Radioactive labelling and characterization of the products of activated mouse lymphocytes.

For chemical characterization of the products of activated lymphocytes a radioactive double-label technique was developed which allows one to distinguish those products synthesized either de novo or in increased amounts by the stimulated culture. Spleen cells from Balb/c mice were cultured in serum-free medium in the presence or absence of concanavalin A and simultaneously labelled with radioactive leucine. Optimal culture conditions were established by determining parameters such as cell density, mitogen concentration, and kinetics of protein synthesis following stimulation. Combined supernatants of stimulated and unstimulated cultures each labelled with either [3H]leucine or [14C]leucine were fractionated on Sephadex G-75. Materials derived from control or stimulated supernatants both yielded a qualitatively similar radiolabelled profile. The isotope ratio of stimulated to nonstimulated culture, however, showed a broad peak at KD 0--.35 (approx. mol. wt 75000-20000) which was further analyzed by isoelectric focusing. Pools of every two fractions were focused in polyacrylamide gels at pH 3.5-10. By determining the isotope ratio, the isoelectric point, and the KD (mol wt), it was possible to distinguish at least 24 molecules which had been produced only, or in greater degree, by the stimulated culture.     

Use of bovine parathyroid hormone, labelled specifically in the N-terminal region by [3H]methyl exchange, to compare hepatic and renal metabolism of this hormone in the rat

[3H]bPTH, prepared by [3H]methyl exchange, was administered intravenously to rats. Analysis of kidney and liver extracts on Sephadex G-50 showed some low mol.wt (peak IV) and some bound (peak Ia) material. A peak of intermediate mol.wt (peak II) was present in kidney but absent from liver. Kidneys that had been perfused with [3H]bPTH contained similar metabolites to those described, indicating that they are generated de novo. Analysis of peak II by reverse-phase h.p.l.c. showed a number of metabolites. The significance of these findings is discussed. This study demonstrates the value of [3H]methyl exchange labelling in investigating peptide metabolism.     

Absence of negative feedback by oxytocin on release from magnocellular neurones in conscious rats.

Peripheral administration of vasopressin (VP) was previously shown to exert a negative feedback influence on its own release and on the release of oxytocin (OT). In this study we examined the possible influence that OT has on the function of hypothalamic magnocellular neurones. Oxytocin was administered intraperitoneally and its effects on release from VP neurones and from OT neurones were determined as indexed by plasma concentrations of vasopressin-associated neurophysin ([VP-RNP]) and oxytocin-associated neurophysin ([OT-RNP]) under basal conditions and conditions of high plasma osmolality (Posm) induced by acute salt loading. Studies were performed on conscious, chronically instrumented Long-Evans rats. Oxytocin (1 nmol or 10 nmol) dissolved in 1 mL of 0.9% saline was administered intraperitoneally to animals 1 h before they received an intravenous infusion of hypertonic saline over 60 min at a rate designed to raise Posm by approximately 0.75 mosmol.min-1. Intraperitoneal injection of vehicle or 1 nmol of OT did not significantly alter [VP-RNP], [OT-RNP], or basal Posm. Administration of 10 nmol OT also had no effect on [VP-RNP] or [OT-RNP], but this dose of peptide significantly lowered basal Posm (299 +/- 2 to 290 +/- 2 mosmol/kg H2O, p less than 0.001). Both doses of OT did not significantly alter the responsiveness of VP neurones to hyperosmotic stimulation.(ABSTRACT TRUNCATED AT 250 WORDS)     

Intra-axonal transport and turnover of neurophysins in the rat. A proposal for a possible origin of the minor neurophysin component

1. Radioactivity associated with the three neurophysins in the neural lobe of the rat was determined at intervals up to 5 weeks after an intracisternal injection of [(35)S]cysteine. 2. The radioactivity associated with the two major neurophysins (one supposedly associated with vasopressin and the other with oxytocin) increased linearly for 12h after the injection and the ratio of the rates of increase in the two proteins was very similar to the ratio of vasopressin to oxytocin in the gland. 3. From 12h onwards the radioactivity associated with each major neurophysin declined exponentially but the half-life of the supposed oxytocin-neurophysin (13.3 days) was shorter than that for the supposed vasopressin-neurophysin (19.8 days). 4. The kinetics of labelling of the minor neurophysin was quite different from that of the two major ones. It became slowly labelled during 3-5 days after injection and the radioactivity hardly decreased during the following 4 weeks. 5. The data could support the hypothesis that the minor neurophysin is a metabolic product of oxytocin-neurophysin. The exponential rate of disappearance of radioactivity from oxytocin-neurophysin and the minor component taken together has a rate constant similar to that for vasopressin-neurophysin (e.g. half-life=18.9 days).     

