HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 31, 27887-27895, August 2, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/31/27887    most recent M202998200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Schröder, J.-M. Articles by Mallet, A. I. Search for Related Content PubMed PubMed Citation Articles by Schröder, J.-M. Articles by Mallet, A. I. Social Bookmarking                      What's this?

Identification of Diacylated Ureas as a Novel Family of Fungus-specific Leukocyte-activating Pathogen-associated Molecules^*

Jens-Michael Schröder^§¶, Robert Häsler^§, Jörg Grabowsky, Barbara Kahlke, and Anthony I. Mallet

From the  Clinical Research Unit "Cutaneous Inflammation," Department of Dermatology, University Hospital Kiel, Schittenhelmstr. 7, D-24105 Kiel, Germany and the  School of Chemical and Life Sciences, University of Greenwich, SE18 6PF London, United Kingdom

Received for publication, March 28, 2002, and in revised form, May 13, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

Polymorphonuclear leukocytes represent primary components of the host's innate immune defenses against fungal infection, suggesting involvement of fungal leukocyte attractants. We have found in various fungi, but not in bacteria or host cells, unstable lipid-like leukocyte chemoattractants, which also induced adherence and degranulation in human neutrophils. Purification from bakers' yeast and structural analyses by electrospray mass spectrometry, ¹H NMR spectroscopy, and chemical synthesis revealed these inflammatory mediators as diacylated ureas, a novel class of unstable lipoids. The N,N'-dipalmitoleyl urea appeared to be the most potent innate immune responses inducing compound eliciting half-maximum neutrophil chemotactic activity at 140 nM. The all-trans isomer, N,N'-dipalmitelaidyl urea, was found to be inactive with respect to stimulation of degranulation in neutrophils, which indicates a ₉ cis-double bond to be essential for bioactivity of these diacyl ureas. N,N'-Dipalmitoleyl urea elicited Ca²⁺ mobilization in neutrophils, which was found to be pertussis toxin-sensitive and sensitive toward a carboxylmethyltransferase inhibitor, indicating that these diacyl ureas activate leukocytes via a putative G_i-protein-coupled receptor. Their isolation exclusively from fungi suggests that these lipoids are fungus-specific pathogen-associated molecules that may alert the human innate immunity system to the presence of a fungal infection.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

The survival of multicellular organisms depends on their ability to recognize invading microbial pathogens and to induce appropriate host defense reactions. Pathogen-associated molecules (PAMs)¹ or molecule patterns signal the presence of pathogens and induce various innate immune responses in order to eliminate the infectious agents (1).

Chemically distinct PAMs such as Gram-negative bacteria-derived lipopolysaccharides, Gram-positive bacteria-derived lipoteichoic acids, bacterial DNA, bacterial filaggrin (1, 2), and double-stranded RNA (3) signal the host to the presence of bacteria or viruses via pathogen pattern recognition receptors such as the Toll-like receptors (4), where they induce antimicrobial peptides and/or recruit professional phagocytes, either indirectly, by PAM-induced host-derived chemoattractants (5), or directly, by release of leukotactic PAMs (6) to eliminate the infectious agents.

Neutrophils represent the predominant tissue-infiltrating type of peripheral blood leukocytes seen upon bacterial infection in vertebrates (7). These are recruited from the blood vessels to the infection foci by chemotactic factors that are either released from the pathogen themselves, generated from locally activated complement, or locally induced in host cells such as macrophages and epithelial cells.

Bacteria release leukocyte-chemotactic PAMs, which have been identified as formylated or nonformylated methionyl peptides (8-11) that were also found in mitochondria of eukaryotic cells (12). These peptides, like virtually all known leukocyte attractants, induce in leukocytes a number of host responses including leukocyte chemotaxis, adhesion to blood vessels, release of proteolytic enzymes from primary granules, formation of reactive oxygen intermediates, and leukotriene B₄ (LTB₄) production via G-protein-coupled formyl peptide receptors (FPRs) (13, 14).

Whereas these leukotactic PAMs appear to be of importance in the very early neutrophil infiltration upon bacterial infection, it is not understood how innate immune responses are initiated in fungal infection and whether host phagocytes are similarly recruited by fungal neutrophil chemotactic PAMs.

Different fungal species have the capacity to recruit neutrophils indirectly via serum-dependent chemotaxinogenic properties, inducing liberation of leukotactic complement split products (15-18), neutrophil-chemotactic LTB₄ production in monocytes via glucan receptors (19), or induction of neutrophil-chemotactic interleukin-8/CXCL8 in blood monocytes by fungal -1,6-glucans (20). These pathways have been commonly used to explain the neutrophil infiltration seen upon fungal infection.

