EP0758393A1 - Mac-1 i-domain protein useful in blocking adhesion and migration of neutrophils - Google Patents

Abstract

An isolated and purified protein of the I-Domain from the human leukocyte β2-integrin Mac-1, expressed in recombinant Escherichia coli as a soluble fusion protein with glutathione S-transferase (GST). The protein, a functional derivative, fragment, analog or chemical derivative of such fragment is useful in the treatment of inflammation by interfering with the cell adhesion mechanism to block adhesion and migration of neutrophils.

Description

Mac-1 1-Domain Protein Useful in Blocking Adhesion and Migration of Neutrophils BACKGROUND OF THE INVENTION

Cell adhesion molecules are essential in a number of cellular processes including immunity and inflammation, cell anchorage and migration, and cell growth and differentiation. Among the large number of various cell adhesion proteins, the leukocyte integrins are involved in mediating the adhesion and endothelial trans-migration of leukocytes into inflamed tissue. These integrins are membrane-anchored proteins on the surface of the leukocytes, serving as receptors to various ligands. These receptors are heterodimer proteins consisting of alpha and beta subunits, and since the 3 known leukocyte integrins share a common beta-2 subunit, they are also called beta-2 integrins. In these 3 β2 integrins, LFA-1 (CDlla/CD18), Mac-1 (CDllb/CD18), and pl50,95 (CDllc/CD18), the β subunit CD 18 interacts with the α subunit CD 11 in a noncovalent manner. Both the α and β subunits of these integrins are large glycoproteins; the

CD 18 subunit is about 95 kd in molecular mass. The 170 kd α subunit CD lib in Mac-1 contains an I-domain (or A-domain) which is also conserved in CD Ila, CD lie, and a few other integrins (Larson et al, J. Cell. Biol., 108:703-12 (1989)). This I- domain of about 200 amino acid residues appears to be an insertion sequence; it is not found in some other integrins. The I-domain exhibits sequence similarity to the domains involved in ligand binding of proteins such as von Willebrand factor, cartilage matrix protein, and complement proteins C2 and factor B.

Thus it is intriguing to consider that the I-domain in the β2 integrins may play a role in the interaction of these integrins to their ligands. Results suggesting the I-domain might function as a recognition region in ligand binding were reported by Diamond et al, J. Cell Biol., 120:1031-43 (1993); they showed that monoclonal antibodies to CDllb I-domain could block the binding of Mac-1 to iC3b, fibrinogen, ICAM-1 and a neutrophil Ugand. It is known that Mac-1 Ugand binding requires the presence of cation. Michishita et al, CeU, 72:857-67 (1993) demonstrated the binding of Mn to a E. coZi-derived recombinant I-domain of CDllb, again suggesting the involvement of the I-domain in Ugand binding.

To provide CDllb I-domain to study its function and structure, E. coli expression of the recombinant protein was required. Unfortunately, expression levels of the I-domain by itself were low, and the protein was difficult to purify. Attempts were made to generate a strain which expressed, at high levels, the I- domain as a fusion to GST (glutathione S-transferase), with a thrombin cleavage site at the fusion junction so that the I-domain could be recovered from the fusion product. However this fusion protein was not susceptible to thrombin proteolysis.

The present invention demonstrates that producing recombinant proteins in E. coli as fusions to GST has some of the foUowing advantages: 1) GST itself is expressed at high levels and therefore should promote high level production of the fusion protein, 2) GST itself is highly soluble and can serve to faciUtate/stabilize the folding of the fusion protein, leading to a soluble and active product, and 3) GST can be used as a purification handle, i.e., one-step purification of the fusion protein by affinity chromatography on immobiUzed glutathione, using a glutathione Sepharose 4B column.

The present construct overcomes problems caused by the physical close proximity of the GST and I-domain, masking the accessibiUty of the cleavage site. The present construct introduces residues such as glycine at the GST and I-domain fusion point to separate the two structures. In addition, the present invention teaches a GST and I-domain fusion with a Factor X^a cleavage site to compare the efficiency of thrombin and Factor X^a proteolysis in such fusion products. The different fusion junctions linking GST and I-domain are shown below. The residues underlined are the recognition sequence for thrombin or Factor X^a with the arrow(- T-) indicating the cleavage point. The residues highUghted in boldface are the extensions at the N-terminus of I-domain after proteolysis. Because of the requirements for recognition by the protease, extra residues in addition to the desired 2 glycines are necessary for extension of the I-domain from the fusion constructs. Thrombin

GST-Leu-Val-Pro-Arg-T-Glv-Ser-(Serl33...Glu337) Thrombin

GST-Leu-Val-Pro-Arg-t-Glv-Ser-GlvGlv-(Serl33...Glu337) Factor X^a GST-Ile-Glu-Glv-Ar_?-T-Glv-Ile-Pro-Glv-Glv-(Serl33...Glu337)

Note that in the subject recombinant product, the CDllb I-domain sequence is that defined by the serine at position #133 ending at the glutamic acid at #337. In the report of Michishita et al (cited, above), their E. coli-άe^ήved recombinant CDllb I-domain (caUed the A domain) starts at glycine at #111 and ends with alanine at #318 (see Figure 1). Without engineering for the exact I-domain sequence, Michishita et al took advantage of the available restriction enzyme sites to clone a sequence coding for the region around the I-domain, thus shifting the I- domain towards the N-terminus of CDllb.

Uniquely, the addition of two glycine residues between the GST and I-domain in the subject construct does aUow thrombin processing of the fusion protein. The construct with the Factor X^a cleavage site also can be processed by Factor X^a efficiently, giving a better yield of I-domain than that from the fusion with thrombin cleavage. The subject invention provides ceU extracts prepared from the E. coli strain producing the fusion protein with the X^a cleavage site for downstream isolation of the CDllb I-domain.

INFORMATION DISCLOSURE STATEMENT

European Patent AppUcation 0365837 (Springer et al.) discloses the general background for interceUular adhesion molecules (ICAM-1) and their function derivatives which may be useful in the treatment of inflammation.

European Patent AppUcation 0391088 (Springer et al.) discloses the interceUular adhesion molecules (ICAM-1) and their function derivatives which may be useful in the treatment of viral infections.

European Patent AppUcation 0364690 (Springer et al.) discloses the leukocyte adhesion receptor Mac-1 alpha subunit and corresponding DNA and derivatives which may be useful in the treatment of inflammation.

European Patent AppUcation 0387668 (Springer et al.) discloses the interceUular adhesion molecules classified as ICAM-2 which are described to be involved in the process where lymphocytes migrate to inflammation sites. Diamond et al., "The I-domain is a Major Recognition Site on the Leukocyte

Integrin Mac-1 (CDllb/CD18) for Four Distinct Adhesion Ligands, J. of CeU Biology, 120, 4, 1031-1043 (1993) discloses that mAbs specific for the I-domain block Mac-1- dependent adhesion.

Michishita et al., "A Novel Divalent Cation-Binding Site in the A Domain of the Beta-2 Integrin CR3 (CD11/CD18) Is Essential for Ligand Binding", CeU, 72, 857-867 (March 26, 1993) discloses that certain mutations in the I-domain region of Mac-1 block adhesion to iC3b.

Zhou et al., "Differential Ligand Binding Specificities of Recombinant CDllb/CD18 Integrin I-Domain", J. Biological Chemistry, 269, 25, 17075-17079 (June 24, 1994) report the expression of a recombinant form of the I-domain of CDllb and that this domain binds fibrinogen and ICAM-1 but that it does not recapitulate the entire CDllb/cdl8 Ugand repertoire.

Kern et al., "The Role of the I Domain in Ligand Binding of the Human Integrin α₁β₁", J. Biological Chemistry, 269, 36, 22811-22816 (September 9, 1994) report that the I Domain plays a central role in Ugand recognition for aU integrin α subunits containing this domain.

Muchowski et al., "Functional Interaction between the Integrin Antagonist NeutrophU Inhibitory Factor and the I Domain of CD lib/CD 18", J. Biological Chemistry, 269, 42, 26419-26423 (October 21, 1994) report that recombinant neutrophU inhibitory factor (rNIF) associates with the about 200 amino acid residue I domain of CD lib/CD 18 and that this intereaction is essential for inhibition of neutrophil function by NIF.

U.S. Patent 5,091,303 to Arnaout et al. discloses a 29kD neutrophiUc protein which binds to autoantibodies present in the sera of individuals afflicted with Wegener's granulomatosis and methods using these autoantibodies to diagnose individuals afflicted with Wegener's granulomatosis.

U.S. Patent 5,200,319 to Arnaout et al. discloses a 29kD neutrophiUc protein which is used in a method of diagnosing pauci-immune nectrotizing and/or crescentic glomerulnephritis in a patient.

SUMMARY OF THE INVENTION

In one aspect the subject invention is a fusion protein of glutathione-S- transferase (GST) and I-Domain derived from human leukocyte B₂-integrin Mac-1 in which the GST and the I-Domain are linked by a peptide segment containing a Factor X^a cleavage site as set forth in ID SEQ NO: I. This fusion protein provides an improved means for handling the I-Domain protein for synthesis and expression and a unique cleavage site which provides accessabiUty for cleavage by a Factor X^a enzyme.

