RASSF2 interaction with Ras
RASSF2 contains a functional RA domain that displays a strong binding to K-Ras, but only weak binding to H-Ras (Vos et al., 2003). RASSF2 associates with the K-Ras effector domain in a GTP-dependent manner thus displaying the basic properties of a genuine Ras effector. RASSF2 growth inhibition is enhanced in the presence of K-RasG12V. Furthermore, siRNA-mediated knock-down of RASSF2 in K-Ras transformed cells enhanced anchorage-independent growth. However, in the absence of K-Ras transformation knock-down of RASSF2 inhibited growth (Akino et al., 2005). These data indicate that RASSF2 mediates some of the growth inhibitory properties of K-Ras and that inactivation of RASSF2 enhances K-Ras-induced transformation.

RASSF2 interaction with the proapoptotic kinases MST1 and MST2
How the interaction of RASSF proteins with Ras results in growth suppression has been the subject of intense investigation. The proapoptotic mammalian Serine/Threonine kinases MST1 and MST2 were identified as RASSF2 interacting partners by yeast two-hybrid (Y2-H) (Khokhlatchev et al., 2002). RASSF1 and RASSF5 were also identified as MST binding partners as part of a novel Ras-regulated signalling pathway. Recently the interaction of RASSF2 and MST1/2 was formally demonstrated in human cells at the endogenous level. Interaction occurs between the SARAH domains found within RASSF2 and MST1/2 (Cooper et al., 2009). RASSF2 appears to have a distinct role in regulating MST2 function. Activation of MST2 is followed by a rapid proteasome-dependent loss of MST2 stability (that is not associated with MST2 cleavage). Interestingly, over expression of RASSF2 results in increased levels of MST2 and provides protection of MST2 from degradation following its activation. In agreement with this loss of RASSF2 protein in colorectal tumours, or in colorectal tumour cell lines in which RASSF2 levels are decreased by shRNAi, also leads to decreases in MST2 levels (Cooper et al., 2009). RASSF2 appears to be a substrate for MST1 and MST2 and co-expression of either kinase with RASSF2 relocalises RASSF2 from the nucleus to the cytoplasm in a manner dependent on kinase activity (Cooper et al., 2009). Since MST2 remains in complex with RASSF2 following its activation these data collectively suggest RASSF2 stabilises active MST2 allowing (or perhaps even targeting) MST2 substrate phosphorylation. This RASSF2-mediated stabilisation also appears to be true for MST1. Thus loss of RASSF2, as is frequently observed in cancer, leads to loss of MST1 and MST2 leading to a decrease in apoptotic potential. That RASSF2 appears to be capable of influencing MST stability so drastically is likely to be due to the observation that the majority of both MST1 and MST2 are in complex with RASSF2 in at least some cell types (Cooper et al., 2009). The interaction of RASSF2 with MST1/2 poses an interesting question with regards to the regulation of RASSF2 localisation. Both RASSF2 and MST1/2 have been shown to contain sequences essential for nuclear import and export (Lee and Yonehara, 2002; Kumari et al., 2009). Both the NLS and NES sequences within MST1/2 and RASSF2 respectively are located very close to or within the SARAH domains and neither are canonical NLS/NES sequences. Mapping of both these sequences were determined by deletion mapping, which would most likely also affect RASSF2-MST1/2 interaction thus it now seems likely that the RASSF2-MST1/2 complex constantly cycles between the nucleus (by virtue of RASSF2 NLS) and cytoplasm (by virture of MST1/2 NES) and disruption of the interaction between RASSF2 and MST1/2 would likely affect the localisation of both proteins. Also, the fact that RASSF2 translocation to the cytoplasm is dependent on ERK2 activity (Kumari and Mahalingam, 2009) suggests the Ras-MEK-ERK pathway may serve to phosphorylate MST1/2, which then phosphorylates RASSF2, translocating it to the cytoplasm and allowing RASSF2 to interact with Ras (figure 4). That nuclear RASSF2 is required for full tumour suppressor activity (Cooper et al., 2008; Kumari et al., 2009) may be explained by the fact that ERK2 translocates to the nucleus upon its activation (Khokhlatchev et al., 1998).

Other functions of RASSF2
Other functions and interacting partners of RASSF2 are extremely likely. Y2-H using RASSF2 as bait implicates NORE1A and RASSF3 in RASSF2 function, although these have not yet been confirmed in mammalian cells (Hesson et al., 2005). These interactions may implicate other RASSF members in modulating RASSF2 function and suggests a complex network of cross-talk between signalling pathways involving RASSF proteins. Also, the exact mechanisms of apoptotic and cell cycle regulation of RASSF2 have yet to be completely defined. Microarray analysis of gene expression before and after exogenous expression of RASSF2 in gastric and OSCC cancer cell lines showed RASSF2 downregulates expression of several inflammatory response genes including the cytokines IL-8, LCN2, CXCL1, CXCL2, CXCL3, CXCL5 and CXCL6, CCL20 and CCL21 and genes involved in immune-cell chemotaxis (Maruyama et al., 2008; Imai et al., 2008). A possible pathway influenced by RASSF2 is the NF-kB pathway since over expression of RASSF2 significantly downregulated NF-kB transcriptional activity (Maruyama et al., 2008; Imai et al., 2008). Of note is the recent observation that pigs experimentally infected with Porcine Circovirus Type 2 (PCV2) show upregulation of several CXCL family cytokines as well as RASSF2 (Fernandes et al., 2009) therefore it is likely a role for RASSF2 in regulating immune response pathways remains to be discovered. There is also evidence that RASSF2 may regulate the actin cytoskeleton since re-expression of RASSF2 leads to loss of stress fibres, cell rounding and the suppression of RhoGTPase activation (Maruyama et al., 2008; Akino et al., 2005). Additionally, RASSF2 upregulation appears to be a cellular response to ionising radiation (Sakamoto-Hojo et al., 2003).