The ADP-ribosylation factor (ARF) family of small GTP-binding proteins are ubiquitously expressed and involved in many cellular events such as cell adhesion, cell migration, neurite outgrowth, cell secretion, endocytosis and exocytosis, and maintenance of the platelet cytoskeleton (D’Souza-Schorey and Chavrier 2006; Ueda et al. 2013; Jang et al. 2016; Urban et al. 2016). In mammals, the ARF family consists of six members (ARFs 1–6) that differ in the localization and function. ARFs 1–5 function at the Golgi, whereas ARF6 regulates cellular events at the plasma membrane (Donaldson and Jackson 2011). ARFs belong to the Ras superfamily of GTPases and therefore act as molecular switches by cycling between inactive GDP-bound and active GTP-bound forms. They depend on Guanine Exchange Factors (GEFs) for activation and GTPase-Activating Proteins (GAPs) for inactivation (Donaldson and Honda 2005). Mammalian cells express 14 ARF GEFs which are classified into the five families: Pleckstrin and SEC7 Domain (PSD) (1–4), Cytohesins (1–4), Brefeldin A-Resistant ARF GEFs (BRAGs) (1–3), Golgi-Specific Brefeldin A-Resistance Factor 1 (GBF1), and Brefeldin A-inhibited guanine nucleotide-exchange proteins for ADP-ribosylation factors (BIGs) (1–2) (Donaldson and Jackson 2011).
In humans, the PSD GEF family consists of four members: PSD1, PSD2, PSD3 (also known as EFA6D, EFA6R, or HCA67), and PSD4, and additional isoforms, generated by alternative splicing, have been described for each member. This family of ARF GEFs shares a common domain organization. They consist of a SEC7 domain (named after the gene product of Saccharomyces cerevisiae), which bears the ARF-GEF activity, a pleckstrin homology (PH) domain, which is responsible for the plasma membrane localization by interacting with phosphatidylinositols (PIs) such as PI 4,5-bisphosphate (PI4,5-P2), and a C-terminal region, which consists of one or more coiled-coil (CC) motif responsible for cytoskeleton rearrangements and interactions with downstream signaling proteins (Kanamarlapudi 2014). Some splice variants of the PSD family also contain a less conserved N-terminal domain, upstream of the SEC7 domain, whose functions are mostly unknown. However, recently, the N-terminal regions of PSD3/PSD4 have been shown to associate with Dynamin2 (Dyn2) during ARF6-induced clathrin-mediated endocytosis (CME) (Okada et al. 2015).
Sites of Expression and Splice Variants of PSD3
In mice, PSD3 gene has initially been shown to express in the brain (Sakagami et al. 2006). Further in situ hybridization analysis has revealed that PSD3 is highly expressed in the hippocampal region of mice brain. Real-time qPCR analysis of various adult mice tissues showed PSD3 expression is highest in the brain, eye, thymus, lung, spleen, and testis while a lower expression in the small intestine, kidney, and heart (Sakagami et al. 2006).
Alternative splicing of the PSD3 gene generates three variants in humans: PSD3 variant A (1048aa, 116 kDa), PSD3 variant B (1047aa, 116 kDa), and PSD3 variant C (513aa, 56 kDa). A recent study in humans have functionally characterized two PSD3 splice variants (Fig. 1a) that differed in the length of their N-terminal region: a long 1047 amino acid variant B and a shorter 513 amino acid variant C. These variants both possess a SEC7, PH, and CC domains (Kanamarlapudi 2014). Based on sequence comparison, the human PSD3B is the counterpart of mouse PSD3 variant 2, whereas the small variant of human PSD3 (PSD3C) does not correspond to any mouse PSD3 variants (Fukaya et al. 2016). However, the human PSD3C is similar, in size, to mouse PSD3 variant 5* (see Fig. 1). In human tissues, thus far, only PSD3C has been detected with high expression levels in the brain, liver, spleen, and testis and lower expression levels in small intestine, kidney, stomach, and ovaries (Kanamarlapudi 2014). Taking both the mouse and human PSD3 expression studies into account, it seems that the expression pattern of PSD3 splice variants may be species-specific.
PSD3 expression has shown to be altered in breast, prostate, kidney, colon, lung, thyroid, and ovarian cancers (see references in Kanamarlapudi 2014). In cell lines, a 56 kDa protein band corresponding to PSD3C has been detected in MCF7, Hela, HEK293, and A549 cell lines (Kanamarlapudi 2014). Recently, we have also observed the expression of PSD3C in many ovarian cancer cell lines (unpublished data).
