Regulation of gene expression via the protein kinase A (PKA) pathway is mediated through Ser133 phosphorylation of the transcription factor (TF), cAMP response element (CRE) binding protein (CREB). Secalonic acid D (SAD), a mycotoxin causing cleft palate (CP), induces phosphorylation of palatal CREB in vivo. SAD-induced increase in phosphoCREB (pCREB), however, is associated with decreased binding of TF to CRE in vivo. Mechanism(s) involved in these two effects of SAD were studied using palatal nuclear extracts (PNE). Stimulation of CREB phosphorylation by SAD was confirmed in vitro in both cell culture and cell-free systems, and this phosphorylation was not altered by currently known CREB kinase (PKA, CaMK, MEK, p38MAPK, PKC) or phosphatase inhibitors. SAD-induced increase in pCREB, however, was associated with decreased TF binding to CRE in vitro. Two-dimensional gel analysis ruled out additional inhibitory phosphorylations. Addition of SAD to PNE following an increase in PKA-phosphorylated CREB resulted in reduced TF binding to CRE. Further, SAD was shown to bind directly to phosphorylated nuclear proteins (pCREB) with greater affinity. In addition, the inhibitory effect of SAD occurred with CRE of proliferating cell nuclear antigen (PCNA) gene. These studies confirm that stimulation of CREB phosphorylation by SAD does not involve sites other than Ser133 and is mediated by a novel kinase. They also indicate that SAD directly binds to CREB to inhibit its binding to CRE of genes such as PCNA. This effect could lead to reduced palatal mesenchymal cell number, smaller palatal shelf, and thus CP.
SAD is a mycotoxin that produces cleft palate (CP) as the only malformation in fetal mice at maternal doses of potential relevance to human health (Reddy and Reddy, 1991). More importantly, SAD has been used as a chemical model to study mechanisms of chemical-induced CP in CD1 mice. SAD-induced CP is associated with smaller palatal shelves that fail to elevate. At the molecular level, SAD has been shown to inhibit adenylate cyclase activity (Reddy et al., 1994), attenuate the cAMP surge occurring at later stages of normal murine palate development (Eldeib and Reddy, 1988), and inhibit the activity of protein kinase A (PKA) both in vitro (Balasubramanian and Reddy, 1998) and in vivo (unpublished data). Associated with these effects, SAD increased phosphorylation of CREB in the palate and attenuated binding of phosphorylated CREB to consensus sequence of CRE and thereby total transcription factor (TF)-CRE complex formation. In contrast, the binding of CREB to CRE was directly proportional to its phosphorylation by PKA at Ser133 at the later stages of normal palate development (Umesh et al., 2000).
Although CREB is primarily described as a TF responsive to cAMP signaling pathway, it is now known to mediate signals from a variety of pathways. Transactivation of CREB by way of its phosphorylation is a convergence point for other signaling pathways involving CaM kinases II and IV (Sun et al., 1994), p38/MAPK kinases (Tan et al., 1996), and PKC (Gonzalez et al., 1989). Although phosphorylation of CREB is directly correlated with recruitment of CREB-binding protein (CBP) to the transcription initiation site (reviewed by Montminy, 1997) beyond doubt, its effect on DNA binding and dimerization is less clear (Boshart et al., 1991; Montminy and Bilezikjian, 1987; Nichols et al., 1992; Wu et al., 1998). While several investigators have shown phosphorylation at Ser133 to be positively correlated with its binding to CRE and its stimulatory function (Bullock and Habener, 1998; Nichols et al., 1992; Weih et al., 1990), phosphorylation at Ser133 along with Ser142 appears to inhibit CREB function (Sheng et al., 1991; Sun et al., 1994).
