KEY WORDS: gel electrophoresis; glycogen phosphorylase; protein fragmentation; ... Preparation of samples for sodium dodecyl sulphate polyacrylamide gel.
Bioscience Reports, Vol. 7, No. 3, 1987
Degradation Artefacts During Sample Preparation for Sodium Dodecyl Sulphate Polyacrylamide Gel Electrophoresis E. Jane Cookson and Robert J. Beynon 1 Received April 29, 1987 KEY WORDS: gel electrophoresis; glycogen phosphorylase; protein fragmentation; western blots.
Preparation of samples for sodium dodecyl sulphate polyacrylamide gel electrophoresis routinely involves heating the protein in solution containing detergent and reducing agent for at least two minutes. Here we show that this treatment causes fragmentation of the protein glycogen phosphorylase, whether purified or as a component of a skeletal muscle preparation. The fragments are detected as minor bands on western blots and represent the products of discrete breakage point in the peptide sequence. Protease inhibitors cannot suppress the fragmentation. Such small amounts of immunoreactive fragments may be incorrectly identified on western blots as contaminants that were originally present in the antigen preparation. They may also be a source of ambiguity in studies that search for degradation intermediates during proteolysis. INTRODUCTION We report that proteins may be degraded, probably by a non-enzymic process, during sample preparation for sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE). Such artefactual intermediates are detected readily on western blots and we caution that care is needed in the interpretation of results obtained by this method. Previous studies have shown that fl-galactosidase is fragmented during sample preparation, when incubated over prolonged periods at 100~ (Kowit and Maloney, 1982). We extend these studies to demonstrate that with detection methods of appropriate sensitivity, fragmentation can be detected over the short incubation periods that are used routinely for sample preparation: Department of Biochemistry, University of Liverpool, PO Box 147, Liverpool L69 3BX. 1 To whom correspondence should be addressed.
2O9 0144-8463/87/0300-0209505.00/0 9 1987 PlenumPublishingCorporation
210
Cookson and Beynon
Intracellular protein degradation requires a sequential series of proteolytic events that result in the conversion of the native protein into its constituent amino acids. It follows that this process might be defined by the intermediate, partially-proteolysed forms (Beynon et al., 1985a; Wilson and Smith, 1985; Reznick et al., 1985). However, because the levels of the intermediates are so low, detection methods are particularly vulnerable to artefacts such as described here.
MATERIALS AND METHODS
Materials Horseradish peroxidase-conjugated anti-mouse and anti-rabbit immunoglobulins were purchased from Dako Ltd., High Wycombe, Bucks, UK. Nitrocellulose (pore size 0.2 #m) was obtained from Schleicher and Schuell, D-3354 Dassel, West Germany. Eupergit C (oxirane acrylic beads) was a generous gift from Rohm Pharma GmbH, D 6108 Weiterstadt, West Germany. Highly purified rabbit phosphorylase (obtained from Professor P. Cohen, University of Dundee) was linked to the Eupergit beads using a published method (Hannibal-Friedrich et al., 1980).
Production of Antibodies A polyclonal antiserum to mouse phosphorylase, purified as in Butler et al. (1984) was raised in rabbits by subcutaneous injection of the protein (200 #g) in an emulsion with Freund's complete adjuvant. Booster injections (200 pg in Freund's incomplete adjuvant) were given 14, 28, 43 and 77 days later. The antibodies were affinity purified on immobilised rabbit phosphorylase before use. A monoclonal antibody to pyridoxal phosphate, E6(4)1, was a kind gift of Dr J. Cidlowski. The preparation and specificity of the antibody is described elsewhere (Viceps-Madore et al., 1983). No further purification of the ascites fluid was necessary.
Samples for Electrophoresis Purified rabbit phosphorylase was diluted to a final concentration of I mg/ml in a buffer consisting of 0.02 M Hepes, 0.14 M NaC1, pH 7.4. Mouse muscle soluble proteins were obtained from hind limb and back muscles of C57BL/6J mice. Muscle was homogenised at 4~ in 4 volumes of 0.02 M Hepes, 0.14 M NaC1, pH 7.4 and the homogenate was centrifuged at 30,000 9/h to yield a supernatant that contained virtually all of the glycogen phosphorylase activity. In some instances, the tissues were homogenised in the presence of one of two mixtures of protease inhibitors: (a) 1 mM phenylmethylsulphonyl fluoride, 10/zM E-64c and 5 mM EDTA or (b) 150#M chymostatin, 50#M leupeptin and 5 m M 1,10 phenanthroline. The pure phosphorylase was also treated with these inhibitors prior to electrophoresis. All preparations of purified enzyme and mouse soluble muscle proteins were prepared for SDS-PAGE in an identical manner. To detect phosphorylase and PLP-containing peptides with the monoclonal antibody to the cofactor it was necessary to reduce and thus stabilise the aldimine linkage between protein and cofactor. Protein samples were
Protein Degradation in vitro
211
treated with 160 mM NaCNBH3 for 15 minutes at 4~ in the presence of 0.25 M imidazole citrate, pH 6.0. Unreduced samples were treated identically except for the omission of the reductant. After reduction, the samples were mixed with an equal volume of sample buffer (0.08 M Tris/HC1, pH 6.8, containing 2% (w/v) SDS, 0.1 M dithiothreitol, 10% (v/v) glycerol and 0.001% (w/v) bromophenol blue) and were placed in a 100~ water bath for 0, 2 or 10 minutes before cooling and applying to the gel. All samples were in the presence of sample buffer for identical times before electrophoresis.
Polyacrylamide Gel Electrophoresis Sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) was performed according to the method of Studier (1973). The 1.5 mm thick gels consisted of 12.5 % running gel and 5 % stacking gel. Electrophoresis was for 3-4 h at 40 mA at 10~ After electrophoresis, parts of the gel were stained for protein or electroblotted into nitrocellulose. The protein stain was 0.1% (w/v) Fast Green in 45% (v/v) methanol, 7 % (v/v) acetic acid.
Western Blotting Proteins separated by SDS-PAGE were electroblotted onto nitrocellulose using a modification of the method of Towbin et al. (1979). The blotting buffer was 25 mM Tris, 192 mM glycine, 20% (v/v) methanol and 0.1% (w/v) SDS, pH 8.3. The transfer was performed at 10~ for 20 h, using a current of 70 mA; Fast Green staining of the gel revealed comprehensive transfer of proteins. The nitrocellulose membrane was stained for protein using Fast Green (see above) or immunostained with anti-PLP monoclonal antibody or affinity-purified anti-mouse phosphorylase antiserum. Protein-bound PLP was detected on nitrocellulose using ascites fluid containing antiPLP antibody. All incubations were at room temperature with continuous shaking. The nitrocellulose was first incubated for 1.5 h in 10 mM sodium phosphate/0.15 M NaC1, pH 7.4, containing 0.2% (v/v) Tween 20 (PBS-T) and ascites fluid diluted 1:250,000. The nitrocellulose was then washed three times in PBS-T (5 min/wash). Second antibody; horseradish peroxidase-conjugated rabbit anti-mouse immunoglobulin, diluted 1:1000 in PBS-T, was incubated with the nitrocellulose for 1.5 h, followed by a further 3 washes in PBS-T (5 min/wash). Second antibody was detected by incubating the nitrocellulose for 10 min in 100 ml PBS-T, containing 0.02 % (w/v) diaminobenzidine and 0.03% (v/v) H202. After thorough washing in water, the stained bands were intensified by washing in 0.5 % (w/v) CuSO4 in 0.15 M NaC1 for 5 min, prior to rinsing in water and drying in air. A similar procedure was employed for the affinity-purified polyclonal antiserum, used at a dilution of 1:500. The second antibody (horseradish peroxidase-conjugated pig anti-rabbit immunoglobulin) was used at a dilution of 1:1000.
