Stromelysin-1 (matrix metalloproteinase-3: MMP-3) occupies a central position in collagenolytic and elastolytic cascades, leading to cutaneous intrinsic and extrinsic aging. We screened extracts of a propolis sample from Algeria with the aim to isolate compounds able to selectively inhibit this enzyme. A butanolic extract (B3) of the investigated propolis sample was found to potently inhibit MMP-3 activity (IC50 = 0.15 ± 0.03 µg/mL), with no or only weak activity on other MMPs. This fraction also inhibited plasmin amidolytic activity (IC50 = 0.05 µg/mL) and impeded plasmin-mediated proMMP-3 activation. B3 was fractionated by HPLC, and one compound, characterized by NMR and mass spectroscopy and not previously identified in propolis, i.e., (+)-chicoric acid, displayed potent in vitro MMP-3 inhibitory activity (IC50 = 6.3 × 10−7 M). In addition, both caffeic acid and (+)-chicoric acid methyl ester present in fraction B3 significantly inhibited UVA-mediated MMP-3 upregulation by fibroblasts.
Les peptides d'élastine (kE) augmentent l'expression des MMP3 et MMP1 dans les fibroblastes gingivaux humains cultivés en 2D sur plastique alors qu'ils n'ont aucune influence sur la production d'uPA, des MMPs2, 13, 14 et des TIMPs1 et 2. L'effet de kE sur la production de MMP3 dépend du récepteur membranaire S-Gal tandis qu'il est inhibé par le lactose et reproduit par des peptides contenant la séquence VGVAPG capable de se fixer sur S-Gal. Afin d'évaluer l'implication des peptides d'élastine dans la collagénolyse, des cultures cellulaires en 3D ou lattis de collagène attaché sont utilisés : kE ou le plasminogène (Plgn) seuls n'induisent pas de façon significative une collagénolyse des lattis de collagène. Toutefois, kE, en présence de Plgn, augmente de façon synergique la dégradation du collagène et l'activation des MMPs. Au sein des lattis en présence de peptides d'élastine, la collagénolyse serait dépendante d'une sur-expression de la cascade protéolytique MMP3/MMP1 par les fibroblastes gingivaux humains, associée à une sous-expression des TIMP1 et de TIMP2. (Med Buccale Chir Buccale 2009 ; 15 : 75-85).
The matrix metalloproteinases gelatinase A (MMP-2) and gelatinase B (MMP-9) are implicated in the physiological and pathological breakdown of several extracellular matrix proteins. In the present study, we show that long-chain fatty acids (e.g. oleic acid, elaidic acid, and cis- and trans-parinaric acids) inhibit gelatinase A as well as gelatinase B with Ki values in the micromolar range but had only weak inhibitory effect on collagenase-1 (MMP-1), as assessed using synthetic or natural substrates. The inhibition of gelatinases depended on fatty acid chain length (with C18 > C16, C14, and C10), and the presence of unsaturations increased their inhibitory capacity on both types of gelatinase. Ex vivoexperiments on human skin tissue sections have shown that micromolar concentrations of a long-chain unsaturated fatty acid (elaidic acid) protect collagen and elastin fibers against degradation by gelatinases A and B, respectively. In order to understand why gelatinases are more susceptible than collagenase-1 to inhibition by long-chain fatty acids, the possible role of the fibronectin-like domain (a domain unique to gelatinases) in binding inhibitory fatty acids was investigated. Affinity and kinetic studies with a recombinant fibronectin-like domain of gelatinase A and with a recombinant mutant of gelatinase A from which this domain had been deleted pointed to an interaction of long-chain fatty acids with the fibronectin-like domain of the protease. Surface plasmon resonance studies on the interaction of long-chain fatty acids with the three individual type II modules of the fibronectin-like domain of gelatinase A revealed that the first type II module is primarily responsible for binding these compounds. The matrix metalloproteinases gelatinase A (MMP-2) and gelatinase B (MMP-9) are implicated in the physiological and pathological breakdown of