Target DNA stabilizes Mycobacterium tuberculosis DevR/DosR phosphorylation by the full-length oxygen sensors DevS/DosS and DosT.

Mycobacterium tuberculosis strongly relies on a latency, or nonreplicating persistence, to escape a human host's immune system. The DevR (DosR), DevS (DosS), and DosT proteins are key components of this process. Like the rhizobial FixL oxygen sensor, DevS and DosT are histidine protein kinases with a heme-binding domain. Like the FixJ partner and substrate of FixL, DevR is a classical response regulator of the two-component class. When activated by DevS or DosT during hypoxia in vivo, DevR induces a dormancy regulon of more than 40 genes. To investigate the contributions of DevS, DosT, and target DNA to the phosphorylation of DevR, we developed an in vitro assay in which the full-length, sensing, DevS and DosT proteins were used to phosphorylate DevR with ATP, in the presence of target DNAs that were introduced as oligonucleotides linked to magnetic nanoparticles. We found that the DevR phosphorylations proceeded only for the deoxy states of the sensors. The reaction was strongly inhibited by O2 , but not CO or NO. The production of phospho-DevR was enhanced sixfold by target consensus DNA or acr-DNA. The phospho-DevR bound tightly to that DNA (Kd ~ 0.8 nm toward acr-DNA), and it was only slightly displaced by a 200-fold excess of unphosphorylated DevR or of a truncated DevR with only a DNA-binding domain. To our knowledge, this represents the first in vitro study of the ligand regulation of DevR phosphorylation by full-length DevS and DosT, and demonstration of a positive effect of DNA on this reaction.

Insights into signal transduction by a hybrid FixL: Denaturation study of on and off states of a multi-domain oxygen sensor.

FixL from Rhizobium etli (ReFixL) is a hybrid oxygen sensor protein. Signal transduction in ReFixL is effected by a switch off of the kinase activity on binding of an oxygen molecule to ferrous heme iron in another domain. Cyanide can also inhibit the kinase activity upon binding to the heme iron in the ferric state. The unfolding by urea of the purified full-length ReFixL in both active pentacoordinate form, met-FixL(FeIII) and inactive cyanomet-FixL (FeIII-CN-) form was monitored by UV-visible absorption spectroscopy, circular dichroism (CD) and fluorescence spectroscopy. The CD and UV-visible absorption spectroscopy revealed two states during unfolding, whereas fluorescence spectroscopy identified a three-state unfolding mechanism. The unfolding mechanism was not altered for the active compared to the inactive state; however, differences in the ΔGH2O were observed. According to the CD results, compared to cyanomet-FixL, met-FixL was more stable towards chemical denaturation by urea (7.2 vs 4.8kJmol-1). By contrast, electronic spectroscopy monitoring of the Soret band showed cyanomet-FixL to be more stable than met-FixL (18.5 versus 36.2kJmol-1). For the three-state mechanism exhibited by fluorescence, the ΔGH2O for both denaturation steps were higher for the active-state met-FixL than for cyanomet-FixL. The overall stability of met-FixL is higher in comparison to cyanomet-FixL suggesting a more compact protein in the active form. Nonetheless, hydrogen bonding by bound cyanide in the inactive state promotes the stability of the heme domain. This work supports a model of signal transduction by FixL that is likely shared by other heme-based sensors.