Phosphorylation-Dependent Nucleocytoplasmic Shuttling of Pancreatic Duodenal Homeobox-1
Pancreatic duodenal homeobox-1 (PDX-1) is a homeodomain protein that plays an important role in the development of the pancreas and in maintaining the identity and function of the islets of Langerhans. It also regulates the expression of the insulin gene in response to changes in glucose and insulin concentrations. Glucose and insulin regulate PDX-1 by way of a signaling pathway involving phosphatidylinositol 3-kinase (PI 3-kinase) and SAPK2/p38. Activation of this pathway leads to phosphorylation of PDX-1 and its movement into the nucleus. To investigate the intracellular trafficking of PDX-1, immunocytochemistry was used to localize PDX-1 in the human β-cell line NesPDX-1, in which PDX-1 is overexpressed, and in MIN6 β-cells. In low-glucose conditions, PDX-1 localized predominantly to the nuclear periphery, with some staining in the cytoplasm. After stimulation with glucose, PDX-1 was present in the nucleoplasm. The translocation of PDX-1 to the nucleoplasm was complete within 15 min and occurred in 5−10 mmol/l glucose. Insulin and sodium arsenite, an activator of the stress-activated pathway, also stimulated PDX-1 movement from the nuclear periphery to the nucleoplasm. When cells were transferred between high glucose− and low glucose−containing medium, PDX-1 rapidly shuttled between the nuclear periphery and the nucleoplasm. Glucose- and insulin-stimulated translocation of PDX-1 to the nucleoplasm was inhibited by wortmannin and SB 203580, indicating that a pathway involving PI 3-kinase and SAPK2/p38 was involved; translocation was unaffected by PD 098959 and rapamycin, suggesting that neither mitogen-activated protein kinase nor p70s6k were involved. Arsenite-stimulated import of PDX-1 into the nucleus was inhibited by SB 203580 but not by wortmannin. Export from the nucleoplasm to the nuclear periphery was inhibited by calyculin A and okadaic acid, suggesting that dephosphorylation of PDX-1 was involved. These results demonstrated that PDX-1 shuttles between the nuclear periphery and nucleoplasm in response to changes in glucose and insulin concentrations and that these events are dependent on PI 3-kinase, SAPK2/p38, and a nuclear phosphatase(s).
Pancreatic duodenal homeobox-1 (PDX-1) is a homeodomain protein present in β-cells and D-cells of islets of Langerhans. It is also expressed in neuroendocrine cells of the gut. PDX-1 plays an important role in the development of the pancreas (1). Thus PDX-1−deficient mice are born without a pancreas (2,3), whereas the heterozygote PDX-1 (+/−) mouse develops a pancreas but becomes mildly diabetic (4). Pancreatic agenesis has also been observed in a human with PDX-1 deficiency resulting from a homozygous mutation in the PDX-1 gene (5). Family members of that individual who were heterozygote carriers of this mutation developed maturity onset diabetes of the young, an autosomal form of type 2 diabetes (6). Other mutations in the PDX-1 gene have been linked to type 2 diabetes (7,8).
In addition to its role in the development of the pancreas, PDX-1 also binds to sequences within and regulates the promoter activity of a number of islet genes, including insulin (9), GLUT-2 (10), glucokinase (11), islet amyloid polypeptide (12,13,14,15), and somatostatin (16). Its exact role in regulating insulin gene transcription is not clear, but it is known not to be essential to the process, as high levels of insulin mRNA have been observed under conditions in which PDX-1 activity is reduced (17) or virtually absent (18,19). There is strong evidence that PDX-1 plays a role in activating the insulin promoter (20,21,22) and increasing insulin mRNA levels (23) in response to glucose. Glucose’s effects on PDX-1 are mediated by way of a cell-signaling pathway that involves phosphatidylinositol 3-kinase (PI 3-kinase) and SAPK2/p38 (24), although atypical protein kinase C−ζ has also been implicated in the mechanisms by which glucose stimulates PDX-1 (25). Insulin can also activate the insulin promoter and PDX-1 DNA-binding activity (26,27).
