Learning Gaussian-Bernoulli RBMs Using Difference of Convex Functions Optimization

The Gaussian-Bernoulli restricted Boltzmann machine (GB-RBM) is a useful generative model that captures meaningful features from the given n -dimensional continuous data. The difficulties associated with learning GB-RBM are reported extensively in earlier studies. They indicate that the training of the GB-RBM using the current standard algorithms, namely contrastive divergence (CD) and persistent contrastive divergence (PCD), needs a carefully chosen small learning rate to avoid divergence which, in turn, results in slow learning. In this work, we alleviate such difficulties by showing that the negative log-likelihood for a GB-RBM can be expressed as a difference of convex functions if we keep the variance of the conditional distribution of visible units (given hidden unit states) and the biases of the visible units, constant. Using this, we propose a stochastic difference of convex (DC) functions programming (S-DCP) algorithm for learning the GB-RBM. We present extensive empirical studies on several benchmark data sets to validate the performance of this S-DCP algorithm. It is seen that S-DCP is better than the CD and PCD algorithms in terms of speed of learning and the quality of the generative model learned.

On Robustness of Neural Architecture Search Under Label Noise

Neural architecture search (NAS), which aims at automatically seeking proper neural architectures given a specific task, has attracted extensive attention recently in supervised learning applications. In most real-world situations, the class labels provided in the training data would be noisy due to many reasons, such as subjective judgments, inadequate information, and random human errors. Existing work has demonstrated the adverse effects of label noise on the learning of weights of neural networks. These effects could become more critical in NAS since the architectures are not only trained with noisy labels but are also compared based on their performances on noisy validation sets. In this paper, we systematically explore the robustness of NAS under label noise. We show that label noise in the training and/or validation data can lead to various degrees of performance variations. Through empirical experiments, using robust loss functions can mitigate the performance degradation under symmetric label noise as well as under a simple model of class conditional label noise. We also provide a theoretical justification for this. Both empirical and theoretical results provide a strong argument in favor of employing the robust loss function in NAS under high-level noise.

Geometric decision tree

In this paper, we present a new algorithm for learning oblique decision trees. Most of the current decision tree algorithms rely on impurity measures to assess the goodness of hyperplanes at each node while learning a decision tree in top-down fashion. These impurity measures do not properly capture the geometric structures in the data. Motivated by this, our algorithm uses a strategy for assessing the hyperplanes in such a way that the geometric structure in the data is taken into account. At each node of the decision tree, we find the clustering hyperplanes for both the classes and use their angle bisectors as the split rule at that node. We show through empirical studies that this idea leads to small decision trees and better performance. We also present some analysis to show that the angle bisectors of clustering hyperplanes that we use as the split rules at each node are solutions of an interesting optimization problem and hence argue that this is a principled method of learning a decision tree.

Conditional probability-based significance tests for sequential patterns in multineuronal spike trains

We consider the problem of detecting statistically significant sequential patterns in multineuronal spike trains. These patterns are characterized by ordered sequences of spikes from different neurons with specific delays between spikes. We have previously proposed a data-mining scheme to efficiently discover such patterns, which occur often enough in the data. Here we propose a method to determine the statistical significance of such repeating patterns. The novelty of our approach is that we use a compound null hypothesis that not only includes models of independent neurons but also models where neurons have weak dependencies. The strength of interaction among the neurons is represented in terms of certain pair-wise conditional probabilities. We specify our null hypothesis by putting an upper bound on all such conditional probabilities. We construct a probabilistic model that captures the counting process and use this to derive a test of significance for rejecting such a compound null hypothesis. The structure of our null hypothesis also allows us to rank-order different significant patterns. We illustrate the effectiveness of our approach using spike trains generated with a simulator.

Statistical significance of sequential firing patterns in multi-neuronal spike trains

Sequential firings with fixed time delays are frequently observed in simultaneous recordings from multiple neurons. Such temporal patterns are potentially indicative of underlying microcircuits and it is important to know when a repeatedly occurring pattern is statistically significant. These sequences are typically identified through correlation counts. In this paper we present a method for assessing the significance of such correlations. We specify the null hypothesis in terms of a bound on the conditional probabilities that characterize the influence of one neuron on another. This method of testing significance is more general than the currently available methods since under our null hypothesis we do not assume that the spiking processes of different neurons are independent. The structure of our null hypothesis also allows us to rank order the detected patterns. We demonstrate our method on simulated spike trains.

Fingerprint Classification Using a Feedback-Based Line Detector

We present a fingerprint classification algorithm in this paper. This algorithm classifies a fingerprint image into one of the five classes: Arch, Left loop, Right loop, Whorl, and Tented arch. We use a new low-dimensional feature vector obtained from the output of a novel oriented line detector presented here. Our line detector is a co-operative dynamical system that gives oriented lines and preserves multiple orientations at points where differently oriented lines meet. Our feature extraction process is based on characterizing the distribution of orientations around the fingerprint. We discuss three different classifiers: support vector machines, nearest-neighbor classifier, and neural network classifier. We present results obtained on a National Institute of Standards and Technology (NIST) fingerprint database and compare with other published results on NIST databases. All our classifiers perform equally well, and this suggests that our novel line detection and feature extraction process indeed captures all the crucial information needed for classification in this problem.

