Structure of the human κ-opioid/JDTic protein-ligand complex.: 2012

Friday, 18 May 2012

What are opioid receptors?

Opioid receptors are class A (rhodopsin-like) subclass γ GPCRs with endogenous and exogenous opioids as ligands. Opioids are psychoactive chemicals which mediate a variety of physiological responses mainly in the central and peripheral nervous system as well as in the gastrointestinal tract. There are four types of opioid receptors, kappa-, delta-, mu- and Nociceptin receptors (See Figure 1). The activation of the first three upon ligand binding is involved in stimulating analgesia, with δ and µ receptors being involved in the development of physical dependence on and withdrawal from drugs, such as the analgesic butorphanol. The first three receptors can be divided into subtypes, there are three subtypes of κ receptors (κ1κ2κ3), two of delta (δ1,δ2) and three of Mu (µ1µ2µ3). The Mu receptor subclasses have very distinct physiological effects and functions. Only one subtype has been identified in Nociceptin receptors called ORL1. Although these four classes of opioid receptors share some common physiological effects on stimulation, they have characteristic functions which differ them from each other, for example, stimulation of the delta opioid receptor is associated with antidepressive effects, kappa-receptors with dysphoria, Mu receptors with euphoria and Nociceptin receptors with anxiety and depression. All of these receptors are found in the brain and all of them, except for delta opioid receptors, are found in the spinal chord, and all of them except for Nociceptin receptors are found in the peripheral nervous system. Only Mu opioid receptors are found in the intestinal tract. All of these receptors act through GABAergic (inhibitory) neurotransmission.

Figure 1. Human Opioid Receptor families^[1].

What are their physiological and therapeutical relevance?

The previously mentioned functions of opioid receptors, as well as a number of other associated effects, reflect their importance in processes like pain, respiratory drive and mood and explain various therapeutic exploitations of opioids for sedation, antidepressants and most commonly, their use as analgesics (ex: Morphine). Opioids can also be used to treat diarrhoea, coughs and anxiety. They do have a number of side effects including respiratory depression, constipation, euphoria, sedation and withdrawal syndromes potentially leading to addiction. The differential, and sometimes opposing, effects caused by the activation of different opioid receptors relies on their differential recognition by their opioid ligands. The main (broad) types of opioid ligands are endogenous opioids, opium alkaloids, semi-synthetic opioids and fully synthetic opioids. Endogenous opioids show a defined preference for their particular subtype of opioid receptors and their structure is key to the differential stimulation of structurally very similar opioid receptors. However, exogenous opioids and synthetic opioids are known to interact much more promiscuously with opioid receptors, probably due to the similarity amoung opioid receptor binding pockets.
The elucidation of the opium receptor-opioid ligand complex has provided important clues on the molecular interactions that drive and stabilize differential receptor recognition and receptor activation events. The data from these studies has provided valuable information for the development of drug agonists or antagonists, which bring about a desired change in phenotype as a result of differential receptor stimulation. However, although medicinal chemistry has lead to the discovery of many reasonably selective opioid receptor ligands to date, there is still a substantial interest in the development of even more subtype-selective K-OR agonists and antagonists that can be exploited in the design of new pharmacological products.

Κ-Opioid Receptors (k-ORs)

The k-Opioid receptor, encoded by the OPRK1 gene, was first identified based on studies with the k-type agonist ketocyclazocine but binds primarily to the endogenous opioid peptide dynorphin, although it does bind other alkaloid ligands. There is evidence to suggest this opioid receptor is involved in addiction and dynorphin is an endogenous agonist to the body’s natural addiction control mechanism and so there is a lot of interest in developing therapeutic applications that target this receptor. Of particular interest is the fact that it has been shown that stressed/abused children display abnormal kappa and Mu opioid receptor distributions, with studies proving k-ORs involvement in human stress responses and their ability to counteract the effects of Mu-opioid receptors. k-OR ligands also have a characteristic diuretic activity due to their negative regulation of antidiuretic hormone (ADH), as well as k-Opioid agonisms being seen to be neuroprotective against hypoxia/ischaemia. Additionally, k-ORs observed involvment in the delevopment of dysphoria suggests that k-OR mediated therapies, with k-OR agonists and antagonists as viable drug candidates, could also provide a potential treatment for depression.

