Elecrrophoresis 1993, 14, 847-851 Native gel eleclroblotting of proteolytic products for sequencing 847

Feng Wang Chang Su Kurt Hollfelder Doreen Waddington Yu-Ching E. Pan

Biotechnology Department of the Molecular Sciences Group, Hoffmann-La Roche Inc., Nutley, NJ

Electroblotting proteolytic products from native gel for direct N-terminal sequence analysis: An approach for studying protein-protein interaction

Proteins which are electroblotted from native gels onto polyvinylidene difluo- ride (PVDF) membranes are suitable for detailed structural analysis. This method, in conjunction with limited proteolysis and N-terminal sequencing, has been used to study the molecular interactions between native protein mol- ecules. The interaction between recombinant interleukin-2 (rIL-2) and its receptor (rIL-2Ra) was examined as a model system. The working strategy consists of (i) proteolysis of rIL-2Ra and rIL-2RalrIL-2 complex, (ii) separ- ation of the major proteolytic products by native polyacrylamide gel elec- trophoresis followed by electroblotting onto PVDF membrane, and (iii) sequence analysis of the blotted protein bands for the identification of peptide regions sensitivc to proteolysis. Results have indicated that the exon 3 encoded region in rIL-2Ra is sensitive to proteolysis regardless whether it is complexed with rIL-2 or not. This suggests that no major conformational changes occur in rIL-2Ra during interaction with rIL-2. This electroblotting approach is, therefore, useful for studying protein-protein interaction in solu- tion.

1 Introduction

Electroblotting has been widely used to recover proteins from sodium dodecyl sulfate (SDS)-polyacrylamide gels onto polyvinylidene difluoride (PVDF) membranes for structural analyses including amino acid composition, N-terminal sequence analyses and peptide mapping [l-31. This method has been expanded to recover pro- tein-protein complexes by electroblotting from native polyacrylamide gels onto PVDF membranes [4]. In native polyacrylamide gel electrophoresis (PAGE), pro- teins migrate according to their charge, size, shape, and other innate structural features [5,6]. By exploiting the nondenaturing aspects of native PAGE, protein-protein interactions in solution can be evaluated. For example, the stoichiometry of protein-protein interaction can be determined through a structural analysis of blotted pro- teins. This was previously demonstrated with the recom- binant forms of interleukin-2 (rIL-2) and the a subunit of the interleukin-2 receptor (rIL-2Ra) [4]. The results have provided direct evidence that rIL-2 binds to rIL- 2Ra in a 1:l ratio. Studying the interaction between rIL- 2 and rIL-2Ra as a model system, this native gel electro- blotting approach has been utilized, in conjunction with limited proteolysis and N-terminal sequence analysis, to identify peptide regions which are sensitive to proteol- ysis. Evaluation of these results provides insight into the molecular interactions between the two molecules. This

Correspondence: Dr. Yu-Ching E. Pan, Biotechnology Department, Hoffmann-La Roche Inc., 340 Kingsland S t . , Nutley, NJ 07110, USA

Abbreviations: CAPS, 3- (cyclohexy1amino)-1-propanesulfonic acid; IL-2, interleukin-2; IL-ZRAa, the a subunit of IL-2 receptor; rIL-2, recombinant IL-2; PAGE, polyacrylamide gel electrophoresis; FTH, phenylthiohydantoin; PVDF, polyvinylidene difluoride; TCA. trichlo- roacetic acid; SDS, sodium dodecyl sulfate; TEMED, N,N,N‘,W-tetra- methylethylenediamine

