Understanding diversity of human innate immunity receptors: analysis of surface features of leucine-rich repeat domains in NLRs and TLRs

Abstract

Background: The human innate immune system uses a system of extracellular Toll-like receptors (TLRs) and intracellular Nod-like receptors (NLRs) to match the appropriate level of immune response to the level of threat from the current environment. Almost all NLRs and TLRs have a domain consisting of multiple leucine-rich repeats (LRRs), which is believed to be involved in ligand binding. LRRs, found also in thousands of other proteins, form a well-defined "horseshoe"-shaped structural scaffold that can be used for a variety of functions, from binding specific ligands to performing a general structural role. The specific functional roles of LRR domains in NLRs and TLRs are thus defined by their detailed surface features. While experimental crystal structures of four human TLRs have been solved, no structure data are available for NLRs.

Results: We report a quantitative, comparative analysis of the surface features of LRR domains in human NLRs and TLRs, using predicted three-dimensional structures for NLRs. Specifically, we calculated amino acid hydrophobicity, charge, and glycosylation distributions within LRR domain surfaces and assessed their similarity by clustering. Despite differences in structural and genomic organization, comparison of LRR surface features in NLRs and TLRs allowed us to hypothesize about their possible functional similarities. We find agreement between predicted surface similarities and similar functional roles in NLRs and TLRs with known agonists, and suggest possible binding partners for uncharacterized NLRs.

Conclusion: Despite its low resolution, our approach permits comparison of molecular surface features in the absence of crystal structure data. Our results illustrate diversity of surface features of innate immunity receptors and provide hints for function of NLRs whose specific role in innate immunity is yet unknown.

Figure 1

Phylogeny of amino acid sequences of C-terminal LRR domains agrees with classification of innate immunity receptors according to their N-terminal effector domains. The most divergent sequence, NAIP, was used to root the tree. Branch lengths are proportional to relative evolutionary distances. Integer numbers indicate bootstrap values (obtained by sampling over 1000 tree realizations) assessing statistical validity of the tree topology.

Figure 2

Surface features of LRR domains from NLRs (homology models) and TLRs (X-ray crystal structures). Columns with three-dimensional structures show different views and representations of LRR domains. First column: cartoon representation colored according to secondary structure. Second, fourth, and sixth columns: molecular surfaces colored according to electrostatic potential. Third, fifth, and seventh columns: molecular surfaces colored according to hydrophobicity.

Figure 3

An example of mapping the RI amino acid sequence into structure and of RI surface partitioning. A, Mapping between an LRR sequence motif and the RI structure. RI sequence was searched for the conserved LRR pattern LaaLXL, and 16 LRRs were identified. The residue X (magenta) was then used as a reference to define residues belonging to the inner concave surface (yellow and magenta), N-terminal side (green), C-terminal side (blue), and outer convex surface (red). Mapping between a general LRR motif sequence and the structure of LRR #13 is shown by coloring. B, Full RI surface partitioned into four parts as described above.

Figure 4

Hierarchical clustering analysis of amino acid hydropathy distributions within LRR domains of NLRs and TLRs. Color matrix shows values for first four moments (m₀,..., m₃; cf. Equation 1) of the hydropathy distribution over four LRR domain surfaces: inner (concave), N-terminal side, C-terminal side, and outer (convex). Red coloring corresponds to positive values of the moments, black to zero, and green to negative values. The measure of similarity between sequences of moment values of LRR surfaces is the Spearman's rank-order correlation coefficient, r_S. The tree on the left is the result of hierarchical clustering of pairwise distances by the complete linkage method. Length of edges is proportional to distances, d_S, between sequences of moment values and is defined as d_S = 1 - r_S. Grouping of LRR domains into clusters indicates overall similarity of their hydrophobicity distributions.

Figure 5

Hierarchical clustering analysis of amino acid charge distributions within LRR domains of NLRs and TLRs. All notations are the same as in Figure 4.