Although there are three subtypes of k-OR receptors only one cDNA has been found, thereby suggesting these different receptor subtypes arise through interactions with different membrane proteins. All k-ORs are then coupled to a cytoplasmic G_i/G₀ protein, which upon receptor activation increases cAMP phosphodiesterase activity resulting in a decrease in physiological cAMP levels and the shutting down of cAMP-dependent sodium channels, thus inhibiting the propagation of any neuronal action potential signal. k-ORs are also coupled to potassium channels and calcium channels, to increase membrane hyperpolarisation and inhibit neurotransmitter release respectively. Like other GPCRs, the activation of the k-OR has also been associated with the activation of MAPKs, strongly suggesting involvement in, or at least crosstalk with, other essential pathways in cell functioning (See Figure 2).

JDTic is an opioid antagonist which specifically blocks the k-OR agonist U50,488-induced antinociception, while not antagonizing mu-OR agonist-induced analgesia. Its effects are long lasting and animal studies with rodents have shown it can have an antidepressive and anti-anxiolytic effects. It is a 4-phenylpiperine derivative which contains piperidine and isoquinoline groups (See Figure3) ,which are important determinants of its high affinity interaction with the kappa opioid receptor that is of particular interest for the development of highly specific theraputic k-OR ligand drugs.

Figure 2. k-OR mediated signalling. This receptor can trigger signalling cascades through cAMP modulation and activation of the PKC pathway and Src-mediated phosphorylation of JNK protein. kOR-mediated signalling can also result in gene regulation through pCREB and zif298 transcription factors. Note the negative feedback loop that operates on the potassium channels and the receptor agonists and antagonists^[2].

Figure 3. Structure of the kappa opioid receptor antagonist ligand, JDTic^[3]. Acidic groups have been depicted in red and basic groups in blue.

Κ-Opioid Receptor General Structure

Κ-Opioid Receptor Structure

The crystal structure of kOR-JDTic ligand complex at a resolution of 2.9 Å revealed this receptor is structurally very similar to receptors of the same γ subclass GPCRs and even other members of the broader class A GPCR family (See Figure 4).

All four opioid receptor subtypes share a GPCR-type structure with 7 alpha helices, which constitute the transmembrane domains, and a G_i/G₀ protein associated with an eighth intracellular α-helix that runs parallel to the membrane. They also contain three extracellular loops (ECLs) and three intracellular loops (ICLs). ECL2 is the largest of these receptor loops and contains a disulphide bond which is thought to stabilize a β-hairpin structure which covers a ligand binding pocket. Opioid receptors share approximately 70% sequence identity in the transmembrane regions and the C-terminus has a high degree of homology. Most of the sequence variation between opioid receptors is found in the ECLs, which are likely to play an important role in selective ligand binding. The crystal structure of the protein revealed k-OR to contain a disulphide bond between Cys131 and Cys 210 bridging ECL2 to the end of helix III. This last feature is present in all opioid receptors and is thought to be a common component of many class A GPCRs. Another canonical, class A GPCR, feature identified in the structure of the k-OR is a ‘NPXXY’ amino acid residue motif in helix VII where the ‘X’ residue varies between different class A GPCR proteins. In the case of k-OR this motif encompasses Asn326, Pro327, Ile328, Leu 329 and Tyr 330. This sequence is thought to be involved in class A GPCR activation by acting as a molecular switch. This particular region of the receptor was observed to have a relatively high structural similarity with β₂-adrenergic receptors and adenosine receptors (A₂AR).

However, a number of structural characteristics were identified in the crystallographic study of k-ORs, which are believed to be unique and are likely to also be determinants of the selectivity of the ligand-receptor interactions. One of these features is the disulphide bond interaction between ECL3 and the N-terminus of the protein which causes helix I to tilt towards the transmembrane bundle. Another example is how regions of the ECL3 seem to be disordered in some residue stretches, resulting in no interpretable electron density in these regions. The crystal structure in conjunction with mutagenesis studies also revealed an important inter-helical hydrogen bond between Arg156 and Thr273 (the latter in helix VI) which is important for the stabilization of the inactive conformation of k-ORs. The degree of conservation of the latter interaction throughout the class A GPCR family is still being assessed.

