Oxygen affinities of DosT and DosS sensor kinases with implications for hypoxia adaptation in Mycobacterium tuberculosis

Abstract

DosT and DosS are heme-based kinases involved in sensing and signaling O₂ tension in the microenvironment of Mycobacterium tuberculosis (Mtb). Under conditions of low O₂, they activate >50 dormancy-related genes and play a pivotal role in the induction of dormancy and associated drug resistance during tuberculosis infection. In this work, we reexamine the O₂ binding affinities of DosT and DosS to show that their equilibrium dissociation constants are 3.3±1.0 μM and 0.46±0.08 μM respectively, which are six to eight-fold stronger than what has been widely referred to in literature. Furthermore, stopped-flow kinetic studies reveal association and dissociation rate constants of 0.84 μM⁻¹s⁻¹ and 2.8 s⁻¹, respectively for DosT, and 7.2 μM⁻¹s⁻¹ and 3.3 s⁻¹, respectively for DosS. Remarkably, these tighter O₂ binding constants correlate with distinct stages of hypoxia-induced non-replicating persistence in the Wayne model of Mtb. This knowledge opens doors to deconvoluting the intricate interplay between hypoxia adaptation stages and the signal transduction capabilities of these important heme-based O₂ sensors.

Graphical Abstract

Graphical Abstract Synopsis

This work reports O₂ affinities of heme-based sensor kinases DosT and DosS, and how their K_d values correlate to distinct stages of hypoxia-induced non-replicating persistence (NRP) in Mycobacterium tuberculosis.

Heme-based sensors play a crucial role in enabling a wide range of microbes to detect and respond to changes in their redox environment [1,2]. Among these, DosT and DosS are two extensively investigated heme-based sensor kinases that play a vital role in the recognition and response mechanism of Mycobacterium tuberculosis (Mtb), the causative agent of tuberculosis (TB), to varying levels of O₂ [3–6]. In low O₂ environments such as those found in TB granulomas, these sensors switch on their histidine kinase activity and initiate the expression of over 50 genes responsible for inducing a state of dormancy (Fig. 1a) [7]. The kinase activity of DosT and DosS is inhibited in their O₂-bound form [8,9]. However, exposure to environments with low O₂ causes these heme-based sensor kinases to lose the O₂ molecule bound to their heme domain and switch to their active ferrous form, resulting in the autophosphorylation of specific amino acid residues, His 392 and His 395 in DosT and DosS, respectively [10]. Remarkably, these heme-based sensors remain active even upon binding to redox-active ligands such as NO and CO that are present in TB granulomas [8,11]. The phosphorylated DosT and DosS subsequently transfer the phosphoryl group to the DosR response regulator, which then binds to DNA to activate dormancy transformation in Mtb [4]. Under dormancy, Mtb’s metabolism significantly slows down rendering it non-replicative and simultaneously resistant to current antibiotic treatments [12].

Figure 1.

a) Function of DosT and DosS sensors in Mtb. In their O₂ bound forms, the kinase domains of sensors are switched off. Under O₂ tension, heme loses O₂ to turn on the kinase domain and phosphorylate His 392 and 395 residues in DosT and DosS, respectively. Phosphorylated forms of DosT/S transfer the phosphoryl group to DosR which binds to DosR regulon and activates dormancy signaling in the bacteria. b) UV-Vis spectral changes in DosT upon binding O₂ at various free O₂ concentrations measured using the optode. c) Difference spectra showing spectral changes when DosT binds O₂. d) O₂ affinity plots for DosT and DosS proteins in dark blue and light blue, respectively (n=3). Gray dashed and dotted lines depict the O₂ concentrations at which NRP-1 and NRP-2 stages are triggered in Mtb, respectively. O₂ concentrations relevant to normoxia, micro-oxia, and nano-oxia are also depicted on the x-axis.

The role of DosT and DosS as two distinct O₂ sensors in the dormancy pathway of Mtb has received considerable attention and has been extensively discussed in literature [3,4,8]. Some studies suggest that the presence of these two sensors aligns with the Wayne’s hypoxia model, which proposes two stages of non-replicating persistence — NRP-1 and NRP-2 [13–16]. Previous research has reported an O₂ K_d value of 26 μM for DosT and a stronger O₂ K_d of 3 μM for DosS, indicating that DosT responds to moderate drops in O₂ levels while DosS responds to deeper hypoxia levels [9]. However, it is noteworthy that NRP-1 occurs at O₂ levels of 2.5 μM (1% O₂ with respect to air saturation) and NRP-2 occurs at O₂ levels of 0.15 μM (0.06% O₂ with respect to air saturation) [17,18], which are significantly different from the reported K_d values for DosT and DosS, respectively. This discrepancy raises questions about the direct correlation between the O₂ sensing capability of these heme-based sensors and the observed stages of non-replicating persistence. Some studies have also suggested that DosT and DosS are switched on much prior to and not during NRP-1 and NRP-2 stages [14]. We note that, while the O₂ K_d value of 3 μM for DosS has been widely referred to in literature, an alternative study has reported a much tighter O₂ K_d of 0.58 μM [19], which better aligns with the O₂ levels for NRP-2. The disparity in reported K_d values highlights the need for reexamining the O₂ K_d values for these heme-based sensors. Ultimately, the O₂ K_d value for DosT and DosS are key parameters that relate their biochemical activity to their physiological function and their accurate determination is crucial to elucidate their biological roles.

