5-HTR3 and 5-HTR4 located on the mitochondrial membrane and functionally regulated mitochondrial functions

5-HT has been reported to possess significant effects on cardiac activities, but activation of 5-HTR on the cell membrane failed to illustrate the controversial cardiac reaction. Because 5-HT constantly comes across the cell membrane via 5-HT transporter (5-HTT) into the cytoplasm, whether 5-HTR is functional present on the cellular organelles is unknown. Here we show 5-HTR3 and 5-HTR4 were located in cardiac mitochondria, and regulated mitochondrial activities and cellular functions. Knock down 5-HTR3 and 5-HTR4 in neonatal cardiomyocytes resulted in significant increase of cell damage in response to hypoxia, and also led to alternation in heart beating. Activation of 5-HTR4 attenuated mitochondrial Ca2+ uptake under the both normoxic and hypoxic conditions, whereas 5-HTR3 augmented Ca2+ uptake only under hypoxia. 5-HTR3 and 5-HTR4 exerted the opposite effects on the mitochondrial respiration: 5-HTR3 increased RCR (respiration control ratio), but 5-HTR4 reduced RCR. Moreover, activation of 5-HTR3 and 5-HTR4 both significantly inhibited the opening of mPTP. Our results provided the first evidence that 5-HTR as a GPCR and an ion channel, functionally expressed in mitochondria and participated in the mitochondria function and regulation to maintain homeostasis of mitochondrial [Ca2+], ROS, and ATP generation efficiency in cardiomyocytes in response to stress and O2 tension.

This accumulation is achieved by activation of mitochondrial Ca 2+ uniporter during ischemia. The mitochondrial Ca 2+ overload results in opening a nonspecific pore in the inner mitochondrial membrane, mitochondrial permeability transition pore (mPTP), which is permeable to small molecules. Opening of mPTP usually occurs at early reperfusion. The cell death mediators such as apoptosis inducing factor (AIF) and cytochrome c are released to cytosol followed by mPTP opening 11 . Increase in inhibition to mPTP opening caused by reperfusion injury provides an obvious target for cardioprotection.
The previous researches has shown the evidence that 5-HT transporter (5-HTT) is expressed in cardiac cells, and the environmental 5-HT concentration affects the heart cells 12 . Besides 5-HT effect through its receptors on the cell membrane 13,14 , the mechanism which we unveiled highlighted another novel path: 5-HT can directly activate its functional receptors on the mitochondrial membrane, and regulate mitochondrial and cellular activities and functions.
We first exploited the location of 5-HTR in the cardiac mitochondria. Semi-quantification PCR revealed that 5-HTR3 and 5-HTR4 were present in both rat and mice hearts (data not shown). 5-HTR3 and 5-HTR4 were shown to be localized on the membrane of cardiomyocyte mitochondria, by co-immunostaining with an inner-mitochondria membrane marker, COX VI (Fig. 1A). These co-localizations were further confirmed by Western Blot results (Fig. 1B). The purity of mitochondria was controlled by using SERCA2 and Lamin which are expressed on the reticulum and nuclear membrane respectively. In new-natal cardiomyocytes, Ca 2+ transient was used to represent the myocyte contraction. Cardiomyocytes of which 5-HTR3 and 5-HTR4 were knocked down by siRNA (37 ± 2.23 min −1 ) exhibited an increase of 59.14% in the contraction frequency comparing with the control group (23.25 ± 1.49 min −1 ) (Fig. 1C).
