The product of C9orf72, a gene strongly implicated in neurodegeneration, is structurally related to DENN Rab-GEFs.
Levine TP, Daniels RD, Gatta AT, Wong LH, Hayes MJ. · 2013
Read The product of C9orf72, a gene strongly implicated in neurodegeneration, is structurally related to DENN Rab-GEFs. by Levine TP, Daniels RD, Gatta AT, Wong LH, Hayes MJ. (2013) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc3570213.txt.
Introduction
Timothy Levine and colleagues tackled an unknown gene in Bioinformatics in 2013. Changes in that gene, C9orf72, are the most common known cause in families with frontotemporal dementia or amyotrophic lateral sclerosis, two nerve diseases that overlap in symptoms and cell damage. Yet no one knew what the gene's protein product actually does inside a cell.
Rather than run bench experiments, the group ran sensitive computer searches comparing the protein's shape and sequence against known families. The match pointed to the DENN family, proteins that switch on Rab-GTPases, the traffic controllers that move cargo between cell compartments. On that basis the authors propose C9orf72 likely helps govern membrane traffic, and suggest renaming it DENNL72. The paper gives later lab work a concrete first hypothesis to test.
Catalog record
| CREATOR | Levine TP, Daniels RD, Gatta AT, Wong LH, Hayes MJ. |
|---|---|
| YEAR | 2013 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 24 KB |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
The product of C9orf72, a gene strongly implicated in neurodegeneration, is structurally related to DENN Rab-GEFs By Timothy P. Levine, Rachel D. Daniels, Alberto T. Gatta, Louise H. Wong, Matthew J. Hayes Published in: Bioinformatics DOI: 10.1093/bioinformatics/bts725 ABSTRACT Motivation: Fronto-temporal dementia (FTD) and amyotrophic lateral sclerosis (ALS, also called motor neuron disease, MND) are severe neurodegenerative diseases that show considerable overlap at the clinical and cellular level. The most common single mutation in families with FTD or ALS has recently been mapped to a non-coding repeat expansion in the uncharacterized gene C9ORF72. Although a plausible mechanism for disease is that aberrant C9ORF72 mRNA poisons splicing, it is important to determine the cellular function of C9ORF72, about which nothing is known. Results: Sensitive homology searches showed that C9ORF72 is a full-length distant homologue of proteins related to Differentially Expressed in Normal and Neoplasia (DENN), which is a GDP/GTP exchange factor (GEF) that activates Rab-GTPases. Our results suggest that C9ORF72 is likely to regulate membrane traffic in conjunction with Rab-GTPase switches, and we propose to name the gene and its product DENN-like 72 (DENNL72). Supplementary information: Supplementary data are available at Bioinformatics online. Contact: [email protected] 1 INTRODUCTION Fronto-temporal dementia (FTD) and amyotrophic lateral sclerosis (ALS) are major neurodegenerative diseases that occur in sporadic and familial forms, which show overlap in clinical presentation and cytopathology (Dion et al., 2009). The commonest cause of familial ALS, accounting for 20–40%, maps to a mutation in the gene C9ORF72 at 9p21 (DeJesus-Hernandez et al., 2011; Renton et al., 2011). Mutation of C9ORF72 is also common in patients with familial FTD (10–30%), sporadic FTD (2–10%) and Alzheimer-like dementia (0–4%) (Morris et al., 2012). A single mutation is found: a massively expanded GGGGCC hexanucleotide repeat in an intronic portion of the 5′ untranslated region. In the general population, alleles have a median of two repeats (range: 0–22), which are expanded in pathogenic alleles to >1000 repeats. The pathogenetic mechanism is unknown, but other neurological syndromes arise from non-coding expansions. Pathology could occur via poisoning of splicing factors, in particular TDP-43, with non-coding repeats inducing toxic RNA foci that sequester splicing factors and other nuclear proteins (Todd and Paulson, 2010), as in Fragile X Tremor Ataxia Syndrome (FXTAS) and myotonic dystrophy. TDP-43 deposits and aberrant cytoplasmic mRNA foci in C9ORF72 mutant cells both support a role for splicing (DeJesus-Hernandez et al., 2011; Renton et al., 2011). However, these are not universal findings (Simon-Sanchez et al., 2012), and alternative mechanisms should be considered. One possibility is that the expansion reduces translation, as in Fragile X, consistent with reduced levels of C9ORF72 protein in mutant cells (DeJesus-Hernandez et al., 2011; Renton et al., 2011). Despite its clinical relevance, the normal function of the C9ORF72 protein has remained obscure. A powerful way to determine protein function is to find homologies to proteins of known function. The