Calreticulin Destabilizes Glucose Transporter-1 mRNA in Vascular Endothelial and Smooth Muscle Cells Under High-Glucose Conditions
http://www.100md.com
循环研究杂志 2005年第11期
the Department of Pharmacology (H.T.-J., Y.R., S.S.), School of Pharmacy, Faculty of Medicine, The Hebrew University of Jerusalem, Israel
The Nephrology (T.N.-M.) and Endocrinology & Metabolism (N.K.) Services, Department of Medicine, Hadassah-Hebrew University Medical Center, Jerusalem, Israel
Department of Clinical Biochemistry and Pathobiochemistry (J.E.), German Diabetes Center, De箂seldorf, Germany.
Abstract
Substrate autoregulation of glucose transporter-1 (GLUT-1) mRNA and protein expression provides vascular endothelial and smooth muscle cells a sensitive mechanism to adapt their rate of glucose transport in response to changing glycemic conditions. Hyperglycemia-induced downregulation of glucose transport is particularly important in protecting these cells against an excessive influx of glucose and consequently increased intracellular protein glycation and generation of free radicals; both are detrimental in the development of vascular disease in diabetes. We aimed to investigate the molecular mechanism of high glucoseeCinduced downregulation of GLUT-1 mRNA expression in primary bovine aortic vascular endothelial (VEC) and smooth muscle (VSMC) cell cultures. Using RNA mobility shift, UV cross-linking, and in vitro degradation assays, followed by mass-spectrometric analysis, we identified calreticulin as a specific destabilizing trans-acting factor that binds to a 10-nucleotide cis-acting element (CAE2181-2190) in the 3'-untranslated region of GLUT-1 mRNA. Pure calreticulin accelerated the rate of GLUT-1 mRNA-probe degradation in vitro, whereas overexpression of calreticulin in vascular cells decreased significantly the total cell content of GLUT-1 mRNA and protein. The expression of calreticulin was augmented in vascular cells exposed to high glucose in comparison with low-glucose conditions. Similarly, increased expression of calreticulin was observed in aortae of diabetic Psammomys obesus in comparison with normoglycemic controls. These data suggest that CAE2181-2190eCcalreticulin complex, which is formed in VSMC and VEC exposed to hyperglycemic conditions, renders GLUT-1 mRNA susceptible to degradation. This interaction underlies the process of downregulation of glucose transport in vascular cells under high-glucose conditions.
Key Words: calreticulin glucose transporter-1 hyperglycemia mRNA turnover vascular smooth muscle cells vascular endothelial cells
Introduction
Hyperglycemia is a major risk factor in the development of cardiovascular complications associated with the metabolic syndrome and diabetes.1,2 Chronic hyperglycemia alters the normal function of endothelial and smooth muscle cells in blood vessels, which undergo morphological and functional modifications, because of an excessive extra- and intracellular protein glycation and an uncontrolled production of free radicals. Collectively, these processes contribute to basement-membrane thickening, vascular occlusion, increased permeability, and initiation and progression of vascular disease and atherosclerosis.3eC6
The transport of glucose across the plasma membrane of cells is the rate-limiting step for subsequent glucose metabolism. A family of glucose transporters (GLUTs) mediates the entry of glucose into cells by facilitated diffusion.7 Vascular endothelial (VEC) and vascular smooth muscle (VSMC) cells express predominantly the ubiquitous GLUT-1 and to a very small extent the insulin-sensitive GLUT-4.8 We showed that these cells autoregulate their rate of glucose transport in response to changes in ambient glucose levels; specifically, they reduce their rate of hexose transport by &50% on exposure to high-glucose levels. Concomitantly, total cell GLUT-1 (but not GLUT-4) protein and mRNA content, as well as its plasma-membrane abundance, are reduced to the same extent.8 This downregulatory mechanism enables the cells to sustain normal glucose metabolism and ATP production similar to cells exposed to normal glucose levels.9
The expression of GLUT-1 is regulated posttranscriptionally in various cell types under diverse physiological and pathophysiological conditions, such as glucose deprivation, inhibition of oxidative phosphorylation, cytokine-, hormone-, and metabolite-dependent stimuli, and neoplastic transformation.10eC12 The 5' and 3' untranslated regions (UTR) of GLUT-1 mRNA are highly conserved among species. The 3' UTR contains a unique and conserved 10-nucleotide (nt) cis-acting stabilizing element (5'-CCAACCACTC; CAE2181 to 2190 in bovine GLUT-1).13,14 Recently, Boado et al showed that CAE2181-2190eCdependent interactions stabilized GLUT-1 mRNA in rat glial C6 tumor cells deprived of glucose and exposed to hypoxic conditions.15 These studies point to CAE2181-2190 as a glucose-sensitive element in GLUT-1 mRNA. Yet, the 3' UTR of GLUT-1 mRNA also contains AU-rich elements that may either stabilize or destabilize the entire molecule.16eC18
Hitherto, the molecular mechanism regulating the turnover rate of GLUT-1 mRNA in vascular cells under hyperglycemic conditions has not been elucidated. This study shows that calreticulin, whose expression is augmented in vascular cells under high-glucose conditions, destabilizes GLUT-1 mRNA by its specific interaction with CAE2181-2190.
Materials and Methods
Cell Cultures
Primary cultures of bovine aortic endothelial and smooth muscle cells were prepared and characterized as described previously.8 GH354 cells (courtesy of Dr J. M. Wilson, University of Pennsylvania, Philadelphia) were grown as described.19
Methods
The preparation of RNA transcripts, expression vectors, and adenoviral constructs, along with detailed descriptions of in vitro degradation assays, RNA mobility shift assays, UV cross-linking assays, real-time PCR, cell extract preparation, Western blot analyses, and immunoprecipitation are given in the online data supplement available at http://circres.ahajournals.org.
Statistical Analysis
Statistical analyses were performed using the nonparametric ManneCWhitney test.
Results
High GlucoseeC and CAE2181-2190eCDependent Degradation of GLUT-1 mRNA Probes
The role of CAE2181-2190 in regulating bovine GLUT-1 mRNA stability in vascular cells was studied in an in vitro degradation assay. [32P]-Labeled 203-nt (5' 2098 to 2300) and 203-ntCAE2181-2190 probes of the 3 UTR of bovine GLUT-1 mRNA were incubated at 37°C with S100 cytosolic proteins of VSMC that had been maintained at 2 or 25 mmol/L glucose. Figure 1A shows a rapid degradation of the intact probe only in the presence of the high-glucose extract. However, deletion of CAE2181-2190 from the probe rendered it stable to degradation under similar conditions.
The role of CAE2181-2190 in regulating the expression of GLUT-1 mRNA in intact cells was also investigated. The 203- or 203CAE2181-2190-nt sequences of GLUT-1 mRNA were inserted into the luciferase expression vector pGL2 at the unique pflM1 restriction site located in the luciferase 3'-UTR region. VSMC were transfected with a green fluorescent protein (GFP)-expression plasmid (pEGFP-N1) without or with the control or modified pGL2 vectors and incubated at 25 mmol/L glucose for 48 hours. GFP-positive cells were collected by fluorescence-activated cell sorting (FACS) and taken for real-time PCR analyses of GLUT-1 and luciferase mRNA. Figure 1B shows an increased abundance of GLUT-1 mRNA in cells transfected with the 203-nteCinserted pGL2 vector in comparison with cells transfected with the unmodified vector; concomitantly, luciferase mRNA level seemed decreased. These data suggest that CAE2181-2190 in luciferase mRNA competed with the same element in the 3' UTR of the endogenous GLUT-1 mRNA for binding of destabilizing factors, thus rendering the latter transcript less susceptible to degradation. However, the expression of luciferase mRNA increased and GLUT-1 mRNA remained unaltered in cells transfected with the 203CAE2181-2190eCinserted pGL2 vector. It appears that in the absence of competing CAE2181-2190 sequences in luciferase mRNA molecules, the same element in endogenous GLUT-1 mRNA interacted with destabilizing factors, as it does in control nontransfected cells. Yet, the increased expression of luciferase mRNA may reflect interactions of cellular stabilizing factors with other elements within the modified luciferase mRNA that were effective in the absence of CAE2181-2190eCmediated destabilizing interactions.
