Wilson disease missense mutations in ATP7B affect metal-binding domain structural dynamics
- 328 Downloads
Wilson disease (WD) is caused by mutations in the gene for ATP7B, a copper transport protein that regulates copper levels in cells. A large number of missense mutations have been reported to cause WD but genotype–phenotype correlations are not yet established. Since genetic screening for WD may become reality in the future, it is important to know how individual mutations affect ATP7B function, with the ultimate goal to predict pathophysiology of the disease. To begin to assess mechanisms of dysfunction, we investigated four proposed WD-causing missense mutations in metal-binding domains 5 and 6 of ATP7B. Three of the four variants showed reduced ATP7B copper transport ability in a traditional yeast assay. To probe mutation-induced structural dynamic effects at the atomic level, molecular dynamics simulations (1.5 μs simulation time for each variant) were employed. Upon comparing individual metal-binding domains with and without mutations, we identified distinct differences in structural dynamics via root-mean square fluctuation and secondary structure content analyses. Most mutations introduced distant effects resulting in increased dynamics in the copper-binding loop. Taken together, mutation-induced long-range alterations in structural dynamics provide a rationale for reduced copper transport ability.
KeywordsWilson disease Missense mutation Copper transport Molecular dynamics Yeast assay
Wilson disease (WD) is a rare autosomal recessively-inherited disorder of copper (Cu) metabolism caused by mutations in the ATP7B gene which codes for a transmembrane Cu transporting ATPase (Park et al. 1991; Mohr and Weiss 2019). The genetics of WD is complex with more than 450 disease-causing mutations confirmed (Mohr and Weiss 2019; Harada et al. 2001). Under physiological conditions, ATP7B has dual function, mediating excretion of Cu into the bile and Cu transport into the Golgi network for loading of Cu-dependent enzymes such as ceruloplasmin (Lo and Bandmann 2017). In patients, impaired Cu excretion leads to increased Cu accumulations, reaching toxic levels, primarily in the liver and eventually in other organs, particularly in the brain. The consequences are a wide range of symptoms such as progressive liver disease and liver failure, neurological syndrome as well as psychiatric symptoms (Mohr and Weiss 2019; Nagano et al. 1998; Rodriguez-Castro et al. 2015). The wide spectrum of phenotypic variations in WD patients often makes it hard to make an accurate and early clinical diagnosis (Jang et al. 2017; Dong et al. 2016; Braiterman et al. 2014).
WD is a rare disorder with an estimated prevalence of symptomatic disease of 1 in 30,000 and a heterozygous ATP7B mutation carrier frequency of 1 in 90 (Mohr and Weiss 2019). However, these numbers have been questioned as they are based on assumptions; it has been postulated that the number of people with disease and carriers are much higher (Mohr and Weiss 2019; Lo and Bandmann 2017). Because early diagnosis is essential to introduce treatment that will prevent future complications, there is an interest in genetic screening of infants for WD (Lo and Bandmann 2017). A number of pilot studies have investigated the possibility of large-scale genetic screening programs for WD (Lo and Bandmann 2017) and, despite technical hurdles, such programs may be reality in the future. Clearly, mutations causing WD should affect ATP7B function, but how this takes place at the molecular-mechanistic level is not known for most missense mutations. Whereas mutations that prematurely truncate the ATP7B polypeptide will intuitively abolish function, it is harder to depict how point mutations in a 1400-residue protein can cause dysfunction. To make best use of genetic screening efforts, it would be required to understand of how different mutations contribute to disease onset, progression and severity. Moreover, many functional aspects of ATP7B are not known and, therefore, laboratory studies of mutated ATP7B variants may also reveal fundamental insights on general mechanisms.
Despite ‘only’ regulatory roles proposed, and the redundancy of domains, it is interesting to note that at least 33 WD causing missense mutations are found in the MBDs (excluding linker regions) (Arioz et al. 2017). The MBDs closest to the membrane-spanning part of ATP7B, MBD5 and MBD6, appear to be a hotspot containing 23 of the 33 missense mutations in the MBDs (Arioz et al. 2017; Gourdon et al. 2012). Only two of the 33 mutations are localized in the Cu-binding motifs. One may thus speculate that most WD-causing mutations in the MBDs will affect MBD structural integrity, MBD–MBD interactions and/or interactions with other ATP7B domains. With the aim to explain dysfunction on a molecular level for WD missense mutations located in MBD5 and MBD6 at different positions within the ferredoxin fold, we here investigated Cu transport ability of four mutated ATP7B proteins. The variants (L492S in MBD5; A604P, R616W, and G626A in MBD6) have all been proposed to cause WD based on genetic analysis of WD patients in different families (Ljubic et al. 2016; Caca et al. 2001; Tuan Pham et al. 2017; Zong and Kong 2015), although patients homozygote for these mutations have not been discussed in the literature. The L492S mutation is situated in the first ß-strand of the ferredoxin-fold, the A604P mutation in the third ß-strand, and the R616W and G626A mutations in different ends of the second -helix of the ferredoxin fold (Fig. 1b). Thus, none of the mutations are near the Cu-binding site in the MBD. We found that three of the variants demonstrated reduced Cu transport efficiency in a yeast assay as compared to the wild type protein. Upon structural dynamics analysis using MD simulations, we found identifiable differences for the mutated domains as compared to the wild-type counterparts, which may be linked to the expected dysfunction of these ATP7B variants in vivo.
