Selecting high-dimensional mixed graphical models using minimal AIC or BIC forests

Background  

Chow and Liu showed that the maximum likelihood tree for multivariate discrete distributions may be found using a maximum weight spanning tree algorithm, for example Kruskal's algorithm. The efficiency of the algorithm makes it tractable for high-dimensional problems.     

Evidence that TET protein functions as a multimer in the inner membrane of Escherichia coli.

The inner membrane TET (TetA) protein, which is involved in Tn10-mediated microbial tetracycline resistance, consists of two domains, alpha and beta, both of which are needed for tetracycline resistance and efflux (M.S. Curiale, L.M. McMurry, and S.B. Levy, J. Bacteriol. 157:211-217, 1984). Since tetracycline-sensitive mutants in one domain can partially complement sensitive mutants in the other domain and since some sensitive mutants show dominance over the wild type, a multimeric structure for TET in the membrane had been suggested. We have studied this possibility by using tetA-phoA gene fusions. We fused all but the last 40 base pairs of the tetA gene with the carboxy terminus of the phoA gene for alkaline phosphatase (PhoA), whose activity requires its dimerization in the periplasm. The tetA-phoA fusion protein was under control of the tetracycline-inducible regulatory system for the tetA gene. Induction led to the synthesis of a 78,000-dalton inner membrane protein. Tetracycline resistance was expressed at reduced levels, consistent with the terminal beta domain deletion. Alkaline phosphatase activity was also present, but at low levels, suggesting that some, but not all, of the fusion proteins had their carboxy-terminal ends in the periplasm. When wild-type or mutant TET proteins were present in the same cell with the fusion protein, the tetracycline resistance level was affected (raised or lowered); however, phosphatase activity was reduced only when TET proteins with intact or near-intact beta domains were present. These findings suggest that TET functions as a multimer and that intact beta domains, on TET molecules in the heterologous multimer, either allow fewer PhoA moieties to project into the periplasm or sterically hinder PhoA moieties from dimerizing.     

Characterization of MOPC 315 IgA oligosaccharide processing intermediates

The structures of alpha 1,2-mannose containing partially processed asparagine-linked oligosaccharides on the alpha-chain of MOPC 315 IgA were characterized using specific glycosidases and acetolysis. Man6GlcNAc2, a substrate for a Golgi alpha 1,2-mannosidase, was found to be a single isomeric structure. Likewise, Man7-9GlcNAc2 were single isomers indicating an ordered sequence of removal of alpha 1,2-linked mannose residues on this murine immunoglobulin heavy chain.     

MORPHOLOGY OF ISOLATED RABBIT HEART MUSCLE MITOCHONDRIA AND THE OXIDATION OF EXTRAMITOCHONDRIAL REDUCED DIPHOSPHOPYRIDINE NUCLEOTIDE

The morphology of rabbit heart muscle mitochondria isolated in several media has been compared by electron microscopy. The internal structure of isolated mitochondria differs from that of in situ mitochondria, with the type and degree of alteration depending on the isolation medium. Examination of the isolated mitochondria after incubation revealed that additional morphological changes occurred during incubation, but these changes were less pronounced when the incubation was conducted in a complete medium containing substrate. The isolated mitochondria have been shown to be capable of catalyzing a slow aerobic oxidation of extramitochondrial reduced diphosphopyridine nucleotide. The rate of DPNH oxidation observed is sufficient to account for the ability of the mitochondria to oxidize lactate in the presence of catalytic amounts of DPNH. The suspensions used were essentially free of mitochondrial fragments, which are known to oxidize DPNH. Possible relationships of these findings to metabolism in situ are discussed. The results indicate the desirability of correlating biochemical activities with the morphology of isolated mitochondria.     

Characterization of the phosphorylated enzyme intermediate formed in the adenosine 5'-phosphosulfate kinase reaction.

