecular weight calculated from ultracentrifugal data PH light scattering 10.34 206,000 190,000 9.43 290,000 330,000 7-65 548,000 450,000 6.29 1,040,000 900,000 ELECTROPHORETIC ExAMmATIoN.-Only a brief outline of electrophoretic results is given here, a more detailed report being planned.19 Electrophoretic diagrams at 20" C of the 40/85,40/65 and 65/85 conarachin fractions in phosphate buffer at I = 0.1, pH = 7.7 are shown in table 10 with mobility and peak area data calculated from the descending patterns.From the mobilities it seems likely that the same 2 components occur in all fractions though in differing amounts. Comparing the 65/85 electrophoretic pattern with that from sedimentation at I = 0.1, pH = 7.7, considerable differences are revealed but there is some similarity with theP .JOHNSON AND W. E. F . NAISMITH 107 sedimentation pattern at I = 0.5, the major electrophoretic component occurring in proportions similar to the sum of the 2 faster sedimenting components. Since the latter appear to be merely different stages of association of the same chemical entity, it seems legitimate to group them together and the conclusion emerges that such association is TABLE 1 ELECTROPHORETIC AND DERIVED DATA FOR DIFFERENT CONARACHIN FRACTIONS IN PHOSPHATE BUFFER (I = 0.1, pH = 7.7) AT 20" C; PROTEIN CONC. = 0.8 (& 0.1) g/100 ml major component minor component -- fraction 7; mobility* area (Yo mobility * A 40185 10.9 79 5-1 21 B 40165 10.7 72 5.2 28 C 65/85 10.8 92 5.2 8 * units : cmz volt-1 sec-1 x l o 5 not detected by electrophoresis.Similarly on varying the pH between 6.5 and 9 and the ionic strength between 0.05 and 0.2 no change in the relative areas of the electrophoreti- cally different components occurred. Thus apparently the very different components observed by the ultracentrifuge have either identical electrophoretic mobilities or are so similar in this respect that no separation occurs in electrophoresis. DISCUSSION Although an ideal separation of protein constituents was not obtained, the 65/85 fraction contained only ca. 10 % of s2 component and ultracentrifuge ex- periments (especially tables 4, 5, 7) demonstrate clearly the inertness of this com- ponent in the reactions being studied.A comparison of sedimentation diagrams for the 40/65 fraction at I = 0.1 and 0-5 confims this point further. The reactions involve only the S8.7 component in the most highly dispersed form of the fraction from which it has been shown that more rapidly sedimenting species only form under a variety of other conditions. From the discrete nature of the new sedi- menting peaks and the constancy of their sedimentation constants over an ap- preciable range of conditions, the phenomena seem, more in the nature of well- defined processes than irregular colloidal aggregation or coagulation. For this reason the terms " association " and " dissociation " have been used throughout this paper in preference to " aggregation " and " disaggregation ".However, while the regular nature of the association reactions is to be stressed, it should also be men- tioned that at low ionic strengths and pH values near 6 where association is con- siderable, solubility is low and the association processes may merge into irregular coagulation. Assuming regular association, the reactions occurring may be summarized in the form : 178.7 . S8.7 + n12.5 . ~ 1 2 . 5 + n20 . s20 + n30 . ~ 3 0 , etc., (1) where ns represents the number of molecules ss of sedimentation constant s which take place in the chain of reactions. In the absence of diffusion or other suitable data to make possible the calculation of unambiguous molecular weights, the n values cannot be fixed. Certain probable features may, however, be mentioned.Large changes in frictional ratio by the di- or trimerization of ellipsoidal units of small axial ratio would not be expected so that the relative magnitudes of the molecular weights in table 2 are probably not greatly in error. Thus under con- ditions where association was considerable, molecular weights are in the ratios 1/2/4/8, the lowest value corresponding with ~12.5. The absence of sg under such conditions is perhaps surprising but may be significant in that the molecular weights108 GROUNDNUT GLOBULINS of sg and ~12.5 seem to be in the ratio 2/3. Possibly the onset of appreciable as- sociation requires a preliminary packing of a basic unit (of molecular weight ca. 100,000) into three’s (~12.5) rather than two’s (sg). Association reactions according to eqn.