Gel chromatography on a Sepharose 4B column: Earlier elution of protein-sodium dodecyl sulfate complexes of low Stokes radii

ANALYTICAL

BI~HEMISTXY

1116,

191-198

(1985)

Gel Chromatography on a Sepharose 48 Column: Earlier Elution of Protein-Sodium Dodecyl Sulfate Complexes of Low Stokes Radii PIERRE WONG,*,’

ANDRE: BARBEAU,*

AND ALLEN D. ROSES

Received May 2, 1984 A procedure is described for the determination of the Stokes radius of a detergent micelie by gel chromatography. It was observed that different lots of Sepharose -4B can exhibit a wide variation in the permeation of their gel pores. It is shown that this variation is due to differences in their pore size distribution. It has been observed that protein-sodium dodecyl sulfate (SDS) complexes of high Stokes radii eluted on a Sepharose 4B column with Stokes radii lower than the theoretical, as it has been previously reported but that protein-SDS complexes of low Stokes radii (~70 A), contrary to what might have been expected. eluted with Stokes radii higher than the theoretical. Evidence was obtained that their anomalous elution is due to an interaction of the detergent SDS with the gel pores of small diameter. cn 198j Academic PRESS.IX KEY WORDS: chromatography: earlier; elution; gel; proteins; SDS.

The Stokes radius of a protein, which includes its bound water and other bound components, is defined in terms of the frictional coefficient by the equation f = 65r& where f and R, are the frictional coefficient and the Stokes radius of an equivalent spherical protein in a solvent of viscosity 17.It is a useful parameter since by its vaiue it characterizes the protein and it can provide, in conjunction with other measurements. some information on the molecular weight and shape of the protein (1). There are different methods that can be used for the determination of the Stokes radius of a protein. One method that is frequently used, due to its convenience, is by column chromatography on gel of Sepharose or that of Sephadex. One must, however, exercise a certain caution when using this method. Some proteins have been shown to absorb on the gel (I). Warshaw and Ackers ’ Present address: Departement de Biochimie, Facultb de Medecine, Universiti de Montreal, Case Postale 6128. Montreal, Quebec. Canada H3C 357

(2) have observed that the correlation between the partition coefficient and Stokes radius is not entirely exact for globular proteins. Nozaki et al. (3) have observed that proteinSDS* complexes and other asymmetrical proteins having large Stokes radii eluted on a Sepharose 4B column with Stokes radii lower than the theoretical: their ,anomalous elution has been attributed to their end on insertion into the gel pores. le Maire et al. (4) have shown that some detergent-solubilized membrane proteins eluted on Sepharose 6B and Sephacryl S-300 columns with Stokes radii higher than the theoretical. Finally, it has been observed that some flexible polymers like polyethylene glycol and dextran eluted with Stokes radii higher than the theoretical on a Sephadex G-200 column (5). The purpose of this article is to report two unexpected observations made during the course of the determination of the Stokes radii of isolated fractions of Band 3 of the human erythrocyte membranes by Sepharose * Abbreviation used: SDS. sodium dodecyl sulfate.

191

~3-2497/85 Copyright

$3.00

&i 1985 hy Academic

All rights of reproducrmn

Press. Inc. in any form reserved.

192

WONG, BARBEAU,

4B chromatography: (a) different lots of Sepharose 4B can be quite different from each other with respect to the permeation of their gel pores and (b) protein-SDS complexes of low Stokes radii elute on Sepharose 4B and Sephacryl S-300 columns with Stokes radii higher than the theoretical. It is shown that their anomalous elution is due to an interaction of the detergent SDS with the gel pores of small diameter. MATERIALS

AND METHODS

Materials. Sepharose 4B (Lot Nos. 12669 and 13897) and Sephacryl S-300 (Lot No. 1A 28553) were purchased from Pharmacia Fine Chemicals. The proteins used, their sources and their assumed Stokes radii in the native state and after complexation with the detergent sodium dodecyl sulfate (SDS) are tabulated in Table 1. Triton X- 100, a trademark of Rohm & Hass, and trizma base were obtained from Sigma Chemical Company;

