Evidence for functional pre-coupled complexes of receptor heteromers and adenylyl cyclase

G protein-coupled receptors (GPCRs), G proteins and adenylyl cyclase (AC) comprise one of the most studied transmembrane cell signaling pathways. However, it is unknown whether the ligand-dependent interactions between these signaling molecules are based on random collisions or the rearrangement of pre-coupled elements in a macromolecular complex. Furthermore, it remains controversial whether a GPCR homodimer coupled to a single heterotrimeric G protein constitutes a common functional unit. Using a peptide-based approach, we here report evidence for the existence of functional pre-coupled complexes of heteromers of adenosine A2A receptor and dopamine D2 receptor homodimers coupled to their cognate Gs and Gi proteins and to subtype 5 AC. We also demonstrate that this macromolecular complex provides the necessary frame for the canonical Gs-Gi interactions at the AC level, sustaining the ability of a Gi-coupled GPCR to counteract AC activation mediated by a Gs-coupled GPCR.

1) Using BiFC they "force" dimerization of either homodimer or the A2AR/D2R heterodimer. 48 hours after transfection they add TAT-TM peptides for 4 hours and test for their ability to disrupt complexes as measured by a loss of fluorescence. They may have confirmed that such tagged versions of their receptors and AC5 remain functional (which they should reference) or if not, they should demonstrate here. In my experience such reconstituted dimers do not come apart. So what do their experiments actually show? Prevention of the formation of new dimers? They should demonstrate this more directly. I am concerned that not all the TAT-TM peptides get into the membrane with the same efficacy-how do the authors control for that? The scrambled peptides that work do not give me confidence in this regard either.
2) If these forced dimers don't come apart, how do we interpret the effects of receptor ligands on fluorescence? BiFC is a crude tool to examine the conformational dynamics of protein complexes in response to ligands unless it can be demonstrated that such complexes can come together and apart in a dynamic equilibrium.
3) I have similar concerns for the PLA experiments-do they suggest the dimers are coming apart in response to TAT-TMs? This goes against the core argument the authors are trying to make. Also, the authors could demonstrate the PLA does not occur in the absence of receptor expression in HEK 293 cells.
4) I don't buy the notion that all heterotetramers are dimers of homodimers as the authors suggest in the introduction. The dynamics of GPCR complexes are highly variable depending on the molecular and cellular context assayed. Personally, I think metastability is the core feature of these complexes and I like the approach taken here to examine it. Perhaps the authors could be more nuanced in the introduction. The question of proximity versus stability is the key issue here-perhaps a third technique like co-immunoprecipitation with and without TAT-TM or agonist could settle the issue?
5) The elephant in the room is what happens to G protein partners in these complexes? Are they important for formation of the R/R/E complexes or merely transducers of conformational information in response to agonist.

Minor comments
1) The word "strong" in the abstract is a relative term.
Reviewer #2 (Remarks to the Author): The manuscript "Functional pre-coupled complexes of receptor heteromers and adenylyl cyclase" aims to address the existence of pre-coupling between GPCRs (A2A and D2R) and transmembrane adenylyl cyclase type 5 using engineered mimetic transmembrane domain peptides to disrupt interactions. Although it is quite accepted that GPCRs and G proteins must be in close proximity to their effectors this remains a fundamental question in the field not much is known on whether these different actors can form a complex in absence of receptor ligand. The study uses fluorescencebased technology to elucidate protein/protein organization in association with computational models and signaling experiments in both heterologous system and primary cells. Overall, this is an elegant work which provides new insights on how GPCR heteromers and AC5 may pre-form a functional complex and how ligand may rearrange these complexes. The authors should consider the following:

1-
Authors assumed that TMs mimetic peptides are all correctly orientated and integrated in the membrane on the basis of the TAT-fused strategy. However, no evidences are provided. This is important to check especially for TM2,3 and 7 of A2A and D2R as they have no apparent effect on Bimolecular Fluorescence complementation.

2-
How were determined the time of pretreatment (4 hours) and the concentration of peptide (4 M)? Moreover, Dose -responses may allow to establish the relative IC50 for each peptide as any other pharmacological tool.

