Báo cáo khoa học: Dopamine D4 receptor oligomerization – contribution to receptor biogenesis doc

Borroto-Escuela2, Wilber Romero-Fernandez2, Kamila Skieterska1, Pieter Rondou1,*, Be´atrice Lintermans1, Peter Vanhoenacker1,, Kjell Fuxe2, Francisco Ciruela3and Guy Haegeman1 1 Laborato

receptor biogenesis

Kathleen Van Craenenbroeck1, Dasiel O Borroto-Escuela2, Wilber Romero-Fernandez2,

Kamila Skieterska1, Pieter Rondou1,*, Be´atrice Lintermans1, Peter Vanhoenacker1,, Kjell Fuxe2, Francisco Ciruela3and Guy Haegeman1

1 Laboratory of Eukaryotic Gene Expression and Signal Transduction (LEGEST), Ghent University Hospital, UZ Gent, Belgium

2 Department of Neuroscience, Karolinska Institutet, Stockholm, Sweden

3 Departament Patologia i Terape`utica Experimental, Facultat de Medicina, Unitat de Farmacologia, IDIBELL-Universitat de Barcelona, L’Hospitalet de Llobregat, Barcelona, Spain

Keywords

bioluminescence resonance energy transfer;

dimerization; dopamine D4receptor;

G protein-coupled receptors; receptor

biogenesis

Correspondence

K Van Craenenbroeck, Laboratory of

Eukaryotic Gene Expression and Signal

Transduction (LEGEST), Ghent

University-UGent, KL Ledeganckstraat 35, 9000 Gent,

Belgium

Fax: +32 (0)9 264 53 04

Tel: +32 (0)9 264 51 35

E-mail:

kathleen.vancraenenbroeck@ugent.be

Present addresses

*Center for Medical Genetics Ghent

(CMGG), Ghent University Hospital – UZ

Gent, Belgium

ActoGeniX, Technologiepark 4, Zwijnaarde,

Belgium

(Received 4 August 2010, revised 23

January 2011, accepted 10 February 2011)

doi:10.1111/j.1742-4658.2011.08052.x

Dopamine D4receptors (D4Rs) are G protein-coupled receptors that play a role in attention and cognition In the present study, we investigated the dimerization properties of this receptor Western blot analysis of the human D4.2R, D4.4R and D4.7R revealed the presence of higher molecular weight immunoreactive bands, which might indicate the formation of receptor dimers and multimers Homo- and heterodimerization of the receptors was confirmed by co-immunoprecipitation and bioluminescence resonance energy transfer studies Although dimerization of a large number

of G protein-coupled receptors has been described, the functional impor-tance often remains to be elucidated Folding efficiency is rate-limiting for

D4R biogenesis and quality control in the endoplasmic reticulum plays an important role for D4R maturation Co-immunoprecipitation and immuno-fluorescence microscopy studies using wild-type and a nonfunctional D4.4R folding mutant show that oligomerization occurs in the endoplasmic reticu-lum and that this plays a role in the biogenesis and cell surface targeting of the D4R The different polymorphic repeat variants of the D4R display dif-ferential sensitivity to the chaperone effect In the present study, we show that this is also reflected by bioluminescence resonance energy transfer sat-uration assays, suggesting that the polymorphic repeat variants have differ-ent relative affinities to form homo- and heterodimers In summary, we conclude that D4Rs form oligomers with different affinities and that dimer-ization plays a role in receptor biogenesis

Structured digital abstract

l D4.4R physically interacts with D4.2R by anti tag coimmunoprecipitation (View interaction)

l D4.xR physically interacts with D4.4R by anti tag coimmunoprecipitation (View interaction)

l D4.xR and D4.xR physically interact by bioluminescence resonance energy transfer (View interaction)

Abbreviations

BRET, bioluminescence resonance energy transfer; CHO, Chinese hamster ovary; DAPI, 4¢,6-diamidino-2-phenylindole; D n R, dopamine D n receptor; ER, endoplasmic reticulum; GPCR, G protein-coupled receptor; HRP, horseradish peroxidase; MP, milk powder; YFP, yellow fluorescent protein.

Although the existence of homo- and⁄ or

hetero-oligo-meric complexes of G protein-coupled receptors

(GPCRs) is generally accepted, their functional

impor-tance often remains to be elucidated [1] Evidence

sug-gests that dimerization or oligomerization is required

for signal transduction by several GPCRs in a fashion

similar to the non-GPCR receptor families, such as

receptor tyrosine kinases In addition, di- or

oligomeri-zation might also be important in biosynthesis

Biosyn-thesis and transport of GPCRs towards the plasma

membrane is a multistep process in which the exit of

GPCRs from the endoplasmic reticulum (ER)

repre-sents a crucial step in the control of their expression at

the cell surface Incompletely folded or misfolded

pro-teins are retained in the ER and targeted for

proteaso-mal degradation; thus, only correctly-folded proteins

are allowed to leave the ER in the direction of the cell

surface The formation of oligomeric complexes

repre-sents an important step in ER quality control because

it may mask retention sequences or hydrophobic

domains that would otherwise result in protein

reten-tion in the ER Several studies, especially of class C

receptors, have shown that GPCR dimerization occurs

in the ER The heterodimerization of c-aminobutyric

acid receptors GABAB1R and GABAB2R is an

obli-gate step in the formation of a functional GABAB

receptor, and masking the retention signal RXR(R) in

the C-terminal of the receptor plays an important role

in intracellular transport [2,3] A role for receptor

dimerization in receptor biogenesis has been reported

for the b2-adrenergic receptor b2-adrenergic receptor

mutants lacking an ER-export motif or receptors

fused to an ER-retention motif still dimerize with the

wild-type b2-adrenergic receptor but trafficking to the

plasma membrane is inhibited [4] Another study with

class A receptors revealed that a truncated dopamine

D3R (D3nf, missing the third intracellular and

sub-sequent regions) [5] prevents cell surface expression of

wild-type dopamine D3Rs upon heteromerization [6]

