Increasingly, we appreciate that biogenic amine transporters typified by SERT are under complex regulatory control mediated by associated proteins and multiple signaling pathways [9–11, 17, 18, 22, 23, 25, 36, 38–50]. Understanding the molecules and contingencies involved in SERT regulation is likely to enhance our abilities to manipulate SERT for therapeutic ends, as well as expand our understanding as to how SERT dysregulation contributes to mental illness [16, 17, 20, 21]. Miller and Hoffman first identified a cGMP-linked pathway promoting rapid enhancement of SERT activity following adenosine receptor activation . Thus, 8-Br-cGMP treatments recapitulate the actions of nonselective adenosine receptor agonists on RBL-2H3 cells (e.g. 5'-N-ethylcarboxamidoadenosine, NECA) and these authors hypothesized that cGMP-stimulated protein kinases might direct phosphorylation of SERT or a SERT-associated protein to trigger increased SERT activity. We now understand that A3ARs and/or 8-Br-cGMP treatments can increase SERT activity in multiple contexts, including the rat raphe cell line RN46A , transfected CHO and HeLa cells [17, 18], rat midbrain and cortical synaptosomes in vitro and the mouse hippocampus in vivo . In all cases, pharmacological inhibition of PKG blocks stimulation of SERT activity. Furthermore, biochemical and surface expression studies reveal that PKG activators lead to both SERT phosphorylation [17, 18, 47] and enhanced SERT insertion into the plasma membrane [17, 19, 23], providing an opportunity for steady-state increases in 5-HT uptake capacity. SERT trafficking regulation leads to p38 MAPK activation that we have shown shifts the transporter to a higher affinity state for 5-HT recognition . Progress has also been achieved in identifying a number of SERT-interacting proteins [9–11, 43, 49] whose associations can impact both SERT trafficking and transport rates. In this report, we provide evidence that PKGIα is responsible for rapid, 8-Br-cGMP-dependent, DT-2-sensitive SERT stimulation. Moreover, we demonstrate that mammalian cells can organize a stable association of PKGIα with SERT, an interaction likely to support the enhanced surface trafficking of the transporter evident after treatment of cells with PKG activators.
In RN46A cells, we established using qualitative and quantitative approaches that endogenous PKGI and SERT colocalize. Although PKGII could not be detected by RT-PCR, low levels of PKGII immunoreactivity were evident via confocal-assisted immunofluorescence studies. However, no evidence for SERT/PKGII colocalization was evident. Interestingly, SERT/PKGI colocalization was nonuniform, being most prominent in permeabilized preparations and absent in the distal segments of nonpermeabilized cells. One possible explanation for these findings is that SERT recruits PKGI to vesicles as part of a coordinated mechanism to achieve regulated plasma membrane trafficking and that, in distal processes, SERT moves to the plasma membrane after dissociation of PKGI. Extensive studies to date in our lab have failed to document consistent interactions between PKGI and SERT (Steiner and Blakely, unpublished findings), indicating a role for an intermediate in formation of SERT/PKGIα complexes. This intermediate could result in a SERT (or PKG) post-translational modification or represent a SERT-associated protein lacking under the conditions of in vitro incubations of purified proteins. Additional studies are clearly needed to elaborate the manner by which transporter/kinase associations are formed. Future studies should also explore dynamic features of the SERT/PKGI association, such as whether these interactions change upon vesicle membrane fusion and localization of SERT to plasma membrane subdomains.
