Flashcards in GPCRs Deck (32):
3 EC loops for ligand binding, 3 IC loops for G proteins, allosteric modulators, other second messengers
Agonist binding alters relative positions of TM3 and TM6
No single active conformation, there are many, which may give rise to biased signalling
Many exist in dimers, which requires cholesterol in a lipid raft
4 types (s AC, i AC, q PLCbeta, 12 p115-RhoGEF
8 human genes
5 bladed propellor arrangement
7 human genes
12 human genes
Modifications to alpha/beta/gamma subunits
alpha and gamma are altered for insertion into the cell membrane - myristoylation on the amino group of the terminal glycine, palmitoylation
Betagamma come as a pair, don't really swap partners. Functional significance of different permutations unknown.
Was previously thought that alpha subunits share betas from a pool, but fusing them together did not prevent translocation or activation, and FRET based studies have shown they don't move apart far on activation
GPCR interaction with G protein
We don't know whether a G protein stays associated with a receptor even when receptor is inactive, but G proteins definitely bind preferentially to certain conformations (link to ligand bias)
But we have more conformations that bias away from Arrestin binding that towards.
Rosenbaum et al 2011 - the same conformation was produced in B2AR from agonist and inverse agonist application!
Single-molecule imaging has shown actin fibres criss-crossing the membrane, that create 'hot-spots' for receptor-G protein interactions. They'll transiently bind when they bump into each other, but only form a lasting interaction when a ligand is bound.
Inhibitors of Galphai
RGS proteins - have 120 aa RGS box, GTPase Activating Proteins
GPR proteins - 20-30 aa GPR motif, Guanine Nucleotide Dissociation Inhibitors
Guanine Exchange Factors
They will only bind inactive G proteins (i.e. won't bind Gi-protein-GTP), which ensures directionality
Girdin aka GIV was the first GEF to have its GEF sequence found
Ric-8a, Ric-8b, Get3 and various others are also GEFs but we haven't found a conserved motif for it.
Ric-8a is essential for normal RTK function, but is also a chaperone protein for betagamma, so how much is its role is due to GEF function vs chaperone function?
Girdin - binding detail
Originally identified in yeast studies
Mainly binds alpha i, binds alpha s a bit, doesn't bind alpha q at all
Has a motif to bind to the G protein, and another to bind to the receptor, so can cause G protein signalling from all sorts of receptors
--SH2 domain, part of receptor binding domain, shows structural plasticity, existing in a disordered state until recruited by an activated RTK
Short aliphatic helix of GEF motif binds between switchII and alpha3 helix of alpha subunit, which explains why it can't bind active G protein (because switch II has moved to occlude the site)
This binding site also may mean it displaces betagamma. How much of its effects are actually due to betagamma signalling downstream?
-but FRET studies have suggested betagamma doesn't need to dissociate to be active, so maybe this is a moot point -
G proteins with non GPCRs
Both beta2 and alpha7 nAChRs can physically associate with alpha i, q, 12, and betagamma.
nAChR can coimmunoprecipitate with Gprin1 and Gai.
Gprin1 controls axon growth. nAChR and Gprin1 form a complex that activates GAP-43, which binds cytoskeleton and promotes growth cone growth
When ACh binds, the growth cone breaks down
Applying betagamma increases nAChR channel open probability 5 fold (perhaps every subunit in the pLGIC can bind betagamma?)
Girdin can bind all sorts of different receptors, and mediate their signalling via G proteins, but we only know its mechanism of binding to RTK
Yevenes et al 2003 - betagamma can bind GlyR via the M3-M4 loop of its alpha subunit. Binding is phosphorylation-independent, but affected by pertussis toxin. Binding increases affinity for Glycine, and may affect amplitude of current
Girdin - in disease
GEF motif has been specifically implicated in increasing metastasis in invasive cancer, and can be used as a predictor of survival for colorectal or breast cancer (Liu et al 2012).
Girdin downregulates antifibrotic pathways (by activating Gi, to reduce cAMP, PKA, CREB) and upregulates profibrotic pathways (e.g. PI3K, FoxO1), thus increasing liver fibrosis. GIV expression is undetectable in hepatocytes, but high in hepatic stellate cells.
In nephrotic syndrome, Girdin is protective, because it reduces apoptosis of podocytes, via the prosurvival pathway PI3K, in response to VEGF.
3 basic properties of G protein signalling
Gain (there's always amplification)
Convergence (many transmitters and receptors will activate the same subset of G protein),
Divergence (one transmitter can activate various receptors, one receptor can work via different G proteins..?)
