Recognition of Pneumococcal Peptidoglycan

Immunity, Vol. 19, 269–279, August, 2003, Copyright 2003 by Cell Press

Recognition of Pneumococcal Peptidoglycan: An Expanded, Pivotal Role for LPS Binding Protein Joerg R. Weber,1,* Dorette Freyer,1 Christian Alexander,1,3 Nicolas W. J. Schro¨der,1,2 Anja Reiss,1 Carsten Ku¨ster,1 Dagmar Pfeil,1,2 Elaine I. Tuomanen,1,4 and Ralf R. Schumann1,2 1 Department of Neurology 2 Department of Microbiology and Hygiene Universitaetsklinikum Charite´ Humboldt University 10117 Berlin Germany 3 Department of Immunochemistry and Biochemical Microbiology Research Center Borstel Center of Medicine and Bio-Sciences 28845 Borstel Germany 4 Department of Infectious Diseases St. Jude Children’s Research Hospital Memphis, Tennessee 38105

Summary Lipopolysaccharide binding protein (LBP) has a wellestablished role in LPS-induced immune responses. Here, we report that LBP also plays an essential role in the innate immune response to Gram-positive pneumococci, specifically to their major inflammatory component, pneumococcal cell wall (PCW). LBP was present in the CSF of patients with meningitis, and LBP-deficient mice failed to develop meningeal inflammation. LBP enhanced PCW-induced cell signaling and TNF-␣ release. LBP bound specifically to PCW multimers, indicating novel lipid-independent binding capability for LBP. We propose the iterative anionic groups along the glycan backbone of the cell wall are a crucial structure for recognition by LBP. Such a function for LBP expands its role to Gram-positive infections. Introduction Streptococcus pneumoniae is the major pathogen causing community-acquired pneumonia and bacterial meningitis in Europe and the US (Dowell et al., 2000; Schuchat et al., 1997). Especially devastating even in the context of modern antibiotics is the 34% mortality of pneumococcal meningitis (de Gans and van de Beek, 2002). Despite its medical importance, the molecular mechanisms of pneumococcal interactions with the innate immune system are still surprisingly poorly understood. Being a Gram-positive pathogen, pneumococcus lacks the familiar endotoxin that drives the majority of Gram-negative immune responses. Rather, the pneumococcal cell surface is composed of membrane-bound lipoteichoic acid (LTA) encased in a cell wall consisting of a multilayered network of peptidoglycan (PG) with *Correspondence: [email protected]

attached teichoic acid (TA) chains. This combination of surface elements induces the especially intense inflammatory response recognized by clinicians to be characteristic of pneumococcal infection. One interaction of the pneumococcal cell wall with the innate immune response involves recognition of the unusual phosphorylcholine residues of the TA by C-reactive protein (CRP) (Winkelstein and Tomasz, 1978). This binding triggers the alternative pathway of complement activation and promotes phagocytosis but is not known to induce proinflammatory signaling (Holzer et al., 1984). Phosphorylcholine on the pneumococcal surface also binds to the platelet activating factor (PAF) receptor (Cundell et al., 1995). However, rather than signaling via G-proteins like the chemokine PAF, bacterial binding to the PAF receptor leads to endocytosis and bacterial invasion of eukaryotic cells (Ring et al., 1998). Thus, the known interactions of PCW with the innate immune system do not explain the majority of features of pneumococcal inflammation. Purified PCW induces meningeal inflammation indistinguishable from meningitis caused by living bacteria in the early phase of the disease (Tuomanen et al., 1985; Weber et al., 1995). In vitro, purified PCW induces the release of TNF-␣ from astrocytes and brain microvascular endothelial cells (BMEC) as well as numerous cytokines and chemokines from microglia (Freyer et al., 1999). The signal transduction events involved in this activation process include phosphorylation and activation of the MAP kinases erk 1/2 and p38 (Schumann et al., 1998), as well as the translocation of the transcription factor NF-␬B (Spellerberg et al., 1996). In astrocytes and in THP-1 cells, the effect of the purified PCW seems to be partially CD14 dependent (Cauwels et al., 1997; Schumann et al., 1998). Recently, it has been shown that heat-killed pneumococci and soluble pneumococcal PG induce signaling by a member of the family of toll-like receptors TLR-2 (Yoshimura et al., 1999). The immunostimulatory effect of soluble, fragmented PCW is clinically of significant importance because of the tremendous increase in its release during antibiotic-induced lysis of bacteria (Tuomanen et al., 1987; Schneider et al., 1999; Fischer and Tomasz, 1984). This process presents a large library of chemically diverse, soluble, and insoluble PCW fragments to the innate immune response apparatus, all of which must be recognized and detoxified or inflammation will persist. Lipopolysaccharide binding protein (LBP) is an acute phase protein synthesized mainly in the liver but also in pulmonary and gastrointestinal epithelia cells (Vreugdenhil et al., 2000). LBP was first recognized for its ability to bind endotoxin (LPS) (Ulevitch and Tobias, 1995; Raetz and Whitfield, 2002) and directs it to host cellular receptors (Schumann et al., 1990). Recently, it has been shown that it also binds to glycolipids of spirochetes, mycobacteria (Hermann et al., 2002; Savedra et al., 1996), and lipoteichoic acid of Gram-positive bacteria (Fan et al., 1999; Schroder et al., 2000, 2003), all by a process

Immunity 270

Figure 1. Requirement for LBP for Induction of Meningeal Inflammation by PCW All mice except controls were challenged intrathecally with PCW at the beginning of the experiment. One group of LBP⫺/⫺ was reconstituted with exogenous, intrathecal murine rLBP (0.5 ␮g) at 0 hr and leukocytes counted at 6 hr (n ⫽ 6). rLBP alone had no effect in LBP⫹/⫺ and LBP⫺/⫺ (n ⫽ 3, n ⫽ 2; ⬍5 leukocytes/␮l as PBS controls). # indicates a difference between the LBP⫹/⫺⫺ ([A] black bar; [B] closed circle) and PCW-injected mice (n ⫽ 7) compared to controls (n ⫽ 5; p ⬍ 0.05); * indicates a statistically significant decrease in LBP⫺/⫺ ([A] white bar; [B] open circle; n ⫽ 7; p ⬍ 0.05, ANOVA, Student-Newman-Keuls test).

directed toward recognition of lipid linked to carbohydrate structures. LBP enhances the binding of one fragment of staphylococcal peptidoglycan to sCD14 (Dziarski et al., 1998). LBP is present at 5–10 ␮g/ml in human serum, and this concentration increases dramatically during the acute phase response (Zweigner et al., 2001), including in Gram-positive sepsis (Blairon et al., 2003). In low concentrations, LBP enhances LPS-effects in vitro. However, high concentrations such as that present in serum during sepsis attenuate LPS inflammation and enable the host to detoxify LPS (Lamping et al., 1998; Zweigner et al., 2001; Hamann et al., 2000). The LBP knockout mouse has been shown to fail to initiate an inflammatory response to LPS (Jack et al., 1997; Wurfel et al., 1997) and is more susceptible to lethal infection with living Gram-negative bacteria (Fierer et al., 2002; Heinrich et al., 2001; Yang et al., 2002). Although the crystal structure of LBP has not been elucidated yet, a series of residues essential for LPS binding has been identified by mutagenesis, and the corresponding LBP mutants lacking LPS binding have been functionally analyzed (Lamping et al., 1996; Schumann et al., 1997). Based on significant overall sequence similarities and mutagenesis data, a model of the threedimensional structure of human LBP has been deduced using the boomerang-like crystal structure of human bactericidal/permeability increasing protein (BPI) (Beamer et al., 1997, 1998b, 1998a).

