Ruolo della proteina OPA3 nella patogenesi di malattie neurodegenerative

OPA3 è una proteina codificata dal genoma nucleare che, grazie a una sequenza di targeting mitocondriale, viene indirizzata ai mitocondri dopo la sua sintesi. Le mutazioni nel gene OPA3 sono associate a due patologie neurodegenerative: la Sindrome di Costeff, causata da mutazioni recessive, e una forma di atrofia ottica dominante che si manifesta con cataratta e spesso sordità. L’esatta funzione e regolazione della proteina non sono ancora state completamente chiarite, così come la sua localizzazione nella membrana mitocondriale esterna o interna. Lo scopo di questa tesi era quello di fare luce sulla funzione della proteina OPA3, con particolare interesse alla dinamica mitocondriale e all’autofagia, sulla sua localizzazione subcellulare ed infine di definire il meccanismo patogenetico nelle patologie neurodegenerative causate da mutazioni in questo gene. A questo scopo abbiamo utilizzato sia una linea di neuroblastoma silenziata stabilmente per OPA3 che linee cellulari primarie derivate da pazienti. I risultati del presente studio dimostrano che la riduzione di OPA3, indotta nelle cellule del neuroblastoma e presente nei fibroblasti derivati dai pazienti, produce alterazioni nel network mitocondriale con uno sbilanciamento a favore della fusione. Questo fenomeno è probabilmente dovuto all’aumento della forma long della proteina OPA1 che è stato riscontrato in entrambi i modelli cellulari. Inoltre, seppur con direzione apparentemente opposta, in entrambi i modelli abbiamo osservato un’alterata regolazione dell’autofagia. Infine, abbiamo confermato che OPA3 localizza nella membrana mitocondriale interna ed è esposta per gran parte nella matrice. Inoltre, un segnale della proteina è stato trovato anche nelle mitochondrial associated membranes, suggerendo un possibile ruolo di OPA3 nel trasferimento dei lipidi tra i mitocondri e il reticolo endoplasmatico. Abbiamo rilevato un’interazione della proteina OPA3 con l’acido fosfatidico che non era mai stata evidenziata fino ad oggi. Queste osservazioni sono compatibili con le alterazioni della dinamica mitocondriale e la disregolazione dell’autofagia documentate nei modelli studiati.

The β-sliding clamp directs the localization of HdaA to the replisome in Caulobacter crescentus.

The initiation of chromosome replication is tightly regulated in bacteria to ensure that it takes place only once per cell cycle. In many proteobacteria, this process requires the ATP-bound form of the DnaA protein. The regulatory inactivation of DnaA (RIDA) facilitates the conversion of DnaA-ATP into replication-inactive DnaA-ADP, thereby preventing overinitiation. Homologues of the HdaA protein, together with the β-clamp of the DNA polymerase (DnaN), are required for this process. Here, we used fluorescence resonance energy transfer experiments to demonstrate that HdaA interacts with DnaN in live Caulobacter crescentus cells. We show that a QFKLPL motif in the N-terminal region of HdaA is required for this interaction and that this motif is also needed to recruit HdaA to the subcellular location occupied by the replisome during DNA replication. An HdaA mutant protein that cannot colocalize or interact with DnaN can also not support the essential function of HdaA. These results suggest that the recruitment of HdaA to the replisome is needed during RIDA in C. crescentus, probably as a means to sense whether chromosome replication has initiated before DnaA becomes inactivated. In addition, we show that a conserved R145 residue located in the AAA+ domain of HdaA is also needed for the function of HdaA, although it does not affect the interaction of HdaA with DnaN in vivo. The AAA+ domain of HdaA may therefore be required during RIDA after the initial recruitment of HdaA to the replisome by DnaN.

Functional diversification of duplicate genes through subcellular adaptation of encoded proteins.

