THE JOURNAL OF BIOLOQICAL CHEXII~TRY Vol. 244, No. 18, Issue of September 25, pp. 5074-5080, 1969

Printed in U.S.A.

Observations on Molecular Weight Determinations

on Polyacrylamide Gel*

(Received for publication, May 5,1969)

A. K. DUNKERI AND ROLAND R. RUECKERT~

From the Biophysics Laboratory and Department of Biochemistry, University of Wisconsin, Madison, Wisconsin 53706

SUMMARY

1. A simple internal calibration technique was used to construct accurate molecular weight-mobility prof?les on polyacrylamide gels with well characterized proteins, and the existence of a “critical point” was observed for small polypeptides. The useful size range limits were determined for several gel concentrations. Apparent molecular weights generally fell within 5 to 6% of literature values, although a few “anomalous” proteins, notably ribonuclease, consider- ably exceeded this margin of error.

2. The effects of intrinsic molecular charge and conforma- tion on electrophoretic behavior in the presence of sodium dodecyl sulfate, evaluated by studies on a set of model pro- teins, were found to be small.

3. The results are in accord with a model in which the protein somehow organizes the sodium dodecyl sulfate anions into a micellar complex of definite size. Although the stoichiometry of the complex is apparently governed not only by the size of the molecule, but also by its state of foldedness, the interplay of anion binding and frictional resistance to passage through the gel is such as to produce a relatively constant log size to mobility ratio.

4. The apparent molecular weights of the five polypeptide chains of Mouse-Elberfeld virus were determined.

In 1967, Shapiro, Vinuela, and Maize1 (1) pointed out that the molecular sizes of polypeptides could be estimated from the relative electrophoretic mobilities of their sodium dodecyl sul- fate complexes on poiyacryiamide geld. The method ‘id rapid and versatile, and requires very little sample (often less than 5 pg of each component). In addition, the tools required are economical and demand relatively little laboratory space or specialized training for application. In spite of the apparent utility of the method, there has been relatively little published

* This work was supported by Grant CA-08662 from the United States Public Health Service.

$ Supported by Predoctoral Training Grant GM-00733-08. Present address, Department of Molecular Biophysics, Yale Universit.y, New Haven, Connecticut 06520.

$ United States Public Health Service Career Development Awardee.

information about its limits of reliability. A need for caution was suggested by the above authors, who pointed out that two well known proteins, lysozyme and ribonuclease, exhibited “anomalous” behavior in this system. We report here the re- sults of our studies on over 20 proteins, including the above anomalous proteins and a number of viral proteins. The origin of the anomalous behavior of lysozyme was identified, and rough limits of accuracy were estimated for the method. In addition, some ways for detecting anomalous behavior by unknown pro- teins are suggested.

MATERIALS AND METHODS

Proteins-ME virus was prepared as described previously (2). Carboxypeptidase A, chymotrypsinogen A, lysozyme, and ovalbumin were obtained from Worthington; cytochrome c, ribonuclease A, myoglobin, 7 S y-globulin, chymotrypsin, tryp- sin, and pepsin from Mann; /3-lactoglobulin, insulin, and bovine serum albumin from Pentex, Inc.; R17-virus, tobacco mosaic virus, and bromegrass mosaic virus were gifts from Paul Kae- berg.

The BSA’ preparation which had been recrystallized in the laboratory of Professor R.,M. Bock in 1964 contained at least five electrophoretically distinct components which evidently represent the monomer and its polymeric forms (3, 4). The mobility of the fastest and most abundant component. was that expected for the BSA monomer (mol wt 66,000). The slower components, which travelled as expected for the successive (i.e. dimer, trimer, etc.) allgo-mers of B&4, !argely~ disappeared from the preparation after treatment with mercaptoethanol.

Chymotrypsin B and C chains were obtained from chymo- trypsin by reduction and denaturation according to the usual procedures of sample preparation (see below). The A chain (molwt 1200 daltons (5)), which is theoretically also produced by this treatment, was not detected, evidently because such small polypeptides stain poorly with the procedure used.

Preparation of Samples for Electrophoresis-The proteins, except as noted below, were dissolved at a concentration of about 2 mg per ml in 1 y0 2-mercaptoethanol, v/v, 4 M urea, and about 1% SDS, and then incubated at 45” for 30 to 60 min. Purified whole virus treated in this way was used without removing the RNA.

