LPS binding protein (LBP) is an acute-phase protein synthesized predominantly in the liver of the mammalian host. It was first described to bind LPS of Gram-negative bacteria and transfer it via a CD14-enhanced mechanism to a receptor complex including TLR-4 and MD-2, initiating a signal transduction cascade leading to the release of proinflammatory cytokines. In recent studies, we found that LBP also mediates cytokine induction caused by compounds derived from Gram-positive bacteria, including lipoteichoic acid and peptidoglycan fragments. Lipoproteins and lipopeptides have repeatedly been shown to act as potent cytokine inducers, interacting with TLR-2, in synergy with TLR-1 or -6. In this study, we show that these compounds also interact with LBP and CD14. We used triacylated lipopeptides, corresponding to lipoproteins of Borrelia burgdorferi, mycobacteria, and Escherichia coli, as well as diacylated lipopeptides, corresponding to, e.g., 2-kDa macrophage activating lipopeptide of Mycoplasma spp. Activation of Chinese hamster ovary cells transfected with TLR-2 by both lipopeptides was enhanced by cotransfection of CD14. Responsiveness of human mononuclear cells to these compounds was greatly enhanced in the presence of human LBP. Binding of lipopeptides to LBP as well as competitive inhibition of this interaction by LPS was demonstrated in a microplate assay. Furthermore, we were able to show that LBP transfers lipopeptides to CD14 on human monocytes using FACS analysis. These results support that LBP is a pattern recognition receptor transferring a variety of bacterial ligands including the two major types of lipopeptides to CD14 present in different receptor complexes.
The innate immune system is represented by both humoral and cellular elements (1, 2). Soluble CD14 (sCD14)4 and LPS binding protein (LBP) are major serum factors with the ability to bind pathogens and initiate innate immune responses (3, 4). The cellular, so-called pattern recognition receptors include the recently identified family of TLRs (2). TLRs are germline-encoded receptors exhibiting homologies to Toll, which is involved in embryogenesis and host defense against Gram-positive bacteria and fungi in Drosophila (5, 6). TLR-2 was originally described to recognize LPS, a major constituent of the outer membrane of Gram-negative bacteria (7, 8), whereas later studies identified TLR-4 as the central transmembrane component of the LPS receptor (9, 10). The reason for this discrepancy was that LPS preparations used contained contaminating bacterial lipoproteins, which later were identified to be highly active TLR-2 agonists (11, 12). TLR-2 also has been shown meanwhile to interact with numerous ligands of mainly bacterial origin, including peptidoglycan and lipoteichoic acid (LTA) of Gram-positive bacteria (13, 14) and spirochetal glycolipids (14, 15).
Lipoproteins and lipopeptides have been identified in a large number of microorganisms, the most prominent ones being mycobacteria, mycoplasms, and spirochetes. They have been found to exhibit both a strong innate inflammatory response in the host and an enduring adaptive immune response in mammalian hosts (16). The strong proinflammatory capacities of lipoproteins were first described for outer surface proteins A and B of Borrelia burgdorferi, which are also highly immunogenic (17) and have lately been the basis for a Lyme disease vaccine development (18). These compounds exhibit an triacylated lipid anchor structure comprising an N-palmitoyl-S-[2,3-bis(palmitoyloxy)-(2RS)-propyl]-(R)-cysteinyl (Pam3Cys) moiety at the N terminus (19), a feature that was previously described for the Braun lipoprotein from Escherichia coli (20). Because the N-terminal Pam3Cys modification is essential for immunoactivation caused by lipoproteins of B. burgdorferi as well as of another spirochete, Treponema pallidum (21), subsequent studies investigating immune responses to spirochetes used synthetic lipopeptides (22). The Pam3Cys moiety was also reported to be present in cytokine-inducing lipoproteins of Mycobacterium and Mycoplasma spp. (23, 24); thus, it can be regarded as a highly conserved molecular motif among different classes of bacteria. In Mycoplasma fermentans, the presence of a macrophage stimulating lipopeptide, termed 2-kDa macrophage-activating lipopeptide (MALP-2), was observed, being stimulatory active at picomolar concentrations (25). This compound, in contrast to the predominant lipopeptide structures present in lipoproteins of E. coli, B. burgdorferi, and mycobacteria, lacks the N-palmitoyl group, thus containing a diacylated (Pam2Cys) lipid anchor structure at the N terminus. Following studies revealed the presence of closely related compounds in other Mycoplasma spp. (26).