Two Timescale Analysis of the Alopex Algorithm for Optimization

Alopex is a correlation-based gradient-free optimization technique useful in many learning problems. However, there are no analytical results on the asymptotic behavior of this algorithm. This article presents a new version of Alopex that can be analyzed using techniques of two timescale stochastic approximation method. It is shown that the algorithm asymptotically behaves like a gradient-descent method, though it does not need (or estimate) any gradient information. It is also shown, through simulations, that the algorithm is quite effective.

Ex vivo expanded allogeneic cord blood progenitor cell transplantation with a non-myeloablative conditioning regimen would cure autoimmune diseases

Loss of self-tolerance is the basis for autoimmune disease. Autologous as well as allogeneic stem cell transplantation is increasingly being considered and used for therapy of autoimmune disease. Non-myeloablative conditioning has been shown in the context of malignant diseases to be feasible. Cord blood transplantation has been associated with significantly less graft-versus-host disease (GVHD) than other forms of allotransplantation. Ex-vivo expansion of this limited resource is increasingly becoming feasible. It is being hypothesized that ex-vivo expanded cord blood progenitor cell transplantation using a non-myeloablative conditioning regimen would be curative in autoimmune diseases.

Transmissible spongiform encephalopathies and the 'rogue share-holder protein' hypothesis

Transmissible spongiform encephalopathies(TSEs) or prion diseases are a closely related group of diseases, whose exact etiology is unknown, but is generally accepted to be related to protease-resistant prion protein PrP. PrPc is normally present in cells and its disease counterpart PrPsc is postulated to occur due to a rare stochastic change. The selfish gene hypothesis is a generally well accepted concept in evolutionary biology. Genes can be likened to the board of a company and proteins can be likened to share-holders. Here it is being hypothesized that a rogue share-holder protein's 'selfish' replicatory tendency might be the explanation for TSEs. The present hypothesis predicts existence of other examples of rogue share-holder protein and also predicts that examples would be found in lower life-forms as well.

Apoptosis and the nervous system

Apoptosis is now recognized as a normal feature in the development of the nervous system and may also play a role in neurodegenerative diseases and aging. This phenomenon has been investigated intensively during the last 6-7 years, and the progress made in this field is reviewed here. Besides a few in vivo studies, a variety of neuronal preparations from various parts of the brain, the majority of which were primary cultures, and some cell lines have been investigated. Several apoptosis-inducing agents have been identified, and these include lack of neurotrophic support, neurotransmitters, neurotoxicants, modulators of protein phosphorylation and calcium homeostasis, DNA-damaging agents, oxidative stress, nitric oxide, and ceramides. The precise signaling cascade is not well established, and there are lacunae in many suggested pathways. However, it appears certain that the Bcl family of proteins is involved in the apoptotic pathway, and these proteins in turn affect the processing of interleukin-1beta converting enzyme (ICE)/caspases. The available evidence suggests that there may be several apoptotic pathways that may depend on the cell type and the inducing agent, and most of the pathways may converge at the ICE/caspases step.

Prevention of progression in chronic myeloid leukemia by altering DNA methylation with a pyridoxine analogue

Altered DNA methylation is one mechanism by which progression might occur in tumors. Progression in chronic myeloid leukemia (CML) was initially thought to be due to mutations. However, altered DNA methylation might be responsible for the same. An agent whose probable target is DNA methylation has been shown to improve hematopoiesis in CML. Pyridoxine is involved in methylation reactions. At present, the link between dietary factors like pyridoxine and cancer is somewhat tenuous. It is hypothesized that a suitable pyridoxine analogue might, by blocking DNA methylation, prevent progression in CML.

Inhibition of TNF-alpha synthesis with thalidomide for prevention of acute exacerbations and altering the natural history of multiple sclerosis

Multiple sclerosis (MS) is a common neurological disorder which has a relapsing/remitting course and is presently incurable. A variety of agents have been tried to prevent excerbations and alter the natural history of the disease. Tumor necrosis factor-alpha (TNF-alpha) has been implicated as the most important cytokine in the pathogenesis of MS. There is evidence that thalidomide is an agent which blocks production of TNF-alpha by a mechanism different from other agents. Hence it is hypothesized that using thalidomide as therapy would prevent acute exacerbations of MS as well as alter its natural history.

Adoptive immunotherapy for cutaneous T-cell lymphomas

The pathology of cutaneous T-cell lymphomas is predominantly in the skin. Allogeneic bone-marrow transplantation results in graft-versus-host disease involving the skin in many cases. It has been suggested that it would be interesting to see the effect of allotransplantation, even though no series has been reported so far. Adoptive immunotherapy with donor lymphocyte infusions has been increasingly used to treat post-transplant relapse and even prevention. The present hypothesis suggests that infusion of human leucocyte antigen (HLA)-matched donor lymphocyte infusions would result in a graft-versus-leukemia/lymphoma (GVL) effect in these disorders. Therefore, clinical trials are suggested.

