Cytoplasmic membranes of the strictly anaerobic sulfate-reducing bacterium Desulfovibrio vulgaris Hildenborough contain two terminal oxygen reductases, a bd quinol oxidase and a cc(b/ o)o 3 cytochrome oxidase (Cox). Viability assays pointed out that single Dbd, Dcox and double DbdDcox deletion mutant strains were more sensitive to oxygen exposure than the WT strain, showing the involvement of these oxygen reductases in the detoxification of oxygen. The Dcox strain was slightly more sensitive than the Dbd strain, pointing to the importance of the cc(b/o)o 3 cytochrome oxidase in oxygen protection. Decreased O 2 reduction rates were measured in mutant cells and membranes using lactate, NADH, ubiquinol and menadiol as substrates. The affinity for oxygen measured with the bd quinol oxidase (K m , 300 nM) was higher than that of the cc(b/o)o 3 cytochrome oxidase (K m , 620 nM). The total membrane activity of the bd quinol oxidase was higher than that of the cytochrome oxidase activity in line with the higher expression of the bd oxidase genes. In addition, analysis of the DbdDcox mutant strain indicated the presence of at least one O 2 -scavenging membrane-bound system able to reduce O 2 with menaquinol as electron donor with an O 2 affinity that was two orders of magnitude lower than that of the bd quinol oxidase. The lower O 2 reductase activity in mutant cells with hydrogen as electron donor and the use of specific inhibitors indicated an electron transfer link between periplasmic H 2 oxidation and membrane-bound oxygen reduction via the menaquinol pool. This linkage is crucial in defence of the strictly anaerobic bacterium Desulfovibrio against oxygen stress.

Although Desulfovibrio vulgaris Hildenborough (DvH) is a strictly anaerobic bacterium, it is able to consume oxygen in different cellular compartments, including extensive periplasmic O 2 reduction with hydrogen as electron donor. The genome of DvH revealed the presence of cydAB and cox genes, encoding a quinol oxidase bd and a cytochrome c oxidase, respectively. In the membranes of DvH, we detected both quinol oxygen reductase [inhibited by heptyl-hydroxyquinoline-N-oxide (HQNO)] and cytochrome c oxidase activities. Spectral and HPLC data for the membrane fraction revealed the presence of o-, b-and d-type haems, in addition to a majority of c-type haems, but no a-type haem, in agreement with carbon monoxide-binding analysis. The cytochrome c oxidase is thus of the cc(o/b)o 3 type, a type not previously described. The monohaem cytochrome c 553 is an electron donor to the cytochrome c oxidase; its encoding gene is located upstream of the cox operon and is 50-fold more transcribed than coxI encoding the cytochrome c oxidase subunit I. Even when DvH is grown under anaerobic conditions in lactate/sulfate medium, the two terminal oxidase-encoding genes are expressed. Furthermore, the quinol oxidase bd-encoding genes are more highly expressed than the cox genes. The cox operon exhibits an atypical genomic organization, with the gene coxII located downstream of coxIV. The occurrence of these membrane-bound oxygen reductases in other strictly anaerobic Deltaproteobacteria is discussed.

Ingestion and killing of bacteria by phagocytic cells protect the human body against infections. While many mechanisms have been proposed to account for bacterial killing in phagosomes, their relative importance, redundancy, and specificity remain unclear. In this study, we used the Dictyostelium discoideum amoeba as a model phagocyte and quantified the requirement of 11 individual gene products, including nine putative effectors, for the killing of bacteria. This analysis revealed that radically different mechanisms are required to kill Klebsiella pneumoniae, Escherichia coli, Pseudomonas aeruginosa, Staphylococcus aureus, and Bacillus subtilis. AlyL, a lysozyme-like protein equipped with a distinct bacteriolytic region, plays a specific role in the intracellular killing of K. pneumoniae, with assistance from BpiC and Aoah, two lipopolysaccharide (LPS)-binding proteins. Rapid killing of E. coli and P. aeruginosa requires the presence of BpiC and of the NoxA NADPH oxidase. No single effector tested is essential for rapid killing of S. aureus or B. subtilis. Overall, our observations reveal an unsuspected degree of specificity in the elimination of bacteria in phagosomes. IMPORTANCE Phagocytic cells ingest and kill bacteria, a process essential for the defense of the human body against infections. Many potential killing mechanisms have been identified in phagocytic cells, including free radicals, toxic ions, enzymes, and permeabilizing peptides. Yet fundamental questions remain unanswered: what is the relative importance of these mechanisms, how redundant are they, and are different mechanisms used to kill different species of bacteria? We addressed these questions using Dictyostelium discoideum, a model phagocytic cell amenable to genetic manipulations and quantitative analysis. Our results reveal that vastly different mechanisms are required to kill different species of bacteria. This very high degree of specificity was unexpected and indicates that a lot remains to be discovered about how phagocytic cells eliminate bacteria.

High-level resistance often evolves when populations of bacteria are exposed to antibiotics, by either mutations or horizontally acquired genes. There is also variation in the intrinsic resistance levels of different bacterial strains and species that is not associated with any known history of exposure. In many cases, evolved resistance is costly to the bacteria, such that resistant types have lower fitness than their progenitors in the absence of antibiotics. Some longer-term studies have shown that bacteria often evolve compensatory changes that overcome these tradeoffs, but even those studies have typically lasted only a few hundred generations. In this study, we examine changes in the susceptibilities of 12 populations ofEscherichia colito 15 antibiotics after 2,000 and 50,000 generations without exposure to any antibiotic. On average, the evolved bacteria were more susceptible to most antibiotics than was their ancestor. The bacteria at 50,000 generations tended to be even more susceptible than after 2,000 generations, although most of the change occurred during the first 2,000 generations. Despite the general trend toward increased susceptibility, we saw diverse outcomes with different antibiotics. For streptomycin, which was the only drug to which the ancestral strain was highly resistant, none of the evolved lines showed any increased susceptibility. The independently evolved lineages often exhibited correlated responses to the antibiotics, with correlations usually corresponding to their modes of action. On balance, our study shows that bacteria with low levels of intrinsic resistance often evolve to become even more susceptible to antibiotics in the absence of corresponding selection.IMPORTANCEResistance to antibiotics often evolves when bacteria encounter antibiotics. However, bacterial strains and species without any known exposure to these drugs also vary in their intrinsic susceptibility. In many cases, evolved resistance has been shown to be costly to the bacteria, such that resistant types have reduced competitiveness relative to their sensitive progenitors in the absence of antibiotics. In this study, we examined changes in the susceptibilities of 12 populations ofEscherichia colito 15 antibiotics after 2,000 and 50,000 generations without exposure to any drug. The evolved bacteria tended to become more susceptible to most antibiotics, with most of the change occurring during the first 2,000 generations, when the bacteria were undergoing rapid adaptation to their experimental conditions. On balance, our findings indicate that bacteria with low levels of intrinsic resistance can, in the absence of relevant selection, become even more susceptible to antibiotics.