Pymol molecular graphics system version 1

Three–dimensional (3D) structures of endoglucanase (PDB codes: 3AMH and 3AMM, Cheng et al., 2011 (link)) and β-glucosidase (PDB codes: 1UZ1, Vincent et al., 2004 (link); 2CBU, Gloster, Madsen & Davies, 2006 (link); 2WBG, Aguilar-Moncayo et al., 2009 (link)) from T. maritima were accessed from the Protein Data Bank (PDB) (Berman et al., 2000 (link)). Structural features relevant to thermostability of the afore-mentioned enzymes were compared with that found in homologous structures from thermophilic and mesophilic organisms. PyMOL (The PyMOL Molecular Graphics System, Version 1.3, Schrödinger, LLC) was used for visualizing 3D structures.

The structure of the Pks13-TE domain was solved by molecular replacement method (MR) using E. coli EntF (PDB: 3tej) structure, as search model. A single MR solution was obtained using Phenix AutoMR (Adams et al., 2010 (link)) which was input into the AutoBuild wizard to generate the initial model for apo-Pks13-TE. The initial model was improved by further manual rebuilding in COOT (Emsley and Cowtan, 2004 (link)). The final model was obtained after iterative cycles of model building and Phenix refinement with simulated annealing yielding a 1.72 Å resolution apo-Pks13-TE model with R_cryst of 16.9% and an R_free of 20.1% with good stereochemistry (Table S2). The final refined apo-model has two chains, designated A and B, a fragment of additive PPG P400 and 471 water molecules in the asymmetric unit. The crystal structures Pks13-TE-inhibitor complex structures, as well as the D1607N mutant structures were refined with simulated annealing (start temperature 5000 K, Phenix). Inspection of electron density maps showed clear |F_o-F_c| positive difference density for the ligands which were fit into the density using Ligandfit routine in Phenix (Terwilliger et al., 2006 (link)). The ligand model and geometry restraint files were created in ELBOW BUILDER of the Phenix suite (Moriarty et al., 2009 (link)). Iterative cycles of model building and NCS-restrained maximum likelihood refinement with simulated annealing yielded high quality models for Pks13-TE-inhibitor complexes (Tables S2 and S3). In all of the structures > 98% of residues are placed in the favored region of the Ramachandran plot (MolProbity, Chen et al., 2010 (link)). Figures of the structures were made with UCSF Chimera package (Pettersen et al., 2004 (link)) and PyMOL Molecular Graphics System version 1.4.1 (Schrodinger, LLC). Structural analysis of Apo Pks13-TE for the identification of tunnels and channels was done using CAVER 3.0 PyMol plugin (Chovancova et al., 2012 (link)). Electrostatic surface potentials were calculated using APBS (Baker et al., 2001 (link)) and displayed using the APBS plugin for PyMOL. Atomic coordinates and structure factors for the reported crystal structures (Tables S2 and S3) have been deposited with the Protein Data Bank under accession codes: PDB: 5V3W (Apo Pks13-TE), PDB: 5V3X (Pks13-TE:TAM1), PDB: 5V3Y (Pks13-TE:TAM16), PDB: 5V3Z (Pks13-TE(D1607N)), PDB: 5V40 (Pks13-TE:TAM6), PDB: 5V41 (Pks13-TE:TAM5), and PDB: 5V42 (Pks13-TE:TAM3).

The study was done in accordance with regulations for studies on human subjects of the hospitals in Lausanne, Seoul, and Rancagua. In addition, approvals from the IRB of the Seoul National Hospital, and of the Ethics Commission of the Lausanne University Hospital, were obtained. Peripheral blood was obtained with informed consent from the patients and their parents and genomic DNA was extracted by routine methods. Fragmented genomic DNA was purified with AMPure XP beads and the quality of the fragmented DNA was assessed with an Agilent Bioanalyzer. Preparation of the exome enriched, barcoded sequencing libraries was perfomed using Agilent SureSelect Human All Exon v4 kit. The final libraries were quantified with a Qubit Fluorometer (Life Technologies) and the correct size distribution was validated with an Agilent Bioanalyzer. Libraries were sequenced on Illumina HiSeq 2000, generating 100 bp paired-end reads. Raw reads were aligned onto the hg19 reference genome with Novoalign (http://www.novocraft.com) and the data cleanup and variant calling were performed according to GATK Best Practices recommendations29 . Variant filtering was made with Annovar30 (link) and with own perl and bash scripts (available on request). Variants identified by this procedure were verified by direct PCR amplification of target exons from genomic DNA and bidirectional Sanger sequencing. A molecular model for the full-length human HSPA9 protein was generated with I-TASSER31 (link). Figures were generated in the PyMOL Molecular Graphics System, Version 1.7.4 (Schrödinger, LLC).