Immuno-cytochemical demonstration of the inability of the homozygous Brattleboro rat to synthesize vasopressin and vasopressin-associated neurophysin

Summary Immuno-enzyme cytochemical investigations have shown that, (1) the hypothalamic supraoptic and paraventricular nuclei of the Brattleboro rat, as in the normal rat, contain separate neurons which produce oxytocin + neurophysin; (2) the hereditary inability of the Brattleboro rat to synthesize vasopressin and its associated neurophysin is due to a biochemical defect of separate neurophysin-vasopressin neurons in the supraoptic and the paraventricular nuclei. These observations strongly support the hypotheses that (1) vasopressin and its associated neurophysin are formed via a common precursor, and (2) the initial point of intracellular appearance of the hereditary defect in the Brattleboro rat lies in the synthesis of this precursor, which occurs on ribosomes.Moreover, observations have demonstrated that, in the Brattleboro rat, in addition to the hereditary inability of the hypothalamic magnocellular neurosecretory system to synthesize vasopressin, there also exists a similar hereditary defect in the hypothetical parvicellular suprachiasmatic-median eminence neurosecretory system.This paper is dedicated to Professor Dr. W. Bargmann, in honour of his 70th birthday.Presented in part at the meeting of the Belgian Society of Endocrinology May 17, 1975 (Vandesande et al., 1975d).     

Biosynthesis and axonal transport of rat neurohypophysial proteins and peptides

35S-cysteine injected adjacent to the supraoptic nucleus (SON) of the rat is rapidly incorporated into proteins. These 35S-cysteine-labeled proteins in the SON (1-24 h after injection) were separated by polyacrylamide gel electrophoresis, and the distribution of radioactive proteins on the gels was analyzed. 1 h after injection, about 73% of the radioactivity appeared in two peaks (both about 20,000 mol wt). With time, these peaks (putative precursors of neurophysin) decreased, as a 12,000 mol wt peak (containing two distinct neurophysins) increased in radioactivity. Both the 20,000- and 12,000-mol wt proteins are transported into the axonal (median eminence) and nerve terminal (posterior pituitary) regions of the rat hypothalamo-neurohypophysial system. Conversion of the larger precursor protein to the smaller neurophysin appears to occur, in large part, intra-axonally during axonal transport. Six distinct 35S-cysteine-labeled peptides (less than 2500 mol wt), in addition to arginine vasopressin and oxytocin, are also synthesized in the SON and transported to the posterior pituitary where they are released together with labeled neurophysin by potassium depolarization in the presence of extracellular calcium. These data provide support for the hypothesis that the neurohypophysial peptides (vasopressin and oxytocin) and neurophysins are derived from the post- translational clevage of protein precursors synthesized in the SON, and that the conversion process can occur in the neurosecretory granule during axonal transport.     

Purification of the 22000- and 20000-mol.wt. forms of human somatotropin and characterization of their binding to liver and mammary binding sites.

Quantitative data concerning the binding of 22000-mol.wt. human somatotropin and its 20000-mol.wt. variant are described using pregnant-rabbit liver and mammary-gland receptors. The purification and the complete chemical characterization of both human somatotropin and its 20000-mol.wt. variant is also presented. Contamination of the 20000-mol.wt.-variant preparation by 22000-mol.wt. hormone was found to be 0.5% by weight as measured in radioimmunoassay using monoclonal antibody. Labelling of human somatotropin and its 20000-mol.wt. variant using the Iodogen method is described as well as the characterization of the binding to pregnant-rabbit liver and mammary-gland receptor preparations. The maximum binding capacity of the 125I-labelled human somatotropin was between 50 and 60% to liver particulate receptor, whereas that of the 20000-mol.wt. variant was 30%. The specificity of binding of both forms to rabbit hepatic and mammary-gland receptor was found to be similar for both proteins in the same system. The affinity constants and capacity were respectively 0.7 X 10(10)M-1 and 815 fmol/mg of protein for human somatotropin and 0.6 X 10(10)M-1 and 1.250 fmol/mg of protein for the 20000-mol.wt. variant. These data suggest that both proteins behave as partial agonists to the receptors studied.     

