INTRODUCTION

Top Abstract Introduction MATERIALS AND METHODS RESULTS AND DISCUSSION References

EPIDERMAL GROWTH FACTOR (EGF) is a very potent 6-kDa polypeptide mitogen that belongs to a family of growth factors that also includes transforming growth factor- (TGF-), heparin-binding EGF-like growth factor (HB-EGF), amphiregulin, and betacellulin. These growth factors bind and activate the intrinsic tyrosine kinase activity of the EGF receptor (EGFR) (10). Most of the known growth factors are derived from soluble precursors. These biologically inactive soluble precursors are confined to cytoplasmic compartments of the secretory pathway where they mature through proteolytic cleavage before being released outside the cell. It is now well established from both cDNA sequence analysis and protein biochemistry that growth factors of the EGF family are also synthesized in the form of precursor molecules. However, they constitute an exception to the general model, since they are thought to be synthesized as transmembrane glycoproteins. For example, cDNA analysis of the EGF precursor predicted proteins of 1,207, 1,217, and 1,133 amino acids, respectively, for human (1), mouse (33, 40), and rat (34), resulting in glycoproteins of molecular mass between 140 and 170 kDa. This molecule is made up of a large extracellular region (~1,000 amino acids) with the EGF sequence (48-53 amino acids) located near the plasma membrane, a transmembrane domain, and an intracellular domain of variable length (19). In tissues such as kidney (2, 33, 37) and mammary gland (3), EGF is present as part of the extracellular portion of the transmembrane precursor. However, in the submaxillary gland from rodents, the EGF precursor molecule is fully and intracellularly processed into EGF (33) and stored in the secretory granules of the granular convoluted tubules (31, 39); thus the ways by which EGF is secreted in the extracellular medium must be completely different. In the first case, the precursor needs to be cleaved extracellularly by an unidentified ectoprotease (12, 13), whereas, in the second case, secretion involves the well-known regulated exocrine secretion through stored secretory granules (6).

Recently, it was shown that human (27, 49), mouse (45), and rat (50) tears contained EGF or an EGF-like immunoreactivity and that the total production of EGF in human tears increased during reflex tearing (46). The EGF precursor mRNA has been detected by Northern blot in mouse (14) and by RT-PCR in human (52) lacrimal glands. Moreover, EGF-like immunoreactivities have been detected in human (26), mouse (14), and rat (48, 50) lacrimal tissues.

The lacrimal gland produces the complex aqueous portion of tears, which contains many components, including electrolytes and proteins. The pH, electrolyte concentration, and protein composition of lacrimal fluids are crucial in maintaining the health of the ocular surface. The proteins synthesized and secreted by the lacrimal glands are very specific and are thought to be mainly involved in the bacteriostatic action of tears. Until now, only a few of them have been identified, but their secretion involves both the constitutive (42) and regulated pathways (8). From the work of Savage and Cohen (35), showing the stimulating effect of EGF on corneal epithelial cell proliferation, the role of EGF in corneal wound healing has been extensively studied. The results indicated that EGF and other growth factors were involved in the stimulation of reepithelization processes as well as keratinocyte and corneal endothelium proliferation (5, 51). Thus the conserved presence of EGF and/or EGF-related molecules in tears and lacrimal glands of both humans and rodents further suggested that it may serve an important function in the periocular environment. In light of these results, it was suggested that the lacrimal gland could be important in the process of corneal wounding by producing growth factors secreted in tears (51). In the human lacrimal gland, EGF secretion could involve a muscarinically regulated pathway (53). However, until now nothing was known about the storage form of EGF in this tissue and consequently about the way by which EGF could be secreted into the tear fluid.

In view of the above hypothesis, and because the lacrimal gland tissue from rat is very easy to obtain compared with its human counterpart, we decided to look for the presence of EGF and to identify the molecular form present in the rat exorbital lacrimal gland. We first examined EGF gene expression by both RT-PCR and Northern blot analysis. Second, EGF-containing proteins were identified and localized using anti-rat EGF and anti-rat EGF precursor antibodies. Results obtained were compared with reference tissues such as rat submaxillary gland and kidney. Our results demonstrate for the first time both EGF precursor gene transcription and EGF precursor protein expression in a lacrimal tissue, i.e., the rat exorbital lacrimal gland. Contrary to previous observations made in both mouse (14) and rat (50) lacrimal gland, we were unable to detect the 6-kDa soluble form of mature EGF in the rat lacrimal gland. Our results demonstrate that EGF is stored only as its full-length membrane-associated precursor and may provide key information to study the regulation of its secretory process.

