Keijzer, Sandra de
Citation
Keijzer, S. de. (2006, April 12). Dynamics of a GPCR studied with single-molecule microscopy.
Retrieved from https://hdl.handle.net/1887/4363
Version:
Corrected Publisher’s Version
One of the fascinating aspects of life is the cells ability to migrate directionally towards a chemoattractant source. Cell movement orchestrates embryonic development, contributes to tissue repair and regeneration and surveillances our immune response, but it also drives the progression in cancer, mental retardation and chronic inflammatory and vascular diseases. Dictyostelium discoideum chemotaxis towards the ligand cAMP, is a highly conserved mechanism and has proven indispensable as a model organism to investigate the processes underlying chemotaxis. Chemotactic cells are extremely sensitive to chemoattractant gradients; a concentration difference of only 2 % across the cell body can direct movement. The ability to chemotaxis in a very shallow gradient is believed to depend on the spatial restriction of the responses along the anterior-posterior axis as displayed by the localization of certain proteins to either the leading or the trailing edge of the cells. The signaling pathway for chemotaxis in Dictyostelium begins with a receptor-G-protein complex, where cAMP binds and the signal is transduced to the downstream signaling cascade and ends with actin polymerization leading to movement. Somewhere in between, the signal becomes strongly localized. Although many years of research have improved our understanding of chemotaxis more knowledge is necessary to develop therapeutic approaches for treating diseases, cellular transplantations and the preparation of artificial tissue. That is why I took a closer look at the G-protein coupled, cAMP receptor, cAR1, which is, for we know, the most important receptor for chemotaxis in D. discoideum. In the research described in this thesis, I followed individual, fluorescent-labeled cAR1s in the plasmamembrane under different physiological conditions with single-molecule microscopy (SMM).
5.1 SMM in live Dictyostelium cells
Minimal interference of autofluorescence in living Dictyostelium cells is necessary to perform high quality SMM measurements. Autofluorescence can reduce the signal to noise ratio and introduce artifacts, leading to the detection of signals which are not originating from the labeled receptor. The main contribution of Dictyostelium autofluorescence was localized in large structures, vacuoles, and shows an excitation maximum at 514 nm. The autofluorescence was reduced by transferring the cells to a low fluorescent medium, 15-20 hours before measuring. This allowed for the exchange of the highly fluorescent culture-medium with the low fluorescent, less nutrient culture-medium in the intracellular vacuoles. The low fluorescence medium did not trigger the developmental program when used on this timescale. Further reduction was achieved by photobleaching the cells with a short laser pulse (10s) of continuous intensity (2 KW/cm2). The remaining autofluorescence was
filtered out with an algorithm that subtracts an inhomogeneous, time-varying background from the images taken of the cells.
Summary and Discussion
- 91 - receptor in car1- cells was similar to the endogenous receptor and rescued the
developmental defect of the mutant. The reduction of the autofluorescence allowed for the detection of single cAR1-eYFP signals in cells with a positional accuracy of ~40 nm and cut-off criteria validated the signals as individual emitting eYFP molecules.
5.2 Stoichiometry of cAR1
GPCRs are thought to coexist in different stoichiometric states (or aggregate-sizes) in the plasmamembrane and oscillating between these states by association-dissociation equilibrium as part of their normal trafficking and function. These aggregate-sizes can be identified by analyzing the fluorescence intensity distribution of the eYFP signals. In vegetative cells, the receptors were mostly monomeric and only a small percentage was forming higher aggregates. However, the photobleaching procedure made it unreliable to identify the different stochiometric states since most cAR1-eYFP molecules from multimers were bleached. The true aggregate-sizes of cAR1 can be investigated when a monomeric red-fluorescent protein becomes available that is bright enough to do SMM experiments with, because photobleaching is not necessary at higher wavelengths. However, photobleaching is a random process and therefore the unbleached cAR1-eYFP population was still a representative subpopulation of all receptors. That is why it was possible to investigate the effect of chemoattractant on the stoichiometry if the photobleaching procedure was held constant. The stoichiometric states of cAR1-eYFP were compared between different global stimulation times with cAMP and there was no significant effect. In chemotaxing cells, no significant difference in aggregate-sizes of the receptors was found between the leading and trailing edge. Therefore I suggest that ligand-induced aggregation or clustering of cAR1 receptors in the plasmamembrane is not an aspect of their function in D. discoideum chemotaxis.
5.3 Increased mobility of ligand-induced cAR1 at the leading edge
The effect of chemoattractant on another aspect of cAR1-function, its mobility, was studied. The positions of the cAR1-eYFP molecules were tracked with a time resolution of ~44 ms and the trajectories in the plasmamembrane were constructed. The cumulative probability distribution of the square displacements of the single molecules was described with a two-population model. Independent of cAMP stimulation, two populations of receptors were found, an immobile and a mobile receptor with D=0.19 µm2/s. Immobile in
this set-up means that receptors moved with smaller steps than the minimal detectable square displacement of 6.4 10-3µm2.
