CROSSTALK BETWEEN CRP AND FICOLINS REGULATES INNATE IMMUNITY 2

Chapter Discussion CHAPTER 4: DISCUSSION Local infection is caused by a wide range of pathogenic microbial agents including pathogenic viruses (Liu et al. 2007), bacteria (Markley 1968), fungi (Marina et al. 2008), protozoa and multicellular parasites (Osnas et al. 2004), imposing great challenges to healthcare management. While this may manifest as a hot-tub rash, failure to limit such localized inflammation may cause life-threatening septic shock to ensue, leading to multiple organ failure and possibly death. These pathogens pose a great challenge to diagnosis and treatment. The corollary of severe infection-inflammation response syndrome, due to an attempt to clear pathogens during an overwhelming systemic infection is the over-activation of the immune system, which can backfire and damage the host’s own tissues. Therefore, therapeutic intervention of the cellular and molecular pathways, particularly, targeting the custodian protein:protein or protein:ligand interaction is clinically valuable, providing a powerful rationale for future drug design and development. Innate immunology has been a fast-growing field in the last two decades with the discovery of the PAMPs and PRRs. With PAMPs and PRRs defined, most of the innate immune molecules that have been isolated can now fit into the functional groups as PRRs. Extensive research efforts have been invested into characterizing the roles of these PRRs with regards to the microbial group that they each PRR targets, and the downstream effectors that they recruit (Holmskov et al. 2003; Leclerc et al. 2004; Takeda et al. 2005) . However, most of the studies have been done with an isolated focus on the role of 176 individual PRRs. Therefore, the potentials that could come from the ensemble of PRRs through PRR:PRR collaborations have been neglected. With the individual PRRs characterised, it is therefore timely to study the interactive roles of these PRRs acting in concert. CRP and ficolins represent two main families of such PRRs and they each trigger and regulate seminal antimicrobial pathways in infection and injury. It is conceivable that insights gained from this PRR:PRR interaction mechanism will offer fresh perspectives on the potential development of complement immune therapies to enhance the eradication of the invading pathogen. 4.1 The functional divergence of L- and M-ficolins in evolution. Ficolins constitute a family of secretory C-type lectins, which play critical roles as pattern-recognition receptors in host defense. Patients with ficolin disorder are susceptible to inflammation brought about by respiratory infections (Atkinson et al. 2004), Behçet's disease (Chen et al. 2006), rheumatoid arthritis (Vander Cruyssen et al. 2007) and necrotizing enterocolitis (Schlapbach et al. 2009). Being structurally and functionally conserved (Zelensky et al. 2005), all three ficolins were shown to bind to GlcNAc which is the sugar moieties of PAMPs displayed on microbes (van Kooyk 2008). This liaison between ficolin:GlcNAc activates complements (Frederiksen et al. 2005; Zhang et al. 2009) and promotes phagocytosis (Jack et al. 2005). Although belonging to the same family, L- and M-ficolins have high homology with each other (76.99%) while H-ficolin has low homology with L- and M-ficolins 177 (42%), which determines the different binding behaviors among three ficolins (Figure 4.1). Among three ficolins, L- and H- ficolins are secreted into the serum by lung and liver and they circulate in the blood (Matsushita et al. 1996; Andersen et al. 2009). In contrast, M-ficolin is the only ficolin isoform secreted by immune cells such as monocytes (Teh et al. 2000), neutrophils (Rorvig et al. 2009) and alveolar epithelial cells (Liu et al. 2005) suggesting the crucial and fundamental roles of M-ficolin in immune responsive cells. Whereas the mouse homologue of L- ficolin was identified to be ficolin A (Endo et al. 2004) and mouse homologue of M-ficolin gene is ficolin B (Endo et al. 2004), the mouse H-ficolin gene was identified to be a pseudogene and orthologous to mouse ficolins A/B and human L-/M- ficolins. This might explain the lack of bindings between CRP and H-ficolin. The dramatic difference between H-ficolin and L-, M-ficolin proteins and the completely different in vivo locations of L- and M-ficolins implies that they have different functions and also explains their non-redundancy and the necessity to conserve all three ficolin isoforms through evolution. Figure 4.1: The alignment of three ficolin isoforms using Bioedit Sequence Alignment Editor. L- and M- ficolin sequences show high homology with each other while L- and Hficolin sequences show low homology to H-ficolin. 178 4.2 CRP:L-ficolin interaction connects the classical and lectin complement pathways and boost the antimicrobial activity Local acidosis, mild hypocalcaemia and high CRP level are characteristic of local inflammation. By defining a typical infection-inflammation condition (pH 6.5 and mM calcium) and normal physiological condition (pH 7.4 and 2.5 mM calcium), we demonstrated crosstalk between CRP and L-ficolin, which resulted in two new autonomous amplification pathways leading to a synergistic level of C3-deposition on the bacteria. This serves to boost the bactericidal activity in the serum under local infection-inflammation condition. L-ficolin was identified to interact with CRP via its fibrinogen-like domain. Here, we showed that the recombinant rL-FBG interacts avidly with CRP at KD of 1.26 x 10-8 M under infection-inflammation condition compared to 1.11 x 10-6 M under normal condition. This represents a 100-fold increase in affinity between the two proteins. A positional preference was evident, where prior anchorage of CRP resulted in a more efficient recruitment of the rL-FBG. Low pH dramatically enhanced CRP:L-ficolin interaction, showing that inflammation-induced acidosis promoted stronger crosstalk between CRP and L-ficolin. Furthermore, lower level of calcium enhanced the CRP:L-ficolin interaction, thus indicating that calcium is a main regulator of the molecular crosstalk between L-ficolin and CRP. Co-IP experiments confirmed that the CRP:L-ficolin complex formation was induced in the serum by infection-inflammation condition (low pH, low calcium level and high CRP level) and ELISA demonstrated that similar to CRP and rL-FBG interaction, full length L- and M- ficolins displayed stronger 179 binding to CRP when the latter was anchored first to bacteria. Taken together, our findings explain why and how a typical infection-inflammation condition provokes crosstalk between CRP and the L-ficolin, which might boost the complement system. This key mechanism is further confirmed by the fact that in infection-inflammation condition, the CRP:L-ficolin amplification pathway caused more C3-deposition. Hitherto, two independent complement pathways involving CRP-activated classical pathway and L-ficolin-mediated lectin pathway are well-established (Kaplan et al. 1974; Fujita et al. 2004). Here, we provide evidence for the inflammation-induced crosstalk between CRP and L-ficolin which impels two new amplification pathways: (1) Amplification pathway PC-beadsÆCRPÆL-ficolinÆMASPsÆC4ÆC3ÆMAC (2) Amplification pathway GlcNAc-beadsÆL-ficolinÆCRPÆC1q,r,sÆC4ÆC3ÆMAC The cleavage of C4 demonstrated that these two amplification pathways were functional only under infection-inflammation condition but are negligible in normal physiological condition. The opsonized particles generated by these amplification pathways were recognized and removed by the immune cells indicating the autonomy and effectiveness of the two new pathways which were further confirmed by quantification of the phagocytic efficiency. Competition assays showed that the amplification pathways not compete against the classical and lectin pathways, but rather, they boost these two well-known complement pathways. The complement system becomes a major microbicidal force in the infection-inflammation condition mediated by the synergistic effect of CRP and ficolins. We showed that CRP interacts with L-ficolin under 180 infection-inflammation condition, enhancing the complement-activated killing of the P. aeruginosa. Although it was shown that purified CRP and L-ficolin did not bind to P. aeruginosa, the presence of serum factors enabled CRP to bind to bacteria indirectly and the bound CRP recruited L-ficolin (Figure 4.2). Figure 4.2: A model to illustrate the mechanism of enhanced antimicrobial activity of the serum in infection and localized inflammation. CRP recruited to the bacterial surface by serum factors activates the classical complement pathway. The infection-inflammation condition triggers the CRP:ficolin interaction, resulting in the immobilization of ficolins to the bacterial surface and the activation of the lectin pathway to amplify the complement response. By crosstalk between CRP:ficolin, two autonomous amplification pathways emerge to boost antimicrobial activity of the classical (black upward arrow) and lectin-mediated (bold black upward arrow) complement pathways: (1) Amplification pathway 1: PC-beadsÆCRPÆL-ficolinÆMASPsÆC4ÆC3ÆMAC (2) Amplification pathway 2: GlcNAc-beadsÆL-ficolinÆCRPÆC1qÆC1r,sÆC4ÆC3ÆMAC. Thus far, we have provided evidence that local infection-inflammation elicits new complement amplification mechanisms to boost the immune system and fight the 181 invading pathogen. This novel antimicrobial mechanism is particularly effective against the P. aeruginosa, an opportunistic pathogen which causes mortality in critically ill and immuno-compromised patients (Goldman et al. 2008). Both the immunoevasive nature (Kharazmi 1991) of P. aeruginosa as well as its acquisition of multi-drug resistance (Zaborina et al. 2008) makes elimination of this microorganism a particular challenge. To date, no effective therapy against P. aeruginosa infection has been found. However, insights gained into the mechanisms of complement amplification shown in the present work are crucial in understanding the host defense to counter the immune evasion by the pathogen, which will contribute to the development of complement-immune therapies. 4.3 M-ficolin connects the extracellular surveillance to the intracellular signal transduction Transmembrane C-type lectins are well-known to transduce signals into the cells independently of, or collaboratively with TLRs. However, the functions and underlying mechanisms of action of the soluble C-type lectins in signal transduction have remained unclear, although other C-type lectins like galectin (Nieminen et al. 2005) and MBL (Ip et al. 2008) were reported to sense danger signals and boost the proinflammatory immune response. We demonstrated that M-ficolin, a representative soluble CL, mediates NF-κB activation and IL-8 secretion upon challenge by pathogens/PAMPs/GlcNAc, which was confirmed by abrogation of the effect in M-ficolin- cells, as well as cells pretreated with M-ficolin antibody. We showed that M-ficolin performs this key role by being localized on the monocyte surface through 182 interaction with its cognate transmembrane receptor partner, GPCR43. Through this liaison, the M-ficolin:GPCR43 complex probably alerts the host cell of microbial invaders, by transducing an infection signal via the activation of NF-κB (Figure 4.3). Although M-ficolin expression in macrophages was significantly lower, it can be upregulated in monocyte/macrophage lineage upon LPS stimulation (Frankenberger et al. 2008), indicating its critical function in immune cells during infection. Interestingly, such interplay of proteins leading to signal transduction was also observed in other membrane-associated molecules such as angiopoietins which coincidentally also contain an FBG domain (personal communications with Dr N.S. Tan, Nanyang Technological University, Singapore), implying the fundamental significance of such a phenomenon. The conformational change of C-type lectins upon pH or ligand-/ proteinbinding, is universal (Zelensky et al. 2005; Menon et al. 2009), although the biological consequence remains unclear. We found that under infection-mediated local acidosis condition, the binding of CRP to M-ficolin:GPCR43 complex was enhanced whereas under normal condition, they associate weakly. As yeast two hybrid has shown M-ficolin to interact with itself to form a homopolymer, it is also possible that CRP and GPCR43 may bind to the same M-ficolin molecule or to two different M-ficolin molecules in the polymer/dimer. The detailed binding pattern between CRP, M-ficolin and GPCR43 needs to be further studied in future. CRP plays a dual reciprocal role in M-ficolin mediated IL-8 secretion: under normal condition, CRP promoted M-ficolin mediated IL-8 secretion whereas under local acidosis, CRP inhibited M-ficolin mediated IL-8 183 secretion. These results clearly support a refined regulatory mechanism where environmental perturbation at the later phase of infection induces the conformational change of M-ficolin and CRP, which regulates its pathogen recognition and IL-8 production. Overall, we have shown how the host exercises plasticity in its immune function-cum-regulation via monocyte-secreted M-ficolin, which exploits its extracellular interacting partners to form a trio-complex to act bidirectionally: (a) to transduce infection signal via GPCR43 into the host cell and (b) to regulate the immune response to restore homeostasis through the reciprocal role of CRP which modulates the M-ficolin to downregulate IL-8 secretion. We envisage that the host can expand its repertoire of immune function-cum-regulation mechanisms by promiscuous protein networking amongst members of the families of soluble C-type lectins, pentraxins and transmembrane proteins, where cues from the infection-mediated environmental perturbation induce the protein complex to immunomodulate and attain homeostasis. Furthermore, our elucidation of the binding interface and the infection-inflammation sensitive regulatory region of the M-ficolin provides insights into the bioactive centre of the M-ficolin molecule. Such insights will be useful for future drug development of immunomodulators. 184 binding nature of the evolutionarily conserved FBG domain, which recognizes a wide spectrum of bacteria [22,23]. Ficolin homologues were found in different species ranging from human, mouse, pig, xenopus to ascidian [21,24,25], each with a common binding specificity for N-acetyl-Dglucosamine (GlcNAc) moiety of various Pathogen Associated Molecular Patterns (PAMPs) displayed on the surface of the invading microbes. It has been shown that the ficolins and TL5A, its homologue in the invertebrate horseshoe crab, Tachypleus tridentatus, which are evolutionarily distant from the human, share a conserved three-dimensional structure in their Ca2+-binding site, and the acetyl group ligand-binding pocket [19]. Although ficolins were demonstrated to have stronger association with its partners [24,26,27] in both the horseshoe crab and human during infection, the underlying mechanism of this functional enhancement remains unknown. It is conceivable that studies of the crosstalk between the FBG of ficolin and its partners under pathological conditions would provide insights into its structure-function conservation in other FREPs, particularly in host-pathogen interaction. C-reactive protein (CRP), an evolutionarily conserved plasma protein, is a 120 kDa major acute phase pentraxin. CRP is implied in diverse pathologic states such as bacterial infection [28], atherosclerosis [29], autoimmune disease [30] and cancer [31]. It is composed of five identical subunits with Ca2+-binding and phosphocholine-binding sites located on the same face of each subunit [32]. CRP was shown to recognize a broad range of ligands including phosphocholine, chromatin, and bacterial antigens [32,33]. As an inflammation marker, CRP enhances phagocytosis [34] and activates complement [35,36]. Our previous studies have shown that under local infection-inflammation condition, CRP interacts with the FBG domain of L-ficolin to boost the immune response through complement crosstalk [21]. Due to the structural and functional similarities between different ficolin isoforms, we reasoned that CRP might also interact with other ficolin isoforms. However, the underlying mechanism and the biological significance are yet to be identified. The solution of the CRP crystal structure with/without calcium has provided a new basis for studying the molecular mechanisms of action of CRP under certain pathophysiological conditions [37,38]. By using the SMART program, we found 1475 potential FREP proteins in different species that possibly contain an FBG domain at the C-terminus. We found that the human FREPs, except for the intelectin family, share high sequence homology with each other, particularly near the calcium-binding region, thus implying a potential conserved responsive element(s) under pathophysiological condition. Using HDMS, we demonstrated the flexible nature of the FBG domain, which underwent dramatic conformational change under pathophysiological condition. The flexible FBG gained a stronger association with the conformationally less mutable CRP, which our group previously found as a dominant interacting partner of both L- and M- ficolins [21]. This might be explained by our further experimental evidence showing that CRP interacted with the C-terminal region of FBG, which was more exposed under pathophysiological condition. By site- directed mutagenesis, one of the critical CRP binding sites was pinpointed to be H284 on the FBG under pathophysiological condition. This site overlaps with the conserved GlcNAc binding pocket [39]. We envisage that high affinity interaction between M-ficolin and CRP under the pathophysiological condition might divert the M-ficolin from GlcNAc/pathogen recognition in order to regain homeostatic control. The findings from our study might provide a fundamental understanding on the mechanism of interaction between FREPs and their partners, occurring specifically under pathophysiological condition. Insights from this study may be applied for future design of immunomodulators. RESULTS Human FREPs represent a big family of proteins with structural similarities. An analysis of the fibrinogen-like domains using SMART, we identified 1475 potential FREP proteins, with 37 members in human and 94 members in Rodentia (Figure 1A). By studying the primary sequence and architecture of human FREPs, we found that all of these proteins contain an FBG domain although some of the FBGs are truncated. The human FREPs with intact FBG domains were selected and a phylogenetic map was constructed to indicate their evolutionary