Biased ligands – aptamers against distinct conformations of G-protein-coupled receptors
Aptamers are the nucleic acid equivalent of antibodies.1 They can be easily selected from pools of random-sequence single-stranded oligonucleotides to bind small molecules, proteins or cellular targets with high specificity and affinity. The SELEX (systematic evolution of ligands by exponential enrichment) process2,3 generally involves around 10 rounds of positive selection against the specific target, but may also include negative selection against rival or non-specific targets. Aptamers generally contain 2´-amino or 2´-fluoro pyrimidines for increased endonuclease resistance alone or in combination with locked nucleic acid or 2´-O-methyl modifications. Spiegelmers are nuclease-resistant aptamers consisting of L-(deoxy)ribose-based nucleotides due to the fact that nucleases are stereo-selective for D-(deoxy)ribose.
The SELEX process can be automated and is generally performed in vitro, so aptamers against new targets can be made within days instead of the several weeks needed for antibodies and large quantities can be produced synthetically. Aptamers have thus found wide application including targeted delivery (for example of siRNA, miRNA and saRNA), as therapeutic inhibitors as well as bioimaging probes and biosensors.4
A recent paper by Kahsai et al.5 has added yet another application – as allosteric stabilizers of active, inactive and ligand-specific conformations of G-protein-coupled receptors (GPCRs).
GPCRs are the largest and most ubiquitous group of plasma membrane receptors that can exist and function as dimers or multimers. They are responsible for transmitting signals from physical and chemical extracellular stimuli such as light, odours, nutrients, metabolites, hormones and neurotransmitters to the cell interior, and as such, they are extremely important therapeutic targets. This is illustrated by the fact that around 40% of drugs currently on the market interact with GPCRs.
GPCRs have seven transmembrane domains with α-helical structure and exhibit multiple conformational states both when ligand-bound and ligand-free (inactive). Particular ligands are thought to stabilize specific conformation subsets, propagating the signal via distinct G-proteins and/or β-arrestins or even by inducing different conformations in these effector proteins. This multi-state GPCR activation model can explain how different antagonists and agonists of the same receptor can effect discrete pharmacological profiles.6,7 The hope is that new “biased” ligands could be selected to induce desired pharmacological effects such as increased efficacy while reducing on-target adverse effects.
Kahsai et al. report the selection of specific aptamers against the purified GPCR β2-adrenoceptor (β2AR) in the inactive state (no ligand) and bound to BI167107, a high affinity full agonist with exceedingly slow off-rate (active state). They performed nine rounds of selection on a 2′-fluoropyrimidine-modified RNA library containing 1015 unique sequences with a central random region of 40 nucleotides. Each round consisted of a negative as well as a positive selection. Additionally, round 5 included a negative selection against the inactive angiotensin receptor subtype 1a (AT1aR).
After nine rounds of selection, aptamer pools showed nanomolar binding affinities for both selection targets (Kd ~100 nM for inactive and Kd ~85 nM for the active states of β2AR). The authors used next-generation sequencing to analyse the aptamer pools from multiple rounds in order to increase the sampling power of the clonal space. They tracked the enrichment of particular sequences during the selection process and then selected the 20 aptamers with the highest enrichment ratios for binding assays with the active and inactive β2AR. According to the binding assay screening, aptamers A1, A2 and A13 showed strong conformational selectivity for the active β2AR while A16 was selective for the inactive receptor.
The aptamers A13 and A1 were able to increase the binding affinity of the β2AR full agonist isoproterenol (ISO), establishing their ability to stabilise the particular receptor conformation induced by ISO binding. As expected, A16 was unable to influence ISO affinity. Surprisingly, aptamer A2 also had no effect on ISO binding affinity, even though it is clearly selective for the active BI167107-bound β2AR conformation. From these data, the authors reasoned that the full agonists ISO and BI167107 induce discrete β2AR conformations and that A2 can distinguish between them.
To further investigate if the selected aptamers were specific for distinct receptor conformations, the authors studied the influence of these aptamers on the binding of a panel of full agonists, partial agonists, antagonists and inverse agonists. The panel included the inverse agonist carvedilol and BI167107, a full agonist, which are known to be modestly biased agonists toward β-arrestin-dependent signalling pathways. Aptamer A1 and A13 binding to β2AR in the presence of antagonists was markedly reduced while binding in the presence of agonists was positively correlated with agonist efficacy. In contrast, there was no such correlation with aptamer A2. In addition to highly effective binding to BI167107-occupied β2AR, A2 showed a unique selectivity for carvedilol-bound β2AR suggesting that these two ligands induce closely related receptor conformations, possibly associated with the β-arrestin signalling bias.
To investigate the functional effect of these aptamers, Kahsai and colleagues then measured the accumulation of cAMP as a consequence of ISO-stimulated activation of adenylyl cyclase activity via the G-protein subunit Gαs. The aptamers A1, A2 and A13 were able to inhibit cAMP accumulation significantly (28, 35 and 48%, respectively), while A16 did not.
As the four selected aptamers had distinct sequences and predicted secondary structures, it was probable that they would bind different epitopes of β2AR and this was indeed found to be the case. Competition studies with the G-protein mimetic nanobody Nb80, which binds the intracellular region of activated β2AR, showed that aptamers A1 and particularly A2, bind to the same epitope as Nb80, while A13 clearly does not. Aptamer A16 does not compete with Nb60, a nanobody recognising the intracellular region of the inactive receptor, suggesting it may bind the extracellular region instead. Transmission electron microscopy and single-particle reconstruction analysis confirmed this and also that A1, A2 and A13 bind the intracellular region of activated β2AR.
The paper clearly demonstrates that it is possible to select aptamers which are able to distinguish between unique receptor conformations and stabilise these conformations. Thus, these aptamers can not only be used to study multi-state GPCR activation but also to stabilise particular conformations for screening against small-molecule drug libraries in order to select biased ligands. They may even be useful as biased ligands themselves.
1. Aptamers as therapeutics.
Keefe AD, Pai S, Ellington A.
Nat Rev Drug Discov. 2010 Jul;9(7):537-50.
2. Systematic evolution of ligands by exponential enrichment: RNA ligands to bacteriophage T4 DNA polymerase.
Tuerk C, Gold L.
Science. 1990 Aug 3;249(4968):505-10.
3. In vitro selection of RNA molecules that bind specific ligands.
Ellington AD, Szostak JW.
Nature. 1990 Aug 30;346(6287):818-22.
4. Aptamers and their biological applications.
Song KM, Lee S, Ban C.
Sensors (Basel). 2012;12(1):612-31.
5. Conformationally selective RNA aptamers allosterically modulate the β2-adrenoceptor.
Kahsai AW, Wisler JW, Lee J, Ahn S, Cahill Iii TJ, Dennison SM, Staus DP, Thomsen AR, Anasti KM, Pani B, Wingler LM, Desai H, Bompiani KM, Strachan RT, Qin X, Alam SM, Sullenger BA, Lefkowitz RJ.
Nat Chem Biol. 2016 Sep;12(9):709-16.
6. Recent updates on GPCR biased agonism.
Pupo AS, Duarte DA, Lima V, Teixeira LB, Parreiras-E-Silva LT, Costa-Neto CM.
Pharmacol Res. 2016 Feb 2. pii: S1043-6618(16)00034-7.
7. Biased ligands at G-protein-coupled receptors: promise and progress.
Violin JD, Crombie AL, Soergel DG, Lark MW.
Trends Pharmacol Sci. 2014 Jul;35(7):308-16.