Cargo trafficking from the trans‐Golgi network towards the endosome

The trans‐Golgi network (TGN) is a major sorting, packing and delivering station of newly synthesised proteins and lipids to their final destination. These cargo molecules follow the secretory pathway, which is a vital part of cellular trafficking machinery in all eukaryotic cells. This secretory pathway is well conserved in all eukaryotes from low‐level eukaryotes, such as yeast, to higher level eukaryotes like mammals. The molecular mechanisms of protein sorting by adaptor proteins, membrane elongation and transport to the final destinations by motor proteins and the cytoskeleton, and membrane pinching‐off by scission proteins must be choreographically managed for efficient cargo delivery, and the understanding of these detailed processes is not yet completed. Functionally, defects in these mechanisms are associated with the pathology of prominent diseases such as acute myeloid leukaemia, Charcot–Marie–Tooth disease, I‐cell disease and Wiskott–Aldrich syndrome. The present review points out the recent advances in our knowledge of the molecular mechanisms involved in the transportation of the cargo from the TGN towards the endosome.     

Amyloid precursor protein traffics from the Golgi directly to early endosomes in an Arl5b‐ and AP4‐dependent pathway

The intracellular trafficking and proteolytic processing of the membrane‐bound amyloid precursor protein (APP) are coordinated events leading to the generation of pathogenic amyloid‐beta (Aβ) peptides. The membrane transport of newly synthesized APP from the Golgi to the endolysosomal system is not well defined, yet it is likely to be critical for regulating its processing by β‐secretase (BACE1) and γ‐secretase. Here, we show that the majority of newly synthesized APP is transported from the trans‐Golgi network (TGN) directly to early endosomes and then subsequently to the late endosomes/lysosomes with very little transported to the cell surface. We show that Arl5b, a small G protein localized to the TGN, and AP4 are essential for the post‐Golgi transport of APP to early endosomes. Arl5b is physically associated with AP4 and is required for the recruitment of AP4, but not AP1, to the TGN. Depletion of either Arl5b or AP4 results in the accumulation of APP, but not BACE1, in the Golgi, and an increase in APP processing and Aβ secretion. These findings demonstrate that APP is diverted from BACE1 at the TGN for direct transport to early endosomes and that the TGN represents a site for APP processing with the subsequent secretion of Aβ.      

Overexpression of trans‐Golgi network t‐SNAREs rescues vacuolar trafficking and TGN morphology defects in a putative tethering factor mutant

The trans‐Golgi network (TGN) is a major site for sorting of cargo to either the vacuole or apoplast. The TGN‐localized coiled‐coil protein TNO1 is a putative tethering factor that interacts with the TGN t‐SNARE SYP41 and is required for correct localization of the SYP61 t‐SNARE. An Arabidopsis thaliana tno1 mutant is hypersensitive to salt stress and partially mislocalizes vacuolar proteins to the apoplast, indicating a role in vacuolar trafficking. Here, we show that overexpression of SYP41 or SYP61 significantly increases SYP41–SYP61 complex formation in a tno1 mutant, and rescues the salt sensitivity and defective vacuolar trafficking of the tno1 mutant. The TGN is disrupted and vesicle budding from Golgi cisternae is reduced in the tno1 mutant, and these defects are also rescued by overexpression of SYP41 or SYP61. Our results suggest that the trafficking and Golgi morphology defects caused by loss of TNO1 can be rescued by increasing SYP41–SYP61 t‐SNARE complex formation, implicating TNO1 as a tethering factor mediating efficient vesicle fusion at the TGN.     

Sphingolipid metabolic flow controls phosphoinositide turnover at the trans‐Golgi network

Sphingolipids are membrane lipids globally required for eukaryotic life. The sphingolipid content varies among endomembranes with pre‐ and post‐Golgi compartments being poor and rich in sphingolipids, respectively. Due to this different sphingolipid content, pre‐ and post‐Golgi membranes serve different cellular functions. The basis for maintaining distinct subcellular sphingolipid levels in the presence of membrane trafficking and metabolic fluxes is only partially understood. Here, we describe a homeostatic regulatory circuit that controls sphingolipid levels at the trans‐Golgi network (TGN). Specifically, we show that sphingomyelin production at the TGN triggers a signalling pathway leading to PtdIns(4)P dephosphorylation. Since PtdIns(4)P is required for cholesterol and sphingolipid transport to the trans‐Golgi network, PtdIns(4)P consumption interrupts this transport in response to excessive sphingomyelin production. Based on this evidence, we envisage a model where this homeostatic circuit maintains a constant lipid composition in the trans‐Golgi network and post‐Golgi compartments, thus counteracting fluctuations in the sphingolipid biosynthetic flow.     

