Figure 1.
Ribbon representation of the hAFP structure.
The positions of ten cysteine residues are labeled. Three conserved disulfide bonds (Cys4-Cys15, Cys32-Cys125, and Cys101-Cys117) are shown in yellow and two unique disulfides bonds (Cys69-Cys100 and Cys89-Cys111) in magenta. The Ca2+ ion is shown in green. This diagram was generated using the programs MOLSCRIPT and Raster3D [39], [40].
Table 1.
Data collection, phasing and refinement statistics.
Figure 2.
Sequence alignment of type II AFPs and most closely related fish-specific lectins.
Similar or identical amino acid residues are highlighted. Residues are colored according to their physicochemical properties. Numbering in the top ruler corresponds to residue numbering of hAFP. Numbers in brackets in the N- and C-terminal part of the alignment indicate how many amino acids have been omitted for the sake of the clarity. Type II AFPs–specific cysteines forming disulfide bonds are indicated by “*”. Residues forming the first Ca2+-binding site in hAFP are indicated with “1”. Residues homologous to residues of the second Ca2+-binding site are indicated as “2”.
Figure 3.
Phylogenetic and evolutionary analyses of type II AFPs.
(A) Maximum likelihood (ML) phylogenetic tree of type II AFPs and representatives of the most closely related fish skin mucus lectins. The values on the nodes represent the bootstrap support for the individual branches. (B) Phylogenetic tree of Teleostei reproduced based on Lundberg, John G. 2006. Teleostei. Version 18 August 2006 (temporary). http://tolweb.org/Teleostei/15054/2006.08.18 in The Tree of Life Web Project, http://tolweb.org/ and classification from the NCBI Taxonomy database (http://www.ncbi.nlm.nih.gov/Taxonomy/). (C) Evolutionary history of type II AFPs. Putative gene losses are indicated. Those taxa for which gene losses could not be stated due to insufficient genomic data are indicated with “?”.
Figure 4.
Maximum likelihood (ML) phylogenetic tree of type II AFPs, fish skin mucus lectins and other closely related lectins.
Values at the nodes represent branch support (probability values) derived with parametric approximate likelihood ratio test for branches (aLRT). A clade postulated in this work to be monophyletic and containing type AFPs is shown in blue. Other clades are shown in green. The position of the root of the tree is hypothetical. Gene duplications responsible for a rise of major clades are indicated.
Figure 5.
Stereo view of Ca2+-coordination sphere.
Black dashed lines indicate the coordination bonds. The Ca2+ ion (green) is coordinated with a water molecule (red), Gln92 Oε1, Asp94 Oδ2, Glu99 Oε1, Asn113 Oδ1 and Asp114 O and Oδ1. This diagram was generated using PyMOL [41].
Figure 6.
Ca2+-binding properties of hAFP and its mutants.
(A) 45CaCl2 overlay assay of WT-6H and its mutants. Lysozyme was used as negative control and β-lactoglobulin was used as positive control. PLMWM represents prestained low molecular weight marker. Both wild-type hAFP and its mutants can bind Ca2+ ions properly. (B) Proteolysis protection assay of WT-6H and its mutants. Endoprotease Glu-C was used to detect conformational changes of hAFP and its mutants as modulated by Ca2+ ions. In the absence of Ca2+ ions, hAFPs were subjected to Glu-C cleavage. Three lanes of each sample from left to right represent hAFP, and hAFP treated with Glu-C, respectively, in the presence, and in the absence, of Ca2+ ions.
Figure 7.
Thermal hysteresis activities and ice crystal morphologies of wild-type hAFP (WT-6H) and its mutants.
(A) Concentration dependent thermal hysteresis activity of (1) WT-6H (•), mutants Ala90Ser (□), Ala90His (▪), Ala91His (▴), and Ala91Thr (▵), and (2) WT-6H (•), mutants Thr95Ala (▴), Thr95Ile (▵), Gln103Ala (□), Gly109Asp (○), and His121Ala (▪), was measured as described under “Materials and Methods” using a Clifton nanoliter osmometer. Values showed represent means of triplicate measurements done on each single sample. (B) Ice crystal morphologies of hAFP and its mutants. Samples were: (1) buffer alone and wild-type hAFP; (2) the mutant that exhibited no effect on thermal hysteresis; (3) mutants that showed reduced thermal hysteresis activities; (4) mutants which retained the ability to modify the ice crystal with no detectable thermal hysteresis activities, (5) mutants that exhibited complete loss of antifreeze activity. The protein concentration used for each sample is also indicated.
Table 2.
Relative thermal hysteresis activities of hAFP mutants (0.4 mM) compared with that of the wild type hAFP WT-6H.
Figure 8.
A model for hAFP binding to [10-10] prism plane of ice.
(A) and (B) show that four ice-binding residues Asp94, Thr96, Thr98, and Glu99 form hydrogen bonds with water molecules of the ice lattice (highlighted in red) through respective side-chain oxygen atoms as indicated by yellow dotted lines. The Ca2+ ion is shown as a green sphere. The water molecule coordinating with the Ca2+ ion is shown as an orange sphere. The orientation of the ice lattice and the [10-10] prism plane are indicated. (C) Yellow spheres represent the water molecules constrained at the hAFP ice-binding face and the ice-water interface. (D) When hAFP binds to ice, constrained water molecules are released. This is considered as an entropy favorable process. (E) With the growth of the second layer of ice on the prism plane, the Ca2+-coordinating water molecule is incorporated into the ice lattice, thus the hAFP-ice interaction is further stabilized. This diagram was generated using PyMOL [41].