212
Cookson and Beynon RESULTS A N D D I S C U S S I O N
The work described herein was initiated as part of an investigation of the degradation of glycogen phosphorylase, defined in terms of the low level degradation products that might represent breakdown intermediates in vivo (Cookson and Beynon, 1985, 1987). At an early stage in these investigations, it became apparent that the levels of such intermediates were vanishingly low and that very sensitive methods, based on monospecific antibodies, would be required for their detection. In contrast to many applications based on western blotting, we required that the detection methods be enhanced such that the parent protein was overstained and the degradation products became visible. Thus, our approach was vulnerable to artefacts such as limited proteolysis that might generate small but significant amounts of degradation products in vitro. Furthermore, this was more likely to occur in heterogeneous, whole tissue preparations than with pure phosphorylase. Pure (recrystallised) phosphorylase was used to define the limits of the detection systems based on the monoclonal PLP-antibody and the polyclonal phosphorylase antiserum. We were surprised to observe that sample preparation for S D S - P A G E resulted in a low degree of fragmentation of the protein, virtually undetectable by
Fig. 1. Fragmentation of glycogenphosphorylaseduring samplepreparation for SDS-PAGE. Highly purified rabbit muscleglycogenphosphorylasewas treated with sodium cyanoborohydrideif required ("NaCNBH3 + "), heated at 100~ for up to 10 minutes in samplebuffer(see Methods) and separated on SDS-PAGE (7.5/~g protein/well). Some lanes were electroblotted onto nitrocellulose and subsequently probed using a monoclonalantibody to pyridoxalphosphate ("Anti-PLP") or an affinity purified polyclonal antibody to phosphorylase ("Anti-phosphorylase")..A separate region of the gel, stained for protein, is included for comparison ("Protein").
Protein Degradation in vitro
213
protein staining of the gel but readily detected by the sensitive immunochemical techniques (Fig. 1). It is very likely that these fragments are derived from phosphorylase as they react with both antibodies and reactivity with the P L P monoclonal requires that the proteins are first treated with sodium cyanoborohydride to reduce and stabilise the aldimine linkage between cofactor and protein (Butler et al., 1985). The fragments are generated during the treatment of the protein with sample buffer at 100~ and become more intense as the incubation time is increased from 2 to 10 minutes. Several fragments react with the polyclonal antibody but not the monoclonal antibody. Such fragments have either lost cofactor or were derived from a region of the polypeptide chain that did not include the PLP-binding residue Lys-680 (Johnson et al., 1987). These apo-fragments are also identifiable by their failure to shift in mobility upon reduction with sodium cyanoborohydride. The increase in apparent molecular weight on SDS-PAGE is a feature of phosphorylase and phosphorylasederived PLP-binding peptides that we have observed repeatedly (Butler et al., 1985; Cookson and Beynon, 1987). Trace contamination by proteases has the potential to cause slight fragmentation, particularly as the substrate is denatured by a combination of SDS and an increase in temperature and subsequently exposed in this more vulnerable conformation to the protease (Beynon, 1987). Mixtures of protease inhibitors were added to the phosphorylase samples prior to preparation for SDS-PAGE, but no change in banding pattern was observed (Fig. 2). The mixtures of inhibitors were chosen as those most
Fig. 2. The effectof protease inhibitors on fragmentation of glycogen phosphorylase.A solution of pure phosphorylasewas supplementedby a cocktail of protease inhibitors (see Methods) beforeundergoingsample preparation for SDS-PAGE (7.5 #g protein/lane).After separation, the proteins were electroblotted onto nitrocellulose and were probed using antibodies to pyridoxalphosphate ("Anti-PLP") or phosphorylase("Anti-phosphorylase"). A separate region of the gel, stained for protein, is included for comparison ("Protein").
214
Cookson and Beynon
effective at suppressing muscle proteases which, in combination with the extremely high degree of purity of the phosphorylase, renders advantitious proteolysis most unlikely as an explanation for the fragmentation. O u r research into phosphorylase degradation focusses on the behaviour of the enzyme in vivo or in complex and heterogeneous tissue preparations. A high speed supernatant from mouse skeletal muscle, containing phosphorylase (approximately 5 ~o of the total soluble protein) and other soluble muscle proteins, was also prepared for SDS-PAGE. Extended boiling (10 minutes) causes significant fragmentation of phosphorylase and major fragments are visible after only 2 minutes treatment (Fig. 3). The fragments are discrete, suggesting some type of site-specific attack upon phosphorylase. However, proteinase inhibitors were without effect on the concentration or nature of the fragments; proteolysis can therefore be dismissed as a cause of the phenomenon, even in crude tissue preparations. Sample preparation requires at least a short treatment at 100~ to eliminate smearing that is otherwise detected by the monoclonal antibody. It is noteworthy that the smearing is not due to PLP-containing proteins because it can be observed whether or not the samples are treated with sodium cyanoborohydride. This implies that smearing is due to nonspecific absorption of the monoclonal antibody.
Protein
Stds
Anti-PLP
97~ 68--
45--
31
100~
m
....
