14-3-3γ binds regulator of G protein signaling 14 (RGS14) at distinct sites to inhibit the RGS14:Gαi–AlF4− signaling complex and RGS14 nuclear localization

Regulator of G protein signaling 14 (RGS14) is a multifunctional brain scaffolding protein that integrates G protein and Ras/ERK signaling pathways. It is also a nucleocytoplasmic shuttling protein. RGS14 binds active Gαi/o via its RGS domain, Raf and active H-Ras–GTP via its R1 Ras-binding domain (RBD), and inactive Gαi1/3 via its G protein regulatory (GPR) domain. RGS14 suppresses long-term potentiation (LTP) in the CA2 region of the hippocampus, thereby regulating hippocampally based learning and memory. The 14-3-3 family of proteins is necessary for hippocampal LTP and associative learning and memory. Here, we show direct interaction between RGS14 and 14-3-3γ at two distinct sties, one phosphorylation-independent and the other phosphorylation-dependent at Ser-218 that is markedly potentiated by signaling downstream of active H-Ras. Using bioluminescence resonance energy transfer (BRET), we show that the pSer-218–dependent RGS14/14-3-3γ interaction inhibits active Gαi1–AlF4− binding to the RGS domain of RGS14 but has no effect on active H-Ras and inactive Gαi1–GDP binding to RGS14. By contrast, the phosphorylation-independent binding of 14-3-3 has no effect on RGS14/Gαi interactions but, instead, inhibits (directly or indirectly) RGS14 nuclear import and nucleocytoplasmic shuttling. Together, our findings describe a novel mechanism of negative regulation of RGS14 functions, specifically interactions with active Gαi and nuclear import, while leaving the function of other RGS14 domains intact. Ongoing studies will further elucidate the physiological function of this interaction between RGS14 and 14-3-3γ, providing insight into the functions of both RGS14 and 14-3-3 in their roles in modulating synaptic plasticity in the hippocampus.

Although RGS14 interaction with G␣ i -GTP has been shown to be enhanced by phosphorylation (40), no clear mechanism for the negative regulation of RGS14 interactions with G␣ i has been identified. Here, we report that native RGS14 forms a stable complex with 14-3-3 in the hippocampus where RGS14 has been shown to negatively regulate synaptic plasticity and, consequently, learning and memory (3). 14-3-3 binds RGS14 at two distinct sites, one phosphorylation-dependent and the other phosphorylation-independent. In HEK 293 cells, we found that RGS14 binds 14-3-3␥ selectively and that this interaction is markedly potentiated by downstream active H-Ras signaling. This interaction is phosphorylation-dependent, and complex formation between RGS14 and 14-3-3 occurs directly, independently of any other interacting partners. 14-3-3 binding is centered at phosphorylated serine 218, and binding at that site inhibits active G␣ i binding at the RGS domain of RGS14 without affecting inactive G␣ i interaction with the GPR motif or active H-Ras binding at the R1 domain. In hippocampal neurons, the cell type in which RGS14 is natively expressed in the brain (13,41), we report that 14-3-3 also binds at a second site in a phosphorylation-independent manner. This 14-3-3 binding is distinct from that at phosphoserine 218. RGS14 is a nucleocytoplasmic shuttling protein (5,12,14), and 14-3-3 prevents RGS14 nuclear import in hippocampal neurons independently of binding to phospho-Ser-218.
Overall, our findings provide clear evidence for distinct phosphorylation-dependent and phosphorylation-independent binding sites for 14-3-3 on RGS14. Binding of 14-3-3 at these sites negatively regulates RGS14 interaction with active G␣ i at the RGS domain and nuclear import, respectively, demonstrating two distinct mechanisms through which 14-3-3 modulates RGS14 function and subcellular localization in hippocampal neurons.

