Intracellular oxygen determined by respiration regulates localization of Ras and prenylated proteins

Reduction of mitochondrial DNA (mtDNA) content induces the reduction of oxidative phosphorylation and dependence on fermentative glycolysis, that is, the Warburg effect. In aggressive prostate cancer (PCa), the reduction of mtDNA reduces oxygen consumption, increases intracellular oxygen concentration, and induces constitutive activation of Ras. Many essential proteins for cell death, growth, differentiation, and development, such as Ras, require prenylation for subcellular localization and activation. Prenylation of a protein is defined as the attachment of isoprenoids to a cysteine residue at or near the C-terminus. 3-Hydroxy-3-methyl-glutaryl-coenzyme A reductase (HMGR) produces isoprenoids, and is posttranslationally regulated by oxygen. We investigated a critical role of intracellular oxygen in membrane localization of prenylated proteins. Localization of prenylated proteins (H-Ras, prelamin A/C, and Rab5a) was observed in poorly differentiated PCa (PC-3) and well-differentiated PCa (LNCaP) cells. PC-3 cells exhibited high intracellular oxygen concentration, and H-Ras, prelamin A/C, and Rab5a were localized to various membranes (Golgi and plasma membrane, nuclear membrane, and early endosomes, respectively). Remarkably, exogenous hypoxia (0.2% O2) in PC-3 cells induced intracellular hypoxia and changed the localization of the prenylated proteins. H-Ras and Rab5a were translocated to cytosol, and prelamin A/C was in the nucleus forming an abnormal nuclear envelope. The localization was reversed by mevalonate indicating the involvement of mevalonate pathway. In contrast, in LNCaP cells, exhibiting low intracellular oxygen concentration, H-Ras and Rab5a were localized in the cytosol, and prelamin A/C was inside the nucleus forming an inadequate nuclear envelope. Exogenous hyperoxia (40% O2) increased the intracellular oxygen concentration and induced Ras translocation from cytosol to the membrane. Prelamin A/C was translocated to the nuclear membrane and formed a proper nuclear envelope. Rab5a was translocated to the early endosomes. The specific localizations of the prenylated proteins were dependent on intracellular oxygen concentration. These results demonstrate that intracellular oxygen concentration regulates the localization and activation of prenylated proteins.

Mitochondrial respiratory function regulates intracellular oxygen concentration. 1 Reduction of mitochondrial DNA (mtDNA) content induces the reduction of oxidative phosphorylation and dependence on fermentative glycolysis, that is, the Warburg effect. 2,3 Reduction of oxidative phosphorylation reduces oxygen consumption, therefore, increases intracellular oxygen concentration. Our previous studies have shown that a reduction of mtDNA induces the aggressive phenotype of prostate cancer (PCa) through increasing oxygen concentration. 4 The results also showed that the increase in oxygen concentration constitutively activated Ras via overexpression of 3-Hydroxy-3-methyl-glutaryl-coenzyme A reductase (HMGR). 4 Ras is an essential protein of signaling pathways in normal and abnormal cellular functions for cell death, growth, differentiation, and development. 5 Ras has been the focus of much attention in cancer biology owing to the substantial amount of genetic and/or functional alterations in human cancers. 5 Ras, a small GTPase, is activated and inactivated by binding to GTP and GDP, respectively. 6 Ras must localize in the membrane in order to be activated and transduce signals. 5 The membrane localization of Ras is mediated by prenylation. Many essential proteins, like Ras, require prenylation for subcellular localization and activation. Prenylation of a protein is defined as the attachment of isoprenoids to a cysteine residue at or near the C-terminus. 6 Most prenylated proteins have a consensus sequence, a CAAX box, at the C-terminus. 7 Others, like some of Rab family proteins, have C-terminus cysteine residue(s) that serve the same function as the consensus sequence. 7 Isoprenoids, farnesylpyrophosphate (FPP) and geranylgeranylpyrophosphate (GGPP), are produced in the mevalonate pathway 8 and regulates prenylation. Prenylated proteins include Ras, nuclear lamins, small GTPases, protein kinases and phosphatase, helicases, and others. The synthesis of FPP and GGPP is regulated by HMGR, a rate-limiting enzyme in the mevalonate pathway. 9 HMGR synthesizes mevalonate from 3-hydroxy-3-methyl-glutaryl-CoA (HMG-CoA). Hypoxia is known to stimulate the degradation of HMGR. 10 We hypothesize that intracellular oxygen concentration determined by mitochondria is a critical regulator of localization and activation of prenylated proteins via control of prenylation.

