**3. PGC transition to the target membrane**

After fission from the TGN, PGCs move to their target membranes. Given that budding PGCs are associated with motor proteins, these can dock onto microtubules and use them as the 'highways' to reach their destination. In this context, different members of the kinesin superfamily (Kamal et al., 2000; Nakata and Hirokawa, 2003; Teng et al., 2005), or even other microtubule motors, such as dynein (Tai et al., 1999), have been shown to drive post-Golgi transport of specific cargo towards their acceptor compartment. Such high fidelity of cargo selection by motors at the TGN, and as a consequence, its further delivery to the correct surface or intracellular domain, might be regulated by interactions of motor proteins directly with the cargo (Kamal et al., 2000; Teng et al., 2005) or with components of the TGN sorting machinery (Nakagawa et al., 2000). For instance, transport of HA and annexin 13b to the apical PM surface in epithelial cells relies on the raft-associated motor protein KIFC3 (Noda et al., 2001). KIF13A has been shown to operate in the other post-Golgi route that directs the mannose-6-phosphate receptor from the Golgi complex to the endosomes (Nakagawa et al., 2000). A number of neuronal proteins, such as bAPP, GAP43 and SNAP25, require KIF5 for their correct targeting (Nakata and Hirokawa, 2003). The microtubule minus-end-directed motor dynein has been described as supporting rhodopsin transport in rod photoreceptors (Tai et al., 1999; Yeh et al., 2006).

Of note, the sorting of specific cargo to either axons or dendrites by different kinesins has been demonstrated in a single individual neuron (Nakata and Hirokawa, 2003). KIF5 has been shown to carry VSVG-containing PGCs to the axon, while KIF17A provides the delivery of the Kv2.1 ion channel to dendrites (Nakata and Hirokawa, 2003). In such cases,

those containing tubular networks and branching tubules, or in thick vacuolar regions). Obviously, the precise points of fission will define not only the composition of a PGC carrier, but also its morphology. If fission occurs close to the tip of the TGN tubule, a carrier will be smaller in size. In contrast, larger PGCs can form by cleavage at the base of a PGC precursor (Fig. 3A). Similarly, endosome-directed PGCs can apparently detach from the TGN as simple cargo-containing vesicles if the fission occurs at the neck of the clathrincoated buds (Fig. 3B). However, many clathrin-positive PGCs have a grape-like morphology (a tubule with several 'buds'), which suggests that entire chunks of TGN membranes that contain 2-3 clathrin-coated buds can be cleaved from the Golgi complex (Polishchuk et al.,

Fission along a budding PGC precursor might also be greatly facilitated by heterogeneous lipid microdomains (Roux et al., 2005). Indeed, liposome tubules have been shown to break into small pieces at the border between phosphatidylcholine-enriched and cholesterolenriched regions (Roux et al., 2005). Given that multiple lipid microdomains can be seen along a forming PGC, it appears that large post-Golgi carriers might contain several regions

Finally, fission of PGCs might be regulated by the cargo proteins themselves. According to a recent hypothesis, in some cases fission appears not to occur until the cargo is completely loaded into the budding tubule (Bard and Malhotra, 2006). For example, if a large and rigid procollagen rod is still present in the neck of a budding tubule, this neck can apparently not

After fission from the TGN, PGCs move to their target membranes. Given that budding PGCs are associated with motor proteins, these can dock onto microtubules and use them as the 'highways' to reach their destination. In this context, different members of the kinesin superfamily (Kamal et al., 2000; Nakata and Hirokawa, 2003; Teng et al., 2005), or even other microtubule motors, such as dynein (Tai et al., 1999), have been shown to drive post-Golgi transport of specific cargo towards their acceptor compartment. Such high fidelity of cargo selection by motors at the TGN, and as a consequence, its further delivery to the correct surface or intracellular domain, might be regulated by interactions of motor proteins directly with the cargo (Kamal et al., 2000; Teng et al., 2005) or with components of the TGN sorting machinery (Nakagawa et al., 2000). For instance, transport of HA and annexin 13b to the apical PM surface in epithelial cells relies on the raft-associated motor protein KIFC3 (Noda et al., 2001). KIF13A has been shown to operate in the other post-Golgi route that directs the mannose-6-phosphate receptor from the Golgi complex to the endosomes (Nakagawa et al., 2000). A number of neuronal proteins, such as bAPP, GAP43 and SNAP25, require KIF5 for their correct targeting (Nakata and Hirokawa, 2003). The microtubule minus-end-directed motor dynein has been described as supporting rhodopsin transport in