Results
Previous work reported that 14-3-3 regulates the functions of several RGS proteins (26, [35][36][37][38][39]. To explore the possibility that 14-3-3 might bind to and regulate the function of RGS14, we immunoprecipitated native RGS14 out of mouse brain and subsequently probed for interactions with 14-3-3 proteins via immunoblotting (Fig. 1). Because RGS14 is highly and selectively expressed in the CA2 area of the hippocampus (41), we used isolated hippocampi as a tissue source for these studies. 14-3-3␥ was highly enriched in samples in which RGS14 was specifically immunoprecipitated as opposed to the beads-only condition where RGS14 was not enriched (Fig. 1). In combination, these data show that endogenously expressed 14-3-3 spe-cifically interacts with native RGS14 from the CA2 region of the mouse hippocampus.
To better understand the subcellular signaling events that affect and control the interaction between RGS14 and 14-3-3, we transitioned to the more manipulation-friendly system of transient transfection in HEK 293 cells. Following FLAG immunoprecipitation from HEK 293 cells expressing FLAGtagged RGS14 (FLAG-RGS14) and 14-3-3␥, we saw little or no interaction between RGS14 and 14-3-3 ( Fig. 2A) when compared with a baseline condition in which the immunoprecipitation was performed in the absence of RGS14 and the presence of 14-3-3. However, when RGS14's binding partner at the R1 domain, active H-Ras(G12V) (6,7), was expressed along with RGS14 and 14-3-3␥, we observed a clear interaction. The specificity of this interaction was further validated by the inclusion of a condition in which 14-3-3␥(K50E), which contains a mutation in the amphipathic helix of 14-3-3 that significantly decreases affinity for known 14-3-3 substrates (42,43), was expressed. As expected, 14-3-3␥(K50E) showed decreased binding when compared with the wildtype (WT) protein ( Fig.  2A), indicating that induction of 14-3-3 association with RGS14 by active H-Ras is specific.
Because active H-Ras has been reported to interact directly with RGS14 (6,7,44,45), it was unclear whether downstream H-Ras signaling or direct H-Ras binding to RGS14 was responsible for its induction of a 14-3-3␥:RGS14 complex. To parse out the role of active H-Ras in relation to 14-3-3␥ and RGS14, we used a mutation, RGS14(R333L), that has been shown to greatly reduce binding of H-Ras at the R1 domain (6,45). Using RGS14(R333L), any observed enhancement of the RGS14/14-3-3␥ interaction by active H-Ras can be attributed only to downstream signaling of H-Ras and not direct interaction with RGS14. We saw no difference in 14-3-3␥ binding to RGS14 or RGS14(R333L) in the presence of active H-Ras (Fig. 2B), indi- Both hippocampi from an adult mouse were isolated and homogenized with lysis buffer to generate cell lysate. Half of the lysate was incubated with anti-RGS14 antibody (IP: RGS14) and protein G-Sepharose beads, and the other half was incubated with beads alone (Beads only). Beads were then washed, and recovered proteins were resolved using SDS-PAGE before being subjected to immunoblotting (IB). The cell lysate immunoblots represent 2% of the whole-cell lysates used for the immunoprecipitation (IP).
To further confirm that downstream signaling by active H-Ras is necessary to promote 14-3-3␥ interaction with RGS14 in HEK 293 cells, we immunoprecipitated FLAG-RGS14 and probed for an interaction with 14-3-3␥ in the presence of active H-Ras(G12V), WT H-Ras, and the dominant-negative mutant H-Ras(S17N). As expected, only H-Ras(G12V) greatly enhanced 14-3-3 binding to RGS14 (Fig. 2C), indicating that downstream signaling through active H-Ras is necessary for the potentiation of the interaction between RGS14 and 14-3-3.
14-3-3 proteins typically bind substrates at specific motifs centered at a phosphorylated serine or threonine (24,25). However, 14-3-3 proteins are also capable of binding target proteins at nonphosphorylated residues (26 -31). To determine whether the H-Ras-potentiated 14-3-3␥ interaction with RGS14 was phosphorylation-dependent, we overexpressed FLAG-RGS14 and 14-3-3␥ and immunoprecipitated RGS14 from HEK 293 cells in the presence or absence of constitutively active H-Ras(G12V). Samples were either collected in the presence of a phosphatase inhibitor mixture to preserve phosphorylation sites or treated with -phosphatase to dephosphorylate any phosphorylated residues on RGS14. Although little interaction between RGS14 and 14-3-3␥ was seen in the absence of A, active H-Ras enhances RGS14 association with 14-3-3␥. 10-cm dishes of HEK 293 cells were transfected with 1.5 g of FLAG-RGS14, 3 g of His-tagged 14-3-3␥ (His-14-3-3␥), 3 g of a binding-incompetent mutant of 14-3-3␥ (14-3-3␥(K50E)), and/or 1.5 g of constitutively active H-Ras (H-Ras(G12V)) as indicated. Whole-cell lysates were subjected to FLAG immunoprecipitation (IP: FLAG), and recovered proteins were subjected to SDS-PAGE followed by immunoblotting (IB). B, active H-Ras does not influence 14-3-3 binding through direct interaction with RGS14. HEK 293 cells were transfected with 1.5 g of FLAG-RGS14, 3 g of His-tagged 14-3-3␥, 1.5 g of active H-Ras, and/or a 1.5 g of a point mutant of RGS14 that demonstrates greatly decreased binding to Ras (FLAG-RGS14(R333L)). Samples were collected and subjected to FLAG immunoprecipitation and SDS-PAGE, and analyzed via immunoblotting. C, H-Ras must be active to potentiate RGS14/14-3-3 interaction. 1.5 g of FLAG-RGS14 and 3 g of His-tagged 14-3-3␥ were transfected into HEK 293 cells along with either 1.5 g of empty vector, 1.5 g of WT, 1.5 g of constitutively active (G12V), or 1.5 g of dominant-negative (S17N) H-Ras. RGS14 and associated proteins were isolated via FLAG immunoprecipitation and subjected to SDS-PAGE before being analyzed using immunoblotting. D, RGS14 selectively interacts with the ␥ isoform of 14-3-3. HEK 293 cells were transfected overnight with FLAG-RGS14, H-Ras(G12V), and His-tagged 14-3-3␥, -␤, -, or -. Samples were then lysed and subjected to FLAG immunoprecipitation before being subjected to SDS-PAGE and analyzed for 14-3-3 interactions with RGS14 by immunoblotting for the hexahistidine tag on the 14-3-3 isoforms. The cell lysate immunoblots represent 2% of the whole-cell lysates used for the immunoprecipitation. These findings are representative of three independent experiments.
