[methods in enzymology] g protein coupled receptors - modeling, activation, interactions and virtual...
TRANSCRIPT
![Page 1: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/1.jpg)
CHAPTER THIRTEEN
Biasing the Parathyroid HormoneReceptor: Relating In Vitro LigandEfficacy to In Vivo BiologicalActivityKathryn M. Appleton*,‡, Mi-Hye Lee*, Christian Alele*,Christine Alele*, Deirdre K. Luttrell†, Yuri K. Peterson‡,Thomas A. Morinelli†,}, Louis M. Luttrell*,},1*Division of Endocrinology, Diabetes & Medical Genetics, Department of Medicine, Medical University ofSouth Carolina, Charleston, South Carolina, USA†Division of Nephrology, Department of Medicine, Medical University of South Carolina, Charleston,South Carolina, USA‡Department of Pharmaceutical & Biomedical Sciences, College of Pharmacy, Medical University of SouthCarolina, Charleston, South Carolina, USA}Research Service, Ralph H. Johnson Veterans Affairs Medical Center, Charleston, South Carolina, USA1Corresponding author: e-mail address: [email protected]
Contents
1.
MetISShttp
Introduction
hods in Enzymology, Volume 522 # 2013 Elsevier Inc.N 0076-6879 All rights reserved.://dx.doi.org/10.1016/B978-0-12-407865-9.00013-3
230
2. Determining the Relative Activity of PTH1R Ligands 2322.1
Defining reference and test ligands 234 2.2 Assaying hPTH1R-mediated cAMP production 236 2.3 Assaying PTH1R-mediated intracellular calcium influx 243 2.4 Assaying PTH1R-mediated ERK1/2 activation 247 2.5 Estimating PTH1R ligand bias 2543.
Discussion 256 Acknowledgments 258 References 259Abstract
Recent advances in our understanding of the pluridimensional nature of GPCR signalinghave provided new insights into how orthosteric ligands regulate receptors, and howthe phenomenon of functional selectivity or ligand “bias” might be exploited in phar-maceutical design. In contrast to the predictions of simple two-state models of GPCRfunction, where ligands affect all aspects of GPCR signaling proportionally, currentmodels assume that receptors exist in multiple “active” conformations that differ in theirability to couple to different downstream effectors, and that structurally distinct ligandscan bias signaling by preferentially stabilizing different active states. The type 1 parathy-roid hormone receptor (PTH1R) offers unique insight into both the opportunities and
229
![Page 2: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/2.jpg)
230 Kathryn M. Appleton et al.
challenges of exploiting ligand bias in pharmaceutical design, not only because numer-ous “biased” PTH analogs have been described but also because many of them havebeen characterized for biological activity in vivo. The PTH1R has pleiotropic signal-ing capacity, coupling to Gs, Gq/11, and Gi/o family heterotrimeric G proteins, and bindingarrestins, which mediate receptor desensitization and arrestin-dependent signaling.Here, we compare the activity of six different PTH1R ligands in a common HEK293 cellbackground using three readouts of receptor activation, cAMP production, intracellularcalcium influx, and ERK1/2 activation, demonstrating the range of signal bias that can beexperimentally observed in a “typical” screening program. When the in vitro activity pro-files of these ligands are compared to their reported effects on bone mass in murinemodels, it is apparent that ligands activating cAMP production produce an anabolicresponse that does not correlate with the ability to also elicit calcium signaling. Para-doxically, one ligand that exhibits inverse agonism for cAMP production andarrestin-dependent ERK1/2 activation in vitro, (D-Trp12, Tyr34)-bPTH(7–34), reportedlyproduces an anabolic bone response in vivo despite an activity profile that is dramat-ically different from that of other active ligands. This underscores a major challenge fac-ing efforts to rationally design “biased” GPCR ligands for therapeutic application. While itis clearly plausible to identify functionally selective ligands, the ability to predict howbias will affect drug response in vivo, is often lacking, especially in complex disorders.
1. INTRODUCTION
Early efforts to model the action of drugs or hormones assumed that
individual receptors behave as binary switches, existing in equilibrium
between an “off” state, which is silent in the assay, and an “on” state, which
is capable of generating a measurable response. In such models, receptor
conformation is the minimal determinant of system response and ligands
act solely by changing the fraction of the receptor population in the on state
(Karlin, 1967; Thron, 1973). The efficacy of a ligand thus becomes a reflec-
tion of its ability to stabilize the on state and can be approximated by two
parameters: the maximal observed response (Emax) and potency (EC50),
the ligand concentration that produces a half-maximal response. In this con-
text, full agonists are ligands that preferentially bind and stabilize the on state,
producing the maximum system response at saturating ligand concentration;
partial agonists are ligands with less conformational selectivity, translating into
a submaximal system response at saturating concentration and potential
attenuation of full agonist activity; true neutral antagonists are ligands with
equal affinity for both the off and on conformations, producing no physio-
logical response but able to block the response to agonists; and inverse agonists
are ligands that preferentially bind the off state, which causes them to appear
![Page 3: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/3.jpg)
231Biasing the Parathyroid Hormone Receptor
as antagonists in systems with low basal activity but with the added property
of reducing receptor-mediated constitutive activity in systems with high
basal tone.
Although readily determined experimentally, measurements of EC50 and
Emax are limited in that they are influenced by system factors external to the
ligand–receptor unit (Ehlert, 2000; Figueroa, Griffin, & Ehlert, 2009). In the
case of G protein-coupled receptors (GPCRs), differences in receptor reserve
and signal amplification can lead to apparent changes in ligand classification
when comparisons are made between different assays. New signaling
responses commonly emerge as the level of receptor expression increases,
permitting less efficiently coupled effectors to reach the detection threshold
of the assay (Zhu, Gilbert, Birnbaumer, & Birnbaumer, 1994). Similarly,
variation in the expression levels of G proteins, arrestins, and downstream
effectors can make ligand activity appear to change between cell types
(Nasman, Kukkonen, Ammoun, & Akerman, 2001). Even in the same cell
background, ligands may appear as full agonists when classified using sig-
nals that are highly amplified, for example, cAMP production, but as par-
tial agonists when assayed for less amplified responses, for example, arrestin
binding and signaling (Rajagopal et al., 2010).
Nonetheless, signal strength arguments cannot account for true reversal of
potency or efficacy, for example, when the rank order of potency for two
ligands acting on the same receptor is opposite in two different assays of cellular
response (Berg et al., 1998). In a two-state model, ligand binding can alter the
fraction of receptors in the on state, but cannot qualitatively change the nature
of that state. Thus, the classification of a ligand as an agonist, antagonist, or
inverse agonist must be independent of the assay used to detect receptor
activation, and the rank order of potency for a series of ligands cannot vary
when two or more assays are employed. Reversal of potency or efficacy
implies that different ligands are activating the same receptor in different ways,
meaning that theymust be generating different active receptor states (Kenakin,
1995). That this phenomenon has now been described for several GPCRs,
among them the serotonin 5-HT2c, pituitary adenylate cyclase-activating
polypeptide, dopamineD2, neurokininNK1,CB1 cannabinoid,b2 adrenergic,angiotensin AT1A, and PTH1Rs, suggests that most, if not all, GPCRs can
adopt multiple active conformations (Luttrell & Kenakin, 2011).
It is nowapparent thatGPCRsignaling is “pluridimensional” (Galandrin&
Bouvier, 2006), meaning that receptors signal by coupling to multiple
G protein and non-G protein effectors. If different active conformations
![Page 4: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/4.jpg)
232 Kathryn M. Appleton et al.
couple the receptor to these downstream effectors with different efficiency,
then the cellular response will be dictated by the distribution of receptors
across the range of achievable active and inactive states. Since there is no
a priori reason that the active conformation(s) favored by one ligand should
be identical either to the spontaneously formedactive state or to those preferred
bya structurallydistinct ligand, thepotential exists for the ligand tobias signaling
in favorof someeffectors at theexpenseofothers.Thus, it is the ligand–receptor
complex, not the receptor alone, that specifies the active state, along with any
other smallmolecule, protein–protein, or lipid–protein interaction that alloste-
rically constrains the conformations available to the receptor (Kenakin &
Miller, 2010). In contrast to a two-state model, wherein agonists and antago-
nists merely control the quantity of receptor activity, the potential of biased
agonism lies in its ability to qualitatively change signaling.
Here, we employ the PTH1R to illustrate some of the issues arising from
pluridimensional efficacy and the challenge of adequately describing ligand
bias. Using a selected panel of PTH analogs and a multiplexed set of cell-
based assays for cAMP, intracellular calcium release, and ERK1/2 activation,
we demonstrate the assay-dependence of efficacy and the range of signaling
responses achievable through biased agonism. These results are then dis-
cussed in the context of the known effects of these same ligands in vivo to
illustrate the difficulty and complexity of using in vitro profiles of ligand
activity to predict biological response.
2. DETERMINING THE RELATIVE ACTIVITY OF PTH1RLIGANDS
For any given GPCR ligand, an activity profile can be generated by
determining its EC50 and Emax across a panel of assays measuring different
indices of receptor activation. These results will be specific for each rec-
eptor–ligand combination, but they will also be subject to system factors
influencing coupling efficiency, such as receptor density, that can cause
ligand classification to appear to vary between assays. Such factors cannot
change the relative order of potency or efficacy for a series of ligands, but
can create the appearance of signal bias. For example, a ligand may appear
to be a full agonist in terms of both potency and efficacy if the signal is tightly
coupled, that is, there is significant “receptor reserve,” or highly amplified,
for example, second messenger production, and as a partial agonist when
assaying a response that is weakly coupled, that is, all receptors must be in
the active state to achieve a maximal response, or unamplified, for example,
stoichiometric GPCR–arrestin binding.
![Page 5: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/5.jpg)
233Biasing the Parathyroid Hormone Receptor
When characterizing a series of ligands across multiple assays of receptor
activation, it is useful to determine intrinsic relative activity (RAi). RAi is a
measure of the microscopic affinity constant of an agonist for the active
state of the receptor expressed relative to that of a reference agonist, which
in its simplest form can be estimated from EC50 and Emax values using the
equation (Griffin, Figueroa, Liller, & Ehlert, 2007):
RAi¼Emax�BEC50�A
Emax�AEC50�B
where Emax is the maximum observed system response, EC50 is the ligand
concentration producing a half-maximal response, and A and B are the ref-
erence and test ligands, respectively. This simple calculation of RAi is valid
in cases where the Hill slope of the agonist concentration–response curve is
close to 1.0 or the Emax values of the agonists are equivalent.
When these conditions are not met, cleaner quantification of ligand bias
can be obtained using the Black/Leff operational model (Black & Leff, 1983),
which relates the equilibrium dissociation constant (KA) of the ligand, a direct
measure of receptor occupancy, to a coupling efficiency factor (t), whichencompasses both the intrinsic efficacy of the agonist and system-dependent
factors such as receptor density and coupling efficiency (Kenakin, 2009). Since
the latter factors are constant for any concentration–response curve deter-
mined for any given signaling pathway in the same cell, the ratio of t valuesfor any two agonists in the same system will yield a ratio of intrinsic efficacy
that is independent of receptor number or coupling efficiency. Because allo-
steric effects exerted by other system components can alter ligand affinity as
well as efficacy, it cannot be assumed that the KA value for a given agonist
will be constant under all conditions, so t/KA ratios must be determined for
each assay system. Once determined for each agonist/pathway of interest,
t/KA ratios can be used to quantify bias relative to a reference agonist.
