Nov 18, 1976 - The bound detergent forms mi- celle-like clusters around thehydrophobic domains of these proteins, and usually the proteins retain their native ...
Proc. Natl. Acad. Sci. USA Vol. 74, No. 2, pp. 529-532, February 1977
Biochemistry
Charge shift electrophoresis: Simple method for distinguishing between amphiphilic and hydrophilic proteins in detergent solution (membrane proteins/protein analysis/Triton X-100/cytochrome bs/Semliki Forest virus)
ARi HELENIUS AND KAI SIMONS European Molecular Biology Laboratory, Postfach 10.2209, 6900 Heidelberg, Federal Republic of Germany
Communicated by John Kendrew, November 18, 1976
ABSTRACT Seventeen hydrophilic proteins and five amphiphilic membrane proteins were subjected to agarose gel electrophoresis in the presence of a nonionic detergent (Triton X-100), a mixture of a nonionic and an anionic detergent (Triton X-100 and sodium deoxycholate), and a mixture of a nonionic and a cationic detergent (Triton X-100 and cetyltrimethylammonium bromide). The electrophoretic mobility of the hydrophilic proteins was unaffected in the three detergent mixtures. However, the mobility of the hi hilic proteins shifted ana e system and cathodally odally in the Triton X-100-deo in the Triton X-100-cetyltrimethylammonium bromide system when compared to the mobility in Triton X-100 alone. The detergent-induced shift in mobility provides a simple, rapid, and sensitive method for distinguishing between hydrophilic and amphiphilic proteins. In the characterization of membrane proteins it is important to be able to decide whether the proteins studied possess hydrophobic domains that anchor them to the hydrocarbon interior of the bilayer or whether they are externally bound to the membrane. The electrophoretic screening method introduced in this paper is based on the fundamental difference in the interaction of hydrophilic and amphiphilic proteins with "mild" detergents such as Triton X-100 (p-t-octylphenylpolyoxyethyleneg-1o). It has been shown in numerous studies that whereas ordinary soluble proteins and peripheral membrane proteins bind little or no Triton X-100, amphiphilic membrane proteins bind large amounts (usually 80-100 mol of Triton per mol of protein) when solubilized from membranes (refs. 1-5; for reviews see refs. 6 and 7). The bound detergent forms micelle-like clusters around the hydrophobic domains of these proteins, and usually the proteins retain their native conformation. Several methods have been described to determine Triton X-100 binding (1-3, 5). Each of these methods can as such be used to differentiate between amphiphilic and hydrophilic proteins, but contrary to the method described here, they require purified protein (usually in milligram amounts) and radioactive detergent. Instead of using Triton X-100 alone we have used mixtures of Triton X-100 and charged detergents. When solubilized with such detergent mixtures, the amphiphilic proteins form detergent-protein complexes containing both neutral and charged detergent molecules (see ref. 6). The net charges of the complexes are thus dependent on the charge of the detergents used, resulting in a clear-cut difference in electrophoretic mobility of the amphiphilic proteins when electrophoresed in cationic and anionic detergent mixtures. The electrophoretic mobility of the hydrophilic proteins, which do not interact with the detergents, remains unaffected by a change in the charge of the detergents used. 529
MATERIALS AND METHODS The amphiphilic membrane proteins used were microsomal cytochrome b5 (a generous gift from Dr. C. Lussan, Talence, France), intestinal brush border aminopeptidase (a generous gift from Dr. S. Maroux, Marseille, France), membrane penicillinase from Bacillus licheniformis (8), and the glycoproteins El and E2 from Semliki Forest virus membrane (9). The following soluble proteins were used: the trypsin-cleaved form of the brush border aminopeptidase (a gift from Drs. S. Maroux and D. Louvard, Marseille, France, ref. 10); the polar hemecontaining domain of cytochrome b5 (obtained by trypsin cleavage of the membrane form, ref. 11); the exopenicillinase (isolated from Bacillus licheniformis, ref. 12); transferrin and gamma globulin (Kabi, Sweden); lysozyme, ferritin, cytochrome c, and myoglobin (Serva, Federal Republic of Germany); chymotrypsinogen and pancreatic ribonuclease (Worthington, USA); bovine serum albumin (Calbiochem, USA); catalase and glutamate dehydrogenase (Miles, Great