distances (Figure 1B). Here, we observed that the human FREPs are composed of several major families, such as intelectins, ficolins, angiopoietins and angiopoietin-related proteins, each of which contains several isoforms. Except for inteletins, all other FREPs are evolutionarily related to one same group, thus implying the functional divergence of intelectins. Furthermore, multiple sequence alignment of the FBG domains of FREPs revealed a conserved sequence profile throughout the FBG domain, implying their functional conservation (Supplementary Figure 1). Interestingly, except for intelectins, all the other human FREPs demonstrated a conserved calcium-binding region, including the ficolins, which represent one of the major families of FREPs. This implies a potential critical function of FREPs under pathophysiological condition where the level of blood calcium is sensitively perturbed. Pathophysiological condition induces structural flexibility of the C-terminus of FBG-containing proteins M-ficolin has been suggested to be responsively released during local infection[40]. Here, we studied the FBG domain of M-ficolin as a representative FREP protein, in order to understand how FREPs behave under pathophysiological condition. Therefore, the FBG domain of M-ficolin was selected for molecular expression and further structural and functional analyses under pathophysiological conditions (herein defined as local acidosis of pH 6.5 and hypocalcaemia of 2.0 mM calcium [21]). As FBG contains the conserved calcium-sensitive region, we hypothesized that a conformational change might occur under pathophysiological conditions where perturbation of the pH and calcium prevails. This might possibly result in distinct behaviours of FREPs. Therefore, to prove this, using hydrogen-deuterium exchange mass spectrometry (HDMS), the conformational changes of The restoration of calcium to the corresponding physiological concentration (2.5 mM) or pathophysiological concentration (2 mM) posed a pHdependent differential effect on the conformation of FBG. It is noteworthy that at pH 6.5, a slight hypocalcaemia of mM calcium only slightly decreased the deuteriumexchange of one region (292-305). However, under pH 7.4, the physiological level of calcium induced a global compaction of the M-ficolin as demonstrated by the decrease in the deuterium-exchange of regions in the FBG domain (175-183, 209-216, 224-236, 284-319, 321326) (Figure 2B, Supplementary Table 1). Under pH 7.4, calcium concentrations of 0, and 2.5 mM induced a dose-dependent effect on the decrease of deuterium incorporation (Supplementary Figure 2), with minimal effect under mild acidosis. This implies that under normal physiological condition, calcium might serve as a controller to switch on/off the FBG molecule, which might affect its interaction with other ligands. However, under pathophysiological condition, this regulatory effect seems to be minimal. Since the P domain of the M-ficolin harbors many ligand-binding sites including GlcNAc, GalNAc and Neu5Ac as well as calcium [39], we aligned the pHand calcium- sensitive regions to the M-ficolin sequence. We found that the pH-sensitive regions (pH-regulatory region and C-terminal region) are mainly located in the P domain whereas the calcium-sensitive region covers both the P and B domains (Figure 2C). The A domain is relatively stable. We also observed that both the pH- and calcium- sensitive regions overlap at 284-326, near the calcium binding loop and the corresponding disulfide bond between Cys270-Cys283 [39,41] (Figure 2C). This is the region which many other ligands are known to bind to [39]. This implies that pH and calcium possibly affect the interaction between M-ficolin and a wide spectrum of ligands. Overall, we have shown that the FBG domain of M-ficolin is very flexible, whereupon pH and calcium perturbations might undergo conformational change. Figure 1. The FREP family encompasses a large number of proteins distributed in different species. (A) Taxonomic distribution of proteins containing FBG domain. The number of FREP hits in different species was calculated from SMART NRDB, a non-redundant database. The human FREPs are highlighted in red. (B) Phylogenetic tree of human FREPs with intact FBG domain(s). The selected proteins were labeled using UniProtKB/Swiss-Prot accession number. The number at the root of the phylogenetic tree indicates a percentage of boostrap. The selected human FREP proteins were divided into several clusters including intelectins, angiopoietins, angiopoietinrelated proteins, tenascins, and ficolins. pH- and calcium- dependent interactions between CRP and M-ficolin FBG domain Recently, we found that CRP associates with L-ficolin to boost the immune response [21] in a pH- and calciumdependent manner. Sharing about 80% amino acid sequence homology with L-ficolin, the M-ficolin was also found to interact with CRP [21,42]. Therefore, it was imperative to understand whether this possible conformational change induced by pH and calcium affects the protein-protein crosstalk between M-ficolin and CRP. Towards this goal, the M-ficolin interaction with CRP was characterized biochemically. As the FBG domain of Mficolin harbors many ligand- and protein- binding sites, we hypothesized that CRP binds to the FBG domain. Therefore, using recombinant FBG in ELISA, we showed that both the CRP and FBG displayed dose-dependent binding to each other when either of them was immobilized on MaxisorpTM plates. However, stronger binding was observed when CRP was immobilized first (Figure 3A,B). This implies that anchored CRP might be the preferred position for CRP:Mficolin interaction. To test whether the pathophysiological FBG under five different pH and calcium conditions were analyzed by HDMS and plotted together: (i) pH 7.4 without calcium, (ii) pH 7.4 with mM calcium, (iii) pH 7.4 with 2.5 mM calcium, (iv) pH 6.5 without calcium and (v) pH 6.5 with mM calcium. A set of 11 peptides covering 45.75% of the primary sequence of FBG have been selected based on their peak quality and intensity (Supplementary Table 1). The peptides located in 284326 in the C-terminal region of FBG showed significant increase in deuterium-exchange under pH 6.5 compared to pH 7.4 (Figure 2A, Supplementary Figure 2), suggesting that this region might be more exposed to the solvent under local mild acidosis. We also observed that peptide 249-255 distinctively showed a significant decrease in deuterium-exchange under pH 6.5, implying that it might become encrypted into the structure under mild acidosis. These corresponding pH-sensitive regions are highlighted in Figure 2A. At pH ≤ 6.5, calcium did not influence FBG:CRP interaction significantly (Figure 3D). However, at physiological pH, increasing calcium concentration dramatically inhibited the protein-protein interaction when CRP was immobilized first (Figure 3D), which was shown to be the preferred position for the interaction of CRP:ficolin. This is consistent with our HDMS results which showed that calcium influences the conformations of M-ficolin FBG with more dramatic effect under pH 7.4 implying that the blood calcium concentration probably fine tunes the conformational change of FBG and thus regulates the recruitment of ficolin to the CRP. To further substantiate the effect of the pathophysiological conditions, we used surface plasmon resonance (SPR) analysis, which demonstrated a binding affinity of 10-8 M under normal physiological condition (Figure 3E), compared to a 100-fold stronger affinity (KD 10-10 M) under local acidosis and hypocalceamia (Figure 3F). pH and calcium affects CRP:FBG interaction, we characterized their binding at various pH and calcium concentrations. Figure 3C shows that the interaction between CRP and FBG was profoundly influenced by pH shift regardless of the order of immobilization of CRP or FBG. Interestingly, the strongest binding between CRP and FBG occurred at pH 6.5, above or below which the interaction diminished implying that CRP:M-ficolin interaction is triggered by mild acidosis whereas under physiological pH or extreme acidosis, the CRP and Mficolin may co-exist but only associate weakly. Similarly, we investigated how calcium affects their interaction. Figure 3. CRP interacts with the FBG domain of M-ficolin in a pHand calcium- dependent manner. ELISA test for the direct interaction between CRP and FBG when (A) 0.8 g CRP or (B) 0.8 g FBG was immobilized on the maxisorp plates. (C) pH effect (5-7.4) on the interaction between CRP and FBG when CRP or FBG was immobilized first. (D) Relative binding of FBG to immobilized CRP at 0-10 mM calcium and at different pHs. For (A)-(D), µg pure binding protein immobilized directly to the plate and detected by the respective antibody, was used as the positive control for 100% binding. Readings from wells without any immobilization served as the negative controls. Readings of OD405 were subtracted off negative controls and expressed as a percentage of the corresponding 100% binding control. The binding affinity between CRP and FBG under (E) pH 7.4, 2.5 mM Ca2+ and (F) pH 6.5, mM Ca2+ was identified by real-time biointeraction using SPR analysis. Figure 2. The conformational change of FBG under different pH and calcium conditions. (A) The pH-sensitive regions are shown in brown and red in the crystal structure of FBG (PDB: 2JHM). Brown