Eukaryotic proteins expressed in Escherichia coli: an improved thrombin cleavage and purification procedure of fusion proteins with glutathione S-transferase

Several systems have been developed to allow for rapid and efficient purification of recombinant proteins expressed in bacteria. The expression of polypeptides in frame with glutathione S-transferase (GST) allows for purification of the fusion proteins from crude bacterial extracts under nondenaturing conditions by affinity chromatography on glutathione agarose (D. B. Smith and K. S. Johnson, 1988, Gene 67, 31-40). This vector expression system has also incorporated specific protease cleavage sites to facilitate proteolysis of the bacterial fusion proteins. In our hands, the cleavage of these fusion proteins at a thrombin cleavage site proceeded slowly. To facilitate the cleavage of fusion proteins, we have introduced a glycine-rich linker (glycine kinker) containing the sequence P.G.I.S.G.G.G.G.G located immediately following the thrombin cleavage site. This glycine kinker greatly increases the thrombin cleavage efficiency of several fusion proteins. The introduction of the glycine kinker into fusion proteins allows for the cleavage of the fusion proteins while they are attached to the affinity resin resulting in a single step purification of the recombinant protein. More than 2 mg of the highly purified protein was obtained from the equivalent of 100 ml of bacterial culture within a few hours when a protein tyrosine phosphatase was employed as a test protein. The vector, pGEX-KG, has also been modified to facilitate cloning of a variety of cDNAs in all reading frames and has been successfully used to express several eukaryotic proteins.     

Expression and characterization of UDPGlc dehydrogenase (KfiD), which is encoded in the type-specific region 2 of the Escherichia coli K5 capsule genes.

Region 2 of the Escherichia coli K5 capsule gene cluster contains four genes (kfiA through -D) which encode proteins involved in the synthesis of the K5 polysaccharide. A DNA fragment containing kfiD was amplified by PCR and cloned into the gene fusion vector pGEX-2T to generate a GST-KfiD fusion protein. The fusion protein was isolated from the cytoplasms of IPTG (isopropyl-beta-D-thiogalactopyranoside)-induced recombinant bacteria by affinity chromatography and cleaved with thrombin. The N-terminal amino acid sequence of the cleavage product KfiD' corresponded to the predicted amino acid sequence of KfiD with an N-terminal glycyl-seryl extension from the cleavage site of the fusion protein. Anti-KfiD antibodies obtained with KfiD' were used to isolate the intact KfiD protein from the cytoplasms of E. coli organisms overexpressing the kfiD gene. The fusion protein, its cleavage product (KfiD'), and overexpressed KfiD converted UDPGlc to UDPGlcA. The KfiD protein could thus be characterized as a UDPglucose dehydrogenase.     

A novel protocol for the production of recombinant LL-37 expressed as a thioredoxin fusion protein

LL-37 is the only cathelicidin-derived antimicrobial peptide found in humans and it has a multifunctional role in host defense. The peptide has been shown to possess immunomodulatory functions in addition to antimicrobial activity. To provide sufficient material for biological and structural characterization of this important peptide, various systems were developed to produce recombinant LL-37 in Escherichia coli. In one previous approach, LL-37 coding sequence was cloned into vector pET-32a, allowing the peptide to be expressed as a thioredoxin fusion. The fusion protein contains two thrombin cleavage sites: a vector-encoded one that is 30-residue upstream of the insert and an engineered one that is immediately adjacent to LL-37. Cleavage at these two sites shall generate three fragments, one of which is the target peptide. However, when the fusion protein was treated with thrombin, cleavage only occurred at the remote upstream site. A plausible explanation is that the thrombin site adjacent to LL-37 is less accessible due to the peptide's aggregation tendency and cleavage at the remote site generates a fragment, which forms a large aggregate that buries the intended site. In this study, I deleted the vector-encoded thrombin site and S tag in pET-32a, and then inserted the coding sequence for LL-37 plus a thrombin site into the modified vector. Although removing the S tag did not change the oligomeric state of the fusion protein, deletion of the vector-encoded thrombin site allowed the fusion to be cleaved at the engineered site to release LL-37. The released peptide was separated from the carrier and cleavage enzyme by size-exclusion chromatography. This new approach enables a quick production of high quality active LL-37 with a decent amount.     