These data strengthen the hypothesis that early

hetero-merization of wild-type and mutant receptors

influ-ences receptor biogenesis expression

The dopamine receptor family can be subdivided

into the dopamine D1-like receptor subfamily

compris-ing the dopamine D1 receptor (D1R) and the D5R,

which mainly couples to Gs⁄ olf and transduces the

sig-nal via activation of adenylyl cyclase, and the

dopa-mine D2-like receptor subfamily, containing D2R, D3R

and D4R, which couples to Gi⁄ o, thus resulting in the

inhibition of adenylyl cyclase Dopamine receptors

have also been reported to associate with themselves,

as well as with other receptors, and form multireceptor networks that may have unique functional properties Several studies suggest the dimerization of D2Rs [7,8] and even show a potential role for D2R dimerization

in the pathology of schizophrenia [9] Ligand-binding studies confirm D2R oligomer formation and indicate its functional relevance [10] Searching for the func-tional role of GPCR oligomerization is more conve-nient when studying heteromerization For example, it has been demonstrated that D2R and D3R form hete-rodimers that possess a reduced affinity for agonists, and an increased potency for coupling with adenylyl cyclase [11,12] Another example of dopamine receptor association is the D1R⁄ D2R heterodimer: both recep-tors are present in the striatum where they are known

to colocalize and show functional synergy in the mod-ulation of striatal activity [13] In addition, the

D1R⁄ D2R complex requires agonist binding to both receptors for G protein activation and intracellular cal-cium release Therefore, D1R⁄ D2R association could

be important in dopamine-mediated synaptic plasticity

in the brain [14,15] Finally, D1R⁄ D2R heterodimers are able to co-internalize upon selective activation of either receptor [16] Furthermore, heterodimerization between the D1R and D3R in the striatum was reported to be involved in receptor internalization [17] and can also enhance the dopaminergic response in striatal neurons co-expressing both receptors [18] Besides associating with themselves, dopamine recep-tors are also able to associate with other receprecep-tors, such as the adenosine 2A receptor [19,20] and the can-nabinoid CB1 receptor [21,22] Furthermore, a combi-nation of bimolecular fluorescence complementation and bioluminescence energy transfer techniques was used to identify the occurrence of D2R–cannabinoid CB1 receptor–adenosine 2A receptor hetero-oligomers

in living cells [23,24] Recently, it was shown that D2R and adenosine 2A receptor can form oligomers with the metabotropic glutamate type 5 receptor and that they co-distribute in the extrasynaptic plasma mem-brane of the same dendritic spines of striatal synapses [25] Other examples of dopamine receptor heterodi-merization occur with the somatostatin-2 [26] and -5 [27] receptors

In the present study, we investigated the human

D4R oligomerization capability This receptor contains

a variable number of tandem repeats in its third intra-cellular loop, denoted as D4.xR, where x is the number

of repeats [28] Most common are the D4.2R, D4.4R and D4.7R, and we show that all these polymorphic variants can form homo- and heterodimers We have

previously reported [29,30] that the folding efficiency is

rate-limiting in the biogenesis of D4R and the addition

of chemical and pharmacological chaperones can

up-regulate receptor expression and even rescue a

non-functional D4R folding mutant (D4.4M345R) This

chap-erone-mediated effect involves the stabilization of

newly-synthesized receptor in the ER [29,30] In

addi-tion, we also reported that different repeat variants of

D4R display differential sensitivity to the chaperone

effect, namely the D4.2R (with only two repeats) is less

up-regulated compared to the D4.4(with four repeats)

[29] Accordingly, in the present study, using

biolumi-nescence resonance energy transfer (BRET) saturation

studies, we show that D4.4R homodimerization is more

efficient compared to the formation of D4.2R

homodi-mers We also provide evidence that oligomerization of

the dopamine D4R plays a role in receptor biogenesis

and, more particularly, in trafficking of the receptor to

the plasma membrane

Results

Investigation of human D4.xR oligomerization by

classical biochemical approaches

We have previously shown by immunoblot experiments

that, in lysates from cells expressing the D4.2R, several

bands for the receptor can be detected [29,31] Briefly,

the band with a molecular mass of 51–54 kDa (Fig 1,

lane 3, open arrow) represents mature, fully processed

receptor, whereas the lower band(s) with a molecular

mass of 46–49 kDa represents immature, ER-retained

receptor of which the N-linked glycosylation is shorter

than the fully processed glycosylation tree on plasma

membrane-expressed D4R (Fig 1, lane 3, black arrow) Similar results are detected for the D4.0R, D4.4R and