Apart from these indirect mechanisms of neutrophil tissue recruitment, some reports (15, 18, 21) have indicated that dermatophytes contain previously uncharacterized neutrophil attractants. Although fungal cell wall -1,3-D-glucans are known to be important in inducing innate immune defense reactions by mediating phagocytosis in macrophages and neutrophils via mannose receptors (22), this molecule class lacks chemotactic activity for neutrophils and thus does not represent the fungal leukotaxin. A recent reinvestigation has shown that dermatophytes, which include Microsporum canis, Epidermophyton floccosum, Trichophyton rubrum, and Trichophyton mentagrophytes, contain leukotactic factors (23).

Partial purification of the dermatophyte-derived neutrophil-chemotactic factor revealed it to be a panleukotactic lipid-like leukocyte activator (LILA) that is distinct in its physicochemical and biological properties from all known lipid-like neutrophil attractants as well as FPR agonists (23). LILA induced secretory responses in human neutrophils, which include the release of primary granule enzymes and superoxide anions (23). Desensitization experiments by dermatophyte-derived LILA and other well characterized chemotaxins revealed only desensitization by LILA itself, indicating a separate putative LILA receptor. Our observation that also bakers' yeast contained LILA (24) prompted the hypothesis that LILA is possibly representing an Ascomycetes-specific neutrophil-chemotactic PAM (29).

The yeast Candida albicans is commonly found as a commensal in the gastrointestinal tract of humans (25), but dissemination is common only in immunosuppressed patients (26-29) with systemic spreading that often leads to lethal septic shock (30). Previous in vitro studies have shown that C. albicans produces neutrophil-chemotactic factors (21, 31-33). In none of the studies were the neutrophil-chemotactic factors characterized, but data from one report implicated FPR as a receptor for a C. albicans-derived neutrophil attractant (33).

In this study, we have characterized yeast-derived neutrophil attractants biologically and biochemically. We have identified a novel class of unstable lipid-like neutrophil attractants, which may represent fungus-specific pathogen-associated molecules.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

Determination of Neutrophil Function-- Neutrophils were freshly prepared from human venous blood as described (34). Neutrophil chemotaxis experiments, adherence of neutrophils to microtiter plates, neutrophil enzyme release, and desensitization measurements were performed as previously described in detail (34). All experiments were performed at least three times in duplicate or triplicate with different donors' cells.

Determination of Ca²⁺ Mobilization-- [Ca²⁺]_i was determined in freshly isolated peripheral blood neutrophils (5 × 10⁶ cells/ml) previously loaded with the fluorescence indicator Fura-2/AM (0.1 nM), according to Ref. 35, that have been subsequently stimulated with N,N'-dihexadecen-(9Z)-oyl urea (C_16:1DAU) under various conditions. Ionomycin (20 µM)- and EGTA (40 mM)-treated cells served as calibrators for maximum (1100 nM Ca²⁺) and minimum signals in all experiments.

Culture of Yeast Cells and Extraction of LILAs-- C. albicans (clinical isolates) was cultured in RPMI 1640 medium (Invitrogen) in the presence of 2.0-20.0 g/liter glucose and 4 g/liter ammonium sulfate at 15-37 °C and pH 5.0-7.0. The controlled addition of glucose and ammonium sulfate at a constant pH of 6.5 and a temperature of 37 °C allowed elongation of the logarithmic growth phase up to 8 h until a cell density of 10⁸ cells/ml was reached. Cells were then centrifuged, washed with phosphate-buffered saline, and immediately used or stored below 70 °C until further use. Mechanical disruption of yeast cells was done in ice/water with a disintegrator (IMA, Zeppelinheim, Germany) at 4000 rpm using glass beads until at least 90% lysis was achieved. Yeast extracts were also obtained from Saccharomyces cerevisiae, Pichia pastoris, or bakers' yeast after mechanical disruption of yeast cells. For extraction of LILAs, freshly prepared yeast cells were treated with 5 volumes of acetone for 1 h at 0 °C. After centrifugation and concentration of the extract by evaporating the acetone, the remaining aqueous phase was further extracted either with diethyl ether or ethyl acetate at 0 °C. Lipid extracts were then depleted from fatty acids by extraction with 0.1 M NaHCO₃, pH 8.0, and separated by reversed phase chromatography (RP-HPLC) (as shown in Fig. 1), or thereafter the remaining organic phase was extracted with aqueous 0.1 M NaOH at 0 °C (a procedure that enriched LILA I-III), and the alkaline phase immediately after separation was neutralized and extracted with diethyl ether or ethyl acetate. The LILA containing organic phase was separated by two or three different steps of RP-HPLC with C8-loaded (J. T. Baker, Inc., Gross Gerau, Germany), C18-loaded (ODS II; Macherey & Nagel, Düren, Germany), and cyanopropyl-loaded (J. T. Baker, Inc.) columns using a gradient of acetonitrile when necessary containing 0.1% trifluoroacetic acid as eluent. 10-30-µl aliquots of HPLC fractions, previously lyophilized in the presence of 10 µl of phosphate-buffered saline containing 0.1% bovine serum albumin were tested for neutrophil myeloperoxidase release and/or neutrophil chemotactic activity using human peripheral blood neutrophils that have been isolated from peripheral blood of healthy donors (34). LILA-containing fractions were further chromatographed until bioactivity corresponded to a single, at 203 nm, absorbing peak and were then immediately lyophilized and stored below 70 °C under argon, where it was found to be stable for several days.

