Ribosomal Protein S6 Phosphorylation in the Nervous System: From Regulation to Function

Since the discovery of the phosphorylation of the 40S ribosomal protein S6 (rpS6) about four decades ago, much effort has been made to uncover the molecular mechanisms underlying the regulation of this post-translational modification. In the field of neuroscience, rpS6 phosphorylation is commonly used as a readout of the mammalian target of rapamycin complex 1 signaling activation or as a marker for neuronal activity. Nevertheless, its biological role in neurons still remains puzzling. Here we review the pharmacological and physiological stimuli regulating this modification in the nervous system as well as the pathways that transduce these signals into rpS6 phosphorylation. Altered rpS6 phosphorylation observed in various genetic and pathophysiological mouse models is also discussed. Finally, we examine the current state of knowledge on the physiological role of this post-translational modification and highlight the questions that remain to be addressed.


INTRODUCTION
The eukaryotic ribosome is composed of the small 40S and the large 60S subunits, comprising together 4 ribosomal RNA species and 79 ribosomal proteins (Kressler et al., 2010). In many organisms, ribosomal proteins undergo various post-translational modifications, including phosphorylation, acetylation, methylation, O-linked β-N-acetylglucosaminylation, and ubiquitylation (Xue and Barna, 2012). Historically, the phosphorylation of the 40S ribosomal protein S6 (rpS6) was the first post-translational modification described (Gressner and Wool, 1974). The presence of phospho-rpS6 (p-rpS6) at different levels in a 2D gel provided the first evidence that rpS6 phosphorylation could occur at several residues (Lastick et al., 1977). Ensuing studies identified five evolutionary conserved and clustered carboxy-terminal phospho-sites, which undergo phosphorylation in an ordered manner, beginning with Ser236 and followed sequentially by Ser235, Ser240, Ser244, and Ser247 (Martin-Pérez and Thomas, 1983;Wettenhall et al., 1992;Meyuhas, 2008Meyuhas, , 2015. Intriguingly, the exact function of the post-translational modification of this indispensable ribosomal protein remains enigmatic. Despite the large debate regarding its physiological role, rpS6 phosphorylation is commonly used as a marker for neuronal activity and a readout of mammalian target of rapamycin complex 1 (mTORC1) activity (Meyuhas, 2008(Meyuhas, , 2015Mahoney et al., 2009;Knight et al., 2012). This review summarizes our current knowledge regarding the molecular mechanisms as well as the variety of stimuli modulating rpS6 phosphorylation in the nervous system.
Although the activation of S6K and the subsequent phosphorylation of rpS6 are commonly used as a readout of mTORC1 activation, several evidences point out the existence of a synergistic crosstalk between mTORC1 and the extracellular signal-regulated kinase (ERK) signaling to control rpS6 phosphorylation. Thus, ERK can promote S6K activation by enhancing its phosphorylation at T421/S424 sites (Mukhopadhyay et al., 1992). When phosphorylated, these sites located in the autoinhibitory domain of the S6K1/2 are thought to prime the activation of S6K, thereby facilitating the subsequent phosphorylation of the other sites of S6K by the upstream kinases (Dennis et al., 1998). Alternatively, ERK can also modulate the mTORC1/S6K cascade upstream of S6K. On the one hand, ERK-mediated inhibitory phosphorylation of Tuberous Sclerosis Complex 2 (TSC2) stimulates the Ras Homolog Enriched in Brain (Rheb) protein, which in turn activates mTORC1 (Roux and Blenis, 2004;Long et al., 2005;Ma et al., 2005). On the other hand, ERK can enhance mTORC1 activation through RSK-mediated phosphorylation of Raptor (Wettenhall et al., 1992). Examples of this synergistic interaction between ERK and mTORC1 in the regulation of rpS6 phosphorylation have been reported in several models and in various brain areas (Kelleher et al., 2004;Antion et al., 2008a;Gobert et al., 2008;Santini et al., 2009;Gangarossa et al., 2014) Interestingly, ERK can also regulate rpS6 phosphorylation at S235/236 through RSK independently of mTORC1/S6K signaling (Roux et al., 2007). Indeed, the increase in pS235/236-rpS6 promoted by tetraethylammonium in cultured cortical neurons is prevented by RSK3 inhibition (Gu et al., 2015). Finally, a recent study performed in hippocampal neurons demonstrated that the cdk5-dependent phosphorylation of S6K at S411 site is also critical in the regulation of S6K activation and the subsequent rpS6 phosphorylation at S235/236 sites (Lai et al., 2015).