color area indicates the region showing decreased deuterium exchange under pH 6.5 and the red color area indicates the region showing increased deuterium exchange under pH 7.4. (B) The calcium-sensitive regions are in purple in the crystal structure of FBG (PDB:2JHM) under pH 7.4 ( upper panel) and pH 6.5 (lower panel). Calcium shows consistent effect in causing the decreased deuterium exchange under both pH7.4 and pH 6.5. However, more calciumsensitive regions become available at pH 7.4. (C) pH-sensitive and calcium-sensitive regions are annotated on the primary sequence of FBG (M-ficolin115 – M-ficolin326). The amino acids of domain A (purple letters), domain B (grey letters) and domain P (blue letters) are indicated. The black boxes are the pH-sensitive regions. The blue underline shows the calciumsensitive regions under pH 7.4 and the red underline shows the calcium-sensitive regions under pH 6.5. The green boxes indicate the conserved calcium binding sites. The disulfide bond (Cys270 – Cys283) that stabilizes the calcium binding site is indicated in orange. The grey shaded region is the secretory peptide. The FBG domain starts from Pro115 and ends at Ala326. Identification of the binding interfaces between Mficolin:CRP As CRP:FBG interaction was enhanced under low pH and low calcium condition, it was pertinent to delineate the underlying mechanism. Towards this, we first identified the CRP:FBG binding interfaces by using HDMS to investigate how the contact points were dynamically affected by pH and calcium under the pathophysiological condition. We found that under pathophysiological condition, peptides in the region of 205-220 and 284-326 showed dramatic decrease in deuterium exchange upon the binding of CRP indicating that CRP might have bound to these regions and thus blocked the deuterium exchange (Figure 4A, Supplemenatry Table 1, Supplementary 3A). Conversely, under physiological condition of pH 7.4 and 2.5 mM calcium, we only observed the decrease of deuterium incorporation in regions: 209-220 and 306-326, suggesting that the regions 205-209 and 284-305 possibly moved away from the binding interface under physiological condition. The exposure of more pathophysiologicallyinduced binding sites, which are otherwise located within the FBG domain (Figure 4A) probably explains the enhanced binding affinity under pH 6.5 and mM calcium, consistent with the SPR result (Figure 3E, F). Notably, the majority of the binding interface under pathophysiological condition is located in the C-terminal region of FBG, coincident with the identified pH- and calcium- sensitive regions. This implies that the loss of CRP binding to 284305 is possibly attributable to the conformational change of FBG, which occurs under pathophysiological condition. The FBG contacts two regions of CRP (53-64, 110-132) under pathophysiological condition. Notably, these CRP regions remain as the FBG-binding sites regardless of the physiological and pathophysiological conditions (Figure 4B, Supplementary Table 2). Although it is conceivable that the weak interaction between CRP:FBG under physiological condition will result in less binding sites on CRP, the invariant binding interface of FBG on CRP might be due to subtle alterations in the exact binding sites within the wider region as evidenced by the difference in the amount of deuterium incorporation between the physiological and pathophysiological conditions (Supplementary Figure 3). The relatively more stable structure of CRP compared to the FBG structure, implies that being the more structurally versatile counterpart, the FBG probably plays a dominant role in guiding the CRP:M-ficolin interaction. To further verify the binding interface identified by HDMS, the peptides spanning the identified binding regions of CRP on the FBG were synthesized. The binding of these peptides to CRP under normal and pathophysiological conditions was analyzed by SPR to rule out any possible secondary structure effect (Figure 4C). It was interesting to note that peptide 205-220, under both conditions, did not show any significant binding to CRP. This indicates that the decrease in the deuterium exchange in region 205-220 might be due to the conformational change of the protein upon CRP binding, which encrypts this region into the M-ficolin structure. Thus we excluded this region from the binding interface. Conversely, although region 284-305 did not demonstrate any significant decrease in deuterium exchange upon CRP binding under physiological condition, the synthesized peptide 284-305 was shown to bind to CRP by SPR analysis. This implies that under normal condition, the secondary structure of FBG possibly embeds the region 284-305 thus preventing its binding to CRP whereas under pathophysiological condition, this region might be exposed, resulting in the association to CRP. This is consistent with our finding that region 284-305 is located in the pH- and calcium- sensitive region, which is more exposed under low pH and low calcium condition (Figure 2A,B). Peptide 306-326, under both conditions, showed significant binding to CRP with higher binding affinity under pathophysiological conditions. This is consistent with the HDMS result wherein under both conditions, region 306326 demonstrated significant decrease in deuterium exchange. Therefore, this region of FBG might represent the constant binding region of CRP. Notably, the SPR results indicated that peptides 284-304 and 306-326 consistently showed higher binding affinity under pathophysiological condition in comparison to the physiological condition. This might possibly be due to the alteration in protein surface electrostatic charge, which occurred under different conditions. This probably also contributes to the enhanced CRP:FBG interaction under pathophysiological condition. Figure 4. Characterization of the binding interfaces between CRP and FBG under physiological and pathophysiological conditions. (A) The CRP binding interface on FBG under physiological and pathophysiological conditions. The green area annotates the identified CRP-binding interface under physiological and pathophysiological condition. The red area annotates the additional CRP binding interface which is exposed under physiological condition. (B) The FBG-binding interface on CRP under physiological or pathophysiological are colored. The binding interface of FBG on the CRP molecule remained unchanged (dark blue and light blue) under both conditions. The dark blue part annotates the binding interface with no difference in the decreased deuterium incorporation upon CRP binding under both physiological or pathophysiological condition. Light blue color area represents the binding interface where the decrease of deuterium incorporation upon CRP binding was reduced under physiological condition compared to pathophysiological condition. (C) Binding of synthesized FBG peptides corresponding to 205-220, 284-305, 306326 amino acids to the CRP molecule was tested under physiological and pathophysiological conditions using SPR. The concentrations (in M) of the synthesized peptides that were injected over the CM5 chip His284 is a critical CRP-binding site on FBG immobilized with CRP are indicated on the graphs. For all the six the curves C-terminal region of FBG alsoof harbors panels, theAs earlier represent the binding curve CRP to PC many such ascurve GlcNAc and theother later ligand-binding curves represent sites the binding of the binding synthesized pocket the calculated region 282-285 it was imperative peptideslocated to CRP.atThe binding [39], affinity of the FBG peptides to CRP is labeled the exact graph. CRP “ns” indicates significant binding for us to identifyonthe bindingnosite and verify accordingCRP to thecompetes valid range of association/dissociation ratewhich constant whether with GlcNAc for the FBG, (BIAcore 2000the manual). would affect pathogen-recognition property of FBG. To locate Overall, by HDMS and SPR, we observed that the CRP binding interface on FBG was located in the Cterminal region of FBG at 284-326 under pathophysiological condition whereas under physiological condition, region 284-305 apparently moved away from the binding interface resulting in a weaker interaction between CRP and FBG. To locate the potential critical sites, we performed a sequence alignment of H-, L- and M- ficolins in the region 282-326 (Figure 5A). It was documented that both the L- and M- ficolins but not the H-ficolin interact with CRP [21]. Thus, amino acids that are conserved in Land M-ficolins but not in H-ficolin likely connote functional significance. Therefore, two such residues (H284 and L293) closest to the GlcNAc binding pocket (282-285) on the CRP binding interface were selected for site-directed mutagenesis. H284 and L293 were mutated to Ala to avoid the drastic structural alteration. The wildtype protein and the mutant proteins (H284A and L293A) were expressed and purified (Figure 5B, upper panel). ELISA was used to test their binding to CRP under different pH and calcium conditions (Figure 5B, lower panel). ELISA showed that at pH 6.5 but not at pH 7.4, the H284A mutation caused a dramatic loss-of-binding to CRP compared to the wildtype proteins implying that H284 might be the probable CRP binding site on FBG. At mM calcium under pH 6.5, the binding of CRP to H284 was apparently retained, indicating that His284 