Transient induction of apoptosis in serum-starved glioma cells by insulin and IGF-1

Insulin has a wide variety of biological effects. One of them is a mitogen-like activity whereby cell proliferation is stimulated. In this study we found a heretofore unreported insulin-elicited transient apoptosis of glioma cells. When serum-starved glioma cells were fed with a fresh regular medium, in the 6- to 12-h post-starvation period, the growth rate as determined by cell number was significantly suppressed by insulin, although cell cycle progression and DNA synthesis were actually accelerated. Increase in apoptosis in those growth-retarded cultures was demonstrable by Hoechst staining, detection of histone-associated DNA fragment, and in situ cell death detection. Apoptosis occurred among cells in all stages of cell cycle. After 24 h post-starvation, insulin increased the total cell number like a typical growth-promoting mitogen. In this regard, IGF-1, but not EGF nor TGF-beta 1, behaved like insulin.

HLA DRB1/DQA1/DQB1 haplotype determines thyroid autoimmunity in patients with insulin-dependent diabetes mellitus

OBJECTIVE

Thyroid autoimmunity is frequently associated with insulin-dependent diabetes mellitus (IDDM). The genetic factors which contribute to thyroid autoimmunity and IDDM have been described but vary between different races. We have therefore investigated the effect of class II HLA genes at both loci and the HLA haplotypes on the presence of autoimmunity in patients with IDDM in Taiwan.

SUBJECTS AND MEASUREMENTS

Eighty-three patients with IDDM and 105 unrelated normal controls were recruited for the measurement of thyroid autoantibodies and for genotyping of HLA DRB1, DQA1 and DQB1 by polymerase chain reaction-based DNA typing techniques.

RESULTS

Among 83 patients with IDDM, 23 (27.7%) were positive for antithyroid autoantibodies. Compared to those without thyroid autoimmunity, there was a female preponderance for IDDM with thyroid autoimmunity (female: male, 3:20 vs 29:31). Among the DR specificities, DR6 was associated with a weak protective effect against thyroid autoimmunity in IDDM patients. Upon detailed analysis of class II HLA haplotypes, the DRB1*0301/ DQA1*0501/DQB1*0201 haplotype was found to be associated with an increased risk of IDDM regardless of thyroid autoimmunity, while DRB1*0405/DQA1*0301/ DQB1*0401 was significantly increased only in the IDDM patients with thyroid autoimmunity. IDDM individuals with the HLA DRB1*0405/DQA1*0301/DQB1*0302 haplotype were not at risk of thyroid autoimmunity.

CONCLUSIONS

Our data indicated that there was a generalized genetic factor within or associated with the DRB1*0301/DQA1*0501/DQB1*0201 haplotype, and a more restricted effect with the DRB1*0405/DQA1*0301/DQB1*0401 haplotype which led to thyroid autoimmunity in patients with insulin-dependent diabetes mellitus.

The cutRS signal transduction system of Streptomyces lividans represses the biosynthesis of the polyketide antibiotic actinorhodin

The nucleotide sequence of a two-component signal transduction operon (cutRS) of Streptomyces lividans TK64 was elucidated. Transcription of the operon was detected during the transition and stationary phases of growth, initiating at a single site upstream of cutR. This promoter region also possessed promoter activity directed away from cutRS, which appears to be responsible for the previously observed suppression of the translational deficiency of a melC1 mutation. Mutations in cutR and cutS were generated by gene replacement. The resulting mutants exhibited accelerated and increased production of the polyketide antibiotic, actinorhodin, which could be reversed by introduction of cutR on a plasmid. cutRS was also shown to repress actinorhodin production in the closely related species, Streptomyces coelicolor A3(2). The cutRS operon is the second two-component system found in Streptomyces that negatively regulates secondary metabolism.

