PROTEIN EXPRESSION AND PURIFICATION 10, 169–174 (1997)ARTICLE NO. PT970726

Isolation and N-Terminal Sequence of Multiple Formsof Granulins in Human Urine

Giulia Sparro, Gabriella Galdenzi, Anna Maria Eleuteri, Mauro Angeletti,Werner Schroeder,* and Evandro FiorettiDepartment of Molecular, Cellular, and Animal Biology, Post-Graduate School in Biochemistryand Clinical Chemistry, University of Camerino, 62032-Camerino (MC), Italy;and *Pharma Research Center, Bayer AG, 42096-Wuppertal, Germany

Received May 21, 1996, and in revised form January 17, 1997

phils (4) and from neural tissue of Locusta migratoriaThree molecular forms of granulins (also known as (5). Four molecular forms (granulins A, B, C, and D)

epithelins) were isolated, for the first time, in human were purified, and sequenced, from human leukocytesurine. Their N-terminal sequences, which have also (1). All the isoforms show a molecular architecturebeen determined, are identical to those of granulins A characterized by the presence of 12 cysteine residuesandB, previously isolated from human leukocytes, and over the 56–57 amino acid ones which constitute theof granulin F, never isolated before butwhose primary mature molecule (1). Granulins, in fact, originate fromstructure is known on the basis of the cDNA sequence. a posttranslational proteolytic processing of a largerThe urinary molecules, which show a molecular precursor containing seven and one-half granulin-likeweight of about 6.5 kDa, are most likely produced by repeats (6–8), with each repeat appearing to be en-a posttranslational proteolytic processing occurring coded by two different exons (9). The particular organi-at the level of the kidney, which appears to be the zation of the granulin gene could allow the generation,organ richest in granulin precursormRNA. Themolec- by alternate splicing, of hybrid granulin-like molecules,ular events underlying the precursor processing are

with peculiar biological properties, in addition to thoseunknown, even though the involvement of the prote-so far isolated. Granulin gene expression, studied inase kallikrein, an enzyme thought to be responsible fordifferent cell types through Northern blot analysis, in-the processing of several polypeptidic growth factordicates an unclear interspecies difference (6–8). Geneprecursors, could be hypothesized. Granulins, how-expression is, however, significantly higher in epithe-ever, do not show antikallikrein activity. The presencelia-rich tissues, with particular reference to the kidney,in human urine of isoform F, previously not identifiedsuggesting for these molecules an important role infrom other human sources, seems to support the hy-renal function (6). It has been suggested that renalpothesis that mature forms of granulins are generatedgranulins could derive from leukocyte entrapment (7),by an organ-specific precursor processing, on the basisbut ‘‘in situ’’ hybridization studies clearly indicate thatof particular physiological requirements, and to sug-

gest also that this isoform may play ‘‘in vivo’’ an im- the kidney epithelial cells express ‘‘in vivo’’ the gra-portant and specific role in the epithelial cells of the nulin mRNA, at least in rat (6). The hybridization sig-human kidney. q 1997 Academic Press nal was particularly marked in the epithelial cells of

the proximal and distal convolute tubules and of Bow-man’s capsule (6). The same cellular districts appear tobe characterized by the presence of the serine proteasekallikrein (10,11), an enzyme thought to play an im-Granulins (also known as epithelins) are cysteine-portant role in renal function, and supposed to be in-rich polypeptides which show a modulating effect onvolved in proteolytic maturation of polypeptidic growthproliferation of the epithelial cells ‘‘in vitro’’ exerting afactor, such as EGF and NGF (12–16) from their pre-growth factor-like activity (1,2). They have been iso-cursors. In our laboratory, it was subsequently shownlated in multiple forms as low-molecular-weight pep-that renal kallikrein activity could be modulated bytides (about 6.5 kDa) from the hematopoietic cells ofprotein inhibitors belonging to the Kunitz family of pro-humans (1), rats (2), and teleost fish (3). Granulin-like

molecules have also been isolated from equine neutro- teinase inhibitors (17,18), immunohistochemically de-

1691046-5928/97 $25.00Copyright q 1997 by Academic PressAll rights of reproduction in any form reserved.