cAMP/PKA Pathway
The enhanced rpS6 phosphorylation in the cerebral cortex following the administration of N6O-2 ′ dibutyryl cAMP was one of the first demonstrations that cAMP could modulate in vivo the state of phosphorylation of rpS6 (Roberts and Morelos, 1979). Despite this evidence, the contribution of the cAMP/PKA pathway in the regulation of rpS6 phosphorylation in the nervous system has been largely neglected. However, recent studies highlighted the importance of cAMP/PKA signaling in the regulation of rpS6 phosphorylation at S235/236 sites. Thus, the direct stimulation of the adenylate cyclase by forskolin increases pS235/236-rpS6 in the striatum and the hippocampus (Gobert et al., 2008;Biever et al., 2015) ( Table 1). Similar results are obtained when the degradation of cAMP is prevented by the administration of papaverine, a potent inhibitor of type 10A phosphodiesterase (Biever et al., 2015). Although the demonstration that PKA directly catalyzes rpS6 phosphorylation in the brain is still lacking, several indirect evidences support its contribution in the control of pS235/236-rpS6 following cAMP elevation. Indeed, forskolin-induced rpS6 phosphorylation in striatal slices is reduced in the presence of a PKA inhibitor (Biever et al., 2015). Moreover, stimulation of PKA activity with the cAMP analog cBIMPS increases pS235/236-rpS6 in striatal culture (Valjent et al., 2011) (Table 1). Finally, the administration of pharmacological agents promoting PKA activation triggers rpS6 phosphorylation in several brain areas (Gobert et al., 2008;Valjent et al., 2011;Knight et al., 2012;Bonito-Oliva et al., 2013;Rapanelli et al., 2014;Biever et al., 2015;Sutton and Caron, 2015).
Although PKA targets selectively S235/236 residues (Moore et al., 2009), recent evidences suggest that PKA also contributes indirectly to rpS6 phosphorylation through a protein phosphatase cascade. This mechanism has been particularly well-studied in the striatum, where the inhibition of PP-1, controlled by the PKA-dependent phosphorylation of dopamine-and cAMP-regulated phosphoprotein, Mr 32,000 (DARPP-32) at T34 (Hemmings et al., 1984;Greengard, 2001), promotes pS235/236-rpS6 induced by d-amphetamine or haloperidol (Valjent et al., 2011;Bonito-Oliva et al., 2013;Biever et al., 2015). This mechanism also contributes to the regulation of rpS6 phosphorylation at S240/244 sites (Bonito-Oliva et al., 2013). These findings highlight the importance of PKA/DARPP-32/PP-1 signaling in the regulation of rpS6 phosphorylation in the striatum and raise the intriguing hypothesis that similar mechanisms could take place in other brain areas.

STIMULI MODULATING rpS6 PHOSPHORYLATION IN THE BRAIN
Recently an increasing number of studies have used rpS6 phosphorylation as a marker for neuronal activation in the context of synaptic plasticity or in response to variety of therapeutic agents in physiological and pathophysiological contexts.   When not specified, a single dose of drug was administered; *repeated doses of drug were administered. NS, not specified; ND, not determined; DG, Dentate Gyrus; MSNs, Medium-sized Spiny Neurons; THC, tetrahydrocannabinol; PTZ, Pentylenetetrazol; 6-OHDA, 6-hydroxydopamine; L-DOPA, levodopa; MPTP, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine.

