Proc. Nati. Acad. Sci. USA Vol. 91, pp. 9362-9366, September 1994 Biochemistry Purification and sequence of rat (//intestine//radlolmmunoassay) NATHAN L. COLLIE*t, JOHN H. WALSHO, HELEN C. WONG*, JOHN E. SHIVELY§, MIKE T. DAVIS§, TERRY D. LEE§, AND JOSEPH R. REEVE, JR.t *Department of Physiology, School of Medicine, University of California, Los Angeles, CA 90024; *Center for Ulcer Research and Education, Gastroenteric Biology Center, Department of Medicine, Veterans Administration Wadsworth Center, School of Medicine, University of California, Los Angeles, CA 90073; and §Division of Immunology, Beckman Institute of City of Hope Research Institute, Duarte, CA 91010 Communicated by Jared M. Diamond, May 26, 1994

ABSTRACT Structural information about rat enteroglu- plus two glucagon-like sequences (GLP-1 and -2) cagon, intestinal containing the pancreatic glucagon arranged in tandem. The present study concerns the enter- sequence, has been based previously on cDNA, immunologic, oglucagon portion of proglucagon (i.e., the N-terminal 69 and chromatographic data. Our interests in testing the phys- residues and its potential cleavage fragments). iological actions of synthetic enteroglucagon peptides in rats Our use of the term "enteroglucagon" refers to intestinal required that we identify precisely the forms present in vivo. peptides containing the pancreatic glucagon sequence. Fig. 1 From knowledge of the proglucagon sequence, we syn- shows two proposed enteroglucagon forms, proglucagon-(1- thesized an enteroglucagon C-terminal octapeptide common to 69) (glicentin) and proglucagon-(33-69) (OXN; see Fig. 1). both proposed enteroglucagon forms, glicentin and oxynto- The primary structures based on amino acid sequence data of modulin, but sharing no sequence overlap with glucagon. We pig glicentin (6), ofdog glicentin and OXN (7), and ofalligator then developed a radilmmunoassay using antibodies raised gar (the holostean fish Lepisosteus spatula) OXN (8) have against the octapeptide that was specific for enteroglucagon been published. By contrast, information in the rat and peptides without cross-reacting with glucagon. Rat intestine human has relied upon a combination of cDNA, immuno- was extracted, and one presumptive enteroglucagon form was logic, and chromatographic techniques (9-11). However, the purified by following the enteroglucagon C-terminal octapep- peptide structure of either enteroglucagon form has not been tide-like immunoreactivity through several HPLC purification fully characterized in the rat and human. steps. Structural characterization ofthe material by amino acid We used a strategy similar to that of Blache et al. (12) to composition, microsequence, and mass spectral analyses iden- generate antibodies to the C-terminal octapeptide ECO com- tified the peptide as rat oxyntomodulin. The 37-residue peptide mon to both enteroglucagon forms, whose sequence was consists of pancreatic glucagon plus the C-terminal extension, deduced from the proglucagon cDNA sequence. This anti- Lys-Arg-Asn-Arg-Asn-Asn-Ule-Ala. This now permits synthesis body would then recognize both presumptive enteroglucagon ofan unambiguous duplicate ofendogenous rat oxyntomodulin forms without cross-reacting with glucagon. Our goal in this for physiological studies. study was to develop such a sequence-specific RIA, to isolate rat enteroglucagon peptides, and then to characterize their Gene cloning techniques have accelerated regulatory peptide peptide structure. research by providing rapid sequence information about potential physiological signals. In particular, we are inter- MATERIALS AND METHODS ested in those signals that regulate intestinal growth and nutrient absorption. Strong but circumstantial evidence has Synthetic Peptides. Two peptides for this study were syn- implicated the enteroglucagon, secreted by the thesized in the Peptide Biochemistry Core Facility of the intestine, as a signal for intestinal adaptation (1-3). Despite Veterans Administration/University of California at Los the cloning of the gene for enteroglucagon's precursor, Angeles Gastroenteric Biology Center: Lys-Arg-Asn-Arg- proglucagon (4), the endogenous forms ofrat enteroglucagon Asn-Asn-Ile-Ala (rat ECO) and Tyr-Lys-Arg-Asn-Arg-Asn- have not been fully characterized. We emphasize that only Asn-Ile-Ala (D-Tyr-ECO). These peptides were made on a the potential structures of the active peptides are derived Biosearch Sam II peptide synthesizer using tert-butoxycar- from the gene sequence because of the numerous processing bonyl-amino acid coupling strategies, cleaved with hydrogen steps that intervene between gene transcription and peptide fluoride, and purified by reverse-phase HPLC. Fractions of end product. Within a single species, tissue-specific differ- >90%o purity were pooled, lyophilized, and shown to have the ences in proglucagon processing, such as that occurring in the correct composition by amino acid and mass spectral anal- and intestine, further add to the structural hetero- ysis. Synthetic peptides for RIA cross-reactivity tests [hu- geneity of secreted peptides. These considerations place a man/rat/porcine vasoactive intestinal peptide (VIP), human premium on full structural characterization of peptides if we gastric inhibitory peptide (GIP), porcine OXN, glucagon, are to determine their precise physiological roles. Hence, our human , human GLP-1-(7-36) fragment, and ] study of enteroglucagon function in rats began with the were from Peninsula Laboratories. objective of unambiguously identifying the molecular forms HPLC Columns. Preparative HPLC C18 and C8 columns of this gut hormone. (21.4 mm x 25 cm; C18 83-223-C and C8 83-323-C) were Fig. 1 summarizes the structure and nomenclature of the obtained from Rainin (Woburn, MA). The semipreparative proglucagon precursor predicted from the cloned rat gene Vydac C4 (10 mm x 25 cm; 214 TP510) and the analytical sequence (4, 5). From the nucleic acid sequence, rat proglu- cagon is a 160-residue polypeptide containing pancreatic Abbreviations: ECO, enteroglucagon C-terminal octapeptide; ECO- LI, ECO-like immunoreactive material; OXN, oxyntomodulin; GLP-1 and -2, glucagon-like peptides 1 and 2. The publication costs of this article were defrayed in part by page charge tTo whom reprint requests should be addressed at the present payment. This article must therefore be hereby marked "advertisement" address: Department of Biological Sciences and Institute for Bio- in accordance with 18 U.S.C. §1734 solely to indicate this fact. technology, Texas Tech University, Lubbock, TX 79409. 9362 Downloaded by guest on October 2, 2021 Biochemistry: Collie et al. Proc. Nati. Acad. Sci. USA 91 (1994) 9363 PROGLU'CAGON were collected for columns of 2.1-cm diameter and larger. Fractions of4 ml were collected for 1.0-cm columns, and 1-ml NH (;ICr,(ENTI]N OXN fractions were collected for all other columns. I1- H Peptide Structurl Analyses. For amino acid analysis, a GRPP small portion (50-200 pmol) of the purified ECO-like immu- 30 noreactive material, which we call ECO-LI, was dried down F in borosilicate tubes (6 x 30 mm, previously heated to 4000C 33 T G I S for 24 hr). The tubes were placed into a hydrolysis chamber L D with 300 /4 of HCl containing 0.2% 2-mercaptoethanol and U heated to 1100C for 24 hr. After cooling, the HCl vapors were C K A removed by vacuum, and the amino acids were dissolved in G L Beckman amino acid dilution buffer and analyzed on a 0 Beckman 6300 amino acid analyzer. Data were collected on N R 61_ R a Nelson analytical data system. 64-- Purified ECO-LI was sequenced on a City of Hope gas- IP-1 69-, F phase microsequencer as described (13). Phenylthiohydan- 72'- FIc( toin derivatives ofamino acids were analyzed as described by Hawke et al. (14). GLP-1 L For mass spectral analysis, the purified peptide (25-100 NA K pmol) was analyzed with a JEOL HX 100HF soluble focusing 1081 T R magnetic sector mass spectrometer operating at 5-kV accel- 110- IP-2 erating potential. Sample ionization was accomplished by 123,. N a 6000-eV xenon atom beam. in R using Dry samples 1.5-ml 126- N polypropylene Microfuge tubes were taken up in 1-2/4 of5% N aqueous acetic acid and added to -1 /4 of the liquid matrix GLP-2 on a 1.5 x 6.0 mm stainless steel stage at the tip of a direct A