Results The Role of LBP in Pneumococcal Meningitis To verify the presence of LBP at a site of pneumococcal inflammation in vivo, LBP concentrations were measured in the CSF of patients with culture-proven pneumococcal meningitis and in controls with normal CSF chemistry. Significantly higher LBP concentrations were found in CSF of meningitis patients (228 ⫾ 193 ng/ml, n ⫽ 5) compared to controls (10 ⫾ 12 ng/ml, n ⫽ 8; p ⫽ 0.003, Mann-Whitney Rank Sum Test). A further indication for the potential role of LBP to be a mediator in this context is the finding that pneumococcal cell wall coprecipitated with LBP in the CSF of a patient with confirmed meningitis (data not shown). To ascribe a biological role for LBP in pneumococcal inflammation, the course of meningitis was compared in heterozygous LBP⫹/⫺ mice and LBP⫺/⫺ mice. Heterozygous LBP⫹/⫺ mice (exhibiting levels of LBP equal to wild-type littermates (Jack et al., 1997), experienced an equivalent and brisk influx of leukocytes into the CSF by 6 hr (Figure 1A), and histopathological analyses confirmed severe meningitis (data not shown). In contrast, leukocyte influx was markedly decreased in LBP⫺/⫺ mice at 6 hr (Figure 1A), an effect that persisted at 12 and 24 hr (Figure 1B). Intrathecal administration of recombinant murine LBP (rLBP) in the LBP⫺/⫺ mice restored PCWinduced leukocyte influx and meningitis to levels in

LBP and Pneumococci 271

Figure 2. Effect of LBP on PCW-Induced Phosphorylation of erk1/2 Kinases and Induction of TNF-␣ Production in Vitro Cells were stimulated with PCW in the presence and absence of LBP. TNF or NO was measured in the supernatants or cells were harvested for MAPK-assay and Western blotting. (A) Assessment of MAP kinase activity in BMEC cells at 30 min. (B) Detection of phosphorylation of erk1/2 in primary BMEC cells at 30 min by Western blot. (C) Amount of TNF-␣ secreted into the supernatant of BMEC and (D) RAW 264.7 cells at 6 hr. * indicates a statistically significant difference p ⬍ 0.01 compared to the stimulation without LBP.

LBP⫹/⫺ animals (Figure 1A). Consistent with this decreased inflammatory response capability, LBP⫺/⫺ mice challenged with living pneumococci showed decreased leukocytosis in the CSF after 24 hr (133 ⫾ 96 leukocytes/ ␮l versus 667 ⫾ 134; p ⫽ 0.001; n ⫽ 5 per group). Infection was confirmed by ⬎106 CFU/ml CSF (no statistically significant difference between the groups); all animals survived the 24 hr experiment. PCW-Induced Cellular Signaling Is Increased by LBP To determine if LBP participates in this arm of pneumococcal-induced signaling, BMEC were stimulated with PCW in the absence and presence of murine rLBP, and tyrosine phosphorylation of erk1/erk2 was assessed by Western blot. LBP enhanced the PCW-induced phos-

phorylation of erk1/erk2 in a dose-dependent manner (Figure 2B). The phosphorylation was accompanied by increased MAP kinase activity assessed by an MBP assay that was rLPB dependent (Figure 2A). Kinase activity reached a maximum at 0.5 ␮g/ml rLBP, a value consistent with the concentration of LBP documented in vivo. Higher concentrations of LBP reduced tyrosine phosphorylation and activity of erk1/erk2 in a fashion similar to that recently observed for LPS-dependent cellular activation (Lamping et al., 1998; Hamann et al., 2000). Effect of LBP on PCW-Induced TNF-␣ and NO Release In order to determine the proinflammatory consequences of LBP-dependent PCW-induced signaling, the

Immunity 272

an apparent KD of 103 ⫾ 20 nM (Figure 4A). The amount of binding also varied as a function of the concentration of PCW attached to the ELISA plate (Figure 4B). LBP binding at 5 and 10 ␮g/ml was comparable to that observed for the anti-phosphorylcholine antibody TEPC15. Intact pneumococcus was also recognized by LBP in a dose-dependent manner (Figures 4C and 4D). LBP Does Not Bind to the Phosphorylcholine Groups of PCW To determine if the phosphorylcholine adduct serves as a substrate for LBP, PCW containing varying amounts of phosphorylcholine were made. LBP binding to PCW was not affected by the absence (ethanolamine substituted PCW) or decrease of choline content (PCW licD2⫺ mutant; Figure 5A). In contrast, anti-choline TEPC-15 Ab binding to the ethanolamine-substituted pneumococcal wall was abrogated and binding to the licD2⫺ mutant PCW was decreased (Figure 5B).

Figure 3. Dependence of Cellular Activation by PCW on TLR-2 HEK293 cells stably transfected with CD14 and MD-2 and bearing the NF-␬B-dependent ELAM promoter driving ␤ galactosidase and luciferase expression were challenged with PCW. (A) Cells transfected with TLR-2 responded to lipoprotein (LP) as a positive control and to PCW (107 CFU equivalents, p ⬍ 0.001 versus control; 106 CFU equivalents yielded 1.06 ⫾ 0.23 rel. light units p ⫽ 0.03 versus control). (B) Cells transfected with TLR-4 did not respond to PCW but did respond to LPS (100 ng/ml).

effect of LBP on PCW-induced TNF-␣ release was determined. rLBP alone had no effect compared to PBStreated cells (RAW cells: 84.6 ⫾ 4.3 versus 78 ⫾ 3.0 pg/ ml; BMEC 3.2 ⫾ 4.5 versus 5.2 ⫾ 6.9). Addition of rLBP significantly enhanced PCW-induced TNF-␣ release both from CD14 negative BMECs (Figure 2C) and to an even greater degree from the CD14 positive murine macrophage cell line RAW 264.7 (Figure 2D). Addition of 0.5 ␮g rLBP shifted the threshold of PCW induced TNF-␣ release one order of magnitude in both cell types. PCW-induced production of nitric oxide was not modified by addition of rLBP in BMEC (no LBP, 63.8 ⫾ 3.2 ␮g/ml nitrite; 0.1 ␮g/ml LBP, 64.9 ⫾ 2.7; 0.5 ␮g/ml LBP, 67.2 ⫾ 1.1; 1 ␮g/ml LBP, 70 ⫾ 4) and was also unchanged in RAW 264.7 cells (data not shown). TLR-2 Transduces the PCW Response HEK293 cells devoid of TLR’s did not respond to PCW, LPS, or lipopeptide (LP; data not shown). Transfection with hTLR-2 resulted in strong activation of NF-␬B after stimulation with PCW and LP, but not LPS (Figure 3A), whereas cells transfected with hTLR-4 responded to LPS, but not LP or PCW (Figure 3B). Binding of LBP to the Pneumococcal Cell Wall and Intact Pneumococci Binding of LBP to immobilized PCW or intact pneumococci was measured in an ELISA format. LBP bound to PCW in a saturable and dose-dependent fashion with