D4.7R (Fig 1) Besides these expected bands of the

D4R monomer, immunoblot analysis of HEK293T transiently transfected with pHA-D4.xR polymorphic variants also revealed bands at a higher molecular weight, probably representing D4R dimers (Fig 1, indicated by *) The specific smear at the top of the blot can represent receptor multimers, receptor forms with different post-translational modification patterns (e.g D4R ubiquitination) [31] or receptor aggregates formed during the denaturation step inherent to this kind of approach It is worth noting that similar results were obtained after receptor immunoprecipita-tion (Fig 1, right); thus, although the concentraimmunoprecipita-tion in the lysates of ER-retained D4R is comparable to the amount of plasma membrane-expressed D4R (Fig 1, left) in the immunoprecipitates, the ER-retained recep-tor is more efficiently immunoprecipitated compared

to the fully glycosylated D4R (Fig 1, right)

We tested whether D4R oligomerization occurred in HEK293T cells transiently expressing these receptors

by performing a co-immunoprecipitation assay First,

we demonstrated heterodimerization of FLAG-D4.4R and HA-D4.2R (Fig 2A) and homodimerization of FLAG-D4.4R and HA-D4.4R (Fig S1) These experi-ments indicate that D4R dimerization already occurs

in the ER because the ER-retained D4R (lower-molec-ular weight band) is clearly isolated To confirm this finding, the same experiment was performed in the presence of brefeldin A, a drug that disrupts the struc-ture and function of the Golgi, preventing protein transport from the ER to the Golgi and thus transport

to the plasma membrane of fully processed receptors

Fig 1 Western blot analysis of HA-tagged D 4 R polymorphic variants HEK293T cells were transiently transfected with pcDNA3(–),

pHA-D4.0R, pHA-D4.2R, pHA-D4.4R and pHA-D4.7R Forty-eight hours after transfection, cells were lysed; 20 lL of lysate was loaded on an 8% gel (left) Immunoprecipitation of the receptor was performed with mouse anti-HA (16B12) (2 lg) and receptor expression was analyzed by 8% SDS ⁄ PAGE and western blotting (right) Next, the blots were probed with rabbit anti-HA sera (dilution 1 : 1000) to detect receptor mono-mers and dimono-mers *D4.x dimers Open arrows indicate mature fully glycosylated D4.xR; black arrows indicate immature ER-retained D4.xR The experiment was repeated three times.

Figure 2B shows that the ER-retained D4Rs form

oligomers; both homodimerization of HA-D4.4R and

FLAG-D4.4R and heterodimerization of HA-D4.2R

and FLAG-D4.4R is demonstrated

To further investigate whether heterodimerization of

D4.2R and D4.4R occurs at the plasma membrane, a

specific membrane co-immunoprecipitation assay was

performed Accordingly, HEK293T cells were

tran-siently transfected with plasmids encoding HA- and

FLAG-tagged D4Rs Forty-eight hours

post-transfec-tion, the cells were first incubated with a primary

anti-HA serum to target specifically plasma membrane-expressed D4Rs Next, cell lysates were made and immunoprecipitation of the FLAG-D4R was per-formed By immunoblotting with anti-HA and anti-FLAG sera, the presence of HA-D4.2and

FLAG-D4.4R, respectively, was visualized in the immuno-precipitates, indicating D4R heterodimerization at the plasma membrane (Fig 2C)

To discriminate between oligomerization in living cells and experimental oligomerization that might occur during lysis, HEK293T cells independently

A

B

C

Fig 2 Dimerization of D4Rs studied by co-immunoprecipitation (A) Dimerization of the D4R in total cell lysates Co-immunoprecipitation studies of FLAG-D 4.4 R and HA-D 4.2 R were performed in HEK293T cells Immunoprecipitation (IP) was performed with mouse anti-HA (16B12) serum (2 lg) After western blot analysis, proteins were visualized with HRP-coupled anti-FLAG M2 or mouse anti-HA (16B12) sera (dilution 1 : 1000) and HRP-coupled anti-mouse (dilution 1 : 3000) Mature, fully processed, plasma membrane (PM) and immature ER-retained (ER) D4R are indicated by an arrow Signal denoting the association of two heavy chains (2 · 50 kDa) (*) or one light chain (25 kDa) (**) of anti-HA sera The same experiment was performed with cells treated for 24 h with brefeldin A (BFA) (B) (C) Dimerization of the D 4 R

at the plasma membrane Immunoprecipitation of membrane D4Rs was performed in HEK293T cells, transiently expressing FLAG-D4.4R and HA-D4.2R, by adding 2 lg mouse anti-HA to the living cells Subsequently, cell lysates were made and membrane-labeled receptors immuno-precipitated, followed by denaturation and SDS ⁄ PAGE Immunoblotting was performed with HRP-coupled anti-FLAG or mouse anti-HA (16B12) sera (dilution 1 : 1000) and HRP-coupled anti-mouse (dilution 1 : 3000) Samples in which cells were independently transfected and mixed post-transfection All experiments were repeated at least three times.