Britain); and thyroglobulin, insulin, and ovalbumin (Sigma, USA). Triton X-100 (Rohm & Haas, USA), sodium deoxycholate (Schwarz/Mann, USA), and cetyltrimethylammonium bromide (Serva, Federal Republic of Germany) were used without purification. The agarose electrophoresis experiments were performed at room temperature (250) in 1 % agarose (BioRad, USA) on glass plates (11 X 20.5 cm) as described by Weeke (13), using a water-cooled chamber (Behringwerke, Federal Republic of Germany), paper wicks, and 1 X 11 mm sample slits. The gel buffer was 0.05 M glycine-NaOH, 0.1 M NaCl at pH 9.0 containing 5 g of Triton X-100 or 5 g of Triton X-100 and 2.5 g of sodium deoxycholate or 5 g of Triton X-100 and 0.5 g of cetyltrimethylammonium bromide per liter. The buffers in the electrode chambers were the same except that in the experiments with Triton X-100 no detergent was added. The plates were first electrophoresed for 15 min at 4.5 V/cm. Samples (12 Al) were applied and electrophoresed for 2 hr at 4.5 V/cm unless otherwise indicated. After electrophoresis the gels were immediately dried under an air fan (40450) and then autoradiographed using Kodak Royal X-Omat film, scanned with a Perkin-Elmer 156 double wavelength spectrophotometer, or stained for protein (45 min in 0.7 g of Coomassie blue, 450 ml of methanol, and 90 ml of acetic acid per liter; rinsed for 15 min in 75 ml of acetic acid and 50 ml of methanol per liter). Immunoelectrophoresis was performed in agarose (1 g/100 ml) gels according to Scheidegger (14), except that the buffer (0.06 M sodium barbital diethylbarbitate-HCI, pH 8.7) contained Triton X-100, deoxycholate, and cetyltrimeth-
530
Proc. Natl. Acad. Sci. USA 74 (1977)
Biochemistry: Helenius and Simons BSA
INSULIN
THYROGLOBULIN
FERRITIN
OVALBUMIN
c2cm
ORIGIN
l-l-4l
-
i
LL Mr ORIGIN
I1
-.
i-
+I
I
I
MYOGLOBIN
TRANSFERRIN
GDH
CATALASE
-
GAMMAGLOBULIN
{
I
I
i
a CD NO DETERGENT CYTOCHROME C a TX-DOC i I I II I a TX _ TX -CTAB
CHYMOTRYPSINOGEN LYSOZYME I I i I I~
RIBONUCLEASE
ORIGIN -10
cz3
CO
S-
a_
_~
FIG. 1. Combined results from agarose gel electrophoresis of soluble proteins in the absence of detergent, in Triton X-100 (TX) and sodium deoxycholate (DOC), Triton X-100 alone, and Triton X-100-cetyltrimethylammonium bromide (CTAB). The samples applied contained 35 jig of protein in 12 Ml. Protein stain: Coomassie blue. BSA, bovine serum albumin; GDH, glutamate dehydrogenase.
ylammonium bromide in the concentrations given above. The protein samples used for agarose electrophoresis and immunoelectrophoresis usually contained 0.7-1.5 mg of protein per ml dissolved in the electrophoretic buffer containing a 4-fold detergent concentration over that used in the gels. The samples were allowed to equilibrate overnight at 00. RESULTS AND DISCUSSION We have tested 17 hydrophilic proteins and 5 amphiphilic membrane proteins. The hydrophilic proteins included acidic and basic proteins and glycoproteins, and some contained several subunits. Of the amphiphilic membrane proteins tested, three could be obtained in a proteolytically cleaved form that contained only the polar moiety: the polar heme-containing EXOPENICIWNASE
domain of cytochrome b5, the trypsin form of aminopeptidise, and the bacterial exopenicillinase. Cytochrome b5 (15) and membrane penicillinase (K. Simons and M. Sarvas, unpublished results) consist of single polypeptide chains when solubilized with Triton X-100, whereas the membrane aminopeptidase contains three noncovalently bound polypeptide chains (16). Fig. 1 shows the electrophoretic patterns obtained for a number of ordinary soluble proteins in the absence of detergent and in the three detergent media. Only minor differences in the electrophoretic mobility were observed. The same was true for exopenicillinase (Fig. 2), the trypsin form of aminopeptidase (Fig. 2), and the polar fragment of cytochrome b5 (Fig. 3). In contrast, all the amphiphilic membrane proteins tested, membrane penicillinase (Fig. 2), membrane aminopeptidase
IIANE PENKMLWHASE
-
ORIGIN
1
TX-DOC
TX
TX-CTAB
TX-DOC
TX
TX-CTAB
~~~~~~d E
TX
C
2
ORIGIN
I
AMINOPEPTIDASE (TRYPSIN FORM)
MEMERANE AMWOPEPTIDASE
FIG. 2. Combined results from agarose gel electrophoresis of exopenicillinase, membrane penicillinase, the trypsin form of aminopeptidase, and membrane awminopeptidase in Triton X-100 and sodium deoxycholate (TX-DOC), in Triton X-100 alone (TX), and in Triton X-100 and cetyltrimethylammonium bromide (TX-GTAB). Protein stain: Coomassie blue.
TX-CTAB 1cm
FIG. 3. Agarose electrophoresis of a mixture of cytochrome b5 (the d-form) and trypsin-cleaved cytochrome b5 (the t-form) in Triton X-100 and sodium deoxycholate (TX-DOC), in Triton X-100 alone (TX), and in Triton X-100 and cetyltrimethylammonium bromide (TX-CTAB). Electrophoresis was performed for 75 min at 4.5 V/cm. The plates were dried and scanned at 415 and 450 nm.
Natl. Acad. Sci. USA 74 (1977) BProc. Helenius and Simons Biochemistry:
531
MEMBRANE PENICILLINASE
ORIGIN ->
I
I
EXOPENICILLINASE
"!9
X.I
0
MEMBRANE PENICILLINASE
TX-DOC
TX
TX-CTAB
FIG. 4. Combined results from agarose electrophoresis of [35Slmethionine-labeled spike glycoprotein El from Semliki Forest virus in the presence of Triton X-100 and deoxycholate (TX-DOG), Triton X-100 alone (TX), and Triton X-100 and cetyltrimethylammonium bromide (TX-CTAB). The samples contained 8000 cpm in 12 Mtl (specific activity 4500 cpm/jug of protein). Electrophoresis was performed for 4 hr at 4.5 V/cm. After drying, the gels were autoradiographed for 6 days.
(Fig. 2), cytochrome b5 (Fig. 3), and the spike glycoproteins El (Fig. 4) and E2 (not shown) displayed a more anodal migration when electrophoresis was performed in the presence of the Triton X-100-deoxycholate mixture than in the presence of Triton X-100 alone. The migration in the Triton X-100-cetyltrimethylammonium bromide mixture was, on the other hand, more cathodal than in the presence of Triton X-100 alone. In all cases the shifts in mobility of the amphiphilic proteins were easily detectable. The hydrophilic as well as the amphiphilic proteins gave single bands in each electrophoretic system, with the exception of aminopeptidase which was heterogeneous in the Triton X-100-cetyltrimethylammonium bromide electrophoresis system (Fig. 2). We have interpreted the emergence of the additional cathodally moving band in this case as an indication of partial dissociation of the subunit structure. Fig. 5 shows the immunoelectrophoretic patterns obtained using both forms of penicillinase in the three detergent mixtures. The shifts in the membrane penicillinase mobility could be seen easily, whereas the exopenicillinase remained unshifted. Similar results were obtained with the trypsin and the membrane forms of aminopeptidase, and with the membrane proteins of Semliki Forest virus using the respective antisera (not shown). Rather than using deoxycholate and cetyltrimethylammonium bromide alone (which may have given a larger shift in mobility for the amphiphilic proteins), we have used these charged detergents in combination with an excess of Triton X-100. This was in order to keep the structure of protein-detergent complexes as constant as possible in all three detergent systems (see ref. 17). Furthermore, cetyltrimethylammonium bromide is known to be a denaturant when used alone. However, when mixed with sufficient Triton X-100, its chemical potential drops below that required for massive binding and denaturation (see ref. 6). The fact that the migration of the hydrophilic proteins remained unchanged and that the antibody-antigen reactions were not affected indicated, indeed, that the cetyltrimethylammonium bromide present did not drastically denature the proteins tested. In preliminary experiments performed at lower ionic strength (0.025 M Tris1HCl, pH 9.0) the basic proteins cytochrome c, lysozyme, and chymotrypsinogen exhibited clear-cut anodal shifts in the Triton X-100-deoxycholate system but no difference in the Triton X-100-cetyltrimethylammonium bromide system when compared to the mobility in Triton X-100 alone. This observation suggests that basic proteins can bind deoxycholate electrostatically (see ref. 18), especially if the ionic
EXOPENICILLINASE
_
MEMBRANE
PENICILLINASE
EXO PENICILLINASE
_
FIG. 5. Immunoelectrophoresis of exopenicillinase and membrane penicillinase in the presence of Triton X-100 and sodium deoxycholate (TX-DOC), Triton X-100 alone (TX), and Triton X-100 and cetyltrimethylammonium bromide (TX-CTAB). Electrophoresis was performed at 3 V/cm for 90 min. The rabbit antiserum used had been raised against exopenicillinase.