insertion probe. The liquid matrix consisted ofdithiothreitol/ 160 dithioerythritol, 5:1 (vol/vol at room temperature), spiked with 6 mM camphorsulfonic acid. Positive-ion spectra were 01l collected under computer control using a JEOL DA5000 data system. Multiple scans over the mass range m/z = 100-3500 FIG. 1. Rat proglucagon (left-most vertical bar) undergoes ex- (cycle time, 45 s) were collected. tensive posttranslational processing in the intestine, yielding four Intetnal Extraction. Distal small intestine (jejunoileum) peptides related in sequence to pancreatic glucagon: glicentin and was at sacrifice from male rats oxyntomodulin (OXN), both containing the entire glucagon se- obtained Sprague-Dawley quence, as well as glucagon-like peptides 1 and 2 (GLP-1, GLP-2), (250- to 350-g body weight), rinsed with ice-cold Kreb's the latter two sharing 50-70%o homology to glucagon. Putative sites bicarbonate Ringer solution (pH 7.4), cut open longitudinally, of proteolytic cleavage are indicated by horizontal parallel lines blotted gently, snap-frozen on dry ice, and stored at -70°C (numbered residues) that border pairs of basic amino acids. The until extraction. Frozen intestinal tissue (107 g) was homog- vertical bars labeled OXN and ECO show the presumptive sequence enized in 4% aqueous CF3COOH (1 liter) at 4°C with a (given in single-letter amino acid code) for OXN deduced from the rat Polytron homogenizer. The extract was then centrifuged at proglucagon gene sequence (see ref. 4). Rat OXN is thus predicted 10,000 x g for 30 min, and the supernatant was applied to a to consist of glucagon plus enteroglucagon C-terminal octapeptide HPLC column. (ECO). GRPP, glicentin-related pancreatic peptide; IP-1 and -2, preparative intervening peptides 1 and 2. Intestinal extracts were similarly prepared from three additional species to test for species specificity in the ECO Vydac C18 and C4 columns (4.9 mm x 25 cm; 218TP54 and RIA: Swiss Webster mice (Mus musculus), golden-mantled 214TP54) were from Western Analytical Products (Hisperia, ground squirrels (Spermophilus lateralis), and New Zealand CA). The phenyl column (3.9 mm x 30 cm; 27198) was from White rabbits (Oryctolagus cuniculus). Before analysis, the a Waters. acetonitrile and water were extract supernatant from the test species was loaded onto HPLC-quality pur- prewashed (5 ml of acetonitrile) C18 Sep-Pak column (Wa- chased from Burdic and Jackson. CF3COOH for HPLC ters), rinsed with 5 ml of buffer A, and eluted with 1 ml of buffers was from Pierce and for extractions was from Sigma. buffer B. Dilutions of this elution buffer (1:10 to 1:10,000 in Peptide Purification. Before each purification step, a col- assay buffer) were tested for cross-reactivity and dilution- umn without injected peptide was run with the gradient to be curve parallelism in the ECO RIA. used for the ensuing step, with effluents assayed to insure ECO RIA. Antibodies to ECO were produced in 6-week- that previous runs of natural or synthetic peptides did not old New Zealand White female rabbits. Synthetic ECO, carry over into subsequent runs. The column effluents were coupled to keyhole limpet hemocyanin (Calbiochem) by the monitored at 280 and 214 nm for UV absorbance and by ECO carbodiimide method (15), was used as the immunogen. Six RIA (described below). Fractions with high immunoreactiv- rabbits were immunized with multiple intradermal injections ity to absorbance ratios were pooled, diluted 1:3 with 0.1% of an emulsion containing equal parts of complete Freund's CF3COOH (buffer A), loaded onto HPLC columns equili- adjuvant (Difco) and conjugated peptide (50-70 ug). Booster brated in the same buffer, and rinsed with 5 volumes of buffer injections of conjugate and incomplete Freund's adjuvant A. All reverse-phase steps followed the same sample dilution, were given at 6-to 8-week intervals. Antibody 8746 produced loading, and rinsing pattern, followed by bringing buffer B after the third booster injection had the highest titer (50%o acetonitrile containing 0.1% CF3COOH) to a starting (1:30,000) and requisite specificity (see Results) for use in all concentration with a 5-min