Projects

There are several projects running in the lab

Identification of new proteins involved in the secretory pathway

Mechanisms of Golgi-to-PM carrier formation at the TGN

Golgi inheritance during mitosis and the Golgi checkpoint

Current Lab projects:

Identification of new proteins involved in the secretory pathway

Recently, a genome-wide RNAi screen has been carried out in our lab in order to identify Drosophila proteins involved in the secretory pathway (Bard et al. 2006). In this screen, a fusion protein between the signal peptide of Drosophila's BiP chaperone and horseradish peroxidase (SS-HRP) was used as a reporter of constitutive protein secretion in fly cells. Out of the 22,000 analyzed double stranded RNAs (dsRNAs), 1,133 were found to block SS-HRP secretion in a significant manner. However, most of these dsRNAs affected SS-HRP secretion in an indirect way, for instance by targeting genes whose products are involved in gene transcription, mRNA translation, cell survival or basic cell metabolism. Thus, from the original 1,133 dsRNAs we ended up selecting 130 whose targets are potential secretory genes. In fact, 26 of these 130 genes encode proteins known to be necessary for protein transport, thus validating our experimental approach. 

With the remaining 104 fruit fly genes, 77 of which have human orthologs, a secondary RNAi screen was performed, this time in order to assess which of these genes are required for the maintenance of Golgi morphology. To this end, a Drosophila cell line stably expressing a fusion protein between the Golgi enzyme mannosidase II and the green fluorescent protein (ManII-GFP) was transfected with each of these 104 dsRNAs and their effects on Golgi structure were monitored by fluorescence microscopy. Of the 104 studied dsRNAs, 13 caused ManII-GFP to relocate to the endoplasmic reticulum (ER) (class A genes), 19 induced the fragmentation of Golgi membranes into smaller vesicles (class B genes) and 6 caused Golgi membranes to swell and lose its usual cisternal appearance (class C genes). The rest of dsRNAs did not affect Golgi morphology as compared to non-transfected cells (class D genes).

Thus, our screen has identified about a hundred new genes potentially involved in protein trafficking. Of these, about 40% also appear to regulate Golgi morphology, further indicating that they probably are bona fide secretory pathway components. More than twenty of these genes have already been cloned in our lab and their intracellular localization been determined in fly cells. Many localize in the ER, Golgi or in vesicular structures, consistent with their presumed role in protein secretion. Many of the genes we have found had not previously been named, so we have "christened" them as the TANGO genes (Transport ANd Golgi Organization). 

Future research in this area will consist in a more detailed characterization of the individual TANGO genes and their mammalian homologs. Some of the major questions we want to address are: why does depletion of these proteins inhibit secretory traffic? (i.e. what is the function of these proteins?), where are these proteins located in the cell?, at what step in the secretory pathway is protein transport blocked when these proteins are depleted?. Finding answers to these and other related questions should shed much light to the still relatively obscure field of secretory membrane traffic.

Mechanisms of Golgi-to-PM carrier formation at the TGN

The trans-Golgi network (TGN) is one of the most important sorting stations in the secretory pathway. From the TGN, transport carriers are formed that will eventually deliver their cargo to either the apical or basolateral plasma membrane (PM), endosomes, lysosomes, secretory granules, the endoplasmic reticulum or the Golgi cisternae that precede the TGN. Given this large number of potential subcellular destinations, it is no wonder that the detailed mechanisms underlying sorting decisions at the TGN are still poorly understood. Our lab is specifically focused on Golgi-to-PM transport carrier formation. 

Previous research carried out in this lab has already unveiled several important players in Golgi-to-PM traffic. In this respect, a crucial finding was that of the role of protein kinase D (PKD) in regulating the fission of PM-bound carriers at the TGN (Liljedahl et al. 2001). Part of the importance of this discovery was that it shed light on the biological relevance of previous data obtained from in vitro Golgi fragmentation assays. As a matter of fact, PKD itself had been found to be required to fragment the Golgi in these assays, which it did after being activated by the beta-gamma subunit of a heterotrimeric G protein (Jamora et al. 1997 & 1999). Therefore, it was now possible to devise a working model of Golgi-to-PM carrier formation according to which the activation in the TGN of a heterotrimeric G protein would lead to the dissociation of its alpha and beta-gamma subunits, whereupon the latter would go on to activate PKD, which would in turn induce the fission of the carrier from the TGN, thus allowing its transport to the PM. 

