614 H. J i cf a / . Electrophoresis 1997, 18, 614-621

Hong Ji' Robert L. Moritz' Gavin E. Reid' Gerd Ritte8 Bruno Catimel' Ed Nice' Joan K. Heath' Sara J. White' Sydney Welt3 Lloyd J. Old3 Antony W. Burgess' Richard J. Simpson'

'Joint Protein Structure Laboratory, Ludwig Institute for Cancer Research (Melbourne Branch) the Walter and Eliza Hall of Medical Research, Parkville, Victoria, Australia 'Ludwig Institute for Cancer Research (Melbourne Branch) 'Ludwig Institute for Cancer Research (New York Branch), Memorial Sloan Kettering

and

Cancer Center, New Yo& NY, USA

Electrophoretic analysis of the novel antigen for the gastrointestinal-specific monoclonal antibody, A33 The murine monoclonal antibody A33 (mAbA33) recognises a human cell membrane-associated antigen selectively expressed in epithelial cells of the lower gastrointestinal tract and > 90% of colonic cancers, but is not detected in a wide range of other normal tissues by immunohistochemical analysis. In phase 1/11 clinical triasl, mAbA33 has been shown to target advanced colon cancers and the humanised version is currently being evaluated in therapy stu- dies. Although the mAbA33 has been well characterised by immunohistoche- mica1 and clinical studies, until recently, the target antigen has remained poorly defined. This was largely attributable to the antigenic determinant recognised by mAbA33 being dependent on the native spatial conformation of the A33 antigen which impeded its identification by conventional two-dimen- sional electrophoresis (2-DE) and immunoblot analysis. We have developed an immunoblot method, based on nonreducinghon-urea precast 2-DE gels, that has facilitated the purification of the detergent (0.3 Yo Triton X-100) solubilised A33 antigen from the human colon cancer cell lines LIM1215 and SW1222. Under these 2-DE conditions, the A33 antigen electrophoreses with an appa- rent Mr - 41000 and p1 5.0-6.0. Attempts to isolate the A33 antigen from 2-DE gels for direct structural analysis were unsuccessful, due to its co-electro- phoresis with actin and cytokeratin proteins. However, using Western blot and biosensor detection the A33 antigen has been purified chromatographically and N-terminal sequence analysis was possible. Using polyclonal antibodies raised against a synthetic peptide corresponding to the N-terminal region of the A33 antigen we have used Western blot analysis to localise the molecule in our master 2-DE protein database for normal human colon crypts and several colon carcinoma cell lines (URL address: http ://www.ludwig.edu.au). Under reducing 2-DE conditions, the A33 antigen electrophoresis as 6 diffe- rentially charged isoforms (pl4.6-4.8) with a single molecular weight species at M. -55000.

1 Introduction

A recently developed murine monoclonal antibody A33 (mAbA33) recognises a cell surface-associated antigen which is selectively expressed on the surface of normal gastrointestinal epithelium and >go% of all primary and secondary colorectal cancers [l , 21. The A33 antigen does not appear to be secreted or shed into the circulation, and cell-bound radiolabelled mAbA33 is rapid interna- lised [3]. Phase 1/11 clinical studies using "'I and '''I labelled mAbA33 demonstrated selective localisation of the radioisotopes to metastatic deposits of colon carci- nomas, modest anti-tumor activity in heavily pretreated patients and a lack of toxicity in the bowel [l, 4, 51.

Although the mAbA33 has been well characterised by immunochemical, immunohistochemical and clinical studies, the identity and the function of the target A33 antigen has remained poorly defined. The purpose of this study was to develop a mAbA33-based immunoblot assay in order to characterise the electrophoretic proper-

Correspondence: Dr. Richard J. Simpson, Ludwig Institute for Cancer Research, PO Box 2008, Royal Melbourne Hospital, Victoria 3050, Aus- tralia (Tel: +61-3-9347-6389; Fax: +61-3-9348-1925; E-mail: simpson@- 1icre.ludwig.edu.au)

Nonstandard abbreviations: CBR-250, Coomassie Brilliant Blue R-250; FCS, fetal calf-serum; mAb, monoclonal antibody

Keywords: Monoclonal antibody A33 / Two-dimensional polyacryl- amide gel electrophoresis I Tandem mass spectrometry I Peptide mass fingerprinting I Human colonic proteins

ties of the target A33 antigen and to facilitate the purifi- cation. Earlier attempts to develop an immunoblot assay for the mAbA33 were unsuccessful since the antigenic determinant recognised by mAbA33 is dependent upon the native spatial conformation of the target antigen. In this study we describe an immunoblot method, based 011 nonreducinghon-urea precast one-dimensional SDS- PAGE and 2-DE gels. Using this assay, we have exam- ined the expression levels of the A33 antigen in a number of human colorectal cell lines, and investigated detergent-extraction conditions of the target antigen from LIM1215 cells. Further, we report the electropho- retic properties of the A33 antigen in nonreducing 2-DE gels using mAbA33 and, using polyclonal rabbit antisera directed towards a synthetic peptide corresponding to an N-terminal region of the target antigen, in conventional reducing 2-DE gels.