expressing FLAG-D4.4R or HA-D4.2R were mixed

post-transfection and immunoprecipitated under

identi-cal conditions As shown in Figs 2C and S1 (lanes

marked), we only obtained a specific signal when both

receptors were co-expressed in the same cells,

indicat-ing that the human D4R does form dimers in living

cells

Study of D4.xR oligomerization by BRET assays

In HEK293 cells, we examined the possibility of direct

receptor–receptor interaction by constructing

quantita-tive BRET1 saturation curves upon co-transfection of

a constant amount of receptor-Rluc construct and

increasing concentrations of the receptor-yellow

fluo-rescent protein (YFP) plasmids Although the curves

generated by fluorescence- and luminescence-directed

measurements provide the theoretical behavior

suffi-cient to predict receptor oligomerization complexes,

they do not provide sufficient information on the

bind-ing parameters required for proper quantitative

analy-sis of receptor–receptor interactions Accordingly, we

decided to perform BRET analysis in a quantitative

fashion To complete this analysis, we conducted

saturation experiments in which the amount of each

receptor effectively expressed in transfected cells was

monitored for each individual experiment by

correlat-ing both total luminescence and total fluorescence with

the number of [3H]-spiperone-binding sites in

permea-bilized cells Total luminescence and total fluorescence

emitted by the Rluc and YFP fusion proteins were

measured after the addition of the Rluc substrate

h-coelenterazine and direct excitation of the YFP at

485 nm Correlation obtained between receptor density

(the number of total binding sites) and either the

lumi-nescence or fluorescence emitted by each of the

recep-tor fusion molecules was linear (Fig S2) The linear

regression equations derived from these data were used

to transform the luminescence and fluorescence values

to the receptor number BRET1 signals were plotted

as a function of the ratio between the

receptor-YFP⁄ receptor-Rluc numbers

As shown inFig 3A, significant quantitative BRET1

signals were observed for each D4.xR homodimer pair,

confirming the co-immunoprecipitation experiments

displayed in Fig 1 In all cases, BRET1 signals

increased as a hyperbolic function of the increased

concentration of the YFP fusion construct, reaching

an asymptote at the highest concentrations used

How-ever, when comparing the BRET1 signals, it is clear

that, at the concentration of the acceptor

correspond-ing to 50% of the maximum energy transfer (BRET50),

the ability to interact is not the same for the different

homodimer pairs (Table S1) These results suggest that

D4.7R homodimer pairs present the highest affinity fol-lowed by increased reduction of affinity by D4.4R and

D4.2R homodimer pairs, respectively In addition, the BRETmax value for each donor–acceptor pair was found to be lower for the D4.7R homodimer This could suggest that the total number of D4.7R homodi-mers is lower than the total number of the other homodimers under the same experimental conditions

or that the relative position between Rluc and eYFP within the D4.7R donor–acceptor pair was less favor-able for energy transfer The only difference between each isoform is the number of repeat sequences in the third cytoplasmic loop The difference in BRET50 val-ues strengthens the hypothesis that the polymorphic repeat region in the third cytoplasmic loop is involved

in folding efficiency Cells co-expressing D2RRluc and

D2RYFP were used as a positive control, in view of previous studies reporting D2R dimerization [7,8] In these cells, a BRET signal was detected that was higher than that for the other D4.xR isoforms In addi-tion, as a negative control, we used cells co-expressing

D2LRRluc with soluble YFP, leading to marginal sig-nals that increased linearly with increasing amounts of YFP added

When comparing the BRET1 saturation curves obtained for the D4.xR homo- and heterodimers (Fig 3A,B and Table S1), different BRET50 values were obtained, indicating that the receptors had differ-ent relative affinities for one another However, BRET50 values of D4.2R–D4.4R heterodimers showed similar affinity with respect to D4.4R homodimers This has important implications because it suggests that, under basal conditions, D4.2R and D4.4R homo- and heterodimers have a similar probability of forming when the two receptors are heterologously expressed Previous studies indicate that heterotrimeric formation between homologous receptors is highly probable [32]

On the other hand, it is very likely that D4.7R sub-types, when co-expressed with other D4.xR variants, will preferably form D4.7R homodimers because the BRET50 values for homo- versus heterodimers are significantly lower

To test whether the BRET signal was indeed a result

of a specific protein–protein interaction, we performed two essential control experiments First, we co-expressed the D4.xRRluc with an increasing concentra-tion of D4.xRYFP in the presence or absence of a fixed and saturated concentration of D4.xR Comparing the saturation curves generated in both cases, we can con-clude that the overexpression of a fixed concentration

of the receptor significantly shifted the saturation curves of the D4.xRRluc–D4.xRYFP pair to the right,

resulting in an increase BRET50value (Fig S3A shows

an example for the D4.2R homodimer and Fig S3B

shows an example for the D4.2R–D4.4R heterodimer)

Second, we overexpressed increasing concentrations of

the D4.xR in combination with the protomers of the

BRET pair (constant ratio 1 : 1) and investigated

whether the wild-type receptor could reduce the BRET

signal Over-expression of D4.xR significantly reduced

the BRET ratio, as shown by the BRET competition

curves (Fig S3)

Finally, to examine the effect of ligands on D4.xR

oligomerization, cells co-expressing D4.2RRluc and

D4.2RYFP were incubated with 10 lm of the full D4R

agonist WAY-100635, the D4R antagonist A-381393

or the inverse D4R agonist FAUC F41 for 10 min Stimulation with any of these ligands failed to promote any consistent change in the BRET1 ratio, indicating that the dimers form constitutively, and that agonist-mediated receptor activation does not affect their oligomerization state (Fig 4)

Functional consequences of D4R oligomerization The data from the BRET and co-immunoprecipitation experiments clearly show that oligomerization already occurs in the ER Therefore, we hypothesized that