Studies on the assembly of the rat lens capsule. Biosynthesis and partial characterization of the collagenous components.

1. Isolated rat lens capsules synthesized hydroxy[3H]proline-containing polypeptides when incubated with [3H]proline. 2. The collagenous polypeptides synthesized during a 2 h incubation were analyzed by both gel-filtration chromatography and sodium dodecyl sulphate/polyacrylamide-gel electrophoresis and shown to have an apparent mol.wt. of approx. 180,000. 3. No evidence was obtained for conversion of these polypeptides into a lower-molecular-weight species in experiments where capsules were labelled for 2 h and chased with non-radioactive proline for up to 22 h. However, a time-dependent incorporation of the 180,000-mol.wt. species into a larger collagenous component was observed and this could be prevented by the inclusion of beta-aminopropionitrile in the incubation medium. 4. The radioactive components synthesized by the capsules correspond to subunits of the intact lens capsule and the direct incorporation of the polypeptide of mol.wt. 180,000 into deoxycholate-insoluble basement membrane was demonstrated.     

Isolation and characterization of a neurophysin from ostrich neurohypophyses

A neurophysin has been isolated from ostrich neurohypophyses using acid acetone extraction, salt fractionation and Sephadex G-75 chromatography. The crude neurophysin eluting from the Sephadex G-75 column was subjected to a) reverse-phase HPLC followed by Sephadex G-75 chromatography, b) DEAE-Sephadex A-50 chromatography or c) isoelectric focusing. The different homogeneous ostrich neurophysin fractions so obtained were compared i.t.o. amino acid composition, spectral properties, N-terminal amino acid residues and PAGE. They all revealed a single N-terminal Ala residue and displayed spectral properties (A280/A260 less than 1) which are typical of mammalian neurophysin-like polypeptides. Ultracentrifugation studies on purified ostrich neurophysin over a range of concentrations revealed a reversible concentration dependent association behaviour characterized by the presence of dimeric complexes at higher concentrations. Partial sequencing from the N-terminus revealed the molecule to be VLDV-like. The purified molecule was also submitted to CNBr fragmentation.     

Purification and partial characterization of high-molecular-weight forms of ectopic calcitonin from a human bronchial carcinoma cell line.

Studies on the high-molecular-weight immunoreactive calcitonin produced ectopically in culture by an epidermoid bronchial carcinoma cell line are reported. In cell-exposed medium, the principal component has a molecular weight of 40000 and molecules of mol.wts. 13000 and 10000 also occur. Only a trace amount of material co-eluting with 35000-mol.wt. human calcitonin is detectable. None of the calcitonins show cross-reactivity with anti-corticotropin serum. The 40000-mol.wt. immunoreactive calcitonin is readily proteolysed to the 13000- and 10000-mol.wt. components, but the 10000-mol.wt. component behaves as a comparatively stable 'core' molecule. By using immunoprecipitation and high-pressure liquid chromatography (h.p.l.c.), it is possible to prepare radiochemically homogeneous 10000-mol.wt. immunoreactive calcitonin from cells grown in the presence of individual 35S- or 3H-labelled amino acids. Peptide mapping of enzymic digests of this material by h.p.l.c. shows that it contains peptides in common with synthetic human calcitonin.     

Immunocytochemical demonstration of separate vasopressin-neurophysin and oxytocin-neurophysin neurons in the human hypothalamus

Summary With the use of immunocytochemistry, it was shown that both the supraoptic and paraventricular hypothalamic nuclei in humans contain at least two different neurophysins. These two human neurophysins are immunologically related to bovine neurophysin I and neurophysin II, respectively. One human neurophysin is associated with vasopressin, the other with oxytocin. Human vasopressin-neurophysin and oxytocin-neurophysin are located separately in two different types of neurons, which correspond respectively to the vasopressinergic and oxytocinergic neurons of both the supraoptic and paraventricular nuclei. The neurophysin of the human vasopressinergic suprachiasmatic neurons appears to be closely related to or identical with neurophysin of the vasopressinergic neurons of the human magnocellular hypothalamic nuclei.This investigation was supported by a grant from the Belgian Nationaal Fonds voor Geneeskundig Wetenschappelijk Onderzoek     

Studies on the fatty acid binding proteins in cytosol of rat liver.