(1) would then remove the trimer as it was formed so that s9 would eventually disappear completely. The effectiveness of high ionic strength or pH in giving maximum dispersion and the rapidity of the dispersion or association processes suggests that electro- static forces are largely responsible for the binding of the smaller units. At the highest pH values studied (10-3), all basic groups except the guanidinium will be almost completely uncharged and the excess negative charge apparently suffices to disperse the protein completely by repulsion of the sg units. As the pH is lowered, +amino groups of lysine become charged positively (the net negative charge now assuming a lower value) and at pH 9.4 and I = 0.1 when this process must be largely complete, it will be recalled (table 6) than an anomalously broad sedimenting boundary was observed.It seems possible that at this stage there is an approximate balance between the electrostatic repulsion of higher pH values and the attraction of lower pH values which takes control when the net negative charge is further reduced. Since at higher ionic strengths such attraction is not effective, it must in part at least be electrostatic. A correct disposition of charged groups on the unit structures wouId make possible such an attraction, but it may also involve Van der Waal’s forces and hydrogen bonds. At lower pH values the attractive forces become more predominant, and the equilibrium moves towards the most rapidly sedimenting species whose boundaries are now of more normal sharpness.Since it has been shown by light scattering that equilibrium is rapidly attained, association and dissociation must occur continuously during ultracentri- fugation, but providing both processes are relatively rapid no noticeable effect on boundary contours would be expected. Nor would observed s values be affected unless the equilibrium position were changed appreciably. The special conditions at pH 9.4, I = 0.1, are probably responsible for the modified contours and sedi- mentation constant observed. To check the above picture it is planned to compare with electrophoretic, analytical and titration data. It is of interest to compare the main features of conarachin with those of other dissociating systems. Dissociation at high pH values seems to occur very generally, being reported for the haemocyanins,l2.13 insulin,l4 thyroglobulin 15 and for arachin,lo the less soluble and quite separate globulin from the groundnut.As to the effect of addition of electrolytes, dissociation or association may occur ac- cording to the protein studied. Dissociation occurs with Helix pomatia haemo- cyanin, whilst association is promoted with haemocyanin from Paludina vivipara.12 The latter effect is probably more general since it has been reported also for arachin,lo insulin,l3 and thyroglobulin.~s As yet no pH effect on reaction velocity, as was observed with arachin, has been detected, but further investigation of this question is required. Nor have any specific ion effects, comparable with those observed for Helix pomatia haemocyanin, been observed.A comparison of sedimentation and electrophoresis diagrams for the 65/85 fraction at I = 0.1 and pH values below 10 reveals striking differences which can only mean that electrophoretic behaviour is largely independent of degree of association. It was shown similarly 16 that the dissociation product of arachin was electrophoretically very similar to the undissociated molecule. Svedberg 17 observed that components of a dissociation would not be distinguishable electro- phoretically but, on the basis of electrophoretic theory, this is not readily explicable as has been pointed out elsewhere.18 The advisability of applying more than one method to the examination of protein systems is clearly indicated. The assumption of identity on the basis of similar electrophoretic mobility at a given pH, as was made by Irving, Fontaine, and Warner,4 is particularly dangerous. The similarity of arachin and conarachin electrophoretically has been confirmed in this work,lg but clear differences in other properties leave no doubt that the proteins are different.P . JOHNSON AND W. E . P . NAISMITH 109 We are grateful to Miss Susan Preger for technical assistance with the ultra- centrifuge and to Messrs. I.C.I. Ltd. (Nobel Division) for seconding one of us (W. E. F. N.) for one year to work on this problem and also for supplies of oil-free ground-nut meal. 1 Ritthausen, Arch. Ges. Physiol., 1880, 21, 81. 2 Johns and Jones, J. Biol. Chem., 1916, 28, 77. 3 Jones and Horn, J. Agric. Res., 1930, 40, 673. 4 Irving, Fontaine and Warner, Arch. Biochem., 1945, 7, 475. 5 Johnson, Trans. Faraday Soc., 1946,42,28. 6 Philpot, Nature, 1938, 141, 283. 7 Goring and Johnson, Trans. Faraday SOC., 1952,48, 367. 8 Goring and Johnson, J. Chem. Soc., 1952, 33. 9 Svedberg and Pedersen, The Ultracentrifuge (Oxford, 1940), p. 58. 10 Johnson and Shooter, Biochim. Biophys. Acta, 1950, 5, 361. 11 Hughes, J. Amer. Cliem. Soc., 1947, 69, 1836. 12 Brohult, J. Physic. Chem., 1947, 51, 206. 13 Pedersen, Cold Spring Harbor Symposia, 1950, 14, 140. 14 Gutfreund, Biochem. J., 1948, 42, 554. 15 Lundgren, Nature, 1939, 143, 896. 16 Johnson, Shooter and Rideal, Biochetn. Biophys. Acta, 1950, 5, 376. 17 Svedberg, Chem. Rev., 1937, 20, 81. 18 Alexander and Johnson, Colloid Science (Oxford, 1949), p. 332. 19 Johnson and Naismith, to be published.