Immunochemical analysis of serologically active lipids of Mycoplasma pneumoniae

Semi-automated curation of metabolic models via flux balance analysis: a case study with Mycoplasma gallisepticum

Primarily used for metabolic engineering and synthetic biology, genome-scale metabolic modeling shows tremendous potential as a tool for fundamental research and curation of metabolism. Through a novel integration of flux balance analysis and genetic algorithms, a strategy to curate metabolic networks and facilitate identification of metabolic pathways that may not be directly inferable solely from genome annotation was developed. Specifically, metabolites involved in unknown reactions can be determined, and potentially erroneous pathways can be identified. The procedure developed allows for new fundamental insight into metabolism, as well as acting as a semi-automated curation methodology for genome-scale metabolic modeling. To validate the methodology, a genome-scale metabolic model for the bacterium Mycoplasma gallisepticum was created. Several reactions not predicted by the genome annotation were postulated and validated via the literature. The model predicted an average growth rate of 0.358±0.12, closely matching the experimentally determined growth rate of M. gallisepticum of 0.244±0.03. This work presents a powerful algorithm for facilitating the identification and curation of previously known and new metabolic pathways, as well as presenting the first genome-scale reconstruction of M. gallisepticum.

Flux balance analysis (FBA) is a powerful approach for genome-scale metabolic modeling. It provides metabolic engineers with a tool for manipulating, predicting, and optimizing metabolism for biotechnological and biomedical purposes. However, we posit that it can also be used as tool for fundamental research in understanding and curating metabolic networks. Specifically, by using a genetic algorithm integrated with FBA, we developed a curation approach to identify missing reactions, incomplete reactions, and erroneous reactions. Additionally, it was possible to take advantage of the ensemble information from the genetic algorithm to identify the most critical reactions for curation. We tested our strategy using Mycoplasma gallisepticum as our model organism. Using the genome annotation as the basis, the preliminary genome-scale metabolic model consisted of 446 metabolites involved in 380 reactions. Carrying out our analysis, we found over 80 incorrect reactions and 16 missing reactions. Based upon the guidance of the algorithm, we were able to curate and resolve all discrepancies. The model predicted an average bacterial growth rate of 0.358±0.12 h⁻¹ compared to the experimentally observed 0.244±0.03 h⁻¹. Thus, our approach facilitated the curation of a genome-scale metabolic network and generated a high quality metabolic model.

Role of Glycolipids and Phosphatidylglycerol in the Serological Activity of Mycoplasma pneumoniae

Mycoplasma pneumoniae lipids were fractionated by combined silicic acid and diethylaminoethyl-cellulose chromatography. The lipid fractions were tested for complement-fixing activity and for blocking of metabolism-inhibiting antibodies in human and rabbit antisera to M. pneumoniae. Thin-layer chromatography and carbohydrate analysis of hydrolysates of the lipid fractions indicated that glycolipids were responsible for the specific serological activities of the lipid extracts of the organism. Phosphatidylglycerol, which accompanied the glycolipids, was serologically inactive when isolated free from glycolipids. Phosphatidylglycerol considerably enhanced the serological activity of the glycolipids, serving as an auxiliary lipid. Because of this auxiliary effect, minute quantities of contaminating glycolipids sufficed to render serological activity to "purified" phosphatidylgylcerol preparations, providing an explanation to the previous identification of the active lipid as a phospholipid.

A processive lipid glycosyltransferase in the small human pathogen Mycoplasma pneumoniae: involvement in host immune response

The human pathogen Mycoplasma pneumoniae has a very small genome but with many yet not identified gene functions, e.g. for membrane lipid biosynthesis. Extensive radioactive labelling in vivo and enzyme assays in vitro revealed a substantial capacity for membrane glycolipid biosynthesis, yielding three glycolipids, five phosphoglycolipids, in addition to six phospholipids. Most glycolipids were synthesized in a cell protein/lipid-detergent extract in vitro; galactose was incorporated into all species, whereas glucose only into a few. One (MPN483) of the three predicted glycosyltransferases (GTs; all essential) was both processive and promiscuous, synthesizing most of the identified glycolipids. These enzymes are of a GT-A fold, similar to an established structure, and belong to CAZy GT-family 2. The cloned MPN483 could use both diacylglycerol (DAG) and human ceramide acceptor substrates, and in particular UDP-galactose but also UDP-glucose as donors, making mono-, di- and trihexose variants. MPN483 output and processitivity was strongly influenced by the local lipid environment of anionic lipids. The structure of a major beta1,6GlcbetaGalDAG species was determined by NMR spectroscopy. This, as well as other purified M. pneumoniae glycolipid species, is important antigens in early infections, as revealed from ELISA screens with patient IgM sera, highlighting new aspects of glycolipid function.