PKG activation has been found to alter actin dynamics and vesicular trafficking [51, 52] as well as promote phosphorylation of a single Thr (Thr276) on SERT whose mutation can eliminate SERT stimulation . Whether Thr276 is phosphorylated directly by PKG, or possibly by p38 MAPK, which PKG is known to activate [23, 36, 53], is unknown. Since Thr276 is positioned within a short intracellular loop connecting transmembrane domains proposed to control 5-HT exit from the transporter [18, 54], it seems most likely that this site has relevance for catalytic modulation, which we have shown is governed by p38 MAPK [23, 36]. PKGI might then contribute to mobilization of SERT vesicles not by phosphorylation of SERT, but by phosphorylation of SERT vesicle tethering proteins. However, it is also possible that additional phosphorylation sites targeted by PKGI may exist; these could support transporter trafficking but be too transient to be captured by standard biochemical methods. In addition to the PKGIα associations we have defined, SERT associates with the phosphatase PP2A . The specific sites in SERT that bind, or are dephosphorylated by PP2A, are not known, but a polymorphism (Gly56Ala) within the hSERT amino terminus leads to constitutive phosphorylation that cannot be further elevated by 8-Br-cGMP. Moreover, the Gly56Ala is a catalytically activated variant [17, 19] that also displays diminished sensitivity to the PP2A specific inhibitor fostriecin . Thus, it is possible that the SERT/PKGIα associations documented here are mutually exclusive with, or are modulated by SERT/PP2A interactions and that these two enzymes govern phosphorylation sites linked to SERT trafficking, as well as catalytic modulation. Interestingly, p38 MAPK is known to phosphorylate and activate PP2A , providing a path by which PKGI activation could lead to both p38 MAPK-dependent transporter catalytic activation and efficient dephosphorylation of trafficking regulatory SERT sites. A mirror image of this idea related to SERT inactivation and internalization has been proposed involving PKC and PP2A by our lab [40, 43] and the work of Ramamoorthy's group . Future studies that elucidate the sites targeted by PP2A and PKC should greatly clarify these ideas. Since OCD and autism-associated SERT variants appear to impact PKG, p38 MAPK and PP2A-dependent transporter regulation , this effort may also have significant impact on our understanding of the origins of 5-HT linked brain disorders.
We have recently proposed by analogy to the classical picture of insulin-dependent glucose transporter regulation  that two modes of SERT trafficking to the plasma membrane exist: one, a finite capacity, regulated pathway overseen by PKG, p38 MAPK, PP2A and PKC, and a second, constitutive pathway that is relatively insensitive to these proteins but which can be more greatly populated by heterologous overexpression. RN46A cells endogenously express a low level of SERT and display substantial elevations (≥ 200%) in SERT activity following 8-Br-cGMP treatment or p38 MAPK stimulation . High-level heterologous expression of SERT in transfected cells, in our hands, leads to SERT proteins that demonstrate greatly reduced, and in some cases, no cGMP-dependent regulation. Due to this characteristic, we routinely constrain SERT expression in transfected cells where SERT regulation is being assessed, including the HeLa model we have used in this study. Our goal is to approach the low level of SERT expression observed in RN46A and RBL-2H3 cells. Unfortunately, such low expression levels make biochemical studies, such as SERT/PKGI co-immunoprecipitation experiments, extremely challenging, if not impossible. Although we have utilized the higher expressing HEK-293T model for our co-immunoprecipitation studies, our goal here was not to parallel functional studies, but rather to determine whether a portion of PKGI can assemble in a SERT complex and whether PKGI isoform specificity is evident. Additionally, we expect the pool of PKGI that associates with SERT to be small, and likely differentially localized, compared to the total PKGI pool. This possibility likely explains why we do not observe ICQ values for SERT and PKGI approaching 0.5, the theoretical maximum, even in permeabilized cells. Our blots of siRNA-treated HeLa extracts demonstrate only a modest reduction in total PKGI level [see Additional file 2] despite full elimination of 8-Br-cGMP triggered SERT stimulation. We also observed a full elimination of p38 MAPK-dependent regulation of SERT after only a modest reduction in p38α MAPK level by p38α MAPK-targeted siRNA treatment . Since siRNAs are cotransfected with hSERT to evaluate transporter regulation, we thought it important to determine whether the full inhibition of regulation could be due to full suppression of kinase expression, though limited to the small percentage of cells transfected. Indeed, when we determined the percentage of cells transfected with siRNA using toxic siRNAs [see Additional file 2], we found a good match to the percent reduction in PKGI protein (~15% in both cases).