2 Materials and methods

2.1 Materials

Acrylamide, Bis, TEMED, DTT, ammonium persulphate, glycine, urea and SDS were purchased from Bio-Rad (Richmond, CA, USA). Formaldehyde was from British Drug House (Poole, UK). Coomassie Brilliant Blue R-250 (CBR-250) was obtained from Pharmacia Biotech (Uppsala, Sweden). Precast IEF slab gels (pH 3-10), 4-2O0T acrylamide gels, linear 10°/oT acrylamide gels, IEF pH 3-10 cathode buffer Cat. # LC5310 (arginine

0 VCH Verlagsgesellschaft mbH, 69451 Weinheim, 1997 0173-0835/97/0304-0614 $17.50+.50/0

Electrophoresis 1997, 18. 614-621 Novcl antigen for the gastrointestinal specific monoclonal antibody, A33 615

and lysine free base), anode buffer Cat. # LC5300 (pho- sphoric acid) and sample buffer Cat. # LC5311 were from Novex (San Diego, CA, USA). RPMI 1640 medium was purchased from Irvine Scientific (Santa Ana, CA, USA). Fetal calf serum (FCS) was from the Common- wealth Serum Laboratory (Melbourne, Australia). Se- quencing-grade modified trypsin (Cat. # V512A, lot No. CM201) was from Promega (Madison, WI, USA). [35S]Methionine and [35S]cystine were from ICN (Seven Hills, NSW, Australia). Nitrocellulose membrane was from Micron Separation Inc. (Westboro, MA, USA). Per- oxidase-labelled anti-mouse antibody and anti-rabbit antibody and the enhanced chemiluminescence detec- tion (ECL) system for immunoblot analysis were from Amersham (Buckinghamshire, UK). The nonionic deter- gent Tween-20 was from Pierce (Rockford, Illinois, USA). The mAbA33, which was raised against a pan- creatic tumour extract, detects a cell surface antigen expressed by 95% of primary or metastatic colon cancer cells and normal colonic epithelium, but not by most other normal tissues and tumour types [I, 41. Murine mAbA33 (IgG2a) was purified from ascites using Pro- tein-A affinity chromatography as described elsewhere [l]. Polyclonal anti-peptide antibody directed towards a synthetic peptide corresponding to part of the N-ter- minal sequence of the A33 antigen (amino acid residues 2-20) [6] was generated in rabbits and immuno-purified using a synthetic A33 antigen peptide affinity column. Anti-G-CSF murine monoclonal antibody (IgG2a) was a gift from Dr. J. Layton (Ludwig Institute Melbourne Branch). Anti-chicken gizzard actin mAbC4, anti-cytoke- ratin mAbK18 were from Boehringer Mannheim (Mann- heim, Germany). Immobiline Drystrips, pH 3-10, were from Pharmacia. PVDF membrane was purchased from Millipore (Milford, MA, USA). Deionized water from a tandem Milli-RO and Milli-Q system (Millipore) was used for all buffers.

2.2 Cell culture and preparation of samples for PAGE

Human colon carcinoma cell line LIM1215 [7] was pas- saged in RPMI 1640 containing 10% v/v FCS. When cells were 80-90% confluent, they were washed three times with PBS, scraped from the tissue culture dishes (150 mm diameter), and then harvested by centrifuga- tion (250 X g). For reducing 2-DE experiments cell pel- lets were lysed by incubation with IEF sample buffer (9 M urea, 2% v/v NP-40, l0h w/v DTT, 2% v/v carrier ampholytes, pH 3-10) for 5 rnin at 25 "C. Insoluble mate- rial was removed by centrifugation (20 000 X g ) for 10 min at 4°C. For one-dimensional SDS-PAGE the cell pellet from a 150 mm culture was directly lysed in 1 mL of nonreducing SDS sample buffer (0.06 M Tris-HC1, pH 6.8, 1% w/v SDS, 10% w/v glycerol) for 5 min at 25°C. Insoluble material was removed by centrifugation as de- scribed above. An enriched preparation of A33 antigen was obtained by extracting the cell pellet from one cul- ture dish (90 mm diameter) of cells with (i) 0.5 mL of 0.5% v/v Tween-20 in 10 mM Tris-HCl, pH 7.5, or (ii) 0.5 mL of 0.3% v/v Triton X-100 in 10 mM Tris-HC1, pH 7.5, for 10 min at 4°C with occasional vortexing (10 s). Inso- luble cell debris was removed by centrifugation (20 000 X g, 10 min, Eppendorf centrifuge) at 4°C. The detergent extracts were then diluted (1:l) with nonreducing IEF

sample buffer (0.5 O/o w/v arginine, 0.7% w/v lysine, 3% v/v glycerol) for 2-DE, or 0.06 M Tris-HC1 (pH 6.8), 1% w/v SDS, 10% glycerol, 0.001% w/v bromophenol blue for nonreducing one-dimensional SDS-PAGE.