Minimal Structure of PCW Required for LBP-Dependent Bioactivity PCW is a complex polymer built from two major parts, peptidoglycan (PG) and teichoic acid (TA). To determine which of the two components was critical to LBP-dependent bioactivity, PCW and TA-free PCW were tested for the ability to induce LBP-dependent TNF-␣ release from RAW 264.7 cells. Both PCW preparations demonstrated equivalent bioactivities, and these activities were equally augmented by LBP (Figure 5C). This implies that the teichoic acid chains are not required for bioactivity or for LBP recognition of PCW. Having eliminated the TA, we enzymatically dissected the PG into components to determine the minimal bioactive structure. Digesting PCW with N-acetylmuramylL-alanine amidase severs the stem peptides from the glycan backbone, yielding high molecular weight glycan chains and a mixture of free monomeric and multimeric peptides. Although amidase-treated PCW induced about half the TNF-␣ compared to whole PCW, addition of LBP still doubled the TNF-␣ release, an amplitude similar to that seen with whole PCW (Figure 5C). This result suggested that a multimeric backbone was important to LBP recognition. This was confirmed by two further results. Degradation of the glycan chains of the amidase product to disaccharide fragments by muramidase greatly decreased bioactivity and LBP augmentation (Figure 5C). Digestion of whole PCW only with muramidase so as to produce a fragmented glycan backbone still maintained as a multimer by crosslinked stem peptides, yielded material of intermediate bioactivity that still showed augmentation by LBP (Figure 5C). Thus, LBP doubled the bioactivity of PCW maintained as a multimer by glycan or peptide crosslinking, or both. Thus, although the inherent bioactivity of PCW without LBP did not survive loss of glycan backbone integrity, the bioactivity of such floppy PG held together only by peptide crosslinks could be salvaged by presentation on the framework of LBP. If LBP provides a scaffold to stabilize multimeric PG, then immobilizing PG fragments would be predicted to at least partially reconstitute recognition of even small fragments. Direct binding of LBP to immobilized PCW

LBP and Pneumococci 273

Figure 4. Direct Binding of LBP to PCW and Intact Pneumococci PCW (A and B) or intact pneumococci (C and D) were bound to a microtiter plate, and the ability to capture LBP was measured by ELISA. Attachment of PCW to the ELISA plate was confirmed by detection of capture of antiphosphorylcholine mAb TEPC-15 (B and D). (A) Binding of LBP to PCW in a saturable and dose-dependent fashion. (B) Binding of LBP to PCW as a function of cell wall and LBP concentration (50 ⫽ 5 ⫻ 106 CFU equivalents). (C) Binding of LBP to intact pneumococci (5 ⫻ 106 CFU equivalents) as a function of LBP concentration. (D) Binding of LBP (1 ␮g/ml LBP) to intact pneumococci as a function of bacterial concentration (100 ⫽ 2 ⫻ 107 CFU equivalents).

fragments was tested in the ELISA assay. The amidasetreated preparation (intact glycan backbone and free peptides) captured as much LBP as whole PCW (97 ⫾ 3.6 versus 100 ⫾ 7.4%), a result consistent with the similar high LBP augmentation of the bioactivity of these preparations. Loss of glycan integrity by muramidase digestion decreased LBP capture to 73 ⫾ 4.6% of whole PCW, a result in agreement with the bioactivity assay. Finally, completely fragmented, double-digested PCW, which was biologically inactive in solution, could capture LBP when immobilized, albeit at a reduced amount (64 ⫾ 11.2% of whole PCW).

cantly reduced the capacity to bind to immobilized PCW in a concentration-dependent fashion (Figure 6A). As binding of human rLBP to LPS has been shown to critically depend on the basic amino acid residues arginine 94 and lysine 95, a mutant LBP containing glutamine substitutions at these sites so as to eliminate binding of LPS (Lamping et al., 1996) was tested for binding to PCW. Binding of the mutant LBP to PCW was 70% less than wild-type LBP (Figure 6B). Taken together with the competition results, these data suggested that the binding site for PCW might overlap that for LPS on LBP. Discussion

The PCW Binding Region on LBP To determine the structural requirement within the LBP protein needed for the LBP-PCW interaction, we first assessed whether PCW could compete with LPS for binding to LBP. Preincubation of LBP with LPS signifi-

The innate immune system recognizes major classes of pathogens based on molecular patterns (Janeway and Medzhitov, 2002). The pattern for Gram-positive pathogens has not been defined even though more than half

Immunity 274

Figure 5. Dependence of Binding of LBP on Structure of the PCW Direct binding of LBP to PCW and intact bacteria assessed in the ELISA assay. (A) LBP (2.5 ␮g/ml) binding to immobilized PCW, choline-free PCW bearing ethanolamine (EA), and PCW lacking one of two choline moieties (LicD2⫺). (B) Anticholine mAb TEPC-15 binding to PCW. Note lack of antibody binding to PCW EA confirms absence of choline. (C) TNF-␣ release at 6 hr from RAW 264.7 cells treated with PCW fragments (5 ⫻ 106/ml CFU equivalents) was measured in the absence (open bars: 100% ⫽ 2860 ⫾ 1024 pg/ml TNF-␣) and presence (solid bars) of LBP (1 ␮g/ml). Fragments immobilized: PCW; PCW treated with trichloroacetic acid to remove the TA (PCWTCA); PCW digested with amidase to sever stem peptides (PCW-A); PCW digested with muramidase to break the glycan backbone (PCW-M); PCW digested with both enzymes (PCW-AM).

of all invasive disease is caused by this class of bacteria. Following the paradigm for LPS, a soluble binding protein capable of recognizing key elements on the Gram-

positive, surface-exposed cell wall would be anticipated to exist. Drosophila has been found to harbor several such peptidoglycan recognition proteins, and one has a potential homolog in humans (Michel et al., 2001; Liu et al., 2001). This report reveals the surprising identity of such a protein to be LBP, a protein historically thought to be dedicated to Gram-negative infection. Here, we report that LBP also plays an essential role in the innate immune response to Gram-positive pneumococci and, specifically, to their major inflammatory component, PCW. The glycan backbone of PG seemed to be a crucial structure for recognition by LBP, a finding that extends the motif recognition capabilities of LBP from glycolipids to complex, multimeric carbohydrates. As an acute phase protein, a high serum concentration of LBP has been reported in severe Gram-positive sepsis and is correlated to a poor outcome of disease (Opal et al., 1999). We found significantly increased concentrations of LBP in the CSF of patients with proven pneumococcal meningitis. This suggests that this response is not just collateral to activation of the acute phase response but participates in specific immune recognition. Leukocyte influx into the CSF, the hallmark of bacterial meningitis, was drastically reduced in LBP null mice compared to heterozygous mice with normal levels of LBP. This was seen in PCW as well as in living pneumococci-induced meningitis. Furthermore, the reduction of leukocytes in the LBP⫺/⫺ mice persisted for over 24 hr after challenge and inflammation was restored by resupplying exogenous rLBP. This is consistent with the lack of symptoms of shock in galactosamine sensitized LBP⫺/⫺ mice injected intraperitoneally with LPS (Jack et al., 1997; Wurfel et al., 1997). Reducing leukocytes in the CSF is believed to be of potential therapeutic benefit because high numbers of CSF leukocytes correlate to an unfavorable clinical outcome and the reduction of leukocytes is associated with a decrease of brain edema, intracranial pressure, hyperemia, neuronal damage, and death (Tuomanen et al., 1989; Weber et al., 1997; Braun et al., 1999). However, decreased inflammatory parameters, such as lower TNF concentrations and leukocytes, result in much higher bacterial numbers and earlier death in LBP null animals with Salmonella sepsis

Figure 6. Competition of LPS with PCW Binding to LBP (A) LBP at 1 ␮g/ml (solid bars) or 2.5 ␮g/ml (open bars) was incubated with the indicated concentration of LPS for 30 min and then allowed to bind to immobilized PCW for 2 hr. Captured LBP was detected by ELISA. (B) The amount of wild-type (solid bar) or mutant (open bar) LBP R94E/K95E bound to immobilized PCW. * indicates a statistically significant difference compared LBP binding without LPS (A) or to normal LBP (B) (p ⬍ 0.05).