0.10

A

120

90

60

30

0.05 0.06

0.04

0.02

0.00

0.04

0.03

0.02

0.01

0.00

0

0.08

0.06

0.04

1 ratio

1 ratio

1 ratio

1 ratio

[Receptor-YFP]/[Receptor-Rluc]

0.02

0.00

D2LR-Rluc/eYFP

D2LR-Rluc/D2LR-YFP D

4.7 R-Rluc/D4.7R-YFP

D4.4R-Rluc/D4.4R-YFP

D4.2R-Rluc/D4.2R-YFP

D4.2R-Rluc/YFP

D4.2R-Rluc/D4.2R-YFP D4.2R-Rluc/D4.7R-YFP

D4.2R-Rluc/D4.4R-YFP

D4.4R-Rluc/D4.7R-YFP

D4.2R-Rluc/D4.7R-YFP

D4.7R-Rluc/eYFP

D4.7R-Rluc/D4.7R-YFP

D2LR-Rluc/eYFP

D2LR-Rluc/D2LR-eYFP D

4.7 R-Rluc/D4.7R-eYFP

D4.4R-Rluc/D4.4R-eYFP

D4.2R-Rluc/D4.2R-eYFP

[Receptor-YFP]/[Receptor-Rluc]

[Receptor-YFP]/[Receptor-Rluc]

[Receptor-YFP]/[Receptor-Rluc]

B

Fig 3 Quantitative analysis of D4.xR homodimerization (A) and heterodimerization (B) BRET 1 donor saturation curves were performed by transfecting HEK293T cells with a constant DNA concentration of acceptor Rluc and increasing concentrations of donor receptor-YFP constructs BRET1ratio, total fluorescence and total luminescence, as well as transformed values into receptor numbers, were deter-mined as described in the Materials and methods The curves represent ten saturation curves that were fitted using a nonlinear regression equation assuming a single binding site.

oligomerization could play a functional role in D4R

maturation We have shown previously that chemical

(dimethylsulfoxide, glycerol) and pharmacological

(receptor ligands) chaperones can help in the folding

procedure of the receptor in the ER, thereby

decreas-ing receptor degradation in the proteasome and

enhancing expression on the plasma membrane The

pharmacological chaperone quinpirole (a D2-like

receptor agonist) clearly enhances the expression not

only of wild-type D4.4R, but also of the folding mutant

D4.4M345R This folding mutant D4.4M345R does not

meet the quality control of the ER and is routed to

the proteasome for degradation [29,30] We used this

folding mutant to investigate the role of

oligomeriza-tion in D4.4M345R folding and subsequent plasma

membrane expression Therefore, Chinese hamster

ovary (CHO) cells stably expressing the folding mutant

FLAG-D4.4M345R [29] were transiently co-transfected

with the control vector pcDNA3 or the vector

encod-ing the wild-type HA-D4.2R (Fig 5) Untreated CHO

FLAG-D4.4M345R cells transfected with the back bone

vector pcDNA3 do not show clear FLAG-D4.4M345R

expression (Fig 5, left) Upon treatment of these cells

with the pharmacological chaperone (quinpirole,

10 lm, 16 h), the receptor is expressed (Fig 5, middle)

This is in agreement with our previous data [29] Note

that not all cells show a clear expression of the

recep-tor, which could be the result of a loss of receptor

expression (e.g silencing of the constitutive

FLAG-D4.4M345R gene transcription) [33] When this CHO

FLAG-D4.4M345R cell line was transiently transfected

with the plasmid coding for a wild-type D4.2R, namely

pHA-D4.2R (Fig 5, right, red), the mutant

FLAG-D4.4M345R is expressed in the CHO cell line (Fig 5,

right, green) In these experiments, receptors on the

membrane were first labeled with FLAG and

anti-HA sera Then cells were fixed, labeled with secondary antibodies and, finally, DNA was stained with 4¢,6-di-amidino-2-phenylindole (DAPI) to visualize the nuclei (blue) These results indicate that receptor oligomeriza-tion can have a chaperone effect In Fig S3, we also included the BRET data of the folding mutant FLAG

D4.4M345R The results confirm the interaction between the mutant D4.4M345R and both the wild-type D4.2R and D4.4R

Discussion

During recent years, the number of studies reporting GPCR dimerization has increased greatly and it is now well accepted that most GPCRs are able to form homodimers The data obtained in the present study demonstrate that the dopamine D4R is no exception Western blot analysis already indicated that the differ-ent polymorphic variants of the dopamine D4R (D4.2R, D4.4R and D4.7R) are interacting with them-selves The present study is an extra element in the research on dopamine receptor oligomerization that, until now, has focused on the dopamine receptors D1,

D2, D3 and D5 By performing traditional co-immuno-precipitation assays, we confirmed both D4 receptor homodimerization (HA-D4.4R and FLAG-D4.4R) and heterodimerization (HA-D4.2R and FLAG-D4.4R) Although some criticisms suggest that GPCR dimeriza-tion might be promoted at relatively high receptor expression levels and hence potentially be at least

Fig 4 Ligand binding effect on D4.2R homodimerization Effects of

10 min of stimulation of 10 l M full agonist WAY-100635, antagonist

A-381393 and the inverse agonist FAUC F41 on the BRET 1 ratios

for the human D4.2R homodimers Ratios are expressed as the

mean ± SEM from at least six experiments.

Fig 5 Role for dimerization of D 4 R in receptor biogenesis CHO cells, stably transfected with pFLAG-D4.4M345 R, were grown on coverslips in six-well plates and transiently transfected with pcDNA3 or a plasmid encoding the wild-type HA-D 4.2 R Thirty-six hours post-transfection, cells were left untreated or treated for

16 h with the D2-like agonist quinpirole (Q, 10 l M ) Membrane-expressed D 4.4 R was first recognized by rabbit anti-FLAG serum (for the mutant FLAG-D4.4M345 R) and mouse anti-HA serum (for the wild-type HA-D4.2R) Subsequently, cells were fixed and samples were incubated with anti-rabbit Alexa 488 (green) and anti-mouse Alexa 594 (red) and the cell nuclei were stained with DAPI (blue) The images shown are representative of the whole experiment (performed in triplicate).

partially an artifact of overexpression, studies of

b2-adrenergic receptor dimerization have indicated that

dimerization is unaltered over a wide range of

expres-sion levels [34] We also obtained evidence with the

quantitative BRET1 technique (keeping receptor

expression near physiological level) indicating that

D4.xRs can form homo- and heteromers, although with

different degrees of efficiency and affinity, with the

D4.7R being the least capable to form heteromers, as

seen from the reduced BRETmax and increased

BRET50 values compared to those obtained with the

D4.2R and D4.4R protomers

From these experiments, we can conclude that the

human D4R does form hetero- and homodimers in

liv-ing cells It is noteworthy that, for the D2R, the

mini-mal signaling unit is suggested to be two receptors and

one G protein [32] A model developed to study D2R

dimerization suggests that the way in which the two

protomers contribute to the active complex with the

G protein is not symmetric and that activation requires

different conformational changes in each protomer

[32] On the other hand, the results of the present

study show only a weak D4R oligomerization at the

plasma membrane, although membrane

immunofluo-rescence studies (Fig 5), whole cell binding assays

(data not shown) and functionality studies (data not

shown) confirm that the D4R is functionally expressed

on the plasma membrane The low amount of receptor

dimerization upon immunoprecipitation of the D4R at

the cell surface can be the result of a transient

interac-tion of the D4R protomers at the plasma membrane,

as recently suggested for several GPCRs [35,36]

The band pattern of the co-immunoprecipitation

specified that receptor dimerization already occurred in

the ER We have studied D4R biogenesis intensively

[29,30] and shown that folding of the D4R in the ER

forms the bottle neck of receptor biogenesis Several

drugs, both chemical and pharmacological chaperones,

can help to enhance folding efficiency in the ER As

soon as membrane proteins are correctly folded, they

can proceed to the Golgi and to the plasma

mem-brane Slow folding of receptors in the ER enhances

ER-associated degradation by the proteasome and

leads to a decrease of mature receptor on the plasma

membrane Because the data from the present study

indicate that receptor dimerization starts in the ER, it

was tempting to assume that this process could

influ-ence receptor biogenesis We obtained data to

strengthen this hypothesis from two independent

experiments: (a) expression of the D4R folding mutant

(D4.4M345R) upon co-expression of wild-type receptor

as visualized with a specific membrane labeling

immu-nofluorescence technique and (b) BRET analysis

indicating that homodimerization of D4.7Rs is more efficient compared to homodimerization of D4.4Rs and

D4.2Rs We do not know whether D4R dimerization involves the masking of a retention signal (as discussed

in the Introduction) because the mutant D4.4M345R is a folding mutant, although we can conclude that dimer-ization helps with D4R biogenesis The acquisition of this role for D4R dimerization does not rule out the possibility that oligomeric D4Rs may have additional functions, once they are brought to the cell surface

In summary, we conclude that the D4R forms hetero-and homodimers This dimerization already occurs in the ER and the quaternary structure enhances the fold-ing process of the receptor, which is linked to receptor

ER export and cell surface trafficking

Materials and methods

Plasmids

The plasmids pFLAG-D4.4R, pFLAG-D4.4M345TR (folding

kindly provided by Dr H Van Tol (University of Toronto, Ontario, Canada) An N-terminal myc (EQKLISEED) epitope-tagged D4R was created by PCR in three steps and plasmids for BRET were made using standard PCR and fragment replacement strategies (Fig S4) The reading frame and PCR integrity of all cloned constructs were confirmed by DNA sequencing

Cell culture, transfection and western blot analysis

Development of the CHO cell line stably transfected with the D4R folding mutant (CHO FLAG-D4.4M345TR), as well

as the transient transfection method using lipofectamine (Invitrogen, Carlsbad, CA, USA), has been described previously [29] HEK293T cells were transiently transfected with 10 lg of plasmid DNA using the poly(ethylenimine) transfection method Therefore, cells were grown in 10 cm dishes until subconfluency in DMEM (Invitrogen) with 10% fetal bovine serum Before transfection, the medium was refreshed with 9 mL of DMEM, supplemented with 2% fetal bovine serum A mixture of 475 lL of serum-free