several extracellular matrix proteins. In the present study, we show that long-chain fatty acids (e.g. oleic acid, elaidic acid, and cis- and trans-parinaric acids) inhibit gelatinase A as well as gelatinase B with Ki values in the micromolar range but had only weak inhibitory effect on collagenase-1 (MMP-1), as assessed using synthetic or natural substrates. The inhibition of gelatinases depended on fatty acid chain length (with C18 > C16, C14, and C10), and the presence of unsaturations increased their inhibitory capacity on both types of gelatinase. Ex vivoexperiments on human skin tissue sections have shown that micromolar concentrations of a long-chain unsaturated fatty acid (elaidic acid) protect collagen and elastin fibers against degradation by gelatinases A and B, respectively. In order to understand why gelatinases are more susceptible than collagenase-1 to inhibition by long-chain fatty acids, the possible role of the fibronectin-like domain (a domain unique to gelatinases) in binding inhibitory fatty acids was investigated. Affinity and kinetic studies with a recombinant fibronectin-like domain of gelatinase A and with a recombinant mutant of gelatinase A from which this domain had been deleted pointed to an interaction of long-chain fatty acids with the fibronectin-like domain of the protease. Surface plasmon resonance studies on the interaction of long-chain fatty acids with the three individual type II modules of the fibronectin-like domain of gelatinase A revealed that the first type II module is primarily responsible for binding these compounds. matrix metalloproteinase tissue inhibitor of metalloproteinases fibronectin type II progelatinase A mutant with amino acids Val191–Gln364 deleted (7-methoxycoumarin-4-yl) acetyl [3-(2′,4′-dinitrophenyl)-l-2,3-diaminopropionyl] p-aminophenylmercuric acetate dimethyl sulfoxide surface plasmon resonance Matrix metalloproteinases (MMPs)1 compose a family of at least 23 related zinc-dependent endopeptidases (1Nagase H. Woessner Jr., J.F. J. Biol. Chem. 1999; 274: 21491-21494Abstract Full Text Full Text PDF PubMed Scopus (3903) Google Scholar) that are collectively able to degrade extracellular matrix proteins such as collagens, laminins, fibronectin, elastin, and proteoglycans. They are consequently implicated in physiological remodeling of connective tissue occurring in embryonic development and repair (2Adler R.R. Brenner C.A. Werb Z. Development. 1990; 110: 211-220PubMed Google Scholar, 3Behrendtsen O. Alexander C.M. Werb Z. Development. 1992; 114: 447-456PubMed Google Scholar, 4Kahari V.M. Saarialho-Kere U. Exp. Dermatol. 1997; 6: 199-213Crossref PubMed Scopus (512) Google Scholar). Most of them are secreted as inactive proenzymes and are then extra- or pericellularly activated by other MMPs or serine proteinases (5Nagase H. Biol. Chem. 1997; 378: 151-160PubMed Google Scholar). Their catalytic activities are strictly controlled by endogenous specific inhibitors designated as tissue inhibitors of metalloproteinases (TIMPs) (6Brew K. Dinakarpandian D. Nagase H. Biochim. Biophys. Acta. 2000; 1477: 267-283Crossref PubMed Scopus (1610) Google Scholar) and also α2-macroglobulin (7Barrett A.J. Starkey P.M. Biochem. J. 1973; 133: 709-724Crossref PubMed Scopus (888) Google Scholar). The balance between activated MMPs and TIMPs determines the overall MMP proteolytic activity and consequently the extent of extracellular matrix degradation. Local disruption of the MMP-TIMP balance can lead to pathological degradative processes including rheumatoid arthritis, atherosclerosis, tumor growth, and metastasis (8Carmeliet P. Collen D. Thromb. Res. 1998; 91: 255-285Abstract Full Text Full Text PDF PubMed Scopus (165) Google Scholar, 9Dean D.D. Martel-Pelletier J. Pelletier J.-P. Howell D.S. Woessner Jr., J.F. J. Clin. Invest. 1989; 84: 678-685Crossref PubMed Scopus (545) Google Scholar, 10Coussens L.M. Werb Z. Chem. Biol. 1996; 3: 895-904Abstract Full Text PDF PubMed Scopus (506) Google Scholar). MMPs are multidomain enzymes containing propeptide, catalytic and, except matrilysin (MMP-7), MMP-23, and endometase/matrilysin-2 (MMP-26), hemopexin-like domains. Gelatinase A (MMP-2) and gelatinase B (MMP-9) contain in addition three tandem copies of a 58-amino acid fibronectin type II (FN-II) module (11Collier I.E. Wilhelm S.M. Eisen A.Z. Marmer B.L. Grant G.A. Seltzer J.L. Kronberger A. He C. Bauer E.A. Goldberg G.I. J. Biol. Chem. 1988; 263: 6579-6587Abstract Full Text PDF PubMed Google Scholar, 12Wilhelm S.M. Collier I.E. Marmer B.L. Eisen A.Z. Grant G.A. Goldberg G.I. J. Biol. Chem. 1989; 264: 17213-17221Abstract Full Text PDF PubMed Google Scholar) inserted within their catalytic domain (13Morgunova E. Tuutila A. Bergmann U. Isupov M. Lindqvist Y. Schneider G. Tryggvason K. Science. 1999; 284: 1667-1670Crossref PubMed Scopus (485) Google Scholar). The basic fold of the FN-II modules is composed of a pair of β sheets that form a hydrophobic pocket accessible to solvent (13Morgunova E. Tuutila A. Bergmann U. Isupov M. Lindqvist Y. Schneider G. Tryggvason K. Science. 1999; 284: 1667-1670Crossref PubMed Scopus (485) Google Scholar). The FN-II repeats confer high affinity binding of these enzymes to gelatin and insoluble elastin (14Banyai L. Tordai H. Patthy L. Biochem. J. 1994; 298: 403-407Crossref PubMed Scopus (70) Google Scholar, 15Murphy G. Nguyen Q. Cockett M.I. Atkinson S.J. Allan J.A. Knight C.G. Willenbrock F. Docherty A.J.P. J. Biol. Chem. 1994; 269: 6632-6636Abstract Full Text PDF PubMed Google Scholar, 16Shipley J.M. Doyle G.A.R. Fliszar C.J. Ye Q.-Z. Johnson L.L. Shapiro S.D. Welgus H.G. Senior R.M. J. Biol. Chem. 1996; 271: 4335-4341Abstract Full Text Full Text PDF PubMed Scopus (177) Google Scholar), a prerequisite for their efficient proteolysis. Long-chain unsaturated fatty acids inhibit the expression and activity of aggrecanases (17Curtis C.L. Hughes C.E. Flannery C.R. Little C.B. Harwood J.L. Caterson B. J. Biol. Chem. 2000; 275: 721-724Abstract Full Text Full Text PDF PubMed Scopus (230) Google Scholar). They bind to serine proteinases such as neutrophil elastase (18Ashe B.M. Zimmerman M. Biochem. Biophys. Res. Commun. 1977; 75: 194-199Crossref PubMed Scopus (72) Google Scholar, 19Tyagi S.C. Simon S.R. J. Biol. Chem. 1991; 266: 15185-15191Abstract Full Text PDF PubMed Google Scholar) and plasmin (20Higazi A.A.-R. Finci-Yeheskel Z. Samara A.A.-R. Aziza R. Mayer M. Biochem. J. 1992; 282: 863-866Crossref PubMed Scopus (22) Google Scholar, 21Higazi A.A.-R. Aziza R. Samara A.A.-R. Mayer M. Biochem. J. 1994; 300: 251-255Crossref PubMed Scopus (20) Google Scholar) and modulate their catalytic activities. Studies by Suzuki et al. (22Suzuki I. Iigo M. Ishikawa C. Kuhara T. Asamoto M. Kunimoto T. Moore M.A. Yazawa K. Araki E. Tsuda H. Int. J. Cancer. 1997; 73: 607-612Crossref PubMed Scopus (84) Google Scholar) show that oleic acid, 18-carbon fatty acid with one double carbon bond in the cis position, partially inhibits the formation of lung metastases from subcutaneous implantation of colon carcinoma cells in athymic mice. We recently reported that oleic acid inhibited in a dose-dependent manner the hydrolysis of the fluorogenic substrate Mca-l-Pro-Leu-Gly-Leu-Dpa-Ala-Arg-NH2 (23Knight C.G. Willenbrock F. Murphy G. FEBS Lett. 1992; 296: 263-266Crossref PubMed Scopus (679) Google Scholar) by gelatinase A (24Emonard H. Marcq V. Mirand C. Hornebeck W. Ann. N. Y. Acad. Sci. 1999; 878: 647-649Crossref PubMed Scopus (38) Google Scholar, 25Polette M. Huet E. Birembaut P. Maquart F.X. Hornebeck W. Emonard H. Int. J. Cancer. 1999; 80: 751-755Crossref PubMed Scopus (22) Google Scholar). We report here an in vitroinvestigation showing that fatty acids, depending on their chain length and degree of unsaturation, inhibit gelatinases A and B with similar efficiency but only have low inhibitory capacity toward collagenase-1. We conclude that interaction between FN-II domain and unsaturated fatty acids leads to gelatinase inhibition because of the following: (i) oleic acid displayed markedly weaker inhibitory capacity toward gelatinase A deleted in FN-II modules; (ii) fatty acids bound avidly to the first FN-II module; and (iii) this module totally prevented the oleic acid-mediated inhibition of full-length gelatinase A. The physiological relevance of our findings was substantiated by ex vivo experiments on human skin tissue sections demonstrating that C18-unsaturated fatty acid efficiently impeded the collagenolytic and elastolytic activities of gelatinases A and B, respectively. Human recombinant progelatinases A and B and natural procollagenase-1 were purchased from Calbiochem. The recombinant pro-ΔVal-191—Gln-364 gelatinase A (pro-ΔFN-II gelatinase A or truncated progelatinase A), corresponding to progelatinase A with FN-II domains deleted (15Murphy G. Nguyen Q. Cockett M.I. Atkinson S.J. Allan J.A. Knight C.G. Willenbrock F. Docherty A.J.P. J. Biol. Chem. 1994; 269: 6632-6636Abstract Full Text PDF PubMed Google Scholar), was kindly provided by Dr. G. Murphy (University of East Anglia, Norwich, UK). Human recombinant TIMP-2 was a gift from Prof. Y. A. DeClerck (Children's Hospital of Los Angeles, Los Angeles, CA). Acid-soluble type I collagen was extracted and purified from guinea pig skin as described previously (26Eeckhout Y. Delaissé J.M. Vaes G. Biochem. J. 1986; 239: 793-796Crossref PubMed Scopus (53) Google Scholar). Recombinant proteins coll 1, coll 2, coll 3, and coll 123 containing different segments of the gelatin-binding site of human gelatinase A were expressed in Escherichia coli and purified by gelatin-Sepharose 4B chromatography as described previously (14Banyai L. Tordai H. Patthy L. Biochem. J. 1994; 298: 403-407Crossref PubMed Scopus (70) Google Scholar,27Banyai L. Tordai H. Patthy L. J. Biol. Chem. 1996; 271: 12003-12008Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar). Human foreskins from healthy young children (10 months to 10 years old) were obtained on parental consent during surgical operations and kept frozen at −80 °C. Serial sections of 8 μm thickness were prepared with a cryostat microtome at −20 °C. The free acid forms of decanoic acid (capric acid, 10:0), tetradecanoic acid (myristic acid, 14:0),cis-9-tetradecenoic acid (myristoleic acid, c9–14:1), hexadecanoic acid (palmitic acid, 16:0), cis-9-hexadecenoic acid (palmitoleic acid, c9–16:1), octadecanoic acid (stearic acid, 18:0), cis-9-octadecenoic acid (oleic acid, c9–18:1), andtrans-9-octadecenoic acid (elaidic acid, t9–18:1), and an alcohol analogue of oleic acid were purchased from Sigma. Thecis-9, trans-11, trans-13,cis-15-octadecatetraenoic acid (cis-parinaric acid, c9, t11, t13, c15–18:4), and all-trans-9, -11, -13, -15- octadecatetraenoic acid (trans-parinaric acid, all-t9, 11, 13, 15–18:4) were from Molecular Probes (Interchim, Montluçon, France). The hydroxamate derivative of oleic acid was prepared as described previously (24Emonard H. Marcq V. Mirand C. Hornebeck W. Ann. N. Y. Acad. Sci. 1999; 878: 647-649Crossref PubMed Scopus (38) Google Scholar). The quenched fluorescent substrate Mca-PLGL(Dpa)-AR-NH2 was from Bachem (Voisins-le-Bretonneux, France). Full-length and ΔFN-II progelatinases A were fully activated by incubating the proenzymes in 150 mm NaCl, 10 mm CaCl2, 50 mm Tris-HCl, pH 7.4, with 1 mm p-aminophenylmercuric acetate (APMA) for 18 h at 4 °C. Active gelatinase B and collagenase-1 were obtained using 2 mm APMA in Tris buffer at 37 °C for 1 h. Full activation of each enzyme was assessed by gelatin zymography or Western blots. Each enzyme was further active site-titrated using a standard preparation of TIMP-2 (28Murphy G. Willenbrock F. Method Enzymol. 1995; 248: 496-510Crossref PubMed Scopus (243) Google Scholar). The inhibitory effect of fatty acids (from C10 to C18, either saturated or cis- ortrans-unsaturated) against gelatinase A (full-length or FN-II-deleted forms), gelatinase B, or collagenase-1 was analyzed using the fluorescent quenched substrate Mca-PLGL-(Dpa)-AR-NH2. Two hundred picomolar of each MMP species were preincubated for 15 min at 22 °C with 0–40 μm fatty acid in a 50 mm HEPES buffer, pH 7.5, containing 150 mmNaCl, and 5 mm CaCl2. The assays were initiated by adding 2 μm Mca-PLGL-(Dpa)-AR-NH2. The final concentration of dimethyl sulfoxide (Me2SO) used to dissolve fatty acid and fluorogenic substrate never exceeded 1% (v/v). The reaction was allowed to proceed at 22 °C for 20 min (gelatinase A), 60 min (gelatinase B), or 180 min (collagenase 1) and then was stopped by adding 10 mm EDTA. Under these experimental conditions, MMPs generated a similar intensity of fluorescence, allowing the comparison of the inhibition results. The effect of substitution of the carboxylic end by alcohol or hydroxamate group on the inhibitory potency of oleic acid against gelatinase A was similarly evaluated. The rate of substrate cleavage was measured in quadruplicate for each fatty acid or derivative concentration examined, using a Perkin Elmer LS 50B spectrofluorimeter with excitation and emission wavelengths of 325 and 387 nm, respectively. Less than 5% of the substrate was hydrolyzed during the rate measurements. Addition of fatty acid after the digestion of fluorogenic substrate had no effect on the fluorescent signal. Nonlinear regression analysis with the Grafit computer software (R. J. Leatherbarrow, Erithacus Software) allowed us to calculate the best estimates of the equilibrium dissociation constant of the enzyme-inhibitor complex or inhibition constant Ki, using the integrated Equation 1(29Bieth J.G. Method Enzymol. 1995; 248: 59-84Crossref PubMed Scopus (190) Google Scholar), vivo=1−([E]o+[I]o+Ki)−{([E]o+[I]o+Ki)2−4[E]o[I]o}1/22[E]oEquation 1 where vi is the rate of substrate hydrolysis in the presence of inhibitor; vo is the rate in its absence; and [E]o and [I]o are the initial concentrations of enzyme and inhibitor, respectively. The inhibitory effect of oleic, elaidic, or stearic acids (0–40 μm) was further evaluated against gelatinase A, using a natural substrate, i.e. heat-denatured [3H]collagen type I. Briefly, 1.6 nmgelatinase A was first mixed with 0–40 μm fatty acid in a 50 mm Tris-HCl buffer, pH 7.4, containing 150 mm NaCl and 10 mm CaCl2 for 15 min at 22 °C before incubation with radiolabeled natural substrate for 20 h at 37 °C (30Marbaix E. Donnez J. Courtoy P.J. Eeckhout Y. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 11789-11793Crossref PubMed Scopus (213) Google Scholar). The 50% inhibitory concentrations (IC50) of fatty acids for gelatinase A were determined using the Grafit computer software. For each enzyme, the mode of inhibition of oleic acid was analyzed with the Grafit computer software, using the graphical methods of either Dixon or Cornish-Bowden (see Ref. 31Knight C.G. Barrett A.J. Salvesen G. Proteinase Inhibitors. Elsevier Science Publishers B.V., Amsterdam1986: 23-51Google Scholar). The trimodular protein coll 123 contains the entire fibronectin-related part of human gelatinase A (27Banyai L. Tordai H. Patthy L. J. Biol. Chem. 1996; 271: 12003-12008Abstract Full Text Full Text PDF PubMed Scopus (46) Google Scholar). Oleic acid and stearic acid-coll 123 interaction studies were performed using a BIAcore X system (Amersham Pharmacia Biotech). Sensor chip HPA (BIAcore AB) was used for all experiments. A flow rate of 5 μl/min was used, and the instrument was thermostated at 25 °C. The surface of sensor chip was washed for 5 min with 40 mm n-octyl-β-d-glucopyranoside in water. Fatty acid (20 μl of an 8 mm solution in Me2SO) was then injected. The process of spontaneous fatty acid adsorption on the sensor chip was monitored. When the sensorgram reading began to level out, the flow rate was briefly increased to 100 μl/min to suppress the multiple lipid layers formed on the sensor chip surface. An additional injection of 10 mm NaOH (10 μl) was performed to regenerate the sensor chip and to reach a stable base line. To