3-
BRET data suggest the formation of a complex between AC5 and D2R or AC5 and A2A. Although non-fluorescent receptor is added in the experiment to allow the formation of the heterotetramer, authors did not verify whether AC5 overexpression may affect the heteromerization of A2A with D2R. This needs to be verified.

4-
In line with previous comment, AC5 TM peptides should be tested on A2A/D2R heteromer formation as negative controls as well as the effect of A2A TMs or D2R TMs on A2A/AC5 and D2R/AC5 respectively.

5-
Authors proposed an "hypothetical" model in which AC5 and heterotetramers (A2A/D2R) associate in a linear manner ( fig. 3g and supp fig. 3) suggesting a certain stochiometry. This is an interesting hypothesis however, this is mostly based on overexpressed proteins. Considering the relative expression of both receptors in AC5 in striatal neurons, authors should comment on how this may also be true in native cells. PLA for AC5/D2R or AC5/A2A in neurons may reveal the preassemble complexes in a relevant cell model.

6-
TM5n impaired association between receptors and AC5 is as good as TM5b, but not TM5 in absence of agonist. Additionally, authors hypothesized that TM5n may be part of intracellular IL2 of AC5 and involved in pre-coupling with receptors. Thus, a scramble IL2 peptide would be a good negative control here.
Minor comments: 1-Line 190. TM4n also behave as a negative control peptide and is not mentioned.

2-
How the association between GPCRs and AC5 would change if agonists were added? Since TMs from M1 domain show a marked increased sensitivity to TM2,3,5 and 6 disrupting peptides, some changes in the association between AC5 and receptors may occur.

3-
Would co-treatment with several peptides have an additive effect on A2A-D2R or AC5/receptors complexes?
Reviewer #3 (Remarks to the Author): In this current study Navarro and co-workers addressed highly important questions in the GPCR field related to homo-or hetero-oligomerization and associated GPCR/effector-complex constitution. Higher order complexes between GPCR protomers, homo-or heteromeric, are in fact known and evidenced for many class A and other classes of GPCRs, but detailed information regarding structural prerequisites or functional consequences/relations of those complexes (GPCR interactome) is rarely available so far. Approaches to obtain experimental data deciphering molecular backgrounds including structural (e.g. active versus inactive state conformations) and functional (signaling) aspects are still challenging with respect to appropriate methods.
To the authors credit they comprehensively studied focused issues, means a systematic "multiplesite" perspective on potentially contacting partners and functional consequences.
The techniques used in this study are appropriate, the received data are not over-interpreted in the conclusions and discussion. The context is described well throughout the manuscript and under consideration of previously and recently published information.
To the reviewers opinion this study can be seen as a progressive contribution and suggestion to several specific but also general aspects in the GPCR field. This concerns the TMH6-TMH6 interface between GPCR homomers, the simultaneous heterooligomeric dimer-dimer interface at TMH4-TMH5, the independence of homodimer-interfaces on the activity state, moreover the concluded pre-coupling or close spatial distance of AC to the inactive receptors, changes in receptor-AC interaction during activation, and finally implications for canonical signaling (Gs and Gi signaling in heteromeric GPCR complexes). The results may also have impact on parameters or mechanisms related to selectivity of GPCR signaling, but definitively they contribute information on pre-coupling partners of GPCRs, at least for the studied receptors.
Moreover, the manuscript is well balanced between the particular sections and the data are presented convincing.
Major points: -Are data available proving on GPCR/AC interaction without available G-protein? Or is a Gprotein/AC pre-complex a prerequisite for interaction with the GPCR (or a GPCR/G-protein precomplex for interaction with AC, respectively)?
Minor points: - The results and suggested scenario of receptor/AC pre-coupling must be strongly dependent (independently from the used method) also on the capacity of the investigated receptors to signal in a ligand-free (constitutive) basal state. It would be helpful to clarify this aspect for A2A and the D2R in the discussion section.
-It is of note, that both homodimers here are characterized by a TMH6-TMH6 interface. A discussion is needed how a TMH6-TMH6 interface in the homodimers may have impact on receptor activation (or how receptor activation can occur in such a constellation), because TMH6 outward movement is a significant feature of GPCR activation from a structural perspective and such a postulated interface may constrain this helix.
1 Answers to Reviewer #1 1. Using BiFC they "force" dimerization of either homodimer or the A2AR/D2R heterodimer. 48 hours after transfection they add TAT-TM peptides for 4 hours and test for their ability to disrupt complexes as measured by a loss of fluorescence. They may have confirmed that such tagged versions of their receptors and AC5 remain functional (which they should reference) or if not, they should demonstrate here. In my experience such reconstituted dimers do not come apart. So what do their experiments actually show? Prevention of the formation of new dimers? They should demonstrate this more directly. I am concerned that not all the TAT-TM peptides get into the membrane with the same efficacy-how do the authors control for that? The scrambled peptides that work do not give me confidence in this regard either.

