Modeling the SDF-1/CXCR4 protein using advanced artificial intelligence and antagonist screening for Japanese anchovy

Abstract

SDF-1/CXCR4 chemokine signaling are indispensable for cell migration, especially the Primordial Germ Cell (PGC) migration towards the gonadal ridge during early development. We earlier found that this signaling is largely conserved in the Japanese anchovy (Engraulis japonicus, EJ), and a mere treatment of CXCR4 antagonist, AMD3100, leads to germ cell depletion and thereafter gonad sterilization. However, the effect of AMD3100 was limited. So, in this research, we scouted for CXCR4 antagonist with higher potency by employing advanced artificial intelligence deep learning-based computer simulations. Three potential candidates, AMD3465, WZ811, and LY2510924, were selected and in vivo validation was conducted using Japanese anchovy embryos. We found that seven transmembrane motif of EJ CXCR4a and EJ CXCR4b were extremely similar with human homolog while the CXCR4 chemokine receptor N terminal (PF12109, essential for SDF-1 binding) was missing in EJ CXCR4b. 3D protein analysis and cavity search predicted the cavity in EJ CXCR4a to be five times larger (6,307 ų) than that in EJ CXCR4b (1,241 ų). Docking analysis demonstrated lower binding energy of AMD3100 and AMD3465 to EJ CXCR4a (Vina score -9.6) and EJ CXCR4b (Vina score -8.8), respectively. Furthermore, we observed significant PGC mismigration in microinjected AMD3465 treated groups at 10, 100 and 1 × 10⁵ nM concentration in 48 h post fertilized embryos. The other three antagonists showed various degrees of PGC dispersion, but no significant effect compared to their solvent control at tested concentrations was observed. Cumulatively, our results suggests that AMD3645 might be a better candidate for abnormal PGC migration in Japanese anchovy and warrants further investigation.

Keywords: Alpha Fold; CB Dock; Colab Fold; Japanese anchovy; PGC; SDF-1/CXCR4.

FIGURE 1

Amino acid sequence analysis for SDF-1/CXCR4 signaling of Japanese anchovy. (A) Alignment of CXCR4 amino acid sequence and motif analysis. Sequences of CXCR4_N Motif are shown in orange color, and 7tm_1 Motif are shown in blue color. (B) Alignment of SDF-1 amino acid sequence and motif analysis. IL8 Motifs are shown in green, four cysteines and RFFESH domains shown in red and yellow, respectively.

FIGURE 2

Various protein 3D models and cavity analysis. Protein models are shown in cartoon and rainbow color (N to C, Blue to Red). (A) 3D model of CXCR4 protein. N-terminal and C-terminals are marked with N and C, respectively. 7 transmembrane α-helices are indicated from I to VII, respectively. The cavity size (ų) obtained at CB Dock2 of each CXCR4 is shown. (B) 3D model of SDF-1 protein. The RFFESH domain, which is required for CXCR4 binding, are additionally shown with the surface, and circled in red.

FIGURE 3

EJ CXCR4 and antagonist binding model of each combination. Vina score, and detail pictures of the binding sites are shown. CXCR4 are shown in cartoons colored with hydrophilicity, and antagonists are shown in the surface colored with blue. In the binding site details, helix numbers with cross-links are shown in green, and non-contacting helices are shown in red.

FIGURE 4

Hatchability (%) at 48 hpf in microinjection experiment. n = 31∼105, One-way ANOVA, Tukey test, *: p < 0.05.

FIGURE 5

Fluorescence observation at 48 hpf of antagonist microinjection experiment. White arrows indicate the location of PGC, that specifically visualized by egfp-nanos3 3′UTR mRNA. PGCs are dispersed throughout the body in the microinjected group.

FIGURE 6

Distance between the Farthest PGCs (DFP, mm) analysis at 48 hpf of antagonist microinjection experiment. A significant increase compare with control was found only for AMD 3465. n = 31∼105, One-way ANOVA, Tukey test, *: p < 0.05.

FIGURE 7

Hatchability (%) at 48 hpf in antagonist immersion experiment. No significant reduction was observed. n = 17, One-way ANOVA, Tukey test.

FIGURE 8

Fluorescence observation at 48hpf of antagonist immersion experiment. White arrows indicate the location of PGC. PGCs are dispersed throughout the body in the antagonist immersion groups.

FIGURE 9

Distance between the Farthest PGCs (DFP, mm) analysis at 48 hpf of antagonist immersion experiment. No significant difference was observed among the groups. n = 6, One-way ANOVA, Tukey test.