Cytokine induction caused by lipoproteins and lipopeptides is mediated by TLR-2 (11, 12). To interact with a rather wide spectrum of ligands, TLR-2 forms heterodimers with TLR-1 and TLR-6 (27). Studies using knockout mice identified TLR-6 as the coreceptor required for the recognition of diacylated lipopeptides corresponding to MALP-2 (28), whereas TLR-1 apparently forms heterodimers with TLR-2 to recognize triacylated lipopeptide structures corresponding to, e.g., B. burgdorferi lipoproteins (29). In contrast to this, a recent study using dominant-negative mutants provided evidence that TLR-6 may also be involved in recognition of these compounds (30).
Activation of immune cells by microbial compounds involves the soluble serum protein LBP (3, 15, 31). LBP is an acute-phase protein synthesized predominantly in the liver exhibiting homologies to other phospholipid binding proteins (3, 32, 33, 34, 35). However, there is growing evidence that it is also released by cells of other compartments, because alveolar and intestinal epithelia have been shown to express this protein (36, 37, 38). LBP was first described to bind LPS from Gram-negative bacteria catalyzing its transfer to CD14, present as a phosphatidylinositol-anchored receptor on monocytes, macrophages, and neutrophils, and secreted in soluble isoforms (sCD14) by a wide variety of cells (4, 39, 40). LBP knockout mice have been shown to be highly susceptible to Salmonella infection (41), apparently due to the impaired recognition of these bacteria. Elevated LBP levels as found during sepsis have been shown to display beneficial effects by inhibiting cytokine response, and protecting the host from septic shock (42, 43). It is well established that CD14 is a pattern recognition receptor recognizing a wide variety of bacterial ligands (44), including LTA and peptidoglycan of Gram-positive bacteria (45, 46) as well as bacterial triacylated lipoproteins (47). In contrast, LBP for some time has been regarded to specifically recognize LPS of Gram-negative bacteria. In previous studies, we were able to show that LBP also recognizes LTA of Gram-positive bacteria as well as structurally similar spirochetal glycolipids (15, 48). Furthermore, LBP was shown to recognize pneumoccal peptidoglycan in a murine meningitis model (31).
In this study, we focus on the interaction of LBP and CD14 with synthetic triacylated and diacylated lipopeptides corresponding to triacylated lipoproteins and MALP-2, respectively. We provide evidence that cellular activation by these compounds is mediated by LBP and CD14 by the use of rLBP and blocking CD14 Abs as well as transfection experiments using Chinese hamster ovary (CHO) cells. In addition, these results were supported by binding studies on microtiter plates, native gel shifts, and FACS analysis. Thus, we provide evidence that LBP enables the innate immune system to recognize not only LPS and glycolipid-like structures such as LTA, but also the N-terminal lipid anchor structures of lipoproteins and lipopeptides, an abundant molecular motif shared by many different bacteria. These data thus provide insight in both function of LBP and pathogenesis of bacterial infections caused by bacteria.