Of note, the sorting of specific cargo to either axons or dendrites by different kinesins has been demonstrated in a single individual neuron (Nakata and Hirokawa, 2003). KIF5 has been shown to carry VSVG-containing PGCs to the axon, while KIF17A provides the delivery of the Kv2.1 ion channel to dendrites (Nakata and Hirokawa, 2003). In such cases,

that are filled with raft and non-raft proteins (Polishchuk et al., 2004) (Fig. 3C, D).

be constricted sufficiently (by dynamin or coat proteins) to trigger fission.

**3. PGC transition to the target membrane** 

rod photoreceptors (Tai et al., 1999; Yeh et al., 2006).

2006).

how do the different motors know where to deliver these specific cargoes? Apparently motor heads of KIF5 and KIF17A can bind with higher affinity to different subsets of microtubules, which will provide directional cues for polarized axonal transport (Nakata and Hirokawa, 2003). Different populations of microtubules have been found also in other cell types, and therefore, these might serve as highways for polarized cargo delivery (Nakata and Hirokawa, 2003; Spiliotis et al., 2008).

In this context, microtubule architecture appears to play a significant role in targeting of TGN-derived PGCs to their correct acceptor membrane(s). Several molecular players, such as Par-1 (Cohen et al., 2004) and septins (Spiliotis et al., 2008), have been shown to regulate microtubule organization in vertical or horizontal arrays during the polarization of columnar (MDCK cells) or planar (hepatocytes) epithelia, respectively. As a consequence of the divers microtubule architecture, same apical proteins can be delivered from the Golgi complex directly to the correct surface domain of MDCK cells, while in hepatocytes, these apical markers first appear at the basolateral surface, and then transcytosis is used for their apical delivery (Cohen et al., 2004).

The other important issue that needs to be addressed is whether any sorting processes take place in the PGCs *en route* to their acceptor compartment. This occurs, for example, when the mannose-6-phosphate receptor is sorted from the maturing secretory granules by clathrincoated vesicles (Klumperman et al., 1998). So several approaches have been used to determine whether similar sorting events occur with PGCs. Mature TGN-derived carriers can be arrested before their fusion with the PM, either by microinjection of an anti-NSF antibody or by treatment with tannic acid (Polishchuk et al., 2003; Polishchuk et al., 2004). In contrast to secretory granules, a comparison of mature with newly formed PGCs did not reveal significant variations in either their ultrastructure or their molecular composition (Polishchuk et al., 2003; Polishchuk et al., 2004). Similarly, mature carriers directed from the Golgi complex to endosomes accumulated in cells upon endosome ablation. However, they did not show any significant transformation, except for a slight reduction in the area covered by clathrin (Polishchuk et al., 2006).

Live-cell imaging of MDCK cells has revealed that PCGs that contain both the basolateral marker VSVG-CFP and the apical marker GPI-YFP do not sort out each of these proteins into separate structures, but deliver both of these proteins to the PM (Polishchuk and Mironov, 2004). GPI-YFP is then sorted from the basolateral surface to the apical surface through transcytosis. On the other hand, segregation of proteins from their common PGC into two separate carriers has also been documented (Jacob and Naim, 2001). This suggests that sorting from the PGC does happen, but that it is likely to depend on the nature of the transported cargo proteins.

The complexity of sorting events in the post-Golgi space became more evident with discovery that certain cargoes can pass through an endosomal intermediate before their arrival at the PM. Such indirect 'through-endosome' trafficking of cargo to the cell surface might be significantly facilitated by close association of the TGN membranes with a number of the endocytic compartments in the perinuclear area of a cell (Pavelka et al., 1998). The list of the proteins using this pathway has been updated recently, and it has now been shown that in MDCK epithelial cells, VSVG, the LDL receptor, and E-cadherin can be detected in endosomes before their exit to the PM (Ang et al., 2004; Lock and Stow, 2005). These findings, however, raise a number of further questions. The first is whether this indirect