AND ROSES

the detergent SDS was obtained from BioRad Laboratories; [ 1,2-3H(N)]cholesterol (40 Ci/mmol) was obtained from New England Nuclear. Other reagents used were of analytical grade. Preparation of the columns. The Sepharose 4B or the Sephacryl S-300 gel was equilibrated in 0.1 M NaCl, 0.01 M Tris-HCl, 0.001% sodium azide buffer, pH 7.5, and poured into a 1.5 X 90-cm Pharmacia column with its upper end fitted with a funnel of 500 ml. The gel was left sedimenting overnight. The excess was removed and the column packed further by an overnight elution with the equilibrating buffer under a head pressure of 20-25 cm. When the column was used in the presence of a detergent it was, before use, eluted with at least 350 ml of the equilibrating buffer having the detergent at a concentration of 0.2%. Preparation of the protein samples. When the column chromatography was carried out in the absence of the detergent SDS a given

TABLE 1 STANDARD

Protein Fibrinogen Phosphorylase Thyrogiobulin @Galactosidase Ferritin Catalase Aldolase Histone H 1 Orosomucoid Bovine serum albumin Ovalbumin Carbonic anhydrase oc-Chymotrypsinogen Myoglobin Cytochrome c

PROTEINS

USED

FOR GEL

CHROMATOGRAPHY

Source

Cat. No.

4 CAY

& (Ajb

Sigma Boehringer Sigma Boehringer Boehringer Boehringer Boehringer Boehringer Sigma Sigma Sigma Sigma Sigma Sigma Sigma

F-4754 108570 T-00 1 105031 20298 106810 102644 223549 G-9885 A-6003 A-5503 C-8137 C-4879 C-0630 C-7752

108 86 69 64 52 46 42’ 41’ 35 27 (30)d 23 17

101 * 11 118 f 20 69+ 5 54+ 3 78f 58_+ 43* 39t 30 + 26

6 3 2 1 0.5

a Stokes radii of the native proteins; values taken from Refs. (l-4). * Stokes radii of the protein-SDS complexes; values taken from Refs. (I ,6). ’ Stokes radius calculated on the basis of the diffusion coefficient of the protein given in Ref. (7). d The value in parenthesis has been observed previously on a Sephadex G-200 column (1). We have observed the same value on a Sephacryl S-300 column, which is a dextran like Sephadex G-200.

CHROMATOGRAPHY

OF

PROTEIN-SODIUM

protein was dissolved or diluted at a concentration of 0.25-l mg/ml with the column equilibrating buffer. One milliliter of the protein solution was then mixed with one drop of glycerol and submitted to column chromatography. When the column chromatography was carried out in the presence of the detergent SDS a given protein was first dissolved or diluted at a concentration of 2 mg/ml with a 2% SDS solution and the protein solution was dialyzed at room temperature and overnight against several changes of a loo-ml 2% SDS solution if the protein had any salt initially present. The protein solution was diluted at a concentration of 1 mg/ml with 0.1 vol of a 0.05 M sodium phosphate buffer, pH 7.5, 0.02 vol of /3mercaptoethanol and 0.88 vol of distilled water, heated at 95°C for 5 min, and cooled at room temperature. Usually 0.5 ml of this prepared protein solution was diluted with 0.5 ml of another protein solution prepared by the same procedure, mixed with one drop of glycerol, and submitted to column chromatography. Volumes of elution of detergent micelles. A procedure that takes advantage of the solubility of hydrophobic compounds in the micelles of detergents was used to determine their volumes of elution. One or two microliters of the [3H]cholesterol solution were evaporated under a stream of nitrogen gas and the residue was dissolved in 1 ml of the column equilibrating buffer having the detergent. This solution was mixed with one drop of glycerol, submitted to column chromatography, and the [3H]cholesterol concentration of the collected fractions measured. The validity of this method for measuring the volumes of elution of detergent micelles was indicated by the derived value of the Stokes radius for the Triton X-100 micelle, which was in agreement with its reported value (see Results). Column chromatography. The column chromatography was carried out under a head pressure of 20-25 cm. Fractions of 1.21.5 ml were collected in polyethylene vials

DODECYL

SULFATE

193

of 3 ml. Each vial was capped within a l-h period after collection of the fraction. The volumes of the fractions were measured by gravimetry as recommended (1) and this was done within a period of 24 h following the beginning of the chromatography. The protein concentration of the different proteins, except that of native @-galactosidase, was determined by measuring the absorbance at 230-235 nm; its concentration was determined by using an enzymatic assay (8). The [3H]cholesterol concentration was determined by counting in duplicate 100 ~11 aliquots with a Beckman LS 7000 Model spectrophotometer. The void volumes and the total volumes of the different gel columns were determined by using blue dextran and /3-mercaptoethanol; their concentrations were determined by measuring the absorbance at 254 and 235 nm, respectively. Plotting the chromatographic data. There are three equations that can be used for plotting the chromatographic data, each having a different basis (9). For comparative purposes with previous gel chromatography analytical works we have selected the equation of Ackers for plotting our chromatographic data (10). This equation is

R, = a0 + b. erff’( 1 - K,,).

