- Volume 143, Issue 1, 1997

Bacterial heat-shock proteins (HSPs) from Escherichia coli (GroES, GroEL and DnaK) were studied for their ability to induce by themselves the expression and release of tumour necrosis factor-a (TNF-a), interleukin-6 (IL-1a), interleukin-6 (IL-6) and intercellular adhesion molecule-1 (ICAM-1) by cultured human keratinocytes. The surface expression of ICAM-1 was also investigated. In the supernatants of untreated cells none or a minimal amount of these molecules was found. After 48 h of stimulation with GroEL significant amounts of TNF-a, IL-1a, IL-6 and soluble ICAM-1 were detected, reaching maximum concentrations at 1 g ml-1. The same effect was elicited by DnaK but to a lesser extent. Treatment of keratinocytes with GroEL and DnaK also increased TNF-a, IL-1a, IL-6 and ICAM-1 mRNA levels. GroES showed significant activity only on the expression and release of IL-6. GroEL and DnaK were also able to up-regulate the surface expression of ICAM-1 on keratinocytes. The effects on ICAM-1 expression seemed to be directly due to HSPs and not mediated via cytokines. Furthermore, these effects were due to the properties of HSPs because they were inhibited by specific monoclonal antibodies. These findings support the potential role of HSPs in modulating cell interactions during immunological and inflammatory responses in the skin.

Two monoclonal antibodies (mAbs) designated 144,H-3 (lgG2a) and 218,C-5 (lgM) were produced after immunization of mice with two different heattreated and sonicated pneumococcal strains. Western blotting, with solubilized proteins from different bacterial genera and from mammalian lymphocytes, showed that both mAbs reacted with a protein of approximately 12 kDa in all 66 strains of eubacteria examined, representing 27 different species. The 12 kDa protein was isolated by immunoaffinity chromatography. Subsequent preparative Western blotting enabled N-terminal amino acid sequence analysis by microsequencing. A high degree of amino acid sequence similarity with eubacterial ribosomal proteins L7/L12 was demonstrated. One of the mAbs (144,H-3) also cross-reacted in Western blotting with a 43 kDa protein, but only from streptococci. The 43 kDa protein carrying the common streptococcal epitope was isolated and sequenced in the N-terminal region. A high degree of amino acid sequence identity was found to elongation factor Ts from Escherichia coli.

The relationship between chemical structure and biological activity of the lipid A from Porphyromonas gingivalis, which we recently isolated and whose complete chemical structure was determined [Kumada et al. (1995). J Bacteriol 177, 2098--2106], was studied. The lipid A exhibited endotoxic activity in all the assay systems tested: Limulus gelation activity, lethal toxicity in galactosamine-sensitized mice, mitogenicity in mouse spleen cells and induction of nitric oxide (NO) and tumour necrosis factor alpha (TNF) release from both mouse peritoneal macrophages and the J774-1 mouse macrophage-like cell line. The activity was, however, about 100-fold less than that of Salmonella minnesota LPS used as a control. The moderate activity of the lipid A may be partially explained by its unique fatty acid composition and the lack of a phosphate group in position 4. In contrast, the lipid A as well as whole LPS of P. gingivalis unexpectedly exhibited an even stronger induction of TNF from the human monocytic THP-1 cell line than control LPS when measured by the minimum stimulatory dose. The difference in sensitivity of human and mouse cells to P. gingivalis lipid A suggests that the recognition mechanism, including that for the receptor for endotoxin, may be regulated in different ways in the two cells.