LBP and Pneumococci 275

Figure 7. Model of LBP Binding to PCW: The Iterative Polyanionic Motif The structural model of human LBP was calculated by the SWISS-MODEL server (Guex and Peitsch, 1997) utilizing the crystal structure of human BPI (Protein Data Bank Code: 1BP1) as the template (Beamer et al., 1997). A series of basic residues in the N-terminal domain of LBP is highlighted in white. Basic side chains in the apical region required for LPS binding are marked in yellow. Inset: mutant LBP (R94E/K95E) with altered residues shown in red. Detailed chemical structures of intact PG (A) and PG digested with amidase (B) are shown with colored arrows indicating the positions of PCW cleavage by trichloroacetic acid (green), amidase (blue), and muramidase (red) treatment, respectively. In the presentation of pneumococcal PG (A), an average muropeptide structure displaying heterogeneity in the C-terminal stem peptide residues (D-Ala or L-Lys) is shown. The ⑀-amino groups of the lysyl side chains have been identified to exist either in an unsubstituted form (R: H), carrying dipeptide substituents (R: L-Ser-L-Ala or L-Ala-L-Ala), or forming direct as well as dipeptide-mediated crosslinking structures to D-alanine residues in stem peptides of neighboring peptidoglycan strands (Garcia-Bustos et al., 1987; Garcia-Bustos and Tomasz, 1990; Severin and Tomasz, 1996). Arrays of anionic residues (shaded in yellow) are iterative in both the repeating PG polymer and the repeating glycan backbone. Such multimeric arrays are proposed to form a motif to contact the N-terminal domain of LBP as indicated by black arrows.

(Fierer et al., 2002; Heinrich et al., 2001; Yang et al., 2002). Thus, attenuating LBP-dependent responses so as to decrease the intensity of CNS inflammation and damage would require preventing bacterial growth by concomitant bactericidal antibiotic therapy to be successful. On the cellular level, we found that augmentation of PCW inflammation by LBP is transduced through activation of the MAPK signaling pathway. Participation of LBP effectively doubled the inherent PCW inflammatory capability. PCW induces the phosphorylation of the MAPK erk1/2 in BMEC’s as shown here and in our previous experiments with astrocytes (Schumann et al., 1998). LBP significantly augmented the release of TNF-␣ but not NO. This difference is consistent with recent findings that inhibition of the erk1/2 pathway decreases TNF-␣, but not NO, release, and conversely, NO, but not TNF-␣, release is blocked by p38 inhibitors (Monier et al., 2002). MAPK signaling and TNF-␣ release were observed in a serum-free system in BMEC that do not express a membrane-bound CD14 receptor. In a previous study using human monocytes, pneumococcal TNF-␣ release

could not be blocked by an anti-CD14 antibody (Heumann et al., 1994). It is tempting to speculate that the observed PCW/LBP binding and presentation to TLR-2 does not absolutely require CD14. However, further experiments are needed to verify the role of CD14 in innate immune activation by PCW and LBP. PCW signaling engages the TLR-2 but not the TLR-4 receptor. Recent reports of TLR-4 involvement for Gram-positive bacteria (Dziarski and Gupta, 2000) were not supported by our experiments. The acute phase pentraxin CRP (Winkelstein and Tomasz, 1978) is an example of a circulating bacterial recognition protein participating in the innate immune response, but its cell wall binding capability is restricted to pneumococcus because PCW contains the unusual element phosphorylcholine. Decreasing or completely eliminating the choline did not affect PCW recognition by LBP. Thus, two acute phase proteins, CRP and LBP, serve to recognize PCW but by distinct mechanisms. It has been shown that PCW inflammatory activity in vivo and in vitro is retained despite degradation of the TA (Tomasz, 1981). Similarly, the enhancing effect of LBP on PCW-induced cytokine release was also inde-

Immunity 276

pendent of TA, suggesting that bioactivity and LBP recognition are both related to the PG component of PCW. Using increasingly refined PCW components, both the inherent bioactivity of PCW and LBP’s ability to augment PCW effects were clearly attributable to an intact PG backbone. Whereas inherent bioactivity was lost by disruption of crosslinking by either the glycan backbone (Majcherczyk et al., 1999) or stem peptides, LBP appeared to bind in such a way as to salvage bioactivity of the floppy PG held together only by peptide crosslinks. Complete loss of both peptide and PG backbone crosslinking was required to eliminate LBP-dependent responses. This suggests that LBP recognizes a multimeric structure and that the motif can be compiled by carbohydrates in the absence of lipid. As typical high molecular weight aggregates of LPS are selectively opsonized by a layer of LBP molecules (Tobias et al., 1995; Iovine et al., 2002), an analogous recognition of multimeric PCW by multiple LBP molecules may occur. In the three-dimensional model of human LBP, the residues in positions 94 and 95 form the central part of a conserved cluster of basic side chains in the apical region of the N-terminal domain (Beamer et al., 1997, 1998b). This cluster includes the highly conserved basic side chains in positions 42, 48, 92, and 99 that have been proposed to be mandatory for recognition of LPS by LBP and BPI proteins (Beamer et al., 1998a; Ferguson et al., 2000). Our findings of LPS competition and a decreased recognition of PCW by the mutant LBP (R94E/K95E) suggest that the PCW binding site on LPB might structurally overlap that for LPS. The phosphate residues or equivalent anionic groups in the lipid A domain of LPS are major determinants for LPS or lipid A recognition by LBP and BPI (Tobias et al., 1989; Gazzano-Santoro et al., 1995). A common feature in the PCW is a polyanionic character that is conferred by free carboxyl groups in C-terminal stem peptide residues or terminal lactosyl groups in muramic acid units exposed following amidase cleavage (Figure 7). Therefore, we propose that the anionic groups found in an iterative arrangement in the stem peptides or along the glycan backbone in polymeric PCW interact with the apical basic side chains of the N-terminal domain of LBP, creating a recognition pattern (Figure 7). A previously unrecognized pivotal role for LBP in immune response to important Gram-positive infections expands its known role beyond LPS. This has several important ramifications. The polyanionic glycan scaffold of PCW recognized by LBP is shared by all bacteria and therefore represents a possible overarching pattern suitable for broad, early immune recognition. This would place LBP at a critical nexus serving as a general soluble bacterial recognition receptor. Complete understanding of these mechanisms of immune activation may create novel opportunities for adjuvant therapy and for diagnosis in severe infections. Experimental Procedures Measurement of LBP Levels in Human CSF Diagnostic CSF samples from five patients suffering from bacterial meningitis and nine samples from controls (i.e., patients in whom CSF studies revealed no inflammatory changes whatsoever) were included in the analysis. For a diagnosis of bacterial meningitis, CSF