is the critical CRP binding site under pathophysiological condition. This also implies that pH is probably the main determinant of the conformation of FBG under pathophysiological condition, consistent with the HDMS result (Figure 2B). To investigate the significance of the H284 residue, we searched the homologues of M-ficolin in NCBI protein database, and aligned the primary sequence of these homologues to the human M-ficolin. Notably, the M-ficolin homologues in 12 different species showed conserved H284. We also found that the pH- and calcium- sensitive CRP-binding interface at 284-305 was highly conserved in these different species. Importantly, the calcium binding site and the His 284 equivalent in many of the FREPs are also conserved. As CRP is also a highly conserved molecule, this implies that the pH- and calcium- sensitive CRP:FBG interaction might be evolutionarily entrenched, indicating the fundamental significance of this protein:protein interaction during infection-/injurymediated pathophysiological perturbations. Taken together, the above findings suggest that His284 is the critical CRP binding site on the FBG domain under pathophysiological condition. As many other ligands of M-ficolin such as GlcNAc, GalNAc or sialic acid also dock onto the same region of M-ficolin [39], it is conceivable that binding of CRP to the conformationally competent M-ficolin might promote/inhibit the PAMP-sugar/pathogen recognition by M-ficolin, which may interfere with pathogen-recognition of M-ficolin. of CRP, an acute phase protein, is dramatically increased Figure 5. His284 on the FBG is a critical residue that binds CRP. (A) The sequences of H-, L- and M- ficolins spanning region (282-326) were aligned to each other and the amino acids closest to the GlcNAc binding pocket (282-285) are boxed in red. (B) ELISA to test the binding affinity between the FBG mutants and CRP at different pH and calcium condition (lower panel). Wild-type FBG, L293A and H284A mutants were purified (upper panel). E1-E6 indicate six sequential eluted fractions from the chromatographic column. One g binding protein directly immobilized onto replicate wells was the positive control. The relative binding was calculated as a percentage of the positive control. Data are representative of three independent experiments presented as means + S.D. (C) Sequence alignment of M-ficolin homologues (from 284-305), with conserved amino acids shaded grey. and the microenvironmental perturbation such as acidosis and hypocalcaemia prevails soon after the onset of infection when excessive leukocyte recruitment [5,44] or overwhelming immune activation [21] occurs, thus creating a need to restore homeostasis. Ficolin represents an important group of FREPs, the homologues of which stretch from horseshoe crab to human over several hundred million years of evolutionary distance [24]. Being a critical group of PRRs, the FBG of ficolin was found to recognize a wide range of microbes via GlcNAc on the microbial PAMPs, playing critical roles in the frontline host defense. It was documented that at physiological pH, FBG is an active conformer for GlcNAc/PAMP whereas under acidic condition, FBG resumes an inactive conformer and becomes less receptive to GlcNAc/PAMP [39]. Here, we showed that under pathophysiological condition, CRP binds to M-ficolin at high affinity via H284, which overlaps with the GlcNAc binding pocket whereas under physiological condition they are separate. Thus, we Biological implications of the pH- and calcium- sensitive reasoned that under pathophysiological condition but not physiological condition, CRP might compete with GlcNAc interaction between CRP:FBG During local infections brought about by trauma-induced in binding to FBG. To experimentally prove this, FBG and CRP and infection [1] and intra-abdominal infection [43], the level the crosslinked CRP:FBG complex were individually injected over GlcNAc immobilized on the chip under physiological or pathophysiological condition. We observed that the CRP:FBG complex showed much lower affinity to GlcNAc compared to FBG under pathophysiological condition. Under physiological condition, they showed similar binding affinity (Figure 6A). As a control, CRP on its own did not show significant binding to GlcNAc (Figure 6A). This implies that the association between CRP:FBG under pathophysiolgical condition but not physiological condition, will inhibit the binding of FBG to GlcNAc. When this occurs, the Mficolin is structurally “switched off” (becomes an inactive conformer) and the subsequent pathogen recognition activity of FBG is inhibited (Figure 6B). Therefore, under the defined pathophysiological condition, FBG is an active form competent for binding CRP but a non-active form for recognising PAMP/sugar. Infection-mediated environmental perturbation (particularly, acidosis) induces a significant conformational change in the M-ficolin FBG domain, docking CRP at high affinity to the GlcNAc binding site and making a molecular switch from pathogenrecognition to CRP-binding. This regulatory mechanism, which kicks in soon after the start of infection, might be critical for the homeostasis of immune response and prevents immune over-activation. DISCUSSION Immune responses are elicited against a wide range of pathogenic microbes, injury-induced inflammation and trauma and malignancy. In an attempt to eradicate the invading pathogens and/or recover from infectious diseases and inflammation, the host immune response is boosted, rendering pathophysiological perturbations in the microenvironment. This leads to modulations in the PRRPRR or PRR-ligand binding patterns. Therefore, a detailed understanding of the structural changes of PRRs and their corresponding binding interfaces under pathophysiological condition, particularly by targeting the custodian proteinprotein interaction is clinically valuable. This molecular strategy can provide a powerful rationale for future drug design and development to immunomodulate inflammation. The FREPs, which contain FBG, represents a common family of proteins in both the invertebrates and vertebrates. Ficolins are a representative group of FREPs, which contain FBG domain. Since the FBG domain is highly conserved, and as a proof of concept, we used Mficolin to demonstrate that FBG undergoes dramatic conformational change under pathophysiological condition, exposing the C-terminal, which might affect the interaction between FBG and its interacting partners. By solid phase protein-protein interaction studies, we demonstrated that CRP binds to the FBG domain of Mficolin, and CRP:FBG interaction is both pH- and calciumsensitive with a 100-fold increase in binding affinity under a defined pathophysiological condition of local acidosis and hypocalcaemia. By HDMS and SPR analyses, we found that residues 284-326 at the C-terminal region of FBG are located at the binding interface with CRP under pathophysiological condition. Under physiological condition, region 284-305 moved away from the binding interface. There appears to be more residues located at the CRP-binding interface under pathophysiological condition compared to physiological condition, which probably explains the enhanced CRP:FBG interaction under pathophysiological conditions. As region 284-305 is also identified as the pH-sensitive region, which is more exposed to the solvent under pathophysiological condition, we reasoned that probably the conformation of FBG under physiological condition makes region 284-305 less accessible to CRP. This hypothesis is further supported by SPR analyses using the synthesized peptides, which shows that without the effect of secondary structure, peptide 284305 binds to CRP under both pathophysiological condition and physiological condition. Although region 205-220 showed decrease in deuterium incorporation upon CRP binding, the lack of binding of the synthesized FBG peptide of 205-220 to CRP was observed. Therefore, this region was excluded from the binding interface. Region 306-326 is consistently located on the binding interface in both the pathophysiological condition and physiological condition, as was also supported by the HDMS and SPR results. The critical function of the histidine residue in M-ficolin has been identified to regulate the pH-sensitive binding of M-ficolin to GlcNAc27. Here, we pinpointed to His284 on the FBG to be the critical CRP-binding site. This site is conserved in most of the M-ficolin homologues, and absent in H-ficolin. This possibly explains our previous observation of the lack of interaction between CRP and H- Figure 6. Biological significance of the pH- and calciumsensitive crosstalk between CRP and M-ficolin. (A) SPR was used to test the binding affinity of individual CRP, FBG and CRP:FBG complex to GlcNAc, under physiological and pathophysiological conditions. “ns” indicates no significant binding according to the valid range of association/dissociation rate constant (BIAcore 2000 manual). (B) A schematic model illustrating CRP:FBG interaction at physiological and pathophysiological conditions. Under pathophysiological condition, the binding interface of CRP (blue):FBG (red) overlaps with GlcNAc-binding site (yellow) on the FBG molecule, whereas under physiological condition, the GlcNAcbinding site on FBG is exposed. Therefore, we hypothesized that under pathophysiological condition brought