Under anaerobic conditions, especially at low pH, Escherichia coli converts pyruvate to D-lactate by means of an NADH-linked lactate dehydrogenase (LDH). This LDH is present in substantial basal levels under all conditions but increases approximately 10-fold at low pH. The IdhA gene, encoding the fermentative lactate dehydrogenase of E. coli, was cloned using λ10E6 of the Kohara collection as the source of DNA. The IdhA gene was subcloned on a 2.8 kb MIuI-MIuI fragment into a multicopy vector and the region encompassing the gene was sequenced. The IdhA gene of E. coli was highly homologous to genes for other D-lactate-specific dehydrogenases but unrelated to those for the L-lactate-specific enzymes. We constructed a disrupted derivative of the IdhA gene by inserting a kanamycin resistance cassette into the unique KpnI site within the coding region. When transferred to the chromosome, the IdhA::Kan construct abolished the synthesis of the D-LDH completely. When present in high copy number, the IdhA gene was greatly overexpressed, suggesting escape from negative regulation. Cells expressing high levels of the D-LDH grew very poorly, especially in minimal medium. This poor growth was largely counteracted by supplementation with high alanine or pyruvate concentrations, suggesting that excess LDH converts the pyruvate pool to lactate, thus creating a shortage of 3-carbon metabolic intermediates. Using an IdhA-cat gene fusion construct we isolated mutants which no longer showed pH-dependent regulation of the IdhA gene. Some of these appeared to be in the pta gene, which encodes phosphotransacetylase, suggesting the possible involvement of acetyl phosphate in IdhA regulation.

To quantify the dimorphic process in wild and mutant strains of Paracoccidioides brasiliensis, we defined a morphology index (Mi) in terms of the maximum cell length (I), maximum cell diameter (d), and septal diameter (s), according to the equation Mi = 2.13 + 1.13 log10(Is/d 2), whose intercept and slope were such that Mi was around 1 for yeast (spherical) cells or 4 for hyphal (elongated) cells. This discriminatory power was used to quantify morphological population mixtures through Mi histograms. During the temperature-induced dimorphic transition (either way), mean Mi (Mi) varied linearly with time, suggesting a continuity in the process. Also, in wild strains and mutants thereof we found an inverse relationship between Mi and content of both cell wall chitin and 1,3-a-glucan.

A stoichiometric model describing the anaerobic metabolism of Saccharomyces cerevisiae during growth on a defined medium was derived. The model was used to calculate intracellular fluxes based on measurements of the uptake of substrates from the medium, the secretion of products from the cells, and of the rate of biomass formation. Furthermore, measurements of the biomass composition and of the activity of key enzymes were used in the calculations. The stoichiometric network consists of 37 pathway reactions involving 43 compounds of which 13 were measured (acetate, CO2, ethanol, glucose, glycerol, NH+ 4, pyruvate, succinate, carbohydrates, DNA, lipids, proteins and RNA). The model was used to calculate the production rates of malate and fumarate and the ethanol measurement was used to validate the model. All rate measurements were performed on glucose-limited continuous cultures in a high-performance bioreactor. Carbon balances closed within 98%. The calculations comprised flux distributions at specific growth rates of 0.10 and 0.30 h−1. The fluxes through reactions located around important branch points of the metabolism were compared, i.e. the split between the pentose phosphate and the Embden-Meyerhoff-Parnas pathways. Also the model was used to show the probable existence of a redox shunt across the inner mitochondrial membrane consisting of the reactions catalysed by the mitochondrial and the cytosolic alcohol dehydrogenase. Finally it was concluded that cytosolic isocitrate dehydrogenase is probably not present during growth on glucose. The importance of basing the flux analysis on accurate measurements was demonstrated through a sensitivity analysis. It was found that the accuracy of the measurements of CO2, ethanol, glucose, glycerol and protein was critical for the correct calculation of the flux distribution.

The approximately 10 kbp region encompassing nprB and argJ at 102 on the Bacillus subtilis chromosome was sequenced, revealing 12 ORFs, four known genes (argJ, argC, ipi and nprB) and two genes, yitY and yitS, whose products respectively display significant homology with L-gulono-?-lactone oxidase of rat and dihydrofolate reductase of Staphylococcus aureus. The data also indicated that nprB mapped to a different position than previously published.