AID PEP 0726 / 6q10$$$621 06-04-97 13:04:41 pepa AP: PEP

SPARRO ET AL.170

tected in the nephron epithelial cells (19). During a 0.1% trifluoroacetic acid (TFA) in water (solvent A) andeluted with a gradient obtained by mixing solvent Astudy addressed to demonstrate the possible excretion

of kallikrein inhibitors in the urinary flow, we found with a solution of 4:1 (v/v) acetonitrile–2-propanol in0.1% TFA (solvent B) at a flow rate of 1 ml/min. Thethat in human urine, in addition to kallikrein inhibi-

tor(s) (manuscript in preparation), three isoforms of samples were prepared by dissolution with solvent Ajust before use. Absorbance of the eluted fractions wasgranulins are present. Their N-terminal sequences,

which have been used for identification, are identical recorded at 254 nm. Chromatographic data were ana-lyzed with the Gold software.to those of granulins A and B, previously purified from

human leukocytes (1), and of granulin F, never isolated Analytical reversed-phase (RP-) and gel permeationchromatography (GPC) were performed with a Smartbefore but whose primary structure was deduced on

the basis of the cDNA sequence (6,7). System apparatus (Pharmacia, Sweden) using a SmartRP-18-column (2.1 1 100 mm) and a Superdex-30 col-umn (3.2 1 300 mm), respectively, both obtained fromMATERIALS AND METHODSPharmacia. The elution conditions are reported in theChemicals legends to Fig. 3 and Fig. 4, respectively.

Porcine pancreas kallikrein and aprotinin were thekind gift of Bayer AG (Germany). The synthetic sub- Antikallikrein Activitystrate S-2266, used for measuring kallikrein activity,

The antikallikrein activity was measured by follow-was obtained from Kabi Vitrum (Sweden) and Sep-Paking the inhibitory effect exerted, at 405 nm, on theC18 cartridges from Millipore (USA). All other chemi-enzymatic hydrolysis of the synthetic substrate S-2266cals were of analytical grade.as previously reported (20). The presence of organicsolvents, which affects the proteolytic activity, was

Source taken into account, in the reference test, by addingThe urine samples utilized for granulin purification an eluate volume from a run performed without the

were usually collected in the morning and treated as inhibitor (20).soon as possible after collection. The samples utilizedfor sequence studies were obtained from about 5 liters SDS–Gel Electrophoresis and Protein Blottingof urine, collected from the same donor, even though

SDS–polyacrylamide gel electrophoresis, under re-samples independently purified from different donorsducing conditions, was performed according to Laem-were analyzed (see Results).mli (21) using a 10–20% polyacrylamide gradient. Forsequence studies the gel was blotted overnight onto a

Isolation of Granulins PVDF membrane according to Gultekin and HeermannIsolation of urinary granulins was obtained through (22). The blotted protein bands were stained with Coo-

a procedure, reported in detail under Results and sum- massie brilliant blue R-250, cut, and sequenced as re-marized in Table 1, which is very similar to the purifi- ported below.cation protocols already utilized for isolation of granul-ins from other sources (1–3). The purification steps N-Terminal Sequence Analysiswere checked from densitometric scans of SDS–poly-

Individual fractions from the Smart System were ap-acrylamide gel electrophoresis as the ratio between theplied onto a Polybren pretreated glass fiber sheet sub-density shown by the 6.5-kDa band and the total pro-sequently loaded on a Model 473A Applied Biosystemstein density (see Table 1).protein sequencer. The fast normal sequencer cycle wasused. The PTH-amino acids were analyzed by on-line

Affinity Chromatography reversed-phase chromatography. Blot sequencing wasThe affinity chromatography column was prepared performed using the fast blot cycle.

immobilizing the porcine pancreas kallikrein to CNBr-activated Sepharose 4B (Pharmacia, Sweden) accord- RESULTSing to the suppliers’ instructions.