ATP, mostly produced in mitochondria, decreased rapidly in cytoplasm after a cell injury or oxygen depletion. As indicated in Fig. 1D, under the normoxic condition, ATP content in 5-HTR3 and 5-HTR4 knocking down cardiomyocytes was not significantly different from the control groups. However, after cultured in anoxic environment for 12 hours, 5-HTR3 and 5-HTR4 knock-down cardiomyocytes exhibited a significant decrease (12.04%) in intracellular ATP content. Knock down 5-HTR3 or 5-HTR4 alone respectively did not result in any significant changes in intracellular ATP content under both normoxic and hypoxic conditions. Lactate dehydrogenase (LDH), a biomarker of cellular damage, was measured under the condition of normoxia and hypoxia. In normoxic condition, 5-HTR3 and 5-HTR4 knock-down cardiomyocytes showed a slight increase (6.76%) in the release of LDH into the culture medium (Fig. 1E). After culture the cells in a hypoxic condition, 5-HTR3 and 5-HTR4 knock-down cardiomyocytes showed a significant elevation (38.48%) in LDH release. Knock down 5-HTR3 or 5-HTR4 alone respectively did not result in any significant changes in LDH release under both normoxic and hypoxic conditions. Under the normal condition, high concentration of Ca 2+ could stimulate mitochondrial Ca 2+ uptake (MCU), thus stimulate mitochondrial bioenergetics (NADH production) to generate ATP 15 . Serotonin hydrochloride (5-HT) as well as serotonin creatinine sulfate at the concentration of 100 μ M significantly reduced 42.13779% and 35.07% of mitochondrial Ca 2+ uptake respectively ( Fig. 2A,B,C), suggesting that 5-HT receptors on mitochondria affected its functions. Moreover, this Ca 2+ uptake in the mitochondria was mediated by a ruthenium red (RuR)-sensitive mechanism (Fig. 2D). Inhibition of mitochondrial Ca 2+ uptake by RuR was in a dose-dependent manner.
Surprisingly, the effect of 5-HT on mitochondrial Ca 2+ uptake under hypoxic condition was opposite to that under the normoxia condition. 3 minutes for hypoxic challenge were employed in most cases as results shown in Fig. 3, 5-HT did enhance mitochondrial Ca 2+ uptake by 87.37357% (Fig. 3F,G). M-chlorophenylbiguanide, a 5-HTR3 agonist, increased mitochondrial Ca 2+ uptake by 31.47068% (Fig. 3G). However, the inhibition of Ca 2+ uptake caused by 5-HTR4 under hypoxia was consistent with that under the normoxia. It reduced 52.08713% of mitochondrial Ca 2+ uptake by zacopride (Fig. 3G). These results suggested that 5-HTR4 did not contribute to the augment effect of 5-HT on mitochondrial Ca 2+ uptake under hypoxia, but 5-HTR3 were involved in.
5-HTR3 belong to the family of ligand-gate cation channels composed of five identical or homologous subunits 17 . In Fig. 3H, single channel currents from mitoplasts derived from cardiomyocytes were recorded at −100 mV. Under the control condition, no single-channel event could be observed. The application of 10 uM 5-HT activated discrete single-channel events of about 1 pA in patches from isolated heart mitoplasts (p < 0.01). As a response to 5-HTR3A antagonist, 3-AQC, the open probability of inward current was significantly reduced (p < 0.01) (Fig. 3J), suggesting that currents was mediated by 5HTR3 18 by its unitary conductance and pharmacological sensitivity.
Scientific RepoRts | 6:37336 | DOI: 10.1038/srep37336 The rates of state 3 and state 4 respirations, and the Respiration Control Rate (RCR) were measured because they display the efficiency of the movement of electrons along the electron transport chain and the coupling of this movement to the production of ATP by oxidative phosphorylation. State 3 respiration rates were determined by the addition of ADP, and state 4 respiration was measured in the presence of adequate substrate but without addition of ADP. Mitochondrial rate of respiration was measured using an O 2 -sensing electrode in the cardiac mitochondria with or without the presence of 5-HTR3 agonists including m-Chlorophenylbiguanide hydrochloride and 1-Phenylbiguanide hydrochloride (Fig. 4). As shown in Fig. 4A, RCRs in mitochondria pre-incubated with 5-HTR3 agonists above were significantly higher than that in control groups (53.4538% and 29.6462%, respectively).
5-HTR4 was activated by different 5-HTR4 agonists respectively, including zacopride, ML 10302 Hydrochloride and BIMU (Fig. 4D). RCRs of mitochondria in which 5-HTR4 was activated, were lower than The opening of mPTP would lead to uncouple oxidative phosphorylation and causing mitochondrial swelling. A high Ca 2+ pulse is a widely accepted method to open mPTP, resulting in mitochondrial swollen. The swollen of mitochondria could be reflected by the decrease of relative UV-light absorbance. CsA, a common mPTP blocker, is therefore shown to prevent this swollen. Either 2 uM CsA or 5-HT3R agonists including 5-HT, m-Chlorophenylbiguanide hydrochloride and 1-Phenylbiguanide hydrochloride prevented mitochondrial swelling followed by a Ca 2+ pulse of 200 uM for 60 s (Fig. 5A,B). (P < 0.05). Inhibitory effect by m-Chlorophenylbiguanide, 1-Phenylbiguanide, and 5-HT on Ca 2+ induced mitochondrial swelling at 300 s and 500 s was shown in Fig. 5B. 5-HTR4 agonist (BIMU), exhibited a similar effect as CsA in preventing 200 uM Ca 2+ induced mitochondrial swelling (Fig. 5B,C). Moreover, inhibition of AC, which is the downstream signal transduction consequent to activation of 5-HTR4, significantly stopped 200 uM Ca 2+ induced mitochondrial swelling. These data are in strong support of the conclusion that 5-HTR4 protects mitochondria via Gα s pathway by attenuating the Ca 2+ uptake and thus, reduce the mPTP opening probability.
5-HT actively comes across the cell membrane via 5-HT transporter (5-HTT) into the cytoplasm 12 . Regulation of intra-mitochondrial Ca 2+ homeostasis is critical in both physiological and pathological functioning of the heart, resulting in the rate of energy production 19 , arrhythmia [20][21][22] , and even cell death 20,21 . Consistent to previous reports [22][23][24] , ruthenium red sensitive MCU accounted upon the majority of Ca 2+ uptake in our study. 5-HT, as the general agonist for all sub categories of 5-HTR, inhibited mitochondria Ca 2+ uptake and O 2 consumption under normoxia, but enhanced mitochondria Ca 2+ uptake and O 2 consumption under hypoxia. Key steps of mitochondrial metabolism in ATP production are Ca 2+ dependent 19 , it is therefore speculated to keep maintenance of ATP production in a relative constant speed by 5-HT under different O 2 tension. Additionally, ROS is also recognized as an important signal molecule 25 and homeostasis of ROS signaling was achieved by 5-HT effect  internal store 28 . Our results is consistent with previous reports that small increases in mitochondrial [Ca 2+ ] takes place followed by a cytosolic [Ca 2+ ] increase induced by hypoxia in cardiomyocytes 29,30 . Augment mitochondria Ca 2+ uptake under hypoxia effectively buffers this elevation in intracellular Ca 2+ , and it limits the damage caused by Ca 2+ , especially in ischemia-reperfusion.
Due to mitochondrial and nuclear genes mutations, arrhythmia is often associated with patients of mitochondrial diseases [31][32][33] . Although abnormality of [Ca 2+ ] in mitochondria buffering is well known to be involved in arrhythmia, the mechanisms are not clear 34,35 . It was known that mitochondria uptakes Ca 2+ rapidly and buffer cytosolic [Ca 2+ ] during E-C coupling, which supports the theory that fluxes in cytosolic [Ca 2+ ] directly lead the increases in mitochondrial [Ca 2+ ]. In 2006, Maack reported his findings that the role of mitochondrial Ca 2+ buffering in cardiac myocytes during contraction is to match the energy/ATP production with demand. Our data supported their conclusion. As shown in Fig. 1D, in the condition of normoxia, ATP concentration remains constant in cardiomyocytes of control and 5-HTR3, 5-HTR4 knock-down groups. Since mitochondrial [Ca 2+ ] increased is a signal to couple energy production in mitochondria when the lack in ATP supply, the larger frequency of cytosolic [Ca 2+ ] transient in 5-HTR3, 5-HTR4 knock-down cardiomyocytes might be due to the maintenance of ATP level in cells. The alternations in frequency of cytosolic [Ca 2+ ] transient accompanies with alternations in heart beating, further relating to arrhythmia. 5-HTR3 and 5-HTR4 are both functionally located on the mitochondrial membrane and exhibit distinct effects on the mitochondrial function. 5-HTR4 is a type of G-protein coupled receptors via Gα s subunits to generate cAMP as the second messenger 17 . 5-HT and 5-HTR4 agonists could significantly increase the concentration of cAMP in isolated heart mitochondria. Consistently, different agonists of 5-HTR4 displayed the similar effect and blockage of AC reversed the effect of activation of 5-HTR4, resulting in credibility of 5-HTR4 function. 5-HTR4 is shown important for maintaining the constant intra-mitochondrial Ca 2+ concentration and lowers the risk of Ca 2+ overload by reducing Ca 2+ influx into mitochondria. 5-HTR3 is the only a ligand-gated ion channels in all 5-HTR families. It is noticed that 5-HTR3 exhibited only augment effect in mitochondria Ca 2+ uptake while under hypoxia, suggesting the other second messengers are required in addition to activation of 5-HTR3.