The mobility was further investigated in cells sensing a cAMP gradient. At the leading edge, the fraction of mobile receptors increased from 38% to 54%, a factor of 1.5. The receptor fractions of immobile and mobile receptors at the trailing edge remained comparable to the control cells. The mobility shift was not caused by a change in membrane viscosity, since the mobility of an inert membrane marker (concanavalin A) was position-independent, nor by a conformational change due to phosphorylation. The latter was assayed with a phosphorylation-deficient mutant, which still showed the cAMP induced mobility shift. Tracking cAR1-eYFP molecules in a gα2- background revealed
that, independent of chemoattractant, the fraction of mobile receptors resembled that of the leading edge in car1- cells. No difference in fraction sizes was observed between the edge
of the cell facing the cAMP source and its opposite side. These two observations suggest that the mobile population found in cAR1-eYFP/car1- cells represent receptors uncoupled
from their associated Gα2 protein. Testing whether the immobile receptor is indeed coupled to Gα2 and the mobile receptor not, can be done with 2-color experiments. Single-molecule dual-color studies would reveal if labeled cAR1 and Gα2 colocalize, when the receptor is immobile.
5.4 A model describing the function of increased cAR1 mobility
One of the crucial questions about chemotaxis is how a cell can sense a very shallow gradient of chemoattractant (<2%) and transduce this into highly polarized intracellular responses. The current models, as described in chapter 1.4, use an excitation and inhibition process in one way or another. In the ‘diffusion-translocation’model (1.4.2), a positive feedback loop results in more effector molecules at the leading edge initiated by the shallow gradient in receptor occupancy along the plasmamembrane. The amplification of the signal can be increased with the production of more second messengers per effector. Here you could also say that the amplification can be increased with more activated effectors per occupied receptor. In the ‘LEGI’model (1.4.3), receptor-occupancy initiates the competing processes of fast excitation and slow inhibition, which controls the membrane binding of PI3K and PTEN. Following the higher receptor occupancy, the fast activation process only exceeds the slow inhibition process at the leading edge. The model assumes the membrane binding of PI3K and PTEN independent processes to acquire high amplification of the extracellular signal. One could argue this assumption with the fact that the activity of PI3K destroys the membrane binding site of PTEN (1.3.3) and membrane localization is required for PTEN activity. In other words, activation and localization of PI3K can be independent of PTEN, via a receptor mediated pathway, but the inhibition of PTEN seems to be mediated via PI3K and therefore not an independent process.
Summary and Discussion
- 93 - receptor-occupancy can not account for the intracellular localized responses. Therefore localization of the proteins must occur after this step. However, there is more to the function of a receptor than localization and its position in the signal transduction pathway. Moreover, only mutant strains that lack either the receptor or the coupled G-protein can not chemotax. Therefore it is tempting to suggest an important role for the receptor in cell polarity.
We showed, with the highly sensitive single-molecule microscopy technique, an increased mobility of cAR1 at the leading edge and this is linked to Gα2 uncoupling. The Gα2 protein in its inactive form is not large enough to immobilize the cAR1 receptor. We propose a model in which the inactive receptor is immobilized by a large protein-protein network and/or cytoskeleton meshwork with the Gα2 protein as linker. Upon cAMP stimulation, Gα2 exchanges its GDP for GTP and uncouples from cAR1 and the βγ subunit of the G-protein. Uncoupling mobilizes the receptor and mobile receptors have the intrinsic ability to find other proteins, Gα2GDP/Gβγ or another signaling molecule, to couple and/or
activate. The collision of receptor and G-protein is stochastically and thus the receptors activate G-proteins in a random walk. The higher mobility of receptors at the leading edge will increase the rate of activation of G-proteins proportionally with the diffusion constant if G-protein activation is solely a diffusion-limited process. When a single occupied receptor activates many G-proteins serially in successive encounters, the mobility produces amplification. The amplification step will lead to a higher local G-protein excitation at the anterior, based on the slightly higher amount of occupied receptors due to the shallow, extracellular cAMP gradient. Since most processes are via a receptor-G-protein mediated pathway, this will lead to the localization of certain proteins to the leading edge and this can be enhanced with the previous mentioned models. There are Gα2βγ-protein-independent pathways which could also be mediated via cAR1 (for example Gα9 or small GTPases) and controlled with this new found mechanism. It would be interesting to see whether cAR1 and G-proteins, and possible other targets of cAR1, reside in microdomains and if these microdomains change size on agonist-stimulation. This would lower the distance between the proteins and therefore the chance of collision increases.