strength is low. It should also be mentioned in this context that pancreatic colipase binds large amounts of deoxycholate and other negatively charged detergents (19, 20), but it does not bind Triton X-100 (21) or positively charged deoxycholate derivatives (B. Borgstrom, unpublished observation). The use of three detergent systems (anionic, nonionic, and cationic) helps to detect such charge-specific binding. Integral membrane proteins with an extensive hydrophobic domain should display both an anodal shift in Triton X-100-deoxycholate and a cathodal shift in Triton X-100-cetyltrimethylammonium bromide as compared to their mobility in Triton X-100 alone. The charge shift electrophoresis may be easily adapted for preparative purposes and for other gel media. When combined with sensitive methods to detect the protein bands it can also be used in the study of trace amounts of protein (see Figs. 4 and 5). Specific detection using antibodies, biological activities, etc. make it feasible to analyze proteins in complex mixtures (unpublished results). In such mixtures a clear-cut shift in one of the protein components indicates either that the protein itself is amphiphilic or that the protein is part of an amphiphilic protein complex. When solubilized whole membranes are analyzed it is important to include an excess of detergent in both the sample and the gel to ensure maximal separation of lipid and protein. If necessary, the lipids and the proteins can be separated beforehand using sucrose gradient centrifugation in the presence of detergent (3). Different dissociation states of oligomeric proteins may sometimes occur in the three different detergent media (see Fig. 2). Our preliminary data indicate that prior crosslinking of the proteins can be used to eliminate this
problem. We thank Bodil Holle and Hilkka. Virta for their skillful assistance. 1. Helenius, A. & Simons, K. (1972) J. Biol. Chem. 247, 36563661. 2. Makino, S., Reynolds, J. A. & Tanford, C. (1973) J. Biol. Chem. 248,4926-4932. 3. Simons, K., Helenius, A. & Garoff, H. (1973) J. Mol. Bwol. 80, 119-133. 4. Clarke, S. (1975) J. Biol. Chem. 250,5459-5469.
532
Biochemistry: Helenius and Simons
5. Fries, E. (1976) Blochim. Blophys. Acta, 455,928-936. 6. Helenius, A. & Simons, K. (1975) Biochim. Biophys. Acta 415, 29-79. 7. Tanford, C. & Reynolds, J. A. (1976) Biochlm. Blophys. Acta, 457, 133-170. 8. Sawai, T. & Lampen, J. 0. (1974) J. Biol. Chem. 249, 62886294. 9. Helenius, A., Fries, E., Garoff, H. & Simons, K. (1976) Biochim. Biophys. Acta, 436,319-334. 10. Maroux, S., Louvard, D. & Baratti, J. (1973) Biochim. Biophys. Acta, 321,282-295. 11. Ito, A. & Sato, R. (1968) J. Biol. Chem. 243,4922-4930. 12. Pollock, M. R. (1965) Blochem. J. 94,666-675. 13. Weeke, B. (1973) "A manual of quantitative immunoelectrophoresis," Scand. J Immunol. 2, Suppl. no. 1, 25-3. 14. Scheidegger, J. J. (1955) Int. Arch. Allergy Appl. Immunol. 7,
Proc. Nati. Acad. Sci. USA 74 (1977) 103-110. 15. Visser, L., Robinson, N. C. & Tanford, C. (1975) Biochemistry 14, 1194-1199. 16. Maroux, S. & Louvard, D. (1976) Biochim. Biophys. Acta 419, 189-195. 17. Tanford, C. (1973) The Hydrophobic Effect (John Wiley and Sons, New York), pp. 814-5. 18. Rowley, R. R. & Wainio, W. W. (1958) J. Am. Chem. Soc. 80, 4384-4386. 19. Borgstr6m, B. & Donner, J. (1975) J. Lipid Res. 16,287-292. 20. Charles, M., Sari, H., Entressangles, B. & Desnuelle, P. (1975) Biochem. Biophys. Res. Commun. 65,740-745. 21. Borgstr6m, B., Donner, J. & Erlanson, C. 01974) in Advances in Bile Acid Research, eds. Matern, S., Hackensmith, Back, P. J. & Gerok, W. (F. K. Schattauer Verlag, Stuttgart-Newr York), pp. 213-217.
