lable at ScienceDirect

Biochemical and Biophysical Research Communications 485 (2017) 241e248

Contents lists avai

Biochemical and Biophysical Research Communications

journal homepage: www.elsevier .com/locate/ybbrc

Molecular determinants of the olfactory receptor Olfr544 activation byazelaic acid

Trung Thanh Thach a, Yu-Jung Hong a, Sangho Lee b, Sung-Joon Lee a, *

a Department of Biotechnology, BK21-PLUS, College of Life Sciences and Biotechnology, Korea University, Seoul 02841, Republic of Koreab Department of Biological Sciences, Sungkyunkwan University, Suwon 16419, Republic of Korea

a r t i c l e i n f o

Article history:Received 17 February 2017Accepted 20 February 2017Available online 21 February 2017

Keywords:Olfactory receptorOlfr544Azelaic acidInteractionActivation

* Corresponding author.E-mail address: junelee@korea.ac.kr (S.-J. Lee).

http://dx.doi.org/10.1016/j.bbrc.2017.02.1040006-291X/© 2017 Elsevier Inc. All rights reserved.

a b s t r a c t

The mouse olfactory receptor Olfr544 is expressed in several non-olfactory tissues and has been sug-gested as a functional receptor regulating different signaling pathways. However, the molecular inter-action between Olfr544 and its natural ligand, azelaic acid (AzA), remains poorly characterized, primarilydue to difficulties in the heterologous expression of the receptor protein on the cell membrane and lackof entire protein structure. In this report, we describe the molecular determinants of Olfr544 activationby AzA. N-terminal lucy-flag-rho tag ensured the heterologous expression of Olfr544 on the Hana3A cellsurface. Molecular modeling and docking combined with mutational analysis identified amino acidresidues in the Olfr544 for the interaction with AzA. Our data demonstrated that the Y109 residue intransmembrane helix 3 forms a hydrogen bond with AzA, which is crucial for the receptor-ligandinteraction and activation. Y109 is required for the Olfr544 activation by AzA which, in turn, stimu-lates the Olfr544-dependent CREB-PGC-1a signaling axis and is followed by the induction of mito-chondrial biogenesis in Olfr544 wild-type transfected Hana3A cells, but not in mock or Y109A mutanttransfected cells. Collectively, these data indicated that a hydrogen bond between Y109 residue and AzAis a major determinant of the Olfr544-AzA interaction and activation.

© 2017 Elsevier Inc. All rights reserved.

1. Introduction

Olfactory receptors (ORs), the largest superfamily of G protein-coupled receptors (GPCRs), are primarily expressed in the olfac-tory epithelium that detects and discriminates a vast number ofodorants [1,2]. Although this family was first identified in 1991 [3],the molecular basis of the interactions between receptor andodorant molecules remain poorly understood, primarily due to thelack of reliable structural information on ORs. ORs belong to therhodopsin-like class A GPCRs and possess seven transmembrane(TM) helices [4,5]. In contrast to small GPCR families that show over80% sequence identity among family members, ORs feature a highvariability in sequence homology and thereforemay adopt differentconformations. Such diversity in the protein sequences of ORs isconsistent with the fact that each OR is recognized and activated bydiverse odor molecules [3,6]. Molecular modeling studies haveshown that amino acid residues in TM3, TM5, TM6 and extracellularloop 2 (EL2) are commonly involved in odorant recognition byforming the OR ligand binding pocket [5e7].

ORs are primarily expressed on the olfactory epithelium but alsoectopically in various non-olfactory tissues and with unknownfunctions [8]. Olfr544 is a medium-chain dicarboxylic acid receptorknown to bind azelaic acid (AzA) and suberic acid, which are C9 andC8 dicarboxylic acids commonly observed in grains [9]. Olfr544 isectopically co-expressed with olfactory marker proteins in eitherbladder interstitial cells or thyroid parafollicular cells, suggestingthe presence of Olfr544-mediated downstream signaling in thesetissues [8]; Olfr544 is also expressed in pancreatic a-cells of mouseislets, where its response to AzA leads to glucagon secretion [9]. Ourprimary data also showed that Olfr544 was prominently expressedin several non-olfactory tissues, such as adipose, liver, and intestine(data not shown). These data raised the possibility that Olfr544could be a functional receptor in non-olfactory tissues. In thecomparative modeling combined with mutational studies, Abaffyet al. have previously reported that Olfr544 interacts with AzA. Inthe study, hydrophobic residues, including I112, V113 (TM3), V202and V206 (TM5), formed part of the ligand pocket and wereinvolved in determining carbon length preference [6,10]. Likewise,using modeling and mutational analysis, Kawasaki et al. showedthat the N-terminal hydrophobic core formed by helix 8 and TM1-2of Olfr544 is crucial for the rapid response via Gaolf [11]. However,both models are ambiguous in quality due to a local disruption in

T.T. Thach et al. / Biochemical and Biophysical Research Communications 485 (2017) 241e248242

the TM4 or TM5 structure [10,11]. In addition, it remains unclearwhich residues are specifically involved in the interaction betweenOlfr544 and AzA.