Materials and methods
Yeast strains and plasmid construction
High-stress resistance yeast S. cerevisiae CEN.PK 113-11C (MATa SUC2 MAL2-8 URA3-52 HIS3-∆1; provided by Dr. P. Kötter, Institute of Microbiology, Johann Wolfgang Goethe-University, Frankfurt, German) was used as a reference strain. The yeast strains were cultivated in liquid YEPD/YPD (Yeast Extract Peptone dextrose media) and grown at 30 °C. The yeast strain bearing CCC2 and ATX1 deletions was used in this study. For the expression of ATP7B and ATOX1, plasmids containing human ATP7B (p426GPD-ATP7B) and ATOX1 (p423GPD-ATOX1) were transformed to the yeast strain with CCC2 and ATX1 knockout (Δccc2Δatx1). The detailed construction of the yeast strain and plasmids p426GPD-ATP7B and p423GPD-ATOX1 were reported previously (Ponnandai Shanmugavel et al. 2017). The disease-causing ATP7B mutants namely L492S, A604P, R616 W, and G626A were constructed. The mutations were introduced into the full-length ATP7B by quickchange site-directed mutagenesis kit using the previously constructed p426GPD-ATP7B plasmid (Ponnandai Shanmugavel et al. 2017) as a template. All constructed plasmids were verified by sequencing (Eurofins). The DNA primers used for the plasmid construction are listed in the Supplementary Table S2. All plasmids were transformed into yeast cells by the standard lithium acetate method (Gietz and Woods 2006).
Yeast complementation assay
The Cu transport activity of the yeast strains was evaluated using growth curve analysis in iron limited medium. A single yeast colony from the plates was inoculated in Iron limited medium (SD medium containing 1.7 g L−1 yeast nitrogen base without Fe and Cu, 50 mM MES buffer pH 6.1, 20 g L−1 glucose, 5 g L−1 ammonium sulfate, complete supplement mixture CSM–Ura–His, 1 mM ferrozine (Fe chelator), 1 µM CuSO4 and 100 µM FeSO4) and incubated overnight at 30 °C and 200 rpm (Morin et al. 2009). Yeast cells from this culture were washed with ice cold deionized water and cultivated in the fresh iron-limited medium at an initial cell density of OD600 = 0.1. The growth of the cells was monitored and growth rates were determined from the exponential phase. All yeast growth experiments were carried out at identical conditions and in six biological replicates.
Protein extraction and western blotting
All yeast strains were grown in iron limited medium for 30 h at 30 °C. Cells were spun down by centrifugation at 2000×g at 4 °C for 10 min. Cells pellets were washed twice with ice-cold water and resuspended in lysis buffer (50 mM HEPES pH 7.5, 150 mM NaCl, 2.5 mM EDTA, 1% v/v Triton X100 and freshly added protease inhibitor). After disruption with glass beads, membranes were collected by centrifugation at 18,000×g (4 °C, 30 min). Samples were re-suspended in SDS loading buffer (0.5 M Tris–HCl, pH 6.8, 10% SDS, 0.5% (w/v) bromophenol blue, 87% glycerol, 100 mM DTT) and 50 mg of membranes were loaded on a 4–12% Bis–Tris gel (Invitrogen) and blotted onto PVDF membranes. ATP7B and Atox1 were detected with monoclonal rabbit ATP7B and Atox1 antibodies, respectively (Abcam, 1:1000 dilution), upon incubation overnight at 4 °C. Next, blots were incubated with horse radish peroxidase conjugated anti-rabbit IgG reagent (Thermo Scientific Pierce) for 15 min, 4 °C. Bands were detected by PierceTM Fast Western Blot Kits, SuperSignalTM West Femto, Rabbit (ThermoScientific Pierce) and visualized with a BioRad ChemiDoc XRSimage analyzer.