Lipid-ion channel interactions: increasing phospholipid headgroup size but not ordering acyl chains alters reconstituted channel behavior

We have recently shown (Chang et al., 1995) that lipid-channel interactions, exemplified by the effects of cholesterol on the calcium-activated potassium (BK) channel, profoundly affect channel properties. The present study further explores such interactions by monitoring changes in BK channel behavior after reconstitution into bilayers where the size of phospholipid (PL) headgroups is increased and where the freedom of motion (inverse order) of fatty acid chains is incremented. Increasing the PL headgroup cross-sectional area, from that of N-meth-DOPE to that of DOPC (an increase from ca. 60 to 70 A2), is associated with a doubling of the channel mean opentime. Channel conductance, however, was unaffected. Increasing the order of the fatty acid chains, from that of DOPE to POPE and to that of DEPE, had no significant effect on channel properties (at 22 degrees C). We interpret the changes reported here to reflect lipid-protein interactions through the induction of structural stress related to the headgroup structures of phospholipids.

HLA DQA1 genotypes and its interaction with HLA DQB1 in Chinese IDDM living in Taiwan

To study the role of the HLA DQA1 gene and its interaction with DQB1 in the susceptibility of IDDM, subjects with insulin-dependent (type 1) diabetes mellitus and non-diabetic unrelated controls were recruited from a Chinese population living in northern Taiwan. HLA DQA1 exon 2 was enzymatically amplified by polymerase chain reaction. HLA DQA1 alleles were diagnosed by dot blotting and hybridization with 11 sequence-specific oligonucleotide probes. Among all the DQA1 alleles, DQA1*0301 and DQA1*0501 were more frequent while DQA1*0102, DQA1*0103 and DQA1*0601 were less frequent in Chinese with IDDM than in controls. Among the DQA1 genotypes, only DQA1*0301/0301 and DQA1*0301/0501 were associated with increased risk to IDDM while DQA1*0301/0601 and DQA1*0102/0103 were protective against IDDM in our population. As the cell surface HLA DQ molecules were formed from each DQA1 and DQB1 alleles either in cis- or trans-position, the numbers of susceptible HLA DQ alpha beta heterodimers were then derived from the genotypes of HLA DQA1/DQB1 in each person. The numbers of the possible diabetogenic DQ alpha beta dimers correlated with the degree of risk to IDDM (r = 0.92) but were not statistically significant (p > 0.05). Subjects with absence of diabetogenic HLA DQ molecules were resistant to developing IDDM while subjects with two or more forms of diabetogenic DQ molecules were associated with increased risk to IDDM. In conclusion, both DQA1 and DQB1 genes, which determine the formation of susceptible DQ alpha beta heterodimers, were significantly associated with IDDM in Chinese subjects living in Taiwan.

Attenuation of channel kinetics and conductance by cholesterol: an interpretation using structural stress as a unifying concept

The ubiquity of cholesterol in cell membranes and changes in its concentration during development, aging and in various diseases suggest that it plays an important role in modulating cell function. We examined this possibility by monitoring the effects of cholesterol on the activity of the calcium-activated potassium (BK) channel reconstituted into lipid bilayers from rat brain homogenates. Increasing the cholesterol concentration to 11% of total lipid weight resulted in a 70% reduction in channel mean open time and a reduction of the open probability of the channel by 80%. Channel conductance was reduced by 7%. Cholesterol is known to change the order state and the modulus of compressibility of bilayers. These physico-chemical changes may be translated into an overall increase in the structural stress in the bilayer, and this force may be transmitted to proteins residing therein. By examining the characteristics of the BK channel as a function of temperature, in the presence and absence of cholesterol, we were able to estimate the activation energy based on Arrhenius plots of channel kinetics. Cholesterol reduced the activation energy of the BK channel by 50% for the open to closed transition. This result is consistent with an increased stress energy in the bilayer and favors the channel moving into the closed state. Taken together, these data are consistent with a model in which cholesterol induces structural stress which enhances the transition from the open to the closed state of the channel. We suggest that this is an important mechanism for regulating the activity of membrane-integral proteins and therefore membrane function, and that the concept of structural stress may be relevant to understanding the modulation of ion channel activity in cell membranes.