Despite the additional data required, the advantages of the operational model
are its ability to quantify the full range of agonism from submaximal effects to
effects in very sensitive systems with receptor reserve, and that t/KA ratios
determined in one system are applicable to all systems without the need to
individually quantify functional selectivity in all systems. Variation in recep-
tor density and coupling efficiency between systems might change the ability
of all agonists targeting a given receptor to activate a particular pathway, but
it will not change the pathway selective bias of different ligands relative to
one another. While this more generalizable approach has proven useful in
discriminating “weak” ligand bias that might otherwise be obscured by sys-
tem factors (Figueroa et al., 2009; Rajagopal et al., 2011), it remains unclear
![Page 6: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/6.jpg)
234 Kathryn M. Appleton et al.
whether the added sensitivity of the operational model will prove useful in
industrial applications.
2.1. Defining reference and test ligandsThe PTH1R is a class 2 GPCR highly expressed in kidney and bone that
mediates the effects of PTH, an 84-amino acid peptide hormone that func-
tions as the primary systemic regulator of calcium homeostasis. Most of its
actions are mediated by classic G protein signaling mechanisms, including
Gs-mediated activation of adenylyl cyclase, resulting in cAMP production
and PKA activation, as well as Gq/11- or Gi/o-mediated activation of
phospholipase-Cb, leading to inositol-1,4,5-trisphosphate production, cal-
cium mobilization, and PKC activation (Gesty-Palmer & Luttrell, 2011).
The PTH1R also engages arrestins, which mediate both receptor desensitiza-
tion and arrestin-dependent signaling (Ferrari, Behar, Chorev, Rosenblatt, &
Bisello, 1999; Gesty-Palmer et al., 2006). In transfected HEK293 cells,
PTH1R-mediated ERK1/2 activation results from two temporally distinct
mechanisms: a conventional G protein-dependent pathway that involves
PKAand/orPKCand aGprotein-independent pathwaymediated by arrestins
(Gesty-Palmer et al., 2006).
The PTH1R has long served as a model for the study of functional selec-
tivity inGPCR signaling, as its pleiotropic downstream signaling is sensitive to
changes in ligand structure. Whereas the C-terminal truncated PTH(1–34)
fragment possesses all of the known biochemical and physiologic properties
of the native hormone, acting as a conventional/full agonist with respect to
activation of Gs and Gq/11 signaling and arrestin-dependent receptor desensi-
tization and internalization, other PTH fragments exhibit marked variations in
coupling PTH1R to downstream effectors. For example, shorter N-terminal
fragments of the PTH peptide, for example, PTH(1–31), activate adenylyl
cyclase in ROS 17/2 rat osteosarcoma cell membranes without stimulating
membrane-associated PKC ( Jouishomme et al., 1994), while N-terminal
truncations, for example, PTH(3–34), activate PKC while failing to activate
adenylyl cyclase (Jouishomme et al., 1992). Further N-terminal truncations,
for example, PTH(7–34), which still possess the structural determinants nec-
essary for relatively high affinity binding but lack the N-terminal residues
needed to stimulate guanine nucleotide exchange, antagonize G protein sig-
naling but still stimulate receptor phosphorylation and internalization
(Sneddon et al., 2004). Other signal-selective PTH analogs include Trp1-
PTHrp(1–36), which has been reported to activate ERK1/2 in HEK293 cells
![Page 7: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/7.jpg)
235Biasing the Parathyroid Hormone Receptor
solely through a Gs/PKA-dependent pathway that is unaffected by PKC
inhibition or downregulation of arrestin expression, suggesting that it is
Gs-selective (Gesty-Palmer et al., 2006), and (D-Trp12, Tyr34)-bPTH
(7–34), which acts as an inverse agonist for adenylyl cyclase activation, yet
is capable of activating ERK1/2 via a arrestin-mediated signaling pathway
(Gardella et al., 1996; Gesty-Palmer et al., 2006).
If the goal is to compare the ability of ligands to bias PTH1R signaling,
prior studies are limited by the fact that they were, in most cases, performed
in different cell backgrounds and/or used different assays to measure effector
coupling. To produce directly comparable data, we chose a panel of PTH
analogs based on literature reports (Table 13.1) and compared them in three
assays of PTH1R receptor activation that reflect most of the downstream
Table 13.1 Reported activity profiles of selected PTH1R ligands
Ligand Kd (nM)G-proteincoupling
Arrestincoupling References
PTH(1–34) 2�1 Gs and Gq/11 Arrestin
2/3
Juppner et al. (1991), Abou-Samra
et al. (1992), Bringhurst et al.
(1993), Pines et al. (1994), Takasu,
Guo, and Bringhurst (1999), Gesty-
Palmer et al. (2006)
PTH(1–31) NDa Gs and Gq/11 ND Whitfield and Morley (1995),
Takasu and Bringhurst (1998),
Sneddon et al. (2004)
Trp1PTHrp
(1–36)
ND Gs only Antagonist Gesty-Palmer et al. (2006)
PTH(3–34) 10 Gs and Gq/11
antagonist
ND Segre, Rosenblatt, Reiner,
Mahaffey, and Potts (1979),
Nussbaum, Rosenblatt, and Potts
(1980), Takasu, Guo, et al. (1999)
PTH(7–34) 58�16 Gs and Gq/11
antagonist
ND Hoare and Usdin (2000), Sneddon
et al. (2004)
D-Trp12,
Tyr34-
bPTH(7–
34)
25�2 Gs inverse
agonist
Gq/11
antagonist
Arrestin
2/3
Gardella et al. (1996), Hoare and
Usdin (2000), Gesty-Palmer et al.
(2006), Gesty-Palmer et al. (2009)
ND, not determined.aPTH(1–31) IC50 for PTH(1–34) binding¼78�8 nm (Barbier et al., 2005).
![Page 8: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/8.jpg)
236 Kathryn M. Appleton et al.
PTH1R actions that are reportedly subject to ligand bias: cAMP production,
intracellular calcium mobilization, and ERK1/2 phosphorylation. All assays
were performed using human PTH1R ectopically expressed in a HEK293
cell background that lacks endogenous PTH1R. For each ligand in each
assay, RAi estimates were generated using human PTH(1–34) as the
reference ligand.
2.2. Assaying hPTH1R-mediated cAMP productionGPCR effects on intracellular cAMP concentration are largely a reflection of
the activation of adenylyl cyclases by Gs and inhibition by Gi/o proteins.
PTH1R effects on cAMP were determined using the Promega GloSensor
cAMP assay, which permits real-time measurement of cAMP concentra-
tion in live cells (Binkowski, Fan, & Wood, 2011). The GloSensor cAMP
reporter is composed of a genetically modified form of Photinus pyralis lucif-
erase fused to a cAMP-binding protein insert. cAMP binding alters the con-
formation of the reporter to increase luciferase activity. To provide adequate
throughput, the GloSensor assay was adapted for use on the FLIPRTETRA
fluorescence imaging plate reader system that enables simultaneous real-time
recording of luciferase activity in 96-well plate format.
2.2.1 Cell culture and transient expression hPTH1R constructsTomaximize assay consistency, cAMPmeasurements were performed using
GloSensor cAMP HEK293 cells supplied by Promega, Inc. that stably
express the reporter. Assays were performed following transient transfection
of hPTH1R expression plasmids. Two hPTH1R constructs were used: wild-
type hPTH1R was used to assess ligand activity in a system with low basal
receptor-catalyzed Gs activity, while the constitutively active H233R mutant
hPTH1R (Gardella et al., 1996) was used to discriminate inverse agonism
from neutral antagonism.
2.2.1.1 Required materialsCell culture and transfection
• Cells: GloSensor cAMP HEK293 cell line (Promega, Inc.)
• Cell growthmedium:Dulbecco’smodifiedEaglesMedium(DMEM) sup-
plemented with 10% fetal bovine serum (FBS), 1% antibiotic–antimycotic
solution, and 50 mg/mL hygromycin B
• cDNA expression plasmids: Wild-type and H223R hPTH1R cDNAs
cloned into the pCMV6-XL6 expression vector (Origene Technologies,
Inc.)
![Page 9: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/9.jpg)
237Biasing the Parathyroid Hormone Receptor
• Transfection reagents and medium: Lipofectamine 2000 Transfection
Reagent (Invitrogen—Life Technologies Corp.); OptiMEM
(Gibco—Life Technologies Corp.); serum-free Minimum Essential
Medium (MEM)
• Low serum medium: MEM supplemented with 1% FBS and 50 mg/mL
hygromycin B
Additional solutions
• Cellgro 0.05% trypsin solution (Mediatech, Inc.)
• Collagen solution: 800 mL of rat tail collagen (Becton, Dickson and
Company) in 49.2 mL of sterile 2% acetic acid in distilled H2O, sterilized
using a 0.2-mm filter sterilization unit
• 1� phosphate-buffered saline (PBS)
Disposables
• 10-cm tissue culture dishes
• 1.5-mL sterile Eppendorf microcentrifuge tubes
• 10-mL sterile culture tubes
• Costar white-walled clear-bottom 96-well plates
2.2.1.2 Culture and transient transfection of GloSensor cAMP HEK293 cellsCell culture. The GloSensor cAMP HEK293 cell line was maintained for up
to 20 passages on 10-cm culture dishes in DMEM growth medium con-
taining 50 mg/mL hygromycin B for selection. Cells were maintained at
37 �C in a 5% CO2 atmosphere and passed by trypsinization every
3–4 days to maintain subconfluence.
Transient transfection of GloSensor cAMPHEK293 cells. On day 1, GloSensor
cAMPHEK293 cellswere split into 10-cm tissue culture dishes at a density suf-
ficient to achieve 50% confluence by day 2. GloSensor cAMP HEK293 cells
were transiently transfected with wild-type or H223R hPTH1R expression
plasmids on day 2. Using sterile Eppendorf tubes, 10 mg of PTH1R or
H223R PTH1R plasmid DNA was added to a final volume of 500 mL of
OptiMEMin tube“A”and2.5 mLof lipofectaminewas added toa final volume
of 500 mL ofOptiMEMin tube “B.”TubesA andBwere vortexed briefly and
left to incubate at room temperature for 5 min. Following incubation, tube B
was added to tube A, inverted twice to mix, and left to incubate at room
temperature for 20 min. The 1 mL mixture of DNA, Lipofectamine, and
OptiMEM was added to a 10-mL conical tube containing 1 mL prewarmed
OptiMEM and 3 mL of serum-free MEM. Following medium aspiration
and 1� PBS wash, the 5 mL mixture was added to each 10-cm plate of
GloSensorcAMPHEK293cells and left to incubate at 37 �Cfor4 h.Following
![Page 10: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/10.jpg)
238 Kathryn M. Appleton et al.
incubation, the transfection medium was aspirated and replaced with growth
medium, and the plates were returned to the tissue culture incubator.
Preparation of assay plates. Prior to cell plating, sterile Costar white-walled
clear-bottom96-well plateswerecollagencoatedbyadding sufficientcollagen/
acetic acidmixture to cover thebottomof eachwell.Theplateswere incubated
at room temperature for 1 h, after which the collagen solution was aspirated,
and each well was washed twice with 100 mL of 1� PBS. On day 3, the trans-
fected cells in 10-cm plates were trypsinized and seeded into collagen-coated
96-well white-walled clear-bottom plates in growth medium at a density of
5�104 cells/well.On day 4, the growthmediumwas aspirated fromeachwell
and replaced with 1% FBS MEM including 50 mg/mL hygromycin B and
returned to the tissue culture incubator overnight.
