Green denotes wild-type residues. Some side chains are shown with the following colour coding:
Blue is known dominant-acting variant described in Motta et al. 2018 ( PMID: 30481304 ).
Red is de novo mutation from the current study.
Orange is a candidate (non-prioritised) variant seen in singleton exome.
Drag with left click to rotate and drag with right click to translate. Mouse-over gives names.
View generated in PyMol with the following commands:
select published, resi 91+119+193+247+284+286+287
select pubrec, resi 121+170+205+217+294
select schwannomatosis, resi 71+122+125+170+187+202+284+286+392+400+404
select candidate, resi 68+145+310+354
select alistair, resi 97+136+244+248
select phoser, resi 244+247
select phothr, resi 111
select photyr, resi 119+141+352
select phospho, phoser or phothr or photyr
color gray40, pubrec & name C*
color gray80, schwannomatosis & name C*
color tv_blue, published & name C*
color tv_orange, candidate & name C*
color tv_red, alistair & name C*
show sticks, published or candidate or alistair or phospho or pubrec or schwannomatosis
color tv_green, name CA+C
hide sticks, name C+HA+H+N+O
PSE converted to NGL viewer via PyMOL to NGL transpiler
Proteins are chains of L-amino acids linked via their backbone atoms (carboxyl and amine of the α carbon). This sequence of backbone atoms is represented by the (green) ribbon-like structure.
When the backbone atoms hydrogen-bond with each other in a consistent way secondary structure arises.
α-Helices are represented as spirals, while β-sheets appear as adjecent flat strands.
Some residues of interest are represented as ball and sticks like akin to an old chemistry set. In these representations the atoms are colour-coded based on the element. Carbon can be any colour, while nitrogens are in blue, oxygens are in red, hydrogens in white, sulfurs in yellow and phosphorous in orange. The latter is found as phosphate groups so will appear as a distinctive red-tipped orange tetrahedral end of a sidechain. Ions are represented by a variety of colours. But are generally represented as spheres of the correct atomic radius.
A recent study by Motta et al performed homology modelling on the LZTR1 protein and showed that the kelch (KT) domains resemble a six-bladed propeller like structure. It was further shown that autosomal dominant acting mutations typically lie in the upper surface of this model. In order to show where the autosomal dominant acting mutations identified in the present study lay, two structural homology models were made. The first was with I-TASSER using threading and ab initio construction with evolutionary constraints. To more closely replicate the work of Motta et al, a second model was made with Phyre2 in one-to-one threading mode. In this second model, a 1.4Å X-ray diffraction structure of Ta-TFP (a thiocyanate-forming protein) from Thlaspi arvense was used as a template (PDB 5A10). PhosphoSite Plus was queried for known phosphorylation sites and these residues were modified in the models by changing the relevant amino acid codes (SER>SEP, THR>TPO, TYR>PTR) and correcting/energy minimising the model with Rosetta Relax. The structures were examined with PyMol 2.2 (Schrödinger Inc.).
As the second model was better structured, we present results from that model (see image below) and show the position of the 5 de novo and one inherited AD-acting mutations. As shown, 4 of the 6 variants identified lie on the upper side of this propeller. The exceptions are p.R97L and p.N145I which both appear to be buried toward the side of the propeller. We note that many of the AD-associated variants lie in close proximity to, or are themselves, known phosphorylated residues. In particular, p.S244C disrupts a known phosphorylation site and is unlikely to disrupt hRAS binding due to the amino-acid change alone.
The hypothesis that disruption of phosphorylation may affect protein interaction may be supported by a comparison to KLHL3, a gene that encodes a homologous kelch-domain protein. Mutations in this gene have been linked to hypertension and electrolyte abnormalities. Although families with both dominant and recessive modes of inheritance were described, it was noted that dominant-acting mutations typically cluster in short segments within the propeller-like structure. A more detailed structural analysis went on show that phosphorylation of serine 433 impeded substrate-binding at the interface, whereas a protein harbouring the dominant, disease-associated p.S433N could not be repressed by this mechanism.
The relative positions of candidate LZTR1 heterozygous variants (including p.N145I) identified in singleton patients from the DDD study are documented in Table S3 and shown in the figure below (orange). Green denotes wild-type residues. Blue side chains show the known dominant-acting variants described in Motta et al.4 Red side chains highlight residues disrupted by de novo mutations identified as part of the current study.