It has been shown previously that glucose stimulates the movement of PDX-1 from the cytoplasm to the nucleus (28,29) and that in low-glucose conditions, PDX-1 might be preferentially associated with the nuclear periphery (15,28). The aim of the present study was to investigate nuclear translocation of PDX-1 in further detail by using immunocytochemistry. The experiments were performed in the human β-like cell line NES2Y stably transfected with human PDX-1 to generate the NesPDX-1 cell line (23) and in MIN6 β-cells. The results demonstrated that in unstimulated cells, PDX-1 preferentially associated with the nuclear periphery, with some staining in the cytoplasm, and that glucose and insulin affected the shuttling of PDX-1 between the nuclear periphery and the nucleoplasm. Movement from the nuclear periphery to the nucleoplasm was inhibited by wortmannin and SB203580, thus supporting the view that PI 3-kinase and SAPK2/p38 were involved. Export from the nucleoplasm to the nuclear periphery was blocked by the phosphatase inhibitors calyculin A and okadaic acid.
RESEARCH DESIGN AND METHODS
Chemicals and reagents.
Inhibitors SB 203580, rapamycin, wortmannin, and calyculin A were purchased from Sigma-Aldrich (Poole, U.K.); PD 098059 and okadaic acid, from Calbiochem (Nottingham, Nottinghamshire, U.K.); and sodium arsenite, from Fisons (Loughborough, Nottinghamshire, U.K.). Anti-PDX-1 antibody was a kind gift from Dr. C. V. Wright (Vanderbilt University, Nashville, TN). Anti−upstream stimulatory factor (USF) antibody was purchased from Santa Cruz Biotechnology (Santa Cruz, CA), and anti−lamin-B1 antibody was from Chemicon International (Harrow, Essex, U.K.). Fluorescein isothiocyanate (FITC)−labeled anti-rabbit IgG antibody was obtained from Diagnostics Scotland, Law Hospital, (Carluke, Lanarkshire, U.K.) and rhodamine-labeled anti-mouse IgG was obtained from Jackson Immunoresearch (Westgrove, PA). All other chemicals were purchased from Sigma-Aldrich.
NesPDX-1 cells that overexpress PDX-1 were generated by stably transfecting NES2Y cells with a cDNA encoding PDX-1. NES2Y cells were derived from islets of Langerhans isolated from the pancreas of a patient with persistent hyperinsulinemic hypoglycemia of infancy, as previously described (18). NES2Y cells express PDX-1 at very low levels that can be detected by reverse transcriptase−polymerase chain reaction, but not by electrophoretic mobility shift assay, Western blot, or immunocytochemistry. NesPDX-1 cells were maintained in RPMI 1640 medium (Life Technologies, Paisley, Strathclyde, U.K.) containing 11 mmol/l glucose and 2 mmol/l glutamine, supplemented with 10% (vol/vol) fetal calf serum (Life Technologies) and 800 μg/ml G418. MIN6 (30) is a pancreatic β-cell line derived from transgenic mice expressing the simian virus 40 large T-antigen under the control of the rat insulin gene promoter. MIN6 cells were grown in Dulbecco’s modified Eagle’s medium containing 5 mmol/l glucose supplemented with 10% (vol/vol) fetal bovine serum in a humidified atmosphere containing 95% air/5% CO2.
Cells were grown in 35-mm plates (Nunclon, Naperville, IL) to 50% confluence. After an overnight incubation in medium containing 10% fetal calf serum and 0.5 mmol/l glucose, the cells were then incubated under conditions described in figure legends. Cells were then washed four times with phosphate-buffered saline (PBS) and fixed in ice-cold methanol for 10 min at 4°C. The cells were again washed four times in PBS before being blocked for 15 min at room temperature in blocking buffer containing 6.7% (vol/vol) glycerol, 0.2% (vol/vol) Tween 20, and 2% (wt/vol) bovine serum albumin (BSA) in PBS. Primary antibodies were added to the cells in 1 ml blocking buffer, and cells were incubated overnight at 4°C. After four washes in a solution containing 6.7% glycerol, 0.4% Tween 20, and 2% BSA in PBS, cells were incubated in a 1:400 dilution, in blocking buffer, of an FITC-conjugated anti-rabbit secondary antibody for 1 h in the dark. The cells were then washed in the same wash buffer for 1–2 h in the dark with gentle agitation, before they were mounted on slides in Vectashield (Vector Laboratories, Burlingame, CA) mounting medium and coverslips were applied. The plates were then studied by confocal microscopy with a 40× water immersion objective. Cells were randomly selected and the confocal image was adjusted to take a section through the entire body of the cell, using a BioRad microradiance confocal scanning system. Images selected were representative of 75–100% of cells present on each plate. Specificity of the reaction was confirmed by demonstrating no staining with pre-immune serum and secondary antibody alone or in NES2Y cells, which do not express detectable levels of PDX-1.
