Abstract
Summary
The molecular functions of TMEM132 genes remain poorly understood and under-investigated despite their mutations associated with non-syndromic hearing loss, panic disorder and cancer. Here we show the full domain architecture of human TMEM132 family proteins solved using in-depth sequence and structural analysis. We reveal them to be five previously unappreciated cell adhesion molecules whose domain architecture has an early holozoan origin prior to the emergence of choanoflagellates and metazoa. The extra-cellular portions of TMEM132 proteins contain five conserved domains including three tandem immunoglobulin domains, and a cohesin domain homologue, the first such domain found in animals. These findings strongly predict a cellular adhesion function for TMEM132 family, connecting the extracellular medium with the intracellular actin cytoskeleton.
Supplementary information
Supplementary data are available at Bioinformatics online.
1 Introduction
Many genes remain experimentally under-investigated not because they are functionally less important but because their discovery came relatively late (Pandey et al., 2014). Our ignorance of aspects of basic biology and disease thus is perpetuated by the serendipitous order by which genes were first characterized. The need to experimentally determine proteins’ normal molecular functions, and their molecular dysfunction in disease, becomes more critical when sequence variants within functionally enigmatic genes are robustly associated with Mendelian or complex disease, or with cancer progression. Determining the molecular functions of such poorly characterized genes is all the more difficult when their protein sequences lack recognizable domains, because these otherwise can reliably provide structural and functional information through homology-based inference. Here, we shed much light on the previously unknown domain structure and functions of the 5 proteins of the human TMEM132 family (TMEM132A, B, C, D and E).
These are genes in which variants are enriched for individuals with hearing loss, panic disorder or cancer. A homozygous missense mutation in human TMEM132E (Arg420Gln) was confirmed using a zebrafish model to cause autosomal-recessive nonsyndromic hearing loss (Li et al., 2015). Common variants within the TMEM132E gene are associated with insomnia symptoms (Lane et al., 2017); common and rare variants near TMEM132D gene are robustly associated with panic disorder (Erhardt et al., 2011, 2012; Hodgson et al., 2016; Howe et al., 2016; Inoue et al., 2015; Quast et al., 2012; Shimada-Sugimoto et al., 2016; Wang et al., 2016); and variants near TMEM132B are associated with excessive daytime sleepiness (Lane et al., 2017). In healthy individuals, some of the TMEM132D non-coding variants exhibit higher anxiety scores and larger volumetric estimates of the amygdala and hippocampus, key neural structures associated with fear and anxiety (Haaker et al., 2014). Furthermore, in cattle the TMEM132D locus appears to have undergone a selective sweep during domestication (Qanbari et al., 2014), and in the mouse, anterior cingulate cortex TMEM132D expression correlates with anxiety-related behaviour (Erhardt et al., 2011). Finally, mutations in TMEM132D are unusually frequent in small-cell lung cancer (Iwakawa et al., 2015; Peifer et al., 2012; Rudin et al., 2012) and in pancreatic cancer (Forbes et al., 2015). Disease mutations near, or functions of, TMEM132A or C have yet to be identified, although TMEM132A is thought to promote neuronal cell survival by regulating stress-related genes (Oh-hashi et al., 2003, 2006, 2010, 2012, 2015). Functions of the single fruit fly (CG14446) or nematode (Y71H2AM.10) orthologous genes are yet to be described.
All five human paralogues encode cell-surface molecules expressed in the brain (Oh-Hashi, 2010; Nomoto et al., 2003; Walser et al., 2011; Uhlen et al., 2010) whose amino acid sequences contain no distinguishable features other than their N-terminal signal peptides and C-terminal proximal transmembrane sequences (Nomoto et al., 2003; Oh-hashi et al., 2010, 2012), with intracellular C-terminal Ser/Thr phosphatase-1 (PP1) docking (Hendrickx et al., 2009; Heroes et al., 2013) and WIRS (WAVE regulatory complex interacting receptor sequence) cytoplasmic motifs (Fig. 1 and Supplementary Fig. S1) (Chen et al., 2014). The latter is consistent with the reported co-localization of TMEM132D with actin filaments (Walser et al., 2011).
Cell–cell junctions in the central nervous system (CNS) are maintained by a variety of transmembrane proteins that signal and physically link between the cytoskeletons of adjacent cells. Many such proteins (e.g. CNTN, LRIG, NCAM, SEMA3 and SIGLEC families) contain one or more immunoglobulin (IG) superfamily domains (Rougon and Hobert, 2003). Here we identify TMEM132 molecules as novel IG domain containing proteins of the CNS.