The interactions between ORs and selective ligands have beeninvestigated with the combined approaches of molecular modelingand functional biological assays of the receptor [5,12]. For suchstudies, the heterologous expression of OR is a prerequisite toexamine the role of odorants as ligands [12e15]. Importantly,Matsunami and colleagues generated a cell system, HEK293T-derived Hana3A cells, that stably expresses molecular chaperonesfor OR trafficking to the plasma membrane and Gaolf, an importantsignaling component of ORs. Thus, the Hana3A cell line provides afunctional heterologous expression system for ORs and has beenapplied to various functional studies of OR, including de-orphanization [15]. In addition, N-terminal tags have shown toimprove cell surface targeting of OR. For instance, N-terminal tagsderived from either the first 20 amino acids of rhodopsin (rho tag)[16] or the 17 amino acid signal peptide from a single-spanningmembrane protein (lucy tag) [14] enhance receptor folding andcell surface expression.

In the present study, we investigated the Olfr544-AzA interac-tion with an improved in silico docking model of Olfr544-AzA andan N-terminal lucy-flag-rho tagged Olfr544 heterologous expres-sion system using Hana3A cells. Our results demonstrated that theY109 residue in the TM3 of Olfr544 was required for AzA-drivenOlfr544 activation. Expression of lucy-flag-rho-Olfr544 stimulatedCRE-luciferase activity in Hana3A cells, but the activity was notobserved in cells transfected with Olfr544 mutants (Y109A orY109F).

2. Materials and methods

2.1. Cell culture and transient transfection

Hana3A cells were obtained as a kind gift from Dr. HiroakiMatsunami, Duke University. Cells were maintained and grownunder standard conditions in Eagles minimal essential medium/Earle's balanced salt solution (MEM/EBSS, HyClone, USA) supple-mented with 10% fetal bovine serum (FBS, HyClone) and 100 units/mL of penicillin together with 100 mg/ml streptomycin (PEST,Welgene Inc., Korea). When the cells reached approximately 80%confluence in 6-well plates, they were transiently transfected with1 mg of each plasmid for 12 h using Lipofectamine 2000 reagent(Invitrogen) according to the manufacturer's instructions. Subse-quently, the transfected cells were treated with AzA (Sigma, USA)for 10 h.

2.2. Total RNA extraction, reverse transcription and quantitativereal-time PCR

The total RNA was isolated from cells using the RNAiso Plusreagent (Takara Bio Inc., Japan). Reverse transcription to synthesizecDNAwas performed from 1 mg total RNA template using the ReverTra Ace® RT Master Mix kit (Toyobo, Osaka, Japan) according to themanufacturer's instructions. Real-time PCR was performed withthe Thunderbird™ SYBR® qPCR Mix reagent (Takara Bio Inc., Japan)and the iQ5 Cycler System (Bio-Rad, USA). The expression levelswere normalized to the level of the housekeeping gene ribosomalprotein L32, as described elsewhere [17].

2.3. Immunoblotting analysis

Cell lysates were prepared using RIPA buffer (10mM Tris-HCl pH7.5, 1% NP-40, 0.1% sodium deoxycholate, 0.1% SDS, 150 mM NaCl,and 1 mM EDTA) supplemented with 1� Halt™ protease and

phosphatase inhibitor reagent (Thermo, USA) at 4 �C. Soluble pro-teins were collected from the supernatant fraction after centrifu-gation at 13,000 rpm for 15 min. Equal protein concentrations,determined by bicinchoninic acid protein assay kit (PierceBiotechnology, USA), were prepared in sample buffer (5% b-mer-captoethanol) for 5 min, separated by SDS-PAGE, and subsequentlytransferred onto a nitrocellulose membrane (Daeillab, Korea). Themembrane was blocked in TBS containing 5% (w/v) non-fat driedmilk and 0.1% Tween 20 (v/v); successively, the membrane wasincubated with specific primary antibodies, as previously described[17]. Anti-flag antibody (1:500) was purchased from Sigma; anti-CREB (1:250), anti-p-CREB (Ser133; 1:500), anti-b-actin (1:1000),and anti-PGC-1a (1:500) antibodies were obtained from Santa CruzBiotechnology (Santa Cruz, USA). Immunoblotting image was ob-tained using the ChemiDoc™ touch imaging system and the ImageLab 5.2 software (Bio-Rad, USA).