* Current address: Cell Growth Factor Department, Imclone Systems Inc., 180 Variack Street, New York, NY 10014, USA

paper is intended as a communication of the potential uses of this approach and not as a detailed characteriza- tion of the interaction between rIL-2 and its receptor. The strategy for analyzing the interaction between rIL- 2Ra and rIL-2 is outlined in Fig. 1. The complex was first subjected to limited proteolysis and the major pro- teolytic products in the digests were then resolved by native PAGE and electroblotted onto PVDF membranes. The protein bands visualized on PVDF were sequenced for the identification of regions sensitive to proteolysis. Noncomplexed rIL-2Ra was also treated in the same manner to recover its major digestion products. Through a comparison of the digestion patterns and sequence results between the complexed and non-complexed rIL- 2Ra, some structural information of rIL-2lrIL-2Ra can

rIG2Ra riL-2Ra I I

I rIG2

! Native PAGE

Electroblotting 1 1 Sequence analysis of proteolytic products on PVDF

Figure 1. Strategy for analyzing proteolytic rIL-Z/rIL-ZRa complex with the use of rIL-2Ra as a control. (rIL-WrlL-2Ra)’ and (rIL-2Ra)’ stand for protein proteolytic digests.

0 VCH Verlagsgesellschaft mbH, 69451 Weinheim, 1993 0173-0835/93/0909-0847 $5.00+.25/0

848 F. Wang er a!. Elearoyhoresk 19Y3, 14, 847-851

be deduced. The approach used in this investigation may be applicable for studying other protein-protein interac- tions.

2 Materials and methods

2.1 Apparatus

An X-cell (Novex, San Diego, CA) was employed for native PAGE. Electroblotting was performed in a Trans- bot Cell (Bio-Rad, Richmond, CA). Sequence analyses of the blotted PVDF protein/peptide bands were per- formed using a model 470A gas-phase sequencer with a model 120A "on-line" phenylthiohydantoin (PTH) amino acid analyzer (Applied Biosystems; Foster City,CA) [7, 81 on either polybrene-treated fiber glass filters or directly in an Applied Biosystems blot-cartridge.

2.2 Protein samples and reagents

Recombinant IL-2 (rIL-2a), cloned and expressed in E. coli., and rIL-2Ra, cloned and expressed in Chinese hamster ovary cells, were purified as previously de- scribed [9, 101. The rIL-2Ra used in this sutdy was a non- glycosylated receptor which was encoded by an expres- sion plasmid containing a 3'-truncated and altered form of IL-2Ra cDNA [l l ] in which the two potential N-glyco- sylation sites (at and Asn6') were replaced by gly- cine. The PVDF membrane used was Immobilon-P (Mil- lipore Bedford, MA). Acrylamide, N,N-methylenebisacryl- amide, ammonium persulfate, riboflavin-5'-phosphate and N,N,N,N-tetramethylethylenediamine (TEMED) were purchased from Bio-Rad, sequencer reagents and solvents were from Applied Biosystems, proteinase K was from Boehringer-Mannheim (Indianapolis, IN), and trypsin from Worthington (Freehold, NJ). Other chemi- cals and solvents were of the highest purity available.

2.3 Formation of rIL-2IrIL-2Ra complex and limited proteolysis procedure

The protein/protein complex was prepared as previously described [4]. Equal moles of rIL-2 and rIL-2Ra with concentrations ranging from 0.05 to 0.5 wg/pL were mixed in phosphate buffered saline (PBS), pH 7.4, for at least 5 min before subjecting to proteolytic treatment. Trypsin and proteinase K digestions of rIL-2Ra and its rIL-2 complex were performed in PBS (pH 8.5) at 25" with an enzyme-to-substrate ratio of 1:20 for 5-10 min. Calcium acetate, which is commonly used as an activator in proteinase K digestion, was omitted in the digestion buffer to decrease the digestion rate. After digestion, the samples were adjusted to a final concentration of 10% sucrose with a 0.05% Bromophenol Blue marker in a total volume of 10-15 pL and immediately subjected to native PAGE.