	MATERIALS AND METHODS

Top Abstract Introduction MATERIALS AND METHODS RESULTS AND DISCUSSION References

Animals. Adult male albino Sprague-Dawley rats were obtained from IFFA CEDO and used throughout this study.

Chemicals. Peroxidase-conjugated and alkaline phosphatase-conjugated goat anti-rabbit IgG, mouse EGF (mEGF), synthetic rat TGF-, trypsin (10,000 N-benzoyl-L-arginine ethyl ester units/mg), soybean trypsin inhibitor (1 mg inhibits 10,000 units trypsin), polyethylene glycol 6000, pepstatin A, leupeptin, chymostatin, antipain, p-nitrophenyl phosphate, p-nitrophenol, and protein molecular mass markers were obtained from Sigma Chemical (St. Louis, MO). Enhanced chemiluminescence (ECL) developer and Hyperfilm were from Amersham France (Les Ulis, France). Bicinchoninic acid protein assay kit and ImmunoPure Ag/Ab immobilization trial kit (SulfoLink coupling gel) were purchased from Pierce (Rockford, IL). Recombinant human EGF (hEGF) was from Preprotech (Washington, MA). Triton X-100 was obtained from Merck (Darmstadt, Germany). Bio-Gel P-10 was from Bio-Rad Laboratories (Ivry/seine, France). Prepacked HiPrep 16/60 Sephacryl S-200 HR, Mono Q HR 5/5, PD-10 desalting columns, deoxynucleotide triphosphates, protein A-Sepharose CL-4B, agarose, and DNA size markers were obtained from Pharmacia Biotech (Orsay, France). Superscript RNase H RT, random primers, and restriction enzymes Pst I, Hae III, and Sst I were from GIBCO BRL (Eragny, France). The random primer DNA labeling kit was obtained from Boehringer Mannheim (Meylan, France). Upstream and downstream oligonucleotide primers and synthetic peptide p437 were synthesized by Eurogentec (Seraing, Belgium). Taq DNA polymerase was from Appligene (Illkirch, France). [¹²⁵I]NaI (100 mCi/ml, 3.7 MBq/ml), [-³²P]dCTP (3,000 Ci/mmol, 111 TBq/mmol), and ¹²⁵I-labeled mEGF (150-200 µCi/µg, 5.6-7.4 MBq/µg) were purchased from New England Nuclear (Les Ulis, France).

Rat submaxillary gland EGF purification. Rat EGF (rEGF) was isolated from submaxillary glands of adult male Sprague-Dawley rats (at least 12 wk old) using rapid HPLC techniques according to Simpson et al. (41). This purification involved the homogenization of the frozen tissue in drastic acidic conditions followed by centrifugation of the homogenate. The rEGF contained in the soluble material was purified by sequential chromatography through a preparative reverse-phase C18 column and an analytical C18 HPLC column and anion exchange on a mono-Q column. During the course of this purification, rEGF was followed by radioreceptor assay (RRA) as described below but using ¹²⁵I-labeled mEGF as radioligand. From 23 g of tissue (wet mass), we obtained ~600 µg of purified rEGF. The purity of the final product was checked by comparison with commercial mEGF and hEGF by SDS-PAGE, automated amino acid sequence analysis, and ability to stimulate HC 11 cell growth (data not shown).

rEGF radiolabeling. Native rEGF was radiolabeled with [¹²⁵I]NaI by the chloramine T method. Briefly, 1 µg rEGF dissolved in 20 µl of 0.25 M sodium phosphate buffer (pH 7.4) was incubated for 45 s in the presence of 0.5 mCi (5 µl) carrier-free [¹²⁵I]NaI and 10 µl chloramine T (2 mg/ml) at room temperature. The reaction was terminated by the addition of 20 µl sodium metabisulfite (2 mg/ml) and 40 µl NaI (2.5 mg/ml) and the mixture was diluted to 500 µl with 50 mM sodium phosphate buffer (pH 7.4) containing 150 mM NaCl and 1% BSA. ¹²⁵I-rEGF was separated from nonincorporated [¹²⁵I]NaI by chromatography on a PD-10 (Sephadex G-25) desalting column. In these conditions, the peptide incorporated 40-60% of the radioactivity, and the eluted iodinated peptide is 95-97% precipitable by 10% TCA.