CAE2181-2190eCCytosolic ProteineCBinding Interactions
RNA electrophoretic mobility-shift assay (REMSA) of the [32P]-labeled 203-nt probe in the presence of increasing protein concentrations of a high-glucose VSMC protein extract was performed. Figure 2A shows 2 complexes, the binding intensity of which was augmented with increasing protein concentrations (upper arrows), reaching maximum with 3 e of extract, and a lower diffused band. The specificity of these binding interactions was investigated by adding 20- to 200-fold molar excess of the unlabeled 203-nt transcript or an unrelated transcript of 3' UTR of parathyroid hormone mRNA to the reaction mixture (Figure 2B).20 As expected, only the former competed effectively with the labeled probe for binding of cytosolic proteins.
The capacity of low- and high-glucose cell extracts to interact with CAE2181-2190eCcontaining probes are shown in Figure 3A. The extracts were prepared from VSMC and VEC that had been maintained at 2 or 25 mmol/L glucose for 14 or 20 hours, respectively. These were the minimal incubation periods at high glucose required to produce maximal binding interactions. The slower response of VEC agrees with our previous report on a delayed downregulatory response of these cells to high glucose.21 Figure 3A shows, in the REMSA of the [32P]-labeled 57-nt probe, 3 distinct complexes that were not apparent with the longer 203-nt probe. Similar REMSAs with CAE2181-2190eCdeleted probes were performed (Figure 3B): the intensities of the 2 upper complexes formed with the [32P]-labeled 57CAE2181-2190 probe were significantly lower in comparison with those formed with the intact 57-nt probe. Yet, the binding interactions of the [32P]-labeled 203CAE2181-2190 probe were similar to those detected with the intact probe, possibly attributable to additional binding sequences for other proteins in the longer 203-nt probe, which masked the effect of CAE2181-2190 deletion. Therefore, we used the 57-nt probe in subsequent experiments.
The molecular mass of the [32P]-labeled 57-nt probe VSMC protein complexes was determined by UV cross-linking analysis. Four complexes corresponding to 112-, 62-, 58-, and 43-kDa, the binding intensities of which were greater with high-glucose than low-glucose VSMC protein extracts, were depicted (Figure 4A, lanes 2 and 3). Unlabeled 57-nt transcript, added at a 2-fold molar excess before UV cross-linking, reduced significantly the formation of the 62- and the 58-kDa complexes and, to a lesser extent, the 112- and 43-kDa complexes (lane 4).
To elucidate which of these complexes required CAE2181-2190 for binding, 20- to 300-fold molar excesses of complementary oligonucleotides (ONDs) were hybridized to the [32P]-labeled 57-nt probe before UV cross-linking with high-glucose VSMC extract (Figure 4B). OND2 was complementary to CAE2181-2190, whereas OND1 and -3 overlapped distal up- or downstream sequences within the probe (Figure 4B). The 58- and 43-kDa complexes were noneCCAE2181-2190 dependent, because all ONDs weakly interfered with their formation. Because OND2 abolished effectively the formation of the 112- and 62-kDa complexes, both could be considered candidates for an interaction with CAE2181-2190. However, the 112-kDa complex was excluded, because its formation was also abolished by OND1. Hence, the 62-kDa complex was characterized as a CAE2181-2190eCdependent complex.
Calreticulin Is a CAE2181-2190eCBinding Protein That Destabilizes GLUT-1 mRNA
High-glucose VSMC protein extract was affinity chromatographed on a Heparin-Fractogel-EMD column with increasing concentrations of KCl. Figure 5A shows REMSA of each KCl-eluted fraction incubated with the [32P]-labeled 57-nt probe. Of the various complexes formed with the original S100 extract (left arrows), 2 were significantly enriched in the 125 and 200 mmol/L KCl eluates (right arrows). Their respective molecular weights were resolved as 43 and 62 kDa in a UV cross-linking assay (Figure 5B).
The 125 and 200 mmol/L KCl eluates were concentrated 10-fold on an Amicon Ultra-4 centrifugal filter device (30-kDa cutoff) and run on SDS-PAGE. The 43- and 62-kDa bands were excised, processed, and analyzed by mass spectrometry. The results showed sequence homologies to rat aldolase A and bovine calreticulin, respectively (online Table I). The calculated molecular mass of bovine calreticulin is 46.6 kDa, but it resolves on gels as an &60-kDa protein, because of a large number of negatively charged carboxylic groups (pI=4.7).22 Because the data in Figure 4B showed that CAE2181-2190 was essential for the formation of the 62-kDa, but not the 43-kDa, complex, we investigated the role of calreticulin in the regulation of GLUT-1 mRNA stability.
The following experiments were conducted to ascertain whether calreticulin was the protein in the 62-kDa complex. Aliquots of high-glucose VSMC protein extract and its 200 mmol/L KCl eluate were incubated with the [32P]-labeled 57-nt probe, UV-cross-linked, separated by SDS-PAGE, and transferred to a nitrocellulose membrane. The membrane was first used for Western blotting of calreticulin and then taken for phosphorimaging to visualize [32P]-labeled RNAeCprotein complexes. Figure 6A depicts 2 colocalized 62-kDa (calreticulin) bands, detected independently by each method of the same gel. In addition, Figure 6B depicts a specific binding interaction of the same probe with pure bovine calreticulin (lane 2), which disappeared in the presence of a 10-fold molar excess of unlabeled 57-nt sequence (lane 3) or following prehybridization of the probe with a 300-fold molar excess of OND2 (lane 4).
Figure 6C compares UV cross-linking profiles of the labeled 57 nt with pure bovine calreticulin and with VSMC cytosolic proteins. Whereas the typical multicomplex profile was apparent with the latter (lane 2), calreticulin formed a single 62-kDa complex (lane 3), which was blocked by prehybridization of the probe with OND2 (lane 4). Because the anti-calreticulin antibody used interacted well with only denatured calreticulin, our attempts to supershift the 62-kDa complex in native REMSAs were unsuccessful. Instead, an aliquot of VSMC protein extract was UV cross-linked with the [32P]-labeled 57-nt probe, immunoprecipitated with anti-calreticulin antibody under denaturing conditions, and electrophoresed. A single band corresponding to the same size as of the complex with pure calreticulin was observed (compare lane 5 with 3). Prehybridization of the probe with OND2 before UV illumination completely eliminated the formation of this complex (not shown). No bands were detected following immunoprecipitation with rabbit nonimmune serum (not shown).
High-Glucose Augments Calreticulin Expression in Vascular Cells In Vitro and In Vivo
VSMC and VEC were incubated in high glucose for 24 or 48 hours, respectively, to induce maximal downregulation of glucose transport and GLUT-1 protein level in VSMC and VEC, respectively, as was shown recently.21 Figure 7A and 7B show a 1.8- to 2.0-fold increase in calreticulin expression and its plasma membrane abundance and reduced GLUT-1 levels in both types of cells exposed to high glucose in comparison with low glucose.
The physiological relevance of these findings in cell cultures was confirmed by determining calreticulin content in aortic segments isolated from normoglycemic and hyperglycemic Psammomys obesus. The respective blood glucose levels of the gerbils were 109±20 (n=3) and 263±29 (n=5) mg/dL. Figure 7C shows a significant 1.99±0.28-fold increased calreticulin content in aortae of diabetic animals in comparison with their normoglycemic controls. However, because of the high level of GLUT-1 content in contaminating erythrocytes trapped within the aortic segments, the analysis of aortic GLUT-1 was unreliable. Yet, calreticulin expression in the erythrocytes was negligible (online Figure I).
Calreticulin Reduces GLUT-1 Expression in VEC and VSMC
Effects of overexpression of calreticulin on GLUT-1 protein level in VSMC and VEC were then determined. VSMC were transfected with GFP-expression plasmid (pEGFP-N1) without or with the calreticulin-expression plasmid pMCC-ZP-CTR, followed by incubations at 2 or 25 mmol/L glucose for 48 hours. GFP-positive cells were then collected by FACS, lysed, and taken for Western blot analyses. Similar to its effect in intact cells (Figure 7A), high-glucose incubation increased calreticulin and reduced GLUT-1 protein levels in control (GFP-expressing) cells in comparison with similar cells maintained at low glucose (Figure 8A, left panel, lanes 1 and 2). Moreover, overexpression of calreticulin in cells exposed to 2 mmol/L glucose mimicked the downregulatory effect of the 25 mmol/L glucose incubation on GLUT-1 content in control (GFP-expressing) cells (compare lane 3 with 2). The level of GLUT-1 protein was similarly reduced in cells overexpressing calreticulin and exposed either to low- or high-glucose conditions (lanes 3 and 4).