To examine the effect of glucose on the intracellular distribution of PDX-1, NesPDX-1 cells were incubated in medium containing low or high glucose concentrations and subjected to immunocytochemistry using an anti−PDX-1 antibody. In cells incubated in 0.5 mmol/l glucose, as previously shown (15,28), PDX-1 was located predominantly around the nuclear periphery, with some staining within the cytoplasm (Fig. 1). After incubation for 30 min in medium containing 20 mmol/l glucose, PDX-1 was localized exclusively within the nucleus (Fig. 1); within the nucleus, PDX-1 staining appeared punctate, which may have reflected association with subnuclear structures, such as nuclear speckles that are known to be sites of increased transcriptional activity and pre-mRNA splicing. The distribution of PDX-1 was similar to that of lamin-B1, a protein localized in the nuclear envelope. The intracellular distribution of lamin-B1, however, was unaffected by glucose (data not shown). In addition, glucose had no effect on the intracellular distribution of another transcription factor, USF, that was localized within the nucleus under all conditions (data not shown).
The glucose-dependent movement of PDX-1 from the nuclear periphery/cytoplasm to the nucleoplasm occurred within 10 min after treatment with high glucose and was complete within 15 min (Fig. 2). The effect of glucose on the intracellular distribution of PDX-1 was dosage dependent. At 0.5 and 3 mmol/l glucose, PDX-1 was predominantly localized to the nuclear periphery. At 5 mmol/l glucose, there was a substantial amount within the nucleus, and at 10 mmol/l glucose, the movement of PDX-1 was complete, with no further effect observed over a range of concentrations up to 30 mmol/l (Fig. 2).
It has been shown previously that insulin and sodium arsenite (a reagent known to stimulate stress-activated pathways) can stimulate PDX-1 DNA binding and insulin promoter activity (24). To investigate their effect on the nuclear trafficking of PDX-1, NesPDX-1 cells were incubated for 30 min in 0.5 mmol/l glucose supplemented with 20 ng/ml insulin or 1 mmol/l sodium arsenite or in 20 mmol/l glucose. Both insulin and arsenite stimulated the movement of PDX-1 from the nuclear periphery/cytoplasm to the nucleoplasm (Fig. 3).
To determine whether the movement of PDX-1 to the nucleoplasm was reversible, NesPDX-1 cells were incubated in 20 mmol/l glucose or 20 ng/ml insulin in the presence of the protein synthesis inhibitor cycloheximide and then incubated for periods of up to 5 h in 0.5 mmol/l glucose. In preliminary experiments, cycloheximide had no effect on the intracellular distribution of PDX-1 in cells incubated in low or high glucose (data not shown). Both glucose and insulin stimulated movement of PDX-1 from the nuclear periphery/cytoplasm to the nucleoplasm within 30 min. Subsequent transfer to low (0.5 mmol/l) glucose resulted in export from the nucleoplasm (Fig. 4).
PDX-1 was further shown to shuttle between the nuclear periphery/cytoplasm and the nucleoplasm when the concentration of glucose in the medium was varied (Fig. 5). Thus, after incubation for 1 h in 20 mmol/l glucose, PDX-1 was present in the nucleus. After a 5-h incubation in 0.5 mmol/l glucose, PDX-1 was present at the nuclear periphery/cytoplasm. When the concentration of glucose in the medium was increased to 20 mmol/l, PDX-1 was localized back in the nucleoplasm within 30 min (Fig. 5).
The effects of both glucose and insulin on the movement of PDX-1 from the nuclear periphery/cytoplasm to the nucleoplasm were inhibited by the SAPK2/p38 inhibitor SB 203580 and the PI 3-kinase inhibitor wortmannin. Rapamycin, an inhibitor of p70s6k, and PD098059, an inhibitor of the mitogen-activated protein kinase pathway, had no effect (Fig. 6). In keeping with the effects of arsenite on PDX-1 DNA binding activity and insulin promoter activity (24), the arsenite-stimulated movement of PDX-1 into the nucleus was inhibited by SB 203580, but was unaffected by wortmannin. The export of PDX-1 from the nucleoplasm was inhibited by calyculin A and okadaic acid, implying that a dephosphorylation event was occurring (Fig. 7).