Results
Determination of H-Ras localization by intracellular oxygen concentration. Our previous report demonstrated that well-differentiated PCa cells (LNCaP) had the greater copy number of mtDNA, the higher oxygen consumption, but Ras expression and activation were reduced as compared with those in poorly differentiated PCa cells (PC-3). 4 Ras is farnesylated on the CAAX-motif located at the C-terminus and is translocated to endomembrane organelles, especially the Golgi apparatus, before arriving at the plasma membrane. 11 Immunofluorescence analysis demonstrated that in LNCaP, Ras localized only in cytosolic fraction, whereas in PC-3, Ras was present in the endomembrane organelles (Golgi) and plasma membrane 11 (Figure 1a). Figure 1a also confirmed the previous results that there was significantly less expression of Ras in LNCaP than that in PC-3. To demonstrate the localization of Ras in live cells, LNCaP and PC-3 cells were transiently transfected with a EGFP-H-Ras fusion protein. 11 In LNCaP cells, cytosolic H-Ras was predominant and was distributed over the entire cell with only a minor amount localized to the plasma membrane ( Figure 1b). In contrast, H-Ras in PC-3 cells accumulated in the plasma membrane as well as in the endomembrane organelles (Figure 1b). Ali et al. 11 showed that H-Ras transits the classical secretory pathway from the cytoplasm to the plasma membrane through the Golgi apparatus. A 3-dimensional video of z-stack data from a cell (Supplementary Movie. 1) demonstrated that H-Ras in LNCaP cells primarily localized to the cytosol, whereas in PC-3 cells, H-Ras localized to the plasma membrane and the Golgi (Supplementary Movie. 2). The membrane localization of Ras in PC-3 cells was significantly inhibited by treatment of lovastatin, an inhibitor of HMGR, 12 and was completely reversed by the addition of mevalonate, a product of HMGR ( Figure 1b and Supplementary Movie. 3). Mevalonate treatment alone did not alter the membrane localization of Ras significantly in PC-3 cells (Figure 1b). These data demonstrate that the localization of Ras is regulated by HMGR and is downstream of the mevalonate pathway.
A recent study showed that hypoxia regulates the expression of HMGR via stimulating proteasomal degradation. 10 LNCaP cells showed notably higher oxygen consumption rate than PC-3 cells. 4 Additionally, the depletion of mtDNA in LNCaP cells led to a reduction of oxygen consumption and an increase in the expression of HMGR, suggesting that mtDNA regulated the oxygen concentration and possibly the expression of HMGR in a cell. 4 In the current study, the oxygen concentration surrounding cells in an open environment (where atmospheric oxygen freely diffuses into the media) was measured by the Oxoplate. The oxygen concentration in medium surrounding LNCaP cells was significantly less than that surrounding PC-3 cells (Figures 1c and d).