PSD3 Structural Motifs and Functions
The alignment of amino acid sequence of PSD3 with that of the other PSD family members suggests that PSD3 shares the core catalytic SEC7 domain, the PH domain, and the CC domain of PSD family GEFs. A recent study has shown that PSD3 preferentially activates ARF6 (Kanamarlapudi 2014). PSD3C has been shown to specifically activate ARF6 when over expressed in mammalian cells. Furthermore, PSD3C is able to induce the loss of actin stress fibers through the ARF6 activation. This GEF activity is dependent on the presence of the catalytically active SEC7 domain (Kanamarlapudi 2014). Using various PSD3 deletion mutants, it has been shown that the PH and CC domains, but not the Sec7 domain, contribute to the targeting of PSD3 to the plasma membrane (Kanamarlapudi 2014). Through its PH domain, PSD3 is able to specifically bind PI4,5-P2, in vitro. Interestingly, the PH domain alone seems to be insufficient for PI4,5-P2 binding, indicating that the adjacent sequences may be necessary for proper folding and binding to membrane PIs. PSD3 localization to the plasma membrane is dependent on the interaction of its PH domain with PI4,5-P2. This localization seems to be further stabilized through the interaction of the CC domain with the actin cytoskeleton (Kanamarlapudi 2014).
PSD3-Interacting Protein Network
In humans, the BioGrid interaction repository has identified four PSD3-interacting partners so far. Using a yeast-two hybrid system, ARAP1 (a member of the ARFGAP family) and MX1 (a dynamin-like large GTPase involved in viral resistance) have been shown to interact with PSD3 (Yoon et al. 2011; Brantis-de-Carvalho et al. 2015). In addition, affinity capture mass spectrometry analysis has identified MLH and PMS2 – which both are components of the postreplicative DNA mismatch repair system – as interacting partners of PSD3 (Cannavo et al. 2007). Although there is no evidence of a physical interaction between PSD3 and tumor protein p53, genetic interaction analysis have shown that mutations in both genes result in cumulative cellular defects in human cancer cells (Xie et al. 2012).
A recent study into the mechanisms of clathrin-mediated endocytosis (CME) has shown that dynamin 2 (Dyn2) mediates the activation of ARF6 by the PSD family GEFs (Okada et al. 2015). In HeLa cells, PSD3 and PSD4 have been shown to colocalize with Dyn2 at the cell periphery (Okada et al. 2015). Moreover, immunoprecipitation studies have shown the physical interaction between Dyn2 and PSD4 in HeLa cells. The interaction occurs through the N-terminal 145–185 amino acid region of PSD4, which shares high homology with that in PSD3. This raises the prospect of similar Dyn2 interactions for PSD3.
PSD3 was originally been identified as a tumor-associated antigen 67 in hepatocellular carcinoma patients and hence named HCA67 (Wang et al. 2002). Consistent with this, PSD3 expression has been shown to be altered in colon, breast, prostate, glioblastoma, and ovarian cancers (see references in Kanamarlapudi 2014). Moreover, PSD3 has been identified as a candidate metastasis suppressor gene in breast cancer (Thomassen et al. 2009). In epithelial ovarian cancer (EOC), PSD3 downregulation in high-grade carcinomas (particularly the serous subtype) compared to low-grade carcinomas significantly impacts survival of patients (Pils et al. 2005). Since PSD3 downregulation correlates with progression of ovarian carcinomas, it can be used as a biomarker of ovarian cancer.
In mouse brain, gene expression analysis has identified PSD3 as a schizophrenia-associated biomarker (Mozhui et al. 2011), while genome-wide association studies have identified PSD3 as a genetic risk factor for systemic sclerosis (Martin et al. 2012; Jin et al. 2014).
PSD3 belong to the PSD family of ARF6-specific GEFs and, as a result, it shares the catalytic Sec7, the PH, and the CC domains of the other PSD family GEFs. PSD3 acts as an ARF6 GEF by localizing at the plasma membrane, which is mediated by its PH domain and stabilized by its CC domain. Alternative splicing of the PSD3 gene has been described in both mouse and humans where the results obtained indicate species- and tissue-specific expression pattern for PSD3. Thus far, four physical interacting partners (MX1, MLH, PMS2, and ARAP1), one highly probable partner (Dyn2), and one genetic interaction (TP53) have been identified. However, no further characterization of these interactions has been undertaken so far. It has been shown recently that Dyn2 is able to activate ARF6 through association with PSD3. PSD3 alterations have been found to be associated with many cancer types as well as other diseases such as schizophrenia and systemic sclerosis.
- Pils D, Horak P, Gleiss A, Sax C, Fabjani G, Moebus VJ, et al. Five genes from chromosomal band 8p22 are significantly down-regulated in ovarian carcinoma: N33 and EFA6R have a potential impact on overall survival. Cancer. 2005;104:2417–29. doi: 10.1002/cncr.21538.PubMedCrossRefPubMedCentralGoogle Scholar
- Sakagami H, Suzuki H, Kamata A, Owada Y, Fukunaga K, Mayanagi H, et al. Distinct spatiotemporal expression of EFA6D, a guanine nucleotide exchange factor for ARF6, among the EFA6 family in mouse brain. Brain Res. 2006;1093:1–11. doi: 10.1016/j.brainres.2006.02.058.PubMedCrossRefPubMedCentralGoogle Scholar
- Xie L, Gazin C, Park SM, Zhu LJ, Debily MA, Kittler EL, et al. A synthetic interaction screen identifies factors selectively required for proliferation and TERT transcription in p53-deficient human cancer cells. PLoS Genet. 2012;8:e1003151. doi: 10.1371/journal.pgen.1003151.PubMedPubMedCentralCrossRefGoogle Scholar