| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 279, Issue 14, 13402-13411, April 2, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
From the
Department of Dermatology, Leiden University Medical Center, Wassenaarseweg 72, 2333 AL Leiden, the Division of Medicinal Chemistry, Free University, De Boelelaan 1083, 1081 HV, Amsterdam, the ¶Department of Parasitology, Leiden University Medical Center, P. O. Box 9600, 2300 RC, Leiden, the ||Department of Molecular Neurobiology, Free University, De Boelelaan 1087, 1081 HV, Amsterdam, and the **Department of Infectious Disease, Leiden University Medical Center, P. O. Box 9600, 2300 RC Leiden, The Netherlands
Received for publication, November 24, 2003 , and in revised form, January 7, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
S binding, Ca2+ mobilization, and chemotaxis assays, we demonstrate that the C-terminally truncated CXCL10 variant is a potent ligand for CXCR3. Moreover, the inverse agonist activity on the virally encoded receptor ORF74 and the direct antibacterial activity of CXCL10 are fully retained. Hence, we have identified furin as a novel chemokine-modifying enzyme in vitro and most probably also in vivo, generating a C-terminally truncated CXCL10, which fully retains its (inverse) agonistic properties. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
In addition to the large number of genes encoding chemokines, processing of these proteins by (a limited number of) enzymes increases the complexity of chemokine signaling. Besides the large body of evidence on chemokine processing by dipeptidylpeptidase IV (CD26) (2), several novel chemokine processing enzymes have been identified in the past few years, e.g. matrix metalloproteinase 2 and 9, cathepsin D and G, and elastase (3–8). Apart from a range of biological effects on chemokine receptor activation, processed chemokines occasionally gain very unexpected, sometimes not even chemokine receptor-related, activities (9, 10).
In general, alterations at the N terminus down-regulate chemokine receptor potencies and are therefore considered important mechanisms to modulate inflammatory responses. Several naturally occurring truncations of chemokines have been identified from which it is still unknown which system is involved in their formation. Interestingly, several of these represent C-terminal truncations, as in the case of CXCL7 (11, 12) and CCL20 (13). One specific C-terminal truncation of CXCL7, found in platelets, involves the removal of the last two amino acids. In contrast to intact CXCL7, this form has potent antibacterial activity (10), pointing to a role central role of the C-terminal amino acids in the innate immune activity of this chemokine.
For the chemokine receptor CXCR3 three ligands are known (CXCL9, -10, and -11). The CXCR3 receptor is highly expressed on activated T-lymphocytes and CXCR3+ T-cells have been implicated in several pathological conditions. For example, CXCR3 and its ligands are involved in allograft rejection (14, 15), wound healing, and angiogenesis (16). In the skin, expression of CXCR3 targeting chemokines by keratinocytes, fibroblasts, and endothelial cells contributes to the influx of T-lymphocytes (17, 18). A recent study has shown that the CXCR3 ligands also possess direct antibacterial activity (19), suggesting a role for these chemokines in the innate immune response.
All three CXCR3 ligands are sensitive to N-terminal truncation by dipeptidylpeptidase IV (20, 21). In addition, we have recently shown that native CXCL10 that is produced upon stimulation of primary human keratinocytes with interferon- is partially processed to a CXCL10 molecule that lacks the last four C-terminal amino acids (22). The aim of the present study was to identify the enzymatic system responsible for the in vivo cleavage of CXCL10 and to determine the biological activity of this novel CXCL10 variant using several different biological assays. We demonstrate that furin can cleave CXCL10 in vitro and that an inhibitor of pro-protein convertases (e.g. furin) prevents processing of CXCL10 in vivo. We therefore identify furin as a novel chemokine-processing enzyme. The resulting CXCL10 variant retains full (inverse) agonistic activities.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
1 and TNF-
were from Strathmann Biotech GmbH (Hannover, Germany). Poros 50 R2 was obtained from Applied Biosystems (Foster City, CA). DEAE-dextran (chloride form) was obtained from Sigma. Dulbecco's modified Eagle's medium (DMEM) was obtained from Invitrogen, fetal bovine serum was purchased from Integro B. V. (Dieren, The Netherlands) and penicillinstreptomycin was from BioWhittaker Europe (Verviers, Belgium). pNF-
B-Luc was obtained from Stratagene (La Jolla, CA.)
Cell Culturing and Stimulation—Keratinocytes were obtained from human foreskin and grown in single cell suspensions in low calcium (0.1 mM) serum-free medium (keratinocyte-SFM, Invitrogen) as described earlier (22). The culture of primary human dermal fibroblasts was established as described earlier (23).
To induce chemokine expression, cells were stimulated with 2000 units/ml human rIFN- (keratinocytes) or 500 units/ml human rIFN- + 500 units/ml human rTNF (fibroblasts), for 6 h. Following this stimulation, cells were washed with phosphate-buffered saline and transferred to MCDB153 medium (Invitrogen) and allowed to secrete chemokines for 18 h at 37 °C.
Application of Proteinase Inhibitors in Keratinocyte Cultures—Human keratinocytes were cultured and stimulated as described above. After stimulation and washing of the cells, keratinocyte-derived chemokines were produced in the presence of either 25 µM decanoyl-RVKR-chloromethylketone or 20 µg/ml Marimastat.
Purification of CXCR3-targeting Chemokines—Supernatants from chemokine expressing cells were filtered through a 0.22-µm filter and applied to a 1-ml HiTrap heparin column (Amersham Biosciences) equilibrated with 10 mM sodium phosphate, pH 7.1. The column was subsequently washed with 5 ml of 10 mM sodium phosphate (pH 7.1) containing 0.35 M NaCl. This was followed by elution of heparin-binding proteins using 2 ml of 1.5 M NaCl in 10 mM phosphate (pH 7.1).
Chemokines were further purified using reverse phase high-performance liquid chromatography (HPLC) on a 2.1 x 250 mm Nucleosil C-18 column (Machery-Nagel, Düren, Germany) with a flow rate of 0.3 ml/min. Proteins were eluted using a continuous gradient from 0 to 60% acetonitrile (in 0.02% HCl) over 60 min and collecting 1-min fractions.
In Vitro Processing and Mass Spectrometric Analysis of Chemokines—Recombinant chemokines at a concentration of 2 µM were incubated with furin (5 units/ml) alone, or in combination with carboxypeptidase B (0.5 units/ml) in a buffer containing 100 mM HEPES and 1 mM CaCl2 (pH 7.6) at 37 °C for 3 h.
To establish the time course of furin-mediated cleavage of human CXCL10, recombinant CXCL10 (0.5 µM) was incubated with 0.1 units of furin in 20 µl of 100 mM HEPES and 1 mM CaCl2, pH 7.6, for 0, 0.5, 1, 5, 30, and 60 min at 37 °C.
After digestion, chemokines were desalted using Poros 50 R2, packed in a pipette tip. Peptides were eluted in 60% acetonitril/0.02% HCl and measured by MALDI-MS (Reflex, Bruker Daltonics, Bremen) using dihydroxybenzoic acid as a matrix.
To determine kinetic parameters of CXCL10 conversion by furin, different concentrations of CXCL10 were incubated with 0.1 units of furin and incubated for 15 min. The ratio of the abundance of the intact and the cleaved forms was measured by MALDI-MS and used to calculate the amount of substrate converted. Data were subsequently fitted to the Michaelis-Menten equation.
For large scale production of C-terminally truncated CXCL10, in vitro digestions were carried out using 25 µg recombinant CXCL10 using 10 units of furin in combination with 0.3 units of carboxypeptidase B, and used this material in several bioassays. As a control, intact recombinant CXCL10 was treated identically, though without the enzymes.
GTPS Binding—CXCR3-expressing CHO cells (24) were scraped in their medium, pelleted, and resuspended in GTPS binding buffer (20 mM HEPES, 100 mM NaCl, 5 mM MgCl2, 0.1% bovine serum albumin, pH 7.4), containing 1 mM EDTA. Cells were kept on ice for 20 min and subsequently sonicated (2 x 15 s, with 2-min interval). Large cellular debris was removed by centrifugation at 200 x g. Cell membranes were pelleted from the supernatant at 48,000 x g and resuspended in GTPS binding buffer and incubated (1 µg) for 20 min at 30 °C with 1 µM GDP in a total volume of 200 µl. Following incubation with various concentrations of chemokines for 10 min at 30 °C, membranes were incubated with 0.3 nM [35S]GTPS (1250 Ci/mmol, PerkinElmer) for 20 min at 30 °C. The reaction was terminated by filtering the membranes through a GF/B filter (Whatman) using ice-cold buffer (20 mM HEPES, 10 mM sodium pyrophosphate (Na4P2O7), pH 7.4). Filters were transferred to counting vials and membrane-bound [35S]GTPS was measured by liquid scintillation in a Wallac 1410.
Ca2+ Mobilization—Intracellular calcium mobilization was measured as described previously (24). In brief, CHO cells stably transfected with human CXCR3 were loaded with 1 mM Fura-2 acetoxymethyl esther (Molecular Probes, Leiden, The Netherlands). Upon receptor stimulation, fluorescence recordings were made in a continuously stirred 1-ml cuvette at 37 °C using a fluorescence spectrophotometer (Photon Technology International, Inc., Lawrenceville, NJ).
Chemotaxis—Human peripheral blood T-cells were isolated from buffy coats of healthy donors and stimulated with IL-2 (100 units/ml) and PHA (1 mg/ml) to induce CXCR3 expression, as described previously (24). Chemotaxis of these cells toward chemokines was studied using a Transwell assay starting with 5 x 106 cells/100 µl for each individual assay.
ORF74 Constitutive Activity—The cDNA of HHV-8-encoded ORF74 (GenBankTM accession number U71368
[GenBank]
with a silent G-to-T mutation at position 927 (a gift from Dr. T. W. Schwartz)) was inserted into pcDEF3 (A gift from Dr. J. A. Langen) after PCR amplification. COS-7 cells were grown at 5% CO2 and 37 °C in Dulbecco's modified Eagle's medium supplemented with 5% fetal bovine serum, 50 units/ml penicillin and 50 µg/ml streptomycin. COS-7 cells were transiently cotransfected with pcDEF3-ORF74 or pcDEF3 (mock) and pNF-B-Luc using the DEAE-dextran method. After transfection, cells were seeded in 96-well white plates (Costar, Corning, NY) in serum-free culture medium in the presence of indicated ligands. 48 h after transfection, cells were assayed for NF-B-driven luciferase expression. Luminescence was measured in a Wallac Victor (PerkinElmer).
Antibacterial Activity—Radial diffusion assays (RDAs) were performed as previously described (19), using Escherichia coli ATCC 054 and Listeria monocytogenes EGD. The underlay consisted of 1% agarose and 1:100 dilution of trypticase soy broth in 10 mM sodium phosphate, pH 7.4, containing 1 x 106 bacteria. Within the solidified agarose, a series of 3-mm diameter wells were punched and 5 µl of chemokines were added. Plates were incubated for 3 h at 37 °C to allow peptide diffusion. The underlay was then covered with overlay (6% trypticase soy broth agar and 1% agarose in H2O), and the plates were allowed to solidify. Antimicrobial activity was determined as a clear zone around the well after 16-h incubation at 37 °C, and are represented in radial diffusion units: (diameter of clear zone in millimeters – well diameter) x 10.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
-stimulated Primary Human Keratinocytes—We have previously shown that stimulation of human keratinocytes with IFN- induces the production of the CXCR3 targeting chemokines CXCL9, -10, and -11 (24). Following heparin affinity chromatography these chemokines can be purified by reversed-phase chromatography (Fig. 1A). Interestingly, upon mass spectrometric analysis of HPLC-purified keratinocyte-derived CXCL10, we found two different molecules of CXCL10, one complete form (1–77) and a C-terminally truncated form (1–73) that lacks the last four amino acids Lys-Arg-Ser-Pro (Fig. 1B).
|
To explore the possibility that CXCL10 is processed by such a system, we first treated IFN--stimulated keratinocytes with the cell permeable pro-protein convertase inhibitor decanoyl-RVKR-chloromethylketone and determined the form of CXCL10 produced by these cells. As predicted, the production of CXCL10-(1–73) was almost completely inhibited by decanoyl-RVKR-chloromethylketone (Fig. 1C).
It has recently been shown that several chemokines are substrates for matrix metalloproteinases (MMP) (3, 25). Several of these MMPs are activated upon cleavage by furin (26). This raised the possibility that inhibition of furin could have indirectly inhibited the production of CXCL10-(1–73) by preventing activation of MMPs. We therefore also treated IFN--stimulated keratinocytes with the broad range MMP inhibitor Marimastat (20 µg/ml). This treatment did not have any effect on the relative amounts of CXCL10-(1–77) and CXCL10-(1–73) that were produced by these cells (Fig. 1D).
To exclude the possibility that the appearance of CXCL10-(1–73) in the medium of IFN- stimulated keratinocytes was an artifact of culturing conditions, we also cultured keratinocytes in medium (Defined SFM, Invitrogen) that did not contain bovine pituitary extract. After stimulation with IFN- these cells produced similar amounts of CXCL10-(1–73) as shown in Fig. 1B.
In Vitro Cleavage of Human CXCL10 with Furin and Carboxypeptidase B—To directly demonstrate that human CXCL10 is a substrate for furin, we performed in vitro digestions with commercially available human CXCL10. This CXCL10 contains an extra N-terminal methionine (Met-CXCL10-(1–78)). We first treated a 2 µM solution of Met-CXCL10-(1–78) with 5 units/ml of furin at 37 °C for 3 h and compared this with CXCL10 left at 37 °C without furin. The mass of Met-CXCL10-(1–78) is 8773 Da (Fig. 2A, Table I). After treatment with furin, the mass of the recombinant CXCL10 shifted to 8589 Da (Fig. 2B, Table I). This matches with a Met-CXCL10-(1–76) molecule lacking the last two amino acids Ser-Pro, resulting from cleavage after the dibasic sequence Lys-Arg preceding these two C-terminal amino acids. This is exactly as would be expected from furin-mediated cleavage of CXCL10. Upon treatment of the furin processed CXCL10 with the basic specific carboxypeptidase B in vitro (0.5 units/ml), this molecule was further processed to Met-CXCL10-(1–74) with a mass of 8305 Da (Fig. 2C). This matches with the removal of the basic amino acids Lys-Arg, the two C-terminal amino acids of CXCL10 after treatment with furin (Table I). Intact human CXCL10 was insensitive to treatment by carboxypeptidase B alone (data not shown). Altogether, these data show that sequential treatment of CXCL10 with furin followed by a basic specific carboxypeptidase results in the formation of CXCL10 lacking the last four amino acids.
|
|
|
CXCL10 Production in Cytokine-stimulated Primary Fibroblasts—Unlike other pro-protein convertases, furin is widely expressed in different cells and tissues. To determine whether processing of CXCL10 is keratinocyte specific or also occurs in another cell type, we stimulated primary human dermal fibroblasts with IFN- and TNF- at concentrations that have previously been shown to highly induce mRNA levels of the CXCR3 targeting chemokines in these cells (28). Following heparin affinity chromatography, chemokines were purified using reversed phase chromatography (Fig. 4A). Mass spectrometric analysis of a CXCL10-containing fraction (indicated with a bar in the chromatogram), showed that also in these cells CXCL10 was (partially) processed to a molecule with a mass of 8173 Da (Fig. 4B), corresponding to CXCL10-(1–73). In contrast to keratinocyte-derived CXCL10, however, most of the fibroblast produced CXCL10 was intact (1–77). In addition, masses corresponding to N-terminally truncated CXCL10-(3–77) and CXCL10-(3–73) were identified, most probably reflecting N-terminal processing of CXCL10 by dipeptidylpeptidase IV.
|
Furin-mediated Cleavage of Other CXCR3-targeting Chemokines—In contrast to human CXCL10, human CXCL9 and -11 do not contain the furin consensus sequence Arg-Xaa-(Lys/Arg)-Arg, although they contain many basic residues in their C-terminal region. As it has been shown that also proteins containing sequences that do not fully fit the consensus sequence can be substrates for furin, we also incubated 2 µM solution of recombinant human CXCL9 and -11 with 5 units/ml furin at 37 °C for 3 h. This demonstrated that human CXCL11 is not a substrate for furin (Table I). On the other hand, human CXCL9 was cleaved at five different positions in the C-terminal part of the molecule, all after a basic amino acid residue (Fig. 5). Under identical conditions used for the complete processing of human CXCL10 (Fig. 2), however, human CXCL9 was only partially processed.
|
Activation of CXCR3—The ability of the novel CXCL10 variant to activate CXCR3 was tested using three different bioassays. First, Met-CXCL10 stimulated GTP binding to G in CXCR3-transfected CHO cells was determined using a GTPS binding assay (Fig. 6A). This revealed that Met-CXCL10-(1–74) retained full activity on CXCR3. Similar results were obtained in Ca2+ mobilization in CXCR3-transfected CHO cells (Fig. 6B) and chemotaxis of primary human T-cells (Fig. 6C). Overall, the above results show that the agonistic properties of Met-CXCL10-(1–78) and -(1–74) for CXCR3 are similar, indicating that the C-terminal part of CXCL10 is not essential for proper CXCR3 activation.
|
Interestingly, CXCL10 has been shown to be an inverse agonist on this receptor (30). Although the N-terminal amino acids of CXCL10 have been shown to be essential for CXCR3 activation (31), this is not the case for the inhibitory effect on the constitutive activity of ORF74 (30), indicating that other determinants might be involved in this activity. Similarly, we found that Met-CXCL10-(1–74) has still potent inverse agonistic activities on ORF74 (Fig. 7).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
It is at present well established that the endopeptidase dipeptidylpeptidase IV (CD26) cleaves many chemokines having a penultimate proline or alanine at the N terminus, including the CXCR3 ligands (2, 22, 31). Recently, the N-terminal region of chemokines has been shown to be sensitive to cleavage by matrix metalloproteinases (MMP). MMP-9- and MMP-2-mediated cleavage of CCL7 and CXCL12, respectively, negatively modulates their chemokine receptor-mediated activity (3, 33) while CXCL8 is highly potentiated after cleavage by MMP-9 (25). Other proteases involved in processing of chemokines include cathepsins (4, 6), elastase (5) urokinase plasminogen activator, and plasmin (8).
In this study we have described furin as a novel chemokine-modifying enzyme. Furin cleaves CXCL10 in vitro at a furin consensus site, resulting in a CXCL10 molecule that is sensitive to a basic amino acid-specific carboxypeptidase. Upon trimming by such a carboxypeptidase in vitro, a CXCL10 molecule is created that lacks the last four amino acids. This C-terminally truncated CXCL10 molecule is also produced in vivo by cytokine stimulated primary keratinocytes and fibroblasts. A pro-protein convertase inhibitor, decanoyl-RVKR-chloromethylketone, prevents the production of this C-terminally truncated CXCL10, indicating that furin is also involved in the in vivo cleavage of CXCL10.
Furin belongs to the group of pro-protein convertases (PCs). These enzymes are involved in the production of many biologically active proteins like e.g. growth factors, hormones, receptors, and matrix metalloproteinases (26, 34). In general, these active proteins result from limited processing at paired basic residues within higher molecular mass precursors. In the case of neuropeptides, proteolysis often occurs in secretory vesicles that are secreted upon the proper stimulus. The expression of furin is found in the trans-Golgi network of many cells and tissues and plays a role in the processing of precursors for constitutively secreted proteins. Some PCs are specific for certain cells and tissues. In the epidermis, expression of furin, PACE4, PC5/6, and PC7/8 has been identified (35). Because the pro-protein convertase inhibitor used in our studies also inhibits other PCs, we cannot exclude the possibility that other PCs are involved in the CXCL10 processing in vivo.
Human and murine CXCL10 (but also e.g. hamster, rat, sheep, goat, and monkey) contain the furin consensus sequence Arg–4-Xaa-Lys–2-Arg–1, and we demonstrated that both are processed by furin. In a survey of the primary structures of other human chemokines we found no perfect match with a furin consensus sequence in the C-terminal region. Our results with the in vitro processing of human CXCL9 however showed, that also sequences that do not fully comply with the consensus sequence are potential furin cleavage sites, though only partial processing was observed. Consequently, more chemokines than solely predicted on this consensus sequence may be substrates for furin in vitro. Using cytokine-stimulated keratinocytes, we only detected full-length CXCL9 (22) and did not find evidence for furin-mediated processing of CXCL9 in vivo. Others have found C-terminal truncations of CXCL9 (36) and although the mechanism by which these are formed is unknown, it most likely does not reflect furin-mediated cleavage.
In the case of CXCL9, furin cleavage in vitro was most efficient at paired basic residues. Interestingly, cleavage also occurred after lysine residues, although site-directed mutagenesis experiment with several substrates and structural analysis of furin suggested that an arginine at this position is essential (26, 37). A basic amino acid at position –4 is also favorable for furin-mediated cleavage. In the case of human CXCL10, the importance of this residue is underscored by the observation that the mutant CXCL10 with the Arg to Met substitution at this position was not cleaved by furin in vitro (Table I). Moreover, mouse CXCL10 contains two pairs of dibasic residues in the C-terminal part of the protein, with Lys71-Arg72 not having a basic residue at position –4. This dibasic sequence was insensitive to furin cleavage in vitro, while the Lys74-Arg75 was fully processed (Table I).
The fact that only 
40% of CXCL10 from cytokine-stimulated keratinocytes is processed indicates that the activity of furin in these cells is limited. In contrast, the fact that we have never identified a native CXCL10 molecule corresponding to CXCL10-(1–75) indicates that the trimming of the basic residues by a carboxypeptidase is not a rate-limiting step in vivo. The cleavage of CXCL10 by furin seems to be very efficient as the Km (5 µM) is much lower then identified for other substrates (38).
In contrast to the N-terminal truncations, there are only a few examples of natural occurring C-terminal truncations of chemokines having effect on chemokine receptor activation. In general, the N terminus and the N-loop are considered to form the most important chemokine binding and activation domains (39). An exception to this rule is CXCL7, for which a more potent CXCR2 agonist was identified after cleavage of several C-terminal amino acids by an as yet unknown mechanism (11). In the case of CXCL10, full activity on CXCR3 was found for the C-terminally truncated variant (Fig. 6). This indicates that furin-mediated processing of CXCL10 does not impair nor enhance the activity on CXCR3. This is in concordance with recent data on the interaction of CXCL10 with its receptor. Mutational analysis studies have focused on specific residues in CXCL10 that are important in the CXCR3 activating activity of CXCL10 (40). Among others, especially the arginine at position 8 was shown to be very important.
Another G-protein-coupled receptor with CXCL10 binding properties is encoded by the Kaposi's-associated human herpes virus 8. This receptor displays a high level of constitutive activity (41), which appears an important determinant in the pathogenesis of Kaposi's sarcoma. Although several chemokines (e.g. CXCL1) are agonists for this receptor, CXCL10 has been shown to be a potent inverse agonist (30). In this case, the N terminus of CXCL10 has been shown to be of minor importance (30) and our studies have now shown that also the C-terminal four amino acids are not essential for this inverse agonistic activity.
Some chemokines possess activities that are not mediated through G-protein-coupled receptor activation. It has for example recently become apparent that certain chemokines possess a direct antibacterial activity, pointing to the fact that these proteins play a role in both the innate and acquired immunity. Similar dual activities have also been assigned to a group of peptides, the 
-defensins, which were at first defined as antibacterial proteins (42). Several overlapping structural domains between chemokines and defensins have been implicated in their common antibacterial activity, which showed that especially the -helix is of major importance (32). Fascinatingly, in the case of CXCL7, a yet undefined mechanism in platelets can release the last two C-terminal amino acids resulting in a very potent antibacterial protein, previously known as the thrombocidins. This points to an essential role of the C terminus in the microbicidal activity of this chemokine (10). Direct antibacterial activity for the CXCR3-targeting chemokines has only recently been recognized (19). In contrast to the results with CXCL7, deletion of amino acid residues at the C terminus of CXCL10 does not alter its antibacterial activity toward E. coli and L. monocytogeneses.