2.2.2 GloSensor cAMP assay using the FLIPRTETRA
To measure cAMP in the FLIPRTETRA using GloSensor cAMP HEK293 cells
transfectedwithwild-type orH223Rmutant PTH1R, 96-well plates of trans-
fected cells were pre-equilibrated in GloSensor cAMP reagent. Drug plates
containing serial dilutions of each test ligand at 5� final concentration were
prepared for dispensing by the FLIPRTETRA. Changes in luminescence were
recorded in real-time following the injection of ligand and normalized signal
intensity was used to generate ligand concentration–response relationships.
2.2.2.1 Required materialsInstrumentation
• FLIPRTETRA fluorescence imaging plate reader system (Molecular Dynam-
ics, Inc.). In luminescencemode, excitation andemission filters are disabled
and the emission filter removed to allow direct luminescence detection.
Reagents and materials
• PTH1R test and reference ligands: hPTH(1–34) (Bachem, Inc.); hPTH
(1–31) (Bachem, Inc.); bPTH(3–34) (Bachem, Inc.); hPTH(7–34)
(Bachem, Inc.); Trp1-hPTHrp(1–36) (American Peptide Co.); D-
Trp12, Tyr34-bPTH(7–34) (Bachem, Inc.), dissolved in sterile distilled
H2O at 0.1 mM and 1 mM stock concentrations and aliquoted into single
use volumes to avoid freeze/thaw of the peptides.
• Clear round bottom 96-well plates
• Single and multichannel pipettors
Additional solutions
• Promega GloSensor cAMP reagent reconstituted according to the man-
ufacturer’s protocol and stored in 200 ml aliquots at �80 �C until use
![Page 11: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/11.jpg)
239Biasing the Parathyroid Hormone Receptor
• Serum-free MEM supplemented with 10 mM HEPES, pH 7.4
• 10 mM forskolin stock solution in dimethylsulfoxide
• 5 mM 3-isobutyl-1-methylxanthine solution (IBMX) in ethanol
Disposables
• Micropipettor tips and solution dispensing trays
• 10-mL sterile culture tubes
• 1.5-mL Eppendorf microcentrifuge tubes
2.2.2.2 Performing the GloSensor cAMP assayLoading the GloSensor cAMP reagent. cAMP assays were performed on day 5
(72 h after transfection). Fresh cAMP reagent medium was prepared by
adding a freshly thawed aliquot of 200 mL of GloSensor cAMP reagent to
10 mL of serum-free MEM buffered with 10 mM HEPES, pH 7.4. The
growth medium was gently aspirated from assay plates and replaced with
100 mL/well of prewarmed cAMP reagent medium using a multichannel
pipettor. Plates were incubated at 37 �C with 5% CO2 for 1 h and then
removed from the incubator and incubated at room temperature in the dark
for an additional 30 min.
Preparation of the ligand dosing plate. Drug plates for the FLIPRTETRA dis-
pensing system were prepared during the cell preincubation period. Fresh
serial dilutions of each PTH1R ligand at 5� working concentration were
prepared in serum-free MEM buffered with 10 mMHEPES.Working con-
centrations spanned the expected active concentration range of each ligand
(10�11 to 10�5 M depending on the ligand). A round-bottom 96-well drug
plate was designed to contain 40 mL of 5� drug concentration per well.
Each concentration was tested in replicates of three on the plate. 10 mMforskolin, which directly stimulates adenyl cyclase, was included on each
drug plate as a positive control.
GloSensor cAMP assay. Excitation and emission filters were removed
prior to initializing the FLIPRTETRA. All assays were run at room tempera-
ture. The instrument was programmed to dispense a total of 25 mL of vehiclecontrol, ligand, or forskolin from each well of the drug plate into the 100 mLof medium in the corresponding well of the assay plate to reach the working
concentration. For measuring luminescence, detection gain was set to
280,000 with exposure time 0.53 s and the gate open 100%. Assay and ligand
dosing plates were loaded into the instrument and luminescence was
recorded every 1 s for 10 reads to establish baseline luminescence, then every
1 s for 50 reads. Thereafter, luminescence was recorded every 2 s for 600
reads (660 total reads over 21 min). In the FLIPRTETRA, the maximum
![Page 12: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/12.jpg)
240 Kathryn M. Appleton et al.
change in luminescence was reached approximately 5 min after the addition
of ligand and remained stable for at least 10 min. Raw data representing the
time–luminescence relationship for each well following ligand addition
were exported to Microsoft Excel for background subtraction and analysis.
The relative expression levels of wild-type and H223R mutant PTH1R
were verified to be similar by immunoblotting. The protocols for wild-type
and H223R hPTH1R were identical, except that for H223R hPTH1R
experiments, IBMX, a nonselective phoshosdiesterase inhibitor, was used
to improve assay sensitivity for detecting inverse agonism. 1 mL of 5 mM
IBMXwas added to each well for the final 10 min of room temperature pre-
incubation in cAMP reagent medium prior to loading the assay plates into
the instrument.
2.2.3 Estimating RAi for PTH1R-mediated cAMP productionFor each ligand, normalized concentration–response curves were generated
from the raw cAMP luminescence data. GraphPad Prism software was used
to calculate Emax, EC50, and Hill slope. RAi was estimated from measured
Emax and EC50 values using PTH(1–34) as the reference ligand.
2.2.3.1 Required materialsComputer software
• Microsoft Excel
• GraphPad Prism
2.2.3.2 Data processing and resultsGenerating ligand concentration–response curves. Measurements were taken from
each well after stable maximum luminescence was attained. UsingMicrosoft
Excel, background luminescence measured in vehicle-treated wells was sub-
tracted from maximal luminescence in ligand-treated wells to yield the net
change in luminescence. The mean net change in luminescence from trip-
licate wells at each ligand concentration was determined, and all values were
normalized to the peak luminescence observed with PTH(1–34). Using
GraphPad Prism, each normalized concentration–response dataset was fit
to a sigmoidal dose–response curve using a variable Hill slope. Emax and
EC50 were determined from these curves. A minimum of three separate
experimental replicates were performed using each ligand.
As shown in Fig. 13.1, experiments performed using wild-type hPTH1R
clearly separated ligands into two groups. hPTH(1–34), hPTH(1–31), and
Trp1-hPTHrp(1–36), all appeared as full agonists, producing a similar
![Page 13: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/13.jpg)
Ligand EC50 Emax RAiPTH (1–34) 1.85 � 10-8
2.79 � 10-8
2.92 � 10-8
1.0
PTH (1–34)
PTH (1–31)
(D-Trp12, Tyr34
)-PTH (7–34)
log, [Ligand] M
-100.0
0.2
0.4
0.6
0.8
1.0
1.2
-8 -6 -4
% M
ax P
TH
(1–
34)
resp
onse
PTH (3–34)
PTH (7–34)
(Trp1)-PTHrp (1–36)
1.0
PTH (1–31) 1.074 0.71
(Trp1)-PTHrp (1–36) 0.94 0.64
PTH (3–34) N.D. N.S. 0
PTH (7–34) N.D. N.S. 0
(D-Trp12,Tyr34)-PTH (7–34) N.D. N.S. 0
A
B
Figure 13.1 Ligand-mediated cAMP production by wild-type PTH1R. (A) Ligandconcentration–response relationships were determined for a panel of six PTH1R peptideligands using GloSensor HEK293 cells transiently overexpressing hPTH1R as described inthe text. The mean net change in cAMP reporter luminescence was normalized to themaximal net response elicited by the reference ligand, PTH(1–34). Sigmoidal dose–response curves and SEM across at least three independent experiments were gener-ated using GraphPad Prism. (B) For each ligand, RAi was estimated from the observedEmax and EC50 values. Ligands that failed to generate a statistically significant change incAMP luminescence were assigned a RAi of zero. NS, no significant response; ND, notdetermined.
241Biasing the Parathyroid Hormone Receptor
maximal response. In contrast, bPTH(3–34), hPTH(7–34), and D-Trp12,
Tyr34-bPTH(7–34) produced no significant change from basal. Figure 13.2
depicts results obtained using the H223R hPTH1R. Normalized basal
cAMP luminescence in cells expressing the H223R hPTH1R was about
0.28 of the maximal PTH(1–34)-induced luminescence observed with
the wild-type hPTH1R, reflecting the constitutive activity of the mutant.
When ligand-induced changes in cAMP luminescence obtained in cells
expressing H223R hPTH1R are normalized to this higher basal, the
concentration–response curves clearly distinguish D-Trp12, Tyr34-bPTH
(7–34) as an inverse agonist for cAMP production, while bPTH(3–34)
![Page 14: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/14.jpg)
Ligand EC50 Emax RAiPTH (1–34) 8.40 � 10–9
1.26 � 10-8
1.79 � 10-8
3.36 ´ 10-9
1.0
PTH (1–34)
PTH (1–31)
PTH (3–34)
PTH (7–34)
(D-Trp12,Tyr34)-PTH (7–34)
log, [Ligand] M
% N
orm
aliz
ed m
ax P
TH
(1–
34)
resp
onse
-10-0.6
-0.4
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
-8 -6 -4
(Trp1)-PTHrp (1–36)
1.0
PTH (1–31) 1.07 0.69
(Trp1)-PTHrp (1–36) 0.85 0.57
PTH (3–34) N.D. N.S. 0
PTH (7–34) N.D. N.S. 0
(D-Trp12,Tyr34)-PTH (7–34) -0.44 -1.12
A
B
Figure 13.2 Ligand-mediated cAMP production by the constitutively active H223Rmutant PTH1R. (A) Ligand concentration–response relationships were determined forthe PTH1R ligand panel using the same protocol described in Fig. 13.1, except thatGloSensor HEK293 cells were transfected with the H223R-mutant PTH1R, which gener-ates constitutive receptor-mediated cAMP production, permitting detection of inverseagonism. Ligand effects are depicted as positive or negative change from basal cAMPluminescence in the constitutively active system. (B) RAi estimates reflect the directionof change in cAMP luminescence relative to the reference agonist. The negative RAi cal-culated for (D-Trp12, Tyr34)-bPTH (7–34) reflects its inverse agonism of receptor–Gs cou-pling. NS, no significant response; ND, not determined.
242 Kathryn M. Appleton et al.
and hPTH(7–34) remain neutral. These results are consistent with literature
reports of the effects of each ligand on PTH1R-Gs coupling (Gardella et al.,
1996; Gesty-Palmer et al., 2006; Jouishomme et al., 1992, 1994; Sneddon
et al., 2004).
Estimating RAi for cAMP production. Because the empirically determined
Hill slopes for PTH1R ligand effects on cAMP production were close to 1.0,
RAi values were estimated using the simplified formula given in Section 2.
Emax/EC50 ratios for each ligand were normalized to that of the reference
ligand, PTH(1–34) (Figs. 13.1 and 13.2). For cAMP production determined
using the wild-type PTH1R, the activity rank order was hPTH(1–34)
![Page 15: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/15.jpg)
243Biasing the Parathyroid Hormone Receptor
¼hPTH(1–31)¼Trp1-hPTHrp(1–36)>>>bPTH(3–34)¼hPTH(7–34)
¼D-Trp12, Tyr34-bPTH(7–34). Using the H223R PTH1R, the observed
activity rank order was hPTH(1–34)>hPTH(1–31)¼Trp1-PTHrp
(1–36)�bPTH(3–34)¼hPTH(7–34)�D-Trp12, Tyr34-bPTH(7–34).
