Search

Anthrax toxin gene expression in Bacillus anthracis is dependent on the presence of atxA, a trans-acting regulatory gene located on the resident 185-kb plasmid pXO1. In atxA+ strains, expression of the toxin genes (pag, lef, and cya) is enhanced by two physiologically significant signals: elevated CO2/bicarbonate and temperature. To determine whether increased toxin gene expression in response to these signals is associated with increased atxA expression, we monitored steady-state levels of atxA mRNA and AtxA protein in cells cultured in different conditions. We purified histidine-tagged AtxA [AtxA(His)] from Escherichia coli and used anti-AtxA(His) serum to detect AtxA in protein preparations from B. anthracis cells. AtxA was identified as a protein with an apparent size of 56 kDa in cytoplasmic fractions of B. anthracis cells. Our data indicate that atxA expression is not influenced by CO2/bicarbonate levels. However, the steady-state level of atxA mRNA in cells grown in elevated CO2/bicarbonate at 37 degrees C is five- to sixfold higher than that observed in cells grown in the same conditions at 28 degrees C. A corresponding difference in AtxA protein was also seen at the different growth temperatures. When atxA was cloned on a multicopy plasmid in B. anthracis, AtxA levels corresponding to the atxA gene copy number were observed. However, this strain produced significantly less pag mRNA and protective antigen protein than the parental strain harboring atxA in single copy on pXO1. These results indicate that increased AtxA expression does not lead to a corresponding increase in pag expression. Our data strongly suggest that an additional factor(s) is involved in regulation of pag and that the relative amounts of such a factor(s) and AtxA are important for optimal toxin gene expression.

Bacillus anthracis plasmid pXO1 carries the structural genes for the three anthrax toxin proteins, cya (edema factor), lef (lethal factor), and pag (protective antigen). Expression of the toxin genes by B. anthracis is enhanced during growth under elevated levels of CO2. This CO2 effect is observed only in the presence of another pXO1 gene, atxA, which encodes a transactivator of anthrax toxin synthesis. Here we show that transcription of atxA does not appear to differ in cells grown in 5% CO2 compared with cells grown in air. Using a new efficient method for gene replacement in B. anthracis, we constructed an atxA-null mutant in which the atxA-coding sequence on pXO1 is replaced with an omega km-2 cassette. Transcription of all three toxin genes is decreased in the absence of atxA. The pag gene possesses two apparent transcription start sites, P1 and P2; only transcripts with 5' ends mapping to P1 are decreased in the atxA-null mutant. Deletion analysis of the pag promoter region indicates that the 111 bp region upstream of the P1 site is sufficient for atxA-mediated activation of this transcript. The cya and lef genes each have one apparent start site for transcription. Transcripts with 5' ends mapping to these sites are not detected in the atxA-null mutant. The atxA-null mutant is avirulent in mice. Moreover, the antibody response to all three toxin proteins is decreased significantly in atxA-null mutant-infected mice. These data suggest that the atxA gene product also regulates toxin gene expression during infection.

The pag gene of Bacillus anthracis, located on plasmid pXO1 (185 kb), encodes protective antigen, a component of the anthrax lethal and edema toxins. Synthesis of protective antigen is enhanced during growth of the organism with elevated levels of CO2. The CO2 effect is at the level of transcription, and pXO1-encoded regulatory factors have been implicated in control of pag expression. We used a Tn917-LTV3 insertion mutant of B. anthracis in which the wild-type pag gene on pXO1 was replaced with a pag-lacZ transcriptional fusion to monitor pag promoter activity. Expression of the pag-lacZ fusion is induced five- to eightfold during growth in 5% CO2 compared with growth in air. Growth in 20% CO2 increases transcription up to 19-fold. By monitoring pag-lacZ expression in atmospheres with different O2 and CO2 concentrations, we demonstrated definitively that the CO2 effect is specific and not simply a result of increased anaerobiosis. The results of 5' end mapping of pag transcripts indicate multiple sites of transcript initiation. We have determined two major apparent start sites, designated P1 and P2, located at positions -58 and -26 relative to the translation initiation codon, respectively. Analysis of total RNA from late-log-phase cells shows comparable initiation from P1 and P2 in wild-type strains grown in aerobic conditions. However, initiation from P1 is increased approximately 10-fold in cultures grown with an elevated level (5%) of CO2. We have identified a locus on pXO1, more than 13 kb upstream from the pag gene, which enhances pag transcription. When added in trans, this locus increases the level of transcripts with 5' ends mapping to P1 but has no effect on the level of transcripts with 5' ends mapping to P2. The CO2 effect on P1 is observed only in the presence of the activator locus.