pC6-2/caspase-6 system to purify glutathione-S-transferase-free recombinant fusion proteins expressed in Escherichia coli

Glutathione-S-transferase (GST) fusion protein expression vectors are often employed for the expression and purification of proteins in Escherichia coli. GST is then removed by site-specific proteolysis using thrombin. However, the presence of internal thrombin cleavage sites in expressed proteins can severely affect the purification of intact proteins. Cysteine-dependent aspartate-specific proteases (caspases) are efficient enzymes with defined substrate specificity. Unlike most of the proteases used for the removal of affinity tags, caspases do not leave any amino acids at the amino-terminus of cleaved proteins. We have engineered the caspase-6 site VEMD in a pGEX vector to give the pC6-2 vector. The caspase-6 can be easily removed after cleavage. Here, we describe the detailed protocol for purifying proteins using our pC6-2/caspase-6 expression and purification system. The cleavage by caspase-6 occurs in <30 min and the entire procedure can be completed in 2 d.     

Highly efficient protein expression and purification using bacterial hemoglobin fusion vector

Recently developed bacterial hemoglobin (VHb) fusion expression vector has been widely used for the production of many target proteins due to its distinctive properties of expressing fusion protein with red color which facilitates visualization of the steps in purification, and increasing solubility of the target proteins. However, after intensive use of the vector, several defects have been found. In this report, we present a modified VHb fusion vector (pPosKJ) with higher efficiency, in which most of the defects were eliminated. First, it was found that thrombin protease often digests target protein as well as inserted thrombin cleavage site, so it was replaced by a TEV cleavage site for more specific cleavage of VHb from target protein. Second, a glycine-rich linker sequence was inserted between 6x his-tag and VHb to improve the affinity of 6x his-tag to Ni-NTA resin, resulting in higher purity of eluted fusion protein. Third, EcoRI and XhoI restriction sites located elsewhere in the vector were removed to make these restriction sites available for the cloning of target protein coding genes. A pPosKJ vector was fully examined with an anti-apoptotic BCL-2 family member of Caenorhabditis elegans, CED-9. A C-terminal VHb fusion expression vector (pPosKJC) was also constructed for stable expression of target proteins that may be difficult to express with an N-terminal fusion. Vaccinia-related kinase 1 (VRK1) was also successfully expressed and purified using the vector with high yield. Taken together, we suggest that the VHb fusion vector may be well suited for high-throughput protein expression and purification.     

Overexpression of active Syrian golden hamster prion protein PrPc as a glutathione S-transferase fusion in heterologous systems.

This article describes a procedure which permits for the first time the isolation of the prion protein PrPc from the Syrian golden hamster in heterologous systems. Using a glutathione S-transferase (GST) fusion approach, milligram amounts of stable, soluble, and homogeneous GST::PrPc protein were obtained in Escherichia coli and with baculovirus-infected insect cells. Authentic PrPc was released from the immobilized fusion protein by direct cleavage with thrombin. GST::PrPc expressed in these two expression systems and also authentic PrPc released by thrombin cleavage were recognized by a polyclonal antibody directed against amino acid 95 to 110 of the golden hamster PrPc protein. GST::PrPc was not detected by a monoclonal antibody recognizing the region encompassing amino acids 138 to 152 of the human prion protein. The fusion protein was sensitive to proteinase K digestion, demonstrating that the cellular rather than the proteinase K-resistant scrapie isoform was produced.     

Discovery of amino acid motifs for thrombin cleavage and validation using a model substrate

Understanding the active site preferences of an enzyme is critical to the design of effective inhibitors and to gaining insights into its mechanisms of action on substrates. While the subsite specificity of thrombin is understood, it is not clear whether the enzyme prefers individual amino acids at each subsite in isolation or prefers to cleave combinations of amino acids as a motif. To investigate whether preferred peptide motifs for cleavage could be identified for thrombin, we exposed a phage-displayed peptide library to thrombin. The resulting preferentially cleaved substrates were analyzed using the technique of association rule discovery. The results revealed that thrombin selected for amino acid motifs in cleavage sites. The contribution of these hypothetical motifs to substrate cleavage efficiency was further investigated using the B1 IgG-binding domain of streptococcal protein G as a model substrate. Introduction of a P(2)-P(1)' LRS thrombin cleavage sequence within a major loop of the protein led to cleavage of the protein by thrombin, with the cleavage efficiency increasing with the length of the loop. Introduction of further P(3)-P(1) and P(1)-P(1)'-P(3)' amino acid motifs into the loop region yielded greater cleavage efficiencies, suggesting that the susceptibility of a protein substrate to cleavage by thrombin is influenced by these motifs, perhaps because of cooperative effects between subsites closest to the scissile peptide bond.     

