Purification and Characterization of Chymodenin A HORMONE-LIKE PEPTIDE FROM PORCINE DUODENUM*

Chymodenin, a hormone-like duodenal peptide which rapidly alters the proportions of secreted pancreatic digestive enzymes to a mixture relatively richer in chymotrypsinogen than that found in basal secretion, has been purified to homogeneity. The starting material was an acidic methanol-soluble, neutral pH-insoluble fraction of an acetic acid extract of porcine duodenum; the purification consisted of cation-ex-change chromatography on SP-Sephadex and CM- Sephadex in ammonium bicarbonate gradients, and gel filtration on Sephadex G-75 in dilute acetic acid. The yield of material was followed by radioimmunoassay. Homogeneity was determined from chymodenin’s be- havior in disc gel electrophoresis in an acidic counter-migration-of-dye system, sodium dodecyl sulfate-urea gels, gel filtration, dansyl-Edman reaction, reversed-phase high pressure liquid chromatography, isotacho- phoresis, and sedimentation equilibrium ultracentrifugation. The electrophoretic mobility, the molecu- lar weight of 9,000-10,000, and the biological activity differed from those of other gastrointestinal peptide hormones. The amino acid composition was unique. Chymodenin is the first purified hormone-like sub- stance reported capable of altering the composition of the mixture of secreted digestive enzymes, independent of the stimulation of massive pancreatic protein output.

protein secretion by the pancreas, since only parallel alterations in enzyme output could be explained by the exocytotic release of zymogen granules with their premixed contents (4).
A recent re-examination of this problem (5) revealed, however, that the relative proportions of several pairs of digestive enzymes were rapidly altered by chymodenin treatment from the proportions found either in the basal or CCK'-stimulated state. Recent studies further showed that the rapid alterations in enzyme proportions observed were accompanied as well by a concomitant increase in the degree of linkage or correlation among the secretory outputs of the prepackaged enzymes. Thus, chymodenin appeared to exert its action by selection of a specific subset of prepackaged proteins for secretion from among several such available pools or subsets; the action of the peptide was also unique in that it appeared capable of regulating enzyme secretion without itself being a major secretory stimulator. The purification, evidence for homogeneity, and chemical characteristics of chymodenin have never been published other than in summary form (6); we present these aspects here.

RESULTS~
Polyacrylamide Gel Electrophoresis of Chymodenin- Fig. 4 is a photograph of patterns of purified and partially purified chymodenin after polyacrylamide disc gel electrophoresis. The gels show progressive degrees of purification from left to right. At the farthest right, the gel was "overloaded" with 100 pg of purified chymodenin to increase the detection limit for minor impurities. When scanned with a densitometer, no impurity bands could be detected, under conditions where 1-2 pg of an impurity were detectable. Electrophoresis of purified chymodenin with authentic secretin, CCK, and vasoactive intestinal peptide (VIP) indicated that chymodenin migrated differently from these gastrointestinal peptide hormones.
Portions of this paper (including "Materials and Methods," parts of "Results," Figs. 1-3 and 6, and Tables I1 and 111) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are available from the Journal of Biological Chemistry, 9650 Rockville Pike, Bethesda, MD 20814. Request Document No. 85M-3706, cite the authors, and include a check or money order for $4.60 per set of photocopies. Full size photocopies are also included in the microfilm edition of the Journal that is available from Waverly Press.

a b
C d e f g FIG. 4. Polyacrylamide disc gel electrophoresis of chyrnodenin purification fractions. a, 100 pg of AMESNI starting material after desalting on Sephadex G-25; b, 100 pg of pooled material from SP-Sephadex column (Fig. 1); c, 100 pg of pooled material from CM-Sephadex column (Fig. 2); d, 100 pg of pooled material from Sephadex G-75 column (Fig. 3); e, f, g, 10 pg, 50 pg, and 100 pg, respectively, of final chymodenin preparation after Sephadex G-75 filtration. Conditions are described in the text. of M, = 9,000 k 500. A single faint impurity band of M, 20,000 was detected in one chymodenin preparation when 20 pg or more of chymodenin per gel was applied, but was absent in other preparations. The level of this impurity was estimated as less than 3% by weight in comparison to the staining intensity of known trace quantities of protein of similar molecular weight scanned by the same procedure. The apparent molecular weight of chymodenin was not changed by the presence or absence of sulfhydryl reducing agent, indicating that the molecule does not contain multiple polypeptide chains linked by disulfide bonds.

