Phospholipid metabolism regulated by a transcription factor sensing phosphatidic acid.

Cells regulate the biophysical properties of their membranes by coordinated synthesis of different classes of lipids. Here, we identified a highly dynamic feedback mechanism by which the budding yeast Saccharomyces cerevisiae can regulate phospholipid biosynthesis. Phosphatidic acid on the endoplasmic reticulum directly bound to the soluble transcriptional repressor Opi1p to maintain it as inactive outside the nucleus. After the addition of the lipid precursor inositol, this phosphatidic acid was rapidly consumed, releasing Opi1p from the endoplasmic reticulum and allowing its nuclear translocation and repression of target genes. Thus, phosphatidic acid appears to be both an essential ubiquitous metabolic intermediate and a signaling lipid.

Mitotic Golgi is in a dynamic equilibrium between clustered and free vesicles independent of the ER.

Golgi inheritance during cell division involves Golgi disassembly but it remains unclear whether the breakdown product is dispersed vesicles, clusters of vesicles or a fused ER/Golgi network. Evidence against the fused ER/Golgi hypothesis was previously obtained from subcellular fractionation studies, but left concerns about the means used to obtain and disrupt mitotic cells. Here, we performed velocity gradient analysis on otherwise untreated cells shaken from plates 9 h after release from an S-phase block. In addition, we used digitonin and freeze/thaw permeabilization as alternatives to mechanical homogenization. Under each of these conditions, approximately 75% of the Golgi was recovered in a population of small vesicles that lacked detectable ER. We also used multilabel fluorescent microscopy with optical sectioning by deconvolution to compare the 3D metaphase staining pattern of endogenous Golgi and ER markers. Although both ER and Golgi staining were primarily diffuse, only the ER was excluded from the mitotic spindle region. Surprisingly, only 2% of the Golgi fluorescence was present as resolvable structures previously characterized as vesicle clusters. These were not present in the ER pattern. Significantly, a portion of the diffuse Golgi fluorescence, presumably representing dispersed 60-nm vesicles, underwent an apparent rapid aggregation with the larger Golgi structures upon treatments that impaired microtubule integrity. Therefore, mitotic Golgi appears to be in a dynamic equilibrium between clustered and free vesicles, and accurate partitioning may be facilitated by microtubule-based motors acting on the clusters to insure random and uniform distribution of the vesicles.

Mitotic phosphorylation of Golgi reassembly stacking protein 55 by mitogen-activated protein kinase ERK2.

The role of the mitogen-activated protein kinase kinase (MKK)/extracellular-activated protein kinase (ERK) pathway in mitotic Golgi disassembly is controversial, in part because Golgi-localized targets have not been identified. We observed that Golgi reassembly stacking protein 55 (GRASP55) was phosphorylated in mitotic cells and extracts, generating a mitosis-specific phospho-epitope recognized by the MPM2 mAb. This phosphorylation was prevented by mutation of ERK consensus sites in GRASP55. GRASP55 mitotic phosphorylation was significantly reduced, both in vitro and in vivo, by treatment with U0126, a potent and specific inhibitor of MKK and thus ERK activation. Furthermore, ERK2 directly phosphorylated GRASP55 on the same residues that generated the MPM2 phospho-epitope. These results are the first demonstration of GRASP55 mitotic phosphorylation and indicate that the MKK/ERK pathway directly phosphorylates the Golgi during mitosis.

Binding relationships of membrane tethering components. The giantin N terminus and the GM130 N terminus compete for binding to the p115 C terminus.

By forming a molecular tether between two membranes, p115, giantin, and GM130 may mediate multiple Golgi-related processes including vesicle transport, cisternae formation, and cisternal stacking. The tether is proposed to involve the simultaneous binding of p115 to giantin on one membrane and to GM130 on another membrane. To explore this model, we tested for the presence of the putative giantin-p115-GM130 ternary complex. We first mapped p115-binding site in giantin to a 70-amino acid coiled-coil domain at the extreme N terminus, a position that may exist up to 400 nm away from the Golgi membrane. We then generated glutathione S-transferase (GST) fusion proteins containing either giantin's or GM130's p115 binding site and tested whether such proteins could bind p115 and GM130 or bind p115 and giantin, respectively. Unexpectedly, GST fusions containing either the giantin or the GM130 p115 binding site efficiently bound p115, but the p115 bound to GST-giantin did not bind GM130, and the p115 bound to GST-GM130 did not bind giantin. To explain this result, we mapped the giantin binding site in p115 and found that it is located at the C-terminal acidic domain, the same domain involved in binding GM130. The presence of a single binding site in p115 for giantin and GM130 was confirmed by demonstration that giantin and GM130 compete for binding to p115. These results question a simple tethering model involving a ternary giantin-p115-GM130 complex and suggest that p115-giantin and p115-GM130 interactions might mediate independent membrane tethering events.

The Golgi and endoplasmic reticulum remain independent during mitosis in HeLa cells.