Biosynthesis and processing of the glycosylphosphatidylinositol anchor in mammalian cells

Many T cell surface proteins are attached to the cell membrane by glycosylphosphatidylinositol (GPI) anchors, which may be involved in cell signaling, protein targeting or protein release. Proteins destined to be GPI-anchored have both N- and C-terminal signal peptides, which permit the nascent polypeptides to transverse the endoplasmic reticulum and to be transferred to preformed GPI anchors in a transamidase reaction. The biosynthetic pathway of the GPI anchor in mammalian cells has been elucidated using a panel of T cell mutants that cannot express GPI-anchored proteins on their cell surfaces. The first step in anchor biosynthesis is the transfer of N-acetylglucosamine (GlcNAc) to a phosphoinositol (PI) acceptor to form GlcNAc-PI, which is then deacetylated to form GlcN-PI. Subsequently, a fatty acid is added to the inositol ring and three mannose residues are transferred to the elongating GPI core. Finally, ethanolamine phosphates are transferred to the mannose residues. Three human cDNAs encoding for GPI synthases (Classes A, F and H) have been identified using expression cloning technique. Mutation in the X-linked class A gene was shown to be the cause of paroxysmal nocturnal hemoglobinuria, an acquired disease affecting hematopoietic stem cells in which the abnormal cells are deficient in GlcNAc-PI formation. The potential involvement of GPI-anchored proteins or GPIs in T cell activation is also discussed.

Glycosyl-phosphatidylinositol anchor synthesis in paroxysmal nocturnal hemoglobinuria: partial or complete defect in an early step

The defect in the biosynthesis of the glycosyl-phosphatidyl inositol (GPI) anchor in paroxysmal nocturnal hemoglobinuria (PNH) appears to be in the initial steps. In biosynthetic studies using [3H]mannose, abnormal granulocytes of eight patients, and B lymphocytes transformed by Epstein-Barr virus of six different patients synthesized dolichyl phosphoryl mannose, but little or no later mannosylated intermediates. When fused with murine cell lines known to be deficient at different biosynthetic steps of the GPI anchor, the GPI-anchor-deficient granulocytes of 21/21 patients and lymphocytes from 6/6 patients complemented all murine cell lines except those of class A; cells of this class are not able to add N-acetylglucosamine to phosphatidylinositol. These studies indicate that the defect in GPI-anchor synthesis in PNH is early in the pathway, and is the same as that of class A mutants, but may be partial in some patients, resulting in the production of small amounts of mannosylated intermediates.

Correction of the class H defect in glycosylphosphatidylinositol anchor biosynthesis in Ltk- cells by a human cDNA clone

Previous attempts to express glycosylphosphatidylinositol-anchored proteins in Ltk- cells have not been successful because Ltk- cells cannot synthesize N-acetylglucosamine-phosphatidylinositol, the first intermediate in anchor biosynthesis. Using complementation cloning, we have identified a human cDNA that corrects the defect in anchor biosynthesis and allows the expression of glycosylphosphatidylinositol-anchored proteins in Ltk- cells. The nucleotide sequence predicts a novel cytosolic protein of 188 amino acids.

Can determination of the proliferative capacity of the nontumor portion predict the risk of tumor recurrence in the liver remnant after resection of human hepatocellular carcinoma?

To test the hypothesis that increased proliferative capacity of cells in a liver remnant is a risk factor for tumor recurrence in patients who have undergone liver resection for hepatocellular carcinoma, DNA flow-cytometric measurement and cell-cycle analysis of the nontumor parts of resected hepatocellular carcinomas (tumor size < 5 cm) were performed. The disease-free survival rates 1, 2, 3 and 4 yr after surgery were 64%, 58%, 43%, and 36%, respectively. Proliferative capacity (fractions of synthetic, postsynthetic and mitotic phases) of the nontumor parts, irrespective of liver pathology, was higher than that of normal liver and statistically lower than that of tumor parts from resected hepatocellular carcinoma specimens. Livers with chronic active hepatitis (+) and with hepatocyte dysplasia (-) had significantly lower proliferative activity than did those with chronic active hepatitis (-) and with hepatocyte dysplasia (+), respectively [corrected]. We saw no significant difference in proliferative capacity between patients with and without cirrhosis. Disease-free-survival analysis showed that the presence of liver pathology (hepatitis B infection, cirrhosis, chronic active hepatitis and hepatocyte dysplasia) was not the factor linked to tumor recurrence in the liver remnant and that a marked increase in proliferative capacity (> or = 18%), regardless of liver pathology, was the risk factor linked to tumor recurrence after liver resection. We conclude that there is some degree of increased proliferative capacity in the nontumor parts of resected hepatocellular carcinomas and that a marked increase in the proliferative capacity (> or = 18%) of the nontumor part is a significant risk factor in predicting tumor recurrence in the liver remnant after liver resection.

