CD44 expression is aberrant in benign Schwann cell tumors possessing mutations in the neurofibromatosis type 2, but not type 1, gene.

Atypical expression of CD44 splice variants has been implicated in the progression of numerous tumors. This abnormal CD44 expression is presumed to result from gene alterations that cause tumorigenic transformation. Two tumor types that have been linked to specific gene alterations are schwannomas, which have mutations in the neurofibromatosis (NF) type 2 (NF2) gene, and neurofibromas, which characteristically possess NF type 1 (NF1) gene mutations. We examined CD44 expression in normal sciatic nerves, in schwannomas with confirmed NF2 mutations, and in neurofibromas and malignant peripheral nerve sheath tumor tissue and cell lines from NF1 patients. Compared to normal nerves, schwannomas express higher total levels of CD44 and additional splice variants, whereas CD44 expression in neurofibromas is unaltered. Malignant peripheral nerve sheath tumor tissue and cell lines express the CD44v6 epitope, which is not expressed by normal Schwann cells or by other Schwann cell tumors. These data indicate that altered CD44 expression correlates strictly with mutations in the NF2 but not NF1 gene and suggest that CD44v6 might be a marker for the malignant transformation of Schwann cells.

Ruffling membrane, stress fiber, cell spreading and proliferation abnormalities in human Schwannoma cells.

Schwannomas are peripheral nerve tumors that typically have mutations in the NF2 tumor suppressor gene. We compared cultured schwannoma cells with Schwann cells from normal human peripheral nerves (NHSC). Both cell types expressed specific antigenic markers, interacted with neurons, and proliferated in response to glial growth factor, confirming their identity as Schwann cells. Schwannoma cells frequently had elevated basal proliferation compared to NHSC. Schwannoma cells also showed spread areas 5-7-fold greater than NHSC, aberrant membrane ruffling and numerous, frequently disorganized stress fibers. Dominant negative Rac inhibited schwannoma cell ruffling but had no apparent effect on NHSC. Schwannoma cell stress fibers were inhibited by C3 transferase, tyrphostin A25, or dominant negative RhoA. These data suggest that the Rho and Rac pathways are abnormally activated in schwannoma cells. Levels of ezrin and moesin, proteins related to the NF2 gene product, merlin, were unchanged in schwannoma cells compared to NHSC. Our findings demonstrate for the first time that cell proliferation and actin organization are aberrant in schwannoma cells. Because NF2 is mutant in most or all human schwannomas, we postulate that loss of NF2 contributes to the cell growth and cytoskeletal dysfunction reported here.

A novel method for analysis of nuclear receptor function at natural promoters: peroxisome proliferator-activated receptor gamma agonist actions on aP2 gene expression detected using branched DNA messenger RNA quantitation.

Peroxisome proliferator-activated receptor-gamma (PPARgamma), a member of the nuclear hormone receptor superfamily, plays an essential role in the mediation of the actions of antidiabetic drugs known as thiazolidinediones (TZDs). PPARgamma activates many target genes involved in lipid anabolism including the adipocyte fatty acid binding protein (aP2). In this study, induction of aP2 gene expression by PPARgamma agonists was examined in both cultured cells and diabetic mice using branched DNA (bDNA)-mediated mRNA quantitation. bDNA technology allows for the direct measurement of a particular mRNA directly within cellular lysate using a 96-well plate format in a time frame comparable to a reporter gene assay. In cultured human subcutaneous preadipocytes, the TZDs, troglitazone and BRL-49653, both rapidly induced aP2 mRNA as detected with the bDNA method. In these cells, the effect of BRL-49653 on aP2 mRNA levels was detectable as early as 30 min after treatment (47% increase) and was maximal after 24 h of treatment (12-fold increase). The effects of troglitazone on aP2 mRNA induction were similar to those of BRL-49653 except that the maximal level of induction was consistently lower (e.g. 24 h treatment = 4-fold increase). Dose-response relationships for both of the TZDs were also determined using the 24-h treatment time point. EC50s for both BRL-49653 and troglitazone were estimated to be 80 nM and 690 nM, respectively. A natural PPARgamma ligand, 15-deoxy-delta12,14-PGJ2, was also active in this assay with a maximal induction of aP2 mRNA of approximately 5-fold when tested at 1 microM. Since the PPARgamma:retinoid X receptor (RXR) heterodimer has been characterized as a permissive heterodimer with respect to RXR ligands, the ability of 9-cis-retinoic acid (9-cis-RA) to induce aP2 mRNA was examined. Although 9-cis-RA had very low efficacy (2-fold induction), the maximal effect was reached at 100 nM. No synergism or additivity in aP2 mRNA induction was detected when 9-cis-RA was included with either of the TZDs used in this study. Significant induction of aP2 mRNA in bone marrow of db/db mice treated with either troglitazone or BRL-49653 was also detected, indicating that the bDNA assay may be a simple method to monitor nuclear receptor target gene induction in vivo.

The effect of histidine modification on the activity of myo-inositol monophosphatase from bovine brain.