VEC were infected with adenoviral vectors expressing GFP or both GFP and calreticulin and incubated at 2 or 25 mmol/L glucose for 48 hours. Figure 8A (right panel, lanes 1 and 2) shows an &50% reduction in GLUT-1 content in control, GFP- expressing cells that were exposed to high glucose in comparison with the low-glucose incubation. Overexpression of calreticulin in cells exposed to 2 mmol/L glucose mimicked the effect of high glucose in control cells (compare lane 3 with 2). Overexpressed calreticulin reduced GLUT-1 levels in VEC exposed either to 2 or 25 mmol/L glucose (lanes 3 and 4).
GLUT-1 and calreticulin mRNA levels in similarly infected VEC were determined by real-time PCR. The level of calreticulin mRNA in cells infected with the calreticulin-expressing virus was 5.4±1.5- and 6.9±2.3-fold higher (n=6) under 2 and 25 mmol/L glucose incubation, respectively, in comparison with cells infected with control GFP-expressing virus. Figure 8B shows 40±17% reduction in GLUT-1 mRNA content in cells expressing GFP only and incubated at 25 mmol/L glucose in comparison with the low-glucose incubation. The level of GLUT-1 mRNA in cells infected with the calreticulin-expressing virus was reduced by 44±16% and 48±16% under 2 or 25 mmol/L glucose, respectively.
Finally, we repeated the in vitro degradation assay shown in Figure 1 and added pure calreticulin to low-glucose VSMC protein extract, which did not degrade the [32P]-labeled 203-nt probe (Figures 1 and 8C). Figure 8C and 8D shows that calreticulin promoted a rapid degradation of the intact probe but not of the [32P]-labeled 203CAE2181-2190 probe.
Discussion
The mechanism of high glucoseeCinduced downregulation of GLUT-1 mRNA expression in VEC and VSMC were investigated. As already reported, high-glucose incubation reduced both GLUT-1 mRNA and protein expression in VSMC and VEC.8 In vitro degradation assays and cell transfection with CAE2181-2190eCcontaining pGL2 plasmid indicate that CAE2181-2190 in the 3' UTR of GLUT-1 mRNA acts as a destabilizing glucose-sensitive element. REMSAs, UV cross-linking assays, and affinity chromatography purification of cell extracts, followed by mass-spectrometric analysis, identified calreticulin as a binding protein in the 62-kDa complex with this CAE. Overexpression of calreticulin in the cells significantly reduced GLUT-1 mRNA and protein content, whereas pure calreticulin promoted rapid degradation of CAE2181-2190eCcontaining transcripts. Furthermore, high-glucose incubations increased calreticulin expression in both types of cells. Similarly, increased calreticulin content was found in aortae of hyperglycemic P. obesus, in comparison with their normoglycemic controls. These findings support a model in which calreticulin interacts with CAE2181-2190 in the 3' UTR of GLUT-1 mRNA and destabilizes the entire transcript, leading to reduced GLUT-1 protein expression and plasma membrane localization and, subsequently, to downregulation of glucose transport. This adaptive mechanism protects vascular cells against damaging effects of an uncontrolled influx of glucose in face of hyperglycemia.
Early studies identified CAE2181-2190 in GLUT-1 mRNA as a glucose-sensitive element, which formed an 80-kDa mRNA stabilizing complex in brain microcapillary endothelial cells and rat C6 glioma cells.13,15 However, no such VSMC- or VEC cytosolic extracteCCAE2181-2190 complex was detected in the present study. Because GLUT-1 mRNA expression was reported to be regulated in an organ-specific manner by diverse proteins,18 GLUT-1 mRNAeCprotein complexes in vascular cells may vary from those formed in other cell types and under various experimental or pathophysiological conditions. Moreover, other CAEs in the 3' UTR of GLUT-1 mRNA can also affect GLUT-1 mRNA stability in concert with CAE2181-2190.16,17,23 Indeed, our findings on an increased expression of luciferase mRNA on transfection of VSMC with the 203CAE2181-2190eCinserted pGL2 vector (Figure 1B) indicate a role for additional stabilizing interactions within the deleted 203-nt sequence that became effective in the absence of CAE2181-2190eCmediated destabilizing effects.
Overexpression of calreticulin in vascular cells exposed to low glucose mimicked the effect of high-glucose incubation on GLUT-1 mRNA and protein levels (Figure 8A and 8B). Yet, the combined effect of high-glucose incubation and overexpression of calreticulin was not additive and was, in fact, comparable to that observed in control cells exposed to high glucose. Thus, it seems that the increased calreticulin content in cells exposed to high glucose was sufficient to induce maximal destabilization of GLUT-1 mRNA, whereas overexpressed calreticulin under these conditions was inconsequential. We showed that transfer of vascular cells from a high-glucose to lower glucose levels induced upregulation of glucose transport and GLUT-1 mRNA expression and that glucose-deprived cells increase the expression of GLUT-1 to ameliorate a metabolically challenging and stressful condition.8,21,24 We also showed that GLUT-1 was a stress-inducible protein belonging to the family of glucose-regulated proteins.24 Thus, the combination of overexpression of calreticulin and high-glucose incubation of the vascular cells could markedly reduce GLUT-1 content and cause glucose deprivationeClike effects, followed by a stress reaction aimed to increase GLUT-1 gene transcription and/or expression of GLUT-1 mRNAeCstabilizing factors. Such mechanisms could also explain the lack of additive effects of calreticulin overexpression and high-glucose incubations on GLUT-1 mRNA and protein in the cells.
Other studies also found specific RNAeCcalreticulin interactions. Destabilization of angiotensin-1 receptor mRNA in VSMC induced calreticulin binding to an AU-rich element in the 3' UTR.25 Calreticulin inhibited the translation of C/EBP transcription factors by its interaction with GCU repeats of C/EBP/ mRNAs.26 In addition, rubella virus replication and/or localization in infected cells was controlled by calreticulin interactions with in the 3' terminal of the viral RNA.27 These studies and the present results indicate the lack of a consensus RNA sequence for calreticulin binding. Indeed, computer-based predictions (http://www.bioinfo.rpi.edu/zukerm/rna) of the 2D structure of the CAE2181-2190 region in the 3'-UTR GLUT-1 mRNA and of the abovementioned RNA regions point to a stem-loop structure, rather than a consensus sequence, as a potential target for calreticulin binding.
The mechanism of high glucoseeCdependent augmented expression of calreticulin in vascular cells has not yet been investigated. We have found that high-glucose levels increase the expression of the enzyme 12-lipoxygenase and the production of its metabolite 12-hydroxyeicosatetraenoic acid (12-HETE) in vascular cells. This eicosanoid mediates high glucoseeCinduced reduction in GLUT-1 mRNA and protein levels and downregulation of hexose transport.8,21 Others have shown that peroxisome proliferatoreCactivated receptors interact with various HETEs to promote gene transcription.28eC30 These studies and present results allude to a peroxisome proliferatoreCactivated receptor interaction with 12-HETE in the mechanism mediating augmented calreticulin expression in vascular cells.
Since the discovery of calreticulin more than 30 years ago as a calcium-binding protein in the sarcoplasmic reticulum, it has become clear that this is a multifunctional protein, whose subcellular localization is not restricted to the sarcoplasmic reticulum.31 Among other functions, calreticulin chaperons proteins in the endoplasmic reticulum, activates integrins, regulates cell adhesion, induces NO synthesis, and regulates steroid-sensitive gene expression.22,32,33 The present results on high glucoseeCdependent increased expression of calreticulin in VEC and VSMC and aortae of diabetic P. obesus (Figure 7A through 7C) suggest a role for calreticulin in modifying cell-matrix interactions, chaperoning and targeting of over-glycosylated, and regulating NO production in diabetic blood vessels.34
Our results also indicate that the gluconeogenic enzyme aldolase A might interact with the 57-nt probe of the 3'-UTR region of GLUT-1 to form a non-CAE2181-2190eCdependent 43-kDa complex (Figure 4 and 5). This enzyme was found to bind specifically to the 3' UTR of myosin heavy chain mRNA and possibly target or anchor it to subcellular compartments of myosin synthesis by association with the cytoskeletal framework.35 It remains to be investigated whether aldolase A shares a similar function in subcellular targeting of GLUT-1 mRNA in vascular cells.
In Summary, this study shows a central role for calreticulin in the mechanism of downregulation of GLUT-1 mRNA in vascular cells under high-glucose conditions, which ultimately protect the cells against excessive glucose influx and its deteriorating effects.
Acknowledgments
This work was supported by grants from the German-Israel Foundation for Scientific Research and Development (I-750-165.2/2002), the Russell Berrie Foundation and Diabetes Care in Israel (D-CURE), the Israel Ministry of Health, the Yedidut Foundation Mexico and the David R. Bloom Center for Pharmacy at the Hebrew University. H.T.-J. received doctoral fellowships from the Hebrew University Center for Diabetes Research and The Israel Ministry of Science. S.S. is a member of the David R. Bloom Center for Pharmacy at the Hebrew University. We thank Dr H. Giladi for assistance in constructing adenoviral vectors.