Cell localization of yeast-expressed ATP7B
Yeast cells were cultured to mid-log phase in the iron-limited medium. Harvested yeast cells were fixed in 5 ml of 50 mM KPO4 (pH 6.5), 1 mM MgCl2 and 4% formaldehyde for 2 h. After fixation, the cells were washed two times in 5 ml of PM buffer (100 mM KPO4 pH 7.5, 1 mM MgCl2) and followed by resuspension in PMST buffer (100 mM KPO4 pH 7.5, 1 mM MgCl2, 1 M sorbitol, 0.1% Triton X-100) to a final OD600 of 10. 100 µl yeast cells were incubated for 20 min in 0.6 µl of β-mercaptoethanol and 1 mg/ml zymolyase (Zymo Research). The spheroplasted cells were washed with PMST buffer and attached to polylysine-coated coverslips. Adherent cells were blocked in PMST–BSA buffer (0.5% BSA in PMST buffer) for 30 min. Next, the adherent cells were incubated overnight at 4 °C with primary antibody (1:500 rabbit monoclonal ATP7B antibody, Abcam) diluted in PMST–BSA buffer. After incubation, the cells were washed three times with PMST–BSA buffer and incubated with secondary antibody (1:1000 anti-rabbit Alexa 488, Abcam) for 3 h at room temperature, and with 0.4 mg/ml DAPI (staining nuclei) for 5 min. Cells were mounted in Vectashield mounting medium (Vector Laboratories). Images were acquired using a Leica DM 2000 inverted microscope and processed with the Leica application suite (LAS-AF lite) software.
In silico model building and molecular dynamics
All in silico systems were prepared based on the solution structure 2EW9 (a two-domain structure of the human ATP7B MBD56) (Achila et al. 2006). The first structure in the NMR structural ensemble were extracted as the starting coordinate for molecular dynamics (MD) simulations. The mutants L492S, A604P, G626A and R616W were in silico mutated based on the extracted wild-type structure (WT). Six single-domain systems were constructed by removing the linker between MBD5 and MBD 6. They are two wild-type proteins, i.e., MBD5WT and MBD6 WT, and four mutant protein systems, i.e., MBD5L492S, MBD6A604P, MBD6G626A, and MBD6R616W. The single-domain structure (wild-type system MBD5WT as the example) was solvated in a 60 Å rhombic dodecahedron water box. Na+Cl− ionic pairs (0.15 M) were added to neutralize the system and mimic the physiological conditions. The system was initially minimized for 10,000 steps under a series of position restraints and constraints, then heated to 300 K and equilibrated under NVT condition (constant volume and temperature) for 5 ns (ns) using the CHARMM program (version 42a1) (Brooks et al. 2009). The production MD simulation was carried out at NPT (constant pressure and temperature) condition, i.e., the pressure at 1 atm and the temperature at 340 K (Nilsson et al. 2013). The simulation lasted for 500 ns on each system using the NAMD program (version 2.12) and three replicas were performed (Phillips et al. 2005). MD snapshots were saved every 20 picoseconds along the trajectories for further analysis. Other details about the MD setting-up can be found elsewhere (Kumar et al. 2017). The trajectory analysis on all systems were done with CHARMM routines. All statistical figures were plotted by MATLAB (Version 2018a MathWorks, Inc.) and structural figures were generated with the PyMOL graphic software (Version 2.2 Schrödinger, LLC.).
Probing ATP7B-mediated Cu-transport in a yeast assay
Localization of ATP7B in yeast cells
The diminished Cu transport may be due to direct effects on Cu transport, reduced protein expression, or protein mis-localization in the yeast. Western blot analysis was used to confirm that the expression levels of all ATP7B variants in yeast were similar to that of wild type ATP7B (Fig. S2). To assess localization of ATP7B variants in yeast cells, we used immunofluorescence. All ATP7B mutants except L492S showed cellular localization patterns similar to that of wild type ATP7B (Fig. 2b) and involved an evenly distribution throughout the yeast cell. However, in yeast expressing the L492S variant, the protein was mainly localized in a confined region in the yeast cells such that it appeared in a scattered pattern concentrated in the endoplasmic reticulum (ER) surrounding the nucleus (Fig. 2b). Thus, for this particular variant, the lack of Cu transport may be mainly due to improper biosynthesis, thereby trapping in the ER.
Molecular dynamics (MD) simulations of individual MBDs
In earlier work, we employed MD simulations to analyze structural dynamics of individual MBDs of ATP7B, as well as two-domain constructs, with a major focus on the effects of Cu loading and domain–domain correlations (Rodriguez-Granillo et al. 2009, 2010). Here, we turned to WD-causing missense mutations in MBDs. Four variants were selected because these mutations’ specific positions in the MBD ferredoxin-like fold did not provide obvious reasons for dysfunction. First we tested the four ATP7B variants (L492S in MBD5 and A604P, R616W, and G626A in MBD6) in a straight-forward Cu transport assay in yeast where yeast growth in the exponential phase directly report on Cu transport (Ponnandai Shanmugavel et al. 2017). Cu transport is the most common way to probe WD mutations, and most studied WD mutations in the MBDs, e.g., G85V (Huster et al. 2012; Li et al. 2019), S406A (Huster et al. 2012), Y532H (Parisi et al. 2018; Hsi and Cox 2004), V456L (Huster et al. 2012), P410L (Lee et al. 2011), A629P (Payne and Gitlin 1998; Lutsenko et al. 2007), S637L (Donsante et al. 2007) and S653H (Kim et al. 2003), displayed reduced or lack of Cu transport in cell-, or vesicle-based, Cu transport assays. For our selected WD mutations, three of the variants showed reduced Cu transport in our yeast assay, but one variant (G626A) had wild-type Cu transport capacity.