The present study was therefore designed with the objectives of identifying the nature of SAD-induced phosphorylation of CREB and the mechanism behind SAD-induced attenuation of TF-CRE complex formation in the palatal tissue, and in addition, whether SAD also inhibits TF binding to CRE of proliferating cell nuclear antigen, a gene involved in regulation of proliferation. This will demonstrate a functional consequence to the effects of SAD on cAMP pathway demonstrated so far. A palatal mesenchymal cell culture system was used to confirm the effect of SAD on CREB phosphorylation and its inhibition of TF binding to CRE. In vitro assays were developed to study the involvement of some of the known CREB kinases or phosphatase in SAD-induced phosphorylation of CREB and to identify inhibitory phosphorylations on CREB, if any. Further, experiments were designed to find out whether SAD was directly involved in alterations of TF-CRE complex formation and the role of phosphorylation on such an alteration.
MATERIALS AND METHODS
SAD was extracted and purified as described by Reddy et al. (1979). Opti-MEM and other tissue culture materials were from Gibco-BRL (Life Technologies, Gaithersburg, MD). The Ser133-phospho-specific CREB antibody was from New England Biolabs (Boston, MA), and CREB antibody was from Santa Cruz Biotech (Santa Cruz, CA). For gel shift assays, double-stranded CREB consensus oligonucleotide was obtained from Santa Cruz Biotech (Santa Cruz, CA) and was end-labeled with γ-P32ATP (NEN Life Science, Boston, MA) using the enzyme T4 polynucleotide kinase (Promega, Madison, WI). PKA-Ca and l-phosphatase were from Calbiochem (La Jolla, CA). Bisindolylmaleimide-I, PD098059, and SB203580 were from Alexis Chemicals (San Diego, CA), and PKA inhibitor (PKI) peptide was from Upstate Biotech (Lake Placid, NY). Compound R24571, okadaic acid, ATP, and other routine chemicals were from Sigma Chemicals (St. Louis, MO). Two-dimensional gel electrophoresis supplies were from Bio-Rad (Hercules, CA).
Animals, cell culture, and treatment.
CD1 mice (Charles River, Wilmington, MA) were housed at 70 ± 2°F and humidity of 60 ± 5% in plastic cages with corn cob as bedding and water and chow (Purina, St Louis, MO) ad libitum. Midpoint of a 2-h cohabitation resulting in a vaginal plug was considered as the beginning of gestation day (GD) zero. For cell culture, GD 13 pregnant females were sacrificed and the embryos were collected. Palates from the embryos were dissected, and palatal mesenchymal (PM) cell cultures were established as described by Pisano and Greene (1999). Upon reaching subconfluency, the cells were exposed for 48 h to SAD at final concentrations of 2.9, 11.7, 47, and 188 mM, which correspond to 16 times lower, 4 times lower, equivalent, and 4 times higher than the in vivo median maternal dosage of 30 mg/ kg. The control groups were exposed to the same volume of the vehicle (5% NaHCO3). The cells were collected and preserved at –80°C until further use.
In vitro phosphorylation.
Palatal CREB-rich fraction was obtained in the nuclear extract prepared as described earlier (Umesh et al., 2000). Phosphorylation of CREB was achieved by incubating 15 mg (10 mg nuclear, 5 mg cytoplasmic protein) of nuclear factor-enriched and reconstituted sample with either 4.7 mM SAD (10-fold higher than the in vivo median maternal dose equivalent) or 1 U of PKA-Ca in the presence of 1 mM ATP and 10 mM MgCl2 in 50 mM MOPS buffer, pH 6.8, for 30 min at 37°C in a water bath. Inhibition studies were carried out separately in the presence of either PKI peptide (2 mM); Calmodulin antagonist compound R24571 (1 mM); PKC inhibitor bisindolylmaleimide-I (10 mM); MEK inhibitor PD098059 (100 mM); p58MAPK inhibitor SB203580 (5 mM), or increasing concentrations (1, 2, and 4 mM) of protein phosphatase 2A (pp2A) inhibitor okadaic acid, with or without SAD. The concentrations of inhibitors used were either equal to or more than those used in cell culture systems to successively block the activities of respective enzymes (Daibata et al., 1994; Lee et al., 1999; Potchinsky et al., 1997; Xing et al., 1998).