Needless to say, this is a very simplified model and all our work since its formulation has gone in the direction of closing some of its gaps. For instance, more recent research has brought new players into the game, including the lipid diacylglycerol, which binds PKD and recruits it to the TGN (Baron & Malhotra 2002) and PKC-eta, which appears to be acting downstream of beta-gamma but still upstream of PKD (Diaz-Añel & Malhotra 2005). Our current research is aimed at characterizing still other proteins or lipids that take part in this process, as well as the interactions between all the components involved. 

In any case, there are still many important questions awaiting an answer. After all, before PKD can induce the fission of carriers at the TGN (how it does so is not yet clear, though it appears to involve synthesis of another phospholipid, PI4P (Hausser et al. 2005)), these carriers must have budded from the TGN, which presumably involves both lipids and proteins that can generate the needed membrane curvature, and must also have been filled with the appropriate cargo. Another unanswered question is what activates the above-mentioned G protein: an appealing possibility is the existence of a Golgi-localized, cargo-activated G protein-coupled receptor. However, this receptor, if it indeed exists, has so far remained elusive.

In summary, understanding the mechanisms whereby Golgi-to-PM carriers bud from the TGN, are filled with cargo and are detached and how these different events are coordinated with each other so that they occur at the right time and to the desired extent represents a magnificent scientific challenge. It is our belief that a reductionist approach based on the identification of all the major players as well as their mutual interactions will eventually lead to the solution of this problem.

Golgi inheritance during mitosis and the Golgi checkpoint.

The mechanisms of Golgi membrane inheritance during mitosis are the subject of intensive research as well as of ongoing controversy (Colanzi et al. 2003a; Shorter & Warren 2002). Such mechanisms have been shown to vary considerably in different eukaryotic organisms (Rossanese & Glick 2001; Pelletier et al. 2002; He et al. 2004 & 2005), but here we will focus on what happens in mammalian cells. The mammalian Golgi apparatus is inherited in three consecutive steps, namely (1) Golgi fragmentation during late prophase and metaphase, (2) partitioning of the resulting Golgi haze into daughter cells during anaphase and (3) Golgi reassembly in telophase.

In our lab, we focus mainly in the first of these steps, i.e. mitotic Golgi fragmentation. This is a complex process involving two distinct phases. In the first one, the lateral connections between Golgi stacks forming the so-called Golgi ribbon are broken and the resulting individual stacks become dispersed throughout the cytoplasm. In the second, Golgi cisternae further fragment and give rise to the mitotic Golgi haze, which some claim arises as a result of Golgi proteins being absorbed into the ER (Altan-Bonnet et al. 2006; Reinke et al. 2004), while others consider it a highly fragmented, ER-independent form of Golgi (Pecot & Malhotra 2004; Axelsson & Warren 2004). Of these two fragmentation phases, only the first one has been shown to be essential for mitotic progression and Golgi inheritance, which has in turn led to the concept of the Golgi checkpoint (Sütterlin et al. 2002). According to this novel concept, cells check the status of their Golgis during the late G2 phase of the cell cycle and only enter cell division once they have ascertained that the Golgi ribbon has fragmented and the stacks have moved away from the pericentriolar region. Thus, only cells that are prepared to equitably partition their Golgi into daughter cells enter into mitosis.

In order to identify components involved in mitotic Golgi fragmentation (MGF), we have previously developed an assay that reconstitutes this process in vitro (Acharya et al. 1998). This assay, which involves incubating purified Golgi membranes with mitotic cytosol and an ATP-regenerating system, has already been used to identify several kinases (Raf1, MEK1 and Plk1) that are needed for MGF to occur (Acharya et al. 1998; Sütterlin et al. 2001; Colanzi et al. 2003b). Also found with this assay to participate in MGF was GRASP65, a peripherally-attached protein of the cis-Golgi membranes. In this case, though, rather than finding an inhibition of MGF upon depletion of GRASP65, what was found was that addition of exogenous GRASP65 inhibits MGF, thus suggesting that GRASP65, rather than being required for MGF, is actually preventing it from happening. Indeed, a recent report has shown that GRASP65 is required for the stability of the Golgi ribbon (Puthenveedu et al. 2006). This, together with the fact that GRASP65 is a known phosphorylation target during mitosis (Preisinger et al. 2005), suggests a model whereby mitotic phosphorylation of GRASP65 is needed in order to break the Golgi ribbon and thus enter mitosis. Currently, our lab is trying to expand this model through the identification of new components involved in MGF and a finer characterization of all the relationships between all the participants in this process.

References: