Flotillin‐1 interacts with the serotonin transporter and modulates chronic corticosterone response

Aberrant serotonergic neurotransmission in the brain is considered at the core of the pathophysiological mechanisms involved in neuropsychiatric disorders. Gene by environment interactions contribute to the development of depression and involve modulation of the availability and functional activity of the serotonin transporter (SERT). Using behavioral and in vivo electrophysiological approaches together with biochemical, molecular‐biological and molecular imaging tools we establish Flotillin‐1 (Flot1) as a novel protein interacting with SERT and demonstrate its involvement in the response to chronic corticosterone (CORT) treatment. We show that genetic Flot1 depletion augments chronic CORT‐induced behavioral despair and describe concomitant alterations in the expression of SERT, activity of serotonergic neurons and alterations of the glucocorticoid receptor transport machinery. Hence, we propose a role for Flot1 as modulatory factor for the depressogenic consequences of chronic CORT exposure and suggest Flotillin‐1‐dependent regulation of SERT expression and activity of serotonergic neurotransmission at the core of the molecular mechanisms involved.


| INTRODUCTION
Serotonin (5-hydroxytryptamine, 5-HT) is a neurotransmitter central to the generation and regulation of emotion and cognition. Alterations of serotonergic neurotransmission have been critically implicated in the etiology of several psychopathologies including mood and anxiety disorders and substance abuse. [1][2][3][4] The serotonin transporter (SERT, SLC6A4) is the major regulator of serotonergic tone in the brain through the control of the reuptake of released serotonin from the synaptic cleft into the presynaptic terminal. As a consequence, SERT directly modulates the intensity and duration of 5-HT-dependent signaling and regulation of SERT expression and activity. SERT has therefore been a central focus of pharmaceutical interest as a druggable target in the therapeutic management of affective disorders. [5][6][7][8] Additionally, genetic variations in the promoter region of the SERT gene have been proposed to be involved in mediating the gene × environment effects involved in the pathogenesis of depression. [9][10][11] Nevertheless, the "serotonin hypothesis" of depression, which has not led to a satisfactory level of understanding of this common psychopathology to this day despite having been proposed several decades ago, 12 should not be considered the only possible explanation for the development of symptoms experienced by MDD patients.
Indeed, aside from the aforementioned studies which support an involvement of the serotonergic system in MDD pathophysiology, there have been numerous publications reporting conflicting evidence (reviewed in References [13][14][15] ) as well as alternative hypotheses. [16][17][18] Hence, focusing on the involvement of the serotonergic system in general, and SERT in particular, may therefore be seen as a particularly dogmatic approach to understanding MDD. However, strong evidence in the literature does indicate that this monoamine system is likely to be involved in the pathophysiology of depression, at least in a subpopulation of patients and further in-depth analyses are therefore warranted. 13,14 This is especially true when considering the large number of possible interactions between different endogenous and exogenous factors and modulatory systems in the context of the pathogenesis of this highly prevalent and excruciating disorder.
For example, life stress has been repeatedly demonstrated as a highly relevant environmental variable determining vulnerability/ resilience to the development of depression. Yet again, the molecular framework governing the intricate web of "nature and nurture" interactions remains incompletely explored.
Here, we identified the lipid raft-associated molecule Flotillin-1 (Flot1) as a novel protein interacting with SERT. Previously, Flot1 had been implicated in the internalization and membrane targeting of the dopamine transporter (DAT), which shares a high structural homology with SERT. 19 As a result of this similarity, Flot1 was considered an obvious candidate protein for interacting with SERT, with the additional potential of this protein-protein interaction being relevant in MDD.
Using a genetic mouse model for Flot1 deficiency and a combined molecular biological/biochemical, electrophysiological and behavioral approach we identified a role for Flot1 in the regulation of stressinduced depression-like behavior in the chronic corticosterone (CORT) paradigm. We provide evidence for a regulation of SERT expression and functional activity of serotonergic neurotransmission as an intermediate phenotype and delineate some of the molecular mechanisms involved.

| METHODS AND MATERIALS
See Appendix S1, Supporting Information for a full description.

| Animals
The generation of Flotillin-1 knockout (Flot1 KO; background strain C57BL6/J) mice has been reported elsewhere. 20 Serotonin transporter knockout (SERT KO) mice used for immunoprecipitation experiments were obtained by heterozygous crossings of the SERT-Cre recombinase knock-in mouse line. 21 Adult KO and wild-type (WT) littermate male and female mice (aged 10-16 weeks) were used for all experiments. All animals were single-housed in standard transparent laboratory cages under standard conditions.

| Preparation of detergent-resistant membrane fractions
Mice were sacrificed by neck dislocation and brains were rapidly dissected on ice. Membranes were subjected to density gradient centrifugation as described 22 with minor modifications.

| Protein isolation, nuclear fraction preparation and immunoprecipitation
Hippocampal tissue was powderized in liquid nitrogen and homogenized in a standard protein lysis buffer. Nuclear fractions were prepared from hippocampal tissue using the Nuclear Extraction Kit (Abcam, Cambridge, UK) by following the manufacturer's supplied protocol.

| Mass spectrometry (LC-MS/MS)
After immunoprecipitation the eluate was subjected to electrophoretic separation of proteins as described earlier. 25 Coomassie-bluestained bands were excised from SDS-PAGE gels digested with 10 ng/μL trypsin (Promega, Madison, Wisconsin) and subjected to LC-MS/MS using an ion trap mass spectrometer (HCT ultra ETD II, Bruker Daltonics, Bremen, Germany) coupled with an Ultimate 3000 nano-HPLC system (Dionex, Sunnyvale, California). MS spectra were recorded followed by 3 data-dependent CID MS/MS spectra generated from 4 highest intensity precursor ions. The MS/MS spectra were interpreted with the Mascot search engine (version 2.4.0, Matrix Science, London, UK) against human the SwissProt database.

| Fluorescence resonance energy transfer
For Fluorescence resonance energy transfer (FRET) analysis, YFP-SERT (Y-SERT) and CFP-Flot1 (C-Flot1) were co-transfected in HEK293 cells using the calcium phosphate co-precipitation method, as described elsewhere. 26 A SERT construct tagged with CFP and YFP on its cytoplasmic N and C termini, respectively (to yield C-SERT-Y), 27 was used as a positive control; CFP-SERT and YFP-myrpalm served as negative controls for FRET experiments.
The "three-filter method" was implemented as previously described. 27 Images were acquired using a 63× oil immersion objective under continuous usage of a gray filter (20% density). All experiments were conducted for individual transfections; 5 to 7 wide-field images were captured during each experiment and 1 to 7 transfected cells per image included in the study. Distances r were calculated based on the Förster equation using the value of 4.92 nm as R 0 for the CFP-YFP FRET pair according to Patterson et al. 28 :

| Behavioral tests
All behavior tests were carried out in order of increasing stressfulness 29 and with at least 24 hours between individual tests (for a timeline see Figure S2). Behavioral tests were performed during the light phase of the light-dark cycle, and mice were allowed to habituate to the experimental room for at least 30 minutes prior to testing.

| Sucrose preference test
The sucrose preference test (SPT) was adapted from Yu et al. 30 (2007). Mice were first introduced to a 2% sucrose solution in tap water before the SPT in a training phase, by replacing the drinking water in their cages with aforementioned solution for 48 hours several days prior to testing. Following a return to tap water for 6 hours, mice were food-and water-restricted for 18 hours before testing.
During the 3 hours testing period, mice were offered the choice between sucrose and water. The sucrose preference (%) exhibited by each mouse, representing anhedonia-like behavior, was calculated from the amount of sucrose consumed relative to total liquid consumption (mL).

| Forced swim test
The forces swim test (FST) was carried out according to a published procedure. [31][32][33] Briefly, mice were placed in a beaker filled with water

| Light-dark box test
The light-dark box (LDB) test was adapted from previously published protocols. 34

| Rotarod
Motor coordination was evaluated using an automated Rotarod system (MedAssociates Inc.) 37 Here, mice were placed onto a rotating drum (speed: 4-40 rpm, steadily increasing over 6 minutes) in 3 consecutive trials, and the latency to fall off was automatically recorded.
The average latency to fall was calculated for each animal and employed as the relevant parameter for motor coordination in this test.

| Flot1 and SERT are present in a complex and interact in vitro
Considering the prominent membrane localization of SERT, we carried out a sucrose gradient centrifugation of mouse brain membranes to identify a possible interaction with Flot1. A striking co-localization of SERT and Flot1 in fraction 5 was observed ( Figure 1A), indicating that in the mouse brain SERT associates with Flot1-positive membrane fractions as previously described for heterologously expressed SERT. 38 We therefore sought to further elucidate the nature of physical contact between Flot1 and SERT using immunoprecipitation (IP) experiments. When using a SERT antibody to immunoprecipitate potentially interacting protein partners from mouse brain lysate, a band corresponding to the immunochemical identity of Flot1 was readily detected by subsequent western blot analysis ( Figure 1B). We next employed a mass spectrometrical approach to additionally confirm the Co-IP result. To this end, SERT complexes were purified from HEK293T cells expressing GFP-SERT by using an anti-GFP antibody and the obtained pull-down fraction was separated by SDS-PAGE. 25 The Coomassie blue-stained bands were excised ( Figure 1C) and in-  Figure 1F). CFP-SERT and YFP-myrpalm served as negative control and yielded a low resonance energy transfer (FRET efficiency 3.14 AE 0.34, Figure 1F).
FRET experiments employing CFP-Flot1 and YFP-SERT provided evidence that the 2 proteins were in a physical vicinity which allowed for resonance energy transfer ( Figure 1F), which was in magnitude significantly different from that of the negative control ( Figure 1G). Interestingly, the examination of CFP-Flot1 and YFP-DAT yielded a comparable mean FRET efficiency (YFP-SERT: 8.85 AE 1.41; YFP-DAT: 7.62 AE 1.11; Figure 1G). This finding is noteworthy considering the high degree of structural similarity between SERT and DAT 39 and the KO mice compared to WT ( Figure 4G). An immunohistochemical approach was selected in order to experimentally address the possibility that this finding merely reflected a reduction in the number of serotonergic neurons. No significant effects of treatment or genotype on the number of serotonergic neurons in Flot1 KO and WT before and after CORT treatment were observed ( Figure 4H,I). Overall, these results suggest a differential response of Flot1 KO in the behavioral and serotonergic responses to long-term chronic CORT exposure.

| The expression of molecular constituents of the glucocorticoid receptor nuclear translocation machinery changes distinctively in response to chronic CORT treatment in Flot1 KO and WT mice
The observed augmented behavioral response of Flot1 KO mice to long-term CORT treatment led us to hypothesize that this enhanced sensitivity might relate to altered expression of glucocorticoid receptors (GR) in Flot1 KO mice. qRT-PCR analysis revealed no effect of genotype on mRNA levels of GR ( Figure 5A). We next considered the possibility that, rather than total gene expression, protein levels and/ or nuclear translocation of GR in response to chronic CORT could show genotype-dependency. We therefore determined the patterns of subcellular localization of GR together with total protein levels in

| DISCUSSION
Since the discovery of SSRIs more than 4 decades ago, the serotonin transporter has undoubtedly been the "poster child" of pharmaceutical companies developing and commercializing antidepressant drugs.
However, only about 40% to 60% of depressed patients exhibit a treatment response after first-line treatment with these drugs with remission rates ranging from 30% to 45%. 44 An increasing body of evidence demonstrates the role of environmental influences as susceptibility factors contributing to the Based upon these observations, we considered that Flot1 might be relevant within the "gene × environment" interaction framework which is largely considered to account for the multifactorial origins of psychiatric disorders, including depression. We therefore tested whether Flot1 deletion may constitute a vulnerability factor for behavioral and serotonergic abnormalities upon exposure to adverse  The fact that CORT treatment had no effect on locomotor activity and motor coordination (OFT and RR) on the other hand, is most relevant in terms of a performance control for the FST as it suggests that the observed effect indeed reflects an increase in despair-like behavior rather than a general impact on motor activity after CORT. Flot1 with the endogenous stress response system. Indeed, Flot1 has been described to co-localize with a discrete pool of GR at the plasma membrane where the two proteins were found to be present in a complex. 51 The low-affinity GR is known to transduce the effects of heightened CORT levels by the translocation of the inactive, cytosolic form of the receptor to the nucleus upon binding of CORT and subsequent rearrangement of the accompanying chaperone machinery. Furthermore, genetic variants of GR have been associated with alterations in the stress response and the risk for development of psychiatric disorders. [52][53][54] Here, we report that the subcellular distribution of the protein changed in response to chronic CORT, with differential enrichment in the nuclear vs the cytosolic fractions in a Flot1-dependent manner. While acute exposure to corticosteroids is known to result in an enhancement of nuclear GR, a cytosolic accumulation of GR has been found to relate to enhanced corticosterone levels and associated depressogenic effects of chronic social stress experience. 55 Accordingly, while in WT mice CORT treatment reduced nuclear GR in relation to the cytosolic protein, the reverse effect was found in Flot1 KO mice which showed more cytosolic GR upon longterm CORT exposure. When we next examined the levels of several

ACKNOWLEDGMENTS
This work was supported by the Austrian Science Fund (FWF) funded grants F3506-B20 and F3516-B20.