A, FixLBj PAS domain (pdb code: 1DRM), with the heme colored grey. C, PAS domain of the M. tuberculosis Rv1364c protein (pdb code: 3KC3), showing the fatty acid in the cavity (in grey). E, cavity of PASHm (pdb code: 3BWL) with the Asp side chains (in yellow) pointing to the 1H-indole-3 carbaldehyde ligand (in grey). In PASBvg (F) the corresponding residues CBL-0137 nmr are Tyr596 and Asn631. We nevertheless tested the possibility that PASBvg harbors a heme co-factor or a related GSK690693 molecular weight molecule when present in the full-length BvgS protein in B. pertussis by replacing His643 with Ala. In bona fide

heme-PAS domains, replacement of the His residue abolishes heme binding [31]. Because B. pertussis is virulent in aerobic growth conditions, we reasoned that O2 would most likely be a positive signal for BvgS if the PAS domain harbored an O2-sensing heme, and therefore that a substitution abolishing heme binding should inactivate BvgS. The mutation was introduced into the chromosome of the B. pertussis Tohama I derivative BPSME705 by allelic exchange, and the activity of BvgAS was assessed by using a lacZ Tozasertib reporter under the control of the ptx promoter, which is positively controlled by

BvgAS. The mutated strain expressed ß-galactosidase activity at a level similar to that of the strain containing wt BvgS (Figure 4). Interestingly, BvgSHis643Ala was insensitive to sulfate and nicotinate (Figure 4). Other negative modulators [32] also failed to modulate the activity of the recombinant strain, even at much higher concentrations than those that modulate wild type BvgS (not shown). Thus, the His643Ala substitution appears to make BvgS unresponsive to modulation.

Figure 4 β-galactosidase activities of the recombinant strains producing the BvgS variants. The β-galactosidase activities of the ptx: lacZ fusion were measured as a function of increasing concentrations of nicotinate Demeclocycline or MgSO4. The basal (non-modulated) activities of the three variants tested were not significantly different (P > 0.1) from that of wild type (WT) BvgS. The BPSMΔbvgS and BPSMΔbvgA variants had hardly detectable levels of β-galactosidase activities in all conditions, and therefore they were not included in the figure. In each panel, one and two asterisks represent significantly different activities (with P < 0.05 and P < 0.01, respectively) than that of the same non-modulated BvgS variant. The His643Ala substitution was also introduced into the N2C3 recombinant protein, and the N2C3 variant was purified. Similar to all soluble proteins produced in this work, N2C3His643Ala was dimeric (not shown). Using the thermal shift assay its Tm was determined to be 7°C lower than its wt counterpart (Table 1). Altogether, our data do not support the notion that PASBvg has a heme cofactor. However, His643 appears to be required for BvgS response to negative signals, indicating its functional importance. It also contributes to the thermal stability of recombinant PASBvg.