pleocytosis of ⬎1000 cells/␮l with ⬎90% neutrophils and a positive culture for S. pneumoniae was required. LBP content was measured by an ELISA using the monoclonal antibodies 1E8 and 2B5 as described previously (Zweigner et al., 2001). CSF cleared of cells was centrifuged (20,000 rpm), 25 ␮l of the supernatant or the pellet resuspended in 200 ␮l PBS, applied to a membrane. The blot was probed with TEPC-15 Ab (1:1000; Sigma, Deisenhofen, Germany) to detect cell wall and with anti LBP mAb (1:1000, a kind gift from XOMA, Berkeley, CA). As a positive control, PCW incubated with rLBP (XOMA) was used. Mouse Model of Meningitis Mouse experiments were conducted using a modification of a previously published model (Tang et al., 1996; Hoffmann et al., 2002). In brief, LBP⫺/⫺ mice (Dr. R. Jack, University of Greifswald, Germany) were bred at the Charite´ animal facility and were genotyped by PCR (Jack et al., 1997). Male LBP⫹/⫺ and LBP⫺/⫺ mice (20 g) were anesthetized, and 50 ␮l saline containing 1 ⫻ 107 CFU equivalents of PCW or 3 ⫻ 104 living pneumococci (D39) was injected into the spinal canal at the level of L2 or L3. CSF was collected after 6, 12, or 24 hr, and leukocytes were counted (Hoffmann et al., 2002). Some of the animals were perfused and the brains were removed and fixed for histological examination. Cell Culture Systems Primary cultures of BMEC were prepared from 3-week-old Wistar rats as previously published (Freyer et al., 1999), substituting the medium with HAM’s F10 with 20% FCS, 2 mM L-Glutamine, 100 U/ml Penicillin, 100 ␮g/ml Streptomycin, 2.5 ␮g/ml Amphotericin B, 0.5 ␮g/ml Vitamin C (Sigma), 8 ␮g/ml ECGF (Roche; Mannheim, Germany), and 80 ␮g/ml heparin. All cell culture media and supplements were purchased from Biochrom (Berlin) unless otherwise mentioned. The endothelial cells formed confluent monolayers after 6–8 days and were subsequently used for the experiments. 0.05% of the cells stained positive for GFAP and 1% for OX-42 and CD14. Approximately 99% of the cultivated cells stained positive for von Willebrandt factor VIII (Dako, Hamburg, Germany). The mouse macrophage cell line RAW 267.4 was grown in RPMI 1640 with 10% of FCS, 100 U/ml Penicillin, 100 ␮g/ml Streptomycin, and 2 mM L-Glutamine. Cells were seeded at a density of 35,000 cells/cm2 in 24-well plates and used 2 days later. Pneumococcal Cell Wall Preparations PCW were purified as described (Tuomanen et al., 1985) with minor modifications. Unencapsulated Streptococcus pneumoniae R6 was cultured overnight in chemically defined media to ⵑ5 ⫻ 108 CFU/ ml. The bacteria were heat-killed and disintegrated two times for 3 min with a Bead Beater apparatus (Roth, Karlsruhe, Germany) by using 0.1 mm beads. The suspension was digested with 10 ␮g/ml DNase (Promega, Mannheim, Germany) and 50 ␮g/ml RNase (USB Corp., Cleveland, OH) for 1 hr at 37⬚C followed by treatment with 100 ␮g/ml trypsin (plus 10 mM CaCl2; Sigma) for 2 hr at 37⬚C. The digest was sedimented by centrifugation (23,000 g, 20 min) and resuspended in 2% SDS (Serva, Heidelberg, Germany) at 90⬚C for 20 min followed by eight cycles of washing. The purified cell walls were resuspended in PBS at OD 620 nm ⫽ 1 (equivalent of 108 CFU) and were stored at ⫺20⬚C. The composition of the cell wall has been described (Garcia-Bustos and Tomasz, 1987). The ethanolamine containing cell wall (R6EA) was prepared by an identical procedure except that choline was replaced by 20 ␮g/ml ethanolamine (Sigma) in the growth medium (Tomasz et al., 1975). The cell wall derived from the licD2⫺ mutant was prepared in the same way except that the bacteria were grown in medium containing erythromycin (1 ␮g/ml; Sigma; Zhang et al., 1999). Teichoic acidfree peptidoglycan was prepared from PCW by treatment with 5% trichloroacetic for 30 min at 100⬚C (Tuomanen et al., 1985). Amidase-Digested PCW Recombinant, affinity-purified pneumococcal N-acetylmuramoylL-alanine amidase (amidase; Garcia-Bustos et al., 1987) was used to hydrolyze the bonds between the glycan chains and the stem peptides. PCW was treated with 100 units/ml of amidase for 18 hr with shaking at 37⬚C. Cell wall hydrolysates were centrifuged to remove the undigested cell wall (15,000 g, 10 min) and the superna-

LBP and Pneumococci 277

tant was then boiled for 10 min to inactivate the enzyme (GarciaBustos and Tomasz, 1987). Muramidase-Digested PCW To hydrolyze the bonds between the N-acetylmuramoyl and N-acetylglucosamine residues, N-acetylmuramoylhydrolase (muramidase, EC 3.2.1.17; Sigma) was used as described (Tomasz, 1981; Majcherczyk et al., 1999). The enzyme was added to a final concentration of 10 ␮g/ml and 1 mg/ml substrate in 0.25 mM potassium phosphate buffer (pH 6.5) containing 5 mM magnesium chloride. The mixture was incubated for 24 hr with gentle shaking at 37⬚C before boiling to inactivate the enzyme. Fully digested material becomes nonsedimentable by centrifugation (Garcia-Bustos and Tomasz, 1987). Double-digested PCW was made by applying first the muramidase followed by the amidase treatment. In all PCW preparations, the LPS concentration was below the detection limit of the Limulus amebocyte lysate test (Bio Whitaker, Walkersville, MD).

Expression and Purification of LBP Murine LBP was expressed in a baculovirus system as described (Lamping et al., 1998). For some experiments, we used commercially available mLBP (Biometec, Greifswald, Germany). The human mutant LBP R94/K95E and the control human LBP used for the experiment presented in Figure 6B were expressed in CHO cells as described (Lamping et al., 1996).

Cell Stimulation Protocol and Statistics On the day of the experiment, BMEC or Raw cell media were replaced by 400 ␮l/well of serum-free media (Invitrogen, Karlsruhe, Germany). For stimulation, 20 ␮l of PCW (108 CFU equivalents/ml) suspension plus 1 ␮g/ml LBP were added unless noted otherwise. The figures show representative experiments from three to four independent experiments each using triplicate or quadruplicate wells. Data are expressed as mean values and standard deviations. Students t tests were performed using SPSS 10 statistical software (SPSS Inc., Chicago, IL) unless otherwise stated.

MAPK p42/44 Electrophoresis and immunoblotting was done as described before (Schumann et al., 1998). For detection of phosphorylated p42/44, membranes were blocked, washed, and incubated with rabbit antip42/44 mAb (New England Biolabs, Schwalbach, Germany) and goat anti-rabbit Ab, conjugated with horseradish peroxidase (Biogenes, Berlin). Blots were detected employing the ECL-system (Amersham Pharmacia Biotech). Estimation of p42/44 activity was performed as described (Schumann et al., 1998) using the Biotrak system (Amersham Pharmacia Biotech). Cell lysates were incubated with a peptide derived from epidermal growth factor receptor highly selective for p42/44, in the presence of 50 mM [␥⫺33P]ATP (10 ␮Ci/test), and incorporated ATP was measured in a scintillation counter.