GPCRs and the genome
3% of genome is GPCRs, most of them olfactory
Olfactory receptors, and others, are part of the rhodopsin family, characterised by a lysine in TM7 for attachment of chromophore.
Other families include Frizzled (24), Adhesion (24), GABA (15), and Secretin (15).
Classification is based on agonist specificity and subtle structural differences
Peptides must bind to loops, monoamines can go down into TM domain, proteases can cleave to activate
GPCRs in healthcare
50-70% of drugs target GPCRs, though much of the phylogenetic tree is uninterrogated - we have not yet found many drugs that will bind
Many diseases are caused by mutations in GPCRs, e.g. single amino acid substitution in V2 vasopressin mutation causes diabetes insipidus
Hard to find specific drugs that bind ligand pocket of GPCRs because of homology between them. Allosteric sites are less conserved, and allosteric modulators modulate physiological responses rather than binary activate or inactivate, so may have fewer side effects. E.g. Cmpd-15 is the first allosteric 'beta-blocker'.
Allosteric modulators that bind at the intracellular side of two chemokine receptors have been found, may be of use in treating chronic inflammatory disorders.
5-HT7 and 5-HT1A are both important in depression, and homodimeris of 5-HT1A are resistant to 5-HT-mediated internalisation, so maybe pattern of dimerisation is important in pathogenesis.
Homologous and heterologous desensitisation
Can cause G protein independent signalling
Heterologous desensitisation - PKA phosphorylates activated receptor, changes preference of B2AR from Gs to Gi, an there's a negative feedback loop AC, cAMP, PKA. Homologous desensitisation - Then GRK can phosphorylate, and recruit arrestin (a new signalling scaffold), then clathrin mediated endocytosis
Signalling can happen from the intracellular compartment, via G proteins or other pathways, e.g. via c-Src and the MAPK pathway
Different GPCRs go through different internalisation pathways. All pass through early endosome, but e.g. GPER (binds steroids) ends up in the perinuclear compartment - transcriptional effects??
Gate-keepers inc PRABs determine whether the receptor gets recycled from endosome or broken down
There's also caveolin-dependent and clathrin and caveolin independent. There's recycling from the endosome, sequence dependent or in bulk.
GPCRs signalling without G proteins
AGTR1 binds Angiotensin II and then can bind JAK-STAT for transcriptional effects
Somatostatin receptor constitutively binds PI3K, but dissociates when somatostatin binds, so PI3K can stimulate the Akt pathway
Arrestin can act as a scaffold for many other second messengers, including ERK and c-Src
Detecting signalling from different compartments
Nb80-GFP can bind only active b2ARs
After 3.5 minutes, most fluorescence is at the cell membrane
After 20 mins, active receptors are found in the endosome. But is this proof they're signalling from there?
Nb37-GFP binds active G proteins
It colocalises with Nb80, suggesting the B2ARs in the endosome are actively signalling
Inhibitin dynein (and hence endocytosis), or expressing a receptor that doesn't endocytose at all, gives you only the rapid and not delayed phase of luminescence from these biosensors.
Endosomal signalling is later and sometimes of longer duration than surface membrane signalling
What have structural studies told us?
-Activation hypotheses, e.g. clam-shell hypothesis was developed by Cheung et al because of the use of hydrogen-deuterium mass spec (which allowed us to see receptors in motion) in conjunction with EM and X ray crystallography
-How the structure's held together, i.e. 24 contacts between TM domains that hold them together like a scaffold around TM3 (maybe because of its extreme tilt angle)
-Ligand binding, e.g. comparative studies found a cradle held together by residues from all TM helices except TM1, which means any TM1 residues that alter binding must be acting indirectly. Also closed vs open pocket conformations helps drug design, e.g. S1P has a side entry because its ligand is hydrophobic
-Oligomerisation of B2ARs when left to their own devices, shown using disulfide trapping. Cholesterol required for dimerisation of 5-HTRs, and DA receptors need 4 cholesterol molecules between their monomers.
-Conformational changes - Rosenbaum et al 2011 found an agonist and an inverse agonist that caused the same B2AR conformation. Manglik et al 2013 - using an agonist, they isolated two active, one intermediate, and one inactive conformation. This predicts ligand bias.
-More efficient drug discovery - allows production of databases that can be searched for drugs that will bind.