linear gradient and by peptide subsequent assays. elution with a slow gradient specific to different column Synthetic D-Tyr-ECO was radioiodinated by using a mod- types. Preparative column gradients and running time were ified chloramine-T method (15). Labeled peptide was first 0-50%o buffer B for 100 min, and other column gradients were separated from free 125I by chromatography on a 1.0 x 10 cm 20-35% buffer B for60 min except the final C4 column, whose Sephadex G-10 column equilibrated with 0.05 M ammonium gradient was 45-65% buffer B for 40 min. Fractions (10 ml) acetate (pH 5.2) buffer containing 0.2% bovine serum albu- Downloaded by guest on October 2, 2021 9364 Biochemistry: CoUie et al. Proc. Natl. Acad. Sci. USA 91 (1994)

min. The peak radiolabeled fraction (1 ml) was further purified by reverse-phase HPLC on a Beckman C18 column 50 with a 40-min linear gradient of 0-50% buffer B in buffer A _% 150 at a flow rate of 1 ml/min. By self-displacement analysis, specific activity of the label measured 360 p.Ci/nmol. Ali- E B ~~~~-40 quots diluted with assay buffer (see below) and stored frozen 0 were EElo100 at -700C stable for >6 weeks. 30 The assay buffer in which standards and samples were Q diluted (1-ml final volume) was 0.05 M sodium phosphate (pH 7.4) with 2% Protenate (5% human plasma fraction, A 20 Baxter Hyland Division, Glendale, CA). Assay tubes con- o 50 C taining 1500 cpm (3 fmol) of label were incubated for 24 hr at LL 80120 160 20 4°C. Bound antigen was then separated from free antigen by adding dextran-coated charcoal suspension (15), centrifuging at 2000 x g for 10 min, and decanting the supernatant. Both 0 - the supernatant and pellet were assayed to determine, re- 40 80 1 20 1 60 200 spectively, the bound and free cpm. TIME (min) FIG. 3. ECO-LI profile of the initial C18 HPLC purification step RESULTS after extraction ofrat intestine. Fractions collected at retention times Rat ECO RIA. Antiserum 8746 exhibited the best titer and A-B (inclusive) were pooled and applied to a C8 HPLC column in a run separate from pooled fractions C-D (indicated by arrows). specificity for an ECO RIA. At afinal dilution of1:30,000, the Immunoreactive fractions from both runs were combined before the Ds50 was 97 fmol/ml with a least detectable dose of 25 fmol C4 HPLC purification step (see Table 1). %B (----) refers to the per tube (Fig. 2). Fig. 2 Upper shows that the antiserum percent of buffer B (50% acetonitrile in 0.1% CF3COOH) used in cross-reacted fully with extracts of mammalian (all column elution. species tested), believed to be the major intestinal site of enteroglucagon synthesis. Synthetic porcine OXN was also EXTRACT DILUTIONS equipotent in displacing labeled ECO from antibody binding 2 1 4f11 04 1 03 1 01 sites (Fig. 2 Lower). However, Fig. 2 Lower shows that 1 .0 neither glucagon nor other related peptides exhibited signif- 0 icant cross-reactivity. Apart from pig OXN, only vasoactive 0.8 intestinal peptide displayed slight cross-reactivity (0.2%). Hence, the ECO RIA was specific for enteroglucagon pep- 0.6 tides containing the ECO sequence. Addition of up to 20 p11 of HPLC buffer B caused no assay interference (data not shown). Thus, HPLC fractions diluted in assay buffer could 0.4 be tested, obviating the need for evaporating and reconsti- I. tuting samples. 0.2 Purification of Rat ECO-LI. Fig. 3 shows that a broad peak ck.* characterized the elution profile of the crude intestinal ex- 0.0 tract 34 IL. initially containing nmol of ECO-LI from 104 g of J tissue (Table 1). Several additional reverse-phase steps were M 0.8 used to further purify the fraction associated with the immu- noreactive peak, as shown in Table 1. The last step resulted 0.6 . V in a single absorbance profile (220 nm) associated with one U ECO-LI peak (Fig. 4). This material was structurally ana- C lyzed as detailed below. Structural Analysis of Purified ECO-LI. Amino acid (Table 0.4 2), microsequence (Table 3), and mass spectral analyses of 'p purified ECO-LI were consistent with the structure for rat OXN shown in Fig. 1. The amino acid composition of 0.2 ECO-LI agrees with that of rat OXN, predicted to comprise Table 1. Recovery of ECO-LI purified from rat jejunoileum 0.0 0.01 0.1 1.0 10 100 Purification ECO-LI, Step nmol % PEPTIDE (pmol/ml) step recovery, Extract 34 FIG. 2. Characterization of antibody 8746 for the ECO RIA. C18 HPLC 30 88 (Upper) Standard curve of bound/free (B/F) radiolabeled ECO as a C8 HPLC 20 67* function of synthetic rat ECO (e) concentration (bottom abscissa) C4 HPLC 15 80 and of various dilutions (top abscissa, number of fold dilution) of Phenyl HPLC 3 20t intestinal extracts from rat (o), mouse (A), ground squirrel (A), and C18 HPLC 7.7 51 rabbit (o). Note that all mammalian intestinal extracts exhibited C4 HPLC 3.1 40 dilution curves parallel to that for synthetic ECO. (Lower) Rat ECO (e) and porcine OXN (o) show identical binding characteristics for Starting material for the 4%o CF3COOH extract was 104 g of small antibody 8746, but other peptides with glucagon-sequence homology intestine (distal half). [A, secretin; v, GLP-1-(7-36); V, glucagon; a, gastric inhibitory *Only 18.7 nmol of ECO-LI fraction was used for next step. peptide; o, vasoactive intestinal peptide] exhibit little or no cross- tApparent low recovery was due to a shift in assay sensitivity, which reactivity in the ECO RIA. was corrected before the next purification step. Downloaded by guest on October 2, 2021 Biochemistry: CoUie et al. Proc. Natl. Acad. Sci. USA 91 (1994) 9365 Table 3. Microsequence analysis of rat ECO-LI Normalized yield, pmol Cycle Amino acid Run 1 Run 2 1 Histidine 4.7 2 Serine 9.3 1.5 3 Glutamine 14.6 2.6 4 Glycine 9.2 5 Threonine 6.3 6 Phenylalanine 17.8 7 Threonine 5.2 8 Serine 3.3 9 Aspartic acid 7.2 10 Tyrosine 10.1 11 Serine 3.3 12 Lysine 7.9 13 Tyrosine 8.1 14 Leucine 9.5 15 Aspartic acid 4.9 16 Serine 2.4 Time (min) 17 Arginine 5.2 18 Arginine 2.3 FIG. 4. Elution profile of the final HPLC purification step listed 19 Alanine 5.7 in Table 1. Absorbance at 220 nm (Ano) revealed one peak coincident 20 Glutamine 3.8 with a single fraction (m) containing ECO-LI. This fraction was 21 Aspartic acid 2.9 subjected to further characterization as described in the text. %B Cycles represent the peptide sequence of the first 21 amino acids refers to the percent of buffer B (---- -) used in column elution. of ECO-LI from its amino terminus. *The amino-terminal histidine was not recovered in run 1 because of glucagon and its carboxyl-terminal extension, Lys-Arg-Asn- HPLC misinjection but was confirmed in run 2 consisting of three Arg-Asn-Asn-Ile-Ala. Some discrepancies in Table 2 be- cycles. tween observed and expected OXN arise from the small peptide amount (65 pmol) analyzed. The amino-terminal 21 DISCUSSION residues of ECO-LI determined by microsequencing (Table We have purified and structurally characterized rat OXN, 3) are identical to residues 1-21 in rat glucagon. Phenylthio- one of two predicted molecular forms of rat enteroglucagon. hydantoin-derivative yields were too low for accurate amino Aided by an enteroglucagon-specific RIA, OXN and glicentin acid assignments beyond the 21st cycle. Mass-spectral anal- have been shown by detailed HPLC profiles to represent the ysis confirmed the average molecular mass of ECO-LI likely major forms ofenteroglucagon in rat intestine, plasma, (4450.78) to be equal, within experimental error, to that ofrat and brain (10, 12, 16). In these tissues, the two forms existed OXN (4450.91) calculated from the rat proglucagon gene in roughly equimolar quantities. Nevertheless, neither form sequence. had been sequenced in the rat. Our study employed a new enteroglucagon C-terminal sequence-specific RIA to follow Table 2. Amino acid composition of rat ECO-LI ECO-LI through a series ofreverse-phase HPLC purification steps. Combining these techniques with amino acid compo- Amino Residues per mol sition, microsequence, and mass-spectral analysis has pin- acid Expected Observed pointed the structure of rat OXN: a 37-residue peptide Ala 2 1.9 containing the full sequence of glucagon (29 amino acids) Arg 4 3.2 extended at its C terminus by the octapeptide Lys-Arg-Asn- Asx 7 3.8* Arg-Asn-Asnll-e-Ala (ECO). Glx 3 2.8 Below we address three topics related to