In certain cases (i.e. insulin, y-globulin, and BSA) where the

l The abbreviations used are: BSA, bovine serum albumin; SDS, sodium dodecyl sulfate.

5074

by guest on April 9, 2019

http://ww

w.jbc.org/

Dow

nloaded from

Issue of September 25, 1969 A. K. Dunker and R. R. Rueckert 5075

dissociation or reshuffling of disultide-linked chains was to be prevented, 2-mercaptoethanol was replaced with 0.002 M iodo- acetamide. The latter reagent was intended to block sulfhydryl groups, adventitious or otherwise, which catalyze disulfide inter- change reactions (6).

Mixed disulfide derivatives of proteins were prepared by incu- bation with 0.1 M solutions of the appropriate disulfide in the presence of catalytic amounts of thiol by a modification of the procedure of Smithies (6). Full details will be described in a forthcoming communication.2 Procedures for the preparation and gel electrophoresis of lysozyme and its mixed disulfide derivatives in 8 M urea have been described by Rueckert and Bock.3

Electrophoresis was carried out with mixtures prepared by combining equal volumes of the protein solutions.

Polyacrylamide Gels-Polyacrylamide gels (5, 10, and 15% final total concentrations) were prepared by mixing the appropri- ate acrylamide solutions with 0.05 volume of 1% ammonium persulfate; columns (6 to 7 cm) were cast in soft glass tubing (inside diameter 6 mm), and the monomer was carefully overlaid with about 0.3 ml of water. For all three concentrations, the acrylamide to methylene-bisacrylamide ratio was 29 : 1 (w/w). Besides acrylamide, the solutions contained 0.1 y0 SDS, 0.1 M so-

dium phosphate, pH 7.2, and O.lO%, v/v, N, N,N’,N’-tetra- methylethylenediamine (7).

After gelation (10 to 15 min at room temperature) a Styrofoam partition (about 6 mm x 10 to 15 mm X 1 mm) was wedged snugly into each tube against the top of the gel column. Each column was then overlaid with 0.1 M sodium phosphate, pH 7.2. Columns prepared in this manner are referred to as “split gels.”

EZectrophoresisThe samples, in dense solutions (4 M urea), were layered under the electrode buffer. Typically, 5 to 30 ~1, containing 5 to 15 pg of each protein, were applied to either side of the partition. Electrophoresis was carried out at 7 to 9 ma per tube (roughly 3 volts per cm) for 2 to 3 hours (5% polyac- rylamide gels), 4 to 5 hours (10% gels), or 6 to 10 hours (1501, gels).

Staining-After electrophoresis, the SDS was leached out and the proteins were precipitated by soaking the gels for 18 to 24 hours in 20% sulfosalicylic acid (7), which was generally more effective for this purpose than 10% trichloracetic acid (8). Next the gels were immersed for 4 to 6 hours in 0.02% Coomassie bril- liant blue R250, freshly diluted in 12.5% trichloracetic acid (8). After decanting the dye, 10% trichloracetic acid was added, and the gels were allowed to stand in faintly blue solutions which further intensified the staining. Gels of 50/, polyacrylamide could usually be photographed within an hour, but lo%, and especially 15%, polyacrylamide gels required 48 hours or longer to develop sufficient contrast for photography.

Measurement of Relative Mobil&es-Stained gels were photo- graphed by transmitted light with a Polaroid MP3 camera and an orange filter as described by Chrambach et al. (8). From the photographs, migration distances were measured, and the con- version to relative mobilities was accomplished by dividing each migration distance by that of chymotrypsinogen. This proce- dure of using a reference protein in each gel automatically cor- rects mobility measurements for changes in gel dimensions due to swelling during the staining operation.

2 C. M. Stoltzfus and R. R. Rueckert, in preparation. 3 R. R. Rueckert and R. M. Bock, from a student manual (1968).

Mimeographed copies are available on request.