In some cases (as in Fig. 6), freshly prepared LILA-containing HPLC fractions were freed from acetonitrile and trifluoroacetic acid prior to the bioassay by partial lyophilization for 15 min and then immediately used in the neutrophil-myeloperoxidase release assay. Under these conditions, typically dose-response curves showed a decrease at a higher LILA/DAU concentration due to the inhibitory effects of the vehicle.

Structural Analyses of Yeast-derived LILAs-- Capillary GC/MS-analyses were performed on a Micromass VSEQ high resolution GC/MS system with acidic aqueous hydrolysates of purified LILAs.

Electrospray mass spectrometric analyses of natural LILAs, which were stored dry for a few days below 70 °C and which were then dissolved in chloroform/methanol (1:1) just prior to the experiments, were performed with a Micromass Platform I mass spectrometer (Micromass, Manchester, UK). All ESI-MS analyses of synthetic DAUs and some investigations of purified natural LILAs were performed with a QTOF II hybrid mass spectrometer (Micromass) in acetonitrile.

COSY-¹H NMR spectra of LILA samples were recorded on a Bruker AMX-600 (¹H, 600.2 MHz) instrument in CDCl₃.

Chemical Synthesis of C_16:1DAU-- For chemical synthesis of C_16:1DAU, 1.19 mmol of palmitoleic acid (ICN Biomedicals, Eschwege, Germany) was first transferred into its chloride by adding 1.79 mmol of thionyl chloride (Sigma) in 2 ml of dichloromethane, and the mixture was then refluxed for 1 h. Thereafter, 25 mg (0.595 mmol) of carbodiimide/cyanamide (Sigma), dissolved in 2 ml of pyridine, was added, and the mixture was refluxed for 2-3 h. After evaporating pyridine, the amorphous precipitate was then separated and washed twice with ice-cold 10% NaHCO₃, and subsequently the remaining half-solid residue was dissolved in acetonitrile (2 ml) and stored below 70 °C until further use. RP-HPLC of the crude DAU preparation using a C18-RP-HPLC column (300 × 7 mm, C18 Nucleosil with end-capping; Macherey & Nagel) and a gradient of acetonitrile as eluent revealed a single peak of a compound absorbing at 203 nm that contained the majority of biological activity.

C_16:1DAU sum formula is C₃₃H₆₀O₃N₂, calculated monoisotopic mass is 532.460, and melting point is ~0.5 °C. The molar extinction coefficient was calculated to be near 10 at 203 nm in acetonitrile. Upon HPLC analyses (in acetonitrile), we found 9 × 10⁶ integration units (at 203 nm with a Spectra Physics SP 1870 integrator), corresponding to 10 µg of C_16:1DAU. Other DAUs were synthesized similarly using in the initial step the appropriate fatty acid.

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

Yeasts Produce LILAs-- In order to investigate whether C. albicans produced LILA(s) we cultured C. albicans under various conditions and extracted both culture supernatants and cells with aqueous buffers and with organic solvents. We observed neutrophil-chemotactic activity in organic extracts, which was not stable in aqueous phases.

When C. albicans was grown under static conditions at pH 6.5, we found 960 ± 80 units (1 unit being defined as the amount of LILA necessary to elicit a half-maximum chemotactic response in neutrophils in the Boyden chamber system (34)) of total LILA activity per 10⁸ cells (n = 5). High LILA activity was only observed at the end of the logarithmic growth phase, after a 10-12-h culture, where it was seen to be cell-associated and subsequently to be secreted. The mycel form of C. albicans showed 40-50% less LILA activity (units/g, fresh weight) than the yeast form.

RP-HPLC analyses of C. albicans lipid extracts revealed in different strains a characteristic LILA activity profile with several chemically distinct peaks of activity (Fig. 1A), like those seen with lipid extracts of nonpathogenic yeasts such as S. cerevisiae (Fig. 1B) or P. pastoris (data not shown). Desensitization experiments of partially purified S. cerevisiae-derived LILAs (LILAs I-V) with neutrophil MPO release as readout and Ca²⁺ mobilization (24) revealed cross-desensitization of all of these LILAs but not with the FPR-ligand formyl-Met-Leu-Phe, indicating that these fungal LILAs may represent chemically distinct but functionally related molecules sharing the same putative cellular receptor and one clearly distinct from the FPR receptor.

View larger version (46K): [in this window] [in a new window]   Fig. 1.   Identification of LILAs in yeast extracts. The chromatography of a C. albicans lipid extract (A) and bakers' yeast (S. cerevisiae) (B) is shown. Fatty acid-depleted organic solvent extracts obtained from 2 g of cultured C. albicans or 200 g of bakers' yeast were separated on a preparative RP-C8-HPLC column, and aliquots of each fraction were tested for induction of MPO release in human neutrophils (shaded bars). Three peaks of activity were detected upon RP-C8-HPLC in both preparations, which could be separated by subsequent cyanopropyl-RP-HPLC followed by RP-C18-HPLC into five different LILAs.