active H-Ras, the sample collected in the presence of both the phosphatase inhibitor mixture and H-Ras(G12V) showed a robust interaction between RGS14 and 14-3-3␥. Notably, the sample collected in the absence of phosphatase inhibitor mixture and treated with -phosphatase showed a greatly reduced interaction compared with the positive control (Fig. 3A), indicating that 14-3-3␥ interacts with RGS14 through a phosphorylated serine or threonine.
One major caveat of using coimmunoprecipitation from whole cells as a technique to examine protein/protein interactions is the inability to determine whether an interaction between two proteins is direct or indirect. In the case of RGS14, it was imperative to determine whether 14-3-3␥ binding to RGS14 was direct because RGS14 is reported to interact with Raf in an active H-Ras-dependent manner (45), and Raf is a well known binding partner of 14-3-3 (20, 24, 31, 42, 46 -49). Therefore, we used a far-Western blotting overlay experiment to determine whether 14-3-3␥ is capable of binding to RGS14 directly under the same conditions in which it interacts with RGS14 using coimmunoprecipitation (Fig. 3B). FLAG-RGS14 expressed in HEK 293 cells in the presence or absence of H-Ras(G12V) was immunoprecipitated to enrich the RGS14 in each sample. The samples were denatured, run on an SDSpolyacrylamide gel, and transferred to a nitrocellulose membrane, which was incubated overnight with pure 14-3-3␥ protein. 14-3-3␥ binding to RGS14 was detected by immunoblotting of the nitrocellulose membrane. We found that purified 14-3-3␥ interacts directly with the denatured RGS14 bands in the membrane at the correct size, 61 kDa (14-3-3␥ is ϳ28 kDa) in the same manner as predicted by previous coimmunoprecipitation experiments. Active H-Ras greatly increased 14-3-3␥ binding to RGS14, whereas treatment of an identical sample with -phosphatase greatly decreased this interaction (Fig. 3B). This indicates that 14-3-3␥ interacts directly with phosphorylated RGS14 with no other protein necessary as part of a complex to mediate the interaction.
Knowing that 14-3-3␥ interacts directly with RGS14 at a typical 14-3-3-binding motif centered at a phosphorylated serine or threonine, we created truncations of RGS14 to map the location of the 14-3-3␥-binding site (Fig. 4A). All truncated proteins were created with an N-terminal FLAG tag so they could be immunoprecipitated equally and their expression could be directly compared. Full-length FLAG-RGS14 and the four truncations were expressed in HEK 293 cells with 14-3-3␥ and active H-Ras and subjected to FLAG immunoprecipitation. Notably, only RGS14 constructs that contained the linker region between the RGS domain and the tandem RBDs (constructs 1, 3, and 4) coimmunoprecipitated with 14-3-3␥ (Fig.  4B), indicating that 14-3-3␥ binds a site within that linker region including RGS14 residues 186 -301 (Fig. 4).
Knowing the specific site at which 14-3-3 binds RGS14, we next attempted to determine the kinase that phosphorylates serine 218 of RGS14 to better understand the role that 14-3-3 interaction with RGS14 plays in relation to both RGS14 and 14-3-3 signaling. We focused on MEK, Akt, and CaMKII as possible kinases that could phosphorylate serine 218 because all of these kinases are downstream of active H-Ras signaling and  3. 14-3-3␥ directly binds RGS14 in a phosphorylation-dependent manner. A, 14-3-3 binds RGS14 in a phosphorylation-dependent manner. HEK 293 cells were transfected with 1.5 g of FLAG-RGS14 and 3 g of His-tagged 14-3-3␥ either in the presence or absence of 1.5 g of overexpressed H-Ras(G12V). Samples were either lysed in buffer containing a phosphatase inhibitor mixture (PI) or treated with -phosphatase following lysis (P). All samples were then incubated with anti-FLAG M2 affinity gel to pull down any proteins associated with FLAG-RGS14. The gel was washed, and samples were denatured in Laemmli buffer before being subjected to SDS-PAGE and immunoblotting (IB). B, 14-3-3 interacts directly with RGS14 in far-Western blotting. HEK 293 cells were transfected with FLAG-RGS14 either in the presence or absence of overexpressed H-Ras(G12V). All samples were then incubated with anti-FLAG M2 affinity gel to separate RGS14 from the whole-cell lysate. Samples were then either denatured in Laemmli buffer or treated with -phosphatase and then subjected to SDS-PAGE and transferred to a nitrocellulose membrane, which was incubated with pure His-tagged 14-3-3␥ overnight and probed for interactions between RGS14 and 14-3-3 via immunoblotting for 14-3-3. The cell lysate immunoblots represent 2% of the whole-cell lysates used for the immunoprecipitation (IP). These findings are representative of three independent experiments.
We next tested whether acute activation of endogenous H-Ras is sufficient to promote 14-3-3␥ binding to RGS14 or whether long-term activation is necessary. To activate endogenous Ras acutely, HEK 293 cells expressing FLAG-RGS14 and 14-3-3␥ were treated with EGF for 0, 10, 60, or 120 min. EGF treatment did not markedly enhance RGS14 pulldown of 14-3-3␥ at the earlier time points, but increased interaction was evident at the 2-h time point, indicating that prolonged H-Ras signaling is necessary (Fig. S2B). This suggests that changes at a protein translation level are likely necessary to enhance the RGS14/14-3-3 interaction in HEK cells. These findings indicate that the kinase responsible for phosphorylating serine 218 of RGS14 is unlikely to be directly downstream of active H-Ras, making the kinase responsible immensely difficult to predict and identify. Therefore, in subsequent efforts to identify a function for 14-3-3␥ interaction with RGS14, we focused on the functional consequences of 14-3-3␥ binding to RGS14 rather than the specific phosphorylation site and responsible kinase.