BACKGROUND: Gene duplication is the primary source of new genes with novel or altered functions. It is known that duplicates may obtain these new functional roles by evolving divergent expression patterns and/or protein functions after the duplication event. Here, using yeast (Saccharomyces cerevisiae) as a model organism, we investigate a previously little considered mode for the functional diversification of duplicate genes: subcellular adaptation of encoded proteins. RESULTS: We show that for 24-37% of duplicate gene pairs derived from the S. cerevisiae whole-genome duplication event, the two members of the pair encode proteins that localize to distinct subcellular compartments. The propensity of yeast duplicate genes to evolve new localization patterns depends to a large extent on the biological function of their progenitor genes. Proteins involved in processes with a wider subcellular distribution (for example, catabolism) frequently evolved new protein localization patterns after duplication, whereas duplicate proteins limited to a smaller number of organelles (for example, highly expressed biosynthesis/housekeeping proteins with a slow rate of evolution) rarely relocate within the cell. Paralogous proteins evolved divergent localization patterns by partitioning of ancestral localizations ('sublocalization'), but probably more frequently by relocalization to new compartments ('neolocalization'). We show that such subcellular reprogramming may occur through selectively driven substitutions in protein targeting sequences. Notably, our data also reveal that relocated proteins functionally adapted to their new subcellular environments and evolved new functional roles through changes of their physico-chemical properties, expression levels, and interaction partners. CONCLUSION: We conclude that protein subcellular adaptation represents a common mechanism for the functional diversification of duplicate genes.

Feasibility of direct mapping of cerebral fluorodeoxy-D-glucose metabolism in situ at subcellular resolution using soft X-ray fluorescence.

Glucose metabolism is difficult to image with cellular resolution in mammalian brain tissue, particularly with (18) fluorodeoxy-D-glucose (FDG) positron emission tomography (PET). To this end, we explored the potential of synchrotron-based low-energy X-ray fluorescence (LEXRF) to image the stable isotope of fluorine (F) in phosphorylated FDG (DG-6P) at 1 μm(2) spatial resolution in 3-μm-thick brain slices. The excitation-dependent fluorescence F signal at 676 eV varied linearly with FDG concentration between 0.5 and 10 mM, whereas the endogenous background F signal was undetectable in brain. To validate LEXRF mapping of fluorine, FDG was administered in vitro and in vivo, and the fluorine LEXRF signal from intracellular trapped FDG-6P over selected brain areas rich in radial glia was spectrally quantitated at 1 μm(2) resolution. The subsequent generation of spatial LEXRF maps of F reproduced the expected localization and gradients of glucose metabolism in retinal Müller glia. In addition, FDG uptake was localized to periventricular hypothalamic tanycytes, whose morphological features were imaged simultaneously by X-ray absorption. We conclude that the high specificity of photon emission from F and its spatial mapping at ≤1 μm resolution demonstrates the ability to identify glucose uptake at subcellular resolution and holds remarkable potential for imaging glucose metabolism in biological tissue. © 2012 Wiley Periodicals, Inc.

Receptor localization of steroid hormones and drugs: discoveries through the use of thaw-mount and dry-mount autoradiography

The history of receptor autoradiography, its development and applications, testify to the utility of this histochemical technique for localizing radiolabeled hormones and drugs at cellular and subcellular sites of action in intact tissues. Localization of diffusible compounds has been a challenge that was met through the introduction of the "thaw-mount" and "dry-mount" autoradiographic techniques thirty years ago. With this cellular receptor autoradiography, used alone or combined with other histochemical techniques, sites of specific binding and deposition in vivo and in vitro have been characterized. Numerous discoveries, some reviewed in this article, provided information that led to new concepts and opened new areas of research. As an example, in recent years more than fifty target tissues for vitamin D have been specified, challenging the conventional view about the main biological role of vitamin D. The functions of most of these vitamin D target tissues are unrelated to the regulation of systemic calcium homeostasis, but pertain to the (seasonal) regulation of endo- and exocrine secretion, cell proliferation, reproduction, neural, immune and cardiovascular responses, and adaptation to stress. Receptor autoradiography with cellular resolution has become an indispensable tool in drug research and development, since information can be obtained that is difficult or impossible to gain otherwise

The N-terminal 33 amino acid domain of Siva-1 is sufficient for nuclear localization

Siva-1 induces apoptosis in multiple pathological processes and plays an important role in the suppression of tumor metastasis, protein degradation, and other functions. Although many studies have demonstrated that Siva-1 functions in the cytoplasm, a few have found that Siva-1 can relocate to the nucleus. In this study, we found that the first 33 amino acid residues of Siva-1 are required for its nuclear localization. Further study demonstrated that the green fluorescent protein can be imported into the nucleus after fusion with these 33 amino acid residues. Other Siva-1 regions and domains showed less effect on Siva-1 nuclear localization. By site-mutagenesis of all of these 33 amino acid residues, we found that mutants of the first 1-18 amino acids affected Siva-1 nuclear compartmentalization but could not complete this localization independently. In summary, we demonstrated that the N-terminal 33 amino acid residues were sufficient for Siva-1 nuclear localization, but the mechanism of this translocation needs additional investigation.