2.4 Native PAGE

Native PAGE analyses of rIL-2Ra, rIL-2, rIL-2RalrIL-2 and the proteolytic digests of rIL-2Ra and rlL-2RalrlL-2 were performed in a discontinuous buffer by published

Table 1. Native gel composition

Resolving gel Stacking gel Acrylamide 5 vo 3 % N,K-Methylenebisacrylamide 0.13 Yo o.osn/o Tris-HCl, pH 8.5 240 m M Tris-phosphate, pH 6.9 40 mM Ammonium persulfate 0.0075 Ya 0.03 % Riboflavin 0.00025~/0 0.001~/0 TEMED 0.2 o/o 0.1 Vn

procedures with some modifications [4, 121. The compo- sitions of the stacking and resolving gels are listed in Table 1. The gels were prepared in Novex casettes (8.5 X 7.5 cm) and acrylamide polymerization was allowed to proceed at room temperature. Both anode and cathode reservoirs contained 65 mM Tris-HC1, pH 7.5. The pro- tein samples were electrophoresed towards the anode (+) at 25'' C for 30 min at 70 V followed by 110 min at 150 V until the dye front reached the bottom of the gel. The samples were then immediately electroblotted onto PVDF.

2.5 Electroblotting and staining

Prior to blotting, the PVDF membranes were rinsed with 100% methanol and then equilibrated in 10 mM cyclohe- xylamino-1-propanesulfonic acid (CAPS), pH 11, without methanol. Two PVDF sheets were used for each transfer. Blotting was conducted at 25" C at constant 400 mA for 3-5 rnin. After blotting, PVDF membranes were stained with Coornassie Brilliant Blue R-250 (0.1% in 50% meth- anol) for 5 min and destained in 50% methanol, 10% acetic acid solution. The native gels were first treated with 12% trichloroacetic acid (TCA) and then stained with Coomassie Brilliant Blue G-250 (0.25% in water) to verify complete protein transfer to the membranes [4, 131.

2.6 Data acquisition, storage, and analysis

Data obtained from sequence analyses were collected, stored and analyzed with an IBM computer using Tur- bochrom I11 software (PE-Nelson, Cupertino, CA).

3 Results and discussion

3.1 Native PAGE and eledroblotting

Native gels were freshly prepared for analyses. The effec- tiveness of riboflavin for acrylamide polymerization is enhanced by adding a small amount of ammonium per- sulfate. The excess persulfate ions apparently do not affect the integrity of native rIL-2, rIL-2Ra, rIL-2IrIL- 2Ra nor their migration through this gel system, as is de- scribed below. The proteins analyzed in this work are stable in the range of pH 7-9. The same Tris-HC1 buffer (pH 9.8) was used in both the upper and lower reservoirs and the samples exhibited similar migration patterns as obtained previously by using a Tris-Gly pH 7.5 buffer in the lower reservoir [4]. Since native proteins are separ- ated on the basis of charge, size, and conformation, the major factors affecting mobility of proteins in native

Electrophoresis 1993, 14, 847-851 Native gel electroblotting of proteolytic products for sequencing 849

PAGE are the pH of the gel and the gel concentration [5 ] . It is essential that the pH under which electro- phoresis is carried out can retain the biological activity and stability of the proteins. Many proteins have isoelec- tric points in the range of 4-7; thus buffers in the pH range of 8-9 are commonly used. Note that the buffer condition used here is not appropriate for analyzing pro- teins with isoelectric points higher than 9. Those pro- teins could migrate in the reverse direction into the upper reservoir. As in the case of SDS-PAGE, the optimal gel concentration used in native PAGE can be determined by measuring the mobility and band width of each protein in a series of gels of different acrylamide concentration.