2.
3. 14-3-3␥ binds RGS14 at serine 218. A, alignment of the human, rat, and mouse RGS14 amino acid sequences shows complete conservation of serine 218 and surrounding amino acids. B, mutation of serine 218 of RGS14 to a phospho-null alanine decreases RGS14's association with 14-3-3. 1.5 g of WT RGS14 or 1.5 g of RGS14(S218A) were expressed in HEK 293 cells in the presence of 3 g of His-tagged 14-3-3␥ and 1.5 g of active H-Ras. Lysates were collected and subjected to FLAG immunoprecipitation before being analyzed by SDS-PAGE followed by immunoblotting (IB). The cell lysate immunoblots represent 2% of the whole-cell lysates used for the immunoprecipitation (IP). These findings are representative of three independent experiments.

14-3-3 regulation of RGS14
domain of RGS3 has been shown to block G protein binding (36). Because of the relatively close proximity of the 14-3-3␥binding site on RGS14 to the RGS domain (33 amino acids away), we examined the effect of 14-3-3␥ binding to RGS14 on its interactions with active H-Ras(G12V) at the R1 domain as well as active G␣ i1 binding at the RGS domain. For this, we examined protein/protein interactions in live cells using bioluminescence resonance energy transfer (BRET).
We next sought to determine whether 14-3-3 affects active G␣ i1 binding at the RGS domain of RGS14. Following incubation of HEK 293 cells expressing H-Ras(G12V), Luc-RGS14, and G␣ i1 -YFP with AlF 4 Ϫ , a nonspecific activator of heterotrimeric G␣ subunits, we measured binding of G␣ i1 at the RGS domain of RGS14. With the addition of overexpressed 14-3-3␥, the maximum measured interaction between G␣ i1 -AlF 4 Ϫ and RGS14 was decreased by ϳ50% (Fig. 7A), indicating that binding of 14-3-3␥ to RGS14 inhibits G protein binding at the RGS domain. We also performed this experiment with Luc-RGS14(S218A) to confirm that 14-3-3␥ binding centered at serine 218, as opposed to a possible off-target effect, was responsible for inhibition of active G␣ i1 binding at the RGS domain. Using this Luc-RGS14(S218A) point mutant, we saw no effect on maximal G protein binding when 14-3-3␥ was overexpressed (Fig. 7B). Furthermore, to examine whether 14-3-3␥ inhibition of G␣ i1 interaction with RGS14 is specific to the RGS domain, we performed the same experiment in the absence of AlF 4 Ϫ with RGS14 -Luc to measure any contributions that 14-3-3␥ binding to RGS14 might have to G␣ i1 binding to the GPR motif at the C-terminal end of the protein. In this case, expression of 14-3-3␥ had no effect (Fig. 7C), indicating that 14-3-3␥ binding to RGS14 only decreases active G␣ i1 interaction with the RGS domain.
After demonstrating that 14-3-3 specifically affects G␣ i1 binding to the RGS domain of RGS14, we next sought to examine the effect of 14-3-3 on RGS14 subcellular localization because 14-3-3 proteins have been shown to often affect the distribution of binding partners within the cell (17,21,26). We and others have previously shown that although RGS14 is predominantly cytosolic it is nonetheless a nucleocytoplasmic shuttling protein (5,12,13) as demonstrated by the fact that treatment of HEK cells with leptomycin B (LMB), which inhibits nuclear export, leads to a buildup of RGS14 in the nucleus as it is constantly being imported but unable to be shuttled out of the nucleus (5,12). Here, we show that YFP-RGS14 shuttles in and out of the nucleus in primary hippocampal neurons, the cell type in which RGS14 is natively expressed in the adult brain (41) (Fig. 8). Notably, coexpression of 14-3-3␥ prevented RGS14 translocation into the nucleus over the course of a 2-h treatment with LMB. Furthermore, coexpression of the binding-null mutant 14-3-3␥(K50E) with RGS14 still allowed RGS14 to be transported into the nucleus (Fig. 8), indicating that 14-3-3␥ is binding specifically to substrates in the cell, presumably RGS14, to block RGS14 nuclear localization.
To explore this idea of a second 14-3-3-binding site and in an effort to make sense of 14-3-3's effect on RGS14 nuclear localization, we reexamined previous data (Fig. 3B), noticing faint bands in the control conditions in the far-Western blot. These findings were consistent with direct 14-3-3 binding to RGS14, thus providing evidence of a potential second, phosphorylation-independent binding site on RGS14 for 14-3-3␥. To test this idea, we again performed a far-Western experiment (Fig. 10), this time comparing 14-3-3 binding to RGS14 immunoprecipitated from HEK cells (not treated with constitutively active H-Ras and therefore not strongly phosphorylated at ser-

14-3-3 regulation of RGS14
ine 218) and RGS14 purified from Escherichia coli, which is assured to be unphosphorylated by mammalian kinases. Notably, 14-3-3␥ directly binds RGS14 in both of these conditions (Fig. 10A), confirming the presence of a phosphorylation-independent 14-3-3-binding site on RGS14. To confirm that we were looking at a 14-3-3-binding site that is completely separate from the that centered at Ser-218 of RGS14, we created two truncation mutants of RGS14, RGS14(N-298) and RGS14(300-C), and partially purified them out of E. coli. We then performed far-Western blotting comparing direct binding of 14-3-3 to these truncations as well as full-length RGS14. 14-3-3␥ interacts with full-length RGS14 and RGS14(300-C) but does not interact with RGS14(N-298) (Fig. 10B), confirming the presence of a second, phosphorylation-independent 14-3-3-binding site on RGS14 that is separate from the phosphorylation-dependent 14-3-3-binding site that requires Ser-218 phosphorylation.

Discussion
Here, we provide evidence that RGS14 directly interacts with 14-3-3 in both a phosphorylation-dependent and phosphorylation-independent manner. The phosphorylation-dependent interaction is markedly enhanced by downstream signaling through active H-Ras and is mediated by phosphorylation of serine 218 of RGS14, located adjacent to the RGS domain, in the linker region between the RGS domain and the R1 domain. Functionally, this interaction blocks active G␣ i binding at the RGS domain but does not affect H-Ras binding to RGS14. Meanwhile, the phosphorylation-independent binding of 14-3-3 is correlated with the blockade of nuclear import of RGS14, thereby trapping it in the cytosol. One or both of these interactions take place within the mouse hippocampus, suggesting that 14-3-3 plays a role in regulating RGS14's suppression of synaptic plasticity in the CA2 region of the hippocampus.

H-Ras signaling promotes the specific phosphorylationdependent interaction between RGS14 and 14-3-3
Our findings (Fig. 2) demonstrate that the 14-3-3 interaction with RGS14 is promoted by downstream signaling of active H-Ras independently of H-Ras binding directly to RGS14. As RGS14 interacts with Raf (45), a well known 14-3-3-binding partner (20,23,31,46,47), it was unclear to us whether 14-3-3 interacts directly with RGS14 or through an indirect interaction bridged by Raf. However, analysis by far-Western blotting showed that 14-3-3 is able to interact directly with RGS14 in the absence of any other RGS14-interacting partners (Fig. 3B). Furthermore, decreased interaction between RGS14 and 14-3-3 following treatment with -phosphatase indicated that 14-3-3 binds at a motif centered at a phosphorylated serine or threonine on RGS14 (Fig. 3). Mapping of the 14-3-3-binding site with truncation mutants of RGS14 followed by creation of phospho-null point mutants at potential 14-3-3-binding sites revealed a 14-3-3-binding site centered around serine 218 on

