the of 268, no. 14, 15, pp. 10425-10432,1993 q 1993 for ... · the journal of biological chemistry...

8
THE JOURNAL OF BIOLOGICAL CHEMISTRY Q 1993 by The American Society for Biochemistry and Molecular Biology, Inc. Vol. 268, No. 14, Iseue of May 15, pp. 10425-10432,1993 Printed in U. S. A. Isolation and Primary Structure of NGAL, a Novel Protein Associated with Human Neutrophil Gelatinase* (Received for publication, December 1, 1992, and in revised form, January 11, 1993) Lars KjeldsenS, Anders H. Johnsen#, Henrik Sengelbv, and Niels Borregaardll From the Granulocyte Research Laboratory, the Departments of Hematology §Clinical Biochemistry, University Hospital, Rigshospitalet, Copenhugen DK-2100,Denmark A 25-kDa protein was found to be associated with purified human neutrophil gelatinase. Polyclonal an- tibodies raised against gelatinase not only recognized gelatinase but also this 25-kDa protein. Specific anti- bodies against the 25-kDa protein were obtained by affinity purification of the gelatinase antibodies. Immunoblotting and immunoprecipitation studies demonstrated the 135-kDa form of gelatinase to be a complex of 92-kDa gelatinase and the 25-kDa protein, and the 220-kDa form was demonstrated to be a homo- dimer of the 92-kDa protein, thus explaining the 220-, 135-, and 92-kDa forms characteristic of neu- trophil gelatinase. The 25-kDa protein was purified to apparent ho- mogeneity from exocytosed material from phorbol my- ristate acetate-stimulated neutrophils. The primary structure of the 25-kDa protein was determined as a 178-residueprotein. It was susceptible to treatment with N-glycanase, and one N-glycosyla- tion site was identified. The sequence did not match any known human protein, but showed a high degree of similarity with the deduced sequences of rat a2- microglobulin-related protein and the mouse protein 24p3. It is thus a new member of the lipocalinfamily. The function of the 25-kDa protein, named neutrophil gelatinase-associated lipocalin (NGAL), remains to be determined. Their activities are balanced by the protein inhibitors, tissue inhibitor of metalloproteinases (TIMP-1 and TIMP-2)’ (1,3- 7). Copurification of TIMP-1 and 92-kDa gelatinase has been observed in SV-40-transformed lung fibroblasts (a), human monocytes (9), and the U937 myelomonocytic (8) and HT1080 fibrosarcoma cell lines (lo), indicating formation of a complex between TIMP-1 and latent 92-kDa gelatinase. Humanneutrophilscontainbothinterstitial collagenase and 92-kDa type IV collagenase (gelatinase) (11). In contrast to gelatinase from other sources, human neutrophilgelatinase has been purified in a TIMP-free form (12), and exists in three molecular mass forms of 220, 135, and 92 kDa, respec- tively, when visualized by SDS-PAGE under nonreducing conditions, but only as one 92-kDa band under reducing conditions (12). An explanation of this phenomenon has been lacking so far. We have confirmed the molecular mass forms of unreduced neutrophil gelatinase, but, in contrast to others, we found that a faint 25-kDa protein is observed in addition to the 92-kDa band under reducing conditions (13). Further- more, antibodies raised against purified gelatinase recognized not only the 92-kDa gelatinase band but also this 25-kDa band. By affinity purification of the gelatinase antibodies, we obtained specific antibodies against both gelatinase and this 25-kDa protein. The aim of the present investigation was to characterize the 25-kDa protein and to determine its associ- ation with neutrophil gelatinase. EXPERIMENTAL PROCEDURES A number of normal human cells and tumor cells are known to synthesize and secrete proteins that are members of the metalloproteinase family (for review see Refs. 1 and 2). These enzymes, including interstitial collagenase, 72- and 92-kDa type IV collagenase (gelatinase), stromelysin, and PUMP-1, are thought to be important participants in the degradation and remodeling of the extracellular matrix (1,2). The enzymes are all secreted as zymogens, and their collagen-degrading activities are initiated by mechanisms not yet fully known. *This work was supported by grants from The Danish Cancer Society, The Danish Medical Research Council, The Lundbeck Fund, Emil C. Hertz’s Fund, The Novo Fund, Amalie J~irgensen’s Fund, BrNhner-Mortensen’s Fund, and Anders Hasselbach‘s Fund. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The nucleotide sequence(s) reported in this paper has been submitted P80188. to the GenBankTM/EMBL Data Bank with accession number(s) $ To whom correspondence should be addressed Granulocyte Re- search Laboratory, Dept. of Hematology L, afsnit 4041, University Hospital, Rigshospitalet, Blegdamsvej 9, DK-2100 Copenhagen, Den- mark. Tel.: 45-3545-4241; Fax: 45-3135-1727. ll Recipient of a Neye-Research Professorship. Preparation of Neutrophikr-Human neutrophils were isolated from buffy coats supplied by the blood bank. Erythrocytes were sedimented by 1:l addition of 2% Dextran T-500 (Pharmacia LKB Biotechnology Inc.) in 0.9% NaCl. The leukocyte-rich supernatant was aspirated, and the cells pelleted at 200 X g for 10 min. Cells were resuspended in saline and neutrophils separated by centrifugation through Lymphoprep (14) (Nygaard, Oslo, Norway) at 400 X g for 30 min. Remaining erythrocytes were removed by hypotonic shock in ice-cold H20 for 30 s. Tonicity was restored with an equal volume of 1.8% NaCl. The neutrophils were then washed once in 0.9% NaCl and subsequently resuspended in the desired buffer. All steps except the dextran sedimentation (room temperature) were performed at 4 “C. Purification of Neutrophil Gelcrtinase-Neutrophil gelatinase was purified essentially as described by Hibbs et al. (12). In short, neutro- phils were prepared from buffy coats and resuspended in Krebs Ringer phosphate: 130 mM NaC1, 5 mM KC1, 1.27 mM MgS04, 0.95 mM CaC12, 5 mM glucose, 10 mM NaH2P04/Na2HP04, pH 7.4. The cells were then stimulated with 2 pg/ml of phorbol myristateacetate (PMA) (Sigma) for 15 min and pelleted by centrifugation. The The abbreviations used are: TIMP-1 and TIMP-2, tissue inhibi- tors of metalloproteinases; PMA, phorbol 12-myristate 13-acetate; NGAL, neutrophil gelatinase-associated lipocalin; PBS, phosphate- buffered saline; PEG, polyethylene glycol; Pipes, 1,4-piperazinedi- ethanesulfonic acid; HPLC, high-performance liquid chromatogra- phy; PAGE, polyacrylamide gel electrophoresis; CAPS, 3-(cyclohex- y1amino)propanesulfonicacid. 10425

Upload: others

Post on 14-Aug-2020

0 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: THE OF 268, No. 14, 15, pp. 10425-10432,1993 Q 1993 for ... · THE JOURNAL OF BIOLOGICAL CHEMISTRY Q 1993 by The American Society for Biochemistry and Molecular Biology, Inc. Vol

THE JOURNAL OF BIOLOGICAL CHEMISTRY Q 1993 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 268, No. 14, Iseue of May 15, pp. 10425-10432,1993 Printed in U. S. A.