Adenosine 5'-phosphosulfate (APS) kinase (ATP:APS 3'-phosphotransferase) catalyzes the ultimate step in the biosynthesis of 3'-phosphoadenosine 5'-phosphosulfate (PAPS), the primary biological sulfuryl donor. APS kinase from Escherichia coli is phosphorylated upon incubation with ATP, yielding a protein that can complete the overall reaction through phosphorylation of APS. Rapid-quench kinetic experiments show that, in the absence of APS, ATP phosphorylates the enzyme with a rate constant of 46 s-1, which is equivalent to the Vmax for the overall APS kinase reaction. Similar pre-steady-state kinetic measurements show that the rate constant for transfer of the phosphoryl group from E-P to APS is 91 s-1. Thus, the phosphorylated enzyme is kinetically competent to be on the reaction path. In order to elucidate which amino acid residue is phosphorylated, and thus to define the active site region of APS kinase, we have determined the complete sequence of cysC, the structural gene for this enzyme in E. coli. The coding region contains 603 nucleotides and encodes a protein of 22,321 Da. Near the amino terminus is the sequence 35GLSGSGKS, which exemplifies a motif known to interact with the beta-phosphoryl group of purine nucleotides. The residue that is phosphorylated upon incubation with ATP has been identified as serine-109 on the basis of the amino acid composition of a radiolabeled peptide purified from a proteolytic digest of 32P-labeled enzyme. We have identified a sequence beginning at residue 147 which may reflect a PAPS binding site. This sequence was identified in the carboxy terminal region of 10 reported sequences of proteins of PAPS metabolism.     

Do Experience and Body Size Play a Role in Responses of Larval Ringed Salamanders,Ambystoma annulatum,to Predator Kairomones? Laboratory and Field Assays

Prey may experience ontogenetic changes in vulnerability to some predators, either because of changes in morphology or experience. If prey match their level of antipredator behavior to the level of predatory threat, prey responses to predators should reflect the appropriate level of threat for their stage of development. For larval salamanders, responses to predators may change with body size because larger larvae are less vulnerable to predation by gape‐limited predators or because fleeing responses by large salamanders may be more effective than for smaller salamanders. In a field experiment, small larval ringed salamanders, Ambystoma annulatum, responded to chemical stimuli (‘kairomones’) from predatory newts, Notophthalmus viridescens, with an antipredator response (decreased activity). Laboratory‐reared larvae decreased their activity following exposure to newt kairomones, indicating that larval ringed salamanders do not require experience with newts to recognize them as predators. In both experiments, larvae distinguished between chemical stimuli from newts and stimuli from tadpoles (non‐predators) and a blank control. In a third experiment, field‐caught (experienced) larvae showed a graded response to newt kairomones based on their body size: small larvae tended to decrease their activity while larger larvae showed no change or an increase in activity. This graded response was not observed for neutral stimuli, indicating that it is predator‐specific. Therefore, ringed salamander larvae exhibit threat‐sensitive ontogenetic changes in their response to chemical stimuli from predatory newts.     

Effect of tunicamycin on IgM, IgA, and IgG secretion by mouse plasmacytoma cells

Tunicamycin, an antibiotic that prevents glycosylation of glycoproteins by blocking the formation of N-acetylglucosamine-lipid intermediates, was used to study the importance of glycosylation for the secretion of immunoglobulins by mouse plasmacytoma lines that produce immunoglobulins of different classes. Biosynthetically labeled secreted and intracellular immunoglobulins were measured by immunoprecipitation assays. Tunicamycin, at a concentration of 0.5 mug/ml produced an 81% inhibition of IgM secretion by MOPC 104E plasma cells without significantly affecting the initial rate of synthesis of intracellular IgM. No increase in the intracellular degradation of nonglycosylated IgM could be demonstrated. Tunicamycin also produced a 64% average inhibition of IgA secretion by several mouse IgA-secreting plasmacytoma lines. In contrast, despite inhibiting the incorporation of D-[14C] glucosamine into newly synthesized IgG, tunicamycin only produced a 28% average inhibition of IgG secretion, which was only slightly more than the nonspecific inhibition of secretion of the normally nonglycosylated lambda2 light chains by variant MOPC 315 plasmacytomas. These data indicate that the extent of inhibition of immunoglobulin secretion produced by tunicamycin depends on the immunoglobulin class produced by the plasma cell.