Báo cáo khoa học: Olfactory receptor signaling is regulated by the post-synaptic density 95, Drosophila discs large, zona-occludens 1 (PDZ) scaffold multi-PDZ domain protein 1

Olfactory receptor signaling is regulated by the post-synaptic density 95, Drosophila discs large, zona-occludens 1 (PDZ) scaffold multi-PDZ domain protein 1 Ruth Dooley1,2,*, Sabrina Baumgart2,*, Sebastian Rasche1, Hanns Hatt1 and Eva M. Neuhaus3

1 Molecular Medicine Lab RCSI, Education & Research Centre Smurfit Building, Beaumont Hospital, Dublin, Republic of Ireland 2 Department of Cell Physiology, Ruhr University Bochum, Germany 3 NeuroScience Research Center, Charite´ , Universita¨ tsmedizin Berlin, Germany

Keywords MUPP1; olfactory neuron; olfactory receptor; PDZ protein; signal transduction

Correspondence E. M. Neuhaus, NeuroScience Research Center, Charite´ , Universita¨ tsmedizin Berlin, 10117 Berlin, Germany Fax: +49 30 450 539 970 Tel: +49 30 450 539 702 E-mail: eva.neuhaus@charite.de

*These authors contributed equally to this work

(Received 9 September 2009, revised 6 October 2009, accepted 12 October 2009)

doi:10.1111/j.1742-4658.2009.07435.x

Structured digital abstract l MINT-7290305: OR2AG1 (uniprotkb:Q9H205) physically interacts (MI:0915) with MUPP1

(uniprotkb:O75970) by anti tag coimmunoprecipitation (MI:0007)

l MINT-7289999, MINT-7290250, MINT-7290063, MINT-7290110: OR2AG1

binds

(uni- protkb:Q9H205) binds (MI:0407) to MUPP1 (uniprotkb:O75970) by peptide array (MI:0081) (uni-

l MINT-7290162: mOR283-1

(uniprotkb:Q9D3U9)

to MUPP1

(MI:0407)

protkb:O75970) by peptide array (MI:0081)

binds

l MINT-7290128: mOR-EG (uniprotkb:Q920P2)

(MI:0407)

to MUPP1

(uni-

protkb:O75970) by peptide array (MI:0081)

Abbreviations CamKII, calcium ⁄ calmodulin-dependent protein kinase II; GABAB, c-aminobutyric acid receptor B; GFP, green fluorescent protein; GST, glutathione S-transferase; HRP, horseradish peroxidase; INAD, inactivation no after potential D; MUPP1, multi-PDZ domain protein 1; OMP, olfactory marker protein; OR, olfactory receptor; OSN, olfactory sensory neuron; PDZ, post-synaptic density 95, Drosophila discs large, zona-occludens 1; RNAi, RNA interference; siRNA, small interfering RNA.

FEBS Journal 276 (2009) 7279–7290 ª 2009 The Authors Journal compilation ª 2009 FEBS

7279

The unique ability of mammals to detect and discriminate between thou- sands of different odorant molecules is governed by the diverse array of olfactory receptors expressed by olfactory sensory neurons in the nasal epithelium. Olfactory receptors consist of seven transmembrane domain G protein-coupled receptors and comprise the largest gene superfamily in the mammalian genome. We found that approximately 30% of olfactory receptors possess a classical post-synaptic density 95, Drosophila discs large, zona-occludens 1 (PDZ) domain binding motif in their C-termini. PDZ domains have been established as sites for protein–protein inter- action and play a central role in organizing diverse cell signaling assem- blies. In the present study, we show that multi-PDZ domain protein 1 (MUPP1) is expressed in the apical compartment of olfactory sensory neurons. Furthermore, on heterologous co-expression with olfactory sen- sory neurons, MUPP1 was shown to translocate to the plasma mem- brane. We found direct interaction of PDZ domains 1 + 2 of MUPP1 with the C-terminus of olfactory receptors in vitro. Moreover, the odor- ant-elicited calcium response of OR2AG1 showed a prolonged decay in MUPP1 small interfering RNA-treated cells. We have therefore elucidated the first building blocks of the putative ‘olfactosome’, brought together by the scaffolding protein MUPP1, a possible central nucleator of the olfactory response.

R. Dooley et al.

PDZ proteins interact with olfactory receptors

binds

l MINT-7290219:

hOR3A1

(uniprotkb:P47881)

(MI:0407)

to MUPP1

(uni-

protkb:O75970) by peptide array (MI:0081)

binds

l MINT-7290191:

hOR1D2

(uniprotkb:P34982)

(MI:0407)

to MUPP1

(uni-

protkb:O75970) by peptide array (MI:0081)

l MINT-7289922: AC3 (uniprotkb:Q8VHH7) and MUPP1 (uniprotkb:O75970) colocalize

(MI:0403) by fluorescence microscopy (MI:0416)

l MINT-7289933, MINT-7289954, MINT-7289978: OR2AG1 (uniprotkb:Q9H205) binds

(MI:0407) to MUPP1 (uniprotkb:O75970) by pull down (MI:0096)

Introduction

division into three discrete functional classes [16], which may not be as strict as initially anticipated because predictions of PDZ domain–peptide interac- tions were recently shown to be evenly distributed throughout selectivity space [17].

[19] acid receptor B (GABAB)

the The multi-PDZ domain protein 1 (MUPP1) is com- posed of thirteen PDZ domains, each diverse with respect to its amino acid sequence. It was first identi- fied through a yeast two-hybrid screening as an interaction partner of the C-terminus of 5-hydroxy- tryptamine receptor type 2C [18]. Subsequently, many diverse interaction partners of MUPP1 have been characterized, including G-protein coupled c-aminobu- and the tyric calcium ⁄ calmodulin-dependent protein kinase II (Cam- KII) [20]. In the present study, we introduce ORs as novel interaction partners of MUPP1.