2.3 Metabolic labelling of LIM1215 cells with P5S1methionine

Metabolic labelling of LIM1215 cells with ["Slmethio- nine was performed as described elsewhere [8]. Briefly, 2 X lo6 cells were plated in RPMI 1640 medium con- taining 10% v/v FCS in 35 cm2 cell culture dishes and incubated at 37°C in a 5 % v/v CO, atmosphere for 16h. The cells were then washed three times with methionine- free Eagle's minimum essential medium (modified) and cultured in the same medium containing 5% v/v dia- lysed FCS and 500 pCi/mL of ["S]methionine for 6 h at 37 "C. [35S]Methionine-labelled cells were washed three times with ice-cold PBS and then lysed with IEF sample buffer.

2.4 Electrophoretic methods

2.4.1 One-dimensional SDS-PAGE

SDS-PAGE was performed using precast polyacrylamide gels from Novex. LIM1215 cells (- 1 X 10') were lysed in 35 pL of SDS sample buffer (0.06 M Tris-HC1, 1% w/v SDS, 10% w/v glycerol, 0.001% w/v bromophenol blue, pH 6.8) in the presence (reducing conditions) or absence (nonreducing conditions) of 0.5 Yo v/v B-mercaptoe- thanol. Electrophoresis (125 V constant voltage) was car- ried out at 25°C in an XellTM Mini-Cell apparatus (Novex) for about 1.5h using Laemmli SDS running buffer (0.0256 M Tris-HC1, 0.192 M glycine, 0.1% SDS, pH 8.3) [9] until the tracking dye reached the bottom of the separation gel.

2.4.2 Nonreducing 2-DE

Triton X-100 extracts of LIM1215 cells (20 pL) were diluted 1:l with IEF sample buffer (0.58% w/v arginine, 0.7% w/v lysine, 30% w/v glycerol) and loaded at the cathodic end of precast IEF gels (Novex) according to the munufacturer's instructions. IEF was carried out at 100 V for 1 h, 200 V for 1 h, and 500 V for 0.5 h, at 25 "C. Prior to the second-dimensional SDS-PAGE, IEF gel strips were equilibrated at 25°C for 20 min in 10 mL of equilibration buffer (0.12 M Tris-HCl, pH 6.8, 2% w/v SDS, 10% w/v glycerol and 0.001% w/v bromophenol blue). The equilibrated IEF gel strips were applied to a second-dimensional 10 O/oT precast polyacrylamide gel (Novex). Electrophoresis (125 V constant voltage) was performed at 25°C in a XellTM Mini-Cell apparatus (Novex) for - 1.5 h using Laemmli SDS-running buffer (0.0256 M Tris-HC1, 0.192 M glycine, 0.1% SDS, pH 8.3) [lo] until the tracking dye reached the bottom of the separation gel. After electrophoresis, proteins were either subjected to immunoblot analysis or visualised by stain- ing the gel with 0.1% w/v CBR-250 in 50% v/v meth- anol/lO% v/v acetic acid/water and destaining with 12 O/o v/v methanol/7% v/v acetic acid/water.

616 H. Ji et ul

2.4.3 Reducing 2-DE

Electrophoresis 1991, 18, 614-62 I

2-DE (reducing conditions) was performed as described elsewhere [ 111. First-dimensional IEF using precast IPG gel strips (Pharmacia) was performed in a Pharmacia Multiphor I1 gel apparatus. Briefly, the IPG gel strips (pH 3-10, linear 18 cm), were rehydrated in 8 M urea, 0.5% v/v Triton X-100, 10 mM DTT and 0.0114% v/v acetic acid for 6 h and then lightly blotted between two sheets of water-saturated filter paper. Cell lysates (50-100 pL), either unlabelled (from 3 X lo6 cells) or "S- labelled (from 5 X los cells) were loaded into each IPG 100 pL sample well. After focusing for 175 000 Vh, the IPG gel strips were equilibrated (2 X 20 min) in 20 mL of equilibration buffer (0.12 M Tris-HC1, pH 6.8, 2% w/v SDS, 1% v/v mercaptoethanol, 20% v/v glycerol and 0.001% bromophenol blue) and then applied to second- dimensional 12%T SDS-polyacrylamide gels (160 X 160 X 1 mm). Electrophoresis was carried out at 25°C in a Bio-Rad Protein I1 multi-cell at 70 V for 16h using Laemmli buffer [9]. After electrophoresis, proteins were visualised using either CBR-250 or by silver staining as described previously [ll]. 35S-labelled proteins were visualised either with a PhosphorImager (Molecular Dynamics, Sunnyvale, CA, USA) or by autoradiography using X-OMATTMAR X-ray film (Kodak, Rochester, NY, USA).

2.5 Immunoblot analysis

SDS-PAGE- or 2-DE-resolved proteins were electro- transferred onto PVDF or nitrocellulose membranes using a Bio-Rad Trans-blot apparatus at 500 mA (con- stant current) for 4 h at 4°C. Nonspecific binding sites were blocked by incubating the membranes in 10% w/v skim milk powder and 0.05% v/v Tween-20 in PBS for l h . Membranes were washed three times in PBS con- taining 0.1% w/v BSA and 0.01% v/v Tween-20, incu- bated with primary antibody (0.7 yg/mL) for l h , and then washed three times (15 rnin per wash) as above. The membranes were then incubated for 1 h with a sec- ondary antibody (horseradish peroxidase-conjugated goat antibody to mouse IgG (Amersham, 1:lO 000 dilu- tion) or rabbit IgG (Amersham, 1:lO 000 dilution), washed three times (as above), and developed using the ECL procedure according to the manufacturer's instruc- tions.

