MOLECULAR FACTORS THAT INFLUENCE THE BINDING OF AGONISTS TO AMPA RECEPTORS

Author

Kyle Everett Montgomery, Southern Illinois University CarbondaleFollow

Date of Award

1-1-2009

Degree Name

Doctor of Philosophy

Department

Pharmacology

First Advisor

Arai, Amy

Second Advisor

Faingold, Carl

Abstract

AMPA receptors mediate excitatory synaptic transmission throughout the central nervous system via activation by their natural agonist glutamate. Several other molecules have been recognized as receptor agonist or antagonist, and recently allosteric modulators have been developed that potentiate the currents generated by these receptors. The goal of this thesis has been to address specific and as yet unresolved questions regarding the binding interactions between the AMPA receptors and these classes of molecules. For instance AMPA receptors are seemingly converted to have lower affinity for agonist as they move towards synapses and we evaluate two hypotheses put forward to explain the molecular mechanisms responsible for this. Additionally, guanine nucleotides competitively inhibit AMPA receptors and a second goal has been to further characterize guanine nucleotide binding, and to create mutations that selectively diminish this so that the function of the inhibition can be evaluated. A third goal has been to characterize the molecular factors that influence the effects of the allosteric modulators in order to explain why their efficacy differs greatly between brain regions. Experiments pertaining to these three goals were carried out sequentially and are described below as Projects 1 (guanine nucleotide inhibition), Project 2 (agonist affinity), and Project 3 (allosteric modulators). Project 1. Guanine nucleotides competitively inhibit AMPA-Rs (AMPA receptors) and because this inhibition is ubiquitous among virtually all types of glutamate receptors from fish to mammals, it likely serves a physiological function. Evaluation of this would be greatly facilitated if nucleotide binding could be eliminated through mutations without altering other aspects of receptor function, or if compounds were discovered that selectively prevent nucleotide binding. It was previously reported that a lysine in the chick kainate binding protein (cKBP) is specifically involved in guanine nucleotide binding. Therefore we mutated the homologous lysine (K445) in AMPA-R subunit GluR1 plus 12 additional residues around the glutamate binding pocket with the expectation that this would reduce nucleotide binding even further. Nucleotide affinity was determined by measuring the displacement of [3H]fluorowillardiine. As expected, the guanine nucleotide affinity was decreased about five-fold in R1-K445A mutants and the agonist affinity was seemingly unchanged. However, when tested by electrophysiology, characteristics of the mutant such as desensitization and the EC50 for glutamate were found to be altered. None of the other mutations were more successful at decreasing nucleotide affinity selectively. Nonetheless, these studies have given new insight into the docking mode of guanine nucleotides. The loss of binding in R1-K445A was much larger for GTP and GDP than for GMP, and guanosine binding, which is much lower, was unaffected by the mutation. These data suggest that the first phosphate of GMP determines the higher affinity of the phosphorylated nucleotides, and that K445 stabilizes the binding of the second and third phosphates of GDP and GTP. This along with various other observations suggest that the guanine base docks deep within the agonist binding pocket and that bulky additions, such as the phosphates, are accommodated by projecting out of the cleft in the vicinity of lysine 445. However, the exact docking mode of guanine nucleotides would have to be determined by crystallography. Project 2. Agonist binding to AMPA-R in brain consists of a high and low affinity components with KDs of 9-28 nM and 190-700 nM. Previous studies have suggested that newly synthesized receptors have high affinity and are converted to lower affinity by a secondary process. Two particular processes have been implicated, namely the conversion of receptor glycosylation from immature to complex, and modulation by receptor associated proteins. Both hypotheses were evaluated in this project using homomeric receptors GluR1-4 expressed in HEK 293 cells. The role of glycosylation was tested mostly with GluR4 receptors because they are expressed in distinct populations that exhibit either immature or complex glycans and their binding consists of high and low affinity components similar to those previously seen in brain