2.4. Fluorescence immunostaining

Hana3A cells were first seeded on a glass poly-D-lysine-coatedcoverslips up to a confluence of 70%e80% and transiently trans-fected with Olfr544 expression vectors for 12 h. The transfectedcells were washed with PBS (pH 7.4) and fixed for 10 min using PBSsupplemented with 4% paraformaldehyde. Fixed cells were thenblocked for 45 min in PBS containing 5% (w/v) bovine serum al-bumin (BSA), incubated with the anti-flag antibody for 2 h at 4 �C,followed by 1 h incubation with the Alexa Fluor 488-conjungatedsecondary antibody (dilution 1:1000; Invitrogen, USA). Subse-quently, cellular nuclei were stained with DAPI (Invitrogen, USA)for 3 min. The coverslip was subsequently embedded in anti-fademounting solution (Thermo Fisher Scientific, USA). Image analysiswas performed using a Zeiss LSM510 META confocal microscopeand the Zeiss LSM700 version 3.2 software (Carl Zeiss, Germany).

2.5. cAMP response element (CRE)-luciferase reporter assay

Hana3A cells were seeded overnight in 24-well plates. Olfr544,CRE-luciferase reporter, and Renilla expression vectors were tran-siently co-transfected into the cells using Lipofectamine 2000(Invitrogen, CA, USA) following the manufacturer's instructions.Eighteen hours after transfection, the cells were stimulated withAzA for 10 h. Luciferase activity was assessed using a Dual-Gloluciferase assay kit (Promega, WI, USA). The luminescence of thesamples was quantified using a Victor X2 multilabel place (Perki-nElmer, USA). The firefly luminescence signal was normalized tothat of Renilla luciferase. EC50 was determined using GraphPadPrism with steady-state analysis.

2.6. Comparative modeling of Olfr544 and Olfr544-AzA docking

We derived a structural model of Olfr544 (residues 19e313)using three crystal structures of GPCR-agonist complexes as tem-plates: k-opioid receptor-JDTic (PDB ID, 4DJH), A2A adenosinereceptor-ZM241385 (PDB ID, 3EML) and M2 muscarinic acetyl-choline receptor-an antagonist (PDB ID, 3UON). Initially, 5 apomodels of Olfr43 were derived based on sequence alignments ob-tained using the Modeller [18] with the best discrete optimizedprotein energy (DOPE) score of �37430, compared to the �34460DOPE score of an average template. Thesemodels were then refinedand submitted to energy minimization using the Galaxy [19] andMolprobility [20], respectively. Olfr544-AzA complexes weregenerated using SwissDock [21]. Three-dimensional graphics wereprepared using PyMol [22]. Interactions between Olfr544 and AzAwere analyzed using LigPlot [23].

T.T. Thach et al. / Biochemical and Biophysical Research Communications 485 (2017) 241e248 243

2.7. Cloning and single amino acid substitution of Olfr544

DNA encoding the full-length Olfr544 sequence was amplifiedfrom an Olfr544 template (MG215211, Origen, USA). The PCRproduct was purified and sub-cloned into a pMV18s plasmid(courtesy of Dr. Jeniffer Pluznick, Johns Hopkins University) be-tween EcoRI and XhoI restriction enzyme sites. The final constructcontained the N-terminal lucy-flag-rho tag followed by Olfr544.The plasmid was amplified by transforming Escherichia coli DH5acells. For site-directed mutagenesis to substitute a single aminoacid, the Olfr544 construct was mutated by PCR with sets of tar-geted mutation primers (Table S1) according to the Quick-changekit protocol supplied (Aligent technologies, USA). The identities ofall constructs were confirmed by DNA sequencing.

2.8. Mitochondrial DNA content and density determination

Relative human mtDNA was assayed through real-time qPCRfrom total DNA as described elsewhere [24,25]. Specific sets ofprimers were used to amplify a 218 bp amplicon on the mito-chondrial NADH dehydrogenase subunit 1 (ND1) spanning from14620 to 14838 in the human mitochondrial genome (Table S1).Mitochondrial density was determined using the MitotrackerGreen probe (Molecular Probes) following the manufacturer's in-structions and our previously described protocol [26]. Briefly, aftertransient transfection with the appropriate expression vectors for12 h, the cells were stimulated with 50 mMof AzA for 10 h. The cellswere then incubated with Green probes (400 nM) for 30 min at37 �C, followed by washing with PBS. Green fluorescence intensitywas measured at excitation and emission wavelength of 490 and516 nm, respectively, using a SpectraMax microplate reader

Fig. 1. Functional expression of Olfr544 in the Hana3A cell system. (A) Expression of Olfr54antibody. (B) Expression of lucy-flag-rho-Olfr544 on the cell membrane assessed by fluorescFluor 488-conjungated secondary antibody. (C) Schematic diagram of CRE-luciferase reportercurves of CRE-luciferase activities in response to AzA.

(Molecular Devices Cor., USA). Images were obtained using aconfocal microscope and the Zeiss LSM700 version 3.2 software(Carl Zeiss, Germany). Figures show representative results obtainedfrom at least triplicate experiments.