2.3. Assaying PTH1R-mediated intracellular calcium influxGPCR effects on intracellular calcium arise largely from activation of
phospholipase-Cb isoforms by Gq/11 and Gbg subunits primarily derived
from Gi/o proteins. The PTH1R regulates phospholipase-Cb through cou-
pling to both Gq/11 and Gi/o in varying proportions depending on cell back-
ground. In osteoblastic cells, PTH-stimulated phospholipase-Cb activation is
primarily Gq/11-dependent, while renal tubular cells exhibit Gi/o-dependent
phospholipase-Cb activation due to expression of the Naþ/Hþ exchanger
regulatory factors 1 and 2, which enhance PTH1R-Gi/o coupling (Mahon,
Donowitz, Yun, & Segre, 2002). Because measurement of intracellular
calcium using calcium-sensitive fluorescent dyes is readily amenable to
high-throughput screening, we chose to assay ligand-induced calcium flux
as a marker for PTH1R regulation of the phospholipase-Cb–inositol-1,4,5-trisphosphate–intracellular calcium–PKC pathway. While this assay predom-
inantly reflects phospholipase-Cb activation in short-term stimulations, it does
not discriminate between the contributions of Gq/11 and Gi/o proteins and
thus is unable to distinguish possible mechanistic ligand bias between these
two effectors.
2.3.1 Cell culture and inducible expression of hPTH1RTo maximize consistency between experiments, calcium assays were per-
formed using a stable HEK293-FlpIn TRex cell line engineered for
tetracycline-inducible expression of wild-type hPTH1R. Assays were
performed following the induction of hPTH1R expression.
2.3.1.1 Required materials• Cells: Tetracycline-inducible hPTH1RHEK293-FlpIn TRex cells were
established by cloning cDNA encoding the hPTH1R into the HEK293-
FlpIn TRex cell line using previously described methods (Zimmerman,
Simaan, Lee, Luttrell, & Laporte, 2009)
• Cell growth medium: Phenol red-free DMEM supplemented with 10%
FBS, 1% antibiotic–antimycotic solution, and 50 mg/mL hygromycin B
plus 5 mg/mL blasticidin for selection
![Page 16: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/16.jpg)
244 Kathryn M. Appleton et al.
• Serum-free medium: Phenol red-free DMEM supplemented with 0.05%
bovine serum albumin
Additional solutions
• Cellgro 0.05% trypsin solution (Mediatech, Inc.)
• 1 mg/mL tetracycline stock solution in sterile distilled H2O
Disposables
• 10-cm tissue culture dishes
• Costar 96-well black-wall clear-bottom plates
2.3.1.2 Cell culture and induction of PTH1R expressionCell culture. hPTH1R HEK293-FlpIn TRex cells were maintained for up to
20 passages on 10-cm culture dishes in phenol red-free DMEM growth
medium containing 50 mg/mL hygromycin B plus 5 mg/mL blasticidin
for selection. The cells were maintained at 37 �C in a 5% CO2 atmosphere
and passed by trypsinization every 3–4 days to maintain subconfluence.
Tetracycline induction of hPTH1R expression. On day 1, hPTH1R HEK293-
FlpIn TRex cells were seeded into dishes at a density sufficient to attain 50%
confluence by day 2. At the time of passage, 1 mg/mL tetracycline was added
to the growth medium to induce PTH1R expression. The cells remained in
tetracycline-supplemented medium until assay 72 h later.
Preparation of assay plates. On day 2, the cells were seeded at a density of
5�104 cells/well into black-wall clear-bottom 96-well plates that were pre-
coated with collagen as described in Section 2.2.1.2 and cultured in growth
medium containing tetracycline, hygromycin B, and blasticidin. On day 4,
the growth medium was aspirated and the cells were serum-starved for 3 h
prior to assay in 100 mL of phenol red-free DMEM supplemented with
0.05% BSA and tetracycline.
2.3.2 FLIPRTETRA calcium assayMeasurements of ligand-induced calcium flux were performed on the
FLIPRTETRA fluorescence imaging plate reader system using the FLIPR
Calcium Assay Kit 5 from Molecular Devices, Inc. Tetracycline-induced
hPTH1R HEK293-FlpIn TRex cells in 96-well plates were preincubated
with a proprietary calcium-sensitive fluorescent dye while drug plates
containing serial dilutions of the test ligands were prepared. Changes in
dye fluorescence were recorded in real-time following the injection
of ligand and normalized signal intensity was used to generate agonist
concentration–response relationships.
![Page 17: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/17.jpg)
245Biasing the Parathyroid Hormone Receptor
2.3.2.1 Required materialsInstrumentation
• FLIPRTETRA fluorescence imaging plate reader system (Molecular
Dynamics, Inc.) with 470–495 nm excitation light-emitting diodes
and 515–575 nm emission filter
Reagents and materials
• PTH1R test and reference ligands
• FLIPR Calcium 5 Assay Kit (Molecular Devices, Inc.)
• Clear round-bottom 96-well plates
• Single and multichannel pipettors
Additional solutions
• 10 mM A23187 stock dissolved in dimethylsulfoxide
• 1� PBS
Disposables
• Micropipettor tips and solution dispensing trays
• 1.5-mL Eppendorf microcentrifuge tubes
2.3.2.2 Performing the fluorescence-based calcium assay using the FLIPRTETRA
Loading the calcium indicator dye. The calcium indicator dye was prepared
according to manufacturer’s protocol and 100 mL was added to the 100 mLof starvation medium in each well for a final volume of 200 mL/well. Cellswere then incubated for 1 h at 37 �C 5% CO2.
Preparation of the ligand dosing plate. During the preincubation, 96-well
clear round-bottom drug plates for the calcium assay were prepared as
described in Section 2.2.2.2 with the exceptions that ligand dilutions were
performed in 1� PBS and a total of 80 mL of 5� drug concentration was
placed in each well. 10 mM of the calcium ionophore A23187 was used as
the positive control on each drug plate.
Performing the calcium assay. 470–495 nm excitation and 515–575 nm
emission filters were installed prior to initializing the FLIPRTETRA. All assays
were run at room temperature. The instrument was programmed to simul-
taneously dispense 50 mL of vehicle control, 5� ligand, or A23187 from the
drug plate into the 200 mL of medium in the corresponding wells of the assay
plate to achieve the final ligand concentration. For measuring calcium fluo-
rescence, excitation intensitywas set to 50%, and detection gainwas set to 2000
with exposure time 0.53 s and the gate open 10.08%. Assay and ligand dosing
plates were loaded into the instrument and fluorescencewas recorded every 1 s
for 10 reads to establish baseline fluorescence, then every 1 s for 300 reads
(310 total reads over 5.17 min). In the FLIPRTETRA, the maximum change in
![Page 18: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/18.jpg)
246 Kathryn M. Appleton et al.
fluorescence was attained approximately 50 s after the addition of ligand
and returned to baseline within approximately 150 s. Raw data representing
the time–fluorescence relationship for each well following ligand addition
were exported to Microsoft Excel for background subtraction and analysis.
2.3.3 Estimating RAi for PTH1R-mediated calcium signalingSimilar to theprocess described inSection2.2.3, background fluorescencemea-
sured in vehicle-treated wells was subtracted from peak fluorescence in each
ligand-treated well to yield the net change in fluorescence. The mean net
change in fluorescence from triplicate wells at each ligand concentration
was determined, and all values were normalized to the peak fluorescence
observed with PTH(1–34). Using GraphPad Prism, each normalized
concentration–responsedatasetwas fit to a sigmoidal dose–responsecurveusing
variableHill slope.Emax andEC50valuesweredetermined fromthesecurves.At
least three separate experimental replicates were performed using each ligand.
As shown in Fig. 13.3, only two of the test ligands generated a significant
increase in intracellular calcium in our assay; hPTH(1–34) and hPTH(1–31).
Both ligands demonstrated equivalent efficacy, and although each was
approximately 20-fold less potent in the calcium assay than the cAMP assay,
hPTH(1–34) remained 1.5� more potent than hPTH(1–31). The finding
that hPTH(1–31) stimulates calcium flux via hPTH1R in our HEK293 cell
background probably reflects a tissue difference in hPTH1R signaling.
Although PTH(1–31) is incapable of activating membrane-associated
PKC in ROS 17/2 rat osteosarcoma cells ( Jouishomme et al., 1994), it
robustly elevates intracellular inositol phosphate levels in murine distal con-
voluted tubule cells (Sneddon et al., 2004). Trp1-hPTHrp(1–36) demon-
strated strong bias toward Gs coupling, as it had no significant effect on
intracellular calcium entry despite its ability to stimulate cAMP production
with potency and efficacy equivalent to PTH(1–31). This is consistent with
literature reports of the coupling selectivity of both Trp1-hPTHrp(1–36)
and a related Bpa1-PTHrp(1–36) peptide (Bisello, Horwitz, & Stewart,
2004; Gesty-Palmer et al., 2006). (D-Trp12,Tyr34)-PTH(7–34) also demon-
strated apparent bias in that its inverse agonism of hPTH1R-Gs coupling was
not evident toward hPTH1R-calcium signaling. This too is consistent with
literature reports (Gardella et al., 1996; Gesty-Palmer et al., 2006). As in the
cAMP assay, both hPTH(7–34) and bPTH(3–34) failed to stimulate calcium
flux. The failure of bPTH(3–34) to signal may reflect a species or tissue dif-
ference in PTH1R signaling. Although PTH(3–34) was reported to selec-
tively activate PKC in murine osteosarcoma cells ( Jouishomme et al.,
1992), amino terminal truncation of hPTH was found to completely
![Page 19: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/19.jpg)
Ligand EC50 Emax RAiPTH (1–34) 4.62 � 10-7
6.37 � 10-7
1.0
PTH (1–34)
PTH (1–31)
PTH (3–34)
PTH (7–34)
(D-Trp12,Tyr34)-PTH (7–34)
log, [Ligand] M
% M
ax P
TH
(1–
34)
resp
onse
-70.0
0.2
0.4
0.6
0.8
1.0
1.2
-6 -5
(Trp1)-PTHrp (1–36)
1.0
PTH (1–31) 0.92 0.67
(Trp1)-PTHrp (1–36) N.D. N.S. 0
PTH (3–34) N.D. N.S. 0
PTH (7–34) N.D. N.S. 0
0(D-Trp12,Tyr34)-PTH (7–34) N.D. N.S.
A
B
Figure 13.3 Ligand-mediated stimulation of intracellular calcium flux by the hPTH1R.(A) Ligand concentration–response relationships were determined for the hPTH1R ligandpanel using tetracycline-inducible hPTH1R HEK293-FlpIn TRex cells as described in thetext. The mean net change in calcium fluorescence was normalized to the maximalnet response elicited by the reference ligand, PTH(1–34). Sigmoidal dose–response curvesand SEM across at least three independent experiments were generated using GraphPadPrism. (B) For each ligand, RAi was estimated from the observed Emax and EC50 values.Ligands that failed to generate a statistically significant change in calcium fluorescencewere assigned a RAi of zero. NS, no significant response; ND, not determined.