Lithospermic acid B as an antioxidant-based protector of cultured ventricular myocytes and aortic endothelial cells of rabbits

Lithospermic acid B, an active principle found in a Chinese herbal medicine for treating various heart ailments, was recently isolated, purified and demonstrated by us to salvage the postischemic rabbit heart from reperfusion injury. In this work, we further report that lithospermate B is able to protect each of two types of rabbit cardiocytes, namely ventricular myocytes and aortic endothelial cells against necrosis inflicted by oxyradicals generated pharmacologically with xanthine oxidase and hypoxanthine. Biochemically, the lithospermate B also inhibits the reduction of cytochrome c by superoxide radical anion. Thus, our in vitro data here are in concord with our earlier in vivo finding that lithospermic acid B is most likely an effective antioxidant-based myocardial protector.

Demonstration of the myocardial salvage effect of lithospermic acid B isolated from the aqueous extract of Salvia miltiorrhiza

Lithospermic acid B has been isolated to > 95% purity by high performance liquid chromatography from the aqueous extract of the roots of Salvia miltiorrhiza. When infused at 5.5 mumoles/kg into the post-ischemic rabbit heart, it reduced by 62 +/- 10% (n = 8) the myocardial damage found in the saline control in a rabbit ischemia-reperfusion model.

Comparison of antitumor activity of vitamins K1, K2 and K3 on human tumor cells by two (MTT and SRB) cell viability assays

Vitamin K (VK) congeners (VK1, VK2, and VK3) have been used as antihemorrhagic agents, while VK3 has also been found to inhibit growth in various rodent and human tumor cells. We have compared the antitumor activities of vitamin K1, K2, and K3 against a panel of human cancer cell lines. For each test agent, a dose-response profile was generated by using an MTT (3-(4,5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide) and an SRB (sulforhodamine B) assay. Both assays yielded similar results. The respective ID50 values of VK3 in five hepatoma cell lines, HA59T, HA22T, PLC, HepG2, and Hep3B, of increasing differentiation state, were 42, 36, 28, 27, and 20 microM. For nasopharyngeal carcinoma (CG1), leukemia (U937), oral epidermoid carcinoma (KB), and breast carcinoma (BC-M1) cells, the ID50 values of VK3 were 26, 15, 25, and 33 microM, respectively. For all the above cells, the ID50 values of VK1 ranged from 6 to 9 mM, and the ID50 values of VK2 ranged from 1 to 2 mM. Thus, the relative potencies of antitumor activity of VK3 compared to VK2 and to VK1 are about 60- and 300-fold, respectively. These results support the preference for use of VK3 over VK1 and VK2 in cancer therapy.

Clinicopathologic comparisons between estrogen receptor-positive and -negative gastric cancers

We report that a modified dextran-coated charcoal (DCC) assay, including the addition of sodium molybdate and 5% DCC stripping of endogenous hormone, detected higher estrogen receptors (ERs) than those by the conventional assay. ERs in 21 gastric adenocarcinoma were determined by the modified DCC assay; 13 patients had ERs of 2.5 to 520.2 fmol/mg protein with a mean dissociation constant of 1.9 x 10(-10) M. The remaining 8 patients had no detectable amount of the receptor. There were no differences between the ER-positive and ER-negative groups in clinico-pathologic characteristics such as age, sex, tumor size, location, gross appearance, invasive depth, invasion of lymph vessel or vein of stomach wall, nodal involvement, peritoneal dissemination, liver metastasis, and curability. Histological analysis, using Japanese, Lauren's, or Ming's classifications, all revealed no differences between the ER-positive and -negative groups. The survival rate was identical for the 2 study groups. These results suggest that the presence or absence of ERs in gastric adenocarcinoma does not correlate to either biologic or clinicopathologic characteristics of this tumor; the role of ERs in human gastric cancer remains to be elucidated.