To measure the oxygen concentration directly inside a cell, acetylacetonatobis [2-(2'-benzothienyl)pyridinato-kN,kC 3 '] iridium(III) (BTP) was used to examine intracellular oxygen concentration in LNCaP and PC-3 cells. 13 BTP is a hypoxiasensitive red phosphorescent molecule, which is quenched by available oxygen in a cell. 13 A significantly higher level of BTP phosphorescence was measured in LNCaP cells compared with PC-3 cells in the atmospheric condition (normoxia) (Figure 1e), clearly demonstrating that LNCaP cells exhibited low intracellular oxygen concentration, whereas PC-3 cells exhibited high intracellular oxygen concentration. LNCaP cells experienced endogenous hypoxia but PC-3 cells did not, possibly because of the significant differences in mitochondrial respiratory activity. LNCaP and PC-3 cells were incubated in exogenous hyperoxic (40% O 2 ) and hypoxic (0.2% O 2 ) conditions, respectively. Exogenous hyperoxia quenched BTP phosphorescence in LNCaP cells and exogenous hypoxia increased that in PC-3 cells (Figure 1e). These exogenous oxygen conditions were sufficient to regulate the intracellular oxygen concentration in LNCaP and PC-3 cells.
Regulation of H-Ras localization to membranes via various oxygen conditions. Protein prenylation is regulated by the generation of FPP and GGPP. 14 As HMGR is a rate-limiting enzyme of the mevalonate pathway and is readily degraded by hypoxia, 10  S1B and S1C. Then, the localizations of the protein in the various oxygen conditions were compared. The relative fluorescence of EGFP-H-Ras in PC-3 cells was notably higher in the membrane than that in the cytosol (Figures 2b and 3a). The exposure to the exogenous hypoxia caused the PC-3 cells to significantly lose the specific membrane localization of the protein, therefore, the relative fluorescence of EGFP-H-Ras in the cytosol was high, and the ratio of membrane to cytosol localization of H-Ras was significantly reduced (Figures 2c and 3a).  (Figures 2e and  3a). The exogenous hyperoxic conditions restored the specific membrane localization of the protein, therefore, the relative fluorescence of EGFP-H-Ras was drastically high in the membrane and the ratio of membrane to cytosolic localization was significantly increased (Figures 2f and 3a).  Figure S1B and S1C C-terminus. 7 In addition to H-Ras, the current study examined the effect of intracellular oxygen concentration on the membrane localization of other CAAX-box proteins, such as prelamin A/C. Prelamin A/C is a precursor of lamin A and C, which are important nuclear envelope proteins. 15,16 The nuclear envelope serves a crucial role in nuclear functions such as cell division and the regulation of transcription and translation. 17 Defects in nuclear envelope are observed in Hutchinson-Gilford Progeria Syndrome, a genetic disease which has a point mutation in prelamin A leading to form an improper nuclear envelope. 18 To monitor the localization of prelamin A/C, cells were transfected with DsRed-prelamin A/C. 19 The majority of prelamin A/C in PC-3 cells was localized to the nuclear membrane, resulting in the appropriate nuclear envelope formation (Figures 4a and b). To examine the effect of changes in oxygen concentration on the nuclear membrane localization of prelamin A/C, DsRed-prelamin A/C-transfected LNCaP and PC-3 cells were incubated in various exogenous oxygen concentrations (Supplementary Figure S2A). In exogenous hypoxia, the formation of the nuclear envelope of PC-3 cells was significantly inhibited (Figure 4a). In addition, prelamin A/C localized inside the nucleus rather than on the nuclear membrane (Figures 4a and c). In contrast, prelamin A/C in LNCaP cells was not only partially localized to the nuclear membrane but also localized inside the nucleus, demonstrating a defect in nuclear envelope formation, similar to that observed in Hutchinson-Gilford Progeria Syndrome (Figures 4d and e). 18 Exogenous hyperoxic conditions induced prelamin A/C localization to the nuclear membrane and formation of a proper nuclear envelope in LNCaP cells (Figures 4d and f). The time-course video analysis demonstrated a significant reduction in prelamin A/C inside the nucleus and the formation of a proper nuclear membrane within 3 h (Figure 4g and Supplementary Movie 7). These results suggest that prelamin A/C localizes on the nuclear membrane and forms a normal nuclear envelope under high intracellular oxygen concentration. In stark contrast, intracellular hypoxia results in poor prelamin A/C processing, leading to abnormal nuclear envelope formation. Collectively, the precise intracellular localization of CAAX-box proteins, such as H-Ras and prelamin A/C, are regulated by intracellular oxygen concentration.