247Biasing the Parathyroid Hormone Receptor
eliminate hPTH1R stimulation of inositol phosphate hydrolysis in renal
epithelial cells (Takasu, Gardella, Luck, Potts, & Bringhurst, 1999).
As with the cAMP assay, RAi values were estimated using the simplified
formula in Section 2 (Fig. 13.3). For intracellular calcium flux, the activity
rank order was hPTH(1–34)>hPTH(1–31)>>>Trp1-PTHrp(1–36)¼bPTH(3–34)¼hPTH(7–34)¼D-Trp12, Tyr34-bPTH(7–34).
2.4. Assaying PTH1R-mediated ERK1/2 activationRegulation of the ERK1/2 cascade is thought to contribute to the biological
actions of PTH1R, affecting diverse processes such as cellular mitogenesis,
embryologic development, and renal tubular phosphate transport
(Lederer, Sohi, & McLeish, 2000; Verheijen et al., 1999). PTH has been
![Page 20: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/20.jpg)
248 Kathryn M. Appleton et al.
reported both to activate ERK1/2 in opossum kidney cells (Lederer et al.,
2000) and to inhibit it in rat osteosarcoma cells (Verheijen & Defize, 1995).
Much of this complexity arises from the fact that ERK1/2 represents a
convergence point where multiple upstream signals are integrated. Recent
efforts to deconvolute the hPTH1R-ERK regulatory network in HEK293
cells and murine calvarial osteoblasts suggest that the pathway is indepen-
dently regulated by conventional G protein-dependent pathways involving
Gs-PKA and/or Gq/11-PKC and a G protein-independent pathway medi-
ated by arrestins (Gesty-Palmer et al., 2006, 2009). Evidence suggests that
these different mechanisms control the formation of spatially, temporally,
and functionally discrete pools of ERK1/2 that determine its ultimate bio-
logical function (Luttrell & Gesty-Palmer, 2010).
Largely because ERK1/2 is activated by a multitude of upstream signals,
it is often used for ligand screening.While it offers a single integrated readout
of receptor activation, it is often necessary to perform additional experiments
using pharmacologic inhibitors or RNA interference to determine the con-
tribution of different effectors, for example, arrestins, to the observed signal
(Gesty-Palmer et al., 2006). In addition, G protein-dependent ERK1/2 acti-
vation signals are often highly amplified, while arrestin-dependent ERK1/2
activation, which relies on the assembly of stoichiometric GPCR–arrestin
“signalsomes,” is unamplified, causing arrestin-dependent agonists to have a
reduced apparent efficacy.
2.4.1 Cell culture and inducible expression of hPTH1RERK1/2 assays were performed using the stable hPTH1R HEK293-FlpIn
TRex cell line. In some assays, the arrestin dependence of hPTH1R-mediated
ERK1/2 activation was assessed by determining the loss of signal resulting
from downregulation of arrestin2/3 expression by RNA interference.
2.4.1.1 Required materials• Cells: Tetracycline-inducible hPTH1R HEK293-FlpIn TRex cells
• Cell growth medium: DMEM supplemented with 10% FBS, 1%
antibiotic–antimycotic solution, and 50 mg/mL hygromycin B plus
5 mg/mL blasticidin for selection
• Serum-free medium: DMEM supplemented with 0.05% bovine serum
albumin
• Double-stranded siRNA oligomers (Qiagen, Inc.): Arrestin2/3: 50-AAACCTGCGCCTTCCGC TATG-30; Scrambled control: 50-AAUUCUCCGAACGUGUCACGU-30
![Page 21: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/21.jpg)
249Biasing the Parathyroid Hormone Receptor
• GeneSilencer™ transfection reagent (Gene Therapy Systems)
Additional solutions
• Cellgro 0.05% trypsin solution (Mediatech, Inc.)
• 1 mg/mL tetracycline stock solution in sterile distilled H2O
Disposables
• 10-cm tissue culture dishes
• 12-well tissue culture plates
2.4.1.2 Cell culture and induction of PTH1R expressionCell culture. hPTH1R HEK293-FlpIn TRex cells were maintained in growth
medium without tetracycline as described in Section 2.3.1.2 until they were
ready for assay plating.
Induction of hPTH1R expression and arrestin expression silencing. On day 1,
hPTH1R HEK293-FlpIn TRex cells were plated on 10-cm dishes in
growth medium containing 0.1 mg/mL tetracycline to induce PTH1R
expression. On day 3, 6�105 cells/well were seeded into 12-well plates
to establish a confluent monolayer for assay on day 4. On day 4, the growth
medium was aspirated and the cells were serum-starved for 3 h prior to assay
in 1 mL of DMEM supplemented with 0.05% BSA and tetracycline.
For experiments involving downregulation of arrestin2/3 expression by
RNA interference, 50% confluent hPTH1R HEK293-FlpIn TRex cells in
10-cm dishes were transfected on day 2 using 20 mg of scrambled control or
arrestin2/3-targeted siRNA oligomers using 50 ml of GeneSilencer transfec-tion reagent according to the manufacturer’s instructions (Lee et al., 2008).
The cells were seeded in 12-well tissue culture plates on day 3 and serum-
starved for assay on day 4 as described above. Silencing of arrestin expression
was confirmed by immunoblotting whole-cell lysates using rabbit polyclonal
anti-arrestin2/3 IgG with horseradish peroxidase-conjugated donkey anti-
rabbit IgG as the secondary antibody.
2.4.2 Assaying ERK1/2 phosphorylationLigand concentration–response curves were generated by immunoblotting
phospho-ERK1/2 in whole-cell lysates of hPTH1R HEK293-FlpIn TRex
cells following treatment with vehicle or PTH1R ligands.
2.4.2.1 Required materialsEquipment
• Invitrogen XCell SureLock Mini-Cell Electrophoresis System
• Invitrogen iBlot and iBlot nitrocellulose gel transfer stacks
• X-ray film developer
• Kodak Image Station for densitometric image analysis
![Page 22: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/22.jpg)
250 Kathryn M. Appleton et al.
Reagents and materials
• PTH1R test and reference ligands
• BCA protein assay kit (Promega)
• Invitrogen NuPAGE Novex precast SDS-PAGE 15-well 4–12% gradi-
ent gels
• Nitrocellulose membranes
• Rabbit polyclonal anti-phosphoERK1/2 IgG (Cell Signaling
Technology)
• Horseradish peroxidase-conjugated goat anti-rabbit IgG (Cell Signaling
Technology)
• SuperSignal West Pico ECL (Promega)
• X-ray film (Kodak)
• X-ray film cassettes
Additional solutions
• 200 mM phorbol myristate acetate (PMA) in dimethylsulfoxide
• 1� SDS lysis buffer: 62.5 mM Tris–HCl (pH 6.8), 2% (w/v) SDS, 10%
glycerol, 50 mM DTT, 0.01% (w/v) bromophenol blue
• Tris-buffered saline with Tween 20 (TBST): 20 mM Tris–HCl, pH 8;
150 mM NaCl; 0.001% Tween 20 in distilled H2O.
• Blocking buffer: 5% nonfat dry milk in TBST
• Primary antibody buffer: 5% BSA in TBST
Disposables
• 1.5-mL Eppendorf microcentrifuge tubes
• Cell scrapers
• Micropipettor tips
2.4.2.2 Measuring PTH1R-regulated ERK1/2 phosphorylationCell stimulation. Serial dilutions of PTH1R ligands were prepared at 1000�desired final concentration. At time 0, 1 mL of vehicle or ligand concentrate
was added to the 1 mL of serum-free media in each well and the plates were
incubated for 5 min at 37 �C. PMA at a final concentration of 200 nM was
added to one well of each plate as a positive control.
Phospho-ERK1/2 Immunoblotting. Five minutes after ligand application,
the medium was aspirated and monolayers were lysed in 100 mL of 1�SDS lysis buffer, harvested in 1.5-mL Eppendorf tubes by scraping, briefly
sonicated, and clarified by microcentrifugation. The protein content of each
sample was determined using a BCA protein assay kit according to manu-
facturer’s protocol. Equal amounts of protein sample (10 mg) were loaded
![Page 23: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/23.jpg)
251Biasing the Parathyroid Hormone Receptor
onto 4–12% SDS-PAGE gradient gels and proteins resolved by electropho-
resis at 120 V constant voltage for 2 h. The proteins were transferred to
nitrocellulose membranes using the iBlot apparatus according to manufac-
turer’s protocol. The membranes were blocked in 5% dry milk in TBST for
1 h at room temperature with gentle shaking and then incubated overnight
at 4 �C with rabbit polyclonal anti-phosphoERK1/2 antibody at 1:1000
dilution in 5% BSA in TBST with gentle shaking. Following 3�5 min
TBST washes, the membranes were incubated for 1 h at room temperature
with HRP-conjugated goat anti-rabbit IgG at 1:10,000 dilution in 5% dry
milk in TBST with gentle shaking. The membranes were again washed
3�10 min in TBST, developed using SuperSignal West Pico ECL reagent
according to manufacturer’s directions, and exposed to Kodak X-ray film.
Exposure times were varied to capture the widest possible range of band
intensities. Quantification of phosphoERK1/2 pixel intensity was per-
formed using a Kodak Image Station.
2.4.3 Estimating RAi for PTH1R-mediated ERK1/2 activationUsing Microsoft Excel, the net change in phospho-ERK1/2 band intensity
from basal was determined at each ligand concentration, and all values were
normalized to the peak band intensity observed with PMA.Using GraphPad
Prism, each normalized concentration–response dataset was fit to a sigmoidal
dose–response curve using variable Hill slope. Emax and EC50 values were
determined from these curves. At least three separate experimental replicates
were performed using each ligand.
As shown in Fig. 13.4, all the PTH1R ligands except hPTH(7–34) gen-
erated a statistically significant increase in ERK1/2 phosphorylation. All
three ligands that were active in the cAMP assay, hPTH(1–34), hPTH
(1–31), and Trp1-hPTHrp(1–36), produced robust ERK1/2 activation with
nanomolar EC50 values. bPTH(3–34), which was without detectable activ-
ity in either the cAMP or calcium assays, also robustly activated ERK1/2.
The Emax of (D-Trp12,Tyr34)-bPTH(7–34), which demonstrated inverse
agonism for cAMP production and was inactive in the calcium assay, was
less than 10% that of PTH(1–34). The ligands that activated ERK1/2 with-
out producing detectable G protein signals were 3- to 12-fold less potent
than PTH(1–34). hPTH(7–34) appeared as a neutral antagonist in all assays,
failing to elicit measureable changes in cAMP production, intracellular
calcium flux, or ERK1/2 activation.