References
Tenenbaum A, Motro M, Schwammenthal E, Fisman EZ. Macrovascular complications of metabolic syndrome: an early intervention is imperative. Int J Cardiol. 2004; 97: 167eC172.
Giles TD, Sander GE. Diabetes mellitus and heart failure: basic mechanisms, clinical features, and therapeutic considerations. Cardiol Clin. 2004; 22: 553eC568.
Brownlee M. Negative consequences of glycation. Metabolism. 2000; 49 (suppl 1): 9eC13.
Brownlee M. Biochemistry and molecular cell biology of diabetic complications. Nature. 2001; 414: 813eC820.
Lorenzi M, Cagliero E, Toledo S. Glucose toxicity for human endothelial cells in culture. Delayed replication, disturbed cell cycle, and accelerated death. Diabetes. 1985; 34: 621eC627.
Natarajan R, Gonzales N, Xu L, Nadler JL. Vascular smooth muscle cells exhibit increased growth in response to elevated glucose. Biochem Biophys Res Commun. 1992; 187: 552eC560.
Joost HG, Thorens B. The extended GLUT-family of sugar/polyol transport facilitators: nomenclature, sequence characteristics, and potential function of its novel members. Mol Membr Biol. 2001; 18: 247eC256.
Alpert E, Gruzman A, Totary H, Kaiser N, Reich R, Sasson S. A natural protective mechanism against hyperglycaemia in vascular endothelial and smooth-muscle cells: role of glucose and 12-hydroxyeicosatetraenoic acid. Biochem J. 2002; 362: 413eC422.
Alpert E, Altman H, Totary H, Gruzman A, Barnea D, Barash V, Sasson S. 4-Hydroxy tempol-induced impairment of mitochondrial function and augmentation of glucose transport in vascular endothelial and smooth muscle cells. Biochem Pharmacol. 2004; 67: 1985eC1995.
Cornford EM, Hyman S, Black KL, Cornford ME, Vinters HV, Pardridge WM. High expression of the Glut1 glucose transporter in human brain hemangioblastoma endothelium. J Neuropathol Exp Neurol. 1995; 54: 842eC851.
Qi C, Pekala PH. The influence of mRNA stability on glucose transporter (GLUT1) gene expression. Biochem Biophys Res Commun. 1999; 263: 265eC269.
Tsukamoto H, Boado RJ, Pardridge WM. Differential expression in glioblastoma multiforme and cerebral hemangioblastoma of cytoplasmic proteins that bind two different domains within the 3'-untranslated region of the human glucose transporter 1 (GLUT1) messenger RNA. J Clin Invest. 1996; 97: 2823eC2832.
Boado RJ, Pardridge WM. Ten nucleotide cis element in the 3'-untranslated region of the GLUT1 glucose transporter mRNA increases gene expression via mRNA stabilization. Brain Res Mol Brain Res. 1998; 59: 109eC113.
Dwyer KJ, Boado RJ, Pardridge WM. Cis-element/cytoplasmic protein interaction within the 3'-untranslated region of the GLUT1 glucose transporter mRNA. J Neurochem. 1996; 66: 449eC458.
Boado RJ, Pardridge WM. Glucose deprivation and hypoxia increase the expression of the GLUT1 glucose transporter via a specific mRNA cis-acting regulatory element. J Neurochem. 2002; 80: 552eC554.
Hamilton BJ, Nichols RC, Tsukamoto H, Boado RJ, Pardridge WM, Rigby WF. hnRNP A2 and hnRNP L bind the 3'UTR of glucose transporter 1 mRNA and exist as a complex in vivo. Biochem Biophys Res Commun. 1999; 261: 646eC651.
Jain RG, Andrews LG, McGowan KM, Pekala PH, Keene JD. Ectopic expression of Hel-N1, an RNA-binding protein, increases glucose transporter (GLUT1) expression in 3T3eCL1 adipocytes. Mol Cell Biol. 1997; 17: 954eC962.
Tsukamoto H, Hamada Y, Wu D, Boado RJ, Pardridge WM. GLUT1 glucose transporter: differential gene transcription and mRNA binding to cytosolic and polysome proteins in brain and peripheral tissues. Brain Res Mol Brain Res. 1998; 58: 170eC177.
Gao GP, Engdahl RK, Wilson JM. A cell line for high-yield production of E1-deleted adenovirus vectors without the emergence of replication-competent virus. Hum Gene Ther. 2000; 11: 213eC219.
Sela-Brown A, Silver J, Brewer G, Naveh-Many T. Identification of AUF1 as a parathyroid hormone mRNA 3'-untranslated region-binding protein that determines parathyroid hormone mRNA stability. J Biol Chem. 2000; 275: 7424eC7429.
Alpert E, Gruzman A, Riahi Y, Blejter R, Aharoni P, Weisinger G, Eckel J, Kaiser N, Sasson S. Delayed autoregulation of glucose transport in vascular endothelial cells. Diabetologia. 2005; 48: 752eC755.
Coppolino MG, Dedhar S. Calreticulin. Int J Biochem Cell Biol. 1998; 30: 553eC558.
Stephens JM, Carter BZ, Pekala PH, Malter JS. Tumor necrosis factor -induced glucose transporter (GLUT-1) mRNA stabilization in 3T3eCL1 preadipocytes. Regulation by the adenosine-uridine binding factor. J Biol Chem. 1992; 267: 8336eC8341.
Wertheimer E, Sasson S, Cerasi E, Ben-Neriah Y. The ubiquitous glucose transporter GLUT-1 belongs to the glucose-regulated protein family of stress-inducible proteins. Proc Natl Acad Sci U S A. 1991; 88: 2525eC2529.
Nickenig G, Michaelsen F, Muller C, Berger A, Vogel T, Sachinidis A, Vetter H, Bohm M. Destabilization of AT1 receptor mRNA by calreticulin. Circ Res. 2002; 90: 53eC58.
Timchenko LT, Iakova P, Welm AL, Cai ZJ, Timchenko NA. Calreticulin interacts with C/EBP and C/EBP mRNAs and represses translation of C/EBP proteins. Mol Cell Biol. 2002; 22: 7242eC7257.
Chen MH, Frey TK. Mutagenic analysis of the 3' cis-acting elements of the rubella virus genome. J Virol. 1999; 73: 3386eC3403.
Huang JT, Welch JS, Ricote M, Binder CJ, Willson TM, Kelly C, Witztum JL, Funk CD, Conrad D, Glass CK. Interleukin-4-dependent production of PPAR- ligands in macrophages by 12/15-lipoxygenase. Nature. 1999; 400: 378eC382.
Muga SJ, Thuillier P, Pavone A, Rundhaug JE, Boeglin WE, Jisaka M, Brash AR, Fischer SM. 8S-Lipoxygenase products activate peroxisome proliferator-activated receptor and induce differentiation in murine keratinocytes. Cell Growth Differ. 2000; 11: 447eC454.
Shappell SB, Gupta RA, Manning S, Whitehead R, Boeglin WE, Schneider C, Case T, Price J, Jack GS, Wheeler TM, Matusik RJ, Brash AR, Dubois RN. 15S-Hydroxyeicosatetraenoic acid activates peroxisome proliferator-activated receptor and inhibits proliferation in PC3 prostate carcinoma cells. Cancer Res. 2001; 61: 497eC503.
Johnson S, Michalak M, Opas M, Eggleton P. The ins and outs of calreticulin: from the ER lumen to the extracellular space. Trends Cell Biol. 2001; 11: 122eC129.
Burns K, Duggan B, Atkinson EA, Famulski KS, Nemer M, Bleackley RC, Michalak M. Modulation of gene expression by calreticulin binding to the glucocorticoid receptor. Nature. 1994; 367: 476eC480.
Michalak M, Corbett EF, Mesaeli N, Nakamura K, Opas M. Calreticulin: one protein, one gene, many functions. Biochem J. 1999; 344: 281eC292.
Hink U, Li H, Mollnau H, Oelze M, Matheis E, Hartmann M, Skatchkov M, Thaiss F, Stahl RA, Warnholtz A, Meinertz T, Griendling K, Harrison DG, Forstermann U, Munzel T. Mechanisms underlying endothelial dysfunction in diabetes mellitus. Circ Res. 2001; 88: e14eCe22.