Antibody production. Purified rEGF was used to generate polyclonal rabbit antibodies. Production was performed according to the Eurogentec custom protocol of immunization. Two rabbits were immunized with 50 µg of rEGF and further boosted twice at 15-day intervals and then once more 1 mo later. The presence of specific anti-rEGF antibodies in sera was tested by their ability to immunoprecipitate ¹²⁵I-rEGF. Both rabbits produced antibodies, but only one (rEGF₂) was retained because of its higher serum titer. The IgG fraction of this antiserum was prepared by chromatography on protein A-Sepharose.

Preparation and fractionation of the rat tissues. Rats were killed by carbon dioxide inhalation. The exorbital lacrimal, parotid, and submaxillary glands, heart, brain, kidney, and liver were rapidly removed. Glandular tissues were trimmed of their fatty and connective tissues, and hearts and livers were extensively washed to eliminate blood contamination as much as possible. All tissues were further fragmented into small pieces and either used for the preparation of lacrimal acini as described previously (20) or frozen in liquid nitrogen and stored at 20°C until processing for RNA (18) or subcellular fraction preparation (see below).

RRA and RIA of soluble and/or membrane-associated EGF precursor molecules. Rat liver membrane fraction prepared as described previously (20) was used as the source of EGFR for RRA. Up to 400 µl of diluted or undiluted soluble fractions from submaxillary gland, lacrimal gland, or kidney (prepared as described above) or known amounts of rEGF were incubated in the presence of ¹²⁵I-rEGF [30,000-50,000 counts/min (cpm), 25 pM] and liver membranes in a final volume of 500 µl of buffer (50 mM sodium phosphate, pH 7.4, and 250 mM sucrose) for 3 h at 20°C. Bound and free ¹²⁵I-rEGF were separated by rapid filtration through glass-fiber filters as described previously (20). The radioactivity retained on the filter was counted on an LKB 1275 mini-gamma-counter. A standard competition curve was generated using increasing concentrations of nonlabeled rEGF ranging from 0 to 100 nM. With the assumption that all activities bind to the EGFR with equal potency, the amounts of EGF-like activities in the various soluble fractions were quantified by comparing the displacement curve with that of authentic rEGF.

Gel filtration analysis of the molecular mass form of membrane-associated EGF precursor molecules. Triton-extracted membrane-associated EGF-containing molecules in both kidney and lacrimal gland were characterized by gel filtration on 1.6 × 60-cm Sephacryl S-200 and 1.6 × 30-cm Bio-Gel P-10 columns. The columns were equilibrated at 4°C in lysis buffer B. Elution was performed in the same buffer at the flow rates of 13 ml/h for Sephacryl S-200 and 8 ml/h for the Bio-Gel P-10 column. Calibration of the Sephacryl S-200 column was performed with ferritin (440 kDa), chicken IgY (190 kDa), BSA (67 kDa), ovalbumin (45 kDa), carbonic anhydrase (29 kDa), cytochrome c (12.4 kDa), and rat submaxillary gland EGF (5.3 kDa). Calibration of the Bio-Gel P-10 column was performed with chicken IgY (190 kDa), rat submaxillary gland EGF (5.3 kDa), and p-nitrophenol. An aliquot of the solubilized membrane proteins from each tissue was analyzed before (Sephacryl S-200) and/or after (Sephacryl S-200 and Bio-Gel P-10) trypsin hydrolysis as described above. In both cases, 1-ml fractions were collected for evaluation of their irEGF content by RIA either directly or after trypsin hydrolysis as indicated in RESULTS AND DISCUSSION.

Immunoprecipitation of membrane-associated EGF precursor molecules. For immunoprecipitation experiments, 1 ml of lacrimal gland lysates and 0.5 ml of kidney membrane lysates prepared in lysis buffer A were used. Immunoprecipitations were carried out overnight at 4°C under constant rocking in the presence of 8 µg of ppEGF₁ in the absence or presence of 20 µg of peptide p437. Immune complexes were precipitated with 30 µl of protein A-Sepharose conjugate for 1 h at 4°C, and the immunoprecipitates were collected by centrifugation for 15 s at 10,000 g. Immunoprecipitates were then washed twice in lysis buffer A and finally in Tris-buffered saline buffer (10 mM Tris · HCl, pH 7.5, and 150 mM NaCl). Immunoprecipitates were then analyzed for the presence of EGF precursor molecules either by RIA or Western blot. For RIA analysis, the washed protein A-Sepharose pellet was first resuspended in 100 µl of trypsin-containing buffer (100 µg trypsin/ml) and incubated for 1 h at 37°C. Protein A-Sepharose was pelleted by centrifugation, and the supernatant was tested for the presence of irEGF. For Western blot analysis, the immunoprecipitate was heated in SDS sample buffer for 5 min at 100°C. Solubilized proteins were separated by 7.5% SDS-PAGE, electrotransferred, and blotted overnight to nitrocellulose membrane (BA 85, Schleicher & Schuell, Dassel, Germany). Blots were then further processed and immunostained with affinity-purified anti-p437 antibody (ppEGF₁; 0.8 µg/ml) and probed with a 1:10,000 dilution of goat anti-rabbit IgG antibody linked to horseradish peroxidase exactly as described previously (18). Blots were developed in ECL according to manufacturer recommendations and visualized by exposure to Amersham Hyperfilm-ECL.