OVALBUMIN

CARBOXYPEPTIDASE A

CHYMOTRYPSINOGEN A TRYPSIN

--iam CYTOCHROME C

\

FIG. 1. Electrophoretogram illustrating the split gel technique for determining relative electrophoretic mobility of proteins on SDS-containing polyacrylamide gels. Mixtures to be compared were applied to opposite sides of a divided sample compartment over a column of 10% acrylamide. Electrophoresis (anode at the bottom), staining, and photography were as described under “Ma- terials and Methods.” The migration distances are normalized with respect to that of chymotrypsinogen A. Each band corre- sponds to about 10 pg of protein.

RESULTS AND DISCUSSION

Method-To estimate the molecular weight of an unknown polypeptide, its mobility is compared with that of polypeptides of known size (markers). This may be done on different gels run in parallel or by coelectrophoresis of markers and unknowns in the same gel. The latter technique minimizes random errors caused by asymmetriesin the apparatus, small bubblesaccumulat- ing at the base of the columns, or slight differences in pol.yacrylam- ide composition at the tops of the columns, etc., and therefore appreciably increases the precision with which such comparisons can be made. The problem of band overlap between markers and unknowns of similar size was circumvented by using a di- vided sample application technique (9). An example of an electrophoretogram prepared in this way is shown in Fig. 1.

Molecular Weight-Mobility Projile and Gel Composition-

Electrophoretograms were prepared on 5, 10, and 15% polyac- rylamide gels, and the relative mobility of each polypeptide was plotted -against the logarithm of its molecular weight (Fig. 2). In each case such a plot yielded a straight line over a molecular weight range which was characteristic of the gel composition. Thus a linear plot was obtained over a molecular weight range of 60,000 to 10,000 daltons for a 1501, gel (Fig. 2C), 100,000 to 10,000 daltons for a 10% gel (Fig. 2B), and 350,000 to 20,000 daltons for a 5% gel (Fig. 2A). The last result is in good agree- ment with the report of Shapiro et al. (l), who were the first to note that migration is inversely related to the logarithm of polypeptide molecular weight in the SDS system. These workers also reported that ribonuclease A and lysozyme deviated from this line on 5 y0 gels, but offered no explanation for this anomalous behavior.

As seen in Fig. 2, the molecular weight-mobility plots undergo a previously unreported inflection when the molecular weight falls below a critical size. For 5 y0 gels the critical size is about 20,000 daltons.

by guest on April 9, 2019

http://ww

w.jbc.org/

Dow

nloaded from

5076 Molecular Weight Determination on Acrylamide Gels Vol. 244, No. 18

24 RELATIVE MOBILITY

500 p 5OO[C

2’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ 21 ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ .4 .8 1.2 1.6 2.0 .4 .8 1.2 I.6 2.0

RELATIVE MOBILITY RELATIVE MOBILITY

FIG. 2. Plot of the logarithm of the molecular weights of a series of proteins versus their relative mobilities on (A), 50/,, (B) 10% and (C) 157& polyacrylamide gels containing 0.1% SDS. Each protein is identified by its number in Table I. The data for each curve were obtained according to the procedure illustrated in Fig. 1 as outlined under “Materials and Methods.” The ordi- nate intercepts obtained by extrapolating to zero mobility are

Although no attempt was made to accurately determine the position of this inflection point on the 10 or 15% gels, the mobility of insulin (mol wt 6,000) suggests that the critical size is lowered to between 6,000 and 10,000 daltons at these higher gel con- centrations. This inflection to steeper slope, which probably reflects less effective sieving of small molecules, accounts for the reported anomalous behavior of lysozyme on 5% polyacrylamide gels, but fails to account for that of ribonuclease A, which moves more slowly than expected, even after taking into account the change in slope. The apparent linearity below 20,000 daltons on the 5% polyacrylamide gels permits extension of their useful range for molecular weight determinations into the 20,000 to 5,000 dalton region. The steepness of the slope, however, necessitates rather accurate determinations of relative mobility; to this end, long gel columns are useful:

For polypeptides weighing between 60,000 and 10,000 daltons, 10% gels were convenient; they produced sharper and more

highly resolved bands than 5% gels (a property which is par- ticularly important for closely spaced bands). Furthermore, 10 y0 gels required -significantly less time for eiectrophoresis and staining than did 15% gels.

