F-type lectins
Beta-strands in the three-stranded sheet are coloured yellow and those in the five-stranded sheet are coloured blue. 310 helices are coloured red. The fucose ligand is shown in purple and the Ca2+ ion in blue. Protein Data Bank structure ID: 1K12.
Structure of F-type domain 1 from S pnemoniae SP2159 with bound blood group A tetrasaccharide
Beta-strands in the three-stranded sheet are coloured yellow, those in the five-stranded sheet are coloured blue, and those in other sub-structures are coloured green. The 310 helix is coloured red. Ca2+ is shown in blue. The ligand is coloured by monosaccharide: yellow, Gal/GalNac; blue, Glc; red, Fuc. Protein Data Bank structure ID: 2J1U.
Evolution of the F-type domain
F-type lectins (fucolectins) are fucose-binding proteins that share a characteristic sequence motif. F-type domains are found in proteins from a range of organisms from bacteria to vertebrates, but exhibit patchy distribution across different phylogenetic taxa, suggesting that F-type lectin genes have been selectively lost even between closely related lineages, and making it difficult to trace the ancestry of the F-type domain. F-type lectins have not been found in plants or fungi and are notably absent in mammals, birds and reptiles. The R-type CRD is the only other sugar binding module from an animal lectin also found in bacterial proteins, but the F-type domain is likely to be present in bacteria as a result of horizontal transfer from metazoans rather than due to early evolutionary origin, because it is present in only a small fraction of bacterial genomes. F-type lectins were first characterized in eels, which are teleost fish. Teleost fish species commonly have a number of F-type lectins (often containing multiple tandem F-type domains) resulting from gene duplication events that have occurred independently in different teleost fish lineages. The F-type domain has clearly gained functional value in fish, whereas in mammals it has become defunct. The duplication and diversification of F-type lectins in the genomes of teleost fish may have expanded the repertoire of pathogen epitopes that can be recognized by these lectins, or allowed for differential spatial expression. Proteins from non-fish species feature the F-type domain in a variety of domain architectural contexts, including multiple tandem F-type domains and/or combination of the F-type domain with other structural modules. The varied domain architecture and temperospatial expression of F-type lectins in different species suggests that the function of the F-type domain has evolved along unique paths in distinct evolutionary lineages, perhaps with the sugar recognition capability of the domain being harnessed in different functional contexts. While the functions of many F-type lectins have not yet been characterized, a pathogen-recognition role has been established for F-type lectin-like proteins in both invertebrates (eg the horseshoe crab Tachypleus tridentatus) and vertebrates (eg the Japanese eel Anguilla japonica).
Anguilla anguilla agglutinin is an F-type lectin from the European eel
Anguilla anguilla agglutinin (AAA) is a fucolectin from the European eel. It has been used widely in blood typing and histochemistry because it binds specifically to terminal fucose residues in certain blood group antigens, but the physiological function of AAA is as a serum pathogen recognition molecule of the innate immune system which recognizes bacterial lipopolysaccharides. AAA is a non-covalent homotrimer in which all three fucose binding sites are oriented in the same direction. This arrangement resembles that of the C-type CRDs in mannose binding protein and may confer high affinity for pathogen surfaces displaying repetitive arrays of oligosaccharide ligands. The AAA F-type CRD has a beta-barrel structure, with one three-stranded beta-sheet and one five-stranded sheet, connected by two disulphide bonds. One end of the barrel features five loop regions (here termed loops 1-5) that form a ring enclosing the site of fucose binding, which is a positively-charged hollow. A Ca2+ ion is bound within a sub-domain that largely lacks regular secondary structure, and stabilization of tertiary or quaternary structure by Ca2+ may enhance sugar binding activity in AAA and other F-type lectins. Within the positively-charged hollow, side chains from a triad of residues characteristic of fucose binding make hydrogen bonds to the 3-OH, 4-OH and ring oxygen of fucose. In AAA these residues - a histidine in loop 3 and two arginine residues in loop 4 - are present in the fucose binding sequence motif H...RGDCCGER, which is found in other F-type domains in the more general form H...RXDXXXX(R/K), where the gap between H and R is around 26 residues. Carbon atoms C1 and C2 of the fucose ring make van der Waals contacts with the adjacent disulphide-bonded cysteine residues in loop 4 (H...RGDCCGER), and the methyl group (C6) is accommodated in a hydrophobic pocket formed by side chains from loops 1 and 2. Other monosaccharide ligands (eg 3-O-methyl-D-galactose, 3-O-methyl-D-fucose) are bound through the same set of interactions. Oligosaccharide ligands, which include the blood group antigens H and Lea, interact with AAA primarily through the terminal fucose, but non-terminal saccharide residues also interact with a subset of the residues involved in fucose binding as well as additional residues from loop regions 1, 2 and 4. In the seven fucolectins from the Japanese eel (Anguilla japonica), various combinations of residue substitutions in loops 1 and 2 open up the methyl binding pocket and/or allow for polar interactions, which may adjust the oligosaccharide binding specificities of these lectins and tune them to recognize different pathogens. Non-eel F-type domains have a shorter loop 1, which may remove some interaction with oligosaccharide ligands and broaden oligosaccharide specificity.