An assay for high-sensitivity detection of thrombin activity and determination of proteases activating or inactivating protease-activated receptors

This paper describes the development of galactosidase protease-activated receptor (GPAR) as a recombinant protein obtained by fusion of beta-galactosidase, the extracellular domains of protease-activated receptors (PARs), and a biotin acceptor domain. Used as an immobilized substrate, this protein allows the detection of thrombin in the sub-picomolar range. A comparative analysis for proteolytic cleavage of murine PAR1, PAR2, and PAR3 and human PAR4 was performed, involving mutated and nonmutated GPAR fusion proteins. Thrombin cleaved GPAR1 (2.6 mol(beta-galactosidase)/(mol(thrombin) * min)), GPAR3 (410 mmol(beta-galactosidase)/(mol(thrombin) * min)), and GPAR4 (4.3 mmol(beta-galactosidase)/(mol(thrombin) * min)) specifically at the proteolytic activation site. A second possible cleavage site for thrombin is present in murine PAR1 and PAR3. Trypsin and plasmin cleaved all receptor fusion proteins with little specificity for the activation site, except for a marked preference of trypsin for cleavage at the activation site of GPAR2. Chymotrypsin cleaves GPAR1 at a rate (58 mmol(beta-galactosidase)/(mol(thrombin) * min)) that suggests the possibility of chymotryptic inactivation of PAR1. Elastase may inactivate PAR1 and PAR3, but probably not PAR2 and PAR4. Neither activated protein C nor the plasminogen activators cleave any GPAR fusion protein at considerable rates.     

Fusion expression of bovine lactoferricin in Escherichia coli

The drug resistance problem has been growing with the utilization of current antibiotics in feed and medical industries. LfcinB, a 25-amino acid antibacterial peptide derived from bovine lactoferrin, is one of potential alternatives of antibiotics. According to the bias of codon utilization of Escherichia coli, a fragment encoding LfcinB has been chemically synthesized, inserted into vector pGEX-4T-2 and expressed in E. coli. The antibacterial peptide was fused with GST with a protease cleavage site located between them. Two constructs with different cleavage sites were made. One construct, pGEX-Th-LfcinB, contains a thrombin cleavage site carried by the vector, and the other, pGEX-Th-Xa-LfcinB, contains a Factor Xa cleavage site which was introduced after the thrombin cleavage site. Fusion protein GST-Th-LfcinB protein was efficiently cleaved by thrombin, yielding recombinant LfcinB showing antibacterial activity. However, fusion protein GEX-Th-Xa-Lfcin B containing Factor Xa recognition site could not be cleaved by Factor Xa at the conditions tried in this study. Successful expression of LfcinB in E. coli provides a possible method to produce LfcinB in large amounts.     

Linkers for improved cleavage of fusion proteins with an engineered alpha-lytic protease.

Addition of an N-terminal fusion partner can greatly aid the expression and purification of a recombinant protein in Escherichia coli. We investigated two genetically engineered proteases designed to remove the fusion partner after the protein of interest has been expressed. Recombinant human insulin-like growth factor-II (hIGF-II) has been produced from E. coli-derived fusion proteins using a novel enzymatic cleavage system that uses a mutant of alpha-lytic protease. Initially, two potential fusion protein linkers were designed, Pro-Ala-Pro-His (PAPH) and Pro-Ala-Pro-Met (PAPM), and were tested as substrates in the form of synthetic dodecapeptides. Using mass spectrometry and reverse-phase HPLC, the position of cleavage was confirmed and the kinetics of synthetic peptide cleavage were examined. Use of the linkers in hIGF-II fusion proteins produced in E. coli was then evaluated. The fusion proteins constructed consist of the first 11 amino acids of porcine growth hormone linked N-terminally to hIGF-II by six amino acids that include the dipeptide Val-Asn followed by a variable tetrapeptide protease cleavage motif. Mass spectrometry and N-terminal sequencing confirmed that proteolytic cleavage of the fusion proteins had occurred at the predicted sites. Using the fusion proteins as substrates, the cleavage of the rationally designed motifs by the alpha-lytic protease mutant was compared. The fusion protein containing the motif PAPM had a k(cat)/K(M) ratio indicating a 1.6-fold preference over the PAPH fusion protein for cleavage by this enzyme. Furthermore, when hIGF-II fusion proteins containing the designed cleavable linkers were processed with the engineered alpha-lytic protease, they gave greatly improved yields of native hIGF-II compared to an analogous fusion protein cleaved by H64A subtilisin. Comparison of the peptide and protein cleavage studies shows that the efficient proteolysis of the cleavage motifs is an inherent property of the designed sequences and is not determined by secondary or tertiary structure in the fusion proteins.     

Thrombin and H64A subtilisin cleavage of fusion proteins for preparation of human recombinant parathyroid hormone

Human parathyroid hormone, hPTH, an 84 amino acid polypeptide, was produced intracellularly inEscherichia coli as a fusion protein, linked to the C-terminus of a 15 kD IgG-binding protein. Approximately 100 mg fusion protein was obtained per liter fermentation medium. To test the efficiency of two alternative enzymatic cleavage methods, two fusion proteins differing only in the linker region were constructed. Cleavage of a Phe-Phe-Pro-Arg linker was obtained with bovine thrombin and cleavage of a Phe-Ala-His-Tyr linker with recombinant H64A subtilisin. Both enzmes yielded the correct N-terminus and cleaved their respective linkers quantitatively, although additional internal cleavage sites in hPTH were detected and characterized. The linker cleavage conditions were optimized and hPTH was purified to homogeneity. Thrombin cleavage resulted in a final yield of 5 mg hPTH/L, while H64A subtilisin cleavage was more specific and gave 8 mg/L. The purified recombinant product was identical to native hPTH and exhibited full biological activity in an adenylate cyclase assay.     

Ligation-independent cloning of glutathione S-transferase fusion genes for expression in Escherichia coli.

A plasmid vector has been constructed that allows the ligation-independent cloning of cDNAs in any reading frame and directs their synthesis in Escherichia coli as glutathione S-transferase-linked fusion proteins. The cloning procedure does not require restriction enzyme digestion of the target sequence and does not introduce any additional sequences between the thrombin cleavage site and the foreign protein. Extended single-stranded tails complementary between the vector and insert, generated by the (3'----5') exonuclease activity of T4 DNA polymerase, obviate the need for in vitro ligation prior to bacterial transformation. This cloning procedure is rapid and highly efficient, and has been used successfully to construct a series of fusion proteins to investigate the sequence requirements for efficient thrombin cleavage.     

Effect of Cleavage Mutants on Syncytium Formation Directed by the Wild-Type Fusion Protein of Newcastle Disease Virus

The effects of Newcastle disease virus (NDV) fusion (F) glycoprotein cleavage mutants on the cleavage and syncytium-forming activity of the wild-type F protein were examined. F protein cleavage mutants were made by altering amino acids in the furin recognition region (amino acids 112 to 116) in the F protein of a virulent strain of NDV. Four mutants were made: Q114P replaced the glutamine residue with proline; K115G replaced lysine with glycine; double mutant K115G, R113G replaced both a lysine and an arginine with glycine residues; and a triple mutant, R112G, K115G, F117L, replaced three amino acids to mimic the sequence found in avirulent strains of NDV. All mutants except Q114P were cleavage negative and fusion negative. However, addition of exogenous trypsin cleaved all mutant F proteins and activated fusion. As expected for an oligomeric protein, the fusion-negative mutants had a dominant negative phenotype: cotransfection of wild-type and mutant F protein cDNAs resulted in an inhibition of syncytium formation. The presence of the mutant F protein did not inhibit cleavage of the wild-type protein. Furthermore, evidence is presented that suggests that the mutant protein and the wild-type protein formed heterooligomers. By measuring the syncytium-forming activity of the wild-type protein at various ratios of expression of mutant and wild-type protein, results were obtained that are most consistent with the notion that the size of the functionally active NDV F protein in these assays is a single oligomer, likely a trimer. That a larger oligomer, containing a mix of both wild-type and mutant F proteins, has partial activity cannot, however, be ruled out.     

Exchange of precursor-specific elements between Pro-sigma E and Pro-sigma K of Bacillus subtilis.

sigma E and sigma K are sporulation-specific sigma factors of Bacillus subtilis that are synthesized as inactive proproteins. Pro-sigma E and pro-sigma K are activated by the removal of 27 and 20 amino acids, respectively, from their amino termini. To explore the properties of the precursor-specific sequences, we exchanged the coding elements for these domains in the sigma E and sigma K structural genes and determined the properties of the resulting chimeric proteins in B. subtilis. The pro-sigma E-sigma K chimera accumulated and was cleaved into active sigma K, while the pro-sigma K-sigma E fusion protein failed to accumulate and is likely unstable in B. subtilis. A fusion of the sigE "pro" sequence to an unrelated protein (bovine rhodanese) also formed a protein that was cleaved by the pro-sigma E processing apparatus. The data suggest that the sigma E pro sequence contains sufficient information for pro-sigma E processing as well as a unique quality needed for sigma E accumulation.     

An evaluation of different enzymatic cleavage methods for recombinant fusion proteins, applied on des(1–3)insulin-like growth factor I

Different enzymatic methods for cleavage of recombinant fusion proteins were compared. To find an efficient cleavage method, five different fusion proteins were produced. The fusion proteins differed only in the linker region between the fusion partner and the desired product, human des(1–3)insulin-like growth factor I. A cleavage study was performed with enterokinase, plasmin, thrombin, urokinase, and recombinant H64A subtilisin. Significant cleavage was obtained using thrombin, H64A subtilisin, and enterokinase. Thrombin cleavage was studied on a larger scale and des(1–3)IGF-I was recovered at a final yield of 3 mg/L growth medium. Thrombin and enterokinase were also studied as immobilized proteases and they cleaved the fusion proteins with retained activity. To further improve thrombin cleavage, a continuous reactor was constructed, consisting of a closed system with a thrombin column and an ion exchange column in series. Here, the fusion protein circulated while free des(1–3)IGF-I was bound to the ion exchange column after release from the fusion protein. In the reactor, thrombin was as efficient as the free enzyme but gave a diminished rate of product degradation.     

Newcastle disease virus fusion protein expressed in a fowlpox virus recombinant confers protection in chickens.

A cDNA copy of the RNA encoding the fusion (F) protein of Newcastle disease virus (NDV) strain Texas, a velogenic strain of NDV, was obtained and the sequence was determined. The 1,792-base-pair sequence encodes a protein of 553 amino acids which has essential features previously established for the F protein of virulent NDV strains. These include the presence of three strongly hydrophobic regions and pairs of dibasic amino acids in the pentapeptide Arg-Arg-Gln-Arg-Arg preceding the putative cleavage site. When inserted into a fowlpox virus vector, a glycosylated protein was expressed and presented on the surface of infected chicken embryo fibroblast cells. The F protein expressed by the recombinant fowlpox virus was cleaved into two polypeptides. When inoculated into susceptible birds by a variety of routes, an immunological response was induced. Ocular or oral administration of the recombinant fowlpox virus gave partial protection, whereas both intramuscular and wing-web routes of inoculation gave complete protection after a single inoculation.     

Expression of recombinant human glucose-dependent insulinotropic polypeptide in Escherichia coli by sequence-specific proteolysis of a protein A fusion protein

Glucose-dependent insulinotropic polypeptide (GIP) is a forty-two amino acid hormone that stimulates the secretion of insulin from the pancreatic B-cells in the presence of elevated glucose concentrations. The human GIP gene with the human A-fibrinopeptide sequence was synthesized and linked to the Staphylococcus aureus protein A gene in the vector pRIT2T. This plasmid was expressed in Escherichia coli, and the resulting fusion protein consisted of three domains: protein A for ease of purification, fibrinopeptide sequence for thrombin cleavage and human GIP. The GIP was subsequently cleaved from the fusion protein with -thrombin. The identity of the recombinant human GIP was confirmed by SDS-PAGE, ELISA, HPLC and amino-terminal amino acid sequence analysis. This recombinant product was shown to have comparable insulinotropic activity to porcine GIP in the isolated perfused pancreas.     