this study: the Gly 3 2.8 need for full structural characterizations of peptides, com- His 1 1.9 parison ofthe ECO RIA with related assays, and the absence ile 1 2.3 of glicentin in our enteroglucagon purification. Leu 2 1.8 Heterogeneity of Enteroglucagon. As su d by Moj- Lys 2 2.1 sov et al. (17), heterogeneity ofrelatedpolypeptides within an Met 1 2.4 animal can arise from gene structure (duplication and diver- Phe 2 1.8 gence), from alternative RNA splicing, and from differential Pro 0 0.8 posttranslational processing. The results from these authors, Ser 4 2.9 showing that rat pancreas and intestine produce identical Thr 3 2.3 mRNA precursors, and from Heinrich et al. (4), showing the Trp 1 NDt presence of one proglucagon gene in the rat genome, narrow Tyr 2 2.4 the focus of enteroglucagon heterogeneity to the third expla- Val 1 1.8 nation: alternative proglucagon processing. Thus, the pan- creas secretes glucagon and the intestine secretes enteroglu- Analysis was based on approximately 65 pmol ofpurified ECO-LI, two an shown above as the observed composition. Expected composition is cagon peptides because the tissues cleave identical that for rat OXN deduced from the rat proglucagon gene sequence. precursor at different sites. *Aspartic acid/asparagine residues typically yield low recoveries on It follows that one cannot predict cleavage sites, and hence analysis. peptide structures, from a knowledge of typical cleavage tND, not determined because of loss during acid hydrolysis. motifs (i.e., pairs of basic amino acids). For example, the Downloaded by guest on October 2, 2021 9366 Biochemistry: CoUie et al. Proc. Natl. Acad. Sci. USA 91 (1994) pancreas cleaves proglucagon at two Lys-Arg pairs (proglu- been degraded into OXN plus the amino-terminal fiagment, cagon residues 31-32 and 62-63) to produce glucagon. Our glicentin-related pancreatic peptide (GRPP; see Fig. 1), dur- study confirms that the intestine ignores the second site and ing purification on the HPLC columns. For example, chole- produces OXN. In addition, proteolysis can occur at sites of cystokinin-58 conversion into smaller peptides has been single basic residues, as occurs in GLP-1 processing (at an observed on HPLC columns (J.R.R., unpublished observa- arginine residue, proglucagon residue 77; ref. 18) and in tion). Hence, glicentin may be more susceptible than is OXN at numerous sites (19). Other posttransla- to column degradation, accounting for the absence ofglicen- tional modifications, including amidation, sulfation, phosphor- tin in our purification. ylation, and glycosylation, further complicate the accurate In conclusion, the diversity of posttranslational modifica- prediction ofpeptide end products from cDNA, immunologic, tions makes deductions about peptide end products from and chromatographic data alone. Thus, unambiguous struc- cDNA sequences difficult. Now that rat OXN's structure has tural information remains essential to studies of peptide ac- been proven chemically, physiological studies in the rat can tion. proceed with confidence using synthetic duplicates of the ECO RIA. Raising antiserum 8746 against the synthetic endogenous peptide. ECO peptide resulted in an assay that recognized OXN without cross-reacting with glucagon. This permitted us to We thank F. J. Ho ofthe Peptide Biochemistry Core for help with track enteroglucagon immunoreactivity by using a single HPLC purifications, Jason Zhu for assistance with graphics, and assay. The more conventional RIA for enteroglucagon, using Kimberly Hammond for providing the ground squirrels. Funding for an N-terminal glucagon antibody cross-reacting with both this study was from the National Institute ofDiabetes, Digestive, and intestinal and pancreatic and then subtracting the Diseases (Grant DK 42973) and from the Veterans Admin- pancreatic glucagon activity measured with a second, gluca- istration Research Service, National Institutes of Health Center gon-specific antiserum, has the obvious disadvantages of Grant (DK 41301). requiring twice the time and reagents of a single assay and of compound error from two assays. 