2.6 Immunoprecipitation with mAbA33

After 3SS-labelling, LIM1215 cells were washed three times in PBS, removed from the culture dishes by scraping, and lysed in 1 mL 0.3% v/v Triton X-100, 100 units/mL Trasylol and 10 p~ leupeptin in PBS. The cell lysate was frozen, thawed once and vortexed briefly. Insoluble cell debris was removed by centrifugation (20 000 X g, 10 rnin). The supernatant was precleared by treatment with 20 yL of protein G-conjugated Sepharose 4B beads (50% v/v in PBS) at 4°C for 40 rnin with rota- tion. The protein G-conjugated Sepharose 4B beads were removed by centrifugation (20 000 X g) and 5 pg of mABA33 was added to the supernatant. This mixture

X I 0 - 3 1 2 3

94 -

67 -

43 -

30 -

20 - 14 -

Figure 1. Immunoblot analysis of A33 antigen. Total LIM1215 cell lysates were fractionated by one-dimensional SDS-PAGE using 4-20VoT gradient acrylamide gels (Novex) operated under reducing (lane 1) or nonreducing (lanes 2 und 3) conditions. Proteins were elec- trotransferred onto a nitrocellulose membrane and immunoblotted with mAbA33 (lanes 1 and 2), or a control mAb (anti-G-CSF mAb) (lane 3). lmmunoblots were developed using the ECL detection system.

was incubated for 2 h at 4°C with slow rotation. Protein G-Sepharose was then added (20 yL, 50% v/v in PBS) and the mixture incubated for an additional 40 rnin at 4°C. The beads were recovered by centrifugation (20 000 X g, 1 min), washed three times with 0.3% v/v Triton X-100 in PBS and the immunoprecipitated proteins were recovered for analysis by the addition of 50 pL IEF sample buffer.

2.7 Electrotransfer of gel-resolved proteins onto PVDF

After 2-DE, gels were equilibrated (5 min at room tem- perature) in transfer buffer (10 mM CAPS buffer, pH 11.0, containing 10% v/v methanol, 1 mM thioglycolic acid) before electrotransfer onto PVDF or nitrocellulose membranes. PVDF membranes were wetted in methanol briefly and then equilibrated in transfer buffer for 5 min. The transfer was performed in a Bio-Rad Trans-blot cell at 500 mA (constant current) at 4°C for 4 h. PVDF meni- brane-bound proteins were visualised by staining with 0.1% w/v CBR-250 in 50% v/v methanol, followed by destaining with aqueous 50 O/o v/v methanol/lO% v/v acetic acid. The membranes were then washed extensive- ly with water, air dried and stored at -20°C.

2.8 Concentration of 2-DE-resolved proteins

CBR-250 stained protein spots from -7 identical 2-DE gels were excised, washed extensively with water, equili- brated with SDS-PAGE sample buffer, loaded in the sample well of a 10%T SDS-polyacrylamide gel and ree- lectrophoresed as described elsewhere [12].

Electrophoresis 1997, 18, 614-621 Novel antigen for the gastrointestinal specific monoclonal antibody, A33 617

Figure 2. Expression of A33 antigen in human colorectal cell lines LIM1215 and SW1222. Human colorectal cancer cells LIM1215 and SW1222 were lysed in nonreducing SDS-PAGE sample buffer and the cell lysates fractionated by SDS-PAGE using precast 10%T acrylamide gels (Novex). Lysates from a number of control human hemopoietic cell lines (Jurkat and AML193) and murine fibroblast cell lines (Balb/ C 3T3 and v-Ha-Ras-transformed NIH 3T3) are shown for compa- rison. Proteins were transferred onto a nitrocellulose membrane and immunoblotted with the mAbA33. Immunoblots were developed using the ECL detection system. " h e position of the M, - 41 000 A33 antigen is indicated by the arrow.

Figure 3. Detergent extraction of A33 antigen from LIM1215 cells. Suhconfluent LIM1215 cells (- 5 X lo6) were washed with PBS, incu- bated in 0.5 mL of 0.5% w/v l k e e n 20 / PBS or 0.3% w/v Triton X-100 / PBS for 15 min on ice, vortexed briefly and centrifuged (20 000 X g) for 10 min at 4OC. The extracts were diluted 1:l with nonreducing, concentrated (2 X) SDS-PAGE sample buffer for immu- noblot analysis. For the control (total cellular proteins), -5 X lo6 LIM1215 cells were treated with 1 mL of SDS-PAGE sample buffer at 25 OC and solubilised proteins were recovered by centrifugation. Sam- ples were resolved by SDS-PAGE using 10%T acrylamide gels, elec- trotransferred to nitrocellulose and immunoblotted with mAbA33 as described in Section 2.5. Proteins were visualised by the ECL detec- tion system. Sample load: 40 VL of each sample (representing proteins prepared from 2 X lo5 cells) was loaded. Total cellular proteins (lane 1); 0.5% Tween-20 extract (lane 2); 0.3% Triton X-100 (lane 3). The position of the M, -41 000 A33 antigen is indicated by the arrow.

