the identification of novel proteins that interact with ... · “the identification of novel...
TRANSCRIPT
The Identification of Novel Proteins that Interact with the GLP-1 Receptor and Restrain its Activity
By
Xinyi Huang (Chinyee)
A thesis submitted in conformity with the requirements for the degree of
Master of Science
Department of Physiology
University of Toronto
© Copyright by Chinyee Xinyi Huang (2013)
(Data are currently in press at Molecular Endocrinology, Huang et al, 2013158)
II
“The Identification of Novel Proteins that Interact with the GLP-1 Receptor and Restrain its Activity”
Xinyi Huang (Chinyee), Masters of Science
Department of Physiology University of Toronto
2013
Abstract
G-protein coupled receptors (GPCRs) have been shown to interact with an array of
accessory proteins that modulate their function. I hypothesize that the GLP-1R, a B-class
GPCR, similarly has interacting proteins that regulate its signaling. An unliganded human
GLP-1R was screened using a membrane-based split ubiquitin yeast two-hybrid (MYTH)
assay and a human fetal brain cDNA prey library to reveal 38 novel interactor proteins.
These interactions were confirmed by co-immunoprecipitation and immunofluorescence.
When co-expressed with the GLP-1R in cell lines, 15 interactors significantly attenuated
GLP-1-induced cAMP accumulation. Interestingly, SiRNA-mediated knock down of
three selected novel interactors, SLC15A4, APLP1 and AP2M1, significantly enhanced
GLP-1-stimulated insulin secretion from the MIN6 beta cells. In conclusion, this present
work generated a novel GLP-1R-protein interactome, identifying several interactors that
suppress GLP-1R signaling; and the inhibition of these interactors may serve as a novel
strategy to enhance GLP-1R activity.
III
Acknowledgements
I would like to express the deepest appreciation to my supervisor Dr. Michael Wheeler,
who provided endless support on multiple levels throughout my MSc. His knowledge and
spirit in his work has inspired me throughout the development of this thesis; and his
guidance on the scientific and personal levels has helped me tremendously.
I would also like to express my heartfelt gratitude to Dr. Feihan Dai for all her support
and encouragements. Her mentorship has guided me not only through this project but also
my personal life.
I would like to express my sincere thanks to Greg Gaisano for initiating this project with
the MYTH screening. I would also like to thank Dr. Ming Zhang and Dr. Junfeng Han for
contributing to this project. Moreover, I would like to thank my respective laboratory
members, Dr. Emma Allister, Dr. Alpana Bhattacharjee, Dr. Alexandre Hardy, Dr. Ying
Liu, Lemieux Luu, Kacey Prentice for their generous help and intellectual support.
Finally, I would like to express my deepest thanks my family and friends, whose endless
support have made everything possible.
IV
Table of Contents
Abstract of Thesis ………………………………………………………..…..………... II
Acknowledgements …………………………………………..………..……………… III
Table of Contents …………………………………………………...…….……………IV
List of Tables ……………………………………………………....………………… VII
List of Figures ………………………………………………………………………... VII
Chapter1: General Introduction
1.1 Glucagon-like Peptide-1 (GLP-1)
1.1.1 Incretin Hormon ..……………………………………………….....……. 1 1.1.2 Proglucagon Expression and Regulation ………………………………... 2 1.1.3 Tissue-specific Proglucagon Posttranslational Products ………..…....….. 4 1.1.4 GLP-1 Secretion ………..……………………………....….…………...... 6 1.1.5 GLP-1 Metabolism and Clearance ……………………………….…........ 8 1.1.6 Biological Actions of GLP-1
1.1.6.1 Pancreas ……………………………….……………………..….. 9 1.1.6.2 Extrapancreatic Tissues: Nervous, Gastrointestinal and
Cardiovascular systems ……………………….…………..….… 10
1.2 Glucagon-like Peptide-1 Receptor (GLP-1R)
1.2.1 Structure and Activation ………...…………………….……...………... 13 1.2.2 Expression, Signal Transduction and Regulation ………...……………. 14
1.3 GLP-1 and GLP-1R as Therapeutic Targets for Type 2 Diabetes (T2D)
1.3.1 Therapeutic Potential of GLP-1/GLP-1R ………...………….....……… 17 1.3.2 GLP-1R Agonists: Exentatide and Liraglutide ………...………………. 18 1.3.3 DPP-4 Inhibitors ………...……………………………………....……... 19 1.3.4 Limitations of the GLP-1-based Therapies ………...…...………...……. 19
1.4 GPCR Interactomes
1.4.1 GPCR Interactomes Regulate Receptor Signaling ………...…….……... 20 1.4.2 Known GLP-1R Interacting Proteins ………...……………..……..….... 22
V
Rational and Hypothesis of Thesis ………...…………………...…………...……….. 23
Chapter 2: Identification of GLP-1R interacting proteins that restrain the receptor
activity
2.1 Introduction ………...………………………………………………………………. 25
2.2 Material and Methods
2.2.1 Construction of yeast and mammalian cell expression GLP-1R plasmids ….. 28 2.2.2 Cell culture and Transfection ………...……………………………………… 29 2.2.3 MYTH Screening ………...………………………………………………….. 29 2.2.4 Immunocytochemistry ………...…………………………………………….. 30 2.2.5 Co-immunoprecipitation and immunoblotting ………...……………………. 31 2.2.6 GLP-1 or Forskolin Stimulated cAMP Formation ………...………………... 32 2.2.7 Insulin Secretion Assay, Reverse Transcription PCR and Quantitative PCR . 32 2.2.8 Statistics ………...…………………………………………………………… 33
2.3 Results
2.3.1 Identification of GLP-1R interacting proteins by MYTH …………………... 33 2.3.2 Validation of the interaction between GLP-1R and interactors in mammalian
cells 2.3.2.1 Co-immunoprecipitation ………………….…………………...……. 39 2.3.2.2 Immunofluorescence ………………………………………………... 40
2.3.3 Interactor proteins reduced the GLP-1R function 2.3.3.1 GLP-1R interactors modulate GLP-1-stimulated cAMP accumulation in
CHO cells …………………………………………………………… 46 2.3.3.2 Overexpression of selected GLP-1R interactors showed no significant
effect on GLP-1 stimulated insulin secretion in MIN6 beta cells ....... 49 2.3.3.3 Reducing expression of selected GLP-1R interactors enhances GLP-1-
stimulated insulin secretion in MIN6 beta cells …………………….. 50
Chapter 3: General Discussion
3.1 Summary of Findings ………...…………………………………………………… 54
3.2 MYTH Identified GLP-1R Interactome ………...………………………………… 54
3.3 Limitations of MYTH ………...…………………………………………………… 55
3.4 Validations of the MYTH Identified Interactors ………...………………………… 57
VI
3.5 MYTH Identified GLP-1R Interaction Proteins Restrained the Receptor Activity .. 60
3.6 Potential Functional Relevance of the Interactor Candidates
3.6.1 SLC15A4 ………...…………………………………………………………. 61 3.6.2 APLP1 ………...………………………………………………………….…. 61
3.7 Protein Interactors as Allosteric Regulators of the GLP-1R and Their Therapeutic
Potential ………...…………………………………………………………………………….. 62
3.8 Alternative Proteomic Screening Strategies – AP-MS ………...…………………... 63
3.9 Alternative Proteomic Screening Strategies – FRET/BRET ………...…………….. 64
Chapter 4: Future Directions and Conclusion
4.1 Future Directions
4.1.1 Access the effect of interactors on MIN6 apoptosis and proliferation ……… 66 4.1.2 Validate PPI using Bioluminescence/Fluorescence Resonance Energy Transfer
Strategies (BRET/FRET) …………………………………………………… 66 4.1.3 Access the functional effects of GLP-1R interactors in mouse
or human islets ………...………………………………………………….…. 66 4.1.4 Access the mechanisms by which the inhibitory effects of SLC15A4, APLP1
and AP2M1 were mediated …………………………………………………. 67 4.1.5 Construct and functionally characterize an islet expressing GLP-1R interactome
……………………………………………………………………………….. 67
4.2 Conclusion …………………………………………………………………………. 68
Abbreviation Footnote …………………………………………………………..…… 70
Reference ……………………………………………………………………………… 73
VII
List of Tables
Table 1. List of MYTH Identified GLP-1R Interactors ………………………………... 38 (Contributions: Chinyee Huang 50%, Greg Gaisano 50%) Table2. Primer Sequences for RT-PCR and qPCR …………………………………..… 69 (Contributions: Chinyee Huang 100%) List of Figures
Figure 1. Structure of Proglucagon Gene, mRNA and Protein ………………………….. 3 (Contributions: Chinyee Huang 100%) Figure 2. Regulation of GLP-1 Secretion by Ingested Nutrients …………………………7 (Contributions: Chinyee Huang 100%) Figure 3. Diverse Actions of GLP-1 on Peripheral Tissues ……………………………. 12 (Contributions: Chinyee Huang 100%) Figure 4. Activation Mechanism of the GLP-1R ………………………………………. 14 (Contributions: Chinyee Huang 100%) Figure 5. GLP-1 Signal Transduction Pathways in the Pancreatic β Cell …………...… 16 (Contributions: Chinyee Huang 100%) Figure 6. Examples of GPCR Binding Proteins ……………………………………..… 22 (Contributions: Chinyee Huang 100%) Figure 7. Split-ubiquitin Based Membrane Yeast Two-hybrid (MYTH) System ……... 28 (Contributions: Chinyee Huang 100%) Figure 8. MYTH Screening for the hGLP-1R …………………………………..…..… 36 (Contributions: Greg Gaisano 100%) Figure 9. GLP-1R Interactome Identified by MYTH …………………………………. 37 (Contributions: Chinyee Huang 100%) Figure 10. Validation of the Interaction Between V5-His-tagged hGLP-1R and the MYTH Identified Interactors in RC2-CHO Cells. …………………………………….. 41 (Contributions: Chinyee Huang 70%, Ming Zhang 30%) Figure 11. Validation of the Interaction Between Flag-tagged hGLP-1R and the MYTH Identified Interactors in CHO cells …………………………………………………….. 43 (Contributions: Chinyee Huang 100%)
VIII
Figure 12. Colocalization Analysis of the GLP-1R and Selected Interactors by Immunofluorescence Imaging ……………………………………………………...….. 44 (Contributions: Chinyee Huang 100%) Figure 13. The Effect of Interactors on GLP-1-induced cAMP Accumulation ……….. 47 (Contributions: Chinyee Huang 100%) Figure 14. Elimination of interactors that have an intrinsic suppressive effect on cAMP production ………………………………………………………………………….…... 48 (Contributions: Chinyee Huang 100%) Figure 15. Validation of GLP-1R and Interaction Protein Expressions in MIN6 Cells .. 49 (Contributions: Junfeng Han 70%, Chinyee Huang 30%) Figure 16. Overexpression of Selected Interaction Proteins Did Not Affect GLP-1-induced Insulin Secretion in MIN6 Cells …………………………………………...….. 50 (Contributions: Chinyee Huang 100%) Figure 17. Knocking Down Selected Interaction Proteins Enhanced GLP-1R Function in MIN6 Cells ………………………………………………………………………....….. 52 (Contributions: Chinyee Huang 100%) Fig 18. Commercially available antibodies are not specific …………………...…...….. 59 (Contributions: Chinyee Huang 100%)
1
Chapter 1. General Introduction
1.1 Glucagon-like Peptide-1 (GLP-1)
1.1.1 Incretin Hormones
In the early 1900s, it was proposed that the gut releases factors in response to nutrient
ingestion, which stimulate the pancreas to secrete a substance capable of reducing blood
glucose1; this pancreatic substance was later identified as insulin2. Insulin is a crucial
hormone that regulates glucose homeostasis. When blood glucose levels rise, insulin is
produced and secreted from the pancreatic beta cells. It acts to lower blood glucose levels
by increasing glucose uptake by the liver, muscle and adipose tissue and decreasing
hepatic gluconeogenesis. The gut factors, on the other hand, are now known as
incretins3,4. Incretins are a group of hormones released from the intestinal mucosa
following food intake. They act to augment glucose-stimulated insulin secretion (GSIS)
from the pancreatic beta cells, and thus aid in maintaining blood glucose levels. This
nutrient-mediated gut response, or “incretin effect”, was confirmed with the development
of radioimmunoassays, which demonstrated that oral glucose ingestion induced a more
robust increase in plasma insulin levels as compared to intravenous administration of the
same amount of glucose5,6. In fact, the incretin effect accounts for at least 50% of the
total insulin secreted after nutrient administration.
Two peptide hormones, glucose-dependent insulinotropic polypeptide (GIP) and
glucagon-like peptide 1 (GLP-1) fulfill the definition of an incretin in humans7. GIP was
discovered in the crude extract of porcine small intestine8. It is mainly expressed in the
stomach and intestinal K-cells, where it is processed and released into the circulation in
2
response to nutrient ingestion. GLP-1, on the other hand, is encoded by the proglucagon
gene9. It is primarily synthesized and secreted by the L-cells of distal ileum and colon.
GIP and GLP-1 have been shown to potentiate GSIS in an additive manner, and together
they account for the full incretin effect7. The importance of GLP-1 in maintaining glucose
homeostasis was demonstrated by a disruption of the murine GLP-1 receptor, which
resulted in mild fasting hyperglycemia and impaired glucose tolerance associated with
defective insulin secretion following both oral and intraperitoneal glucose challenge7.
1.1.2 Proglucagon Expression and Regulation
The proglucagon gene comprises 6 exons and 5 introns9 (Fig. 1). It encodes a single
180 amino acid precursor protein, which can be processed into several products including
glicentin-related polypeptide (GRPP) contained in exon 2, glucagon and oxyntomodulin
(OXM) in exon 3, and GLP-1 in exon4 and GLP-2 in exon 5. Proglucagon mRNA is
expressed in three types of cells in humans: alpha cells of the pancreatic islet, L-cells of
the intestine and neurons located in the caudal brainstem and the hypothalamus10,11. At
least 6 homeodomain proteins have been identified to interact with the proglucagon
promoter and regulate its activation13. While deletion of some of these transcription
factors (Cdx-2, Brn-4 and Nkx6.2) showed no apparent effect on proglucagon production
in vivo, suggesting functional redundancy and compensation, expression of the others
(Pax-6 and Pdx-1) are proved to be essential for the proper development of proglucagon-
producing cells13. Mice with a dominant-negative Pax-6 mutation, for instance, exhibits
disrupted islet development and lacked enteroendocrine cells14. Tissue-specificity has
also been demonstrated at the level of transcriptional regulation of the proglucagon gene.
For example, Wnt signaling-mediated activation of beta-catenin stimulates proglucagon
3
expression in the intestine but not in the islet cells15. In addition, DNA sequences within
the proglucagon promoter also contribute to tissue specificity. A combination of cell line
and transgenic mouse studies have demonstrated that approximately 1.6 kb of the human
proglucagon gene 5’-flanking sequences can direct proglucagon transcription to the brain
and intestine, whereas additional upstream sequences extending to the first 6kb are
required for expression in the pancreatic cells16. Interestingly, within this 6kb region, one
conserved intron sequence, designated ECR3, has also been identified to regulate
proglucagon expression specifically in the pancreatic alpha cells.
1.1.3 Tissue-specific Proglucagon Postranslational Products
Proglucagon yields tissue-specific products in the brain, intestine and pancreas. The
proglucagon mRNA is identically expressed in these tissues10,17; thus, its specificity is
mainly achieved through posttranslational modifications by prohormone convertase (PC)
enzymes PC1/3 and PC218 (Fig. 1). Studies have shown that PC1/3 is both necessary and
sufficient to complete the proglucagon processing in L-cells to produce glicentin,
oxyntomodulin, intervening peptide-2, GLP-1 and GLP-2 19,20,21. PC1/3 null mice exhibit
Figure 1. Structures of (A) the proglucagon gene, (B) mRNA, and (C) protein. (D) Tissue-specific posttransla- tional processing of proglucagon in the pancreas leads to the generation of Glicentin-related polypeptide (GRPP), glucagon (GLUC), intervening peptide-1 (IP-1), and major proglucagon fragment (MPGF), whereas glicentin, oxyn- tomodulin (OXM), intervening peptide-2 (IP-2), and GLP-1 and GLP-2 are liberated after proglucagon processing in the intestine and brain.
4
increased levels of intestinal proglucagon accompanied by decreased levels of its
cleavage products. PC enzymes responsible for the posttranslational processing of
proglucagon in the central nervous system (CNS) are not well established. However, high
levels of PC1/3 and PC2 are present throughout the CNS, suggesting a similar processing
mechanism to the intestine7.
Glicentin is the major intestinal peptide that corresponds to proglucagon residues
number 1-69 (PG 1-69). Although its exact function is not well defined, studies have
shown that it exerts trophic effects on rodent intestinal mucosa, slows down gastric
emptying and inhibits extraintestinal invasion of enteric bacteria22-24. Oxyntomodulin
(OXM), a cleaved product of Glicentine (PG 33-69), is a dual agonist of the GLP-1 and
glucagon receptors25. It promotes satiety, inhibits gastrointestinal secretion and motility,
and stimulates pancreatic enzyme secretion and intestinal glucose uptake23,26.
Interestingly, injections of OXM in humans caused a significant reduction in weight and
appetite as well as increased energy expenditure27.
The two glucagon-like peptides, GLP-1 and GLP-2, are contained in a stretch of
proglucagon residues known as the major proglucagon fragment (MPGF, PG 72-160)28.
These two peptides are flanked by pairs of basic amino residues, which serve as canonical
cleavages sites for PC enzymes. The full length GLP-1 (PG 72-108) is biologically
inactive. Further truncation is required to liberate the active form, GLP-1 7-37 (PG 78-
108), which acts as a potent stimulator of glucose-stimulated insulin secretion29.