In order to analyze whether LILAs are fungus-specific PAMs, we then investigated whether bacteria as well as human cells produced similar LILAs. Extraction at optimized conditions (the aqueous extraction into a pH 13 medium followed by reextraction at low pH into another organic solvent, was found to be a characteristic property of the LILA activity) revealed no or no similar LILA activity in various strains of bacteria such as E. coli, P. aeruginosa, S. aureus, and S. epidermidis as well as human cells (keratinocytes, neutrophils, fibroblasts, mononuclear leukocytes, and its culture supernatants (data not shown)).

In order to demonstrate similarities of biological activities of purified LILA preparations obtained from different fungi (C. albicans, S. cerevisiae, T. mentagrophytes) we compared induction of migratory and secretory activities and found these to be similar in neutrophils. Furthermore, cross-desensitization experiments in chemotaxis with 20 units/ml dermatophyte-derived LILA revealed, after subsequent stimulation with either dermatophyte-derived LILA, C. albicans-derived LILA, or S. cerevisiae-derived LILA, 35 ± 10, 35 ± 15, and 30 ± 15% chemotactic response, respectively, of a buffer-pretreated control (100%), which we interpreted as a functional and possibly structural similarity of LILAs obtained from different fungi.

Adherence to the substratum is a prerequisite of chemotactically migrating leukocytes and is induced by known chemotactic factors (36). We tested the adherence of neutrophils to plastic microtiter plates and observed 25 ± 5% of the total number of neutrophils adhering upon stimulation with 1 unit of purified LILA I. 10 units of LILA I led to 90 ± 10% adherence, whereas the medium control gave 4 ± 3%, and the positive control formyl-Met-Leu-Phe (10⁸ M) led to 85 ± 7% adherence (n = 5), indicating that LILA I, like other attractants, also induces adherence.

Taken together, fungal LILA activates a number of proinflammatory functions in human neutrophils that are known to be elicited in these phagocytes when stimulated with well defined host-derived neutrophil attractants (37) that include C5a, LTB₄, platelet-activating factor, interleukin-8/CXCL8, and bacterial FPR ligands.

Fungal LILAs Are Structurally Diacylated Urea Derivatives-- In order to characterize the structures of the major fungal LILAs, we analyzed bakers' yeast (S. cerevisiae)-derived LILA, which behaved similarly to dermatophyte-derived LILA (23), being short lived under physiologic conditions (estimated half-life time at ambient temperature in aqueous media was <1 h), sensitive toward oxygen, and very sensitive toward storage in methanol. We observed that LILA activity is also sensitive toward treatment with diazomethane (data not shown), indicating that LILA contains acidic protons. This hypothesis was further supported by the observation that LILA can be extracted from organic solvents at high pH (pH 13-14). We used this unique physicochemical property to enrich LILAs, to separate them from neutral lipids, and to purify them by subsequent RP-HPLC analyses to homogeneity.

pH 13-soluble bakers' yeast lipids were first separated on a RP-C8-HPLC column, and fractions were tested for biological activity. Three activity peaks inducing MPO release in neutrophils were identified that were seen after further RP-HPLC purification to consist of five chemically distinct and unstable LILAs (Fig. 1B). The third activity peak, containing LILA I, LILA II, and LILA III (see Fig. 1B), was further separated by cyanopropyl-RP-HPLC. Here we could separate LILA III from LILA I and II, which coeluted in this system (data not shown). The peak containing LILA I and LILA II was then further separated by RP-C18-HPLC and isocratic elution, where now two activity peaks (LILA I and LILA II) could be separated (A). The principal and most potent species (LILA I) and two other minor and less potent LILAs (LILAs II and III) were purified by cyanopropyl-RP-HPLC (data not shown) followed by RP-C18-HPLC of MPO release-inducing HPLC fractions (Fig. 2A) to apparent homogeneity as revealed by analyses of bioactivity that corresponded to a single, at 203 nm, absorbing peak. LILAs IV and V could not be purified to homogeneity and structurally characterized, probably on account of their chemical instability.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Purification of LILAs. A, RP-C18-HPLC purification of LILA I and LILA II. pH 13-soluble bakers' yeast lipids (obtained from 200 g of yeast) were first separated on a RP-C8-HPLC column using an increasing gradient of acetonitrile, and the third activity peak (see Fig. 1B) was further separated by cyanopropyl-RP-HPLC. The peak containing LILA I and LILA II was then further separated by RP-C18-HPLC and isocratic elution, where now two activity peaks (LILA I and LILA II) could be separated (A). A typical HPLC run is shown. B and C, electrospray (ESI) MS spectra of LILA I (B) and LILA II (C). Lyophilized LILA preparations were dissolved immediately prior to experiments in chloroform/methanol (1:1) and then analyzed by ESI-MS using a Micromass Platform I mass spectrometer at a cone voltage of 60 V. A cone voltage of 35 V led to additional prominent masses at 309, 325, and 339 Da for LILA I and 279, 335, 351, and 365 Da for LILA II together with prominent masses below 100 Da.