RNA extraction and RT-PCR analysis. Total RNAs from brain, heart, liver, kidney, parotid glands, submaxillary glands, whole lacrimal glands, and lacrimal acinar cells from rats were prepared and subjected to reverse transcription as described previously (18). Oligonucleotide primers (22-mer) were used for amplification of the EGF precursor mRNA by PCR. The sense and antisense oligonucleotide sequences were obtained from a published cDNA sequence of the rat prepro-EGF (Ref. 34; GenBank no. M63585). They are listed in the 5'-to-3' direction with the following coordinates: sense positions 3084-3105 (ATGTCTGCCAATGCTCAGAAGG) and antisense positions 3679-3700 (TAGGACCACAAACCAAGGTTGGG). After 30 cycles of amplification (1 min at 94°C, 1 min at 60°C, and 1 min at 72°C) performed in a Perkin-Elmer thermal cycler, amplified cDNAs were analyzed by electrophoresis on a 2% agarose gel in buffer containing 89 mM Tris, 89 mM boric acid, and 1 mM EDTA (pH 8) and identified by ethidium bromide staining as previously described (18). Negative controls were carried out either with reverse transcription performed in the absence of RNA templates or with RNA incubated in the absence of RT. cDNAs were further transferred to Zeta-probe membranes (Bio-Rad) by capillary blotting overnight under high ionic strength (10× SSC buffer = 1.5 M sodium chloride and 0.15 M sodium citrate supplemented with 0.5% SDS wt/vol) and fixed covalently to the membrane by intense ultraviolet (UV) illumination.

Southern blot analysis of PCR amplification products. To verify the specificity of the amplification products, the hybridization of the Southern blots with a specific cDNA probe was performed. The probe used was the 0.4-kb Pst I fragment derived from cDNA clone pmEGF-26F12 of the mouse prepro-EGF (11) (American Type Culture Collection, ref. no. 37486). The probe was labeled by random priming with [-³²P]dCTP. Southern blots were prehybridized and hybridized with the labeled probe (20 ng, 3 × 10⁶ cpm/ml) and washed under increasing stringency from 2× SSC-0.1% SDS at 45°C for 15 min to 0.1× SSC-0.1% SDS at 65°C for 15 min, as previously described (18). Autoradiographs were obtained by exposure to Amersham Hyperfilm using two intensifying screens at 80°C.

Restriction mapping of the amplification products. To further confirm the specificity of amplification, restriction enzyme analysis of the amplified cDNA fragments from rat lacrimal gland cells and rat submaxillary gland was also performed. Three different restriction enzymes were used, Pst I, Sst I, and Hae III. The incubations were carried out in a final volume of 30 µl containing 14 µl of amplification products, 1 µl of each restriction enzyme (10 units), 3 µl of the 10× buffer supplied by the manufacturer, and sterile distilled water (up to 30 µl) for 3 h at 37°C. Incubations were stopped by dilution with 80 µl of Tris-EDTA buffer and immediate precipitation at 20°C with ethanol in the presence of sodium acetate. Digestion products were analyzed by electrophoresis on 2.5% agarose gel and visualized as described above.

Northern blot analysis of poly(A)⁺ RNA from rat tissues. Poly(A)⁺ RNA (mRNA) was purified from total RNA through oligo(dT)-cellulose affinity chromatography. Samples of mRNA (20 µg) were first size separated by electrophoresis in agarose (1.5%) and then transferred and covalently fixed to Zeta-probe membranes as previously described (18). The 617-bp cDNA fragment obtained by RT-PCR amplification of the rat submaxillary gland RNA (see above) was used to probe the Northern blot. This probe was purified by extraction from agarose gel and labeled by random priming in the presence of [-³²P]dCTP. The Northern blot was prehybridized, hybridized overnight in the presence of the labeled probe (3 × 10⁶ cpm/ml), washed under high stringency, and autoradiographed by exposure to Amersham Hyperfilm using two intensifying screens at 80°C as described previously (18). To test for the integrity of the poly(A)⁺ RNA preparations from the different rat tissues, Northern blot analysis for -actin mRNA was carried out with a 1150-bp fragment of the mouse -actin cDNA in a manner similar to that described above.