about by infection, the crosstalk between CRP and M-ficolin might block the pathogen-recognition site of FBG and inhibit further interaction with the pathogen, and therefore, switches to homeostasis and restoration of the physiological condition. ficolin, regardless of pH and calcium shift [21]. It is reported that the conserved PAMP/sugar binding pocket is also located in the region 282-285 and under acidic condition, the FBG will present as an inactive form for PAMP/sugar-binding. This hypothesis was further strengthened by SPR analysis which indicates that the crosslinked CRP:FBG complex showed lower binding affinity compared to FBG under pathophysiological condition. Therefore, we envisage that the stronger binding of CRP to the FBG His284 under pathophysiological condition will further dissociate the PAMP/sugar and switch the protein’s structure-activity towards downregulation of the antimicrobial immune response in order to restore homeostasis. Our elucidation of the binding interface and the infection-inflammation sensitive regulatory region of the M-ficolin provide insights into the bioactive centre of the M-ficolin molecule, which might be useful for future drug development for immunomodulation. Overall, we have found that as a representative of FREPs, the FBG of M-ficolin, exploits its flexible structure and modifies its binding pattern with its interacting partner, CRP (an acute phase protein), under different pathophysiological conditions. This structurefunction switch probably diverts the pathogen-recognition activity of the FBG to the regulation of the host immune response, thus, restoring homeostasis. times with 300 µl of TBST, and 100 µl of the peroxidase substrate, ABTS, was added. After 10-min incubation at room temperature in the dark, the OD405 nm was read. MBS was commonly used to adjust the acidic condition of the serum. Thus MBS containing 25 mM MES, 145 mM NaCl adjusted to pH 6.5 and below, and TBS containing 25 mM Tris-HCl, 145 mM NaCl, adjusted to pH 7.0 and above, were used as the binding buffers to examine the effect of pH on the protein-protein interaction. Varying concentrations of calcium was supplemented to the binding buffer to examine the effect of calcium. Surface Plasmon Resonance BIAcore 2000 was used to demonstrate real-time biointeraction. PC-HPA chip was prepared by immobilizing 1-palmitoyl-2-oleoyl-phosphatidylcholine, PC (Avanti Polar Lipids, Birmingham, AL) on the HPA chip (GE Healthcare). Then 100 nM CRP was injected over the surface of PC-HPA followed by separate injections of different amounts of FBG or synthesized peptides at a flow rate of 30 µl/min. The amount of all the proteins was calculated as monomer. The dissociation was for 180 s at the same flow rate. The PC-HPA chip was regenerated by injection of 15 µl of 0.1 M NaOH at 30 µl/min. Injection of binding buffer without CRP was used as the negative control. MBS containing 25 mM MES, 145 mM NaCl adjusted to pH 6.5 and below, and TBS containing 25 mM Tris-HCl, 145 mM NaCl, adjusted to pH 7.4, were used as the running and diluting buffers to simulate physiological condition and pathophysiological condition. For the synthesized peptide binding to CRP, as the peptides were dissolved in water, which might dilute the binding buffer, buffer composed of same ratio of running buffer to water was made and used as negative controls. Protein binding signals were calculated by subtracting the negative control. BIAevaluation 3.2 software was used to calculate the KD using a Langmiur 1:1 binding model. According to BIAcore 2000 manual (GE healthcare, Uppsala, Sweden) the valid range of association rate constant Ka is 103 to 106 M-1S-1 and the valid range of dissociation rate constant Kd is 10-5 to 10-1. The valid range of equilibrium constant KD is 10-4 to 10-11 M. When the calculated parameters of a binding curve fall out of the valid range of Ka , Kd and KD, we consider that there is no significant binding observed. MATERIALS AND METHODS In silico analysis of sequence of FREPs FREP proteins were identified using domain search on the SMART database [45,46]. Human FREPs were selected and the primary sequence and architectural domains of the human FREPs were checked one by one in SMART’s NRDB, non-redundant database. Proteins with intact FBG domain were chosen. Multiple sequence alignment was carried out using Accelrys discovery studio 2.5 (Accelrys, Inc., San Diego, CA). A phylogenetic tree was then constructed based on the sequence alignment with a bootstrap value of 10000. Solid phase protein-protein interaction assay (ELISA) To investigate the effects of pH and calcium on the Mficolin:CRP interaction, ELISA was conducted to test the binding between wild-type/mutant FBG and CRP. CRP at 0.8 µg/well in coating buffer (50 mM NaHCO3/Na2CO3 buffer, pH 9.6 containing 0.9% (w/v) NaCl) was immobilised onto the 96-well FalconTM plates (BD Biosciences, San Jose, CA) at oC overnight. Then, the wells were rinsed three times with 300 µl of TBS and blocked with 200 µl of 3% (w/v) bovine serum albumin (BSA) in TBS at 37 oC for h. The wells were rinsed four times post-blocking with 300 µl of TBS containing 0.05% (v/v) Tween 20 (TBST) before varying concentrations of wild-type/mutant FBG were added. The reactions were incubated at 37 oC for h followed by rinsing times with TBST. Next, anti-myc antibodies (Invitrogen) were added at 1:3000 dilution (in TBST with 3% (w/v) BSA) for the detection of myc-his-tagged wild-type and mutant FBG. Subsequently, the secondary anti-mouse antibodies conjugated with horseradish peroxidase (1:3000 dilution in TBST with 3% (w/v) of BSA; Dako) were added and incubated at 37 oC for h. Finally, the wells were rinsed HDMS HDMS reaction was initiated by combining µl of protein solutions each at concentrations > 2.5 g/ml with 18 µl of one of the following deuterated buffers: (i) pH 7.4 without calcium, (ii) pH 7.4 with 2.5 mM calcium, (iii) pH 6.5 without calcium and (iv) pH 6.5 with mM calcium. After 0, 0.5, 1, 2, 5, or 10 min, the hydrogen-deuterium exchange reaction was quenched by the addition of 180 µl of ice-cold 0.1% (v/v) trifluoroacetic acid (TFA) (SigmaAldrich) at pH 2.5. Then, an aliquot of 100 µl of the quenched reaction was mixed with 50 µl of pepsin bead slurry (Pierce) which was previously activated by washing three times in 500 µl of 0.1% TFA at 4oC. The mixture was incubated on ice with occasional vortexing for with 30 s intervals. The exchange mixture was then centrifuged for at 7000 g at °C, divided into three aliquots, flash-frozen in liquid N2 and stored at -80°C until analyzed. The pepsin-digested protein was analyzed by mass spectrometry with ionization method of matrixassisted laser desorption/ionization (MALDI) using 4800 Plus MALDI TOF/TOF™ Analyzer (Applied Biosystems, Foster City, CA). The back-exchange that occurred during the analysis was determined by carrying out control experiments where FBG and CRP were individually deuterated for 24 h at 25 °C. The spectra were viewed and calibrated by Data Explorer® V4.9 (Applied Biosystems). Changes in deuterium incorporation of > +10% were considered significant. Although a higher level of deuterium exchange is expected at higher pH [47], our data showed unexpected higher level of deuterium exchange at lower pH. This is indicative of the significance of our finding. The MALDI-TOF-TOF spectra were analyzed using the theoretical mass of two prominent peptides (theoretical m/z = 803.34 and 1637.82). The average mass of a peptide was calculated by determining the centroid of its isotopic envelope using software, Decapp (Applied Biosystems). The difference between the average masses of the deuterated and non-deuterated peptide was the number of deuterions incorporated. The side chain exchange was determined to be 4.5% of fast exchanging side chain hydrogen atoms based on dilution factors. MALDI data analysis corrected for the side chain deuteration was carried out prior to back-exchange correction. Finally, a correction factor was applied to account for the amount of back exchange. The correction factor was calculated using the following formula: was transfected into HEK 293 cells by lipofectamine 2000 (Invitrogen) according to the product instruction. The medium was collected at 48 h after transfection. 