To eliminate possible interactions of granulins withother urinary proteins, the freshly collected urine sam-

HPLC ple, usually 200–250 ml, was added to a solution of 1M glycine, in order to achieve a 50 mM concentrationHPLC analyses were performed with a System Gold

apparatus (Beckman, USA) using a Supelcosil LC18-DB of the amino acid, and brought to pH 3.0 with a concen-trated solution (37%) of HCl. The mixture was made(25 1 4.6 cm) column equipped with a Supel-Guard

LC18-DB (2.5 1 4.6 cm) guard column, both obtained 500 mM in NaCl by addition of a 4 M solution of thesalt and then was maintained for 1 h, at 47C, underfrom Supelco (USA). The column was equilibrated in

AID PEP 0726 / 6q10$$$621 06-04-97 13:04:41 pepa AP: PEP

HUMAN URINE GRANULINS 171

stirring. This solution was made 2.5% in trichloroaceticacid by slow addition, under stirring, of a concentratedsolution (50%) of the acid. The supernatant obtainedafter centrifugation was brought to pH 8.6 with a con-centrated solution of NaOH and applied to a porcinepancreas kallikrein–Sepharose 4B column (2 1 10 cm)equilibrated in 200 mM Tris–HCl buffer, pH 8.6. Thecolumn was washed with the equilibrium buffer andthen eluted with 250 ml of a solution of HCl at pH 2.0.The eluate was neutralized at pH 7.0 with a 1 M Trissolution and loaded on a freshly activated Sep-Pak C18

cartridge, which was eluted with a solution obtainedby mixing 60% of solvent A and 40% of solvent B (seeMaterials and Methods). The fraction obtained fromthe cartridge (about 3 ml) was concentrated to drynessand stored at 0207C. In order to have enough material

FIG. 2. SDS–polyacrylamide gel electrophoresis (10–20%) of Peakfor the sequence studies (see below), a sufficiently high I and Peak II recovered from the RP-C18 column (see Fig. 1). Thenumber of these fractions (corresponding to about 5 analysis was performed on the material obtained from three indepen-liters of starting urine) were dissolved in the smallest dently purified preparations (see text). Lane 1: protein markers

(phosphorilase 97 kDa, bovine serum albumin 68 kDa, ovoalbuminquantity of the mixture utilized for the Sep-Pak C1843 kDa, carbonic anidrase 31 kDa, trypsin inhibitor 21.5 kDa, lysoz-cartridge elution (see above), collected, and dried againime 14 kDa). Peaks I and the corresponding Peaks II were run in

using a Speedvac centrifuge. This pooled fraction was lanes 3, 6, and 8 and in lanes 2, 4, and 7, respectively. A mixture ofdissolved in 50 ml of solvent A and loaded on a RP-C18 Peak I plus Peak II, obtained from the same preparation, and a pool

of Peaks II were run in lanes 5 and 9, respectively. Lane 10: aprotinincolumn. The column was equilibrated in solvent A and(6.5 kDa). The protein bands were visualized by silver staining.eluted with a gradient obtained by mixing solvent A

and solvent B at a flow rate of 1 ml/min. Even thoughurinary granulins do not show antikallikrein activity(see Discussion), owing to our primary interest in the at a concentration of solvent B ranging between 29 and

32% (Peak I) and a second one those eluted betweensearch for urinary kallikrein inhibitors (see introduc-tion), the eluted fractions were assayed for their anti- 32 and 48% of solvent B (Peak II). As can be seen in Fig.

2, where, under reducing conditions, an electrophoretickallikrein activity and, on this basis, two fractions werecollected (Fig. 1). A first one contained the tubes eluted run of three independently purified preparations (ob-

tained by different urine donors) is shown, both of thesefractions contain proteins displaying a molecularweight very similar to that of bovine aprotinin (MW6.513 kDa) in addition to proteins at higher molecularweight. In order further to purify the low-molecular-weight proteins, both of the pooled fractions recoveredfrom the first C18 column (Peak I and Peak II) wereseparately subjected to analytical reversed-phase andgel permeation chromatography on a Pharmacia SmartSystem. Figure 3 shows the elution profiles obtainedwhen Peak I (Fig. 3A) and Peak II (Fig. 3B) were sub-jected to reversed-phase chromatography on a micro-bore RP-18 column.