Isolation and Primary Structure of NGAL, a Novel Protein Associated with Human Neutrophil Gelatinase*

(Received for publication, December 1, 1992, and in revised form, January 11, 1993)

Lars KjeldsenS, Anders H. Johnsen#, Henrik Sengelbv, and Niels Borregaardll From the Granulocyte Research Laboratory, the Departments of Hematology §Clinical Biochemistry, University Hospital, Rigshospitalet, Copenhugen DK-2100, Denmark

A 25-kDa protein was found to be associated with purified human neutrophil gelatinase. Polyclonal an- tibodies raised against gelatinase not only recognized gelatinase but also this 25-kDa protein. Specific anti- bodies against the 25-kDa protein were obtained by affinity purification of the gelatinase antibodies.

Immunoblotting and immunoprecipitation studies demonstrated the 135-kDa form of gelatinase to be a complex of 92-kDa gelatinase and the 25-kDa protein, and the 220-kDa form was demonstrated to be a homo- dimer of the 92-kDa protein, thus explaining the 220-, 135-, and 92-kDa forms characteristic of neu- trophil gelatinase.

The 25-kDa protein was purified to apparent ho- mogeneity from exocytosed material from phorbol my- ristate acetate-stimulated neutrophils.

The primary structure of the 25-kDa protein was determined as a 178-residue protein. It was susceptible to treatment with N-glycanase, and one N-glycosyla- tion site was identified. The sequence did not match any known human protein, but showed a high degree of similarity with the deduced sequences of rat a2- microglobulin-related protein and the mouse protein 24p3. It is thus a new member of the lipocalin family. The function of the 25-kDa protein, named neutrophil gelatinase-associated lipocalin (NGAL), remains to be determined.

Their activities are balanced by the protein inhibitors, tissue inhibitor of metalloproteinases (TIMP-1 and TIMP-2)’ (1,3- 7). Copurification of TIMP-1 and 92-kDa gelatinase has been observed in SV-40-transformed lung fibroblasts (a), human monocytes (9), and the U937 myelomonocytic (8) and HT1080 fibrosarcoma cell lines (lo), indicating formation of a complex between TIMP-1 and latent 92-kDa gelatinase.

Human neutrophils contain both interstitial collagenase and 92-kDa type IV collagenase (gelatinase) (11). In contrast to gelatinase from other sources, human neutrophil gelatinase has been purified in a TIMP-free form (12), and exists in three molecular mass forms of 220, 135, and 92 kDa, respec- tively, when visualized by SDS-PAGE under nonreducing conditions, but only as one 92-kDa band under reducing conditions (12). An explanation of this phenomenon has been lacking so far. We have confirmed the molecular mass forms of unreduced neutrophil gelatinase, but, in contrast to others, we found that a faint 25-kDa protein is observed in addition to the 92-kDa band under reducing conditions (13). Further- more, antibodies raised against purified gelatinase recognized not only the 92-kDa gelatinase band but also this 25-kDa band. By affinity purification of the gelatinase antibodies, we obtained specific antibodies against both gelatinase and this 25-kDa protein. The aim of the present investigation was to characterize the 25-kDa protein and to determine its associ- ation with neutrophil gelatinase.

EXPERIMENTAL PROCEDURES

A number of normal human cells and tumor cells are known to synthesize and secrete proteins that are members of the metalloproteinase family (for review see Refs. 1 and 2). These enzymes, including interstitial collagenase, 72- and 92-kDa type IV collagenase (gelatinase), stromelysin, and PUMP-1, are thought to be important participants in the degradation and remodeling of the extracellular matrix (1,2). The enzymes are all secreted as zymogens, and their collagen-degrading activities are initiated by mechanisms not yet fully known.

*This work was supported by grants from The Danish Cancer Society, The Danish Medical Research Council, The Lundbeck Fund, Emil C. Hertz’s Fund, The Novo Fund, Amalie J~irgensen’s Fund, BrNhner-Mortensen’s Fund, and Anders Hasselbach‘s Fund. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted

P80188. to the GenBankTM/EMBL Data Bank with accession number(s)

$ To whom correspondence should be addressed Granulocyte Re- search Laboratory, Dept. of Hematology L, afsnit 4041, University Hospital, Rigshospitalet, Blegdamsvej 9, DK-2100 Copenhagen, Den- mark. Tel.: 45-3545-4241; Fax: 45-3135-1727.

ll Recipient of a Neye-Research Professorship.

Preparation of Neutrophikr-Human neutrophils were isolated from buffy coats supplied by the blood bank. Erythrocytes were sedimented by 1:l addition of 2% Dextran T-500 (Pharmacia LKB Biotechnology Inc.) in 0.9% NaCl. The leukocyte-rich supernatant was aspirated, and the cells pelleted at 200 X g for 10 min. Cells were resuspended in saline and neutrophils separated by centrifugation through Lymphoprep (14) (Nygaard, Oslo, Norway) at 400 X g for 30 min. Remaining erythrocytes were removed by hypotonic shock in ice-cold H20 for 30 s. Tonicity was restored with an equal volume of 1.8% NaCl. The neutrophils were then washed once in 0.9% NaCl and subsequently resuspended in the desired buffer. All steps except the dextran sedimentation (room temperature) were performed at 4 “C.

Purification of Neutrophil Gelcrtinase-Neutrophil gelatinase was purified essentially as described by Hibbs et al. (12). In short, neutro- phils were prepared from buffy coats and resuspended in Krebs Ringer phosphate: 130 mM NaC1, 5 mM KC1, 1.27 mM MgS04, 0.95 mM CaC12, 5 mM glucose, 10 mM NaH2P04/Na2HP04, pH 7.4. The cells were then stimulated with 2 pg/ml of phorbol myristate acetate (PMA) (Sigma) for 15 min and pelleted by centrifugation. The

The abbreviations used are: TIMP-1 and TIMP-2, tissue inhibi- tors of metalloproteinases; PMA, phorbol 12-myristate 13-acetate; NGAL, neutrophil gelatinase-associated lipocalin; PBS, phosphate- buffered saline; PEG, polyethylene glycol; Pipes, 1,4-piperazinedi- ethanesulfonic acid; HPLC, high-performance liquid chromatogra- phy; PAGE, polyacrylamide gel electrophoresis; CAPS, 3-(cyclohex- y1amino)propanesulfonic acid.