Ferlins Show Tissue‐Specific Expression and Segregate as Plasma Membrane/Late Endosomal or Trans‐Golgi/Recycling Ferlins

Ferlins are a family of transmembrane‐anchored vesicle fusion proteins uniquely characterized by 5–7 tandem cytoplasmic C2 domains, Ca²⁺‐regulated phospholipid‐binding domains that regulate vesicle fusion in the synaptotagmin family. In humans, dysferlin mutations cause limb‐girdle muscular dystrophy type 2B (LGMD2B) due to defective Ca²⁺‐dependent, vesicle‐mediated membrane repair and otoferlin mutations cause non‐syndromic deafness due to defective Ca²⁺‐triggered auditory neurotransmission. In this study, we describe the tissue‐specific expression, subcellular localization and endocytic trafficking of the ferlin family. Studies of endosomal transit together with 3D‐structured illumination microscopy reveals dysferlin and myoferlin are abundantly expressed at the PM and cycle to Rab7‐positive late endosomes, supporting potential roles in the late‐endosomal pathway. In contrast, Fer1L6 shows concentrated localization to a specific compartment of the trans‐Golgi/recycling endosome, cycling rapidly between this compartment and the PM via Rab11 recycling endosomes. Otoferlin also shows trans‐Golgi to PM cycling, with very low levels of PM otoferlin suggesting either brief PM residence, or rare incorporation of otoferlin molecules into the PM. Thus, type‐I and type‐II ferlins segregate as PM/late‐endosomal or trans‐Golgi/recycling ferlins, consistent with different ferlins mediating vesicle fusion events in specific subcellular locations.      

Functional expression of PHO1 to the Golgi and trans-Golgi network and its role in export of inorganic phosphate

Arabidopsis thaliana PHO1 is primarily expressed in the root vascular cylinder and is involved in the transfer of inorganic phosphate (Pi) from roots to shoots. To analyze the role of PHO1 in transport of Pi, we have generated transgenic plants expressing PHO1 in ectopic A. thaliana tissues using an estradiol-inducible promoter. Leaves treated with estradiol showed strong PHO1 expression, leading to detectable accumulation of PHO1 protein. Estradiol-mediated induction of PHO1 in leaves from soil-grown plants, in leaves and roots of plants grown in liquid culture, or in leaf mesophyll protoplasts, was all accompanied by the specific release of Pi to the extracellular medium as early as 2-3 h after addition of estradiol. Net Pi export triggered by PHO1 induction was enhanced by high extracellular Pi and weakly inhibited by the proton-ionophore carbonyl cyanide m-chlorophenylhydrazone. Expression of a PHO1-GFP construct complementing the pho1 mutant revealed GFP expression in punctate structures in the pericycle cells but no fluorescence at the plasma membrane. When expressed in onion epidermal cells or in tobacco mesophyll cells, PHO1-GFP was associated with similar punctate structures that co-localized with the Golgi/trans-Golgi network and uncharacterized vesicles. However, PHO1-GFP could be partially relocated to the plasma membrane in leaves infiltrated with a high-phosphate solution. Together, these results show that PHO1 can trigger Pi export in ectopic plant cells, strongly indicating that PHO1 is itself a Pi exporter. Interestingly, PHO1-mediated Pi export was associated with its localization to the Golgi and trans-Golgi networks, revealing a role for these organelles in Pi transport.     

Retromer guides STxB and CD8-M6PR from early to recycling endosomes, EHD1 guides STxB from recycling endosome to Golgi