In a nutshell, the crystallographic study of k-OR revealed that the structure of this receptor was highly similar to that of members of the same GPCR subclass and even family, and so were the molecular interactions that held the structure together. This is not a surprise since GPCRs are a good example of how a range of functions evolved from a particular protein topology. The similarity of k-OR structure with that of proteins from the same family was high even though many of these regions contained a significant difference in amino acid sequence (mutagenesis experiments confirmed this where amino acid substitution for that present in other proteins such as the β₂-adrenergic receptor didn’t cause a significant change in structure or affinity). However the crystal structure did confirm that most of sequence diversity was present in the ECLs, where the ligand binding pocket was present. This suggested that the ability of the opioid ligands to distinguish between very similar receptors and selectively bind to them was likely to lie in the chemical properties of the receptor binding pocket and the chemical structure of the ligand itself, rather than the general structure of the k-OR, which is more of a stabilizing framework for this interaction to occur and for further signal transduction to be possible through interaction with intracellular G-proteins.

Figure 4. Structure of the opioid receptor classes. A) Structure of two dimerized kappa-opioid receptors with associated G-proteins. This is the assymetric subunit used to elucidate the structure of the receptor through X-ray crystallography. B) Structure of the Mu opioid receptor bound to a morphinian antagonist (red). C) Structure of a delta opioid receptor bound to naltrindole (red). D) Structure of Nociceptin opioid receptor in association with a peptide mimetic. Note that although the receptors do have a similar overall structure, especially in the transmembrane bundle, the ECL regions on the ligand binding sites can be seen to differ significantly.

The Ligand Binding Pocket

The ligand binding pocket of k-ORs was seen to have a mixture of unique and more common structural properties. The k-OR ligand binding pocket was seen to have an Asp138 deep in the cavity, which formed an ionic interaction with the bound JDTic ligand. This residue is present in most aminergic GPCRs and opioid receptors and provides a contact for stabilizing interactions with a protonated amine group on the ligand. Along with the ECL2 β-hairpin structure covering the binding pocket, this Asp138 supposes the main shared feature (with chemokines and aminergic receptor families) in the binding pocket of the k-OR. The binding pocket was narrower and deeper (see Figures 5 and 6) than that of closely related proteins such as CXCR4 and was also seen to differ significantly in the amino acid residues lining the cavity. As compared to CXCR4, the shape of the pocket was significantly different as a result of ‘an approximately 4.5A˚ inward shift of the extracellular tip of helix VI in the k-OR’.

Figure 5. Three-Dimensional structure of the k-OR binding pocket.

Figure 6: Structure of the ligand binding pocket from above.

JDTic interactions in the binding pocket illustrate selectivity

The JDTic antagonist, bound to the ligand binding pocket in the crystallographic structure of the kOR-JDTic complex, displayed high affinity interaction with the pocket residues (Ki=0.32nM), this property may also explain the potency and long duration of its binding as determined by structure-activity relationship assays (SARs). These studies also showed this ligand displays a 1000-fold selectivity for human k-OR as compared to other human opioid receptors. The interactions that comprise this selective JDTic-kOR interaction in the binding pocket have been characterized and comprise forms of polar, ionic and hydrophobic interactions with the receptor.