F-type lectins in vertebrates
Proteins from fish and amphibians with a range of domain organizations contain F-type domains, most of which are likely to bind fucose based on the conservation of the key residues described above. The single-domain F-type lectins found in eels, which represent an early branch of teleost fish, are generally not found in modern teleosts, having been replaced by tandem two-domain proteins in species including striped bass (Morone saxatilis), zebrafish (Danio rerio), pufferfish (Tetraodon nigroviridans and Fugu rubripes) and stickleback (Gasterosteus aculeatus). In two-domain F-type lectins, both domains feature the fucose and Ca2+-binding residues, but the N-terminal domain lacks one inter-strand disulfide bond, and the the C-terminal domain lacks the adjacent cysteines in loop 4, which may affect sugar binding. In four-domain tandem F-type lectins, which are exclusive to trout species (eg Oncorhynchus mykiss), the first and second domains lack the adjacent cysteines, and the second domain has strikingly lost all three residues from the His/Arg/Arg triad, suggesting that fucose binding has been lost in this domain. FBP32 is a recently-characterized two-domain F-type lectin from striped bass, which is expressed abundantly by the liver, and to a minor extent in a small number of other tissues, is present in serum, and is upregulated by inflammatory challenge. Gene duplication has produced at least one FBP32 paralogue (the more widely expressed FBP32II) in striped bass, while multiple two-domain F-type lectins are present in other fish species as a result of independent gene duplications. FBP32 binds specifically, in a Ca2+-independent manner, to terminal fucose, but does not share the specificity of AAA for the H and Lea antigens. FBP32 is monomeric and the presence of two F-type domains within one polypeptide may serve to increase ligand binding affinity, as an alternative mechanism to the oligomerization seen in AAA. Single-domain eel fucolectins which terminate in a Cys residue may form covalent homodimers, reproducing the effect of tandem domains. Xenopus species possess proteins containing F-type domains in copies of 1 (X laevis X-epilectin), 2 (X tropicalis II-FBPL), 3 (X tropicalis III-FBPL) or 4 (X laevis II-FBPL), as well as 5 F-type domains in combination with a pentraxin domain (X laevis PXN-FBPL). Xenopus F-type domains sometimes lack the fucose-binding His residue and/or a small number of Ca2+-binding residues. The fate of F-type domains in higher vertebrates is unclear. Two genes encoding three-domain F-type proteins are predicted in the genome of the opossum (Monodelphis domestica), an early-branching mammal. The genes are present in tandem within a genomic region that is absent in placental mammals and in birds, but which in Xenopus tropicalis contains a gene encoding a three-domain F-type lectin ~40% identical to the opossum sequences. In humans this region would lie within chromosome 12, between the conserved genes Wbp11 and Foxj2. Some remnants of F-type lectin genes are evident in this region of the human genome, and translate into polypeptide sequences ~40% identical to the opossum sequences but interrupted by in-frame stop codons.
F-type lectins in invertebrates
The F-type domain is found in a range of invertebrate species, often within lineage-specific protein contexts. Sugar binding has been demonstrated in an F-type lectin from the Japanese horseshoe crab Tachypleus tridentatus. Like many invertebrates, the horseshoe crabs (which are related to arachnids) have an open circulatory system: hemolymph fluid, containing solutes and hemocytes, bathes the internal organs. The vast majority of horseshoe crab hemocytes are a single type of granular cell, which mounts an innate immune response upon recognition of bacterial lipopolysaccharide. Tachylectin-4 is an oligomeric hemocyte lectin consisting of a single F-type domain with an N-terminal extension, which exhibits Ca2+-dependent binding to terminal fucose. Tachylectin-4 recognizes bacterial lipopolysaccharide, probably through binding to fucose-like sugars such as colitose, a 3-deoxy form of fucose, and abequose, the D-isomer of colitose. Furrowed and CG9095 from Drosophila melanogaster contain single F-type domains within an architecture that also includes a C-type lectin-like domain and a number of complement control repeats. These proteins, which have homologues in other insects (eg bee and mosquito), are the only F-type proteins known to have transmembrane domains. Signalling by these receptors is influenced by O-fucosylation, but the F-type domains in the receptors do not contain complete His/Arg/Arg triads (the C-type lectin-like domain in these receptors is also unlikely to bind sugar). It has been stated that F-type domains are not present in the nematode worm Caenorhabditis elegans, but a highly divergent F-type domain, which is very unlikely to bind fucose, is present in the CG9095 orthologue, C54G4.4. The divergence of F-type domains between CG9095 proteins in different species is suggestive of a non-sugar-binding function of the F-type domain in these receptors. In the sea urchin Strongylocentrotus purpuratus, an invertebrate which is relatively closely related to chordate organisms, single F-type domains are present in complex protein architectures, as well as in a simple single-domain fucolectin. The complex proteins include CRL, which is involved in the complement system, and a protein containing a CCP and an EGF domain, both of which have very similar F-type domains, and a protein containing scavenger receptor Cys-rich, Kringle and other domains. F-type domains in the sea urchin are distinctive due to a number of insertions and the absence of the fucose-binding His residue and adjacent cysteines.
F-type lectins in bacteria
The F-type domain is found in a small number of diverse bacterial species, in a variety of architectural contexts. Bacterial F-type domains frequently conserve the fucose binding motif, but they do not include disulphide bonds. Sugar binding has been demonstrated in the F-type lectin SP2159 from the Gram-positive pathogen Streptococcus pneumoniae TIGR4. SP2159 is a virulence factor encoded in the fucose utilization operon, and features three tandem F-type domains preceded by an family 98 glycoside hydrolase (GH98) domain. The role of the protein as a virulence factor is not related to nutrition, but may be connected to the ability of the GH98 domain to cleave blood group A and B trisaccharides from glycoconjugates. The F-type domains in SP2159, all three of which are required for maximal affinity, bind to a distinct set of fucosylated oligosaccharides compared to AAA: ligands include the H, A and B antigens and, optimally, the Ley epitope, but not Lex or Lea. SP2159 binds to lung tissue in a fucose-dependent manner, suggesting a role for the protein in infection (the converse role to host defence, a function of the F-type domain in eukaryotes). A second F-type protein from Streptococcus pneumoniae TIGR4, SP0833, contains a partial F-type domain preceded by a lipase class 3 domain. Proteins equivalent to SP2159 and SP0833 are found in the related strain Streptococcus pneumoniae R6, whereas in Streptococcus mitis the F-type domain is present in conjunction with a thiol-activated cytolysin domain in human platelet aggregation factor. Among other Gram-positive bacteria, an F-type domain is found in a very large protein from Solibacter usitatus Ellin6076, combined with a number of bacterial immunoglobulin-like domains and FG-GAP repeats. Among Gram-negative purple bacteria, an F-type domain follows the structurally similar coagulation factor 5/8 domain in a protein from Saccharophagus degradans 2-40, an organism which degrades a range of polysaccharide substrates, whereas in Acidiphilium cryptum JF-5 an F-type domain is present at the C-terminus of the fucolectin tachylectin-4 pentraxin-1 protein, and in Gluconobacter oxydans 621H F-type domains are present in the proteins GOX0967 and GOX0982. Among bacteria of the planctomycetes phylum, an F-type domain is present at the centre of a large protein that is a periplamic component of a sugar transport system in Blastopirellula marina DSM 3645, and a more distantly related domain is present in a probable cytochrome c precursor (NP_868124) in Rhodopirellula baltica SH 1.
Structural similarity of the F-type fold to other protein domains
Structurally, the F-type CRD resembles domains from a number of proteins of unrelated function, which do not exhibit significant similarity to the F-type domain at the primary structure level. These include domains from human coagulation factors V and VIII, a fungal galactose oxidase, a bacterial sialidase, the human ubiquitin ligase APC10/DOC1, the DNA repair protein XRCC1, the cell surface receptor neuropilin, and the yeast catabolic enzyme allantoicase, pointing to an ancient origin for the F-type fold. A few of these proteins display conservation of some of a small number of key structural and ligand binding residues characteristic of F-type lectins. Sialidase and galactose oxidase have conserved the calcium binding motif between loops 1 and 2, and have bound cations, and sialidase binds monosaccharides by an analogous mechanism to AAA.
Domain organization in selected F-type domain-containing proteins