2.9 In situ tryptic digestion of 2-DE-resolved proteins

In-gel proteolytic digestion of 2-DE-resolved proteins was performed as described previously [13].

empty membrane-compatible column (Hewlett-Packard # G144GA) and subjected to Edman degradation using the routine 3.0 sequencer program.

2.11 Peptide mapping and mass spectrometry

2.10 Microsequence analysis

N-terminal amino acid sequencing of proteins and pep- tides was carried out by automated Edman degradation using a Hewlett-Packard (model G1005A) protein se- quencer operating with the routine 2.2 sequencer pro- gram. A Hewlett-Packard model HP1090 liquid chroma- tographyer was used for PTH amino acid analysis [13]. For proteins electrotransferred onto PVDF, CBR-250- stained protein spots were excised, positioned in an

Peptide mass and tandem mass spectrometric analysis was performed using a Finnigan-MAT triple quadrupole mass spectrometer (model TSQ-700; San Jose, CA, USA) equipped with an electrospray (ESI) ionisation source as previously described [lo, 131.

2.12 Peptide mass fingerprinting

Peptide mass fingerprinting was performed using either the MOWSE algorithm accessible via e-mail to "mow-

618 H . Ji pf a/ Electrophoresis 1997, 18, 614-621

se@dl.ac.uk" or the MS-FIT algorithm developed at the UCSF Mass Spectrometry Facility using the World Wide Web (U RL: http :/ hafael .ucsf.edu/ M S -Fit .html).

3 Results and discussion

3.1 Immunoblot analysis using mAbA33-recognition of a conformational epitope

While mAbA33 has been used successfully for immuno- histochemical analysis, immunostaining of human colo- rectal cancer [2] and in vivo imaging for colonic cancers [l, 4, 51, attempts to use this mAb in Western blot anal- ysis of total colonic cell proteins resolved by conven- tional SDS-PAGE were unsuccessful. This finding sug- gested that mAbA33 may recognise a conformational epitope which is disrupted during the process of reduc- ing SDS-PAGE. In an effort to maintain the conforma- tional integrity of the A33 antigen epitope recognised by mAbA33, nonreducing gel electrophoresis conditions for immunoblot analysis were examined. Figure 1 compares an immunoblot analysis of a total cell lysate of the colo- rectal cancer cell line LIM1215 that was electrophoresed under reducing SDS-PAGE conditions (Fig. 1, lane l), and nonreducing conditions (Fig. 1, lane 2); a control for the 'non-reducing' immunoblot conditions utilising anti- granulocyte - colony stimulating factor mAb is shown in Fig. 1, lane 3. As shown in Fig. 1, an M, - 41 000 protein was specifically recognised by mAbA33 under nonre- ducing conditions but not by the control antibody (anti- G-CSF mAb). These data indicate that the antigenic determinant survives the extraction and electrophoretic procedures as long as there is no reducing agent present. Presumably, disulfide bonding is critical in maintaining the conformational integrity of the mAbA33 epitope.

3.2 Expression of the A33 antigen in various cell lines

In previous immunohistochemical studies, mAbA33 was shown to specifically stain tissue and cell lines that origi- nated from gastrointestinal tissue [2]. In the present study the expression of the A33 antigen in a limited number of cell lines was investigated by immunoblot analysis. In Fig. 2 it can be seen that the A33 epitope is expressed in the human colorectal carcinoma cell lines, LIM1215 and SW1222, but not in the human haemato- poietic tumor cell lines Jurkat and AML193, or murine fibroblast cell lines (Balb/C 3T3 and v-Ha-Ras-trans- formed NIH 3T3).

3.3 Detergent extraction of A33 antigen from LIM1215 cells

It has been shown previously that the mAbA33 recog- nises an integral rnembrane-associated protein that is rapidly internalised [3]. As a first Step in designing a PUri- fication strategy for the A33 antigen, two detergent ex traction conditions were evaluated in order to enrich se lectively cell surface proteins from ~ 1 ~ 1 2 1 5 cells. For these studies, subconfluent LIM1215 cells (- 106) were washed with ice-co1d PBS and incubated with PBS, containing detergent for 15 min on ice and the extracts examined by immunoblot analysis for A33 antigen con-

Figure 4. Immunoblot analysis of IM1215 proteins resolved by non- reducing 2-DE. Triton X-100 solubilized protein fractions of LIM1215 were resolved by 2-DE using precast linear pH 3-10 IEF gels (Novex) under nonreducing conditions and electroblotted onto nitrocellulose membranes. The locations of (A) A33 antigen, (B) cytokeratin K18, and (C) actin in the 2-DE gels were identified using their respective monoclonal antibodies, as described in Section 2.5. The immunoblots were developed using the ECL detection system. The plvalues shown on the abscissa are based on the assumption that the IEF gel has a linear pH 3.5-8.5 gradient as reported by the manufacturer.