Electroblotting of protein bands from a native gel can be achieved within a few minutes with protein/peptide sam- ples ranging from 0.4 to 4 pg/well, being retained on a single PVDF sheet [4]. This quantity is usually enough for obtaining structural information using micro-analy- tical techniques. Prolonging electroblotting time will not result in an “over-transfer” of proteins onto a second PVDF sheet. This high efficiency transfer can be attrib- uted to the absence of SDS in native gels and the transfer buffer. It is consistent with the recent finding that a low SDS concentration, in SDS gels and transfer buffer, permits effective binding of proteins to the PVDF membrane [14]. Methanol is eliminated in the electro- blotting step to avoid the loss of protein and the disso- ciation of the protein complex. When staining the gel a 12% TCA was first used to precipitate proteins to pre- vent gradual resolubilization. The precipitated protein bands were then stained by water-based Coomassie Blue G-250 [13]. Several ligand-receptor complexes, such as interferon-y/receptor and IL-S/receptor complexes, have been analyzed by the same native gel system, and satis- factory structural information has been obtained [15, 161.

3.2 Native PAGE analysis and electroblotting of rIL- 2Ra, IL-ZIrIL-2Ra complex and their proteolytic digestion products

Recombinant IL-2, rIL-2Ra, rIL-2IrIL-2Ra complex, and the major proteolytic fragments of rIL-2Ra, rIL-2Ra/rIL- 2 after proteinase K and tryptic digestions were resolved by native PAGE and electroblotted onto PVDF. As dis- cussed above, in native PAGE the protein mobility in the electric field depends upon the net charge, size, and comformation of the molecule, along with other innate structural features. As shown in Fig. 2, rIL-2 (band A) migrates slower than rIL-2Ra (band B) in native PAGE even though it has a lower molecular weight (the apparent molecular weights of rIL-2 and rIL-2Ra in SDS PAGE are 15 and 30 kDa, respectively). As expected, the rIL-2hIL-2Ra complex (Fig. 2, band C) migrates between the uncomplexed protein molecules. The iden- tity of this complex had been confirmed previously by amino acid and N-terminal sequence analyses [4]. Protein- ase K fragmentation of the rIL-2Ra and rIL-2Ra/rIL-2 complex produced a single band in each digest (Fig. 2, bands D and E). Each product migrated slower than the corresponding intact molecules (bands B and C). In the case of tryptic products, two new bands were detected in both complexed and noncomplexed rIL-2Ra digests (Fig. 2, bands F through I) with mobilities slightly dif- ferent from those of the intact protein molecules. These mobility changes indicate that cleavages occurred in both samples during limited proteolysis. The different mobility of these proteolytic bands is due to the loca- tions of various sensitive cleavage sites of the samples (for further structural information see Section 3.4).

A

NO DIGESTION

C

D E

F H G I

Figure 2. Native PAGE of rIL-2, rIL-2Ra, and rIL-ZhIL-2Ru as well as the proteo- lytic digests of rIL-2Ra and rIL-2IrIL- 2Ra complex after electroblotting onto PVDF membrane. Detailed procedures

PROTEINASE K TRYPSIN can be found in Section 2.

850 F. Wang e t a / . Elecrrophoresis 1993, 14, 847-85 I

3.3 Sequence analyses of PVDF samples

In order to identify the electroblotted bands, each was sequenced (Table 2). For the intact proteins, single sequences of Ma-A'-P-T-, and E'-L*-(C)-D- were detected from rIL-2, and rIL-2Ra, respectively (Fig. 2, bands A and B); two sequences, corresponding to those of rIL- 2Ra and rIL-2 were obtained from the rIL-2hIL-2Ra complex (Fig. 2, band C). The first two PTH amino acid chromatograms of rIL-2Ra and the complex are shown in Fig. 3 (plots I and 11). In the major bands resulting from the proteinase K digestion of rIL-2Ra and its rIL-2 complex, more than one sequence was detected (Table 2). In the rIL-2Ru digest (Fig. 2, band D), a second sequence, beginning at A9*-S-L-, was detected along with the N-terminal sequence of rIL-2Ru. This indicates the