Materials and Methods
Reagents and plasmids
Triacylated lipopeptides (Pam3CSK4; lipopeptide (LP)3) and diacylated lipopeptides (Pam2CSK4; LP2), first described by Mühlradt et al. (26), as well as synthetic MALP-2 were purchased from EMC Microcollections (Tübingen, Germany). LPS from Salmonella minnesota R595 was purchased from Sigma-Aldrich (St. Louis, MO) and found to be devoid of any lipoprotein contaminations in previous studies. Plasmids encoding human (h)CD14, hTLR-2, hTLR-4, and hTLR-1 as well as β-galactosidase and the ELAM NF-κB reporter plasmid were kindly provided by C. Kirschning (University of Munich, Munich, Germany). Plasmids encoding for hMD-2 and hTLR-6 were kindly provided by K. Miyake (University of Tokyo, Tokyo, Japan) and O. Takeuchi (University of Osaka, Osaka, Japan), respectively. Recombinant human LBP as well as rabbit polyclonal Ab against human LBP were kindly provided by Xoma (Berkeley, CA). The blocking CD14 Ab My4 was purchased from Beckman Coulter (Krefeld, Germany). IgG2 isotype control Ab was from BD Pharmingen (Hamburg, Germany). The blocking CD14 Ab MEM-18 (49) was a kind gift from V. Horejsi (Institute of Molecular Genetics, Prague, Czech Republic). IgG1 isotype control was from DakoCytomation (Hamburg, Germany).
Transfection of HEK293 and CHO cells
Isolation and stimulation of human peripheral blood monocytes
Blood from healthy volunteers was drawn with heparin (5 U/ml) and diluted 1/2 in RPMI 1640 (PAA) supplemented with glutamine and antibiotics. A volume of 30 ml was layered on 15 ml of Pancoll (Pan Biotech, Aidenbach, Germany) and centrifuged at 400 × g and 21°C for 45 min. The resulting interphase containing leukocytes was isolated and washed twice with RPMI 1640 at 400 × g and 21°C for 10 min. Remaining platelets were removed by centrifugation at 100 × g for 15 min. Resulting leukocytes were cultivated in 96-well plates in RPMI 1640 containing 2% human albumin (Sigma-Aldrich) for 2 h, followed by two washing steps to remove nonadherent cells. Cells were stimulated with the indicated compounds in the presence or absence of rhLBP. As indicated, the anti-CD14 Ab My4 or its isotype control were added to the cells at a concentration of 5 μg/ml for 30 min before stimulation. After 4 h, supernatants were collected, and TNF-α levels were estimated. To this end, 96-well Immunomaxi plates (TPP, Buchs, Switzerland) were coated with 5 μg/ml rabbit anti-human TNF-α Ab (BD Pharmingen) diluted in 100 mM NaHCO3. Plates were blocked with PBS containing 10% FCS for 2 h at room temperature, followed by incubation with the samples as well as TNF-α standard at 4°C overnight. After washing, plates were incubated with biotinylated TNF-α Ab (BD Pharmingen) for 1 h, followed by incubation with peroxidase-conjugated streptavidin (1 μg/ml; Sigma-Aldrich) for 30 min. Detection of TNF-α was conducted using ortho-phenylen-diphosphate (Sigma-Aldrich).
Phast Gel native and SDS gel electrophoresis
LBP microtiter binding assay
LBP-binding assays were performed using a published protocol (50) with some modifications. Rabbit anti-human LBP Ab was biotinylated using sulfosuccinimidyl biotin (Sigma-Aldrich). The Ab was dialyzed against PBS, and 100 μg were mixed with 30 μg of biotin dissolved in DMSO (Roth, Braunschweig, Germany) in a total volume of 80 μl and incubated at 4°C overnight. Immunomaxi plates were coated with LPS, LP2, LP3, or MALP-2 at 50 μg/ml in 20 mM EDTA and 100 mM Na2CO3 (pH 9.2) for 3 h at room temperature. Blocking was performed with 150 mM NaCl and 50 mM HEPES (pH 7.4), supplemented with 10 mg/ml BSA (Roth) for 30 min at 37°C. Plates were then incubated with rhLBP (100 ng/ml) with or without LPS (50 μg/ml) in the same buffer supplemented with 1 mg/ml BSA at 37°C for 2 h. Detection of bound LBP was performed using the biotinylated anti-LBP Ab (1 μg/ml) and peroxidase-coupled streptavidin as described above. For certain experiments, LBP was added to microplates coated with LPS together with increasing concentrations of LPS, LP2, LP3, or MALP-2.