Western blot analysis.
To demonstrate the extent of phosphorylation of CREB, Western analyses were carried out as described previously for palate tissue (Balasubramanian et al., 2000). Briefly, nuclear extracts (30 mg protein) after respective in vitro treatments were mixed with equal volume of 2X sample buffer and heated in boiling water for 5 min. The samples were resolved by electrophoresis on 12% SDS-polyacrylamide gel and electrophoretically transferred onto nitrocellulose membrane. The nonspecific sites on the membrane were blocked by incubating with 5% nonfat milk for 1 h at room temperature. The membrane was then washed with PBS-Tween and incubated with either anti-CREB antibody (1:500) at room temperature for 3 h or anti-phosphoCREB antibody (1:2000) at 4°C overnight. The membrane was further washed with PBS-T and incubated with HRP-conjugated secondary antibody (1:5000; Santa Cruz Biotech, Santa Cruz, CA). The protein-antibody complexes were detected by enhanced chemiluminescence (ECL Plus, Amersham, Piscataway, NJ) and visualized on radiographic film.
Electrophoretic mobility shift assay.
Electrophoretic mobility shift assays (EMSA) were performed as described previously (Umesh et al., 2000). Briefly, double-stranded CRE consensus oligonucleotide 5‘-AGAGATTGCCTGACGTCAGAGAGCTAG-3‘ with a single palindrome of consensus CRE (bold lettering) was end-labeled with γ32P-ATP by a kinase reaction with T4 polynucleotide kinase and later gel purified. Ten milliliters of DNA-protein mixture was established using reaction buffer containing 60 mM KCl, 10 mM MgCl2, 1 mM DTT, 1 mM EDTA in 10 mM HEPES (pH, 7.9), 0.1% NP-40, 5% glycerol, 10 mg nuclear extract, and 1 mg poly(dI-dC) (Amersham, Piscataway, NJ) per lane and incubated at room temperature for 15 min. To demonstrate the presence of CREB in the complex, 1 mg of either anti-CREB antibody or nonspecific anti-IgG was incubated with nuclear extract for 30 min on ice before the addition of radiolabeled probe. After the addition of 20,000 cpm radiolabeled CRE probe, the protein-DNA binding was allowed to proceed for 30 additional minutes. DNA-protein complexes were electrophoretically resolved from free oligonucleotides on a 5% nondenaturing polyacrylamide gel using Tris borate EDTA buffer (pH 8.0) at a constant 200 V for 2 h at 4°C. The gels were then dried and DNA-protein complexes were visualized by autoradiography. Mouse PCNA-CRE complementary oligomers (5‘-ATCAGCGCTGTGGCGTCATGACCTCGCTGACAG-3‘) (Feuerstein et al., 1995) were synthesized (Biosource International, Camarillo, CA). Complementary strands were annealed by heating to 90°C and cooling slowly to room temperature; 5‘ ends were dephosphorylated by calf intestinal alkaline phosphatase and radiolabeled with T4 polynucleotide kinase as described before. The DNA-protein reaction and electrophoresis conditions were essentially the same as described for consensus CRE.
2D protein electrophoresis.
Phosphorylation sites on CREB were identified by Western analysis following 2D protein separation. For isoelectric focusing (IEF), samples (100 mg protein) following in vitro treatments were desalted by centrifugal concentrator with a nominal molecular weight limit of 3 kD (Microcon, Millipore Corp., Bedford, MA) and reconstituted in a buffer containing 8 M urea, 0.5% CHAPS, 20 mM DTT, and 0.2% Biolyte (pH 3–10; Bio-Rad, Hercules, CA). With the same buffer, precast immobilized pH gradient gel (IPG) strips (11 cm, pH 3–10) were rehydrated, and samples were loaded per manufacturer’s instructions. IEF was carried out in the Bio-Rad Protean IEF cell with the focusing conditions essentially as per the directions of the manufacturer (Bio-Rad, Hercules, CA). Following IEF, the IPG strips were equilibrated for 20 min with SDS-PAGE equilibration buffer (6 M urea, 2% SDS, and 20% glycerol in 0.375 M Tris; pH 8.8) with 2.5% iodoacetamide and aligned on top of 12% precast SDS gels for the 2D electrophoresis. After the 2D electrophoresis, the gels were treated as for Western analyses (described above) using anti-phosphoCREB antibody.