	RESULTS AND DISCUSSION

Top Abstract Introduction MATERIALS AND METHODS RESULTS AND DISCUSSION References

As stated in the introduction, the presence of EGF as well as its mRNA in lacrimal gland tissues and the presence of EGF in tears were recently documented. Moreover, it is now known that EGF mRNA encodes a high-molecular-mass precursor molecule from which EGF may be released by proteolytic cleavage. EGF may be present as a part of the extracellular portion of the transmembrane precursor (as in the kidney), or this precursor could be fully and intracellularly processed into EGF (as in the submaxillary gland). Until now, despite the demonstration of the presence of an EGF immunoreactivity in some lacrimal gland tissues, little attention has been given to the identification of the molecular form present in such tissues. The following experiments were thus performed with the rat exorbital lacrimal gland to answer this question. We first used the highly sensitive technique of RT-PCR, in conjunction with specific oligonucleotide primers for the rEGF, to investigate the expression of the EGF mRNA in the lacrimal gland and for comparison with other rat tissues.

RT-PCR analysis of the tissue expression of EGF mRNA. As a control for the integrity of the RNA that serves as substrate for the RT-PCR analysis, agarose gel electrophoresis was performed, and the RNA was visualized with ethidium bromide under UV illumination (data not shown, but see Ref. 18). Equal amounts of highly preserved RNA appear to be present in the fractions. These RNAs were first reverse transcribed into cDNA and then amplified by PCR using specific sense and antisense primers deduced from the sequence of the rEGF precursor (34). Amplification primers were chosen to amplify a region of the precursor mRNA (nt 3084-3700) that overlaps the sequence encoding the EGF molecule (nt 3308-3466) (Fig. 1A). A single amplified product of the expected size (617 bp) was visible on an ethidium bromide-stained agarose gel (Fig. 1B). Strong signals were obtained with RNA from the submaxillary gland, lacrimal gland, and lacrimal acinar cells, as well as from kidney. Weaker signals were observed with both parotid and liver RNA, and no amplification signals were seen with brain and heart RNA. Moreover, no amplification products could be detected when the RNA template was omitted from the reaction (control) or when the amplification was performed with non-reverse-transcribed RNA (data not shown). These results show first that RNA templates were necessary to observe the amplification product and second that this product originates from mRNA rather than from potentially contaminating genomic DNA.

View larger version (44K): [in this window] [in a new window]   Fig. 1.   Identification by RT-PCR of epidermal growth factor (EGF) precursor mRNA from rat tissues. A: schematic representation of rat 4.8-kb prepro-EGF mRNA as described in Ref. 34. Box (nt 389-3790) indicates protein-coding portion of mRNA. Mature EGF coding region (nt 3308-3466) is represented as shaded boxes. Thin lines correspond to untranslated 5' and 3' regions of mRNA. RT-PCR-amplified portion of mRNA (nt 3084-3700) is enlarged to show relative positions of different restriction sites for Pst I, Sst I, and Hae III. Also shown is hybridizing position of 0.4-kb Pst I fragment derived from cDNA clone pmEGF-26F12 used as Southern probe. B: rat tissue RNA was subjected to RT-PCR using EGF precursor-specific primers, electrophoresed through 2% agarose gel, stained with ethidium bromide, and visualized under ultraviolet light as described in MATERIALS AND METHODS. First and last lanes contain mass markers as indicated at right, in bp. Size of predicted amplified product is indicated at left, in bp. Rat SM, rat submaxillary gland; gl, gland. C: Southern blot of RT-PCR products from B. Agarose gel was transferred to nylon membrane and hybridized with a 32P-labeled mouse EGF (mEGF) precursor cDNA probe. After washing at high stringency, blot was exposed to X-ray film with intensifying screens at 70°C. Images in B and C were digitized, and, for B, an inverted image is shown for easier visualization.