5.5 Phosphorylation dependent internalization of cAR1
cAR1-eYFP molecules in the membrane-compartment and the cytosol. The single-molecule experiments revealed that the fraction of cytosolic receptors increased after persistent, global cAMP stimulation with a t1/2=5 min. This was confirmed in experiments on cell
populations stimulated for different times with the ligand. The internalized receptors can be either recycled back to the plasmamembrane or targeted to lysosomes for degradation. The highly sensitive single-molecule technique showed that part of the internalized cAR1s were degraded. The observed internalization process was abolished in the phosphorylation mutant, cm1234-eYFP, or when cells where stimulated with the antagonist Rp-cAMPS, which binds to cAR1 but does not induce receptor phosphorylayion. This indicates that ligand-induced internalization requires phosphorylation and this result was confirmed by confocal imaging and biochemical assays.
The cAR1-mediated responses to abrupt cAMP increases are transient and return to prestimulus levels in 30s, which is called adaptation. This process is not phosphorylation-dependent and probably controlled via inhibition. Mutant cells lacking PTEN don’t show adaptation and yet they do desensitize to the signal after a few minutes. Therefore, the timescale on which ligand-induced internalization occurs suggests an additional desensitization mechanism. There are some remaining questions to be answered in the future like the role of the cytoskeleton in cAR1 internalization, which can be addressed by monitoring cAR1 in cytoskeleton mutants or in cells treated with an actin-polymerization inhibiting drug. It can also be interesting to see if the internalized receptor is still occupied with cAMP as can be shown with dual color imaging with labeled ligand. The internalized receptor could have intracellular functions. Preliminary data (C.A. Parent, ASCB meeting 2005) showed that adenylyl cyclase A and maybe also CRAC are localized in vesicles. It would be interesting to see if the internalized receptor has any link with these vesicles.
5.6 Concluding remarks
Samenvatting
- 95 -
Nederlandse samenvatting
Één van de meest fascinerende aspecten van het ‘leven’ is de gerichte migratie van cellen naar een bron van signaalstof. De beweging van cellen dirigeert embryonale ontwikkeling, het draagt bij aan weefselherstel en regeneratie en het controleert ons afweersysteem. Maar het is ook de drijvende kracht achter de progressie in kanker, mentale degeneratie en chronische ontstekings- en vaatziekten. De chemotaxis van Dictyostelium
discoideum naar cAMP vertoont grote gelijkenis met de mechanismen van celbeweging in
hogere organismen en heeft zich onmisbaar bewezen als modelsysteem om de processen van chemotaxis te bestuderen. Chemotaxende cellen zijn extreem gevoelig voor gradiënten in de signaalstof; een concentratieverschil van slecht 2% over het cellichaam is voldoende voor de beweging naar de bron. Het vermogen om te chemotaxen is hoogst waarschijnlijk afhankelijk van de locale beperking van respons langs de lichaamsas van de cel, zoals de lokalisatie van bepaalde eiwitten aan de voor- óf achterkant van de cel. De signaalcascade voor Dictyostelium chemotaxis begint bij het receptor-G-eiwit complex, waar de signaalstof aan bindt en het signaal wordt afgegeven aan de intracellulaire cascade, en eindigt bij actine polymerisatie, waardoor beweging plaatsvindt. Ergens tussen deze stappen in de signaalcascade vindt sterke lokalisatie van het signaal plaats. In vele jaren van onderzoek is er veel kennis vergaard over chemotaxis, maar een nog beter begrip is nodig voor de ontwikkeling van therapieën om ziekten te genezen, cellen te transplanteren en kunstmatig weefsel te produceren. Met dit in gedachte heb ik een gedetailleerd onderzoek gedaan naar de cAMP receptor (cAR1), welke, voorzover wij weten, de belangrijkste receptor is voor chemotaxis in D. discoideum. In deze studie, zoals in dit proefschift beschreven, heb ik het gedrag van individuele, fluorescent gelabelde receptoren gevolgd in de plasmamembraan onder verschillende fysiologische omstandigheden met single-molecule miroscopy (SMM).
Om enkele cAR1 moleculen te kunnen volgen in de plasmamembraan, zijn de receptoren gefuseerd met een geel-fluorescent eiwit (eYFP). Dit fusie-eiwit werd ingebracht in cellen die de oorspronkelijke receptor niet meer hebben en het fusie-eiwit bleek dit defect te complementeren en was daarom functioneel. Voordat er SMM metingen gedaan konden worden was het nodig om de optimale condities te vinden waarin enkele moleculen gedetecteerd kunnen worden in levende D. discoideum cellen. Dit hield voornamelijk in de fluorescentie die de cellen zelf hebben te verminderen, zodat dit niet opgepikt kon worden als signaal van de receptor met eYFP. Hierna was het mogelijk om enkele cAR1 moleculen te zien in de plasmamembraan van levende cellen met een nauwkeurigheid van 40 nm.
chemotaxis. Dit geeft aan dat de receptor en het G-eiwit zéér belangrijk zijn en lijkt het intuïtief vreemd dat ze geen rol spelen in de lokalisatie van het signaal.
