Introduction: 

Psychotic disorders, such as schizophrenia, have long been understood to be caused by imbalances in dopamine levels (1). Dopamine, a neurotransmitter, binds to the dopamine family of receptors which are comprised of  “D1-like” and a “D2-like” receptor subfamilies. D2-like receptors (D2, D3, and D4) are GPCRs bound to Gi/o proteins that inhibit cyclic AMP synthesis by preventing adenylate cyclase formation (1). Recently, the structures of the D3 and D4 receptors were discovered in complex with the antipsychotic drugs Nemonapride and Eticlopride, therebyeticlopride. This finding providing more information as how to develop future drugs that enhance specificity for the binding site (2,3). Even more recently, Wang et al. (2018) discovered the structure of DR2 in complex with the antipsychotic risperidone, finding that the drug displayed a unique mode of binding by forming a subpocket not found in other dopamine receptors in complex with their drugs (4). Identification of the subpocket highlighted key residues in the receptor necessary for increasing the residue time of antipsychotic drugs. While these structural discoveries have provided great insight into increasing the specificity of drugs, researchers have yet to provide a structure for the D1-like receptor family. However, increased knowledge of how to create a thermostable mutant through various methods, such as alanine scanning, now provide a means for creating a thermostable DRD1 variant. Knowledge of a thermostable mutant can increase the likelihood of creating a crystal to understand DRD1 drug selectivity, thereby allowing researchers a way to decrease the chance of a patient to experience extrapyramidal effects, such as kinesthesia. As a result, using established protocols, I will:

1. Identify a thermostable D1 receptor mutant for crystallization
2. Observe which residues in the dopamine receptors determine selectivity for antipsychotics
3. Determine, if it is possible, how to create a D1- Gi/o protein complex

Specific Aim 1: Identify a thermostable D1 receptor mutant for crystallization

Background and hypothesis: Previous studies have demonstrated that thermostable mutants for dopamine receptors may be developed by targeting the intracellular loop 3 of the protein (2–4). Other studies have used the technique known as phage display proteolysis in order to select for thermostable mutants of the four helix bundle protein apoctychrome b562 (3). Using this approach, Wang et al. (2017) created a D4 construct by replacing residues 228–336 of ICL3 with thermostable apocytochrome b562RIL (BRIL) to create their thermostabilized mutant.

In contrast, Chien et al. (2010) used a point mutation in the transmembrane domain (Leu1193.41) to create the thermostable mutant (2). Additionally, they incorporated a T4-lysozyme shown to increase thermostability into intracellular loop 3 (ICL 3). Wang et al. (2018) used similar methods for their DRD2 construct; however, they also developed a new approach to developing thermostable mutants since alanine scanning led to a DRD2 variant with thermostable mutations (I1223.40A, L3756.37A and L3796.41A), indicating a probable method for determining a thermostable construct for a D1 receptor that had  not previously been found (4). Thermostability was tested through a temperature gradient; results revealed that the thermostable mutant bound a higher percentage of [3H]-N-methylspiperone, a radioactive ligand, at higher temperatures than the wildtype D2 receptor. Furthermore, binding assays demonstrated that the DRD2 construct had similar mean Ki for multiple antipsychotic drugs as the wildtype.

Currently, there is no proposed structure for the D1 receptor. Given this, alanine scanning will be used in conjunction with thermostability and binding assays to determine a thermostable D1 receptor. 

Figure 1: Schematic proposal for identifying a thermostable construct of DRD1, as well as identifying key structures in DRD1 necessary for drug interaction and selectivity.