Pam3CSK4-FLAG and Pam2CSK4-FLAG (LP3-FLAG and LP2-FLAG) were obtained from EMC Microcollections and dissolved in distilled water. Human PBMC were isolated from heparinized blood of healthy adult donors by Ficoll-Isopaque density gradient centrifugation and incubated with LP3-FLAG or LP2-FLAG (10 nM) for 15 min at 37°C in serum-free RPMI 1640 with or without recombinant human LBP (0.5 μg/ml). After two washing steps with PBS, cells were incubated with a murine anti-FLAG M2 Ab (Sigma-Aldrich) at 5 μg/ml for 1 h. Cells were again washed twice, followed by incubation with Cy5-conjugated secondary goat anti-murine IgG Ab (Sigma-Aldrich) for 1 h. Cells were washed twice and fixed in 3% paraformaldehyde, and gated monocytes were analyzed by FACS (FACSCalibur; BD Biosciences, San Jose, CA). For experiments investigating CD14, mononuclear cells were incubated with the CD14 blocking Ab MEM-18 or IgG1 control at 10 μg/ml for 30 min at 37°C before incubation with the lipopeptides and LBP. After two washing steps with PBS, lipopeptides bound to the cells were detected using biotinylated anti-FLAG-M2 Ab (Sigma-Aldrich) at 5 μg/ml, for 1 h at 37°C, followed by two washing steps and detection with streptavidin-Cy5 (Dianova, Hamburg, Germany). Cells were washed, fixed, and analyzed as described above.
As indicated, experimental data were analyzed for statistical significance using Student’s t test for unpaired samples. Significant effects are indicated by asterisks.
Interaction of synthetic lipopeptides with different TLRs
The first set of experiments was aimed at confirming that the synthetic compounds used in this study were similar to their naturally occurring counterparts regarding interaction with TLRs. HEK293 cells stably transfected with human CD14 were transiently transfected with β-galactosidase, the ELAM NF-κB reporter plasmid and TLR-4/MD-2 or TLR-2 alone, followed by stimulation with LPS, LP3, and LP2. Cells were unresponsive to any ligands tested when no TLR was transfected (Fig. 1⇓A), whereas transfection with TLR-4/MD-2 led to activation by LPS only (data not shown). TLR-2 was transfected without additional transfection of MD-2, because MD-2 failed to exhibit any effect toward recognition of TLR-2 ligands in our previous studies (31, 48). Only upon transfection with TLR-2, HEK293 cells could be activated by both synthetic lipopeptides (Fig. 1⇓B), identifying them as TLR-2 agonists. Because no activation of TLR-4/MD-2-transfected HEK293 cells by LP2 or LP3 was detectable, the presence of significant amounts of contaminating LPS was ruled out.
TLR-2 is known to form heterodimers with either TLR-1 or -6, and MALP-2 has been identified to interact with TLR-2/TLR-6 heterodimers (28). In contrast, triacylated lipopeptides have been shown to be recognized via TLR-2/TLR-1 (29). Therefore, we tested whether the synthetic lipopeptides studied here could also be discriminated by these receptors. Because HEK293 cells are known to constitutively express both TLR-1 and TLR-6 (51), we overexpressed these receptors and determined whether this had any effect on activation by the lipopeptides. Overexpression of TLR-6 significantly increased the responsiveness to LP2, whereas TLR-1 had the same effect on LP3 (Fig. 1⇑C). However, we also observed a trend toward an increased stimulation of HEK293 by LP3 upon overexpression of TLR-6 (Fig. 1⇑C) confirming a previous report (30).