Reversed-phase High Pressure Liquid Chromatography-
This procedure was carried out in an aqueous acetonitrile gradient containing 0.1% trifluoroacetic acid and revealed a single homogeneous peak which eluted at a 40% acetonitrile concentration. The detection limit for impurities in this system was estimated at 2-3%. Amino Acid Composition of Chymodenin-A summary of the amino acid composition of chymodenin is presented in Table I. The expected number of residues per molecule was estimated from the amino acid composition, assuming a M, = 10,000. This assumed molecular weight gave very close to integral numbers of residues for almost all the amino acids. In turn, the minimum molecular weight calculated from these integral numbers was 10,270-10,670. This number was somewhat greater than the molecular weight of 9,000 f 500 ob- served on SDS-urea-polyacrylamide gel electrophoresis, and was consistent with the molecular weight calculated from sedimentation equilibrium ultracentrifugation. Chymodenin appears to contain two disulfide bonds and no free sulfhydryl groups. I~do['~C]acetic acid was not incorporated into chymodenin in the absence of reducing agent, under reaction conditions specific for sulfhydryl groups.

Purification and Homogeneity of Chymodenin
The degree of homogeneity of chymodenin purified by the procedure described in this paper was tested by several analytical methods. Levels of detection of any single peptide impurity of chymodenin could be estimated for each method. With SDS-polyacrylamide gel electrophoresis, for example, 0.5 pg of a single impurity band whose molecular weight differed from that of chymodenin by ?lo00 could be detected in our system. With 20 pg of chymodenin in the sample, a minor contaminant peptide of about 20,000 was detectable at a level corresponding to 0.5 pg. We therefore estimated no single peptide impurity of different molecular weight was present a t a level greater than about 2-3% by weight of chymodenin. Similarly, high resolution polyacrylamide gel electrophoresis (18) demonstrated that less than 2% of any single impurity of different mobility was present. No peptide contaminant with an unblocked amino-terminal residue was observed after dansylation of the chymodenin preparation under conditions which would detect an estimated 1% impurity. In addition, potential contaminating peptides having NH2-terminal pyroglutamyl residues, blocked to dansylation but susceptible to cleavage by pyrrolidonecarboxylyl peptidase, would have had to be present at levels below 10% or else be followed by a half-cystine residue in their amino acid sequence to have been unobserved. Peptides with NHs-terminal pyrrolidonecarboxylic acid followed by proline, which the enzyme could not cleave, would not have been detected (24). Peptide impurities with blocked NHp-termini which were not susceptible to cleavage by pyrrolidonecarboxylyl peptidase might have been present and not detectable by the methods used. Such peptides have not been observed thus far among known gastrointestinal peptides. Finally, reversed-phase high pressure liquid chromatography of the whole molecule gave a single homogeneous peak, without detectable impurities being found.
Other sensitive methods which provided qualitative estimates of homogeneity (isotachophoresis, ultracentrifugation, gel filtration) did not detect peptide contaminants. Taken together, these results strongly suggest that chymodenin has been purified to homogeneity at the detection limits of existing techniques.
Chymodenin, by all criteria employed, including disc gel electrophoresis, amino acid composition, NH2terminal pyrrolidonecarboxylic acid and molecular weight, is distinct from other known gastrointestinal hormones. Radioimmunoassay showed no cross-reactivity with secretin, CCK, or VIP (26).
The radioimmunoassay reacted weakly with gastric inhibitory polypeptide (1/1000 as strongly) and glucagon (1/10,000 as strongly). Partial sequencing revealed sequence homology to gastric inhibitory polypeptide and glucagon in a limited region, but chymodenin was different in the regions immediately preceding and following (27).
Amino acid composition and molecular weight of chymodenin also distinguished the substance from other gastrointestinal and pancreatic peptides studied so far. Gastric inhibitory polypeptide, VIP, secretin, gastrin, pancreatic polypeptide, motilin, glucagon, and insulin, for example, have molecular weights below about 6000. The fact that chymodenin was isolated from an acidic methanol-soluble fraction, which might be expected to contain only peptides below M, = 5000, suggests it has unusual solubility properties. Chymodenin's amino acid composition showed a high content of aromatic amino acids and proline. It has four half-cystine residues, with no free sulfhydryl groups detected by reaction with iodoacetic acid. NH2-terminal pyrrolidonecarboxylic acid, found in chymodenin, is shared only by gastrin among the gastrointestinal peptides.