If these forced dimers don't come apart, how do we interpret the effects of receptor ligands on fluorescence? BiFC is a crude tool to examine the conformational dynamics of protein complexes in response to ligands unless it can be demonstrated that such complexes can come together and apart in a dynamic equilibrium.
In these two points, the reviewer expresses his/her concerns about two assumptions regarding BiFC. First, that fusion of the two complementary YFP fragments "forces" the intermolecular interaction between the proteins to which they are separately attached. In fact, it is the complex formation between the two interacting proteins what forces the fusion of the YFP fragments (Rose et al., 2010, Brit J Pharmacol, 159, 738 Schmidt et al. (2003, Mol Cell 12, 1287, in their study about the differential association and dissociation of specific complexes of NF-κB with two different isoforms of the I-κB inhibitor; also by Guo et al. (2005, J Biol Chem, 280, 1438, in experiments about the βγ-dependent dissociation of complexes of phospholipases Cβ 2 and Cδ 1 ; and by Anderie andSchmid (2007, Cell Biol Int, 31, 1131), with visualizing in vivo of actin cytoskeleton dynamics (reversible actin/actin BiFC complexes). But also, our results constitute a significant demonstration of reversible BiFC. And this lies on the specificity of the peptide approach, demonstrated by parallel experiments of BiFC and PLA. However, the reviewer also expresses concern about a possible differential efficacy of the peptides that could depend on a differential efficacy to integrate in the plasma membrane. We agree with the reviewer about the possibility that this integration can be different among peptides, as we particularly discuss in the manuscript when evaluating hydropathy values in order to explain the positive effect of the two differently oriented peptides mimicking putative TM 5 of AC5 (Table 1) Jastrzebska et al., J Biol Chem, 2015, 290, 25728; cited in the text). Fig. 1) Supplementary Information (Supplementary Fig. 3a and  3b)

, we have also included the controls indicated by the reviewer, showing the lack of A2AR-D2R complexes in HEK cells only expressing A2AR or D2R.
4. I don't buy the notion that all heterotetramers are dimers of homodimers as the authors suggest in the introduction. The dynamics of GPCR complexes are highly variable depending on the molecular and cellular context assayed. Personally, I think metastability is the core feature of these complexes and I like the approach taken here to examine it. Perhaps the authors could be more nuanced in the introduction. The question of proximity versus stability is the key issue here-perhaps a third technique like co-immunoprecipitation with and without TAT-TM or agonist could settle the issue? Fig. 7).

The elephant in the room is what happens to G protein partners in these complexes? Are they important for formation of the R/R/E complexes or merely transducers of conformational information in response to agonist.
This is an important question already addressed in the first version of the manuscript, that obviously needs to be improved. In fact, the same point was also raised by reviewer 6. Minor comment: The word "strong" in the abstract is a relative term.