Role of feedback in mammalian vision: a new hypothesis and a computational model

This paper presents a novel hypothesis on the function of massive feedback pathways in mammalian visual systems. We propose that the cortical feature detectors compete not for the right to represent the output at a point, but for exclusive rights to abstract and represent part of the underlying input. Feedback can do this very naturally. A computational model that implements the above idea for the problem of detection is presented and based on that we suggest a functional role for the thalamo-cortical loop during perception of lines. We show that the model successfully tackles the so called Cross problem. Based on some recent experimental results, we discuss the biological plausibility of our model. We also comment on the relevance of our hypothesis (on the role of feedback) to general sensory information processing and recognition.

Gonadal function following ABVD therapy for Hodgkin's disease

To assess the effect of combination chemotherapy with doxorubicin, bleomycin, viablastine, and decarbazine (ABVD) on gonadal function in patients treated for Hodgkin's disease, we assessed 38 male patients with Hodgkin's disease who were > 15 years of age and in complete remission for the development of secondary sexual characteristics, sexual habits, and fatherhood after treatment. Semen analysis and serum hormone level estimation of follicle-stimulating hormone (FSH), leutinising hormone (LH), and testosterone (T) were done in all cases. Twenty-six patients received ABVD therapy and 12 received a combination of ABVD with COPP or MOPP (cyclophosphamide or nitrogen mustard, vincristine, procarbazine, and prednisone). Radiation of the pelvic region was done in one case. Median time between completion of therapy and assessment of gonadal function was 34 months (range, 12-68 months). Secondary sexual characteristics developed normally in all patients. Azoospermia was seen in one patient from the ABVD group and 10 patients from the COPP/ABVD group (p < 0.001). Serum FSH levels were significantly higher in the COPP/ABVD group than in the ABVD group (23.5 versus 4.7 mlu/ml; p < 0.001) The levels were in the normal range in 23 patients from the ABVD group, as compared to four in the COPP/ABVD group (88.5% versus 33.3%; p < 0.001). Three patients treated with ABVD fathered children post-therapy. We conclude that ABVD is associated with relatively better preservation of gonadal function.

Antilymphocyte globulin (ALG) or antithymocyte globulin (ATG) with methylprednisone and oxymethalone in aplastic anaemia

From 1986 to 1994 we treated 26 patients of aplastic anaemia between 6 to 61 years age group with ATG/ALG, Methylprednisone and Oxymethalone. Five had very severe aplastic anaemia, 16 had severe and 5 nonsevere disease. Disease was associated with hepatitis in 5 patients and with pregnancy and drug use in 2 patients each. In others no cause could be ascertained. A total of 31 courses of treatment were given (range 1-3 courses per patient). Nine patients had complete response (34.62%) and 3 had partial response (11.54%) with an overall response rate of 46.16%. Four patients died within 2 months of starting the treatment. The median follow up was 24 months (range 6-102 months) with an overall survival probality of 45% at 2 yr. At the time of evaluation 12 patients have died, 9 are alive disease-free and 5 are alive with disease. The side effects associated with therapy were tolerable and did not require cessation of therapy in any patient. We conclude that ATG/ALG with Methylprednisone and Oxymethalone is beneficial to significant number of patients with aplastic anaemia.

Serum tumor necrosis factor for monitoring response of hepatic veno-occlusive disease to pentoxiphyllin--a case report

Hepatic veno-occlusive disease (VOD) is the second most common cause of death after autologous bone marrow transplantation (ABMT). A patient with multiple myeloma undergoing ABMT developed classic features of hepatic VOD. He responded to treatment with pentoxiphyllin. Serum tumor necrosis factor (TNF) levels showed remarkable correlation with the severity of VOD and response to therapy.

Pulmonary microsporidial infection in a patient with CML undergoing allogeneic marrow transplant

Very few cases of human microsporidial infection have been reported. The advent of AIDS has changed this. There is increasing recognition that microsporidia are important opportunistic pathogens. However, the number of cases reported in the non-HIV population is small. We report here a case of microsporidial infection in a female patient with chronic myeloid leukemia undergoing allogeneic bone marrow transplantation. There was also an associated fungal infection. The diagnosis could be reached only after postmortem and was confirmed by electron micrography. We suggest that transplant patients are another group of patients who are susceptible to this group of opportunistic pathogens.

Rapid cytoreduction in childhood leukemic hyperleukocytosis by conservative therapy

Childhood leukemic hyperleukocytosis poses a serious threat to life because of its associated metabolic complications. The present prospective trial utilized conservative management of childhood acute lymphoblastic leukemia with hyperleukocytosis (total white cell count equal or > 100 x 10(9)/L) by intravenous hydration, urinary alkalinization, and allopurinol presenting without severe life-threatening complications. The median reduction in WBC count was 81.51% (range: 66-98.8%) within a median period of 36 hours (range: 12-60 hours) following hospitalization. There were no failures or treatment related complications. Thus we conclude that in childhood acute lymphoblastic leukemia, hyperleukocytosis can be managed safely and effectively with intravenous hydration, urinary alkalinization, and allopurinol before starting any specific anti-leukemic chemotherapy avoiding risk-associated cranial irradiation, leukapheresis, and exchange transfusion.