The inhibition of mPTP opening induced by 5-HTR3 and 5-HTR4 could be physiologically important to protect mitochondria from loss of mitochondrial membrane integrity, and ability of generation of ATP. In addition, it could also reduce the pro-apoptotic proteins released from mitochondria and protect the heart from apoptosis. Heart is a high energy demanding organ. Shortage in O 2 and substrate e.g. ischemia causes the dysfunction and even damage of cardiomyocytes. There is therefore a constant status for cardiomyocytes to keep homeostasis of mitochondrial [Ca 2+ ], ROS, and ATP generation efficiency. Served as "guards", 5-HTR3 and 5-HTR4 together protected cardiomyocytes by increasing the mitochondrial Ca 2+ uptake when facing anoxia, which kept [Ca 2+ ] homeostasis in the cytoplasm. At the same time, they cooperated to reduce the mPTP opening caused by increased [Ca 2+ ] in mitochondria. As a result, the damage of ischemia/reperfusion injury on cardiomyocytes was minimized. 5-HT via its receptors 5HTR3 and 5-HTR4 on the mitochondria membrane contributes to this homeostasis. Our data contributes to a better understanding in how 5-HT, as a neurotransmitter and a potent activator in the blood stream, participates in the regulation of cardiac behaviors and how cardiomyocytes preform in response to O 2 tension by using 5-HT in mediating its mitochondrial activities to pass through the stress.

Methods and Materials
Animal experiments. All of the animal experiments were approved by the Ethics Review Board for Animal Studies of IMM, Peking University. All methods were performed in accordance with the relevant guidelines and regulations.
Mitochondria isolation and mitoplast preparation. Cardiac mitochondria were isolated from BALB/c mice by using a differential centrifugation method that retains mitochondrial structure and functions such as respiration and Ca 2+ uptake 36,37 . After thoracotomy, heart were rapidly excised into an ice-cold isolation buffer (210 mM annitol, 70 mM sucrose, 5 mM HEPES, 1 mM EGTA, 0.5% BSA, pH7.2 adjusted with KOH), and washed 3 times. Atria and fat were removed and heart was chopped into small pieces. Heart was homogenized using glass homogenizer. The samples were the centrifuged for 10 min at 1000 g at 4 °C. The supernatant containing the mitochondria fraction was further centrifuged at 10000 g for 10 min at 4 °C (Centrifuge 5417 R, Eppendrof). Finally, the pellet was re-suspended in the isolation buffer and the protein concentration was determined. Samples were stored on ice before experiments and would be used up within 4 hours after isolation.
Mitoplast preparation was previously introduced by Kirichok et al. with modifications 24 . In order to obtain mitoplasts, the isolated mitochondria were subject to osmotic shock for 5 min in hypotonic solution consisting Scientific RepoRts | 6:37336 | DOI: 10.1038/srep37336 5 mM sucrose, 5 mM HEPES and 1 mM EGTA (pH 7.2 with KOH). The sample was centrifuged for 5 min at 3920 g and subsequently re-suspended in the solution containing 750 mM KCl, 100 mM HEPES and 1 mM EGTA (pH 7.2 with KOH). The isolated mitoplasts were stored on ice before experiments and would be used up within 4 hours after isolation.