Similar results were observed for endogenous PDX-1 in MIN6 cells. Thus, glucose stimulated the movement of PDX-1 from the nuclear periphery/cytoplasm with a similar time course (Fig. 8) and dosage response to glucose (3−5 mmol/l; data not shown), as observed for the exogenous PDX-1 in NesPDX-1 cells. In MIN6 cells, insulin and arsenite also stimulated the intranuclear transport of PDX-1. As for NesPDX-1 cells, the effects of glucose (Fig. 8) and insulin (data not shown) on the movement of PDX-1 from the nuclear envelope to the nucleoplasm were inhibited by wortmannin and SB203580, but were unaffected by PD 098059 and rapamycin. The arsenite-stimulated movement of PDX-1 to the nucleoplasm was inhibited by SB203580, but was unaffected by wortmannin (data not shown). PDX-1 was further shown to shuttle between the nuclear periphery/cytoplasm and nucleoplasm as the cells were transferred between stimulatory media (i.e., that containing 20 mmol/l glucose [Fig. 8], insulin, or arsenite [data not shown]) and nonstimulatory media (i.e., that contained 0.5 mmol/l glucose) conditions. As for NesPDX-1 cells, okadaic acid and calyculin A (Fig. 8) inhibited the export of PDX-1 from the nucleus in MIN6 cells. It should be noted that the morphology of MIN6 cells is different from that of NesPDX-1 cells; the former are smaller, more rounded, and have a tendency to group together.
The findings that PDX-1 undergoes phosphorylation in response to glucose (22,29) and that wortmannin and SB 203580 inhibit its intranuclear translocation are compatible with the view that movement of PDX-1 from the nuclear periphery/cytoplasm to the nucleoplasm is triggered by phosphorylation. In turn, phosphatase-mediated dephosphorylation of PDX-1 appears to be involved in the movement of PDX-1 from the nucleoplasm to the nuclear periphery. PDX-1 therefore joins a growing number of transcription factors that function by a regulated nuclear translocation mechanism to control specific gene expression. These include signal transducers and activators of transcription, nuclear factors of IgκB cells, nuclear factors of activated T-cells (NFATs), p53, and steroid receptors (31). In each case, an activating signal causes a modification to the protein, which is then rapidly transported to the nucleus. PDX-1 is similar to these other examples as it undergoes a modification (phosphorylation), but it differs in that its movement is predominantly between the nuclear periphery and the nucleoplasm rather than the cytoplasm and the nucleus.
The nuclear/cytoplasmic transport of proteins >∼40 kDa is tightly regulated and energy dependent (32). Many transcription factors shuttle between the nucleus and the cytoplasm by a process that relies on their interaction with soluble shuttling receptors that recognize specific nuclear localization sequences and nuclear export sequences (33). A functional nuclear localization signal within PDX-1 (197RRMKWKK) has been identified (34,35). It remains to be determined whether a nuclear export sequence is involved in the movement of PDX-1 from the nucleoplasm to the nuclear periphery. Potential leucine-rich nuclear export sequences within PDX-1 occur at 82LHHLPAQLALP and 178VELAVMLNL.
In keeping with the findings of a previous report (28), in which real-time imaging was used to study the effect of glucose on the intracellular distribution of an myc-tagged PDX-1 construct in MIN6 cells, these results clearly demonstrated that in unstimulated cells, PDX-1 is localized predominantly around the nuclear periphery, with some staining in the cytoplasm. This localization may be the consequence of a physical interaction between PDX-1 and the nuclear envelope or sequestration in a storage zone. Current studies are under way to distinguish between these possibilities by investigating interactions between PDX-1 and purified turkey erythrocyte nuclear membranes. Proteins known to associate with the nuclear envelope include cyclooxygenase 1, cyclooxygenase 2, 5-lipoxygenase, 5-lipoxygenase−activating protein, leukotriene synthase, microsomal glutathione s-transferase, and the lamins. In the case of lamins, tethering to the nuclear envelope depends on palmitoylation and the presence of a basic amino acid–rich nuclear localization sequence (36). Prenylation occurs at the CaaX motif at the COOH-terminus of proteins. PDX-1 does not contain a consensus prenylation sequence nor is there any evidence that it undergoes prenylation. It is possible that the previously identified nuclear localization sequence (34,35) alone may be responsible for the stable association of PDX-1 with the nuclear envelope.