Furthermore, it was found that the last residue of full length GLP-1 (PG 108), Gly, serves
as a substrate for amidation by the action of peptidylglycine alpha-amidating
monooxygenase to generate GLP-1 (7-36)NH230. GLP-1 (7-36)NH2 is equipotent to GLP-
5
1 7-37 and has similar metabolism31. It exhibits slightly improved stability towards
degradation and is the predominant form in humans32,33. The sequence of naturally
occurring GLP-2 corresponds exclusively to PG 126-158 with no sign of further
posttranslation modifications34. GLP-2 has a large number of target tissues including the
gastrointestinal tract and the CNS. It stimulates cell proliferation and inhibits apoptosis in
the intestinal crypt compartment35. It also up-regulates intestinal glucose transport,
improves intestinal barrier function, and inhibits food intake, gastric emptying, and acid
secretion36,37. In the CNS, GLP-2 improves neuronal proliferation and survival38.
In pancreatic alpha-cells, the proglucagon precursor is mainly processed into glicentin-
related polypeptide (GRPP), glucagon, intervening peptide-1, and MPGF19. Glucagon is a
hormone that counteracts insulin. It is crucial for maintaining glucose homeostasis in the
fasting state by increasing hepatic glucose production via activations of glycogenolysis
and gluconeogenesis39. GRPP, intervening peptide-1 and MPGF do not have a known
function to date. PC2 is at least partly responsible for the alpha-cell proglucagon
processing. Mice that lack active PC2 exhibit deficient proglucagon processing and
develop phenotypes that are similar to mice that have a glucagon receptor deletion: mild
hypoglycemia, improved glucose tolerance and alpha cell hyperplasia40.
1.1.4 GLP-1 Secretion
Although GLP-1 producing enteroendocrine cells can be found throughout the entire
small intestine in pigs, rats and humans, its main synthesizer, the L-cells, are located
primarily in the distal ileum and colon12. L-cells are open-type entero-epithelial endocrine
cells that have two surfaces: apical and basolateral41. The apical surface faces the
6
intestinal lumen where it is in direct contact with the luminal nutrients, whereas the
basolateral surface faces the neural and vascular tissue. Thus, a variety of nutrient, neural
and endocrines factors are involved in regulating GLP-1 secretion from the L-cells. Meal
ingestion is the primary physiological stimulus for GLP-1 secretion, which occurs in a
biphasic manner in rodents and humans42. The first phase of secretion occurs rapidly after
nutrient ingestion (within 15-30 min) and is followed by a prolonged second phase at 90-
120min post-ingestion. Since most L-cells are located in the distal small intestine, it is
unlikely that the early phase of GLP-1 secretion is mediated by direct nutrient contact. In
fact, placement of glucose or fat in the proximal duodenum could induce an immediate
stimulation of L-cells that resembles the early phase of GLP-1 secretion43, suggesting a
cross-talk between the proximal and distal small intestine (Fig. 2). The autonomic
nervous system has been proposed to mediate this cross-talk as direct activation of the
vagus nerve increased fat-induced GLP-1 secretion in rats, whereas bilateral
subdiaphragmatic vagotomy completely blocked the secretion44. Both M1 and M2
muscarinic-receptors have been demonstrated to play crucial roles in this process since
muscarinic-receptor agonists enhanced GLP-1 release while antagonists inhibited it45.
Acetylcholine and gastrin-releasing peptide have been identified as the key
neurotransmitters that mediate this proximal-distal loop45,46 47. In the canines and rodents,
the incretin hormone GIP secreted from proximal K-cells also contributes to this early
phase of GLP-1 secretion 43 (Fig.2).
7
In contrast to the indirect mechanisms that mediate first phase GLP-1 release, the
second phase is believed to occur in response to direct interaction of the distal L-cells
with luminal nutrients as they move down the small intestine (Fig. 2B). Indeed,
placement of nutrients into the distal ileum stimulates GLP-1 release43. Fat has been
proposed to be the more important physiological regulator for GLP-1 secretion as
compared to glucose since glucose does not reach distal gut in high concentration. In
addition, intestinal somatostatin, neurotransmitter GABA and alpha- and beta-adrenergic
agonists, leptin and insulin hormones, have also been demonstrated to modulate GLP-1
secretion through positive and negative feedback loops12. It is of interest that insulin
resistance in the L-cells decreases both basal and stimulated GLP-1 release, which may in
part account for the reduced GLP-1 secretion observed in obesity and Type 2 Diabetes
(T2D) patients.
Figure 2. Regulation of GLP-1 secretion by ingested nutrients. A. After a meal, nutrients in the duodenum activate a proximal-distal neuroendocrine loop, which stimulates the first phase GLP-1 secretion from L-cells in the ileum and colon. In rodents, GIP released from K-cells activates vagal afferents, which subsequently causes GLP-1 secretion through vagal efferents and enteric neurons that release acetylcholine (Ach) and gastrin-releasing peptide (GRP). B. Movement of nutrients toward more distal sections of the intestine leads to the direct interaction of nutrients with L-cells, which stimulates the second phase GLP-1 secretion.
8
1.1.5 GLP-1 Metabolism and Clearance
As mentioned previously, GLP-1 presents in multiple forms in vivo. The active forms,
GLP-1 7-37 and GLP-1 (7-36)NH2, are generated from their full-length precursors by the
actions of PC1/348. Bioactive GLP-1 contains a penultimate alanine residue in position 2,
which is a target site for the enzyme dipeptidyl peptidase-4 (DPP-4) 49. DPP-4 is a serine
protease that specifically cleaves dipeptides from the amino terminus of proteins, thereby
inactivating them. It is widely expressed in multiple tissues and can be found in brush
border enterocytes lining the small intestine, as well as endothelial cells positioned
directly adjacent to the GLP-1 producing sites. Owing to the rapid actions of DPP-4,
bioactive GLP-1 has an extremely short half-life of less than 2 minutes in the circulation.
In fact, ex vivo studies have shown that majority of GLP-1 secreted by the L-cells is
already degraded into inactive metabolites before even leaving the gut50. Only about 25%
of the newly secreted GLP-1 reaches the systemic circulation in the intact, active form.
Conversely, the half-life of GLP-1 could be prolonged by inhibiting DPP-4 in both
animals and humans50. Other than DPP-4, neutral endopeptidase 24.11, a membrane-
bound zinc metallopeptidase, has also been demonstrated to degrade GLP-1 via c-
terminal cleavage51. Inactive GLP-1 metabolites are subsequently cleared by the kidney
with a half-life of 4-5 minutes52. Fasting plasma levels of bioactive GLP-1 range between
5 and 10pmol/L in humans and upon feeding, the levels increase approximately 2 to 3
fold depending on the size and nutrient composition of the meal. Postprandial levels of
GLP-1 are reduced with obesity and the development of Type 2 Diabetes (T2D) 53.
9
1.1.6 Biological Actions of GLP-1
1.1.6.1 Pancreas
GLP-1 controls important and diverse physiological functions in multiple organs
including the pancreas, brain, gastrointestinal and cardiovascular systems7 (Fig. 3). Its
actions are most extensively studied in the pancreatic beta cells, where it potentiates
glucose-stimulated insulin secretion (GSIS), improves glucose sensing, insulin gene
transcription and biosysnthesis, and increases cell survival and proliferation. The primary
effector of GLP-1-enhanced GSIS is cAMP. Upon ligand binding to the GLP-1 receptor
(GLP-1R), cAMP accumulates to activate two distinct pathways: PKA-dependent
phosphorylation and PKA-independent activation of Epac25,6,29. Through these two
pathways, GLP-1 acutely depolarizes the beta cells by increasing ATP synthesis in the
mitochondria and promoting closure of the KATP channels. It also prevents repolarization
of the plasma membrane by closing the Kv channels. As a result, voltage-gated Ca2+
channels are activated, elevating intracellular Ca2+ levels to subsequently increase insulin
exocytosis. The glucose-dependency of this GLP-1-induced insulin secretion is not
completely understood; however, studies have suggested that KATP and Kv channels may
be involved7.
Besides enhancing GSIS, GLP-1 has many other beneficial effects in the beta cells: 1)
it increases insulin gene transcription and biosynthesis through cAMP-induced PKA-
dependent and independent signaling pathways; 2) it restores glucose sensitivity to
previously glucose-resistant beta cells via up-regulation of the glucose transporters and
glucokinases; 3) it attenuates ER stress by reducing eIF2α levels; and 4) it preserves beta
10
cell mass by either directly activating the GLP-1R to promote beta cell proliferation and
inhibit apoptosis or indirectly through the stimulations of insulin, which improves the
metabolic environment for beta cell survival 54.55,56. These actions of GLP-1 are mediated
by multiple signal transduction mechanisms including transactivation of the epidermal
growth factor receptor (EGFR), activation of the cAMP/PKA, PI-3K/PKC/Akt and
MAPK pathways, and up-regulation of insulin receptor substrate 2 expressions causing
the subsequent FoxO1 nuclear exclusion57,58.
GLP-1 also lowers blood glucose by inhibiting glucagon secretion from the alpha
cells. This effect may be direct via GLP-1Rs expressed on alpha cells or indirect via
stimulation of insulin and somatostatin secretion59,60. Although the exact mechanism
whereby GLP-1 inhibits glucagon is unclear, it is known to be glucose-dependent61. In
other words, GLP-1-induced glucagonostatic actions will be relieved when blood glucose
falls to the normal or hypoglycemic range.
1.1.6.2 Extrapancreatic Tissues: Nervous, Gastrointestinal (GI) and
Cardiovascular Systems
GLP-1R signaling in the CNS exerts important regulatory effects on feeding
behaviour, gastric motility, blood glucose levels and cardiovascular functions7 (Fig. 3).
There are three sources of central GLP-1: GLP-1 produced within the CNS, peripheral
GLP-1 that diffused across the blood-brain barrier and peripheral GLP-1 that
communicates with the CNS via sensory afferent vagal neurons62. Rodent studies have
shown that central or peripheral administrations of GLP-1R agonists, Exendin-4, reduced
short-term food and water intake, and consequently decreased body weight63-65. These
11
GLP-1-induced anorectic effects also apply to human subjects7. They are likely to be
mediated by direct activation of the GLP-1Rs localized at the hypothalamic CNS and
vagal afferent nerve fibers. Surgical transection of the brainstem-hypothalamic pathway
and bilateral subdiaphragmatic vagotomy completely blocked the GLP-1-induced
anorexia in rat66. Ablation of the vagus nerve by capsaicin similarly abolished the
anorectic effect of peripherally administered exendin-4 in mice confirming a central role
for GLP-1 in affecting food intake67. In addition to inhibiting food intake, GLP-1R
signaling in the murine CNS has also been implicated to modulate whole-body glucose
homeostasis by promoting insulin secretion, inhibiting muscle glucose usage and
supporting enhanced hepatic glycogen storage under hyperglycemic conditions68.
Furthermore, central infusion of GLP-1R agonists reduced amyloid-β peptide in vivo and
enhanced neurite and PC12 cell differentiation and outgrowth in vitro, indicating that
GLP-1 exerts protective actions on neuronal cells69. Finally, GLP-1 has been shown to be
important for learning and memory since GLP-1R-/- mice showed impaired learning
behaviour, which was corrected by hippocampal Glp1r gene transfer.
GLP-1R agonists inhibit meal-induced gastric acid secretion and gastric emptying in
the gastrointestinal (GI) system. Attenuation of gastric emptying is important for
normalizing postprandial blood glucose elevation especially in T2D patients, as the
transit of nutrients from stomach to small intestine is decelerated70. When the GLP-1-
dependent inhibition of gastric emptying is antagonized by enrythromycin in humans, its
glucose-lowering effects are also reduced71. Mechanisms whereby GLP-1R agonists exert
their GI effects are highly complex. Rodent studies have suggested that GLP-1 signaling
in the nervous system may play a role since vagal afferent denervation abolished the
12
GLP-1-induced attenuation of gastric emptying and acid secretion7. Direct mechanisms
may also be involved since GLP-1R expression has been reported in the parietal cells of
the stomach.
GLP-1 has been demonstrated to exhibit protective effects in the heart72. Intravenous
administration of GLP-1R agonists increased systolic and diastolic blood pressure and
heart rate in rodents. Although these effects are not observed in humans, a study in T2D
patients revealed that GLP-1 infusion improved endothelial function73. Interestingly,
GLP-1 treatment has been demonstrated to increase cardiac output, reduce left ventricular
end diastolic pressure, and improve myocardial insulin sensitivity and glucose uptake in
dog models of cardiomyopathy74. It also reduces infarct size and improves overall
performance of isolated rat heart and animal models of myocardial ischemia75,76. GLP-
1Rs expressed in the heart and CNS have been suggested to mediate these cardiovascular
effects through the activation of cAMP, PI-3K/Akt and p44/42 MAPK75.
Figure 3. Diverse actions of GLP-1.
13
1.2 Glucagon-like Peptide-1 Receptor (GLP-1R)
1.2.1 Structure and Activation
The GLP-1R is a 463-amino-acid, 7-transmembrane-spanning protein that belongs to
the B-class of G-protein coupled receptors (GPCRs)77. This class of GPCRs also includes
the receptors for glucagon, GLP-2 and GIP. The GLP-1R possesses a large hydrophilic,
extracellular N-terminus containing one α-helical region and five β–strands, which form
two antiparallel β–sheets and six cysteine residues allowing for disulfide interactions78.
These structures give GLP-1R the classic ‘short consensus repeats’ conserved in the B-
family GPCRs that support the N-terminal stability. The large GLP-1R N-terminus is
often described as an “affinity trap” that recognizes and binds peptide ligands, as
supported by X-ray crystal structures79. The binding specifically adopts a widely accepted
two-domain model, in which the C-terminus of the ligand interacts with the N-terminus
of the receptor for ligand recognition, and the N-terminus of the peptide associates with
the receptor core for conformational rearrangement of the receptor structure78 (Fig. 4).
This conformational change elicits a shift in the receptor intracellular loops, which is
important for the subsequent G-protein recruitment and signal transduction activity. The
two-domain activation model is experimentally supported by studies using chimeric
receptors, mutagenesis analysis and photolabile peptide cross-linking80. The intracellular
C-terminus of GLP-1R, on the other hand, contains three pairs of serine residues that are
phosphorylation sites important for agonist-induced homologous and heterologous
receptor desensitization and internalization.
14
1.2.2 Expression, Signal Transduction and Regulation
The GLP-1R is expressed in a wide range of tissues in rodents and humans including
alpha, beta and delta cells of the pancreas, the lung, heart, kidney, stomach, intestine,
pituitary, skin, ganglion neurons of the vagus nerve, and the CNS7. It is a pleiotropically
coupled receptor capable of signaling through multiple G-proteins: Gαs, Gαq, Gαi, and
Gαo81. Among them, Gαs coupling is the most documented. Upon ligand binding, the
GLP-1R undergoes conformational changes that recruit the G-protein complex and
catalyze the release of GDP from the Gαs (Fig. 4). Gαs subsequently binds GTP and
dissociates from the Gβγ. Liberated Gαs-GTP activates membrane-bound adenyl cyclase,
which in turn catalyzes the conversion of ATP into cAMP. cAMP is the main effector of
GLP-1R signaling80 (Fig. 5) and it turns on both PKA and Epac2 pathways that are
responsible for a number of GLP-1 actions. Other than cAMP, the ligand-bound GLP-1R
Figure 4. Activation mechanism of the GLP-1 receptor. Biochemical and structural studies have led to a model of class B1 GPCR activation by peptide hormones referred to as the “two-step” mechanism. A. In the unliganded state, the GLP-1 receptor (GLP-1R) is in a predominately inactive conformation. The natural ligand GLP-1(7-36)-NH2 is freely diffusible in solution and likely has substantial intrinsic α-helical structure. B. An initial binding event between the globular ectodomain at the N-terminal of the GLP-1R and the C- terminal of the GLP-1(7-36)-NH2 peptide occurs. This “low affinity” interaction acts as a tether or “affinity trap” to localize GLP-1 at the GLP-1R. C. The weak affinity of the N-terminus of GLP-1(7-36)-NH2 is then able to productively engage with transmembrane domain and loop residues of the receptor to induce a high affinity interaction and likely a conformational change in the GLP-1R. D. The conformational change of GLP-1R recruits the G-protein heterotrimer, initiating downstream signaling cascade. E. The GLP-1R stimulates guanine nucleotide exchange on the α-subunit of the G-protein, leading to G-protein dissociation and independent or synergistic activation of effector proteins by liberated Gα·GTP and Gβγ.
15
also activates phospatidylinositol-3 kinase (PI-3K) through transactivating the epidermal
growth factor receptor (EGFR)82 (Fig. 5). PI-3K in turn stimulates PKC ζ and Akt/PKB
signaling pathways that are responsible for suppressing Kv currents and inducing beta-
cell proliferation. In addition, reports have demonstrated that GLP-1R activation could
mobilize intracellular Ca2+ through phospholipase C81. The effect of GLP-1 on Ca2+
mobilization is commonly assumed as a result of Gαq coupling. However, recent studies
have shown that GLP-1R activation could induce Ca2+ mobilization in GLP-1R-
expressing HEK cells without activating Gαq83. Thus, the role of alternative Gα subunits
in mediating GLP-1R signaling remains inconclusive. Moreover, GLP-1R activation has
been associated with signaling pathways that involve ribosomal S6 kinases and mitogen-
and stress-activated protein kinase family of CREB kinases, which at least in part mediate
the enhancement of insulin gene transcription via basic region-leucine transcription
factors81; as well as pathways that involve β-arrestin-1 and pERK1/2, which aid in
promoting beta cell survival. It is apparent that GLP-1R signaling differs according to the
cellular context of its expression78. Although this phenomenon is common among all
GPCRs, the underlying mechanism is unclear.
The GLP-1R undergoes rapid and reversible homologous and heterologous
desensitization in islet cell lines upon ligand stimulation84. Although the exact
mechanisms remain elusive, phosphorylation of specific residues in the GLP-1R C-
terminus is clearly required. Recruitment of caveolin-1, G protein-coupled receptor
kinase and β-arrestin has been suggested to mediate this GLP-1R desensitization even
though it is controversial. Recently, computer simulations have suggested that the GLP-
1R desensitization undergoes two sequential states85: a first ligand-bound state followed
16
by a second transition state. Although GLP-1R desensitization is well recognized in vitro,
it has not been observed in vivo adding controversies to the field84.
