Published January 10, 2011 People & Ideas

Sean Munro: Revealing the Golgi’s true identity Munro studies how proteins localize within the secretory pathway.

lthough cargo proteins move I learnt about the revolu- through the different organelles of tion that was happening, and that all seemed A the secretory pathway, other pro- very exciting. I studied at teins remain in place to give each compart- Oxford as an undergraduate. Molecular bi- ment its own unique identity and function. ology hadn’t quite reached Oxford yet, but , from the MRC Laboratory of there was a small amount going on, and I Molecular Biology in Cambridge, England, became interested in gene . has been interested in how proteins ! nd their place within the secretory pathway since his How did that interest lead you to join PhD studies with Hugh Pelham, when he Hugh Pelham’s lab for your PhD? discovered a heat-shock protein that was re- He was working on the transcription of tained in the ER lumen due to a speci! c heat-shock genes at that time. But he was Downloaded from “KDEL” sequence at its C terminus (1). trying to diversify a bit when I got there, so OF MRC CAMBRIDGE IMAGE COURTESY Sean Munro After a brief detour into receptor biology he said, “Why don’t you work on heat-shock as a postdoc with Tom Maniatis, Munro be- proteins and try to understand what they’re gan his own lab and as an independent group actually doing?” He wanted me to study that a protein is not randomly distributed— leader moved from studying the ER to the Hsp70, but a competing lab was reluctant to it’s clearly in a speci! c place. That tells you Golgi. He’s demonstrated that peripheral send us antibodies. So Hugh had to devise that the protein has a signal on it, and that Golgi proteins are targeted to the organelle another way of studying the protein and there’s something that recognizes that signal. jcb.rupress.org through a combination of lipid and small came up with the idea of epitope tagging. It’s a particularly fascinating issue for pro- GTPase binding (2, 3), where- My ! rst paper showed that teins in the endomembrane system, where as Golgi-resident enzymes de- this approach worked. It took proteins are constantly moving around. Resi- pend on their transmembrane “My intention a while to catch on as a tech- dent proteins have to remain in place whilst

domains for their localization wasn’t to nique, but eventually it spread cargo proteins traf! c through the system. on December 13, 2012 (4). More recently, Munro con- say that rafts very widely. ! rmed a long-held theory of I then used epitope tagging LOCALIZING TO THE GOLGI Why focus on the Golgi? his that transmembrane do- don’t exist, to identify and localize new I was looking around for an interesting cell main length correlates with the but that one members of the Hsp70 family biological problem, and there were two localization of integral mem- and found one—BiP/Grp78— has to be things that inspired me. One was cell move- brane proteins (5), possibly that localized to the lumen of ment, so I started to look at neutrophil due to variations in lipid com- cautious.” the ER and was associated chemotaxis. During my postdoc with Tom position that alter membrane with unassembled immuno- Maniatis I’d developed methods to clone re- thickness. He has also proposed that Golgi- globulin heavy chains. This provided early ceptors, so I tried to identify the chemotac- resident coiled-coil proteins act as “tentacles” evidence for a role of heat-shock proteins in tic receptor. I accidentally cloned some that catch incoming vesicles (6) and has protein biogenesis. We also wondered how other neutrophil receptors instead, one of poked holes in the “lipid raft” theory of mem- BiP/Grp78 was retained in the ER as it lacked which turned out to be related to the brain brane microdomains (7). In a recent inter- a transmembrane domain. Luckily, it was the receptor for cannabinoids. I spent quite a view, Munro discussed his scienti! c identity second lumenal ER protein to be cloned and long time proving that this was a second and his targets for the future. sequenced, and we noticed this conserved cannabinoid receptor, but I stopped working KDEL sequence at the C terminus, which I EARLY IDENTITY on it after that ! rst paper. Ironically, that’s then showed was an ER retention signal. Where did you grow up? still my most cited independent paper be- I was born in Cambridge, England, and I’ve So that was the beginning of your cause of the huge pharmaceutical interest in lived here most of my life. My dad, Alan, interest in organelle identity and protein cannabinoid derivatives, and it’s the only was a cellular immunologist at Cambridge localization… paper I’ve ever published that the media and University and then in biotech before he re- Exactly. It really introduced me to the ! eld my non-scientist friends were interested in. tired. He was certainly a strong in" uence on of , which at the time was only So perhaps I should’ve stayed with that, me, but I was always interested in biology. I just starting to become molecular. I’m fasci- but it wasn’t clear where it was going to lead collected caterpillars and went ! shing and nated by the whole question of how a cell is at that point. In the meantime, I’d been look- bird watching as a boy. organized and how proteins localize to dif- ing at these Golgi enzymes. They were just I decided that I didn’t want to go into ferent parts of the cell, because it’s very visu- beginning to be cloned, and I wondered how immunology like my dad. But through him ally appealing. You can look at cells and see they were targeted to the Golgi. No one had