Retrograde trafficking transports proteins, lipids and toxins from the plasma membrane to the Golgi and endoplasmic reticulum (ER). To reach the Golgi, these cargos must transit the endosomal system, consisting of early endosomes (EE), recycling endosomes, late endosomes and lysosomes. All cargos pass through EE, but may take different routes to the Golgi. Retromer-dependent cargos bypass the late endosomes to reach the Golgi. We compared how two very different retromer-dependent cargos negotiate the endosomal sorting system. Shiga toxin B, bound to the external layer of the plasma membrane, and chimeric CD8-mannose-6-phosphate receptor (CI-M6PR), which is anchored via a transmembrane domain. Both appear to pass through the recycling endosome. Ablation of the recycling endosome diverted both of these cargos to an aberrant compartment and prevented them from reaching the Golgi. Once in the recycling endosome, Shiga toxin required EHD1 to traffic to the TGN, while the CI-M6PR was not significantly dependent on EHD1. Knockdown of retromer components left cargo in the EE, suggesting that it is required for retrograde exit from this compartment. This work establishes the recycling endosome as a required step in retrograde traffic of at least these two retromer-dependent cargos. Along this pathway, retromer is associated with EE to recycling endosome traffic, while EHD1 is associated with recycling endosome to TGN traffic of STxB.     

A distinct class of vesicles derived from the trans‐Golgi mediates secretion of xylogalacturonan in the root border cell

Root border cells lie on the surface of the root cap and secrete massive amounts of mucilage that contains polysaccharides and proteoglycans. Golgi stacks in the border cells have hypertrophied margins, reflecting elevated biosynthetic activity to produce the polysaccharide components of the mucilage. To investigate the three‐dimensional structures and macromolecular compositions of these Golgi stacks, we examined high‐pressure frozen/freeze‐substituted alfalfa root cap cells with electron microscopy/tomography. Golgi stacks in border cells and peripheral cells, precursor cells of border cells, displayed similar morphological features, such as proliferation of trans cisternae and swelling of the trans cisternae and trans‐Golgi network (TGN) compartments. These swollen margins give rise to two types of vesicles larger than other Golgi‐associated vesicles. Margins of trans‐Golgi cisternae accumulate the LM8 xylogalacturonan (XGA) epitope, and they become darkly stained large vesicles (LVs) after release from the Golgi. Epitopes for xyloglucan (XG), polygalacturonic acid/rhamnogalacturonan‐I (PGA/RG‐I) are detected in the trans‐most cisternae and TGN compartments. LVs produced from TGN compartments (TGN‐LVs) stained lighter than LVs and contained the cell wall polysaccharide epitopes seen in the TGN. LVs carrying the XGA epitope fuse with the plasma membrane only in border cells, whereas TGN‐LVs containing the XG and PGA/RG‐I epitopes fuse with the plasma membrane of both peripheral cells and border cells. Taken together, these results indicate that XGA is secreted by a novel type of secretory vesicles derived from trans‐Golgi cisternae. Furthermore, we simulated the collapse in the central domain of the trans‐cisternae accompanying polysaccharide synthesis with a mathematical model.     

Arabidopsis RABA1 GTPases are involved in transport between the trans‐Golgi network and the plasma membrane,and are required for salinity stress tolerance

RAB GTPases are key regulators of membrane traffic. Among them, RAB11, a widely conserved sub‐group, has evolved in a unique way in plants; plant RAB11 members show notable diversity, whereas yeast and animals have only a few RAB11 members. Fifty‐seven RAB GTPases are encoded in the Arabidopsis thaliana genome, 26 of which are classified in the RAB11 group (further divided into RABA1–RABA6 sub‐groups). Although several plant RAB11 members have been shown to play pivotal roles in plant‐unique developmental processes, including cytokinesis and tip growth, molecular and physiological functions of the majority of RAB11 members remain unknown. To reveal precise functions of plant RAB11, we investigated the subcellular localization and dynamics of the largest sub‐group of Arabidopsis RAB11, RABA1, which has nine members. RABA1 members reside on mobile punctate structures adjacent to the trans‐Golgi network and co‐localized with VAMP721/722, R‐SNARE proteins that operate in the secretory pathway. In addition, the constitutive‐active mutant of RABA1b, RABA1b^Q72L , was present on the plasma membrane. The RABA1b ‐containing membrane structures showed actin‐dependent dynamic motion . Vesicles labeled by GFP–RABA1b moved dynamically, forming queues along actin filaments. Interestingly, Arabidopsis plants whose four major RABA1 members were knocked out, and those expressing the dominant‐negative mutant of RABA1B, exhibited hypersensitivity to salinity stress. Altogether, these results indicate that RABA1 members mediate transport between the trans‐Golgi network and the plasma membrane, and are required for salinity stress tolerance.     