Results

MUPP1 is expressed at the sites of olfactory signal transduction

Detection of an odorant is initiated by activation of a fraction of many hundreds of G protein-coupled odorant receptors (ORs) expressed in olfactory sen- the mammalian olfactory sory neurons (OSNs) of epithelium [1]. Signal transduction begins when an odorant molecule binds to an OR, resulting in the activation of adenylyl cyclase type III [2] via the olfactory G protein Gaolf [3]. cAMP then binds to a cyclic nucleotide-gated channel [4–7], allowing it to such as Na+ ⁄ Ca2+. The calcium conduct cations ions then bind to a calcium-gated chloride channel [8], further depolarizing the cell. How these diverse signaling molecules find each other in the complex and densely-packed environment of the cell, avoiding cross-talk with other signaling pathways, in order to signaling, ensure rapidity and specificity of remains an unanswered question. The idea of the existence of an ‘olfactosome’, or highly ordered multi-component protein network involving the olfac- tory signal transducing molecules, has been previ- ously discussed [9–11]; however, until now, no concrete evidence has been provided. Scaffolding net- in the visual works have been investigated in detail system of Drosophila melanogaster, where inactivation no after potential D (INAD), made up of five post- synaptic density 95, Drosophila discs large, zona- occludens 1 (PDZ) domains, has the ability to bind to various molecules in the signal transduction cas- cade, thereby bringing them into close proximity and ensuring a rapid and specific signal transduction [9,12].

ORs are expressed on the ciliary membranes of OSNs, the first point of contact of the sensory cell with incoming odorant molecules. To investigate whether a receptor centered multi-component protein network might exist, we tested for the expression of PDZ scaffolding proteins in the olfactory epithelium using RT-PCR (Fig. 1A). We detected robust expression of MUPP1 and weak expression of ZO-1, but could not detect Patj, Erbin or DLG-2. Because MUPP1 mRNA was highly abundant, we examined the expression of the protein by western blotting. Upon fractionation of the olfactory epithelium [21], we found MUPP1 to be present to a greater extent in the cilia-enriched fraction compared to the remaining cell fractions (Fig. 1B).

FEBS Journal 276 (2009) 7279–7290 ª 2009 The Authors Journal compilation ª 2009 FEBS

7280

Using olfactory marker protein (OMP)-green fluo- rescent protein (GFP) transgenic mice, expressing GFP in every mature olfactory sensory neuron [22], we investigated the cellular localization of MUPP1 in the to be epithelium and found MUPP1 olfactory PDZ domains are modular protein–protein interac- tion domains, which are amongst the most abundant protein interaction domains in organisms from bacteria to mammals, and have been implicated in various pro- including clustering, targeting and routing of cesses, their binding partners [13–15]. PDZ target specificity is usually dependent on the extreme carboxyl-terminal amino acid sequence of the interacting protein; how- ever, for some ligands, residues as far back as the )10 position may influence binding energy [16]. Peptide- binding preferences of PDZ domains led to their

R. Dooley et al.

PDZ proteins interact with olfactory receptors

A

B

C

D

Fig. 1. MUPP1 expression in olfactory sen- sory neurons. (A) Expression of mRNA of different PDZ scaffolding proteins in the olfactory epithelium by RT-PCR. *Weak band for ZO-1. (B) Fractional preparation of whole olfactory epithelium shows MUPP1, at 220 kDa, enriched in the cilia fraction (1) compared to the remaining cell fractions (2–4); a Coomassie-stained gel is shown as a loading control. (C) MUPP1 is co-localized with adenylyl cyclase 3 in the apical layer of the olfactory epithelium. Immunohistochemi- cal staining of 14 lm cryosections of OMP-GFP mouse olfactory epithelium using specific antibodies against MUPP1 (green) and adenylyl cyclase 3 (red). Overlay shows mature OSNs in blue. White arrow denotes the apical layer. Scale bars = 50 lm. (D). Higher magnification image of MUPP1 ⁄ ade- nylyl cyclase 3 stained olfactory epithelium. The arrow shows MUPP1 expression in cilia and in dendritic knobs. Scale bar = 5 lm.

expressed in the apical part of OSNs, mainly in the cilia layer (Fig. 1C). Double immunolabeling showed co-localization with adenylyl cyclase 3, a central mole- cule in the olfactory signal transduction cascade in the cilia of OSNs (Fig. 1C, D).

Interaction of PDZ domains 1 + 2 of MUPP1 with OR2AG1 in vitro

We performed co-immunoprecipitation experiments to verify the ability of ORs containing a PDZ interac- tion motif to bind to MUPP1. Hana3A cells were transfected with HA-OR2AG1. These cells express members of the RTP and REEP family, which com- prise molecular chaperones known to promote the expression of ORs [23]. Antibodies to the HA tag were used to co-immunoprecipitate MUPP1, as shown by a band of 220 kDa in western blots (Fig. 2B, immuno- precipitation: a-HA), whereas no MUPP1 immuno- could be observed in precipitates of reactivity nontransfected cells (Fig. 2B, control). These results indicated an interaction between OR2AG1 and MUPP1 in the recombinant expression system.

FEBS Journal 276 (2009) 7279–7290 ª 2009 The Authors Journal compilation ª 2009 FEBS

7281

PDZ domain interactions have been well characterized and modes of binding have been grouped into three main classes of PDZ binding motifs, occurring at the C-terminus of the interacting proteins [16]. We scanned the human olfactory receptor repertoire for putative binding motifs and discovered them in the extreme C-termini of approximately 30% of human ORs, with examples from each of the three classes being outlined to date (7% Class I, 12% Class II and 10% Class III; Fig. 2A). Intriguingly, this suggested that a subset of ORs could have the ability to interact with PDZ domains of MUPP1. MUPP1 is made up entirely of thirteen different PDZ domains, each being diverse in sequence. We set out to determine which of these PDZ domains were involved in the molecular interaction with ORs. We created a glutathione S-transferase (GST) fusion pep- tide of the C-terminus of OR2AG1 and in vitro trans- lated the PDZ domains of MUPP1, in pairs (1 + 2,

R. Dooley et al.

PDZ proteins interact with olfactory receptors

array positives scored as 3 + 4, etc.) (Fig. 2C). Interaction assays were then carried out by incubating different pairs of PDZ domains as in vitro translation products with OR2AG1 C-terminus GST fusion peptides. A specific binding of PDZ domains 1 + 2 was determined via western blot- in vitro translated PDZ ting, whereas, for example, domains 3 + 4 did not have the ability to bind to the C-terminus of OR2AG1 (Fig. 2D). None of the PDZ domains could bind to GST alone. We then tested binding of single PDZ domains 1 + 2, and found that both could bind to the OR C-terminus (Fig. 2D).

