Figure Legends:
Fig 1 Bias plot showing the recruitment of β-arrestin-2 (Tango
assay) as the abscissae and activation of a genetically engineered
firefly luciferase cAMP biosensor (GloSensor; Promega, Madison, WI) as
the ordinates. Four agonists for the Κ-opioid receptor were compared:GR89696 ,
4-([3,4-dichlorophenyl]acetyl)-3-(1-pyrrolidinylmethyl)-1-piperazinecarboxylic
acid methyl ester fumarate salt, ICI 199,441 ,
2-(3,4-dichlorophenyl)-N-methyl-N-([1S]-1-phenyl-2-[1-pyrrolidinyl]ethyl)acetamide
hydrochloride; RB 48, 22-thiocyanatosalvinorin A and RB
64 , 22-bromosalvinorin A. Plot shows that the salvinorin analogues RB64
and RB48 are biased toward G protein while ICI199441 and GR89696 are
biased toward β-arrestin. Data from White et al, 2014.
Fig 2 A. Angiotensin receptor II type 1 receptor downstream
signaling pathways and BRET sensors for G protein and β-arrestin
activation, specifically heterotrimeric G proteins and activation of
βarr2. cAMP, adenosine 3′,5′-cyclic monophosphate. B. Heatmap of
relative activity (ΔLog(τ/KA) (Kenakin et al, 2012)
values for angiotensin analogs showing activation of various G proteins
and β-arrestin. Data from Namkung et al (2018)
Fig 3. Theoretical scheme for bias testing showing how
secondary bias assays may further differentiate and better characterize
an initial bias estimate.
Fig 4: Concentration-response (CR) curves for two agonists
(denoted by black and red curves) A. Theoretical model showing receptor
R interacting with an agonist A and two G proteins G1 and G2 (Kenakin,
2003). B. Non biased agonist parameters: Agonist1 (black
curve): α = 10, γ1 = γ2 =2,
KG1= KG2 = 3: Agonist2(red curve): α = 10, γ1 = γ2 =10,
KG1= KG2 = 1. Composition of Cell Types:
CellA = [G1]= 1, [G2]= 10: CellB
= [G1]= 10, [G2] = 1. C. Biased
agonist parameters: Agonist1 (black curve): α = 10,
γ1 =1, γ2 =10, KG1= 3,
KG2 = 1: Agonist2 (red curve): α = 10,
γ1 = 10, γ2 =1, KG1 = 1,
KG2 = 3. Cell type composition as in Panel B.
Fig 5. Bias at different levels within the cell. A. Schematic
diagram showing three general zones where assays may access bias.
Usually it is detected at the level of the receptor. As the
receptor-mediated signals spread through the cell various other
components become involved which can modify bias in different areas.
Finally, whole cell response integrates these signals to yield a
quantity and quality of efficacy. B. Data for seven dopamine agonists
producing six responses in the cell. Radar plot shows the relative bias
of the agonists for each pathway (color coded) through
ΔΔLog(max/EC50) values. Note how the bias for the
agonists differs depending on which part of the signaling pathway is
assessed. Data from Klein Herrenbrink et al, 2016.