When bound to protein, SAD shows higher fluorescence, with the amount of SAD-protein complex formed directly correlating with the intensity of fluorescence (Nakamura et al., 1983). To determine whether SAD binds to palatal nuclear proteins preferentially when they are phosphorylated, fluorescent spectroscopic analysis was carried out with an excitation wavelength of 380 nm and an emission wavelength of 535 nm. Phosphorylated and dephosphorylated palatal nuclear proteins were prepared by incubating nuclear extracts (10 mg protein) with 1 U of PKA-Ca with 1 mM ATP and 1 kU of l-phosphatase in 20 mM MOPS, respectively, at 37°C for 30 min. Total protein content in the mixtures were maintained the same by addition of heat-inactivated PKA-Ca and l-phosphatase and ATP to the respective samples. SAD was added to the mixtures at a final concentration of 4.7 mM and maintained at room temperature for 15 min. The samples were diluted 500-fold with 5% NaHCO3 and read in quartz cuvettes (4 ml, 1 cm path length). The intensity of fluorescence recorded was considered indicative of the extent of SAD-protein complex in the sample.
SAD Induces Phosphorylation of CREB
Palatal mesenchymal cell cultures exposed for 48 h to SAD exhibited optimally stimulated phosphorylation of CREB (Fig. 1A) at 11.7 mM (four-fold lower than the in vivo maternal dose equivalent). Higher concentrations of SAD, however, produced a gradual reduction of phosphorylation from this optimal level to below that of controls at 188 mM (four-fold higher than the in vivo maternal dose equivalent). CREB levels remained same at the concentrations tested (Fig. 1B). Further, fraction enriched with nuclear factors derived from whole palatal shelves also showed an increase in phosphorylation of CREB (Fig. 1C) when incubated with SAD at a final concentration of 4.7 mM (ten-fold lower than the in vivo maternal dose equivalent) in the presence of ATP.
SAD-Induced CREB Phosphorylation Is Not Mediated by PKA, CaMK, MEK, p53MAPK, or PKC
In vitro phosphorylation reactions carried out in the presence of SAD with or without the specific inhibitors of various kinases known to phosphorylate CREB yielded results indicating that SAD phosphorylated CREB even in the presence of PKI peptide, CaMK inhibitor compound R24571 (Fig. 2A), in the presence of MEK inhibitor SB203580, p53MAPK inhibitor PD098059, or PKC inhibitor bisindolylmaleimide I (Fig. 2B).
SAD-Induced CREB Phosphorylation Is Not Mediated by Phosphatase
Figure 3 presents data from the in vitro phosphorylation reactions carried out in the presence of SAD with or without increasing amounts of okadaic acid, a protein phosphatase 2A (pp2A) inhibitor. Levels of phosphoCREB increased in the presence of SAD independent of the presence of okadaic acid.
Presence of SAD Decreases TF Binding to CRE Despite Increase in CREB Phosphorylation
The TF in the complex was confirmed to be CREB by using a CREB antibody, which ablated the TF-CRE complex (Fig. 4A). Total complex formation with CRE decreased in a dose-dependent manner in nuclear extracts prepared from cells exposed to increasing concentrations of SAD (Fig. 4B). To evaluate the effect of SAD-induced CREB phosphorylation on the binding of CREB to CRE, EMSA was carried out on samples containing phosphoCREB (phosphorylated in vitro separately with PKA-Ca and SAD). Phosphorylation of CREB and binding of TF to CRE in SAD-stimulated sample were similar to the phosphorylation and binding of TF in PKA-Ca–stimulated sample (Figs. 4C and 4D). However, addition of SAD to the reaction mixture of PKA-stimulated sample resulted in a decrease in TF-CRE complex formation (Fig. 4E).