To estimate the contribution of arrestin-dependent signaling to the
ERK1/2 response, single concentration assays (1 mM) were performed in
![Page 24: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/24.jpg)
Ligand EC50 Emax RAiPTH (1–34) 3.76 � 10-9
PTH (1–34)
PTH (1–31)
PTH (3–34)
(Trp1)-PTHrp (1–36)
(D-Trp12,Tyr34)-PTH (7–34)
log, [Ligand] M
-120.0
0.2
0.4
0.6
0.8
1.0
0.00-10 -8 -6
0.02
0.04
0.06
0.08
0.10
-10 -8 -6 -4
% P
MA
max
res
pons
e
% P
MA
max
res
pons
e
log, [Ligand] M
PTH (7–34)
1.08 � 10-8
4.36 � 10-8
3.42 � 10-9
8.12 � 10-9
0.91 1.0
PTH (1–31) 0.81 0.41
(Trp1)-PTHrp (1–36) 0.85 1.02
PTH (3–34) 0.68 0.26
PTH (7–34) N.D. N.S. 0
(D-Trp12,Tyr34)-PTH (7–34) 0.07 0.007
A
B
Figure 13.4 Ligand-mediated ERK1/2 activation by the hPTH1R. (A) Ligand concentra-tion–response relationships were determined for the hPTH1R ligand panel usingtetracycline-inducible hPTH1R HEK293-FlpIn TRex cells as described in the text. Thenet change in ERK1/2 was normalized to the net response elicited by a maximally effi-cacious dose of PMA. Sigmoidal dose–response curves and SEM across at least threeindependent experiments were generated using GraphPad Prism. (B) For each ligand,RAi was estimated from the observed Emax and EC50 values. Ligands that failed to gen-erate a statistically significant change in ERK1/2 phosphorylation were assigned a RAi ofzero. NS, no significant response; ND, not determined.
252 Kathryn M. Appleton et al.
hPTH1R HEK293-FlpIn TRex cells after arrestin2/3 expression was
downregulated >80% by RNA interference. Figure 13.5 depicts the sensi-
tivity of basal- and ligand-stimulated ERK1/2 activation to arrestin silenc-
ing. Consistent with the literature reports suggesting that arrestin signaling
contributes to ERK1/2 activation by conventional ligands (Ahn, Shenoy,
Wei, & Lefkowitz, 2004; Gesty-Palmer et al., 2006), hPTH(1–34) and
hPTH(1–31) responses were modestly attenuated, demonstrating approxi-
mately 20% reduction of the peak ERK1/2 activation observed at 5 min
stimulation. In contrast, ERK1/2 activation by bPTH(3–34) and
(D-Trp12,Tyr34)-bPTH(7–34), which showed antagonist or inverse agonist
activity in assays of G-protein signaling, were 70–80% sensitive to arrestin
![Page 25: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/25.jpg)
siRNA Arrestin 2/3
pERK IB
Scrambled
NS
% I
nhib
itio
n of
ER
Kph
osph
oryl
atio
n
PTH (1
–34)
0
20
40
60
80
100+ -- +
+ -- +
+ -- +
+ -- +
+ -- +
+ -- +
PTH (1
–31)
PTH (3
–34)
(D-T
rp12 ,T
yr34 )-P
TH (7
–34)
(Trp1 )-P
THrp (1
–36)
Figure 13.5 Sensitivity of ligand-mediated ERK1/2 activation to downregulation ofarrestin2/3 expression. To elucidate the arrestin dependence of ERK1/2 activation bydifferent hPTH1R ligands, hPTH1R HEK293-FlpIn TRex cells were transfected with scram-bled control or arrestin2/3-targeted siRNA prior to determining the ERK1/2 response to amaximally effective concentration (1 mM) of each ligand that produced significant ERK1/2activation. The upper panel shows a representative immunoblot taken from one of threeindependent experiments. The bar graph presents the percentage inhibition of thephosphoERK1/2 signal in arrestin2/3 siRNA-treated cells normalized to the responsein control siRNA-treated cells. As shown, ligands that stimulated ERK1/2 withoutdetectably stimulating cAMP production were more sensitive to arrestin silencing thanthose producing a robust G protein-mediated response.
253Biasing the Parathyroid Hormone Receptor
downregulation. Interestingly, Trp1-PTHrp(1–36), which has been described
as a “nondesensitizing” Gs-selective hPTH1R ligand, was intermediate in its
sensitivity to arrestin knockdown. Overall, these data are consistent with
reports that arrestin-dependent signaling accounts for the majority of G
protein-independent ERK1/2 activation by GPCR ligands (Luttrell &
Gesty-Palmer, 2010).
As with the preceding assays, RAi values were estimated using the sim-
plified formula in Section 2. For ERK1/2 activation, the activity rank order
was hPTH(1–34)>Trp1-PTHrp(1–36)>hPTH(1–31)>bPTH(3–34)>D-Trp12, Tyr34-bPTH(7–34)>hPTH(7–34). Because (D-Trp12,Tyr34)-
bPTH(7–34) exhibited less than 1/10th the apparent efficacy and a 12-fold
lower potency than PTH(1–34), the estimated RAi for ERK1/2 activation
is markedly lower than that for the other active ligands.
![Page 26: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/26.jpg)
254 Kathryn M. Appleton et al.
2.5. Estimating PTH1R ligand biasFigure 13.6A presents a simple graphic representation of the activity profiles of
our ligand panel based onRAi estimates. As shown, hPTH(1–34) behaves as a
full agonist in all three assays, while hPTH(1–31) appears as a relatively
A BpERK1/2
pERK1/2
Ca2+i–1.0
–0.8
–0.6
–0.4
–0.2
–0.2 0.2 0.4 0.6 0.8 1.0–0.4–0.6–0.8–1.0 0
–0.2
0
0.2
0.4
0.6
0.8
1.0
–0.4
–0.6
–0.8
–1.0
Ca2+i
cAMP cAMP
0.2
1.0
0.8
0.6
0.4pER
K
0.2
1.0
D C
0.8PTH(1–34)PTH(1–31)(Trp1)-PTHrp (1–36)PTH (3–34)PTH (7–34)(D-Trp12,Tyr34)-PTH (7–34)
0.6
0.4pER
K
0.2
0.00.0 0.2 0.4
Ca2+0.6 0.8 1.0 –1.0 –0.5 0.0
cAMP
0.5 1.0
–1.0 –0.5 0.0
cAMP
0.5 1.0
0.4
Ca2
+
0.6
0.8
1.0
0.2
0.4
0.6
0.8
Figure 13.6 Graphic analysis of PTH1R ligand bias. (A) Multiaxial representation of PTH1Rligand activity in the cAMP, calcium, and ERK1/2 assays. Estimated RAi values for eachligand are plotted on each axis to represent themagnitude and direction of effect in eachsignaling response. Ligands with “balanced” efficacy, such as hPTH1-34 and hPTH(1–31),show effects of similar amplitude and direction on all three axes, while those demonstrat-ing “bias” show disproportionate activity in one or more pathways or reversal of efficacy.(B–D) Pairwise comparison of ligand activity at equimolar concentration in cAMP versuscalcium signaling (B), cAMP versus ERK1/2 phosphorylation (C), and calcium signaling ver-sus ERK1/2 phosphorylation (D). In each plot, deviation from the line of unity reflects dif-ferences in coupling efficiency between the receptor and its downstream effectors, forexample, cAMP and calcium signaling. Ligands whose coupling efficiency between effec-tor pathwaysmatches that of the reference ligand, for example, PTH(1–34) and PTH(1–31),are unbiased, while those that deviate significantly exhibit functional selectivity. (D-Trp12,Tyr34)-bPTH(7–34), which acts as an inverse agonist for cAMP production, is neutral forcalcium signaling and is a partial agonist for ERK1/2 activation, demonstrates assay-dependent reversal of efficacy, a hallmark of biased GPCR agonism.
![Page 27: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/27.jpg)
255Biasing the Parathyroid Hormone Receptor
“balanced” partial agonist. hPTH(7–34), which was inactive in all assays,
appears as a “neutral” antagonist. The other three ligands exhibit apparent sig-
nal bias. Trp1-PTHrp(1–36) appears to be selective for Gs-cAMP signaling,
since it fails to elicit a threshold calcium signal at concentrations that robustly
activate cAMP production and ERK1/2, whereas hPTH(1–34) and hPTH
(1–31) exhibit proportional RAi in all three assays. bPTH(3–34) exhibitsmore
drastic bias, retaining substantial activity in the ERK1/2 assay in the absence of
a detectable cAMP or calcium signal. (D-Trp12,Tyr34)-bPTH(7–34) exhibits
frank reversal of efficacy, one of the hallmarks of biased agonism, appearing as
an inverse agonist for cAMP production, neutral for calcium signaling, and a
partial agonist for ERK1/2 activation.
A useful way of visualizing ligand bias is to plot the response observed in
two different assays at equal ligand concentrations against one another
(Gregory, Hall, Tobin, Sexton, & Christopoulos, 2010). The resultant plot
is a direct comparison of the efficiency of signaling through the two path-
ways. Figure 13.6B shows the relationship between cAMP and calcium sig-
naling for our panel of six PTH1R ligands. In our HEK293 cell background,
the reference ligand hPTH(1–34) is more efficiently coupled to cAMP sig-
naling, in that a significant increase in cAMP production occurs before the
rise in the calcium response, that is, the calcium concentration–response
curve is right-shifted relative to the cAMP curve. hPTH(1–31) appears as
a balanced agonist, in that its cAMP-calcium signal coupling overlaps that
of the reference ligand. bPTH(3–34) and hPTH(7–34) appear as balanced
antagonists, producing no significant change in either assay. Both Trp1-
hPTHrp(1–36) and (D-Trp12,Tyr34)-bPTH(7–34) show apparent signal bias.
Trp1-hPTHrp(1–36) exhibits Gs-selective agonist bias, in that it achieves
cAMP production activity equivalent to the reference agonist while failing
to significantly increase calcium flux. Similarly, (D-Trp12,Tyr34)-bPTH
(7–34) exhibits Gs-selective inverse agonist bias contrasting with neutral
antagonism of calcium signaling.
Figure 13.6C plots the relationship between cAMP and ERK1/2 activa-
tion. Here, the strongly Gs-coupled ligands, hPTH(1–34), hPTH(1–31), and
Trp1-hPTHrp(1–36) appear unbiased, and hPTH(7–34) remains a neutral
antagonist. Both bPTH(3–34) and (D-Trp12,Tyr34)-bPTH(7–34) show signal
bias. bPTH(3–34) robustly activates ERK1/2 without producing a significant
cAMP signal, while (D-Trp12,Tyr34)-bPTH(7–34) exhibits reversal of effi-
cacy, appearing as an inverse agonist for cAMP production and a partial ago-
nist for ERK1/2. The ability of bPTH(3–34) and (D-Trp12,Tyr34)-bPTH
(7–34) to activate ERK1/2 in the absence of significant G protein-signaling
correlates with a high degree of sensitivity to arrestin2/3 knockdown,
![Page 28: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/28.jpg)
256 Kathryn M. Appleton et al.
suggesting that these ligands rely on arrestin coupling to activate ERK1/
2. Figure 13.6D shows the relationship between calcium signaling and
ERK1/2 activation. The two balanced ligands, hPTH(1–34) and
hPTH(1–31), appear similar, while Trp1-hPTHrp(1–36), bPTH(3–34),
and (D-Trp12,Tyr34)-bPTH(7–34) appear to be ERK1/2 biased, with par-
tial or full agonist activity in the ERK1/2 assay in the absence of a detectable
calcium response. hPTH(7–34), which has no significant activity in any of
the assays, still appears as a neutral antagonist.