Because the stretches of basic amino acids in the C terminus of chemokines are generally believed to be involved in the affinity for heparin, one possibility could be that the truncated CXCL10, that has lost two basic amino acids, has altered heparin binding affinities. Recent studies performed with a large set of CXCL10 mutants however, suggests that the last four amino acids are not essential for heparin binding (40). Moreover, in our own chemokine purification procedure we have also applied salt gradients to elute chemokines from the heparin affinity column. In these cases, both CXCL10 variants always co-eluted (data not shown).
Truncated CXCL10 might have an altered affinity for the recently described CXCR3B receptor (43), which is implicated in the angiogenic activities of the CXCR3 ligands. However, CXCR3B differs from CXCR3 only in its N-terminal part and as this region is probably only involved in recognizing the N terminus of the ligands this possibility seems less plausible. Moreover, genomic data from the CXCR3 locus in mouse indicates that there is no CXCR3B in this species and therefore a similar mechanism, expected for the conserved processing site, would be absent in this species.
Since all ligands for CXCR3 are natural antagonists for CCR3 (44), truncated CXCL10 might have an increased or impaired affinity for CCR3.
Finally, truncated CXCL10 might affect hitherto unexpected biological responses like recently shown for CXCL12, where MMP-2-mediated cleavage gives rise to a highly neurotoxic variant implicated in (CXCR4-independent) neurodegeneration (9). Interestingly, there is evidence for a CXCL10 receptor that is not CXCR3 on a set of different cell lines (45). Its cloning and characterization would open the possibility to study the role of the C-terminally processed CXCL10 in activating this novel receptor.
In conclusion, we have identified furin as a novel chemokine-modifying enzyme in vitro and most probably also in vivo. This conversion generates a CXCL10 variant that retains full agonistic activity for CXCR3, inverse agonistic activities on ORF74 and antibacterial properties toward E. coli and L. monocytogenes. The biological significance of furin-mediated CXCL10 truncation is as yet not elucidated. The high conservation of the furin consensus site in CXCL10 from different species implies a specific function for furin-processed CXCL10.
|
FOOTNOTES |
|---|
To whom correspondence should be addressed: Leiden University Medical Center, Sylvius Laboratory, Department of Dermatology, Wassenaarseweg 72, 2333 AL Leiden, The Netherlands. Tel.: 31-71-5271903; Fax: 31-71-5271910; E-mail: C.P.Tensen{at}lumc.nl.
1 The abbreviations used are: IFN-, interferon , CHO, Chinese hamster ovary; GTPS, guanosine 5'-3-O-(thio)triphosphate; ORF, open reading frame; TNF, tumor necrosis factor; PC, pro-protein convertases.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  P. Proost, T. Loos, A. Mortier, E. Schutyser, M. Gouwy, S. Noppen, C. Dillen, I. Ronsse, R. Conings, S. Struyf, et al. Citrullination of CXCL8 by peptidylarginine deiminase alters receptor usage, prevents proteolysis, and dampens tissue inflammation J. Exp. Med., August 18, 2008; (2008) jem.20080305. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. H. Cox, R. A. Dean, C. R. Roberts, and C. M. Overall Matrix Metalloproteinase Processing of CXCL11/I-TAC Results in Loss of Chemoattractant Activity and Altered Glycosaminoglycan Binding J. Biol. Chem., July 11, 2008; 283(28): 19389 - 19399. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Hasan, L. Mazzucchelli, M. Liebi, M. Lis, R. E. Hunger, A. Tester, C. M. Overall, and M. Wolf Function of Liver Activation-Regulated Chemokine/CC Chemokine Ligand 20 Is Differently Affected by Cathepsin B and Cathepsin D Processing. J. Immunol., June 1, 2006; 176(11): 6512 - 6522. [Abstract] [Full Text] [PDF]
|
|
| ||||||