Regulation of localization of Rab5a by intracellular oxygen concentration. GGPP is the product from the mevalonate pathway that is transferred to prenylated proteins via geranylgeranyltransferases (GGTs). 23 There are two types of GGTs: GGTI and GGTII (also known as Rab-geranyltransferase). Rab proteins are small GTPases and play an important role in vesicle formation and transport. 21,22 Most Rab proteins do not contain the CAAX sequence; instead, they have a single cysteine residue targeted by GGTII. 20 Rab5a gets geranylgeranylated before being transported to the membrane. 23 We transiently transfected cells with pEGFP-Rab5a, and monitored the localization of the protein (Supplementary Figure S2B). In endogenously normoxic PC-3 cells, Rab5a was observed mostly in endomembrane organelles (Figures 5a and b). These endomembrane organelles are the early endosomes, shown in other studies done by Ali et al. 11 Next, the effect of oxygen concentration on the localization of Rab5a in PC-3 and LNCaP cells was examined (Supplementary Figure  S2B). In contrast to the control, exogenous hypoxia induced Rab5a in PC-3 cells to be in the cytosol, indicating These results indicate that the localization of prenylated proteins without CAAX sequences is also regulated by intracellular oxygen concentration. Rab proteins are greatly involved in vesicle transport as they regulate the fusion of vesicles to target membranes. Intracellular oxygen concentration may be able to regulate membrane localization of Rab proteins leading to control vesicle transport events.
Determination of the role of intracellular oxygen concentration controlled by mitochondrial respiration in prenylated proteins. Hypoxia stimulates the degradation of HMGR in the mevalonate pathway. The mevalonate pathway produces FPP and GGPP to induce prenylation. We investigated whether the oxygen, determined by mitochondrial respiration, is the critical molecule to regulate localization (prenylation) via HMGR. Our previous studies have shown that Rotenone, a mitochondrial respiratory inhibitor, increased intracellular oxygen concentration 4,24 and increased HMGR expression and Ras activation. 4 In the current report, Rotenone drastically induced the membrane localization of H-Ras, indicating that the mitochondrial respiratory activity controlled the oxygen concentration and regulated membrane localization of H-Ras (Figure 6a). pEGFP-H-Ras-and pEGFP-Rab5a transfected-PC-3 cells were incubated in the exogenous hypoxia with or without mevalonate, the product of HMGR. As described above, hypoxic conditions inhibited the membrane localization of EGFP-H-Ras in the PC-3 cells, whereas, mevalonate enhanced its membrane localization slightly (Figure 6b). When the cells were treated with mevalonate and incubated in hypoxic conditions for 6 h, H-Ras was localized in the Golgi and the plasma membrane, indicating that the inhibition of H-Ras membrane localization by the exogenous hypoxia was completely reversed by mevalonate (Figure 6b). Similarly, EGFP-Rab5a was predominantly in the cytosol, exhibiting a diffused staining pattern, after the exogenous hypoxia treatment (Figure 6c). Mevalonate increased the localization in the early endosomes. The combination of the exogenous hypoxic conditions and mevalonate treatment restored the localization of Rab5a to the early endosomes (Figure 6c). The results suggest that intracellular oxygen is the key molecule regulating HMGR by which the isoprenoids for prenylation was produced in the mevalonate pathway.

