Preso1, mGluR5 and the machinery of pain

Abstract

By bringing mGluR1/5 and proline-directed kinases together, the scaffold protein Preso1 stabilizes the interaction between mGluR1/5 and Homer. This mechanism may attenuate calcium influx into spinal neurons and reduce pain.

The type I metabotropic glutamate receptors mGluR1 and mGluR5 (mGluR1/5) regulate complex behaviors ranging from pain to emotion and cognition^1–3. Consequently, molecules that regulate their function are considered to be promising drug targets for major neuropsychiatric illnesses, such as chronic pain, affective disorders, autism and schizophrenia³. At a cellular level, the actions of mGluR1/5 are determined by their ability to induce localized changes in calcium (Ca²⁺) concentration^4,5 and protein kinase signaling⁶. Association with different isoforms of Homer proteins determines the activation state of mGluR1/5 and thereby constrains or enhances mGluR1/5-induced signal propagation. In this issue, Hu et al.⁷ identify the scaffold protein Preso1 (ref. 8), encoded by the FERM and PDZ domain containing 4 gene (Frmpd4), as a regulator of interactions between mGluR1/5 and Homer.

Scaffold proteins are crucial coordinators of receptor function. By anchoring specific kinases or phosphatases close to membrane receptors, scaffold proteins incorporate input specificity into a system in which a limited number of signaling pathways is required for multiple functional outcomes⁹. There are several known scaffolds that regulate mGluR1/5 activity, localization and signaling. Shank3 links mGluR1/5 to NMDA receptors via PSD-95 (ref. 10), tamalin (GRASP) regulates mGluR1/5 trafficking¹¹, and Na⁺/H⁺ exchanger regulatory factor 2 prolongs mGluR5-mediated Ca²⁺ mobilization¹². Most extensively studied, however, are the interactions between mGluR1/5 and Homer, which determine the effects of mGluR1/5 on Ca²⁺ homeostasis⁴, protein kinase signaling and behavior¹.

Hu et al.⁷ now demonstrate that another scaffold protein, Preso1, interacts with both mGluR1/5 and Homer. To address the function of Preso1 in mGluR1/5/Homer interaction, Hu et al.⁷ first conducted a series of biochemical experiments demonstrating that Preso1 interacts with mGluR1/5 and the proline-directed protein kinases cyclin-dependent kinase 5 (CDK5) and extracellular signal–regulated kinase (ERK). Preso1 contains one WW, one FERM and two PDZ domains, which associate with various intracellular effectors regulating dendritic spines⁸. Hu et al.⁷ found that the FERM domain of Preso1 is critical for binding to the C-terminal tail of mGluR5 and to proline-directed kinases. The binding site (amino acid sequence between 920 and 1,020) is upstream of the proline-rich motif (TPPSPF) of mGluR1/5, which interacts with the N-terminal EVH1 (Ena-VASP homology) domain of Homer. The mGluR1/5-Preso1-kinase complex was found to trigger phosphorylation of mGluR1/5 and enhance its interaction with the long Homer isoforms (Homer 1b, 1c, 2 and 3).

Next, Hu et al.⁷ investigated the effect of Preso1 on mGluR1/5 activity by monitoring the Ca²⁺ response of dorsal spinal neurons to stimulation by glutamate. Preso1-deficient neurons showed a greater mGluR1/5-dependent increase of Ca²⁺ through L-type voltage-sensitive Ca²⁺ channels than did wild-type neurons. This effect was reversed when exogenous Preso1 was introduced into the cells. To determine whether proline-directed kinases regulate this mGluR1/5-Preso1 function, Hu et al.⁷ tested whether inhibitors of these kinases could affect Ca²⁺ influx. Consistent with their hypothesis, inhibitors of ERK or CDK5, which induced a late rise in Ca²⁺ in wild-type neurons, did not further increase Ca²⁺ in Preso1-lacking neurons. From these findings, the authors conclude that Preso1 and proline-directed kinases suppress Ca²⁺ influx and suggest that this is a result of the newly identified interaction between mGluR1/5, Preso1 and Homer.

In a final set of experiments, Hu et al.⁷ examined the physiological function of Preso1 in mouse models of inflammatory pain. mGluR5 activity is thought to enhance sensitization to pain⁵, as do two of its main downstream effectors, Ca²⁺ (ref. 5) and ERK¹³. Mice lacking Preso1 showed stronger pain responses than their wild-type littermates, an effect that could have been a result either of weakened mGluR1/5-Homer interactions or possibly of other actions of Preso1. Supporting the former possibility, the authors found that expression of a mutant mGluR5 that could not bind to Homer also resulted in enhanced inflammatory pain, as did deletion of Homer2 plus Homer3.

Taken together, these findings indicate that Preso1 constrains mGluR5-mediated Ca²⁺ influx and inflammatory pain (Fig. 1a). Hu et al.’s⁷ findings are consistent with the postulated roles of mGluR5 and Ca²⁺ signaling in pain⁵. However, the data also add to the complexity and controversy on the role of Homer proteins in mGluR5-mediated signaling and pain regulation.