Protein subcellular targeting inTrichoderma aggressivum f. aggressivum

Trichoderma aggressivum f. aggressivum is a filamentous soil fungus. Green mold disease of commercial mushrooms caused by this species in North America has resulted in millions of dollars in lost revenue within the mushroom growing industry. Research on the molecular level of T aggressivum have jus t begun with the goal of understanding the functions of each gene and protein, and their expression control. Protein targeting has not been well studied in this species yet. Therefore, the intent of this study was to test the protein localization and production levels in T aggressivum with green fluorescent protein (GFP) with an intron and tagged with either nuclear localization signal (NLS) or an endoplasmic reticulum retention signal (KDEL). Two GFP constructs (with and without the intron) were used as controls in this study. All four constructs were successfully transferred into T aggressivum and all modified strains showed similar growth characteristics as the wild type non-transformed isolate. GFP expression was detected from all modified T aggressivum with confocal microscopy and the expression was similar in all four strains. The intron tested in this study had no or very minor effects as GFP expression was similar with or without it. The GFP signal increased over a 5 day period for all transformants, while the GFP to total protein ratio decreased over the same period for all transformants. The GFP-KDEL transformant showed similar protein expression level and localization as did the control transformant lacking the KDEL retention signal. The GFP-NLS transformant similarly failed to localize GFP into nucleus as fluorescence with this strain was virtually identical to the GFP transformant lacking the NLS. Thus, future research is required to find effective localization signals for T aggressivum.

DivIVA-Mediated Polar Localization of ComN, a Posttranscriptional Regulator of Bacillus subtilis

ComN (YrzD) is a small, 98-amino-acid protein recently shown to be involved in the posttranscriptional control of the late competence comE operon in Bacillus subtilis. We show here that ComN localizes to the division site and cell poles in a DivIVA-dependent fashion. Yeast two-hybrid and glutathione S-transferase pulldown experiments showed that ComN interacts directly with DivIVA. ComN is not essential for the polar assembly of the core competence DNA uptake machinery. Nevertheless, polar localization of ComN should play some role in competence acquisition because delocalization of ComN leads to a small reduction in competence efficiency. We found that ComN promotes the accumulation of its target comE mRNA to septal and polar sites. Thus, we speculate that localized translation of ComE proteins may be required for efficient competence development. Our results underscore the versatility of DivIVA as a promoter of the differentiation of bacterial poles and demonstrate that the repertoire of polarly localized molecules in B. subtilis is broad, including a regulator of gene expression and its target mRNA. Moreover, our findings suggest that mRNA localization may play a role in the subcellular organization of bacteria.

Cellular localization and mechanism of amyloid precursor protein (APP) homodimer formation in an oxidizing environment