A soluble preparation from developing groundnut seeds (Arachis hypogaea) catalyzes de novo synthesis of long chain fatty acids

A 100,000 x g supernatant fraction prepared from developing groundnut seeds (30-35 days after flowering) catalyzed the synthesis of fatty acids from [l-14C]acetate at a rate of 120nmoles of acetate incorporated per hr per gram fresh weight of tissue. 90% of this incorporated label was associated with fatty acids. The major fatty acids formed were stearic- (77%) and palmitic acids (14%) with 4% of oleic acid. The fatty acid synthetase activity was stable when stored at 0-4 degrees C for at least fifteen days. It is concluded from these results that acetyl-coA carboxylase and all the enzymes of fatty acid synthetase from developing groundnut seeds are soluble.

Purification and partial characterization of acyl carrier proteins from developing oil seeds of pisa (Actinodaphne hookeri) and ground nut (Arachis hypogaea)

Acyl carrier proteins (ACP) were purified to homogeneity in the active form from developing seeds of pisa (Actinodaphne hookeri) which synthesizes exclusively trilaurin and from ground nut (Arachis hypogaea) which synthesizes triacylglycerols containing long chain fatty acids. Two major isoforms of ACPs were purified from developing pisa seeds using DEAE-cellulose, Superose-6 FPLC and C4 reversed phase HPLC chromatographic methods. In contrast, only a single form of ACP was present in ground nut seeds which was purified by anion-exchange and activated thiol-Sepharose 4B affinity chromatography. The two isoforms of ACPs from pisa showed nearly the same specific activity of 6,706 and 7,175 pmol per min per mg protein while ground nut ACP showed a specific activity of 3,893 pmol per min per mg protein when assayed using E. coli acyl-ACP synthetase and [1-14C]palmitic acid. When compared with E. coli ACP, the purified ACPs from both the seeds showed considerable difference in their mobility in native PAGE, but showed similar mobility in SDS-PAGE under reducing conditions. In the absence of reducing agents formation of dimers was quite prominent. The ACPs from both the seed sources were acid- and heat-stable. The major isoform of pisa seed ACP and the ground nut ACP contain 91 amino acids with M(r) 11,616 and 1,228 respectively. However, there is significant variation in their amino acid composition. A comparison of the amino acid sequence in the N-terminal region of pisa and ground nut seed ACPs showed considerable homology between themselves and with other plant ACPs but not with E. coli ACP.

Diacylglycerol kinase is stimulated by arachidonic acid in neural membranes

The effect of arachidonic acid (AA) on the activity of diacylglycerol (DG) kinase in neural membranes was investigated. When rat brain cortical membranes were incubated with 0.5 mM dipalmitin and [gamma-32P]ATP, formation of phosphatidic acid (PA) was observed. It was linear up to 5 min, and the initial rate was approximately 1.0 nmol/min/mg of protein. The DG kinase activity was stimulated twofold by 0.25 mM AA. The stimulation was apparent at the earliest time point measured (1 min) and with the lowest concentration of AA tested (62.5 microM). The stimulation was proportional to the concentration of AA up to 250 microM. AA was the most potent stimulator of DG kinase, and linolenic acid showed approximately 40% stimulation. Oleic acid showed no effect, whereas linoleic and the saturated fatty acids tested were inhibitory. AA stimulation of DG kinase was observed only with membranes of cerebrum, cerebellum, and myelin and not with brain cytosol or liver membranes. AA also stimulated the formation of PA in the absence of added dipalmitin (endogenous activity) with membranes prepared from whole brain. DG kinase of neural membranes was extracted with 2 M NaCl, which on dialysis yielded a precipitate. Both the precipitate and the supernatant showed DG kinase activity, but only the enzyme in the precipitate was stimulated by AA at concentrations as low as 25 microM. It is suggested that AA, through its effect on DG kinase, regulates the level of DG in neural membranes, which in turn regulates protein kinase C activity.

Regulation of diacylglycerol kinase in rat brain membranes by docosahexaenoic acid

The effect of docosahexaenoic acid (DHA) on the diacylglycerol kinase (DG kinase) activity in rat brain membranes was investigated. DHA at 500 microM concentration, stimulated the enzyme activity by about 2 fold. This effect was concentration- and time-dependent and was observed after very short periods of incubation (one min). DHA stimulation of DG kinase was observed only with rat brain membranes, and not with rat brain cytosol or rat liver membranes. Treating the rat brain membranes with phospholipase A2 which released free fatty acids including DHA, significantly stimulated the DG kinase activity. It is concluded that DHA through its stimulatory effect on DG kinase may regulate the signalling events in growth-related situations in the brain such as synaptogenesis.

Topology of Wolfgram proteins and 2', 3'-cyclic nucleotide 3'-phosphodiesterase in CNS myelin: studies with proteases

The topological disposition of Wolfgram proteins (WP) and their relationship with 2', 3'-cyclic nucleotide 3'-phosphodiesterase (CNPase) in human, rat, sheep, bovine, guinea pig and chicken CNS myelin was investigated. Controlled digestion of myelin with trypsin gave a 35KDa protein band (WP-t) when electrophoresed on dodecyl sulfate-polyacrylamide gel in all species. Western blot analysis showed that the WP-t was derived from WP. WP-t was also formed when rat myelin was treated with other proteases such as kallikrein, thermolysin and leucine aminopeptidase. Staining for CNPase activity on nitrocellulose blots showed that WP-t is enzymatically active. Much of the CNPase activity remained with the membrane fraction even after treatment with high concentrations of trypsin when WP were completely hydrolysed and no protein bands with M.W > 14KDa were detected on the gels. Therefore protein fragments of WP with M.W < 14KDa may contain CNPase activity. From these results, it is suggested that the topological disposition of all the various WP is such that a 35KDa fragment is embedded in the lipid bilayer and the remaining fragment exposed at the intraperiod line in the myelin structure which may play a role in the initiation of myelinogenesis.