Pathophysiology and principles of pain management in rheumatic diseases

Pain is a major presenting symptom in patients with rheumatic diseases. It is difficult, however, to define pain precisely because there are both objective and subjective components to the experience of pain. Thus, patients with the same degree of tissue inflammation may experience different levels of pain. The variation in individual pain experience often presents us with a diagnostic and therapeutic dilemma. Pain management, therefore, must be based on a solid understanding of the pathophysiology of pain and on a careful review of the risks, benefits, cost, and efficacy of available treatments.

The nontransformed progesterone and estrogen receptors in gastric cancer

Contents of the progesterone receptors (PgR) and estrogen receptors (ER) in 18 gastric adenocarcinoma tissues were determined using both the dextran-coated charcoal (DCC) assay and enzyme immunoassay (EIA). PgR were found in 15 cancer tissues (range, 1.0-58.8 fmol/mg protein) and 12 normal mucosal tissues (range, 1.4-26.8 fmol/mg protein) by DCC assay, whereas only 6 cancer tissues (ranged, 0.2-3.3 fmol/mg protein) and 7 normal mucosal tissues (range, 0.1-0.8 fmol/mg protein) had measurable PgR by EIA analysis. Similar results were observed for ER. DCC assay found ER in 12 cancer tissues (range, 2.9-112.6 fmol/mg protein) and 12 normal mucosal tissues (range, 1.2-36.6 fmol/mg protein), whereas EIA measured ER in 16 cancer tissues (range, 0.1-3.5 fmol/mg protein) and 15 normal mucosal tissues (range, 0.1-4.8 fmol/mg protein). No significant correlation between DCC and EIA was observed for either PgR or ER. DCC assay and its modified procedures including 5% DCC stripping of cytosol and/or the addition of sodium molybdate in buffer were simultaneously measured in 5 gastric adenocarcinoma tissues and 1 gastritis cystica polyposa tissue (a precancerous lesion). Higher receptor levels were found by the modified procedures than by conventional method. Using the DCC procedure with addition of sodium molybdate in buffer for receptor analysis, PgR and ER were found in gastric tissues in six patients, with significantly increased levels of measurable PgR. The results suggest that PgR and ER may be involved in the physiology of normal and gastric cancer tissues; their clinical implications are worthy of further study.

Identification of a missing link in glycosylphosphatidylinositol anchor biosynthesis in mammalian cells

A large number of mammalian proteins are anchored to the cell membrane by a glycosylphosphatidylinositol (GPI) anchor. Biosynthetic intermediates of the GPI anchor have been identified in mammalian cells. The early GPI precursors are sensitive to phosphatidylinositol (PI)-specific phospholipase C (PLC). However, all of the later GPI precursors, which contain 1 or more mannose residues, are PI-PLC-resistant, suggesting that there is another unidentified precursor. Here, we report the identification of this missing link. This GPI precursor can only be labeled with glucosamine and inositol, and is resistant to PI-PLC but sensitive to GPI-phospholipase D. It accumulates in large quantity only in mutants which are defective in the addition of the first mannose residue to the elongating GPI core. Thus, fatty acylation of glucosaminylphosphatidylinositol, to render it PI-PLC-resistant, is an obligatory step in the biosynthesis of mammalian GPI anchor precursors.

Differential expression of glycosylphosphatidylinositol-anchored proteins in a murine T cell hybridoma mutant producing limiting amounts of the glycolipid core. Implications for paroxysmal nocturnal hemoglobinuria

A T cell hybridoma mutant, which expressed a markedly reduced level of glycosylphosphatidylinositol (GPI)-anchored proteins on the cell surface, was characterized. The surface expression level of Thy-1 was approximately 17% of the wild-type level, whereas the surface expression of Ly-6A was approximately 2.4% of the wild-type level. We show here that these cells synthesized limiting amounts of the GPI core and that the underlying defect in these cells was an inability to synthesize dolichyl phosphate mannose (Dol-P-Man) at the normal level. The defect in Ly-6A expression could be partially corrected by tunicamycin, which blocked the biosynthesis of N-linked oligosaccharide precursors and shunted Dol-P-Man to the GPI pathway. Full restoration of Thy-1 and Ly-6A expression, however, required the stable transfection of a yeast Dol-P-Man synthase gene into the mutants. These results revealed that when the GPI core is limiting, there is a differential transfer of the available GPI core to proteins that contain GPI-anchor attachment sequences. Our findings also have implications for the elucidation of the defects in paroxysmal nocturnal hemoglobinuria.