Discussion
Here, we showed that the intracellular oxygen concentration regulated membrane localization and activation of various prenylated proteins. The regulation of the prenylation is through the hypoxia-stimulated degradation of HMGR in the mevalonate pathway. The exogenous oxygen conditions were able to manipulate intracellular oxygen concentration and the localization of the CAAX-box-containing proteins, H-Ras and prelamin A/C, and non-CAAX-box protein, Rab5a. Reduction of mitochondrial respiration increases intracellular oxygen concentration. 4 Several mechanisms can inhibit oxygen consumption, such as changes of mtDNA 25 and mutations to specific nuclear genes. 26,27 Another study implies that changes in expression of glycolytic genes can lead to the reduction of oxidative phosphorylation and an increase in glycolysis. 28 We showed here a direct link between the reduction of mitochondrial respiration and the membrane localization of prenylated proteins. The reduction of oxidative phosphorylation increases intracellular oxygen concentrations, leading to the prenylation and membrane localization of the proteins (Figure 7).
Our data showed that intracellular oxygen concentration regulated the membrane localization and activation of Ras, suggesting that the Warburg effect was the cause for Ras activation. Several studies indicate that Ras is greatly involved in the reprogramming of metabolism of cancer. 5 Ras plays an important role in an increase of glycolysis by inducing glucose uptake via GLUT1 and increasing transcription and translation of glycolytic enzymes, 5 and is considered the cause for the Warburg effect. In addition, a recent study reported that K-Ras transformation(G12V) led to mitochondrial dysfunction and a metabolic switch from oxidative phosphorylation to glycolysis. 29 Reduction of oxidative phosphorylation can lead to membrane localization and subsequent activation of Ras. The activated Ras then triggers the reprogramming of metabolism of cancer cells to increase glycolysis, which further amplifies the activation of Ras. The Warburg effect and the activation of Ras may generate positive feedback. Prelamin A/C and Rab5a are essential proteins which are prenylated to localize at the final places. Mutations on the farnesylated CaaX site of prelamin A/C are indicated to cause HGPS. Here we discovered that hypoxia impaired the formation of nuclear envelope and led prelamin A/C to localize inside of the nuclear lumen, mimicking the nuclear envelope in the HGPS patients. 30 We also observed that early endosome localization of Rab5a was inhibited by hypoxia, and the hyperoxic conditions or mevalonate were able to reverse the inhibitory effect of the intracellular hypoxia. The detailed mechanism needs to be elucidated.
Prenylation is an important posttranslational modification for biological functions. There are more than 300 CAAX-box proteins in a cell and they are essential proteins such as G-proteins, helicases, phosphatases, kinases, phospholipases, etc. 31 Small GTP-binding proteins control many other important intracellular processes, including trafficking of membrane compartments within eukaryotic cells and cytoskeleton function. 32 Prelamin and helicases regulate gene transcription and translation. 18 Prenylated proteins, such as Ras, play a central role for cancer development. 5 Reduction of mitochondrial respiration enhances cancer to exhibit more aggressive phenotypes 4 suggesting that endogenous normoxia-induced prenylation may induce a more aggressive phenotype of cancer. Therefore, the prenylated proteins could be targeted for anti-cancer drug development. Epidemiological analysis showed that in statin users, the number of cancer-related deaths is reduced. Inhibitors of prenyltransferase (PTI) have been subjected in preclinical and clinical studies for its efficacy as cancer treatment. 33,34 Regulation of intracellular oxygen concentration may enhance the efficacy of PTIs as intracellular oxygen concentration can regulate induction of FPP and GGPP via HMGR. Translocation of H-Ras and Rab5a by enhancing intracellular oxygen was quite rapid (30 min) but that of prelamin A/C was slow (3 h). This could be that the prenylated prelamin requires more time to localize to the nucleus and then translocate to the nuclear membrane.