Figure 5. GLP-1 signal transduction pathways in the pancreatic β cell. Signaling in pancreatic β-cells via the classical GLP-1R-coupled Gαs pathway mediates increases in cAMP to up-regulate PKA and Epac2 (exchange protein activated by cAMP). These pathways increases intracellular Ca2+ concentrations through 1) enhancing intracellular Ca2+ mobilization, 2) inhibition of K (potassium) channels and 3) acceleration of Ca2+ influx through VDCCs (voltage-dependent Ca2+ channels). Together, they lead to increased insulin biosynthesis and secretion (green). Activation of proto-oncogene tyrosine kinase src (c-src) and subsequent transactivation of epidermal growth factor receptor (EGFR) aid in increasing PI3K, IRS-2 (insulin receptor substrate 2) and PKB (Akt) to enhance β-cell proliferation (orange). This is also facilitated in part by PKA-mediated increases in MAPKs (mitogen-activated protein kinases), β-catenin and cyclin-D1. Inhibition of caspases, FoxO1 (forkhead box protein O1) and NF-κB (nuclear factor κB), in addition to regulation of CREB (cAMP-response-element-binding protein) and subsequent inhibition of Bcl-2-associated death promoter (BAD), aid in the inhibition of apoptosis (blue), a process also mediated by β-arrestin-1 and the pERK1/2 (phosphorylation of ERK1/2, as well as the IGF-1R/IGF-2 (insulin-like growth factor-1 receptor/ insulin-like growth factor2) autocrine loop. ER (endoplasmic reticulum) stress reduction (black) involves the up-regulation of multiple transcription factors, including ATF-4 (activating transcription factor-4), CHOP [C/EBP (CCAAT/enhancer-binding protein)-homologous protein], which inhibits the dephosphorylation of eIF2α (eukaryote initiation factor 2 α). Cross-talk exists between most pathways, including the regulation of the important promotor of insulin gene transcription, synthesis and secretion, Pdx-1 (pancreas duodenum homeobox-1) via both cAMP-dependent and IRS-dependent mechanisms.
17
1.3 GLP-1 and GLP-1R as Therapeutic Targets for Type 2 Diabetes (T2D)
1.3.1 Therapeutic Potential of GLP-1/GLP-1R
Type 2 Diabetes (T2D) is a metabolic disease marked by chronically elevated levels of
blood glucose or hyperglycemia in the context of insulin resistance and deficiency. This
disease is often associated with complications including increased risk of cardiovascular
disease, lower limb amputation, nontraumatic blindness, cognitive dysfunction and
kidney failure. In 2008/2009, almost 2.4 million Canadians are living with T2D (Diabetes
in Canada, Public Health Agency of Canada). Estimates suggest that this number will go
up to 3.7 million by 2018/2019 if incidence and mortality rates continue at levels seen in
the 2008/09 data. Studies have shown that the nutrient-stimulated insulin secretion or
incretin effect is largely reduced in T2D patients86. This impaired incretin effect is
associated with modestly but significantly diminished levels of nutrient-induced GLP-1
secretion. However, the glucoregulatory actions of GLP-1 are preserved in the subjects
with T2D. Thus, efforts have been focused on developing “replacement” therapies of
GLP-1 and analogs to treat T2D. In fact, many studies have demonstrated that continuous
administration of native GLP-1 or three times daily supplementation could significantly
reduce meal-related glycemic excursion in the context of increased plasma insulin and
reduced glucagon levels87,88. Inhibition of gastric emptying, appetite and food intake have
also been observed in subjects infused with GLP-1. Despite the remarkable anti-diabetic
pharmacology demonstrated by GLP-1 infusion, its short biological half-life presents a
major obstacle limiting its pharmaceutical potential. To combat this shortcoming,
numerous efforts have been focused on developing GLP-1 analogs with improved
metabolic properties.
18
1.3.2 GLP-1R Agonists: Exentatide and Liraglutide
To date, two degradation resistant GLP-1R agonists, Exenatide and Liraglutide, have
been approved for marketing as treatments of T2D78. These agonists demonstrate an
efficacy comparable to insulin treatment but with minimal risk of hypoglycaemia since
their actions are glucose-dependent. Exenatide is the synthetic version of GLP-1R agonist
exendin-4, a bioactive peptide extracted from the poisonous venom of Heloermatidae
lizards 89. It is fully efficacious and exerts the same physiologic glucoregulatory actions
as native GLP-190,91. However, Exenatide is a poor DPP-4 substrate and is protected from
its degradation92. Instead, it is mainly cleared through glomerular filtration in the
kidney93. Thus, Exenatide has a much longer biological half-life of approximately 4 hours
as compared to native GLP-1. Pre-clincal trials of Exenatide therapy reported
significantly reduced levels of HbA1c (a measure that correlates to blood glucose levels),
improved parameters of beta cell function, decreased fasting and postprandial glucose
concentrations and loss of weight in T2D subjects94-96. In April of 2005, Exenatide
became the first incretin-based therapeutic approved by the United States Food and Drug
Administration (FDA) for the treatment of T2D78.
Liraglutide is the second GLP-1 receptor agonist approved to treat T2D. Fatty acid
derivatization strategy, which attaches a fatty acid moiety to native GLP-1 facilitating its
binding to serum albumin, was used to prolong the action of this molecule in vivo97.
Binding to albumin sterically protects Liraglutide from DPP-4 degradation, consequently
increasing its half-life to 11-15 hours. Thus, this molecule is suitable for once-weekly
administration. Liraglutide received marketing approval in January of 2010 for the
treatment of T2D under the brand name Victoza.
19
1.3.3 DPP-4 Inhibitors
As an alternative to utilizing synthetic GLP-1R agonists as therapy, another option to
treat T2D is by attenuating endogenous incretin degradation using DPP-4 inhibitors.
Sitagliptin was the first DPP-4 inhibitor approved for clinical use in October 2006
followed by Vildagliptin in Europe and Saxagliptin in the US markets98. These agents can
be used as monotherapy or in combination with other agents. They increase plasma levels
of GLP-1 and GIP after food intake, and thus enhance pancreatic GSIS. DPP-4 inhibitors
are well tolerated, weight neutral, and are not associated with hypoglycemia99-101.
1.3.4 Limitations of the GLP-1-based Therapies
There is no doubt that GLP-1-based therapeutics are effective glucose-lowering
agents. However, several concerns have been raised regarding their side effects. For
example, mild to moderate nausea, vomiting and diarrhea are often associated with the
Exenatide therapy102. The fact that Exenatide needs to be injected twice daily also causes
a lot of pain to the patients, possibly resulting in its relatively high withdrawal rate of
~20%94. Liraglutide, on the other hand, is suitable for once-weekly dosing. However,
gastrointestinal side effects are also common. In addition, Liraglutide treatment may give
rise to welling or lump in the throat area, weakness, confusion, increased thirst and signs
of infections such as fever102. Minor gastrointestinal complaints and nasopharyngitis
appear to be the most common side effects for the DPP-4 inhibitors.
In contrast to the fairly minor short-term side effects, long-term consequences of the
GLP-1-based therapies are far more worrisome. Given GLP-1’s prominent role in
increasing beta cell proliferation, panreatitis or the enlargement of the pancreas may
20
result from the use of these therapeutics. In fact, studies have shown that Exenatide
treatment is associated with a significantly higher reporting rate for acute pancreatitis as
compared to the other non-GLP-1 based therapies103; and the FDA issued its first alert for
this potential risk in October 2007. Moreover, pancreatitis is well known to predispose
individuals to pancreatic cancer. Thus, GLP-1-based treatments may also increase the risk
of pancreatic cancer. Furthermore, recent reports have suggested that GLP-1-based
therapies may impose an increased risk of thyroid cancer similarly due to its proliferative
effects in the thyroid102.
1.4 GPCR Interactomes
1.4.1 GPCR Interactomes Regulate Receptor Signaling
Guanine nucleotide-binding protein-coupled receptors (GPCRs) are the largest class of
transmembrane signaling molecules that mediate a myriad of complex physiological
processes104. GPCRs are commonly categorized into 5 groups based on conserved
structures: class-A rhodopsin-like, class-B secretin-like, class-C glutamate-like, Adhesion
and Frizzled/Taste2. GLP-1R, together with the other glucagon-subfamily of receptors,
belongs to the B-class of GPCRs77. Despite intensive research effort, little is known about
the precise mechanism of GPCR activation and signaling. In fact, an activated GPCR is
highly versatile as they interact with a large variety of accessory proteins commonly
referred to as interacting proteins or interactors105; and together they form signaling
networks known as GPCR interactomes (Fig. 6). The protein interactors predominantly
target the C-terminus and sometimes the third intracellular loop of the GPCRs. They
serve to modulate a number of GPCR activities including cross-linking to cytoskeleton,
21
targeting and trafficking between cellular compartments, assembly with other membrane
proteins and consequent allosteric regulation and fine-tuning of downstream signaling
activities106.
The GPCR interactors can be both transmembrane and cytosolic proteins.
Interestingly, some of the transmembrane interactors are themselves GPCRs that form
homo and/or heterodimers. For example, the C-terminus of the D5 dopamine receptor has
been found to interact with the intracellular second loop of the GABAA receptor gamma2
subunit, which results in a mutually inhibitory interaction between the two receptors107.
Similarly, reports have shown that the D1 receptor physically interacts with the NMDA
receptor subunits and the GIP receptor interacts with the GLP-1 receptor108,109. Other
transmembrane interactors include ion channels, ionotropic receptors and single
transmembrane proteins. The discovery of receptor activity modifying proteins (RAMPs)
has revealed a new concept of GPCR pharmacology as RAMPs not only assist receptor
folding and trafficking, but also contribute directly to the activation and signaling of
some B family GPCRs110. There are more than 50 cytosolic GPCR protein interactors that
have been identified. They include accessory ion channels subunits, protein kinases,
small G proteins, cytoskeletal proteins and adhesion molecules105. It is evident from these
studies that the protein interactors to which a GPCR binds at least in part determines its
signaling specificity.
22
1.4.2 Known GLP-1R Binding Proteins
To date, a limited number of interacting molecules have been identified that regulate
GLP-1R function. β-arrestin-1, for example, is a cytosolic scaffolding protein that has
known interactions with many GPCRs111. Interestingly, it has been reported to interact
with the GLP-1R in an agonist-dependent manner using bioluminescence resonance
energy transfer (BRET) technologies112. The GPCR recruitment of β-arrestin is
classically associated with an induction of GPCR desensitization. Emerging data has also
implicated that β-arrestin may mediate GPCR signaling independent of G-proteins113. In
fact, β-arrestin recruitment is required for GLP-1-induced cAMP production and insulin
Figure 6. Examples of GPCR binding proteins (GBPs). G-proteins are signaling proteins that include large heterotrimeric G-proteins with α and βγ subunits and small GTPases such as Rho and ADP-ribosylation factor (ARF). Adapter proteins, such as Nck and Grb2, mediate specific protein-protein interactions (PPI) that drive the formation of protein complexes. Scaffolding proteins, such as β-arrestins and Homer proteins, facilitate GPCR receptome formation by anchoring multiple proteins to one cellular area and regulating signal transduction pathways. Protein tyrosine phosphatases (PTP), like SHP-2, are signaling molecules that regulate a variety of cellular processes including cell growth, differentiation, mitotic cycle, and oncogenic transformation. PDZ-domain containing GBPs are the most abundant GBPs, one of which is Na+/H+ exchanger regulator factor (NHERF). It is required for receptor recycling to the plasma membrane. Moreover, GPCRs form heterodimers with other GPCRs.
23
secretion from INS-1 beta cells, even though the underlying mechanisms remain
unclear114. Also, studies using MIN6 beta cells have presented β-arrestin as a cAMP
independent signaling axis that mediates the prosurvival effects of GLP-1 through the
activation of ERK and BAD115. Caveolin-1, another scaffolding protein, has also been
shown to interact with the GLP-1R at the plasma membrane116. This interaction is
required for the GLP-1R localization to lipid rafts, as caveolin-1 deletion completely
abolished GLP-1 binding and activity in vitro. Small ubiquitin-related modifier protein
(SUMO), on the other hand, interacts with the GLP-1R to down-regulate receptor
trafficking to the plasma membrane, and thus reduces GLP-1 signaling117. Since SUMO
expression is increased in mouse islets exposed to high glucose, its interaction with the
GLP-1R may in part contribute to the reduced incretin responsiveness in T2D. Moreover,
interesting interactions have been characterized within the Glucagon Subfamily
Receptors. For example, GIPR heterodimerizes with the GLP-1R in a ligand-dependent
manner109. This interaction decreased GLP-1-induced β arrestin-1 recruitment to the
GLP-1R and consequently shifted calcium influx, suggesting a potential allosteric
regulation between the GSRs.
1.5 Rational and Hypothesis of Thesis
GLP-1R agonists regulate diverse and important functions in multiple organs,
especially the pancreas. Despite the numerous efforts that have been invested to delineate
the complex signaling cascades of GLP-1R activation, the exact mechanism remains
largely unknown. It is well accepted that GLP-1R signaling differs according to the cell
type where it is expressed in; a common phenomenon shared among many GPCRs. Also,
studies have shown that GPCR interaction partners have profound effects at modulating
24
receptor signaling and as discussed, a few GLP-1R interacting proteins have already been
revealed suggesting that the GLP-1R may have many other unknown interactors.
Therefore, I hypothesize that the GLP-1R has undiscovered interacting proteins capable
of modulating the receptor function, specifically in pancreatic beta cells. These
interaction proteins may at least in part account for the increasingly diverse actions of
GLP-1 and the complex nature of GLP-1R signaling. Also, manipulation of these
interacting proteins may serve as novel strategies to enhance GLP-1/GLP-1R activity and
thus, its therapeutic potential. Furthermore, revealing and studying the GLP-1R
interactome may provide valuable insights into understanding B-family GPCR signaling
in general, advancing the field of GPCR research.
25
Chapter 2: Identification of GLP-1R interacting proteins that restrain the receptor
activity
2.1 Introduction
Glucagon-like peptide 1 (GLP-1) is an incretin hormone mainly produced and
secreted from the enteroendocrine L-cells upon feeding7,118. Acting through its native
glucagon-like peptide 1 receptor (GLP-1R), it controls diverse and important
physiological functions. GLP-1 is best known to enhance glucose-stimulated insulin
secretion (GSIS) from the pancreatic beta cells. Interestingly, this activity is maintained
in patients suffering from Type 2 Diabetes (T2D)119, a metabolic disorder characterized
by chronically elevated blood glucose. Therefore, therapeutic strategies targeting GLP-1
and GLP-1R have been developed to treat T2D. In addition to augmenting GSIS, GLP-1
improves glucose sensing; insulin gene transcription and biosynthesis; and survival and
proliferation of the beta cells7,118. GLP-1 also has complex actions in extra-pancreatic
tissues, including being a potent satiety factor that signals the central nervous system to
inhibit food intake, decreasing gastric emptying and modulating gastric acid secretions.
Moreover, GLP-1R agonists have been demonstrated to protect myocardial cells and
improve heart function72.
GLP-1R, together with other members of the glucagon subfamily of receptors,
glucagon-like peptide 2 receptor (GLP-2R), glucose-dependent insulinotropic
polypeptide receptor (GIPR), and glucagon receptor (GCGR), belongs to the secretin-like
class B family of guanine nucleotide-binding protein-coupled receptors (GPCRs)77,78.
These receptors signal primarily through the activation of a Gs-protein complex and the
second messenger cAMP upon ligand binding. A wide array of accessory proteins have
26
been shown to interact directly with the GPCRs to form functional protein
interactomes105,106,120. These accessory proteins modulate multiple receptor activities such
as receptor folding, subcellular trafficking and intracellular signaling. To date, a limited
number of interacting molecules have been identified to regulate the GLP-1R function. β-
arrestin-1, for example, is a scaffolding protein shown to be required for the stimulation
of GLP-1-induced cAMP production and insulin secretion from INS-1 beta cells114.
Caveolin-1, another scaffolding protein, interacts with the GLP-1R to stabilize its
localization at the plasma membrane116. Interesting interactions within the GSR
subfamily have also been characterized. For example, GIPR heterodimerizes with GLP-
1R to decrease GLP-1-induced β arrestin-1 recruitment to the GLP-1R, suggesting a
potential allosteric regulation between the GSRs109. These findings collectively raise the
possibility that the GLP-1R has many other undiscovered accessory proteins capable of
modulating the receptor function. It is, therefore, important to reveal the GLP-1R
interactome, which may account for the increasingly diverse actions of GLP-1 and the
complex nature of GLP-1R signaling.
The membrane based split-ubiquitin yeast two hybrid (MYTH) system is a modified
yeast two-hybrid screen that does not require the protein of interest to be expressed in the
nucleus121,122 (Fig. 7). Therefore, it allows the study of full-length proteins in situ at the
cell membrane. This screen is designed based on the finding that an ubiquitin protein can
be split into two halves: Cub and Nub. By homologous recombination, the membrane
protein of interest is fused with the Cub linked with an artificial transcription factor (TF)
as bait, while the interactor protein candidate is fused with the NubG as prey. NubG is a
mutated form of Nub to prevent spontaneous reconstitution with the Cub. The bait and
27
prey vectors are sequentially transformed into a competent yeast host, Saccharomyces
cerevisiae strain THY.AP4. Interaction of bait and prey proteins will bring Cub and
NubG into close proximity, leading to resconstitution of the two to form a pseudo-
unbiquitin, which acts as a substrate for ubiquitin-specific protease (UBP). UBP
consequently cleaves the pseudo-ubiquitin to release the TF from the bait construct. The
liberated TF translocates into the nucleus to activate downstream reporter genes for
nutrient (HIS, ADE2) and colorimetric (LacZ) selections. As a result, yeast colonies
expressing positive interactions can survive Adenine and Histidine depleted SD-WL
media (SD-WLAH) and will appear blue. MYTH has already been used to identify
protein interactors of GPCRs, including the Caenorhabditis elegans D2-like dopamine
receptor, the mammalian M3 muscarinic receptor and mu-opioid receptor, and the human
frizzled 1 receptor123-125. It has also been useful for studying interactions of other
membrane bound proteins and transporters such as the glucose transporter 1126.
