ANALYTICAL BIOCHEMISTRY 174,204-208 (1988)

Detection of Cellulose-Binding Proteins in Electrophoresis Gels by Filter Paper Affinity Blotting

LARRY MONTGOMERY AND YUNG-KANG Fu

Department ofAnimaiSciences, University oflllinois. Urbana, Illinois 61801

Received April 4, 1988

Proteins separated in electrophoresis gels were tested for the ability to bind cellulose by a simple blotting procedure. Proteins were blotted onto Whatman No. 1 filter paper by diffusion or by electrophoretic transfer and detected by Coomassie blue staining. Certain proteins released into culture supernatant by Bacteroides succinogenes NR9 (ATCC 43854) adhered strongly to cellulose, but were not found to have carboxymethylcellulase activity. Boiling of samples prior to electrophoresis eliminated the ability of proteins to bind to cellulose. Proteins that did not adhere to filter paper cellulose were detected on a nitrocellulose membrane placed behind the filter paper during electrophoretic transfer. The technique, referred to as filter paper affinity blotting, detects cellulose-binding proteins with great sensitivity. Q 1988 Academic press. hc.

KEY WORDS: gel electrophoresis, proteins: binding assays; lectins; cellulases (glycosidases); zymograms.

Enzymes of the cellulase complex are ex- pected to adsorb to their solid substrate (1); this property has been exploited in purifica- tion of some such enzymes (2). There have also been reports of cellulases that do not ad- here to cellulose under conditions tested (3). Other proteins, which do not have hydrolytic activity, may serve as affinity factors binding cells or hydrolytic enzymes to cellulose (4,5).

Cellulose adsorption of cellulases in pro- tein mixtures has generally been tested in batch experiments, with subsequent column chromatography to determine which en- zymes have been adsorbed (3). Cellulose has also been used as a column packing in affinity chromatography, with differential elution to separate proteins on the basis of adsorption (2). This report describes a simple method for detecting cellulose-binding proteins in pro- tein mixtures separated by gel electrophore- sis, analogous to the detection of enzymes in gels using zymograms.

MATERIALS AND METHODS

Proteins. Bacteroides succinogenes strain NR9 (ATCC 43854) was grown in 2XAB

medium (6) with cellobiose as the energy source. After removal of cells by centrifuga- tion, the supernatant was passed through an H 1 P 100-20 hollow-fiber filter (molecular weight cutoff, 100,000; Amicon, Danvers, MA) to remove membrane fragments. Pro- teins in the filtrate were concentrated on an H 1 Pl O-20 filter (cutoff, 10,000) and further concentrated on a PM10 filter (cutoff, 10,000: Amicon) in a stirred cell. Culture me- dium was replaced by diafiltration with 100 mM sodium phosphate buffer, pH 6.8. Pro- teins were estimated by the Lowry method (7). Molecular weight standards for electro- phoresis were from Sigma Chemical Co. (St. Louis, MO).

Electrophoresis and zymograms. Sodium dodecyl sulfate-polyacrylamide gel electro- phoresis (SDS-PAGE)’ (8) was performed with 11% (2.7% crosslinking) polyacrylamide gels ( 14 cm X 16 cm X 1.5 mm). Except where noted, samples were treated at 37°C

’ Abbreviations used: CMC, carboxymethylcellulose; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; CBP, cellulose-binding protein.

0003-2697188 $3.00 Copyright CI 1988 by Academic Press. Inc. All rights of reproduction in any form reserved.

204

FILTER PAPER AFFINITY BLOTTING OF CELLULOSE-BINDING PROTEINS 205

for 20 min in sample buffer (8). Proteins in PAGE gels were visualized by silver staining (9) or by staining with 0.125% Coomassie brilliant blue R-250 (Sigma) in methanol: H20:acetic acid (50:40: lo), with destaining in methanol:H20:acetic acid (50:40: 10; Destain I) for 1 h and then in H*O:acetic acid:metha- no1 (88:7:5; Destain II).

Carboxymethylcellulases were detected by a modification of the zymogram technique of Beguin (10). The gel was rinsed in 100 mM sodium phosphate (pH 6.8) and then incu- bated at 38°C for 3 h in contact with a slab containing 1.25% carboxymethylcellulose (CMC, Sigma low viscosity) and 1% agar in 100 mM sodium phosphate (pH 6.8). The slab was stained for 20 min with 1% Congo red (Sigma) and rinsed with 1 M NaCl lo visu- aIize bands of CMC hydrolysis.