The amyloid precursor protein (APP) is a type I transmembrane glycoprotein, which resembles a cell surface receptor, comprising a large ectodomain, a single spanning transmembrane part and a short C-terminal, cytoplasmic domain. It belongs to a conserved gene family, with over 17 members, including also the two mammalian APP homologues proteins APLP1 and APLP2 („amyloid precursor like proteins“). APP is encoded by 19 exons, of which exons 7, 8, and 15 can be alternatively spliced to produce three major protein isoforms APP770, APP751 and APP695, reflecting the number of amino acids. The neuronal APP695 is the only isoform that lacks a Kunitz Protease Inhibitor (KPI) domain in its extracellular portion whereas the two larger, peripheral APP isoforms, contain the 57-amino-acid KPI insert. rnRecently, research effort has suggested that APP metabolism and function is thought to be influenced by homodimerization and that the oligomerization state of APP could also play a role in the pathology of Alzheimer's disease (AD), by regulating its processing and amyloid beta production. Several independent studies have shown that APP can form homodimers within the cell, driven by motifs present in the extracellular domain, as well as in the juxtamembrane (JM) and transmembrane (TM) regions of the molecule, whereby the exact molecular mechanism and the origin of dimer formation remains elusive. Therefore, we focused in our study on the actual subcellular origin of APP homodimerization within the cell, an underlying mechanism, and a possible impact on dimerization properties of its homologue APLP1. Furthermore, we analyzed homodimerization of various APP isoforms, in particular APP695, APP751 and APP770, which differ in the presence of a Kunitz-type protease inhibitor domain (KPI) in the extracellular region. In order to assess the cellular origin of dimerization under different cellular conditions, we established a mammalian cell culture model-system in CHO-K1 (chinese hamster ovary) cells, stably overexpressing human APP, harboring dilysine based organelle sorting motifs at the very C-terminus [KKAA-Endoplasmic Reticulum (ER); KKFF-Golgi]. In this study we show that APP exists as disulfide-bound, SDS-stable dimers, when it was retained in the ER, unlike when it progressed further to the cis-Golgi, due to the KKFF ER exit determinant. These stable APP complexes were isolated from cells, and analyzed by SDS–polyacrylamide gel electrophoresis under non-reducing conditions, whereas strong denaturing and reducing conditions completely converted those dimers to monomers. Our findings suggested that APP homodimer formation starts early in the secretory pathway and that the unique oxidizing environment of the ER likely promotes intermolecular disulfide bond formation between APP molecules. We particularly visualized APP dimerization employing a variety of biochemical experiments and investigated the origin of its generation by using a Bimolecular Fluorescence Complementation (BiFC) approach with split GFP-APP chimeras. Moreover, using N-terminal deletion constructs, we demonstrate that intermolecular disulfide linkage between cysteine residues, exclusively located in the extracellular E1 domain, represents another mechanism of how an APP sub-fraction can dimerize within the cell. Additionally, mutational studies revealed that cysteines at positions 98 and 105, embedded in the conserved loop region within the E1 domain, are critical for interchain disulfide bond formation. Using a pharmacological treatment approach, we show that once generated in the oxidative environment of the ER, APP dimers remain stably associated during transport, reaching the plasma membrane. In addition, we demonstrate that APP isoforms, encompassing the KPI domain, exhibit a strongly reduced ability to form cis-directed dimers in the ER, whereas trans-directed cell aggregation of Drosophila Schneider (S2)-cells was isoform independent, mediating cell-cell contacts. Thus, suggesting that steric properties of KPI-APP might be the cause for weaker cis-interaction in the ER, compared to APP695. Finally, we provide evidence that APP/APLP1 heterointeractions are likewise initiated in the ER, suggesting a similar mechanism for heterodimerization. Therefore, dynamic alterations of APP between monomeric, homodimeric, and possibly heterodimeric status could at least partially explain some of the variety in the physiological functions of APP.rn

Molecular analysis of $\beta$1,4-galactosyl-transferase localization to the cell surface and participation in cell adhesion

$\beta$1,4-Galactosyltransferase (GalTase) is unusual among the glycosyltransferases in that it is found in two subcellular compartments where it performs different functions. In the trans-Golgi complex, GalTase participates in oligosaccharide biosynthesis as do other glycosyltransferases. GalTase is also found on the cell surface, where it associates with the cytoskeleton and functions as a receptor for extracellular oligosaccharide ligands. Although we know much regarding GalTase function on the cell surface, little is known about the mechanisms underlying its transport to the plasma membrane. Cloning of the GalTase gene revealed that there are two GalTase proteins (i.e., long and short) with different size cytoplasmic tails. This raises the possibility that differences in the cytoplasmic domain of GalTase may influence its subcellular distribution. The object of this study was to examine this hypothesis directly through the use of molecular, immunological, and biochemical approaches.^ To examine whether the two GalTase proteins are targeted to different subcellular compartments, F9 embryonal carcinoma cells were transfected with either long or short GalTase cDNAs and intracellular and cell surface enzyme levels measured. Cell surface GalTase activity was enriched in cells overexpressing the long, but not the form of short GalTase. Furthermore, a dominant negative mutation in cell surface GalTase was created by transfecting cells with GalTase cDNAs encoding a truncated version of long GalTase devoid of the extracellular catalytic domain. Overexpressing the complete cytoplasmic and transmembrane domains of long GalTase led to a loss of GalTase-dependent cellular adhesion by specifically displacing surface GalTase from its cytoskeletal associations. In contrast, overexpressing the analogous truncated protein of short GalTase had no effect on cell adhesion. Finally, chloramphenicol acetyltransferase (CAT) reporter proteins were used to determine directly whether the cytoplasmic domains of long and short GalTase were responsible for differential subcellular distribution. The cytoplasmic and transmembrane domains of long GalTase led to CAT expression on the ceil surface and its association with the detergent-insoluble cytoskeleton; the analogous fusion protein containing short GalTase was restricted to the Golgi compartment. These results suggest that the cytoplasmic domain unique to long GalTase is responsible for targeting a portion of this protein to the cell surface and associating it with the cytoskeleton, enabling it to function as a cell adhesion molecule. ^