Gel filtration on Sephadex G-75 of crude rat liver supernatant preincubated with [1-14C]oleic acid yields three peaks of radioactivity which are attributed to the presence in these fractions of fatty acid binding proteins. We have confirmed these observations with binding assays by phase partition, polyacrylamide gel electrophoresis, and thin layer electrofocusing. Peak I (mol. wt. 60,000 pI 5.01 was shown to be albumin, which mainly arises from a contamination of the liver preparation by blood. Peak II (mol. wt. 10,000, pI 5.9) is a fatty acid binding protein. Finally peak III (mol. wt. 1500, pI 5.7) is a fatty acid binding component, the chemical nature of which was not elucidated. These fatty acid binding fractions have no effect on the reaction of acyl-CoA synthetase whereas the crude liver supernatant does stimulate the activation of fatty acid as shown earlier. In consequence, the physiological role of these fatty acid binding fractions is not yet elucidated.     

Binding of bile acids by 100 000g supernatants from rat liver.

1. The binding of glycocholic acid, chenodeoxycholic acid and lithocholic acid to rat liver 1000 000g supernatants was studied by equilibrium dialysis. 2. The binding characteristics of the bile acids suggest that the binding components are involved in bile acid transport. 3. When mixtures of [14C]lithocholic acid and liver supernatants were eluted from columns of Sephadex G-75, a prominent peak of [14C]lithocholic acid appeared with proteins of mol.wt. approx. 40000. A second, smaller, peak of [14C]lithocholic acid was eluted with proteins of mol.wt. approx. 100000. 4. The inclusion of cholic acid, glycocholic acid or chenodeoxycholic acid in the eluting buffer decreased the amount of [14C]lithocholic acid that was eluted with the higher-molecular-weight component.     

Electrophoretic and immunological charactorization of rat neurophysin

1. The electrophoretic properties of rat posterior pituitary proteins have been compared on starch gel with those of bovine and porcine neurophysins. 2. [(35)S]-Cysteine was injected into the supraoptic nucleus of male rats and 16-24h later the distribution of labelled neural-lobe protein in starch and polyacrylamide gels was determined. In both systems a single major protein component was found to contain more than 80% of the total recovered radioactivity. Between 5 and 10% of the radioactivity was found in a minor component in polyacrylamide gel. 3. In agar, microimmuno-diffusion and -electrophoresis of the rat neural-lobe proteins gave a single arc with neurophysin antiserum, and after starch-gel electrophoresis this arc was shown to be due to the major labelled component. 4. The molecular weights of the rat neural-lobe proteins were estimated by polyacrylamide-gel electrophoresis in the presence of sodium dodecyl sulphate. The molecular weight of the major labelled component was found to be 12000. 5. It is concluded that the rat neurophysin consists of one major and possibly one minor component.     

Immunocytochemical study of the hypothalamo-neurohypophysial system

Summary Antisera, with cross reactive antibodies removed by affinity chromatography, were used in the immunoperoxidase-bridge technique to study the distribution of oxytocin and vasopressin together with neurophysin in the hypothalamo-neurohypophysial system of the rat. The hormones were demonstrated in different areas of the supraoptic nucleus (SON) and paraventricular nucleus (PVN), in neurosecretory fibres of the hypothalamoneurohypophysial tract, median eminence, and in nerve terminals of the neurohypophysis. Intact normal and rats with hereditary hypothalamic diabetes insipidus (Brattleboro strain), and rats dehydrated by the administration of oral hypertonic saline were studied. In dehydrated rats the hormone concentration in the neurons, and the number of neurons containing hormone varied according to the time of dehydration stress.The observations support the hypotheses that: 1) oxytocin and oxytocinneurophysin, and vasopressin and vasopressin-neurophysin are synthesised in different neurons and are transported along different axons; 2) the SON and PVN are functionally indistinguishable in that neurons containing oxytocin or vasopressin are present in both nuclei; and 3) the two types of neurons respond to osmotic stimulation in a way that is qualitatively the same but quantitatively different.This work was supported by a grant from the Medical Research Council of New Zealand     

Characterization of estrogen binding proteins in mouse liver cytosol: absence of sexual dimorphism