Synthesis of trilaurin by developing pisa seeds (Actinodaphne hookeri)

The developing seeds of Actinodaphne hookeri were investigated to delineate their ability to synthesize large amounts of trilaurin. Until 88 days after flowering the embryos contained 71% neutral lipids (NL) and 29% phospholipids (PL) and both these components contained C16:0, C18:0, C18:2, and C18:3 as the major fatty acids (FA). At 102 days after flowering the seeds began to accumulate triacylglycerols (TAG) and to synthesize lauric acid (C12:0). By 165 days after flowering, when the seeds were mature, they contained about 99% NL and 1% PL. At this stage the TAG contained exclusively C12:0, while the PL consisted of long-chain fatty acids (LCFA) only. Leaf lipids in contrast did not contain any C12:0. Experiments on [1-14C]acetate incorporation into developing seed slices showed that at 88 days after flowering only 4% of the label was in TAG, 1% in diacylglycerols (DAG), and 87% in PL. One hundred two days after flowering seeds incorporated only 2% of the label into TAG, 30% into DAG, and 64% into PL. In contrast at 114 days after flowering 71% of the label was incorporated into TAG, 25% into DAG, and only 2% into PL. Analysis of labeled FA revealed that up to 102 days after flowering it was incorporated only into LCFA, whereas at 114 days after flowering it was incorporated exclusively into C12:0. Furthermore, 67% of the label in PL at 114 days after flowering was found to be dilaurylglycerophosphate. Analysis of the label in DAG at this stage showed that it was essentially in dilaurin species. These observations indicate the induction of enzymes of Kennedy pathway for the specific synthesis of trilaurin at about 114 days after flowering. Homogenates of seeds (114 days after flowering) incubated with labeled FA in the presence of glycerol-3-phosphate and coenzymes A and ATP incorporated 84% of C12:0 and 61% of C14:0, but not C16:0, C18:2, and C18:3, into TAG. In contrast the LCFA were incorporated preferentially into PL. It is concluded that, between 102 and 114 days after flowering, a switch occurs in A. hookeri for the synthesis of C12:0 and trilaurin which is tissue specific. Since the seed synthesizes exclusively C12:0 at 114 days after flowering onwards and incorporates specifically into TAG, this system appears to be ideal for identifying the enzymes responsible for medium-chain fatty acid as well as trilaurin synthesis and for exploiting them for genetic engineering.

Regulation of the activity of glyceraldehyde 3-phosphate dehydrogenase by glutathione and H2O2

The activity lost during storage of a solution of muscle glyceraldehyde 3-phosphate dehydrogenase was rapidly restored on adding a thiol compound, but not arsenite or azide. On treatment with H2O2, the enzyme was partially inactivated and complete loss of activity occurred in the presence of glutathione. Samples of the enzyme pretreated with glutathione followed by removal of the thiol compound by filtration on a Sephadex column showed both full activity and its complete loss on adding H2O2, in the absence of added glutathione. Most of the activity was restored when the H2O2-inactivated enzyme was incubated with glutathione (25 mM) or dithiothreitol (5 mM) whereas arsenite or azide were partly effective and ascorbate was ineffective. The need for incubation for a long time with a strong reducing agent for restoration of activity suggests that the oxidized group (disulfide or sulfenate) must be in a masked state in the H2O2-inactivated enzyme. Analysis by SDS-PAGE gave evidence for the formation of a small quantity of glutathione-reversible disulfide-form of the enzyme. Circular dichroic spectra indicated a decrease in alpha-helical content in the inactivated form of the enzyme. The evidence suggest that glutathione and H2O2 can regulate the active state of this enzyme.

Inverse relationship of the dehydrogenase and ADP-ribosylation activities in sodium-nitroprusside-treated glyceraldehyde-3-phosphate dehydrogenase is coincidental

Incubation of glyceraldehyde-3-phosphate dehydrogenase (GAPD) with sodium nitroprusside (SNP) decreased its activity in concentration- and time-dependent fashion in the presence of a thiol compound, with DTT being more effective than GSH. Both forward and backward reactions were effected. Coinciding with this, HgCl2-sensitive labelling of the protein by [32P]NAD+ also increased, indicating the stimulation of ADP-ribosylation. Treatment with SNP of GAPD samples from rabbit muscle, sheep brain and yeast inactivated the dehydrogenase activity of the three, but only the mammalian proteins showed ADP-ribosylation activity. The SNP-modified protein of rabbit muscle GAPD, freed from the reagent by Sephadex filtration showed a concentration-dependent restoration of the dehydrogenase activity on preincubation with DTT and GSH. Such thiol-treated preparations also gave increased ADP-ribosylation activity with DTT, and to a lesser extent with GSH. The SNP-modified protein was unable to catalyze this activity with the native yeast enzyme and native and heat-inactivated muscle enzyme. It was possible to generate the ADP-ribosylation activity in muscle GAPD, by an NO-independent mechanism, on dialysis in Tris buffer under aerobic conditions, and on incubating with NADPH, but not NADH, in muscle and brain, but not yeast, enzymes. The results suggest that the inverse relationship of the dehydrogenase and ADP-ribosylation activities is coincidental but not correlated.