receptors, Next, we investigated the ability of PDZ domains 1 + 2 to bind to receptor C-termini of 15 amino acids in length, which were spotted on microarrays. The arrays were probed four times with PDZ domains 1 + 2 fused to the HA tag for subsequent analysis of binding by antibody incubation and chemiluminescence detection (Fig. 2E). Interactions were analyzed in every experiment with positive (HA tag spotted directly) and negative (FLAG tag spotted directly and A1 ⁄ A2) control spots. Interactions that yielded robust interactions of PDZ domains 1 + 2 with the olfactory receptors hOR2AG1 (S-T-L), mOR283-1 (A-T-V) and hOR3A1 (S-L-A), which all contain PDZ domain binding motifs in their C-ter- mini, were (Fig. 2E). hOR1D2 and mOR-EG, which are olfactory receptors that do not contain classical PDZ interaction motifs in their C-termini, also showed positive interactions but, in the case of hOR1D2, only in two out of four experiments. Other olfactory such as mOR167-4, mOR199-1, M71, M72 and mOR241-1

A

B

C

D

E

Fig. 2. The C-terminus of OR2AG1 interacts with MUPP1 in vitro. (A) Pie chart illustrating the abundance of classical PDZ motifs in human OR C-termini. (B) MUPP1 was immunoprecipitated in HA-OR2AG1 expressing Hana3A cells using a-HA antibodies, detection was performed with a-MUPP1 (*MUPP1) and a control was performed with identical amounts of nontransfected cell lysates from Hana3A cells. The blot shown is representative of three independent immunoprecipitation experiments. (C) Western blot using HA-specific antibodies showing the in vitro translation products of PDZ domain pairwise constructs. Three nonspecific bands appear at 170, 70 and 30 kDa. Specific bands at the correct molecular weights are outlined (white asterisk). (D) PDZ domains 1 + 2 both interact with OR2AG1_GST in vitro. Interaction assay using in vitro translated PDZ domains 1 + 2, PDZ domains 3 + 4, PDZ domain 1 and PDZ domain 2 with GST alone or C-terminus OR2AG1_GST. The blots shown are representative of four independent experiments for each interaction assay described. (E) Peptide micro- array with C-termini of different receptors incubated with PDZ domains 1 + 2 fused to HA; chemiluminescence detection on film after incu- bation with a-HA antibodies and HRP-coupled secondary antibodies. The array shown is representative of four independent experiments; peptide sequences for spots A1–A12 (row 1), A13–A24 (row 2) and B1 (FLAG tag) and B2 (HA tag, positive control) are listed in Table S1. A1 and A2 serve as negative controls.

FEBS Journal 276 (2009) 7279–7290 ª 2009 The Authors Journal compilation ª 2009 FEBS

7282

R. Dooley et al.

PDZ proteins interact with olfactory receptors

A

B

C

D

Fig. 3. MUPP1 plasma membrane translocation on co-expression of odorant receptors. (A) MUPP1-GFP expressed alone in Hana3A cells exhibits a diffuse cytosolic expression. Co-transfection of OR2AG1 with MUPP1-GFP leads to a predominant plasma membrane expression of MUPP1-GFP with clustering apparent. Truncation of OR2AG1 from the final eight amino acids leads to a cytosolic expression of MUPP1- GFP; at least five independent experiments were performed for each condition. (B) Higher magnification of the plasma membrane of the cells shown in (A). (C) In vitro interaction properties of truncated hOR2AG1 C-terminus. Western blot showing HA-PDZ1 + 2 probed with 2AG1-GST and trunc8-GST, at 55 kDa, using a-HA antibodies. The experiment was repeated three times with similar results being obtained. (D) Co-expression of hOR1D2 and hOR3A1 also resulted in translocation of MUPP1-GFP to the plasma membrane; at least five independent experiments were performed for each receptor. Scale bars = 20 lm.

and mGluR2, as well as the olfactory cyclic nucleo- tide-gated ion channel subunit A2, did not show any interaction with the PDZ domains investigated.

We furthermore investigated the binding determi- nants in the C-terminus of hOR2AG1 by spotting pep- tides that correspond to mutated or shortened receptor C-termini. Truncation of the last amino acids abol- ished binding of the C-terminus of hOR2AG1 to PDZ domains 1 + 2. hOR2AG1 constructs where the last four amino acids H-S-T-L were mutated to H-A-T-A [OR2AG1_deltaPDZ(A)] still bound to PDZ domains 1 + 2, whereas mutation to H-W-T-W [OR2AG1_del- taPDZ(W)] completely abolished binding (Fig. 2E).

MUPP1 shows plasma membrane localization upon co-expression of ORs

FEBS Journal 276 (2009) 7279–7290 ª 2009 The Authors Journal compilation ª 2009 FEBS

7283

MUPP1 is a cytosolic protein, and MUPP1-GFP, simi- lar to endogenous MUPP1, shows a homogenous, predominantly cytosolic distribution when expressed Interestingly, when (Fig. 3A). in Hana3A cells co-expressed with hOR2AG1, MUPP1 exhibited a lar- gely plasma membrane expression in a subset of cells, forming clusters at surface (Fig. 2A, B). the cell Approximately 5% of transfected cells exhibited this translocation effect of MUPP1-GFP. This apparently low proportion of cells reflects the notoriously low expression rate of olfactory receptors in heterologous systems and the relatively high amount of cells express- ing MUPP1-GFP. However, the results obtained in the present study correlate with various studies showing a similar proportion of transiently transfected cells responding to odorant in ratiometric calcium imaging experiments [24–26]. This alteration in the subcellular distribution of MUPP1 upon co-expression with ORs supported the finding of a physical association between MUPP1 and ORs in the heterologous expression system. We hypothesized that, by deleting the components of the PDZ binding motif, we could disrupt MUPP1 translocation. Truncation of the receptor from the final eight amino acids (amino acids 309–316) did result in a clear abrogation of the association, as outlined by the predominantly cytosolic localization of MUPP1- GFP (Fig. 3A). We then investigated the in vitro bind- ing properties of the truncated C-terminal mutant of OR2AG1 by creating a GST fusion construct. The

