The Ensembl Database shows several TRIO transcripts, only one of which encodes the full length protein and two, which are translated into parts of the protein (Figure 1). In addition, 6 splice variants have been identified and characterized at the protein level (Figure 2) (Portales-Casamar et al., 2006).

The TRIO gene consists of 58 exons encoding for a 10244 bp transcript.

TRIO mRNA shows a ubiquitous expression pattern in adult human tissues (Debant et al., 1996). It is also present in embryonic tissues, as in mice it is detectable starting from around embryonic day 10 (E10) (Portales-Casamar et al., 2006).

The human TRIO gene encodes a protein of 3097 amino acids, with a predicted molecular weight of 340.6 kD. The TRIO protein harbors distinct domains (Figure 3).
Six shorter isoforms have been described, which are listed in Figure 4 with their respective molecular weights.
TRIO is a complex protein containing three catalytic domains (hence the name "TRIO"): two tandem Dbl Homology (DH) - pleckstrin homology (PH) domains, and a Serine/Threonine kinase domain (Debant et al., 1996). The DH-PH module activates the small GTPases of the Rho family by mediating guanine nucleotide exchange. The DH1PH1 domain (also named GEF1) activates the GTPases RhoG and Rac1, while DH2PH2 (GEF2) activates RhoA (Bellanger et al., 1998; Blangy et al., 2000). So far, the kinase domain has not been characterized further. In addition to these domains, TRIO harbors numerous other motifs, including a Sec14 lipid-binding domain (also called CRAL-TRIO domain), eight spectrin repeats, two Src Homology 3 domains (SH3) and an Immunoglobulin domain. The presence of these numerous domains suggests that TRIO functions as an integrator of multiple upstream inputs and as an activator of multiple downstream pathways.

The full length TRIO protein is ubiquitously expressed, although at a low level, but its shorter isoforms have a more specific expression pattern. Isoforms TRIO A, B and D are strongly and exclusively expressed in various regions of the brain, throughout development (starting at E10) and in the adult, while the expression of the isoform TRIO C is specific to the cerebellum (Portales-Casamar et al., 2006; Sun et al., 2006). To date, Isoform TRIO E has only been detected in human neuroblastoma cell lines (Portales-Casamar et al., 2006). The Tgat isoform has been detected at the mRNA level in cells from patients with adult T-cell leukemia (ATL) (Yoshizuka et al., 2004).

TRIO protein is primarily located in the cytoplasm of cells, but it is also found partly associated with the plasma membrane, as in the case of the TRIO C isoform (Sun et al., 2006), or colocalizing with F-actin at the tip of the growth cone in PC12 cells (Estrach et al., 2002). The link between TRIO and the actin cytoskeleton is further strengthened by its interactions with the actin-binding proteins Filamin A and Tara (Bellanger et al., 2000; Seipel et al., 2001), as well as its association with Focal Adhesion Kinase (FAK) (Medley et al., 2003).