Filter paper ajinity blotting. To detect proteins binding specifically to cellulose, buffer-rinsed SDS-PAGE gels were placed in contact with sheets of Whatman No. 1 filter paper. For diffusion blotting, filter paper dampened with phosphate buffer was placed on top of the gel, which in turn was placed on top of the CMC agar slab. For electrophoretic transfer, the gel was rinsed for 30 min in transfer buffer (0.3% Tris, 1.44% glycine, 20% methanol) and placed in an electrophoretic transfer apparatus (Bio-Rad, Richmond, CA) with a filter paper sheet, then a nitrocellulose sheet, on the anodic side of the gel. Proteins were transferred for 20-24 h at 30 V (4°C).

Nitrocellulose blots were dried at room temperature; filter paper blots were washed for 30 min in 100 mM sodium phosphate buffer (pH 6.8) with constant agitation and kept moist until stained. Blots were stained for 20 min (with agitation) in Coomassie blue stain as used for PAGE gels. After rinsing in distilled H20, blots were destained in Destain I until protein bands were distinct, ca. 1 h. Blots were then washed in distilled HZ0 and air-dried. Optimal destaining time must be determined empirically; used Destain I re- moves stain more gradually, perhaps giving more control over the final contrast.

To obtain good contrast, blots were photo- graphed through a red filter (25A) with Ko- dak Technical Pan film which was processed in Kodak D19 developer. Kodak TMax 100 film was used for silver-stained gels.

RESULTS

As part of an investigation of cellulose deg- radation by the intestinal bacterium B. succi- nogenes NR9 (6), enzymes degrading CMC were detected using CMC-agar zymograms following separation of extracellular proteins by SDS-PAGE (samples were not boiled). We anticipated that CMCases would bind to cellulose and speculated that they could then be detected by staining. This was tested by placing Whatman No. 1 filter paper, wetted with buffer, in contact with the electrophore- sis gel during its 3-h incubation on CMC- agar. With soluble extracellular proteins as the sample, staining of the filter paper with Coomassie blue under empirically deter- mined conditions revealed two strongly stained bands (Fig. 1A) which corresponded with major protein bands in the silver-stained gel (Fig. IC), but not with the CMCases de- tected in the zymogram (Fig. 1B).

Because evidence of enzymatic activity is absent, these proteins are referred to as cellu- lose-binding proteins (CBPs). This designa- tion describes a property of the proteins, but does not impute to them any physiological function in cellulose binding, although such a function may later be discovered. Under other conditions (e.g., absence of SDS) these proteins might exhibit CMCase or other en- zymatic activity. Several other CBPs, present at lower levels in culture supernatant, also lacked detectable CMCase activity (Fig. 1 A, lane 5). With this sample, no CBPs were de- tected at locations in the gel correlated with the migration of CMCase activity. In fact, only one of the CMCases was associated with a substantial protein band, as visualized by silver staining. Loading more protein, or opti- mizing conditions for binding, might have al- lowed detection of cellulose binding by these CMCases.

206 MONTGOMERY AND FU

A 6 C 1 2 3 4 5 16 6 14 5

FIG. 1. Comparison of proteins detected by different methods after SDS-PAGE. (A) Cellulose-binding proteins were detected by Midwestern blotting (by diffusion, Materials and Methods). (B) CMCases were detected by Congo red staining ofCMC agar zymogram. (C) Proteins in the gel were silver-stained. Soluble proteins from B. succinogenes NR9 culture supematant were present in the following amounts: lanes 1 and 6, 3.5 ggug; lanes 2 and 3. 7 pgrg; lanes 4 and 5, 14 pg. Samples in lanes 2,4, and 6 were boiled in sample buffer prior to electrophoresis. Locations ofstandard proteins are indicated at right, with molecular weights given in thousands.

Binding of the CBPs was sufficiently strong that increasing the wash time of the blot from 30 min to 2.5 h did not significantly reduce the levels of CBPs detected. To establish that binding to cellulose is a significant and funda- mental property of CBPs, it was desirable to rule out trivial explanations for the phenome- non. When the soluble supernatant proteins were boiled in SDS-PAGE sample buffer, CBPs were essentially undetectable on the blot (Fig. 1A). Silver staining of the gel re- vealed that electrophoretic mobilities of one or both predominant CBPs were decreased by boiling in SDS (Fig. lC), as observed for some other proteins (11).

Because proteins were still detectable by Coomassie blue staining of the gel after 3 h of diffusion blotting (not shown), electropho- retie transfer was used to ensure that all pro- teins would contact the filter paper. This seemed particularly important because the relatively small CBPs might diffuse from the gel more rapidly than larger proteins. After electrophoretic transfer, very low levels of binding to filter paper were detectable for

boiled culture supernatant CBPs and even for molecular weight standards (Fig. 2A). How- ever, all of those proteins were in much higher concentrations on the nitrocellulose membrane which was “behind” the filter pa- per during electrophoretic transfer (Fig. 2B), demonstrating that larger amounts of these proteins passed through the filter paper than adhered to it. In contrast, the major CBP in unboiled samples was undetectable on the ni- trocellulose membrane, indicating that the electrophoretic current did not remove it from underivatized cellulose. Comparison of the stained bands on the filter paper blot indi- cated that binding of CBPs was at least lo- fold lower after boiling (Figs. IA and 2A).