I II T

- 20 40

, STD

I I

presence of a sensitive proteolytic site between residues Qg7 and A9*. In the digest of rIL-2IrIL-2Ra (Fig. 2, band E), three sequences were detected. One corresponded to the N-terminal of rIL-2Rn and the other two from internal cleavage of the complex. One of them was de- rived from cleavage of rIL-2 between residues S 5 and S6 and the other one was obtained from the same rIL-2Ra site (between Q" and Ay8) as previously detected. The first two PTH amino acid chromatograms of the corre- sponding PVDF samples are shown in Fig. 3 (plots 111 and IV). The sequence results of the PVDF bands from the tryptic digestions are also given in Table 2 (PTH amino acid chromatograms not shown). Multiple sequences were detected in the upper and lower bands (Fig. 2, bands F and G) of the tryptic rIL-2Ra. In addi- tion to the N-terminal sequence of rIL-2Ra, both sam-

111

PK rlL-2Ro

DHAgs sg9

STD

20 40 20

Time (min)

IV

PK rlL-2/rlL-2Ra ;

STD

w 20 40

Figure 3. PTH amino acid chromatograms of the first two sequence cycles of PVDF samples: rIL-2Ra (plot I), rIL-Z/rIL-ZRa complex (plot 11), major proteolytic products after proteinase K treatments of rIL-2Ru (plot III), and rIL-2IrlL-2Ra (plot IV). The chromatograms are obtained from samples equivalent to the bands shown in Fig. 2. Varying amount of samples were applied to sequencer and thercfore results are not quantitative. DHA (PTH-dehydroalanine) is a PTH derivative of serine after Edman degradation.

Electrophoresis 1993, 14, 847-851 Native gel electroblotting of proteolytic products for sequencing 85 1

Table 2. Sequence results of PVDF bands 4 Concluding remarks

Electrophoresis through polycrylamide gels is one of the simplest and most widely used methods for character- izing protein molecules. In the absence of detergents and denaturing agents, the native PAGE system de- scribed in this investigation can be used to recover a pro- tein-protein complex by the electroblotting technique. In conjunction with limited proteolysis and electroblotting, proteolytic digestion products can be recovered for fur- ther analyses including N-terminal sequencing and the information can be used to gain insight into the molec- ular interactions between two protein molecules.

We would like to thank Mr. P. Bailon and Dr. F: Khan f o r the supply of rIL-2Ra and rlL-2 and Dr. A. Stern .for reviewing the manuscript.

PVDF band Digestion From rIL-2 From rIL-2Ra A None MU-A-P-T-~'

B None E~-L-(c)-D-~ C None MO-A-P-T- E'-L-(c)-D- D Proteinase K E'-L-(c)-D-

E Proteinase K S6-T-K-K- E'-L-(c)-D-

F Trypsin E'-L-(c)-D-

G Trypsin E'-L-(C)-D-

H Trypsin MO-A-P-T- E'-L-(c)-D-

I Trypsin MO-A-P-T- E~-L-(c)-D-

A9E-S-L-P-

A98-S-L-P-

Ks4-T-?'-E-

K84-T-T- E. TE5-T-E-M-

Ka4-T-T-E-

KS4-T-T-E- TS5-T-E-M-

a) Mo in rIL-2 was not removed during the post-ribosomal processing. b) PTH-cysteine can not be positively identified.

ples contained a sequence derived from a susceptible tryptic site at RS3. An additional sequence was detected in the lower band (band G), indicating the presence of an accessible site at the C-terminal side of KS4. The two bands from the digested complex (Fig. 2, bands H and I) also contain multiple sequences. In addition to the sequences previously detected in the noncomplexed rIL- 2Ra (bands F and G), a sequence corresponding to the N-terminus of rIL-2 was also detected.