Biological activities of monoclonal antibodies to Mycoplasma pneumoniae membrane glycolipids

A purified preparation of membranes was obtained by using a unique method of treating Mycoplasma pneumoniae with the ATPase inhibitor, diethylstilbestrol. This method was shown to yield highly purified membranes with little or no cytoplasmic contamination. These membranes were used to immunize mice for subsequent productions of monoclonal antibodies (MAbs). Hybridoma culture supernatants were screened by enzyme-linked immunosorbent assay with whole-cell M. pneumoniae and lipid extract antigens. Four stable MAbs were obtained and characterized. MAb CP3-46F5 reacted with a protein of a molecular weight of approximately 52,000 as determined by Western blot (immunoblot). MAbs CP3-50C2, CP3-53C5, and CP3-53C8 did not react with any antigens on Western blots but did bind to at least 10 distinct glycolipid bands as determined by orcinol staining on thin-layer chromatograms of M. pneumoniae lipid extracts. The MAbs did not react with similarly prepared lipid extracts from Mycoplasma genitalium, Mycoplasma neurolyticum, and Mycoplasma gallisepticum. These MAbs did not inhibit M. pneumoniae metabolism or attachment to WiDr cell cultures. The anti-glycolipid MAbs recognize determinants specific to M. pneumoniae, unlike polyclonal hyperimmune sera against M. pneumoniae, which cross-react with lipid extracts of M. genitalium.

An ELISA to detect monoclonal antibodies specific for lipid determinants of Mycoplasma pneumoniae

An enzyme-linked immunosorbent assay (ELISA) was developed for analysis of monoclonal antibodies directed towards lipid determinants of Mycoplasma pneumoniae. Chloroform-methanol lipid extracts as well as chromatographed lipid fractions of M. pneumoniae were bound to polyvinylchloride microtiter wells. Of 293 clones positive to M. pneumoniae as detected by a whole cell ELISA, a total of 78 clones produced antibodies which selectively bound to lipid extracts of M. pneumoniae. The simplicity and reproducibility of the assay permit rapid screening for detection of antibodies directed against non-protein cellular antigens.

Phospholipid interconversions in Mycoplasma capricolum

Mycoplasma capricolum cells increase their phospholipid content by incorporating exogenous phospholipids from the growth medium. Growing the cells in media with increasing serum concentrations resulted in a massive incorporation of phosphatidylcholine and sphingomyelin (up to about 50% of total phospholipids) into the cell membrane. The incorporation of the exogenous phospholipids had essentially no effect on the rate of cell growth and did not decrease the overall phospholipid biosynthesis of the cells. Thus, the ratio of phospholipid to protein in membranes from cells grown with 5% horse serum was 0.5 (mumol/mg) compared to 0.3 (mumol/mg) in cells grown without serum, and the relative content of charged polar lipids was apparently decreased. The consequence of the incorporation of exogenous phosphatidylcholine was an alteration in the relative amount of the major end-products of the de novo phospholipid biosynthesis; a marked increase in the ratio of diphosphatidylglycerol to phosphatidylglycerol was observed. The possibility that the increase in the ratio of diphosphatidylglycerol to phosphatidylglycerol is part of a control mechanism to maintain a mixture of bilayer and non-bilayer lipids is discussed.