10425

Page 2: THE OF 268, No. 14, 15, pp. 10425-10432,1993 Q 1993 for ... · THE JOURNAL OF BIOLOGICAL CHEMISTRY Q 1993 by The American Society for Biochemistry and Molecular Biology, Inc. Vol

10426

A r- B Mr (lo3) Mr (lo3)

200 > 200 >

97.4 > (I, 0 97.4 ' 69 > 69 >

46 3

30 '

21.5 >

14.3 >

46 >

NGAL Protein

m C Mr (lo3)

200 >

97.4 >

69 '

46 >

30 > 30 >

21.5 >

14.3 >

21.5 >

14.3 '

200 >

97.4 >

69 >

46 >

30 >

c

21.5 >

14.3 >

1 2 1 2 1 2 FIG. 1. Purification of gelatin- and generation of specific polyclonal antibodies against gelatin- and the 26-kDa protein

(p26). A, 6 pg of the purified gelatinase was subjected to electrophoresis under nonreducing ( l a n e 1 ) and reducing conditions (hne 2). E , Western blotting of a postnuclear supernatant from human neutrophils. Five pg of postnuclear supernatant was subjected to SDS-PAGE under reducing conditions and transferred to a nitrocellulose membrane as described under "Experimental Procedures." Primary antibodies were obtained after immunization of rabbits with purified gelatinase. The antibody was used in a dilution of 1:1000. C. separation of ge1atina.w and p25 by gel filtration. Exocytosed material from cells, stimulated with PMA, was concentrated and subjected to gel filtration on Sephadex G-200. A fraction (6 pg) corresponding to the gelatinase peak ( l a n e 1 ) and another fraction (6 pg) eluting at a lower molecular weight ( l a n e 2), were applied to electrophoresis under reducing conditions. Western blotting was performed as described, using the antibody reacting to both gelatinase and p25 in a dilution of 1:1000. D, Western blotting of a postnuclear supernatant with antibodies against p25 and gelatinase. In each lane a total of 5 pg of protein from a postnuclear supernatant was electrophoresed under reducing conditions and transferred to nitrocellulose. Lune 1, primary antibody was antibody eluted from the affinity column to which "gelatinase-free" p2S had been coupled. This antibody was diluted 2000-fold. Lane 2, primary antibody was anti-gelatinase-antibody, depleted of anti-p25 antibody (diluted 1:2000).

resulting supernatant was collected and subjected to ion exchange chromatography on a DE-52 column (Whatman) followed by affinity chromatography on a CNBr-Sepharose 4B column (Pharmacia) to which heat-denatured type I collagen (gelatine) had been coupled.

Production of Polyclonal Antibodies toward Gelatinuse and the 25- M a Protein-Immunization was performed as described (13). The antiserum was affinity-purified (13) on CNBr-activated Sepharose- 4B to which 8 mg of purified gelatinase had been coupled.

The specificity of the antibodies was tested by immunoblotting of a postnuclear supernatant from disrupted human neutrophils. This showed that the antibody reacted with both gelatinase and with a 25- kDa protein. To remove antibody with specificity for the 25-kDa protein, the following procedure was performed: neutrophils (10 X 10') a t 1 X lo8/ml in Krebs Ringer phosphate were stimulated with PMA (2 pg/ml) to exocytose granule proteins. After addition of phenylmethanesulfonyl fluoride (1 mM) (Sigma), the exocytosed ma- terial was collected and concentrated approximately 15-fold using an Amicon ultrafiltration cell. In order to separate the 25-kDa protein from gelatinase, the concentrated material was subsequently sub- jected to gel filtration on a Sephadex G-200 (Pharmacia) column (2 X 100cm) equilibrated in phosphate-buffered saline (PBS) (137 mM NaCI, 2.7 mM KCI, 9.5 mM sodium phosphate, pH 7.4) containing 0.1% sodium azide. Fractions containing the 25-kDa protein, but free of gelatinase, as evidenced by Western blotting, were pooled ( total volume: 60 ml) and concentrated 8-fold. The resulting material was coupled to CNBr-activated Sepharose-4B. The anti-gelatinase anti- bodies were depleted of antibodies against the 25-kDa protein by passage through this column. Antibodies against the 25-kDa protein were obtained by subsequent elution with 3 M KSCN in PBS.

Purification of 25-kDa Protein from Human Neutrophils-Neutro- phils were isolated from 20-25 buffy coats as described. Cells were resuspended in Krebs Ringer phosphate a t 5 X 1@/ml and preincu- bated for 5 min at 37 "C followed by stimulation with 5 pg/ml of PMA for 15 min. Cells were pelleted, the supernatant aspirated, and phenylmethanesulfonyl fluoride (1 mM), aprotinin (Bayer) (100 kal- likrein inactivator unita/ml), and EGTA (1 mM) added. After cen- trifugation at 30,000 X g for 10 min at 4 "C, polyethylenglycol 6000

(PEG) (Merck) in 20 mM Pipes, pH 7.2, waa added to a final concentration of 18% (w/v). The sample was rotated end over end for 1 h at 4 'C. The reeultingprecipitate was pelleted by centrifugation a t 1500 X g for 10 min at 4 'C. and the supernatant, containing the majority of the 25-kDa protein, was aspirated. K2HP04 was added to a final concentration of 0.5 M, followed hy incubation for 30 min at 4 "C. This induced a phase separation with PEG in the upper face and more than 90% of the protein in the lower face (15). The lower phase was harvested and chromatographed on Sephadex G-25 (Phar- macia) to obtain a change of buffer to 50 mM sodium acetate, pH 6.0. This material was subjected to cation exchange on MonoS using fast- protein liquid chromatography (Pharmacia). Round material was eluted with a gradient of NaCl from 0 to 1 M in 50 mM Rodium acetate, pH 6.0. Two peaks containing most of the 25-kDa protein were pooled separately, and concentrated 8-fold in a Centricon 10 microconcentrator (Amicon) to a final volume of 250 pl each. These samples were subjected to gel filtration on Superose 12 using fast- protein liquid chromatography. The column was equilibrated in 25 mM Tris-HCI, pH 8.0, 100 mM NaCI. Fractions were visualized by SDS-PAGE to a s w their content of 25-kDa protein.

Carboxymethylation-The purified protein was reduced and car- boxymethylated with iodoacetamide as described (16). Thereafter, the carboxymethylated protein was desalted and repurified by re- versed-phase HPLC on a 2.1 x 150-mm 5-pm C4 column (Vydac), which was eluted by a gradient from 0.1% trifluoroacetic acid in water to 0.1% trifluoroacetic acid in acetonitrile. The instrument used waa a 1090 HPLC from Hewlett-Packard.

Amino Acid Analysis-The carboxymethylated protein waa hydro- lyzed for 20 h in 6 N HCI gas phase at 110 'C under argon in pyrolyzed tapered microvials (100 pl, Hewlett-Packard). The hydrolysate waa dried and redissolved in 20 pl of 0.4 M sodium borate, pH 10.4. Amino acid analysis was performed on 6 pl using a Hewlett-Packard Amin- oquant analyzer (17) (precolumn derivatization with o-phthaldialde- hyde followed by 9-fluorenylmethylchloroformate, both reagentn RUP- plied by Hewlett-Packard).