Figure 1.

Signaling and pain mediated by mGluR1/5-Homer interaction. (a) A mouse model of inflammatory pain and underlying signaling pathways identified by Hu et al.⁷ using constitutive global knockout of Frmpd4 (Preso1). After injection of an irritant into the footpad, pain signals reach the spinal synapses and are subsequently relayed to the pain centers in the brain (left). This process requires mGluR1/5 activity (right). Preso1 interacts with mGluR1/5, CDK5, ERK and Homer. This complex favors phosphorylation of mGluR1/5, which in turn enhances interaction with Homer1, 2 and 3. Consequently, Ca²⁺ influx through L-type voltage-sensitive calcium channels (VSCC) and inflammatory pain are reduced (right). (b) An alternative model proposed by Tappe et al.⁵ on the basis of genetic manipulations of Homer1a in the spinal cord. Interactions between mGluR1/5 and Homer 1–3 enhance Ca²⁺ influx and pain, whereas Homer1a has the opposite effect. Binding of mGluR1/5 to Homer1, 2 and 3 also enhances protein kinase signaling^1,3,6. IP₃R, inositol trisphosphate receptor; PSD-95, postsynaptic density protein 95; RyR, ryanodine receptor; SAPAP, synapse-associated protein 90 (also known as GKAP (guanylate kinase–associated protein) and postsynaptic density-95–associated protein); SHANK, SH3 and multiple ankyrin repeat domains protein.

Although it is broadly acknowledged that Homer proteins exhibit region-, cell type–and stimulation-specific roles in mGluR-dependent Ca²⁺ signaling^4,5,14, Hu et al.’s⁷ findings contrast with earlier observations that mGluR5-Homer interactions mediate enhanced, rather than reduced, Ca²⁺ influx in dorsal spinal cord neurons⁵. Similarly, disruption of mGluR5-Homer signaling by Homer1a, the dominant-negative short isoform of Homer, has been implicated in suppression of Ca²⁺ signaling and pain (Fig. 1b), whereas Hu et al.’s⁷ results propose the opposite: that Homer1a enhances both processes (Fig. 1a). Although several key differences in methodological approaches, such as constitutive versus inducible and complete versus spinally localized manipulations of Homer1a, could be responsible for some of the reported differences, it is remarkable that the reported effects are quite opposite. Several lines of evidence, as discussed below, may be particularly relevant to resolving the discrepancies in the role of mGluR-Homer inter actions in Ca²⁺ signaling and pain.

Compartmentalization of Ca²⁺ and protein kinase signaling by scaffolding proteins provides critical specificity to neuronal inputs⁹. mGluR1/5 regulates several sources of Ca²⁺, encompassing intracellular stores^4,10, NMDA receptors¹⁰ and L-type voltage sensitive Ca²⁺ channels⁷ (Fig. 1). Thus, if Preso1 and Homer target non-overlapping sources of Ca²⁺ influx, their effects on both Ca²⁺ and pain may be distinctly different. In addition to Ca²⁺, the neuronal plasticity underlying inflammatory pain depends on protein kinase signaling. Notably, strengthening the mGluR1/5-Homer interaction facilitates the ERK and phosphatidylinositol-3-OH kinase (PI3K) signaling pathways^3,6, which markedly contribute to pain¹⁵. Thus, an important goal of future studies will be to establish whether Preso1 downregulates both Ca²⁺ and protein kinase signaling or has divergent effects. The former possibility would be consistent with reduced pain.

The discovery of Preso1 as a scaffold for proline-directed kinases and mGluR1/5 highlights the importance of compartmentalized signal transduction in behavior. mGluR1/5 docks to many proteins, and it is likely that some, but not all, of their actions involve Preso1 or Homer. In turn, Preso1 has several binding motifs for other molecules, and it is unlikely that all of its effects are mediated by mGluR1/5. Delineating the roles of individual scaffolds of mGluR1/5 in the specific cellular and regional context of neuronal function will provide a more complete view of the role of Preso1 and mGluR1/5 in regulation of pain. For example, how Preso1 interacts with the short Homer1a, and what the functional implications of this interaction are, remain to be established. This is particularly important because binding to Homer1a may decrease ligand-dependent⁵ or enhance ligand-independent^2,4 activity of mGluR1/5, resulting in opposing changes in Ca²⁺ and possibly in pain.

Pain is regulated at multiple levels of the peripheral and central nervous systems. In addition to the spinal cord, mGluR1/5 regulates pain and associated affective states through cortical and limbic brain circuits^5,14. It is therefore highly relevant to determine how scaffolding to Preso1 contributes to the regional specificity of mGluR1/5 actions not only in the spinal cord, but also in limbic and cortical brain areas. Given the range of behaviors requiring mGluR1/5 activity, the newly discovered interaction with Preso1 has important implications for the fine-tuning of mGluR1/5 signals underlying complex behavior.