Acetylation has also been implicated in the tethering of proteins to the nuclear envelope. Heterochromatin protein-1 (HP-1) proteins are a family of proteins that function as chromatin modifiers or regulators of gene expression. When microinjected into HeLa cells, recombinant forms of HP-1 were efficiently transported into the nucleus, but en route, they transiently associated with the nuclear envelope. This transient association was blocked when the cells were treated with the deacetylase inhibitors trichostatin A or sodium butyrate (37). However, in the case of PDX-1, acetylation is unlikely to be involved because there is no exact match to the consensus acetylation motif (GKXXP/GKXP) (38), and neither trichostatin A nor sodium butyrate affected the association of PDX-1 with the nuclear envelope (data not shown).
A number of kinases and phosphatases have been implicated in the control of nuclear shuttling. In the case of NFATs, the Ca2+-calmodulin−dependent protein phosphatase calcineurin dephosphorylates NFAT proteins, leading to the unmasking of nuclear localization signals and translocation of NFATs to the nucleus (39). Activation of c-Jun NH2-terminal kinase by overexpression of MKK7 opposes the effects of calcium by promoting nuclear exclusion of NFAT-4 in calcineurin-stimulated baby hampster kidney cells (40). Casein kinase-1α and MEKK-1 inhibit NFAT-4 nuclear translocation (41), whereas glycogen synthase kinase-3 plays a role in the nuclear export of NFAT-3 in neurones (42). Protein kinase-B/Akt has been shown to phosphorylate the forkhead transcription factor FKHR and trigger its export from the nucleus (43). In the present study, the results with wortmannin and SB 203580 lend further support to a model whereby PI 3-kinase and SAPK2/p38 regulate PDX-1 location and DNA-binding activity as well as insulin-promoter activity in response to stimulation by glucose and insulin. It is of interest therefore that SAPK2/p38 has also been shown to phosphorylate and affect the intracellular location of NFATp in intact lymphocytes (44).
It will be important to determine whether glucose mediates its effects on nuclear shuttling of PDX-1 via secreted insulin feeding back on the β-cell or through generation of a metabolic intermediate that triggers the signaling pathway involved. An example of metabolic regulation of nuclear shuttling is provided by glucokinase. Glucokinase lacks a nuclear localization signal and cannot enter the nucleus by itself. It enters the nucleus as part of a complex with glucokinase regulatory protein (GRP), which does contain a nuclear localization signal. On the other hand, glucokinase contains a nuclear export sequence, and on release from binding to GRP in response to metabolic cues, glucokinase is exported from the nucleus by an active process (45).
The results reported here are at variance with those reported by Moede et al. (34). The latter researchers monitored the intracellular location of a green fluorescent protein–tagged PDX-1 in transfected MIN6 β-cells and failed to show any cycling of PDX-1 between the nucleus and cytoplasm (or nuclear periphery) in response to glucose. The reason for this discrepancy may be related to the use of a green fluorescence protein-tagged PDX-1 versus the intact expressed protein used in the present study or to differences in cell culture conditions.
In conclusion, we demonstrated here that glucose and insulin stimulate shuttling of PDX-1 between the nuclear periphery and the nucleoplasm by a phosphorylation-dependent process involving PI 3-kinase, SAPK2/p38, and a nuclear phosphatase(s). The results may have important implications for a better understanding of the mechanisms through which nutrients and hormones regulate insulin gene expression in the islets of Langerhans.
This work was supported by the Wellcome Trust. L.J.E. was funded by a Biotechnology and Biological Science Research Council studentship. We thank Chris Wright for providing the anti−PDX-1 antibody.
Address correspondence and reprint requests to Professor Kevin Docherty, Department of Molecular and Cell Biology, University of Aberdeen, Institute of Medical Sciences, Foresterhill, Aberdeen, AB25 2ZD, U.K. E-mail:.
Received for publication 29 March 2001 and accepted in revised form 3 July 2001.
BSA, bovine serum albumin; FITC, fluorescein isothiocyanate; GRP, glucokinase regulatory protein; HP-1, heterochromatin protein-1; NFAT, nuclear factor of activated T-cell; PBS, phosphate-buffered saline; PDX-1, pancreatic duodenal homeobox-1; PI 3-kinase, phosphatidylinositol 3-kinase; USF, upstream stimulatory factor.