Functional Screen of Wilson Disease ATP7B Variants Reveals Residual Transport Activities

Abstract

Wilson disease is a disorder of copper (Cu) homeostasis caused by the malfunction of Cu transporter ATP7B and associated Cu accumulation in tissues. The existence of over 700 disease-associated variants in the ATP7B gene and a broad spectrum of disease manifestations complicate the analysis of genotype-phenotype correlations and the development of better treatments for this disorder. To assist such studies, we screen 101 variants of ATP7B for expression and Cu transport activity in human fibroblasts lacking active ATP-dependent Cu transporters. The ClinVar database classified 59 of these as variants of uncertain significance or having conflicting pathogenicity classifications; six variants were not in the database. Thirty-three of the variants have been previously characterized by other assays. Only three variants (S657R, G1061E, and G1266R) resulted in the complete inactivation of Cu transport. The in silico analysis of these mutants was used to rationalize this drastic effect on ATP7B activity. The remaining ATP7B variants showed a range of Cu transport activities. Coexpression of variants with different properties yielded activity values different from the simple average. The advantages and limitations of this functional screen are discussed.

Keywords: ATP7B; Wilson disease; copper; mutations; variants.

Figure 1

(a) AlphaMissense pathogenicity heat map of ATP7B showing all reported amino acid alterations at the given position and (b) specific missense variants experimentally tested in this work. Variants later identified to completely abolish ATP7B Cu transport activity (S657R, G1061E, and G1266R) are labeled in bold with the asterisk below each panel.

Figure 2

Expression and activity of representative ATP7B variants in YST cells. Top panels: The expression of GFP-tagged ATP7B visualized using GFP fluorescence (scale bar: 50 μm). Bottom panels: The Cu transport activity of ATP7B was evaluated via the tyrosinase activation assay. The formation of black eumelanin pigment indicates active tyrosinase and therefore Cu transport by ATP7B (scale bar: 50 μm). (a) WT ATP7B and D1027A are the positive and negative controls, respectively. (b) Representative mutants illustrating a broad range of protein levels and activities under identical experimental conditions. Complete Cu transport inactivation was observed for G1266R, G1061E, and S675R variants. ⁣^nsp value > 0.05; ⁣^∗∗∗∗p value < 0.0001.

Figure 3

ATP7B variants that show a stronger reduction in Cu transport activity when compared to H1069Q mutant. Multiple images for WT ATP7B and each ATP7B variant were collected; pigment intensity was quantified in individual cells and plotted. H1069Q did not show a statistically significant difference in pigment intensity compared to WT ATP7B. ⁣^nsp value > 0.05; ⁣^∗∗∗p value < 0.001; ⁣^∗∗∗∗p value < 0.0001.

Figure 4

ATP7B sequence conservation in the regions where activity-abolishing variants S657R, G1061E, and G1266R occur. (a) Representative multiple sequence alignment (MSA) of ATP7B orthologs from model organisms and sequences for which experimental structures are available. The regions of primary sequence are centered around human ATP7B residues S657, G1061, and G1266. (b) WebLogo representation shown for the same sequence segments was generated from large scale MSA by ConservFold with human ATP7B as the input sequence and MMseqs2 as an algorithm to search cluster sequence data [52, 62]. (c) Variant effect prediction by AlphaMissense [57] for human ATP7B residues 657, 1061, and 1266 [57]. Crossed squares represent residues found in the wild-type sequence. ⁣^∗ indicates the positions of residues 657, 1061, and 1266.

Figure 5

Structural model of in silico S657R substitution. Human ATP7B AlphaFold 2 model colored by estimated evolutionary conservation (red-to-blue from most conserved to least conserved). S657R variant was virtually introduced by ChimeraX. Mutated R657 is shown as sticks colored in black. (a) Rotamer with the best evaluation criteria protrudes towards Cu entry site. (b) Visualization of six R657 rotamers with the least number of clashes; clashes are shown in yellow dashed line.

Figure 6

Structural model of in silico G1061E substitution. The AlphaFold 3 prediction of human wild-type ATP7B-Mg²⁺-ATP complex. ATP (dark purple) binding mode in the human ATP7B model closely fits the binding mode of adenylyl methylene-diphosphonate (AMP-PCP; light purple) in the experimental structure of the CopA PN-domain fragment [54] (1.163 and 2.483 Å RMSD between 185 pruned atom pairs and across all pairs, respectively). In silico mutated residue E1061 is shown as sticks colored in black G1061E virtual mutation was introduced into the model by ChimeraX, with the best-fitting rotamer shown.

Figure 7

Structural model of the in silico G1266R substitution introduced into Xenopus tropicalis ATP7B structure (PDB ID 7SI3 [66]). Catalytically important motifs DKTG and TGE in the P- and A-domain, respectively, are labeled with X. tropicalis human sequence numbering. The structure mimics the so-called E2-P_i catalytic state by inhibiting the enzyme with aluminum fluoride complex (AlF_x). The virtual R1266 mutation (human numbering) introduced in ChimeraX is shown as sticks colored in black.

Figure 8

Coexpression of variants with different functional properties. (a) GFP fluorescence (scale bar: 50 μm) and tyrosinase assay (scale bar: 50 μm) of individual and coexpressed plasmids. (b) Immunoblotting of individual and coexpressed plasmids. (c) Quantification of tyrosinase assay signals for WT, R919W, L1299F, and R919W coexpressed with L1299F.