Actiuity of Chymodenin The activity of chymodenin was distinct with respect to the exocrine pancreas; the molecule had no secretin-like activity on pancreatic fluid flow in the cat and exhibited no massive nonspecific stimulation of enzyme secretion, such as would be provoked by CCK or cholinergic stimulation. The major known activity of purified chymodenin is the enrichment of chymotrypsinogen secretion from the rabbit pancreas which has been reported earlier (1,2), by changing the preset mixture of secreted digestive enzymes to one richer relatively in chymotrypsinogen ( 5 ) . In the course of working out the purification procedure for chymodenin, several distinct biological effects were found and attributed to the molecule. It is advantageous to briefly discuss these in the order in which they were found, in order to place them in context.
Release of Enzymes from Pancreatic Zymogen Granules in Vitro-This activity was first described by Adelson and Ehrlich (10) in crude extracts of porcine duodena which were prepared commercially by an ethanol extraction method (28). This activity was employed in the search for a suitable starting material for the purification described above. Although amylase activity was used for convenience of assay as the principal enzyme to indicate the release of the zymogen granule contents in vitro, it was also demonstrated that other enzymes, including proteases, were released (10). This zymogen granule enzyme releasing activity was present in pure chymodenin (Fig. 6), but as discussed above, the activity was not specific to chymodenin. Although neither secretin nor CCK elicited the release of zymogen granule contents (IO), several peptides present in the AMESNI mixture did do SO and it was not possible to accomplish the purification of chymodenin on the basis of this activity alone. We now view the release of zymogen granule contents as an unusual but purely physicochemical activity present in a number of peptides which exist in the AMESNI fraction, and we are presently unaware of whether this activity has any particular biological significance. The observation that this activity occurs over at least four orders of magnitude with no plateau (Fig. 6) supports the likelihood that the chymodenin-elicited release of enzymes from the zymogen granules is not a receptor-mediated phenomenon.
Increase in Rat Pancreatic Fluid Output Following Intrauenous Administration of Chymodenin-This activity was first found employing AMESNI, and later described for the nearly pure molecule (29). We report the activity of the pure molecule here (Table 111). Several other polypeptides have been shown to increase basal pancreatic fluid output in the rat or other species including secretin, CCK, and VIP. The rat apparently responds to these substances by increasing the output of both pancreatic fluid and protein (30), while in other species such as the cat, pancreatic response is observed to be more specific, with increased fluid output (e.g. secretin and VIP (31), or increased protein output (e.g., CCK), but not both). This fluid-stimulating biological activity was present in AMESNI, and in conjunction with the zymogen granule releasing activity and gel electrophoresis, it provided a basis for developing the initial purification; the method was, however, subject to a high degree of variance and was therefore only marginally reliable. On several occasions this inadequacy led to the selection of fractions for further purification which were later proven to be inactive on retesting.
Stimulation of Chymotrypsinogen-enriched Secretion-The effect of chymodenin on the secretion of chymotrypsinogen by the pancreas was first demonstrated using the pure molecule following its purification on the basis of the earlier criteria described above (32). Characterization of the activity in enrichment of pancreatic chymotrypsinogen content by purified chymodenin has been published previously. This effect has, however, been controversial (33, 34) because the authors of the earlier work interpreted the data to mean that chymotrypsinogen alone was secreted from the acinar cell (2). Since the pancreatic enzymes are secreted via exocytosis of the contents of the zymogen granule which contains a complete mixture of the pancreatic enzymes, the exocytotic mechanism could not be expected to directly account for the type of nonparallel or enzyme-specific secretion of chymotrypsinogen observed following chymodenin treatment. More recently, however, Adelson and Miller reported (5) that the mechanism underlying nonparallel secretion is selective stimulation of secretion by exocytosis from chymotrypsinogenrich heterogeneous pools of preformed enzyme within the gland. This mechanism is consistent with exocytosis as the fundamental process underlying secretion by the pancreas.