Morphogenesis and Dynamics of Post-Golgi Transport Carriers 33

with basolateral markers within the same carriers docked onto the lateral membranes of the MDCK cells (Polishchuk et al., 2004). Only after their arrival at the basolateral surface of the PM were these proteins sorted to the apical domain of the PM by transcytosis (Polishchuk et al., 2004). However, after 4 days in culture, MDCK cells apparently start to switch delivery of GPI-anchored proteins from a transcytotic to direct route (Paladino et al., 2006). Similarly, in thyroid epithelial cells, the delivery of dipeptidylpeptidase-IV to the apical PM surface changes from transcytosis to a direct route as the epithelial monolayer matures (Zurzolo et al., 1992). This might happen because the target patch for the PGC fusion forms at the lateral surface of epithelial cells earlier then at the apical targeting patch (Mostov et al., 2003; Nelson, 2003; Rodriguez-Boulan et al., 2005). Indeed, Sec6 undergoes recruitment to the sites of cell-to-cell junctions as soon as subconfluent cells start to contact each other (Yeaman et al., 2004), while expression of the apical sorting machinery components occurs later during the process of cell polarization (Halbleib et al., 2007). Thus, during the early stages of polarization, most of PGCs fuse near junctional complexes at the lateral domain of the PM (Kreitzer et al., 2003; Polishchuk et al., 2004), which contributes to fast vertical elongation of a cell within the epithelial sheet (Mostov et al., 2003; Nelson, 2003; Rodriguez-Boulan et al.,

The precise spatial targeting of PGCs to the correct PM area might be important for processes that contribute to correct tissue development, such as, for example, the parallel alignment of collagen fibers in a tendon (Canty et al., 2004). The cellular mechanism of this alignment is thought to involve the assembly of intracellular collagen fibrils within PGCs. The PGCs carrying procollagen subsequently connect to the extracellular matrix via fingerlike projections of the PM, which are known as fibripositors, and which are oriented along the axis of the tendon (Canty et al., 2004). Interestingly, actin filament disassembly results in the rapid loss of fibripositors and in the subsequent disappearance of intracellular fibrils. In this case, a significant proportion of collagen fibrils are found to no longer be orientated with the long axis of the tendon. This suggests an important role for the actin cytoskeleton in the alignment of PGC delivery and in the further organization of the collagenous

In brain tissue, transformation of the contact between an axon and a dendrite into a synapse is accompanied by the accumulation of the synaptic machinery at the site, with these delivered in TGN-derived carriers. In cultured hippocampal neurons, PGCs are linked via spectrin to clusters of the neural cell adhesion molecule (NCAM) in the PM. These complexes are trapped at sites of initial neurite-to-neurite contact within several minutes of the formation of the initial contact. The accumulation of PGCs at contacts with NCAMdeficient neurons is reduced when compared with wild-type cells, which suggests that NCAM mediates the anchoring of intracellular organelles in nascent synapses (Sytnyk et al.,

The extensive characterization of PGC morphology by the combination of live-cell imaging and EM has provided significant advances in our understanding of the mechanisms that operate during PGC morphogenesis and the other steps of the PGC lifecycle. It appears that the type and size of a cargo can strongly impact on the architecture of a PGC, its path to the

extracellular matrix in the embryonic tendon (Canty et al., 2006).

2005).

2002).

**5. Concluding remarks** 

transport route exists in different cells. The second is whether different cargoes move through the same endosomal compartment on their way to the cell surface in epithelial cells. This appears not to be a case, as a number of proteins (such as VSVG and the LDL receptor) have been reported to use a Rab8-positive sub-population of endosomes as an intermediate station on their way to the basolateral PM in epithelial cells (Ang et al., 2004), while others cargoes (such as E-cadherin, for example) move to the PM through a Rab11-containing endocytic compartment (Lock and Stow, 2005). Furthermore, ablation of the different endocytic compartments through horseradish peroxidase (HRP)-catalyzed crosslinking has revealed a number of apical proteins to take various through-endosome routes to reach the cell surface (Weisz and Rodriguez-Boulan, 2009). It remains to be understood, however, whether any cross-talk between these routes exists. Unfortunately, detailed characterization of PGCs that operate either to or from these intermediate endocytic stations is still missing. Moreover, directionality and selectivity of the post-Golgi routes, and therefore the PGC properties, might change upon cell differentiation. For example, at the early stages in MDCK cells, GPI-anchored proteins are transported to the basolateral surface of the PM, from where that tend to be transcytosed to the apical surface (Polishchuk et al., 2004). Later, however, the cells tend to divert GPI-anchored proteins into a direct Golgi-to-apical surface route (Paladino et al., 2006). Further efforts need to be made to understand this interplay of the different mechanisms that define the PGC path from the TGN to the target compartment.