Northern blot analysis of poly(A)⁺ RNA. The 617-bp PCR product was used as a probe after ³²P labeling by random priming. Northern blots were hybridized with the labeled probe and washed under high-stringency conditions as described in MATERIALS AND METHODS. Hybridization revealed one specific 5-kb transcript in our control tissues, i.e., in submaxillary gland and kidney as well as in lacrimal and liver preparations (Fig. 2A). The size of this transcript is close to that of the prepro-EGF mRNA identified by Northern blot in mouse kidney and submaxillary gland (14) and only slightly longer than the rat prepro-EGF cDNA (4801 bp) cloned from rat kidney (34). This Northern blot analysis clearly shows that the lacrimal gland contains significant amounts of prepro-EGF transcript. The tissue content could be estimated to be about one-tenth of that present in both submaxillary gland and kidney. Contrary to the results of Kasayama et al. (14) obtained with mouse tissues, we did not find any difference in the size of the transcript between the lacrimal gland and kidney.

View larger version (39K): [in this window] [in a new window]   Fig. 2.   Northern blot analysis of poly(A)+ RNA from rat tissues. Poly(A)+ RNA from lacrimal gland and acinar cells, submaxillary gland, brain, liver, and kidney were prepared as described in MATERIALS AND METHODS; 20 µg of each mRNA preparation were separated by electrophoresis on a 1.5% formamide-containing agarose gel and transferred to nylon membranes. Blot was probed either with 32P-labeled 617-bp cDNA of rat EGF (rEGF) precursor (A) or with 32P-labeled 1150-bp fragment of mouse -actin cDNA (B). After hybridization and washing under high stringency, blot was exposed to X-ray film at 70°C with intensifying screens. Positions of different RNA size markers are indicated at left, in kb.

Assay of EGF-related molecules in soluble fractions from submaxillary gland, lacrimal gland, and kidney by RRA. The next experiments were designed to determine the presence of EGF or EGF-related molecules in the rat lacrimal gland. We first looked for the presence of soluble EGFR binding proteins in the soluble fractions from control tissues, i.e., submaxillary gland and kidney, and compared the results with those of the lacrimal gland. In these experiments, serial dilutions of the soluble fractions from the different tissues were tested for their ability to compete with ¹²⁵I-rEGF for binding to the rat liver EGFR. As can be seen in Fig. 3, all three soluble fractions compete with ¹²⁵I-rEGF for binding to the EGFR and show displacement curves parallel to that of purified rEGF. Comparison of these curves with that of standard EGF allowed an estimation of the amounts of EGF-like molecules present in the different tissues. As could be predicted, submaxillary glands contain the highest amount of soluble EGF-like activity (5,700 pg/mg tissue), whereas kidney (340 pg/mg tissue) and lacrimal glands (200 pg/mg tissue) contain lower and comparable activities. Taking into account our knowledge that EGF is not the only EGFR binding molecule, these results only indicate that the lacrimal gland contains soluble EGF-related activities. EGF might be only one of these activities, since TGF- has been demonstrated to be present in the rat lacrimal gland (47) and since we have recently identified both TGF- and HB-EGF transcripts in addition to EGF in the same tissue by RT-PCR (data not shown). Thus the determination of the specific contribution of EGF in this EGFR binding activity requires development of a specific and sensitive RIA for rEGF.

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Radioreceptor assay of soluble EGF-like growth factors in submaxillary gland, lacrimal gland, and kidney from rat. Soluble fractions from different rat tissues were prepared as described in MATERIALS AND METHODS and subjected to serial dilutions (up to 1:10,000) in phosphate-sucrose buffer. [rEGF], rEGF concentration (from 1 × 1012 to 1 × 106 M). Aliquots of either purified rEGF or soluble fractions of submaxillary gland (SM), lacrimal gland (LAC), and kidney were assayed for their ability to compete with 125I-labeled rEGF (25 pM, 50,000 counts/min) in binding to rat liver membrane EGF receptor as described in MATERIALS AND METHODS. Data points are means of triplicate determinations. Nonspecific binding was determined in presence of 100 nM rEGF and subtracted from total binding to give specific binding.

RIA for rEGF. Polyclonal antibodies were produced by immunizing rabbits against a purified and native preparation of rat submaxillary gland EGF (rEGF). The IgG fraction of the antiserum containing the highest titer was purified through protein A-Sepharose chromatography as described in MATERIALS AND METHODS. An RIA was thus developed using ¹²⁵I-rEGF and the anti-rEGF antiserum rEGF₂. As shown in Fig. 4, for an antibody dilution of 1:5,000, rEGF inhibits >95% of ¹²⁵I-rEGF binding, with a half-maximal inhibition at an rEGF concentration between 0.1 and 0.2 nM. These experimental conditions allowed the detection of 60 pg/assay of rEGF. This antibody is highly selective for rEGF. Although mEGF has very high sequence homology (79%) with rEGF, 100 times more of this growth factor is required for comparable displacement. Moreover, mEGF displaced ¹²⁵I-rEGF in a nonparallel fashion compared with rEGF, indicating the nonidentity of the antigenic determinants. The antibody rEGF₂ only weakly recognized hEGF, despite its 67% homology with rEGF, and did not recognize rat TGF- (35% homology) at concentrations as high as 100 nM. Taken together, these results show that this antibody directed against native rEGF clearly cross-reacts with closely related molecules as a function of their sequence homology with rEGF. These results are in striking opposition to those obtained with RRA experiments in which we observed that all these molecules compete with equal potency when binding to the EGFR (data not shown).