Mitogenic activity of cytoplasmic membranes isolated from L-forms of Staphylococcus aureus

Cytoplasmic membranes of L-forms of Staphylococcus aureus exerted a strong mitogenic effect on splenocytes of athymic nude mice as well as normal mice, while a cytoplasmic fraction of the same bacteria did not show definite mitogenicity. The mitogenic principle(s) of the membrane fraction was resistant to treatment with trypsin and was heat stable (at 100 C for 10 min). The active principle(s) in the insoluble residue of the membrane fraction digested with trypsin was not extracted with cold acetone, but could be solubilized by extraction with a cold chloroform-methanol mixture (2:1, v/v). The mitogenic principle(s) in the extract was fractionated by silicic acid column chromatography. Among five fractions separated by chromatography, fractions eluted with chloroform-methanol mixtures (1:1 and 1:20, v/v) were found to be strongly mitogenic. The cytoplasmic membranes of the L-forms also exerted a definite mitogenic effect on guinea pig splenocytes, but not on the thymocytes.

Phospholipid and cholesterol uptake by Mycoplasma cells and membranes

The ability of growing mycoplasma cells and their isolated membranes to take up exogenous phospholipids was correlated with their ability to take up cholesterol. Horse serum or vesicles made of phosphatidylcholine and cholesterol served as lipid donors. Growing cells of five Mycoplasma species took up significant quantities of phosphatidylcholine and sphingomyelin as well as free and esterified cholesterol. In contrast, growing cells of three Acholeplasma species failed to take up any of the exogenous phospholipids, and only incorporated low amounts of free cholesterol and no esterified cholesterol. Hence, the ability of mycoplasmas to take up large quantities of cholesterol appears to be correlated with an ability to take up exogenous phospholipids. Isolated membranes of Mycoplasma capricolum and Acholeplasma laidlawii took up lower amounts of cholesterol than did membranes of growing cells and did not take up phospholipids. Inhibition of M. capricolum growth decreased the ability of the cells to take up exogenous phospholipids and cholesterol. The possibility that the contact between the lipid donors and the membrane involves specific receptors best exposed in actively growing cells is discussed.

Binding of plant lectins to mycoplasma cells and membranes

The binding of iodinated wheat germ agglutinin, Ricinus communis agglutinin, and concanavalin A to mycoplasma cells and membranes was examined. All mycoplasmas studied specifically bound concanavalin A or R. communis agglutinin and, to a lesser degree, wheat germ agglutinin. The binding of lectins to whole cells was similar to that recorded for membranes, suggesting that significant binding only occurred on the outer surface of the mycoplasma membrane. Proteolysis of the membrane almost always increased the capacity to bind lectins, which indicates that additional carbohydrate groups on the mycoplasma membrane are masked by a protein layer or protein complexes on the membrane. The observation that carbohydrates are apparently exposed on the surface of mycoplasma membranes should stimulate more concentrated study on the isolation and chemical characterization of these substances since it is quite likely that they are responsible for a variety of reactions between mycoplasmas and host cells.

Studies on delayed hypersensitivity of antigens isolated from Mycoplasma pneumoniae cells

Cells of Mycoplasma pneumoniae Mac strain were fractionated into acetone-soluble and insoluble fractions. Acetone-insoluble fractions were digested with pronase and further purified by chromatography on Sephadex G-75, yielding three water-soluble fractions which were free from lipid and consisted mainly of polysaccharide-protein complex. All these water-soluble fractions possessed eliciting antigenicity to delayed hypersensitivity for M. pneumoniae as measured by skin reactions and macrophage migration inhibition tests, but not to complement-fixing antibodies. In contrast, the acetone-soluble fraction was reactive with the complement-fixing antibodies but not for the delayed hypersensitivity.

Distribution and composition of lipopolysaccharides from mycoplasmas

Polymeric carbohydrates containing glycerol and fatty acids were isolated from whole cells and membranes of mycoplasmas by hot aqueous phenol extraction and gel filtration. Lipopolysaccharides were found to occur in four species of Acholeplasma, two of Anaeroplasma, and in Mycoplasma neurolyticum. None were detected in Spiroplasma citri or in five species of Mycoplasma. All lipopolysaccharides contained both neutral and N-acylated amino sugars in ratios varying from 1:1 to 3:1. The neutral sugars found in varying distribution were glucose, galactose, and mannose. The amino sugars included fucosamine, an unidentified deoxyhexosamine, galactosamine, and glucosamine. Fucosamine and glucose were the only sugars common to all lipopolysaccharides. The fatty acids were similar to those found in the lipids of each organism.