The interactions that stabilize JDTic binding to the k-OR pocket occur through 22 residues lining the pocket wall, which provide the contacts for these interactions to occur, all of them within 4.5Å of the ligand. These include salt bridges formed between the conserved Asp 138 residue at the bottom of the pocket and the protonated amines in both piperidine and isoquinoline moieties in the ligand. These two amino groups will anchor the ligand to the bottom of the cleft, making it acquire a V-shaped structure. The crystal structure of the JDTic complex also revealed that the ligand is partly stabilized though water-mediated polar interactions involving the distal hydroxyl groups in the piperidine and isoquinoline moieties (this data is confirmed by other relevant SAR studies). It was also noted that the majority of the residues lining the ligand binding pocket create a largely hydrophobic environment which is likely to play an important role in the stabilization of the aromatic rings in the two JDTic moieties. One of the most important of these hydrophobic interactions between the ligand and the binding pocket involved a conserved Trp287 and the ligand’s isopropyl group (it is thought this conserved amino acid residue plays an important role in blocking the activating conformational changes in class A GPCRs). Of all the numerous stabilizing interactions that are involved in the binding of the JDTic ligand to the k-OR pocket, arguably the most significant ones are those occurring in four residues which differ even between closely related opioid receptors. These residues are Val 108, Val 118, Ile 294 and Tyr 312 and are likely to play an important role in the selectivity of JDTic ligand binding of k-ORs. The valine and isoleucine residues provide hydrophic stabilization of the ligand whilst the tyrosine residue provides a polar interaction with JDTic. Analysis of JDTic ligand structure alone and bound to k-OR pocket revealed a degree of flexibility in the conformations it can acquire thanks to the available bond rotations allowed by the planar groups. This flexibility allows the ligand to explore a number of different conformations that facilitates its passage through the narrow and deep k-OR cleft and also allows the formation of a V-structure to correctly align the molecule in the pocket and stabilize its interactions. See some of the stabilizing interactions in Figure 7.

Figure 7. Some important molecular interactions between JDTic ligand and the binding pocket lining residues. Note the polar contacts with Asp138 and the hydrogen bonds involving the terminal hydroxyl groups (seem to disappear but are actually water-mediated). Note the Val 118 and Tyr312 residues which form characteristic interactions with the JDTic ligand and Trp287 which provides hydrophobic stabilisation of the ligand. Also labelled are Ile294 which is involved in forming important hydrophobic interactions with nor-BNI and GNTI morphine analogues and Tyr139 which will also form a hydrogen bond with both of these morphine analogues.

Binding of k-OR selective morphine opioids

Morphine is one of the most physiologically relevant opioid molecules known and its use as a treatment of acute and chronic pain is well established. It also serves as a precursor for other opioid molecules in the pharmaceutical industry. Various binding models and mutagenesis studies suggested mechanisms of action of this potent alkaloid but the elucidation of the crystal structure of k-OR provided ultimate evidence on its structure-function relationship. The same authors that published the crystal structure of the JDTic/k-OR complex also characterized, in the same publication, the binding mechanism of morphine analogues nor-BNI and GNTI (Figure 8), with the aim to clarify the interaction events which bring about the physiological changes caused by morphine and laying out the ground work for the development of synthetic analogues, which exploit these characterized interactions to improve binding selectivity, reduce side-effects or provide a tighter control over opioid receptor activity.

The crystal structure of nor-BNI/k-OR and GNTI/k-OR showed a number of shared and unique interactions in the ligand binding pocket of the k-OR. They both formed the canonical salt bridge with Asp138 and a hydrogen bond with Tyr139 (both of these interactions are conserved in morphine-binding sites). Both morphine-analogues contain a basic moiety, which form a salt-bridge with Glu297 (located on the entrance of the ligand binding pocket). Both of these molecules also displayed the same hydrophobic interactions at the kOR-specific Ile294 residue as well as a highly complementary Van der Waals interface through their naltrexone moiety. A number of different molecular interactions were also seen between the different morphine analogues and the residues in the pocket. For example, nor-BNI displayed polar interactions with Glu209 and Ser211 in the ECL2 whilst GNTI didn’t.