Kiri A, Goldspink G. RNA-protein interactions of the 3' untranslated regions of myosin heavy chain transcripts. J Muscle Res Cell Motil. 2002; 23: 119eC129.(Hana Totary-Jain, Tally N)
The Nephrology (T.N.-M.) and Endocrinology & Metabolism (N.K.) Services, Department of Medicine, Hadassah-Hebrew University Medical Center, Jerusalem, Israel
Department of Clinical Biochemistry and Pathobiochemistry (J.E.), German Diabetes Center, De箂seldorf, Germany.
Abstract
Substrate autoregulation of glucose transporter-1 (GLUT-1) mRNA and protein expression provides vascular endothelial and smooth muscle cells a sensitive mechanism to adapt their rate of glucose transport in response to changing glycemic conditions. Hyperglycemia-induced downregulation of glucose transport is particularly important in protecting these cells against an excessive influx of glucose and consequently increased intracellular protein glycation and generation of free radicals; both are detrimental in the development of vascular disease in diabetes. We aimed to investigate the molecular mechanism of high glucoseeCinduced downregulation of GLUT-1 mRNA expression in primary bovine aortic vascular endothelial (VEC) and smooth muscle (VSMC) cell cultures. Using RNA mobility shift, UV cross-linking, and in vitro degradation assays, followed by mass-spectrometric analysis, we identified calreticulin as a specific destabilizing trans-acting factor that binds to a 10-nucleotide cis-acting element (CAE2181-2190) in the 3'-untranslated region of GLUT-1 mRNA. Pure calreticulin accelerated the rate of GLUT-1 mRNA-probe degradation in vitro, whereas overexpression of calreticulin in vascular cells decreased significantly the total cell content of GLUT-1 mRNA and protein. The expression of calreticulin was augmented in vascular cells exposed to high glucose in comparison with low-glucose conditions. Similarly, increased expression of calreticulin was observed in aortae of diabetic Psammomys obesus in comparison with normoglycemic controls. These data suggest that CAE2181-2190eCcalreticulin complex, which is formed in VSMC and VEC exposed to hyperglycemic conditions, renders GLUT-1 mRNA susceptible to degradation. This interaction underlies the process of downregulation of glucose transport in vascular cells under high-glucose conditions.
Key Words: calreticulin glucose transporter-1 hyperglycemia mRNA turnover vascular smooth muscle cells vascular endothelial cells
Introduction
Hyperglycemia is a major risk factor in the development of cardiovascular complications associated with the metabolic syndrome and diabetes.1,2 Chronic hyperglycemia alters the normal function of endothelial and smooth muscle cells in blood vessels, which undergo morphological and functional modifications, because of an excessive extra- and intracellular protein glycation and an uncontrolled production of free radicals. Collectively, these processes contribute to basement-membrane thickening, vascular occlusion, increased permeability, and initiation and progression of vascular disease and atherosclerosis.3eC6
The transport of glucose across the plasma membrane of cells is the rate-limiting step for subsequent glucose metabolism. A family of glucose transporters (GLUTs) mediates the entry of glucose into cells by facilitated diffusion.7 Vascular endothelial (VEC) and vascular smooth muscle (VSMC) cells express predominantly the ubiquitous GLUT-1 and to a very small extent the insulin-sensitive GLUT-4.8 We showed that these cells autoregulate their rate of glucose transport in response to changes in ambient glucose levels; specifically, they reduce their rate of hexose transport by &50% on exposure to high-glucose levels. Concomitantly, total cell GLUT-1 (but not GLUT-4) protein and mRNA content, as well as its plasma-membrane abundance, are reduced to the same extent.8 This downregulatory mechanism enables the cells to sustain normal glucose metabolism and ATP production similar to cells exposed to normal glucose levels.9
The expression of GLUT-1 is regulated posttranscriptionally in various cell types under diverse physiological and pathophysiological conditions, such as glucose deprivation, inhibition of oxidative phosphorylation, cytokine-, hormone-, and metabolite-dependent stimuli, and neoplastic transformation.10eC12 The 5' and 3' untranslated regions (UTR) of GLUT-1 mRNA are highly conserved among species. The 3' UTR contains a unique and conserved 10-nucleotide (nt) cis-acting stabilizing element (5'-CCAACCACTC; CAE2181 to 2190 in bovine GLUT-1).13,14 Recently, Boado et al showed that CAE2181-2190eCdependent interactions stabilized GLUT-1 mRNA in rat glial C6 tumor cells deprived of glucose and exposed to hypoxic conditions.15 These studies point to CAE2181-2190 as a glucose-sensitive element in GLUT-1 mRNA. Yet, the 3' UTR of GLUT-1 mRNA also contains AU-rich elements that may either stabilize or destabilize the entire molecule.16eC18
Hitherto, the molecular mechanism regulating the turnover rate of GLUT-1 mRNA in vascular cells under hyperglycemic conditions has not been elucidated. This study shows that calreticulin, whose expression is augmented in vascular cells under high-glucose conditions, destabilizes GLUT-1 mRNA by its specific interaction with CAE2181-2190.
Materials and Methods
Cell Cultures
Primary cultures of bovine aortic endothelial and smooth muscle cells were prepared and characterized as described previously.8 GH354 cells (courtesy of Dr J. M. Wilson, University of Pennsylvania, Philadelphia) were grown as described.19
Methods
The preparation of RNA transcripts, expression vectors, and adenoviral constructs, along with detailed descriptions of in vitro degradation assays, RNA mobility shift assays, UV cross-linking assays, real-time PCR, cell extract preparation, Western blot analyses, and immunoprecipitation are given in the online data supplement available at http://circres.ahajournals.org.
Statistical Analysis
Statistical analyses were performed using the nonparametric ManneCWhitney test.
Results
High GlucoseeC and CAE2181-2190eCDependent Degradation of GLUT-1 mRNA Probes
The role of CAE2181-2190 in regulating bovine GLUT-1 mRNA stability in vascular cells was studied in an in vitro degradation assay. [32P]-Labeled 203-nt (5' 2098 to 2300) and 203-ntCAE2181-2190 probes of the 3 UTR of bovine GLUT-1 mRNA were incubated at 37°C with S100 cytosolic proteins of VSMC that had been maintained at 2 or 25 mmol/L glucose. Figure 1A shows a rapid degradation of the intact probe only in the presence of the high-glucose extract. However, deletion of CAE2181-2190 from the probe rendered it stable to degradation under similar conditions.
The role of CAE2181-2190 in regulating the expression of GLUT-1 mRNA in intact cells was also investigated. The 203- or 203CAE2181-2190-nt sequences of GLUT-1 mRNA were inserted into the luciferase expression vector pGL2 at the unique pflM1 restriction site located in the luciferase 3'-UTR region. VSMC were transfected with a green fluorescent protein (GFP)-expression plasmid (pEGFP-N1) without or with the control or modified pGL2 vectors and incubated at 25 mmol/L glucose for 48 hours. GFP-positive cells were collected by fluorescence-activated cell sorting (FACS) and taken for real-time PCR analyses of GLUT-1 and luciferase mRNA. Figure 1B shows an increased abundance of GLUT-1 mRNA in cells transfected with the 203-nteCinserted pGL2 vector in comparison with cells transfected with the unmodified vector; concomitantly, luciferase mRNA level seemed decreased. These data suggest that CAE2181-2190 in luciferase mRNA competed with the same element in the 3' UTR of the endogenous GLUT-1 mRNA for binding of destabilizing factors, thus rendering the latter transcript less susceptible to degradation. However, the expression of luciferase mRNA increased and GLUT-1 mRNA remained unaltered in cells transfected with the 203CAE2181-2190eCinserted pGL2 vector. It appears that in the absence of competing CAE2181-2190 sequences in luciferase mRNA molecules, the same element in endogenous GLUT-1 mRNA interacted with destabilizing factors, as it does in control nontransfected cells. Yet, the increased expression of luciferase mRNA may reflect interactions of cellular stabilizing factors with other elements within the modified luciferase mRNA that were effective in the absence of CAE2181-2190eCmediated destabilizing interactions.
CAE2181-2190eCCytosolic ProteineCBinding Interactions
RNA electrophoretic mobility-shift assay (REMSA) of the [32P]-labeled 203-nt probe in the presence of increasing protein concentrations of a high-glucose VSMC protein extract was performed. Figure 2A shows 2 complexes, the binding intensity of which was augmented with increasing protein concentrations (upper arrows), reaching maximum with 3 e of extract, and a lower diffused band. The specificity of these binding interactions was investigated by adding 20- to 200-fold molar excess of the unlabeled 203-nt transcript or an unrelated transcript of 3' UTR of parathyroid hormone mRNA to the reaction mixture (Figure 2B).20 As expected, only the former competed effectively with the labeled probe for binding of cytosolic proteins.