receptors. Cells were treated with castanospermine or deoxymannojirimycin to decrease the proportion of receptors with complex glycosylation, or with cycloheximide plus chloroquine to increase the number of receptors with complex glycosylation. Although 70% of receptors from cells treated with cyloheximide/chloroquine exhibited complex glycans compared to <5% with other treatments, the affinity decreased at most 2-fold. Also, the low affinity component was nearly 80% of the total binding in receptors that exhibited virtually no complex glycans. Taken together these data indicate that complex glycosylation is not the key factor that confers low affinity. To test the second hypothesis GluR1i or GluR2i were co-expressed with stargazin which associates to receptors in neurons and affects their kinetics and trafficking. Considering the affinities of the two components seen in brain, we expected stargazin to cause a 20-fold or greater decrease in binding affinity. This was not the case, however our results did suggest that stargazin caused the appearance of a low affinity component but this was small and remained largely masked by the more abundant high affinity component. Recently, experiments with brain membranes have revealed preliminary evidence that an associated protein of ~85kDa may cause receptors to have low affinity. This hypothesis is currently under investigation. Project 3. Ampakines are cognitive enhancers that potentiate AMPA receptor currents at excitatory synapses. The efficacy of these drugs varies substantially among neurons in different brain regions, being for example about three times larger in the hippocampus than in the thalamus. Binding assays have shown that these compounds also increase the affinity of receptors for agonists. Importantly, the efficacy of these drugs to increase synaptic responses and agonist binding exhibit a positive correlation. Indeed, we have found that the increase in agonist binding (Emax) induced by the prototypical ampakine CX546 is highly variable across eight brain regions and that there is a 3-fold difference between the hippocampus and the thalamus which is similar to the difference reported for physiological efficacy. Therefore, binding assays or receptor autoradiography can potentially be used to predict the physiological efficacy of these drugs in a particular brain region. An important goal of this project has been to identify factors that may be responsible for the regionally different efficacies. Ampakines show some preference for receptor subunits but various considerations suggest that other factors must be involved. In this project we evaluated the role of a novel class of proteins called TARPs (transmembrane AMPA receptor regulatory proteins) that have recently been discovered to be tightly associated with AMPA receptors and to regulate their kinetics. Four of these proteins, named lambda;2(stargazin),λ3,λ4,and λ8 are abundant in the brain, but they exhibit highly selective regional distribution. We determined the maximum increase in agonist binding (Emax ) caused by saturating CX546 in three different AMPA receptor subunits, GluR1i, GluR2i, and GluR4i without and with co-expression of the four TARPs. Without TARPs, both Glu2i and GluR4i showed an Emax value of 100% over baseline binding. Co-expression of TARPs increased the Emax in GluR2i and this was largest for λ3 and λ8 (~130%). However, TARPs decreased the Emax of CX546 in GluR4i and this was most notable with λ2 and λ4 (~72%). Agonist binding in GluR1i was increased by only 15% and it was not significantly changed by TARPs. The expression patterns of TARPs and AMPA-R subunits in the brain have been partially characterized in the literature. Thus, it was previously reported that GluR4i transcripts are abundant in the thalamus but minor in the hippocampus. Using western blots we confirmed that this is also true for protein content; in the thalamus expression of GluR1, GluR2, GluR3, and GluR4 was 4%, 33%, 40%, and 147% respectively, of that in the hippocampus. When considering the known expression patterns of TARP variants, the hippocampus can be described as being enriched in GluR2, λ3 and λ8 while GluR4, λ2 and λ4 are prevalent in the thalamus. In comparison between these specific subunit/TARP combinations, the Emax values for those representative of the hippocampus (GluR2i/λ3 or λ8) were ~2-times larger than the Emax values of thalamic combinations (R4i/λ2 or λ4). Thus we can conclude that the differences in the expression of both TARP variants and AMPA-R subunits are critical factors for determining the variable efficacy of ampakines across brain regions.