Capillary GC/MS analyses of acidic aqueous hydrolysates of purified LILA I (S. cerevisiae) revealed the presence of a major component (>98%) giving a mass spectrum identical with that of palmitoleic acid (C_16:1). Minor compounds were C_16:0, C_18:1, and C_18:2 (<2%). Capillary GC/MS analyses of acidic aqueous hydrolysates of LILA II (S. cerevisiae) revealed the presence of a major component (>92%) giving a mass spectrum identical with that of linoleic acid (C_18:2). Impurities were C_16:0, C_16:1, and C_18:1 (<8%). Analyses of LILA III revealed oleic acid (C_18:1) as a major (>96%) component. Impurities were C_16:0, C_16:1, and C_18:2 (<4%).

We then analyzed LILA I and LILA II that have been stored dry for a few days below 70 °C (where it seemed to be stable) by electron impact mass spectrometry (MS) and by positive ionization electrospray ionization (ESI) MS analyses with a Micromass Platform I mass spectrometer. Electron impact-MS analyses gave no clear results, probably on account of LILA decomposition. ESI-MS analyses of LILAs were found to strongly depend on the conditions of analyses, in particular the solvent and the cone voltage used. However, when purified LILAs were dissolved in chloroform/methanol/formic acid (60:40:0.5) just prior to the experiments, we saw at 60 V cone voltage characteristic ions at m/z 533.7 for LILA I (Fig. 2B) and at m/z 585.5 for LILA II (Fig. 2C), which we interpreted as molecular ions ((M + H)⁺). The signals for ions 14 mass units higher in each case (m/z 547.7 and 599.5) were only observed when the samples were analyzed in chloroform/methanol, probably due to LILA derivatization by adventitious methylation. The mass difference between both putative (M + H)⁺ peaks of LILA I and LILA II was found to be 52 Da, which is twice the mass difference between C_18:2 (280.4 Da) and C_16:1 (254.4 Da). Furthermore, abundant masses of LILA fragment ions formed by loss of 254 and 280 Da from (M + H)⁺ peaks of putatively methylated LILA I and LILA II (m/z 547.7 and 599.5) that correspond to one molecule of fatty acid were seen at m/z 293.6 for LILA I and at m/z 319.5 for LILA II (Fig. 2, B and C). Both pairs showed a mass difference of 26 Da, that is equal to the mass difference between C_18:2 and C_16:1. These data could be interpreted by the presence of two identical fatty acid acyl groups in each purified LILA. Possibly due to the use of different cone voltages, partial degradation, and oxidation during storage and/or in the ionization source, samples sometimes showed the presence of signals for the free fatty acid, the fatty acid methyl ester (C_16:1 for LILA I or C_18:2 for LILA II), oxidized DAU metabolites, and other unknown ions.

In order to get further support of our interpretation of MS data, we also analyzed products formed during decomposition of LILA. LILA biological activity is known to be extremely sensitive toward treatment with methanol (23), a finding that did not allow us the use of methanol for RP-HPLC. Therefore, we investigated the products formed during storage of LILA. An RP-HPLC analysis of LILA II (S. cerevisiae) that was stored for several months in methanol/chloroform (1:1) below 70 °C revealed, apart from a UV-absorbing peak corresponding to LILA II, two major additional peaks. The more polar peak revealed upon ESI-MS analyses in the positive ionization mode a dominant signal at m/z 323 corresponding to the mass of a mono-C_18:2-urea (calculated (M + H)⁺ at m/z 323). Storage of this HPLC fraction in methanol/chloroform for 1 week at 4 °C lead to a MS signal at m/z 295 corresponding to C_18:2-methyl ester (calculated (M + H)⁺ at m/z 295), a finding that was further supported by capillary GC/MS analyses. These observations indicate that the intermediate LILA II fragment contains an additional C_18:2 group and that the decomposition product is probably a monoacyl urea. Thus, LILA II may form mono-C_18:2-urea and C_18:2-methyl ester upon storage in methanol. We obtained similar results with LILA I (data not shown).

In order to solve the LILA structure, we also performed ¹H NMR analyses. Samples were freshly prepared and were free of metabolites including fatty acids. The 600-MHz COSY ¹H NMR spectra of LILA I (Fig. 3), LILA II (data not shown), and LILA III (data not shown) show the presence of protons belonging to only one set of cis-unsaturated fatty acid protons, C_16:1 for LILA I (Fig. 3), C_18:2 for LILA II, and C_18:1 for LILA III. ESI-MS analyses of LILAs performed immediately after the ¹H NMR analyses were found to be identical to those obtained from freshly isolated material, indicating the absence of decomposition during ¹H NMR analyses. Thus, a pair of fatty acid acyl residues should be substituted to an unknown 58-Da structural element in each LILA that must be symmetrical. The only elementary composition that fulfills these criteria is CH₂ON₂, which represents urea as the backbone of each LILA.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   COSY 1H NMR spectrum of LILA I. A, a two-dimensional (COSY) 1H NMR spectrum of LILA I, recorded on a Bruker AMX-600 (1H, 600.2 MHz) instrument in CDCl3 is shown. Signals are consistent with a C16 fatty acid with one double bond. B, double bond proton signal section of the one-dimensional 1H NMR spectrum of LILA I. The symmetric multiplet and coupling constants clearly indicate the presence of a cis-double bond. C, 1H NMR signal attachment to protons in LILA I.

Interpretation of the physicochemical data therefore led to the proposal of C_16:1DAU, N,N'-dilinoleyl urea (C_18:1DAU), and N,N'-dioleyl urea (C_18:2DAU), as yet not described in chemical data bases, as chemical structures for LILA I, LILA II, and LILA III, respectively (Fig. 4).

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Fungal LILAs are structurally DAUs. Shown are chemical structures of LILA I (C16:1DAU) (a), an all-trans isomer of LILA I (all-trans-C16:1DAU) (b), LILA II (C18:2DAU) (c), and LILA III (C18:1DAU) (d).

We interpret the failure to detect in 600-MHz COSY ¹H NMR spectra of LILA I (Fig. 3), LILA II, and III (not shown) signals for imide protons to be a result of bridging of protons to two carbonyl residues, which is known to cause a low field shift together with broadening of the signals, often resulting in undetectability. To prove our structure hypothesis, we chemically synthesized the respective DAUs.

Data bank search revealed only a few reports (mainly patents) about synthesis and properties of open-chain diacyl ureas (38-40), where some short carbonic acid chain derivatives have been described. These were reported to be soluble under alkaline conditions but short lived and easily decomposed, even at pH 7, to the free fatty acid and the monoacyl urea. Diacyl ureas can be synthesized by adding the carboxylate organic acid chloride in the presence of traces of sulfuric acid to solutions of urea in benzene. Using this method with long chain unsaturated fatty acids, we could isolate some material having the properties of diacyl ureas, but the recovery was found to be very low (<0.01% with respect to urea). The recovery could be increased by changing the conditions and using a tertiary amine such as triethylamine or pyridine and the fatty acid chloride in the absence of sulfuric acid. The addition of pyridine has been reported to catalyze the rearrangement of an O-acylated urea (which is otherwise formed as a major product) to an N-acylated urea. After optimizing different methods, we identified as the best method to produce N,N'-diacylated ureas its synthesis from carbodiimide/cyanamide and fatty acid chlorides (41) in pyridine, which is a good solvent for carbodiimide/cyanamide. The recovery of the DAU, however, was found to be 2%.

When we separated a crude C_16:1DAU synthesis preparation by RP-HPLC and analyzed HPLC fractions by ESI-MS analyses (QTOF II; Micromass) in the positive ionization mode in acetonitrile, we identified as products C_16:1DAU ((M + H)⁺ at m/z 533.4; calculated: 533.46; (M + Na + H₂)²⁺: 278.65) and an early eluting compound, which we identified as the C_16:1-cyanamide ((M + H)⁺ at m/z 279.2, calculated: 279.24).

It is important to note that QTOF II ESI-MS analyses of C_16:1DAU were performed immediately after HPLC separation, with an only moderately heated source in the positive ionization mode (cone voltage, 100 V; capillary voltage, 2.5 kV, very low collision energy). We found that ESI-MS results strongly depended on experimental conditions, where in our hands the use of a heated source, for example, usually led to the loss of the (M + H)⁺ peak.

The product ion MS-MS spectrum obtained following collision-induced dissociation of the (M + H)⁺ from C_16:1DAU in acetonitrile revealed a dominant positive fragment ion at m/z 296.3, which corresponds to (M + H)⁺ of mono-N-palmitoleyl urea (calculated (M + H)⁺: 296.25) and another ion at m/z 279.2, which corresponds to (M + H)⁺ of N-palmitoleyl cyanamide (calculated mass: 279.24), a likely fragmentation product of C_16:1DAU by loss of one molecule fatty acid, which indicates the loss of a mass of 254.2, that corresponds to C_16:1 (calculated monoisotopic mass: 254.23) (Fig. 5B). Negative ionization ESI-MS of the C_16:1DAU gave a single mass at m/z 277.2, corresponding to N-palmitoleyl cyanamide anion (calculated (M  H): 277.23).

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Comparison of natural LILA I and chemically synthesized C16:1DAU. A, RP-C18-HPLC of purified natural LILA I (left panel), synthetic C16:1DAU (middle panel), and a mixture of both (right panel). B, product-positive ion MS2 spectrum obtained following collision-induced dissociation of the (M + H)+ at m/z = 533 of a natural LILA I preparation (lower panel) and synthetic C16:1DAU (upper panel) determined by positive ionization on a QTOF II mass spectrometer in acetonitrile. C, comparison of myeloperoxidase-releasing activity of freshly purified natural LILA I () and synthetic C16:1DAU () (normalized for the integration units at 203 nm, determined after RP-HPLC) in human neutrophils. Typical results out of three independent experiments are shown.

When we analyzed the stability of purified synthetic DAU we observed the same physicochemical properties seen in natural LILAs (23), which include sensitivity toward methanol, alkaline, and acidic conditions as well as sensitivity toward oxygen. Therefore, DAUs were stored in the absence of acids under dry conditions under argon below 70 °C, where we found them to be stable for days to weeks.

When we compared purified synthetic C_16:1DAU with natural LILA I in several systems, we found both to be identical with respect to its retention time upon RP-HPLC analysis (Fig. 5A), its exact molecular mass on ESI-MS, its collision-induced ion dissociation ESI tandem mass spectrum of the (M + H)⁺ peak (Fig. 5B), and its dose-response curves of neutrophil-stimulating properties, such as neutrophil MPO release (Fig. 5C).

Furthermore, when ESI-MS analyses of synthetic C_16:1DAU were performed using the QTOF II mass spectrometer in chloroform/methanol/formic acid, as done with natural LILA I (Fig. 2B), similarly we identified a putative (M + H)⁺ at m/z 547.45, which showed upon collision-induced dissociation MS² analyses a fragment ion at m/z 293.23 indicating the loss of 254.25 Da, which corresponds to palmitoleic acid, as observed with natural LILA I. The exact (M + H)⁺ mass (547.4529 Da) corresponded to the elementary composition of a methylated and protonated C_16:1DAU (C₃₄H₆₃O₃N₂, calculated mass: 547.4838 Da).

As is the case with known chemoattractants (37), synthetic C_16:1DAU exhibited a bell-shaped dose-response curve for neutrophil chemotaxis with activity at higher nanomolar concentrations (ED₅₀, 140 nM) and chemotaxis efficacy (percentage input migrating cells) similar to FNLP-treated cells (Fig. 6A).

View larger version (20K): [in this window] [in a new window]   Fig. 6.   Neutrophil-stimulating activity of structurally different DAUs. A, in vitro chemotactic activity of C16:1DAU (), C18:1DAU (), and C18:2 DAU () in the Boyden chamber system. B, in vitro release of MPO in cytochalasin B-pretreated human neutrophils by C16:1DAU (), all-trans-C16:1DAU (), C18:1DAU (), and C18:2DAU (). FNLP served as positive control for chemotaxis (108 M) and for MPO release (107 M). Typical results out of at least three independent experiments performed in duplicate are shown.

C_18:2DAU, corresponding to LILA II, and C_18:1DAU, corresponding to LILA III, were both found to be nearly 40-fold less active in both chemotaxis and degranulation bioassays (Fig. 6, A and B), a finding also seen with the minor natural LILA II and III (data not shown), indicating that C_16:1DAU is the most potent and principal neutrophil activator among the three presently characterized fungal LILAs.

All LILAs we had structurally characterized showed as a constant structural element the presence of a cis-double bond at carbon atoms 9 and 10 in both fatty acid residues (Fig. 4). The observation that a cis-double bond can be essential for bioactivity in host-derived leukotactic lipids such as LTB₄ (42) raises the question of whether this structural element similarly might be important for biological activity of DAUs. A chemically synthesized all-trans derivative of LILA I, N,N'-dipalmitelaidyl urea (Fig. 4) did not induce MPO release in neutrophils (Fig. 6B), indicating that in fungal DAUs a ₉ cis-double bound is essential for bioactivity.

DAUs Activate Neutrophils via a Putative G-protein-coupled Receptor-- Neutrophil chemoattractants activate neutrophil function (including Ca²⁺-signaling) via G-protein-coupled receptors, which are sensitive toward pertussis toxin pretreatment (43, 44). Our previous studies revealed that natural LILA preparations can mobilize [Ca²⁺]_i in human neutrophils, which can be desensitized when neutrophils are pretreated with LILA but not with other chemoattractants including formyl-Met-Leu-Phe, LTB₄, platelet-activating factor, or 5-oxo-ETE (24).

In support of these data, we found that synthetic C_16:1DAU induced Ca²⁺ signaling in neutrophils in a dose-dependent fashion. This was desensitized after the repeated addition of 20 ng of C_16:1DAU but did not desensitize FNLP (100 nM)-induced Ca²⁺-signaling (Fig. 7, upper panel). These findings clearly indicate that the C. albicans-derived and not yet characterized neutrophil-chemotactic factor previously described (33), which activates neutrophils via the FPR, is distinct from DAUs.