R. Dooley et al.

PDZ proteins interact with olfactory receptors

A

B

C

D

E

F

G

H

I

Fig. 4. Functional role of MUPP1 in OR signaling. (A) Western blot showing MUPP1 expression in Hana3A cells (control) compared to 48 and 72 h after siRNA (exon5) transfection. (B) Representative ratiometric calcium imaging responses of transiently transfected Hana3A cells [siRNA(1)]; the arrow represents the beginning of application of amylbutyrate, lasting for 10 s. (C) Western blot showing MUPP1 expression in Hana3A cells (control) compared to scrambled siRNA, siRNA against exon 5 of MUPP1 and siRNA against exon 45 of MUPP1, 72 h after transfection. (D) Bar chart showing the rise time (10–90%) of Hana3A cells responding to amylbutyrate, transfected with OR2AG1 (ctrl) (n = 15) or siRNA(exon5)-treated Hana3A cells transfected with OR2AG1 (RNAi) (n = 15). (E) Time of decay from 90% of peak amplitude to 10% of average baseline (n = 15) for each condition and the percentage of cell responses decaying to basal levels within the time-frames outlined. Cell responses not decaying within the time-frame of experiment were included in the > 20 s section; n = 15 for control, n = 27 for RNAi(exon5). (F) Bar chart showing the rise time (10–90%) for siRNA(exon45) transfected Hana3A cells; n = 12 for control, n = 12 for RNAi(exon45). (G) Bar chart showing the time of decay (90–10%) and percentages of cell responses for siRNA(exon45) transfected Hana3A cells; n = 12 for control, n = 12 for RNAi(exon45). (H) Bar chart showing the rise time (10–90%) for Hana3A cells transfected with scrambled siRNA. (I) Bar chart showing the time of decay (90–10%) and percentages of cell responses for scrambled siRNA transfected Hana3A cells. Error bars show the SEM. **P < 0.01, ***P < 0.001.

Consistent with this observation is the fact

FEBS Journal 276 (2009) 7279–7290 ª 2009 The Authors Journal compilation ª 2009 FEBS

7284

C-terminal mutant peptide was incubated with PDZ domains 1 + 2 of MUPP1 and failed to interact with truncated mutant (Fig. 3C). that other olfactory receptors showing interactions with PDZ domains 1 + 2 (hOR1D2, hOR3A1) also caused

R. Dooley et al.

PDZ proteins interact with olfactory receptors

A

B

C

D

E

F

Fig. 5. Interaction of MUPP1 with OR2AG1 is important for controlled signal decay. (A) Immunocytochemistry (a-HA antibody) showing sta- ble expression of MUPP1-PDZ1 + 2-HA in Hana3A cells. Scale bar = 20 lm. (B) Representative ratiometric calcium imaging traces for Hana3A cells (control) and MUPP1-PDZ1 + 2-HA cells (1 + 2) transiently transfected with OR2AG1. Arrows denote amylbutyrate application. (C) Bar chart showing the rise time (10–90%) for Hana3A cells stably expressing MUPP1-PDZ1 + 2-HA, transiently transfected with OR2AG1; n = 13 for control, n = 24 for PDZ domains 1 + 2. (D) Decay of response (90–10%) (n = 13 for control, n = 24 for PDZ domains 1 + 2) and the percentage of responses to amylbutyrate decaying to basal levels within the given time-frames. (E) Transient expression of a truncated version of OR2AG1 missing the last eight amino acids (trunc8). Bar chart showing the rise time (10–90%) for Hana3A cells expressing the truncated receptor (n = 12 for control, n = 17 for OR2AG1-trunc8). (F) Decay of response (90–10%) (n = 12 for control, n = 17 for OR2AG1-trunc8) and the percentage of responses to amylbutyrate decaying to basal levels within the given time-frames. Error bars show the SEM. **P < 0.01, ***P < 0.001.

MUPP1-GFP translocation to the plasma membrane in the Hana3A cells (Fig. 3D).

MUPP1 controls the duration of Ca2+ signaling mediated by recombinant ORs

(Fig. 4B, D). Using another

FEBS Journal 276 (2009) 7279–7290 ª 2009 The Authors Journal compilation ª 2009 FEBS

7285

To determine whether MUPP1 has the ability to regulate OR function, we investigated the role of this scaffolding protein in hOR2AG1-mediated Ca2+ mobilization by performing ratiometric Ca2+ imaging in Hana3A cells. These cells express MUPP1 endoge- nously and, by reducing the amount of MUPP1 by RNA interference, we aimed to investigate the func- the interaction between MUPP1 and tionality of hOR2AG1. Transfection of small interfering RNA (siRNA) against Mupp1 led to an almost complete knockdown of the MUPP1 protein, as shown by wes- tern blotting (Fig. 4A). Next, we monitored the response of transiently transfected OR2AG1 to its spe- cific odorant ligand, amylbutyrate [24] via ratiometric calcium imaging in siRNA-treated cells (Fig. 4B). the OR-elicited When MUPP1 was largely absent, response exhibited a similar rise time to that of the control cells, 5.54 ± 0.87 s for RNA interference (RNAi) compared to 4.27 ± 0.65 s for the control (Fig. 4B, C). However, the response failed to decay within the normal average time-frame in siRNA-trea- ted cells (19.3 ± 2.9 s) compared to the control cells (7.2 ± 2.1 s) siRNA directed against an alternative exon of Mupp1, similar results were obtained. The rise time of the OR-elicited the control cells response was similar to that of

R. Dooley et al.

PDZ proteins interact with olfactory receptors

of

(Fig. 4C), although the decay was prolonged signifi- cantly (Fig. 4D). Cells transfected with a scrambled version of the siRNA did not show any significant the differences the OR-elicited kinetics in Ca2+ response compared to nontransfected cells (Fig. 4E, F).

transduction. With its Interaction with MUPP1 is important for controlling OR-mediated Ca2+ signaling