1. Bloom, S. R. & Polak, J. M. (1982) Scand. J. Gastroenterol. Blache et al. (12) used a similar approach to develop an 72, 93-103. RIA 2. Lund, P. K., Ulshen, M. H., Rountree, D. B., Selub, S. E. & enteroglucagon C-terminal specific (antiserum "LEG"). Buchan, A. M. (1990) 46, 66-73. Their antiserum and ours exhibit similar specificity toward 3. Rountree, D. B., Ulshen, M. H., Selub, S., Fuller, C. R., OXN, comparable affinity constants (ca. 1010 M-1; N.L.C., Bloom, S. R., Ghatei, M. A. & Lund, P. K. (1992) Gastroen- unpublished observation), and similar titers (respectively, terology 103, 462-468. 1:20,000 and 1:30,000). Two advantages of our assay, par- 4. Heinrich, G., Gros, P., Lund, P. K., Bentley, R. C. & ticularly for screening large numbers of HPLC fractions, Habener, J. F. (1984) Endocrinology 115, 2176-2182. were the short incubation time (24 hr vs. 4 days for the LEG 5. Conlon, J. M. (1988) Diabetologia 31, 563-566. antiserum) and the lack of interference from HPLC sample 6. Thim, L. & Moody, A. J. (1981) Regul. Pept. 2, 139-150. buffer, which obviates the need to dry and reconstitute 7. Shinomura, Y., Eng, J. & Yalow, R. S. (1988) Regul. Pept. 23, fractions before assay. 299-308. An additional utility of the ECO RIA is its insensitivity to 8. Pollock, H. G., Kimmel, J. R., Ebner, K. E., Hamilton, J. W., species differences. Fig. 2 shows that porcine OXN and Rouse, J. B., Lance, V. & Rawitch, A. B. (1988) Gen. Comp. Endocrinol. 69, 133-140. mouse, rabbit, and squirrel intestinal extracts cross-react 9. 0rskov, C., Holst, J. J., Poulsen, S. S. & Kirkegaard, P. (1987) fully with antiserum 8746. Porcine (6), canine (7), and bovine Diabetologia 30, 874-881. (20) ECO differ from the rat and human sequence at a single 10. Kervran, A., Blache, P. & Bataille, D. (1987) Endocrinology position (i.e., proglucagon residue 65 is lysine in the former 121, 704-713. and arginine in the latter species). Thus, our assay appears 11. Fuller, P. J., Beveridge, D. J. & Taylor, R. G. (1993) Gastro- uniformly valid for all mammalian species examined, despite enterology 104, 459-466. single-residue substitutions. 12. Blache, P., Kervran, A., Martinez, J. & Bataille, D. (1988) OXN, but No Glicentin. At the outset ofthe purification, we Anal. Biochem. 173, 151-159. to recover at least two to 13. Shively, J. E., Miller, P. & Ronk, M. (1987) Anal. Biochem. expected peptides, corresponding 163, 517-529. rat glicentin as well as the OXN eventually purified. Possible 14. Hawke, D., Yuan, P. M. & Shively, J. E. (1982) Anal. Bio- reasons for our failure to purify glicentin include our extrac- chem. 120, 302-311. tion conditions, lack ofglicentin cross-reactivity in the ECO 15. Walsh, J. H. & Wong, H. C. (1987) in Radioimmunoassay in assay, and degradation or losses during purification. We can Basic and Clinical Pharmacology, eds. Patrono, C. & Peskar, tentatively rule out the first two explanations, because we B. A. (Springer, Berlin), pp. 315-334. used a similar extraction protocol to analyze rat ileum 16. Blache, P., Kervran, A. & Bataille, D. (1988) Endocrinology enteroglucagon forms on a Sephadex G-50 column. The 123, 2782-2787. elution profile of immunoreactivity showed two ECO-LI 17. Mojsov, S., Heinrich, G., Wilson, I. B., Ravazzola, M., Orci, peaks with estimated molecular weights corresponding to L. & Habener, J. (1986) J. Biol. Chem. 261, 11880-11889. 18. Kreymann, B., Yiangou, Y., Kanse, S., Williams, G., Ghatei, glicentin and OXN (N.L.C., F. J. Ho, and J.R.R., unpub- M. A. & Bloom, S. R. (1988) FEBS Lett. 242, 167-170. lished observation). Thus, both peptides were detected by 19. Eysselein, V. E., Ebrlein, G. A., Schaeffer, M., Grandt, D., our ECO assay. However, we were unable to make a direct Goebell, H., Niebel, W., Rosenquist, G. L., Meyer, H. E. & cross-reactivity comparison in the ECO RIA between glicen- Reeve, J. R., Jr. (1990) Am. J. Physiol. 258, G253-G260. tin and OXN because glicentin is currently unavailable in 20. Lopez, L. C., Frazier, M. L., Su, C.-J., Kumar, A. & Saun- pure form. The final possibility is that glicentin may have ders, G. F. (1983) Proc. Natl. Acad. Sci. USA 80, 5485-5489. Downloaded by guest on October 2, 2021