Mapping and characterization of the Xlsirt P11 RNA cis-acting localization element

In many organisms, polarity of the oocyte is established post-transcriptionally via subcellular RNA localization. Many RNAs are localized during oogenesis in Xenopus laevis, including Xlsirts ( Xenopus laevis short interspersed repeat transcripts) [Kloc, 1993]. Xlsirts constitute a large family defined by highly homologous repeat units 79–81 nucleotides in length. Endogenous Xlsirt RNAs use the METRO (Message Transport Organizer) pathway of localization, where RNAs are transported from the nucleus to the mitochondrial cloud in stage I oocytes. Secondly, RNAs anchor at the vegetal pole in stage II oocytes. Exogenous Xlsirt RNAs can also utilize the Late pathway of localization, which involves localization to the vegetal cortex during stage III of oogenesis and results in RNAs anchored in the cortex of the entire vegetal hemisphere. ^ The Xlsirts localization signal is contained within the repeat region. This study was designed to test the hypothesis that there are cis -acting localization elements in Xlsirts, and that higher order structure plays a role. Results of experiments on Xlsirt P11, a 1700 basepair (bp) family member, led to the conclusion that a 137-bp fragment of the repetitive region is necessary and sufficient for METRO and Late pathway localization. This analysis definitively demonstrates that the Xlsirt localization signal for the METRO and Late pathways reside within the repetitive region and not within the flanking regions. Analysis of Xlsirt linker scanning mutations revealed two METRO-pathway specific subelements, and one Late-pathway specific subelement. Functional, computer, and biochemical evidence relates the higher order structure of this element to its ability to function as a localization element. ^ Xlsirt 137 is 99% identical to the Xlsirt consensus sequence identified in this study, suggesting that it is the localization element for all localized Xlsirt family members. The repeat unit was reframed based on function, rather than arbitrarily based on sequence. This work supports the hypothesis presented in 1981 by George Spohr, who originally isolated the Xlsirts, which stated that the highly conserved repetitive elements must be constrained from variability due to some unknown function of the repeats themselves. These studies shed light on the mechanism of RNA localization, linking structure and function. ^

Mitochondrial localization of a NADR-dependent isocitrate dehydrogenase isoenzyme by using the green fluorescent protein as a marker

In this work, we describe the isolation of a new cDNA encoding an NADP-dependent isocitrate dehydrogenase (ICDH). The nucleotide sequence in its 5′ region gives a deduced amino acid sequence indicative of a targeting peptide. However, even if this cDNA clearly encodes a noncytosolic ICDH, it is not possible to say from the targeting peptide sequence to which subcellular compartment the protein is addressed. To respond to this question, we have transformed tobacco plants with a construct containing the entire targeting signal-encoding sequence in front of a modified green fluorescent protein (GFP) gene. This construct was placed under the control of the cauliflower mosaic virus 35S promoter, and transgenic tobacco plants were regenerated. At the same time, and as a control, we also have transformed tobacco plants with the same construct but lacking the nucleotide sequence corresponding to the ICDH-targeting peptide, in which the GFP is retained in the cytoplasm. By optical and confocal microscopy of leaf epiderm and Western blot analyses, we show that the putative-targeting sequence encoded by the cDNA addresses the GFP exclusively into the mitochondria of plant cells. Therefore, we conclude that this cDNA encodes a mitochondrial ICDH.