assess the extent of coverage of the sensor chip surface by both fatty acids, we injected 10 μl of 0.1 g/liter bovine serum albumin in 4.4 mm Na2HPO4 buffer, pH 7.4, 130 mm NaCl, 3 mm KCl. The amount of bovine serum albumin bound to the sensor chip surface corresponded to 43 resonance units, a value much lower than that typically obtained with a surface fully coated with dimyristoyl phosphatidylcholine or palmitoyloleoyl phosphatidylcholine. Binding experiments with coll 123 were performed in the same buffer. After each cycle, the sensor chip was regenerated by injecting 10 μl of 10 mm NaOH. These regeneration conditions allowed to restore the base-line level observed before each injection. The coll 123 binding kinetics were measured at six different concentrations of analyte, i.e. 3.125, 6.25, 12.5, 25, 50, and 100 μm. We similarly studied oleic acid interactions with the isolated FN-II modules of gelatinase A, i.e. coll 1, coll 2, or coll 3. The binding curves were analyzed using the nonlinear data fitting software BIAevaluation to obtain the rate constants of association (ka) and dissociation (kd) and the equilibrium constants of dissociation (KD). Oleic acid (5 μm) and gelatinase A (200 pm) were preincubated for 15 min at 22 °C with molar fold excess (1–100-fold in comparison with gelatinase A) of the trimodular protein coll 123 or the single module proteins coll 1, coll 2, or coll 3, and residual enzymatic activity was measured with Mca-PLGL- (Dpa)-AR-NH2 (2 μm), as described above. Whatever the concentration used, coll 123, coll 1, coll 2, or coll 3 did not exhibit inhibitory effect against gelatinase A. A set of three skin tissue sections was laid on a coated polylysine microscopic slide (Bio-Rad) and overlaid with 10 μl of 50 mm Tris-HCl buffer, pH 7.4, containing 150 mm NaCl and 5 mmCaCl2 (assay buffer) or the same buffer containing 50 nm APMA-activated gelatinase A or gelatinase B. The preparations were then incubated for 4 h at 37 °C in a moist chamber. After incubation, the tissue sections were rinsed and stained with polyphenolic catechin-fuschin and red sirius for staining of elastic and collagen fibers, respectively (32Berton A. Godeau G. Emonard H. Baba K. Bellon P. Hornebeck W. Bellon G. Matrix Biol. 2000; 19: 139-148Crossref PubMed Scopus (56) Google Scholar). Elastic fibers appeared in deep blue-black, and the background was poorly stained and collagen fibers were revealed in red-orange. This allowed a further quantitative estimation of the area (AA) and volume (VV) fractions occupied by these fibers. For elaidic acid inhibition assays, gelatinase A or gelatinase B were first incubated with 0, 1, or 10 μm fatty acid for 15 min at 22 °C; alternatively, tissue sections were preincubated with 1 or 10 μm elaidic acid or Me2SO-containing assay buffer as negative control for 30 min at 37 °C before adding gelatinase A or gelatinase B. The microscopic slides were then observed under a Zeiss standard 14 microscope equipped with CF 126 PHR video camera, and computerized analyses of the elastic and collagenic fibers were carried out as described previously (32Berton A. Godeau G. Emonard H. Baba K. Bellon P. Hornebeck W. Bellon G. Matrix Biol. 2000; 19: 139-148Crossref PubMed Scopus (56) Google Scholar). Briefly, three skin serial sections were used for each assay, and 10 fields of dimensions 0.7 × 0.7 mm were analyzed for each skin tissue section. Black and white images generated by the video camera were converted into 256 different gray levels using a Sophretec NUM 600 image memory, transferred to a BFM 186 microcomputer, and finally analyzed using a software for mathematical morphology. The area fraction (AA) occupied by the elastic and collagen fibers was automatically calculated. The area fraction represents the surface of fibers as a function of the tissue area analyzed. For the elastic fibers, the volume fraction (VV) was also determined and corresponds to AA × k, wherek <1 represents the Weibel correction factor (k = d/d + t, with dand t, elastic fiber diameter, and section thickness, respectively, in μm). The average diameter of the elastic fibers was obtained semiautomatically using skin tissue sections treated or not and a calibrated slide. The effect of fatty acids varying in chain length and degree of unsaturation was first evaluated on gelatinases A and B using the fluorogenic Mca-PLGL-(Dpa)-AR-NH2 substrate. TableI shows that both enzymes were inhibited with a similar efficiency by fatty acids. Inhibition depended on their alkyl chain length (with C18 > C16, C14, and C10). The presence of unsaturations in fatty acids increased the inhibition of both gelatinases. Furthermore, gelatinases A and B inhibition did not depend on the cis-trans configuration of double bond(s) since oleic acid and elaidic acid, its trans-counterpart, displayed similar Ki values; also, no difference was observed between cis- and trans-parinaric acids. Such polyunsaturated fatty acids were, however, more efficient as inhibitors toward gelatinase B versus gelatinase A (TableI).Table IInhibition of gelatinases A and B activities by fatty acidsLong-chain fatty acidsTrivial namesKiGelatinase AGelatinase Bμm10:0CapricNI aNI, no or less than 10% of inhibition, at [I]o = 40 μm.NI14:0MyristicNINIc9–14:1MyristoleicNINI16:0PalmiticNINIc9–16:1Palmitoleic29.3 ± 5.534.3 ± 4.918:0Stearic35.2 ± 5.647.4 ± 7.7c9–18:1Oleic4.3 ± 0.46.4 ± 0.4t9–18:1Elaidic4.4 ± 0.35.8 ± 0.3c9, t11, t13, c15–18:4cis-Parinaric8.1 ± 1.20.8 ± 0.3All t9, 11, 13, 15–18:4trans-Parinaric5.7 ± 0.41.9 ± 0.6Fatty acids were first incubated with gelatinases A and B before adding fluorogenic substrate. The equilibrium dissociation constantKi and its standard error were calculated by nonlinear regression analysis using Equation 1, and the given value corresponds to the best estimate obtained from one fit curve among three independent experiments; the S.D. between the three experiments is less than 15%.a NI, no or less than 10% of inhibition, at [I]o = 40 μm. Open table in a new tab Fatty acids were first incubated with gelatinases A and B before adding fluorogenic substrate. The equilibrium dissociation constantKi and its standard error were calculated by nonlinear regression analysis using Equation 1, and the given value corresponds to the best estimate obtained from one fit curve among three independent experiments; the S.D. between the three experiments is less than 15%. When they were tested against collagenase-1, only unsaturated fatty acids with 18 carbon atoms exhibited a low inhibitory effect, with aKi value equal to 59.6 ± 5.7 μmfor oleic acid. Contrary to data obtained with gelatinases, atrans-configuration of the unsaturation slightly improved collagenase-1 inhibition; 40 μm of oleic or elaidic acids inhibited by 30 and 40%, respectively, the degradation of fluorogenic substrate by collagenase-1. Changing carboxylic end group of oleic acid to hydroxamic group, a more potent bidentate ligand for catalytic zinc, did not dramatically improve gelatinase A inhibition, with the Ki value only decreasing from 4.3 ± 0.4 to 1.8 ± 0.2 μm(Fig. 1). Also, replacement of carboxylic group by an alcohol group, not considered as a zinc ligand, did not strikingly impair the inhibitory activity of oleic acid, with theKi value only rising from 4.3 ± 0.4 to 6.3 ± 0.3 μm (Fig. 1). Our data suggested that FN-II repeats, present in gelatinases A and B but absent in collagenase- 1, could represent selective targets for fatty acids. We therefore evaluated the inhibitory capacity of oleic, elaidic, and stearic acids against gelatinase A deleted in FN-II repeats. Fig.2 illustrated the weak inhibition of oleic and elaidic acids toward truncated gelatinase A; as shown above for collagenase-1, oleic acid was less efficient than itstrans-counterpart, elaidic acid, with Kivalues of 32.5 ± 3.0 and 12.3 ± 1.6 μm, respectively. Again, stearic acid did not inhibit truncated gelatinase A (Fig. 2). The mode of inhibition of each enzyme