Low Rho activity in hepatocytes prevents apical from basolateral cargo separation during trans‐Golgi network to surface transport

Hepatocytes, the main epithelial cells of the liver, organize their polarized membrane domains differently from ductal epithelia. They also differ in their biosynthetic delivery of single‐membrane‐spanning and glycophosphatidylinositol‐anchored proteins to the apical domain. While ductal epithelia target apical proteins to varying degrees from the trans‐Golgi network (TGN) to the apical surface directly, hepatocytes target them first to the basolateral domain, from where they undergo basolateral‐to‐apical transcytosis. How TGN‐to‐surface transport differs in both scenarios is unknown. Here, we report that the basolateral detour of a hepatocyte apical protein is due, in part, to low RhoA activity at the TGN, which prevents its segregation from basolateral transport carriers. Activating Rho in hepatocytic cells, which switches their polarity from hepatocytic to ductal, also led to apical‐basolateral cargo segregation at the TGN as is typical for ductal cells, affirming a central role for Rho‐signaling in different aspects of the hepatocytic polarity phenotype. Nevertheless, Rho‐induced cargo segregation was not sufficient to target the apical protein directly; thus, failure to recruit apical targeting machinery also contributes to its indirect itinerary.     

Oxysterol‐binding protein recruitment and activity at the endoplasmic reticulum‐Golgi interface are independent of Sac1

Oxysterol‐binding protein (OSBP) localizes to endoplasmic reticulum (ER)‐Golgi contact sites where it transports cholesterol and phosphatidylinositol 4‐phosphate (PI‐4P), and activates lipid transport and biosynthetic activities. The PI‐4P phosphatase Sac1 cycles between the ER and Golgi apparatus where it potentially regulates OSBP activity. Here we examined whether the ER‐Golgi distribution of endogenous or ectopically expressed Sac1 influences OSBP activity. OSBP and Sac1 co‐localized at apparent ER‐Golgi contact sites in response to 25‐hydroxycholesterol (25OH), cholesterol depletion and p38 MAPK inhibitors. A Sac1 mutant that is unable to exit the ER did not localize with OSBP, suggesting that sterol perturbations cause Sac1 transport to the Golgi apparatus. Ectopic expression of Sac1 in the ER or Golgi apparatus, or Sac1 silencing, did not affect OSBP localization to ER‐Golgi contact sites, OSBP‐dependent activation of sphingomyelin synthesis, or cholesterol esterification in the ER. p38 MAPK inhibition and retention of Sac1 in the Golgi apparatus also caused OSBP phosphorylation and OSBP‐dependent activation of sphingomyelin synthesis at ER‐Golgi contacts. These results demonstrate that Sac1 expression in either the ER or Golgi apparatus has a minimal impact on the PI‐4P that regulates OSBP activity or recruitment to contact sites.      

Calcium release from the Golgi apparatus and the endoplasmic reticulum in HeLa cells stably expressing targeted aequorin to these compartments

Extracellular agonists mobilize Ca2+ from SERCA-comprising intracellular Ca2+ stores located in both the Golgi apparatus and the endoplasmic reticulum. Ca2+ release from both these compartments was studied in HeLa cells stably expressing the luminescent Ca2+ indicator aequorin specifically targeted to these compartments. Changes in lumenal [Ca2+] as detected by the aequorin measurements were correlated with parallel changes in total Ca2+ content of the stores. The latencies and initial rates of Ca2+ release from the Golgi apparatus and the endoplasmic reticulum were quite similar. However, maximal Ca2+ release measured with Golgi-targeted aequorin terminated faster than that from the endoplasmic reticulum. The rate and extent of Ca2+ depletion from both compartments correlated well with the peak amplitude of the cytosolic [Ca2+] rise. Time-course experiments further revealed that the peak of the cytosolic Ca2+ response occurred before the lumenal [Ca2+] reached its lowest level. We conclude that both the Golgi apparatus and the endoplasmic reticulum contribute to the rise in cytosolic [Ca2+] upon agonist stimulation, but the kinetics of the Ca2+ release are different.     