4 JCB • VOLUME 192 • NUMBER 1 • 2011 Published January 10, 2011

Text and Interview by Ben Short [email protected]

to say that rafts don’t exist, but that one has and genetics to look at their function, to be cautious. I think it was useful because mostly in Drosophila. it forced the ! eld to realize that there were Then we’re trying to work out how trans- problems with some of their methodologies, membrane domains get sorted by length. particularly detergent resistance. I think raft One thing I’ve become interested in is the experts appreciated the problem, but a lot of idea that proteins are at a very high density in people had picked up this method and biological membranes—much higher than in

IMAGE COURTESY OF KATJA RÖPER OF KATJA IMAGE COURTESY thought it provided an unequivocal way of the textbook cartoon of the odd potato " oat- Munro grapples with the Golgi. showing that a protein was clustered in ing around in a sea of lipids. So the proteins microdomains at the cell sur- themselves may have a pro- done anything on that. We found quite face. I think the review en- “Golgins… found effect on a bilayer’s quickly that the transmembrane domain couraged people to be a bit thickness and other physical (TMD) was involved in their retention in more critical and use better surround the properties. What’s appealing the Golgi, and I’ve been working on Golgi imaging methods to look for Golgi in a about this is that maybe you protein targeting ever since. protein clustering. network of could sort lipids by collecting I think the original idea of proteins of certain properties Downloaded from You recently validated an idea you’ve rafts as relatively large, long- tentacles.” together. So, protein cluster- had for a long time about TMD length lived platforms sitting in the ing during vesicle formation and protein targeting… plasma membrane isn’t really tenable any- could explain quite a lot about lipid sorting, I noticed that Golgi protein TMDs seemed more. But the idea has evolved a bit—the which is poorly understood at the moment. shorter than those of similar proteins that current model seems to be that rafts are very went to the cell surface. There weren’t that Can you imagine what you might be if small and very transient. That may be the

many genes cloned at the time, but from the you weren’t a scientist? jcb.rupress.org case, but I don’t think that’s been de! nitive- small data set we could collect, the differ- My mum worked as a careers advisor, and ly proven either. So I remain skeptical. ence appeared to be statistically signi! cant. she was quite keen that I didn’t become a sci- So Mark Bretscher and I published this idea FUTURE TARGETS entist. I think she got a bit fed up waiting for and discussed how it might relate to changes What about peripheral membrane my dad to come home late from the lab every in lipid composition and bilayer thickness. protein targeting? night. So she encouraged me to become a And then I did some experiments with syn- That’s become the main focus of my lab. barrister, because I always liked arguing with on December 13, 2012 thetic TMDs of varying lengths that targeted We’re particularly interested in golgins— her. But really I feel lucky that I went into proteins to different places. coiled-coil proteins that localize to the cell biology at a time when the molecular Now, we’ve gone back and analyzed Golgi—because there are a lot of them, and revolution happened. There’s been such much bigger data sets, and it’s clear that people have suggested they have various dif- amazing progress since I ! rst started in sci- the original differences in length that we ferent functions. We identi! ed the GRIP ence. I’m lucky to have been born at the right observed are absolutely true. We also found domain, which targets several coiled-coil time to have taken part in it. differences in the composition proteins to the Golgi by binding the GTPase of TMDs in proteins targeted to different Arl1. We recently found that many golgins 1. Munro, S., and H.R. Pelham. 1987. Cell. 48:899–907. organelles, which may contribute to pro- also contain multiple binding sites for Rab 2. Levine, T.P., and S. Munro. 2002. Curr. Biol. tein sorting as well. GTPases. These don’t seem to be involved in 12:695–704. But we’ve yet to prove that the plasma Golgi targeting, but we think it allows gol- 3. Panic, B., et al. 2003. Curr. Biol. 13:405–410. membrane really is thicker and that this is gins to capture membranes containing dif- 4. Munro, S. 1991. EMBO J. 10:3577–3588. the mechanism of protein sorting. It’s a very ferent Rabs. We propose that the golgins act 5. Sharpe, H.J., et al. 2010. Cell. 142:158–169. tricky thing to do because you can’t manip- together to surround the Golgi in a network 6. Sinka, R., et al. 2008. J. Cell Biol. 183:607–615. ulate lipid composition in cells without in- of tentacles studded with Rab-binding sites, 7. Munro, S. 2003. Cell. 115:377–388. ducing many indirect effects. The other to capture incoming vesicles or other Golgi complication is that cholesterol and sphin- cisternae. I actually wanted to call it a furry golipids—which can change bilayer thick- Golgi, but the reviewers didn’t like that! ness—have also been suggested to form lipid rafts at the plasma membrane. That’s What are you working on at the moment? slightly incompatible with our view in We’re working on these Golgi tentacles, which the plasma membrane is essentially trying to show that that’s what golgins are one big raft that’s thick and highly ordered. really doing. In addition, having found these Rab-binding sites on the golgins,

Is that why you’re skeptical of lipid rafts? we’ve become interested in some of the IMAGE COURTESY OF ALISON GILLINGHAM AND SEAN MUNRO Yes. A few years ago I wrote a review that less-well characterized members of the The golgin giantin (red) surrounds Golgi was critical of the ! eld. My intention wasn’t Rab family. So we’re doing biochemistry stacks (green).

People & Ideas • THE JOURNAL OF CELL BIOLOGY 5