Interaction of synthetic lipopeptides with CD14
Although triacylated lipopeptides have been shown to interact with CD14 to cause cytokine induction (47), there are no reports on the potential involvement of CD14 in immune responses caused by diacylated lipopeptides. Studies with other Mycoplasma lipoproteins reported a stimulatory pathway independent of CD14 (52, 53). Therefore, we tested whether cytokine induction in human monocytes can be blocked by the CD14-blocking Ab My4. Human monocytes isolated from peripheral blood were stimulated, and TNF-α-production was assessed after 4 h. Both synthetic lipopeptides led to a comparable increase in cytokine production by these cells in the presence of serum; however, LP2 was ∼10 times more active compared with LP3 (Fig. 2⇓, A and B). When human monocytes were preincubated with My4, no cytokine induction could be observed, indicating that cellular activation caused by diacylated lipopeptides is also CD14 dependent (Fig. 2⇓, A and B).
In a different approach, we aimed at verifying this finding by studying the potential direct interaction of LP2 with CD14 using PhastGel native gel analysis. Soluble CD14, when preincubated with LP2 at a molar excess, displayed a shift toward lower electrophoretic mobility, indicating the formation of complexes of CD14 with LP2 (Fig. 2⇑, C and D).
To further verify these results, we used CHO cells expressing neither CD14 nor functional TLR-2, the latter due to a frameshift mutation that causes a C-terminally truncated protein (54). These cells were transfected with human TLR-2, β-galactosidase, the ELAM NF-κB reporter plasmid, and either human CD14 or its corresponding empty vector. Baseline NF-κB activity in CD14-transfected cells was ∼30% higher compared with controls (not shown); that is why data were analyzed as fold increase in luciferase activity. Stimulation with both LP3 and LP2 induced a marked increase in NF-κB-mediated cellular activation, which was significantly greater when CD14 was present (Fig. 3⇓), indicating that cellular recognition of both compounds is facilitated by this coreceptor. However, in these cells, in contrast to the human monocytes, LP3 turned out to be more active compared with LP2, possibly due to different expression ratios of TLR-1 and -6 compared with monocytes.
Cytokine induction in human monocytes is increased in the presence of human LBP
In the first experiments on cytokine induction caused by synthetic lipopeptides, we observed an increase in cellular activation in the presence of serum (data not shown). To test whether this increase was caused by LBP present in serum, we used human monocytes stimulated with lipopeptides in the presence or absence of 1 μg/ml recombinant human LBP. Both lipopeptides induced TNF-α release by human monocytes in the absence of serum; however, cytokine levels induced by LP3 were lower as compared with LP2 (Fig. 4⇓, A and B). In the presence of LBP, cytokine levels caused by both compounds were significantly increased. Both lipopeptides led to a marked cytokine production at concentrations of 0.17 ng/ml when LBP was present, whereas in the absence of LBP, higher concentrations of lipopeptides were required for detectable induction of TNF-α in serum-free medium (Fig. 4⇓, A and B). When these higher concentrations of LP3 and LP2 were used, cytokine levels were increased ∼25-fold in the presence of LBP (Fig. 4⇓, A and B).
LBP has been shown by us to inhibit cytokine induction caused by LPS both in vivo and in vitro when present at high concentrations similar to serum concentrations found during sepsis (32, 42). In subsequent studies, we were able to demonstrate that high-dose LBP also is able to inhibit the immunostimulatory activity brought about by LTA of Gram-positive bacteria (48). Thus, we wanted to know whether cellular activation caused by lipopeptides is also differentially influenced by varying concentrations of LBP. To this end, we stimulated human monocytes with different concentrations of lipopeptides in the presence of increasing concentrations of LBP (Fig. 4⇑, C and D). Surprisingly, we failed to observe any inhibitory effect of high-dose LBP on cellular activation by lipopeptides even if these compounds were administered at low concentrations. As expected, inhibitory LBP effects were also absent when LP2 and LP3 were added at higher concentrations (data not shown).