CBL‐mediated targeting of CIPKs facilitates the decoding of calcium signals emanating from distinct cellular stores

During adaptation and developmental processes cells respond through nonlinear calcium‐decoding signaling cascades, the principal components of which have been identified. However, the molecular mechanisms generating specificity of cellular responses remain poorly understood. Calcineurin B‐like (CBL) proteins contribute to decoding calcium signals by specifically interacting with a group of CBL‐interacting protein kinases (CIPKs). Here, we report the subcellular localization of all 10 CBL proteins from Arabidopsis and provide a cellular localization matrix of a plant calcium signaling network. Our findings suggest that individual CBL proteins decode calcium signals not only at the plasma membrane and the tonoplast, but also in the cytoplasm and nucleus. We found that distinct targeting signals located in the N‐terminal domain of CBL proteins determine the spatially discrete localization of CBL/CIPK complexes by COPII‐independent targeting pathways. Our findings establish the CBL/CIPK signaling network as a calcium decoding system that enables the simultaneous specific information processing of calcium signals emanating from different intra‐ and extracellular stores, and thereby provides a mechanism underlying the specificity of cellular responses.     

Characterization and regulation of constitutive transport intermediates involved in trafficking from the trans-Golgi network

Transport vesicles or containers (TCs) mediate constitutive protein transport between the trans-Golgi network (TGN) and the plasma membrane. A key question is the nature and regulation of these transport containers or intermediates. We have used a trans-Golgi network resident, TGN38, to investigate TC formation. TGN38 is a recycling membrane glycoprotein that moves to the cell surface via constitutive membrane traffic and returns via the endosomal pathway. An in vitro assay to measure TC formation was devised using rat liver Golgi membranes, cytosolic factors and ATP. Transport intermediates containing TGN38 were produced and found to be smooth vesicles and tubules of up to 200 nm in length. These membrane-enclosed structures contain different constitutively secreted membrane glycoproteins, including molecules involved in immune functions such as MHC Class I and the polymeric Ig receptor, showing that these intermediates correspond to TCs that have been previously identified in vivo. Importantly, TC formation can be stimulated or inhibited by protein kinase and phosphatase inhibitors, showing regulation by intracellular signalling pathways.     

Ins(1,4,5)P3 receptor type 1 associates with AKAP9 (AKAP450 variant) and protein kinase A type IIβ in the Golgi apparatus in cerebellar granule cells

Background information. Interconnections between the Ca²⁺ and cAMP signalling pathways can determine the specificity and diversity of the cellular effects mediated by these second messengers. Most cAMP effects are mediated by PKA (protein kinase A), which is anchored close to its membranous substrates by AKAPs (A kinase‐anchoring proteins). In many cell types, the activation of InsP₃R (inositol 1,4,5‐trisphosphate receptor), an endoplasmic reticulum Ca²⁺ channel, is a key event of Ca²⁺ signalling. The phosphorylation of InsP₃R1 by PKA stimulates Ca²⁺ mobilization. This control is thought to be tight, involving the association of PKA with InsP₃R1. The InsP₃R1 isoform predominates in central nervous tissue and its concentration is highest in the cerebellar microsomes. We investigated the complex formed by InsP₃R1 and PKA in this fraction, vith a view to identifying its components and determining its distribution in the cerebellar cortex. Results. Immunoprecipitation experiments showed that InsP₃R1 associated with PKA type IIβ and AKAP450, the longer variant of AKAP9, in sheep cerebellar microsomes. The co‐purification of AKAP450 with InsP₃R1 on heparin‐agarose provided further evidence of the association of these proteins. Immunohistofluorescence experiments on slices of cerebellar cortex showed that AKAP450 was colocalized with InsP₃R1 and RIIβ (regulatory subunit of PKA IIβ) in granule cells, but not in Purkinje cells. AKAP450 was localized in the Golgi apparatus of these two cell types whereas InsP₃R1 was detected in this organelle only in granule cells. Conclusions. Taken together these results suggest that InsP₃R1 forms a complex with AKAP450 and PKAIIβ, localized in the Golgi apparatus of cerebellar granule cells. In contrast, the association of InsP₃R1 with PKA in Purkinje cells would require a different macromolecular complex.     