Dopamine receptors in human foetal brains: characterization, regulation and ontogeny of [3H]spiperone binding sites in striatum

Eighteen corpora striata from normal human foetal brains ranging in gestational age from 16 to 40 weeks and five from post natal brains ranging from 23 days to 42 years were analysed for the ontogeny of dopamine receptors using [3H]spiperone as the ligand and 10 mM dopamine hydrochloride was used in blanks. Spiperone binding sites were characterized in a 40-week-old foetal brain to be dopamine receptors by the following criteria: (1) It was localized in a crude mitochondrial pellet that included synaptosomes; (2) binding was saturable at 0.8 nM concentration; (3) dopaminergic antagonists spiperone, haloperidol, pimozide, trifluperazine and chlorpromazine competed for the binding with IC50 values in the range of 0.3-14 nM while agonists--apomorphine and dopamine gave IC50 values of 2.5 and 10 microM, respectively suggesting a D2 type receptor. Epinephrine and norepinephrine inhibited the binding much less efficiently while mianserin at 10 microM and serotonin at 1 mM concentration did not inhibit the binding. Bimolecular association and dissociation rate constants for the reversible binding were 5.7 x 10(8) M-1 min-1 and 5.0 x 10(-2) min-1, respectively. Equilibrium dissociation constant was 87 pM and the KD obtained by saturation binding was 73 pM. During the foetal age 16 to 40 weeks, the receptor concentration remained in the range of 38-60 fmol/mg protein or 570-1080 fmol/g striatum but it increased two-fold postnatally reaching a maximum at 5 years.(ABSTRACT TRUNCATED AT 250 WORDS)

Agonist-Stimulated Inositol Polyphosphate Formation in Cerebellum

The accumulation of inositol polyphosphates in the cerebellum in response to agonists has not been demonstrated. Guinea pig cerebellar slices prelabeled with [3H]inositol showed the following increases in response to 1 mM serotonin: At 15 s, there was a peak in 3H label in the second messenger inositol 1,4,5-trisphosphate [Ins(1,4,5)P3], decreasing to a lower level in about 1 min. The level of 3H label in the putative second-messenger inositol 1,3,4,5-tetrakisphosphate [Ins(1,3,4,5)P4] increased rapidly up to 60 s and increased slowly thereafter. The accumulation of 3H label in various inositol phosphate isomers at 10 min, when steady state was obtained, showed the following increases due to serotonin: inositol 1,3,4-trisphosphate [Ins(1,3,4)P3], eight-fold; Ins(1,3,4,5)P4, 6.4-fold; Ins(1,4,5)P3, 75%; inositol 1,4-bisphosphate [Ins(1,4)P2], 0%; inositol 3,4-bisphosphate, 100%; inositol 1-phosphate/inositol 3-phosphate, 30%; and inositol 4-phosphate, 40%. [3H]Inositol 1,3-bisphosphate was not detected in controls, but it accounted for 7.2% of the total inositol bisphosphates formed in the serotonin-stimulated samples. The fact that serotonin did not increase the formation of Ins(1,4)P2 could be due to the fact that Ins(1,4)P2 is rapidly degraded or that Ins(1,4,5)P3 is metabolized primarily by Ins(1,4,5)P3-3'kinase to form Ins(1,3,4,5)P4. In the presence of pargyline (10 microM), [3H]Ins(1,3,4,5)P4 and [3H]Ins(1,3,4)P3 levels were increased, even at 1 microM serotonin. Ketanserin (7 microM) completely inhibited the serotonin effect, indicating stimulation of serotonin2 receptors. Quisqualic acid (100 microM) also increased the levels of [3H]Ins(1,4,5)P3, [3H]Ins(1,3,4,5)P4, and [3H]Ins(1,3,4)P3, but the profile of these increases was different.(ABSTRACT TRUNCATED AT 250 WORDS)

Effect of thiocarbamate, urea, and uracil herbicides on lipid metabolism in groundnut (Arachis hypogaea) leaves

The effect of thiocarbamates (S-ethyldipropylthiocarbamate and diallate), substituted ureas (monuron and diuron), and uracils (bromacil and terbacil) on lipid metabolism in groundnut (Arachis hypogaea) leaves was investigated under nonphotosynthetic conditions. The uptake of [1-14C]acetate by leaf disks was inhibited by the thiocarbamates and marginally by the substituted ureas, but not by the uracil herbicides. The uptake of [methyl-14C]choline was inhibited to a lesser extent by thiocarbamates, while the other herbicides showed a slight stimulation. The thiocarbamates almost completely inhibited uptake of [32P]orthophosphate at 1.0 mM concentration, while diuron and terbacil showed significant inhibition. [1-14C]Acetate incorporation into lipids was inhibited only by diallate. [methyl-14C]Choline incorporation into the choline phosphoglycerides was inhibited by diallate, diuron, and bromacil. The incorporation of [32P]orthophosphate into phospholipids was substantially inhibited (over 90% at 1.0 mM) by the thiocarbamates, but not by the other herbicides. [35S]Sulfate incorporation into sulfoquinovosyl diglycerides was markedly inhibited only by the thiocarbamates. Fatty acid synthesis by isolated chloroplasts was inhibited 40–85% by thiocarbamates, substituted ureas, and bromacil, but not by terbacil. The inhibitory effect of the urea derivatives was reversible, but that of thiocarbamates was irreversible. sn-Glycerol-3-phosphate acyltransferase(s) of the chloroplast and microsomal fractions were profoundly inhibited by thiocarbamates, but not by the other two groups of herbicides. Phosphatidic acid phosphatase was insensitive to all the herbicides tested.Key words: herbicides, thiocarbamates, substituted ureas, uracils, effects on plant lipids.