LBP binds to diacylated and triacylated lipopeptides
We next aimed at verifying the data obtained for cytokine induction by binding studies with LBP and lipopeptides. Because both compounds did not cause a marked shift in electrophoretic mobility of LBP when investigated by PhastGel analysis (data not shown), we performed an LBP-binding assay with different microbial ligands coated to microtiter plates. Plates were coated with LPS, LP3, and LP2, as well as its naturally occurring counterpart, MALP-2, followed by incubation with recombinant human LBP and subsequent detection using biotinylated anti-LBP Abs and streptavidin peroxidase. We observed a marked increase in absorption when the wells were coated with LP2, LP3, or MALP-2, comparable to the increase seen with LPS, indicating that LBP binds to these lipopeptide ligands (Fig. 5⇓A). This increase was not present when LBP was added to the plates with LPS, suggesting a competition of LPS with lipopeptides for the binding to LBP. In another set of experiments, plates coated with LPS were incubated with LBP together with increasing concentrations of LP2, LP3, and MALP-2 (Fig. 5⇓B). In these assays, the two diacylated lipopeptides, LP2 and MALP-2, inhibited binding of LBP to LPS immobilized on the plates in a comparable fashion to LPS itself (Fig. 5⇓B), indicating a competition for the same binding site. However, triacylated LP3 did not display this inhibition, even at the highest concentration tested.
Transfer of FLAG-tagged lipopeptides to human peripheral blood monocytes by LBP
Next, we investigated whether LBP was able to mediate the transfer of lipopeptides to CD14-bearing cells. For this purpose, we used FLAG-tagged derivatives of LP2 and LP3 sharing the Pam2Cys and Pam3Cys motifs with the original nonderivatized lipopeptides while containing an additional FLAG epitope at their C-terminal lysine residues. The FLAG sequence did not alter the biologic activities of these compounds in comparison to LP2 and LP3 (data not shown). Human peripheral blood monocytes were incubated with 10 nM LP3-FLAG or LP2-FLAG (14 and 15 ng/ml, respectively) in the presence or absence of recombinant human LBP. Then, cells were incubated with anti-FLAG Abs and subsequently with Cy5-labeled anti-IgG Abs, and fluorescence of gated monocytes was assessed by FACS analysis. In the presence of both lipopeptides, a small increase in fluorescence was observed (Fig. 6⇓, A and B), which was stronger for LP2-FLAG compared with LP3-FLAG, in line with the data on cytokine induction indicating a lower potency of LP3-FLAG to interact with cells in serum-free medium (Figs. 2⇑, A and B; 4⇑, A and B). However, in the presence of LBP, a profound increase in fluorescence was observed for both lipopeptides in a comparable manner (Fig. 6⇓, A and B), indicating that LBP is able to transfer lipopeptides irrespective of their acylation pattern to monocytes. To verify that the transfer of lipopeptides observed was specific, we used a blocking CD14 Ab. In these experiments, no transfer of either LP3-FLAG and LP2-FLAG occurred in the presence of MEM-18 (Fig. 6⇓C), indicating that LBP transfers both lipopeptides to CD14.
Lipoproteins are well-defined molecular patterns of a large variety of pathogenic microorganisms with the capability to stimulate APCs. They have long been known as strong cytokine inducers and have recently become the basis for vaccine development, making them key molecules of innate and adaptive immunity (16, 55). We show in this study that a host acute-phase protein, LBP, which previously has been known to be a key molecule for pathogen recognition via LPS, LTA, and peptidoglycan, is also crucial for lipoprotein recognition. Moreover, we obtained several lines of evidence that CD14 is also involved in cellular responses to triacylated and diacylated lipopeptides. In confirmation of prior studies, lipopeptides displayed a pattern of recognition via TLR-2 in differential combinations with TLR-1 and -6 related to their different degree in acylation. Thus, LBP apparently has the ability to recognize a wide variety of microorganisms and to mediate the initiation of signal transduction not only via TLR-4/MD-2, but also via TLR-2/-1 or TLR-2/-6 as suggested in Fig. 7⇓.