Experimental Approach: Determination of a thermostable D1 construct

a) Use alanine scanning to create point mutations and introduce T-4 lysozyme

Studies have highlighted that the intracellular loop 3 of dopamine receptors is important for the development of a thermostable mutant (2,4). For D2-like dopamine receptors, the function of the intracellular loop 3 is to determine Gi coupling specificity (5). Studies have found that within ICL 3, residues 228-336 are important targets for identifying thermostable constructs in the adrenergic receptors (3). The T4 phage lysozyme has also been found to improve the thermostability of GPCRs, with its replacement of the residues increasing the thermostability of both DRD2 and DRD3 variants (3). As a result, alanine scanning will be used to create random point mutations along each of these residues and incorporate the T4-lysozyme to develop the gene for the D1 thermostabilized construct. Heterologous expression will be used to express the chosen variant in a selected species so that further functional assays may be conducted. The consequence of mutating ICL 3 should also be examined, however. Molecular dynamics and docking studies revealed that a serine to cysteine mutation at position 311 alone was enough to affect cAMP signaling (6). While this does seem to be of concern, it is important to note that thermostable mutants for dopamine receptors have only been discovered by mutating ICL 3,since it is a membrane bound protein. Alterations to other regions of the protein do not confer as much thermostability as those targeting ICL 3. Thus, recent understanding of thermostable mutations has provided insight into different methods of creating a thermostable construct. Lastly, if the use of alanine scanning fails to provide a thermostable mutant, phage display proteolysis maycan be used instead to create the thermostable mutant, though this method has been found to be more time-consuming and less successful than the former (6).

b) Determine stability of each construct with thermostability assays

Each of the D1 receptor mutants will be tested for thermostability with the use of [3H]-N-methylspiperone as a bound ligand along an increasing temperature gradient and compare it with bound ligand to the wildtype receptor, which will act as a negative control. In this case, the DRD2 thermostable variant found by Wang et al. (2018) will serve as the positive control, while the experimental groups will be the mutant DRD1 created through alanine scanning. Previous thermostability assays for a DRD2 receptor showed that the mutated receptor containing alanine mutations and the T4-lysozyme increased the amount of temperature necessary to unbind the ligand from the receptor when compared to the WT (4). As a result, expected findings for the thermostability assays should highlight that a D1 construct containing alanine mutations and a T-4 lysozyme requires more temperature to unbind [3H]-N-methylspiperone than the WT DRD2 and DRD3 negative controls.

To further this aim, I will also conduct ligand binding assays. The assays  will be used to determine the binding affinity of each identified thermostable mutant; the affinities will then be compared to the D2 wildtype receptor’s activity to ensure that the alanine mutations did not inhibit binding. Previous studies on the DRD2 and DRD3 thermostable constructs revealed that ligand binding in new thermostable constructs for each receptor was not perturbed by the thermostable mutations to ICL 3 (2,4). Given this, expected findings would also reveal that the D1 receptor had a similar mean Ki across a multitude of antipsychotics, such as risperidone, N-methylspiperone, and nemonapride. Just as in the thermostability assays, there are no current D1 receptors that could serve as a positive control for ligand binding; however, T-4 stabilized D2 and D3 receptors could be used since the lysozyme is incorporated into ICL 3, ensuring that the conditions are the same for each receptor. In this experiment, WT DRD2 and WT DRD1 will serve as negative controls, while the alanine-mutated DRD1 construct will be the experimental groups, with the last group, thermostable DRD2, serving as the positive control. In case of failure, a possible substitute for ligand binding assays could be DSF-GTP. Using green fluorescent protein (GFP),can help detect a change in the environment of the fluorescent protein through unfolding or denaturing of the targeted protein, in this case DRD1 (7). As a result, ligand binding can be measured to DRD1 tagged with GFP, and this can also reconfirm the thermostability of the protein. 

Summary of Aim 1: Each of these experiments will help identify a thermostable D1 receptor construct that holds the same ligand binding properties as a wildtype receptor but is capable of being crystallized.

Specific Aim 2: Determine residues that affect drug selectivity

Background and hypothesis: The crystal structure of DRD2 revealed helices important for the increased selectivity of risperidone. In the structure, helices III, V, and VI form an extended binding pocket because of the Cys118, Thr119, Ser197, Phe198, Phe382, Phe390, and Trp386 residues (4). In comparison, DRD3’s EBP is formed from Extracellular loops 1 and 2, and helices II, III, and VII, while DRD4’s EBP forms from a junction between TMs II and III (2,3). Importantly, the residue Trp100EL1, which is stabilized by Ile184 and Leu94, is conserved among aminergic receptors and in DRD2 it interacts with the antipsychotic risperidone to enhance residence time (4). Trp100 rearrangement that enables conformational changes of risperidone are absent in DRD3 and DRD4 complexes, indicating that each receptor has different residues important for ligand binding, though D1033.32 and serine residues remain conserved in aminergic receptors. There is no available structure of DRD1 to understand drug selectivity, thus understanding this could provide more information to further drug development in the future that targets this receptor

Figure 2: Key residues found by Wang et al. (2018) shown to be important for risperidone binding for DRD2, with the orange line indicating formation of the subpocket not seen in DRD3 and DRD4.

