SR-A, MARCO and TLRs differentially recognise selected surface proteins from Neisseria meningitidis: an example of fine specificity in microbial ligand recognition by innate immune receptors

Abstract

Macrophages express various classes of pattern recognition receptors involved in innate immune recognition of artificial, microbial and host-derived ligands. These include the scavenger receptors (SRs), which are important for phagocytosis, and the Toll-like receptors (TLRs) involved in microbe sensing. The class A macrophage scavenger receptor (SR-A) and macrophage receptor with a collagenous structure (MARCO) display similar domain structures and ligand-binding specificity, which has led to the assumption that these two receptors may be functionally redundant. In this study we show that SR-A and MARCO differentially recognise artificial polyanionic ligands as well as surface proteins from the pathogenic bacterium Neisseria meningitidis. We show that, while acetylated low-density lipoprotein (AcLDL) is a strong ligand for SR-A, it is not a ligand for MARCO. Of the neisserial proteins that were SR ligands, some were ligands for both receptors, while other proteins were only recognised by either SR-A or MARCO. We also analysed the potential of these ligands to act as TLR agonists and assessed the requirement for SR-A and MARCO in pro-inflammatory cytokine induction. SR ligation alone did not induce cytokine production; however, for proteins that were both SR and TLR ligands, the SRs were required for full activation of TLR pathways.

Fig. 2

SR-A and MARCO differentially bind to N. meningitidis surface proteins. Wells of duplicate 96-well plates were coated with 10 µg/ml of various purified His- or GST-tagged N. meningitidis surface proteins. a SR-A binding was detected using Mt lysates from WT and SR-A–/– BMMϕ along with the SR-A-specific monoclonal antibody 2F8. PBS and AcLDL were used as a negative and a positive control for binding to SR-A, respectively (inset). b Binding to MARCO was detected using recombinant soluble MARCO and the anti-MARCO monoclonal antibody ED31. Recombinant soluble EMR1 acted as a control for non-specific binding. ChSO₄ and DxSO₄ were used as a negative and a positive control for binding to MARCO, respectively (inset).

Fig. 3

Polyanionic ligands do not stimulate TLR-mediated cytokine induction. Bg-PMϕ from WT and MyD88–/– mice were stimulated with 1 µg/ml of each polyanionic ligand and the induction of IL-6 (a) and TNF-α (b) was measured. Cells were separately stimulated with E. coli LPS (100 ng/ml), which served as a positive control for cytokine induction.

Fig. 4

Selected N. meningitidis surface proteins stimulate TLR-mediated cytokine induction. Bg-PMϕ from WT and MyD88–/– mice were stimulated with 1 µg/ml of the purified recombinant N. meningitidis surface proteins and the induction of IL-6 (a) and TNF-α (b) was measured. Cells were separately stimulated with E. coli LPS (100 ng/ml), which served as a positive control for cytokine induction.

Fig. 5

Cytokine induction in SR-A–/–, MARCO–/– and double knock-out Mϕ. RPMϕ from WT, SR-A–/–, MARCO–/– and SR-A-MARCO–/– double knock-out animals were stimulated with 1 µg/ml of the N. meningitidis proteins that were showntobeTLR agonists in figure 4 and the induction of IL-6 (a) and TNF-α (b) was measured. Each strain of Mϕ was also stimulated with E. coli LPS (100 ng/ml), which served as a positive control for cytokine induction.

Fig. 1

SR-A and MARCO bind to various polyanionic ligands. a Known ligands and non-ligands for SR-A were coated into wells of a 96-well plate and overlaid with post-nuclear cell lysate from WT and SR-A knock-out (SR-A^−/−) BMMϕ. Binding to SR-A was detected with a rat anti-mouse SR-A antibody, 2F8, and visualized with a secondary anti-rat HRP antibody and TMB reagent. b In parallel, wells of a 96-well plate were coated with the same series of polyanionic molecules as in a and overlaid with recombinant sMARCO or recombinant sEMR1, an unrelated Mϕ surface receptor, as a control for non-specific binding. Binding of sMARCO and sEMR1 was detected by using their specific monoclonal antibodies ED31 and BC9, respectively, together with an HRP-conjugated secondary antibody.