Triacylglycerol synthesis in developing seeds of groundnut (Arachis hypogaea): pathway and properties of enzymes of sn-glycerol 3-phosphate formation

The enzymatic pathway for the synthesis of sn-glycerol 3-phosphate was investigated in developing groundnut seeds (Arachis hypogaea). Glycerol-3-phosphate dehydrogenase was not detected in this tissue but an active glycerokinase was demonstrated in the cytosolic fraction. It showed an optimum pH at 8.6 and positive cooperative interactions with both glycerol and ATP. Triosephosphate isomerase and glyceraldehyde-3-phosphate phosphatase were observed mainly in the cytosolic fraction while an active glyceraldehyde reductase was found mainly in the mitochondrial and microsomal fractions. The glyceraldehyde 3-phosphate phosphatase showed specificity and positive cooperativity with respect to glyceraldehyde 3-phosphate. The glyceraldehyde reductase was active toward glucose and fructose but not toward formaldehyde and showed absolute specificity toward NADPH. It is concluded that in the developing groundnut seed, sn-glycerol 3-phosphate is synthesized essentially by the pathway dihydroxyacetone phosphate----glyceraldehyde 3-phosphate Pi----glyceraldehyde NADPH----glycerol ATP----glycerol 3-phosphate. All the enzymes of this pathway showed activity profiles commensurate with their participation in triacylglycerol synthesis which is maximal during the period 15-35 days after fertilization. Glycerokinase appears to be the rate-limiting enzyme in this pathway.

The biosynthesis of sulfoquinovosyldiacylglycerol: studies with groundnut (Arachis hypogaea) leaves

The biosynthetic pathway of sulfoquinovosyldiacylglycerol (SQDG) was investigated using groundnut (Arachis hypogaea) leaf discs and 35S-labeled precursors. [35S]SO4(2-) was actively taken up by the leaf discs and rapidly incorporated into SQDG. After 2 h, 1.5% of the [35S]SO4(2-) added to the incubation medium was taken up, of which 28% was incorporated into SQDG. The methanol-water phases of the lipid extracts of the leaf discs were analyzed for the 35S-labeled intermediates. Up to 2 h of incubation, cysteic acid, 3-sulfopyruvate, 3-sulfolactate, 3-sulfolactaldehyde, and sulfoquinovose (SQ) which have been proposed as intermediates [Davies et al. (1966) Biochem. J. 98, 369-373] were not labeled. Only a negligible amount of radioactivity was observed in these compounds after incubation for 4 h and more. Addition of sodium molybdate inhibited the uptake of [35S]SO4(2-) as well as its incorporation into SQDG by the leaf discs, suggesting that 3'-phosphoadenosine-5'-phosphosulfate may be involved in the biosynthesis of SQDG. Addition of unlabeled cysteic acid to the incubation medium enhanced the uptake of [35S]SO4(2-) but did not affect its incorporation into SQDG. 35S-labeled cysteic acid was taken up by the leaf discs and metabolized to sulfoacetic acid but not incorporated into SQ or SQDG. These results show that cysteic acid is not an intermediate in SQDG biosynthesis. [35S]SQ was taken up by the leaf discs and incorporated into SQDG in a time-dependent manner. [35S]Sulfoquinovosylglycerol was also taken up by the leaf discs but not incorporated into SQDG. It is concluded that SQDG is not biosynthesized by the proposed sulfoglycolytic pathway in higher plants. Though [35S]SQ was converted to SQDG, the rates are much lower compared to [35S]SO4(2-) incorporation, which suggests that a more direct pathway involving sulfonation of a lipid precursor may exist in higher plants.