The primary source of olfactory signaling and the sites of expression of ORs are the ciliary structures of the OSNs. Interestingly, we found MUPP1 to be pre- dominantly expressed in the apical compartment of OSNs and enriched in the cilia fraction of a prepara- tion of whole murine olfactory epithelium. We hypoth- esize that MUPP1, through its multivalent capabilities, could play a key role as a central nucleator of olfac- thirteen PDZ tory signal domains, each diverse in its amino acid sequence, MUPP1 holds great potential for organizing signal transduction molecules into defined protein networks and thereby regulating signaling events.

compared cells to

petides. Moreover, compared to control cells

Because PDZ domains 1 + 2 interact with MUPP1 (Fig. 2), we generated a cell line over-expressing these two PDZ domains (Fig. 5A). Interestingly, the pro- longed signal decay observed in the siRNA experi- ments was mirrored in the OR-dependent responses of Hana3A cells stably over-expressing PDZ domains 1 + 2 of MUPP1 (Fig. 5B). When monitoring the response of transiently transfected OR2AG1 to amy- lbutyrate, we found that the OR-elicited response did not decay within the normal average time-frame in cells over-expressing PDZ domains 1 + 2 (26.2 ± (10.2 ± 1.9 s) control 1.8 s) (Fig. 5D). As in siRNA-treated cells, the rise time after odorant stimulation was almost indistinguishable in cells over-expressing PDZ domains 1 + 2 (4.87 ± (4.71 ± 0.69 s) 0.2 s) (Fig. 5C). In summary, when the interaction between MUPP1 and OR2AG1 is inhibited by over-expressing the interacting PDZ domains, the resulting response of OR2AG1 to odorant is modified in that the rapid decay of signal is impaired.

eight the receptor (amino acids amino acids

We finally examined the effect of deletion of the components of the PDZ binding motif on the signal- in Ca2+ imaging the receptor ing properties of from the experiments. Truncation of final 309–316) resulted in a prolonged signal decay, similar to that observed in the siRNA experiments and in the cells over-expressing PDZ domains 1 + 2, whereas the rise time of the signals was again not affected (Fig. 5E, F). indicating that

Discussion

MUPP1 has previously been found to interact with a diverse array of molecules, including G protein-cou- pled receptors such as the 5-hydroxytryptamine recep- tor type 2C [18,27] and the GABAB receptor [19]. We postulated that MUPP1 could directly interact with the olfactory receptor itself. We scanned the entire human olfactory receptor repertoire and discovered that up to 30% of receptors contain putative PDZ binding motifs in their C-termini, following the previ- ously outlined rules of binding [16]. PDZ domains 1 + 2 of MUPP1 indeed showed direct interaction with OR C-terminal upon co-expression of ORs containing classical PDZ bind- ing motifs in their C-termini, MUPP1-GFP exhibited a translocation from the cytosol to the plasma mem- brane in the heterologous expression system, suggest- ing a physical association between both proteins within the cell. We found this translocation to be dependent on the final amino acids of the receptor protein. Similar to the other PDZ domain interactions that have been shown to be abolished by mutating amino acids at position 0 and )2 from the C-termi- nus [16], we found that binding of the hOR2AG1 C- terminus to PDZ domains 1 + 2 did not occur when positions 0 and )2 were mutated to tryptophans, which are not present in the PDZ binding motifs of other membrane proteins [17]. We also found interac- tion of PDZ domains 1 + 2 with ORs showing no classical PDZ binding motifs, the the exact molecular rules of OR understanding of PDZ interaction will require further analysis. Previous work with other proteins has already indicated that it is highly likely that a large number of PDZ domain interactions will not fit into the confined class defini- tions and that PDZ domains may have been opti- mized across to minimize the proteome in order cross-reactivity [17].

FEBS Journal 276 (2009) 7279–7290 ª 2009 The Authors Journal compilation ª 2009 FEBS

7286

We observed that a reduction of MUPP1 resulted in a significant increase in the duration of Ca2+ responses evoked by the activation of recombinantly expressed Until now, the involvement of PDZ domain scaffold- ing proteins in olfactory signal transduction has gone unstudied. It has previously been suggested that such scaffolding networks or ‘olfactosomes’ may exist [9,10] but, to date, no evidence for this phenomenon has been outlined. In the present study, we have uncovered a PDZ protein as a novel interaction partner of olfac- tory receptors and have elucidated the molecular details of this interaction.

R. Dooley et al.

PDZ proteins interact with olfactory receptors

regulator of

signaling receptors and a putative processes in OSNs. It is tempting to speculate that a so-called ‘olfactosome’ exists in the cilia of olfactory sensory neurons, organizing the vast array of signaling molecules and ensuring the specificity of signaling. How exactly MUPP1 could carry out such an impor- tant task remains to be elucidated, although the answer may lie in the remaining and as yet unidentified interac- tion partners of MUPP1 in the olfactory sensory cell.

Experimental procedures

the effect of DNA constructs and primers

[28]. response Interestingly,

(Erbin),

and

pCDNA3_MUPP1-GFP and in vitro translation tandem PDZ domain constructs in vector pBAT were provided by H. Lu¨ bbert (Ruhr-University, Bochum, Germany). pCDNA3_OR2AG1 was cloned as described previously [33]. C-terminal mutant constructs of OR2AG1 [^PDZ_ pCDNA3 (S314A, L316A), trunc4_pCDNA3 (amino acids 313–316) and trunc8_pCDNA3 (amino acids 309–316)], were obtained by PCR using varying 3¢ primers and pCDNA3_OR2AG1 as a template. GST fusion constructs of the C-terminus of OR2AG1, and mutant thereof (trunc8) were created by cloning the region between amino acids 293 and 316 from the receptor into pGEX-3X vector (Amersham Pharmacia Biotech, Piscataway, NJ, USA), using varying reverse primers. PDZ domains 1 + 2 were cloned using pBAT_1 + 2_HA as a template. The stable cell line construct, pCMV ⁄ Bsd_PDZ1 + 2_HA, was cloned using pCDNA3_MUPP1-GFP as a template. All constructs were verified by sequencing. For RT-PCR, mRNA was extracted from adult mouse olfactory epithelium, and the primers were used were: 5¢-CAAAACGCTCTACAGGC (ZO-1), TCC-3¢, 5¢-GAAGAGCTGGACAGAGGTGG-3¢ 5¢-TTATGGGCCACCGGATATTA-3¢, 5¢-GGAGAGTCA CTGAAGGCTGG-3¢ (DLG-2), 5¢-AAGCTAAGAGGCA CGGAACA-3¢, 5¢-TCCTTATTGCCAGCGAGACT-3¢ (Patj), 5¢-GGCCACTT 5¢-TTGCAGACGGAAGAGGTTCT-3¢, 5¢-GCGGATCCGCAT TCAGCATCAAAT-3¢ GTTGGAAACCATAGAC-3¢ 5¢-GCGAATTCGA CATTTTTAGTGAGTTCCAC-3¢ (MUPP1).