To probe underlying structural dynamics effects in the variants, and search for (perhaps more subtle) alterations in the G626A variant, we performed MD simulations of individual MBDs with mutations introduced. The MD simulations demonstrated clearly that the effects of individual mutations extended throughout the domains and, in all cases, affected the dynamics of the Cu-binding loop, although this loop is distant from the site of mutation in each case. The structural dynamics in the L492S domain was most increased among the variants; in fact, the whole polypeptide was affected here. The high domain dynamics due to this mutation may thus explain why this variant (as a full length protein), did not fold properly in the yeast cells and became retained in the ER. The observed altered structural dynamics in the Cu-binding loops of A604P and R616W variants may explain these variants’ reduced Cu transport activity in the yeast assay. Increased Cu-loop dynamics may affect efficiency of Cu delivery to MBD6, from other MBDs or Atox1, as well as Cu discharge from MBD6 to the Cu entry site in the ATP7B membrane-spanning part. For the G626A mutation, the in silico data revealed little effects on structural dynamics (if anything, the fold became more ordered), which agrees with this variant being able to transport Cu as well as wild-type ATP7B in the yeast assay. The reason that the G626A mutation is linked to WD in patients cannot be established from our work but may be due to modulation of other, more subtle, aspects of ATP7B function.
There are no biophysical-molecular studies of WD-causing mutations in the MBDs, except for a recent study with introduced mutations in MBD4 (Kumar et al. 2017). However, there are limited data on functional properties of the here selected WD mutations when introduced in full-length ATP7B. In similarity to our observation in the yeast assay, in vitro vesicle-reconstitution studies reported low Cu transport activity for L492S and R616W variants but only somewhat decreased such activity for G626A ATP7B (A604P has not been investigated before our work) (Huster et al. 2012). In contrast to the Cu transport effects, ATP7B interactions with the regulatory protein COMMD1 appeared normal for L492S and R616W variants, but was found decreased for G616A and A604P ATP7B variants (de Bie et al. 2007). Notably, mutations in the ATP7B homolog, ATP7A, corresponding to the here studied A604P, R616W, and G626A mutations, have been linked to Menke’s disease (Gourdon et al. 2012), another Cu metabolism disease which involves mutations in ATP7A. Moreover, the I492T mutation in ATP7A (Tumer 2013), found at a position that corresponds to L492 in ATP7B, and F281N in ATP7A (Caicedo-Herrera et al. 2018) (at a similar position to L492 in the ferredoxin fold, but in MBD3) were also reported to cause Menke’s disease. In analogy to the Cu transport results for G626A in ATP7B, a similar mutation in ATP7A was reported to exhibit normal Cu transport to the Golgi, but the ATP7A variant showed trafficking problems at high Cu levels (Kim et al. 2003). However, the trafficking ability of G626A ATP7B has been reported to be wild-type like (Braiterman et al. 2014). Therefore, the mechanism of dysfunction due to G626A appears to involve subtle changes in interactions with regulatory proteins, such as COMMD1 (de Bie et al. 2007).
Taken together, the yeast assay provides a platform to screen putative disease-causing mutations in ATP7B for Cu transport ability. From this data (L492S showing no Cu transport, due to ER trapping, but G626A exhibiting wild-type Cu transport) one may predict that the L492S mutation will confer more severe WD than the G626A mutation, although clearly many additional factors are of importance. However, the yeast assay may miss to identify WD mutations (e.g., G626A) that promote more subtle dysfunction and/or affect other aspects of ATP7B function. To reveal molecular-mechanistic reasons for mutation-induced ATP7B dysfunction, MD simulations are helpful as it can probe mutation-induced local structural perturbations with atomic resolution. To ultimately predict disease progression in WD patients, much more basic knowledge of molecular mechanisms through which mutations cause ATP7B dysfunction is required.
Open access funding provided by Chalmers University of Technology. This work is supported by the Swedish Research Council, the Knut and Alice Wallenberg foundation and the International Postdoc Grant funded by the Swedish Research Council (to Y.L).
Compliance with ethical standards
Conflict of interest
The authors declare no conflict of interest.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.