Partitioning of the mammalian Golgi apparatus during cell division involves disassembly at M-phase. Despite the importance of the disassembly/reassembly pathway in Golgi biogenesis, it remains unclear whether mitotic Golgi breakdown in vivo proceeds by direct vesiculation or involves fusion with the endoplasmic reticulum (ER). To test whether mitotic Golgi is fused with the ER, we compared the distribution of ER and Golgi proteins in interphase and mitotic HeLa cells by immunofluorescence microscopy, velocity gradient fractionation, and density gradient fractionation. While mitotic ER appeared to be a fine reticulum excluded from the region containing the spindle-pole body, mitotic Golgi appeared to be dispersed small vesicles that penetrated the area containing spindle microtubules. After cell disruption, M-phase Golgi was recovered in two size classes. The major breakdown product, accounting for at least 75% of the Golgi, was a population of 60-nm vesicles that were completely separated from the ER using velocity gradient separation. The minor breakdown product was a larger, more heterogenously sized, membrane population. Double-label fluorescence analysis of these membranes indicated that this portion of mitotic Golgi also lacked detectable ER marker proteins. Therefore we conclude that the ER and Golgi remain distinct at M-phase in HeLa cells. To test whether the 60-nm vesicles might form from the ER at M-phase as the result of a two-step vesiculation pathway involving ER-Golgi fusion followed by Golgi vesicle budding, mitotic cells were generated with fused ER and Golgi by brefeldin A treatment. Upon brefeldin A removal, Golgi vesicles did not emerge from the ER. In contrast, the Golgi readily reformed from similarly treated interphase cells. We conclude that Golgi-derived vesicles remain distinct from the ER in mitotic HeLa cells, and that mitotic cells lack the capacity of interphase cells for Golgi reemergence from the ER. These experiments suggest that mitotic Golgi breakdown proceeds by direct vesiculation independent of the ER.

Stimulation of the DNA-dependent protein kinase by RNA polymerase II transcriptional activator proteins.

The DNA-dependent protein kinase (DNA-PK) phosphorylates RNA polymerase II and a number of transcription factors. We now show that the activity of DNA-PK is directly stimulated by certain transcriptional activator proteins, including the human heat shock transcription factor 1 (HSF1) and a transcriptionally active N-terminal 147 amino acid GAL4 derivative. Stimulation of DNA-PK activity required specific sequences in the activator proteins outside the minimal DNA binding domains. The stimulation of DNA-PK activity also required DNA and was greater with DNA containing relevant activator binding sites. Comparison of different HSF binding fragments showed that optimal stimulation occurred when two HSF binding sites were present. Stimulation with HSF and GAL4 was synergistic with Ku protein, another regulator of DNA-PK activity. DNA-PK is tightly associated with the transcriptional template, and an increase in its activity could potentially influence transcription through the phosphorylation of proteins associated with the transcription complex.

DNA-dependent protein kinase (Ku protein-p350 complex) assembles on double-stranded DNA.

The Ku protein is an autoantigen that consists of 70- and 80-kDa polypeptides. It associates with double-stranded DNA at free ends. In the present study, we examined the ability of anti-Ku antibodies to immunoprecipitate various structures from extracts of HeLa cells prepared at different salt concentrations. Under physiological conditions, these antibodies identified a complex containing the Ku protein and the 350-kDa component (p350) of DNA-dependent protein kinase (DNA-PK), which appeared to be closely associated on the DNA strand. In reconstitution experiments with cell extracts and biochemically purified components, the Ku protein-p350 complex formed only in the presence of double-stranded DNA. The reconstituted complex was catalytically active. Together with previous studies, these results indicate that the Ku protein interacts with DNA to create a binding site for p350 as the DNA-PK holoenzyme assembles.

5-Azacytidine-induced demethylation of DNA to senescent level does not block proliferation of human fibroblasts.

IMR-90 human diploid fibroblasts (HDF) lose from 30-50% of their genomic 5-methyldeoxycytidine (5mdC) during the cellular aging process. In contrast, immortal SV40-transformed IMR-90 maintain a constant level of 5mdC in culture. Precrisis SV40-transformed HDF (AG3204) represent a stage in between normal cell aging and immortalization because these cells still have a finite proliferative lifespan, but it is longer than that of normal HDF and ends in cell death rather than in G1-arrest. We find that AG3204 cells continue to lose from 12-33% of their 5mdC after a population has become 99% positive for SV40 T-antigen. Both IMR-90 cells and AG3204 cells have similar levels of 5mdC (average of 2.25%) at the end of lifespan. We investigated whether this level of 5mdC is an absolute block to further proliferation by treating IMR-90 and AG3204 cells with 5-azacytidine (5azaC) to reduce their 5mdC levels below the terminal level normally achieved at end of lifespan. We find that both IMR-90 and AG3204 cells undergo extensive proliferation with subterminal levels of 5mdC and that the lifespans of both cell types are shortened by 5azaC treatment. These studies indicate that random genomic DNA demethylation to a specific level of 5mdC is not a direct cause of finite proliferative lifespan. However, the correlation between accelerated DNA demethylation and accelerated aging still suggests that these two phenomena are related. Two ways to explain this relationship are: (1) DNA demethylation during aging is not random, and/or (2) both DNA demethylation and other independent aging processes cooperate to produce finite lifespan. In both cases, accelerated random DNA demethylation could accelerate aging, but not necessarily in direct relationship to the final genomic level of 5mdC achieved during the normal aging process.