Super-resolution Investigations of COPI-coated Vesicles in Thin Sections

COPI-coated vesicles play a central role in the secretory pathway. They have been implicated in protein trafficking since they were first identified as putative secretory transport carriers. The protein contents of COPI vesicles have been extensively debated due to limited approaches for studying individual COPI vesicles. These vesicles are smaller than 250nm, the highest resolution of a conventional light microscope. In the past, the only imaging-based method to study individual vesicles was electron microscopy. Now with advances in super-resolution, the resolution of fluorescence microscopy has been dramatically improved, thus fluorescence microscopy can be applied to studying small cellular structures. I have developed a technique to label, image, and identify individual COPI vesicles within cells. I imaged COPI structures using both localization (PALM/STORM) and STED microscopy, which are super-resolution methods that have approximately 20nm and 70nm lateral resolution respectively. By embedding and cutting 70nm thin sections of cells before imaging them, I additionally improved the axial resolution of my images and confidence that the structures were individual vesicles. The combination of sectioning with STED microscopy yielded images where individual COPI vesicles could be studied and their cargo examined. Using STED imaging of thin sections, I determined if proposed cargos localize to COPI-coated vesicles. To analyze the STED images, I developed a custom computer program to identify COPI vesicles in a STED image by their pixel intensity and shape. A corresponding confocal image of the putative COPI cargo was then examined to determine if the COPI cargo colocalizes with the vesicle. As expected, I found the V-SNARE Bet1 and the COPI scaffold SCYL1 strongly colocalize with COPI vesicles, with 47% and 21.2±5.3% colocalization respectively. Golgi matrix proteins such as GM130 and GRASP65 do not colocalize, with 0.5% and 1.8% colocalization respectively. I found that both the retrograde cargo KDELR, 12.1±3.0% colocalization, and the anterograde cargo hGH, 9.8±2.8% colocalization, localize to a small percent of identified COPI vesicles, but the Golgi glycosylation protein Mannosidasell does not, with 2.0% colocalization. My STED super-resolution microscopy approach supports that COPI vesicles carry both anterograde and retrograde cargo. I examined the localization of gamma1 and gamma2 COP isoforms using the same sectioning and STED imaging approach. An immunoelectron microscopy study by the Wieland lab (Moelleken, 2007) concluded that gamma1 and gamma2 COP are spatially separated at opposing sides of the Golgi stack, with gamma1 enriched at the cis Golgi and gamma2 enriched at the trans Golgi. This lead to the hypothesis that gamma1 and gamma2 are on two different populations of COPI vesicles, which are present at either the cis or trans Golgi. My results disprove this hypothesis by showing that the gamma isoforms colocalize to the same COPI vesicles, with a Pearson's correlation coefficient of 0.7376. This supports that the gamma COP isoforms copolymerize to form a mixed gamma1 and gamma2 COPI coat. While the gamma COP isoforms likely have some differences between them, they do not coat two different populations of vesicles. Lastly I examine if COPI vesicles can bud from unstacked Golgi cisternae, using an inducible FKBP-FRB dimerization system to recruit the Golgi tether GRASP55 to mitochondria. I used STED imaging of thin sections and serial thin sections to determine the location of Golgi proteins after recruiting GRASP55 to mitochondria. I show that inducing GRASP55 recruitment fragmented the Golgi ribbon and separated cis and trans Golgi cisternae. I found that the mitochondrial anchored Golgi cisternae, called "land-locked" Golgi, are capable of trafficking anterograde cargo. Land-locked Golgi are also surrounded by COPI buds and vesicles. Therefore the Golgi stack is not necessary for secretory protein transport