Mitochondrial [Ca
Co-immunofluorenscence. Mitochondria centrifuged onto glass coverslips were fixed with freshly made paraformaldehyde (2%) in PBS for 20 min at room temperature. After rinsed 3 times, mitochondria were permeabilized with PBS containing 0.1% Triton X-100 for 5 min. The samples were blocked with 2% BSA in PBS. The coverslips were probed overnight with combination of an anti-5-HTR4 (1:100) and anti-COX IV antibody, or anti-5-HTR3 (1:100) (Santa Cruz Biotechnology, INC., Santa Cruz, CA, USA) and anti-COX IV (1:750) (Cell Signaling Technology, Danvers, MA, USA) Technology antibody, at 4 °C. After washing away the primary antibodies, the coverslips were incubated with combinations of secondary antibodies with two different wavelengths: anti-goat Texas red (1:500) and FITC anti-rabbit (1:500). The samples were rinsed and observed using Nikon N-SIM microscopy.
Mitochondria swelling measurement. Mitochondria swelling and opening of mitochondrial permeability transition pore (mPTP) were measured by studying the rate of absorbance change at 520 nM using UV-vis Spectrophotomiter (UV-1750, Shimadzu) connected to a computer. Incubation conditions and inductions of mPTP opening were indicated in the legends of figures and results.
Cell Culture. Neonatal mouse ventricular cardiomyocytes were cultured as previously introduced by Iwaki et al. with modifications 38 . Cardiomyocytes were isolated from 1-to 3-day-old BALB/c mice. Hearts removed and rapidly trisected and digested with pancreatin (0.1%, Gibco) and collagenase type II (0.03%, Worthington) for 6 minutes at 37 °C. The cell was collected from supernatant after centrifugation (1000 rpm, 5 minutes), and the pellet was resuspended in the buffer containing pancreatin and collagenase. The above steps were repeated for several times until the tissue was fully digested. The myocytes were cultured in medium containing Bulbecco's modified Eagle's medium (DMEM, Gibico) with fetal bovine serum and penicillin/streptomycin (100 ug/ml, Gibco).
Scientific RepoRts | 6:37336 | DOI: 10.1038/srep37336 8 × 10 6 cells were plated on 100-mm plastic culture plate. The cells were cultured 24 hours before transfection. Physiological level 4 uM of 5-HT was added to medium 24 hours prior to the experiment. siRNA. siRNA against 5Htr3a and 5Htr4 were designed and purchased from GenePharma Inc.(Beijing, PR China). Three separated siRNA were designed against distinct regions of each gene and the knock-down efficiency was measured by Semi-quantification PCR and Western Blot. The ones with best efficiency were used for following experiments. siRNA against 5Htr3a is 5′ -GAGGUGAAGUUCAGAACUATT-3′ , scrambled is: 5′ -GGGUUAGCACGAAAUAAGUTT-3′ .
siRNA against 5Htr4 is CUGGCCUAUUACCGAAUCUTT-3′ , scrambled is 5′ -GCGUACUCACUGCAUC UAUTT-3′ Lactate dehydrogenase (LDH) and Intracellular ATP content measurements. LDH released to cultured medium was quantified by colorimetric measurements using Lactate Dehydrogenase detection kit (Shanghai Gensource Co. Ltd., Shanghai, P.R. China) according to the manufacturer's instructions. Extinctions of cells in different groups were measured by mirotiter plate reader (Promega Corporation, Madison, USA) at 570 nm. The intencity values were normalized to control for time points correspondingly.
CellTiter-Glo Luminescent Cell Viability Assay (Promega Corporation, Madison, USA) was used according to the manufacturer's instructions. Luminescence signal was measured using synergy HT fluorometer (Bio Tek, Winooski, VT, USA). The values were normalized to control for each corresponding point. LDH releases from cardiomyocytes into the cultured medium and intracellular ATP content were analyzed after cultured in a hypoxic condition at 37 °C for 12 hours in comparison to normoxic groups.