Electrophoresis 1997, 18, 614-621 Novel antigen for the gastrointestinal specific monoclonal antibody, A33 619

Figure 5. Nonreducing 2-DE pattern of detergent-extracted proteins of LIM1215 cells. Triton X-100 (0.3°/o)/10 mM Tris-HCI, pH 7.5 solubilized protein fractions of LIM1215 cells were mixed with IEF sample buffer without reducing reagent and resolved by 2-DE using precast Novex gels. Proteins were visualised by CBR-250 staining. The p l values shown on the abscissa are based on the assumption that the precast IEF gel (Novex) has a linear pH 3.5-8.5 gradient as indicated by the manufacturer. The positions of the A33 antigen, as judged by immunoblot analysis, is indicated. Actin and cytokeratin K18 were identified by immunoblot analysis (see Fig. 4) and a combination of peptide-mass fingerprinting [lo] and Edman degradation.

tent. Figure 3 shows that under these conditions 0.3% Triton X-100 is approximately 2-fold more efficient than 0.5% Tween-20 for extracting A33 antigen from LIM1215 cells.

3.4 Affinity purification and SDS-PAGE

Initial attempts to purify the A33 mAb antigen involved a combination of affinity chromatography and SDS- PAGE. Since Triton X-100 (0.3 O/o) extracts were shown to be positive using a biosensor (BIAcore) [14] on which mAbA33 had been immobilised onto the sensor surface, biosensor analysis and immunoblot analysis (nonre- ducing conditions) were used to monitor the purifica- tion. An affinity column was constructed by coupling the mAbA33 to Affi-10. The 0.3% Triton X-100 extract was

applied to the column, which was then washed sequen- tially with 1 and 3 M NaCl in 10 mM HEPES, pH 7.4 (which had been shown by biosensor analysis to cause no desorption) followed by elution with 10 mM NaOH. SDS-PAGE analysis of the 10 mM NaOH fraction indi- cated a major M, 41000 protein, which was not present in a control experiment in which the Afi-10 had been blocked with Tris buffer. However, when the M, 41000 band was transferred onto PVDF and subjected to N-terminal sequence analysis no sequence data was obtained, suggesting that the protein was N-terminally blocked. Internal sequence analysis following “in-gel” tryptic digestion and micropreparative RP-HPLC [ 131 revealed two actin peptides.

When a preparation of actin was injected over the bio- sensor surface, strong positive binding to immobilised

620 H. J i rf nl Electrvphuresis 1991, 18, 614-621

3 I

Figure 6. Electrophoretic analysis of A33 antigen by reducing 2-DE. Total cellular proteins of [35S]methionine-labelled LIM1215 cells were electrophoresed, under reducing conditions, using IPG (Pharmacia linear pH 3-10 gels) in the first dimension and SDS-PAGE in the second dimension. (A) Position of the A33 antigen as judged by immunoblot analysis using polyclonal antiserum directed towards a synthetic N-terminal peptide of the A33 antigen; (B) autoradiograph of the same gel (the position of the A33 antigen is shown by a broken circle). (C) Immunoblot analysis (using anti-peptide polyclonal anti- serum) of the A33 antigen that was prepared by immunoprecipitation (using mAbA33) from a total cellular lysate of LIM1215 cells. (D) Immunoblot analysis of a mixture of affinity-purified A33 antigen (- 50 ng; Ritter eral., manuscript in preparation) and total cellular lysate of LlM1215 cells (5 X lo6). The p l values shown on the abscissa were based on the manufacturer’s report that immobilised IPG gel has a linear 3-10 pH gradient.

mAbA33 was observed. However, similar binding was also observed when the actin preparation was passed over nonrelated antibodies of the same immunoglobulin subclass (IgG2A). By comparison, when the A33 Fab’ fragment was immobilised onto the sensor surface, binding with actin was no longer observed, suggesting that actin was binding to the Fc portion of the antibody. ?is cross-reactivity was confirmed by Western blot anal- ysis; actin, purified from rabbit muscle, was recognised with mAbA33 under nonreducing and reducing condi- tions. To avoid complications due to actin cross-reac- tivity, a Green-Sepharose HE4BD column was included in the purification scheme. Actin bound to the ligand dye column, while the A33 antigen was recovered in the column breakthrough, and eventually purified to homo- geneity using a multi-dimensional chromatographic puri- fication protocol [ 6 ] . N-terminal sequence analysis of the purified A33 antigen revealed a unique sequence:

QWD. A sequence similarity search of the available pro- tein databases failed to reveal any significant sequenc,e identity.