The capacity of low- and high-glucose cell extracts to interact with CAE2181-2190eCcontaining probes are shown in Figure 3A. The extracts were prepared from VSMC and VEC that had been maintained at 2 or 25 mmol/L glucose for 14 or 20 hours, respectively. These were the minimal incubation periods at high glucose required to produce maximal binding interactions. The slower response of VEC agrees with our previous report on a delayed downregulatory response of these cells to high glucose.21 Figure 3A shows, in the REMSA of the [32P]-labeled 57-nt probe, 3 distinct complexes that were not apparent with the longer 203-nt probe. Similar REMSAs with CAE2181-2190eCdeleted probes were performed (Figure 3B): the intensities of the 2 upper complexes formed with the [32P]-labeled 57CAE2181-2190 probe were significantly lower in comparison with those formed with the intact 57-nt probe. Yet, the binding interactions of the [32P]-labeled 203CAE2181-2190 probe were similar to those detected with the intact probe, possibly attributable to additional binding sequences for other proteins in the longer 203-nt probe, which masked the effect of CAE2181-2190 deletion. Therefore, we used the 57-nt probe in subsequent experiments.
The molecular mass of the [32P]-labeled 57-nt probe VSMC protein complexes was determined by UV cross-linking analysis. Four complexes corresponding to 112-, 62-, 58-, and 43-kDa, the binding intensities of which were greater with high-glucose than low-glucose VSMC protein extracts, were depicted (Figure 4A, lanes 2 and 3). Unlabeled 57-nt transcript, added at a 2-fold molar excess before UV cross-linking, reduced significantly the formation of the 62- and the 58-kDa complexes and, to a lesser extent, the 112- and 43-kDa complexes (lane 4).
To elucidate which of these complexes required CAE2181-2190 for binding, 20- to 300-fold molar excesses of complementary oligonucleotides (ONDs) were hybridized to the [32P]-labeled 57-nt probe before UV cross-linking with high-glucose VSMC extract (Figure 4B). OND2 was complementary to CAE2181-2190, whereas OND1 and -3 overlapped distal up- or downstream sequences within the probe (Figure 4B). The 58- and 43-kDa complexes were noneCCAE2181-2190 dependent, because all ONDs weakly interfered with their formation. Because OND2 abolished effectively the formation of the 112- and 62-kDa complexes, both could be considered candidates for an interaction with CAE2181-2190. However, the 112-kDa complex was excluded, because its formation was also abolished by OND1. Hence, the 62-kDa complex was characterized as a CAE2181-2190eCdependent complex.
Calreticulin Is a CAE2181-2190eCBinding Protein That Destabilizes GLUT-1 mRNA
High-glucose VSMC protein extract was affinity chromatographed on a Heparin-Fractogel-EMD column with increasing concentrations of KCl. Figure 5A shows REMSA of each KCl-eluted fraction incubated with the [32P]-labeled 57-nt probe. Of the various complexes formed with the original S100 extract (left arrows), 2 were significantly enriched in the 125 and 200 mmol/L KCl eluates (right arrows). Their respective molecular weights were resolved as 43 and 62 kDa in a UV cross-linking assay (Figure 5B).
The 125 and 200 mmol/L KCl eluates were concentrated 10-fold on an Amicon Ultra-4 centrifugal filter device (30-kDa cutoff) and run on SDS-PAGE. The 43- and 62-kDa bands were excised, processed, and analyzed by mass spectrometry. The results showed sequence homologies to rat aldolase A and bovine calreticulin, respectively (online Table I). The calculated molecular mass of bovine calreticulin is 46.6 kDa, but it resolves on gels as an &60-kDa protein, because of a large number of negatively charged carboxylic groups (pI=4.7).22 Because the data in Figure 4B showed that CAE2181-2190 was essential for the formation of the 62-kDa, but not the 43-kDa, complex, we investigated the role of calreticulin in the regulation of GLUT-1 mRNA stability.
The following experiments were conducted to ascertain whether calreticulin was the protein in the 62-kDa complex. Aliquots of high-glucose VSMC protein extract and its 200 mmol/L KCl eluate were incubated with the [32P]-labeled 57-nt probe, UV-cross-linked, separated by SDS-PAGE, and transferred to a nitrocellulose membrane. The membrane was first used for Western blotting of calreticulin and then taken for phosphorimaging to visualize [32P]-labeled RNAeCprotein complexes. Figure 6A depicts 2 colocalized 62-kDa (calreticulin) bands, detected independently by each method of the same gel. In addition, Figure 6B depicts a specific binding interaction of the same probe with pure bovine calreticulin (lane 2), which disappeared in the presence of a 10-fold molar excess of unlabeled 57-nt sequence (lane 3) or following prehybridization of the probe with a 300-fold molar excess of OND2 (lane 4).
Figure 6C compares UV cross-linking profiles of the labeled 57 nt with pure bovine calreticulin and with VSMC cytosolic proteins. Whereas the typical multicomplex profile was apparent with the latter (lane 2), calreticulin formed a single 62-kDa complex (lane 3), which was blocked by prehybridization of the probe with OND2 (lane 4). Because the anti-calreticulin antibody used interacted well with only denatured calreticulin, our attempts to supershift the 62-kDa complex in native REMSAs were unsuccessful. Instead, an aliquot of VSMC protein extract was UV cross-linked with the [32P]-labeled 57-nt probe, immunoprecipitated with anti-calreticulin antibody under denaturing conditions, and electrophoresed. A single band corresponding to the same size as of the complex with pure calreticulin was observed (compare lane 5 with 3). Prehybridization of the probe with OND2 before UV illumination completely eliminated the formation of this complex (not shown). No bands were detected following immunoprecipitation with rabbit nonimmune serum (not shown).
High-Glucose Augments Calreticulin Expression in Vascular Cells In Vitro and In Vivo
VSMC and VEC were incubated in high glucose for 24 or 48 hours, respectively, to induce maximal downregulation of glucose transport and GLUT-1 protein level in VSMC and VEC, respectively, as was shown recently.21 Figure 7A and 7B show a 1.8- to 2.0-fold increase in calreticulin expression and its plasma membrane abundance and reduced GLUT-1 levels in both types of cells exposed to high glucose in comparison with low glucose.
The physiological relevance of these findings in cell cultures was confirmed by determining calreticulin content in aortic segments isolated from normoglycemic and hyperglycemic Psammomys obesus. The respective blood glucose levels of the gerbils were 109±20 (n=3) and 263±29 (n=5) mg/dL. Figure 7C shows a significant 1.99±0.28-fold increased calreticulin content in aortae of diabetic animals in comparison with their normoglycemic controls. However, because of the high level of GLUT-1 content in contaminating erythrocytes trapped within the aortic segments, the analysis of aortic GLUT-1 was unreliable. Yet, calreticulin expression in the erythrocytes was negligible (online Figure I).
Calreticulin Reduces GLUT-1 Expression in VEC and VSMC
Effects of overexpression of calreticulin on GLUT-1 protein level in VSMC and VEC were then determined. VSMC were transfected with GFP-expression plasmid (pEGFP-N1) without or with the calreticulin-expression plasmid pMCC-ZP-CTR, followed by incubations at 2 or 25 mmol/L glucose for 48 hours. GFP-positive cells were then collected by FACS, lysed, and taken for Western blot analyses. Similar to its effect in intact cells (Figure 7A), high-glucose incubation increased calreticulin and reduced GLUT-1 protein levels in control (GFP-expressing) cells in comparison with similar cells maintained at low glucose (Figure 8A, left panel, lanes 1 and 2). Moreover, overexpression of calreticulin in cells exposed to 2 mmol/L glucose mimicked the downregulatory effect of the 25 mmol/L glucose incubation on GLUT-1 content in control (GFP-expressing) cells (compare lane 3 with 2). The level of GLUT-1 protein was similarly reduced in cells overexpressing calreticulin and exposed either to low- or high-glucose conditions (lanes 3 and 4).
VEC were infected with adenoviral vectors expressing GFP or both GFP and calreticulin and incubated at 2 or 25 mmol/L glucose for 48 hours. Figure 8A (right panel, lanes 1 and 2) shows an &50% reduction in GLUT-1 content in control, GFP- expressing cells that were exposed to high glucose in comparison with the low-glucose incubation. Overexpression of calreticulin in cells exposed to 2 mmol/L glucose mimicked the effect of high glucose in control cells (compare lane 3 with 2). Overexpressed calreticulin reduced GLUT-1 levels in VEC exposed either to 2 or 25 mmol/L glucose (lanes 3 and 4).