[II

where R, is the Stokes radius of the protein, which includes its bound water and other bound components: a0 and b. are two constants; KD is the coefficient of partition of the protein as defined by Laurent and Killander (11) and erff’( 1 - KD) is the inverse error function of the value in parentheses. According to this equation a plot of R, as a function of erff’( I - KD) should yield a linear curve. The curves that were observed for the columns of Sepharose 4B and Sephacryl S-300 used were not linear. This lack of linearity has been previously observed by le Maire et al. (4) and has been attributed to the existence of a heterogenous distribution of the size of the gel pores.

194

WONG,

BARBEAU,

RESULTS

Dlrerences between Sepharose 4B lots. We have observed that different Sepharose 4B lots can be quite different from each other in the permeation of their gel pores. The calibration curves R, vs erf-‘( 1 - &) of globular proteins were determined for two different Sepharose 4B lots after packing into columns (Fig. 1). Both curves were observed to be at different positions; one showed a very sharp transition around 50 A, similar to that observed by le Maire et al. (4), while the other increased almost in a linear fashion as the value of the Stokes radius increased.

AND

ROSES

The calibration curves were reproductible. They were redetermined at different time intervals over a period of 5 months when the columns were frequently used. Their shapes did not change while their positions remained unchanged for a period of 3 months and then shifted slightly toward higher values of erf-‘( 1 - &). Repacking the columns was also not observed to have any effect on their shapes and positions. The observed differences of the calibration curves with different Sepharose 4B lots raised the question as to whether the determination of the Stokes radius of a protein could be influenced by the Sepharose 4B lot. In order thyr

Ol 0,O

I 0.2

0.4

e r f -‘(l

- K,,)

FIG. I. The calibration curves of globular proteins obtained with two different lots of Sepharose 4B. The elution of the Sepharose 4B columns was done with a 0.1 M NaCI, 0.01 M Tris-HC1. 0.001% sodium azide buffer, pH 7.5. The calibration curves with the triangles and the squares correspond to the Sepharose 4B lot Nos. 12669 and 13897, respectively.

CHROMATOGRAPHY

OF PROTEIN-SODIUM

to provide an answer to this question the Stokes radii of the globular micelle of the nonionic detergent Triton X-100, of orosomucoid, a strongly acidic protein, and of histone HI, a strongly basic protein were determined for the two Sepharose 48 lots. The Stokes radius of the Triton X- 100 micelle was determined after the Sepharose 4B columns were equilibrated at the detergent concentration of 0.2%. a concentration having no effect on the volumes of elution of the globular proteins used for calibration, consistent with previously reported data that indicated that nonionic detergents have no effect on the Sepharose 4B column characteristics (3.4). The results that were obtained were the following. The Stokes radius of the Triton X-100 micelle was the same for the two Sepharose 4B columns: its average value was 49 +- 1 A, which is the same as that determined by light scattering (12). The Stokes radii of orosomucoid and of histone H 1 were also the same for the two Sepharose 4B columns used: their respective average values were 39 + 2 and 42 & 2 A, which are in agreement with those calculated on the basis of their diffusion coefficient (Table 1). It can be concluded from the above results that the Sepharose 4B lot has no influence on the determination of the Stokes radius of a protein and that the observed differences of the calibration curves of the globular proteins with different Sepharose 4B lots reflect simply differences in pore size distribution of the Sepharose 4B lots. El&m qj’ protein-SDS complexes, It has been previously observed that protein-SDS complexes of high Stokes radii as well as asymmet~~al proteins of high Stokes radii such as fibrinogen and myosin eluted on a Sepharose 4B column with Stokes radii lower than the theoretical {3). This anomalous eiution has been attributed to their end on insertion into the gel pores. On the basis of the above data we had expected that proteinSDS complexes of low Stokes radii would elute like the protein-SDS complexes of high Stokes radii or would elute with their theo-