Both fractions contained several protein peaks whichwere identified through N-terminal microsequencing(see Table 2) performed as reported under Materialsand Methods. Figures 4A and 4B report the resultsFIG. 1. HPLC elution profile and antikallikrein activity (s) of theobtained when the same fractions (Peak I and Peak II)concentrated fraction recovered from the Sep-Pak C18 cartridge. Col-were subjected to gel permeation chromatography onumn: Supelcosil LC18-DB (25 1 4.6 cm). Flow rate: 1.0 ml/min. Detec-

tion wavelength: 254 nm. Solvent systems: (A) 0.1% trifluoroacetic a Superdex-30 column. The column was calibrated withacid (TFA) in water and (B) 40% of a solution of 4:1 (v/v) acetonitrile– standard proteins (MW ranging between 19.5 and 0.182-propanol in 0.1% TFA. The antikallikrein activity was measured kDa) and the fractions showing a retention time similaras reported under Materials and Methods. Peak I and Peak II refer

to that of bovine aprotinin (MW 6.513 kDa), also usedto the pooled fractions subsequently used for granulin isolation (seetext). as a standard, were sequenced. The results obtained

AID PEP 0726 / 6q10$$$621 06-04-97 13:04:41 pepa AP: PEP

SPARRO ET AL.172

From a quantitative standpoint, we are unable toestablish the exact granulin content of human urine,as a direct and specific method for granulin detectionis lacking. However, the granulin amount found at thesequence step (see Table 2) and the purification yield(see Table 1) allow us to estimate that isoforms A, B,and F, found in human urine, are, most likely, ex-pressed in a ratio of 1:1:3, respectively, for a totalamount of about 3 mg per liter of urine. These data,however, could be grossly affected by the parameter weused to check the purification steps (see Table 1), owingto the presence of low-molecular-weight polypeptidesin urine. This could also explain the apparently verylow yield of the purification procedure (see Table 1) andthe low levels of granulins found at the sequence stepFIG. 3. Smart System RP-chromatography of Peak I (A) and Peak

II (B). Column: RP-18 (5 mm material, 2.1 1 100 mm, 120 A pore (see Table 2).diameter). Solvent systems: 0.1% TFA in water (solvent A), 60%acetonitrile in 0.1%TFA (solvent B). Other conditions: room tempera-

DISCUSSIONture; flow rate, 0.2 ml/min; detection wavelength, 215 nm. Fractionsof 100 ml were collected and those showing a retention time included The data reported in this paper show, for the firstbetween 15 and 25 min were sequenced. The retention times for time, that human urine contains multiple forms of gra-granulins were: 16 min for granulin A , 19 min for granulin B, and

nulins. Their amino acid sequences are identical to22 min for granulin F.those of granulins A and B, previously isolated fromhuman leukocytes (1), and of granulin F, never before

indicated that granulin A was found when Peak I was isolated but whose primary structure was deduced onloaded on the column, while the isoforms B and F were the basis of the cDNA sequence (6,7). We were unablefound when Peak II was loaded. N-terminal sequence to demonstrate, in human urine, the presence of otheranalyses were performed also on the material showing granulin-like domains, previously isolated from humana molecular weight of about 6.5 kDa, separated under leukocytes (granulins C and D) (1) or predictable onreducing conditions on a 10–20% SDS polyacrylamide the basis of the cDNA sequence (6,7). This can be ac-gel (see Fig. 2), and subsequently blotted on a PVDF counted for either by the purification procedure utilizedmembrane (Fig. 5). The results obtained indicate that, during this study or by a particular posttranslationalindependently of the analytical method utilized for gra- processing of the primary gene expression product. Innulin purification (i.e., RP- and GP-chromatography or fact, human granulins are polypeptidic molecules (56–protein blotting), both Peak I and Peak II (see Fig. 57 amino acid residues) generated by a proteolytic post-1) of the three preparations, performed using urine of translational cleavage of a precursor protein (593different donors, contain granulin-like peptides. In par- amino acids) encoded by a single copy gene localizedticular, granulin A appears to be the only isoform pres- on chromosome 17 (9). The urinary molecules probablyent in Peak I, while isoforms B and F predominate inPeak II (40 and 50% of the total amount of granulins,respectively). As can be seen in Table 2, where the TABLE 1sequence results are summarized, all the granulins iso-