GLUT-1 and calreticulin mRNA levels in similarly infected VEC were determined by real-time PCR. The level of calreticulin mRNA in cells infected with the calreticulin-expressing virus was 5.4±1.5- and 6.9±2.3-fold higher (n=6) under 2 and 25 mmol/L glucose incubation, respectively, in comparison with cells infected with control GFP-expressing virus. Figure 8B shows 40±17% reduction in GLUT-1 mRNA content in cells expressing GFP only and incubated at 25 mmol/L glucose in comparison with the low-glucose incubation. The level of GLUT-1 mRNA in cells infected with the calreticulin-expressing virus was reduced by 44±16% and 48±16% under 2 or 25 mmol/L glucose, respectively.
Finally, we repeated the in vitro degradation assay shown in Figure 1 and added pure calreticulin to low-glucose VSMC protein extract, which did not degrade the [32P]-labeled 203-nt probe (Figures 1 and 8C). Figure 8C and 8D shows that calreticulin promoted a rapid degradation of the intact probe but not of the [32P]-labeled 203CAE2181-2190 probe.
Discussion
The mechanism of high glucoseeCinduced downregulation of GLUT-1 mRNA expression in VEC and VSMC were investigated. As already reported, high-glucose incubation reduced both GLUT-1 mRNA and protein expression in VSMC and VEC.8 In vitro degradation assays and cell transfection with CAE2181-2190eCcontaining pGL2 plasmid indicate that CAE2181-2190 in the 3' UTR of GLUT-1 mRNA acts as a destabilizing glucose-sensitive element. REMSAs, UV cross-linking assays, and affinity chromatography purification of cell extracts, followed by mass-spectrometric analysis, identified calreticulin as a binding protein in the 62-kDa complex with this CAE. Overexpression of calreticulin in the cells significantly reduced GLUT-1 mRNA and protein content, whereas pure calreticulin promoted rapid degradation of CAE2181-2190eCcontaining transcripts. Furthermore, high-glucose incubations increased calreticulin expression in both types of cells. Similarly, increased calreticulin content was found in aortae of hyperglycemic P. obesus, in comparison with their normoglycemic controls. These findings support a model in which calreticulin interacts with CAE2181-2190 in the 3' UTR of GLUT-1 mRNA and destabilizes the entire transcript, leading to reduced GLUT-1 protein expression and plasma membrane localization and, subsequently, to downregulation of glucose transport. This adaptive mechanism protects vascular cells against damaging effects of an uncontrolled influx of glucose in face of hyperglycemia.
Early studies identified CAE2181-2190 in GLUT-1 mRNA as a glucose-sensitive element, which formed an 80-kDa mRNA stabilizing complex in brain microcapillary endothelial cells and rat C6 glioma cells.13,15 However, no such VSMC- or VEC cytosolic extracteCCAE2181-2190 complex was detected in the present study. Because GLUT-1 mRNA expression was reported to be regulated in an organ-specific manner by diverse proteins,18 GLUT-1 mRNAeCprotein complexes in vascular cells may vary from those formed in other cell types and under various experimental or pathophysiological conditions. Moreover, other CAEs in the 3' UTR of GLUT-1 mRNA can also affect GLUT-1 mRNA stability in concert with CAE2181-2190.16,17,23 Indeed, our findings on an increased expression of luciferase mRNA on transfection of VSMC with the 203CAE2181-2190eCinserted pGL2 vector (Figure 1B) indicate a role for additional stabilizing interactions within the deleted 203-nt sequence that became effective in the absence of CAE2181-2190eCmediated destabilizing effects.
Overexpression of calreticulin in vascular cells exposed to low glucose mimicked the effect of high-glucose incubation on GLUT-1 mRNA and protein levels (Figure 8A and 8B). Yet, the combined effect of high-glucose incubation and overexpression of calreticulin was not additive and was, in fact, comparable to that observed in control cells exposed to high glucose. Thus, it seems that the increased calreticulin content in cells exposed to high glucose was sufficient to induce maximal destabilization of GLUT-1 mRNA, whereas overexpressed calreticulin under these conditions was inconsequential. We showed that transfer of vascular cells from a high-glucose to lower glucose levels induced upregulation of glucose transport and GLUT-1 mRNA expression and that glucose-deprived cells increase the expression of GLUT-1 to ameliorate a metabolically challenging and stressful condition.8,21,24 We also showed that GLUT-1 was a stress-inducible protein belonging to the family of glucose-regulated proteins.24 Thus, the combination of overexpression of calreticulin and high-glucose incubation of the vascular cells could markedly reduce GLUT-1 content and cause glucose deprivationeClike effects, followed by a stress reaction aimed to increase GLUT-1 gene transcription and/or expression of GLUT-1 mRNAeCstabilizing factors. Such mechanisms could also explain the lack of additive effects of calreticulin overexpression and high-glucose incubations on GLUT-1 mRNA and protein in the cells.
Other studies also found specific RNAeCcalreticulin interactions. Destabilization of angiotensin-1 receptor mRNA in VSMC induced calreticulin binding to an AU-rich element in the 3' UTR.25 Calreticulin inhibited the translation of C/EBP transcription factors by its interaction with GCU repeats of C/EBP/ mRNAs.26 In addition, rubella virus replication and/or localization in infected cells was controlled by calreticulin interactions with in the 3' terminal of the viral RNA.27 These studies and the present results indicate the lack of a consensus RNA sequence for calreticulin binding. Indeed, computer-based predictions (http://www.bioinfo.rpi.edu/zukerm/rna) of the 2D structure of the CAE2181-2190 region in the 3'-UTR GLUT-1 mRNA and of the abovementioned RNA regions point to a stem-loop structure, rather than a consensus sequence, as a potential target for calreticulin binding.
The mechanism of high glucoseeCdependent augmented expression of calreticulin in vascular cells has not yet been investigated. We have found that high-glucose levels increase the expression of the enzyme 12-lipoxygenase and the production of its metabolite 12-hydroxyeicosatetraenoic acid (12-HETE) in vascular cells. This eicosanoid mediates high glucoseeCinduced reduction in GLUT-1 mRNA and protein levels and downregulation of hexose transport.8,21 Others have shown that peroxisome proliferatoreCactivated receptors interact with various HETEs to promote gene transcription.28eC30 These studies and present results allude to a peroxisome proliferatoreCactivated receptor interaction with 12-HETE in the mechanism mediating augmented calreticulin expression in vascular cells.
Since the discovery of calreticulin more than 30 years ago as a calcium-binding protein in the sarcoplasmic reticulum, it has become clear that this is a multifunctional protein, whose subcellular localization is not restricted to the sarcoplasmic reticulum.31 Among other functions, calreticulin chaperons proteins in the endoplasmic reticulum, activates integrins, regulates cell adhesion, induces NO synthesis, and regulates steroid-sensitive gene expression.22,32,33 The present results on high glucoseeCdependent increased expression of calreticulin in VEC and VSMC and aortae of diabetic P. obesus (Figure 7A through 7C) suggest a role for calreticulin in modifying cell-matrix interactions, chaperoning and targeting of over-glycosylated, and regulating NO production in diabetic blood vessels.34
Our results also indicate that the gluconeogenic enzyme aldolase A might interact with the 57-nt probe of the 3'-UTR region of GLUT-1 to form a non-CAE2181-2190eCdependent 43-kDa complex (Figure 4 and 5). This enzyme was found to bind specifically to the 3' UTR of myosin heavy chain mRNA and possibly target or anchor it to subcellular compartments of myosin synthesis by association with the cytoskeletal framework.35 It remains to be investigated whether aldolase A shares a similar function in subcellular targeting of GLUT-1 mRNA in vascular cells.
In Summary, this study shows a central role for calreticulin in the mechanism of downregulation of GLUT-1 mRNA in vascular cells under high-glucose conditions, which ultimately protect the cells against excessive glucose influx and its deteriorating effects.
Acknowledgments
This work was supported by grants from the German-Israel Foundation for Scientific Research and Development (I-750-165.2/2002), the Russell Berrie Foundation and Diabetes Care in Israel (D-CURE), the Israel Ministry of Health, the Yedidut Foundation Mexico and the David R. Bloom Center for Pharmacy at the Hebrew University. H.T.-J. received doctoral fellowships from the Hebrew University Center for Diabetes Research and The Israel Ministry of Science. S.S. is a member of the David R. Bloom Center for Pharmacy at the Hebrew University. We thank Dr H. Giladi for assistance in constructing adenoviral vectors.