Compiled in Table I are the molecular weight values of the polypeptides used in this study and their apparent molecular weights determined from the plots of Fig. 2. In general, the molecular weights found were independent of the gel composi- tion and agreed within 5% of the literature values. With the notable exception of ribonuclease A, the maximum deviation between literature- and SDS-determined molecular weights was about 11%.

Apparent molecular weights were readily determined with a precision of 2%; it is evident from Table I, however, that dis- crepancies in the determined values frequently exceeded this uncertainty.,

characteristic of gel composition and therefore may be useful for comparing results from different laboratories. Such extrapola- tions, however, probably do not accurately describe the behavior of polypeptides larger than the ones presented, especially since the slow moving large polypeptides spend an appreciable portion of their time passing through an atypically porous zone, which is formed at the water-polymer interface during gel casting.

Such discrepancies, therefore, arise not from random experi- mental errors, but rather from small but significant variations in the log of molecular weight to mobility ratios of different pro- teins. Such variations might reasonably be due to inherent differences in the conformation or intrinsic charge of different proteins, or to differences in the weight proportion of detergent anions bound to each type of polypeptide.

Effect of Intrinsic Charge and Chain Unfolding-To establish quantitatively the effect of differences in intrinsic charge (i.e. the net charge contributed by the ionizing groups of the polypeptide chain as distinguished-from the net charge of the protein-SDS complex) on the mobility of a protein in the SDS system, it would be advantageous to study polypeptide chains differing, insofar as possible, only in charge but not in amino acid sequence.

A model set of proteins closely approxrmating this requirement was prepared by substituting lysozyme with appropriately charged side chains by the method of disulfide interchange (6). This procedure, which involves reacting the disulfide groups of the protein (I) with a large molar excess of low molecular weight --.. - ̂ ...

drsulndes (II); yieids a mixed -dBuif?de product. (III) which is

p/Y S-SR

/ ’ + RS-SR RS- P

‘A \

S-SR (1) (II) (III)

stable in the absence of thiol anion. Three homologous mixed disulfide derivatives bearing negative (R is CH2CH.#02), zero (R is CHzCH,OH), or positive (R is CH2CH2NH3+) charge groups were examined (Fig. 3A). When fully substituted, the derivatives bear eight additional negative, zero, or positive groups, respectively, in each chain. All three were insoluble in ordinary. buffers, but dissolved readily in the presence of 8 M urea.

by guest on April 9, 2019

http://ww

w.jbc.org/

Dow

nloaded from

Iisue of September 25, 1969 ̂ 5077

Protein

1. Bovine serum albumin (pentamer). .............. 2. Bovine serum albumin (tetramer). ............... 3. Bovine serum albumin (trimer). ................. 4. Y-Globulin. ..................................... 5. Bovine serum albumin (dimer) .................. 6. Bovine serum albumin ........................... 7. Ovalbumin. .................................... 8. Pepsin. ........................................ 9. Carboxypeptidase A. ............................

10. Chymotrypsinogen A. .......................... 11. Trypsin. ....................................... 12. Bromegrass mosaic virus ......................... 13. p-Lactoglobulin .................................. :4: :v!l~+Yk&i: .................................... 15. Tobacco mosaic virus ............................

16. Lysoayme ......................................

17. p-Hydroxyethyl lysozyme ........................ 18. Ribonuclease A. .................................

19. Acetylcystaminyl ribonuclease A ................. 20. Chymotrypsin B chain ........................... 21. R17virus ....................................... 22. Cytochrome c. .................................

23. Chymotrypsin C chain. ........................ 24. Insulin ..........................................

-

_-

.

-

TABLE I - I

330,000”

264,000” 198,000” 160,000” 132,000”

66,OOOd 46,000d 35,500c 34,400d

25,741” 23,800”

20,300f 18, OOOh ;7 ; ~(J(-p.

17,400d 14,400d

15,000i

13, 680c 14,620i

13,927” 13,729” 12,400c 10,157s

5,700d

a Calculated assuming multimers of bovine serum albumin (see “Materials and Methods”).

b These polypeptides were too large to penetrate the 10% and 15y0 polyacrylamide gels.

c Obtained from a set of molecular weight markers from Mann.

d See Reference 10. e See Moroux and Rovery (5). f See Stubbs and Kaeberg (11).