The Importance of Exosite Interactions for Substrate Cleavage by Human Thrombin

Thrombin is a serine protease of the chymotrypsin family that acts both as a procoagulant and as an anticoagulant by cleaving either factor VIII, factor V and fibrinogen or protein C, respectively. Numerous previous studies have shown that electropositive regions at a distance from the active site, so called exosites, are of major importance for the cleavage by human thrombin. Upstream of all the known major cleavage sites for thrombin in factor VIII, factor V and fibrinogen are clusters of negatively charged amino acids. To study the importance of these sites for the interaction with the exosites and thereby the cleavage by thrombin, we have developed a new type of recombinant substrate. We have compared the cleavage rate of the minimal cleavage site, involving only 8-9 amino acids (typically the P4-P4’ positions) surrounding the cleavage site, with the substrates also containing the negatively charged regions upstream of the cleavage sites. The results showed that addition of these regions enhanced the cleavage rate by more than fifty fold. However, the enhancement was highly dependent on the sequence of the actual cleavage site. A minimal site that showed poor activity by itself could be cleaved as efficiently as an optimal cleavage site when presented together with these negatively charged regions. Whereas sites conforming closely to the optimal site were only minimally enhanced by the addition of these regions. The possibility to mimic this interaction for the sites in factor V and factor VIII by recombinant substrates, which do not have the same folding as the full size target, indicates that the enhancement was primarily dependent on a relatively simple electrostatic interaction. However, the situation was very different for fibrinogen and protein C where other factors than only charge is of major importance.     

Comparison of substrate specificities against the fusion glycoprotein of virulent Newcastle disease virus between a chick embryo fibroblast processing protease and mammalian subtilisin-like proteases

The fusion (F) protein precursor of virulent Newcastle disease virus (NDV) strains has two pairs of basic amino acids at the cleavage site, and its intracellular cleavage activation occurs in a variety of cells; therefore, the viruses cause systemic infections in poultry. To explore the protease responsible for the cleavage in the natural host, we examined detailed substrate specificity of the enzyme in chick embryo fibroblasts (CEF) using a panel of the F protein mutants at the cleavage site expressed by vaccinia virus vectors, and compared the specificity with those of mammalian subtilisin-like proteases such as furin, PC6 and PACE4 which are candidates for F protein processing enzymes. It was demonstrated in CEF cells that Arg residues at the -4, -2 and -1 positions upstream of the cleavage site were essential, and that at the -5 position was required for maximal cleavage. Phe at the +1 position was also important for efficient cleavage. On the other hand, furin and PC6 expressed by vaccinia virus vectors showed cleavage specificities against the F protein mutants consistent with that shown by the processing enzyme of CEF cells, but PACE4 hardly cleaved the F proteins including the wild type. These results indicate that the proteolytic processing enzymes of poultry for virulent NDV F proteins could be furin and/or PC6 but not PACE4. The significance of individual contribution of the three amino acids at the -5, -2 and +1 positions to cleavability was discussed in relation to the evolution of virulent and avirulent NDV strains.     

Identification of candidate substrates for ectodomain shedding by the metalloprotease-disintegrin ADAM8

ADAM proteases are type I transmembrane proteins with extracellular metalloprotease domains. As for most ADAM family members, ADAM8 (CD156a, MS2) is involved in ectodomain shedding of membrane proteins and is linked to inflammation and neurodegeneration. To identify potential substrates released under these pathologic conditions, we screened 10-mer peptides representing amino acid sequences from extracellular domains of various membrane proteins using the ProteaseSpot system. A soluble ADAM8 protease containing a pro- and metalloprotease domain was expressed in E. coli and purified as active protease owing to autocatalytic prodomain removal. From 34 peptides tested in the peptide cleavage assay, significant cleavage by soluble ADAM8 was observed for 14 peptides representing membrane proteins with functions in inflammation and neurodegeneration, among them the beta-amyloid precursor protein (APP). The in vivo relevance of the ProteaseSpot method was confirmed by cleavage of full-length APP with ADAM8 in human embryonic kidney 293 cells expressing tagged APP. ADAM8 cleaved APP with similar efficiency as ADAM10, whereas the inactive ADAM8 mutant did not. Exchanging amino acids at defined positions in the cleavage sequence of myelin basic protein (MBP) revealed sequence criteria for ADAM8 cleavage. Taken together, the results allowed us to identify novel candidate substrates that could be cleaved by ADAM8 in vivo under pathologic conditions.