sensitivity of the basic local alignment search tool (BLAST) has been increased in several ways. Position-specific iterative (PSI)-BLAST creates a profile of key residues from BLAST, and then searches with increasingly large profiles to converge on a final stable alignment (Altschul et al., 1997). Sensitivity is improved by profile–profile comparison, using profiles for every database entry (Soding, 2005). Because structure diverges far slower than primary sequence (Soding et al., 2005), a further enhancement is to include secondary structural prediction in all profiles. The tool HHpred allocates 15% of the profile weighting at each residue to predicted structure: helix sheet or loop. Importantly, HHpred searches targets not only in the Protein Database (PDB) of solved structures, other databases of domains, but also through proteome-wide libraries of model eukaryotes. Its enhanced profile–profile searches can find homologues that lack significant amino acid conservation even where no structure is solved, and no domain recognized (Hayes et al., 2011). DENN domain proteins (standing for ‘differentially expressed in normal and neoplastic cells’) are highly conserved Rab-GEFs, with large numbers of homologues in some species: humans have 16 in 7 subgroups (Marat et al., 2011; Yoshimura et al., 2010). Using PSI-BLAST, several DENN-like families have been identified: DENND6 (FAM116), AVL9, Fam45A, LCHN, Afi1p and Anr2p, some of which have GDP/GTP exchange factor (GEF) activity (Linford et al., 2012). All DENN-like proteins contain a domain of 400–500 aa (Marat et al., 2011), the structure of which consists of two lobes that associate intimately with each other and Rab. The tripartite ‘domains’ of u-DENN, c-DENN and d-DENN previously identified in silico had no bearing on the structure, and we do not use this nomenclature (Wu et al., 2011). The C-terminal domain makes 16 of 21 Rab contacts, and this lobe is the active GEF. The N-terminal lobe of DENNs adopts a longin domain (LD) ββαβββαα fold, also found in SNAREs, small and medium coat subunits and three types of GEF: three components of the TRAPP complex, both of MON1/CCZ1 and fuzzy (FUZ) (Brooks and Wallingford, 2012; Gerondopoulos et al., 2012; Kinch and Grishin, 2006). LDs, particularly as dimers, are highly conserved platforms for GTPases, deriving from their circularly permuted ancestral roadblock domain, found in prokaryotes and eukaryotes (Miertzschke et al., 2011). Recently folliculin (FLCN), a tumour suppressor mutated in Birt–Hogg–Dubé syndrome leading to renal cancers, which had no homologies according to PSI-BLAST, was identified as a structural and functional homologue of DENN (Nookala et al., 2012; Wu et al., 2011). Thus, DENN is the founder of a family that is wider than can be appreciated using PSI-BLAST. Here, we show that C9ORF72 is one of several previously unrecognized full-length homologues of DENN proteins, suggesting that C9ORF72 is a member of the DENN-like superfamily, and hence a regulator of membrane traffic. 2 METHODS Online tools used were PSI-BLAST (NCBI), ClustalW2 (EBI), PHYML (T-rex), HHalign, HHpred and CLANS (MPG, Tuebingen). Default settings were used, except in HHpred, which was set to eight iterations, and where query-template alignments were realigned by a Maximum Accuracy algorithm. To optimize HHpred specifically for queries with C9ORF72 and FLCN, parameters were varied for individual searches (see Supplementary Table S2). Protein structure was visualized in QtMG (CCP4). 3 RESULTS 3.1 Domain structure of C9ORF72 PSI-BLAST seeded with human C9ORF72 converged on 103 sequences including some in protists, indicating that C9ORF72 was present in the last common eukaryotic ancestor (Supplementary Fig. S1A). Insects, plants and most fungi have no C9ORF72, most other species have one, but there are highly expanded sub-families in Entamoeba (7) and Trichomonas (10). All C9ORF72s have 400–750 aa, and there are no partial homologues. The most conserved residues are distributed throughout the protein (Supplementary Figs S1 and S2). Such conservation in a single block implies that C9ORF72 acts as one functional unit. Because the primary sequence of C9ORF72 shows no homologies, we used six independent structure prediction tools to model C9ORF72 by comparison with the PDB library. Four engines suggested multiple specific homologies, with the only domains that appeared in the results of more than one tool being: M16 peptidases (4), cytochrome bc1 (2), glycohydrolases (2) and DENN (5) (Supplementary Table S1). DENN is represented only twice in the database (n = 85, 68 and 476 for the other folds), but was identified by the greatest number of the tools, suggesting a similarity of C9ORF72 to DENN. 3.2 C9ORF72 is