14-3-3 regulation of RGS14
RGS14 in the linker region between the RGS and R1 domains (Figs. 4 and 5). This serine and the surrounding sequence are fully conserved among human, rat, and mouse RGS14, making it a likely candidate for 14-3-3 interaction and subsequent regulation of RGS14 activity (Fig. 5).
We next attempted to determine the kinase downstream of H-Ras that facilitates the interaction of 14-3-3 and RGS14 through the phosphorylation of serine 218 of RGS14. Although Scansite3 set on minimum stringency predicted the phosphorylation of serine 218 by CaMKII and Akt (50), which are both 8. 14-3-3␥ blocks RGS14 import into the nucleus of hippocampal neurons. A, primary neurons were transfected with 1 g of YFP-RGS14 and 1 g of empty vector, 1 g of His-tagged 14-3-3␥, or 1 g of His-tagged 14-3-3␥(K50E) per coverslip. Twenty-four hours later, neurons were treated with 40 nM leptomycin B (ϩLMB) or vehicle for 2 h before being fixed in 4% paraformaldehyde in PBS. YFP-RGS14 was visualized by the intrinsic fluorescence of YFP (green). 14-3-3␥ immunostaining was performed using an anti-14-3-3␥ primary antibody and Alexa Fluor 546 (red) secondary antibody, and nuclei were visualized by staining DNA with Hoechst 33258 (blue). Images are representative of at least three separate experiments. Scale bars represent 10 m. B, relative fluorescence intensity of YFP-RGS14, 14-3-3, and nuclear staining was measured through a cross-section of each cell as indicated by the white line through each cell in A.

14-3-3 regulation of RGS14
activated downstream of H-Ras, treatment of cells with selective MEK, PI3K, and CaMKII inhibitors to block ERK, Akt, and CaMKII kinase activity, respectively, had no effect (Fig. S2A). Additional experiments showed that treatment of cells with EGF to activate kinases downstream of activated Ras, including ERK, does not begin to potentiate the interaction between RGS14 and 14-3-3 until approximately 2 h after treatment even though ERK activation occurs very rapidly (Fig. S2B). Taken 9. 14-3-3␥ does not bind at serine 218 of RGS14 to block RGS14 import into the nucleus of hippocampal neurons. A, primary neurons were transfected with 1 g of YFP-RGS14 and 1 g of empty vector, 1 g of His-tagged 14-3-3␥, or 1 g of His-tagged 14-3-3␥(K50E) per coverslip. 24 h later, neurons were treated with 40 nM leptomycin B (ϩLMB) or vehicle (Veh.) for 2 h before being fixed in 4% paraformaldehyde in PBS. YFP-RGS14(S218A) was visualized by the intrinsic fluorescence of YFP (green). 14-3-3␥ immunostaining was performed using an anti-14-3-3␥ primary antibody and Alexa Fluor 546 (red) secondary antibody, and nuclei were visualized by staining DNA with Hoechst 33258 (blue). Images are representative of at least three separate experiments. Scale bars represent 10 m. B, relative fluorescence intensity of YFP-RGS14(S218A), 14-3-3, and nuclear staining was measured through a cross-section of each cell as indicated by the white line through each cell in A.

14-3-3 regulation of RGS14
together, these data suggest that active H-Ras enhancement of the interaction between RGS14 and 14-3-3 likely takes place due to long-term H-Ras activation, likely independent of ERK signaling, that influences changes in transcription and translation in the cell. These data coupled with the failure of the most highly predicted kinases to phosphorylate the 14-3-3-binding site on RGS14 makes identification of a kinase that phosphorylates serine 218 of RGS14 extremely difficult. Although the Scansite3 database takes many kinases into account, not all kinases are included, potentially excluding the kinase relevant to this discussion. Furthermore, there are many kinases for which there is no clear phosphorylation motif, making prediction of phosphorylation by those kinases difficult. H-Ras has many reported effectors that engage many more downstream kinases (56). Thus, at this time, we are unable to identify the kinase responsible for phosphorylating RGS14 at serine 218.

14-3-3 interaction with RGS14 inhibits active G␣ i binding at the RGS domain
The interaction of 14-3-3 with target substrate proteins can elicit various effects, including conformational change, change in subcellular localization, and/or enhancement or inhibition of protein/protein interactions (16,17,19). Looking specifically at RGS proteins, 14-3-3 has been shown in several cases to inhibit G␣ binding at the RGS domain (26, 35). These previously reported effects of 14-3-3 on RGS functions, coupled with the proximity of the RGS14 14-3-3-binding site adjacent to the RGS domain, led us to examine effects of 14-3-3 binding on RGS14 interactions with its known binding partners, active G␣ i/o , active H-Ras, and inactive G␣ i1 . For these studies, ideally, we would examine the effects of 14-3-3 on direct protein/protein interactions using pure proteins to gain insight into the function of the RGS14/14-3-3 interaction. However, our inabil-ity to phosphorylate RGS14 in vitro with the appropriate kinase to promote 14-3-3 interaction limited this approach. In an attempt to mimic phosphorylation of RGS14 at serine 218, we replaced the residue with a phosphomimetic aspartate as well as glutamate. However, these phosphomimetic residues have been shown to be a poor mimic for a phosphate with respect to 14-3-3 binding (25,57,58), and these residue changes failed to promote 14-3-3 binding to RGS14 (data not shown). Therefore, with such limited options, we were forced to analyze interactions between RGS14 and associated proteins in live cells using BRET.
We first determined the effect of 14-3-3 interaction on the binding of active G␣ i at the RGS domain and found that 14-3-3 does indeed decrease active G␣ i interaction with the RGS domain, presumably inhibiting RGS14's negative regulation of G␣ i downstream signaling. Because H-Ras signaling enhances 14-3-3 binding to RGS14 and 14-3-3 binds within the linker region between the RGS and Ras-binding domains of RGS14, we also examined the effect of 14-3-3 on H-Ras interaction with RGS14. Here, we saw no effect, indicating that 14-3-3 interactions with RGS14 selectively inhibit only active G␣ i binding at the RGS domain, leaving intact RGS14 interaction with H-Ras at the R1 domain and inactive G␣ i at the GPR motif. This selective inhibition by 14-3-3 of active G␣ i/o interaction could function as a regulatory mechanism in which RGS14 could promote or enhance signaling through interactions at the R1 domain and GPR motif while selectively silencing RGS14 activity through the RGS domain.

14-3-3 regulation of RGS14
we have previously shown that RGS14 shuttles in and out of the nucleus (5,12,14), we wanted to examine the effect of 14-3-3 interaction on RGS14 nuclear localization. Surprisingly, we found that 14-3-3 prevents RGS14 from translocating to the nucleus but that 14-3-3 binding at the phosphorylation-dependent binding site at serine 218 is not responsible for this effect (Figs. 8 and 9). Further examination showed that 14-3-3 is able to interact with RGS14 in a phosphorylation-independent manner at a unique 14-3-3-binding site on RGS14 (Fig. 10B), providing a bifunctional mechanism by which 14-3-3 can modulate RGS14 nuclear localization at one binding site while regulating active G␣ i binding at another (Fig. 11). However, we should note that the phosphorylation-independent binding of 14-3-3 is only correlated with blockade of RGS14 nuclear import, and we cannot rule out the possibility that 14-3-3 may be interacting with other proteins to indirectly inhibit RGS14 transport into the nucleus. Because this RGS14/14-3-3 interaction is not mediated by phosphorylation of a typical 14-3-3-binding site, there remains the question of how this interaction is regulated in the cell. RGS14 levels in the CA2 region of the hippocampus increase into adulthood and remain stable (41). However, 14-3-3 expression has been shown to change based on conditions in the cell (59). It is possible that expression levels of 14-3-3 may be a unique mechanism by which RGS14 nuclear localization is regulated. Insight into the actual function of 14-3-3's blockade of RGS14 nuclear localization is limited at this time because a function for RGS14 in the nucleus has not yet been determined.

Endogenous RGS14 interacts with 14-3-3 in mouse hippocampus
Within the brain, RGS14 is selectively expressed in the CA2 subregion of the hippocampus (41) where it has been shown to suppress synaptic plasticity and associated learning and memory (3). Importantly, we found that endogenous RGS14 and 14-3-3 can form a stable complex in this part of the brain as shown via coimmunoprecipitation of 14-3-3 with RGS14 from isolated mouse hippocampus (Fig. 1). 14-3-3␥, the 14-3-3 isoform that selectively interacts with RGS14 (Fig. S1), is highly expressed in hippocampal neurons as well (60), putting this isoform in position to regulate RGS14 function. Of note, 14-3-3 proteins are also linked to synaptic plasticity. Functional knockout of 14-3-3 has been shown to inhibit hippocampal LTP and associative learning and memory (18). Additionally, 14-3-3 proteins have been shown to be necessary for activation of the Ras/Raf/MEK/ERK signaling pathway (20,47), which we show here indirectly promotes the phosphorylation-dependent interaction between RGS14 and 14-3-3. Furthermore, this signaling cascade has been shown to be necessary for normal synaptic plasticity and the expression of LTP (61) behind mammalian associative learning (62), which native RGS14 has been shown to suppress in the CA2 region (3). Additionally, it is unclear where and how RGS14 is acting to suppress LTP. Phosphorylation-independent interaction between RGS14 and 14-3-3 could regulate the balance of RGS14 in the cytoplasm and nucleus, changing RGS14's distribution to either reduce or enhance its suppression of LTP in the hippocampus. Because of the effect of 14-3-3 on RGS14's subcellular localization (Figs. 8 and 9) and interaction with functional binding partners, we speculate that 14-3-3 interactions could regulate RGS14 suppression of synaptic plasticity.

Conclusions and future directions
Here, we show that RGS14 interacts with a novel binding partner, 14-3-3, in both a phosphorylation-dependent and phosphorylation-independent manner. The phosphorylation-dependent interaction is promoted by signaling downstream of activated H-Ras. This interaction specifically inhibits interaction of active G␣ i/o at the RGS domain while leaving association with active H-Ras and inactive G␣ i unperturbed. Meanwhile, 14-3-3 also binds within the C-terminal half of RGS14 in a phosphorylation-independent manner and blocks (either directly or indirectly) RGS14 import into the nucleus of neurons. These interactions provide novel mechanisms of regulation of RGS14 interactions and localization, respectively, that give us further insight into RGS14 function in its native environment, the hippocampus, where we demonstrated that endogenous RGS14 interacts with native 14-3-3. Furthermore, both H-Ras and 14-3-3 have been shown to play critical roles in the proper expression of LTP with RGS14 suppressing this form of synaptic plasticity. Future studies will be directed at determining whether and how H-Ras signaling in CA2 pyramidal neurons functions to control phosphorylation-dependent RGS14/14-3-3 interactions. Also of great interest is understanding the role of RGS14 in the nucleus and 14-3-3's role in blocking RGS14 localization there. To our knowledge, this is the first report In resting cells, RGS14 exists in the cytoplasm, capable of shuttling in and out of the nucleus as well as interacting with active G␣ i -GTP at the plasma membrane (PM) via its RGS domain. Long-term H-Ras activation following an extracellular stimulus leads to the up-regulation and/or activation of an unidentified kinase, leading to phosphorylation of serine 218 of RGS14 and binding of 14-3-3␥ at the motif centered around that residue. This phosphorylation-dependent binding disrupts RGS14 interaction with active G␣ i at the plasma membrane while leaving interaction with other binding partners and nuclear localization unaffected (black arrows). RGS14 also interacts directly with 14-3-3␥ in a phosphorylation-independent manner, possibly interacting directly with RGS14 to prevent RGS14 nuclear localization while not affecting interaction with active G␣ i -GTP (red arrows), although it is possible that 14-3-3 interacts with another binding partner to indirectly inhibit RGS14 nuclear localization.