In the present study, several novel interacting proteins have been identified to
construct the first GLP-1R interactome. Using MYTH, 38 interacting candidates were
revealed for the receptor in absence of ligand stimulation. These interactors were
validated by co-immunoprecipiation and immunofluorescence analyses. I continued to
investigate the functional effects of these interactors and found that many significantly
attenuated GLP-1R signaling when over-expressed in Chinese Hamster Ovarian (CHO)
cells. Three beta cell expressing membrane-bound interactors, solute carrier family 15
member 4 (SLC15A4), amyloid beta A4 precursor-like protein 1 (APLP1) and adaptor-
related protein complex 2 mu 1 subunit (AP2M1), were further selected for individual
knock down in MIN6 beta cells using siRNAs. GLP-1-induced insulin secretion was
28
significantly enhanced when these genes were silenced, strongly suggesting that these
interactor proteins attenuate GLP-1R activity.
2.2 Material and Methods
2.2.1 Construction of yeast and mammalian cell expression GLP-1R plasmids
The coding sequence of GLP-1R was amplified from human islet cDNA by PCR
using Pfu Turbo DNA polymerase (Stratagene Invitrogen, Burlington, Ontario, Canada).
The GLP-1R was cloned into a yeast expression pCWW-ste vector fused with C-terminal
ubiquitin (Cub) and artificial transcription factor LexA-VP16 (TF) (Dualsystems Biotech,
Zurich, Switzerland) by homologous recombination. The human fetal brain cDNA library
Figure 7. Membrane-based split-ubiquitin yeast two-hybrid system (MYTH). (A) The MYTH system is composed of bait and prey proteins. The bait protein is a transmembrane protein of interest fused with a C-terminal ubiquitin (Cub) and an artificial transcription factor comprising LexA and VP16 (TF). The prey protein is a potential interactor fused with an N-terminal ubiquitin (NubG). (B) On the interaction of the bait and prey proteins, Cub and NubG reconstitute a pseudo form of ubiquitin, which leads to the ubiquitin-specific protease (UBP) dependent proteolytic release of the downstream TF. The liberated TF translocates into the nucleus and induces the expression of HIS3, ADE2, and LacZ reporter genes in the yeast. HIS3 and ADE2 allow the yeast to grow on synthetic dropout minus tryptophan/leucine/adenine/histidine (SD-WLAH) plates and β-galactosidase (LacZ) expression is required for colorimetric selection.
29
was cloned into a pPR3-N vector fused with point-mutated N-terminal ubiquitin (NubG).
This prey NubGx library was commercially made by Dualsystems Biotech (Zurich,
Switzerland). All of the constructs were validated by sequencing. Alternatively, the
coding sequence of the receptor was cloned into mammalian cell expression vectors
pcDNA3.1D/V5-His-TOPO and pcDNA3.1D/FLAG-TOPO for functional validation.
The cDNAs of identified interactors on the other hand were cloned into
pcDNA3.1D/3XHA vector. Plasmids were prepared using GenElute plasmid Midiprep kit
(Sigma, Oakville, Canada)
2.2.2 Cell culture and Transfection
HEK293T cells and CHO cells were maintained in DMEM (Gibco Invitrogen,
Burlington, ON, Canada), containing 10% FBS and 100 U/ml penicillin (P/S) at 37°C in
5% CO2. MIN6 cells were maintained in DMEM containing 15% FBS, 100U/ml P/S and
1.7µL beta-mercaptoethanol per 500ml. These cells were transfected with the GLP-1R,
interactors and siRNAs via Lipofectamine 2000 following the manufacturer’s protocol
(Invitrogen, Burlington, Ontario, Canada). A human GLP-1R expressing CHO cell line
(RC2-CHO) was established by selection using 1mg/mL G418 (Gibco, Invitrogen,
Burlington, Ontario, Canada) and was maintained in 0.1mg/mL G418.
2.2.3 MYTH Screening
Membrane based split ubiquitin yeast two-hybrid (MYTH) screen is designed based
on the finding that an ubiquitin protein can be split into two halves: Cub and Nub (Fig.
7)127. By homologous recombination, Glp-1r is fused with the Cub-TF as bait, while the
interactor protein is fused with non-interacting NubG (for MYTH screening) or
30
spontaneously reconstituting NubI (for control experiments) as prey. During MYTH
screening, Cub and NubG reconstitute when the bait and prey proteins interact, leading to
proteolytic cleavage of the TF by an ubiquitin-specific protease. The liberated TF then
translocates into the nucleus to activate reporter genes for nutrient and colorimetric
selections. MYTH was performed as previously described127. Briefly, bait and prey
vectors described above were sequentially transformed into the competent host
Saccharomyces cerevisiae strain THY.AP4 following Gietz and Woods’s method128. The
yeast two-hybrid selection was achieved on synthetic dropout
tryptophan/leucine/adenine/histidine deficient (SD-WLAH) plates. All positive colonies
were inoculated in yeast. The plasmids harboring interactor sequences were extracted and
amplified in Escherichia coli, and purified using mini-prep kits (Qiagen, Toronto,
Ontario, Canada). All the plasmids were sequenced and the sequences were further
blasted in the NCBI database for potential interactor candidates. To validate the
interaction, each interactor was individually transformed back into the yeast
overexpressing the bait (bait dependency test). Interactors which passed the test were
screened through various online databases to determine expression at the gene, protein
and mRNA levels in specific tissues (Human Protein Atlas (HPA) www.proteinatlas.org),
cellular localization (Human Protein Reference Database (HPRD) www.hprd.org),
functional category and known physiological and/or pathophysiological relevance
(PubMed Central www.ncbi.nlm.nih.gov/pubmed and BioGPS www.biogps.org) and the
Beta Cell Biology Consortium (www.betacell.org) to understand the possible role of the
interactors in pancreatic beta cells.
2.2.4 Immunofluorescence
31
Yeast and CHO cells were grown onto poly-L-lysine glass slide and fixed for 30 min
in 4% paraformaldehyde at room temperature. The yeast slides were washed in 1 ml SK
solution (1.8mM K2HP04, 1.7mM KH2PO4, 3.6mM sorbitol, pH7.5) and then re-
suspended in 100-200 µl SK solution. The CHO cells were washed three times with PBS
and then permeabilized with 0.1% TritontmX-100. In both cases, blocking was performed
with 3% BSA in PBS for 30 min at RT in a humidified chamber, followed by
counterstaining with rabbit anti-VP16 (1:200 dilution; Sigma, Oakville, Ontario, Canada),
rabbit anti-Flag (1:250 dilution; Cell Signaling, Whitby, Ontario, Canada), and rat anti-
HA (1:500 dilution; Roche, Mississauga, Ontario, Canada) antibodies for 1 hour at room
temperature. Secondary antibodies used were Alexa Fluor®488 Goat Anti-Rabbit and
Alexa Fluor®546 Goat Anti-Rat IgG (Invitrogen, Burlington, Ontario, Canada). Slides
were imaged using spinning disk confocal microscopy and Pearson’s correlation
coefficient (PCC) was calculated using Volocity 3D Image Analysis Software (Perkin
Elmer, Vaughan, Ontario, Canada). PCC has a range of +1 (perfect correlation) to -1
(perfect negative correlation), with 0 denoting the absence of a relationship129.
2.2.5 Co-immunoprecipitation and immunoblotting
Interactor overexpressed RC2-CHO cells were washed with ice-cold PBS and
harvested in lysis buffer (1%digitonin, 5mM imidazole, 1 Protease Inhibitor Cocktail
Tablet per 10mL, PBS). The cell lysate was centrifuged and the supernatant was extracted
and incubated with Ni2+ beads that were equilibrated with the wash buffer (0.1%
digitonin, 5mM imidazole, 1 Protease Inhibitor Tablet, PBS) for 1h at 4°C. The
precipitation lysate was washed and eluted in PBS containing 1X SDS loading buffer.
The precipitated proteins were analysed by western immunoblotting. 40µl of the
32
precipitated proteins from each sample was loaded and separated on a 10% BioRad TGX
gel and transferred using the BioRad Transfer system (BioRad, Mississauga, Ontario,
Canada). 1:2500 Anti-V5 (Invitrogen, Burlington, Ontario, Canada) and 1:5000 anti-HA
primary antibodies (Covance, Montreal, Quebec, Canada) and HRP conjugated mouse
secondary antibody were used for detection. Signal was detected by chemiluminescence
(GE Healthcare, Inc., Vaughan, Ontario, Canada) and images were acquired using the
Kodak image station 4000pro (Carestream Health, Inc, Vaughan, Ontario, Canada). The
same procedure was used for anti-FLAG tag CO-IP, in which anti-FLAG M2 Affinity
beads (Sigma, Oakville, Ontario, Canada) were used instead of Ni2+ beads.
2.2.6 GLP-1 or Forskolin Stimulated cAMP Formation
Intracellular cAMP content was measured using homogeneous time resolved
fluorescence (HTRF) technology according to the manufacturer’s protocol (Cisbio US,
Bedford, Massachusetts, United States). Briefly, interactor transfected RC2-CHO cells
were plated onto 384-well cell culture plates at a confluence of 3000cells/well. Cells were
preincubated with DMEM (0.5%BSA) for 30min and then incubated with indicated
concentrations of GLP-1 (Abcam, Toronto, Ontario, Canada) or forskolin (Sigma,
Oakville, Ontario, Canada) in DMEM (0.5%BSA, 1µM 3-isobutyl-1-methylxanthine) for
30min. After treatment, cells were incubated with lysis buffer containing cAMP-D2 and
cryptate conjugate tracers for 1 hour and counted using PHERAstar Plus microplate
reader (BMG Labtech, Guelph, ON)
2.2.7 Insulin Secretion Assay, Reverse Transcription PCR and Quantitative
PCR
33
MIN6 cells transfected with siRNAs (Dharmacon, Thermo Fisher Scientific Inc.,
Waltham, Massachusetts, United States) were equilibrated in Krebs-Ringer Bicarbonate
Buffer buffer (in mM, NaCl 115, KCl 5, NaHCO3 24, CaCl2 2.5, MgCl2 1, HEPES 10,
BSA 0.1%) for 1.5h and stimulated with indicated concentration of glucose or GLP-1 for
1h. The supernatant was collected for insulin measurement using HTRF following the
manufacture’s protocol (Cisbio US, Bedford, Massachusetts, United States). Reverse
transcription PCR (RT-PCR) and quantitative PCR (qPCR) was performed as previously
described130. Briefly, cells were lysed and total RNA was extracted using an RNeasy Plus
Kit (Qiagen, Toronto, Ontario, Canada) and converted to cDNA using a reverse
transcription kit (Sigma, Oakville, Ontario, Canada). Power SYBR green PCR master
mix was used for qPCR following the manufacturer’s protocol (Applied Biosystems,
Invitrogen, Burlington, Ontario, Canada). The software used was ViiA™ 7 Real-Time
PCR System (Applied Biosystems, Invitrogen, Burlington, Ontario, Canada). The
sequences of all the qPCR primers were listed in Table 2.
2.2.8 Statistics
Paired t-tests and two-way ANOVA were performed to determine statistical
significance between intracellular cAMP and insulin measurements. P<0.05 was
considered statistically significant.
2.3 Results
2.3.1 Identification of GLP-1R interacting proteins by MYTH
GLP-1R is an integral membrane receptor, thus, MYTH screen was chosen to detect
its interacting proteins in a membrane environment. The interactor candidates were
34
identified from a NubG-fused human fetal brain cDNA prey library because GLP-1R is
expressed in the CNS, where it mediates important functions as I have discussed
previously. To generate a bait construct compatible with MYTH, human GLP-1R cDNA
was subcloned into a yeast-expression pCCw-ste vector fused with the Cub and LexA-
VP16 transcription factor (TF) tag at the C-terminus. Proper expression and localization
of the receptor at the yeast plasma membrane was validated by immunofluorescence (Fig.
8A). The functionality of the receptor was tested by subcloning the same cDNA into
pcDNA3.1, a mammalian-expression vector. GLP-1R overexpressed HEK293T cells
showed a dose-dependent increase of cAMP accumulation upon GLP-1 stimulation (Fig.
8B), indicating that the cloned GLP-1R was biologically functional.
Protein interaction capability within the MYTH system and self-activation of the bait
constructs were verified by NubG/NubI test using non-interacting plasma membrane and
ER resident proteins, Ost1p and Fur4p, as positive (pOst1-NubI and pFur4-NubI) and
negative (pOst1-NubG and pFur4-NubG) prey controls. Co-transformation of the control
prey plasmids and the bait did not affect yeast growth on normal SD-WL (Trp and Leu
deficient) media (Fig. 8C). However, when the co-transformants were plated on selection
Adenine and Histidine depleted SD-WLAH media, yeast growth was only observed with
the NubI-fused proteins but not with the NubG-fused proteins (Fig. 8C). These results
indicated that the GLP-1R was properly transported through the secretory pathway
(positive pOst1-NubI, ER), expressed at the plasma membrane (positive pFur4-NubI) and
not self-activating in the yeast (negative pOst1-NubG and pFur4-NubG). Incubation with
a competitive inhibitor of the His3 reporter gene, 3’-aminotriazole (3'AT) (10mM), was
chosen to minimize inherent leaky His3 reporter expression. At this concentration,
35
growth with the negative NubG control prey was entirely abolished but not with the
positive NubI control prey (Fig. 8C). Before library screening, the GLP-1R expression
was also confirmed in the human fetal brain by RT-PCR (Fig. 8D).
Three independent assay screens using the unliganded hGLP-1R bait and fetal brain
cDNA library prey have been performed, covering approximately 31% of the library with
an average of 1.81 x 106 clones/screen. A bait dependency test was carried out to validate
the identified interactions in yeast. In the test, interactors were individually transformed
back into the GLP-1R expressing yeast and subjected to selection. To pass the test, at
least 2 of 3 colonies of the bait-prey co-expressing yeast must grow on the selection
media in three independent trials. TSPAN13, SYNGR3 and NRGN represented
interactors that passed the test, while NPTN failed (Fig. 8E). Using this strategy, 38 novel
interactors were identified for the GLP-1R (Table 1, Fig. 9). Searching for the proteins in
various databases (NCBI, HPA, HPRD, SymAtlas, BCC) revealed 7 of them to be
associated with cell signaling, 6 with metabolism (energy, protein, nucleic acids,
carbohydrates), 5 with cell proliferation and development, 5 with transmembrane
transportation, 4 with trafficking, endo and exocytosis, 3 with protein processing and
degradation, 2 with transcription regulation, 2 with immune response and 4 with other or
unknown functions (Fig. 9). Thus, a diverse group of novel receptor interactor has been
identified.
36
Figure 8. MYTH screening using an unliganded Cub- and TF (LexA-VP16)-fused human GLP-1R as bait and a NubG-fused human fetal brain cDNA library as prey. (A) Proper expression of the GLP-1R in yeast. Yeast cells transformed with the human GLP-1R were fixed, permeabilized, and probed for VP16. Immunofluorescence indicated the proper expression of the constructed bait vector in yeast. Bar, 10 µm. (B) Characterization of the GLP-1R in HEK293T cells. GLP-1-stimulated cAMP accumulation was measured in HEK293T cells transfected with the bait (n=3). (C) NubG/NubI self-activation test. The bait vector harboring GLP-1R was cotransformed into yeast with control constructs of Ost1p (target endoplasmic reticulum) or Fur4p (target plasma membrane) fused to NubG or NubI. The cotransformants were selected on either tryptophan/leucine (SD-WL) or tryptophan/leucine/adenine/histidine (SD-WLAH) depleted synthetic dropout plates with increasing concentrations (10 mM, 25 mM, 50 mM) of 3’-aminotriazole. Each row of the plate showed a serial dilution of the yeast cells from 1 to 10-4. (D) GLP-1R expression in the human fetal brain. GLP-1R was amplified from the human fetal brain tissue by RT-PCR. Human islets were used as a positive control. (E) Validation by the bait-dependency test. The interactors identified from MYTH were cotransformed with the Cub-TF human GLP-1R into yeast and the cotransformants were selected on SD-WLAH plates with 10 mM 3’-aminotriazole to validate receptor-protein interactions.
37
Fig. 9: GLP1-R interactome identified by MYTH. MYTH revealed 38 interactors for the GLP-1R. Among them, 7 are associated with cell signaling, 6 with metabolism (energy, protein, nucleic acids, carbohydrates), 5 with cell proliferation and development, 5 with transmembrane transportation, 4 with trafficking, endo and exocytosis, 3 with protein processing and degradation, 2 with transcription regulation, 2 with immune response and 4 to have other or unknown functions.
38
Table 1. List of MYTH identified GLP-1R Interactors
39
2.3.2 Validation of the interaction between GLP-1R and interactors in
mammalian cells
2.3.2.1 Coimmunoprecipitation
It was important to examine whether the MYTH-identified receptor-protein
interactions occur in the context of mammalian cells, so co-immunoprecipitation (CO-IP)
was performed. A stable human GLP-1R (V5-6XHistidine tagged) overexpressing CHO
cell line (RC2) was created for these studies. The 38 MYTH-identified GLP-1R
interacting candidates were subcloned into HA-tagged pcDNA3.1 vectors and
independently introduced into the RC2 cells by transient transfection. Protein complexes
were precipitated using Ni2+ beads targeting the His-tag on the GLP-1R, and western
blotting was performed to detect the interactors. For each of the interacting proteins,
intensity of the elute/precipitate band was quantified and expressed as a ratio to the lysate
band (E:L ratio). The E:L ratio was used to estimate the binding affinity of interactors to
the GLP-1R; the closer the E:L ratio of a particular interactor that was to 1, the stronger
its interaction with the GLP-1R. Among the 38 proteins, 6 had high E:L ratios (E:L≥0.7;
SVOP, AP2M1, MFSD5, GABBR2, SLC15A4, ELOVL1), 19 had intermediate E:L
ratios (E:L=0.3-0.7; CEND1, C18orf32, GPR37, STMN3, SYNGR3, CFI, ACSF3,
PIGG, CAND2, PAQR6, SLC31A2, HPN, C5orf32, IFITM3, PH4, CD81, TMEM147,
APH1A, STOML2), 10 had low E:L ratios (E:L=0.1-0.3; MRPL22, APLP1, MYL6,
FDFT1, C14orf166, ATP6V0B, ERP29, ATP13A1, TSPAN13, NUCKS1) and only 3
had insignificant E:L ratios and failed to interact with the GLP-1R (E:L≤0.1; ARFRP1,
STMN1, NRGN) (Fig. 10A,B). To confirm these results, I transiently co-transfected
Flag-tagged GLP-1R and selected interacting candidates into a native CHO cell line and
40
performed co-immunoprecipitation using anti-Flag tag antibodies (Fig. 11). Since the E:L
ratios obtained from anti-FLAG tag CO-IP were similar to that of using Ni2+ beads, it is
confident that the interacting molecules bind to the bait. These results provided validation
of the MYTH system for identifying GLP-1R-protein interactions.