DODECYL

SULFATE

195

retical Stokes radii. However, contrary to our expectations they eluted with Stokes radii higher than the theoretical (Fig. 2). The validity of the data was indicated by three controls: (a) the volume of elution of /3mercaptoethanol was not observed to change during the determination of the volumes of elution of the protein-SDS complexes and was the same as that observed before the addition of the SDS to the column; (b) protein-SDS complexes of high Stokes radii as well as fibrinogen eluted as previously reported (Fig. 2); and (c) the volumes of elution of globular proteins that were determined after the removal of the SDS were the same as those determined before its addition to the column. There are two possible explanations for the above results: (a) the elution of globular proteins and of asymmet~cal proteins are function of their shape and size; and (b) there is an interaction of the detergent SDS with the gel material; the interaction is limited to gel pores of small diameter since fib~nogen, an asymmetrical protein of high Stokes radius, eluted like protein-SDS complexes of comparable Stokes radii (Fig. 2). Three observations that we have made subsequently indicated that it is the second one which is correct. These observations were the following: (a) The Stokes radius of the globular SDS micelle, which is 22 8, in 0.1 M NaCl at 25°C (131, was observed to be 35 W on the Sepharose 4B column equilibrated at the concentration of SDS used to study the elution of protein-SDS complexes (0.2%); the same value was also observed when the column was equilibrated at a SDS concentration five times lower (0.04%). (b) The elution of protein-SDS complexes was different on a Sephacryl S-300 column, a gel material of different chemical structure and mechanical strength than that of Sepharose 4B: the enhancement of the elution of the protein-SDS complexes was significantly less on the Sephacryl S-300 column and the intersection of the curves of globular proteins and proteinSDS complexes occurred at a Stokes radius

196

WONG, BARBEAU,

AND ROSES

1oa

80

40

20-

01 0.0

0:2

1

014

0.6

erf -’ (l- Kd) FIG. 2. The calibration curves of globular proteins and of protein-SDS complexes obtained on a Sepharose 4B column. The curve with the circles is for the globular proteins while that with the squares is for the protein-SDS complexes. The data point for fibrinogen has been obtained in the absence of SDS during the chromatography. The elution of the Sepharose 4B column was done with a 0.1 M NaCI, 0.01 M Tris-HCI, 0.001% sodium azide buffer, pH 7.5, in the absence or the presence of SDS at a concentration of 0.2%.

of 60 A on the Sephacryl S-300 column while it occurred at a Stokes radius of around 70 A on the Sepharose 4B column (Figs. 2 and 3). (c) The Stokes radius of the globular SDS micelle was observed to be 27 A on the Sephacryl S-300 column which is different of that observed on the Sepharose 4B column and is 5 A higher than the reported value.

DISCUSSION

One observation that we have made is that different Sepharose 4B lots can be quite different with respect to the permeation of their gel pores (Fig. 1). This observation was not, however, totally unexpected since calibrations of Sepharose 4B columns done at

CHROMATOGRAPHY

OF PROTEIN-SODIUM

J

DODECYL

SULFATE

197

mucoid, a strongly acidic protein, and of histone H 1, a strongly basic protein for the phospho 100. Sepharose 4B lots used indicated that its value is not influenced by the Sepharose 4B lot and that the observed differences of the thyr calibration curves with different Sepharose 4B lots are due simply to differences in pore size distribution of the Sepharose 4B lots. The second observation that we have made, which is of greater interest, is that the deter60, gent SDS interacts with the Sepharose 4B 02 and Sephac~l S-300 gel pores of small diu* ameter. The impIication is the Stokes radii of protein-SDS complexes cannot be determined by Sepharose 4B or Sephacryl S-300 40 chromatography using the calibration curve of globular proteins. It is, however, possible to have an estimation of their Stokes radii provided that standard protein-SDS com20. plexes are used for the column calibration since they elute in a predictable manner (Figs. 2 and 3). There are two modes of interaction of the detergent SDS with the gel 0 I t a2 of6 1.0 1.4 pores of small diameter that can be proposed. One is related to the residual sulfate groups err-‘(1-I$) known to be covalently bound to Sepharose FIG. 3. The calibration curves of the globular proteins and of the protein-SDS complexes obtained on a Se- gels, which are responsible for the earlier phacryl S-300 column. The curve with the circles is for elution of proteins when the ionic strength the globular protein while that with the squares is for of the medium is decreased (4): the residual the protein-SDS complexes. The data point for fibrinogen groups bound to the part of the gel forming has been obtained in the absence of SDS during the the gel pores of small diameter prevent the chromatography. The elution of the Sephacryl S-300 penetration of the protein-SDS complexes column was done with a 0.1 M NaCl, 0.01 M Tris-HCI. into the gel pores by an electrostatic repulsion 0.001% sodium azide buffer, pH 7.5, in the absence or in the presence of SDS at a concentration of 0.2%. of the sulfate group of the detergent. The other mode of interaction involves a binding of the detergent to the part of the gel forming different times have indicated that possibility the gel pores of small diameter. The binding (3). The above observation is of interest since couid cause the earlier elution of the proteinit can explain the differences of the calibration SDS complexes either by changing the gel curves of globular proteins that have been pores characteristics or by preventing their obtained by different groups of investigators penetration into the gel pores by an electro(3,4). The differences of the calibration curves static repulsion. Of these two alternatives with different Sepharose 4B lots raised the the second seems the most likely since the miquestion as to whether the value of the Stokes celle of the nonionic detergent T&on X-100, radius of a protein could be influenced by orosomucoid, a strongly acidic protein and the Sepharose 4B lot. ~e~urcments of the of histone H 1, a strongly basic protein, elute Stokes radii of the globular micelle of the with their expected Stokes radii and that the nonionic detergent T&on X-100, of oroso- detergent SDS has been shown to complex