Measurement of the activity resulting in chymotrypsinogen enrichment of the pancreatic fluid and the increased correlation among enzymes was done by painvise regression analysis of chymotrypsinogen uersus lipase outputs in a series (usually 6-8) of experimental animals compared with an equivalent number of time-paired controls. As such, this determination was sufficiently laborious that it could not be expected to be routinely applicable for use in monitoring purification fractions for activity. At present, this is the most specific unique activity ascribed to the molecule, and the only one thought by us to potentially represent a true physiological role for the molecule (1, 2, 5).
Chymodenin provides a link between the intestine and the exocrine pancreas as a potential hormonal mediator of the composition of the digestive enzyme mixture in response to the presence of a specific (and as yet uncharacterized) stimulus in the lumen of the intestine. The demonstration of the existence of chymodenin greatly increases the likelihood that the rapid alteration of the enzyme proportions of digestive enzymes is of functional importance in digestion. Although this activity in altering pancreatic secretion is the most intriguing of the potential functional roles which may be ascribed to chymodenin, it is by no means clear at present that this is indeed chymodenin's primary action or role. Such demonstration awaits more physiological characterization. It should be remarked that the enzyme proportion-altering activity would be extraordinarily difficult to detect until the degree of purification of the molecule from crude duodenal extracts was sufficient to separate it from virtually all the CCK, and other possible protein secretion-stimulating substances also present in these extracts, e.g. VIP, which may well mask the enzyme-specific alterations. Fortunately, the other earlier-observed effects of the molecule upon enzyme release from isolated zymogen granules, and the flow of pancreatic fluid in the rat, did provide a basis for this purification. An analogous situation may well exist in other systems where a single regulatory substance present in a crude glandular extract produces a response of such a magnitude or nature as to mask the presence of other important regulatory substances.
Acknowledgments-We would like to dedicate this paper to the memory of Prof. J. Eric Jorpes of the Karolinska Institute in Stockholm in appreciation of his long-term efforts with the hormones of the gastrointestinal tract, and his early interest in chymodenin. We thank Prof. V. Mutt, Stockholm, Sweden, for criticism, advice, and help with the starting material and purification; and Prof. E. R. Blout, Boston, MA, for his interest and support during the initial stages of the purification work. a f t e r , GARGG addition. Labeled chymodenin was prepared by reaction with Ne l p 5 1 in the preaence of lactoperoxidase end l i m i t i n g q u a n t i t i e s of hydrogen peroxide, uaing a modification of the method of Thorell and Johansson (12). Antiserum to chymodenin was prepared in rabbits using hemocyanin-conjugated chymodenin as antigen (13,14).1

Note Added in
Chemical and enzyme aesaye -Amylase was eaaeyed by the procedure described   s t a r t i n g m a t e r i a l and the final product released the contents of pancreatic Release of enzymes from p a n c r e a t i c zymogen granules -Both the AMESNI of pancreatic hydrolytic enzyme8 f r m t h e zymogen granules was assayed by zymogen granules in a logarithmic dose-dependent fashion (Fig.  6 ) . The release demonstrated (10,25). AMOUNT ADDED (ny) Fig. 6 . E f f e c t o f i n c r e a s i n g amounts of chymodenin and "AMESNI" s t a r t i n g m a t e r i a l upon a n y l a s e r e l e a s e from i s o l a t e d p a n c r e a t i c zymogen g r a n u l e s i n v i t r o . T r i p l i c a t e i n c u b a t i o n s were run; conditions are described in the text.
Effect of chymodenin on p a n c r e a t i c f l u i d s e c r e t i o n i n t h e r a t -R a t s i n j e c t e d with nanagrm mounts of chymodenin showed a s i g n i f i c a n t i n c r e a s e i n t h e f l o w r a t e from the pancreas. Data from experiments with eeven animals, presented in Table 111