### **4. Docking and fusion of PGCs with acceptor membranes**

To complete their movement across a cell, PGCs need to fuse with their acceptor membrane to deliver their contents. Docking and fusion of PGCs with the cell surface has been studied by total internal reflection (TIRF) microscopy, which allows the visualization of intracellular events within very narrow (50-150 nm) distance from the PM. This TIRF analysis demonstrated that after the docking heterogeneously sized PGCs, these usually fuse completely with the PM (Schmoranzer et al., 2000; Toomre et al., 2000). However, occasionally, larger tubular PGCs can fuse with the PM only partially, at their tips (Schmoranzer et al., 2000; Toomre et al., 2000), using a kind of 'kiss-and-run' mechanism. Interestingly, some PGCs also remain attached to microtubules even as their fusion with the PM initiates (Schmoranzer and Simon, 2003).

PGC fusion with the PM might not be randomly distributed, and might instead be concentrated at several 'hot spots' along the PM (Toomre et al., 2000). In general, the delivery of the PGCs is frequently directed to rapidly growing membrane surfaces. In motile cells, this process is restricted to the leading edge of the cell (Polishchuk et al., 2004; Schmoranzer et al., 2003). Visualization of exocytosis in polarized cells has revealed that the vertical growth of epithelia relies on the directed delivery of PGCs to the lateral surface of the columnar cells, where the tethering factors, such as mammalian exocyst or *Drosophila* DLG, reside (Kreitzer et al., 2003; Lee et al., 2003; Polishchuk et al., 2004). Similarly, the exocyst can define PGC docking sites in neurite growth cones of differentiating neurons (Vega and Hsu, 2001). Interestingly, during tissue biogenesis, a number of cells (e.g. epithelial cells) show an incredible flexibility in terms of their transport routes (Mostov et al., 2003; Rodriguez-Boulan et al., 2005). In epithelia growing on a filter support for 2-3 days, GPI-anchored proteins (which are normally apically targeted) have been found together

transport route exists in different cells. The second is whether different cargoes move through the same endosomal compartment on their way to the cell surface in epithelial cells. This appears not to be a case, as a number of proteins (such as VSVG and the LDL receptor) have been reported to use a Rab8-positive sub-population of endosomes as an intermediate station on their way to the basolateral PM in epithelial cells (Ang et al., 2004), while others cargoes (such as E-cadherin, for example) move to the PM through a Rab11-containing endocytic compartment (Lock and Stow, 2005). Furthermore, ablation of the different endocytic compartments through horseradish peroxidase (HRP)-catalyzed crosslinking has revealed a number of apical proteins to take various through-endosome routes to reach the cell surface (Weisz and Rodriguez-Boulan, 2009). It remains to be understood, however, whether any cross-talk between these routes exists. Unfortunately, detailed characterization of PGCs that operate either to or from these intermediate endocytic stations is still missing. Moreover, directionality and selectivity of the post-Golgi routes, and therefore the PGC properties, might change upon cell differentiation. For example, at the early stages in MDCK cells, GPI-anchored proteins are transported to the basolateral surface of the PM, from where that tend to be transcytosed to the apical surface (Polishchuk et al., 2004). Later, however, the cells tend to divert GPI-anchored proteins into a direct Golgi-to-apical surface route (Paladino et al., 2006). Further efforts need to be made to understand this interplay of the different mechanisms that define the PGC path from the TGN to the target

**4. Docking and fusion of PGCs with acceptor membranes** 

PM initiates (Schmoranzer and Simon, 2003).