(Sigma Aldrich, St Louis, MO, USA) was added to a solu-tion of 500 lL of serum-free medium containing 10 lg of DNA Upon mixing thoroughly and incubation for 10 min

at room temperature, the DNA⁄ poly(ethylenimine) mixture was added to the cells Six hours later, the medium was refreshed with DMEM, supplemented with 10% fetal bovine serum Forty-eight hours after transfection, the cells were washed twice with NaCl⁄ Pi, collected by scraping,

centrifuged and frozen at)70 C for at least 1 h RIPA

ly-sates were performed as described previously [31] and

loaded on a 10% SDS⁄ PAGE gel Proteins were transferred

onto a nitrocellulose membrane (Schleicher & Schuell

Bio-sciences, Dassel, Germany) Subsequently, membranes were

Tween 20, pH 7.6) overnight, after which the membranes

were incubated for 1 h with primary antibodies (dilution

1 : 1000) mouse anti-HA (clone 16B12; Covance Research

Products, Berkley, CA, USA), rabbit anti-HA (Genetex,

Irvine, CA, USA) or mouse anti-FLAG (clone M2; Sigma

Aldrich) in 5% MP⁄ NaCl ⁄ Tris ⁄ Tween 20 Thereafter, the

blots were incubated with secondary antibodies (dilution

1 : 2000) anti-mouse or anti-rabbit, horseradish peroxidase

(HRP)-linked (Amersham Biosciences, Piscataway, NI,

Western Lightning Cheminuluminescence Reagent Plus

(PerkinElmer Life Sciences, Wellesley, MA, USA) detection

system

Co-immunoprecipitation

Co-immunoprecipitation studies were performed as

previ-ously described [37] In short, HEK293T cells were grown

in 10 cm dishes and transiently transfected as described

above For control experiments, cells were independently

transfected with plasmids encoding only one receptor type

and mixed after transfection Cells were collected and

fro-zen at)70 C after which the cells were lysed in 400 lL of

Triton-X100, 0.5% sodium deoxycholate, 0.2% SDS and

was used for immediate testing of protein expression by

western blot analysis To the rest of the lysate, 2 lg of either

primary antibody mouse HA 16B12 or mouse

anti-FLAG M2 was added After rotation for 4 h at 4C, 20 lL

of protein A Trisacryl beads (Pierce, Rockford, IL, USA)

washing the beads three times with RIPA buffer, the beads

were denatured at 37C for 10 min in SDS-loading buffer

bromophenol blue) + 20 mm dithiothreitol (freshly added)

Proteins were separated on a 10% SDS⁄ PAGE gel and

trans-ferred onto a nitrocellulose membrane Dilutions of 1 : 1000

mouse anti-HA 16B12 and 1 : 1000 mouse anti-FLAG M2

HRP (Sigma) were used as primary antibodies and 1 : 2000

HRP-linked anti-mouse (Amersham Biosciences) as

secon-dary antibody

For isolation of the receptor at the plasma membrane,

cells, transiently transfected with dopamine

receptor-encod-ing plasmids, were incubated with 2 lg of antibody for 1 h

at 37C in serum-free medium before lysis The remainder

of the protocol is similar to that described above

Immunofluorescence microscopy

CHO FLAG-D4.4M345R cells were seeded in wells with cov-erslips and transfected using lipofectamine Thirty-six hours later, cells were treated with quinpirole (10 lm, 16 h) Membrane receptors were labeled by adding primary anti-body [rabbit anti-FLAG (Sigma) and mouse anti-HA 16B12], diluted 1 : 500 in serum-free medium supplemented with Hepes, to the cells for 1 h at 37C After labeling, cells were fixed (150 mm NaCl, 10 mm sodium phosphate,

pH 7.4, 3.7% formaldehyde) for 15 min at room tempera-ture After washing, cells were quenched in 50 mm glycine for 15 min and washed again Cells were permeabilized with Blotto⁄ Triton (3% MP, 1 mm CaCl2, 0.1% Triton X-100,

50 mm Tris HCl, pH 7.5) for 20 min at room temperature After washing, cells were incubated for 5 min with Blotto

with secondary antibody (anti-rabbit Alexa Fluor 488 and anti-mouse Alexa Fluor 594; Invitrogen) diluted 1 : 500 in Blotto for 20 min Nuclei were visualized by incubating cells for 5 min with DAPI Samples were analyzed using the Axiocam 200 microscope (Zeiss, Thornwood, NY, USA)

BRET1assays

HEK293T cells were transiently transfected with a constant (1 lg) amount of cDNA encoding D4.2RRluc, D4.4RRluc or

D4.2RYFP, D4.4RYFP or D4.7RYFP cDNA’s Forty-eight hours after transfection, HEK293T cells were rapidly washed twice in NaCl⁄ Pi, detached, and resuspended in the same buffer Cell suspensions (20 lg of protein) were dis-tributed in duplicate into 96-well microplates (either black clear-bottomed or white opaque, Corning 3651 or 3600; Corning Inc., Lowell, MA, USA) for fluorescence and luminescence determinations The total fluorescence of cell suspensions was quantified and then divided by the back-ground (mock-transfected cells) in a POLARstar Optima plate-reader (BMG Lab-Technologies, Offenburg, Ger-many) equipped with a high-energy xenon flash lamp, using

a 10 nm bandwidth excitation filter at 485 nm, and a

10 nm bandwidth emission filter corresponding to 535 nm Total bioluminescence was determined on samples incu-bated for 10 min with 5 lm h-coelenterazine (Molecular Probes, Eugene, OR, USA) The background values for total luminescence were negligible and subtracted from

substrate was added at a final concentration of 5 lm, and readings were performed 1 min later using the POLARstar Optima plate-reader, which allows the sequential integra-tion of the signals detected with two filter settings [485 nm (440–500 nm) and 530 nm (510–560 nm)] The BRET ratio

is defined as described previously [38] Ligands-promoted

obtained in the absence of ligand (agonist or antagonist)

addition from that obtained in the presence of the ligands

of ligand incubation

Luminescence and fluorescence levels of several

receptor-RLuc and receptor-YFP fusion proteins have been found

to be linearly correlated with receptor numbers [33]