2 Results and discussion
2.1 Protein sequence analysis: three tandem immunoglobulin domains
We initiated our analyses by performing a JackHMMER iterative search (Finn et al., 2015) starting from the human TMEM132A protein sequence of the UniRef50 database (Wu, 2006). Whilst characterising the TMEM132 family we identified full-length homologous proteins across essentially all of the animal kingdom, including nematodes (Caenorhabditis elegans) and hexapods (Drosophila melanogaster). As input for our analysis we used a full-length multiple sequence alignment generated with T-Coffee (Notredame et al., 2000). Using extensive profile-to-sequence and profile-to-profile comparison analyses (Finn et al., 2015; Söding et al., 2005) we identified a repeated pattern of conserved amino-acids in the region lying between positions 400 and 767 for human TMEM132A, corresponding approximately to the conserved region used to define the family in Pfam (Family TMEM132, accession: PF16070) (Punta et al., 2012). This region is conserved among animals and some premetazoan proteins that are additionally rich in cadherin domains (Supplementary Fig. S2) (Abedin and King, 2008; Nichols et al., 2012). Profile-versus-sequence and profile-versus-profile comparisons of this conserved repeated pattern allowed the identification of three consecutive repeated regions, each of which independently yielded statistically significant E-values of sequence similarity with the same fold, the bacterial immunoglobulin-like (BIG) domain (Mei et al., 2015; Ptak et al., 2014). HHpred searches against the PDB70 profile database. (Söding et al., 2005) using TMEM132 repeats 1, 2 and 3 as input (corresponding to amino acids 400–491, 495–630 and 641–767 of human TMEM132A) detected the BIG domain from Leptospira interrogans (PDB ID: 2mh4), for example, with E-values of 0.031, 0.018 and 0.022, respectively. Moreover in all three HHpred results, in support of the first match, the next most statistically significant matches corresponded to additional members of the immunoglobulin superfamily (Supplementary Figs S3 and S4). The PDB70 database contains profile hidden Markov models (HMMs) for representative sequences, clustered to 70% maximum pairwise sequence identity to reduce redundancy, drawn from the PDB database (Söding et al., 2005).
BIG domains are widely distributed among bacteria, archaea and eukaryotes (Pfam family Big_2, accession: PF02368) (Punta et al., 2012). This domain adopts a beta-sandwich fold composed of nine strands organized in three sheets. Two of these sheets (composed of seven strands) contribute the immunoglobulin-like core of BIG domains. These seven strands are labelled ‘a’ to ‘g’ in Figure 1 and Supplementary Figure S3, following an established convention in the immunoglobulin fold (Bork et al., 1994). BIG domains have been described with diverse functions, usually relating to matrix, protein–ligand, or protein–protein interactions and are mainly extracellular (Mei et al., 2015; Ptak et al., 2014).
Close 3D proximity and evolutionary conservation of four cysteines allow us to identify two putative disulphide bridges, one that is internal to BIG1 and another that is inter-domain between BIG2 and BIG3 domains (Fig. 1; Supplementary Figs S3, S4 and S7). Disulphide bridges are commonly found in different IG folds and contribute to their structural stability (Bork et al., 1994).
2.2 First animal cohesin domain
Identification of the three tandem BIG domains, then allowed our analyses to be focused on the TMEM132 family N-terminal region taking advantage of iterative profile-versus-sequence searches against the UniRef50 protein sequence database (Wu, 2006). These resulted in the identification of two additional domains, each of which is present not only in animals but also among more diverse eukaryotes, including members of the Coherin family in choanoflagellates and sponges (Nichols et al., 2012). The domain preceding the BIG domains was discovered as the first cohesin homology domain in vertebrates (HHpred E-value < 5 × 10−3) (Supplementary Fig. S5). Cohesin domains are found widely in prokaryotes but, in eukaryotes, were previously thought to be restricted to choanoflagellate and sponge proteins (Pfam accession: PF00963) (Abedin and King, 2008; Nichols et al., 2012; Peer et al., 2009). These are not to be confused with the cohesin complex that regulates the separation of sister chromatids. Rather, cohesin domains are highly specialized protein-protein interaction modules that bind dockerin domains together forming the core that glues together the Cellulosome complex, a multi-enzymatic complex present in cellulolytic bacteria specialized in degrading cellulose (Adams et al., 2008; Artzi et al., 2017; Bras et al., 2016; Pinheiro et al., 2008; Tavares et al., 1997). Bacterial cohesin-dockerin rupture forces (>120 pN) are among the highest ever reported for a receptor-ligand system (Nash et al., 2016; Stahl et al., 2012). It is unclear whether the TMEM132 cohesin domain mediates such a strong interaction, in part because dockerin domain homologues are not detectable in vertebrate proteins.