Currently, methods for determining protein structure in the absence of any crystals requires the use of modeling. Prior to its discovery in 2018, the DRD2 structure was identified through various homology directed models between the known D2 sequence and DRD3 receptor structure (8). The evolutionary history of D1 and D2-like receptors is distinct in that their dopamine-binding properties arose independently (9). While previous models indicated the relationship between D2 and D3 receptors, the recently found structure of D2 by Wang et al. (2018) presents a new opportunity to model and compare D1 to the determined crystal structures of D3 and D2, potentially identifying residues for selectivity that are different from the D2-like receptors. Given this, I hypothesize that while many conserved residues, such as D103 and other residues of TM V are important for the binding of drugs in each receptor, the difference in rearrangement of the ECLs compared with DRD2 will lead to variation in the orientation of the EBP of D2R.

Figure 3: Molecular docking simulations conducted by Platania et al. (2012) revealed distinct binding modes for the drug pramipexole by both DRD2 (left) and DRD3 (right).

Experimental Approach: Using established protocols of molecular docking simulations for aminergic receptors (8), I will use the crystal structure of the human D2 aminergic receptor taken from the Protein Data Bank in complex with risperidone as a template for the D1 receptor, since studies have shown that D1 and D2-like receptors have high sequence homology at the binding site (9).12 I anticipate that the docking simulations of risperidone will reveal no movement of TM VI, since I am comparing D1 to the inactive state of D2. Additionally, it is expected that a deep binding pocket formed by risperidone through interactions with Cys118, Thr119, and Ser197, each within TM III, V, and VI , will be nonexistent since Wang et al. (2018) found this feature to be unique to DRD2. Mutations to F288 and F289 in DRD1 have also revealed that ligand binding is inhibited, and ECL 2 has been implicated in DRD2 and DRD3 selectivity (10). Thus, it would be expected that these residues hold key interactions with the bound ligand to DRD1.

Compared with other methods, molecular docking simulations provides the best result in identifying a potential structure for D1. Other techniques for modeling, such as protein threading, for example, can only compare two proteins if they do not share high sequence homology (11); however, both D1 and D2 share over 50% homology in this region, indicating that protein threading cannot be used with much success. Even with this, molecular docking simulations may still be weak given that D1 and D2 receptors belong to different dopamine families. This suggests that the model produced will be more biased towards the template of the D2 receptor rather than be indicative of a D1-like receptor. If the D1 or D4 receptor crystal structure is identified in the future, then either one could then be used for molecular docking simulations to reveal ligand binding site interactions with greater confidence.

Summary of Aim 2: Molecular docking simulations reveal receptor selectivity

These simulations will identify the residues in the D1 receptor that determine receptor selectivity for antipsychotics.  

Specific Aim 3: Determine if a D1-Gi/o complex can be formed

Background: All GPCRs share a common structure: each contains seven transmembrane domains. Within this large family of GPCRs also lies two subsets of dopamine receptors. D1-like receptors bind to a Gi/0 subunit on the cytoplasmic side of the membrane, inhibiting adenylate cyclase activity. In contrast, D2-like receptors bind to Gs, thereby activating the enzyme. It has recently been proposed that intracellular loop 3 may be important for determining coupling specificity (5). While this proposal has shed new light on possible selection for a G protein, it remains to be seen whether this is the only loop responsible for GPCR-G protein complex formation and selectivity. Understanding GPCR coupling specificity could bring more insight into understanding how the intracellular pathways of cAMP begin. Given this, I hypothesize that the cytoplasmic interactions between a D1 receptor with a Gi/o subunit through ICL 3 is what confers selectivity to the receptor.