2.3.2.2 Immunofluorescence
Subcellular distribution of selected interactors was examined using confocal laser
scanning microscopy. In CHO cells, Flag-tagged GLP-1R and HA-tagged interactors
were transiently co-expressed and immuno-stained for their respective tags. The
Pearson’s Correlation Coefficient (PCC) was calculated to quantify the degree of co-
localization for each interactor with the receptor. CHO cells overexpressing either the
GLP-1R or the interactor, NRGN, alone were used to control for the background
fluorescence and antibody specificity (Fig. 12A-12B). All of the 13 interactors showed a
degree of co-localization with the GLP-1R, with a PCC ranging from 0.21 to 0.89 (Fig.
12F). Interactors such as MRPL22 and ACSF3 had PCC values that fell below the level
of NRGN (0.27), the protein that failed to co-precipitate with the GLP-1R, were deemed
to have marginal co-localization with the receptor.
41
42
B
Figure 10. Validation of the interaction between GLP-1R and the MYTH-identified interactors in RC2-CHO cells. HA-tagged interactors were independently transfected into the human GLP-1R (V5-His tagged) expressing RC2-CHO cells. CO-IPs were performed using Ni2+ beads targeting the His-tagged GLP-1R and Western blots were performed to detect the precipitants. Intensities of the protein bands in the elute were expressed as a ratio to the bands in the lysate (E:L) to estimate the binding strength for each interactor. (A) Western blots (anti-HA) showing protein interactors with high (***), medium (**), low (*), and insignificant E:L ratios. L, lysate; W3, wash 3; E, elute. (B) Plot of E:L ratios estimating binding strength. Three independent CO-IPs were performed for each interactor.
43
B
Fig. 11: Validation of the interaction between Flag-tagged hGLP-1R and the MYTH identified interactors in CHO cells. Flag-tagged GLP-1R and HA-tagged interactors were co-transfected into CHO cells. CO-IP was performed using Anti-Flag beads targeting the Flag-tagged GLP-1R and western blots were performed to detect the precipitants. A. Western-blots (anti-HA) showing protein interactions with high (***), medium (**), low (*) and insignificant E:L ratios. L = Lysate, W3 = Wash 3, E = Elute. B. Intensity of the protein band in the elute was expressed as a ratio to the band in the lysate (E:L) to estimate the binding strength for each interactor. Two independent CO-IPs were performed for each interactor.
44
45
F
Fig. 12: Co-localization analysis of the GLP-1R and selected interactors by immunofluorescent imaging. CHO cells were over-expressed with the Flag-tagged GLP-1R and HA-tagged interacting proteins and stained for their respective tags. Pearson’s correlation coefficient was calculated to indicate the degree of co-localization. A. CHO cells overexpressing the GLP-1R alone were used as a control for non-specific anti-HA (interactor) binding. Bar=26µm B. CHO cells overexpressing the NRGN alone were used as a control for non-specific anti-Flag (GLP-1R) binding. Bar=26µm C-E. CHO cells coexpressing the GLP-1R and selected interactors with high, intermediate and low PCCs. Two representative images were presented for each interactor protein. Bar=26µm. F. Plot of Pearson’s coefficient indicating degree of colocalization. PCC = +1 (perfect correlation); PCC = 0 (absence of a relationship). Interactors with PCCs lower than NRGN were deemed to have low co-localization with the GLP-1R (n=25-30 cells from 3 independent studies).
46
2.3.3 Interactor proteins reduced the GLP-1R function
2.3.3.1 GLP-1R interactors modulate GLP-1-stimulated cAMP accumulation in
CHO cells
cAMP is a principal signaling molecule generated by the binding and activation of B-
class GPCRs77,78. Therefore, I quantitatively assessed GLP-1R signaling by measuring
cAMP accumulation. CO-IP-validated interactors were expressed in RC2 cells under
basal (0 GLP-1) and stimulated (0.25nM GLP-1) conditions. The dose of 0.25nM of
GLP-1 was chosen for the assay because it is lower than the EC50 concentration
0.5nM131, and thus would not saturate the receptor. Of the 35 validated proteins, 17
showed no significant effects while surprisingly, the other 18 all reduced GLP-1-induced
cAMP accumulation (GABBR2, GPR37, PAQR6, ACSF3, ELOLV1, STMN3,
SLC15A4, SLC31A2, SVOP, APLP1, AP2M1, TMEM147, ERP29, HPN, IFITM3,
C15orf32, TSPAN13 and PIGG)(Fig. 13). Interestingly, the majority of these suppressive
interactors clustered in the functional groups of signal transduction (Group A, 3 out of 7),
transmembrane transportation (Group D, 3 out of 5) and intracellular trafficking (Group
E, 3 out of 4). No interactor was however found to significantly increase cAMP
accumulation. To ensure the observed suppressive effects were occurring at the level of
GLP1-R activation and not simply related to the overexpression of the interactors, I
treated the cells with forskolin, which bypasses the receptor to stimulate cAMP
formation. SVOP, TMEM147 and GPR37 significantly attenuated forskolin-induced
cAMP (Fig. 14) indicating an intrinsic ligand-independent suppression. These three
proteins were therefore not further studied, although these observations are of interest.
47
Fig. 13: The effect of interactors on GLP-1-induced cAMP. RC2-CHO cells over-expressing the individual interactors were stimulated with 0 and 0.25nM of GLP-1. The corresponding cAMP accumulation was measured by HTRF and NRGN was used as the negative control. The interactors were organized into their function groups: A=Signal Transduction, B=Metabolism, C=Cell Proliferation & Development, D=Transportation, E=Trafficking, F=Protein Processing, G=Transcription Regulation, H=Immune Response, I=Others & Unknown. (n=4), * = p<0.05, ** = p<0.01, *** = p<0.001
48
Fig. 14: Elimination of interactors that have an intrinsic suppressive effect on cAMP production. RC2-CHO cells over-expressing the individual interactors were stimulated with 0 and 5µM of forskolin. The corresponding cAMP accumulation was measured by HTRF. NRGN was used as the negative control (n=4), * = p<0.05
49
2.3.3.2 Overexpression of selected GLP-1R interactors showed no significant
effect on GLP-1 stimulated insulin secretion in MIN6 beta cells
The GLP-1R is expressed in pancreatic beta cells, where it is known to augment
insulin secretion in a glucose-dependent manner upon ligand binding7. Therefore, I
continued to examine whether these interactors modulate GLP-1R function in the beta
cells. I first confirmed GLP-1R expression in the mouse insulinoma (MIN6) beta cells by
RT-PCR and qPCR (Fig. 15A-B). Next, I overexpressed selected GLP-1R interactors into
the MIN6 cells and measured their respective effects on GLP-1-induced insulin secretion
at 20mM glucose. Surprisingly, no significant change in insulin secretion was observed
consequent to the overexpressions (Fig. 16).
A B
Fig 15: Validation of GLP-1R and interaction protein expressions in MIN6 cells. A. GLP-1R was amplified from MIN6 by RT-PCR. B. Endogenous expression level of the GLP-1R and the interactors was assessed by qPCR. Interactor transcript quantities were normalized to the level of Kcnj11, a gene that codes for a subunit of beta cell KATP channel. Proteins with transcript levels of 90% Kcnj11 or higher were deemed to have high expression in the MIN6.
50
2.3.3.3 Reducing expression of selected GLP-1R interactors enhances GLP-1-
stimulated insulin secretion in MIN6 beta cells
Overexpression of the interacting proteins reduced GLP-1-induced cAMP
accumulation in CHO cells but did not affect insulin secretion in MIN6. Thus, I
continued to examine whether knocking down these interactors could elicit an effect. I
first validated endogenous expression of the interactors, which were originally identified
from the fetal brain cDNA library, in MIN6 (Fig. 15B). Of the 15 interactors that
attenuated GLP-1R signaling in CHO, 7 (SLC15A4, APLP1, AP2M1, TSPAN13,
Fig. 16: Overexpression of selected interaction proteins did not affect GLP-1-induced insulin secretion in MIN6 beta cells. Selected interactors were overexpressed in MIN6 cells. Cells were treated with 10nM of GLP-1 at high glucose (20mM) and incubated for 1 hour. Cell medium was collected for insulin measurement by HTRF. n=3
51
ELOVL1, ERP29, STMN3) were expressed in MIN6 at levels higher than or comparable
(at least 90%) to the gene, Kcnj11. Kcnj11 encodes Kir6.2, subunit of the well
characterized beta-cell adenosine triphosphate dependent potassium (KATP) channel that
is integral for the regulation of insulin release132; therefore, I used its transcriptional level
as a reference for choosing the proteins of interest. Since the GLP-1R functions at the cell
membrane, I selected 3 membrane-bound interacting proteins, SLC15A4, APLP1,
AP2M1 out of the 7 candidates for knock down studies.
MIN6 cells were treated with either non-targeting siRNA (scrambled control) or
siRNAs targeting each of the three interactors. The transcript expression level of each
interactor was knocked down by over 85% as compared to the control (Fig. 17A). I
stimulated the tested cells with 10nM of GLP-1 at both low and high glucose conditions
(0mM and 20mM) and measured insulin secretion (Fig. 17B). In the scrambled control,
GLP-1 treatment increased insulin secretion by 22.0±4% at 20mM glucose (Fig. 17C) but
not at 0mM. This is expected, because GLP-1 response is known to be glucose
dependent. When the three interactor proteins, SLC15A4, APLP1 and AP2M1 were
individually knocked down, GLP-1-induced insulin secretion was significantly increased
to 56.0±10% (p<0.001), 49.2±7% (p<0.001) and 47.9±8% (p<0.05) respectively at
20mM glucose (Fig. 17C). However, in the absence of GLP-1, there was no difference in
insulin secretion at low or high glucose in any group (Fig. 17B). Taken together, these
data strongly suggest that the MYTH-identified interactors attenuate ligand-stimulated
GLP-1R activity in beta cells at high glucose.
52
A
B
53
C
Fig. 17: Knocking down selected interactors enhanced GLP-1R function in MIN6 cells. SLC15A4, APLP1 and AP2M1 were individually knocked down in MIN6 cells using siRNA. Cells were treated with 10nM of GLP-1 at low (0mM) and high (20mM) glucose and incubated for 1 hour. Cell medium was collected for insulin measurement by HTRF. MIN6 cells transfected with non-targeting siRNA were used as a control (Scramble). A. qPCR showing efficient knock-down of SLC15A4, APLP1 and AP2M1 by siRNA as compared to the control. B. GLP-1 induced insulin secretion at low and high glucose. C. Percent increase of insulin secretion upon GLP-1 treatment at high glucose (HG, 20mM). (n=4), * = p<0.05, *** = p<0.001
54
Chapter 3: General Discussion
3.1 Summary of Findings
Recent studies have revealed a large range of accessory proteins that interact directly
with the intracellular domains of the GPCRs, forming functional protein units or
interactomes105. In this study, we screened a human fetal brain cDNA library to produce
the first GLP-1R interactome using the MYTH system (Fig. 9). Bait dependency test in
yeast and CO-IP and immunofluorescence analysis in mammalian CHO cells were used
to validate the protein-receptor interactions. Bioinformatic analysis revealed the
interactors to be involved in signal transduction, metabolism, cell growth and trafficking.
Using cAMP and insulin secretion assays, we further showed that a subset of the
interacting proteins attenuated the receptor function. Thus, inhibiting these interactors
may serve as a novel strategy to enhance GLP-1R activity.
3.2 MYTH Identified GLP-1R Interactome
Using the MYTH screen, we have for the first time identified an interactome for the
GLP-1R (Fig. 9). GLP-1R is a GPCR best known to maintain global glucose homeostasis
and promote pancreatic beta cell proliferation. Thus, it is interesting to find the majority
of GLP-1R interacting candidates to be involved in cell signal transduction, metabolism,
proliferation and development (Fig. 9). However, none of these proteins has been
previously recognized as an interactor of the GLP-1R. The inactive state of GLP-1R
during the MYTH screen may in part explain why well-known interaction partners of the
GPCRs, such as G-proteins, are not identified. Also, some interaction proteins may be
neglected from the screen as false negatives as discussed previously. Furthermore, three
55
MYTH screens only covered about 31% of the human fetal brain cDNA library in this
study, leaving the remaining 69% unstudied.
Of the 38 MYTH identified GLP-1R interaction proteins, 7 are functionally clustered
into cell signaling. These proteins include small GTPases such as ARFRP1. GTPases are
a family of hydrolase enzymes that can bind and hydrolyze guanosine triphosphate (GTP)
and they play essential roles in mediating GPCR signaling133. Thus, it is interesting to
reveal such GTPases as novel interacting partners of the GLP-1R. GLP-1 has marked
effects on nutrient metabolism; therefore, it is not surprising that its receptor interacts
with metabolic factors. Six interactors identified from the screen are involved in
metabolism, which include ELOVL1 and ACSF3. ELOVL1 is a condensing enzyme that
catalyzes the synthesis of both saturated and monounsaturated very long chain fatty acids,
whereas ACSF3 activates the breakdown of complex fatty acids (UniProt). Revealing
these factors suggests an action of GLP-1 on fat metabolism. Proteins that are involved in
cell differentiation and proliferation have also been identified to interact with the GLP-
1R, which corresponds to its role in enhancing beta and neuronal cell survival and
proliferation. CEND1, for instance, is known to enhance pig neuroblastoma cell
differentiation in vitro and may be involved in neuronal differentiation in vivo134,135.
Furthermore, in agreement with previous findings, this study showed that GPCRs could
interact with one another. In fact, we have identified multiple GPCRs in the GLP-1R
interactome including GABBR2 and GPR37.
3.3 Limitations of MYTH
The split-ubiquitin MYTH is a novel, high-throughput method for studying protein-
protein interactions (PPIs) developed by Dr. Igor Stagljar’s laboratory at the University of
56
Toronto127. It is often the method of choice for analyzing full-length integral membrane
proteins such as GPCRs because it is not restricted by their hydrophobicity. Positive
interactors are easily identified in MYTH through growth selection, colorimetric or
fluorescent phenotype and prey plasmid DNA is readily accessible for sequencing.
Furthermore, MYTH can detect interacting proteins expressed at very low abundance in
vivo. However, every technology has its limitations. One shortcoming of the MYTH
assay is that the bait, GLP-1R in this case, is not activated during the screens because
peptide ligands cannot penetrate through the yeast cell wall to stimulate receptors
expressed on the plasma membrane136. Thus, MYTH does not detect protein complexes
that would be recruited by agonist activation. In order to study ligand dependency of the
receptor interctome, additional innovative strategies will be required to mimic the
receptor activation process.
Artifacts within the prey interactor datasets are also limiting factors of MYTH. These
artifacts include some bona fide interactors that are not picked up by the screen, so-called
“false negatives”. Bona fide interactors may arise from a number of possibilities: 1)
failure of subcellular trafficking of these interacting proteins to the receptor; 2) disruption
in the sequence of protein trafficking, protein modification (i.e. phosphorylation) and
sequential assembly and disassembly, that collectively lead to the eventual functional
multimeric receptor complex; 3) lack of required post-translational modification in the
yeast host organism; 4) improper folding of the Nub fusion prey protein; or 5)
interference of key binding domains of the receptor or interactor by fused ubiquitin
tags137. Moreover, partial cDNA clones that make up the cDNA prey libraries may also
lack all the domains necessary to interact with the bait receptor as they normally would at
57
full-length. Finally, these proteins may be underrepresented or even absent from the
cDNA library used for yeast two hybrid (YTH) screening.
False positives are also common in yeast two hybrid (YTH) screens138. There are two
types of false positives. The first type results from the binding of study protein to the bait
protein of interest only in the context of YTH assay, but not in a true physiologic setting.
An explanation for this phenomenon is that this study protein contains certain requisite
recognition sequences for the bait, but does not normally interact with the bait protein in
normal cellular conditions. The second type of false positive occurs when the study
protein induces TF activity independent of any meaningful PPI. This phenomenon may
be a consequence of overexpression of bait proteins, a process that can inadvertently
cause endogenous expression of ubiquitin tag leading to self-activation. Furthermore,
this artificial process can induce plasmid rearrangements or copy number changes that
generate auto-activators, or alter reporter genes in a manner that causes constitutive
expression137. Finally, MYTH is restricted to identifying only binary interactions,
leading to an underrepresentation of those proteins that interact only within entire protein
complexes.
3.4 Validations of the MYTH Identified Interactors
As mentioned above, artefacts are common in MYTH screening. Thus, I have
performed a series of validation assays to ensure the accuracy of the screen. The bait
dependency test was first used to validate the interactors (Fig. 8E). This test is very useful
at eliminating false positives generated by MYTH, however, it should be noted that it is
not 100% accurate and is performed in yeast but not mammalian cells. To further validate
the interactor proteins in a mammalian setting, CO-IPs were performed using a stable-
58
hGLP-1R expressing RC2-CHO cell line. Thirty-five of the 38 GLP-1R interactor
candidates co-precipitated with the receptor, highlighting the dependability of MYTH
screen to examine mammalian protein-protein interactions for membrane-spanning
proteins. It should be noted however that the CO-IPs were carried out in a model system,
and overexpressing some proteins may promote artificial interactions. Since globally, the
proteins identified interacted with the receptor with differing strengths (Fig. 10B), and
some completely failed to bind the bait (NRGN, STMN1 and ARFRP1), I am confident
that these CO-IP interactions were not simply artifacts of overexpression. The
complementary anti-Flag CO-IP (Fig. 11) and co-localization (Fig. 12) results also
confirmed the positive binding between GLP-1R and the MYTH-identified interactors.
However, it would be ideal to show endogenous interactions, which would require good
quality antibodies for the native proteins. Unfortunately, it has been reported that
commercially available GLP-1R antibodies are not specific, which is consistent with my
data as the anti-GLP-1R antibody (Abcam 39072) failed to identify the overexpressed
GLP-1R in CHO cells (Fig. 18A). To address the interactors, I tested three antibodies
that are commercially available (Sigma SAB3500044 for AP2M1, Sigma SAB2102167
for SLC15A4 and Thermo Scientific PA5-23720 for GABBR2) for western blotting (Fig.
18B) and immunofluorescence (Fig. 18C). Unfortunately, the results are again negative
for overexpressed interactors with the antibodies in question.
59
Fig 18. Commercially available antibodies are not specific. A. Anti-GLP-1R antibody was tested by western blotting. Lysates of untransfected CHO cells were loaded in lane 1; lysates overexpressed with Flag-tagged GLP-1R were loaded in lane2 and 3. Anti-Flag antibody was used as the positive control. B. Anti-AP2M1, anti-APLP1 and anti-GABBR2 antibodies were tested by western blotting. In each blot, lysates of untransfected MIN6 were loaded in lane 1; lysates overexpressed with the respective HA-tagged interactors were loaded in lane2 and 3. Anti-HA antibody was used as the positive control. C. MIN6 cells were transfected with HA-tagged GABBR2 and HA-tagged SLC15A4 in the first two columns and stained with anti-HA and anti-GABBR2/anti-SLC15A4 antibodies respectively. In the last column, untransfected MIN6 cells were stained with anti-GABBR2 or anti-SLC15A4 antibodies for endogenous proteins. Bar=26µm
60
3.5 MYTH Identified GLP-1R Interaction Proteins Restrained the Receptor
Activity
Interestingly, 15 out of the 35 CO-IP validated GLP-1R interactors attenuated cAMP
accumulation in the RC2 cells upon GLP-1 stimulation, but none enhanced it (Fig. 13).
Since the initial MYTH screening was performed using an unliganded inactive GLP-1R,
it is possible that the identified interactors suppress receptor activation. These putative
interactors may act to attenuate receptor activity, stabilizing the GLP1-R in a low-activity
state in absence of ligand binding. Although there is no direct evidence supporting this
idea, knocking down SLC15A4, APLP1 and AP2M1 significantly enhanced GLP-1-
stimulated insulin secretion from the MIN6 cells (Fig. 17B, C). I did not observe any
significant change to the basal insulin secretion in absence of GLP-1 at both low and high
glucose when the interactors were knocked down individually (Fig. 17B). This further
suggests a functional redundancy and compensation among the protein interactors, which
act as a whole to keep the GLP-1R quiescent in its unliganded state. The inability to
detect interactors capable of enhancing cAMP accumulation may also be explained by
effects on the cAMP-independent signaling cascades of the GLP-1R. For example,
stimulation of the GLP-1R in pancreatic beta cells has been reported to activate pathways
that involve transactivation of the EGFR and subsequent PI3 kinases. The identification
of interactors that work through such pathways will require employing different screening
assays. It was to my surprise that overexpression of the protein interactors did not affect
GLP-1-induced insulin secretion in MIN6 cells (Fig. 16). This could be explained by the
endogenous expression of these proteins in MIN6, which may be sufficient to keep the
61
receptor inactive. Thus, increasing the amount of the interactors did not give rise to
further effects.
3.6 Potential Functional Relevance of the Interactor Candidiates
3.6.1 SLC15A4
The interactors that we identified may modulate GLP-1R function through distinct
mechanisms. SLC15A4 is a proton oligopeptide co-transporter that transports free
histidine, protons and certain di- and tri-peptides across cell membranes139. Protonation
has long been known to affect multiple GPCR properties, including the rate of ligand-
induced conformational change and the receptor-mediated G-protein activation140. By co-
transporting protons and peptides, SLC15A4 may couple the GLP-1R activity to the
extracellular nutrient state through modulating the local pH. Interestingly, a single
nucleotide polymorphism in the SLC15A4 gene has been reported to associate with the
development of T2D141, supporting the regulatory role of this transporter on GLP-1R.
The exact in vivo function of SLC15A4 is however largely unknown and its investigation
is beyond the scope of this study.
3.6.2 APLP1
APLP1 is a membrane-associated glycoprotein that belongs to a highly conserved
amyloid precursor protein gene family. It has been previously reported to interact with
the α2A-adrenergic receptor, a Gi coupled GPCR, and the voltage gated calcium channel
Cav2.3142,143. In both cases, a considerable shift of the receptor/channel localization from
the plasma membrane to the intracellular compartments was observed. Clathrin is
suggested to play a role in this APLP1-mediated re-localization because APLP1 contains
a highly conserved NPxY motif that is known to mediate endocytosis via clathrin coated
62
pits. APLP1 may similarly induce GLP-1R endocytosis via the same mechanism, which
would explain the decrease in GLP-1R activity observed in our study. The fact that I
show AP2M1, a subunit of the Adaptor Protein 2 (AP2) complex known to regulate
GPCR recycling via clathrin144, also attenuated GLP-1R signaling further supports this
idea (Fig. 13, Fig. 17B,C). Interestingly, localization of the GLP-1R shifted from the
plasma membrane to the intracellular space in CHO cells when AP2M1 was over-
expressed (Fig. 12D).
Upon intracellular cleavage of APLP1 by proteases in the secretase family, a
cytoplasmic fragment is released, which may act as a transcriptional activator to mediate
apoptosis145. Similarly, islet amyloid polypeptide (or amylin) can induce apoptotic cell
death of pancreatic β-cells, which has been implicated in the development of type 2
diabetes146. Most interestingly, Needham et al. (2008) demonstrated that APLP1 and
APLP2 double knocked out mice exhibit both hypoglycmia and hyperinsulinemia, which
strongly suggests that these proteins play a role in regulating glucose homeostasis147.
Thus, it is possible that GLP-1R-APLP1 interactions contribute to the cytoprotective and
glucoregulatory effects of GLP-1.
3.7 Protein Interactors as Allosteric Regulators of the GLP-1R and Their
Therapeutic Potential
The interacting proteins that are identified in this study act allosterically, that is at sites
distinct to the endogenous ligand. From a therapeutic perspective, targeting the receptors
allosterically have several major advantages including: 1) enhanced receptor subtype
selectivity; 2) the ability to simultaneously bind to the receptor with the endogenous
ligand (restoring physiologically relevant temporal control); 3) induce a new repertoire of
63
receptor conformations and therefore influencing receptor activity; and 4) in particular for
peptide-activated receptors, introducing the potential for oral administration80. At present,
very few allosterically acting ligands have been identified for the GLP-1R including the
Novo Nordisk compounds 2-(2′-methyl)thiadiazolylsulfanyl-3-trifluoromethyl-6,7-
dichloroquinoxaline (compound 1) and 6,7-dichloro-2-methylsulfonyl-3-t-
butylaminoquinoxaline (compound 2)148, the latter of which demonstrates glucose-
dependent insulin release via the GLP-1R148,149. My thesis study implicates that
interacting partners of the GLP-1R can regulate the receptor activity allosterically; thus,
they may serve as new target sites for drug development. Since all of the interactors that
are identified restrained the receptor signaling, antagonization of these proteins may serve
as novel strategies to enhance the insulintropic effects of GLP-1R.
3.8 Alternative Proteomic Screening Strategies – AP-MS
Many biochemical approaches have been developed to detect PPI including affinity
precipitation and mass spectroscopy (AP-MS), protein chip technology using tagged gene
arrays, protein fragment complementation in mammalian cells, and fluorescence (FRET)
and bioluminescence resonance energy transfer (BRET) approaches150,151. AP-MS is a
well-established high-throughput technology comparable to MYTH for allowing studies
of full-length GPCR as the bait152. This approach isolates interacting proteins by affinity
pull-down using an affinity-tagged bait protein of interest. Thus, AP-MS can be
performed in mammalian cells providing an in vivo environment for interactions. The
isolated protein complexes are separated by gel electrophoresis and identified by mass
spectrometry. The advantage of AP-MS is that it allows the detection of higher order
interactions as it captures entire protein complexes153. Also, only one fused artificial tag
64
is present in AP-MS, which lessens the chance of interference with PPI. However, the
strategy also has potential limitations: 1) the difficulty in distinguishing between specific
and nonspecific interactors; 2) the unreliable quantification of changes in the abundance
of protein complexes or their components; 3) weak hydrostatic/transient interactions
between proteins may be disrupted by harsh biochemical conditions during the
purification step; 4) often is difficult to identify low-abundance interacting proteins; and
5) overexpression of bait protein can lead to artifact or false-positive interactions154.
Currently, a post-doctoral fellow in my lab has used AP-MS and purified the GLP-1R
interactome in both CHO and MIN6 cells. Very interestingly, he observed differential
interaction profile for the GLP-1R in the stimulated and non-stimulated state
(Unpublished). However, further validation tests are required.
3.9 Alternative Proteomic Screening Strategies – FRET/BRET
FRET is a useful tool for analyzing interactions between two molecules155. Here, the
interacting partners are labeled with two fluorophores, the “donor” and the “acceptor”
that share overlapping emission/absorption spectrum, respectively. When the two partners
interact, the fluorophores are brought into close proximity, allowing for an energy
transfer from the donor to the acceptor in the form of photons through nonradiative
dipole–dipole coupling. The efficiency of this energy transfer or FRET efficiency can be
used to determine the exact distance between the two binding partners and quantify the
interaction. One common fluorophore pair is a cyan fluorescent protein (CFP) – yellow
fluorescent protein (YFP) pair. Both of these fluorephores are color variants of the green
fluorescent protein (GFP). A major limitation of FRET is the requirement for external
illumination to initiate the energy transfer, which can lead to photobleaching and/or
65
background noise due to direct excitation of the acceptor. To avoid this drawback, BRET
has been developed156. BRET uses a bioluminescent luciferase rather than CFP to
produce an initial photon emission compatible with YFP. One major advantage of
FRET/BRET over other tools is that they allow for real-time analysis of protein-protein
interactions in live cells. Also, FRET/BRET is suitable for assaying interactions in
different subcellular compartments or specific organelles of the native cell. To date,
BRET technology has already been used extensively to study oligomerization between
B1 class of GPCRs including the secretin receptor, the glucagon receptor, and the
glucagon-like peptide-1 and -2 receptors (GLP-1R and GLP-2R)157.
66
Chapter 4: Future Directions and Conclusion
4.1 Future Directions
4.1.1 Assess the effects of interactors on MIN6 apoptosis and proliferation.
As I have mentioned previously, GLP-1R activation not only enhances insulin
secretion but also promotes proliferation and inhibits apoptosis of the pancreatic beta
cells. Thus, it would be of interest to examine whether overexpressing or knocking down
selected interactor candidates of interest may affect these two parameters using MIN6.
Viability assays such as the caspase assay can be used.
4.1.2 Validate PPI using Bioluminescence/Fluorescence Resonance
Energy Transfer Strategies (BRET/FRET)
Up to this stage, all of the PPI validations, CO-IP or IF, were performed in
overexpression model systems statically in vitro. It would be ideal to examine
endogenous interactions between the GLP-1R and interacting proteins in actual cellular
conditions. However, due to the quality issues of the commercially available antibodies
that I have mentioned previously, it is technically difficult to address this question at the
current stage. Assessing PPI using BRET/FRET is an interesting alternative as it allows
the study of PPI temporally and spatially in living cells. Using BRET/FRET, ligand or
glucose stimulatory effects on GLP-1R-protein interactions can also be assessed in real-
time.
4.1.3 Assess the functional effects of GLP-1R interactors in mouse or human
islets.
All of the functional assays that have been performed were carried out in cell lines,
CHO or MIN6. Thus, it would be of interest to examine whether these effects similarly
67
occur in the context of mouse or human islets. Specifically, interacting proteins can be
introduced or knocked down from the islets using viral approach with ShRNAs. cAMP
and insulin secretion assays can then be performed.
4.1.4 Assess the mechanisms by which the inhibitory effects of SLC15A4, APLP1
and AP2M1 were mediated.
Using cAMP and insulin secretion assays, I have shown that SLC15A4, APLP1 and
AP2M1 restrained the GLP-1R activity in MIN6 cells. It is important to discover the
mechanisms that underlie these inhibitory effects. As I have discussed previously, APLP1
and AP2M1 are trafficking proteins that may affect the localization of GLP-1R via
clathrin-mediated endocytosis. Thus, ligand-binding assays can be performed to quantify
the amount of GLP-1R expression on the plasma membrane under APLP1 and AP2M1
overexpression or knock down conditions. Clathrin can be further knocked down to
assess whether it is indeed involved in this trafficking event.
4.1.5 Construct and functionally characterize an islet expressing GLP-1R
interactome
As mentioned above, GPCR interactomes are tissue specific. In other words, the
interaction proteins that are revealed from the fetal brain tissue may not represent the
interactome profile in the pancreatic beta cells. Given the beneficial effects that GLP-1
has on insulin biosynthesis and secretion in the beta cells, it would be ideal to identify an
islet-specific GLP-1R interactome using an islet prey library. An islet is ~60% beta-cells,
thus, it will likely to contain a full complement of GLP-1R interacting proteins for normal
beta-cell function.
68
4.2 Conclusion
This present study is the first example of an interaction screen using the full-length
GLP-1R. I show that the unliganded GLP-1R interacts with a large and diverse group of
proteins. The functional assays suggest that these interactors act collectively to attenuate
GLP-1R activity. Taken together, my data sheds light on the allosteric regulation of GLP-
1R, which may also apply to other GSRs or GPCRS in general. Furthermore, some of the
newly identified GLP-1R accessory proteins may allow for selective manipulation of the
GLP-1R activity.
69
Table2. Primer Sequences for RT-PCR and qPCR
Gene Name Sequence
mPAQR6 (qPCR) Forward: 5’-CTATGCCGCCTACTCCATGC-3’ Reverse 5’-CTTGCTGAACCCGGGGTATT-3’
mERP29 (qPCR) Forward: 5’-GCCCTGGCAGGCGAGTTCAG-3’ Reverse: 5’-ATCCGGGCCATCTCAGAGGCTG-3’
mSTMN3 (qPCR) Forward: 5’-CGGCGTGAGCATGAACGCGA-3’ Reverse: 5’-CAAGTGCGCCTCGCGGATCT-3’
mPIGG (qPCR) Forward: 5’-GCTGCCAGCTGCGTAGCGAT-3’ Reverse: 5’-GGCGCTGGGGTCTCTGCATC-3’
mACSF3 (qPCR) Forward: 5’-GCCTCCGTCAGGGGGTTGGA-3’ Reverse: 5’-TGGTCTCAGGTTCGGCCGCT-3’
mTSPAN13 (qPCR) Forward: 5’-CGTGCGCCCCCATAATCGGA-3’ Reverse: 5’-TGGCGCGAGGGTCTTTCTGG-3’
mSLC31A2 (qPCR) Forward: 5’-GAGCCTGCCTGCCACCAAGAG-3’ Reverse: 5’-AGCCGAGCCCAGGACCACAC-3’
mIFITM3 (qPCR) Forward: 5’-GCCTTCATCACCGCTGCCAGT-3’ Reverse: 5’-GTGCGGTGCCCCCATCTCAG-3’
mHPN (qPCR) Forward: 5’-TGGGGCTCACAGAGGAGGTTCC-3’ Reverse: 5’-GCAAGCCGTGAGTCCCCTGG-3’
mELOVL1 (qPCR) Forward: 5’-ACCTGGCGCTGTGACCCCAT-3’ Reverse: 5’-AGAGCCAGGCCACTCGAACCA-3’
mGABBR2 (qPCR) Forward: 5’-ACGTGACCAGCCCAACGTGC-3’ Reverse: 5’-CGTGACGTGCTCGCCTGGTT-3’
mTMEM147 (qPCR) Forward: 5’-GGCTGGGCCACTGCTGAACT-3’ Reverse: 5’-AACAGCAGGCCGGAAGGTGTG-3’
mMRPL22 (qPCR) Forward: 5’-GATCTGGGCAACGCAGACGCT-3’ Reverse: 5’-TCTCCGAGGTTCGCCAGGCA-3’
mSLC15A4 (qPCR) Forward: 5’-GGCTCACCATGTTTGATGGG-3’ Reverse: 5’-ACATCCCCACAGCAATCCTC-3’
mGPR37 (qPCR) Forward: 5’-GACTGGGTTCCTGACCAAGG-3’ Reverse: 5’-CGGTCCCAAAGATCACCACA-3’
mAPLP1 (qPCR) Forward: 5’-AACGGCAACGCCTGGTGGAG-3’ Reverse: 5’-GGCCACGTGCTGGTAGTGCC-3’
mAP2M1 (qPCR) Forward: 5’-CTGGCGGCGAGAAGGCATCA-3’ Reverse: 5’-CCGGCCTGACACATGGGCAC-3’
mKCNJ11 (qPCR) Forward: 5’-AAGCTGGGTTGGGGGCTCAGT-3’ Reverse: 5’-TTGGCCCCGGGAAGGGGTTT-3’
mGLP-1R (qPCR) Forward: 5’-GCCTTCTCGGTGTTCTCGG-3’ Reverse: 5’-TGGGCAATCGGATGATGAGC-3’
hGLP-1R (Homologous Recombination for Bait Construct)
Forward: 5’-CGGCCAGGCCTTTAATTAAGGCCGCCTCGGCCATC-ATGGCCGGCGCCCCCGGCCC-3’ Reverse: 5’-CGACGGTATCGATAAGCTTGATATCGAATTCCTGCA-GCTGCAGGAGGCCTGGCAAG-3’
mGLP-1R (RT-PCR, Fig. 2D) Forward: 5’-CTGGCCTTCTCGGTGTTCTC-3’ Reverse: 5’-ATGAAGCGTAGGGTTCCTCG-3’
mGLP-1R (RT-PCR, Fig. 8A) Forward: 5’-GTACCACGGTGTCCCTCTCA-3’ Forward: 5’-CAGCATTTCCGAAACTCCAT-3’
70
Abbreviations Footnote
Ach acetylcholine
AP-‐MS affinity precipitation and mass spectroscopy
AP2 Adaptor protein 2
AP2M1 Adaptor-‐related protein complex 2 µ1 subunit
APLP1 Amyloid β precursor-‐like protein 1
ATF-‐4 activating transcription factor-‐4
BAD Bcl-‐2-‐associated death promoter
BRET bioluminescence resonance energy transfer
c-‐src proto-‐oncogene tyrosine kinase src
CHO Chinese hamster ovarian cells
CHOP C/EBP (CCAAT/enhancer-‐binding protein)-‐homologous protein
CNS central nervous system
CO-‐IP Coimmunoprecipitation
CREB cAMP-‐response-‐element-‐binding protein
Cub C-‐terminal ubiquitin
E:L ratio The ratio of the intensity of the elute ti the lysate band
EGFR epidermal growth factor receotpr
eIF2α eukaryote initiation factor 2 α
Epac2 exchange protein activated by cAMP
ER endoplasmic reticulum
Ex-‐4 exendin 4
FDA United States Food and Drug Administration
71
FRET Förster resonance energy transfer
GI gastrointestinal system
GIP glucose-‐dependent insulinotropic polypeptide
GLP-‐1 glucagon-‐like peptide 1
GBP GPCR binding protein
GPCRs G-‐protein coupled receptors
GRP gastrin-‐releasing peptide
GRPP glicentin-‐related polypeptide
GSIS glucose-‐stimulated insulin secretion
GSR Glucagon subfamily of receptors
HTRF Homogeneous time-‐resolved fluorescence
IRS-‐2 insulin receptor substrate 2
MAPK mitogen-‐activated protein kinases
MIN6 Mouse insulinoma β cells
MPGF major proglucagon fragment
MYTH membrane-‐based split ubiquitin yeast two-‐hybrid assay
NubG Point mutated N-‐terminal ubiquitin
OXM oxyntomodulin
PC1/3, 2 prohomrone convertase 1/3, 2
PCC Pearson's correlation coefficient
PDX-‐1 pancreas duodenum homeobox-‐1
PI3K phosphoinositide 3-‐kinase
PKA protein kinase A
72
PKB protein kinase B
PKC protein kinase C
PPI protein-‐protein interaction
RAMPs receptor activity modifying proteins
RC2-‐CHO Human GLP-‐1R expressing CHO cell line
REEP2 Receptor accessory protein 2
SD-‐WL Synthetic dropout minus tryptophan/leucine
SD-‐WLAH Synthetic dropout minus tryptophan/leucine/adenine/histidine
siRNAs small interfering RNAs
SLC15A4 Solute carrier family 15 member 4
T2D type-‐2 diabetes
TF Artifical transcription factor comprised of LexA-‐VP16
UBP Ubiquitin-‐specific protease
VDCCs voltage-‐gated calcium channels
73
References
1. Bayliss, W.M. & Starling, E.H. The mechanism of pancreatic secretion. The Journal of Physiology 28, 325–353 (1902).
2. Best, C.H. A BRIEF REVIEW OF CERTAIN PHYSIOLOGICAL PROPERTIES OF INSULIN. Can Med Assoc J 23, 141–145 (1930).
3. MCINTYRE, N., HOLDSWORTH, C.D. & TURNER, D.S. NEW INTERPRETATION OF ORAL GLUCOSE TOLERANCE. Lancet 2, 20–21 (1964).
4. ELRICK, H., STIMMLER, L., HLAD, C.J. & ARAI, Y. PLASMA INSULIN RESPONSE TO ORAL AND INTRAVENOUS GLUCOSE ADMINISTRATION. J. Clin. Endocrinol. Metab. 24, 1076–1082 (1964).
5. Mojsov, S., Weir, G.C. & Habener, J.F. Insulinotropin: glucagon-like peptide I (7-37) co-encoded in the glucagon gene is a potent stimulator of insulin release in the perfused rat pancreas. J. Clin. Invest. 79, 616–619 (1987).
6. Kreymann, B., Williams, G., Ghatei, M.A. & Bloom, S.R. Glucagon-like peptide-1 7-36: a physiological incretin in man. Lancet 2, 1300–1304 (1987).
7. Baggio, L.L. & Drucker, D.J. Biology of incretins: GLP-1 and GIP. Gastroenterology 132, 2131–2157 (2007).
8. Brown, J.C., Dryburgh, J.R., Ross, S.A. & Dupré, J. Identification and actions of gastric inhibitory polypeptide. Recent Prog. Horm. Res. 31, 487–532 (1975).
9. White, J.W. & Saunders, G.F. Structure of the human glucagon gene. Nucleic Acids Research 14, 4719–4730 (1986).
10. Mojsov, S. et al. Preproglucagon gene expression in pancreas and intestine diversifies at the level of post-translational processing. J. Biol. Chem. 261, 11880–11889 (1986).
11. Larsen, P.J., Tang-Christensen, M., Holst, J.J. & Orskov, C. Distribution of glucagon-like peptide-1 and other preproglucagon-derived peptides in the rat hypothalamus and brainstem. Neuroscience 77, 257–270 (1997).
12. Lim, G.E. & Brubaker, P.L. Glucagon-Like Peptide 1 Secretion by the L-Cell: The View From Within. Diabetes 55, S70–S77 (2006).
13. Jin, T. Mechanisms underlying proglucagon gene expression. J. Endocrinol. 198, 17–28 (2008).
14. Trinh, D.K.Y., Zhang, K., Hossain, M., Brubaker, P.L. & Drucker, D.J. Pax-6 activates endogenous proglucagon gene expression in the rodent gastrointestinal epithelium. Diabetes 52, 425–433 (2003).
15. Yi, F., Brubaker, P.L. & Jin, T. TCF-4 mediates cell type-specific regulation of proglucagon gene expression by beta-catenin and glycogen synthase kinase-3beta. J. Biol. Chem. 280, 1457–1464 (2005).
16. Nian, M., Drucker, D.J. & Irwin, D. Divergent regulation of human and rat proglucagon gene promoters in vivo. Am. J. Physiol. 277, G829–37 (1999).
17. Drucker, D.J. & Asa, S. Glucagon gene expression in vertebrate brain. J. Biol. Chem. 263, 13475–13478 (1988).
18. Rouillé, Y., Martin, S. & Steiner, D.F. Differential processing of proglucagon by the subtilisin-like prohormone convertases PC2 and PC3 to generate either glucagon or glucagon-like peptide. J. Biol. Chem. 270, 26488–26496 (1995).
74
19. Holst, J.J. Glucagon and glucagon-like peptides 1 and 2. Results Probl Cell Differ 50, 121–135 (2010).
20. Ugleholdt, R. et al. Impaired intestinal proglucagon processing in mice lacking prohormone convertase 1. Endocrinology 145, 1349–1355 (2004).
21. Zhu, X. et al. Disruption of PC1/3 expression in mice causes dwarfism and multiple neuroendocrine peptide processing defects. Proc. Natl. Acad. Sci. U.S.A. 99, 10293–10298 (2002).
22. Myojo, S., Tsujikawa, T., Sasaki, M., Fujiyama, Y. & Bamba, T. Trophic effects of glicentin on rat small-intestinal mucosa in vivo and in vitro. J. Gastroenterol. 32, 300–305 (1997).
23. Pellissier, S., Sasaki, K., Le-Nguyen, D., Bataille, D. & Jarrousse, C. Oxyntomodulin and glicentin are potent inhibitors of the fed motility pattern in small intestine. Neurogastroenterol. Motil. 16, 455–463 (2004).
24. Chiba, M. et al. Glicentin inhibits internalization of enteric bacteria by cultured INT-407 enterocytes. Pediatr. Surg. Int. 23, 551–554 (2007).
25. Schjoldager, B., Mortensen, P.E., Myhre, J., Christiansen, J. & Holst, J.J. Oxyntomodulin from distal gut. Role in regulation of gastric and pancreatic functions. Dig. Dis. Sci. 34, 1411–1419 (1989).
26. Baggio, L.L., Huang, Q., Brown, T.J. & Drucker, D.J. Oxyntomodulin and glucagon-like peptide-1 differentially regulate murine food intake and energy expenditure. Gastroenterology 127, 546–558 (2004).
27. Wynne, K. et al. Oxyntomodulin increases energy expenditure in addition to decreasing energy intake in overweight and obese humans: a randomised controlled trial. Int. J. Obes. Relat. Metab. Disord. 30, 1729–1736 (2006).
28. Orskov, C. et al. Glucagon-like peptides GLP-1 and GLP-2, predicted products of the glucagon gene, are secreted separately from pig small intestine but not pancreas. Endocrinology 119, 1467–1475 (1986).
29. Holst, J.J., Orskov, C., Nielsen, O.V. & Schwartz, T.W. Truncated glucagon-like peptide I, an insulin-releasing hormone from the distal gut. FEBS Lett. 211, 169–174 (1987).
30. Orskov, C., Bersani, M., Johnsen, A.H., Højrup, P. & Holst, J.J. Complete sequences of glucagon-like peptide-1 from human and pig small intestine. J. Biol. Chem. 264, 12826–12829 (1989).
31. Orskov, C., Wettergren, A. & Holst, J.J. Biological effects and metabolic rates of glucagonlike peptide-1 7-36 amide and glucagonlike peptide-1 7-37 in healthy subjects are indistinguishable. Diabetes 42, 658–661 (1993).
32. Wettergren, A., Pridal, L., Wøjdemann, M. & Holst, J.J. Amidated and non-amidated glucagon-like peptide-1 (GLP-1): non-pancreatic effects (cephalic phase acid secretion) and stability in plasma in humans. Regul. Pept. 77, 83–87 (1998).
33. Orskov, C., Rabenhøj, L., Wettergren, A., Kofod, H. & Holst, J.J. Tissue and plasma concentrations of amidated and glycine-extended glucagon-like peptide I in humans. Diabetes 43, 535–539 (1994).
34. Hartmann, B. et al. Structure, measurement, and secretion of human glucagon-like peptide-2. Peptides 21, 73–80 (2000).
35. Drucker, D.J., Erlich, P., Asa, S.L. & Brubaker, P.L. Induction of intestinal
75
epithelial proliferation by glucagon-like peptide 2. Proc. Natl. Acad. Sci. U.S.A. 93, 7911–7916 (1996).
36. Tang-Christensen, M., Larsen, P.J., Thulesen, J., Rømer, J. & Vrang, N. The proglucagon-derived peptide, glucagon-like peptide-2, is a neurotransmitter involved in the regulation of food intake. Nat. Med. 6, 802–807 (2000).
37. Lovshin, J., Estall, J., Yusta, B., Brown, T.J. & Drucker, D.J. Glucagon-like peptide (GLP)-2 action in the murine central nervous system is enhanced by elimination of GLP-1 receptor signaling. J. Biol. Chem. 276, 21489–21499 (2001).
38. Lovshin, J.A., Huang, Q., Seaberg, R., Brubaker, P.L. & Drucker, D.J. Extrahypothalamic expression of the glucagon-like peptide-2 receptor is coupled to reduction of glutamate-induced cell death in cultured hippocampal cells. Endocrinology 145, 3495–3506 (2004).
39. Quesada, I., Tudurí, E., Ripoll, C. & Nadal, A. Physiology of the pancreatic alpha-cell and glucagon secretion: role in glucose homeostasis and diabetes. J. Endocrinol. 199, 5–19 (2008).
40. Furuta, M. et al. Defective prohormone processing and altered pancreatic islet morphology in mice lacking active SPC2. Proc. Natl. Acad. Sci. U.S.A. 94, 6646–6651 (1997).
41. Eissele, R. et al. Glucagon-like peptide-1 cells in the gastrointestinal tract and pancreas of rat, pig and man. Eur. J. Clin. Invest. 22, 283–291 (1992).
42. Unger, R.H., Ohneda, A., Valverde, I., Eisentraut, A.M. & Exton, J. Characterization of the responses of circulating glucagon-like immunoreactivity to intraduodenal and intravenous administration of glucose. J. Clin. Invest. 47, 48–65 (1968).
43. Roberge, J.N. & Brubaker, P.L. Regulation of intestinal proglucagon-derived peptide secretion by glucose-dependent insulinotropic peptide in a novel enteroendocrine loop. Endocrinology 133, 233–240 (1993).
44. Rocca, A.S. & Brubaker, P.L. Role of the vagus nerve in mediating proximal nutrient-induced glucagon-like peptide-1 secretion. Endocrinology 140, 1687–1694 (1999).
45. Anini, Y., Hansotia, T. & Brubaker, P.L. Muscarinic receptors control postprandial release of glucagon-like peptide-1: in vivo and in vitro studies in rats. Endocrinology 143, 2420–2426 (2002).
46. Anini, Y. & Brubaker, P.L. Muscarinic receptors control glucagon-like peptide 1 secretion by human endocrine L cells. Endocrinology 144, 3244–3250 (2003).
47. Persson, K. et al. Reduced GLP-1 and insulin responses and glucose intolerance after gastric glucose in GRP receptor-deleted mice. Am. J. Physiol. Endocrinol. Metab. 279, E956–62 (2000).
48. Dhanvantari, S., Seidah, N.G. & Brubaker, P.L. Role of prohormone convertases in the tissue-specific processing of proglucagon. Molecular Endocrinology 10, 342–355 (1996).
49. Deacon, C.F. et al. Both subcutaneously and intravenously administered glucagon-like peptide I are rapidly degraded from the NH2-terminus in type II diabetic patients and in healthy subjects. Diabetes 44, 1126–1131 (1995).
50. Hansen, L., Deacon, C.F., Orskov, C. & Holst, J.J. Glucagon-like peptide-1-(7-
76
36)amide is transformed to glucagon-like peptide-1-(9-36)amide by dipeptidyl peptidase IV in the capillaries supplying the L cells of the porcine intestine. Endocrinology 140, 5356–5363 (1999).
51. Plamboeck, A., Holst, J.J., Carr, R.D. & Deacon, C.F. Neutral endopeptidase 24.11 and dipeptidyl peptidase IV are both mediators of the degradation of glucagon-like peptide 1 in the anaesthetised pig. Diabetologia 48, 1882–1890 (2005).
52. Vilsbøll, T., Agersø, H., Krarup, T. & Holst, J.J. Similar elimination rates of glucagon-like peptide-1 in obese type 2 diabetic patients and healthy subjects. J. Clin. Endocrinol. Metab. 88, 220–224 (2003).
53. Vilsbøll, T., Krarup, T., Deacon, C.F., Madsbad, S. & Holst, J.J. Reduced postprandial concentrations of intact biologically active glucagon-like peptide 1 in type 2 diabetic patients. Diabetes 50, 609–613 (2001).
54. Wang, H. et al. Suppression of Pdx-1 perturbs proinsulin processing, insulin secretion and GLP-1 signaling in INS-1 cells. Diabetologia 48, 720–731 (2005).
55. Holz, G.G., Kühtreiber, W.M. & Habener, J.F. Pancreatic beta-cells are rendered glucose-competent by the insulinotropic hormone glucagon-like peptide-1(7-37). Nature 361, 362–365 (1993).
56. Yusta, B. et al. GLP-1 receptor activation improves beta cell function and survival following induction of endoplasmic reticulum stress. Cell Metabolism 4, 391–406 (2006).
57. Buteau, J., Foisy, S., Joly, E. & Prentki, M. Glucagon-like peptide 1 induces pancreatic beta-cell proliferation via transactivation of the epidermal growth factor receptor. Diabetes 52, 124–132 (2003).
58. Buteau, J., Spatz, M.L. & Accili, D. Transcription factor FoxO1 mediates glucagon-like peptide-1 effects on pancreatic beta-cell mass. Diabetes 55, 1190–1196 (2006).
59. Heller, R.S., Kieffer, T.J. & Habener, J.F. Insulinotropic glucagon-like peptide I receptor expression in glucagon-producing alpha-cells of the rat endocrine pancreas. Diabetes 46, 785–791 (1997).
60. Fehmann, H.C. & Habener, J.F. Functional receptors for the insulinotropic hormone glucagon-like peptide-I(7-37) on a somatostatin secreting cell line. FEBS Lett. 279, 335–340 (1991).
61. Nauck, M.A. et al. Effects of glucagon-like peptide 1 on counterregulatory hormone responses, cognitive functions, and insulin secretion during hyperinsulinemic, stepped hypoglycemic clamp experiments in healthy volunteers. J. Clin. Endocrinol. Metab. 87, 1239–1246 (2002).
62. Kastin, A.J., Akerstrom, V. & Pan, W. Interactions of glucagon-like peptide-1 (GLP-1) with the blood-brain barrier. J. Mol. Neurosci. 18, 7–14 (2002).
63. Turton, M.D. et al. A role for glucagon-like peptide-1 in the central regulation of feeding. Nature 379, 69–72 (1996).
64. Tang-Christensen, M. et al. Central administration of GLP-1-(7-36) amide inhibits food and water intake in rats. Am. J. Physiol. 271, R848–56 (1996).
65. Meeran, K. et al. Repeated intracerebroventricular administration of glucagon-like peptide-1-(7-36) amide or exendin-(9-39) alters body weight in the rat. Endocrinology 140, 244–250 (1999).
77
66. Abbott, C.R. et al. The inhibitory effects of peripheral administration of peptide YY(3-36) and glucagon-like peptide-1 on food intake are attenuated by ablation of the vagal-brainstem-hypothalamic pathway. Brain Research 1044, 127–131 (2005).
67. Talsania, T., Anini, Y., Siu, S., Drucker, D.J. & Brubaker, P.L. Peripheral exendin-4 and peptide YY(3-36) synergistically reduce food intake through different mechanisms in mice. Endocrinology 146, 3748–3756 (2005).
68. Knauf, C. et al. Brain glucagon-like peptide-1 increases insulin secretion and muscle insulin resistance to favor hepatic glycogen storage. J. Clin. Invest. 115, 3554–3563 (2005).
69. Perry, T., Haughey, N.J., Mattson, M.P., Egan, J.M. & Greig, N.H. Protection and reversal of excitotoxic neuronal damage by glucagon-like peptide-1 and exendin-4. J. Pharmacol. Exp. Ther. 302, 881–888 (2002).
70. Meier, J.J. et al. Normalization of glucose concentrations and deceleration of gastric emptying after solid meals during intravenous glucagon-like peptide 1 in patients with type 2 diabetes. J. Clin. Endocrinol. Metab. 88, 2719–2725 (2003).