-Realtime conformational changes - using solid state NMR in fully hydrated liquid crystal phase
Difficulties in structural studies
Getting enough protein (often the receptors aren't strongly expressed, and so samples are unstable and impure)
Conformations is dependent on environment, and it's hard to crystallise a protein in lipid
Isolated receptors are flexible
Solutions to difficulties in structural studies
Express the protein in bacterial cells, or insect cells using a baculovirus vector to get more of it, or use xFEL (x-Ray Free Electron Laser) or serial femtosecond crystallography so you damage less and need a smaller sample
Make chimaera to stabilise the proteins, e.g. ICL3 for b2AR (though Venkatakrishnan et al 2013 said this fusion affected mobility of some segments, so may prevent investigation into conformational changes), or use antibodies or fragments of, or thermostabilise.
Developed lipid cubic crystallography, or lipid bicelles
Solid state NMR in fully hydrated liquid crystal phase means physiological temperature and pH, no detergents, can alter composition of lipid, small molecules can be added, no mutations or truncations or insertions of foreign proteins necessary.
Cryo-EM results in less damage than X-ray crystallography (and was awarded a Nobel prize)
Class C GPCRs
Venus flytrap region, cysteine rich region, then 7TM
ligands bind in VFR, transduced through CRR
Obligate dimers, via VFR and CRR, but also maybe ICD
e.g. GABA receptors are heterodimers of B1 and B2 subunits. B1 binds ligand, B2 binds G protein. B2 is needed to compete with PRAF2 (gatekeeper) for B1, to allow translocation to membrane.
Comps-Agrar et al - GABA receptors exist in an equilibrium between heterodimers and oligomers, with a preference for tetramers and octamers, though the latter are transient couplings while dimers are formed from strong non-covalent bonds. Different oligomeric state gives preference for different G proteins.
e.g. mGluR, 6 cholesterols between monomers, monomers are affected by allosteric modulators (bind TM region, prevent signal transduction), dimers are affected by glutamate.
e.g. Knockout T1R1, attenuate response to umami. KO T1R2, attenuate response to sweet. KO T1R3, attenuate response to either. So umami = T1R1+T1R3, sweet = T1R2+T1R3
Signalling mechanisms learnt from structural studies
-37 class A GPCR structures available, only 5 have both active and inactive structures identified
-common set of rearrangements near the G protein binding region
-G protein binding region evolved separately from ligand binding pocket
-ionic lock identified in rhodopsin between Arg3.50 and Glu6.30, must be broken for activation. But not found in all inactive structures, and known to be plastic, so may contribute to basal activity
-Contact found in all inactive states, and contact found in all active states
-Existence of intermediate conformations suggested concept of induced fit
-active conformations are many and varied, from tissue to tissue and even cell to cell, and dependent upon binding of agonist and intracellular species inc G protein.
Drug discovery in GPCRs is made harder by homology in the ligand binding pocket - hard to get specificity. Allosteric sites may allow greater specificity in drugs.
Allosteric modulators that bind at the intracellular side of two chemokine receptors have been found, may be of use in treating chronic inflammatory disorders.
GPCRs are 'allosteric machines', controlled in part by ions inc sodium, water, and membrane components like cholesterol.
In Adenosine 2A receptor:
Continuous Internal Water Pathway gives flexibility to change shape
Binding sodium to allosteric site causes efficacy switch, from MAPK signalling to G protein increase of cAMP
How do we detect dimers?
Structural studies - cystallography shows structural interfaces between certain TM segments. But normal caveats - samples are impure, maybe stabilising foreign proteins were added in.
Coimmunoprecipitation - tag one subunit with HA and another with YFP. Immunoprecipitate with anti-HA, then look for YFP. If the two were dimerised, you won't find any. Found 5-HT7 and 5-HT1A dimerised in mouse hipppocampus.
FRET - Tag one dimer with donor chromophore CFP and another with acceptor YFP. When coexpressed, the emission spectrum will shift to the right if there was donation. After bleaching the acceptor, the donor will fluoresce brighter and the acceptor will have a dark patch where they were close enough for energy transfer. This found 5-HT7 is close to 5-HT1A
Functional studies - Heterodimerisation of 5-HT7---5-HT1A receptors inhibits GIRK in Xenopus oocytes. Effect increased with more 5-HT7 cRNA. 5-HT7 siRNA, to decrease expression, increased GIRK current. 5-HT7 antagonist did not block the effect, so it's due to presence of receptors, not their activity
Proximal ligation studies - raise an antibody against each, in separate species. Add secondaries with oligonucleotides attached. Add hybridise connector to make a circle, which a probe can recognise and amplify. Tells us dimers exist, but not about their functionality.