Arabidopsis MPK6 is involved in cell division plane control during early root development,and localizes to the pre‐prophase band,phragmoplast, trans‐Golgi network and plasma membrane

The proper spatial and temporal expression and localization of mitogen‐activated protein kinases (MAPKs) is essential for developmental and cellular signalling in all eukaryotes. Here, we analysed expression, subcellular localization and function of MPK6 in roots of Arabidopsis thaliana using wild‐type plants and three mpk6 knock‐out mutant lines. The MPK6 promoter showed two expression maxima in the most apical part of the root meristem and in the root transition zone. This expression pattern was highly consistent with ‘no root’ and ‘short root’ phenotypes, as well as with ectopic cell divisions and aberrant cell division planes, resulting in disordered cell files in the roots of these mpk6 knock‐out mutants. In dividing root cells, MPK6 was localized on the subcellular level to distinct fine spots in the pre‐prophase band and phragmoplast, representing the two most important cytoskeletal structures controlling the cell division plane. By combining subcellular fractionation and microscopic in situ and in vivo co‐localization methods, MPK6 was localized to the plasma membrane (PM) and the trans‐Golgi network (TGN). In summary, these data suggest that MPK6 localizing to mitotic microtubules, secretory TGN vesicles and the PM is involved in cell division plane control and root development in Arabidopsis.     

In the grass species Brachypodium distachyon,the production of mixed‐linkage (1,3;1,4)‐β‐glucan (MLG) occurs in the Golgi apparatus

Mixed‐linkage (1,3;1,4)‐β‐glucan (MLG) is a glucose polymer with beneficial effects on human health and high potential for the agricultural industry. MLG is present predominantly in the cell wall of grasses and is synthesized by cellulose synthase‐like F or H families of proteins, with CSLF6 being the best‐characterized MLG synthase. Although the function of this enzyme in MLG production has been established, the site of MLG synthesis in the cell is debated. It has been proposed that MLG is synthesized at the plasma membrane, as occurs for cellulose and callose; in contrast, it has also been proposed that MLG is synthesized in the Golgi apparatus, as occurs for other matrix polysaccharides of the cell wall. Testing these conflicting possibilities is fundamentally important in the general understanding of the biosynthesis of the plant cell wall. Using immuno‐localization analyses with MLG‐specific antibody in Brachypodium and in barley, we found MLG present in the Golgi, in post‐Golgi structures and in the cell wall. Accordingly, analyses of a functional fluorescent protein fusion of CSLF6 stably expressed in Brachypodium demonstrated that the enzyme is localized in the Golgi. We also established that overproduction of MLG causes developmental and growth defects in Brachypodium as also occur in barley. Our results indicated that MLG production occurs in the Golgi similarly to other cell wall matrix polysaccharides, and supports the broadly applicable model in grasses that tight mechanisms control optimal MLG accumulation in the cell wall during development and growth. This work addresses the fundamental question of where mixed linkage (1,3;1,4)‐β‐glucan (MLG) is synthesized in plant cells. By analyzing the subcellular localization of MLG and MLG synthase in an endogenous system, we demonstrated that MLG synthesis occurs at the Golgi in Brachypodium and barley. A growth inhibition due to overproduced MLG in Brachypodium supports the general applicability of the model that a tight control of the cell wall polysaccharides accumulation is needed to maintain growth homeostasis during development.     

E-cadherin transport from the trans-Golgi network in tubulovesicular carriers is selectively regulated by golgin-97

E-cadherin is a cell-cell adhesion protein that is trafficked and delivered to the basolateral cell surface. Membrane-bound carriers for the post-Golgi exocytosis of E-cadherin have not been characterized. Green fluorescent protein (GFP)-tagged E-cadherin (Ecad-GFP) is transported from the trans-Golgi network (TGN) to the recycling endosome on its way to the cell surface in tubulovesicular carriers that resemble TGN tubules labeled by members of the golgin family of tethering proteins. Here, we examine the association of golgins with tubular carriers containing E-cadherin as cargo. Fluorescent GRIP domains from golgin proteins replicate the membrane binding of the full-length proteins and were coexpressed with Ecad-GFP. The GRIP domains of p230/golgin-245 and golgin-97 had overlapping but nonidentical distributions on the TGN; both domains were on TGN-derived tubules but only the golgin-97 GRIP domain coincided with Ecad-GFP tubules in live cells. When the Arl1-binding endogenous golgins, p230/golgin-245 and golgin-97 were displaced from Golgi membranes by overexpression of the p230 GRIP domain, trafficking of Ecad-GFP was inhibited. siRNA knockdown of golgin-97 also inhibited trafficking of Ecad-GFP. Thus, the GRIP domains of p230/golgin-245 and golgin-97 bind discriminately to distinct membrane subdomains of the TGN. Golgin-97 is identified as a selective and essential component of the tubulovesicular carriers transporting E-cadherin out of the TGN.