In another aspect, the invention is an I-Domain protein defined by the amino acid sequence Gly 226 through Glu 435, inclusive, as set forth in ID SEQ NO: I. This sequence contains the I-Domain and the above mentioned special cleavage site. In yet another aspect, the invention is a pharmaceutical composition comprising the recombinant I-Domain protein derived from human leukocyte B₂- integrin Mac-1 defined by the amino acid sequence Gly 226 through Glu 435, inclusive, as set forth in ID SEQ NO: I; and a pharmaceuticaUy acceptable carrier or excipient. The pharmaceutical composition can also consist of a fragment, analog or chemical derivative of the aforementioned I-Domain protein.

In yet another aspect, the invention is a method for treating inflammation comprising the administration to a patient suffering from an inflammatory condition a pharmaceuticaUy effective amount of an anti-inflammatory agent comprising a recombinant I-Domain protein derived from human leukocyte B₂-integrin Mac-1 defined by the amino acid sequence Gly 226 through Glu 435, inclusive, as set forth in ID SEQ NO: I. The method can also consist of a fragment, analog or chemical derivative of the aforementioned recombinant I-Domain protein. The fusion protein, I-domain protein (which can contain the special cleavage segment) and the pharmaceutical compositions prepared therefrom are aU recombinant proteins substantiaUy free of contaminants or other biological impurities.

BRIEF DESCRIPTION OF THE DRAWING

Figure 1. Schematic representation of the A-domain described by Michishita et al., "A Novel Divalent Cation-Binding Site in the A Domain of the Beta-2 Integrin CR3 (CD11/CD18) Is Essential for Ligand Binding", CeU, 72, 857-867 (March 26, 1993) and the I-domain of the present invention. Note that the A-domain is about the same size as the I-domain, but begins about 20 amino acids upstream in the sequence of Mac-1 (Gly^). Accordingly, it ends at Ala_31g, some 20 residues upstream of the C-terminal residue in the I-domain (Glu₃₃₇). The A-domain has a C-terminal extension of Asn-Ser-Ser, introduced as part of the cloning strategy. The I-domain is N-terminaUy extended by H₂N-Gly-Ile-Pro-Gly-Gly-.. , a sequence required to promote cleavage by Factor X^a . The first Mac-1 residue in the I-domain construct is the serine corresponding to Ser₁₃₃.

DETAILED DESCRIPTION OF THE INVENTION

The present invention is an isolated and purified protein corresponding to the I-domain from the human leukocyte β₂-integrin Mac-1, expressed in recombinant Escherichia coli as a soluble fusion protein with glutathione S-transferase (GST). The protein is useful in the treatment of inflammation by interferring with the ceU adhesion mechanism to block adhesion and migration of neutrophUs.

The fusion protein corresponding to the glutathione-S-tranferase (GST) and the I-Domain from the human leukocyte β₂-integrin Mac-1, is shown as a complete amino acid sequence (SEQ ID NO: I). The I-Domain protein amino acids are numbered according to their location in the fusion protein. The NH₂-terminal portion of the molecule (residue 220) is glutathione-S-transferase (GST), a fusion partner which helps increase the level of soluble expression, and faciUtates purification by GSH-affinity chromatography. Residues from 221-230 indicate the segment linking GST to the I-domain and the site of cleavage by Factor X^a at Arg225 - Gly226 is dictated by the sequence; [Ile-Glu-Gly-Arg-i-Gly-Ile-Pro-]. The boldface Gly-Gly sequence represents a spacer to aUow accommodation of the proteinase to the site of cleavage at Arg-Gly. The Ser at 231 foUowing the Gly-Gly sequence corresponds to the beginning of the I-domain, Ser₁₃₃ in processed Mac-1, and the construct ends with Glu₃₃₇. Accordingly, the I-domain begins with the pentapeptide sequence: Gly-Ue-Pro-Gly-Gly- required to aUow removal of the GST by Factor X^a, foUowed by the I-domain of Mac-1 (residues 133 through 337). This structure is schematicized in Fig. 1.

The isolated protein is useful for the treatment of inflammation and related conditions in human patients and other warm blooded animals by either parenteral or oral administration.

Pharmaceutical compositions of this invention may be prepared by combining the protein of the I-domain from the human leukocyte β₂-integrin Mac-1 (SEQ ID NO: I), fragment, variant, analog or chemical derivative thereof of this invention with pharmaceuticaUy acceptable carrier, pharmaceuticaUy acceptable adjuvants or excipients employing standard and conventional techniques.

Preferably, the pharmaceutical composition is prepared using conventional techniques in unit dosage form containing an anti-inflammatory effective or appropriate amounts of the active ingredient, protein, that is, I-domain from the human leukocyte β₂-integrin Mac-1 (SEQ ID NO: I), fragment, variant, analog or chemical derivative thereof.

The quantity of active component, according to this invention, in the pharmaceutical composition and unit dosage form thereof may be varied or adjusted widely depending upon the particular appUcation, the potency of the particular compound or the desired concentration. In therapeutic use for treating, or combatting inflammation or any of its related symptoms in patients that can be diagnosed with such, the isolated and purified recombinant protein or pharmaceutical compositions thereof will be administered oraUy and/or parenteraUy at a dosage to obtain and maintain a concentration, that is, an amount, or blood- level of active component in the patient undergoing treatment which wiU be anti- inflammatoriaUy effective. GeneraUy, such pharmaceuticaUy effective amount of dosage of active component will be evidenced by monitoring of the inflammatory site. It is to be understood that the dosages may vary depending upon the requirements of the patient, the severity of the inflammation being treated, and the particular component being used. Also, it is to be understood that the initial dosage administered may be increased beyond a predetermined upper level in order to rapidly achieve the desired blood-level or the initial dosage may be smaUer than the optimum and the daily dosage may be progressively increased during the course of treatment depending on the particular situation. If desired, the daily dose may also be divided into multiple doses for administration, e.g., two to four times per day. The compositions of this invention can be administered parenteraUy, i.e., by injection, for example, by intravenous injection or by other parenteral routes of administration. Pharmaceutical compositions for parenteral administration will generaUy contain a pharmaceuticaUy acceptable amount of the recombinant I- domain protein from the human leukocyte β -integrin Mac-1 (SEQ ID NO: I), fragment, variant, analog or chemical derivative thereof mixed in a pharmaceuticaUy acceptable Uquid carrier such as, for example, water-for-injection and a buffer to provide a suitably buffered isotonic solution.

The active component wiU be admixed in the carrier in an amount sufficient to provide a pharmaceuticaUy acceptable injectable concentration. The resulting Uquid pharmaceutical composition wiU be administered so as to obtain the above- mentioned anti-inflammatory effective amount of dosage.

The fusion protein construct of the subject invention utiUzes the GST partner to faciUtate high levels of expression of soluble protein, and to aUow purification by affinity chromatography on immobilized glutathione (GSH). It was also designed to contain an accessible Factor X^a-sensitive site in the region Unking GST to the I- domain so as to permit removal of the N-teπninaUy attached GST moiety. Affinity chromatography of E. coli extracts over a column of GSH-Sepharose aUowed separation of the bound GST/I-domain from contaminating proteins which passed directly through the column. Removal of the GST/I-domain was effected by washing the column with buffer containing GSH. The fusion protein thus purified was hydrolyzed with Factor X^a which cleaved specificaUy in the linker region, and the resultant GST and I-domain proteins were separated by ion-exchange chromatography on a column of S-Sepharose. The final I-domain product was shown to be >99% pure by sequence and compositional analysis, and by electrospray ionization mass spectroscopy which gave a molecular weight in agreement with that expected for the I-domain (25,767). The fact that the I-domain was resistant to trypsin, coupled with physical characterization by circular dichroism and two- dimensional NMR provided evidence that the recombinant I-domain has a folded and ordered three dimensional structure.

Because of the enormous size and complexity of the β₂-integrins, they are not amenable to structural analysis by conventional methods. The I-domain, however, represents a reasonable target; it is relatively smaU in size, has no disulfide cross links, only low-level of glycosylation and, most importantly, it appears to be functionaUy relevant. In preparation for preparing the subject I-domain product, the I-domain of Mac- 1 in E. coli was cloned and expressed. Two I-domain constructs were designed. One had an extension at the COOH-terminus of six histidines (6His tag) to faciUtate purification by immobiUzed metal ion affinity chromatography. This construct begins with H₂N-Ser₁₃₃-Asp-Ile-Ala-Phe-Leu..., and ends with ..Ser- Gln-Glu₃₃₇-Ile-Leu-Gly-Arg-l-His-His-His-His-His-His-COOH, where the numbers indicate residues in the Mac-1 α-subunit (SEQ ID NO: I, Fig. 1), and the arrow indicates a bond designed, by virture of the preceeding 4 amino acids, to be a cleavage point for Factor X^a. Throughout this description, the numbering system will be that of the processed Mac-1, lacking the signal peptide, a nomenclature consistent with that of Michishita et al, CeU, 72:857-67 (1993). These findings helped promote interest in the recombinant I-domain. Another construct was made which contained GST as a fusion protein. The general layout of this molecule is as shown below:

GST-Leu-Val-Pro-Arg -Gly-Ser-Ser₁₃₃ Glu₃₃₇-Ile-Glu-Gly-Arg-T-(His)₆ where i indicates a thrombin cleavage sequence, and T indicates a Factor X^a cleavage sequence. Accordingly, the I-domain is defined, as for the 6His construct, as the sequence from Ser₁₃₃ and Glu₃₃₇. This construct, therefore, had two handles for purification, GST and the 6His tag. Unfortunately, this protein could not be cleaved with thrombin, Factor X^a, or even trypsin. It appeared that the thrombin site was Umited in accessibiUty, due to hindrance in the inter-domain region.