TAT peptide - peptide that can cross membranes, and prevent dimerisation if fused to sequence thought to be at interface. GLUN1 and D1R subunits of NMDAR were thought to dimerise in cocaine addiction. TAT peptide with GLUN1 decreased FRET signal, costimulation increased it. Preventing dimerisation reduced LTP, and prevented D1R-mediated enhancement of NMDAR currents, and hence reduced of ERK
NMDARs ARE NOT GPCRS!!!!
Evidence for class A dimers
Controversial as to whether dimers are obligatory
Structural studies have shown us the interface for various receptors' dimerisation (but remember these are just snapshots)
Heterodimerisation of 5-HT7---5-HT1A receptors inhibits GIRK in Xenopus oocytes
Effect increased with more 5-HT7 cRNA
5-HT7 siRNA, to decrease expression, increased GIRK current
5-HT7 antagonist did not block the effect, so it's due to presence of receptors, not their activity
Co-IP and FRET have shown that at least 5-HT7 and 5-HT1A subunits exist v close together
Other effects of disrupting dimerisation
Using TAT peptide acts as competitive inhibition of dimer formation inside cell (remember it often happens in ER, before insertion into membrane):
Disrupting SecretinB and AT1R dimerisation in rat brain reduced hyperosmolarity-induced drinking
Disrupting D1R and D2R dimerisation had an antidepressant effect in rats, as seen in forced swim test.
Disrupting mu and delta opioid receptor dimerisation caused opioid tolerance and increased thermal analgesia
Evidence for ligand bias
Different pharmacological profile and kinetics depending on second messenger pathway assayed
Agonist specific receptor conformation ensembles -
Different agonists generate different ensembles of conformations, some of which show constitutive activity, and which have unique second messenger preferences, though it's unlikely we can detect these functionally.
Agonist specific G protein coupling preference - Different GPCRs have their own 'fingerprints' of preference for different G proteins. These are agonist-specific too.
Allosteric coupling between ligand and G protein - When the G protein is bound, the alpha subunit tail remains buried in the receptor, stabilising conformational change at the ligand binding pocket, that impedes entry and exit of ligand from the binding site, potential mechanism for increased agonist affinity when G protein is bound.
Conformation-specific pharmacological profile - Stabilising 'high affinity' and 'low affinity' states gave different pharmacological profiles for different agonists
Ligand bias vs functional selectivity
functional selectivity’ is a more general term for different effects in any kind of assay under any different condition.
‘ligand bias’ is specifically due to receptor conformation.
Properties of a well-designed in vitro assay for ligand bias:
Control for receptor specificity of ligand
Either screen proximal signals - high throughput, precise, but may miss functions of the receptor - or look at downstream, integrative signals - thorough, but hard to prove target specificity
Matching assay conditions between different second messenger systems being assayed
Matching assay duration to prevent complex degradation or altered kinetics, or time-dependent switching e.g. signalling from endosome
Examples of biased ligands - clinical
Carvedilol is particularly useful in heart failure. Binds B2AR. It was found that it inhibits Gs activity but stimulates beta-Arrestin activity, and ERK1/2 activity, and enhances receptor internalisation.
TRV027 - ATR2, biased towards beta-Arrestin, increases cardiac contractility and reduces apoptosis without causing vasoconstriction from G protein effects. beta-Arrestin measured by chemiluminescent beta-galactosidase activity, G protein by accumulation of IP1. Shown in rats, then dogs, now humans.
TRV130 - mu opioid receptor, biased towards G-proteins, to avoid nausea, constipation, sedation, respiratory depression, addictive effects from beta-Arrestin
cAMP biased pepducins - membrane-penetrating palmitoylated/myristoylated peptides, modelled on an intracellular loop, can mimic interactions with G proteins, or with other receptors, or interact with other regions of the same receptor. Used to treat asthma, causing bronchodilation without beta-arrestin-mediated desensitisation
Examples of biased ligands - native
PACAP receptor (pituitary adenylyl cyclase activating receptor) binds several endogenous ligands, which each have different concentration-effect curves for cAMP production and IP production
Drosophila Tyr/Oct receptors expressed in Chinese Hamster Ovary cells show that tyrosine is biased towards Gi-like response, and octopamine is biased towards Gq-like response. Tyr and Oct differ only by one OH group.