Defective glycosylphosphatidylinositol anchor synthesis in paroxysmal nocturnal hemoglobinuria granulocytes

To investigate the biosynthesis of the glycosylphosphatidylinositol (GPI) anchor in the granulocytes of paroxysmal nocturnal hemoglobinuria (PNH), the glycolipids of granulocytes from PNH patients and normal volunteers were biosynthetically labeled with [3H]mannose in the presence of tunicamycin. Extracted glycolipids were examined by thin-layer chromatography and compared with known biosynthetic intermediates. Normal granulocytes consistently showed [3H]mannose incorporation into the complete GPI core, several GPI biosynthetic intermediates, and dolichol phosphate mannose (DPM). The granulocytes of 10 patients with PNH that had no expression of CD55 and CD59 on greater than 95% of the cells were able to incorporate [3H]mannose into DPM, but were not able to incorporate detectable amounts into the complete GPI core. THus, PNH granulocytes do not synthesize detectable amounts of the complete GPI core and this defect likely accounts for the absence of GPI-linked membrane proteins on hematopoietic cells in this syndrome.

Prediction of relapse or survival after resection in human hepatomas by DNA flow cytometry

To investigate the change of DNA content and the effect of synthetic phase (S-phase) fraction on hepatocytes and hepatomas, DNA content and S-phase fraction were measured by flow cytometry in human livers and hepatoma tissues. The ploidy status of nontumor parts of resected hepatoma, fetal liver, and focal nodular hyperplasia were diploid, similar to that of the normal liver. Three patterns of DNA ploidy in human hepatoma cells were newly classified, namely, pattern I, diploid tumors; pattern II, aneuploid tumors with single G0/G1 peak; and pattern III, aneuploid tumors with more than one G0/G1 peaks. Among the 130 resectable hepatomas measured for DNA ploidy status, 84 (64.6%) were pattern I, 20 (15.4%) pattern II, and 26 (20%) pattern III. Multivariate analyses for those 130 patients who underwent hepatic resection showed that, in addition to tumor size, DNA ploidy was another prognostic factor in predicting overall survival and disease-free survival. Patients with small tumors (less than 5 cm) had a significantly higher overall survival rate than those with large tumor (greater than 5 cm). Patients with pattern III hepatomas had a significantly lower overall survival rate and a higher recurrent rate than did those with pattern I or pattern II tumors. The S-phase fraction was a significant predictor of overall survival rate in patients with pattern II, but not with pattern I, tumors. We conclude that DNA flow-cytometric measurements of ploidy and S-phase fraction are potential important prognostic predictors in patients with resectable hepatomas.

Functional analysis of T-cell mutants defective in the biosynthesis of glycosylphosphatidylinositol anchor. Relative importance of glycosylphosphatidylinositol anchor versus N-linked glycosylation in T-cell activation