0 N.D., not determined. h See Townsend, Weinberger, and Timasheff (12). i These values fall below the critical inflection point on 50/, gels,

and they were determined on standard (not long) columns.

j Calculated assuming fully substituted proteins. k Calculated from the amino acid composition presented by

Weber (13).

The electrophoretic mobility of these derivatives on polyac- would be expected to encounter greater frictional resistance to rylamide gels containing 8 M urea (no SDS) was in the order passage through the polyacrylamide matrix; furthermore, it is expected from the charge of their added R groups (Fig. 3A). not unlikely that unfolding may expose previously buried ioniz- Thus, the most positively charged B-aminoethyl lysozyme ing groups, thereby changing the net charge on the molecule. derivative moved more rapidly toward the anode than did the The large effect of such conformational changes on electrophoretic homologous P-hydroxyethyl lysozyme derivative bearing an un- mobility in a gel is most clearly seen by comparing lysozyme with charged substituent. The most negatively charged /%carboxy- its hydroxyethyl derivative to which no additional charge groups ethyl lysozyme derivative, in turn, moved only very slowly have been added. In this particular example, the effect exceeded toward the anode. Unexpectedly, all three derivatives moved in magnitude that due to a difference of 16 electrostatic charge markedly slower than unreacted lysozyme. units.

The decrease in electrophoretic mobility of lysozyme associated with cleavage and substitution of its disulfide bonds is much too large to be accounted for by the 4 to 5% increase in molecular weight expected from addition of eight R groups to the lysozyme molecule. However, lysozyme is unusually stable and resistant to unfolding even in 8 M urea (14). Viscometric studies in this laboratory have confirmed this result and indicated, in addition, that the molecule becomes extensively unfolded in 8 M urea, as its disulfide bonds are cleaved during the interchange reaction.

Thus, it seems likely that the decreased mobility of the deriva- tives reflects uncoiling and extension of the lysozyme molecule, as its intrachain disulfide bonds are substituted and cleaved. The unfolded chains, because of their increased Stokes radii,

The behavior of these model polypeptides in SDS-containing gels (Fig. 3B) stands in sharp contrast to that in urea-containing gels. In the presence of the detergent each of the polypeptides migrated in the anodic, rather than the cathodic, direction. Furthermore, their mobilities were distinguishable only with the aid of the high resolution afforded by the split gel technique. Thus the effects of charge and of chain unfolding on electro- phoretic mobility, so evident in urea-containing gels, were vir- tually obliterated. The highly anionic behavior of the poly- peptides is presumably caused by binding of large amounts of detergent, amounts sufficient to swamp out even relatively large differences in intrinsic charge (1).

Assuming for a moment that (a) all the derivatives bind

Data from 5% gels

Apparent molecular weight

305,000

260,000 210,000 164,000 134,000

62,000 46,500 37,500

31,000 26,300 24,500 N.D.s

17,oooi ?7&9@.

N.D. 13,800; N.D.

18,500i N.D. 14,lOOi

N.D. 12,000” 9,500i

6,000i

-

D .-

-

eviation

% 7.6

1.5 6.0 2.5 1.5

6.5 1.1 5.6

9.9 2.0 2.9

5.5 Q..E..

4.2

36.0

1.4

3.6 6.0

5.3

- I Data from 10% gels T Data from 15% gels

Apparent lId?Clll~~ weight

6 b

b b

b b

b b b b

62,000 6.5 61,000 46,400 1.0 46,800 37,000 4.2 38,000

31,300 9.0 32,000 26,300 2.1 26,200 24,100 1.3 24,200

20,800 2.4 N.D. 16,000 11.0 16,200 17,500. $5. 17,590. 18,600 6.0 N.D. 13,900 3.5 13,600 13,700 8.7 N.D. 16,500 21.0 15,800 15,500 6.0 15,000 14,000 0.7 14,100 13,500 1.6 N.D. 13,600 9.7 13,300 10,500 3.9 10,600

-T--

D

-

eviation

%

Apparent molecular weight

Deviation

%

8.2 1.7 7.0

7.4 1.9 1.7

10.0 s..r..

5.5

15.0

2.6 1.44

7.3 4.8

by guest on April 9, 2019

http://ww

w.jbc.org/

Dow

nloaded from

5078 Molecular Weight Determination on Acrylamide Gels Vol. 244, No. 18

FIG. 3. Comparison of the electrophoretic mobility of lysozyme (L) and three mixed disulfide derivatives, p-carboxyethyl @C-L), p-hydroxyethyl (BH-L), and p-aminoethyl (BA-L). A, in 8 M urea-O.1 M cacodylate, pH 6.5, 7.5% polyacrylamide gel, 12 volts per cm per 90 min; B, in 0.1% SDS, 0.1 M phosphate, pH 7.2, 10% polyacrylamide. The two bands in the BC-L derivative seen in the urea gels suggest that a fraction was either incompletely sub- stituted or formed discrete aggregates. Note the opposite polarity of the electrodes in the two systems.

equivalent weights of detergent anions, and that (6) the 12% difference in mobility of the carboxyl and amino derivatives of lysosyme is determined solely by a charge difference of 16 units, then the amount of SDS bound may be calculated to bc of the order of 16/0.12, or about 133 molecules of SDS per lysozyme molecule. This corresponds to 2.5 mg of SDS per mg of protein. Such a value exceeds by almost 2-fold that measured for other denatured proteins which (at about the detergent concentrations, pH, and ionic strength used in these experiments) typically bind about 1.4 mg of SDS per mg of protein (15). Accepting as more reasonable the latter value of 1.4, which corresponds to about 70 SDS molecules per chain, we conclude that the effective charge difference between the derivatives is not 16 but about 8. This discrepancy can be reconciled if the derivatives do not, in fact, bind equivalent amounts of SDS, but rather, if the P-carboxy- et,hyl lysozyme derivative binds about 8 fewer molecules than does fi-aminoethyl lysozyme.

The model set of lysozyme derivatives suggests that the in- trinsic charge of a polypeptide may slightly modify its SDS- binding capacity; specifically, a negatively charged polypeptide may tend to bind slightly less SDS and a positively charged polypeptide slightly more SDS than a comparable neutral poly- peptide. Further support for this notion comes from comparing electrophoretic mobilities and intrinsic charges. For example,

TABLE II Effect of reduction or S-substitution on apparent molecular

weight of some proteins

Each figure represents a single molecular weight determination based upon the relative mobility of the indicated polypeptide in SDS-containing polyacrylamide electrophoretograms as de- scribed in Figs. 1 and 2. Each value was corrected for mass con- tribution of the S-substituted derivatives, assuming the following number of cysteine residues per chain: 35 for BSA; 6 for oval- bumin; 10 for chymotrypsinogen A; 8 for lysozyme, and 8 for ribonuclease A. Deviation from the literature value is in paren- theses.

Protein

Bovine serum albumin (mol wt 66,000) Reduced ................................. P-Hydroxyethyl derivative. ............. Not reduced .............................

Ovalbumin (mol wt, 46,000) Reduced ................................. P-Hydroxyethyl derivative. ............. Not reduced .............................

Chymotrypsinogen A (mol wt 25,700) Reduced ................................. p-Hydroxyethyl derivative. ............. Not reduced .............................

Lysozyme (mol wt 14,400) Reduced ................................. p-Hydroxyethyl derivative. ............. Not reduced .............................

Ribonuclease A (mol wt 13,680) Reduced ................................. p-Hydroxyethyl derivative. ............. IA derivative. .......................... Acetylcystaminyl derivative ............. Not reduced ............................

Apparent molecular weight

65,000 (-2%) 61,300 (-7%) 62,000 (-6%)

48,000 (+4%) 46,500 (+I%) 46,600 (+I%)

26.000 (+lyQ 24,300 (-5%) 25,500 (-1%)

13,900 (-4%) 13,700 (-9%) 13,900 (-4%)

16,500 (+21%) 14,900 (flO%,) 15,200 (+12%,) 14,600 (+6%) 15,800 (+15%)

titration studies indicate that denatured pepsin should have a charge of about -35 at pH 7.2 (10). Assuming that pepsin binds a typical complement of SDS (about 170 molecules based on 1.4 mg of SDS bound per mg of protein (15)), it would be expected to migrate about (170 + 35)/170, or about 1.2 times faster than a neutral protein of the same size. Such a mobility should lead to an apparent molecular weight of about 30,000 to 32,000 by the SDS-gel procedure. The 37,000 to 38,000 actually found (Table I) can be explained if less than a typical amount of SDS is bound by pepsin. Differential SDS binding can be simi- larly invoked to explain the unexpectedly small effects of intrin- sic charge on the mobilities of several other polypeptides, includ- ing for example, lysozyme, P-lactoglobulin, and BSA.

Although the state of foldedness can have a relatively large effect on the electrophoretic mobility of a protein in non-SDS gels, it has a surprisingly small effect in SDS-containing gels. This is illustrated by comparing the mobilities of lysozyme and its P-hydroxyethyl derivative in urea gels (Fig. 3A) and in SDS gels (Fig. 3B). All other polypeptides tested for the conse- quences of chain unfolding, with the notable exception of ribo- nuclease, showed little or no change in apparent molecular weight following disruption of intramolecular disulfide bonds (Table II). Again, as in the preceding discussion of intrinsic charge, differen- tial SDS binding may be invoked to explain the unexpectedly small effect of chain unfolding on mobility in SDS gels; in the

by guest on April 9, 2019

http://ww

w.jbc.org/

Dow

nloaded from

Issue of September 25, 1969 A. K. Dunker and R. R. Rueckert 5079

I I I I I I I 06 0.8 I*0 I*2 I*4 I.6

Relative Mobility FIG. 4. Determination of the molecular weights of the poly-

peptide components of ME virus using a 10% polyacrylamide gel; ME virus proteins were enriched in the minor components, 6 and E, by recovering the precipitate formed during incubation for 10 min at 37” in 0.1 M sodium chloride and 0.05 M sodium citrate, pH 5.7. The relative mobility of each known (0) was plotted against its apparent molecular weight. The apparent molecular weight of each ME virus polypeptide was then determined from its mobility with the aid of this plot.

present case it is necessary to suppose that a substantially greater amount of SDS is bound by unfolded as compared with folded polypeptides. The resulting greater negative charge could then counteract the increased frictional resistance ex- pected for the unfolded state. That such a substantial increase in SDS binding actually occurs is supported by equilibrium dialysis studies, which demonstrate a large (40 to 50%!) increase in SDS binding accompanying the disruption and blockade of structure-stabilizing disulfide bonds (15, 16). The nearly exact balance between the consequences of chain unfolding and en- hanced SDS binding must be regarded as fortuitous.

Of all the proteins examined in the study, the most anomalous was pancreatic ribonuclease, which migrated more slowly than expected in the SDS gels. The reason for this behavior remains unclear. It was aggravated by reduction, but improved by alkylation or by disulfide interchange (Table II). One possi- bility, that the SDS-binding capacity of either the folded or the unfolded forms differs appreciably from that of other proteins, is not borne out by the binding measurement in free solution (15) ; however, it remains possible that this situation is somehow modified by the presence of the polyacrylamide matrix. Alter- natively, it is possible that the low mobility of ribonuclease is the expression of a rapid reversible dimerization in solution; its apparent size would then be increased in proportion to the frac- tion of time the molecules exist in the aggregated state.

The wide variation in apparent molecular weights of the var- ious RNase derivatives suggests that comparative electrophoresis of several different derivatives may be a useful diagnostic pro- cedure for detecting anomalous behavior in uncharacterized polypeptides.

Molecular Weight of Virus Capsid Proteins-One of the primary motives for initiating this study was interest in evaluating and using SDS gels to determine the molecular weights of the five polypeptide components of ME virus, a small RNA-containing

TABLE III

Apparent molecular weight of polypeptide chains of ME virus

Chain Reduced Mixed disulfide~

33,000 (5)* 30,500 (5) 25,000 (5) 10,500 (2)c 41,000 (2)

32,800 (5) 30,500 (5) 26,500 (5)

N.D. N.D.

- Q Corrected for mass contribution of the substituted acetyl-

cystaminyl groups assuming the following number (18) of cysteine residues per chain, LI! (5), p (a), and y (3).