300 l NiNTA beads (Qiagen, Valencia, CA) were added to 30 ml medium and incubated at oC overnight. Then the beads were packed into the 10 ml poly-prepTM chromatography column (Biorad, Hercules, CA). Fifty milliliters of washing buffer containing 20 mM Tris, pH 8.0, 500 mM NaCl and 20 mM imidazole (Sigma) was passed through the column. Using elution buffer (20 mM Tris, pH 8.0, 500 mM NaCl and 250 mM imidazole), the protein was eluted in fractions of 120 l each. The purity of the recombinant proteins was verified by SDS PAGE. Since the mature FBG contains 326 amino acids, the extra amino acids are attributable to the cloning strategy of the recombinant proteins. Therefore the extra amino acids were not accounted for in the FBG protein. and Crosslinking and purification of CRP:FBG complex CRP:FBG complex was formed by adding excessive FBG to CRP and incubating for h at room temperature in the corresponding buffers to simulate physiological condition and pathophysiological condition. Then the protein was fixed by adding paraformaldehyde solution to a final concentration of 0.2% for 10 according to the method described previously [48,49]. Protein complex was purified by PC-Sepharose (Pierce, Rockford, IL) and eluted with the same binding buffer containing mM EDTA. The protein buffer was exchanged to TBS or MBS buffer using Vivaspin column (Sartorius Stedim Biotech, Aubagne, France). Peptide design and synthesis The pepsin digested fragment of FBG identified to locate in CRP binding interface on FBG was chosen and vertified by hydrophilicity and solubility values. Three peptides were selected from regions spanning: 205-220 (RVDLVDFEGNHQFAKY), 284-305 (HASNLNGLYLMGPHESYANGIN), 306-321 (WSAAKGYKYSYKVSEMKVRPA). The peptides were synthesized by Genemed Synthesis Inc., California, USA, and purified to >95% under pyrogen-free condition. Peptides were dissolved in water to make a mM stock solution, aliquoted and stored at -20 oC. Kinetic plots of deuteration best fit were made to a single exponential model accounting for deuterions exchanging at a rapid rate (mainly solvent-accessible amides). The best fit was implemented in GraphPad Prism® V5 (GraphPad Software, San Diego, CA). Site-directed mutagenesis of M-ficolin FBG The site-directed mutagenesis was performed using QuickChange® XL Site-directed Mutagenesis Kit (Stratagene, La Jolla, CA) according to the manufacturer’s instructions. PCR was carried out with initial denaturation at 95 oC for min, followed by 18 cycles of denaturation at 95 oC for 50 s, annealing at 60 oC for 50 s and extension at 68 oC for before the last extension step of 68 oC for min. Following the temperature cycling, the product was treated with 10 Units of DpnI endonuclease at 37 oC for h to digest the parental DNA template. The nicked vector DNA incorporating the desired mutation was then transformed into competent E. coli TOP10 (Invitrogen) and plated on LB-Amp (100 µg/ml) plates. Colonies were randomly selected and isolated. DNA sequencing was carried out to verify the plasmid sequence. ACKNOWLEDGEMENTS We thank Ms Yoong Sia Lee and Ms Mok Lim Sum (Proteins & Proteomics Center, NUS) for assistance with HDMS. This research was supported by BMRC A*STAR (08/1/21/19/574), MoE (T208B3109) and SMA-CSB. Jing Zhang is a research scholar of the NUS Graduate School for Integrative Science and Engineering. DISCLOSURE OF CONFLICTS OF INTEREST We declare no competing interest. REFERENCES Expression and purification of wildtype/mutant FBG proteins The mutant proteins were expressed and purified by affinity chromatography using Ni-NTA beads (Qiagen, Hilden, Germany) as described previously [21]. Briefly, the clone 1. Baranov D, Neligan P (2007) Trauma and aggressive homeostasis management. Anesthesiol Clin 25: 49-63, viii. 2. Geborek P, Saxne T, Pettersson H, Wollheim FA (1989) Synovial fluid acidosis correlates with radiological joint destruction in rheumatoid arthritis knee joints. J Rheumatol 16: 468-472. 3. Ricciardolo FL, Gaston B, Hunt J (2004) Acid stress in the pathology of asthma. J Allergy Clin Immunol 113: 610-619. 4. Gerweck LE, Vijayappa S, Kozin S (2006) Tumor pH controls the in vivo efficacy of weak acid and base chemotherapeutics. Mol Cancer Ther 5: 1275-1279. 5. Issekutz AC, Bhimji S (1982) Role for endotoxin in the leukocyte infiltration accompanying Escherichia coli inflammation. Infect Immun 36: 558-566. 6. Martinez D, Vermeulen M, Trevani A, Ceballos A, Sabatte J, et al. (2006) Extracellular acidosis induces neutrophil activation by a mechanism dependent on activation of phosphatidylinositol 3-kinase/Akt and ERK pathways. J Immunol 176: 1163-1171. 7. Beers M (2000) The Merck Manual of Geriatrics 1507 p. 8. Feske S (2007) Calcium signalling in lymphocyte activation and disease. Nat Rev Immunol 7: 690702. 9. Iwasaki A, Medzhitov R (2004) Toll-like receptor control of the adaptive immune responses. Nat Immunol 5: 987-995. 10. Cairns SP, Westerblad H, Allen DG (1993) Changes in myoplasmic pH and calcium concentration during exposure to lactate in isolated rat ventricular myocytes. J Physiol 464: 561-574. 11. Trevani AS, Andonegui G, Giordano M, Lopez DH, Gamberale R, et al. (1999) Extracellular acidification induces human neutrophil activation. J Immunol 162: 4849-4857. 12. Vermeulen M, Giordano M, Trevani AS, Sedlik C, Gamberale R, et al. (2004) Acidosis improves uptake of antigens and MHC class I-restricted presentation by dendritic cells. J Immunol 172: 3196-3204. 13. Ackerman AL, Cresswell P (2004) Cellular mechanisms governing cross-presentation of exogenous antigens. Nat Immunol 5: 678-684. 14. Guermonprez P, Amigorena S (2005) Pathways for antigen cross presentation. Springer Semin Immunopathol 26: 257-271. 15. Runza VL, Schwaeble W, Mannel DN (2008) Ficolins: novel pattern recognition molecules of the innate immune response. Immunobiology 213: 297-306. 16. Bitto A, Minutoli L, Galeano MR, Altavilla D, Polito F, et al. (2008) Angiopoietin-1 gene transfer improves impaired wound healing in genetically diabetic mice without increasing VEGF expression. Clin Sci (Lond) 114: 707-718. 17. Johnson SA (1958) Recent advances in blood coagulation with regard to anticoagulant therapy, fibrinogen, hemophilia and platelet disorders. Henry Ford Hosp Med Bull 6: 291-297. 18. Wang X, Zhao Q, Christensen BM (2005) Identification and characterization of the fibrinogen-like domain of fibrinogen-related proteins in the mosquito, Anopheles gambiae, and the fruitfly, Drosophila melanogaster, genomes. BMC 10 Genomics 6: 114. 19. Gokudan S, Muta T, Tsuda R, Koori K, Kawahara T, et al. (1999) Horseshoe crab acetyl grouprecognizing lectins involved in innate immunity are structurally related to fibrinogen. Proc Natl Acad Sci U S A 96: 10086-10091. 20. Matsushita M, Endo Y, Taira S, Sato Y, Fujita T, et al. (1996) A novel human serum lectin with collagen- and fibrinogen-like domains that functions as an opsonin. J Biol Chem 271: 24482454. 21. Zhang J, Koh J, Lu J, Thiel S, Leong BS, et al. (2009) Local inflammation induces complement crosstalk which amplifies the antimicrobial response. PLoS Pathog 5: e1000282. 22. Teh C, Le Y, Lee SH, Lu J (2000) M-ficolin is expressed on monocytes and is a lectin binding to N-acetyl-D-glucosamine and mediates monocyte adhesion and phagocytosis of Escherichia coli. Immunology 101: 225-232. 23. Rorvig S, Honore C, Larsson LI, Ohlsson S, Pedersen CC, et al. (2009) Ficolin-1 is present in a highly mobilizable subset of human neutrophil granules and associates with the cell surface after stimulation with fMLP. J Leukoc Biol. 24. Ng PM, Le Saux A, Lee CM, Tan NS, Lu J, et al. (2007) C-reactive protein collaborates with plasma lectins to boost immune response against bacteria. EMBO J 26: 3431-3440. 25. Endo Y, Liu Y, Kanno K, Takahashi M, Matsushita M, et al. (2004) Identification of the mouse H-ficolin gene as a pseudogene and orthology between mouse ficolins A/B and human L-/M-ficolins. Genomics 84: 737-744. 26. Low DH, Frecer V, Le Saux A, Srinivasan GA, Ho B, et al. (2010) Molecular interfaces of the galactosebinding protein Tectonin domains in hostpathogen interaction. J Biol Chem 285: 98989907. 27. Low DH, Ang Z, Yuan Q, Frecer V, Ho B, et al. (2009) A novel human tectonin protein with multivalent beta-propeller folds interacts with ficolin and binds bacterial LPS. PLoS One 4: e6260. 28. Mold C, Rodic-Polic B, Du Clos TW (2002) Protection from Streptococcus pneumoniae infection by Creactive protein and natural antibody requires complement but not Fc gamma receptors. J Immunol 168: 6375-6381. 29. Bhakdi S, Torzewski M, Paprotka K, Schmitt S, Barsoom H, et al. (2004) Possible protective role for C-reactive protein in atherogenesis: complement activation by modified lipoproteins halts before detrimental terminal sequence. Circulation 109: 1870-1876. 30. Szalai AJ, Weaver CT, McCrory MA, van Ginkel FW, Reiman RM, et al. (2003) Delayed lupus onset in (NZB x NZW)F1 mice expressing a human Creactive protein transgene. Arthritis Rheum 48: 1602-1611. 31. Deodhar SD, James K, Chiang T, Edinger M, Barna BP (1982) Inhibition of lung metastases in mice bearing a malignant fibrosarcoma by treatment with liposomes containing human C-reactive protein. Cancer Res 42: 5084-5088. 32. Volanakis JE (2001) Human C-reactive protein: expression, structure, and function. Mol Immunol 38: 189-197. 33. Black S, Kushner I, Samols D (2004) C-reactive Protein. J Biol Chem 279: 48487-48490. 34. Kilpatrick JM, Volanakis JE (1991) Molecular genetics, structure, and function of C-reactive protein. Immunol Res 10: 43-53. 35. Pepys MB, Baltz ML (1983) Acute phase proteins with special reference to C-reactive protein and related proteins (pentaxins) and serum amyloid A protein. Adv Immunol 34: 141-212. 36. Volanakis JE (1982) Complement activation by Creactive protein complexes. Ann N Y Acad Sci 389: 235-250. 37. Shrive AK, Cheetham GM, Holden D, Myles DA, Turnell WG, et al. (1996) Three dimensional structure of human C-reactive protein. Nat Struct Biol 3: 346-354. 38. Ramadan MA, Shrive AK, Holden D, Myles DA, Volanakis JE, et al. (2002) The three-dimensional structure of calcium-depleted human C-reactive protein from perfectly twinned crystals. Acta Crystallogr D Biol Crystallogr 58: 992-1001. 39. Garlatti V, Martin L, Gout E, Reiser JB, Fujita T, et al. (2007) Structural basis for innate immune sensing by M-ficolin and its control by a pH-dependent conformational switch. J Biol Chem 282: 3581435820. 