TRIO has originally been identified in human in 1996, in a two-hybrid screen using the cytoplasmic fragment of the tyrosine phosphatase LAR as bait (Debant et al., 1996). TRIO is a RhoGEF for the GTPases RhoG/Rac1 on one hand (via its GEF1 domain) and RhoA on the other hand (via its GEF2 domain). TRIO is implicated in the regulation of different physiological processes, most notably in neuronal physiology, but also in myogenesis and phagocytosis, processes that all require the rearrangement of the actin cytoskeleton. At the pathological level, evidence accumulates that TRIO is implicated in disease, mainly cancer, most likely by regulating actin cytoskeleton dynamics in migrating or invasive cells.
The TRIO Rho-GEF and neuronal physiology
The first evidence for a role of TRIO in neuronal physiology came from genetic analyses of TRIO orthologues in Caenorhabditis elegans (unc-73) and in Drosophila (D-TRIO), which established the TRIO family as a key component of the intracellular signaling pathway that regulates axon guidance and cell migration, mainly in the nervous system (Steven et al., 1998; Awasaki et al., 2000; Bateman et al., 2000; Liebl et al., 2000; Newsome et al., 2000).
Gene targeting studies in mouse have then shown that TRIO is required for late embryonic development, probably by playing essential roles in the development of the nervous system and in fetal skeletal muscle formation (OBrien et al., 2000). Consistently, mammalian TRIO has been shown to be required for RhoG-mediated neurite outgrowth in PC12 cells in response to NGF (Estrach et al., 2002). A direct interaction between TRIO and the integral membrane protein Kidins220/ARMS was further shown in this pathway, which might regulate membrane localization of TRIO in response to NGF (Neubrand et al., 2010). And more recently, a role for TRIO in mediating netrin-1-induced axon guidance has been found (Briançon-Marjollet et al., 2008). During development of the nervous system, Netrin-1 has been shown to promote axonal outgrowth through the binding of its receptor DCC. The binding of netrin-1 to DCC activates the small GTPase Rac1, through a Rho-GEF which had not been identified. Briançon-Marjollet and colleagues have shown that TRIO is the GEF responsible for this Rac activation signal elicited by Netrin-1/DCC, during axon outgrowth and guidance (Briançon-Marjollet et al., 2008). Further studies have also shown that TRIO regulates the organization of neuronal clusters in the hindbrain in vivo (Backer et al., 2007).
The ultimate evidence for the essential role of TRIO in the development of the nervous system has been brought recently with the generation and characterization of mice with a neural specific knock-out of TRIO (TrioNKO) (Peng et al., 2010). About 90% of TrioNKO mice died within 1 day after birth. Surviving mice showed a reduced body weight and smaller brain, with defective cerebella and severe signs of ataxia, probably due to the absence of granule cells in the internal granule cell layer (Peng et al., 2010).
TRIO and muscle development
In addition to neuronal defects, TRIO KO mice show also defects in secondary myogenesis, suggesting a role for TRIO in this late developmental process (OBrien et al., 2000). More recently, Charrasse and colleagues have clarified this role using C2C12 cells (Charrasse et al., 2007). Myoblast fusion requires M-Cadherin-dependent Rac1 activation. It appears that TRIO is complexed with Rac1 and M-Cadherin prior to myoblast fusion, leading to Rac1 activation. In contrast TRIO knockdown by shRNA inhibits this fusion process.
TRIO and phagocytosis of apoptotic cells
TRIO is implicated in phagocytosis of apoptotic cells both in mammals and in worms (deBakker et al., 2004). In mammals, the ELMO/DOCK180/CrkII complex is able to activate Rac and to induce cytoskeletal modifications leading to phagocytosis. TRIO activates RhoG, which in turn interacts with and activates this ELMO/DOCK180/Rac complex. This pathway is conserved in C.elegans.
TRIO and cancer
This topic is dealt with in the following section concerning diseases TRIO is implicated in.

TRIO has originally been identified in human in 1996, as an interactor of the tyrosine phosphatase LAR (Debant et al., 1996). Since then, TRIO orthologs have been identified in invertebrates, C. elegans and Drosophila, called Unc-73 and DTRIO respectively (Steven et al., 1998; Lin and Greenberg, 2000). In mammals, there exists a paralog of TRIO, called Kalirin that was originally identified in rat (Alam et al., 1997). Interestingly, the structural organization of TRIO family members is very well conserved through evolution. Human TRIO and human Kalirin share 61.4% identity, and both show a similar identity to DTRIO (43.6% and 43.2% respectively) and to Unc-73 (29.3% and 29.6% respectively). Identity between orthologs is even higher if the domains are considered separately (up to 92% for TRIO and Kalirin GEF1 for example).
TRIO function is also conserved through evolution, since all orthologs are implicated in neuronal specific functions, ranging from neuronal development to cell migration, axon outgrowth and guidance, and synaptic plasticity.

A systematic sequencing of cancer genomes led to the identification of more than 1000 somatic mutations in the coding exons of 518 protein kinase genes. In this study, 9 mutations have been found in the TRIO gene (Greenman et al., 2007). More detailed information of this study can be found on the following website: http://www.sanger.ac.uk/genetics/CGP/Studies/

None reported.

A systematic sequencing study of cancer genomes obtained from 210 diverse human cancers (primary tumors and immortalized cell lines) has been performed to find possible mutations in the 518 kinases of the human genome (Greenman et al., 2007). The screened cancers included breast, lung, colorectal, gastric, testis, ovarian, renal, melanoma, glioma and acute lymphoblastic leukemia. Over 1000 somatic mutations were detected in 321 kinases, including mostly single substitutions, and a few small insertions, deletions or complex changes. Statistical analysis allowed determining a factor indicating whether these mutations are "driver" or "passenger" mutations. TRIO having a kinase domain in its C-terminus was among the genes sequenced in this study. Nine mutations were identified in the TRIO gene, spanning the whole sequence (Figure 5). Interestingly, 6/9 mutations lied within the catalytic domains of TRIO (both DH-PH domains and the kinase domain). The cancer types these mutations were found in were diverse, with a prevalence of melanoma however. The "driver/passenger" selection pressure factor was 0.78 for TRIO, suggesting that the mutations are passenger mutations.