Cleauage with CNBr-2 nmol of the carboxymethylated protein

Page 3: THE OF 268, No. 14, 15, pp. 10425-10432,1993 Q 1993 for ... · THE JOURNAL OF BIOLOGICAL CHEMISTRY Q 1993 by The American Society for Biochemistry and Molecular Biology, Inc. Vol

NGA L Protein 10427

A Mr (lo3) fl

200 ' Complex -b

97.4 '

69 '

46 '

30 > '

21.5 >

1 2 3 4 5

B Mr (lo3) 200 ' F

97.4 ' 4

69 '

46 '

30 ' p26 -b

21.5 >

1 2 3 4 5 FIG. 2. Immunoprecipitation of purified gelatinase with an-

tibodies against p25. Twenty pg of purified gelatinase in PBS (total volume 170 pl) was incubated for 1 h with either buffer, anti- p25 antibodies (30 pl) , or preimmune rabbit IgG in equivalent amounts. 100 pl of protein A-Sepharose 4B (200 mg/ml in PBS) was added and rotated end over end for 1 h. The Sepharose particles were pelleted by centrifugation. The resulting supernatants were aspirated. After washing twice in PBS, the Sepharose pellets were resuspended in PBS to the initial volume. Supernatants and pellets were applied to SDS-PAGE under nonreducing ( A ) or reducing conditions ( B ) . Lane I , gelatinase (control); lane 2, gelatinase, supernatant after immunoprecipitation with anti-p25 antibodies; lone 3, gelatinase, supernatant after immunoprecipitation with preimmune rabbit IgC; lane 4, Sepharose pellet from gelatinase after immunoprecipitation with anti-p25-antibodies; lane 5, Sepharose pellet from gelatinase after immunoprecipitation with preimmune rabbit I&.

was cleaved with 100-fold excess of CNBr in 70% HCOOH BB de- scribed (16).

Proteolytic Cleavages-1-3 nmol of the carboxymethylated protein (see above) was digested with three enzymes. All were sequencing grade from Boehringer Mannheim. Cleavage with 1 pg of trypsin was performed in 400 pl of 100 mM Tris (pH 8.5) for 18 h at 37 'C. Cleavage with 1 pg of endoproteinase Glu-C was performed in 400 pl of 25 mM ammonium acetate, pH 4.0, for 18 h at room temperature. Cleavage with 1 pg of endoproteinase Asp-N was performed in 400 pl of 50 mM sodium phosphate, pH 8.0, for 18 h at 37 'C. Finally, a tryptic fragment was digested with 100 pg of pyroClu aminopeptidase (Boehringer Mannheim) for 30 min at 37 "C in 200 pl buffer (100 mM sodium phosphate, 10 mM EDTA, 5 mM dithiothreitol, 5% glycerol, pH 7.8). The resulting peptides were isolated on a Hewlett-Packard 1090 HPLC on 2.1 X 150-mm 5-pm Ce and Cl8 columns from Vydac, by eluting with gradients from 0.1% trifluoroacetic acid in water to

97.4 * 97.4 a

09

09 *

40 *

30 I

21.6

14.3

40 *

I 30 9

21.6

14.3 *

1 2

FIG. 3. Aeeociation of p26 with gelatin- and other neutro- phil proteins. A, Western blotting of purified gelatinase with anti- gelatinase- or anti-p25-antibodies. Blotting was performed as de- scribed under "Experimental Procedures." l ~ n e l , 3 pg of gelatinase probed with anti-gelatinase antibodies depleted of reactivity toward p25 (1/1000); lane 2,9 pg of gelatinase probed with anti-p25 antihodies (l/8000). E , Western blotting of postnuclear supernatant with anti- bodies against p25 under nonreducing conditions. A total of 5 pg of protein from a postnuclear supernatant was electrophoresed and blotted onto nitrocellulose. Primary antibody was anti-p25 antibody diluted 2000-fold.

0.1% trifluoroacetic acid in acetonitrile. Sequence Analysk-The amino acid sequences of the purified

peptides were determined using an automatic protein sequenator (475A, Applied Biosystems) equipped with an on-line HPLC system for detection of the amino acid phenylthiohydantoins. The thiohy- dantoins were separated on a Cls-DB column (5-7943; Supelco). All chemicals and solvents were sequence or HPLC grade and delivered by Applied Biosystems.

Mass Spectrometry-An aliquot of each of the isolated peptides was analyzed in a Bio-Ion 20 plasma desorption time-of-flight mass spectrometer (Applied Biosystems). The samples were either applied directly as 2 X 5-pl aliquots to aluminized Mylar foil (coated with nitrocellulose) and evaporated, or dried and redissolved in 10 pl of 0.05% trifluoroacetic acid in 50% methanol before application. The spectra were recorded for 1-6 X lb primary ions. The method has an accuracy of 0.1%.

Endoglycosidase Treatment-Samples of the purified 25-kDa pro- tein (8.25 pg) were dried in a Speed-Vac centrifuge (Savant) and redissolved in 0.5% SDS and 1% &mercaptoethanol, boiled for 6 min. and diluted 5-fold into either 0.2 M NaHP04/NaH2P04. pH 8.6, 10 mM 1,lO-phenantroline, 1.25% Nonidet P-40 (Sigma), or 20 mM Tris- maleat, pH 6.0, 10 mM D-galactonolactone, 1.25% Nonidet P-40 and incubated with either 10 units/ml of N-glycanase (Genzyme) or 32 milliunits/ml 0-glycanaee (Genzyme), respectively, for 16 h at 37 'C (8). Samples were diluted in SDS-PAGE sample buffer and subjected to SDS-PAGE. A proteolytic fragment was also digested with N- glycanase, as described above but without SDS and mercaptoethanol, and purified by HPLC.

SDS-PAGE-SDS-PAGE was performed eseentially as described by Laemmli (18), employing 5-20% gradient gels with 3% stacking gels.

Postnuclear Supernatant-A postnuclear supernatant of neutro- phils was obtained using nitrogen cavitation followed by centrifuga- tion to remove nuclei and unbroken cells as described (19).

Zrnmumblotting-Protein was transferred from SDS-PAGE slabs to 0.2-pm nitrocellulose filters (Bio-Rad) essentially as described by Towbin (20) in a Bio-Rad trans-blot vertical system at 60 V, 210 mA for 4 h. Transfer buffer was either 192 mM glycine, 25 mM Tria, pH 8.3, 20% (v/v) methanol, or 10 mM CAPS, pH 11.0, 10% methanol. Additional binding sites were blocked by incubating the nitrocellulose filters for 1 h in either 2% Tween-20 in PRS, or 3% bovine serum albumin and 10% goat serum (Dakopatts) in PRS. After three washes in PBS, 0.05% Tween, the blots were incubated with primary antibody overnight. Primary antibodies were labeled with peroxidase-conju- gated swine anti-rabbit antibody (Dakopatts (P217)) diluted 1:lOoO

Page 4: THE OF 268, No. 14, 15, pp. 10425-10432,1993 Q 1993 for ... · THE JOURNAL OF BIOLOGICAL CHEMISTRY Q 1993 by The American Society for Biochemistry and Molecular Biology, Inc. Vol

10428 NGAL Protein

t r n O.’