Hypoxia-stimulated proteasomal degradation of HMGR is mediated by the accumulation of sterols in the mevalonate pathway. 10 The final product of the mevalonate pathway is cholesterol, and enzymes catalyzing cholesterol from lanosterol are oxygen-requiring enzymes. During hypoxia, the oxygen-requiring enzymes are inhibited; therefore, cholesterol intermediate molecules are accumulated in a cell. The accumulation of the sterols triggers insig-mediated ubiquitination of HMGR, leading to the degradation of the protein. 35 Direct measurement of amount for lanosterol and its derivative molecules in a cell will be necessary to determine the role of intracellular oxygen concentration in the expression of HMGR. 35 Several reports indicate that HMGR is transcriptionally upregulated. In HepG2 cells, weak hypoxia led to an increase in transcription of HMGR. 36 The effect of hypoxia in murine models has been investigated using ischemic/hypoxic conditions, and it increased HMGR expression transcriptionally. 37 In contrast, hypoxia affected only lipid metabolism without affecting cholesterol synthesis. 38 Weak hypoxia (2% O 2 ) may enhance the mRNA level of HMGR though activation of SREBP by HIF-1. Our data showed that low intracellular oxygen concentration induced the posttranscriptional modification (i.e., degradation of the protein via ubiquitination), possibly through Insig activation by accumulation of sterol intermediates. 10 Therefore, it is very likely that weak hypoxia may induce transcriptional activation of HMGR and strong hypoxia (0.2% O 2 ) will induce proteolytic degradation of HMGR.
The current report investigated the roles of intracellular oxygen concentration in the activation of prenylated proteins. Many prenylated proteins were involved in numerous essential pathways for cell growth and death, differentiation, and proliferation. The main focus of this paper is to elucidate the link between oxygen changes and the activation of Ras and other prenylated proteins, and it will provide better insight to progression of the many human diseases, such as cancer and aging. Cell culture and transfection. PC-3 was purchased from ATCC (Manassa, VA, USA) and LNCaP was purchased from UROCOR (Minnetonka, NM, USA). LNCaP and PC-3 were maintained in RPMI (Life Technologies, Grand Island, NJ, USA) plus 5% fetal calf serum (Life Technologies). All cell lines were maintained at 37°C with room air plus 5% CO 2 unless noted for specific oxygen experiments. PC-3 and LNCaP cells were plated and grown in glass-bottom 35 mm dish to be approximately 90% confluency then transfected with EGFP-H-Ras, 10 EGFP-Rab5a, 10 or DsRed-Lamin. 19 The transiently transfected cells were subjected to confocal microscopy.
Laser confocal microscopy. Cells were maintained in a stage-mounted atmospheric box (Pathology Devices, Westminster, MD, USA) at 37°C, 5% CO 2 , and 75% humidity during the course of the experiments. All samples were analyzed on an Olympus Fluoview FV1000 laser confocal microscope (Olympus, Center Valley, PA, USA). For detection of intracellular oxygen concentration, acetylacetonatobis [2-(2′-benzothienyl)pyridinato-kN,kC3']iridium(III) (BTP) (515 excitation/620 emission) was utilized in the indicated experiments at a concentration of 5 μM in RPMI medium. BTP has been described in detail by Zhang et al. 13 BTP is a phosphorescent compound that is phosphorescent in low-oxygen conditions and is quenched in the presence of oxygen. The extent of quenching is dependent upon intracellular oxygen concentration. Samples were incubated in the presence of BTP for 1 h before imaging. In all experiments, cells were plated and grown overnight at a cell density of 1 × 10 5  Determination of oxygen concentration surrounding cells using oxoplate. Oxoplate OP96F plates were purchased from PreSens (Innovative Instruments, Indian Trail, NC, USA). The Oxoplate system has been described in detail by Cook et al. 4 LNCaP and PC-3 were grown in 10 cm culture dishes under normal culture conditions until the cells had reach approximately 80% confluency and then were trypsinized to obtain cell suspensions. Cells were added to the Oxoplate at a concentration of 1 × 10 6 cells/ml in RPMI plus 5% fetal calf serum (200 μl final volume). Fluorescence in each well was then measured every 5 min for 3 h at 37°C in a plate reader (Synergy HT, BIO-TEK, Winooski, VT, USA). The oxygen concentration in the wells at each time point was calculated. All samples were run in triplicate and the indicated error bars were standard error of each sample.