The Pam3Cys motif of lipoproteins is abundant among different bacteria and has been described for Enterobacteriaceae (20), spirochetes (19, 21), mycobacteria (23), and mycoplasms (24). In contrast, the Pam2Cys motif exhibiting two fatty acids has only been shown to be present in mycoplasms (25). Because the Pam3Cys motif is present in such a large variety of microorganisms and corresponding molecular structures are absent in higher vertebrate species, it may represent a common structure discriminating pathogens from the host. In this study, we investigated the biological characteristics of synthetic triacylated lipopeptides corresponding to triacylated lipoproteins, as well as diacylated lipopeptides corresponding to MALP-2. These derivatives have been previously described to induce cytokines in immune cells in a dose-dependent fashion comparable to their naturally occurring counterparts (26), and therefore appear to be suitable for studying innate immune responses to lipopeptides in general. Although the Pam3CSK4 lipopeptide was previously demonstrated to resemble naturally occurring lipoproteins regarding use of TLRs (28, 29), no data have been available for the Pam2CSK4 lipopeptide. We show in this study that both lipopeptides interact with TLR-2 in synergy with TLR-1 or -6 as previously reported. We were able to confirm the data obtained with MALP-2 (28) with its synthetic counterpart LP2: We found that TLR-2 was essential for stimulation of HEK293 cells by this compound, and its stimulatory capacity could be significantly enhanced by overexpression of TLR-6. Cotransfection of both TLR-1 and, to a lesser extent, TLR-6 increased stimulation of HEK293 cells by triacylated lipopeptides. This finding is partially contradictory to previous reports identifying TLR-1 as the sole coreceptor (29), while supporting results of another group using a dominant-negative mutant of TLR-6 (30).
We provide evidence that innate immune recognition of lipoproteins is mediated by LBP, a molecule that until recently was thought of as interacting with LPS exclusively. Cytokine induction in human monocytes caused by diacylated and triacylated lipopeptides was greatly enhanced in the presence of LBP, and the interaction of LBP with the lipopeptides could be verified by microplate binding assays. LBP bound to plates coated with diacylated and triacylated lipopeptides as well as MALP-2 in a manner comparable to LPS, and the specificity of this binding could be demonstrated by inhibition of LBP binding to the plates when LPS was added. However, when microtiter plates coated with LPS were incubated with lipopeptides, only LP2 and MALP-2 were able to compete with LPS, whereas LP3 did not block binding of LBP to LPS immobilized on the plate. At this point, it cannot be ruled out that this was caused by a lesser affinity of LP3 to LBP compared with LP2 and MALP-2. However, it is tempting to speculate that the lack of competition of LP3 with LPS revealed in these experiments is caused by the use of different binding sites within LBP. In contrast, LP2 and MALP-2 obviously bind to LBP via the same or at least adjacent binding sites within the N-terminal part (50), a feature that we previously demonstrated for two glycolipids exhibiting a lipid anchor with two fatty acids, LTA of Gram-positive bacteria and Treponema glycolipids (15, 56). As demonstrated by FACS analysis, LBP transfers both triacylated and diacylated lipopeptides to CD14 because transfer was blocked by an anti-CD14 Ab. Our results imply that the function of LBP is not restricted to structures bearing carbohydrate polymers, such as LPS, LTA (15, 48), mycobacterial lipomannan (57), or muropeptides linked to teichoic acid (31), but also includes interaction with lipoproteins or lipopeptides, both of the triacylated and diacylated subtypes. Additionally, by using inhibiting CD14 Abs, native shift gel electrophoresis, transfection studies, and FACS analysis, we were able to demonstrate for the first time that CD14 is also involved in recognition of diacylated lipopeptides, a feature that was previously demonstrated only for triacylated lipopeptides (47).