Because this correlation is an intrinsic characteristic of

each fusion protein, correlation curves have to be

estab-lished for each construct HEK293 cells were transfected

with increasing cDNA concentrations of the Receptor-Rluc

or YFP fusion protein constructs Maximal luminescence

and fluorescence was determined as described above and

correlated with relative receptor number determined in the

same cells as described in the radioligand binding

experi-ments (Table S2) Luminescence and fluorescence were

both plotted against binding sites, and linear regression

curves were generated The standard curves generated for

each single experiment were used to transform fluorescence

and luminescence values into fmol of receptor Thus, the

(receptor-YFP)⁄ (receptor-RLuc) ratios, which allowed us

control the number of cells and also to express receptor

numbers in fmolÆmg)1of total cell protein, protein

concen-tration was determined using a Bradford assay kit

(Bio-Rad, Hercules, CA, USA)

Data analysis

All binding data were analyzed using graphpad prism,

ver-sion 4.0 (GraphPad Prism, San Diego, CA, USA) BRET

saturation curves were analyzed using graphpad prism

Isotherms were fitted using a nonlinear regression equation

or luminescence and receptor density was analyzed by a

linear regression curve fitting with the same software For

statistical evaluation, and unless otherwise specified,

one-way analysis of variance was used

Acknowledgements

K.V.C has a postdoctoral fellowship from FWO

(Fonds voor Wetenschappelijk Onderzoek) This work

was supported by grants SAF2008-01462 and

Con-solider-Ingenio CSD2008-00005 from Ministerio de

Ciencia e Innovacio´n to F.C.; by European Social

Foundation and Gobierno de Catalunya

FI2004-BE2006 to D.O.B.-E.; and from the Swedish Research

Council (04X-715), Torsten and Ragnar So¨derberg

Foundation to KF The authors would like to thank

Hubert Van Tol for helpful discussion at the beginning

of the study

References

1 Milligan G (2009) G protein-coupled receptor hetero-dimerization: contribution to pharmacology and function Br J Pharmacol 158, 5–14

2 Galvez T, Duthey B, Kniazeff J, Blahos J, Rovelli G, Bettler B, Prezeau L & Pin JP (2001) Allosteric interac-tions between GB1 and GB2 subunits are required for optimal GABA(B) receptor function EMBO J 20, 2152–2159

3 Margeta-Mitrovic M, Jan YN & Jan LY (2000) A trafficking checkpoint controls GABA(B) receptor heterodimerization Neuron 27, 97–106

4 Salahpour A, Angers S, Mercier JF, Lagace M, Marullo

S & Bouvier M (2004) Homodimerization of the beta2-adrenergic receptor as a prerequisite for cell surface targeting J Biol Chem 279, 33390–33397

5 Schmauss C, Haroutunian V, Davis KL & Davidson M (1993) Selective loss of dopamine D3-type receptor mRNA expression in parietal and motor cortices of patients with chronic schizophrenia Proc Natl Acad Sci USA 90, 8942–8946

6 Karpa KD, Lin R, Kabbani N & Levenson R (2000) The dopamine D3 receptor interacts with itself and the truncated D3 splice variant d3nf: D3-D3nf interaction causes mislocalization of D3 receptors Mol Pharmacol

58, 677–683

7 Callier S, Snapyan M, Le Crom S, Prou D, Vincent JD

& Vernier P (2003) Evolution and cell biology of dopa-mine receptors in vertebrates Biol Cell 95, 489–502

8 Zawarynski P, Tallerico T, Seeman P, Lee SP, O’Dowd

BF & George SR (1998) Dopamine D2 receptor dimers

in human and rat brain FEBS Lett 441, 383–386

9 Wang M, Pei L, Fletcher PJ, Kapur S, Seeman P & Liu

F (2010) Schizophrenia, amphetamine-induced sensi-tized state and acute amphetamine exposure all show a common alteration: increased dopamine D2 receptor dimerization Mol Brain 3, 25

10 Strange PG (2005) Oligomers of D2 dopamine recep-tors: evidence from ligand binding J Mol Neurosci 26, 155–160

11 Scarselli M, Novi F, Schallmach E, Lin R, Baragli A, Colzi A, Griffon N, Corsini GU, Sokoloff P, Levenson

R et al (2001) D2⁄ D3 dopamine receptor heterodimers exhibit unique functional properties J Biol Chem 276, 30308–30314

12 Maggio R & Millan MJ (2010) Dopamine D2-D3 receptor heteromers: pharmacological properties and therapeutic significance Curr Opin Pharmacol 10, 100– 107

13 Aizman O, Brismar H, Uhlen P, Zettergren E, Levey

AI, Forssberg H, Greengard P & Aperia A (2000) Ana-tomical and physiological evidence for D1 and D2 dopamine receptor colocalization in neostriatal neurons Nat Neurosci 3, 226–230