The pH dependence of myo-inositol monophosphatase may indicate a role for histidine residues in the catalytic mechanism (Ganzhorn, A. J., and Chanal, M.-C. (1990) Biochemistry 29, 6065-6071). This possibility was investigated by chemical modification. At pH 6.0 and 25 degrees C, the enzyme was inactivated by diethylpyrocarbonate in a pseudo-first order reaction with a bimolecular rate constant of 0.37 M-1 s-1. Two histidines were modified rapidly with no effect on enzyme activity, while 3 residues were modified at a slower rate corresponding to the rate of inactivation. No noticeable changes in the secondary structure of the enzyme were observed by comparison of circular dichroic spectra before and after modification. Treatment of myo-inositol monophosphatase with diethylpyrocarbonate in the presence of inositol 1-phosphate, Mg2+, and Li+ protected 2 residues from modification and decreased the inactivation rate by about 5-fold. Spectrophotometric analysis, the restoration of enzyme activity by hydroxylamine, and the lack of any inhibitory effect with alkylating agents suggest that inactivation is due solely to modification of histidine. We conclude that a histidine residue is essential for activity and may act as a base catalyst during hydrolysis of the substrate.

Kinetic characterization of enzyme forms involved in metal ion activation and inhibition of myo-inositol monophosphatase.

Activation and inhibition of recombinant bovine myo-inositol monophosphatase by metal ions was studied with two substrates, D,L-inositol 1-phosphate and 4-nitrophenyl phosphate. Mg2+ and Co2+ are essential activators of both reactions. At high concentrations, they inhibit hydrolysis of inositol 1-phosphate, but not 4-nitrophenyl phosphate. Mg2+ is highly selective for inositol 1-phosphate (kcat. = 26 s-1) compared with the aromatic substrate (kcat. = 1 s-1), and follows sigmoid activation kinetics in both cases. Co2+ catalyses the two reactions at similar rates (kcat. = 4 s-1), but shows sigmoid activation only with the natural substrate. Li+ and Ca2+ are uncompetitive inhibitors with respect to inositol 1-phosphate, but non-competitive with respect to 4-nitrophenyl phosphate. Both metal ions are competitive inhibitors with respect to Mg2+ with 4-nitrophenyl phosphate as the substrate. With inositol 1-phosphate, Ca2+ is competitive and Li+ non-competitive with respect to Mg2+. Multiple inhibition studies indicate that Li+ and Pi can bind simultaneously, whereas no such complex was detected with Ca2+ and Pi. Several sugar phosphates were also characterized as substrates of myo-inositol monophosphatase. D-Ribose 5-phosphate is slowly hydrolysed (kcat. = 3 s-1), but inhibition by Li+ is very efficient (Ki = 0.15 mM), non-competitive with respect to the substrate and competitive with respect to Mg2+. Depending on the nature of the substrate, Li+ inhibits by preferential binding to free enzyme (E), the enzyme-substrate (E.S) or the enzyme-phosphate (E.Pi) complex. Ca2+, on the other hand, inhibits by binding to E and E.S, in competition with Mg2+. The results are discussed in terms of a catalytic mechanism involving two metal ions.

CENSOR--a program for identification and elimination of repetitive elements from DNA sequences.

CENSOR is a program designed to identify and eliminate fragments of DNA sequences homologous to any chosen reference sequences, in particular to repetitive elements. CENSOR is based on two principal algorithms of Smith & Waterman (1981) [J. Mol. Biol. 147, 195] and Wilbur & Lipman (1983) [Proc. Natl. Acad. Sci. U.S.A. 80, 726]. It includes several pre-set sensitivity levels based on both biological and statistical criteria which help to distinguish between aligned pairs of homologous and non-homologous sequences. CENSOR has been implemented in C/C + + in the SUN/UNIX environment.

PPARgamma activation induces the expression of the adipocyte fatty acid binding protein gene in human monocytes.

The peroxisome-proliferator activated receptor gamma (PPARgamma), a member of the nuclear receptor superfamily of ligand activated transcription factors, plays a key role in the anti-diabetic actions of the thiazolidinediones (TZDs). PPARgamma induces the expression of many genes involved in lipid anabolism, including the adipocyte fatty acid binding protein (aP2), and is a key regulator of adipocyte differentiation. PPARgamma is also expressed in hematopoietic cells and is up-regulated in activated monocytes/macrophages. Activation of PPARgamma may play a role in the induction of differentiation of macrophages to foam cells that are associated with atherosclerotic lesions. We report that both natural and synthetic PPARgamma agonists induce time- and dose-dependent increases in aP2 mRNA in both primary human monocytes and the monocytic cell line, THP-1. These data suggest that PPARgamma activation may play a role in monocyte differentiation and function analogous to its well-characterized role in adipocytes.

The contribution of lysine-36 to catalysis by human myo-inositol monophosphatase.

The role of lysine residues in the catalytic mechanism of myo-inositol monophosphatase (EC 3.1.3.25) was investigated. The enzyme was completely inactivated by amidination with ethyl acetimidate or reductive methylation with formaldehyde and cyanoborohydride. Activity was retained when the active site was protected with Mg2+, Li+, and D,L-myo-inositol 1-phosphate. Using radiolabeling, peptide mapping, and sequence analysis, Lys-36 was shown to be the protected residue, which is responsible for inactivation. Replacing Lys-36 with glutamine produced a mutant protein, K36Q, with similar affinities for the substrate and the activator Mg2+, but a 50-fold lower turnover number as compared to the wild-type enzyme. Crystallographic studies did not indicate any gross structural changes in the mutant as compared to the native form. Initial velocity data were best described by a rapid equilibrium ordered mechanism with two Mg2+ binding before and a third one binding after the substrate. Inhibition by calcium was unaffected by the mutation, but inhibition by lithium was greatly reduced and became noncompetitive. The pH dependence of catalysis and the solvent isotope effect on kcat are altered in the mutant enzyme. D,L-myo-Inositol 1-phosphate, 4-nitrophenyl phosphate, and D-glucose 6-phosphate are cleaved at different rates by the wild-type enzyme, but with similar efficiency by K36Q. All data taken together are consistent with the hypothesis that modifying or replacing the lysine residue in position 36 decreases its polarizing effect on one of the catalytic metal ions and prevents the efficient deprotonation of the metal-bound water nucleophile.

Inhibition of myo-inositol monophosphatase isoforms by aromatic phosphonates.

alpha-Hydroxyphosphonates are moderately potent (Ki = 6-600 microM) inhibitors of the enzyme myo-inositol monophosphatase (McLeod et al., Med. Chem. Res. 1992, 2, 96). Hydroxy-[4-(5,6,7,8-tetrahydronaphtyl-1-oxy)phenyl]methyl phosphonate (3) was resynthesized and its inhibitory potency towards the recombinant bovine brain enzyme confirmed (Ki = 20 microM). Similar aromatic difluoro-, keto-, and ketodifluorophosphonates (5, 7, 9) were inactive. Compound 3 was 15-fold less active on the human as compared to the bovine enzyme. Molecular modeling suggested that the hydrophobic part of the inhibitor interacts with amino acid side chains that are located at the interface between the enzyme subunits in an area (amino acids 175-185) with low similarity between the two isozymes. Phe-183 in the human enzyme was replaced with leucine, the corresponding residue in the bovine isoform. The three isozymes (human wild-type, bovine wild-type and human F183L) had similar kinetic properties, except that the bovine enzyme was less effectively inhibited by high concentrations of the activator Mg2+. The F183L mutant enzyme had a twofold increased affinity for compound 3 as compared to the human wild-type form. We conclude that residue 183 contributes to the binding of aromatic hydroxyphosphonates to IMPase, but it is not the only determining factor for inhibitor specificity with respect to different isozymes.

7Li nuclear-magnetic-resonance study of lithium binding to myo-inositolmonophosphatase.

The interaction of Li+ with myo-inositol monophosphatase was studied by 7Li-NMR spectroscopy. Li+ binding to the enzyme induces a downfield shift and broadening of the 7Li-NMR signal. Changes of the chemical shift were used to follow the titration of the enzyme with lithium and to determine a dissociation constant, Kd = (1.0 +/- 0.1) mM. Only one major binding site/enzyme subunit was inferred. The complex forms independently of the presence of inorganic phosphate. Metals from the group IIa of the periodic table compete with Li+ binding with the affinity increasing in the order Mg2+ < Ca2+ < Be2+. In contrast to lithium, their binding is enhanced by phosphate.

Interaction between two isoforms of the NF2 tumor suppressor protein, merlin, and between merlin and ezrin, suggests modulation of ERM proteins by merlin.

The product of the neurofibromatosis type II (NF2) tumor suppressor gene, merlin, is closely related to the ezrin-radixin-moesin (ERM) family, a group of proteins believed to link the cytoskeleton to the plasma membrane. Mutation in the NF2 locus is associated with Schwann cell tumors (schwannomas). The two predominant merlin isoforms, I and II, differ only in the carboxy-terminal 16 residues and only isoform I is anti-proliferative. Merlin lacks an actin-binding domain conserved among ezrin, radixin and moesin. Because merlin, ezrin and moesin are co-expressed in Schwann cells, and all homodimerize, we have examined whether merlin and ezrin dimerize with one another. We found by immunoprecipitation and yeast two-hybrid assays that both merlin isoforms interact with ezrin. The interaction occurs in a head-to-tail orientation, with the amino-terminal half of one protein interacting with the carboxy-terminal half of the other. The two merlin isoforms behave differently in their interaction with ezrin. Isoform I binds only ezrin whose carboxy-terminus is exposed, whereas isoform II binds ezrin regardless of whether ezrin is in the open or closed conformation. The heterodimerization of merlin is a much stronger interaction than the interaction between either merlin isoform and ezrin, and can inhibit merlin-ezrin binding. This suggests that, in vivo, merlin dimerization could regulate merlin-ERM protein interaction, and could thus indirectly regulate other interactions involving ERM proteins.