Metabolism of the plant sulfolipid--sulfoquinovosyldiacylglycerol: degradation in animal tissues

Metabolism of the plant sulfolipid--sulfoquinovosyldiacylglycerol (SQDG)--was studied in animal tissues. In vivo experiments with [35S]SQDG in guinea pigs showed that this lipid is not absorbed intact in the gastrointestinal tract. In these experiments, 3 h after administration of [35S]SQDG, the intestinal mucosa contained 1 to 5% of the radioactivity as SQDG, while the remainder was in a water-soluble form. Analysis of the water-soluble components showed that about 60% of the radioactivity was present as sulfoquinovosylglycerol (SQG) and the remainder was present as free SO4(2-). In the blood, 99% of the radioactivity was present as SO4(2-), SQG was not observed. In liver, only very little radioactivity was observed and appeared to be mainly in the form of SO4(2-). Experiments with everted intestinal sacs of guinea pigs confirmed the formation of SQG, SO4(2-), and, in addition, sulfoquinovosylmonoacylglycerol (SQMG) in this tissue. In vitro experiments with saline extracts of acetone powders of pancreas and intestinal mucosa of guinea pig, sheep, and rat showed that [35S]SQDG was deacylated to SQMG (sulfolipase A activity) and SQG (sulfolipase B activity). It is concluded that animal tissues deacylate SQDG in a stepwise manner to SQG. It is further metabolized to yield free SO4(2-) by cleavage of the C-S bond which appears to be brought about by the intestinal microflora. Sheep pancreatic sulfolipases were characterized. Bile salts, sodium dodecyl sulfate, and Triton X-100 inhibited the pancreatic sulfolipases, while CaCl2 activated them. Substrate competition experiments and investigations on substrate specificity with a partially purified preparation indicated that relatively specific sulfolipase(s) may exist in pancreas. Among the species tested, guinea pig tissues showed the highest sulfolipase A and B activities followed sheep and rat tissues. Pancreatic enzymes were 18 to 60 times more active than intestinal enzymes.

Susceptibility of the Wolfgram proteins and stability of 2',3'-cyclic nucleotide 3'-phosphodiesterase of rat brain myelin to limited proteolytic digestion

The susceptibility of proteins in the myelin membrane to proteases was studied. Lyophilized rat brain myelin suspended in water was subjected to controlled proteolytic digestion with pure trypsin (N-tosyl-L-phenylalanine chloromethyl ketone treated, 5 units/mg of myelin), and proteins remaining in the pellet were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Under these conditions, large basic protein (LBP) was completely hydrolyzed in 5-10 min, proteolipid proteins remained largely intact until 60 min, whereas Wolfgram protein (WP) was progressively degraded from 10 min onward with the simultaneous appearance of a new protein band with a molecular weight of 35K. A similar pattern was obtained on treatment with chymotrypsin or subtilisin. The 35K protein band was shown to be derived from WP by its immunological cross-reactivity with WP antibodies. Western blot analysis showed that 35K protein is the only major breakdown product of WP under these conditions. Treatment with higher concentrations of trypsin (greater than 20 units/mg of myelin) resulted in the degradation of all myelin proteins. Essentially all the 2',3'-cyclic nucleotide 3'-phosphodiesterase (CNP) activity was observed in the myelin pellet after controlled or drastic digestion with trypsin. It is concluded that the major fragment of WP (35K) is located in the hydrophobic milieu of the bilayer, relatively inaccessible to trypsin, whereas a portion (20K) of the WP is exposed to the cytoplasmic side (major dense line), like LBP, and that peptide fragments (less than 14K) that remained in the myelin membrane lipid bilayer after trypsin digestion could exhibit CNP activity.

Relaxation labeling with learning automata

Relaxation labeling processes are a class of mechanisms that solve the problem of assigning labels to objects in a manner that is consistent with respect to some domain-specific constraints. We reformulate this using the model of a team of learning automata interacting with an environment or a high-level critic that gives noisy responses as to the consistency of a tentative labeling selected by the automata. This results in an iterative linear algorithm that is itself probabilistic. Using an explicit definition of consistency we give a complete analysis of this probabilistic relaxation process using weak convergence results for stochastic algorithms. Our model can accommodate a range of uncertainties in the compatibility functions. We prove a local convergence result and show that the point of convergence depends both on the initial labeling and the constraints. The algorithm is implementable in a highly parallel fashion.

Cholinergic muscarinic receptors in human fetal brain: ontogeny of [3H]quinuclidinyl benzilate binding sites in corpus striatum, brainstem, and cerebellum

The ontogeny of muscarinic receptors was studied in human fetal striatum, brainstem, and cerebellum to investigate general principles of synaptogenesis as well as the physiological balance between various chemical synapses during development in a given region of the brain. [3H]Quinuclidinyl benzilate ([3H]QNB) binding was assayed in total particulate fraction (TPF) from various parts of brain. In the corpus striatum, QNB binding sites are present at 16 weeks of gestation (average concentration 180 fmol/mg protein of TPF), slowly increase up to 24 weeks (average concentration 217 fmol/mg protein), and rapidly increase during the third trimester to 480 fmol/mg protein of TPF. In contrast, dopaminergic receptors exist as two subpopulations, one with low affinity and the other with high affinity up to the 24th week of gestation; all of them acquire the high-affinity characteristic during the third trimester. In brainstem, the muscarinic receptors show maximum concentration by 16 weeks of age (360 fmol/mg protein of TPF). Subsequently the muscarinic receptor concentration shows a gradual decline in the brainstem. In cerebellum, except for a slight increase at 24 weeks (average concentration 90 fmol/mg protein of TPF), the receptor concentration remained nearly constant at about 60-70 fmol/mg protein of TPF throughout fetal life. This study demonstrates that the ontogeny of muscarinic receptors varies among the different regions, and the patterns observed suggest that receptor formation occurs principally in the third trimester. Also noteworthy is the finding that the QNB binding sites decreased in all regions of the human brain during adult life.