Purification of Human Urinary Granulinslated in human urine show N-terminal sequences iden-tical to those previously isolated in leukocytes (isoforms

TotalA and B) (1) or predicted (isoform F) on the basis of the Volume proteins PurificationcDNA sequence (6,7). The minute amount of urinary Fraction (ml) (mg) (%)granulins recovered (see below) did not allow us to iden-

Urine 310 35.03 5tify the C-terminal portion of the molecules. TraceTCA supernatant 386 13.43 35amounts of an isoform of granulin B (about 3% of theSep-Pak C18 4.3 0.038 nd

mature form) which showed a serine residue as N-ter- C18 HPLCa 22 0.020 5minal [residue -1 adopting the cDNA sequence (6,7)] Superdex-30 column 0.4 0.020 ú99and a truncated form of granulin A (about 10% of the

Note. The protein content of the fractions was determined ac-mature form) lacking the N-terminal residue Asp werecording to Bradford (23), while granulin purification was estimatedalso identified. In all Peaks II analyzed, we also foundfrom densitometric scans of Coomassie blue SDS–PAGE gels as the

a porcine pancreas kallikrein glycopeptide most likely ratio of the total 6.5-kDa density and the total protein density.stemming from the affinity column utilized for granulin a This fraction was obtained collecting several preparations (see

text). The data reported refer to 6 liters of starting urine.purification.

AID PEP 0726 / 6q10$$$621 06-04-97 13:04:41 pepa AP: PEP

HUMAN URINE GRANULINS 173

TABLE 2

N-Terminal Sequence of Urinary Granulins

Amounta

Source N-terminal sequence (pmol) % Protein

Urine (Peak I) DVKCDMEVSCPDGYTCCRLQSGAWGCCPFTQAV... 152 89 grnAVKCDMEVSCPDGYTCCRL.............. 19 11 grnA

Urine (Peak II) VMCPDARSRC............................ 231 20 grnBSVMCPDARSRC........................... 35 3 grnBAIQCPDSQFECPDFST.................. 461 40 grnFDVKCDMEVSCP................... 92 8 grnA

Leukocytes DVKCDMEVSCPDGYTCCRLQSGAWGCCPFTQAVCCEDHIHCCPAGFTCDTQKGTCE grnAVMCPDARSRCPDGSTCCELPSGKYGCCP............................ grnB

cDNA AIQCPDSQFECPDFSTCCVMVDGSWGCCPMPQASCCEDRVHCCPHGAFCDLVHTRCI grnF

Note. Peak I and Peak II refer to the protein fractions recovered from the first RP-HPLC column (see Fig. 1). The analyses were performedon individual fractions recovered from the Smart System reversed-phase and gel permeation chromatography columns and from proteinblotting (see text). The N-terminal sequences of granulins A and B, purified from human leukocytes, and of granulin F, deduced from thecDNA, are also reported for comparison.

a The granulin amount at sequence step refers to 1 liter of starting urine.

arose from the kidney, which appears to be the organ thought to be responsible for the processing of severalpolypeptidic growth factor precursors (12–16). How-richest in granulin precursor mRNA (6,7) and to pos-

sess, at least in rat, the machinery for both precursor ever, the precise nature of interaction between granul-ins (or their precursor) and kallikrein is currently verytranslation and processing (2,6). Most likely, they are

excreted in the urinary flow as low-molecular-weight obscure, since granulins do not show antikallikrein ac-tivity (W. Schroder and E. Fioretti, unpublished obser-polypeptides, as is suggested by both their molecular

weight (about 6.5 kDa) and their N-terminal sequence, vations) even though they are recovered from the elu-ate of an affinity column performed with this immobi-reported here, after a posttranslational proteolytic pro-