40. Liu Y, Endo Y, Iwaki D, Nakata M, Matsushita M, et al. (2005) Human M-ficolin is a secretory protein that activates the lectin complement pathway. J Immunol 175: 3150-3156. 11 41. Tanio M, Kondo S, Sugio S, Kohno T (2007) Trivalent recognition unit of innate immunity system: crystal structure of trimeric human M-ficolin fibrinogen-like domain. J Biol Chem 282: 38893895. 42. Tanio M, Wakamatsu K, Kohno T (2009) Binding site of C-reactive protein on M-ficolin. Mol Immunol 47: 215-221. 43. Simmen HP, Battaglia H, Giovanoli P, Blaser J (1994) Analysis of pH, pO2 and pCO2 in drainage fluid allows for rapid detection of infectious complications during the follow-up period after abdominal surgery. Infection 22: 386-389. 44. van Zwieten R, Wever R, Hamers MN, Weening RS, Roos D (1981) Extracellular proton release by stimulated neutrophils. J Clin Invest 68: 310-313. 45. Schultz J, Milpetz F, Bork P, Ponting CP (1998) SMART, a simple modular architecture research tool: identification of signaling domains. Proc Natl Acad Sci U S A 95: 5857-5864. 46. Jones DT (1999) Protein secondary structure prediction based on position-specific scoring matrices. J Mol Biol 292: 195-202. 47. Bai Y, Milne JS, Mayne L, Englander SW (1993) Primary structure effects on peptide group hydrogen exchange. Proteins 17: 75-86. 48. Carlson OWNaGM (2007) Protein interactions captured by chemical cross-linking: one-step cross-linking with formaldehyde. Cold Spring Harb Protoc. 49. Owen W. Nadeau GMC (2007) Protein Interactions Captured by Chemical Cross-linking: Simple Cross-linking Screen Using Sulfo-MBS. Cold Spring Harb Protoc. NGS Inaugural Symposium 2009 Date: 26th February 2009 (6.30pm-9pm) and 27th February 2009 (8.30am-6pm) Location: University Hall Auditorium More details of this symposium can be found here http://blog.nus.edu.sg/ngssymposium/ Selected student speakers as well as overseas keynote speakers (from both the biological sciences as well as engineering fields) will present their exciting work and results. This is a good opportunity for your students to improve their academic experience as well as to mingle around with students from various backgrounds and to interact with distinguished scientists from overseas. All are welcome and we would like you to strongly encourage your students to submit an abstract for either the poster or oral presentations. Attractive prizes will be awarded to the best speakers and posters from both categories. Oral Presentation Abstract (limited to 250 words or less): Crosstalk amongst pathogen recognition receptors enhances host defense against acute infection Zhang Jing1,2, Sethi S3, Ho B4,† and Ding JL1,† Department of Biological Sciences, 2NUS Graduate School for Integrative Science and Engineering, 3Department of Pathology, 4Department of Microbiology, National University of Singapore, Singapore, 117543. We recently demonstrated that in the invertebrates which only harbors innate immunity, the C-reactive protein (CRP) associates indirectly with a carcinolectin (CL5c), a homologue of human ficolins. Their association during a systemic infection was suggested to enhance immune response. This prompted us to examine how the human CRP and ficolins interact with each other during an acute infection, and to determine the consequence of their molecular communication to human innate immune defense. Here, we report that the direct interactions between CRP:L-ficolin and CRP:M-ficolin were enhanced in human serum at low pH and low calcium concentration, a condition prevalent in local infection and inflammation. By ELISA and surface plasmon resonance (SPR) analyses, we showed that the infection-mediated collaborative partnership between CRP:L-ficolin led to crosstalk between the classical and lectin complement pathways, resulting in the upregulation of C4-cleavage and phagocytosis to amplify the antimicrobial activity. On the other hand, CRP:M-ficolin interaction activated the NF-B pathway, enhancing the secretion of IL-8 by the monocytes, thus boosting the immune response. Importantly, by real-time imaging and flow cytometry, we found that P. aeruginosa, a persistent opportunistic human pathogen prevalent at the site of local infection-inflammation, is effectively killed by CRP:L-ficolin crosstalk in the serum of infected patients under inflammation condition. Overall, our findings provide new insights into the host immune response against invading pathogens, with potential for development of new therapeutic strategies against bacterial infection, particularly in a localized wound site. [...]... Chem 26 6(33): 22 459 -22 464 Inohara, N., Y Ogura, et al (20 03) "Host recognition of bacterial muramyl dipeptide mediated through NOD2 Implications for Crohn's disease." J Biol Chem 27 8(8): 5509-55 12 Ip, W K., K Takahashi, et al (20 09) "Mannose-binding lectin and innate immunity. " Immunol Rev 23 0(1): 9 -21 20 4 Ip, W K., K Takahashi, et al (20 08) "Mannose-binding lectin enhances Toll-like receptors 2 and. .. anti-fungal innate immune response." Nature 420 (69 12) : 1 82- 186 Garlatti, V., N Belloy, et al (20 07) "Structural insights into the innate immune recognition specificities of L- and H -ficolins. " Embo J 26 (2) : 623 -633 Garlatti, V., L Martin, et al (20 07) "Structural basis for innate immune sensing by M-ficolin and its control by a pH-dependent conformational switch." J Biol Chem 28 2(49): 35814-35 820 Garred,... paralogon groups, and fingerprints." Mol Pharmacol 63(6): 125 6- 127 2 Fujita, T (20 02) "Evolution of the lectin-complement pathway and its role in innate immunity. " Nat Rev Immunol 2( 5): 346-353 Fujita, T., M Matsushita, et al (20 04) "The lectin-complement pathway its role in innate immunity and evolution." Immunol Rev 198: 185 -20 2 Gallin J, S R e (1999) "Inflammation: Basic Principles and Clinical Correlates."... protein and formation of its complex with vitamin K-dependent protein S." Biochem J 20 9(3): 847-856 Dempsey, P W., M E Allison, et al (1996) "C3d of complement as a molecular adjuvant: bridging innate and acquired immunity. " Science 27 1( 524 7): 348-350 Dempsey, P W., S A Vaidya, et al (20 03) "The art of war: Innate and adaptive immune responses." Cell Mol Life Sci 60( 12) : 26 04 -26 21 Dixit, V and T W Mak (20 02) ... Genis, et al (20 05) "Membrane type 1-matrix metalloproteinase is regulated by chemokines monocyte-chemoattractant protein-1/ccl2 and interleukin-8/CXCL8 in endothelial cells during angiogenesis." J Biol Chem 28 0 (2) : 129 2- 129 8 Garcia-Garcia, E and C Rosales (20 02) "Signal transduction receptor-mediated phagocytosis." J Leukoc Biol 72( 6): 10 92- 1108 during Fc Garlanda, C., E Hirsch, et al (20 02) "Non-redundant... et al (20 09) "MBL2, FCN1, FCN2 and FCN3-The genes behind the initiation of the lectin pathway of complement." Mol Immunol 46(14): 27 37 -27 44 20 2 Geissmann, F., S Jung, et al (20 03) "Blood monocytes consist of two principal subsets with distinct migratory properties." Immunity 19(1): 71- 82 Gewurz, H., C Mold, et al (19 82) "C-reactive protein and the acute phase response." Adv Intern Med 27 : 345-3 72 Girardin,... Thiel, et al (20 03) "Collectins and ficolins: humoral lectins of the innate immune defense." Annu Rev Immunol 21 : 547-578 Holmskov, U., S Thiel, et al (20 03) "Collectins and ficolins: Humoral lectins of the innate immune defense." Annual Review of Immunology 21 : 547-578 Holmskov, U., S Thiel, et al (20 03) "Collections and ficolins: humoral lectins of the innate immune defense." Annu Rev Immunol 21 : 547-578... R and B Dahlback (1990) "C4b-binding protein, a regulatory component of the classical pathway of complement, is an acute-phase protein and is elevated in systemic lupus erythematosus." Complement Inflamm 7 (2) : 71-77 Basset, C., J Holton, et al (20 03) "Innate immunity and pathogen-host interaction." Vaccine 21 Suppl 2: S 12- 23 Beers, M (20 00) The Merck Manual of Geriatrics Beutler, B (20 04) "Innate immunity: ... adhesion-recognition receptor essential for innate immunity. " Microbes Infect 2( 3): 28 9 -29 4 20 0 Eisen, D P and R M Minchinton (20 03) "Impact of mannose-binding lectin on susceptibility to infectious diseases." Clin Infect Dis 37(11): 1496-1505 Endo, Y., Y Liu, et al (20 04) "Identification of the mouse H-ficolin gene as a pseudogene and orthology between mouse ficolins A/B and human L-/M -ficolins. " Genomics 84(4): 737-744... and complement receptors in induction and regulation of immunity. " Annu Rev Immunol 16: 545-568 Carroll, M C (20 04) "The complement system in regulation of adaptive immunity. " Nat Immunol 5(10): 981-986 Cerenius, L., B L Lee, et al (20 08) "The proPO-system: pros and cons for its role in invertebrate immunity. " Trends Immunol 29 (6): 26 3 -27 1 Chen, S C., C H Yen, et al (20 01) "Biochemical properties and . infection-inflammation condition (pH 6.5 and 2 mM calcium) and normal physiological condition (pH 7.4 and 2. 5 mM calcium), we demonstrated crosstalk between CRP and L-ficolin, which resulted in two. interaction between CRP and H-ficolin, regardless of pH and calcium shift (Zhang et al. 20 09). It is reported that the conserved PAMP/sugar binding pocket is also located in the region 28 2 -28 5 and. awaiting 1 92 elucidation and understanding. Taken together, the differential interactions between (a) L-ficolin :CRP, (b) M-ficolin :CRP and (c) CRP: M-ficolin:GPCR43 under normal and infection-inflammation