20 40 60 80

Volume (ml)

r I I

FIG. 4. Fast-protein liquid chromatography purification of p26 on MonoS and Superose 12. The supernatant from neutro- phils stimulated with PMA was precipitated with PEG in a final concentration of 18% (w/v). Nonprecipitated protein was applied on a MonoS cation exchange column, equilibrated in 50 mM sodium acetate, pH 6.0, and eluted with a gradient of NaCl (broken straight line). Elution was monitored by absorbance at 280 nM. The peaks, P, and PPI contain the majority of the eluted p25. These peaks were concentrated separately and each subjected to gel filtration on Su- perose 12. The elution profile of PI is shown in the middle panel, PZ in the lower panel. The major peaks of both chromatograms contain p25 as shown in Fig. 5.

and incubated for 2 h. The filters were then washed three times in PBS, 0.05% Tween and once in 50 mM Tris, pH 7.6, and developed in 50 mM Tris, pH 7.6, containing diaminobenzidine tetrahydrochlo- ride chromogen (Dakopatts) 0.20 mg/ml and 0.03% HzOz.

Bio-Rad with catalase as standard (21). Assays-Total protein was measured using a commercial kit from

RESULTS Human neutrophil gelatinase was purified as described by

Hibbs et al. (12) from material exocytosed from PMA-stimu-

lated neutrophils. SDS-PAGE of the purified preparation is shown in Fig. 1A. In accordance with the findings of Hibbs et al. (12), three bands of 220, 135, and 92 kDa were seen under nonreducing conditions, and one major 92-kDa band under reducing conditions. A faint 25-kDa band, not described by Hibbs et ul. (12), was observed in addition to the 92-kDa band, but only under reducing conditions, indicating the existence a covalent complex between this protein (henceforth termed p25) and gelatinase.

The purified gelatinase was used for immunization of rab- bits. This resulted in generation of polyclonal antibodies that possessed reactivity toward both gelatinase and p25, as dem- onstrated by Western blotting of a postnuclear supernatant of disrupted human neutrophils (Fig. 1B). In order to separate gelatinase and p25 for subsequent use in affinity purification of the antibodies, we performed gel filtration of material exocytosed from neutrophils in response to stimulation with PMA. For this separation to be successful, p25 has to be exocytosed not only in association with gelatinase, but also in a “gelatinase-free” form. This is indeed the case, as shown in Fig. lC, which demonstrates immunoblotting of two differ- ent fractions eluted from the Sephadex G-200 column, probed with the antibodies reacting with both gelatinase and p25. One fraction contained gelatinase, whereas the other fraction contained p25, but was devoid of gelatinase. Fractions con- taining gelatinase-free p25 were coupled to CNBr-activated Sepharose, and the gelatinase antibodies were applied to this column. Antibodies directed against p25 bound to the column, and were subsequently eluted, whereas the antibodies with specificity toward 92-kDa gelatinase did not bind and were collected in the void. By this procedure, specific antibodies against p25 and gelatinase were obtained as evidenced by Western blotting (Fig. 1D).

A very faint 69-kDa band is noticed in SDS-PAGE of purified reduced gelatinase (Fig. L4, lune 2) and also in Western Blotting using antibodies obtained after immuniza- tion of rabbits with the purified gelatinase (Fig. 1, B and C ) . The nature of this 69-kDa protein is not known.

The p25 antibody reacted selectively with the 135-kDa form of unreduced gelatinase both by immunoprecipitation (Fig. 2, lanes 2 and 4 ) and by immunoblotting (Fig. 3A), indicating that the 135-kDa gelatinase is a complex of p25 and 92-kDa gelatinase. No reactivity of the 220-kDa gelatinase with anti- p25-antibodies was observed. The 220-kDa gelatinase is there- fore either a homodimer of 92-kDa gelatinase or a complex of two 92-kDa gelatinase and one p25 molecule in which p25 is hidden from the antibodies. This latter possibility is less likely since reduction of the 220-kDa band (after immunoprecipi- tation of the 135-kDa form) did not reveal any p25 (Fig. 2, lune 2, lower panel). Thus, the association of 92-kDa gelatin- ase with p25 explains the peculiar 135-kDa form of “pure” gelatinase from human neutrophils.

To investigate the association of p25 with other neutrophil proteins, Western blotting of a postnuclear supernatant was performed under nonreducing conditions, using anti-p25 an- tibodies. This demonstrated four bands of 135, 70, 46, and 25 kDa, respectively (Fig. 3B). The 25- and 46-kDa bands rep- resent p25 in its mono- and dimeric forms, as shown below. The 135-kDa band could be accounted for as a complex between 92-kDa gelatinase and p25. The nature of the 70- kDa band is unknown, but it is possible that p25 even exists in a trimeric form.

The protein was purified from exocytosed material from PMA-stimulated human neutrophils. After precipitation by 18% PEG, the supernatant (including p25) was subjected to cation exchange chromatography on MonoS. p25 eluted in

Page 5: THE OF 268, No. 14, 15, pp. 10425-10432,1993 Q 1993 for ... · THE JOURNAL OF BIOLOGICAL CHEMISTRY Q 1993 by The American Society for Biochemistry and Molecular Biology, Inc. Vol

NGAL Protein 10429

, Mr (10’) Mr (10’) 200 ’

97.4 ’ 69

46 >

30 ’

21.5 ’ 14.3 >

200 ’

97.4 ’ 69 >

46 >

I I 30 ’

C 4 14.3 ’

1 2 3 4 5 6 7 a 9 10 FIG. 5. SDS-PAGE protein profiles during purification of p25. Lane I , supernatant from PMA-stimulated neutrophils (62 pg); lane

2, PEG supernatant (75 pg); lane 3, P, from MonoS (4.5 pg); lane 4, P2 from MonoS (7.5 pg); lane 5. peak fraction from gel filtration of PI (4.25 pg); lane 6, peak fraction from gel filtration of Pz (4.25 pg); lane 7, p25 incubated with N-glycanase; lane 8, p25 incubated with 0- glycanase. Lane 9, peak fraction from gel filtration of PI (4.25 pg, nonreducing conditions); lane IO, peak fraction from gel filtration of Pz (4.25 pg, nonreducing conditions). Unless otherwise indicated electrophoresis is run under reducing conditions.

Mr (lo3)

200 ’

97.4 ’ 69 >

46

30 >

21.5 >

14.3 ’

subjected to gel filtration on Superose 12 (Fig. 4, lower panels). Gel filtration of PI resulted in the elution of a major protein peak. SDS-PAGE of this peak revealed that p25 had been purified to apparent homogeneity (Fig. 5, lanes *5 and 9 ) . Gel filtration of Pi! resulted in the purification of p25 in i t s dimeric form, with a molecular mass of 46 kDa on SDS-PAGE in the absence of reducing agent (Fig. 5, lanes 6 and IO). The purified protein was recognized by the anti-p25-antibodies as shown by Western blotting (Fig. 6).

The yield of purified p25, combining the mono- and dimeric forms, was 350 pg from a total of 140 mg of protein present in exocytosed material from approximately 15 x loe PMA- stimulated neutrophils.

Amino acid sequence analysis of the native p25 failed, presumably due to blockage of the NHderminal amino acid. This was even the case after reduction and carboxymethyla- tion of the protein. The native protein was also resistant to digestion with trypsin (not shown). However, after reduction and carboxymethylation, the protein could be cleaved by trypsin as well as by endoprotease Glu-C, endoprotease Asp- N and CNBr. The resulting peptides were isolated by HPLC.

1 2 3 4 FIG. 6. Western blotting of purified p25 with anti-p26-an-

tibodies. Blotting was performed as described under “Experimental Procedures.” Primary antibodies were anti-p25 antibody diluted 2000- fold ( l o n e 2-3) or preimmune rabbit I& (diluted to equal amounts of I&). Lane 1 , purified monomeric p25 (0.5 pg); lone 2, purified dimeric p25 (1 pg, reducing conditions); lane 3, purified dimeric p25 (1 pg, nonreducing conditions); lone 4, purified monomeric p25 (0.5 pg) probed with preimmune I&.

two peaks (PI and P2), as illustrated in Fig. 4 (upper panel). Most of the eluted p25 was present within these two peaks as evidenced by SDS-PAGE of individual fractions (not shown). SDS-PAGE profiles of the supernatant after PEG precipita- tion and of PI and Pz are shown in Fig. 5 (lanes 2 4 ) . Fractions from each peak were pooled and concentrated separately and

After sequence analysis of these fragments, it was possible to determine the t o t a l amino acid sequence of p25 (Fig. 7 ) . The NHz-terminal sequence of p25 was determined by identifica- tion of a tryptic fragment (Fig. 7 , T-1 ) resistant to sequence analysis. Following digestion with pyro-glutamate aminopep- tidase, sequence analysis of the resulting peptide was possible. This shows the NHz-terminal residue to be pyroglutamate. This explains the resistance to sequence analysis of the intact carboxymethylated protein (Fig. 7 ) . The COOH terminus of p25 as shown in Fig. 7 was deduced from the two fragments (7’-21 and E-IO) both terminating in a Gly, since neither trypsin nor endoprotease Glu-C are known to cleave at the COOH-terminal side of Gly.

The total amino acid composition of the carboxymethylated p25 was determined by amino acid analysis and was found to be in agreement with the composition deduced from the total

Page 6: THE OF 268, No. 14, 15, pp. 10425-10432,1993 Q 1993 for ... · THE JOURNAL OF BIOLOGICAL CHEMISTRY Q 1993 by The American Society for Biochemistry and Molecular Biology, Inc. Vol

10430 NGAL Protein

IpEDSTSDL I PA PPLSKVPLQQ NFQDNQFQGK WYVVGLAGNA I LREDKDPQK MYAT I YELKE" II

T- 1 -

T-2 T-3 T-6

E-2 1

I t D-3'

I D-6

c-2

"'VFFKKVSQNR EYFK I TLYGR TKELTSELKE NF I RFSKSLG LPENH I VFPV P I DQC I DG178 - H- I I T-15 T-16 T-17" T-19 T-2 1 -

E-6 I

I

II E-9 E-1 0

t c-5

FIG. 7. Amino acid sequence of p25. The sequence was deduced from fragments obtained by cleavage of the reduced and carboxymeth- ylated protein by trypsin (T-I to T-21), endoprotease Glu-C (E-2 to E-IO), endoprotease Asp-N (0-3' to 0-T), and CNBr (C-2 to C-5). The fragments are numbered starting from the NH, terminus according to all hypothetical cleavage points. Note that CNBr induced cleavage both after M and W (occasionally leaving out a potential cleavage point). Fragments including a hypothetical cleavage point are named by the first included fragment and marked by a hyphen (0-3 ' , C-3', T-IO', and T-17'). Horizontal lines indicate residues identified by sequence analysis. Vertical lines indicate the NH, terminus of a fragment as determined by sequence analysis and the COOH terminus according to the measured M , (Table II), respectively. * denotes glycosylation site as determined from 0 - 7 before and after deglycosylation with N - glycanase (0-T) . Every 10 amino acids in the sequence are grouped in order to improve the clarity of the figure.

TABLE I Amino acid analysis of 60 pmol of carboxymethylated p25

Numbers in brackets are values determined by counting residues found in the sequence.

Amino acid Number of residues

Asx 18.7 (19) Glx" 22.6 (20) Ser 10.7 (12) His 2.5 (2) GlY 10.2 (10) Thr 9.3 (10) Ala 5.3 (5) '4% 7.2 (7) TYr 9.1 (10) Val 12.0 (13) Met 2.2 (2) Ile 9.4 (10) Phe 11.0 (11) Leu 15.3 (15) LY s 18.3 (16) Pro 11.6 (11) Trpb (2) Cys' (3)

' CM-Cys elutes as two peaks, one which coelutes with Glx causing

Not determined, due to destruction, but the presence confirmed

'Not determined due to great variability but the presence con-

slight overestimation.

from second derivative of the UV spectrum.

firmed by specific peak in the chromatogram.

amino acid sequence of the protein (Table I). Furthermore, the molecular mass of most peptides depicted in Fig. 7 was measured by mass spectrometry. These values were in agree- ment with those calculated from the sequence (Table 11).

The protein was subjected to treatment with N- and 0- glycanase. 0-Glycanase was without effect, whereas N-glycan- ase resulted in a partial reduction in the molecular mass to approximately 21 kDa (Fig. 5, lunes 7 and 8) , indicating posttranslational addition of N-linked sugar residues. The molecular mass of the deglycosylated protein is in agreement with the molecular mass calculated from the 178 amino acid residues of the protein backbone (Mr = 20,542)(Fig. 7 and Table I). Sequence analysis of all fragments containing resi- due 65 resulted in blank cycles for this particular residue. This indicated glycosylation of residue 65 in accordance with the consensus sequence for N-glycosylation, -Am-Xaa-Ser/ Thr-. To verify this, an aliquot of the fragment D-7 (residues 61-76, Fig. 7 ) was digested with N-glycanase. The resulting peptide was less hydrophilic than the original peptide as indicated by the following RP-HPLC purification. Subse- quent sequence analysis showed residue 65 to be an Asp, confirming this residue to be a glycosylated Asn in the native protein.

DISCUSSION

We report here the purification and identification of p25, a 25-kDa human neutrophil protein, that is in part associated with gelatinase from human neutrophils. In search for pro- teins with similarity to the amino acid sequence of p25 (Fig.

Page 7: THE OF 268, No. 14, 15, pp. 10425-10432,1993 Q 1993 for ... · THE JOURNAL OF BIOLOGICAL CHEMISTRY Q 1993 by The American Society for Biochemistry and Molecular Biology, Inc. Vol

NGAL Protein 10431

7) 74,551 protein sequences included in the MIPSX data base at EMBL (release October 1992) were screened. The deduced sequences of the rat a2-microglobulin-related protein (22) and the mouse 24p3 protein (23), which are 80% identical, showed the highest degree of similarity with p25 (63.5 and 62% identity, respectively), indicating that p25 is a human ana- logue of these proteins. The function of these rodent proteins is not known. Fig. 8 shows alignment of p25 with the deduced rat a2-microglobulin-related protein precursor. The first 20 residues of the hypothetical mouse protein 24p3 (23) fulfill the criteria for a consensus sequence of a signal peptide (24). This indicates that residue 21 of the precursor is the first residue in the mature protein. In contrast, the deduced se- quence of the rat protein contains an Arg in position -3 from

TABLE I1 Molecular masses of fragments used for identification of p25

The fragments are identified in Fig. 7. The measured M , values are compared to those calculated from the sequences, given in brackets. No M, was obtained for the glycosylated fragments E-4, D-7, and C- 3’ nor for the deglycosylated D-7* of which too little was left. Furthermore, D-3’ evaded mass determination.

Fragment” Residues M .

T-1A 1-15 1553.1 (1553.4) T-1B 2-15 1442.1 (1440.6) T-2 16-30 1792.3 (1790.9) T-3 31-43 1432.4 (1431.7) T-6 51-59 1131.9 (1131.3) T-10‘ 75-81 1040.5 (1040.2) T-11 82-98 1866.3 (1865.1) T-12 99-109 1255.6 (1255.4) T-13 110-125 1914.2 (1913.2) T-15 131-134 586.2 (585.7) T-16 135-140 722.6 (721.9) T-17‘ 141-149 1049.1 (1048.2) T-19 150-154 678.7 (677.8) T-21 158-178 2321.4 (2320.6) E-2 45-57 1602.3 (1601.8) E-4 E-6 E-9 151-163 1508.5 (1507.7) E-10 164-178 1726.4 (1724.0)

61-? 132-143 1520.6 (1518.8)

D-3‘ 6-? D-6 47-60 1730.4 (1729.0) D-7* c-2

61-? 32-51 2170.4 (2169.5)

c-3’ 52-? c -4 80-120 4572.7 (4569.2) c -5 121-178 6883.4 (6878.9)

T-lA, before digestion with pyroGlu aminopeptidase; T-lB, after digestion with pyroGlu aminopeptidase.

the potential signal peptide cleavage site. This violates the rule which predicts the residue in this position to be nonpolar (24), indicating a sequencing error.

The rat az-microglobulin-related protein and the mouse 24p3 protein both belong to a functionally heterogeneous family of proteins named lipocalins (25, 26). Human proteins belonging to the lipocalins include brain prostaglandin D synthase (32.7% identity with p25), a,-microglobulin (24.6% identity with p25), retinol-binding protein, and apolipoprotein D (27-30). The latter two demonstrated only a low degree of similarity with p25. Despite a low degree of similarity regard- ing overall amino acid sequence between the different mem- bers of this family, certain motifs within the structure of the proteins are conserved (25, 29, 31). These motifs are also present in p25, indicating that p25 is a novel lipocalin mem- ber. The motifs are situated very close together in the proteins and are responsible for their common tertiary structure, with one a-helix and a P-barrel surrounding a hydrophobic core, in which the ability to bind small lipophilic ligands resides (28,29,31). For many lipocalins the lipophilic ligand is retinol, but for other members including p25, the lipophilic ligand is unknown.

In addition to similarity regarding the primary structure, p25 shows several traits that are characteristic of the lipocalin family of proteins. First, their protein backbones are all of a size around 20 kDa (32). Second, their compact structure makes them resistant to proteases (26). p25 was hardly de- graded by trypsin, unless the protein was reduced and carbox- ymethylated. Furthermore, their common a-helix is believed to be responsible for their tendency for self-association (28), as also observed with p25. The ability to associate with other proteins has been described for other lipocalins, namely al- microglobulin (28), which is covalently bound to IgA, and apolipoprotein D which is bound to 1ecithin:cholesterol acyl- transferase (30). In agreement with this, we have demon- strated that p25 is covalently bound to gelatinase. This 25- kDa protein copurifies with neutrophil gelatinase, forming a covalent complex with a molecular mass of 135 kDa. This association explains the unusual molecular mass forms of “pure” unreduced neutrophil gelatinase. We therefore suggest the protein be named neutrophil gelatinase-associated lipo- calin (NGAL) .

The existence of a homodimer of 92-kDa gelatinase and a complex between 92-kDa gelatinase and an unidentified 25- kDa protein resulting in the 135-kDa form of neutrophil gelatinase has recently been described by others (33,34). One would have expected this 25-kDa protein to be TIMP-1 or an

az MGLGVLCLALVLLGVLQRQAQDSTQNLIPAPPLISVPLQPGFWTERFQGR I I I I : I I I I I I I I I I I : I : I l l :

P2 5 ~DSTSDLIPAPPLSKVPLQQNFQDNQFQGK

FIG. 8. Alignment of p25 with the a2 WFVVGLAANAVQKERQSRFTMYSTIYELQEDNSYNVTSILVRGQGCRYWI deduced precursor of rat a*-micro- I : I I I I I : I I : : : I : : : I I : I I I I I : I I : I I I I I I : I : I : I I l l globulin-related protein (az). Solid p25 WYWGLAGNAILREDKDPQKMYATIYELKEDKSYNVTSVLFRKKKCDYWI lines represent identity of amino acids (63.5%), whereas “:” represents amino acid substitutions, resulting from single a2 RTFVPSSRPGQFTLGNIHSYPQIQSYDVQVADTDYDQF~VFFQKTSENK base substitutions (27.0%). I I I I I : : I I : I I I I I I I l l : I I I : I : 1:I:l I I I I I : I I : I :

p25 RTFVPGCQPGEFTLGNIKSYPGLTSYLVRWSTNYNQHAMVFFKKVSQNR

a2 QYFKVTLYGRTKGLSDELKERFVSFAKSLGLKDNNIVFSVPTDQCIDN

p25 EYFKITLYGRTKELTSELKENFIRFSKSLGLPENHIVFPVPIDQCIDG : 1 1 1 : 1 1 1 1 1 1 1 : 1 : I I I I I : I I I I I I : I : I I I : I I : I I I I I

50

30

100

80

150

130

198

178

Page 8: THE OF 268, No. 14, 15, pp. 10425-10432,1993 Q 1993 for ... · THE JOURNAL OF BIOLOGICAL CHEMISTRY Q 1993 by The American Society for Biochemistry and Molecular Biology, Inc. Vol

10432 NGAL Protein

analogous protein, since association of gelatinase with TIMP- 1 has been reported in other cell types like SV-40-transformed fibroblast, HT 1080 fibrosarcoma cells, U937 myelomonocytic cells, and human monocytes (8-10). Neutrophil gelatinase is thus unique in being TIMP-free and in part covalently asso- ciated with this 25-kDa protein, which is not a homologue of TIMP-1. As proposed by others (8), the fact that neutrophil gelatinase is TIMP-free may explain its high specific activity compared to gelatinases from other sources. This could reflect the need for a specialized function of neutrophil gelatinase, in providing the environment necessary for diapedesis of the cell.