cessing of the precursor molecule occurring at the kid- lized enzyme (see above). The presence in human urineof granulin F, apparently not expressed in leukocytesney level. The molecular events underlying precursor

processing are unknown, although various reported ev- (1), seems to support the hypothesis that mature formsof granulins are generated by organ-specific precursoridence (10–19) (see also introduction) seems to indicate

involvement of the protease kallikrein, an enzyme

FIG. 4. Smart System gel permeation chromatography of Peak I(B) and Peak II (A). Column: Superdex-30 (7 mm material, 3.2 1 300 FIG. 5. Protein blotting of Peak II (lane 4) separated under reduc-

ing conditions on a 10–20% SDS–polyacrylamide gel (see Fig. 2).mm). Mobile phase: 50 mM phosphate buffer, pH 7.1, containing 150mM NaCl. Detection wavelength, 215 nm. The column was calibrated Lanes 1 and 3: prestained protein markers (myosin 200 kDa, phos-

phorilase 97 kDa, bovine serum albumin 68 kDa, ovoalbumin 43with the following markers: interferon (19.5 KDa), cytochrom c (12.3kDa), aprotinin (6.5 kDa), glucagon (3.5 kDa), and tyrosine (0.18 kDa, chymotrypsinogen 26 kDa, and lactoglobulin 18 kDa). Lane 2:

aprotinin. The blot was stained using Coomassie brilliant blue R-kDa). Fractions of 200 ml were collected and those showing a reten-tion time similar to that of aprotinin (about 30 min) were sequenced. 250. The sequence of the purified granulin band showed the presence

of the isoforms B and F.For other details see text.

AID PEP 0726 / 6q10$$$621 06-04-97 13:04:41 pepa AP: PEP

SPARRO ET AL.174

marrow reveals tandem cysteine-rich granulin domains. Proc.processing (7), on the basis of physiological require-Natl. Acad. Sci. USA 89, 1715–1719.ments. Furthermore, the relative abundance of this iso-

8. Plowman, G. D., Green, J. M., Neubauer, M. G., Buckley, S. D.,form in urine suggests also that it may play in vivo anMcDonald, V. L., Todaro, G. J., and Shoyab, M. (1992) The epi-

important and specific role in kidney, or in urinary thelin precursor encodes two proteins with opposing activitiesspace, function. In conclusion, the identification and on epithelial cell growth. J. Biol. Chem. 267, 13073–13078.purification, using a relatively simple protocol, of mul- 9. Bhandari, V., and Bateman, A. (1992) Structure and chromo-

somal location of the human granulin gene. Biochem. Biophys.tiple forms of granulins in an easily available biologicalRes. Commun. 188, 57–63.fluid such as urine will certainly help promote studies

10. Barajas, L., Powers, K., Carretero, O., Scicli, A. G., and Inagami,on these interesting molecules, even though their com-T. (1986) Immunocytochemical localization of renin and kalli-

plete characterization appears currently hindered by krein in the rat renal cortex. Kidney Int. 29, 965–970.the minute amount that can be purified from human 11. Xiong, W., Chao, L., and Chao, J. (1989) Renal kallikrein mRNAsources. Furthermore, the results presented here can localization by in situ hybridization. Kidney Int. 35, 1324–1329.

12. Angeletti, R. H., and Bradshaw, R. A. (1971) The amino acidcontribute to a better understanding of the role thatsequence of 2.5S mouse submaxillary gland nerve growth factors.granulins exert in vivo, with particular reference toProc. Natl. Acad. Sci. USA 68, 2417–2420.their growth modulatory function on human epithelial

13. Taylor, J. M., Cohen, S., and Mitchell, W. M. (1970) Epidermalcells and to the regulative processes that can controlgrowth factor: High and low molecular weight forms. Proc. Natl.