The Tail of a Yeast Class V Myosin, Myo2p, Functions as a Localization Domain

Myo2p is a yeast class V myosin that functions in membrane trafficking. To investigate the function of the carboxyl-terminal-tail domain of Myo2p, we have overexpressed this domain behind the regulatable GAL1 promoter (MYO2DN). Overexpression of the tail domain of Myo2p results in a dominant–negative phenotype that is phenotypically similar to a temperature-sensitive allele of myo2, myo2–66. The tail domain of Myo2p is sufficient for localization at low- expression levels and causes mislocalization of the endogenous Myo2p from sites of polarized cell growth. Subcellular fractionation of polarized, mechanically lysed yeast cells reveals that Myo2p is present predominantly in a 100,000 × g pellet. The Myo2p in this pellet is not solubilized by Mg++-ATP or Triton X-100, but is solubilized by high salt. Tail overexpression does not disrupt this fractionation pattern, nor do mutations in sec4, sec3, sec9, cdc42, or myo2. These results show that overexpression of the tail domain of Myo2p does not compete with the endogenous Myo2p for assembly into a pelletable structure, but does compete with the endogenous Myo2p for a factor that is necessary for localization to the bud tip.

Dual Localization of Squalene Epoxidase, Erg1p, in Yeast Reflects a Relationship between the Endoplasmic Reticulum and Lipid Particles

Squalene epoxidase, encoded by the ERG1 gene in yeast, is a key enzyme of sterol biosynthesis. Analysis of subcellular fractions revealed that squalene epoxidase was present in the microsomal fraction (30,000 × g) and also cofractionated with lipid particles. A dual localization of Erg1p was confirmed by immunofluorescence microscopy. On the basis of the distribution of marker proteins, 62% of cellular Erg1p could be assigned to the endoplasmic reticulum and 38% to lipid particles in late logarithmic-phase cells. In contrast, sterol Δ24-methyltransferase (Erg6p), an enzyme catalyzing a late step in sterol biosynthesis, was found mainly in lipid particles cofractionating with triacylglycerols and steryl esters. The relative distribution of Erg1p between the endoplasmic reticulum and lipid particles changes during growth. Squalene epoxidase (Erg1p) was absent in an erg1 disruptant strain and was induced fivefold in lipid particles and in the endoplasmic reticulum when the ERG1 gene was overexpressed from a multicopy plasmid. The amount of squalene epoxidase in both compartments was also induced approximately fivefold by treatment of yeast cells with terbinafine, an inhibitor of the fungal squalene epoxidase. In contrast to the distribution of the protein, enzymatic activity of squalene epoxidase was only detectable in the endoplasmic reticulum but was absent from isolated lipid particles. When lipid particles of the wild-type strain and microsomes of an erg1 disruptant were mixed, squalene epoxidase activity was partially restored. These findings suggest that factor(s) present in the endoplasmic reticulum are required for squalene epoxidase activity. Close contact between lipid particles and endoplasmic reticulum may be necessary for a concerted action of these two compartments in sterol biosynthesis.

A Role for the Lumenal Domain in Golgi Localization of the Saccharomyces cerevisiae Guanosine Diphosphatase

Integral membrane proteins (IMPs) contain localization signals necessary for targeting to their resident subcellular compartments. To define signals that mediate localization to the Golgi complex, we have analyzed a resident IMP of the Saccharomyces cerevisiae Golgi complex, guanosine diphosphatase (GDPase). GDPase, which is necessary for Golgi-specific glycosylation reactions, is a type II IMP with a short amino-terminal cytoplasmic domain, a single transmembrane domain (TMD), and a large catalytic lumenal domain. Regions specifying Golgi localization were identified by analyzing recombinant proteins either lacking GDPase domains or containing corresponding domains from type II vacuolar IMPs. Neither deletion nor substitution of the GDPase cytoplasmic domain perturbed Golgi localization. Exchanging the GDPase TMD with vacuolar protein TMDs only marginally affected Golgi localization. Replacement of the lumenal domain resulted in mislocalization of the chimeric protein from the Golgi to the vacuole, but a similar substitution leaving 34 amino acids of the GDPase lumenal domain intact was properly localized. These results identify a major Golgi localization determinant in the membrane-adjacent lumenal region (stem) of GDPase. Although necessary, the stem domain is not sufficient to mediate localization; in addition, a membrane-anchoring domain and either the cytoplasmic or full-length lumenal domain must be present to maintain Golgi residence. The importance of lumenal domain sequences in GDPase Golgi localization and the requirement for multiple hydrophilic protein domains support a model for Golgi localization invoking protein–protein interactions rather than interactions between the TMD and the lipid bilayer.