2.3 TMEM132 domain architecture is ancient
The conserved region (corresponding to amino acids 127–239 of human TMEM132A) preceding the cohesin domain in TMEM132 is also evident in choanoflagellate and sponge proteins (HMMER E-value < 0.005) (Supplementary Fig. S6). Strikingly, despite each of the five TMEM132 domains (Fig. 1) being identified independently in these choanoflagellate and sponge proteins, all five are both present and in the identical order in the three cadherin protein families—lefftyrins, coherins and hedglings—that were contained in the last common ancestor of choanoflagellates and metazoans (Abedin and King, 2008; Nichols et al., 2012) (Supplementary Fig. S2). The TMEM132 domain architecture is thus ancient, preceding the emergence of early metazoans, and a repeated constituent of ancient cadherin domain-containing proteins with roles connecting the actin cytoskeleton with neighbouring cells and the extracellular matrix (Brieher and Yap, 2013; Ratheesh and Yap, 2012).
2.4 Disease and biological relevance
Eleven proteins are currently known to contain missense mutations within IG domains associated with 23 different disorders (Letunic et al., 2015). To these now can be added a twelfth, TMEM132E, whose R420Q missense mutation, mapped to its second BIG domain (Supplementary Fig. S1), has been validated using a zebrafish model to cause autosomal-recessive nonsyndromic hearing loss (Li et al., 2015). A TMEM132B nonsynonymous variant that replaces a serine conserved in TMEM132B-E in a putative phosphorylation motif has been associated with intra-cranial aneurysm (Farlow et al., 2015) although this variant also occurs rarely (frequency 1 × 10−4) in the general population (Lek et al., 2016). This lies adjacent to the Ser/Thr PP1 docking (Hendrickx et al., 2009; Heroes et al., 2013) and WIRS cytoplasmic motifs (Fig. 1), the latter which is found in a variety of neurological and other proteins including protocadherins (Chen et al., 2014). The newly identified domains, and conserved domain architecture, of TMEM132 proteins now should facilitate detailed experimental investigation of these proteins' domain and molecular functions and how these are modulated by sequence variants.
3 Conclusion
Their ancient ancestry and their associations with neurological disease suggest that TMEM132 genes have been undeserving of their relative obscurity. Our identification of these proteins as CNS-expressed IG domain superfamily adhesion molecules now places them in a more appropriate perspective as a putative key connection between the extracellular matrix and the actin-based cell cytoskeleton, with major roles in regulating changes in neuronal cell morphology, motility and migration. These findings should precipitate more detailed experimental and structural characterization of the TMEM132 family, and assist in formulating hypotheses concerning the cellular mechanisms by which sequence variants in these genes contribute to neurological disease.
Funding
This work has been supported by the Medical Research Council UK.
Conflict of Interest: none declared.