* The number of independent determinations are indicated in parentheses. The maximum deviation of any set of multiple determinations was about 2%.

c 6 falls in the zone of the critical size inflection point on 10% gels, and its molecular weight is therefore subject to greater un- certainty than that of the other chains. However, experiments on 20-cm long, 5% gels (Cf. Fig. 2A) confirmed the molecular weight value obtained on 10% gels.

dN.D., not determined.

animal virus. For this purpose, whole virus was dissociated in urea-SDS solutions into its polypeptide and ribonucleate com- ponents, and the entire solution was subjected to electrophoresis. Similar studies on three well characterized, RNA-containing viruses, tobacco mosaic virus, bromegrass mosaic virus, and R17 virus showed that their protein subunits behaved typically on SDS gels (Table I) after disruption of the virus particle, and that no special steps were necessary to remove viral RNA. A typical molecular weight determination on the polypeptide components of ME virus is illustrated in Fig. 4. The results of several such determinations on mercaptoethanol-reduced polypeptides are summarized in Table III. Similar results were obtained for the acetyl-cystaminyl mixed disulfide derivative of the viral poly- peptides, indicating no evidence of unusual behavior.

That the values obtained are actually rather good is indicated from two sources. First, they are in good agreement with mini- mum molecular weight values calculated from the ammo acid compositions of each of three components.2 Second, we have isolated, by controlled dissociation of the virion, a 5 S subunit, which apparently contains one each of the CC, /3, and y chains. The molecular weight for this subunit was found by sedimenta- tion equilibrium to be 86,400 (17, 18). This value agrees well with that calculated from the sum (88,500) of the gel-determined sizes of the three chains.

1.

2.

3.

4.

5.

;: 8.

9.

REFERENCES

SHAPIRO, A. L., VINUELA, E., AND MAIZEL, J. V., Biochem. Biophys. Res. Commun., 28, 815 (1967).

RUECKERT, R. R., AND DUESBERG, P. H., J. Mol. Biol., 17,490 (1966).

SOGAMI, M., AND FOSTER, J. F., J. Biol. Chem., 237, 2514 (1962).

SPAHR, P. F., AND EDSALL, J. T., J. Biol. Chem., 239, 850 (1964).

MOROUX, S., AND ROVERY, M., Biochim. Biophys. Acta, 113, 126 (1966).

SMITHIES, O., Science, 160, 1595 (1965). MAIZEL, J. V., Science, 161, 988 (1966). CHRAMBACH, A., REISFELD, R. A., WYCKOFF, M., AND ZACCARI,

J., Anal. Biochem., 20, 150 (1967). TRAUT, R. R., MOORE, P. B., DELUIS, H., AND TISSIERES, A.,

Proc. Nat. Acad. Sci. U. S. A., 67, 1294 (1967).

by guest on April 9, 2019

http://ww

w.jbc.org/

Dow

nloaded from

5080 Molecular Weight Determination on Acrylamide Gels Vol. 244, No. 18

10. NEURATH, H. (Editor), The proteins, Vol. I, Academic Press, 15. PITT-RIVERS, R., AND IMPIOMBATO, F. S. A., Biochem. J., 109,

New York, 1963. 825 (1968).

11. STUBBS, J. D., AND KAEBERG, P., J. Mol. Biol., 8, 314 (1964). 16. HUNTER, M. J., AND MCDUFFIE, F. C., J. Amer. Chem. Sot.,

12. TOWNSEND, R., WEINBERGER, L., AND TIMASHEFF, S. N., J. 81, 1400 (1959).

Amer. Chem. Sot., 82, 3175 (1960). 17. DUNKER, A. K., AND RUECKERT, R. R., Fed. PTOC., 28, 850

13. WEBER, K., Biochemistry, 6, 3144 (1967). (1969).

18. RUECKER- R R 14. Foss, J. G., Biochim. Biophys. Acta. 43, 300 (1960).

*A) IY. ‘Y.) DUNKER, A. K., AND STOLTZFUS, C. M., Proc. 1 Vat. Acad. Sci. U. S. A., 62,912 (1969).

by guest on April 9, 2019

http://ww

w.jbc.org/

Dow

nloaded from

A. K. Dunker and Roland R. RueckertObservations on Molecular Weight Determinations on Polyacrylamide Gel

1969, 244:5074-5080.J. Biol. Chem. 

  http://www.jbc.org/content/244/18/5074Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/244/18/5074.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on April 9, 2019

http://ww

w.jbc.org/

Dow

nloaded from