XSVETPQDVLRASQGKSVTLPXTYHTSXXXREGLI-

3.5 Nonreducing 2-DE of A33 antigen

Cell surface proteins from LIM1215 cells were extracted with 0.3% Triton X-100/5 mM Tris-HC1, pH 7.5, resolved by nonreducing 2-DE, and then electrotransferred onto nitrocellular membranes. The position of A33 antigen was localised in the 2-DE map by immunoblot analysis using mAbA33 (Fig. 4A). For a number of reasons, it was not possible to compare directly the nonreducing 2-DE protein distribution pattern with our master 2-DE gel of human colonic proteins reported elsewhere [l 11. Firstly, the resolution of precast Novex gels is restricted due to their small size (80 X 80 mm). Secondly, these gels were operated under nonreducinghonurea conditions that resulted in distinct differences in the mobility of several protein [lo]. Our attempts at maintaining the integrity of conformational epitopes by running precast IPGs (Phar- macia) with sample buffer lacking reducing reagent, as well as our efforts at carrier ampholyte focusing using the O’Farrell sample buffer system [15], which lacks urea and reducing agent, were unsuccessful (data not shown). Since cytokeratin 18 and actin were known to migrate in the nonreducing 2-DE gel with similar pl/M, loci to that of the A33 antigen [lo], immunoblot analysis of these proteins was performed using their respective mAbs.

It can be seen in Fig. 4 that A33 antigen (Fig. 4A, p l 5.0-6.0, M, 41 000), electrophoreses in a similar posi- tion to actin (Fig. 4C, pZ4.0-5.3, M, 41000) and cytoke- ratin 18 (Fig. 4B, pl5.5 - 6.5, M, 42500). These pZ and M, values for cytokeratin K18 and actin are in close agreement with those reported by Celis and co-workers [16-181 in their human AMA cell (transformed amnion cells) 2-DE protein database. In Fig. 5 the position of the A33 antigen, cytokeratin 18 and actin in a nonre- ducing 2-DE gel stained with Coomassie blue are indi- cated (CJ: Fig. 4). The identity of these and a number of other proteins in our master nonreducing 2-DE gel (Novex nonreducing IEF, pH 3-10), together with their M, and p l values are reported elsewhere [19].

Electrophoresis 1997, 18, 614-621 Novel antigen for the gastrointestinal specific monoclonal antibody, A33 62 1

Since cytokeratin 18 and actin electrophorese in very close proximity with the A33 antigen under these condi- tions and overlap the A33 antigen area, and since the A33 antigen is much less abundant (< 1%) than these cytoskeletal proteins, it is unlikely that it would be possible to analyse its structure from 2-DE gels. As expected, our attempts to isolate the A33 antigen region from the gel and identify peptides by conventional in-gel digestion/microsequencing procedures [ 10, 131 yielded only cytokeratin 18 and actin peptides. Indeed, both cyto- keratin 18 and actin proteins (Fig. 5) were contaminated with each other [lo] and there was no evidence of tryptic peptides from other proteins. The problem of coelectro- phoresing protein spots in 2-DE gels is well recognised [16-181 and, as emphasized in the present study, caution needs to be exercised when attempting to identify pep- tides from low-abundance proteins previously detected by the immunoblotting of complex mixtures. For this reason it is important to characterise as many of the high-abundance proteins and peptides as possible, so that as we gain in experience and data accumulation these proteins and peptides can be subtracted from the analysis - hopefully, revealing the low-abundance pro- teinlpeptides.

3.6 Reducing 2-DE gel localisation of A33 antigen

Localisation of the A33 antigen in our master 2-DE pro- tein database for the LIM1215 cell line [19] was accom- plished using a combination of immunoblot analysis and 35S-radiolabelling. First, total LIM1215 cellular proteins were intrinsically radiolabelled with [3SS]methionine and subjected to reducing 2-DE (18 cm IPG strips, linear pH 3-10 range) and then electrotransferred onto PVDF membrane. Immunoblot analysis was performed using a polyclonal antbA33-peptide antibody raised against a synthetic peptide corresponding to residues 2-20 of A33 antigen (Ser-Val-Glu-Thr-Pro-Gln-Asp-Val-Leu-Arg-Ala- Ser-Gln-Gly-Lys-Ser-Val-Thr-Leu) [6]. After protein vis- ualisation using the ECL detection system, 3SS-labelled proteins were detected by autoradiography. It can be seen by immunoblotting (Fig. 6A) that under reducing conditions the A33 antigen from a total cell lysate of [35S]methionine-labelled LIM12 15 colorectal cancer cells (-- 2 X lo6) electrophoreses (Pharmacia 18 cm IPG strips, linear pH 3-10 range) as several isoforms with a similar molecular weight (M, -55000) and pls in the range 4.6-4.8 in a position similar to that of p tubulin [19], but slightly more acidic. An autoradiograph of this gel indi- cates that the 3sS-labelled A33 antigen is of very low abundance compared to the major "S-labelled proteins (Fig. 6B). Both immunoprecipitated A33 antigen (using mAbA33, from total LIM1215 cell lysate) and affinity- purified A33 antigen (Ritter et al., manuscript in prepara- tion) can be resolved into as many as six distinct iso- forms (Fig. 6C and 6D, respectively), under the IPG/2- DE reducing conditions used in this study. These iso- forms result, presumably, from different extents of post- translational modification.