3. DISCUSSION
The principal targets of PTH in vivo are kidney and bone, where its
actions promote a rise in serum calcium. In the kidney, PTH stimulates
reabsorption of filtered calcium by the cortical thick ascending limb of
the loop of Henle and distal convoluted tubule. In the proximal tubule, it
stimulates expression of the 1a-hydroxylase that converts 25(OH)-vitamin
D to its active form 1,25(OH)2-vitamin D, which in turn enhances intestinal
calcium absorption. The actions of PTH in bone are complex. PTH directly
stimulates osteoblasts, accelerating bone formation by increasing osteoblast
number and activity, promoting the deposition of new bone matrix and
accelerating the rate of mineralization (Dobnig & Turner, 1995; Schmidt,
Dobnig, & Turner, 1995). At the same time, PTH accelerates bone degra-
dation and the release of calcium locked in the mineralized skeleton by pro-
moting the recruitment, differentiation, and activity of bone-resorbing
osteoclasts. The effects of PTH on osteoclasts are indirect. Osteoclasts lack
PTH receptors, but respond to factors secreted by osteoblasts in response to
PTH, such as receptor activator of nuclear factor kB ligand. Because the ana-
bolic and catabolic effects of PTH are coupled, the net effect of PTH on
bone is dependent upon the kinetics of receptor activation, with intermit-
tent exposure leading to a net increase in bone formation, while continuous
exposure produces net bone loss and possible hypercalcemia (Dobnig &
Turner, 1995; Hock & Gera, 1992; Qin, Raggatt, & Partridge, 2004).
The skeletal effects of several conventional and biased PTH1R ligands
have already been determined in murine models, permitting some degree
of direct comparison between their in vitro efficacy profile and in vivo bio-
logical activity. PTH(1–34), which elicits the full range of PTH1R signaling
in vitro, also reproduces the full spectrum of PTH action in vivo. Mice given
daily injections of PTH(1–34) show increased indices of bone formation and
a net increase in bone mass (Dobnig & Turner, 1995). Osteoblast number,
![Page 29: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/29.jpg)
257Biasing the Parathyroid Hormone Receptor
osteoid surface, serum osteocalcin level, and mineral apposition rates, all
increase, as does trabecular bone volume and cortical thickness. At the same
time, indices of osteoclastic bone resorption rise, including osteoclast num-
ber and urinary deoxypyrodinoline level. Serum and urine calcium levels
also rise, reflecting the net effect of PTH(1–34) on bone resorption, intes-
tinal calcium absorption, and renal tubular calcium retention. PTH(1–31),
which in our screen appears to be a balanced partial agonist, also resembles
PTH(1–34) in vivo (Mohan et al., 2000). With several weeks treatment,
PTH(1–31) increases markers of bone formation as effectively as PTH
(1–34). The increase in bone resorption parameters is less dramatic
for PTH(1–31), yet the increase in bone density is also smaller. PTHrp
(1–36), which like Trp1-hPTHrp(1–36) is reportedly Gs pathway-selective,
is also anabolic in vivo, suggesting that Gq/11 signaling may be dispensable for
PTH1R actions in bone. In an ovariectomized rat model of bone loss, both
PTH(1–34) and PTHrp(1–36) increase indices of bone formation, bone
mass, and bone strength (Stewart et al., 2000). PTH(1–34), which produces
more sustained Gs-cAMP activation than PTHrp(1–36) in vitro, is more effi-
cacious in this model, where neither agent produces significant increases in
osteoclast activity.
UnlikeGq/11, thecapacity toactivateGs signalingappears tobe linked to the
anabolic effects of conventional PTH1R ligands. In mice, the N-terminal
truncated fragment, PTH(2–34),which is dramatically impaired inGs coupling
is far less efficacious than either PTH(1–34) or PTH(1–31) (Mohan et al.,
2000). Similarly, comparison of the ligand series, PTH(1–38), PTH(2–38),
and PTH(3–38) in rats further supports the conclusion that Gs signaling
is critical. Despite the capacity to activate PKC and simulate mitogenesis
in rat osteoblastic cells in vitro, PTH(3–38) produced no detectable anabolic
or catabolic effects on bone in vivo (Hilliker, Wergedal, Gruber, Bettica, &
Baylink, 1996).
However, results obtained with the putatively arrestin pathway-selective
biased agonist (D-Trp12, Tyr34)-bPTH(7–34) seem to contradict the general
conclusion that Gs activation is requisite for the anabolic actions of PTH.
Arrestin3 null mice exhibit higher basal rates of bone turnover and an
impaired anabolic response to PTH(1–34), with blunted increases in trabec-
ular bone volume and no change in cortical thickness compared to controls.
The attenuated response is associated with smaller changes in osteoblast
number and osteoid deposition, but preserved or exaggerated increases in
osteoclast number and urine deoxypyrodinoline (Bouxsein et al., 2005;
Ferrari et al., 2005; Gesty-Palmer et al., 2009). While this supports the
![Page 30: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/30.jpg)
258 Kathryn M. Appleton et al.
conclusion that PTH1R-mediated osteoclast activation is Gs-dependent, it
does not address whether the diminished anabolic PTH(1–34) response in
arrestin3 null mice reflects exaggerated cAMP signaling in the setting of
impaired arrestin-mediated desensitization or loss of arrestin-mediated signal-
ing. Paradoxically, intermittent administration of (D-Trp12, Tyr34)-bPTH
(7–34) to wild-type mice results in increased bone mass, characterized by
greater trabecular bone volume and increased osteoblast number, osteoid
surface, serum osteocalcin, and mineral apposition rate, with no significant
effect on osteoclast number or bone turnover markers (Gesty-Palmer et al.,
2009). All responses to (D-Trp12, Tyr34)-bPTH(7–34) are either absent or
reversed in arrestin 3 null mice, consistent with the hypothesis that arrestin sig-
naling in vivo contributes to the anabolic response to PTH, andwhen activated
in isolation is sufficient to promote osteoblastic bone formation but not to
stimulate osteoclastic bone resorption.
Results such as these, wherein two GPCR ligands with dramatically dif-
ferent in vitro activity profiles produce qualitatively similar biological responses
in vivo, suggest that there is something different about “arrestin-biased”
PTH1R activation at the tissue level that is not predicted by conventional
ligand screening. At present, it is unclear whether this means that changing
the balance of GPCR coupling is sufficient to produce markedly different
responses at the tissue level, where the strength/duration of signals originating
from proximal effectors are integrated to produce a response, or whether
biased ligands can couple GPCRs in “unnatural” ways, generating qualita-
tively different proximal efficacy. Either way, the finding that biasing GPCR
signaling can engender new/unexpected signaling outcomes has significant
implications for the rationale design of functionally selective drugs. While
it is clearly possible to tailor ligands to elicit specific efficacy profiles, we
are in many cases left with the quandary of not knowing which downstream
signals to favor and which to avoid. Thus, the greatest challenge at present is
not in detecting ligand bias, but in determining what efficacy profile is needed
to produce the optimal therapeutic response in any given setting.
ACKNOWLEDGMENTSThe authors thank Dr. Diane Gesty-Palmer (Duke University Medical Center) for helpful
advice and criticism and Allie Pinosky for technical assistance. The work was supported by
National Institutes of Health Grant R01 DK55524 (L. M. L.), the MUSC fluorescence
imaging plate reader (FLIPRTETRA) facility (S10 RR027777; L. M. L./T. A. M.), and the
Research Service of the Charleston, SC Veterans Affairs Medical Center (L. M. L./T. A. M.).
The contents of this chapter do not represent the views of the Department of Veterans Affairs
or the United States Government.
![Page 31: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/31.jpg)
259Biasing the Parathyroid Hormone Receptor
REFERENCESAbou-Samra, A. B., Juppner, H., Force, T., Freeman,M.W., Kong, X. F., Schipani, E., et al.
(1992). Expression cloning of a common receptor for parathyroid hormone and parathy-roid hormone-related peptide from rat osteoblast-like cells: A single receptor stimulatesintracellular accumulation of both cAMP and inositol trisphosphates and increases intra-cellular free calcium. Proceedings of the National Academy of Sciences of the United States ofAmerica, 89, 2732–2736.
Ahn, S., Shenoy, S. K., Wei, H., & Lefkowitz, R. J. (2004). Differential kinetic and spatialpatterns of beta-arrestin and G protein-mediated ERK activation by the angiotensin IIreceptor. The Journal of Biological Chemistry, 279, 35518–35525.
Barbier, J. R., Gardella, T. J., Dean, T., MacLean, S., Potetinova, Z., Whitfield, J. F., et al.(2005). Backbone-methylated analogues of the principle receptor binding region of humanparathyroid hormone. Evidence for binding to both the N-terminal extracellular domainand extracellular loop region. The Journal of Biological Chemistry, 280, 23771–23777.
Berg, K. A., Maayani, S., Goldfarb, J., Scaramellini, C., Leff, P., & Clarke, W. P. (1998).Effector pathway-dependent relative efficacy at serotonin type 2A and 2C receptors: Evi-dence for agonist-directed trafficking of receptor stimulus. Molecular Pharmacology, 54,94–104.
Binkowski, B. F., Fan, F., &Wood, K. V. (2011). Luminescent biosensors for real-timemon-itoring of intracellular cAMP. Methods in Molecular Biology, 756, 263–271.
Bisello, A., Horwitz, M. J., & Stewart, A. F. (2004). Parathyroid hormone-related protein:An essential physiological regulator of adult bone mass. Endocrinology, 145, 3551–3553.
Black, J. W., & Leff, P. (1983). Operational models of pharmacological agonist. Proceedings ofthe Royal Society London B Biological Sciences, 220, 141–162.
Bouxsein, M. L., Pierroz, D. D., Glatt, V., Goddard, D. S., Cavat, F., Rizzoli, R., et al.(2005). Beta-arrestin2 regulates the differential response of cortical and trabecular boneto intermittent PTH in female mice. Journal of Bone and Mineral Research, 20, 635–643.
Bringhurst, F. R., Juppner, H., Guo, J., Urena, P., Potts, J. T., Jr., Kronenberg, H. M., et al.(1993). Cloned, stably expressed parathyroid hormone (PTH)/PTH-related peptidereceptors activate multiple messenger signals and biological responses in LLC-PK1 kid-ney cells. Endocrinology, 132, 2090–2098.
Dobnig, H., & Turner, R. T. (1995). Evidence that intermittent treatment with parathyroidhormone increases bone formation in adult rats by activation of bone lining cells.Endocrinology, 136, 3632–3638.
Ehlert, F. J. (2000). Ternary complex model. In A. Christopolous (Ed.), Biomedical applicationsof computer modeling (pp. 21–85). Boca Raton: CRC Press.
Ferrari, S. L., Behar, V., Chorev, M., Rosenblatt, M., & Bisello, A. (1999). Endocytosis ofligand-human parathyroid hormone receptor 1 complexes is protein kinase C-dependentand involves beta-arrestin2.Real-timemonitoring by fluorescencemicroscopy.The Journalof Biological Chemistry, 274, 29968–29975.
Ferrari, S. L., Pierroz, D. D., Glatt, V., Goddard, D. S., Bianchi, E. N., Lin, F. T., et al.(2005). Bone response to intermittent parathyroid hormone is altered in mice null forbeta-arrestin2. Endocrinology, 146, 1854–1862.
Figueroa, K. W., Griffin, M. T., & Ehlert, F. J. (2009). Selectivity of agonists for the activestate of M1 to M4 muscarinic receptor subtypes. The Journal of Pharmacology and Exper-imental Therapeutics, 328, 331–342.
Galandrin, S., & Bouvier, M. (2006). Distinct signaling profiles of beta1 and beta2 adrenergicreceptor ligands toward adenylyl cyclase and mitogen-activated protein kinase reveals thepluridimensionality of efficacy. Molecular Pharmacology, 70, 1575–1584.