The predominant biologic effect of LBP initially was thought to be an amplification of cytokine response during sepsis, potentially leading to deleterious effects such as septic shock (58, 59). Studies with LBP knockout mice indicated that LBP serves in preventing bacterial infections, because LBP−/− mice are more susceptible to infection by Gram-negative bacteria compared with their wild-type littermates (41). Additionally, our previous results showed that LBP could protect mice from an otherwise lethal septic shock induced by LPS or whole bacteria (42). These findings most likely are based on the fact that LBP is able to bind bacterial ligands and transfer them from the cellular receptor into other compartments of the host for detoxification (60). However, in the experiments described in this paper, lipopeptides differed from LPS and LTA regarding the fact that cytokine induction could not be inhibited by high concentrations of LBP.
LBP is not only synthesized in hepatocytes, but has also been found to be expressed in alveolar epithelial cells (36) and intestinal cells (37). Thus, LBP is apparently present in compartments of the organism exposed to large quantities of microorganisms, and may play a key role in early recognition of pathogens. Lipoproteins and lipopeptides bearing the Pam2Cys or Pam3Cys motif are abundant among pathogens of the respiratory tract, such as Mycoplasma pneumonia and Mycobacterium tuberculosis. Our own previous studies have revealed that major cell wall components of Streptococcus pneumoniae and Staphylococcus aureus causing pneumonia, are also recognized by LBP (31, 48). Thus, recognition of molecular patterns by LBP may be an important step in the initiation of a host response during pneumonia. In addition, our data suggest that LBP, by recognizing lipoproteins, may also play a role in chronic or cyclic infections: With the exception of Mycoplasma-caused pneumonia, clinically important pathogens carrying lipoproteins represent causes of slow, cyclic or chronic infectious diseases, e.g., Lyme disease or tuberculosis. Even mycoplasm pneumonia, although frequent in children, is often afebrile and self-limiting in contrast to pneumonia caused by other pathogens (61). Our data suggest that recognition of these slow-acting microorganisms may also be accomplished by LBP. Synthetic lipopeptides have been the basis for broad vaccine development lately for a wide variety of diseases including HIV, malaria, CMV, and HSV (reviewed in Ref. 16). Understanding of the molecular mechanisms of host recognition of lipopeptides and lipoproteins in more detail may further help to further improve these innovative strategies.
We acknowledge the excellent technical assistance of Sabine Bobbe and Diana Woellner (Berlin, Germany) as well as Nina Grohmann and Ina Goroncy (Borstel, Germany).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 This work was supported in part by grants given by the Bundesministerium für Bildung und Forschung to R.R.S. (CAPNetz), by Clinique La Prairie Research to C.A., and the Deutsche Forschungsgemeinschaft to R.R.S. and N.W.J.S. (Schr 726/1-1) as well as H.H. (He 2758/3-1).
↵2 Current address: Pediatrics Infectious Diseases, Cedars-Sinai Medical Center, Davis Research Building, 110 George Burns Road, Los Angeles, CA 90048.
↵3 Address correspondence and reprint requests to Dr. Ralf R. Schumann, Institut für Mikrobiologie und Hygiene, Charité-Universitätsmedizin Berlin, Dorotheenstrasse 96, D-10117 Berlin, Germany. E-mail address:
↵4 Abbreviations used in this paper: sCD14, soluble CD14; LBP, LPS binding protein; LTA, lipoteichoic acid; MALP-2, 2-kDa macrophage-activating lipopeptide; h, human; LP, lipopeptide.
- Received December 16, 2003.
- Accepted June 14, 2004.
- Copyright © 2004 by The American Association of Immunologists