Immunoblotting
Protein samples in Laemmli buffer were loaded onto 11% acrylamide gels, and proteins were resolved via SDS-PAGE before being transferred onto a nitrocellulose membrane. Membranes were blocked at room temperature for 1 h at 4°C overnight in buffer containing 5% nonfat milk (w/v), 0.1% Tween 20, and 0.02% sodium azide in 20 mM TBS, pH 7.6. Membranes were then incubated with primary antibody in the same buffer, except for anti-FLAG, for 2 h at room temperature or overnight at 4°C. Membranes were washed three times for 8 min with TBS containing 0.1% Tween 20 (TBST) before being incubated with either anti-mouse or anti-rabbit HRP-conjugated secondary antibody diluted in TBST (1:5,000 and 1:25,000, respectively) for 45 min at room temperature. Membranes being blotted for FLAG were incubated with anti-FLAG HRP-conjugated primary antibody for 45 min immediately after blocking. Following incubation with HRP-conjugated primary or secondary antibody, membranes were washed three times for 8 min each in TBST before visualizing protein bands using enhanced chemiluminescence and exposing membranes to X-ray films.
Ponceau red stain stock was made from 20 mg/ml Ponceau S in water. Membranes were stained in a 1:100 stock-to-water dilution for 5 min before being rinsed in water and imaged.

14-3-3 regulation of RGS14
Cell culture and transfection HEK 293 cells were incubated in 1ϫ Dulbecco's modified Eagle's medium with phenol red pH indicator supplemented with 10% fetal bovine serum and 100 units/ml each penicillin and streptomycin. Cells were maintained at 37°C in 5% CO 2 . Transfections were carried out using polyethyleneimine (PEI) as described previously (73).
Our protocol for culturing neurons was adapted from Beaudoin et al. (74). Brains were removed from E18 -19 embryos obtained from a timed pregnant Sprague-Dawley rat (Charles River). All animal housings and procedures were approved by the Emory University Institutional Animal Care and Use Committee (IACUC), and all procedures were approved by IACUC protocol. The meninges were removed, and the hippocampi were isolated in calcium-and magnesium-free HBSS (Invitrogen) supplemented with 1ϫ sodium pyruvate (Invitrogen), 0.1% glucose (Sigma-Aldrich), and 10 mM HEPES (Sigma-Aldrich), pH 7.3. Isolated hippocampi were washed with the HBSS solution and then dissociated using the same buffer containing 0.25% trypsin (Worthington) for 8 minutes at 37°C. Trypsinized hippocampi were washed two times with the same HBSS buffer before being triturated five to six times with a fire-polished glass Pasteur pipette in basal medium Eagle (Invitrogen) supplemented with 10% fetal bovine serum (VWR), 0.45% glucose (Sigma-Aldrich), 1ϫ sodium pyruvate (Invitrogen), 1ϫ GlutaMAX (Invitrogen), and 1ϫ penicillin/streptomycin (HyClone). Neurons were counted and plated at a density of 80,000 cells/cm 2 in basal medium Eagle-based buffer on coverslips that had been etched with 70% nitric acid (Sigma-Aldrich) before being coated with 1 mg/ml poly-L-lysine (Sigma-Aldrich) in borate buffer. Cells were allowed to adhere for 1-3 h before medium was changed to Neurobasal (Invitrogen) supplemented with 1ϫ B27 (Invitrogen) and 1ϫ GlutaMAX (Invitrogen). Neurons were kept in a 37°C, 5% CO 2 incubator, and half of the medium was replaced with new Neurobasal every 3-4 days until neurons were used for experiments.
Neurons were transfected using a calcium phosphate transfection kit (Clontech) with a protocol adapted from Jiang and Chen (75). Briefly, calcium solution containing DNA was slowly added to an equal volume of HEPES-buffered saline to form calcium phosphate precipitate containing the DNA. This precipitate was added to neurons on coverslips that had been placed in new wells in new medium. Neurons were incubated with precipitate for 1.5 h before dissolving the precipitate with slightly acidified medium that had been equilibrated in a 10% CO 2 incubator for 15-20 min. Cells were placed back in the 5% CO 2 incubator. After this incubation, coverslips were placed back in their original well in their original culture medium.

RGS14 immunoprecipitation from mouse hippocampus
Brains from C57BL6 WT mice were removed, and hippocampi were isolated by dissection. Hippocampi were homogenized using a Dounce homogenizer in lysis buffer containing 50 mM Tris, pH 7.05, 150 mM NaCl, 1 mM EDTA, 2 mM DTT, 5 mM MgCl 2 , 1 EDTA-free cOmplete mini protease inhibitor tablet (Roche Applied Science) at 1ϫ concentration, 1ϫ Halt phosphatase inhibitor mixture (Thermo Scientific), and 1% Triton X-100. Brain lysate was incubated on a rotator at 4°C for 1 h and then cleared by spinning it down at 13,000 rpm for 10 min. Protein G-Sepharose beads (GE Healthcare) were preblocked with 3% bovine serum albumin (BSA) before lysate was added. Half of the brain lysate was incubated with 50 l of preblocked beads in the absence of RGS14 antibody, and the other half of the lysate was incubated with 50 l of beads along with 5 l (1.63 g) of anti-RGS14 (Proteintech) antibody for 2 h on a rotator at 4°C. Immunocomplexes were then washed four times with PBS containing 0.1% Tween 20 before being boiled off the beads in 2ϫ Laemmli buffer for 5 min. Samples were then resolved via SDS-PAGE and subjected to immunoblotting for analysis. All animal housings and procedures were approved by the Emory University IACUC, and all procedures were approved by IACUC protocol.