2.9. Statistical analysis

All data are presented as the mean ± SEM. Student's t-test orone-way ANOVA followed by Tukey's HSD test were used for two ormultiple group comparisons, respectively, and performed using theGraphPad Prism. A value of P < 0.05, denoted by *, was consideredsignificant. Values that are denoted by different letters are signifi-cantly different (P � 0.05).

3. Results and discussion

3.1. Functional expression of Olfr544 in Hana3A cells

Heterologous expression of ORs on the cell surface is a prereq-uisite for thephysiological analysis of these receptors. Cleavable lucytag has been shown to promote OR protein trafficking or endo-plasmic reticulum exit that stabilizes the receptor and preventsmisfolding [14]. In addition, rho tag enhancesORprotein folding andsurface expression [16]. Therefore, we constructed an N-terminallucy-flag-rho-fused Olfr544 sequence cloned in a pMV18s plasmid.As mentioned earlier, the Hana3A cell line stably expresses RTP1L,RTP2, REEP1 andGaolf, which assist trafficking the OR proteins to theplasma membrane [27]. Hence, N-terminal tagged Olfr544 wastransfected intoHana3Acells. Immunoblotting analysis showed thatOlfr544 proteins are expressed on the plasma membrane (Fig. 1A)and fluorescence immunostaining with an anti-flag antibody

4 fused with N-terminal lucy-flag-rho tag analyzed by immunoblot using an anti-flagence immunostaining using an anti-flag antibody followed by incubation with an Alexaassay in AzA-activated Olfr544-dependent PKA-CREB signaling axis. (D) Dose-response

T.T. Thach et al. / Biochemical and Biophysical Research Communications 485 (2017) 241e248244

confirmed the expression of Olfr544 on the cell surface (Fig. 1B).Next, we performed a functional assay for Olfr544 in Hana3A

cells. Odorant ligand binding to OR changes the receptor confor-mation, which activates Gaolf [28] and increases the intracellularcAMP concentration [29,30], subsequently activates the PKA-CREBsignaling axis. For this reason, we performed a CRE-luciferaseassay to determine the functional relevance of Olfr544 expressedin Hana3A cells (Fig. 1C). Forskolin, an adenylyl cyclase activator,was used as a positive control. We showed that with the trans-fection of lucy-flag-rho-Olfr544 into Hana3A cells, AzA trans-activated CREB resulting in an increased CRE-luciferase activity. TheEC50 value of AzA was 16.8 ± 5.1 mM in Hana3A cells transfectedwith N-terminal tagged Olfr544; at the same time, AzA did notincrease the CRE-luciferase activity in cells that expressed the lucy-flag-rho tag only (mock). The EC50 value of forskolinwas 88 ± 14 nM(Fig. 1D). These data suggested that the activation of Olfr544 by AzAstimulated the PKA-CREB signaling pathway in Hana3A cells.

3.2. Structural modeling of the Olfr544-AzA complex

To characterize the Olfr544-AzA molecular interaction, wederived a structural model of the Olfr544-AzA complex using theModeller [18]. The most crucial step in comparative modeling is tofind the optimal sequence alignment between a target protein andthe template structure. ORs share not more than 25% sequenceidentity with rhodopsin-like GPCR, thereby the generated modelsmay be suboptimal in quality [31]. To overcome this obstacle, we

Fig. 2. Olfr544-AzA complex model. (A) Olfr544 is represented by a ribbon diagram. N- an(Insert) The ligand binding pocket is enlarged showing the residues that interact with AzAOlfr544 and AzA complex are indicated as dotted lines with the relative distances labeled. (Binterpretation of the references to colour in this figure legend, the reader is referred to the

truncated flexible regions in the N- and C-termini of Olfr544 andused multiple template structures in order to reduce such ambi-guity. Overall, the structure of Olfr544 resembles that of arhodopsin-like GPCR family member and consists of seven trans-membrane helices named TM1-7, which are followed by an intra-cellular C-terminal helix 8 (residues 301e312) that is highlyconserved among GPCR family members. There is no local disrup-tion in all seven TM sequences (Fig. S1A). An important feature ofour model is a disulfide bond formed between the C101 residue inTM3 and the C183 residue in the extracellular loop 2 (EL2) that ishighly conserved among rhodopsin-like receptors [32]. Ram-achandran plot showed 290 allowed residues among a total of 293in the structure (99%), indicating that our model is reliable(Fig. S1B) [33]. The superposition between our Olfr544 model andtwo other models previously reported by Bavan et al. and Kawasakiet al., showed differences in the mobile EL2 and intracellular loop 3(IL3) with root mean square distances (RMSD) between Ca atoms of3.6 and 6.3 Å, respectively (Fig. S1C) [11,34]. By contrast, super-position of the 7 TMs showed a high level of conservation with anRMSD value ranging between 1.0 and 1.8 Å (Fig. S1C), demon-strating that our models qualify as a reliable OR structure forfurther analysis.