TNF-␣ and Nitrite TNF-␣ bioactivity was measured employing a modified L 929 cytotoxicity assay, and nitrite, a primary reaction product of NO in aqueous solution, was determined by the Griess reaction (Freyer et al., 1999).

Transfection and Stimulation of HEK293 Cells The tranfection procedure was described before (Opitz et al., 2001). HEK293/CD14 cells were transfected with plasmids encoding for hTLR-2 (2 ng), hTLR-4 (2 ng), hMD2 (2.5 ng), ␤-galactosidase (0.1 ␮g), and the ELAM NF-␬B luciferase reporter plasmid (0.25 ␮g; all plasmids were a kind gift of Dr. C. J. Kirschning, Munich) employing 4 ␮l/well Lipofectamine transfection reagent (Invitrogen, Karlsruhe, Germany). Stimulation was performed with LPS of S. minnesota (100 ng/ml; Sigma), Pam3Cys Lipopeptide (10 ng/ml; EMC, Tu¨bingen, Germany), or PCW (107 and 106 CFU equivalents/ml) in DMEM without FCS for 16 hr, followed by lysis and measurement of ␤-galactosidase and luciferase activity using a kit based on chemiluminescence (Roche, Mannheim, Germany).

PCW-LBP Binding Assay Nunc MaxiSorb ELISA plates were coated with different PCW preparations or heat-killed pneumococci, respectively, by incubating the plate with a coating buffer consisting of 100 mM Na2CO3, 20 mM EDTA adjusted at a pH of 9.6, and 50 ␮l/ml (2 ⫻ 107 CFU equivalents) cell wall suspension for 3 hr at room temperature followed by three times washing with H2O and drying. In order to block nonspecific binding, BSA (5 mg/ml) dissolved in incubation buffer (50 mM HEPES, 150 mM NaCl, pH ⫽ 7.4) was added overnight at 4⬚C. After removing the buffer, the coated plate was incubated with anticholine TEPC-15 Ab (Sigma) diluted 1:125 in incubation buffer containing BSA (1 mg/ml) for 2 hr in order to confirm PCW or pneumococci binding to the plates. Detection was carried out with a conjugated anti-IgA-FITC labeled secondary antibody (1:250; Sigma). In order to assess LBP binding, recombinant murine LBP (1 ␮g/ml unless otherwise noted; see above), was added to the PCW, or pneumococci-coated plates. After incubation for 2 hr at 37⬚C and repeated washing with washing buffer, a monoclonal anti-LBP antibody (1:125; big 33, Biometec, Greifswald, Germany) was added and incubated for another 2 hr, followed by detection with an anti-IgG FITC labeled secondary antibody (1:250; Sigma). After repeated washing, fluorescence was quantified at 485 and 530 nm in a fluorescence reader (Cyto Fluor, Perseptive Biosystems, Framingham, MA). Acknowledgments We dedicate this work to the memory of our longstanding coworker Dr. Dagmar Pfeil. We thank R.S. Jack, C. Schu¨tt, and M. Freudenberg for providing LBP knockout mice; C.J. Kirschning for help with the HEK 293 system; F. Creutzburg, N. Siegemund, C. Mahrhofer, S.H. Smith, and B. Mann for excellent technical assistance; and Na-Ri Oh for LBP measurements. Grant support: DFG WE-1981/1-1 (J.R.W., R.R.S.), Schr 7261/1-1 (N.W.J.S., R.R.S.), Wilhelm Sander-Stiftung (J.R.W.), BMBF 01KI0104 (R.R.S.) and NIH AI 27913 (E.I.T.), and the Cancer Center CORE grant CA-21765 (E.I.T.) as well as the Hermann and Lilly Schilling Foundation and the American Lebanese and Syrian Associated Charities (ALSAC). Received: January 29, 2003 Revised: June 2, 2003 Accepted: June 11, 2003 Published: August 19, 2003 References Beamer, L.J., Carroll, S.F., and Eisenberg, D. (1997). Crystal structure of human BPI and two bound phospholipids at 2.4 angstrom resolution. Science 276, 1861–1864. Beamer, L.J., Carroll, S.F., and Eisenberg, D. (1998a). The BPI/LBP family of proteins: a structural analysis of conserved regions. Protein Sci. 7, 906–914. Beamer, L.J., Fischer, D., and Eisenberg, D. (1998b). Detecting distant relatives of mammalian LPS-binding and lipid transport proteins. Protein Sci. 7, 1643–1646. Blairon, L., Wittebole, X., and Laterre, P.F. (2003). Lipopolysaccharide-binding protein serum levels in patients with severe sepsis due to gram-positive and fungal infections. J. Infect. Dis. 187, 287–291. Braun, J.S., Novak, R., Herzog, K.H., Bodner, S.M., Cleveland, J.L., and Tuomanen, E.I. (1999). Neuroprotection by a caspase inhibitor in acute bacterial meningitis. Nat. Med. 5, 298–302. Cauwels, A., Wan, E., Leismann, M., and Tuomanen, E. (1997). Coexistence of CD14-dependent and independent pathways for stimulation of human monocytes by gram-positive bacteria. Infect. Immun. 65, 3255–3260. Cundell, D.R., Gerard, N.P., Gerard, C., Idanpaan-Heikkila, I., and Tuomanen, E.I. (1995). Streptococcus pneumoniae anchor to activated human cells by the receptor for platelet-activating factor. Nature 377, 435–438. de Gans, J., and van de Beek, D. (2002). Dexamethasone in adults with bacterial meningitis. N. Engl. J. Med. 347, 1549–1556. Dowell, S.F., Kupronis, B.A., Zell, E.R., and Shay, D.K. (2000). Mortal-

Immunity 278

ity from pneumonia in children in the United States, 1939 through 1996. N. Engl. J. Med. 342, 1399–1407.

sults in differential localization of C3 and stimulation of phagocytosis. J. Immunol. 133, 1424–1430.

Dziarski, R., and Gupta, D. (2000). Role of MD-2 in TLR2- and TLR4mediated recognition of Gram-negative and Gram-positive bacteria and activation of chemokine genes. J. Endotoxin Res. 6, 401–405.

Iovine, N., Eastvold, J., Elsbach, P., Weiss, J.P., and Gioannini, T.L. (2002). The carboxyl-terminal domain of closely related endotoxinbinding proteins determines the target of protein-lipopolysaccharide complexes. J. Biol. Chem. 277, 7970–7978.