Gardella, T. J., Luck,M. D., Jensen, G. S., Schipani, E., Potts, J. T., Jr., & Juppner, H. (1996).Inverse agonism of amino-terminally truncated parathyroid hormone (PTH) andPTH-related peptide (PTHrP) analogs revealed with constitutively active mutantPTH/PTHrP receptors. Endocrinology, 137, 3936–3941.
![Page 32: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/32.jpg)
260 Kathryn M. Appleton et al.
Gesty-Palmer, D., Chen, M., Reiter, E., Ahn, S., Nelson, C. D., Wang, S., et al. (2006).Distinct conformations of the parathyroid hormone receptor mediate G protein andbeta-arrestin dependent activation of ERK1/2. The Journal of Biological Chemistry, 281,10856–10864.
Gesty-Palmer, D., Flannery, P., Yuan, L., Corsino, L., Spurney, R., Lefkowitz, R. J., et al.(2009). A beta-Arrestin biased agonist of the parathyroid hormone receptor (PTH1R)promotes bone formation independent of G protein activation. Science TranslationalMedicine, 1, 1ra1.
Gesty-Palmer, D., & Luttrell, L. M. (2011). ‘Biasing’ the parathyroid hormone receptor: Anovel anabolic approach to increasing bone mass? British Journal of Pharmacology, 164,59–67.
Gregory, K. J., Hall, N. E., Tobin, A. B., Sexton, P. M., & Christopoulos, A. (2010). Iden-tification of orthosteric and allosteric site mutations in M2 muscarinic acetylcholinereceptors that contribute to ligand-selective signaling bias. The Journal of Biological Chem-istry, 285, 7459–7474.
Griffin,M. T., Figueroa, K.W., Liller, S., & Ehlert, F. J. (2007). Estimation of agonist activityat G protein-coupled receptors: Analysis of M2 muscarinic receptor signaling throughGi/o, Gs, and G15. The Journal of Pharmacology and Experimental Therapeutics, 321,1193–1207.
Hilliker, S., Wergedal, J. E., Gruber, H. E., Bettica, P., & Baylink, D. J. (1996). Truncationof the amino terminus of PTH alters its anabolic activity on bone in vivo. Bone, 19,469–477.
Hoare, S. R., & Usdin, T. B. (2000). Tuberoinfundibular peptide (7–39) [TIP(7–39)], anovel, selective, high-affinity antagonist for the parathyroid hormone-1 receptor withno detectable agonist activity. The Journal of Pharmacology and Experimental Therapeutics,295, 761–770.
Hock, J. M., & Gera, I. (1992). Effects of continuous and intermittent administration andinhibition of resorption on the anabolic response of bone to parathyroid hormone. Jour-nal of Bone and Mineral Research, 7, 65–72.
Jouishomme, H., Whitfield, J. F., Chakravarthy, B., Durkin, J. P., Gagnon, L., Isaacs, R. J.,et al. (1992). The protein kinase-C activation domain of the parathyroid hormone. Endo-crinology, 130, 53–60.
Jouishomme, H., Whitfield, J. F., Gagnon, L., MacLean, S., Isaacs, R., Chakravarthy, B.,et al. (1994). Further definition of the protein kinase C activation domain of the para-thyroid hormone. Journal of Bone and Mineral Research, 9, 943–949.
Juppner, H., Abou-Samra, A. B., Freeman, M., Kong, X. F., Schipani, E., Richards, J., et al.(1991). AG protein-linked receptor for parathyroid hormone and parathyroid hormone-related peptide. Science, 254, 1024–1026.
Karlin, A. (1967). On the application of “a plausible model” of allosteric proteins to thereceptor for acetylcholine. Journal of Theoretical Biology, 16, 306–320.
Kenakin, T. (1995). Agonist-receptor efficacy. II. Agonist-trafficking of receptor signals.Trends in Pharmacological Sciences, 16, 232–238.
Kenakin, T. P. (2009). 7TM receptor allostery: Putting numbers to shapeshifting proteins.Trends in Pharmacological Sciences, 30, 460–469.
Kenakin, T., & Miller, L. E. (2010). Seven transmembrane receptors as shapeshifting pro-teins: The impact of allosteric modulation and functional selectivity on new drug discov-ery. Pharmacological Reviews, 62, 265–304.
Lederer, E. D., Sohi, S. S., & McLeish, K. R. (2000). Parathyroid hormone stimulates extra-cellular signal-regulated kinase (ERK) activity through two independent signal transduc-tion pathways: Role of ERK in sodium-phosphate cotransport. Journal of American Societyof Nephrology, 11, 222–231.
![Page 33: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/33.jpg)
261Biasing the Parathyroid Hormone Receptor
Lee,M.H., EI-Shewy, H.M., Luttrell, D. K., & Luttrell, L.M. (2008). Role of beta-arrestin-mediated desensitization and signaling in the control of angiotensin AT1a receptor-stimulated transcription. Journal of Biological Chemistry, 283, 2088–2097.
Luttrell, L. M., & Gesty-Palmer, D. (2010). Beyond desensitization: Physiological relevanceof arrestin-dependent signaling. Pharmacological Reviews, 62, 305–330.
Luttrell, L. M., & Kenakin, T. P. (2011). Refining efficacy: Allosterism and bias in G protein-coupled receptor signaling. Methods in Molecular Biology, 756, 3–35.
Mahon, M. J., Donowitz, M., Yun, C. C., & Segre, G. V. (2002). Na(þ)/H(þ) exchangerregulatory factor 2 directs parathyroid hormone 1 receptor signaling. Nature, 417,858–861.
Mohan, S., Kutilek, S., Zhang, C., Shen, H. G., Kodama, Y., Srivastava, A. K., et al. (2000).Comparison of bone formation responses to parathyroid hormone(1–34), (1–31), and(2–34) in mice. Bone, 27, 471–478.
Nasman, J., Kukkonen, J. P., Ammoun, S., & Akerman, K. E. (2001). Role of G-proteinavailability in differential signaling by a2-adrenoceptors. Biochemical Pharmacology, 62,913–922.
Nussbaum, S. R., Rosenblatt, M., & Potts, J. T., Jr. (1980). Parathyroid hormone—Renalreceptor interactions. Demonstration of two receptor-binding domains. The Journal ofBiological Chemistry, 255, 10183–10187.
Pines, M., Adams, A. E., Stueckle, S., Bessalle, R., Rashti-Behar, V., Chorev, M., et al.(1994). Generation and characterization of human kidney cell lines stably expressingrecombinant human PTH/PTHrP receptor: Lack of interaction with a C-terminalhuman PTH peptide. Endocrinology, 135, 1713–1716.
Qin, L., Raggatt, L. J., & Partridge, N. C. (2004). Parathyroid hormone: A double-edgedsword for bone metabolism. Trends in Endocrinology and Metabolism, 15, 60–65.
Rajagopal, S., Ahn, S., Rominger, D. H., Gowen-MacDonald, W., Lam, C. M.,Dewire, S. M., et al. (2011). Quantifying ligand bias at seven-transmembrane receptors.Molecular Pharmacology, 80, 367–377.
Rajagopal, S., Rajagopal, K., & Lefkowitz, R. J. (2010). Teaching old receptors new tricks:biasing seven-transmembrance receptors. Nature Reviews Drug Discovery, 9, 373–386.
Schmidt, I., Dobnig, H., & Turner, R. (1995). Intermittent parathyroid hormone treatmentincreases osteoblast number, steady state messenger ribonucleic acid levels forosteocalcin, and bone formation in tibial metaphysis of hypophysectomized female rats.Endocrinology, 136, 5127–5134.
Segre, G. V., Rosenblatt, M., Reiner, B. L., Mahaffey, J. E., & Potts, J. T., Jr. (1979). Char-acterization of parathyroid hormone receptors in canine renal cortical plasma membranesusing a radioiodinated sulfur-free hormone analogue. Correlation of binding with ade-nylate cyclase activity. The Journal of Biological Chemistry, 254, 6980–6986.
Sneddon, W. B., Bisello, A., Magyar, C. E., Willick, G. E., Syme, C. A., Galbiati, F., et al.(2004). Ligand-selective dissociation of activation and internalization of the parathyroidhormone receptor. Conditional efficacy of PTH peptide fragments. Endocrinology, 145,2815–2823.
Stewart, A. F., Cain, R. L., Burr, D. B., Jacob, D., Turner, C. H., &Hock, J. M. (2000). Six-month daily administration of parathyroid hormone and parathyroid hormone-relatedprotein peptides to adult ovariectomized rats markedly enhances bone mass and biome-chanical properties: A comparison of human parathyroid hormone 1–34, parathyroidhormone-related protein 1–36, and SDZ-parathyroid hormone 893. Journal of Boneand Mineral Research, 15, 1517–1525.
Takasu, H., & Bringhurst, F. R. (1998). Type-1 parathyroid hormone (PTH)/PTH-relatedpeptide (PTHrP) receptors activate phospholipase C in response to carboxyl-truncatedanalogs of PTH(1–34). Endocrinology, 139, 4293–4299.
![Page 34: [Methods in Enzymology] G Protein Coupled Receptors - Modeling, Activation, Interactions and Virtual Screening Volume 522 || Biasing the Parathyroid Hormone Receptor](https://reader035.vdocuments.site/reader035/viewer/2022080407/575096111a28abbf6bc755cb/html5/thumbnails/34.jpg)
262 Kathryn M. Appleton et al.
Takasu, H., Gardella, T. J., Luck, M. D., Potts, J. T., & Bringhurst, F. R. (1999). Amino-terminal modifications of human parathyroid hormone (PTH) selectively alter phospho-lipase C signaling via the type 1 PTH receptor: Implications for design of signal-specificPTH ligands. Biochemistry, 38, 13453–13460.
Takasu, H., Guo, J., & Bringhurst, F. R. (1999). Dual signaling and ligand selectivity of thehuman PTH/PTHrP receptor. Journal of Bone and Mineral Research, 14, 11–20.
Thron, C. D. (1973). On the analysis of pharmacological experiments in terms of an allostericreceptor model. Molecular Pharmacology, 9, 1–9.
Verheijen, M. H., & Defize, L. H. (1995). Parathyroid hormone inhibits mitogen-activatedprotein kinase activation in osteosarcoma cells via a protein kinase A-dependent path-way. Endocrinology, 136, 3331–3337.
Verheijen, M. H., Karperien, M., Chung, U., vanWijuen, M., Heystek, H., Hendriks, J. A.,et al. (1999). Parathyroid hormone-related peptide (PTHrP) induces parietal endodermformation exclusively via the type I PTH/PTHrP receptor. Mechanisms of Development,81, 151–161.
Whitfield, J. F., & Morley, P. (1995). Small bone-building fragments of parathyroid hor-mone: New therapeutic agents for osteoporosis. Trends in Pharmacological Sciences, 16,382–386.
Zhu, X., Gilbert, S., Birnbaumer, M., & Birnbaumer, L. (1994). Dual signaling potential iscommon among Gs-coupled receptors and dependent on receptor density. MolecularPharmacology, 46, 460–469.
Zimmerman, B., Simaan, M., Lee, M.-H., Luttrell, L. M., & Laporte, S. A. (2009). c-Src-mediated phosphorylation of AP-2 reveals a general mechanism for receptors internal-izing through the clathrin pathway. Cellular Signalling, 21, 103–110.