To solve this problem, a new set of constructs was designed in which the 6His tag was removed, and where additional amino acids were inserted in the Unker region connecting GST and the I-domain, thus:

Thrombin construct: GST-Leu-Val-Pro-Arg4-Gly-Ser-Gly-Gly-Ser₁₃₃....Glu₃₃₇- COOH Factor X^a construct: GST-Ile-Glu-Gly-Arg-T-Gly-πe-Pro-Gly-Gly-Ser₁₃₃...Glu₃₃₇- COOH

Again, arrows indicate sites of processing by the two proteases and therefore the constructs had N-terminal extensions highUghted in boldface, but had no COOH- terminal extensions. Both fusion proteins were found to be processed by thrombin and Factor X^a, respectively, but the latter proved to be much better in its yield of I- domain. Cloning and Expression of Two Fusion Proteins of CDllb I-domain to GST

Two constructs were made for the expression of the CDllb I-domain as a fusion to GST in the cytoplasm of E. coli. One construct has the recognition site for Factor X^a cleavage at the fusion junction, while the other construct has the thrombin proteolysis site. After purification of the fusion proteins, the I-domain can be recovered from the fusion protein by Factor X^a or thrombin processing. In order to separate the I-domain from the GST structure to aUow accessibility of the proteolytic site, two glycine residues were introduced into the fusion segment. Due to the proteolytic recognition sequence it was necessary to introduce other residues in addition to the two glycine residues. After proteolytic processing, the CDllb I- domain has the N-terminus extension shown in boldface. The CDllb I-domain contains amino acid residues from Serl33 to Glu337. Factor X^a cleavage product: Gly-D.e-Pro-Gly-Gly-(Serl33....Glu337) Thrombin cleavage product: Gly-Ser-Gly-Gly-(Serl33....Glu337)

Two DNA sequences coding for the CDllb I-domain were cloned by PCR from plasmid pET-CDllb/I(His)₆, as described in the EXAMPLE. These BamHI and EcoRI digested 660 bp DNA fragments were cloned into the Pharmacia expression vectors pGEX-3X and pGEX-2T which were digested with BamHI and EcoRI. The 660 bp fragment from PCR primers KAC250 and KAC251 was cloned into pGEX-3X and the one from primers KAC249 and KAC251 was cloned into pGEX-2T. The resulting expression plasmid vectors carrying the I-domain fused to GST are named pGST-X^a- CDllb/I and pGST-Throm-CDllb/I, with the Factor X^a cleavage site and thrombin process site, respectively. In the expression plasmids, the fusion protein is under the control of the strong tac promoter which can be induced by IPTG. The plasmids also carry the lacfl sequence to repress the tac promoter activity in the absence of an inducer such as IPTG. The presence of the repressor lacr¹ is important for minimizing background expression of the recombinant protein, especiaUy in cases where the recombinant protein is detrimental to the host ceU. If the lacfi repressor sequence is not present in the plasmid, then one would want to maintain the plasmid in a E. coli host carrying the repressor, usually on the F' episome. Since the above vectors carry lacfl, the E. coli DH1 used for plasmid construction and E. coli K12S used for expression do not contain the repressor sequence. To test the expression of the fusion protein from the above expression vectors, pGST-X^a-CDllb/I and pGST-Throm-CDllb/I were transformed into the 3 E. coli hosts JM103, JM109 and K12S. These strains were induced for expression of the fusion protein by lxlO^"3M IPTG as described in the EXAMPLE. When ceU extracts were analyzed by SDS-PAGE, a prominent protein band corresponding to the expected size of the fusion protein (about 45 kd) was observed. This protein band is absent in ceU extracts derived from strains expressing GST without the I-domain fusion. As expected, GST is expressed at high levels in these control extracts. The expression of the fusion protein in these 3 hosts with induction by 1x10 M IPTG was repeated more than once. It appears that JM109 and K12S are better hosts than JM103, producing high levels of the fusion protein, leading to the accumulation of fusion protein as 30% of total ceU protein.

By inducing expression with lxlO^"3M IPTG, the majority of the fusion product accumulates in an insoluble, aggregated form, i.e., inclusion bodies. In some cases a recombinant protein can remain in soluble form if the promoter is turned on slowly, aUowing protein synthesis which might lead to a more favorable condition for the folding of the recombinant protein into soluble state. The distribution of the GST I-domain fusion protein in soluble and insoluble forms under promoter induction at various levels of IPTG was examined. As the IPTG level is lowered from lxlO^"3 M, lxlO^"4 M, 5xl0^"5 M to lxlO^"5 M, more of the fusion product accumulates in the soluble state. At lxlO^"5 M IPTG, about 60-80% of the fusion protein produced are in the soluble state.

Using the E. coli strain K12S(pGST-X^a-CDllb/I) which produces the I- domain fusion to GST with the Factor X^a cleavage site, more than 25 Uters of ceUs can be grown in shake-flasks with lxlO^"5M IPTG to induce expression. From these ceUs, ceU extracts are prepared by sonication and supernatant extracts free of inclusion bodies can be obtained by centrifugation to remove the insoluble material. These supernatant extracts can be purified for the I-domain. Over 10 mg of purified I-domain can be obtained from 1 Uter of ceUs. This amount of production is significant, since shake-flasks are used to grow E. coli cultures, which reach a ceU density of only about 1.5 A550.

In order to label the fusion protein with N and N/ C for isolation of labeled I-domain for NMR analysis the ceU growth conditions were tested to determine what would efficiantly but economically produce the labeled protein. PUot experiments involved using the Celtone-N medium (without N or C, see

EXAMPLE). The unlabeled Celtone-N medium was dUuted with unlabeled M9 salts medium at a ratio of 1:1, 1:2 and 1:3 and used to culture strain K12S(pGST-X^a- CDllb I) with lxlO^"5M IPTG for induction of expression. Celtone medium itself appears to give lower expression levels (24% of total ceU protein) than the Luria Broth (30% of total ceU protein). Dilution of the Celtone medium with 1 part or 2 parts of M9 salts medium does not significantly decrease expression, while dilution with 3 parts of M9 did lower expression. Under these culture conditions, about 60% of the fusion product remains in the soluble state.

Using the ¹⁵N- and ¹⁵N/¹³C-containing Celtone medium dfluted with 2 parts of ¹⁵N-and ¹⁵N/¹³C-containing M9 salts medium (see, EXAMPLE), several Uters of the K12S(pGST-X^a-CDllb I) ceUs were grown (the strain expressing I-domain fusion to GST with Factor X^a cleavage site) and prepared supernatant extracts free of inclusion bodies. The ⁵N- and N/ C-labeled I-domain molecules were isolated for NMR analysis.

The recombinant I-domain of the α-subunit CDllb of human leukocyte integrin Mac-1 was produced in E. coli as a fusion to GST (glutathione S- transferase), under the control of the tac promoter in a pBR-based vector background. High level expression leads to the accumulation of the fusion product to at least 30% of total ceU protein. Induction of the tac promoter with high levels of IPTG results in the majority of the fusion product in insoluble, aggregated inclusion body form. By lowering the IPTG level to lxlO^"5M, over 60% of the fusion protein produced remains in a soluble form in the cytoplasm of E. coli.

Two fusion proteins were produced, each with a specific proteolytic cleavage site (Factor X^a or thrombin) at the fusion junction to aUow the recovery of the I- domain from the fusion product by digestion with either Factor X^a or thrombin. To assure that the I-domain and GST are not in too close physical proximity to mask the accessibiUty of the cleavage site to the protease, glycine residues were introduced into the fusion junction to separate the two polypeptide structures. After cleavage, the resulting I-domain has 4-5 extra residues at the N-terminus. To provide starting material for purification of the I-domain, multi-Uters of E. coli ceUs were cultured in shake-flasks, and inclusion body-free ceU extracts prepared.