Immunological properties of glycolipids from membranes of Acholeplasma laidlawii

Glycolipids, the predominant class of lipids in the membranes of Acholeplasma laidlawii, are the haptenic determinants that react with anti-A. Laidlawii serum to fix complement. The predominant complement-fixing activity of the membrane glycolipids was associated with the monoglucoysyl diglyceride, diglucosyl diglyceride, glycerlphosphoryl diglucosyl diglyceride (GPDD), and an unknown lipid B, which did not react with ninhydrin but release glucose and glycerol and traces of phosphorus upon hydrolysis. The glycolipids monoglucosyl diglyceride and diglucosyl diglyceride or GPDD and unknown lipid B were paired as a result of their cross-reactions with selective antisera prepared with the aid of reconstituted membrane complexes containing membrane lipids. Reconstituted membrane complexes assembled from [14C]monoglucosyl diglyceride and delipidated membrane proteins gave optimal complement fixation titers before saturation of the complexes with the ]14C]monoglucosyl diglyceride. The phosphoglycolipid of the membrane, GPDD, was anticomplementary as a pure lipid, a cholesterol liposome, and a reconstituted membrane complex. This anticomplementary activity, which was caused by 3 mug of pure GPDD, affected both human and guinea pig complement. Although human C1, C4, C3, and C5 were not inhibited by GPDD, C2 was inhibited 10-fold by reconstituted membrane complexes containing 150 mug of GPDD. A role for this phosphoglycolipid is discussed in the hypothetical mechanism of inhibition of C2 attachment to SAC1, 4 sites.

Persistent cerebellar symptoms after infection with Mycoplasma pneumoniae

Mycoplasma pneumoniae infection in a 47-year-old man is reported. Symptoms of upper respiratory tract infection were followed by pneumonia and meningoencephalitis. In contrast to published cases with neurological manifestations of infection with M. pneumoniae, the patient was disabled by persistent cerebellar symptoms with generalized ataxia and atactic dysarthria. Some possible pathogenic mechanisms of the neurological manifestations of infection with M. pneumoniae are considered.

Immunological analysis of glycolipids and lipopolysaccharides derived from various mycoplasmas

The lipids and lipopolysaccharides of five mycoplasmas were examined for complement-fixing activity to antimembrane rabbit sera. Total glycolipid fractions and the aqueous phenol fractions (lipopolysaccharides) from the membranes of Acholeplasma laidlawii, A. modicum, A. axanthum, and Mycoplasma neurolyticum exhibited significant antigenic activity. The glycolipids and phosphoglycolipids from Thermoplasma acidophilum were either anticomplementary or did not react due to their extreme hydrophobic nature. High activity was found using the lipopolysaccharide of T. acidophilum. The pure glycolipids and phosphoglycolipids of the Acholeplasma and Mycoplasma species also exhibited significant complement-fixing activity. Acylation of the sugar residues of these lipids reduced or negated complement-fixing activity. Double cross-reactions between the glycolipids of A. laidlawii and A. modicum appeared to be due to mono- and diglucosyl diglycerides of identical structure. Specificity of glycolipid structure was noted by the absence of cross-reactions between A. laidlawii and M. neurolyticum, the glycolipids of which differ only in the nature of the glucose linkages. The existence of lipopolysaccharides in the membranes of mycoplasmas and their complement-fixing activity in the presence of antimembrane sera suggest their possible importance as specific antigenic determinants.

Chemical and serological properties of Mycoplasma hyorhinis fractions

Mycoplasma hyorhinis organisms were fractionated into lipid, carbohydrate, and protein fractions by chemical and physiochemical procedures. When all fractions were tested for complement fixation and inhibition of growth inhibition activity, only the protein fractions possessed activity. Thus, it appears that the major antigenic component of M. hyorhinis is a protein.