Experimental Approach: Using methods established in previous studies for domain swapping in GPCRs (12), I will test whether a new GPCR-G protein complex can be formed through domain swapping of ICL 3 of the D1 receptor with ICL 3 of the D2 receptor. I expect that domain swapping of ICL3 will reveal a GPCR-Gi/o complex for DRD1 and the GPCR-Gs/o complex for DRD2 because ICL3 is what was found to be important for this G-protein coupling specificity.

Since GPCR-G protein complexes are categorized as Gi or Gs through enzyme function, the function of the newly domain-swapped D1 receptor will need to be determined through adenylate cyclase functional assays to ensure that it is indicative of a GPCR-Gi complex. Prior to this, , the domain swap needs to be detected through other methods. Immunofluorescence staining has been shown to be successful in identifying proteins with the use of antibodies (13). As a result, I will use antibodies specific for each complex to identify their presence. The experimental controls for immunofluorescence will be the newly formed complexes while the negative control will need to be a knockout that does not express a functional gene, thereby not creating a protein of interest. It is expected that antibodies will label their respective proteins with immunofluorescence that can be examined under a microscope, but the negative control that uses a knockout mutant not expressing either protein will not be stained. Thus, immunofluorescence can be used to detect a successful swap, but it cannot show activity since that must be measured through other means.

To measure activity for a newly formed GPCR, adenylate cyclase activity assays must be used to detect cAMP levels. As described by Degorce et al. (2008), an HTRF assay can be used to study cAMP levels for DRD1 (14). This technique has been used to successfully determine the activation and inhibition of adenylate cyclase in 5-HT6 and 5-HT4: two serotonin receptors stimulated by agonists and antagonists (15,16). In this method, labeled cAMP will compete with cAMP produced in vivo and the more cAMP produced by the cell, the less FRET signal will occur. In this respect, increased cAMP levels are inversely related to FRET levels. It would be expected that if the DRD1-Gi complex formed correctly, then cAMP levels would be lower, increasing the FRET signal observed. The DRD1-Gs could serve as a positive control since it is known to activate adenylate cyclase, while the DRD2-Gi complex could be used as a negative control because it inhibits the enzyme. The experimental groups will be the DRD1-Gi and DRD2-Gs complexes. 

There are many benefits for the proposed experiment. For one, the cAMP assay allows for direct measurement of cell lysate. Additionally, the assay is easy to setup since it can be miniaturized with minimal difficulty in a culture medium (14). And domain swapping of ICL 3 reveals whether it is a key structure important for GPCR coupling specificity. One weakness of this study, however, is that it is possible that domain swapping of ICL3 alone will not be sufficient for transferring D2-Gs properties to D1 and the functional assays will not be able to detect other domains necessary for coupling specificity. Furthermore, it is necessary to know the coupling mechanism, making this a more difficult option to use for orphan GPCRs, or proteins that do not have as much information available regarding their coupling specificity (15). If the HTRF does not remain a viable option, a GTPγS binding assay may be used instead. This assay can measure the guanine nucleotide exchange, with Gi/o complexes exchanging the nucleotide at a higher rate than Gs complexes. As a result, the coupling specificity of a GPCR can be determined by examining the rate of exchange. 

Conclusion:

Ultimately, the importance of identifying the crystal structure of DRD1 cannot be understated. Currently, extrapyramidal effects, or effects that affect motor control such as kinesthesia, are still a common problem with antipsychotics even with advancement in technology (16). Certain steps must be taken before this problem can be solved. Firstly, determining a thermostable mutant of DRD1 through alanine scanning would increase the likelihood of finding the crystal structure that has been absent in the D1-like family of dopamine receptors. Second, understanding what residues in the protein play a central role in interacting with drugs through molecular docking simulations will enhance our understanding of the selectivity of the receptor, thereby helping the development of future drugs that can better target the receptor, at least until a crystal structure can be used for these studies. Lastly, identifying the role of ICL 3 in G protein coupling specificity would allow for greater insight into the necessary components for intracellular signaling of the receptor, potentially allowing for the development of drugs that target ICL 3 in the cytoplasmic side of the membrane rather than the extracellular region. Each of these findings have a high impact in creating new therapies for patients that suffer from extrapyramidal effects.