#2 (2 nd minor comment) and #3 (major point). The following sentences have been added or modified in the Results and Discussion sections (and a new reference included): "The results indicate TM peptides that disrupt heteromerization do not disrupt the individual functional interactions between the receptors and AC5 in striatal cells in
We deleted "strong" from the abstract Answers to Reviewer #2 1. Authors assumed that TMs mimetic peptides are all correctly orientated and integrated in the membrane on the basis of the TAT-fused strategy. However, no evidences are provided. This is important to check especially for TM2,3 and 7 of A2A and D2R as they have no apparent effect on Bimolecular Fluorescence complementation. He et al. (2011, Neuron, 69, 120;previously

cited in the Methods section and now explicitly mentioned in the Results section), when determining the interface of the mu-delta-opioid receptor heteromer. Their method implied the immuno-cytochemical detection of TM peptides with GST fused to the N-terminus and TAT fused to either the N-or the C-terminus in permeabilized and non-permeabilized neuronal preparations. The localization of the TAT sequence determined the localization of GST, facing the intra-or the extracellular space.
The corresponding TAT-TM peptides where always integrated in the membrane, but when facing the cytosol, GST could only be visualized in the permeabilized preparation. Therefore, we believe it should not be necessary to replicate these findings, which should require a large number of experiments. In addition, as mentioned in the response to point 1 by reviewer #1 (and also now mentioned in the Results section), we strongly believe that our qualitative results demonstrate the selectivity of the peptide approach. Indeed, the specificity of these results provided the information for only one solution in the computational modeling of the heterotetramer and the very same peptides that disrupted A2AR-D2R heteromerization (as demonstrated by two methods, BiFC and PLA) were the only ones able to disrupt the functional canonical interaction at the AC5 level. Additional positive results with the negative TM peptides (TM2, 3 and TM7) would have probably yielded no computational solution.
2. How were determined the time of pretreatment (4 hours) and the concentration of peptide (4 μM)? Moreover, Dose-responses may allow to establish the relative IC50 for each peptide as any other pharmacological tool.

About the justification for the optimal concentration and time of incubation of the TM peptides, we had in fact performed concentration-and time-dependent response experiments of the possible BiFC disruption by all seven TM peptides of the A2AR in cells transfected with A2AR-nYFP and A2AR-cYFP. The results are now included in Suplementary Fig. 2, and show that the selective disruptive effect of TM6 of A2AR is optimal at the concentration of 4 μM and with a pretreatment of 4 hours. In view of
their very similar physicochemical properties, we considered that the same parameters could be used for TM peptides of the D2R and AC5. In fact, we obtained differential specific positive effects with those peptides in BiFC experiments which allowed totally congruent results with PLA and cAMP experiments.

BRET data suggest the formation of a complex between AC5 and D2R or AC5 and A2A.
Although non-fluorescent receptor is added in the experiment to allow the formation of the heterotetramer, authors did not verify whether AC5 overexpression may affect the heteromerization of A2A with D2R. This needs to be verified.
4. In line with previous comment, AC5 TM peptides should be tested on A2A/D2R heteromer formation as negative controls as well as the effect of A2A TMs or D2R TMs on A2A/AC5 and D2R/AC5 respectively.
Following the reviewer's comment, as additional negative controls, we also tested AC5 TM1 to TM12 peptides on A2AR-nYFP-D2R-cYFP complementation and all the D2R TM and A2AR TM peptides on AC5-nYFP-A2AR-cYFP and AC5-nYFP-D2R-cYFP complementation, respectively, in the absence of ligands (Supplementary Fig. 5). Sentence included in the Results Section. 5. Authors proposed an "hypothetical" model in which AC5 and heterotetramers (A2A/D2R) associate in a linear manner ( fig. 3g and supp fig. 3) suggesting a certain stochiometry. This is an interesting hypothesis however, this is mostly based on overexpressed proteins. Considering the relative expression of both receptors in AC5 in striatal neurons, authors should comment on how this may also be true in native cells. PLA for AC5/D2R or AC5/A2A in neurons may reveal the pre-assemble complexes in a relevant cell model.
Following the reviewer's advice, PLA experiments were also performed with a recently characterized AC5 antibody (Xie et al., 2015). A 2A R-AC5 and D 2 R-AC5 complexes could be revealed as red punctate staining in neuronal cells (Supplementary Fig. 7). This information and the corresponding additional methods have been added in the text.
6. TM5n impaired association between receptors and AC5 is as good as TM5b, but not TM5 in absence of agonist. Additionally, authors hypothesized that TM5n may be part of intracellular IL2 of AC5 and involved in pre-coupling with receptors. Thus, a scramble IL2 peptide would be a good negative control here.
Following the reviewer's advice, a scrambled TM5-TM5n peptide was synthesized (AC5-TM5s in Supplementary Table 1). As expected, it did not decrease AC5-nYFP-A2AR-cYFP or AC5-nYFP-D2R-cYFP complementation in the absence of ligands. This information is now included in the Results Section.