Serological cross-reaction between lipids of Mycoplasma pneumoniae and Mycoplasma neurolyticum

The complement-fixing activity of crude lipid extracts of 10 Mycoplasma species was compared with that of whole organism antigens employing immune rabbit serum. Five species (M. pneumoniae, M. neurolyticum, M. granularum, M. laidlawii, and M. fermentans) showed serological activity, whereas the remaining five species (M. canis, M. felis, M. gallisepticum, M. hyorhinis, and M. pulmonis) did not show significant activity in their lipid fractions. The lipid fractions of the five species which had serological activity in their lipid fractions showed three groups of serological specificity. M. pneumoniae cross-reacted with M. neurolyticum, M. granularum cross-reacted with M. laidlawii, and M. fermentans showed specific activity. Acute and convalescent sera from human pneumonia patients from whom M. pneumoniae had been isolated showed antibody increases which could be measured nearly as well by lipids of M. neurolyticum as by those of M. pneumoniae. Only a few human convalescent sera showed antibody measureable by lipids of M. granularum, M. pneumoniae did not cross-react with M. neurolyticum by other serological parameters such as growth inhibition on agar or double immunodiffusion, indicating that only the lipid antigens of these two species cross-react.

Morphology and ultrastructure of Mycoplasma pneumoniae spherules

During growth in fluid medium, most strains of Mycoplasma pneumoniae produce free-floating granules which become larger with time. We have called these granules "spherules." This study describes the morphological and ultrastructural features of M. pneumoniae strain AP-164 spherules, both free and in association with HeLa cells in cell culture. In thin section, spherules were composed of lobulated cells, connected together by membranes, and ring-shaped cells. The two-dimensional morphology observed varied according to the plane of section and to the age of the culture. In HeLa cell cultures, mycoplasmata were found attached to plasma membranes of cells; in sections, individual mycoplasmata were often aligned in radial apposition to the membranes. Mycoplasmata were not found intracellularly. The three-dimensional morphology of spherules was examined by the critical point method and by scanning electron microscopy. Both methods revealed lobulated forms, ring-shaped forms, and star-shaped forms. Treatment of the spherules with crude porcine pancreatic lipase effectively released large numbers of free organisms. Phosphotungstic acid preparations of these uncentrifuged forms revealed a morphology in agreement with the other methods used. Lobulated ring forms with "beaded" filaments were prominent. In respect to morphology, M. pneumoniae under our conditions resembles that of the type species M. mycoides.

Antibodies to brain and other tissues in cases of Mycoplasma pneumoniae infection

Complement-fixing (CF) antibodies against brain tissue antigens were demonstrated in seven cases of Mycoplasma pneumoniae infection with neurological complications. However, forty-three out of fifty-four cases (80%) without neurological symptoms also acquired antibodies to brain during the course of M. pneumoniae infection. 75% of the patients with antibodies to brain also had antibodies against human lung and liver, but the antibody titres to the latter tissue antigens were usually lower than against brain. The reactivity of M. pneumoniae antisera to brain and other tissues was significantly reduced by prior absorption with M. pneumoniae antigen, indicating the presence of a related antigen in M. pneumoniae organisms and these tissues. Fractionation by density gradient centrifugation of thirty-three sera from twenty-six cases showed the antibodies to brain to be of the IgM class exclusively in twenty-four cases while the CF antibodies to M. pneumoniae were associated with both IgM and IgG.

The brain antigen reacting in the CF test with M. pneumoniae antisera could be extracted with chloroform–methanol in the same way as the crude lipid CF antigen of M. pneumoniae.

Immunological analysis of Mycoplasma membranes

The antigens responsible for the production of antibodies to Mycoplasma laidlawii and M. gallisepticum causing growth and metabolic inhibition of these organisms were localized in the cell membrane. Various membrane fractions were tested for serological activity. Membrane lipids were completely or almost completely inactive, whereas several preparations of defatted membrane proteins retained some serological activity, shown by their ability to stimulate metabolic inhibition antibody in rabbits and to adsorb metabolic inhibition antibody and form precipitation lines with an antiserum to the membrane. When the membranes were heated to 65 C for 1 hr, they virtually lost their ability to adsorb metabolic inhibition antibody, which suggests that the antigenic determinants are proteins. Serological activity was retained in reaggregated membranes obtained by dialysis against Mg(2+) of membranes solubilized in sodium dodecyl sulfate. The amount of solubilized membrane protein and lipid incorporated into the reaggregated membranes could be regulated by varying the Mg(2+) concentration. As the serological tests indicated that the various membrane antigens were selectively incorporated into the different reaggregated membranes, the use of controlled reaggregation of solubilized membranes is suggested as a new tool for the fractionation and antigenic analysis of membrane proteins.