De functie van een GPCR kan bepaald worden door clustering van meerdere receptoren en dit kan beïnvloed worden door binding van de signaalstof. De grootte van deze clusters kan bepaald worden door naar de sterkte van het fluorescente signaal te kijken. Echter, door het fotobleken was het signaal van de meeste receptoren weg en kon er niet meer gezegd worden hoeveel receptoren er daadwerkelijk in de cluster zaten. Toch kon het effect van signaalstofbinding onderzocht worden door de intensiteit van het fluorescente signaal te vergelijken na verschillende stimulatieperiodes met cAMP. Na globaal toedienen van de signaalstof, dus zonder gradient, was er geen effect te zien op de clustering. En ook in chemotaxende cellen was de clustering niet groter voor receptoren aan de voor- of achterkant. Hieruit concludeer ik dat clustering van receptoren door stimulatie met signaalstof geen rol speelt in chemotaxis van D. discoideum.
Daarentegen was er wel een effect van cAMP stimulatie op de mobiliteit van receptoren. De positie van de moleculen werd gevolgd met een tijdresolutie van 44 ms. Hiermee kon de weg, die de receptor aflegde in de plasmamembraan, vastgelegd worden. Ongeacht of de cellen gestimuleerd werden, waren er altijd twee populaties receptoren te vinden, mobiele en immobiele receptoren. Het bleek dat na cAMP stimulatie er meer mobiele receptoren waren in de voorkant van de cellen. Deze verandering in mobiliteit kwam niet door een verandering in de membraan of een vormverandering van de receptor door phosphorylatie. Het gebruik van mutant-cellijnen toonde aan dat wanneer de receptor ontkoppeld is van het Gα2 subunit van het G-eiwit, hij mobiel is. Gebaseerd op deze vindingen is een model opgesteld. Het G-eiwit is zelf te klein van massa om een receptor te immobiliseren en daarom is het Gα2 subunit een link die de receptor aan het cytoskelet en/of een groot eiwitnetwerk verbindt. Dit kan de receptor wel immobiel maken. Wanneer cAMP aan de receptor bindt, wordt het G-eiwit geactiveerd en ontkoppeld van de receptor. De link is nu weg en de receptor wordt mobiel en kan andere G-eiwitten tegenkomen om te activeren. Één geactiveerde receptor kan meerdere G-eiwitten activeren en dit is amplificatie van het signaal. Aangezien er iets meer cAMP bindt aan de receptoren aan de kant van hoogste concentratie, de voorkant, zal er door de amplificatie snel een verschil optreden in geactiveerde G-eiwitten tussen de voor- en achterkant van de cel. Veel van de chemotaxis signaalcascades worden geactiveerd door het G-eiwit en daardoor zal er lokalisatie van het signaal optreden. Dit leidt er uiteindelijk toe dat het actine aan de kant van de hoogste cAMP concentratie gepolymeriseerd wordt met als gevolg dat de cellen de goede kant uitlopen.
Samenvatting
- 97 - het kan door meerdere oorzaken optreden; de receptoren kunnen verdwijnen uit het membraan en afgeschermd worden van cAMP of van G-eiwitten. Internalisatie van de receptoren kan hieraan bijdragen, echter tot nu toe zijn de bevindingen van cAR1 internalisatie tegenstrijdig. Met de zeer gevoelige SMM techniek was het mogelijk om individuele receptoren te tellen in de plasmamembraan en intracellulair. Het bleek dat met globale cAMP stimulatie de receptoren internaliseerde met een half-waarde tijd van 5 minuten. Bovendien werden receptoren alleen geïnternaliseerd als ze gephosphoryleerd waren. Een deel van de respons op cAMP is tijdelijk en keert snel terug naar de beginwaarden (30s), dit ongevoeligheidproces wordt adaptatie genoemd en wordt waarschijnlijk door een inhibitiecascade veroorzaakt. Mutant-cellijnen die een eiwit missen uit deze inhibitiecascade vertonen geen adaptatie, maar na een paar minuten vertonen ze toch de ongevoeligheid. Daarom suggereert de tijdschaal waarop cAR1 geïnternaliseerd wordt dat dit een tweede mechanisme is wat een rol speelt in de aanpassing aan het signaal.