References
Tenenbaum A, Motro M, Schwammenthal E, Fisman EZ. Macrovascular complications of metabolic syndrome: an early intervention is imperative. Int J Cardiol. 2004; 97: 167eC172.
Giles TD, Sander GE. Diabetes mellitus and heart failure: basic mechanisms, clinical features, and therapeutic considerations. Cardiol Clin. 2004; 22: 553eC568.
Brownlee M. Negative consequences of glycation. Metabolism. 2000; 49 (suppl 1): 9eC13.
Brownlee M. Biochemistry and molecular cell biology of diabetic complications. Nature. 2001; 414: 813eC820.
Lorenzi M, Cagliero E, Toledo S. Glucose toxicity for human endothelial cells in culture. Delayed replication, disturbed cell cycle, and accelerated death. Diabetes. 1985; 34: 621eC627.
Natarajan R, Gonzales N, Xu L, Nadler JL. Vascular smooth muscle cells exhibit increased growth in response to elevated glucose. Biochem Biophys Res Commun. 1992; 187: 552eC560.
Joost HG, Thorens B. The extended GLUT-family of sugar/polyol transport facilitators: nomenclature, sequence characteristics, and potential function of its novel members. Mol Membr Biol. 2001; 18: 247eC256.
Alpert E, Gruzman A, Totary H, Kaiser N, Reich R, Sasson S. A natural protective mechanism against hyperglycaemia in vascular endothelial and smooth-muscle cells: role of glucose and 12-hydroxyeicosatetraenoic acid. Biochem J. 2002; 362: 413eC422.
Alpert E, Altman H, Totary H, Gruzman A, Barnea D, Barash V, Sasson S. 4-Hydroxy tempol-induced impairment of mitochondrial function and augmentation of glucose transport in vascular endothelial and smooth muscle cells. Biochem Pharmacol. 2004; 67: 1985eC1995.
Cornford EM, Hyman S, Black KL, Cornford ME, Vinters HV, Pardridge WM. High expression of the Glut1 glucose transporter in human brain hemangioblastoma endothelium. J Neuropathol Exp Neurol. 1995; 54: 842eC851.
Qi C, Pekala PH. The influence of mRNA stability on glucose transporter (GLUT1) gene expression. Biochem Biophys Res Commun. 1999; 263: 265eC269.
Tsukamoto H, Boado RJ, Pardridge WM. Differential expression in glioblastoma multiforme and cerebral hemangioblastoma of cytoplasmic proteins that bind two different domains within the 3'-untranslated region of the human glucose transporter 1 (GLUT1) messenger RNA. J Clin Invest. 1996; 97: 2823eC2832.
Boado RJ, Pardridge WM. Ten nucleotide cis element in the 3'-untranslated region of the GLUT1 glucose transporter mRNA increases gene expression via mRNA stabilization. Brain Res Mol Brain Res. 1998; 59: 109eC113.
Dwyer KJ, Boado RJ, Pardridge WM. Cis-element/cytoplasmic protein interaction within the 3'-untranslated region of the GLUT1 glucose transporter mRNA. J Neurochem. 1996; 66: 449eC458.
Boado RJ, Pardridge WM. Glucose deprivation and hypoxia increase the expression of the GLUT1 glucose transporter via a specific mRNA cis-acting regulatory element. J Neurochem. 2002; 80: 552eC554.
Hamilton BJ, Nichols RC, Tsukamoto H, Boado RJ, Pardridge WM, Rigby WF. hnRNP A2 and hnRNP L bind the 3'UTR of glucose transporter 1 mRNA and exist as a complex in vivo. Biochem Biophys Res Commun. 1999; 261: 646eC651.
Jain RG, Andrews LG, McGowan KM, Pekala PH, Keene JD. Ectopic expression of Hel-N1, an RNA-binding protein, increases glucose transporter (GLUT1) expression in 3T3eCL1 adipocytes. Mol Cell Biol. 1997; 17: 954eC962.
Tsukamoto H, Hamada Y, Wu D, Boado RJ, Pardridge WM. GLUT1 glucose transporter: differential gene transcription and mRNA binding to cytosolic and polysome proteins in brain and peripheral tissues. Brain Res Mol Brain Res. 1998; 58: 170eC177.
Gao GP, Engdahl RK, Wilson JM. A cell line for high-yield production of E1-deleted adenovirus vectors without the emergence of replication-competent virus. Hum Gene Ther. 2000; 11: 213eC219.
Sela-Brown A, Silver J, Brewer G, Naveh-Many T. Identification of AUF1 as a parathyroid hormone mRNA 3'-untranslated region-binding protein that determines parathyroid hormone mRNA stability. J Biol Chem. 2000; 275: 7424eC7429.
Alpert E, Gruzman A, Riahi Y, Blejter R, Aharoni P, Weisinger G, Eckel J, Kaiser N, Sasson S. Delayed autoregulation of glucose transport in vascular endothelial cells. Diabetologia. 2005; 48: 752eC755.
Coppolino MG, Dedhar S. Calreticulin. Int J Biochem Cell Biol. 1998; 30: 553eC558.
Stephens JM, Carter BZ, Pekala PH, Malter JS. Tumor necrosis factor -induced glucose transporter (GLUT-1) mRNA stabilization in 3T3eCL1 preadipocytes. Regulation by the adenosine-uridine binding factor. J Biol Chem. 1992; 267: 8336eC8341.
Wertheimer E, Sasson S, Cerasi E, Ben-Neriah Y. The ubiquitous glucose transporter GLUT-1 belongs to the glucose-regulated protein family of stress-inducible proteins. Proc Natl Acad Sci U S A. 1991; 88: 2525eC2529.
Nickenig G, Michaelsen F, Muller C, Berger A, Vogel T, Sachinidis A, Vetter H, Bohm M. Destabilization of AT1 receptor mRNA by calreticulin. Circ Res. 2002; 90: 53eC58.
Timchenko LT, Iakova P, Welm AL, Cai ZJ, Timchenko NA. Calreticulin interacts with C/EBP and C/EBP mRNAs and represses translation of C/EBP proteins. Mol Cell Biol. 2002; 22: 7242eC7257.
Chen MH, Frey TK. Mutagenic analysis of the 3' cis-acting elements of the rubella virus genome. J Virol. 1999; 73: 3386eC3403.
Huang JT, Welch JS, Ricote M, Binder CJ, Willson TM, Kelly C, Witztum JL, Funk CD, Conrad D, Glass CK. Interleukin-4-dependent production of PPAR- ligands in macrophages by 12/15-lipoxygenase. Nature. 1999; 400: 378eC382.
Muga SJ, Thuillier P, Pavone A, Rundhaug JE, Boeglin WE, Jisaka M, Brash AR, Fischer SM. 8S-Lipoxygenase products activate peroxisome proliferator-activated receptor and induce differentiation in murine keratinocytes. Cell Growth Differ. 2000; 11: 447eC454.
Shappell SB, Gupta RA, Manning S, Whitehead R, Boeglin WE, Schneider C, Case T, Price J, Jack GS, Wheeler TM, Matusik RJ, Brash AR, Dubois RN. 15S-Hydroxyeicosatetraenoic acid activates peroxisome proliferator-activated receptor and inhibits proliferation in PC3 prostate carcinoma cells. Cancer Res. 2001; 61: 497eC503.
Johnson S, Michalak M, Opas M, Eggleton P. The ins and outs of calreticulin: from the ER lumen to the extracellular space. Trends Cell Biol. 2001; 11: 122eC129.
Burns K, Duggan B, Atkinson EA, Famulski KS, Nemer M, Bleackley RC, Michalak M. Modulation of gene expression by calreticulin binding to the glucocorticoid receptor. Nature. 1994; 367: 476eC480.
Michalak M, Corbett EF, Mesaeli N, Nakamura K, Opas M. Calreticulin: one protein, one gene, many functions. Biochem J. 1999; 344: 281eC292.
Hink U, Li H, Mollnau H, Oelze M, Matheis E, Hartmann M, Skatchkov M, Thaiss F, Stahl RA, Warnholtz A, Meinertz T, Griendling K, Harrison DG, Forstermann U, Munzel T. Mechanisms underlying endothelial dysfunction in diabetes mellitus. Circ Res. 2001; 88: e14eCe22.
Kiri A, Goldspink G. RNA-protein interactions of the 3' untranslated regions of myosin heavy chain transcripts. J Muscle Res Cell Motil. 2002; 23: 119eC129.(Hana Totary-Jain, Tally N)