Dziarski, R., Tapping, R.I., and Tobias, P.S. (1998). Binding of bacterial peptidoglycan to CD14. J. Biol. Chem. 273, 8680–8690. Fan, X., Stelter, F., Menzel, R., Jack, R., Spreitzer, I., Hartung, T., and Schu¨tt, C. (1999). Structures in Bacillus subtilis are recognized by CD14 in a lipopolysaccharide binding protein-dependent reaction. Infect. Immun. 67, 2964–2968. Ferguson, A.D., Welte, W., Hofmann, E., Lindner, B., Holst, O., Coulton, J.W., and Diederichs, K. (2000). A conserved structural motif for lipopolysaccharide recognition by procaryotic and eucaryotic proteins. Struct. Fold. Des. 8, 585–592. Fierer, J., Swancutt, M.A., Heumann, D., and Golenbock, D. (2002). The role of lipopolysaccharide binding protein in resistance to Salmonella infections in mice. J. Immunol. 168, 6396–6403. Fischer, H., and Tomasz, A. (1984). Production and release of peptidoglycan and wall teichoic acid polymers in pneumococci treated with beta-lactam antibiotics. J. Bacteriol. 157, 507–513. Freyer, D., Manz, R., Ziegenhorn, A., Weih, M., Angstwurm, K., Docke, W.D., Meisel, A., Schumann, R.R., Schonfelder, G., Dirnagl, U., and Weber, J.R. (1999). Cerebral endothelial cells release TNF-␣ after stimulation with cell walls of Streptococcus pneumoniae and regulate inducible nitric oxide synthase and ICAM-1 expression via autocrine loops. J. Immunol. 163, 4308–4314. Garcia-Bustos, J.F., and Tomasz, A. (1987). Teichoic acid-containing muropeptides from Streptococcus pneumoniae as substrates for the pneumococcal autolysin. J. Bacteriol. 169, 447–453. Garcia-Bustos, J., and Tomasz, A. (1990). A biological price of antibiotic resistance: major changes in the peptidoglycan structure of penicillin-resistant pneumococci. Proc. Natl. Acad. Sci. USA 87, 5415–5419. Garcia-Bustos, J.F., Chait, B.T., and Tomasz, A. (1987). Structure of the peptide network of pneumococcal peptidoglycan. J. Biol. Chem. 262, 15400–15405. Gazzano-Santoro, H., Parent, J.B., Conlon, P.J., Kasler, H.G., Tsai, C.M., Lill-Elghanian, D.A., and Hollingsworth, R.I. (1995). Characterization of the structural elements in lipid A required for binding of a recombinant fragment of bactericidal/permeability-increasing protein rBPI23. Infect. Immun. 63, 2201–2205. Guex, N., and Peitsch, M.C. (1997). SWISS-MODEL and the SwissPdbViewer: an environment for comparative protein modeling. Electrophoresis 18, 2714–2723. Hamann, L., Schumann, R.R., Flad, H.D., Brade, L., Rietschel, E.T., and Ulmer, A.J. (2000). Binding of lipopolysaccharide (LPS) to CHO cells does not correlate with LPS-induced NF-kappaB activation. Eur. J. Immunol. 30, 211–216.

Jack, R.S., Fan, X., Bernheiden, M., Rune, G., Ehlers, M., Weber, A., Kirsch, G., Mentel, R., Furll, B., Freudenberg, M., et al. (1997). Lipopolysaccharide-binding protein is required to combat a murine gram-negative bacterial infection. Nature 389, 742–745. Janeway, C.A., Jr., and Medzhitov, R. (2002). Innate immune recognition. Annu. Rev. Immunol. 20, 197–216. Lamping, N., Hoess, A., Yu, B., Park, T.C., Kirschning, C.J., Pfeil, D., Reuter, D., Wright, S.D., Herrmann, F., and Schumann, R.R. (1996). Effects of site-directed mutagenesis of basic residues (Arg 94, Lys 95, Lys 99) of lipopolysaccharide (LPS)-binding protein on binding and transfer of LPS and subsequent immune cell activation. J. Immunol. 157, 4648–4656. Lamping, N., Dettmer, R., Schroder, N.W., Pfeil, D., Hallatschek, W., Burger, R., and Schumann, R.R. (1998). LPS-binding protein protects mice from septic shock caused by LPS or gram-negative bacteria. J. Clin. Invest. 101, 2065–2071. Liu, C., Xu, Z., Gupta, D., and Dziarski, R. (2001). Peptidoglycan recognition proteins: a novel family of four human innate immunity pattern recognition molecules. J. Biol. Chem. 276, 34686–34694. Majcherczyk, P.A., Langen, H., Heumann, D., Fountoulakis, M., Glauser, M.P., and Moreillon, P. (1999). Digestion of Streptococcus pneumoniae cell walls with its major peptidoglycan hydrolase releases branched stem peptides carrying proinflammatory activity. J. Biol. Chem. 274, 12537–12543. Michel, T., Reichhart, J.M., Hoffmann, J.A., and Royet, J. (2001). Drosophila Toll is activated by Gram-positive bacteria through a circulating peptidoglycan recognition protein. Nature 414, 756–759. Monier, R.M., Orman, K.L., Meals, E.A., and English, B.K. (2002). Differential effects of p38- and extracellular signal-regulated kinase mitogen-activated protein kinase inhibitors on inducible nitric oxide synthase and tumor necrosis factor production in murine macrophages stimulated with Streptococcus pneumoniae. J. Infect. Dis. 185, 921–926. Opal, S.M., Scannon, P.J., Vincent, J.L., White, M., Carroll, S.F., Palardy, J.E., Parejo, N.A., Pribble, J.P., and Lemke, J.H. (1999). Relationship between plasma levels of lipopolysaccharide (LPS) and LPS-binding protein in patients with severe sepsis and septic shock. J. Infect. Dis. 180, 1584–1589. Opitz, B., Schroder, N.W., Spreitzer, I., Michelsen, K.S., Kirschning, C.J., Hallatschek, W., Zahringer, U., Hartung, T., Gobel, U.B., and Schumann, R.R. (2001). Toll-like receptor-2 mediates Treponema glycolipid and lipoteichoic acid-induced NF-␬B translocation. J. Biol. Chem. 276, 22041–22047.

Heinrich, J.M., Bernheiden, M., Minigo, G., Yang, K.K., Schutt, C., Mannel, D.N., and Jack, R.S. (2001). The essential role of lipopolysaccharide-binding protein in protection of mice against a peritoneal Salmonella infection involves the rapid induction of an inflammatory response. J. Immunol. 167, 1624–1628.

Raetz, C.R., and Whitfield, C. (2002). Lipopolysaccharide endotoxins. Annu. Rev. Biochem. 71, 635–700.

Hermann, C., Spreitzer, I., Schroder, N.W., Morath, S., Lehner, M.D., Fischer, W., Schutt, C., Schumann, R.R., and Hartung, T. (2002). Cytokine induction by purified lipoteichoic acids from various bacterial species—role of LBP, sCD14, CD14 and failure to induce IL-12 and subsequent IFN-gamma release. Eur. J. Immunol. 32, 541–551.

Savedra, R., Jr., Delude, R.L., Ingalls, R.R., Fenton, M.J., and Golenbock, D.T. (1996). Mycobacterial lipoarabinomannan recognition requires a receptor that shares components of the endotoxin signaling system. J. Immunol. 157, 2549–2554.

Heumann, D., Barras, C., Severin, A., Glauser, M.P., and Tomasz, A. (1994). Gram-positive cell walls stimulate synthesis of tumor necrosis factor alpha and interleukin-6 by human monocytes. Infect. Immun. 62, 2715–2721. Hoffmann, O., Keilwerth, N., Bille, M.B., Reuter, U., Angstwurm, K., Schumann, R.R., Dirnagl, U., and Weber, J.R. (2002). Triptans reduce the inflammatory response in bacterial meningitis. J. Cereb. Blood Flow Metab. 22, 988–996. Holzer, T.J., Edwards, K.M., Gewurz, H., and Mold, C. (1984). Binding of C-reactive protein to the pneumococcal capsule or cell wall re-

Ring, A., Weiser, J.N., and Tuomanen, E.I. (1998). Pneumococcal trafficking across the blood-brain barrier. Molecular analysis of a novel bidirectional pathway. J. Clin. Invest. 102, 347–360.