their biological activity. Acad. Sci. USA 67, 164–171.14. Frey, P., Forand, R., Maciag, T., and Shooter, E. M. (1979) The

biosynthetic precursor of epidermal growth factor and the mech-ACKNOWLEDGMENTSanism of its processing. Proc. Natl. Acad. Sci. USA 76, 6294–6298.The authors thank Mr. S. Polzoni for his invaluable technical assis-

15. Blaber, M., Isakson, P. J., Marster, J. C., Jr., Burnier, J. P., andtance and Bayer AG (Wuppertal, Germany) for the logistical andBradshaw, R. A. (1989) Substrate specificities of growth factortechnical support. This work was supported, in part, by grants fromassociated kallikreins of the mouse submandibular gland. Bio-the University of Camerino (E.F.).chemistry 28, 7813–7819.

16. Edwards, R. H., Selby, M. J., Garcia, P. D., and Rutter, W. J.REFERENCES (1988) Processing of the native nerve growth factor precursor to

form biologically active nerve growth factor. J. Biol. Chem. 263,1. Bateman, A., Belcourt, D., Bennet, H. P. J., Lazure, C., and Solo-6810–6815.mon, S. (1990) Granulins, a novel class of peptide from leuko-

17. Laskowski, J. M., and Kato, I. (1980) Protein inhibitors of pro-cytes. Biochem. Biophys. Res. Commun. 173, 1161–1168.teinases. Annu. Rev. Biochem. 49, 593–626.

2. Shoyab, M., McDonald, V. L., Byles, C., Todaro, G., and Plow-18. Fritz, H., and Wunderer, G. (1983) Biochemistry and applica-man, G. D. (1990) Epithelins 1 and 2: Isolation and characteriza-

tions of aprotinin, the kallikrein inhibitor from bovine organs.tion of two cysteine-rich growth-modulating proteins. Proc. Natl.Drug Res. 33(Suppl.), 479–494.Acad. Sci. USA 87, 7912–7916.

19. Nori, S. L., De Renzis, G., Fumagalli, L., Fioretti, E., Barra, D.,3. Belcourt, T., Lazure, C., and Bennett, H. P. J. (1993) Isolation and Ascoli, F. (1992) Bovine pancreatic trypsin inhibitor and

and primary structure of the three major forms of granulin-like homologous polypeptide inhibitors in nephron cells. Peptides 18,peptides from hematopoietic tissues of a teleost fish (Cyprinus 365–371.carpio). J. Biol. Chem. 268, 9230–9237.

20. Fioretti, E., Eleuteri, A. M., Angeletti, M., and Ascoli, F. (1993)4. Couto, M. A., Harwing, S. S., Cullor, J. S., Hughes, J. P., and High-performance liquid chromatographic separation of aproti-

Leher, R. I. (1992) Identification of eNAP-1, an antimicrobial nin-like inhibitors and their determination in very smallpeptide from equine neutrophils. Infect. Immun. 60, 3065–3071. amounts. J. Chromatogr.–Biomed. Appl. 617, 308–312.

5. Nakakura, N., Hietter, H.,VanDorsselaer,A., and Luu, B. (1992) 21. Laemmli, U. K. (1970) Cleavage of structural proteins duringIsolation and characterization of three peptides from the insect the assembly of the head of the bacteriophage T4. Nature 227,Locusta migratoria. Eur. J. Biochem. 204, 147–153. 680–685.

6. Bhandari, V., Giaid, A., and Bateman, A. (1993) The complemen- 22. Gultekin, H., and Heermann, K. (1988) The use of polyvinyli-tary deoxyribonucleic acid sequence, tissue distribution and cel- denedifluoride membranes as a general blotting matrix. Anal.lular localization of the rat granulin precursor. Endocrinology Biochem. 172, 320–329.133, 2682–2689. 23. Bradford, M. M. (1976) A rapid and sensitive method for the

quantitation of microgram quantities of protein utilizing the7. Bhandari, V., Palfree, R. G. E., and Bateman, A. (1992) Isolationand sequence of the granulin precursor cDNA from human bone principle of protein-dye ligand. Anal. Biochem. 72, 248–254.

AID PEP 0726 / 6q10$$$621 06-04-97 13:04:41 pepa AP: PEP