To complete their movement across a cell, PGCs need to fuse with their acceptor membrane to deliver their contents. Docking and fusion of PGCs with the cell surface has been studied by total internal reflection (TIRF) microscopy, which allows the visualization of intracellular events within very narrow (50-150 nm) distance from the PM. This TIRF analysis demonstrated that after the docking heterogeneously sized PGCs, these usually fuse completely with the PM (Schmoranzer et al., 2000; Toomre et al., 2000). However, occasionally, larger tubular PGCs can fuse with the PM only partially, at their tips (Schmoranzer et al., 2000; Toomre et al., 2000), using a kind of 'kiss-and-run' mechanism. Interestingly, some PGCs also remain attached to microtubules even as their fusion with the

PGC fusion with the PM might not be randomly distributed, and might instead be concentrated at several 'hot spots' along the PM (Toomre et al., 2000). In general, the delivery of the PGCs is frequently directed to rapidly growing membrane surfaces. In motile cells, this process is restricted to the leading edge of the cell (Polishchuk et al., 2004; Schmoranzer et al., 2003). Visualization of exocytosis in polarized cells has revealed that the vertical growth of epithelia relies on the directed delivery of PGCs to the lateral surface of the columnar cells, where the tethering factors, such as mammalian exocyst or *Drosophila* DLG, reside (Kreitzer et al., 2003; Lee et al., 2003; Polishchuk et al., 2004). Similarly, the exocyst can define PGC docking sites in neurite growth cones of differentiating neurons (Vega and Hsu, 2001). Interestingly, during tissue biogenesis, a number of cells (e.g. epithelial cells) show an incredible flexibility in terms of their transport routes (Mostov et al., 2003; Rodriguez-Boulan et al., 2005). In epithelia growing on a filter support for 2-3 days, GPI-anchored proteins (which are normally apically targeted) have been found together

compartment.

with basolateral markers within the same carriers docked onto the lateral membranes of the MDCK cells (Polishchuk et al., 2004). Only after their arrival at the basolateral surface of the PM were these proteins sorted to the apical domain of the PM by transcytosis (Polishchuk et al., 2004). However, after 4 days in culture, MDCK cells apparently start to switch delivery of GPI-anchored proteins from a transcytotic to direct route (Paladino et al., 2006). Similarly, in thyroid epithelial cells, the delivery of dipeptidylpeptidase-IV to the apical PM surface changes from transcytosis to a direct route as the epithelial monolayer matures (Zurzolo et al., 1992). This might happen because the target patch for the PGC fusion forms at the lateral surface of epithelial cells earlier then at the apical targeting patch (Mostov et al., 2003; Nelson, 2003; Rodriguez-Boulan et al., 2005). Indeed, Sec6 undergoes recruitment to the sites of cell-to-cell junctions as soon as subconfluent cells start to contact each other (Yeaman et al., 2004), while expression of the apical sorting machinery components occurs later during the process of cell polarization (Halbleib et al., 2007). Thus, during the early stages of polarization, most of PGCs fuse near junctional complexes at the lateral domain of the PM (Kreitzer et al., 2003; Polishchuk et al., 2004), which contributes to fast vertical elongation of a cell within the epithelial sheet (Mostov et al., 2003; Nelson, 2003; Rodriguez-Boulan et al., 2005).

The precise spatial targeting of PGCs to the correct PM area might be important for processes that contribute to correct tissue development, such as, for example, the parallel alignment of collagen fibers in a tendon (Canty et al., 2004). The cellular mechanism of this alignment is thought to involve the assembly of intracellular collagen fibrils within PGCs. The PGCs carrying procollagen subsequently connect to the extracellular matrix via fingerlike projections of the PM, which are known as fibripositors, and which are oriented along the axis of the tendon (Canty et al., 2004). Interestingly, actin filament disassembly results in the rapid loss of fibripositors and in the subsequent disappearance of intracellular fibrils. In this case, a significant proportion of collagen fibrils are found to no longer be orientated with the long axis of the tendon. This suggests an important role for the actin cytoskeleton in the alignment of PGC delivery and in the further organization of the collagenous extracellular matrix in the embryonic tendon (Canty et al., 2006).

In brain tissue, transformation of the contact between an axon and a dendrite into a synapse is accompanied by the accumulation of the synaptic machinery at the site, with these delivered in TGN-derived carriers. In cultured hippocampal neurons, PGCs are linked via spectrin to clusters of the neural cell adhesion molecule (NCAM) in the PM. These complexes are trapped at sites of initial neurite-to-neurite contact within several minutes of the formation of the initial contact. The accumulation of PGCs at contacts with NCAMdeficient neurons is reduced when compared with wild-type cells, which suggests that NCAM mediates the anchoring of intracellular organelles in nascent synapses (Sytnyk et al., 2002).