EXAMPLE

Enzvmes and Basic Molecular Bioloyv Techniques

Restriction endonucleases, other DNA modifying enzymes and T4 DNA Ugase were from New England Biolabs or Boehringer Mannheim. AU enzymes were used according to the manufacturer's instructions. AU plasmids used carry the ampicillin-resistance marker and were constructed and maintained in E. coli DH1 in the presence of 100 μg/ml ampicilUn in either Luria broth or on Antibiotic Medium #2 (Difco) agar plates. Isolation of DNA fragments, transformation, smaU and large scale plasmid preparation, and other basic molecular biology techniques were according to those described by

Sambrook et al , Molecular Cloning: A Laboratory Manual, 2nd Edition, Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press (1989). Verification of the plasmid constructs was by restriction enzyme analysis and DNA sequencing using the Sequenase kit from United States Biochemical Corporation. Construction of Vectors for Expression of CDllb I-domain Fused to GST

The vectors, pGEX-3X and pGEX-2T, used for constructing plasmids with the CDllb I-domain sequence fused to GST, were purchased from Pharmacia. Plasmid pGEX-3X carries the Factor X^a cleavage recognition site and pGEX-2T the thrombin cleavage site in the polylinker region downstream from the GST sequence. These vectors contain the tac promoter [inducible by IPTG (isopropylthiogalactoside)] to control GST expression, the lacjfl repressor sequence to minimize the tac promoter activity in the absense of IPTG, a polyUnker region for gene cloning as a fusion to GST, the ampicillin-resistance marker, and the repUcation origin derived from pBR322. The CDllb I-domain sequence was cloned from a plasmid containing human

CDllb cDNA which was originaUy constructed by Corbi et al, J. Biol. Chem., 263:12403-11 (1988). With the pET-CDllb/I(His)₆ as the PCR template, oUgonucleotides named KAC250 and KAC251 were used to clone the I-domain sequence for fusion with GST in pGEX-3X, whUe KAC249 and KAC251 were used to clone the I-domain for fusion with GST in pGEX-2T. The sequences of these oUgonucleotide primers are shown below. The underUned portions of the primers overlap with the I-domain sequence (see, ID SEQ NO: I). The other portions in KAC249 and KAC250 provide the N-terminus extension sequence for the I-domain and the restriction enzyme BamHI site for cloning of the I-domain into the vector pGEX-3X or pGEX-2T. The unUned portion in KAC251 provides a protein synthesis stop codon TAG to terminate the I-domain sequence at glu337 and the restriction enzyme EcoRI site for cloning into the vectors.

KAC249: 5' GGCTCGGATCCGGTGGCAGTGACATTGCCTTCTTGATTGATGGCTCT KAC250: 5' GGCTCGGGATCCCCGGTGGCAGTGACATTGCCTTCTTGATTGATGGC

KAC251: 5' GAGCCTGAATTCTATTCCTGAGACATCTCATGCTCAAAGGAGCT The PCR reaction mixture, in a final volume of 100 ul, was composed of 68 ul water, 10 ul lOx reaction buffer (100 mM Tris-HCl, pH 8.3, 500 mM KC1, 15 mM MgCl₂, 0.01% gelatin), 16 ul of a mixture of the four deoxynucleotide triphosphates (each at 1.25 mM), 1 uM each of the appropriate oUgonucleotide primer, 1 ng pET- CDllb/I(His)₆ plasmid DNA, and 1 ul of Taq polymerase. The reaction was carried out by appUcation of the GeneAmp PCR system (Cetus) in the foUowing manner: The sample was heated at 94°C , then treated for 30 cycles of 94 C for 30 seconds, 65 C for 30 seconds, and 72°C for 45 seconds, foUowed by a 5 minute anneal at 65 C and a 5 minute extension at 72°C, and then cooled to 4°C. An aUquot of the PCR reaction mixture was analyzed by 1.5% agarose gel electrophoresis to confirm that the size of the DNA fragment is about 660 bp. The DNA fragment was then purified by phenol extraction and precipitated from the PCR reaction mixture by the addition of 100 ul of 0.3 M sodium acetate and 400 ul ethanol. This 660 bp DNA fragment was treated with BamHI and EcoRI, purified by 1.5% agarose gel electrophoresis, purified by phenol extraction and ethanol precipitation, and stored at -20°C in 10 mM Tris-HCl, pH 7.4 and 1 mM EDTA until use for cloning into expression vectors. Expression of the CDllb I-domain Fusion to GST

Expression vectors pGST-X^a-CDllb/I and pGST-Throm-CDllb/I were transformed into 3 different E. coli hosts (JM103 and JM109 are commerciaUy avaflable; K12S was isolated to test for expression of the fusion protein). Induction of the tac promoter to express the fusion protein was by the addition of IPTG (isopropylthiogalactoside) at lxlO^"3 M or as specified. The E. coli ceUs were grown at 37°C with aeration in growth medium (Luria Broth with 100 μg/ml ampiciUin). SpecificaUy, an overnight culture of E. coli ceUs was diluted 50-100 fold with growth medium to A550 of about 0.1, aUowed to grow until mid-log phase (A550 0.6-0.8), and then induced by the addition of IPTG. At the end of induction (2-4 hours) ceUs were coUected by centrifugation, resupended to A550 of 20 with 10 mM Tris-HCl, pH 7.4 and 1 mM EDTA, and stored at -20 C until further use. Sonicated ceU extracts were prepared by sonication of the ceU suspension in a Branson Sonifier.

To determine the level of expression, sonicated ceU extracts were analyzed by SDS-PAGE and the gel with the Coomassie stained protein bands was scanned by a Shimadzu densitometer. The amount of the fusion protein in the ceU extract was calculated as a percent of total ceU protein. SDS-PAGE was carried out using a gel system where the cross Unker is N,N'-diaUytartar diamide (Morse et al, Anatomy of herpes simplex virus DNA, J. Virol., 26:389-410 (1978) at a polyacrylamide concentration of 17%. Protein molecular weight standards were purchased from

Amersham.

Preparation of Inclusion Bodv-Free Extracts

The soluble fusion protein found in the cytoplasm of the E. coli ceUs was used as the starting material for downstream purification of the I-domain. To remove the insoluble, aggregated form of the fusion protein (in the inclusion bodies), sonicated ceU extracts were centrifuged in the SorvaU RC-5B centrifuge with the SS34 rotor at 13,000xg for 5 minutes at 4°C. The inclusion body-free supernatant was stored at - 20°C until further use. Preparation of Medium for ^lδN and ¹⁵N/¹³C Labeling of I-domain

E. coli ceUs to produce N- and ¹⁵N¹³C-labeled I-domain, were cultured in medium purchased form Martek Bioscience Corp., Columbia, Maryland. The medium was dfluted with M9 salts medium, Sambrook et al. (cited, above). To test the approriate dilution ratio, the Martek non-labeled medium Celtone-U was used with M9 salts medium without supplement of amino acids. AmpicilUn at 100 μg/ml was used and ceU growth was at 37°C. To label the proteins with N, ceUs were grown in Martek Celtone-N medium (>98% N) dfluted with 2 parts of M9 salts medium in which 2 gm/1 of N-ammonium chloride was used as the sole ammonium source. To label the proteins with ¹⁵N and C, ceUs were grown in Martek Celtone-CN medium (>98% ¹³C and ¹⁵N) dUuted with 2 parts of M9 salts medium in which 2 gm/1 of N-ammonium chloride and 4 gm/1 of C-glucose were used as the sole ammonium and carbon source. Methods: Expression of GST-I-domain

The subject I-domain, generated from this fusion protein by cleavage with Factor X^a, corresponds exactly to residues 133 through 337 of mature Mac-1, plus a pentapeptide N-terminal extension, Gly-Ile-Pro-Gly-Gly, required to enable the proteolytic cleavage. A schematic comparison of the I-domain structure with that pubUshed by Michishita et al. (cited, above) for the A-domain, is given in Figure 1. The subject protein begins and ends about 20-residues downstream from the A- domain (Fig. 1). This is a significant difference, and can account for differences which might be seen in the function of these proteins. Moreover, the subject I- domain has an N-terminal extension, whereas the A-domain has a COOH-terminal segment, neither of which belong to the protein in question. Purification of I-domain: Glutathione Sepharose Chromatography Glutathione-Sepharose and S-Sepharose Fast Flow were purchased from

Pharmacia LKB. Reduced glutathione (GSH) was purchased from Sigma. Sequencing-grade Factor X^a was obtained from Boehringer Mannheim. Ultrafiltration membranes were from Amicon and polyacrylamide gels were purchased from ISS Enprotech. AU other reagents were of the highest quaUty commerciaUy available. E. coli ceU homogenates were centrifuged at 40,000 x g and subsequently filtered through a 0.45 um filter. The filtrate was loaded directly onto a Glutathione Sepharose column (50-ml; 1.6cm i.d. x 26cm length) pre-equiUbrated in PBS/0.2% β-octyl glucoside, pH 7.4. The column was loaded at 1.0 ml/min and protein in the effluent was monitored at 280 nm. The column was washed with 4 column volumes (CV) of equiUbration buffer and step-eluted with 2 CV of 50 mM Tris buffer, pH 8.0 containing 20 mM reduced glutathione. The eluent was coUected in 9-ml fractions. AU process fractions were subjected to SDS-PAGE under non- reducing conditions (NR). Fractions containing GST/I-domain were pooled and stored at 4 °C. Factor X^a Cleavage Purified fusion protein was subjected to digestion with Factor X^a (1% by weight) for 16-24 hours at 23°C. SDS-PAGE was always run to confirm that digestion was complete. S-Sepharose Fast Flow Chromatography

The protein digest was loaded directly onto an S-Sepharose Fast Flow column (50-ml; 1.6cm i.d. x 26cm length) pre-equiUbrated with 50 mM sodium phosphate buffer, pH 6.5. The column was run at 1.0 ml/min and the effluent monitored at 280 nm. The column was washed with 4 CV of equiUbration buffer, and eluted with a 2 CV linear salt gradient run from 0-1.0 M NaCl in the presence of 50 mM sodium phosphate, pH 6.5. Column fractions were subjected to SDS-PAGE (NR). Purified I- domain was observed as a single band at approximately 25 kD, and eluted from the column at approximately 150 mM NaCl. Ultrafiltration

Purified I-domain was concentrated to 20 mg/ml using an Amicon stirred-ceU ultrafiltration module containing a YM05 membrane. Nitrogen pressure was maintained at 60 psi during the concentration. Characterization of I-domain: Analytical Methods

Proteins were sequenced by automated Edman degradation in an AppUed Biosystems Model 470 Gas Phase Sequencer fitted with an on-Une HPLC analyzer (Model 120A) for identification and quantitation of phenylthiohydantoin (PTH) amino acids. Integration of the peaks from the HPLC was performed with a Nelson Analytical 3000 Series chromatography data system connected in paraUel with the recorder to the output of the Model 120A HPLC system.

Amino acid analysis was performed with the aid of a Beckman Model 6300 analyzer. Samples were hydrolyzed for 24 hours in vacuo in 6N HCl at 110 °C, foUowed by drying in a Speed Vac Concentrator (Savant). Resulting residues were reconstituted in buffer (NaS; Beckman) and appUed to the analyzer.

Reversed-Phase (RP) HPLC was performed on an HP 1090 Uquid chromatograph with a 4.6 x 250 mm Vydac C4 column. Gradient mobfle phases were water and acetonitrile, each containing 0.15% trifluoroacetic acid (TFA). The gradient was run from 0-70% acetonitrile in 70 minutes with a flow rate of 1.0 ml/min. The effluent was monitored simultaneously at 220 nm and 280 nm. The I- domain eluted as a single peak, and was coUected in 2-ml polypropylene microfuge tubes, dried in a Speed Vac, and analyzed. SDS-PAGE was run as described by LaemmU, Nature, 277:680-85 (1970), or using the tricine buffer system as described by Schagger and von Jagow, Anal. BioChem., 166:368-79 (1987). Gels were fixed and subsequently stained with 0.1% Coomassie BriUiant Blue R-250.

AU electrospray ionization (ESI) mass spectra were recorded on the Vestec 201A mass spectrometer. AUquots (5 ul) of a solution of the I-domain in acetonitrile/water/0.1% TFA (about 0.05 to 0.5 ug of protein isolated by HPLC in the acetonitrile TFA gradient system) were injected via a loop injector into the ion source. The mass spectrometer was scanned from m/z 500 to 2000 at 2 sec/scan. The data were acquired with the Teknivent Vector 2 data system. Ten scans were averaged and transfered to the Harris 800 computer for further processing. The average molecular weights (av. M.W.) were determined using programs developed in-house. For single components, the centroid program and the deconvolution program were used. For mixtures, the deconvolution program yielded better results. The experimentaUy obtained average M.W. were then compared with the theoretical av. M.W. of the various samples of I-domain.

The circular dichroism (CD) spectrum of I-domain ( 1.0 mg/ml) was measured at room temperature (20-22 °C) on a Jasco Model J-720 CD spectropolarimeter from 260-190 nm in a 0.086 mm ceU. The spectropolarimeter was caUbrated at 290 nm with D-10-camphorsulfonic acid. Molar intensities were computed from the concentration of the protein sample and a mean residue molecular weight of 113.2. The secondary structure was calculated using the method of Compton and Johnson against a data base of 16 proteins. Purification of I-domain

The E. coli expression system with plasmid pG-3x-CDllb provided reasonable levels of GST/I-domain fusion protein. Purification of the GST/I-domain construct over immobilized GSH proved to be very straightforward, as good yields were observed with extremely high purities. In fact, the only contaminant observed upon SDS-PAGE analysis appeared to be a low level (<5%) of free GST, which would be expected to co-purify with the fusion protein. RP-HPLC of the protein recovered from the GSH-column yielded a single peak which, upon N-terminal sequence analysis yielded a single sequence corresponding to the N-terminus of GST. The absence of secondary sequences was good evidence that there was Uttle or no proteolytic degradation of the fusion protein occurring during the first process step. Amino acid analysis of the RP-HPLC peak shows exceUent correlation with the expected composition of intact GST/T-domain fusion protein. Based on these analytical results, the purity of the intact fusion protein was estimated to be >95%.

Liberation of the I-domain from the fusion protein was achieved enzymaticaUy via incubation with Factor X^a. Factor X^a is a serine protease which, by design of the construct, should specificaUy cleave the fusion protein at the Arg- Gly bond leading into the Gly-Ile-Pro-Gly-Gly- N-terminal extension of the I-domain (SEQ. ID. NO. I , Fig. 1). Factor X^a was added directly to the purified fusion protein to a final concentration of 1% (by weight), and incubation was carried out at room temperature. An SDS-PAGE profile of the enzymatic timecourse of digestion with Factor X^a shows that digestion of the fusion protein is complete between 16-24 hours. Since GST and I-domain are nearly identical in molecular weight, these protein products are not resolved by this method. Nevertheless, SDS-PAGE provided a useful technique for monitoring the extent of digestion, since the loss of fusion protein could be foUowed easUy. Because the isoelectric points of GST (pl=6.5) and Factor X^a (pl=4.3) are significantly different from one another, and from that of the intact I-domain (pl=9.0), ion exchange chromatography was the final step in the purification of I- domain. The Factor X^a digest was loaded directly onto the cation exchange resin S- Sepharose Fast Flow under conditions that would aUow only I-domain to bind. The I-domain eluted as a single homogeneous peak at around 150 mM NaCl in the gradient, resulting in its recovery in physiologic-type buffer. Efficiency of binding was monitored by RP-HPLC, which resolves GST from I-domain. The S-Sepharose load showed two peaks upon RP-HPLC (corresponding to GST and I-domain), whUe the flowthrough, wash, and pooled fractions each contained predominantly one peak corresponding to GST, GST, and I-domain, respectively. Characterization of I-domain

N-terminal sequence analysis and amino acid analysis performed on the final product resulted in very high correlations to expected results with no hint of contamination or degradation. Upon overloading a sample of this pool for SDS- PAGE, two minor bands are observed (=9kD and =15kD). RP-HPLC analysis also shows the presence of two early-eluting peaks at low levels (2-3%). CoUection, lyophilization, and SDS-PAGE analysis of these peaks confirms that they are the =»9kD and «15kD species. N-terminal sequencing and amino acid analysis of both peaks shows that they are degradation products of the I-domain resulting from a single internal cleavage at Arg residue 306 (SEQ ID NO: I). Based on these results, the purity of I-domain derived from this process has been determined to be >95%. Material of this high quaUty has been obtained consistently by means of this purification protocol (n=7).

Electrospray ionization mass spectrometry was used to further assess the integrity of the I-domain preparations. Purified I-domain (2-3 nmoles) was coUected from RP-HPLC, lyophilized, and reconstituted in a smaU volume of 50% acetonitrile prior to appUcation to the spectrometer probe. A series of molecular ions corresponding to mass per unit charge (m/z) was obtained, from which the average molecular weight can be determined. AU Factor X^a-derived I-domain preparations yielded essentiaUy identical results. The spectra identified two species with distinct molecular weights. The molecular weight of the major species (80-95%) was determined to be 23,767, which matches the theoretical value exactly, whUe the minor species yielded a mass 164 daltons lower. Generation of this species does not appear to be due to C-terminal heterogeneity or typical post-translational modification based on this mass difference.

The monoclonal antibody (mAb) 3H5 directed against CDllb was generated, isolated and characterized. Fibrinogen was purchased from Kabi Pharmacia (FrankUn, Ohio) and was fibronectin depleted utiUzing gelatin-Sepharose 4B (Pharmacia; Uppsala, Sweden). CeU Preparation

Polymorphonuclear leukocytes (PMNs) were isolated from heparinized blood by dextran-sedimentation and FicoU-Hypaque gradient centrifugation as previously described (Smith, C.W., et al., J. CUn. Invest., 83:2008-17 (1989)). PMNs were washed 2 times in phosphate-buffered saline (1.2 mM phosphate, 138 mM NaCl, pH 7.4; PBS) containing 1 mM MgCl₂ and 1 mM CaCl₂. PMNS were labeled with 2', 7'-6is-(2~carboxyethyl)-5 (and -6)-carboxyfluorescein (BCECF; Molecular Probes, Inc., Eugene, OR; 1 mg/ml in 90% DMSO). Cell Adhesion Assays:

Neutronhil Adhesion to Protein Substrates in Microtiter Plates (Tables 1. 2. & 3) To evaluate neutrophil adhesion to iC3b Ugands, 96 weU microtiter plates (Immulon 2, Dynatech Labs, ChantiUy, VA) were coated with 200 ul of PBS containing 170 μg/ml BSA and 75 μg/ml human polyclonal IgG (Calbiochem, La JoUa, CA). FoUowing incubation for 1 hour at 37°C, plates were washed three times with 200 ul PBS. The remaining sites on the plastic were blocked by addition of 1% gelatin in PBS. Plates were washed once with 200 ul of PBS containing 1 mM CaCl₂ and 1 mM MgCl₂ (PBS/Ca²⁺/Mg²⁺) foUowed by the addition of 200 ul of human serum (dfluted 1:3 in PBS/Ca²⁺/Mg²⁺. Plates were washed three times with 200 ul of PBS containing 0.05% Tween 20 and 0.01% thimerosal.

Labeled neutrophUs were stimulated with n-formyl-methione-leucine- phenylalanine (fMLP; 10^'7 M) for 10 min at 37°C. Fifty ul of the anti CDllb mAb 3H5 (positive control) (20 μg/ml) or I-domain protein (0.5, 1, 5, 10, or 20 uM) was added to some of the weUs before addition of 50 ul of neutrophUs (10⁷/ml). FoUowing 20 minutes incubation at 37°C, nonadherent ceUs were removed by flicking the plates. Plates were then washed three times with 200 ul PBS/Ca ⁺/Mg²⁺. The amount of fluorescence in the weUs was measured at 485/535 nm using a Pandex fluorescence concentration analyzer (Baxter Healthcare Corp., Mundelein, IL).

To evaluate neutrophU adhesion to fibrinogen, 96 weU microtiter plates were coated with 50 ul of fibrinogen (100 μg/ml) for 2 hours at 37°C. The remaining sites on the plastic were blocked by addition of 1% gelatin in PBS. Plates were washed three times with PBS/Ca²⁺/Mg2⁺. Fifty μl of mAb 3H5 (20 μg/ml) or I-domain protein (0.5, 1, 5, 10, or 20 μM) was added to some of the weUs before addition of ceUs. fMLP stimulated labeled neutrophUs were added to the weUs and incubated for 20 minutes at 37°C, and ceU adherence was determined as described above. In order to evaluate neutrophU adherence to ICAM-1, this protein was immunopurified from 70 grams of human placenta using a modification of a procedure as described (Mariin and Springer, Purified interceUular adhesion molecule-l(ICAM-l), CeU., 51:813-19 (1987)). The anti ICAM-1 mAb 8.4 (*) was used for these affinity chromatography procedures. For neutrophU adhesion assays, microtiter plates were coated with purified ICAM-1 dfluted 1:15 (vol./vol.) in PBS/Ca²⁺/Mg²⁺. After a 2 hour incubation at 37°C, ICAM-1 substrates were blocked by addition of 1% gelatin in PBS. Plates were washed x 3 with PBS/Ca²⁺/Mg²⁺. Fifty μl of mAb 3H5 (20 μl/ml) of I-domain protein (0.5, 1, 5, 10 or 20 μM) was added to some weUs before addition of ceUs. fMLP stimulated labeled neutrophUs were added to weUs and incubated for 20 minutes at 37°C and ceU adherence was determined as described above. NeutrophU adhesion to protein substrates in Smith-HoUers adhesion chambers

(Table 4)

In addition to the microtiter plate assays (above), adhesion of chemotacticaUy stimulated neutrophUs to Mac-1 substrates was assayed in Smith-HoUers chambers as previously described (Anderson et al., Abnormal mobiUty of neonatal polymorphonucler leukocytes, J. CUn. Invest. 68:863-74 (1981)). Use of these adhesion chambers faciUtates analysis of adhesion in the absence of shear stress, conditions which enhance β₂ integrin-dependence (Lawrence et al., Effect of venous shear stress on CD- 18..., Blood, 75:227-37 (1989)). NeutrophUs (10⁷/ml) were preincubated in suspensions containing fMLP (10 nM) and I-domain (2.5 - 25 μM) or in the CDllb mAb 3H5 (10 μg/ml) for 5 minutes at 21°C prior to incorporation in adhesion chambers (0.5 x 10 ). For preparation of protein substrates, KLH (1 mg/ml) or fibrinogen (10 mg/ml) were incubated on glass coversUps for 30 min at 37°C and washed prior to assay. The iC3b substrates were prepared by incubating a human polyclonal IgG (75 μg/ml) - BSA (170 μg/ml) mixture for 1 hour at 37°C on glass coversUps prior to adding human serum (dUutes 1:3 in PBS) and incubating for an additional 1 hour at 37°C. The percentages of adherent neutrophUs contacting protein substrates were calculated as described (Anderson et αZ,(1981), cited above). Neutrophil-endotheUal adhesion and migration assays (Table 5. 6)

Assessments of neutrophU adhesion to and migration through confluent monolayers of human umbiUcal vein endotheUal ceUs (HUVEC) were performed in Smith-HoUers adhesion chambers as described (Smith et al. , Cooperative Interactions of LFA-1 and Mac-1 with interceUular adhesion molecule- 1..., J. CUn. Invest. 83:2008-17 (1989)). For the representative experiments lustrated in Tables 4 and 5, HUVEC were preincubated with TNF (100 U/ml, 4 h, 37°C) to induce high levels of ICAM-1 expression prior to incorporation into chambers. NeutrophUs

(10 /ml) were preincubated with selected concentrations of CDllb I-domain (Factor X_a-cleaved GST fusion protein), GST, or the anti CDllb mAb 3H5 (10 μg/ml) for 5 min at 37°C prior to incorporation in adhesion assays (0.7 ml/chamber). The percentage of adherent neutrophUs contacting HUVEC and the percentage of ceUs exhibiting transendotheUal migration were determined as previously described (Smith, C.W. et al., J. CUn. Invest. 82:1746-1756, (1988). ICAM-1 Purification

ICAM-1 was purified from 70 grams of human placenta by a modified procedure described (Marlin, S., Springer, T.A., CeU, 51:813-19 (1987)). The anti- ICAM-1 monoclonal antibody 8.4 was used for the affinity chromatography purification.

TABLE 1 Adhesion of NeutrophUs to iC3b

I-Domain Cone. (μM) 6-His I-domain Factor X 6-His peptide

20 627 35484 38500

10 730 37561 41216

5 762 32640 35450

1 15027 28402 36387

.5 24456 32177 40912

0 27322 N.D. 36876

3H5 monoclonal Ab.

8570

Table 1. Adhesion of neutrophUs to iC3b.

Fluorescently labeled neutrophUs and I-domain constructs (at the indicated concentrations) were added to microtiter weUs that had been coated with iC3b as described under "Materials and Methods". FoUowing a 20 minute incubation, nonadherent neutrophUs were removed by flicking and the microtiter weUs were washed three times with assay buffer. Adherent ceUs were quantitated by measuring the fluorescent intensity using a Pandex fluorescence concentration analyzer. The adhesion of neutrophUs was also determined in the presence of mAb 3H5 (anti-CD lib). Each value (expressed as fluorescence intensity) represents the mean ± standard error of three determinations. Results are representative of three experiments. The 6-His I-Domain is an extension at the COOH-terminus of six histidines (6His tag) to facUitate purification by immobiUzed metal ion affinity chromatography. This construct begins with H N-Ser₁₃₃-Asp-Ile-Ala-Phe-Leu..., and ends with ..Ser-Gln-Glu₃₃₇-Ile-Leu-Gly-Arg--l-His-His-His-His-His-His-COOH, where the numbers indicate residues in the Mac-1 α-subunit (from, Michishita et al, CeU, 72:857-67 (1993), correspondes to Ser 231-Glu 435, SEQ ID NO: 1), and the arrow indicates a bond designed, by virture of the proceeding 4 amino acids, to be a cleavage point for Factor X . The 6-His peptide is a control for the 6-His I-Domain starting Ser-Gbi-Glu-Ile-Leu-Gly-Arg-i-His-His-His-His-His-His-COOH to assure that the observed activity of the 6-His I-Domain is not an artifact. These results indicate that the 6-His-I-Domain protein inhibits neutrophU adhesion to iC3b in a dose-dependent manner. Half maximal inhibition was observed at a concentration of 1 micromolor. This inhibition does not appear to be due to the 6-His tag itself, since the control peptide did not inhibit adhesion. In contrast, the Factor X (never frozen) cleaved I-Domain protein did not inhibit neutrophU adhesion to iC3b even at the highest concentration tested (20 micromolar). Attachment was specific, since it was inhibited by mAb 3H5 (anti- CDllb).

TABLE 2 Adhesion of NeutrophUs to Fibrinogen

I-Domain cone. (μM) 6-His I-Domain Factor X* Factor X** 6-His peptide

20 34917 52056 16790 44616

10 30949 43282 15110 47685

5 38270 54150 19461 51523

1 37002 42670 26798 46818 .

.5 47019 46423 48530 47230

0 40854 41588 49273 38952

3H5 monoclonal Ab. 18906

* previously frozen material **never frozen

Table 2. Adhesion of neutrophUs to fibrinogen.

Fluorescently labeled neutrophUs and I-domain constructs (at the indicated concentrations) were added to microtiter weUs that had been coated with fibrinogen

(100 μl/ml). FoUowing a 20 minute incubation, nonadherent neutrophUs were removed by flicking and the microtiter weUs were washed three times with assay buffer. Adherent ceUs were quantitated by measuring the fluorescent intensity using a Pandex fluorescence concentration analyzer. The adhesion of neutrophUs was also determined in the presence of mAb 3H5 (anti-CDllb). Each value (fluorescence intensity) represents the mean ± standard error of three determinations. Results are representative of three experiments.

The results indicate that the Factor X cleaved GST-I-Domain fusion protein inhibited neutrophU adhesion in a dose-dependent manner such that half maximal inhibition was observed at a concentration of 1 micromolar. In contrast, neutrophU adhesion to fibrinogen was only marginaUy inhibited by the 6-His I-Domain fusion protein. The control peptide had no effect on adhesion and the previously frozen and thawed Factor X was inactive. Attachment was specific, since it was inhibited by mAb 3H5 (anti-CDllb). TABLE 3 Adhesion of NeutrophUs to ICAM-1

I-domain Factor X* Factor X** cone. (μM) 6-His I-domain 6-His peptide

20 8480 22156 11274 23274

10 9758 19187 9846 18008

5 13772 22040 12370 20502

1 18476 20895 14292 19977

.5 19401 19081 15599 16546

0 19977 20587 18258 14894

3H5 monoclonal

Ab.

10465

*previously froz< ϊn material; ** never frozen

Table 3. NeutrophU adhesion to ICAM-1.

Fluorescently labeled neutrophUs and I-domain constructs (at the indicated concentrations) were added to microtiter weUs that had been coated with purified placental ICAM-1. FoUowing a 20-minute incubation, nonadherent neutrophUs were removed by flicking and the microtiter weUs were washed three times with assay buffer. Adherent ceUs were quantitated by measuring the fluorescent intensity using a Pandex fluorescence concentration analyzer. The adhesion of neutrophUs was also determined in the presence of mAb 3H5 (anti-CDllb). Each value (fluorescence intensity) represents the mean ± standard error of three determinations.

The results demonstrate that both the Factor X and the 6-His tag I-Domain fusion proteins inhibited neutrophU adhesion to purified ICAM-1 in a dose- dependent manner. Half maximal inhibition was observed at concentrations of 1 micromolar and 3 micromolar, respectively. The control peptide had no effect on adhesion and the frozen and thawed Factor X was inactive. Attachment was specific since it was inhibited by mAb 3H5 (anti-CDllb).

In summary, Tables 1-3 show that CDllb I-Domain proteins have differential effects on neutrophU adhesion. The Factor X I-Domain inhibits neutrophU adhesion to fibrinogen and ICAM-1 in a dose-dependent manner, but has no inhibitory effect on neutrophU adhesion to iC3b. Half maximal inhibition to fibrinogen and ICAM-1 is observed at a concentration of 1 micromolar for both proteins. In contrast, the 6- His tag I-domain protein blocks neutrophU adhesion to Ugands iC3b and ICAM-1 in a dose-dependent manner such that half maximal inhibition occurs at concentrations of 1 and 3 micromolar, respectively. A control peptide containing hexahistidine had no effect on neutrophU adhesion. These results further support the importance of the I-domain in Ugand binding function of CDllb/CD18.

TABLE 4

Effect of CDllb I-Domain (Factor X_a cleaved GST fusion protein) on neutrophU adhesion to protein substrates

% Inhibition

Neutrophil- Adhesion (%)** by I-domain

Substratet Pretreatment* (n = 2 or mAb

KLH fMLP 10 nM, PBS 81

, I-Domain (25 μM) 60 26

, I-Domain (10 μM) 71 14

, I-Domain (2.5 μM) 61 25

, 3H5 mAb (10 μg/ml) 23 71

iC3b , PBS 97

, I-Domain (25 μM) 98

, I-Domain (2.5 μM) 60 39

, 3H5 mAb (10 μg/ml) 47 51

Fibrinogen PBS 72

, I-Domain (25 μM) 49 32

, I-Domain (10 μM) 44 39

, I-Domain (2.5 μM) 34 53

, 3H5 mAb (10 μg/ml) 21 71 + KLH (1 mg ml) or fibrinogen (10 μg ml) were incubated on glass coverslips for 30 min. at 37°C and washed prior to assay. The iC3b substrates were prepared by incubating an IgG (75 μg/ml) - BSA (170 μg ml) mixture for 1 hr. at 37°C on glass coverslips prior to adding a 1:3 dilution of human serum and incubating an additional 1 hr. at 37°C. * NeutrophUs were preincubated in fMLP and I-domain or mAb for 5 min. at 21°C prior to incorporation into Smith-Hollers adhesion chambers. Final concentrations of proteins are designated.

** NeutrophU adhesion was assayed as described (Anderson et al., Abnormal mobility of neonatal polymorphonucler leukocytes, J. Clin. Invest. 68:863-74 (1981)). Results of the representative experiment (above) indicate the capacity of the CD1 lb I-

Domain protein to inhibit stimulated neutrophU adhesion to KLH, iC3b and/or fibrinogen substrates when assessed in the absence of shear stress. This is especially apparent with respect to iC3b and fibrinogen binding, where 2.5 μM and or 10 μM concentrations of I-domain inhibit adhesion by 39-53% as compared to stimulated controls. This level of inhibition is comparable to that of the blocking mAb 3H5 directed at Mac-1 (CD1 lb) when used at saturating concentrations (10 μg/ml). An apparent concentration-independent relationship for I-Domain inhibition is observed. The optimum inhibitory concentration of I-Domain is 2.5 μM.

TABLE 5

Effect of CDllb I-Domain (Factor X_a-cleaved GST fusion protein) on neutrophU-endotheUal adhesion

Pretreatment Conditions Total Adhesion (%)** % Inhibition

HUVEC Ntmtrophils*

- -- 11 ± 2

TNF 100 U/ml, 4 hτ, ~ 88 ± 6

37°

"n I I--DDoommaaiinn 6 688 ±± 44 24 (25 μM)

I-Domain 56 ± 4 37 (5 μM)

π 3H5 anti CDllb mAb 39 ± 3 56 (10 μg/ml)

* NeutrophUs were preincubated with I-domain or mAb for 5 min. at 21°C prior to incorporation into Smith-Hollers adhesion chambers. Final concentrations of proteins are desginated.

** Assay performed as described by Smith, C.W. et al, J. Clin. Invest. 82:1745-1756, (1988).

Results of the representative experiment shown above demonstrate the capacity of CDl lb - I-Domain to inhibit the adhesion of unstimulated neutrophUs to endotheUal monolayers pretreated with TNF to maximalUy induce expression of ICAM-1 and other neutrophU adhesion ligands. As shown, final I-Domain concentrations of 5 μM or 25 μM inhibited adhesion (as compared to stimulated controls) by 37% or 24%, respectively, while the positive control 3H5 mAb inhibited adhesion by 56 when used at saturating concentrations. The remaining adhesion (unblocked by 3H5 mAb) represents the contribution of E-selectin and other endotheUal ligands elicited by TNF and is not expected to be impacted by I-Domain protein in this assay. TABLE 6

Effect of CDl lb I-Domain (Factor X_a-cleaved GST fusion protein) on neutrophil-endothelial adhesion: Experiment #2

Pretreatment Total Transendothelial Conditions Adhesion (%)** Migration (%)**

HUVEC NeutrophUs

1±2

TNF — 79 ±7 20 ±4

100 U/ml, hrs, 37°C

I-Domain 75 ±6 5±2

(25 μM)

I-Domain 62 ±5 15 ±3

(10 μM)

I-Domain 49 ±4 16 ±4

(5μM)

I-Domain 18 ±3 5±2

(lμM)

3H5 (anti CDllb) mAb 21 ±3 5± 1

(10 μg/ml)

GST(25μM) 82 ±6 24 ±6

GST (5 μM) 88 ±6 20 ±4 * NeutrophUs preincubated with I-domain, mAb or GST protein for 5 min at 21°C prior to incorporation in Smith-HoUers adhesion chambers. Final concentrations of proteins are designated.

** NeutrophU adhesion and transendothelial migration assayed as described by Smith, C.W. et al, J. CUn. Invest. 82: 1746-1756, (1988).

Results of the representative experiment shown above illustrate the capacity of CDllb - I-Domain to inhibit both neutrophU adhesion to and migration through endotheUal monolayers prestimulated with TNF. An inverse relationship between I-domain concentration and degree of inhibition of adhesion is seen. The optimal inhibitory concentration is 1 μM and in this experiment a 25 μM concentration is non inhibitory for adhesion. Importantly, a 1 μM concentration of I-Domain inhibits adhesion to the same extent as seen with the 3H5 mAb when used at saturating concentrations. The GST (control) fusion protein fragment is non inhibitory. Somewhat different relationships were observed in migration assessments. A bimodal dose response effect is apparent in which high concentrations (25 μM) of I-Domain block neutrophU motility independent of adhesion, and low concentrations (1 μM) block migration associated with diminished neutrophil-endothelial adhesion. In each case, transendothelial migration is potently inhibited to the same extent as effected by the 3H5 mAb. These findings indicate distinct inhibitory influences of CDllb - I-Domain protein on neutrophU motility (chemotaxis) and on adhesion per se. Such observations are entirely consistent with the known multifunctional properties of Mac-1 (Anderson et al, Contributions of the Mac-1 glycoprotein family to adherence-dependent granulocyte functions..., J. Immunol. 137:15-27 (1986) and Smith et al, Transendothelial migration, In Adhesion, pp 85-115, W.E.Freeman and Co. NY (1992).

SEQUENCE LISTING

(1) GENERAL INFORMATION:

(i) APPLICANT: The Upjohn Company

Heinrikson, Robert L. Anderson, Donald C. Tomich, Che-Shen C. Fairbanks, Michael B.

Bajt, Mary L.

(ii) TITLE OF INVENTION: MAC-1 I-DOMAIN PROTEIN USEFUL IN BLOCKING ADHESION AND MIGRATION OF NEUTROPHILS

(iii) NUMBER OF SEQUENCES: 9

(iv) CORRESPONDENCE ADDRESS:

(A) ADDRESSEE: The Upjohn Company, Intellectual Property Law (B) STREET: 301 Henrietta

(C) CITY: Kalamazoo

(D) STATE: MI

(E) COUNTRY: USA

(F) ZIP: 49001

(v) COMPUTER READABLE FORM:

(A) MEDIUM TYPE: Floppy disk

(B) COMPUTER: Gateway 2000 P5-90

(C) OPERATING SYSTEM: PC-DOS/MS-DOS (D) SOFTWARE: PatentIn Release #1.0, Version #1.25

(Vi) CURRENT APPLICATION DATA:

(A) APPLICATION NUMBER:

(B) FILING DATE: (C) CLASSIFICATION:

(viii) ATTORNEY/AGENT INFORMATION:

(A) NAME: Darnley, James D. , Jr.

(B) REGISTRATION NUMBER: 33,673 (C) REFERENCE/DOCKET NUMBER: 4767.P CN1

(ix) TELECOMMUNICATION INFORMATION:

(A) TELEPHONE: 616/385-5210

(B) TELEFAX: 616/385-6897 (C) TELEX: 224401

(2) INFORMATION FOR SEQ ID NO:l: (i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 435 amino acids

(B) TYPE: amino acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: protein

( i) SEQUENCE DESCRIPTION: SEQ ID NO:l:

Met Ser Pro lie Leu Gly Tyr Trp Lys lie Lys Gly Leu Val Gin Pro 1 5 10 15 Thr Arg Leu Leu Leu Glu Tyr Leu Glu Glu Lys Tyr Glu Glu His Leu

20 25 30 Tyr Glu Arg Asp Glu Gly Asp Lys Trp Arg Asn Lys Lys Phe Glu Leu 35 40 45

Gly Leu Glu Phe Pro Asn Leu Pro Tyr Tyr lie Asp Gly Asp Val Lys 50 55 60

Leu Thr Gin Ser Met Ala lie lie Arg Tyr lie Ala Asp Lys His Asn 65 70 75 80

Met Leu Gly Gly Cys Pro Lys Glu Arg Ala Glu lie Ser Met Leu Glu

85 90 95

Gly Ala Val Leu Asp lie Arg Tyr Gly Val Ser Arg lie Ala Tyr Ser 100 105 110

Lys Asp Phe Glu Thr Leu Lys Val Asp Phe Leu Ser Lys Leu Pro Glu

115 120 125 Met Leu Lys Met Phe Glu Asp Arg Leu Cys His Lys Thr Tyr Leu Asn

130 135 140

Gly Asp His Val Thr His Pro Asp Phe Met Leu Tyr Asp Ala Leu Asp 145 150 155 160

Val Val Leu Tyr Met Asp Pro Met Cys Leu Asp Ala Phe Pro Lys Leu 165 170 175

Val Cys Phe Lys Lys Arg lie Glu Ala lie Pro Gin lie Asp Lys Tyr 180 185 190

Leu Lys Ser Ser Lys Tyr lie Ala Trp Pro Leu Gin Gly Trp Gin Ala 195 200 205

Thr Phe Gly Gly Gly Asp His Pro Pro Lys Ser Asp Leu lie Glu Gly 210 215 220

Arg Gly lie Pro Gly Gly Ser Asp lie Ala Phe Leu lie Asp Gly Ser 225 230 235 240

Gly Ser lie lie Pro His Asp Phe Arg Arg Met Lys Glu Phe Val Ser 245 250 255

Thr Val Met Glu Gin Leu Lys Lys Ser Lys Thr Leu Phe Ser Leu Met 260 265 270

Gin Tyr Ser Glu Glu Phe Arg lie His Phe Thr Phe Lys Glu Phe Gin 275 280 285 Asn Asn Pro Asn Pro Arg Ser Leu Val Lys Pro lie Thr Gin Leu Leu 290 295 300

Gly Arg Thr His Thr Ala Thr Gly lie Arg Lys Val Val Arg Glu Leu

305 310 315 320

Phe Asn lie Thr Asn Gly Ala Arg Lys Asn Ala Phe Lys lie Leu Val

325 330 335

Val lie Thr Asp Gly Glu Lys Phe Gly Asp Pro Leu Gly Tyr Glu Asp 340 345 350

Val lie Pro Glu Ala Asp Arg Glu Gly Val lie Arg Tyr Val lie Gly 355 360 365 Val Gly Asp Ala Phe Arg Ser Glu Lys Ser Arg Gin Glu Leu Asn Thr 370 375 380 lie Ala Ser Lys Pro Pro Arg Asp His Val Phe Gin Val Asn Asn Phe 385 390 395 400

Glu Ala Leu Lys Thr lie Gin Asn Gin Leu Arg Glu Lys lie Phe Ala 405 410 415

lie Glu Gly Thr Gin Thr Gly Ser Ser Ser Ser Phe Glu His Glu Met 420 425 430

Ser Gin Glu 435

(2) INFORMATION FOR SEQ ID NO:2:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 6 amino acids

(B) TYPE: amino acid

(C) STRANDEDNESS: single (D) TOPOLOGY: linear

(ii) MOLECULE TYPE: protein

( i) SEQUENCE DESCRIPTION: SEQ ID NO:2:

Leu Val Pro Arg Gly Ser 1 5

(2) INFORMATION FOR SEQ ID NO:3:

(i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 8 amino acids (B) TYPE: amino acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: protein

(Xi) SEQUENCE DESCRIPTION: SEQ ID NO: 3: Leu Val Pro Arg Gly Ser Gly Gly 1 5

(2) INFORMATION FOR SEQ ID NO:4: (i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 9 amino acids

(B) TYPE: amino acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: protein

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:4: lie Glu Gly Arg Gly lie Pro Gly Gly 1 5 (2) INFORMATION FOR SEQ ID NO:5:

(i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 5 amino acids (B) TYPE: amino acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear (ii) MOLECULE TYPE: protein

(Xi) SEQUENCE DESCRIPTION: SEQ ID NO:5:

Gly lie Pro Gly Gly 1 5

(2) INFORMATION FOR SEQ ID NO:6:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 47 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single (D) TOPOLOGY: linear

(ii) MOLECULE TYPE: protein

(Xi) SEQUENCE DESCRIPTION: SEQ ID NO:6: GGCTCGGATC CGGTGGCAGT GACATTGCCT TCTTGATTGA TGGCTCT 47 (2) INFORMATION FOR SEQ ID NO:7:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 47 base pairs

(B) TYPE: nucleic acid (C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: protein

(Xi) SEQUENCE DESCRIPTION: SEQ ID NO:7: GGCTCGGGAT CCCCGGTGGC AGTGACATTG CCTTCTTGAT TGATGGC 47

(2) INFORMATION FOR SEQ ID NO:8:

(i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 44 base pairs (B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: protein

(Xi) SEQUENCE DESCRIPTION: SEQ ID NO: 8: GAGCCTGAAT TCTATTCCTG AGACATCTCA TGCTCAAAGG AGCT 44

(2) INFORMATION FOR SEQ ID NO:9:

(i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 10 amino acids

(B) TYPE: amino acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear (ii) MOLECULE TYPE: protein

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:9:

He Leu Gly Arg His His His His His His 1 5 10

Claims

WHAT IS CLAIMED:

1. A fusion protein of glutathione-S-transferase (GST) and I-Domain derived from human leukocyte B₂-integrin Mac-1 in which said GST and said I-Domain are linked by a peptide segment containing a Factor X^a cleavage site as set forth in ID SEQ NO: I.

2. An I-Domain protein defined by the amino acid sequence Gly 226 through Glu 435, inclusive, as set forth in ID SEQ NO: I.

3. A pharmaceutical composition comprising: a) a recombinant I-Domain protein derived from human leukocyte B₂-integrin Mac-1 defined by the amino acid sequence Gly 226 through Glu 435, inclusive, as set forth in ID SEQ NO: I; and b) a pharmaceutically acceptable carrier or excipient.

4. The pharmaceutical composition of Claim 3 wherein said recombinant I-Domain is a fragment, analog or chemical derivative of said protein.

5. A method for treating inflammation comprising: administering to a patient suffering from an inflammatory condition a pharmaceutically effective amount of an anti-inflammatory agent comprising a recombinant I-Domain protein derived from human leukocyte B₂-integrin Mac-1 defined by the amino acid sequence Gly 226 through Glu 435, inclusive, as set forth in ID SEQ NO: I.

6. The method of Claim 5 wherein said recombinant I-Domain protein is a fragment, analog or chemical derivative of said protein.