by oleic acid was evaluated. In both cases, Dixon plots (data not shown) were consistent with oleic acid acting as either competitive or mixed inhibitor, with Ki values graphically estimated to 1.6 and 26.3 μm for full-length and truncated gelatinases A, respectively. Cornish-Bowden plot allowed us to distinguish between these two possibilities (Fig. 3). The intersection of lines above the abscissa in the Cornish-Bowden plot was consistent with a mixed mode of inhibition of oleic acid against the full-length gelatinase A (Fig. 3 A), whereas it appeared to act as a weak competitive inhibitor toward truncated gelatinase A (Fig.3 B). Potential interactions between fatty acids and the gelatinase A FN-II repeats were studied by SPR analyses using a recombinant peptide corresponding to the three FN-II modules (coll 123) of gelatinase A and oleic or stearic acids. Fatty acids were linked to the surface of a sensor cell as described under "Experimental Procedures." Solutions of coll 123 were allowed to bind to the immobilized oleic acid or stearic acids, and data were analyzed as a function of time. The results of typical binding assays are shown in Fig.4, A and B. The association rate constant (ka), dissociation rate constant (kd), and dissociation equilibrium constant (KD) of coll 123 for oleic acid or stearic acid were calculated from SPR analyses, using the nonlinear data fitting software BIAevaluation (Table II). Oleic and stearic acids bound coll 123, with KD of 41 ± 1 and 36.1 ± 2 μm, respectively. Furthermore, SPR analysis demonstrated that oleic acid interacted with the first FN-II module, coll 1, with a KD value of 62 ± 4 μm, close to that obtained with the entire FN-II domain, coll 123; on contrary, the second and third FN-II modules weakly bound oleic acid with dissociation constants of 210 ± 90 and 3900 ± 1400 μm, respectively.Table IIKinetic and equilibrium constants of FN-II repeats of gelatinase A with oleic or stearic acidsFatty acidskakdKDm−1s−1s−1μmOleic acid3.9 ± 0.1 × 10216.0 ± 0.3 × 10−341 ± 1Stearic acid6.1 ± 0.1 × 10222.0 ± 0.7 × 10−336.1 ± 2Binding of the trimodular protein coll 123 to oleic or stearic acids was monitored by SPR. Binding constants were calculated using the BIAcore evaluation software. The errors for the kaand kd rate constants are expressed as the standard deviation of the slope and the standard deviation of six analyte concentrations, respectively, from two separate experiments. The errors for the KD values represent the sum of the errors from the ka and kd values. Open table in a new tab Binding of the trimodular protein coll 123 to oleic or stearic acids was monitored by SPR. Binding constants were calculated using the BIAcore evaluation software. The errors for the kaand kd rate constants are expressed as the standard deviation of the slope and the standard deviation of six analyte concentrations, respectively, from two separate experiments. The errors for the KD values represent the sum of the errors from the ka and kd values. We next investigated whether the ability of oleic acid to bind FN-II modules was related to its gelatinase inhibitory capacity. Gelatinase A activity was measured in the presence of increasing concentrations of trimodular protein coll 123 and an oleic acid concentration producing 50% inhibition, i.e. 5 μm. Increasing concentrations of coll 123 relieved the inhibition of gelatinase A, and the inhibitory effect of oleic acid was totally abolished by a 100-fold molar excess of coll 123 (Fig.5 A). In the same way, we analyzed the effect of isolated FN-II modules, i.e. coll 1, coll 2, and coll 3, on gelatinase A inhibition by oleic acid. A 50-fold molar excess of coll 1 restored up to 95% gelatinase A activity in the presence of 5 μm oleic acid (Fig. 5 B). In contrast, only a slight or no restoration of gelatinase A activity was observed in the presence of coll 2 or coll 3. Denatured collagen (gelatin) represents physiological substrate for gelatinase A. We also evaluate