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Specificity of rEGF RIA. RIA was performed in absence or presence of increasing concentrations of EGF from rat, mouse, and human (hEGF) or transforming growth factor- from rat (rTGF-) in a final volume of 250 µl of lysis buffer in presence of anti-rat EGF antibody (rEGF2, at a final dilution of 1:5,000) and 125I-rEGF. Antigen-antibody complexes were recovered as described in MATERIALS AND METHODS. Data points are means of duplicate determinations. Nonspecific interactions (~5% of total binding) were determined in presence of 30 nM rEGF and subtracted from total binding to give specific binding.

Assay of EGF in soluble fractions from submaxillary gland, lacrimal gland, and kidney by RIA. To measure the level of irEGF in the soluble fractions from the different tissues, serial dilutions were assessed for their ability to compete with ¹²⁵I-rEGF in binding to the rEGF₂ antibody. As was proposed by Schaudies et al. (36), to detect the presence of EGF-containing molecules (precursor) that may have antigenic properties different from those of mature rEGF, samples of lacrimal gland and kidney were tested both before and after tryptic digestion as described in MATERIALS AND METHODS. Mature EGF is insensitive to trypsin (38), and tryptic digestion of EGF precursor molecules releases EGF in a form that is both immunologically and biologically indistinguishable from mature EGF (Ref. 36 and see Fig. 7).

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Immunoreactive EGF in soluble fractions. RIA of soluble fraction from submaxillary gland, lacrimal gland, and kidney. Soluble fractions from submaxillary gland, lacrimal gland, and kidney were prepared, digested (+) or not () with trypsin as described in MATERIALS AND METHODS, and subjected to serial dilutions (up to 1:10,000) in phosphate-sucrose buffer; 200-µl aliquots of diluted samples and increasing concentrations of rEGF were tested in a final volume of 250 µl phosphate-sucrose buffer for their ability to compete with 125I-rEGF for binding to antibody rEGF2 (1:5,000). Antigen-antibody complexes were recovered as described in MATERIALS AND METHODS. Data points are means of duplicate determinations. Nonspecific interactions (~5% of total binding) were determined in presence of 30 nM rEGF and subtracted from total binding to give specific binding.

Assay of EGF in solubilized membrane fractions from lacrimal gland and kidney by RIA. Membrane fractions from both the lacrimal gland and kidney were prepared and solubilized in a Triton-containing buffer. Serial dilutions of Triton-solubilized membranes were then tested for the presence of immunoreactive EGF both before and after trypsin hydrolysis as described above. As can be seen in Fig. 6, samples from both tissues only poorly (lacrimal gland) or moderately (kidney) compete with ¹²⁵I-rEGF for binding to the rEGF₂ antibody. The displacement curves generated by the Triton-solubilized membranes were not parallel to the curve obtained with purified rEGF, indicating the nonidentity of the immunoreactive materials. As discussed above for the soluble fractions, tryptic digestion of the sample was used to test for the presence of EGF precursor in the Triton X-100 extracts. As can be seen in Fig. 6, trypsin hydrolysis of both lacrimal gland and kidney results in a dramatic increase in the ability of the samples to compete for binding to the antibody. Opposite to what was observed with the untreated samples, the trypsin-treated samples generated displacement curves that were parallel to the standard curve, thus suggesting the generation of mature irEGF from precursor molecules. After trypsin hydrolysis, the level of irEGF detected rose from 1.6 to 25.7 pg/mg of tissue in the lacrimal gland and from 9.5 to 110 pg/mg of tissue in the kidney. As stated above, it is now known that the rat kidney contains membrane-associated EGF precursor molecules (37). Thus, by analogy with the kidney, the presence of the trypsin-sensitive EGF-containing molecules in the Triton-solubilized membrane fraction from the rat lacrimal gland strongly suggests the existence of membrane-associated EGF precursor molecules in this tissue.

