Rapid evolution of binding specificities and expression patterns of inhibitory CD33-related Siglecs in primates

Abstract

Siglecs are sialic acid-binding Ig-like lectins that recognize sialoglycans via amino-terminal V-set domains. CD33-related Siglecs (CD33rSiglecs) on innate immune cells recognize endogenous sialoglycans as "self-associated molecular patterns" (SAMPs), dampening immune responses via cytosolic immunoreceptor tyrosine-based inhibition motifs that recruit tyrosine phosphatases. However, sialic acid-expressing pathogens subvert this mechanism through molecular mimicry. Meanwhile, endogenous host SAMPs must continually evolve to evade other pathogens that exploit sialic acids as invasion targets. We hypothesized that these opposing selection forces have accelerated CD33rSiglec evolution. We address this by comparative analysis of major CD33rSiglec (Siglec-3, Siglec-5, and Siglec-9) orthologs in humans, chimpanzees, and baboons. Recombinant soluble molecules displaying ligand-binding domains show marked quantitative and qualitative interspecies differences in interactions with strains of the sialylated pathogen, group B Streptococcus, and with sialoglycans presented as gangliosides or in the form of sialoglycan microarrays, including variations such as N-glycolyl and O-acetyl groups. Primate Siglecs also show quantitative and qualitative intra- and interspecies variations in expression patterns on leukocytes, both in circulation and in tissues. Taken together our data explain why the CD33rSiglec-encoding gene cluster is undergoing rapid evolution via multiple mechanisms, driven by the need to maintain self-recognition by innate immune cells, while escaping 2 distinct mechanisms of pathogen subversion.

Keywords: N-acetylneuraminic acid; N-glycolylneuraminic acid; innate immunity; sialic acids.

Figure 1.

Binding of Siglec-Fc chimeras to GBS are divergent among the 3 species. A) Diagram of various types of GBS CPS. B) Differences in binding patterns of human (Hu), chimpanzee (Ch) and baboon (Bab) Siglec-3-Fc and Siglec-9-Fc to GBS. Top panel: schematic diagram of WT vs. AUS-treated GBS. WT GBS CPS carries terminal Sias that are represented as spikes over a circle. When such bacteria are treated with AUS, the Sia are peeled off; hence, the desialylated bacteria are represented as circles without spikes. Left panel: human Siglec-3 showed binding to GBS type Ia and type III in a Sia-dependent manner, whereas neither chimpanzee nor baboon Siglec-3 showed significant GBS binding. Right panel: human and chimpanzee Siglec-9 showed Sia-independent (type Ib), partially Sia-dependent (type Ia), or Sia-dependent (type III) binding to GBS, whereas the dependence on Sia in the binding to type Ia differed between human and chimpanzee Siglec-9. Baboon Siglec-9 showed no significant binding to GBS. Ex, excitation; Em, emission. C) Effects of AUS pretreatment of Siglec-Fc chimeras on binding to GBS. Siglec-Fc chimeras carry sialylated glycans and can potentially bind to each other when found in proximity (cis binding). However, when the Siglec-Fc chimeras are pretreated with AUS, the terminal Sias are removed and their binding sites are free to bind other sialylated targets (trans binding).

Figure 2.

Ganglioside binding patterns of Siglec-Fc-chimeras are divergent between the 3 species. Binding patterns of Fc chimeras of CD33rSiglec (Siglec-3 and -9) from human, chimpanzee (chimp), and baboon were tested by ELISA using multiple types of GM3-based gangliosides presenting different types of Sias. Top panel: Siglec-3 of baboon showed robust binding to most GM3s tested, with strong preference for Neu5Gc. Human and chimpanzee Siglec-3 also showed some preference toward Neu5Gc, but not as marked as that of baboon Siglec-3. Bottom panel: Siglec-9 of human and baboon showed similar patterns of binding, whereas chimpanzee Siglec-9 showed weaker binding. Ac, Neu5Ac; Gc, Neu5Gc.

Figure 3.

Sialoglycan microarray analysis shows variable species-specific Siglec binding patterns. A) Binding of human, chimpanzee, and baboon Siglec-Fc chimeras was assayed at 10, 20, and 40 μg/ml and detected with 1.5 μg/ml Cy3-conjugated goat anti-human IgG. Binding was ranked (43) as percentage of maximal signal in each array [rank=100×(glycan RFU/RFU Max at each concentration)] and the relative ranking was then averaged to obtain the average rank of each glycan. Each Siglec-Fc average rank is presented as a heat map (red, white, and blue represent the maximum, 50th percentile, and minimum, respectively). Major findings include: all Siglecs showed preference toward non-9-O-acetylated sialic acids; all Siglecs showed preferences toward Neu5Gc over Neu5Ac to differing extents, with baboon Siglec-3 the most extreme; and binding preferences of human and baboon Siglec-9 were similar to each other, whereas chimpanzee (chimp) Siglec-9 showed unique properties, confirming the result shown in Fig. 2. B) Neu5Ac vs. Neu5Gc binding preferences of human, chimpanzee, and baboon Siglec-3-, Siglec-5-, and Siglec-9-Fcs were demonstrated after combining and averaging RFU values for all the non-O-acetylated-Neu5Ac glycans and the non-O-acetylated-Neu5Gc glycans and are presented as percentage stacked columns. Human Siglecs showed increased “tolerance” toward Neu5Ac.

Figure 4.

Intra- and interspecies variability in blood leukocyte cell type expression patterns and expression levels in flow cytometry analysis. A) Representative antibody staining patterns of monocytes. Gray-shaded area represents human monocytes stained with the negative control antibody. B) Overall summary of CD33-related Siglec (CDrSiglec) expressions on monocytes and granulocytes, represented as percentage of positive cells (top panels) and MFI (bottom panels) of each population. Each point represents an individual donor. We found a significant difference in MFI between human and chimpanzee monocytes for Siglec-3 (P=0.004) and for Siglec-9 (P=0.02). Furthermore, there was a significant difference in granulocyte MFI and percentage of positive cells between human and chimpanzee (P=0.008 and P=0.02, respectively; Kruskal-Wallis test).

Figure 5.

Variability in expression patterns of Siglecs between human, chimpanzee, and gorilla spleens. A) Frozen sections of human, chimpanzee, and gorilla spleen were overlaid with anti-Siglec-3, anti-Siglec-5, or anti-Siglec-9 followed by biotinylated anti-mouse or anti-rabbit IgG, alkaline phosphatase streptavidin, and developed using a substrate kit as described in the Materials and Methods. The blue color indicates binding of the antibody. Sections were counterstained using a nuclear stain (red) (×40). B) Comparison of the number of pixels per ×40 field of frozen human (Hu, black bars), chimpanzee (Ch, white bars), and gorilla (Gor, gray bars) spleen sections overlaid with anti-Siglec-3, anti-Siglec-5, or anti-Siglec-9 (blue). *P < 0.05; **P < 0.005; ***P < 0.0005.

Figure 6.

Immune cell expression patterns of Siglecs between human, chimpanzee, and gorilla spleens. A) Siglec expression on spleen macrophages. Frozen spleen sections from human, chimpanzee, and gorilla were stained with anti-CD68 (a pan-macrophage marker; detected with Cy3-conjugated IgG; red fluorescence), and anti-Siglec-3, anti-Siglec-5, or anti-Siglec-9 (detected by biotinylated IgG and Alexa Fluor 488-conjugated streptavidin; green fluorescence). Siglec-3 was expressed on macrophages of all 3 species, whereas Siglec-5 was not expressed on macrophages. Siglec-9 was only detectable on a subset of human macrophages but not on chimpanzee or gorilla macrophages. ×200; scale bar = 100 μm. B) Siglec-5 expression on spleen neutrophils. Frozen spleen sections were stained with antineutrophil elastase (a neutrophil marker; detected by Cy3-conjugated IgG; red fluorescence) and anti-Siglec-5 (detected by biotinylated IgG and subsequently by Alexa Fluor 488-conjugated streptavidin; green fluorescence). Double positive cells are orange/yellow. ×200; scale bar = 100 μm.

Figure 7.

Probable evolutionary chain of Red Queen effects involving Sias and CD33rSiglecs. See text for discussion. [Based on Varki and Angata (7). Originally modified by T.A. for Glycoforum/Glycowords (http://www.glycoforum.gr.jp/science/word/evolution/ES-C04E.html), and further modified here.]