An interesting common interaction motif between JDTic and these morphine analogues was discovered where the cyclopropyl moiety of both nor-BNI and GNTI posses the same position as the isopropyl moiety of JDTic, making hydrophobic contact with conserved residue Trp287. These studies strongly suggested that although morphine analogues shared some common interaction features with those occurring in the elucidated JDTic/k-OR complex, like hydrophobic contacts with His291 in the aromatic cluster or polar contacts with Asp138, the interactions of these analogues with the k-OR pockets differed to an extent that suggested a high degree of selectivity was conserved even between significantly different ligands and thus illustrated the facilitated recognition multivalency of the k-OR through the residues lining the ligand-binding pocket. The most representative example is the importance of the Glu297 residue for the anchoring of nor-BNI and GNTI and its low significance in maintaining the 1000-fold selectivity of JDTic binding to kOR. Another example is the coinciding behaviour of the morphine analogues with the Schwyzer's message-address model for peptide recognition elements and the apparent inapplicability of this concept to JDTic interaction with k-OR. Most of the characterized interactions between the morphine analogues and the receptors coincided with or were confirmed by mutagenesis experiments.

Figure 8. Structures of morphine analogues. A) nor-BNI. B) GNTI. Note how such significant structures (also compare with that of JDTic) share binding interactions with the ligand binding pocket of the kappa opioid receptor and are yet highly selective in their binding^[4][5].

Biological Implications

What is the relevance of these findings?

These results provided strong clues about the likely mechanism of binding of another physiologically relevant molecule, the k-OR agonist and hallucinogenic Salvinorin A. Mutagenesis and SAR experiments suggested this molecule brings about its effects through 2-acetoxy moiety interaction with Cys315 as well as other electrophilic interactions through its thiocyanate group. The crystallization of ligand/receptor complex and testing involving SalA analogues such as 22-thiocyanatesalvinorinA will provide important evidence regarding the interactions, which drives this high affinity and selective binding event. The crystal structure of k-OR has revealed a number of shared and unique structural features which provide multiple anchoring points that allow the interaction of these opioid receptors with a wide variety of structurally different ligands, which interact with the receptor in a highly selective manner but through different mechanisms. Studies on the energetic conformations and flexibility of the ligands in the binding pocket and the different potential interactions with the pocket residues provide an important groundwork for the structure-based development of new drug agonists and antagonists, with new and more personalized therapeutic potential, which target opioid receptors.

Methods

Most of the narrated information in this blog entry comes from the paper Wu et al. (2012) where X-ray crystallography was used to elucidate the structure of JDTic/kOR, nor-BNI/kOR and GNTI/kOR, ligand/receptor complexes. The space group used was P2₁2₁2₁. The asymmetric unit consists of two receptors forming a parallel dimer, with an interface of 1,100 Å, involving helices I, II and VII. The R-value and R-free values obtained were 0.228 and 0.265 respectively. Unit cell descruptions are displayed in Figure 9. The orientation of both copies in the asymmetric unit differs in orientation by 60º and in fact, the slightly differing structural features between the two receptors in the asymmetric unit, for example like differences in the degree of tilt in the ECL2, are likely to be important clues as to the plasticity of these regions of the receptor. The authors also point out: ‘structural studies were carried out using an engineered human k-OR construct and crystallized in cholesterol-doped monoolein lipid cubic mesophase’. They also ensured that the pharmacological behavior of the isolated k-OR protein was similar to that of the native receptor expressed on HEK293T cells. For further detail on the methods used please refer to the primary literature.

Figure 9. Data on angles and lengths of the assymetric subunit from which the X-ray crystallographic data was interpreted from.

Unit Cell:
	Length [Å]	Angles [°]
	a = 54.90	α = 90.00
	b = 147.30	β = 90.00
	c = 205.29	γ = 90.00

References

Bibliography:

[1] http://en.wikipedia.org/wiki/Opioid_receptor

[2] Bruchas, M. R. & Chavkin, C. Kinase cascades and ligand-directed signaling at the kappa opioid receptor. Psychopharmacology 210, 137-147, doi:10.1007/s00213-010-1806-y (2010).

[3] http://en.wikipedia.org/wiki/JDTic

[4] http://en.wikipedia.org/wiki/File:Nor-BNI.png

[5] http://en.wikipedia.org/wiki/File:5'-GNTI.png

All the data in this blog, unless cited otherwise, was based on the following publication: Wu, H., D. Wacker, et al. (2012). "Structure of the human [kgr]-opioid receptor in complex with JDTic." Nature 485(7398): 327-332.