FLAG immunoprecipitation
Twenty-four hours following transfection, cells were washed twice in ice-cold PBS and collected in lysis buffer of the same composition used for RGS14 immunoprecipitation from mouse brain. Lysates were rotated at 4°C for 1 h. Lysates were cleared by spinning down samples at 13,000 rpm for 10 min with 30 l put aside to determine total cell lysate expression of proteins of interest, and the remaining supernatant was incubated with anti-FLAG M2 affinity gel (Sigma), which was previously blocked in 3% BSA in PBS for 1 h. Cleared lysates were rotated at 4°C for 1.5 h. Immunocomplexes were then washed three times with 1 ml of PBS containing 0.1% Tween 20 before being suspended in 2ϫ Laemmli buffer, boiled for 5 min, resolved by 11% SDS-PAGE, and transferred to a nitrocellulose membrane for immunoblotting.

Purification of recombinant proteins
RGS14 and truncated RGS14 (N-298 and 300-C) were expressed in E. coli and partially purified as described previously (9). Recombinant hexahistidine-14-3-3␥ was expressed from a pDEST17-based plasmid in BL21(DE3) E. coli cells and purified using Ni 2ϩ -affinity chromatography in the manner described previously (76) with minor modifications. Specifically, cells were lysed using sonication instead of a French pressure cell. Purified protein was snap frozen in liquid nitrogen and stored in aliquots at Ϫ80°C until use.
-Phosphatase treatment of lysates FLAG immunoprecipitation was performed as described above with the exception of the addition of 1ϫ Halt phosphatase inhibitor mixture. Before washing the immunocomplexes and eluting samples, the anti-FLAG affinity gel was resuspended in fresh lysis buffer with the addition of 2 mM MnCl 2 and 900 units of Mn 2ϩ -dependent -phosphatase and rotated at room temperature for 1 h to cleave phosphate groups from potentially phosphorylated serine and threonine residues. Following treatment with phosphatase, lysates were washed, eluted, and analyzed as described above.

Blot overlay "far-Western blotting"
HEK 293 cells were transfected with FLAG-RGS14 in the presence or absence of active H-Ras using PEI. Twenty-four

14-3-3 regulation of RGS14
hours after transfection, FLAG-RGS14 and associated proteins were isolated via FLAG immunoprecipitation, resolved by SDS-PAGE, and transferred to a nitrocellulose membrane. The membrane was then treated with pure His-tagged 14-3-3␥ at 7.5 g/ml in a buffer containing 50 mM Tris, 50 mM NaCl, and 2 mM DTT on a rocking incubator at 4°C overnight. The membrane was then immunoblotted as described above for 14-3-3 to determine whether 14-3-3 directly bound the RGS14 in the membrane.

BRET
BRET experiments were performed as described previously (6,9). Briefly, HEK 293 cells were transfected with a donor and acceptor as well as any additional constructs using PEI. To determine whether 14-3-3 affects active H-Ras association with RGS14, we expressed H-Ras(G12V)-Venus with RGS14 -Luc or RGS14(S218A)-Luc in the presence or absence of 14-3-3␥. To measure YFP-G␣ i1 interactions with the RGS14 RGS domain, Luc-RGS14 was expressed, whereas YFP-G␣ i1 interactions with the RGS14 GPR motif were measured using RGS14 -Luc because the RGS domain and GPR motif are closer to the N and C termini of RGS14, respectively. Luc-RGS14(S218A) was used as a negative control to determine the effect of 14-3-3 interaction with RGS14 on active YFP-G␣ i1 binding at the RGS domain. Twenty-four hours after transfection, cells were suspended in Tyrode's solution containing 140 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 1 mM CaCl 2 , 0.37 mM NaH 2 PO 4 , 24 mM NaHCO 3 , 10 mM HEPES, and 0.1% glucose, pH 7.4. Cells treated with AlF 4 Ϫ were harvested in Tyrode's solution containing the addition of 10 mM NaF, 9 mM MgCl 2 , and 30 M AlCl 3 for 30 min before readings were taken. Cells were plated onto white 96-well OptiPlates (PerkinElmer Life Sciences). Acceptor expression was determined by measuring fluorescence using a TriStar LB 941 plate reader (Berthold Technologies) with 485-and 530-nm excitation and emission filters, respectively. BRET was measured using 485-and 530-nm emission filters after a 2-min application of 5 M coelenterazine H (Nanolight Technologies). The change in BRET (net BRET) was determined by dividing the 530 nm (YFP) signal by the 485 nm (Luc) signal and subtracting the signal from Luc-RGS14 or RGS14 -Luc alone. All data were collected using MikroWin 2000 software and analyzed in Microsoft Excel followed by GraphPad Prism.

Kinase inhibitor treatments
HEK 293 cells were treated with the MEK inhibitor U0126 (Sigma) at 10 M, the PI3K inhibitor LY294002 (Sigma) at 50 M, or the CaMKII inhibitor KN-93 (Sigma) at 100 M for 2 h before cells were collected and lysed. Samples were then submitted to FLAG immunoprecipitation before being analyzed via immunoblotting.

EGF treatment of cells
HEK 293 cells were treated with EGF (Sigma) at 100 ng/ml for 0, 10, 60, or 120 min before being collected and lysed for FLAG immunoprecipitation.

Leptomycin B treatment of cells and immunocytochemistry
Primary hippocampal neurons were treated with LMB (Sigma) at 40 ng/ml or vehicle (70% ethanol) for 2 h before being fixed for 15 min in 4% paraformaldehyde in 1ϫ PBS. Cells were then permeabilized for 10 min in PBS containing 0.1% Triton X-100 and incubated in 8% BSA in PBS blocking buffer for 1 h before being incubated with primary antibody in 4% BSA in PBS for 2 h. Following three washes with 0.05% Triton X-100 in PBS, coverslips were incubated with secondary antibody in 4% BSA in PBS for 1 h before being stained with 1.6 M Hoechst dye (Thermo Fisher) for 4 min. Coverslips were then washed three times in PBS before being mounted on slides using ProLong Diamond antifade mountant (Thermo Fisher). Images were taken on an Olympus FV1000 inverted confocal microscope at 100ϫ and then processed using ImageJ software. Approximately 10 images per condition were obtained, and representative images are shown.