To characterize the interaction between Olfr544 and AzA, wedocked AzA into the structural model of Olfr544 using the Swiss-Dock [21]. The final model of the Olfr544-AzA complex revealedthat AzA sits nicely in the ligand binding pocket of Olfr544 (Fig. 2A).Analysis using LigPlot showed that two residues, M105 and Y109

d C-termini are labeled. The AzA ligand is depicted as a ball-and-stick model in green.as stick models. Two hydrogen bonds formed by the S263 and Y109 residues in the) LigPlot analysis showing the interactions between the Olfr544 residues and AzA. (Forweb version of this article.)

Fig. 3. Mutational analysis of the Olfr544-Aza interaction and subsequent receptor activation. (A) Expression of lucy-flag-rho-Olfr544 mutants analyzed by immunoblot using ananti-flag antibody. (B) Dose-response curves of CRE-luciferase activities in response to AzA. (C) EC50 values obtained from the dose-response curves.

T.T. Thach et al. / Biochemical and Biophysical Research Communications 485 (2017) 241e248 245

positioned on TM3, interact with AzA via hydrophobic andhydrogen bonds, respectively (Fig. 2B). A conserved H185 residue inEL2 forms a hydrophobic interaction with AzA. T259 and S263residues on TM6 bind to AzA via hydrophobic and hydrogen bonds,respectively (Fig. 2A&B). These inter-molecular interactions be-tween residues in Olfr544 and AzA are characteristic of the OR-ligand interaction in which TM3, TM6 and EL2 constitute theodorant-binding sites [5,35].

Table 1EC50 value of lucy-flag-rho-Olfr544 variances with AzA.

Olfr544 EC50 (mM) R 2c

WT 16.8 ± 5.1 0.98mock n.d.a e

M105A 34.1 ± 25.1 0.95Y109A v.w.b e

Y109F v.w. e

H185A 20.2 ± 3.4 0.99T259A 22.2 ± 7.8 0.93S263A 46.1 ± 13.2 0.98

a n.d., not determined.b v.w., very weak.c Goodness of fit (R2� 0.90), defined as howwell a curve fits a set of observations.

3.3. Mutational analysis of Olfr544-AzA interaction and activation

Using our Olfr544-AzA complex model, we selected 5 residues(M105, Y109, H185, T259, S263) and examined their requirementsor AzA binding bymutation analysis. After cloning 5 mutated formsof Olfr544, the relative protein expressions were determined byimmunoblotting with an anti-flag antibody to ensure similarexpression levels among the mutant proteins and the wild-typeform (Fig. 3A). Quantitative analysis of luciferase activity showedthat all mutants showed a reduced activation of Olfr544 by AzA(Table 1& Fig. 3B&C). Among them, H185A and T259A did not affectEC50 values and M105A slightly reduced EC50. However, S263Areduced the Olf544 activation by 2.5-fold. Importantly, the Y109Amutation showed virtually no activation. These results wereconsistent with our rigid body complex model indicating that Y109and S263 residues interact with AzA. Furthermore, to validatewhether Y109 binds AzA via a hydrogen bond, we mutated thisresidue (Y109F mutant) and repeated the luciferase assay. Phenyl-alanine (F) is identical to tyrosine (Y), except for missing a eOHgroup for hydrogen-bond formation. Similar to the Y109A muta-tion, the Y109F variant exhibited a very weak luciferase activityfollowing the Olfr544 activation by AzA (Fig. 3B&C). Taken together,these data demonstrated that the interaction via hydrogen bondbetween Y109 and AzA is crucial for the Olfr544 activation.

3.4. Tyr-109 residue is required for the Olfr544 activation by AzA

Since the Y109 residue in the Olfr544 sequence is critical for theinteraction and activation by AzA (Fig. 3), we examined whetherthe same residue is required for the AzA-activation of Olfr544 in thePKA-CREB cellular signaling axis by introducing an Y109A mutantreceptor in the Hana3A cell system. Since the EC50 value of Olfr544bound to AzA is 16.8 ± 5.1 mM, we stimulated the cells with AzA at a50 mM concentration. Not surprisingly, we observed that pCREB(Ser133) was significantly induced in Olfr544 (WT)-transfectedcells, but its expression was not affected in mock and AzA-stimulated cells transfected with the Olfr544 (Y109A) mutant(Fig. 4A). In accordance with the fact that CREB activity is increasedby phosphorylation, we examined whether the peroxisomeproliferator-activated receptor g coactivator-1a (PGC-1a) wasinduced. Since the PGC-1a promoter possesses a conserved bindingsite for CREB spanning from �133 to �116 (human) or �146to �129 (mice), CREB activation highly up-regulates PGC-1a tran-scription [36,37]. Quantitative real-time PCR results showed thatthe expression of PGC-1amRNA level was significantly increased by

Fig. 4. The Tyr-109 residue is required for AzA-activated Olfr544 functions. (A) Protein expression levels of total CREB, pCREB (Ser133), PGC-1a, and b-actin analyzed by immunoblotusing appropriate primary antibodies. (B) Gene expression level of Pgc-1a, Tfam assessed by qPCR. (C) mtDNA content was measured by qPCR. (D) Mitochondrial density wasmeasured by immunofluorescence-based confocal imaging using a Mitotracker probe. Scale bar, 50 mm.