View larger version (21K): [in this window] [in a new window]   Fig. 7.   DAUs induce in neutrophils a pertussis toxin-sensitive increase of intracellular Ca2+ concentrations ([Ca2+]i) via a putative G-protein-coupled receptor. C16:1DAU induces a ligand-specific [Ca2+]i mobilization in human neutrophils, which is desensitized by the repeated addition of C16:1DAU (upper panel). Middle panel, [Ca2+]i mobilization is pertussis toxin-sensitive. Neutrophils, pretreated for 2 h at 37 °C with 4 µg/ml Bordetella pertussis toxin (Calbiochem), were first stimulated with C16:1DAU (12 nM in each challenge) and then with FNLP (107 M) as control. No [Ca2+]i mobilization is visible. Lower panel, acetylgeranylgeranyl cysteine (AGGC) inhibits C16:1DAU-mediated [Ca2+]i mobilization. Neutrophils were first stimulated with 10 nM C16:1DAU, and after the addition of AGGC (5 µM) cells were challenged again with 20 nM C16:1DAU. No [Ca2+]i mobilization was observed with either C16:1DAU or FNLP (107 M), which served as control. Typical results out of at least three independent experiments are shown.

We then investigated whether the DAU signal is pertussis toxin-sensitive. As shown in Fig. 7, middle panel, no induction of Ca²⁺ signaling was seen with C_16:1DAU. These findings support the hypothesis that a putative G_i-protein-coupled receptor is involved in the DAU-mediated activation of neutrophils.

The membrane-associated carboxylmethyltransferase is important for methylesterification of G-proteins. Inhibition of this enzyme causes inhibition of G-protein-coupled receptor signaling (45). When neutrophils were pretreated with the inhibitor of the membrane-associated carboxylmethyltransferase, acetylgeranylgeranyl cysteine (35), and subsequently stimulated with C_16:1DAU followed by FNLP as positive control, Ca²⁺ mobilization in neutrophils was inhibited (Fig. 7, lower panel), further supporting the hypothesis that in neutrophils the LILA signaling depends on low molecular weight G-proteins but not toll-like receptors, which are not signaling via G-proteins (46).

In summary, we described a novel class of potentially proinflammatory mediators, which appeared to be fungus-specific PAMs, that were identified as long chain unsaturated fatty acid DAUs. These lipoids elicit in phagocytes a number of innate immune responses via as yet unknown putative receptors that at least in neutrophils are G-protein-coupled.

DAUs represent candidate mediators that may contribute to the inflammatory reaction seen upon fungal infection by recruiting phagocytes directly. These unstable lipoids, which are unique for fungi and have not previously been found in nature, and their putative cellular receptors, together with DAU-dependent signaling in host cells and the enzymes involved in the fungal DAU-biosynthesis, could be targets for a novel anti-inflammatory therapy in fungal infection.

	ACKNOWLEDGEMENTS

We thank D. Tiaden and C. Mehrens for expert technical assistance, H. Toms and the University of London's Intercollegiate Service at Queen Mary and Westfield College for performing ¹H NMR spectroscopic analyses, J. Brasch for advice in culturing fungi, and E. Christophers for stimulating discussions.

	FOOTNOTES

^* This work was supported by Deutsche Forschungsgemeinschaft Grants SFB 415 and Schr 305/2-1.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ These authors contributed equally to this work.

^¶ To whom all correspondence should be addressed: Dept. of Dermatology, University Hospital Kiel, Schittenhelmstrasse 7, D-24105 Kiel, Germany. Tel.: 49-431-597-1536; Fax: 49-431-597-1611; E-mail: jschroeder@dermatology.uni-kiel.de.

Published, JBC Papers in Press, May 22, 2002, DOI 10.1074/jbc.M202998200

	ABBREVIATIONS

The abbreviations used are: PAM, pathogen-associated molecule; DAU, diacyl urea; MS, mass spectrometry; ESI-MS, electrospray ionization mass spectrometry; FNLP, formyl-Nle-Leu-Phe; FPR, formyl peptide receptor; LILA, lipid-like leukocyte activator; MPO, myeloperoxidase; LTB₄, leukotriene B₄; HPLC, high pressure liquid chromatography; RP-HPLC, reversed phase HPLC; GC, gas chromatography; C_16:1DAU, N,N'-dihexadecen-9Z-oyl urea; C_18:1DAU, N,N'-dilinoleyl urea; C_18:2DAU, N,N'-dioleyl urea.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				M. C. Noverr, J. R. Erb-Downward, and G. B. Huffnagle Production of Eicosanoids and Other Oxylipins by Pathogenic Eukaryotic Microbes Clin. Microbiol. Rev., July 1, 2003; 16(3): 517 - 533. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) All Versions of this Article: 277/31/27887    most recent M202998200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Schröder, J.-M. Articles by Mallet, A. I. Search for Related Content PubMed PubMed Citation Articles by Schröder, J.-M. Articles by Mallet, A. I. Social Bookmarking                      What's this?