Supplementary Material
References
- Abedin M., King N. (2008) The premetazoan ancestry of cadherins. Science, 319, 946–948. [DOI] [PubMed] [Google Scholar]
- Adams J.J. et al. (2008) Structural basis of toxin complex formation. Proc. Natl. Acad. Sci. USA, 105, 12194–12199. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Artzi L. et al. (2017) Cellulosomes: bacterial nanomachines for dismantling plant polysaccharides. Nat. Rev. Microbiol., 15, 83–95. [DOI] [PubMed] [Google Scholar]
- Bork P. et al. (1994) The immunoglobulin fold. Structural classification, sequence patterns and common core. J. Mol. Biol., 242, 309–320. [DOI] [PubMed] [Google Scholar]
- Brieher W.M., Yap A.S. (2013) Cadherin junctions and their cytoskeleton(s). Curr. Opin. Cell Biol., 25, 39–46. [DOI] [PubMed] [Google Scholar]
- Bras J.L. et al. (2016) Diverse specificity of cellulosome attachment to the bacterial cell surface. Sci. Rep., 6, 38292. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chen B. et al. (2014) The WAVE regulatory complex links diverse receptors to the actin cytoskeleton. Cell, 156, 195–207. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Erhardt A. et al. (2012) Replication and meta-analysis of TMEM132D gene variants in panic disorder. Transl. Psychiatry, 2, e156. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Erhardt A. et al. (2011) TMEM132D, a new candidate for anxiety phenotypes: evidence from human and mouse studies. Mol. Psychiatry, 16, 647–663. [DOI] [PubMed] [Google Scholar]
- Farlow J.L. et al. (2015) Lessons learned from whole exome sequencing in multiplex families affected by a complex genetic disorder, intracranial aneurysm. PLoS One, 10, e0121104. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Finn R.D. et al. (2015) HMMER web server: 2015 update. Nucleic Acids Res., 43, W30–W38. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Forbes S.A. et al. (2015) COSMIC: exploring the world's knowledge of somatic mutations in human cancer. Nucleic Acids Res., 43, D805–D811. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Haaker J. et al. (2014) Higher anxiety and larger amygdala volumes in carriers of a TMEM132D risk variant for panic disorder. Transl. Psychiatry, 4, e357.. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hendrickx A. et al. (2009) Docking motif-guided mapping of the interactome of protein phosphatase-1. Chem. Biol., 16, 365–371. [DOI] [PubMed] [Google Scholar]
- Heroes E. et al. (2013) The PP1 binding code: a molecular-lego strategy that governs specificity. FEBS J., 280, 584–595. [DOI] [PubMed] [Google Scholar]
- Hodgson K. et al. (2016) Genome-wide significant loci for addiction and anxiety. Eur. Psychiatry, 36, 47–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Howe A.S. et al. (2016) Candidate genes in panic disorder: meta-analyses of 23 common variants in major anxiogenic pathways. Mol. Psychiatry, 21, 665–679. [DOI] [PubMed] [Google Scholar]
- Inoue A. et al. (2015) Association of TMEM132D, COMT, and GABRA6 genotypes with cingulate, frontal cortex and hippocampal emotional processing in panic and major depressive disorder. Int. J. Psychiatry Clin. Pract., 19, 192–200. [DOI] [PubMed] [Google Scholar]
- Iwakawa R. et al. (2015) Expression and clinical significance of genes frequently mutated in small cell lung cancers defined by whole exome/RNA sequencing. Carcinogenesis, 36, 616–621. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lane J.M. et al. (2017) Genome-wide association analyses of sleep disturbance traits identify new loci and highlight shared genetics with neuropsychiatric and metabolic traits. Nat. Genet., 49, 274–281. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lek M. et al. (2016) Analysis of protein-coding genetic variation in 60, 706 humans. Nature, 536, 285–291. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Letunic I. et al. (2015) SMART: recent updates, new developments and status in 2015. Nucleic Acids Res., 43, D257–D260. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Li J. et al. (2015) Whole-exome sequencing identifies a variant in TMEM132E causing autosomal-recessive nonsyndromic hearing loss DFNB99. Hum. Mutat., 36, 98–105. [DOI] [PubMed] [Google Scholar]
- Mei S. et al. (2015) Solution structure of a bacterial immunoglobulin-like domain of the outer membrane protein (LigB) from Leptospira. Proteins, 83, 195–200. [DOI] [PubMed] [Google Scholar]
- Nash M.A. et al. (2016) Single versus dual-binding conformations in cellulosomal cohesin-dockerin complexes. Curr. Opin. Struct. Biol., 40, 89–96. [DOI] [PubMed] [Google Scholar]
- Nichols S.A. et al. (2012) Origin of metazoan cadherin diversity and the antiquity of the classical cadherin/beta-catenin complex. Proc. Natl. Acad. Sci. USA, 109, 13046–13051. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Nomoto H. et al. (2003) Molecular cloning of a novel transmembrane protein MOLT expressed by mature oligodendrocytes. J. Biochem., 134, 231–238. [DOI] [PubMed] [Google Scholar]