Minor comments:
1-Line 190. TM4n also behave as a negative control peptide and is not mentioned.
The following sentence has been added in the text: "The same as TM4, TM4n did not have a significant effect, and TM2n, TM3n and TM6n did behave as negative controls to their opposite-oriented peptides, since they did not decrease AC5-nYFP-A 2A R-cYFP or AC5-nYFP-D2R-cYFP complementation..."

2-How the association between GPCRs and AC5 would change if agonists were added?
Since TMs from M1 domain show a marked increased sensitivity to TM2,3,5 and 6 disrupting peptides, some changes in the association between AC5 and receptors may occur. Answers to Reviewer #3 1. Are data available proving on GPCR/AC interaction without available G-protein? Or is a G-protein/AC pre-complex a prerequisite for interaction with the GPCR (or a GPCR/G-protein pre-complex for interaction with AC, respectively)?

We have extended the information previously included in the sentences
Please, see answer to the same point raised by reviewer #1 (point 5).

Minor points:
-The results and suggested scenario of receptor/AC pre-coupling must be strongly dependent (independently from the used method) also on the capacity of the investigated receptors to signal in a ligand-free (constitutive) basal state. It would be helpful to clarify this aspect for A2A and the D2R in the discussion section.
This is an important point we are in fact investigating. We have obtained experimental results that indicate that the previously described constitutive activity of the A2AR (Fernando-Duenas et al., ACS Chem Biol, 2014, 9, 2496 is absent in the A2AR-D2R heteromer. We would obviously want to keep this information for our manuscript in preparation. Nevertheless, we have added the following short sentence in the Discussion: "It can also be predicted that the constrains imposed by oligomerization should alter the constitutive activity of A2AR and D2R". -It is of note, that both homodimers here are characterized by a TMH6-TMH6 interface. A discussion is needed how a TMH6-TMH6 interface in the homodimers may have impact on receptor activation (or how receptor activation can occur in such a constellation), because TMH6 outward movement is a significant feature of GPCR activation from a structural perspective and such a postulated interface may constrain this helix.
heterodimerization interfaces provide a frame for the understanding of allosteric communications through the protomers in GPCR oligomers. For instance, we have suggested that formation of a very stable four-helix bundle via the TMs 5/6 interface, as observed in the crystal structure of the μ-opioid receptor, causes cross-antagonism (Viñals et al., PLoS Biol, 2015, 13, e1002194), because agonists cannot surmount this very stable, high surface complementarity dimer interface, and trigger the outward movement of TMs 5 and 6 for receptor activation. In contrast, in the constructed models of the present work, TM 6 in the inactive closed conformation of the unliganded protomer interacts with TM 6 in the active open conformation of the agonist-bound protomer (Supplementary Figure 4). The reviewer is totally right when noting that simultaneous outward movements of TM 6 in the homodimer is not feasible due to a steric clash between active open conformation of both TM 6. Likewise, simultaneous binding of two G proteins to the homodimer would not be possible due to a steric clash between both bulky G proteins. Thus, in the models of the A 2A R and D 2 R homodimers only one protomer (active open TM 6 conformation) binds both the agonist and the α5 helix of the G-protein.
-Are there any indications that beside AC and here tested GPCRs other secondary effectors interact in pre-complexes with their respective GPCRs?
We are grateful for this comment of the reviewer, which allow us to generalize our findings to other putative pre-coupled complexes of GPCR, G proteins and effectors. We have therefore expanded the end of the Discussion with the following sentence (and additional references): "Therefore, we postulate that pre-coupling should not only apply to other Gs-Gi-AC-coupled heteromers, but also to heteromers coupled to other G proteins and effectors, such as the well-stablished Gi-Gq-coupled metabotropic glutamate receptor mGlu 2 receptor-serotonin 5-HT 2A receptor heteromer, which could be pre-coupled to potassium channels.