On the basis of visual comparisons, the lower limits of detection for the CBP with the highest electrophoretic mobility were similar for affinity blotting and the silver stain proce- dure used with PAGE gels. The surprising sensitivity of the former method is probably due to the high concentration of protein at the surface of the filter paper. In contrast, Coomassie blue staining of the nitrocellulose

FILTER PAPER AFFINITY BLOTTING OF CELLULOSE-BINDING PROTEINS 207

A B 12 3 4 5 6 123456

FIG. 2. Comparison of cellulose-binding and -nonbinding proteins. Proteins were electrophoretically transferred from an SDS-PAGE gel to filter paper (A); proteins which passed through filter paper adhered to the nitrocellulose membrane behind (B). Soluble proteins from B. succinogenes supernatant were pres- ent in the following amounts: lanes I and 4, I4 pg; lanes 2 and 5, 7 wg; lanes 3, 3.5 wg. Samples in lanes 4 and 5 were boiled prior to electrophoresis. Standard proteins (23 Fg. total) were in lanes 6; molecular weights of standards are as indicated in Fig. I.

membrane gave lower sensitivity than silver staining of the gel for all proteins: difficulty in destaining the membrane may have led to excessive destaining of protein bands.

DISCUSSION

Filter paper affinity blotting (also referred to as Midwestern blotting) is a simple and sensitive technique for screening proteins for the capacity to bind to cellulose. It is more rapid and convenient than column methods, particularly if PAGE is already being used to analyze protein composition of samples. Fur- thermore, the blot provides a durable record of results.

Major CBPs in the soluble-protein fraction of B. sztccinogenes NR9 culture supernatant did not exhibit CMCase activity, nor did CM- Cases bind to cellulose, within the limits of detection. As with enzyme assays, optimiza- tion of conditions would be necessary to test these preliminary conclusions and to screen for additional CBPs. In particular, it seems possible that SDS, included in the PAGE sys-

tem because membrane-containing samples were also under study, may interfere with cel- lulose binding by some as-yet undiscovered CBPs. Further experimentation will be neces- sary to ascertain whether CBPs have undis- covered enzymatic activity.

Specificity of cellulose binding by CBPs is indicated by destruction of that capability by boiling and by the observation that other pro- teins, e.g., PAGE standards, bound weakly, if at all. Discovery of the cellulose-binding properties of these proteins raises questions regarding their possible physiological roles in cellulose degradation. Certain CBPs may have undiscovered enzymatic activity, or may serve as lectins with the physiological function of binding cells or cellulases to cellu- lose. The blotting procedure, or modifica- tions thereof, will be useful in seeking such functions and in studying factors affecting CBP production by B. succinogenes. It may also be useful in determining which (if any) of the 14 polypeptides in the Clostridium thermoceflum cellulosome (5) mediate bind- ing of the cellulosome to cellulose.

208 MONTGOMERY AND FU

ACKNOWLEDGMENTS

We thank Bryan White and Walter Hurley for helpful comments on the manuscript. This work was supported by Project No. 20-0324 from the Agricultural Experi- ment Station, College of Agriculture. University of Illi- nois at Urbana-Champaign.

REFERENCES

I. Ljungdahl. L. G., and Eriksson, K.-E. (1985) in Ad- vances in Microbial Ecology (Marshall, K. C., Ed.), Vol. 8, pp. 237-299. Plenum Press, New York.

2. Halliwell, A., and Griffin, M. (1978) Biochem. J. 169,713-715.

3. Stutzenberger, F. (1987) Lett. Appl. Microbial. 5, I- 4.

4. Leatherwood, J. M. (1973) Fed. Proc. 32, 1814- 1818.

5. Bayer, E. A., Setter, E., and Lamed, R. (1985) J. Bac- teriol. 163,552-559.

6. Montgomery, L., and Macy, J. M. (1982) Appl. Envi- ron. Microbial. 44, 1435- 1443.

7. Lowry. 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) J. Biol. Chem. 193, 265- 275.

8. Laemmli. U. K. (1970) Narure (London) 227,680- 685.

9. Gerton. G. L., and Millette, C. F. (1986) J. Biol. Re- prod. 35, 1025-1035.

10. Beguin, P. (1983) Anal. Biochem. 131,333-336. I 1. Kent, N. E., and Wisnieski, B. J. (1983) J. Bacterial.

156,956-96 1.