SAD-Induced CREB Phosphorylation Does Not Involve a Second Site
Western analysis of phosphorylated CREB following 2D electrophoresis of PKA-Ca–stimulated (Fig. 5A) and SAD-stimulated (Fig. 5B) samples showed a single pCREB protein and similar pattern of migration of pCREB in both samples. PKA-Ca also phosphorylated ATF-1, which cross-reacts with pCREB antibody (second spot in panel 5A). SAD, however, failed to phosphorylate this TF (panel 5B). The pCREB antibody recognized the Ser133-phosphorylated form of both CREB and ATF-1.
SAD Directly Interferes with Binding of TF to CRE
SAD inhibited TF-CRE complex formation in a dose-dependent manner, with 4.7 mM SAD almost completely abolishing the complex formation (Fig. 6). Analysis of SAD-protein complex by fluorescent spectroscopy revealed severalfold greater intensity of fluorescence when SAD was added to phosphorylated palatal nuclear extract compared with the dephosphorylated extract (Figs. 7A and 7B). The inset in panel 7B shows that phosphatase used in our experiment effectively dephosphorylated pCREB.
SAD Directly Interferes with TF Binding with PCNA-CRE
As an example of a CRE-driven gene, PCNA-CRE was used to test the relevance of the observed effect with consensus CRE. The presence of SAD completely abolished palatal TF binding to PCNA-CRE (Fig. 8).
Although the expression of CREB in the developing murine palate remains constant throughout, its phosphorylation status undergoes dynamic changes during both normal (Potchinsky et al., 1997) and SAD-altered development (Umesh et al., 2000). Whereas increased phosphorylation of CREB occurring at the later stage (GD 14) is associated with normal development of palate (Potchinsky et al., 1997; Umesh et al., 2000), increased phosphorylation of CREB on GD 12 in the presence of SAD is associated with abnormal palate development. Results of this study using the PM cell culture system showed that this effect of SAD on GD 12 likely occurs in PM cells. Further, the results of in vitro studies showed that this stimulation of phosphorylation of CREB was mediated by direct interaction of SAD with the cytoplasmic and/or nuclear components of PM cells and is not receptor or cell membrane mediated.
In contrast to the general belief that CREB is phosphorylated by PKA, SAD-induced phosphorylation of CREB on GD 12 cannot be mediated by PKA, as suggested by earlier studies which demonstrated that SAD inhibits PKA activity in vitro (Balasubramanian and Reddy, 1998; Wang and Polya, 1995). SAD also inhibits Ca++-dependent calmodulin (CaM) activity, suggesting interference with CaM-dependent enzyme activity (Pala et al., 1999). Therefore, CaMKII stimulation by SAD is also less likely. These two possibilities were confirmed by our in vitro studies, where the presence of PKI peptide and calmodulin antagonist compound R24571 were unable to prevent SAD-induced phosphorylation of CREB in the palatal extract.
PKC, p38MAPK, and MEK, known to phosphorylate CREB, belong to PKC and MAPK pathways active during normal palate development (Balasubramanian and Reddy, 2000; Hehn et al., 1998). However, the involvement of these kinases in enhancing CREB phosphorylation was ruled out by our in vitro studies wherein the presence of respective inhibitors did not prevent SAD-induced phosphorylation of CREB in palatal extracts. Increase in pCREB levels can also be a result of phosphatase inhibition. However, this possibility was also ruled out by our in vitro study wherein pp2A (CREB phosphatase) inhibitor failed to alter SAD action. Together, these results suggest that SAD-induced CREB phosphorylation is likely mediated by a kinase yet to be identified. A similar phosphorylation of CREB, not mediated by the known CREB kinases, has been reported to occur in PC12 cells under moderate hypoxia (Beitner-Johnson et al., 2000).