71. Meier, J.J., Kemmeries, G., Holst, J.J. & Nauck, M.A. Erythromycin antagonizes the deceleration of gastric emptying by glucagon-like peptide 1 and unmasks its insulinotropic effect in healthy subjects. Diabetes 54, 2212–2218 (2005).
72. Ussher, J.R. & Drucker, D.J. Cardiovascular biology of the incretin system. Endocr. Rev. 33, 187–215 (2012).
73. Nyström, T. et al. Effects of glucagon-like peptide-1 on endothelial function in type 2 diabetes patients with stable coronary artery disease. Am. J. Physiol. Endocrinol. Metab. 287, E1209–15 (2004).
74. Nikolaidis, L.A. et al. Recombinant glucagon-like peptide-1 increases myocardial glucose uptake and improves left ventricular performance in conscious dogs with pacing-induced dilated cardiomyopathy. Circulation 110, 955–961 (2004).
75. Bose, A.K., Mocanu, M.M., Carr, R.D., Brand, C.L. & Yellon, D.M. Glucagon-like peptide 1 can directly protect the heart against ischemia/reperfusion injury. Diabetes 54, 146–151 (2005).
76. Gros, R. et al. Cardiac function in mice lacking the glucagon-like peptide-1 receptor. Endocrinology 144, 2242–2252 (2003).
77. Brubaker, P.L. & Drucker, D.J. Structure-function of the glucagon receptor family of G protein-coupled receptors: the glucagon, GIP, GLP-1, and GLP-2 receptors. Recept. Channels 8, 179–188 (2002).
78. Willard, F.S. & Sloop, K.W. Physiology and emerging biochemistry of the glucagon-like peptide-1 receptor. Exp Diabetes Res 2012, 470851 (2012).
79. Underwood, C.R. et al. Crystal structure of glucagon-like peptide-1 in complex with the extracellular domain of the glucagon-like peptide-1 receptor. Journal of Biological Chemistry 285, 723–730 (2010).
80. Koole, C. et al. Recent advances in understanding GLP-1R (glucagon-like peptide-1 receptor) function. Biochem. Soc. Trans. 41, 172–179 (2013).
81. Montrose-Rafizadeh, C. et al. Pancreatic glucagon-like peptide-1 receptor couples to multiple G proteins and activates mitogen-activated protein kinase pathways in Chinese hamster ovary cells. Endocrinology 140, 1132–1140 (1999).
78
82. Kodama, S. et al. Enhanced expression of PDX-1 and Ngn3 by exendin-4 during beta cell regeneration in STZ-treated mice. Biochem. Biophys. Res. Commun. 327, 1170–1178 (2005).
83. Coopman, K. et al. Comparative effects of the endogenous agonist glucagon-like peptide-1 (GLP-1)-(7-36) amide and the small-molecule ago-allosteric agent “compound 2” at the GLP-1 receptor. J. Pharmacol. Exp. Ther. 334, 795–808 (2010).
84. Baggio, L.L., Kim, J.-G. & Drucker, D.J. Chronic exposure to GLP-1R agonists promotes homologous GLP-1 receptor desensitization in vitro but does not attenuate GLP-1R-dependent glucose homeostasis in vivo. Diabetes 53 Suppl 3, S205–14 (2004).
85. Takeda, Y. et al. Systems analysis of GLP-1 receptor signaling in pancreatic -cells. AJP: Cell Physiology 301, C792–C803 (2011).
86. Nauck, M., Stöckmann, F., Ebert, R. & Creutzfeldt, W. Reduced incretin effect in type 2 (non-insulin-dependent) diabetes. Diabetologia 29, 46–52 (1986).
87. Todd, J.F., Edwards, C.M., Ghatei, M.A., Mather, H.M. & Bloom, S.R. Subcutaneous glucagon-like peptide-1 improves postprandial glycaemic control over a 3-week period in patients with early type 2 diabetes. Clin. Sci. 95, 325–329 (1998).
88. Meneilly, G.S. et al. Glucagon-like peptide-1 (7-37) augments insulin release in elderly patients with diabetes. Diabetes Care 24, 964–965 (2001).
89. Eng, J., Kleinman, W.A., Singh, L., Singh, G. & Raufman, J.P. Isolation and characterization of exendin-4, an exendin-3 analogue, from Heloderma suspectum venom. Further evidence for an exendin receptor on dispersed acini from guinea pig pancreas. J. Biol. Chem. 267, 7402–7405 (1992).
90. Göke, R. et al. Exendin-4 is a high potency agonist and truncated exendin-(9-39)-amide an antagonist at the glucagon-like peptide 1-(7-36)-amide receptor of insulin-secreting beta-cells. J. Biol. Chem. 268, 19650–19655 (1993).
91. Fehmann, H.C. et al. Stable expression of the rat GLP-I receptor in CHO cells: activation and binding characteristics utilizing GLP-I(7-36)-amide, oxyntomodulin, exendin-4, and exendin(9-39). Peptides 15, 453–456 (1994).
92. Thum, A. et al. Endoproteolysis by isolated membrane peptidases reveal metabolic stability of glucagon-like peptide-1 analogs, exendins-3 and -4. Exp. Clin. Endocrinol. Diabetes 110, 113–118 (2002).
93. Simonsen, L., Holst, J.J. & Deacon, C.F. Exendin-4, but not glucagon-like peptide-1, is cleared exclusively by glomerular filtration in anaesthetised pigs. Diabetologia 49, 706–712 (2006).
94. Fineman, M.S. et al. Effect on glycemic control of exenatide (synthetic exendin-4) additive to existing metformin and/or sulfonylurea treatment in patients with type 2 diabetes. Diabetes Care 26, 2370–2377 (2003).
95. Heine, R.J. et al. Exenatide versus insulin glargine in patients with suboptimally controlled type 2 diabetes: a randomized trial. Ann. Intern. Med. 143, 559–569 (2005).
96. Nauck, M.A. et al. A comparison of twice-daily exenatide and biphasic insulin aspart in patients with type 2 diabetes who were suboptimally controlled with sulfonylurea and metformin: a non-inferiority study. Diabetologia 50, 259–267
79
(2007). 97. Knudsen, L.B. et al. Potent derivatives of glucagon-like peptide-1 with
pharmacokinetic properties suitable for once daily administration. J. Med. Chem. 43, 1664–1669 (2000).
98. Nisal, K., Kela, R., Khunti, K. & Davies, M.J. Comparison of efficacy between incretin-based therapies for type 2 diabetes mellitus. BMC Med 10, 152 (2012).
99. Ahrén, B. et al. Inhibition of dipeptidyl peptidase-4 reduces glycemia, sustains insulin levels, and reduces glucagon levels in type 2 diabetes. J. Clin. Endocrinol. Metab. 89, 2078–2084 (2004).
100. Charbonnel, B. et al. Efficacy and safety of the dipeptidyl peptidase-4 inhibitor sitagliptin added to ongoing metformin therapy in patients with type 2 diabetes inadequately controlled with metformin alone. Diabetes Care 29, 2638–2643 (2006).
101. Aschner, P. et al. Effect of the dipeptidyl peptidase-4 inhibitor sitagliptin as monotherapy on glycemic control in patients with type 2 diabetes. Diabetes Care 29, 2632–2637 (2006).
102. Butler, P.C., Elashoff, M., Elashoff, R. & Gale, E.A.M. A Critical Analysis of the Clinical Use of Incretin-Based Therapies: Are the GLP-1 therapies safe? Diabetes Care 36, 2118–2125 (2013).
103. Gale, E.A.M. Collateral damage: the conundrum of drug safety. Diabetologia 52, 1975–1982 (2009).
104. Vassilatis, D.K. et al. The G protein-coupled receptor repertoires of human and mouse. Proc. Natl. Acad. Sci. U.S.A. 100, 4903–4908 (2003).
105. Bockaert, J., Fagni, L., Dumuis, A. & Marin, P. GPCR interacting proteins (GIP). Pharmacology & Therapeutics 103, 203–221 (2004).
106. Bockaert, J. et al. GPCR-interacting proteins (GIPs): nature and functions. Biochem. Soc. Trans. 32, 851–855 (2004).
107. Liu, F. et al. Direct protein-protein coupling enables cross-talk between dopamine D5 and gamma-aminobutyric acid A receptors. Nature 403, 274–280 (2000).
108. Lee, F.J.S. et al. Dual regulation of NMDA receptor functions by direct protein-protein interactions with the dopamine D1 receptor. Cell 111, 219–230 (2002).
109. Schelshorn, D.W. et al. Lateral Allosterism in the Glucagon Receptor Family: GLP-1 Induces GPCR Heteromer Formation. Molecular Pharmacology (2011).doi:10.1124/mol.111.074757
110. Morfis, M., Christopoulos, A. & Sexton, P.M. RAMPs: 5 years on, where to now? Trends Pharmacol. Sci. 24, 596–601 (2003).
111. Reiter, E. & Lefkowitz, R.J. GRKs and β-arrestins: roles in receptor silencing, trafficking and signaling. Trends in Endocrinology & Metabolism 17, 159–165 (2006).
112. Jorgensen, R., Kubale, V., Vrecl, M., Schwartz, T.W. & Elling, C.E. Oxyntomodulin differentially affects glucagon-like peptide-1 receptor beta-arrestin recruitment and signaling through Galpha(s). J. Pharmacol. Exp. Ther. 322, 148–154 (2007).
113. Rajagopal, S., Rajagopal, K. & Lefkowitz, R.J. Teaching old receptors new tricks: biasing seven-transmembrane receptors. Nat Rev Drug Discov 9, 373–386
80
(2010). 114. Sonoda, N. et al. Beta-Arrestin-1 mediates glucagon-like peptide-1 signaling to
insulin secretion in cultured pancreatic beta cells. Proceedings of the National Academy of Sciences 105, 6614–6619 (2008).
115. Quoyer, J. et al. GLP-1 mediates antiapoptotic effect by phosphorylating Bad through a beta-arrestin 1-mediated ERK1/2 activation in pancreatic beta-cells. Journal of Biological Chemistry 285, 1989–2002 (2010).
116. Syme, C.A., Zhang, L. & Bisello, A. Caveolin-1 Regulates Cellular Trafficking and Function of the Glucagon-Like Peptide 1 Receptor. Molecular Endocrinology 20, 3400–3411 (2006).
117. Rajan, S., Torres, J., Thompson, M.S. & Philipson, L.H. SUMO downregulates GLP-1-stimulated cAMP generation and insulin secretion. Am. J. Physiol. Endocrinol. Metab. 302, E714–23 (2012).
118. Drucker, D.J. The biology of incretin hormones. Cell Metabolism 3, 153–165 (2006).
119. Lovshin, J.A. & Drucker, D.J. Incretin-based therapies for type 2 diabetes mellitus. Nat Rev Endocrinol 5, 262–269 (2009).
120. Tilakaratne, N. & Sexton, P.M. G-Protein-coupled receptor-protein interactions: basis for new concepts on receptor structure and function. Clin. Exp. Pharmacol. Physiol. 32, 979–987 (2005).
121. Iyer, K. et al. Utilizing the split-ubiquitin membrane yeast two-hybrid system to identify protein-protein interactions of integral membrane proteins. Sci. STKE 2005, pl3 (2005).
122. Kittanakom, S. et al. Analysis of membrane protein complexes using the split-ubiquitin membrane yeast two-hybrid (MYTH) system. Methods Mol. Biol. 548, 247–271 (2009).
123. Pandey, P. & Harbinder, S. The Caenorhabditis elegans D2-like dopamine receptor DOP-2 physically interacts with GPA-14, a Gαi subunit. Journal of Molecular Signaling 7, 3 (2012).
124. Rosemond, E. et al. Regulation of M₃ muscarinic receptor expression and function by transmembrane protein 147. Molecular Pharmacology 79, 251–261 (2011).
125. Jin, J. et al. Interaction of the mu-opioid receptor with GPR177 (Wntless) inhibits Wnt secretion: potential implications for opioid dependence. 1–15 (2010).doi:10.1186/1471-2202-11-33
126. Ohno, H. et al. 4F2hc stabilizes GLUT1 protein and increases glucose transport activity. AJP: Cell Physiology 300, C1047–54 (2011).
127. Snider, J. et al. Detecting interactions with membrane proteins using a membrane two-hybrid assay in yeast. Nat Protoc 5, 1281–1293 (2010).
128. Gietz, R.D. & Woods, R.A. Transformation of yeast by lithium acetate/single-stranded carrier DNA/polyethylene glycol method. Meth. Enzymol. 350, 87–96 (2002).
129. Adler, J. & Parmryd, I. Quantifying colocalization by correlation: the Pearson correlation coefficient is superior to the Mander's overlap coefficient. Cytometry A 77, 733–742 (2010).
130. Hardy, A.B. et al. Characterization of Erg K+ channels in alpha- and beta-cells of
81
mouse and human islets. Journal of Biological Chemistry 284, 30441–30452 (2009).
131. J. Biol. Chem.-2006-Dai-21942-53. 1–12 (2006). 132. Bennett, K., James, C. & Hussain, K. Pancreatic β-cell KATP channels:
Hypoglycaemia and hyperglycaemia. Rev Endocr Metab Disord 11, 157–163 (2010).
133. Ba, W., van der Raadt, J. & Nadif Kasri, N. Rho GTPase signaling at the synapse: Implications for intellectual disability. Exp. Cell Res. (2013).doi:10.1016/j.yexcr.2013.05.033
134. Politis, P.K., Akrivou, S., Hurel, C., Papadodima, O. & Matsas, R. BM88/Cend1 is involved in histone deacetylase inhibition-mediated growth arrest and differentiation of neuroblastoma cells. FEBS Lett. 582, 741–748 (2008).
135. Politis, P.K., Thomaidou, D. & Matsas, R. Coordination of cell cycle exit and differentiation of neuronal progenitors. Cell Cycle 7, 691–697 (2008).
136. Nekhotiaeva, N. et al. Cell entry and antimicrobial properties of eukaryotic cell-penetrating peptides. FASEB J. 18, 394–396 (2004).
137. Fields, S. High-throughput two-hybrid analysis. The promise and the peril. FEBS J. 272, 5391–5399 (2005).
138. Bork, P. Comparative analysis of protein interaction networks. Bioinformatics 18 Suppl 2, S64 (2002).
139. Kottra, G. & Daniel, H. The proton oligopeptide cotransporter family SLC15 in physiology and pharmacology. Pfl�gers Archiv European Journal of Physiology 447, 610–618 (2004).
140. Ghanouni, P. et al. The effect of pH on beta(2) adrenoceptor function. Evidence for protonation-dependent activation. J. Biol. Chem. 275, 3121–3127 (2000).
141. Takeuchi, F. et al. Search for type 2 diabetes susceptibility genes on chromosomes 1q, 3q and 12q. J Hum Genet 53, 314–324 (2008).
142. Weber, B. et al. Interaction of the amyloid precursor like protein 1 with the α2A-adrenergic receptor increases agonist-mediated inhibition of adenylate cyclase. Cellular Signaling 18, 1748–1757 (2006).
143. Radhakrishnan, K., Krieger, A., Dibué, M., Hescheler, J. & Schneider, T. APLP1 and Rab5A interact with the II-III loop of the voltage-gated Ca-channel Ca(v)2.3 and modulate its internalization differently. Cell. Physiol. Biochem. 28, 603–612 (2011).
144. Owen, D.J., Collins, B.M. & Evans, P.R. Adaptors for clathrin coats: structure and function. Annu. Rev. Cell Dev. Biol. 20, 153–191 (2004).
145. Yankner, B.A. & Lu, T. Amyloid beta-protein toxicity and the pathogenesis of Alzheimer disease. J. Biol. Chem. 284, 4755–4759 (2009).
146. Lorenzo, A., Razzaboni, B., Weir, G.C. & Yankner, B.A. Pancreatic islet cell toxicity of amylin associated with type-2 diabetes mellitus. Nature 368, 756–760 (1994).
147. Needham, B.E. et al. Identification of the Alzheimer's disease amyloid precursor protein (APP) and its homologue APLP2 as essential modulators of glucose and insulin homeostasis and growth. J. Pathol. 215, 155–163 (2008).
148. Knudsen, L.B. et al. Small-molecule agonists for the glucagon-like peptide 1 receptor. Proc. Natl. Acad. Sci. U.S.A. 104, 937–942 (2007).
82
149. Irwin, N., Flatt, P.R., Patterson, S. & Green, B.D. Insulin-releasing and metabolic effects of small molecule GLP-1 receptor agonist 6,7-dichloro-2-methylsulfonyl-3-N-tert-butylaminoquinoxaline. Eur. J. Pharmacol. 628, 268–273 (2010).
150. Figeys, D. Mapping the human protein interactome. Cell Res. 18, 716–724 (2008).
151. Suter, B., Kittanakom, S. & Stagljar, I. Interactive proteomics: what lies ahead? BioTechniques 44, 681–691 (2008).
152. Varjosalo, M. et al. Interlaboratory reproducibility of large-scale human protein-complex analysis by standardized AP-MS. Nat. Methods 10, 307–314 (2013).
153. Cai, B., Wang, H., Zheng, H. & Wang, H. Detection of protein complexes from affinity purification/mass spectrometry data. BMC Syst Biol 6 Suppl 3, S4 (2012).
154. Auerbach, D., Thaminy, S., Hottiger, M.O. & Stagljar, I. The post-genomic era of interactive proteomics: facts and perspectives. Proteomics 2, 611–623 (2002).
155. Sun, Y., Rombola, C., Jyothikumar, V. & Periasamy, A. Förster resonance energy transfer microscopy and spectroscopy for localizing protein-protein interactions in living cells. Cytometry A (2013).doi:10.1002/cyto.a.22321
156. Lam, M.H.Y. & Stagljar, I. Strategies for membrane interaction proteomics: no mass spectrometry required. Proteomics 12, 1519–1526 (2012).
157. Roed, S.N., Orgaard, A., Jorgensen, R. & De Meyts, P. Receptor oligomerization in family B1 of G-protein-coupled receptors: focus on BRET investigations and the link between GPCR oligomerization and binding cooperativity. Front Endocrinol (Lausanne) 3, 62 (2012).
158. Huang, X., Dai F. F., Gaisano. G. et al. The identification of novel proteins that interact with the GLP-1 receptor and restrain its activity. Molecular Endocrinology. Epub ahead of print.