3.4 Sensitive proteolytic sites and their structural implications

The interaction of IL-2 with the three subunits of IL-2R (a, f3 and y) is a key regulatory event during an immune response [17, 181. The detailed structural information of the interaction between rIL-2 and its receptor will contribute to a better understanding of this system. In a previous investigation using limited proteolysis, an IL-2 binding core of a glycosylated rIL-2Ra had been identi- fied, fully characterized, and was found to consist of two major homologous peptide domains joined by five disul- fide bonds [19]. A hydrophilic peptide segment encoded by exon 3, covering residues 65-101, was found to be sensitive to proteolysis and not required for IL-2 binding. Similar results were obtained in this study. As given in Table 2, in this non-glycosylated rIL-2Ra mole- cule, the sensitive proteolytic sites are also located in the exon 3-encoded region. Furthermore, using this native PAGE electroblotting approach, various proteolytic digestion products of the rIL-2IrIL-2Ra complex were also recovered, and the exon 3-encoded region in rIL- 2Ra was shown to be sensitive to proteolysis. Although many potential proteolytic sites are present throughout the sequence of rIL-2Ra, the results indicate that they are conformationally protected from proteolysis, regard- less whether this receptor subunit is complexed with rIL- 2 or not. Therefore, it can be concluded that no major conformational change occurred in the rIL-2Ra while interacting with rIL-2.

Received February 8, 1993

5 References

Matsudaira, P., J. Biol. Chem. 1987, 262, 10035-10038. Hulmes, J. D., Miedel, M. C. and Pan, Y.-C. E., in: Hugli, T. E. (Ed.), Techniques in Protein Chemistry, Academic Press, San Diego, CA 1989, pp. 7-16. Aebersold, R. H., Leavitt, J. , Saavedra, R. L. and Hood, L. E., Proc. Natl. Acad. Sci. USA 1987, 84, 6970-6974. Wang, F. and Pan, Y.-C. E., Anal. Biochem. 1991, 198, 285-291. Hames, B. D., in: Hames, B. D. and Rickwood, D. (Eds.), Gel Elec- truphoresis of Proteins, IRL Press., Oxford 1981, pp. 1-91. Goldenberg, D. P., in: Creighton, T. E. (Ed.), Protein Structure, IRL Press, Oxford 1990, pp. 225-250. Hewick, R. M., Hunkapiller, M. W., Hood, L. and Dreyer, W. J., J. Biol. Chem. 1981, 256, 7990-7997. Hunkapiller, M. W., Granlund-Moyer, K. and Whiteley, N. W., in: Shively, J. E. (Ed.), Methods of Protein Characterization, The Humana Press, Clifton, NJ 1986, pp. 315-327. Bailon, P., Weber, D. V., Keeney, R. F., Fredericks, J. E., Smith, C. Familletti, P. C. and Smart, J.E., BidTechnologv 1987, 5, 1195- 1198. Weber, D. V., Keeney, R. F., Familletti, P. C. and Bailon, P., J . Chromatogr. 1988, 431, 55-63. Cullen, B., Podlaski, F. J., Peffer, N. J., Hosking, J . B. and Greene, W. C., J. B i d . Chem. 1988, 263, 4900-4906. Davis, B. J., Ann. NY Acad. Sci 1964, 121, 404-427. Diezel, W., Kopperschlager, G. and Hofmann, E., Anal. Biochem.

Mozdzanowski, J., Hembach, P. and Speicher, D. W., Electro- phoresis 1992, 13, 59-64. Fountoulakis, M., Lorenzo, E. T., Juranville, J.-F. and Manneberg, M., Anal. Biochem. 1993, 208, 270-276. Devos, R., Guisez, Y., Cornelis, S., Verhee, A,, Van der Heyden, J., Manneberg, M.? Lahm, H.-W., Fiers, W., Tavernier, J. and Plae- tinck, G., J. Biol. Chern. 1993, 268, 6581-6587. Waldmann, T. A., J. B i d . Chem. 1991, 266, 2681-2684. Takeshita, T., Asao, H., Ohtani, K., Ishii, N., Kumaki, S., Tanaka, N., Munakata, H., Nakamura, M. and Sugamura, K., Science 1992,

Miedel, M. C., Hulmes, J. D. and Pan, Y.-C. E., J . Biol. Chem.

1912, 48, 617-620.

257, 379-382.

1989, 264, 21097-21105.