An increase in cAMP levels or PKA activity resulting in increased phosphorylation of CREB is directly related to CREB binding to CRE (Asanuma et al., 1996; Herring et al., 1998). In agreement with these reports, our present in vitro studies demonstrated PKA-phosphorylated CREB to be associated with increased TF binding to CRE. However, SAD-induced pCREB was associated with decreased TF binding to CRE in previous studies in palatal tissue in vivo (Umesh et al., 2000). In our previous study, total TF binding to CRE, as well as CREB binding to CRE, was reduced in the presence of SAD compared with controls on GD 12, despite a drastic increase in phosphorylation of CREB (Umesh et al., 2000). In this study, a dose-dependent decrease in TF binding to CRE was observed in palatal mesenchymal cells in culture exposed to increasing concentrations of SAD. Published reports suggest that phosphorylation of CREB at sites other than the PKA phosphorylation site (Ser133) may lead to inhibitory effects downstream of this pathway (Sheng et al., 1991; Sun et al., 1994). However, 2D gel analysis and comparison of pCREB induced by SAD to pCREB induced by PKA ruled out the possibility of involvement of phosphorylation sites other than Ser133.
The possibility that SAD inhibits TF binding to CRE by a direct interaction with CREB and other TFs in the TF-CRE complex was confirmed by in vitro studies wherein the presence of SAD in the reaction mixture inhibited TF-CRE complex formation despite high levels of CREB existing in its PKA-phosphorylated form. The unique ability of SAD to fluoresce upon binding to proteins (Nakamura et al., 1983) was used to study the direct interaction of SAD with components in the palatal nuclear extract. The fact that SAD bound to phosphorylated nuclear proteins with greater affinity compared with the dephosphorylated preparation and that only CREB and not ATF-1 was phosphorylated by SAD (2D gel results) indicated that SAD bound predominantly to pCREB to inhibit its binding to CRE. Although the exact nature of the physicochemical interaction of SAD with proteins and their phosphorylated forms is not known, such an interaction has been proposed to occur with PKC in vitro (Balasubramanian and Reddy, 2000). Reduction in total TF-CRE complex in vivo and in vitro and reduced participation of CREB in complex formation with CRE in vivo despite CREB existing in its phosphorylated form (Umesh et al., 2000) can be explained by the ability of SAD to preferentially bind to phosphorylated nuclear proteins and keep Ser133-phosphorylated CREB from binding to CRE. SAD-induced phosphorylation of CREB in vitro is at a concentration that is 10-fold lower than the in vivo maternal dose equivalent. This suggests that SAD is available in vivo as well as in vitro at a much higher concentration to bind phosphorylated CREB, exceeding the concentration required for inducing phosphorylation of CREB, and thus may trap all the available CREB, resulting in decreased TF-CRE complex formation.
The requirement of CREB binding to CRE for induction of CRE-containing genes is well documented. Reduction in binding of CREB to CRE leads to a selective decrease in expression of CRE-driven genes such as PCNA by rapamycin (Feuerstein et al., 1995) and dl-propranolol (Hong et al., 1997) affecting proliferation of cells. Actively proliferating PM cells, which form the bulk of the palatal shelves during early development, are required for the palatal shelves to increase in size and elevate, allowing opposing shelves to meet, fuse, and form a normal palate. The SAD phenotype, which has the characteristic small shelves, suggests an alteration in PM cell proliferative function during early development. Our recent results with PM cells confirmed these antiproliferative effects (Reddy et al., 2001). A decrease in PCNA gene expression, shown earlier in vivo and in vitro (Reddy et al., 2001; Umesh et al., 2000), may be due to reduced TF binding to PCNA-CRE in SAD-exposed tissues, as shown in this study. This may explain the reduced proliferative potential of PM cells, with resultant smaller shelves that fail to elevate and fuse at the midline and lead to persistent cleft (Fig. 9). However, the role of other CRE-promoted genes in palate development and the interference of SAD in their expression and function need to be studied.
We thank Dr. Bimal Ray for his technical help in carrying out mobility shift assays.