Estrogen binding proteins in mouse liver cytosol were characterized by separation on Sephadex G-75 columns, by Scatchard plot analysis, and by hormonal competition studies. A high affinity receptor (56-70 fmol/mg cytosolic protein) with a mol. wt greater than 75,000, Kd of 5.7-8.4 X 10(-10) M was identified in male and female C3H liver. A second high capacity low affinity (HCLA) binder (200-300 fmol/mg cytosolic protein) with a mol. wt of about 50,000, Kd of 1.7-7.2 X 10(-8) was also identified. Following partial purification of the estrogen binders by ammonium sulfate precipitation, Scatchard plot analysis revealed selective removal of HCLA. On Sephadex G-75 filtration, the purification also resulted in selective removal of the 17 beta-estradiol binding component with a mol. wt of 50,000. Comparison with rat cytosol separations show that the sexual dimorphism in HCLA binding proteins (5 times higher in male than female rat liver) was absent in the mouse liver. These studies document the presence of a specific high affinity estrogen binding protein in mouse liver and indicate that the sexual dimorphism in HCLA proteins is not a universal feature of all rodent species.     

Conversion of biosynthetic human proinsulin to partially cleaved intermediates by collagenase proteinases adsorbed to isolated rat adipocytes.

Studies of the biological activity of proinsulin have resulted in widely varying conclusions. Relative to insulin, the biological activity of proinsulin has been reported from less than 1% to almost 20%. Many of the assays in vitro for the biological potency of proinsulin have utilized isolated rat adipocytes. To examine further the interaction of proinsulin with rat adipocytes, we prepared specifically-labelled proinsulin isomers that were iodinated on tyrosine residues corresponding to the A14, A19, B16 or B26 residue of insulin. These were incubated with rat adipocytes and their metabolism was examined by trichloroacetic acid precipitation, by Sephadex G-50 chromatography, and by h.p.l.c. chromatography. By trichloroacetic acid-precipitation assay, there was little or no proinsulin degradation. By G-50 chromatography and subsequent h.p.l.c. analysis, however, we found that the labelled proinsulin isomers were converted rapidly and almost completely to materials which eluted differently on h.p.l.c. from intact proinsulin. This conversion was due primarily to proteolytic activity which adsorbed to the fat cells from the crude collagenase used to isolate the cells. Two primary conversion intermediates were found: one with a cleavage at residues 23-24 of proinsulin (the B-chain region of insulin), and one at residues 55-56 in the connecting peptide region. These intermediates had receptor binding properties equivalent to or less than intact proinsulin. These findings show that isolated fat cells can degrade proinsulin to intermediates due to their contamination with proteolytic activity from the collagenase used in their preparation. Thus the previously reported range in biological activities of proinsulin in fat cells may have arisen from such protease contamination. Finally, the present findings demonstrate that a sensitive assay for degradation of hormones is required to examine biological activities in isolated cells.     

Isolation of a tripeptide showing weak acetylthiocholine hydrolysing activity from a soluble form of monkey basal ganglia acetylcholinesterase by limited trypsin digestion

Acetylcholinesterase was purified from the soluble supernatant of monkey (Macaca radiata) brain basal ganglia by a three-step affinity purification procedure. The purified enzyme showed two major protein bands corresponding to molecular weights of approximately 65 kDa and approximately 58 kDa which could be labelled by [3H]diisopropylfluorophosphate. When the purified enzyme was subjected to limited trypsin digestion followed by gel filtration on Sephadex G-75 or Sephadex G-25 column, a peptide fragment of molecular weight approximately 300 Da having a weak acetylthiocholine hydrolysing activity was isolated. The amino acid sequence analysis of this peptide showed a sequence of Gly-Pro-Ser. When the [3H]DFP labelled enzyme was subjected to limited trypsin digestion and Sephadex G-75 column chromatography, a labelled peptide corresponding to approximately 430 Da was isolated. The kinetics, inhibition characteristics and binding characteristics to lectins of this peptide were compared with the parent enzyme. A synthetic peptide of sequence Gly-Pro-Ser was also found to exhibit acetylthiocholine hydrolysing activity. The kinetics and inhibition characteristics of the synthetic peptide were similar to those of the peptide derived from the purified acetylcholinesterase, except that the synthetic peptide was more specific towards acetylthiocholine than butyrylthiocholine. The specific activity (units/mg) of the synthetic peptide was about 123700 times less than that of the purified AChE.     