0
2
10
10
NaCNBH3 ............... Inhibitors ...............
_
_
+
10
0
2
10
0
2
10
+
+
+
+
+
+
+
-
+
+
+
+
+
10
Fig. 3. Fragmentation of phosphorylasein crude tissue preparations derived from mouse skeletal muscle. A high speed supernatant from mouse skeletal muscle was prepared in the presence or absenceof a cocktailof proteaseinhibitors (seeMethods) beforeundergoing samplepreparation for SDS-PAGE (50#g protein/lane). After separation, the proteins were electroblotted onto nitrocellulose and were probed using antibodies to pyridoxal phosphate ("Anti-PLP"). A separate region of the gel, stained for protein, is included for comparison ("Protein").
Protein Degra~tation in vitro
215
The amount of protein fragmented during the preparation of samples for SDSP A G E is minute, the intensity of the parent protein band is comparable in samples of pure phosphorylase heated for 10 minutes to untreated samples (Fig. 1). Although this phenomenon is likely to be a characteristic of all studies involving S D S - P A G E it only creates problems when a technique is employed which has the specificity and sensitivity to detect the fragments. Different fragments of phosphorylase, containing PLP, are present in the same sample of soluble mouse muscle proteins when heated for different lengths of time (Fig. 3). A band at approximately 40,000 Mr is probably due to transaminase(s); these appear to require at least a short incubation at 100~ in order to form a discrete band. The band at 55,000 Mr is also enhanced by a short period of heating, but as this band is present in all lanes it is not a PLP-protein, although it too appears to be degraded after 10 minutes incubation. Furthermore, the fragments produced on heating are in the size range and abundance expected of degradation intermediates formed in vivo (Beynon et al., 1985b). Thus care is needed in the interpretation of immuno-reactive bands, of lower molecular weight than the native protein, as fragments derived from the intracellular degradation of the protein. S D S - P A G E is commonly used to assess the purity of proteins and, in conjunction with western blotting, to determine the specificity of antisera. We caution that fragments produced during sample preparation might be misinterpreted as contaminants.
ACKNOWLEDGEMENTS This work was supported by grants from the Medical Research Council (G840/7575SB) and the Muscular Dystrophy G r o u p of Great Britain (RA3/162). The monoclonal antibody to P L P was a generous gift from Dr J. Cidlowski.
REFERENCES
Beynon, R. J. (1987). In: Methods in Molecular Biology, Vol. 3 (J. Wilson, Ed.), in press. Beynon, R. J., Cookson, E. J. and Butler, P. E. (1985a). Biochem. Soc. Trans. 13:1005-1007. Beynon, R. J., Place, G. A. and Butler, P. E. (1985b). Biochem. Soc. Trans. 13:306-308. BUtler, P. E., Fairhurst, D. and Beynon, R. J. (1984). Anal. Biochem. 141:494-498. Butler, P. E., Cookson, E. J. and Beynon, R. J. (1985). Biochim. Biophys. Acta 847:316-323. Cookson, E. J. and Beynon, R. J. (1985). Biochem. Soc. Trans. 13:1168-1169. Cookson, E. J. and Beynon, R. J. (1987). In: Adaptive Mechanisms of Muscle (I. Sohar and J. W. C. Bird, Eds.), in press. Hannibal-Friedrich, O., Chun, M. and Sernetz, M. (1980). Biotechnol. Bioeng. 22:157-175. Johnson, L. N., Hajdu, J., Acharya, K. R., Stuart, D. I., McLaughlin, P. J., Oikonomakos, N. G. and Barford, D. (1987). In: Allosteric Proteins (G. Heive, Ed.), CRC Press, in press. Kowit, J. D. and Maloney, J. (1982). Anal. Biochem. 123:86-93. Reznick, A. Z., Rosenfelder, L., Shpund, S. and Gershon, D. (1985). Proc. Natl. Acad. Sci. USA 82:61146118. Studier, F. W. (1973). J. Mol. Biol. 79:237-248. Towbin, H., Staehelin, T. and Gordon, J. (1979). Proc. Natl. Acad. Sci. USA 76:435ff4354. Viceps-Madore, D., Cidlowski, J. A., Kittler, J. M. and Thanassi, J. W. (1983). J. Biol. Chem. 258:26892696. Wilson, J. E. and Smith, A. D. (1985). J. Biol. Chem. 260:12838-12843.