- Notredame C. et al. (2000) T-Coffee: a novel method for fast and accurate multiple sequence alignment. J. Mol. Biol., 302, 205–217. [DOI] [PubMed] [Google Scholar]
- Oh-Hashi K. et al. (2006) GRP78-binding protein regulates cAMP-induced glial fibrillary acidic protein expression in rat C6 glioblastoma cells. FEBS Lett., 580, 3943–3947. [DOI] [PubMed] [Google Scholar]
- Oh-Hashi K. et al. (2010) Knockdown of transmembrane protein 132A by RNA interference facilitates serum starvation-induced cell death in Neuro2a cells. Mol. Cell Biochem., 342, 117–123. [DOI] [PubMed] [Google Scholar]
- Oh-Hashi K. et al. (2012) Characterization of the expression and cell-surface localization of transmembrane protein 132A. Mol. Cell Biochem., 370, 23–33. [DOI] [PubMed] [Google Scholar]
- Oh-Hashi K. et al. (2003) Cloning and characterization of a novel GRP78-binding protein in the rat brain. J. Biol. Chem., 278, 10531–10537. [DOI] [PubMed] [Google Scholar]
- Oh-Hashi K. et al. (2015) Transcriptional and post-transcriptional regulation of transmembrane protein 132A. Mol. Cell Biochem., 405, 291–299. [DOI] [PubMed] [Google Scholar]
- Pandey A.K. et al. (2014) Functionally enigmatic genes: a case study of the brain ignorome. PLoS One, 9, e88889.. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Peer A. et al. (2009) Noncellulosomal cohesin- and dockerin-like modules in the three domains of life. FEMS Microbiol. Lett., 291, 1–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Peifer M. et al. (2012) Integrative genome analyses identify key somatic driver mutations of small-cell lung cancer. Nat. Genet., 44, 1104–1110. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Pinheiro B.A. et al. (2008) The Clostridium cellulolyticum dockerin displays a dual binding mode for its cohesin partner. J. Biol. Chem., 283, 18422–18430. [DOI] [PubMed] [Google Scholar]
- Ptak C.P. et al. (2014) NMR solution structure of the terminal immunoglobulin-like domain from the leptospira host-interacting outer membrane protein, LigB. Biochemistry, 53, 5249–5260. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Punta M. et al. (2012) The Pfam protein families database. Nucleic Acids Res., 40, 290–301. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Qanbari S. et al. (2014) Classic selective sweeps revealed by massive sequencing in cattle. PLoS Genet., 10, e1004148.. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Quast C. et al. (2012) Rare variants in TMEM132D in a case-control sample for panic disorder. Am. J. Med. Genet. B Neuropsychiatr. Genet., 159B, 896–907. [DOI] [PubMed] [Google Scholar]
- Ratheesh A., Yap A.S. (2012) A bigger picture: classical cadherins and the dynamic actin cytoskeleton. Nat. Rev. Mol. Cell Biol., 13, 673–679. [DOI] [PubMed] [Google Scholar]
- Rougon G., Hobert O. (2003) New insights into the diversity and function of neuronal immunoglobulin superfamily molecules. Annu. Rev. Neurosci., 26, 207–238. [DOI] [PubMed] [Google Scholar]
- Rudin C.M. et al. (2012) Comprehensive genomic analysis identifies SOX2 as a frequently amplified gene in small-cell lung cancer. Nat. Genet., 44, 1111–1116. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Söding J. et al. (2005) The HHpred interactive server for protein homology detection and structure prediction. Nucleic Acids Res., 33, W244–W248. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Shimada-Sugimoto M. et al. (2016) Polymorphisms in the TMEM132D region are associated with panic disorder in HLA-DRB1*13: 02-negative individuals of a Japanese population. Hum. Genome Var., 3, 16001.. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Stahl S.W. et al. (2012) Single-molecule dissection of the high-affinity cohesin-dockerin complex. Proc. Natl. Acad. Sci. USA, 109, 20431–20436. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Tavares G.A. et al. (1997) The crystal structure of a type I cohesin domain at 1.7 A resolution. J. Mol. Biol., 273, 701–713. [DOI] [PubMed] [Google Scholar]
- Uhlen M. et al. (2010) Towards a knowledge-based Human Protein Atlas. Nat. Biotechnol., 28, 1248–1250. [DOI] [PubMed] [Google Scholar]
- Walser S. et al. (2011) TMEM132D – a putative cell adhesion molecule involved in panic disorder. Pharmacopsychiatry, 21, A117. [Google Scholar]
- Wang M.H. et al. (2016) A fast and powerful W-test for pairwise epistasis testing. Nucleic Acids Res., 44, e115. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wu C.H. (2006) The Universal Protein Resource (UniProt): an expanding universe of protein information. Nucleic Acids Res., 34, D187–D191. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.