Single-step isolation and sequencing of vasopressin and oxytocin precursors.

The pre- and post-Golgi processing of preprovasopressin and prepro-oxytocin was evaluated by microsequencing for incorporated radiolabel. 35S-Cysteine and 3H-fucose were microinjected into rat supraoptic nuclei (SON), and proteins and peptides related to the biosynthesis of vasopressin (VP) and oxytocin (OT) were isolated at various times from the supraoptic nuclei and neural lobe by employing a one-step procedure of high performance liquid chromatography (HPLC). These proteins and peptides were recognized through their binding to specific antibodies against VP, OT, and rat neurophysins (RNPs), and by their binding to ConA-Sepharose. Two immunoreactive glycoproteins related to VP biosynthesis were recovered from the SON and both contained fucose and had a 35S-cysteine placement consistent with the location of the hormone sequence at the N-terminus. SDS-electrophoresis revealed the major protein form to be 21,000 daltons and the minor protein form to be 19,000 daltons. One nonglycosylated protein of 16,000 daltons related to oxytocin biosynthesis was recovered from the SON, and this protein also had a 35S-cysteine placement consistent with an N-terminal OT sequence. These data provide the first sequential evidence that prior to, or shortly after, packaging in the Golgi the preprohormones of VP and OT have lost their entire leader-peptide structures.     

Comparison of the substrate specificities of protein phosphatases involved in the regulation of glycogen metabolism in rabbit skeletal muscle.

Muscle extracts were subjected to fractionation with ethanol, chromatography on DEAE-cellulose, precipitation with (NH4)2SO4 and gel filtration on Sephadex G-200. These fractions were assayed for protein phosphatase activities by using the following seven phosphoprotein substrates: phosphorylase a, glycogen synthase b1, glycogen synthase b2, phosphorylase kinase (phosphorylated in either the alpha-subunit or the beta-subunit), histone H1 and histone H2B. Three protein phosphatases with distinctive specificities were resolved by the final gel-filtration step and were termed I, II and III. Protein phosphatase-I, apparent mol.wt. 300000, was an active histone phosphatase, but it accounted for only 10-15% of the glycogen synthase phosphatase-1 and glycogen synthase phosphatase-2 activities and 2-3% of the phosphorylase kinase phosphatase and phosphorylase phosphatase activity recovered from the Sephadex G-200 column. Protein phosphatase-II, apparent mol.wt. 170000, possessed histone phosphatase activity similar to that of protein phosphatase-I. It possessed more than 95% of the activity towards the alpha-subunit of phosphorylase kinase that was recovered from Sephadex G-200. It accounted for 10-15% of the glycogen synthase phosphatase-1 and glycogen synthase phosphatase-2 activity, but less than 5% of the activity against the beta-subunit of phosphorylase kinase and 1-2% of the phosphorylase phosphatase activity recovered from Sephadex G-200. Protein phosphatase-III was the most active histone phosphatase. It possessed 95% of the phosphorylase phosphatase and beta-phosphorylase kinase phosphatase activities, and 75% of the glycogen synthase phosphatase-1 and glycogen synthase phosphatase-2 activities recovered from Sephadex G-200. It accounted for less than 5% of the alpha-phosphorylase kinase phosphatase activity. Protein phosphatase-III was sometimes eluted from Sephadex-G-200 as a species of apparent mol.wt. 75000(termed IIIA), sometimes as a species of mol.wt. 46000(termed IIIB) and sometimes as a mixture of both components. The substrate specificities of protein phosphatases-IIA and -IIB were identical. These findings, taken with the observation that phosphorylase phosphatase, beta-phosphorylase kinase phosphatase, glycogen synthase phosphatase-1 and glycogen synthase phosphatase-2 activities co-purified up to the Sephadex G-200 step, suggest that a single protein phosphatase (protein phosphatase-III) catalyses each of the dephosphorylation reactions that inhibit glycogenolysis or stimulate glycogen synthesis. This contention is further supported by results presented in the following paper [Cohen, P., Nimmo, G.A. & Antoniw, J.F. (1977) Biochem. J. 1628 435-444] which describes a heat-stable protein that is a specific inhibitor of protein phosphatase-III.