hOR2AG1. Similarly, over-expression of the OR-inter- acting PDZ domains 1 + 2 of MUPP1 also resulted in odorant-evoked Ca2+ responses that persisted longer than those in control cells. When over-expressed, PDZ domains 1 + 2 may bind to the C-terminus of hOR2AG1, thus having a blocking effect on the bind- ing of the less abundant endogenous MUPP1. A trun- cated receptor no longer containing a PDZ motif in the C-terminus showed the same effect of prolonged signal duration. Thus, all of the experiments revealed that the association of hOR2AG1 with MUPP1 regu- lates signal duration. To a certain extent, the impaired the signal desensitization resembles absence of the multi-PDZ domain protein INAD in the Drosophila visual signal transduction cascade. Flies lacking INAD exhibit a profound reduction of the INAD is also light required for normal deactivation of visual signaling by positioning eye protein kinase C in close proximity to TRP to facilitate its phosphorylation, ultimately result- ing in deactivation of the channel [12,29,30]. On the other hand, in contrast to the findings of the present study, MUPP1 was shown to prolong the duration of GABAB receptor signaling and increase the stability of the receptor [19]. However, we must note that the situ- ation in the recombinant expression system is different from that in the olfactory neurons, where alternative binding partners of MUPP1 presumably exist. It is therefore possible that MUPP1 could exhibit alterna- tive effects on the dynamics of calcium responses induced by ORs in the neurons compared to those induced by heterologously expressed ORs. The observed effects can therefore only be taken as proof of the functional significance of the observed interac- tion. Further studies are necessary to shed light on the function of this interaction in the in vivo situation.

Antibodies

polyclonal

anti-GFP,

such as

Primary antibodies used were: anti-adenylyl cyclase 3, rab- bit polyclonal (Santa Cruz Biotechnology, Santa Cruz, CA, USA), directly labeled using DyLight(cid:2)549 Microscale Antibody Labeling Kit (Pierce, Rockford, IL, USA); anti- MUPP1, rabbit polyclonal (provided by H. Lu¨ bbert, Ruhr- University); (#ab290-50; rabbit Abcam, Cambridge, MA, USA); a-HA antibody, mouse monoclonal (#H9658; Sigma, St Louis, MO, USA). Second- ary antibodies used were goat anti-rabbit Alexa546nm

By influencing the duration of the Ca2+ signal of ORs in the cilia of the sensory neurons, MUPP1 could have a strong influence on the olfactory signaling path- way. An interesting interaction partner of MUPP1 out- lined to date is CamKII, which is known to play an important role in olfactory adaptation [20]. By phos- phorylation of adenylyl cyclase 3 in OSNs, CamKII provides an important mechanism for the attenuation of odorant-stimulated cAMP increases [31]. Alterna- tively, because different pathways, those involving phosphoinositide 3-kinase [32], are ultimately engaged after OR stimulation, MUPP1 may control OR activity by acting as a scaffold to link different signaling pathways.

FEBS Journal 276 (2009) 7279–7290 ª 2009 The Authors Journal compilation ª 2009 FEBS

7287

In conclusion, we have outlined a novel aspect of the olfactory signal transduction cascade by uncovering a previously unknown interaction partner of olfactory

R. Dooley et al.

PDZ proteins interact with olfactory receptors

standard cages at room temperature. Each cage was sur- rounded by a Perspex chamber with ventilation suction to maintain a constant air-flow.

(Molecular Probes, Carlsbad, CA, USA) and horseradish peroxidase (HRP) coupled goat anti-mouse and goat anti- rabbit IgGs (Bio-Rad, Hercules, CA, USA).

All tissue culture media and related reagents were purchased from Invitrogen (Carlsbad, CA, USA). Hana3A cells [23] (provided by H. Matsunami, Duke University Medical Cen- ter, Durham, NC, USA), were maintained in DMEM plus 10% fetal bovine serum and 1% penicillin ⁄ streptomycin, at 37 (cid:3)C and 5% CO2, and transfections were carried out using a standard calcium phosphate precipitation technique. MUPP1-GFP and OR plasmid DNAs were transfected in a ratio of 1 : 10, with approximately 2 lg of total DNA per dish. All images were acquired using a Zeiss LSM 510 Meta confocal microscope (Carl Zeiss, Oberkochen, Germany). Hana3A cells were stably transfected with pCMV ⁄ Bsd plas- mid (Invitrogen) containing tandem PDZ domains 1 + 2 along with an HA tag. Positive clones were selected for using blasticidin (10 lgÆmL)1) and stable transfection was confirmed by immunocytochemistry.

The C-terminal region of OR2AG1 was found to lie between amino acids 293 and 316, as predicted by tmhmm, a trans- membrane helices prediction program based on a hidden Markov model [34]. OR2AG1 C-terminus GST fusion pro- teins and mutant construct thereof (trunc8-GST) were pro- duced in Escherichia coli XL1 blue and purified on glutathione sepharose beads (Becton-Dickinson Biosciences, Franklin Lakes, NJ, USA). PDZ domains were in vitro trans- lated using the TNT(cid:4)T3 Coupled Reticulocyte Lysate Sys- tem (Promega, Madison, WI, USA). Interaction assays were carried out by incubating 10 lL of in vitro translation prod- uct with 50 lL of GST fusion peptide bead slurry for 2 h at 4 (cid:3)C with gentle shaking. After a series of washing steps using Buffer S (20 mm Hepes, 100 mm KCl, 0.5 mm EDTA, 1 mm dithiothreitol, pH 7.9), specific interactions were assessed via immunoblotting. GST alone was used as a negative control.

Cell culture and transfection GST fusion peptides and in vitro interaction assays

(Intavis AG, Cologne, CelluSpots(cid:2) Peptide Arrays Germany) were blocked for 2 h at room temperature with 5% skimmed milk in NaCl ⁄ Tris ⁄ Tween. The arrays were incubated with a PDZ1 + 2_HA fusion protein (produced as described in E. coli) overnight at 4 (cid:3)C. CelluSpots(cid:2) were incubated with the a-HA antibody for 4 h at room temper- ature. Detection was performed with HRP coupled second- ary antibody and using ECL western blotting detection reagent (GE Healthcare).

Cell membrane preparation and western blotting Peptide microarray

The olfactory epithelium of CD1 mice was fractionated by mechanical agitation as described previously [21]. Equal amounts of protein from each fraction were loaded on an SDS gel and subjected to immunoblotting on poly(vinyli- dene difluoride) membrane (Millipore, Billerica, MA, USA) and Coomassie staining. Detection was performed using the ECL western blotting detection system (GE Healthcare, Milwaukee, WI, USA). For co-immunoprecipitation, Hana3A cells were transfected with OR2AG1-HA and MUPP1_pCDNA3 or untransfected. The nucleus-free cell lysates were incubated with biotinylated a-HA antibody and precipitated protein collected using Dynabeads (Invi- trogen) and DynaMag (Invitrogen). Any interaction was detected using MUPP1 antibodies.

Mupp1 siRNA

(Invitrogen),

2000

in

Pre-synthesized and tested Mupp1 siRNA (identification 1 #107246 and 2 #216971) and a custom designed scrambled version of Mupp1 (CUGACUGUGUAUCGAACGGtt) were purchased from Ambion (Austin, TX, USA). siRNA was transfected 72 h prior to calcium imaging using Lipo- serum-free medium fectamine (Opti-Mem; Invitrogen). Forty-eight hours prior to calcium imaging, 2 lg of OR2AG1 plasmid DNA were transfected per dish using ExGen 500 transfection reagent (Fermentas, Glen Burnie, MD, USA). Medium was exchanged for fresh DMEM 24 h post-transfection.

Immunohistochemistry

Mice were raised and maintained according to governmen- tal and institutional care instructions. Immunohistochemis- try was carried out on 14 lm horizontal sections, and fluorescence images were obtained with a confocal micro- scope (Zeiss LSM 510 Meta) with ·40 objective. Control in the absence of any primary antibody experiments revealed a very low level of background staining. For odor- ant exposure experiments, OMP-GFP mice were exposed to a mixture of 100 different odorant molecules (Henkel KGaA, Du¨ sseldorf, Germany) for specified amounts of time. Control mice were housed in a separate room free from artificial odorant stimulation. All mice were held in

Stably ⁄ transiently transfected Hana3A cells were incubated with 7.5 lm FURA-2 AM (Invitrogen). Ratiometric

FEBS Journal 276 (2009) 7279–7290 ª 2009 The Authors Journal compilation ª 2009 FEBS

7288

Ratiometric Ca2+ imaging in heterologous cells

R. Dooley et al.

PDZ proteins interact with olfactory receptors

8 Frings S, Reuter D & Kleene SJ (2000) Neuronal Ca2+

-activated Cl) channels – homing in on an elusive channel species. Prog Neurobiol 60, 247–289.

9 Huber A (2001) Scaffolding proteins organize multimo- lecular protein complexes for sensory signal transduc- tion. Eur J Neurosci 14, 769–776.

10 L’Etoile ND & Bargmann CI (2000) Olfaction and odor discrimination are mediated by the C. elegans guanylyl cyclase ODR-1. Neuron 25, 575–586.

11 Paysan J & Breer H (2001) Molecular physiology of

calcium imaging was performed as described previously [25] using a Zeiss inverted microscope equipped for ratiometric imaging. Cells were exposed to 100 lm amylbutyrate (Hen- kel GmbH), which is a typical odorant concentration used for heterologously expressed ORs [23,35,36], employing a specialized microcapillary application system. The rise time was calculated as the time in seconds from 10% of peak response, starting from the average baseline value, to 90% of peak amplitude. The response decay duration was calcu- lated as the time in seconds between 90% and 10% of the maximum amplitude.

odor detection: current views. Pflugers Archiv – Eur J Physiol 441, 579–586.

12 Tsunoda S & Zuker CS (1999) The organization of

Acknowledgements

INAD-signaling complexes by a multivalent PDZ domain protein in Drosophila photoreceptor cells ensures sensitiv- ity and speed of signaling. Cell Calcium 26, 165–171. 13 Harris BZ & Lim WA (2001) Mechanism and role of

PDZ domains in signaling complex assembly. J Cell Sci 114, 3219–3231.

14 Nourry C, Grant SG & Borg JP (2003) PDZ domain

proteins: plug and play! Sci STKE 2003, RE7.

15 Sheng M & Sala C (2001) PDZ domains and the orga- nization of supramolecular complexes. Ann Rev Neuro- sci 24, 1–29.

for

We thank H. Bartel and J. Gerkrath for their excellent technical assistance; H. Matsunami (Duke University Medical Center, Durham, NC, USA) for the donation of Hana3A cells; P. Mombaerts (MPI Biophysics, Frankfurt, Germany) for the donation of OMP-GFP transgenic mice; and H. Lu¨ bbert ⁄ X. Zhu (Ruhr Uni- versity, Bochum, Germany) the donation of MUPP1 antibodies and constructs. This work was sup- ported by the International Max-Planck Research in Chemical Biology (IMPRS-CB), the Stud- School ienstiftung des deutschen Volkes, the Heinrich und Anna Vogelsang Stiftung and the Deutsche Fors- chungsgemeinschaft (SFB 642).

References

16 Songyang Z, Fanning AS, Fu C, Xu J, Marfatia SM, Chishti AH, Crompton A, Chan AC, Anderson JM & Cantley LC (1997) Recognition of unique carboxyl-termi- nal motifs by distinct PDZ domains. Science 275, 73–77. 17 Stiffler MA, Chen JR, Grantcharova VP, Lei Y, Fuchs D, Allen JE, Zaslavskaia LA & MacBeath G (2007) PDZ domain binding selectivity is optimized across the mouse proteome. Science 317, 364–369.