T.T. Thach et al. / Biochemical and Biophysical Research Communications 485 (2017) 241e248246

2.5-fold in Olfr544 (WT) but not in the Y109A mutant (Fig. 4B).Importantly, immunoblot results confirmed that the PGC-1a pro-tein was induced by 1.9-fold in Olfr544 (WT) compared with mock,whereas the expression of PGC-1a did not change in AzA-stimulated cells transfected with the Y109A mutant (Fig. 4A).

As a major transcriptional regulator of cellular energy mecha-nisms, PGC-1a directly affects mitochondrial biogenesis and res-piratory function [38,39]. Hana3A cells were generated fromHEK293T cells, a human embryonic kidney cell line. PGC-1a ishighly expressed in kidneys, which contain a high number ofcellular mitochondria and levels of oxygen consumption similar tothat of the heart [40,41]. Mitochondrial biogenesis has been shown

to be a crucial process in renal recovery from injury due to aging,tissue hypoxia, and other acute or chronic kidney diseases [40,42].Olfr544 activation by AzA induces PGC-1a expression both at themRNA and protein level, suggesting that it may also induce mito-chondrial biogenesis in Hana3A cells. Mitochondrial transcriptionfactor A gene (tfam), a key activator of mitochondrial transcriptionand a transcriptional target of PGC-1a [43], was increased by 2.3-fold in Olfr544 expressing cells (WT) compared to mock, whilethere was no virtual difference between the Y109A mutant and themock group (Fig. 4B). Further quantitative real-time PCR resultsshowed that mitochondrial abundance, assessed as mtDNA con-tent, was induced, approximately by 2.5-fold, in Olfr544 (WT)

T.T. Thach et al. / Biochemical and Biophysical Research Communications 485 (2017) 241e248 247

compared to mock and the Y109A mutant (Fig. 4C). This resultsuggested that Olfr544 activation induces mitochondrial DNAsynthesis and transcription as well as biogenesis in Hana3A cells.Consistently, our mitotracker-probed mitochondria imagesdemonstrated that Olfr544 (WT) overexpression substantiallyenhanced mitochondrial density in Hana3A cells, but not in theY109A mutant (Fig. 4D). Together, these data indicated that theY109 residue is required for Olfr544 activation by AzA, which resultin the stimulation of the CREB-PGC-1a signaling axis in Hana3Acells.

In summary, using molecular modeling with multiple templatesfor Olfr544, we proposed an Olfr544 model that was improvedcompared with those previously reported. Based on Olfr544-AzAstructural information, together with CRE-luciferase reporter as-says in Hana3A cells, we demonstrated that the Y109 residue in theOlfr544 amino acid sequence is a key element required for theinteraction with AzA via hydrogen bond interaction. Furthermore,we demonstrated that AzA-driven Olfr544 activation plays a sub-stantial role in the induction of mitochondrial biogenesis throughstimulation of the CREB-PGC-1a signaling axis. These results sug-gested that the ectopic expression of Olfr544 in metabolic tissuesmay regulate cellular energy metabolism. Further in vivo studiesshould be performed to investigate the effect of AzA-driven Olf544activation.

Conflict of interest

The authors declare no conflict of interest.

Acknowledgements

This study was supported by a National Research Foundation ofKorea grant (No. NRF-2013R1A2A2A01016176) funded by the SouthKorean government (MSIP).

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.bbrc.2017.02.104.

References

[1] P.A. Godfrey, B. Malnic, L.B. Buck, The mouse olfactory receptor gene family,PNAS 101 (2004) 2156e2161.

[2] B. Malnic, P.A. Godfrey, L.B. Buck, The human olfactory receptor gene family,PNAS 101 (2004) 2584e2589.

[3] L.B. Buck, A novel multigene family may encode odorant receptors, Soc. Gen.Physiol. Ser. 47 (1992) 39e51.

[4] P. Mombaerts, F. Wang, C. Dulac, S.K. Chao, A. Nemes, M. Mendelsohn,J. Edmondson, R. Axel, Visualizing an olfactory sensory map, Cell 87 (1996)675e686.

[5] G. Antunes, F.M. Simoes de Souza, Olfactory receptor signaling, Methods CellBiol. 132 (2016) 127e145.