s/

fibri

198

WONG, BARBEAU,

reversibly with amylose, a polysaccharide ( 14); advantage has been taken of this property to separate different fractions of amylose by electrophoresis (14,15). Our observation on the interaction of the detergent SDS with the Sepharose 4B gel is in disagreement with that of Fish et al. (9) and Tanford et al. (1). This is not clearly understood but possible explanations for this discrepancy are the following. They have studied the elution of protein-SDS complexes on a Sepharose 4B column with respect to that of proteins denatured with guanidine hydrochloride and this was done at a very low ionic strength which causes an earlier elution of the proteins (4) and may affect the interaction of the detergent SDS with the Sepharose 4B gel. Furthermore, they have compared the elution of globular proteins and of protein-SDS complexes on a Sephadex G-200 column. It is possible that this crosslinked dextran does not interact with the detergent SDS since Sephacryl S-300, which is a different type of crosslinked dextran, interacts less with the detergent than Sepharose 4B, a gel of agarose. le Maire et al. (4) have observed that detergent-solubilized membrane proteins eluted on Sepharose 6B and Sephacryl S-300 columns with Stokes radii higher than the theoretical. An examination of the data does not indicate that their earlier elution could be due to an interaction of the detergent with the part of the gel material forming the gel pores of small diameter since none of the detergent used affected the elution of globular proteins and that the values of the Stokes

AND ROSES

radii for the solubilized membrane examined were about the same for both types of columns, contrary to what we have observed for the protein-SDS complexes.

We thank Professors C. Tanford and J. R. Reynolds with whom intem~~on has inspired this work, and /Carolyn Rancourt and Lise Pomerleau for their expert secretarial assistance.

REFERENCES 1. Tanford, C., Nozaki, Y., Reynolds, J. A., and Makino, S. (1974) Biochemistry 13,2369-23X. 2. Wamhaw, H. S., and Ackers, G. K. (1971) Ataaf. Biochem. 42, 405-42 1. 3. Nozaki, Y., Schechter, N. M., Reynolds, J. A.: and Tanford, C. (1976) Biochemistry 15, 3884-3890. 4. le Maire, M., Rivas, E., and Moller, J. V. (1980) Anal. Bioehem. 106, 12-21. 5. Jorgensen, K. E., and Moller, J. V. (1979) Amer. J. Physiol. 236, F103-Fill. 6. Fish, W. W. (1975) ~eth~s Membr. Bid. 4* 189276. 7. Handbook of Biochemistry (1970) C-12 and C-36 2nd Ed (Sober, H. A., ed.), CRC Press. 8. Craven, G. R., Steers, E., Jr., and Anfinsen, C. B. (1965) J. Biol. Chem. 240,2468-2470. 9. Fish, W. W., Mann, K. G., and Tanford, C. (1969) J. Bid. Chem. 244,4989-4934. 10. Ackers, G. K. (1967) J. Bid Chem. 242, 32373238. 11. Laurent, T. C., and Killander, J. (1964) J. Chvomatogr. 14, 3 17-330. 12. Yedgar, S., Barenholz. Y.. and Cooper. V. G. (1974) Biochim. Biophys. Acta 363, 98-11 i. 13. Tanford, C. (1978) J. Phys. Chem. 78, 2469-2479. 14. Nishimura, N.. and Janado, M. (1975) J. Biochem. 77,42 t-426. 15. Shimada. K., Kido, S.. and Janado, M. (1976) Anal. Biochem. 72.664-668.