18 Ullmer C, Schmuck K, Figge A & Lubbert H (1998)

1 Buck L & Axel R (1991) A novel multigene family may encode odorant receptors – a molecular-basis for odor recognition. Cell 65, 175–187.

2 Bakalyar HA & Reed RR (1990) Identification of a

Cloning and characterization of MUPP1, a novel PDZ domain protein. FEBS Lett 424, 63–68.

19 Balasubramanian S, Fam SR & Hall RA (2007)

specialized adenylyl cyclase that may mediate odorant detection. Science 250, 1403–1406.

3 Jones DT & Reed RR (1989) Golf – an olfactory

GABA(B) receptor association with the PDZ scaffold Mupp1 alters receptor stability and function. J Biol Chem 282, 4162–4171.

neuron specific G-protein involved in odorant signal transduction. Science 244, 790–795.

20 Krapivinsky G, Medina I, Krapivinsky L, Gapon S & Clapham DE (2004) SynGAP-MUPP1-CaMKII synap- tic complexes regulate p38 MAP kinase activity and NMDA receptor-dependent synaptic AMPA receptor potentiation. Neuron 43, 563–574.

21 Washburn KB, Turner TJ & Talamo BR (2002)

4 Bradley J, Li J, Davidson N, Lester HA & Zinn K (1994) Heteromeric olfactory cyclic nucleotide-gated channels: a subunit that confers increased sensitivity to cAMP. Proc Natl Acad Sci USA 91, 8890–8894. 5 Bradley J, Reisert J & Frings S (2005) Regulation of cyclic nucleotide-gated channels. Curr Opin Neurobiol 15, 343–349.

6 Dhallan RS, Yau KW, Schrader KA & Reed RR

(1990) Primary structure and functional expression of a cyclic nucleotide-activated channel from olfactory neurons. Nature 347, 184–187.

Comparison of mechanical agitation and calcium shock methods for preparation of a membrane fraction enriched in olfactory cilia. Chem Senses 27, 635–642. 22 Potter SM, Zheng C, Koos DS, Feinstein P, Fraser SE & Mombaerts P (2001) Structure and emergence of specific olfactory glomeruli in the mouse. J Neurosci 21, 9713–9723.

23 Saito H, Kubota M, Roberts RW, Chi QY & Matsun- ami H (2004) RTP family members induce functional

7 Liman ER & Buck LB (1994) A second subunit of the olfactory cyclic nucleotide-gated channel confers high- sensitivity to cAMP. Neuron 13, 611–621.

FEBS Journal 276 (2009) 7279–7290 ª 2009 The Authors Journal compilation ª 2009 FEBS

7289

R. Dooley et al.

PDZ proteins interact with olfactory receptors

32 Spehr M, Wetzel CH, Hatt H & Ache BW (2002)

expression of mammalian odorant receptors. Cell 119, 679–691.

24 Neuhaus EM, Mashukova A, Zhang WY, Barbour J & Hatt H (2006) A specific heat shock protein enhances the expression of mammalian olfactory receptor proteins. Chem Senses 31, 445–452.

25 Spehr M, Gisselmann G, Poplawski A, Riffell JA,

3-Phosphoinositides modulate cyclic nucleotide signaling in olfactory receptor neurons. Neuron 33, 731–739. 33 Mashukova A, Spehr M, Hatt H & Neuhaus EM (2006) beta-arrestin2-mediated internalization of mammalian odorant receptors. J Neurosci 26, 9902– 9912.

Wetzel CH, Zimmer RK & Hatt H (2003) Identification of a testicular odorant receptor mediating human sperm chemotaxis. Science 299, 2054–2058.

26 Wetzel CH, Oles M, Wellerdieck C, Kuczkowiak M,

34 Krogh A, Larsson B, von Heijne G & Sonnhammer EL (2001) Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes. J Mol Biol 305, 567–580.

Gisselmann G & Hatt H (1999) Specificity and sensitiv- ity of a human olfactory receptor functionally expressed in human embryonic kidney 293 cells and Xenopus laevis oocytes. J Neurosci 19, 7426–7433.

27 Becamel C, Figge A, Poliak S, Dumuis A, Peles E,

35 Katada S, Hirokawa T, Oka Y, Suwa M & Touhara K (2005) Structural basis for a broad but selective ligand spectrum of a mouse olfactory receptor: map- ping the odorant-binding site. J Neurosci 25, 1806– 1815.

36 Spehr M, Schwane K, Riffell JA, Barbour J, Zimmer

Bockaert J, Lubbert H & Ullmer C (2001) Interaction of serotonin 5-hydroxytryptamine type 2C receptors with PDZ10 of the multi-PDZ domain protein MUPP1. J Biol Chem 276, 12974–12982.

28 Tsunoda S, Sierralta J, Sun YM, Bodner R, Suzuki E,

RK, Neuhaus EM & Hatt H (2004) Particulate adenyl- ate cyclase plays a key role in human sperm olfactory receptor-mediated chemotaxis. J Biol Chem 279, 40194– 40203.

Supporting information

Becker A, Socolich M & Zuker CS (1997) A multivalent PDZ-domain protein assembles signalling complexes in a g-protein-coupled cascade. Nature 388, 243–249. 29 Hardie RC & Raghu P (2001) Visual transduction in

Drosophila. Nature 413, 186–193.

30 Popescu DC, Ham AJL & Shieh BH (2006) Scaffolding protein INAD regulates deactivation of vision by pro- moting phosphorylation of transient receptor potential by eye protein kinase C in Drosophila. J Neurosci 26, 8570–8577.

The following supplementary material is available: Table S1. Peptide microarray. This supplementary material can be found in the online version of this article.

31 Wei J, Zhao AZ, Chan GC, Baker LP, Impey S, Beavo JA & Storm DR (1998) Phosphorylation and inhibition of olfactory adenylyl cyclase by CaM kinase II in neu- rons: a mechanism for attenuation of olfactory signals. Neuron 21, 495–504.

FEBS Journal 276 (2009) 7279–7290 ª 2009 The Authors Journal compilation ª 2009 FEBS

7290

Please note: As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.