Schneider, O., Michel, U., Zysk, G., Dubuis, O., and Nau, R. (1999). Clinical outcome in pneumococcal meningitis correlates with CSF lipoteichoic acid concentrations. Neurology 53, 1584–1587. Schroder, N.W., Opitz, B., Lamping, N., Michelsen, K.S., Zahringer, U., Gobel, U.B., and Schumann, R.R. (2000). Involvement of lipopolysaccharide binding protein, CD14, and Toll-like receptors in the initiation of innate immune responses by Treponema glycolipids. J. Immunol. 165, 2683–2693. Schroder, N.W., Morath, S., Alexander, C., Hamann, L., Hartung, T., Zahringer, U., Gobel, U.B., Weber, J.R., and Schumann, R.R. (2003). Lipoteichoic acid (LTA) of S. pneumoniae and S. aureus activates

LBP and Pneumococci 279

immune cells via toll-like receptor (TLR)-2, LPS binding protein (LBP) and CD14 while TLR-4 and MD-2 are not involved. J. Biol. Chem. 278, 15587–15594.

Winkelstein, J.A., and Tomasz, A. (1978). Activation of the alternative complement pathway by pneumococcal cell wall teichoic acid. J. Immunol. 120, 174–178.

Schuchat, A., Robinson, K., Wenger, J.D., Harrison, L.H., Farley, M., Reingold, A.L., Lefkowitz, L., and Perkins, B.A. (1997). Bacterial meningitis in the United States in 1995. Active Surveillance Team. N. Engl. J. Med. 337, 970–976.

Wurfel, M.M., Monks, B.G., Ingalls, R.R., Dedrick, R.L., Delude, R., Zhou, D., Lamping, N., Schumann, R.R., Thieringer, R., Fenton, M.J., et al. (1997). Targeted deletion of the lipopolysaccharide (LPS)-binding protein gene leads to profound suppression of LPS responses ex vivo, whereas in vivo responses remain intact. J. Exp. Med. 186, 2051–2056.

Schumann, R.R., Leong, S.R., Flaggs, G.W., Gray, P.W., Wright, S.D., Mathison, J.C., Tobias, P.S., and Ulevitch, R.J. (1990). Structure and function of lipopolysaccharide binding protein. Science 249, 1429–1431. Schumann, R.R., Lamping, N., and Hoess, A. (1997). Interchangeable endotoxin-binding domains in proteins with opposite lipopolysaccharide-dependent activities. J. Immunol. 159, 5599–5605. Schumann, R.R., Pfeil, D., Bu¨rger, W., Lamping, N., Kirschning, C.J., Goebel, U., and Weber, J.R. (1998). Lipopolysaccharide and pneumococcal cell wall components activate the mitogen activated protein kinases (MAPK) erk-1, erk-2, and p 38 in astrocytes. Glia 22, 295–305. Severin, A., and Tomasz, A. (1996). Naturally occurring peptidoglycan variants of Streptococcus pneumoniae. J. Bacteriol. 178, 168–174. Spellerberg, B., Rosenow, C., Sha, W., and Tuomanen, E.I. (1996). Pneumococcal cell wall activates NF-kappa B in human monocytes: aspects distinct from endotoxin. Microb. Pathog. 20, 309–317. Tang, T., Frenette, P.S., Hynes, R.O., Wagner, D.D., and Mayadas, T.N. (1996). Cytokine-induced meningitis is dramatically attenuated in mice deficient in endothelial selectins. J. Clin. Invest. 97, 2485– 2490. Tobias, P.S., Soldau, K., and Ulevitch, R.J. (1989). Identification of a lipid A binding site in the acute phase reactant lipopolysaccharide binding protein. J. Biol. Chem. 264, 10867–10871. Tobias, P.S., Soldau, K., Gegner, J.A., Mintz, D., and Ulevitch, R.J. (1995). Lipopolysaccharide binding protein-mediated complexation of lipopolysaccharide with soluble CD14. J. Biol. Chem. 270, 10482– 10488. Tomasz, A. (1981). Surface components of Streptococcus pneumoniae. Rev. Infect. Dis. 3, 190–211. Tomasz, A., Westphal, M., Briles, E.B., and Fletcher, P. (1975). On the physiological functions of teichoic acids. J. Supramol. Struct. 3, 1–16. Tuomanen, E., Liu, H., Hengstler, B., Zak, O., and Tomasz, A. (1985). The induction of meningeal inflammation by components of the pneumococcal cell wall. J. Infect. Dis. 151, 859–868. Tuomanen, E., Hengstler, B., Rich, R., Bray, M.A., Zak, O., and Tomasz, A. (1987). Nonsteroidal anti-inflammatory agents in the therapy for experimental pneumococcal meningitis. J. Infect. Dis. 155, 985–990. Tuomanen, E.I., Saukkonen, K., Sande, S., Cioffe, C., and Wright, S.D. (1989). Reduction of inflammation, tissue damage, and mortality in bacterial meningitis in rabbits treated with monoclonal antibodies against adhesion-promoting receptors of leukocytes. J. Exp. Med. 170, 959–969. Ulevitch, R.J., and Tobias, P.S. (1995). Receptor-dependent mechanisms of cell stimulation by bacterial endotoxin. Annu. Rev. Immunol. 13, 437–457. Vreugdenhil, A.C., Snoek, A.M., Greve, J.W., and Buurman, W.A. (2000). Lipopolysaccharide-binding protein is vectorially secreted and transported by cultured intestinal epithelial cells and is present in the intestinal mucus of mice. J. Immunol. 165, 4561–4566. Weber, J.R., Angstwurm, K., Bu¨rger, W., Einha¨upl, K.M., and Dirnagl, U. (1995). Anti ICAM-1 (CD54) monoclonal antibody reduces inflammatory changes in experimental bacterial meningitis. J. Neuroimmunol. 63, 63–68. Weber, J.R., Angstwurm, K., Rosenkranz, T., Lindauer, U., Freyer, D., Burger, W., Busch, C., Einhaupl, K.M., and Dirnagl, U. (1997). Heparin inhibits leukocyte rolling in pial vessels and attenuates inflammatory changes in a rat model of experimental bacterial meningitis. J. Cereb. Blood Flow Metab. 17, 1221–1229.

Yang, K.K., Dorner, B.G., Merkel, U., Ryffel, B., Schutt, C., Golenbock, D., Freeman, M.W., and Jack, R.S. (2002). Neutrophil influx in response to a peritoneal infection with Salmonella is delayed in lipopolysaccharide-binding protein or CD14-deficient mice. J. Immunol. 169, 4475–4480. Yoshimura, A., Lien, E., Ingalls, R.R., Tuomanen, E., Dziarski, R., and Golenbock, D. (1999). Cutting edge: recognition of Gram-positive bacterial cell wall components by the innate immune system occurs via Toll-like receptor 2. J. Immunol. 163, 1–5. Zhang, J.R., Idanpaan-Heikkila, I., Fischer, W., and Tuomanen, E.I. (1999). Pneumococcal licD2 gene is involved in phosphorylcholine metabolism. Mol. Microbiol. 31, 1477–1488. Zweigner, J., Gramm, H.J., Singer, O.C., Wegscheider, K., and Schumann, R.R. (2001). High concentrations of lipopolysaccharide-binding protein in serum of patients with severe sepsis or septic shock inhibit the lipopolysaccharide response in human monocytes. Blood 98, 3800–3808.