4 Concluding remarks Monoclonal antibody A33 recognises a novel membrane- associated protein that has been classified as an organ-

specific differentiation antigen for both normal and transformed gastrointestinal epithelium. The mAbA33 recognises a conformation-dependent epitope that is rapidly destroyed by reducing reagents. Under nonre- ducing 2-DE conditions (precast Novex gels), the A33 antigen electrophoresis with an apparent M, - 41 000 and a pl5.0-6.0. Under conventional reducing 2-DE con- ditions (linear pH 3-10 IPG from Pharmacia), the A33 antigen electrophoreses as six differentially charged iso- forms (pl 4.6-4.8) with an apparent M, -55000.

Received September 18, 1996

5 References

[I] Welt, S., Divgi, C. R., Real, F. X., Yel, S. D., Garin-Chesa, P., Finsad, C. L., Sakamoto, J., Cohen, A., Sigurdson, E. R., Kemeny, N., Carswell, E. A,, Oettgen, H. F., Old, L. J., J . Clin. Oncol. 1990, 8, 1894-1960.

[2] Garin-Chesa, P., Sakamoto, J., Welt, S., Real, F. X., Rettig, W. J., Old, L. J., In?. J . Oncol. 1996, 9, 465-471.

[3] Daghighian, F., Barendswaard, E., Welt, S., Humn, J., Scott, A., Willingham, M. C., McGufie, E., Old, L. J., Larson, S. M., J. Nucl. Med. 1996, 37, 1052-1057.

[4] Welt, S., Divgi, C. R., Kemeny, N., Finn, T. D., Scott, A. M., Graham, M., St. Germain, J., Richards, E. C., Larson, S. M., Oettgen, H. F., Old, L. J., J. Clin. Oncol. 1994, 12, 1561-1571.

151 Welt, S., Scott, A. M., Divgi, C. R., Kemeny, N. F., Finn, R. D., Daghighian, F., St. Germain, J., Carswell Richards, E., Larson, S. M., Old, L. J., J. Clin. Oncol. 1996, 14, 1787-1797.

[6] Catimel, B., Titter, G., Welt, S., Old, L. J., Cohen, L., Nerrie, M. A., White, S. J., Heath, J. K., Demediuk, B., Domagala, T., Lee, F. T., Scott, A. M., Tu, G. F., Ji, H., Moritz, R. L., Simpson, R. J., Burgess, A. W., Nice, E. C., J. Biol. Chem. 1996,271,25664- 25670.

[7] Whitehead, R. H., Macrae, F. A,, St. John, J., Ma, J., J. Nut. Cancer Inst. 1995, 74, 759-765.

[8] Ji, H., Baldwin, G. S., Burgess, A. W., Moritz, R. L., Ward, L. D., Simpson, R. J., J. Biol. Chern. 1993, 263, 13396-13405.

191 Laernmli, U. K., Nature 1970, 227, 680-685. [lo] Reid, G. E., Ji, H., Eddes, J. S., Moritz, R. L., Simpson, R. J.,

[ I l l Ji, H., Whitehead, R. H., Reid, G. E., Moritz, R. L., Ward, L. D.,

1121 Ward, L. D., Hong, J., Whitehead, R. H., Simpson, R. J., Electro-

1131 Moritz, R. E., Eddes, J. S., Reid, G. E., Simpson, R. J., Elecfro-

1141 Malmqvist, M., Nature 1993, 361, 186-187. 1151 O'Farrell, P. H., J. B i d . Chem. 1975, 250, 4007-4021. 1161 Celis, J. E., Gesser, B., Rasmussen, H. H., Madsen, P., Leffers, H.,

Dejgaard, K., Honore, B., Olsen, E., Ratz, G. , Lauridsen, J. B., Basse, B., Mouritzen, S., Hellerup, M., Andersen, A., Walbum, E., Celis. A,, Bauw, G., Puype, M., Van Damme, J., Vandekerckhove, J., Elecfrophoresis 1990, 11, 989-1071.

1171 Celis, J. E., Leffers, H., Rasmussen, H. H., Madsen, P., HonorC, B., Gesser, B., Dejgaard, K., Olsen, E., Ratz, G. P., Lauridsen, J. B., Basse, B., Andersen, A. H., Walbum, E., Brandstrup, B., Celis, A., Puype, M., Van Damme, J., Vandekerckhove, J., Electro- phoresis 1991, 12, 765-801.

[18] Celis, J. E., Rasrnussen, H. H., Madsen, P., Leffers, H., Honore, B., Dejgaard, K., Gesser, B., Olsen, E., Gromov, P., Hoffmann, H. J., Nielsen, H., Anderson, A. H., Walbum, E., Kjaergaard, I., Puype, M., Van Darnme, J., Vandekerckhove, J., Electrophoresis 1992, 13, 893-959.

[19] Ji, H., Moritz, R. L., Reid, G. E., Eddes, J . S., Burgess, A. W., Simpson, R. J., Electrophoresis 1997, 18, 605-613.

Electrophoresis 1995, 16, 1120-1 130.

Simpson, R. J., Electrophoresis 1994, IS, 391-405.

phoresis 1990, 11, 883-891.

phoresis 1996, 17, 907-917.