Synaptic Plasticity
Increased rpS6 phosphorylation during synaptic plasticity was reported for the first time by Klann and colleagues in 1991 using a synthetic peptide of rpS6 containing the residues 222-249. Since then, enhanced rpS6 phosphorylation has been observed in several electrical or chemical models of synaptic plasticity. Thus, pS235/236-rpS6 increases in the CA1 subfield of the hippocampus during long-term potentiation following high frequency stimulation (Antion et al., 2008b) or forskolin application (Kelleher et al., 2004;Antion et al., 2008b;Gobert et al., 2008). Similarly, mGluR-dependent long-term depression induced by application of the mGluR1 agonist [3,5-RS] dihydroxyphenylhydrazine (DHPG) is associated with marked increases in pS235/236-and pS240/244-rpS6 in hippocampal slices (Antion et al., 2008a). Interestingly, a recent study reported that the state of phosphorylation of rpS6 at S240/244 sites could be used to estimate the neuronal activity state of striatal cholinergic interneurons (Bertran-Gonzalez et al., 2012).

Pharmacological Stimuli
A large number of pharmacological stimuli have been described to promote rpS6 phosphorylation in neurons (Tables 1, 2). Indeed, several ex vivo studies performed in slices or neuronal cultures showed that rpS6 phosphorylation is enhanced by stimuli triggering multiple forms of neuronal activity ( Table 1). In vivo, the phosphorylation of rpS6 has been assessed following a single or repeated administration of a large variety of pharmacological agents in various brain areas ( Table 2). Thus, proconvulsant drugs such as kainate, pilocarpine, pentylenetetrazol (PTZ), or the dopamine D1R agonist SKF81297 lead to a massive increase in pS235/236-and pS240/244-rpS6 in principal cells in the hippocampus and in various cortical areas (Table 2). Moreover, the administration of drugs of abuse (cocaine, d-amphetamine, methamphetamine, morphine, and tetrahydrocannabinol) as well as antipsychotics (haloperidol, clozapine, and olanzapine) also regulates rpS6 phosphorylation at multiple sites in several brain areas including  the striatum, the nucleus accumbens, the cortex, and the hippocampus ( Table 2). Finally, several hormones involved in the regulation of energy balance enhance rpS6 phosphorylation in hypothalamic nuclei ( Table 2).

Physiological and Pathophysiological Conditions
Since the pioneering report demonstrating that rpS6 phosphorylation was enhanced in the CA1 subfield of the hippocampus in mice trained to contextual fear conditioning (Kelleher et al., 2004), rpS6 phosphorylation has been used as a marker of neuronal and circuit activation following physiological conditions (Table 3). Thus, phospho-rpS6 levels oscillate in the hippocampus and the suprachiasmatic nucleus of the hypothalamus along the circadian cycle (Table 3). rpS6 phosphorylation is also strongly modulated in the amygdala and the hypothalamus when defensive behaviors (freezing, escape or attacks) are engaged or in the hypothalamus in response to nutritional perturbations (Table 3). Finally, altered rpS6 phosphorylation has been reported in rodents following spontaneous seizures and traumatic brain injury and in humans in several neurodevelopmental disorders including Down syndrome, Tuberous sclerosis, Autism, and Rett syndrome ( Table 3).

Genetic Mouse Models Displaying Altered rpS6 Phosphorylation
Most of the full or conditional knockout mice for the key components of the mTORC1 pathway display altered rpS6 phosphorylation (Table 4). Interestingly, the vast majority of mutant mice in which dysregulation of rpS6 phosphorylation has been demonstrated correspond to mouse models for various neurological and neurodevelopmental disorders, including Tuberous sclerosis, Down syndrome, Rett syndrome, Angelman syndrome, and Fragile X syndrome, among others. Most of these pathologies share common features such as autism, intellectual disability, and epilepsy, which might be rescued by mTORC1 inhibitors under certain circumstances. The phosphorylation of rpS6 is also altered in neurodegenerative diseases such as Huntington disease and in mouse models of psychiatric disorders such as schizophrenia ( Table 4). Other genetic mouse models showing hormonal perturbations as leptin deficiency also display altered rpS6 phosphorylation in the hypothalamus ( Table 4).

Role in Overall mRNA Translation
Despite the growing number of reports analyzing the phosphorylation of rpS6, its biological significance still remains controversial. One of the first hypotheses put forward suggested that rpS6 phosphorylation played a role in translation initiation. Thus, an early in vitro study reported a correlation between the phosphorylation of rpS6 and enhanced translation under certain experimental conditions (Thomas et al., 1982). Moreover, the 40S subunit with a highest proportion of phosphorylated rpS6 was preferentially found into polysomes compared to subpolysomal fractions (Duncan and McConkey, 1982). However, several studies rapidly called into question this hypothesis. Indeed, although localized at the mRNA/tRNA binding site junction between the small and large ribosomal subunits (Nygård and Nilsson, 1990), increased rpS6 phosphorylation is not sufficient to mobilize small ribosomal subunits into protein synthesis (Kruppa and Clemens, 1984;Tas and Martini, 1987). Finally, the generation of rpS6 knockin mice, in which the five phosphorylated serines were replaced by alanines, constituted a valuable tool to determine whether rpS6 phosphorylation and the protein synthesis were causally linked (Ruvinsky et al., 2005). Unexpectedly, protein synthesis is increased in mouse embryo fibroblasts (MEF) of phospho-deficient mice. Moreover, a similar (Ruvinsky et al., 2005) or even increased (Chauvin et al., 2014) proportion of ribosomes engaged in translation were found in the liver of rpS6 knockin mice. Together, these puzzling observations suggest a negative role of rpS6 phosphorylation on  global protein synthesis or the presence of feedback mechanisms taking place in this mouse model.

Role in TOP mRNA Translation
The phosphorylation of rpS6 through the mTORC1/S6K axis was believed for many years to exert an effect on the translation of a specific subset of mRNAs bearing a 5 ′ terminal oligopyrimidine tract (TOP). However, this long-lasting model has been challenged by subsequent studies showing that MEFs from the double mutant S6K1/2 as well as from the rpS6 knockin mouse lines exhibit normal TOP translation (Tang et al., 2001;Stolovich et al., 2002;Ruvinsky et al., 2005). Further work demonstrated that insulin-induced TOP translation requires the PI3K/TSC/Rheb/mTOR pathway but is independent of the S6K/rpS6 axis (Patursky-Polischuk et al., 2009). Despite these evidences, the involvement of rpS6 phosphorylation in the control of translation in the nervous system is still controversial. Indeed, several findings in hippocampal neurons and slices correlated increased rpS6 phosphorylation with enhanced global and TOP-encoded protein synthesis following different forms of synaptic plasticity (Kelleher et al., 2004;Klann and Dever, 2004;Tsokas et al., 2005Tsokas et al., , 2007Antion et al., 2008a,b) or in a mouse model of fragile X syndrome . By contrast, such correlations have not been observed in vivo in the striatum where the pharmacologically-induced rpS6 phosphorylation by damphetamine, haloperidol, or papaverine relies on the activation of the cAMP/PKA/DARPP-32 pathway (Biever et al., 2015). Interestingly, the direct binding of mRNA to the small ribosomal subunit decreases or increases after cAMP-dependent or cAMPindependent phosphorylation of rpS6, respectively (Burkhard and Traugh, 1983). Therefore, one cannot exclude that when an upstream signaling cascade other than PKA is engaged, for example mTORC1/S6K or ERK/RSK, rpS6 phosphorylation might have different physiological roles. Nevertheless, despite compelling studies indicating that rpS6 phosphorylation is dispensable for efficient global and TOP mRNA translation, the role of the phosphorylation in the translation of a specific subset of mRNAs remains to be fully addressed by high-throughput sequencing analyses of total and polysomal RNAs combined with proteomic approaches.

Extratranslational Functions
Another possibility is that rpS6 phosphorylation, within or outside the ribosome, exerts functions unrelated to mRNA translation, for example by interacting with other cellular proteins, which might become active or inactive upon the binding with rpS6. Indeed, co-immunoprecipitation studies suggest either a direct or indirect interaction of rpS6 with several extraribosomal proteins, including heat-shock protein 90 (Kim et al., 2006), alphavirus non-structural protein (Montgomery et al., 2006), DAPK (Schumacher et al., 2006), huntingtin (Culver et al., 2012), and mTOR complex 2 (mTORC2) (Yano et al., 2014). In the latter, rpS6 phosphorylation has been proposed to have a role in cardioprotective signaling by amplifying mTORC2-mediated Akt phosphorylation (Yano et al., 2014). In the mouse liver, Chauvin and colleagues recently uncovered the involvement of rpS6 phosphorylation in the control of the ribosome biogenesis (RiBi) transcriptional program by S6Ks (Chauvin et al., 2014). This program regulates the expression of nucleolar proteins required for ribosomal RNA synthesis, cleavage, post-transcriptional modifications, ribosome assembly, and export (Lempiäinen and Shore, 2009). Whether all these translation-unrelated responses occur in the brain merits further study.

Extraribosomal Functions
Surprisingly, extraribosomal functions have been attributed to several ribosomal proteins (Wool, 1996;Warner and McIntosh, 2009). For instance, the ribosomal protein rpL13a, when phosphorylated, is released from the 60S ribosomal subunit and acts as a silencer of targeted mRNAs (Mazumder et al., 2003). In this regard, few studies suggested an extraribosomal role of rpS6 phosphorylation (Kim et al., 2014;Son et al., 2015;Xiao et al., 2015). Recent work in the plant Arabidopsis Thaliana proposes a role of rpS6 in rRNA synthesis and rDNA transcription via its interaction with the histone deacetylase AtHD2B (Kim et al., 2014) and the histone chaperon AtNAP1 (Son et al., 2015), respectively, an effect that might be dependent on the phosphorylation state of rpS6. Finally, the ubiquitylation and proteasomal degradation of phosphorylated rpS6 following its subsequent interaction with PALL has been identified as a critical mechanism regulating efferocytosis in drosophila (Xiao et al., 2015). To date, such extraribosomal functions of rpS6 in the nervous system have not been described.

CONCLUDING REMARKS
Since the pioneering studies performed four decades ago, many progresses have been made regarding the identification of signaling events leading to rpS6 phosphorylation. Although rpS6 phosphorylation is still considered as a readout of mTORC1/S6K activity, caution should be taken with this assumption since other intracellular cascades largely contribute to the regulation of rpS6 phosphorylation, as exemplified with the PKA/DARPP-32/PP-1 pathway in the striatum. One should also keep in mind that the different sites of phosphorylation can be regulated independently in various brain areas or different cell-types within a same brain region. Although rpS6 phosphorylation has been and will remain a valuable hallmark of neuronal activity, understanding its biological role in the brain is undoubtedly one of the major challenges of the coming years.

AUTHOR CONTRIBUTIONS
AB, EV, and EP wrote the manuscript.

ACKNOWLEDGMENTS
This work was supported by Inserm, Fondation pour la Recherche Médicale (EV), and a NARSAD Young Investigator Grant from the Brain and Behavior Research Foundation (EP). AB is supported by the Fonds National de la Recherche, Luxembourg (Grant 3977033). EP is a recipient of Marie Curie Intra-European Fellowship IEF327648. The authors apologize to all colleagues whose work was not discussed in this Review.