[6] T. Abaffy, H. Matsunami, C.W. Luetje, Functional analysis of a mammalianodorant receptor subfamily, J. Neurochem. 97 (2006) 1506e1518.

[7] O. Baud, S. Etter, M. Spreafico, L. Bordoli, T. Schwede, H. Vogel, H. Pick, Themouse eugenol odorant receptor: structural and functional plasticity of abroadly tuned odorant binding pocket, Biochemistry 50 (2011) 843e853.

[8] N. Kang, H. Kim, Y. Jae, N. Lee, C.R. Ku, F. Margolis, E.J. Lee, Y.Y. Bahk, M.S. Kim,J. Koo, Olfactory marker protein expression is an indicator of olfactoryreceptor-associated events in non-olfactory tissues, PLoS One 10 (2015)e0116097.

[9] N. Kang, Y.Y. Bahk, N. Lee, Y. Jae, Y.H. Cho, C.R. Ku, Y. Byun, E.J. Lee, M.S. Kim,J. Koo, Olfactory receptor Olfr544 responding to azelaic acid regulatesglucagon secretion in alpha-cells of mouse pancreatic islets, Biochem. Biophys.Res. Commun. 460 (2015) 616e621.

[10] T. Abaffy, A. Malhotra, C.W. Luetje, The molecular basis for ligand specificity ina mouse olfactory receptor: a network of functionally important residues,J. Biol. Chem. 282 (2007) 1216e1224.

[11] T. Kawasaki, T. Saka, S. Mine, E. Mizohata, T. Inoue, H. Matsumura, T. Sato, TheN-terminal acidic residue of the cytosolic helix 8 of an odorant receptor isresponsible for different response dynamics via G-protein, FEBS Lett. 589

(2015) 1136e1142.[12] H. Zhuang, H. Matsunami, Evaluating cell-surface expression and measuring

activation of mammalian odorant receptors in heterologous cells, Nat. Protoc.3 (2008) 1402e1413.

[13] K. Touhara, Deorphanizing vertebrate olfactory receptors: recent advances inodorant-response assays, Neurochem. Int. 51 (2007) 132e139.

[14] B.D. Shepard, N. Natarajan, R.J. Protzko, O.W. Acres, J.L. Pluznick, A cleavableN-terminal signal peptide promotes widespread olfactory receptor surfaceexpression in HEK293T cells, PLoS One 8 (2013) e68758.

[15] H. Saito, M. Kubota, R.W. Roberts, Q. Chi, H. Matsunami, RTP family membersinduce functional expression of mammalian odorant receptors, Cell 119(2004) 679e691.

[16] D. Krautwurst, K.W. Yau, R.R. Reed, Identification of ligands for olfactory re-ceptors by functional expression of a receptor library, Cell 95 (1998) 917e926.

[17] C. Wu, Y. Jia, J.H. Lee, Y. Kim, S. Sekharan, V.S. Batista, S.J. Lee, Activation ofOR1A1 suppresses PPAR-gamma expression by inducing HES-1 in culturedhepatocytes, Int. J. Biochem. Cell Biol. 64 (2015) 75e80.

[18] B. Webb, A. Sali, Comparative protein structure modeling using MODELLER,Curr. Protoc. Bioinforma. 47 (2014) 1e32, 5 6.

[19] J. Ko, H. Park, L. Heo, C. Seok, GalaxyWEB server for protein structure pre-diction and refinement, Nucleic Acids Res. 40 (2012) W294eW297.

[20] V.B. Chen, W.B. Arendall 3rd, J.J. Headd, D.A. Keedy, R.M. Immormino,G.J. Kapral, L.W. Murray, J.S. Richardson, D.C. Richardson, MolProbity: all-atomstructure validation for macromolecular crystallography, Acta Crystallogr. D.Biol. Crystallogr. 66 (2010) 12e21.

[21] A. Grosdidier, V. Zoete, O. Michielin, SwissDock, a protein-small moleculedocking web service based on EADock DSS, Nucleic Acids Res. 39 (2011)W270eW277.

[22] W.L. DeLano, The PyMOL Molecular Graphics System, DeLano Scientific LLC,San Carlos, CA, 2002.

[23] A.C. Wallace, R.A. Laskowski, J.M. Thornton, LIGPLOT: a program to generateschematic diagrams of protein-ligand interactions, Protein Eng. 8 (1995)127e134.

[24] S. Ayala-Torres, Y. Chen, T. Svoboda, J. Rosenblatt, B. Van Houten, Analysis ofgene-specific DNA damage and repair using quantitative polymerase chainreaction, Methods 22 (2000) 135e147.

[25] J.P. Rooney, I.T. Ryde, L.H. Sanders, E.H. Howlett, M.D. Colton, K.E. Germ,G.D. Mayer, J.T. Greenamyre, J.N. Meyer, PCR based determination of mito-chondrial DNA copy number in multiple species, Methods Mol. Biol. 1241(2015) 23e38.