The glycosylphosphatidylinositol (GPI) anchor, potentially capable of generating a number of second messengers, such as diacylglycerol, phosphatidic acid, and inositol phosphate glycan, has been postulated to be involved in signal transduction in various cell types, including T-cells. We have identified a panel of T-cell hybridoma mutants that are defective at various steps of GPI anchor biosynthesis. Since they were derived from a functional T-T hybridoma, we were able to determine the precise role of the GPI anchor in T-cell activation. Two mutants were chosen for this analysis. The first mutant is defective at the first step of GPI anchor biosynthesis, i.e. in the transfer of N-acetylglucosamine to a phosphatidylinositol acceptor. Thus, it cannot form any GPI precursors or GPI-like compounds. Interestingly, this mutant can be activated by antigen, superantigen, and concanavalin A in a manner comparable to the wild-type hybridoma. These data strongly suggest that the GPI anchor, its precursor, or its potential cleavage product, inositol phosphate glycan, is not required for the early phase of T-cell activation. The second mutant is able to synthesize the first two GPI precursors, but is not able to add mannose residues to them due to a deficiency in dolichol-phosphate-mannose (Dol-P-Man) biosynthesis. Unexpectedly, all of the Dol-P-Man mutants are defective in activation by antigen, suprantigen, and concanavalin A despite normal T-cell receptor expression. Here, we show that the activation defect was due to a pleiotropic glycosylation abnormality because Dol-P-Man is required for both GPI anchor and N-linked oligosaccharide biosynthesis. When the yeast Dol-P-Man synthase gene was stably transfected into the mutants, full expression of surface GPI-anchored proteins was restored. However, N-linked glycosylation was either partially or completely corrected in different transfectants. Reconstitution of activation defects correlates well with the status of N-linked glycosylation, but not with the expression of GPI-anchored proteins. These results thus reveal an unexpected role of N-linked glycosylation in T-cell activation.

Identification of defects in glycosylphosphatidylinositol anchor biosynthesis in the Thy-1 expression mutants

A number of eukaryotic proteins are anchored to the membrane by glycosylphosphatidylinositol (GPI), of which the core structure is conserved from protozoan to mammalian cells. Here, we used a panel of thymoma mutants, which synthesize Thy-1 but cannot express it on the cell surface, to study the GPI biosynthetic pathway in mammalian cells. These mutants have been assigned into six complementation classes (A, B, C, E, F, H) by the technique of somatic cell hybridization. Using a combination of metabolic labeling and chemical/enzymatic tests, the biosynthetic defects were mapped to four different steps. Class A, C, and H mutants cannot transfer N-acetylglucosamine (GlcNAc) to a phosphatidylinositol acceptor, suggesting that the first step of GPI synthesis is regulated by at least three genes. The Class E mutant does not synthesize dolichol-phosphate-mannose, the donor for the first mannose residue transferred to the GPI core, and thus cannot form any mannose-containing GPI precursors. Class B and F mutants are defective in the addition of the third mannose residue or ethanolamine phosphate, respectively, to the elongating GPI core. Our findings have implications for the biosynthesis and attachment of the mammalian GPI anchor.

Miltirone, a central benzodiazepine receptor partial agonist from a Chinese medicinal herb Salvia miltiorrhiza

Ten diterpene quinones, which inhibited the binding of [3H]flunitrazepam to central benzodiazepine receptors with IC50s ranging from 0.3 to 36.2 microM, were isolated from the ethereal extract of the roots of Salvia miltiorrhiza. Among these natural products, miltirone has the highest potency (IC50 = 0.3 microM). It was orally active in an animal model used to predict clinical tranquilizing effects. Unlike diazepam, miltirone behaved as a partial agonist in the central benzodiazepine receptor binding and behavioural tests. Moreover, it produced no acute muscle relaxant effect and did not induce drug dependence and withdrawal reactions after chronic administration in mice.

Structure-activity relationship of miltirone, an active central benzodiazepine receptor ligand isolated from Salvia miltiorrhiza Bunge (Danshen)

Twenty one o-quinonoid-type compounds and one coumarin-type compound related to miltirone (1) have been synthesized with the aim to identify the key structural elements involved in miltirone's interaction with the central benzodiazepine receptor. On the basis of their inhibition of [3H]flunitrazepam binding to bovine cerebral cortex membranes, it is apparent that ring A of miltirone is essential for affinity. Although increasing the size of ring A from six-membered to seven- and eight-membered is well-tolerated, the introduction of polar hydroxyl groups greatly reduces binding affinity. The presence of 1,1-dimethyl groups on ring A is, however, not essential. On the other hand, the isopropyl group on ring C appears to be critical for binding as its removal decreases affinity by more than 30-fold. It can, however, be replaced with a methyl group with minimal reduction in affinity. Finally, linking ring A and B with a -CH2CH2- bridge results in analogue 89, which is 6 times more potent than miltirone at the central benzodiazepine receptor (IC50 = 0.05 microM).