View larger version (20K): [in this window] [in a new window]   Fig. 6.   RIA of immunoreactive rEGF in detergent-solubilized membrane fractions from lacrimal gland and kidney. Membrane-enriched fractions from lacrimal gland and kidney were solubilized in lysis buffer A, and detergent-soluble fractions from both tissues were digested (+) or not () with trypsin as described in MATERIALS AND METHODS. Serial dilutions (up to 1:100) were performed in lysis buffer A, and 200-µl aliquots of diluted samples and increasing concentrations of rEGF were tested in a final volume of 250 µl of lysis buffer A for their ability to compete with 125I-rEGF in binding to antibody rEGF2 (1:5,000). Antigen-antibody complexes were recovered as described in MATERIALS AND METHODS. Data points are means of duplicate determinations. Nonspecific interactions (~5% of total binding) were determined in presence of 30 nM rEGF and subtracted from total binding to give specific binding.

Characterization of the detergent-extracted membrane-associated EGF precursor molecules by size exclusion chromatography. The following experiments were performed to determine the size of the EGF-containing molecules that generate immunoreactive EGF on trypsin treatment as well as the size of the material released by trypsin. Triton-solubilized membrane fractions from both the kidney (Fig. 7, A and C) and the lacrimal gland (Fig. 7, B and D) were analyzed by size exclusion chromatography either on Sephacryl S-200 (Fig. 7, A and B) or Bio-Gel P-10 (Fig. 7, C and D) as described in MATERIALS AND METHODS.

View larger version (32K): [in this window] [in a new window]   Fig. 7.   Gel filtration analysis of molecular mass form of membrane-associated EGF precursor molecules. Triton-extracted membrane-associated EGF-containing molecules in both kidney (A and C) and lacrimal gland (B and D) were characterized by gel filtration on Sephacryl S-200 (A and B) and Bio-Gel P-10 (C and D) columns. Before being loaded onto column, membrane-enriched fractions were prepared in lysis buffer B as described in MATERIALS AND METHODS. Aliquots of 0.5 ml (A-C) or 0.75 ml (D) were either injected directly (filled symbols) or subjected to trypsin hydrolysis before injection (open symbols). Aliquots of individual fractions eluting from each run were assayed directly for their immunoreactive rEGF (ir-rEGF) content by RIA (open symbols) or digested with trypsin to generate low-molecular-mass EGF and were subsequently assayed by RIA (filled symbols). Calibration of Sephacryl S-200 column was performed with ferritin (440 kDa; Vo), chicken IgY (190 kDa), BSA (67 kDa), ovalbumin (45 kDa), carbonic anhydrase (29 kDa), cytochrome c (12.4 kDa), and rat submaxillary gland EGF (sm-EGF; 5.3 kDa). Calibration of Bio-Gel P-10 column was performed with chicken IgY (190 kDa; Vo), rat submaxillary gland EGF (5.3 kDa), and p-nitrophenol (Vt).

Immunoprecipitation and Western blot analysis of the detergent-extracted membrane-associated EGF precursor molecules. The above size exclusion chromatographic analysis of the detergent-solubilized membrane-associated EGF immunoreactivity only provided rough estimates of the EGF precursor(s) molecular mass(es). So we tried to obtain more accurate values by using an antibody (ppEGF₁) raised against a synthetic peptide (p437) that corresponds to a sequence of the rat prepro-EGF that is predicted to be located in its intracellular juxtamembrane domain. Because of the location of this antigenic determinant, ppEGF₁ antibody is postulated to identify only membrane-associated precursor molecules.

View larger version (52K): [in this window] [in a new window]   Fig. 8.   Immunoprecipitation of membrane-associated EGF precursor molecules. Lacrimal gland lysate (1 ml) and kidney membrane lysate (0.5 ml) prepared in lysis buffer A were incubated overnight in absence or presence of 8 µg of ppEGF1 in absence or presence of 20 µg of peptide p437, and immune complexes were recovered as described in MATERIALS AND METHODS. Immunoprecipitates were dissolved and analyzed by Western blot with affinity-purified anti-p437 antibody (ppEGF1; 0.8 µg/ml) and probed with goat anti-rabbit IgG antibody linked to horseradish peroxidase as described in MATERIALS AND METHODS. Blots were developed with enhanced chemiluminescence (ECL) and visualized by exposure to Amersham Hyperfilm-ECL. Molecular mass markers are indicated at left, in kDa. Immunoprecipitated proteins are indicated at right.