Only a part of neutrophil gelatinase is covalently complexed to NGAL. Also, we find that NGAL can be purified in a mono- and dimeric form not complexed with gelatinase. The exist- ence of these uncomplexed forms was confirmed by Western blotting of a postnuclear supernatant. Thus, in spite of the clear association between NGAL and gelatinase, both proteins seem to exist mainly in forms unassociated with each other. An explanation of this might be that the uncomplexed forms of neutrophil gelatinase and NGAL are segregated, residing in different subcellular compartments. Neutrophil gelatinase is mainly located in a subcellular compartment of its own (35, 36), which is characterized by a high degree of mobilization upon weak stimulation of the neutrophil. This compartment is different from specific and azurophil granules. The subcel- lular localization of NGAL remains to be determined.

The present isolation and identification of a new member of the lipocalin family opens opportunities to determine its function. The partial association of NGAL with neutrophil gelatinase suggests that NGAL may exert modulatory actions on gelatinase. However, NGAL is mostly exocytosed from the neutrophil in a form not complexed to gelatinase. Further- more, NGAL is a member of the lipocalin family, and may therefore possess the ability to bind small lipophilic sub- stances. This indicates that NGAL may serve other functions yet to be determined.

Acknowledgments-The expert technical assistance of Charlotte Horn, Pia Linnk Olsen, Ronnie Danielsen, and Allan Kastrup is gratefully appreciated.

Note Added in Proof-While this paper was under editorial evalu-

ation, a report appeared that confirmed the existence of a complex of neutrophil gelatinase with a lipocalin (Triebel, S., Blaser, J., Reinke, H., and Tschesche, H. (1992) FEBS Lett. 314, 386-388).

REFERENCES

2. Emonard, H., and Grimaud, J.-A. (1990) Cell. Mol. Biol. 36, 131-153 1. Docherty, A. J. P., and Murphy, G. (1990) Ann. Rheum. Dis. 49,469-479

3. Reynolds, J. J. (1986) Ann. Biol. Clin. 44, 188-194 4. Mercer, E., Cawston, T. E., De Silva, M., and Hazlemann, B. L. (1985)

5. Stricklin, G. P., and Welgus, H. G. (1983) J. Biol. Chem. 268,12252-12258 6. Goldhere. G. I.. Marmer. B. L.. Grant. G. A.. Eisen. A. Z.. Wilhelm. S.. and

Biochem. J. 231,505-510

He, Cy(1989) Proc. Natl. Acad. Sci.'U. S. A . 86,8207-8211 . ,

7. Stetler-Stevenson. W. G.. Krutzsch. H. C.. and Liotta. L. A. (1989) J. Biol. ~Chem. 264,17374-17378

and Goldberg, G. I. (1989) J. Biol. Chem. 264, 17213-17221

(1991) FEBS Lett. 284,73-78

Cancer Res. SO, 6162-6170

~I I , ,

8. Wilhelm, S. M., Collier, I. E., Marmer, B. L., Eisen, A. Z., Grant, G. A.,

9. Opdenakker, G., Masure, S., Proost, P., Billiau, A,, and van Damme, J.

10. Moll, U. M., Youngleib, G. L., Rosinski, K. B., and Quigley, J. P. (1990)

11. Murphy, G., Bretz, U., Baggiolini, M., and Reynolds, J. J. (1980) Biochem.

12. Hihbs, M. S. , Hasty, K. A,, Seyer, J. M., Kang, A. H., and Mainardi, C. L.

13. Kjeldsen, L., Bjerrum, 0. W., Hovgaard, D., Johnsen, A. H., Sehested, M., (1985) J. B i d . Chem. 260,2493-2500

14. Boyum, A. S (1968) Scand. J. Clin. Lab. Inuest. 21, (suppl.) 77-89 and Borregaard, N. (1992) Eur. J. Haematol. 49,180-192

15. Bush , T. F., and Ingham, K. C. (1980) Vox. Sang. 39,93-100 16. Chargonneau, H. (1989) in A Practical Guide to Protein and Peptide

Pur i$+on for Microsequencing (Matsudaira, P. T., ed), pp. 15-30, Aca emlc press, New York

J. 192,517-525

17. Shuster, R. (1988) J. Chromatogr. 431,271-284 18. Laemmli, U. K. (1970) Nature 227,680-695 19. Borre aard N., Heiple, J. M., Simons, E. R., and Clark, R. A. (1983) J. Cell

20. Towbin, H. T., Staehelin, T., and Gordon, J. (1979) Proc. Natl. Acad. Sci. Biof 97,'52-61

r J . s. A . 76.43.50-4354 21. Bradford, M. M. (1976) Anal. Biochem. 72,248-254 22. Chan, Y.-L., Paz, V., and Wool, I. G. (1988) Nucleic Acids Res. 16,11368 23. Hraha-Renevev. S.. Turler. H.. Kress. M.. Salomon, C.. and Weil. R. (1989)

- . ~ . .. ~ ~ . . ~~~~

. . . . Oncogene 4,v601-607

. .

24. von Hei'ne G (1986) Nucleic Acids Res. 14,4683-4690 25. Flower, b. R.,North, A. C. T., and Attwood, T. K. (1991) Biochern. Biophys.

26. Peitsch, M. C., and Boguski, M. S. (1991) Trends Biochem. Sci. 16,363 27. Nagata, A., Suzuki, Y., Igarashi, M., Eguchi, N., Toh, H., Urade, Y., and

28. Akerstrom B., and Logdberg, L. (1990) Trends Biochern. Sci. 16,240-243 Hayaishi, 0. (1991) Proc. Natl. Acad. Sci. U. S. A. 88,4020-4024

29. Pervaiz, S.: and Brew, K. (1987) FASEB J. 1,209-214 30. Drayna, D., Fielding, C., McLean, J., Baer, B., Castro, G., Chen, E.,

Comstock, L., Henzel, W., Kohr, W., Rhee, L., Wion, K., and Lawn, R. (1986) J. Biol. Chem. 261,16535-16539

31. Cowan. S. W.. Newcomer. M. E.. and Jones. T. A. (1990) Proteins Stuct.

Res. Commun. 180,69-74

Funit. Genet. 8,44-61 '

32. Godovac-Zimmermann, J. (1988) Trends Biochem. Sci. 13,64-66 33. Goldberg, G. I., Strongin, A., Collier, I. E., Genrich, T., and Marmer, B. L.

34. Triebel, S., BlLer, J., Reinke, H., Knauper, V., and Tschesche, H. (1992)

35. Deztld, B., Bretz, U., and Baggiolini, M. (1982) J. Clin. Inuest. 70, 518-

(1992) J. Biol. Chum. 267,4583-4591

FEBS Lett. 298,280-284

36. Kjeldsen, L., Bjerrum, 0. W., Askaa, J., and Borregaard, N. (1992) Biochem. DZD

J. 287,603-610