[26] T.T. Thach, C.K. Lee, H.W. Park, S.J. Lee, S.J. Lee, Syringaresinol induces mito-chondrial biogenesis through activation of PPARbeta pathway in skeletalmuscle cells, Bioorg. Med. Chem. Lett. 26 (2016) 3978e3983.

[27] H. Saito, M. Kubota, R.W. Roberts, Q.Y. Chi, H. Matsunami, RTP family mem-bers induce functional expression of mammalian odorant receptors, Cell 119(2004) 679e691.

[28] L. Belluscio, G.H. Gold, A. Nemes, R. Axel, Mice deficient in G(olf) are anosmic,Neuron 20 (1998) 69e81.

[29] T. Nakamura, Cellular and molecular constituents of olfactory sensation invertebrates, Comp. Biochem. Physiol. A Mol. Integr. Physiol. 126 (2000)17e32.

[30] H. Yamada, K. Nakatani, Odorant-induced hyperpolarization and suppressionof cAMP-activated current in newt olfactory receptor neurons, Chem. Senses26 (2001) 25e34.

[31] B. Nagarathnam, S.D. Karpe, K. Harini, K. Sankar, M. Iftekhar, D. Rajesh, S. Giji,G. Archunan, V. Balakrishnan, M.M. Gromiha, W. Nemoto, K. Fukui,R. Sowdhamini, DOR - a database of olfactory receptors - integrated repositoryfor sequence and secondary structural information of olfactory receptors inselected eukaryotic genomes, Bioinforma. Biol. Insights 8 (2014) 147e158.

[32] P.A. Hargrave, Rhodopsin structure, function, and topography the Friedenwaldlecture, Invest Ophthalmol. Vis. Sci. 42 (2001) 3e9.

[33] G.N. Ramachandran, C. Ramakrishnan, V. Sasisekharan, Stereochemistry ofpolypeptide chain configurations, J. Mol. Biol. 7 (1963) 95e99.

[34] S. Bavan, B. Sherman, C.W. Luetje, T. Abaffy, Discovery of novel ligands formouse olfactory receptor MOR42-3 using an in silico screening approach andin vitro validation, PLoS One 9 (2014) e92064.

[35] S.E. Hall, W.B. Floriano, N. Vaidehi, W.A. Goddard 3rd, Predicted 3-D structuresfor mouse I7 and rat I7 olfactory receptors and comparison of predicted odorrecognition profiles with experiment, Chem. Senses 29 (2004) 595e616.

[36] S. Herzig, F. Long, U.S. Jhala, S. Hedrick, R. Quinn, A. Bauer, D. Rudolph,G. Schutz, C. Yoon, P. Puigserver, B. Spiegelman, M. Montminy, CREB regulateshepatic gluconeogenesis through the coactivator PGC-1, Nature 413 (2001)179e183.

[37] S.F. Schmidt, S. Mandrup, Gene program-specific regulation of PGC-1a activ-ity, Genes Dev. 25 (2011) 1453e1458.

[38] Z. Wu, P. Puigserver, U. Andersson, C. Zhang, G. Adelmant, V. Mootha, A. Troy,S. Cinti, B. Lowell, R.C. Scarpulla, B.M. Spiegelman, Mechanisms controllingmitochondrial biogenesis and respiration through the thermogenic coac-tivator PGC-1, Cell 98 (1999) 115e124.

[39] T. Wenz, Mitochondria and PGC-1alpha in aging and age-associated diseases,J. Aging Res. 2011 (2011) 810619.

[40] J.M. Weinberg, Mitochondrial biogenesis in kidney disease, J. Am. Soc.Nephrol. 22 (2011) 431e436.

[41] D.J. Pagliarini, S.E. Calvo, B. Chang, S.A. Sheth, S.B. Vafai, S.E. Ong, G.A. Walford,C. Sugiana, A. Boneh, W.K. Chen, D.E. Hill, M. Vidal, J.G. Evans, D.R. Thorburn,

T.T. Thach et al. / Biochemical and Biophysical Research Communications 485 (2017) 241e248248

S.A. Carr, V.K. Mootha, A mitochondrial protein compendium elucidatescomplex I disease biology, Cell 134 (2008) 112e123.

[42] M.T. Tran, Z.K. Zsengeller, A.H. Berg, E.V. Khankin, M.K. Bhasin, W. Kim,C.B. Clish, I.E. Stillman, S.A. Karumanchi, E.P. Rhee, S.M. Parikh, PGC1alpha

drives NAD biosynthesis linking oxidative metabolism to renal protection,Nature 531 (2016) 528e532.

[43] R.C. Scarpulla, R.B. Vega, D.P. Kelly, Transcriptional integration of mitochon-drial biogenesis, Trends Endocrinol. Metab. 23 (2012) 459e466.