To gain a better understanding of how the O₂ affinities of DosT and DosS correlate with Mtb’s hypoxia adaptation, we conducted detailed equilibrium and kinetic investigations of O₂ binding to these heme enzymes. We employed a novel method integrating a UV-Vis spectrometer and an O₂ optode that enables precise O₂ K_d measurements (low nanomolar to high micromolar range) of heme proteins [20]. Briefly, the fraction of O₂-bound heme is monitored via hypsochromic shifts in the heme Soret band, the free dissolved O₂ is simultaneously measured in situ using an O₂ optode, and a simple Hill’s fit with n=1 for fraction O₂-bound protein vs free O₂ concentration yields the K_d value. This approach is superior to titration-based approaches that employ total added O₂ as a measurement parameter which are prone to significant errors from O₂ escaping into the headspace given the 750-fold [21] higher tendency for O₂ to exist as gas than dissolved in water. Starting from ferrous DosT with a Soret maximum at 430 nm, adding increasing amounts of O₂ resulted in systematic hypsochromic shifts with increasing O₂-bound fractions to a Soret maximum of 414 nm corresponding to the ferrous-oxy DosT (Fig. 1b). The O₂-bound DosT fractions are calculated from delta absorbance curves (Fig. 1c) as a normalized sum of the magnitude change in absorbances at 410 and 434 nm. A Hill’s fit with n=1 for the O₂-bound fraction vs the corresponding free O₂ measured using the optode gives a K_d value of 3.3 ± 1.0 μM for DosT (dark blue curve, Fig. 1d) which is 8-fold tighter than what has been previously reported [9]. We also note that this value correlates well with O₂ levels for the onset of NRP-1 (dashed gray line, Fig. 1d) as per the Wayne’s model of hypoxia adaptation in Mtb. Next, we measured O₂ affinity of DosS using the same method (Fig. S3), and we determined a K_d value of 0.46 ± 0.08 μM for DosS (light blue curve, Fig. 1d). This value is 6.5-fold tighter than the widely referred O₂ K_d value of 3 μM for DosS reported by Sousa et al. [9], but matches well with the K_d of 0.58 μM reported by Ioanoviciu et al [19]. Again, our measured O₂ K_d value of 0.46±0.08 μM for DosS correlates well with O₂ levels that drive the onset of NRP-2 (dotted gray line, Fig. 1d). In all, our O₂ affinity measurements for both DosT and DosS reveal that their respective K_d values lie well within the nano-oxic regime (O₂ levels of 0.13 – 6 μM) that are typical for TB granulomas [22]. Next, we conducted stopped-flow kinetic investigations of O₂ binding to DosT and DosS to determine their ligand association and dissociation rates (Fig. S5–6). Upon reacting ferrous DosT with 10 μM O₂, we observe a rather slow association rate (k_on = 0.84 μM⁻¹s⁻¹) that equilibrates to about 60% ferrous-oxy DosT in the hundred millisecond timescale (dark blue curve, Fig. 2). Upon reacting ferrous DosS with 10 μM O₂, we note a faster association rate (k_on = 7.2 μM⁻¹s⁻¹) that goes to near completion in the tens of millisecond timescale (light blue curve, Fig. 2). We note that these k_on values match well with previous reports by Sousa et al. and Ioanoviciu et al [9,19]. Based on k_on values from these kinetic investigations and the K_d values from equilibrium affinity measurements, we calculate the dissociation rate constant (k_off) for DosT and DosS as 2.8 s⁻¹ and 3.3 s⁻¹, respectively.

Figure 2.

O₂ binding kinetics for DosT and DosS proteins shown in dark blue and light blue, respectively (n=3).

In summary, our studies rectify a substantial mischaracterization and misconception regarding the O₂ affinities of DosT and DosS — two important heme-based O₂ sensors responsible for hypoxia adaptation and dormancy signaling in Mtb. We note that the O₂ K_d value of 3.3 ± 1.0 μM for DosT is 8-fold tighter than previous reports and correlates with O₂ levels at the onset of NRP-1 (~2.5 μM O₂). During this early stage, DosT is activated and phosphorylates DosR to kickstart the induction of DosR-regulon which includes dosS, dosR and a host of other genes responsible for driving a metabolic shift towards the non-replicative state [4,23–25]. The induction of DosR regulon elevates DosS levels in the cells. As the O₂ levels drop down to NRP-2 (~0.15 μM) which is lower than the O₂ K_d of DosS (0.45 ± 0.08 μM), DosS transitions to an active form that is capable of phosphorylating DosR and drives further induction of the DosR regulon. Given that the DosR regulon includes both dosS and dosR genes, the activation of DosS creates a positive feedback loop. Such a positive feedback loop embedded in complex cellular signaling networks are known to drive digital responses of gene regulation programs needed for orchestrating adaptive immunity in cells [26–28]. In all, the stagewise role of heme-based O₂ sensors DosT and DosS play an important role in Mtb’s hypoxia adaptation. Previous studies exploring the O₂ K_d of DosT and DosS utilized a direct titration method and measured O₂-binding as a function of total dissolved O₂ added during titration steps [9]. Using total added ligand concentration as a variable for binding affinity assays can work well for soluble non-gaseous ligands and when assay protein concentrations are chosen to be lower than the K_d value to be determined. When the same approach is used for sparingly soluble gaseous ligands like O₂ which have a 750-fold preference for escaping as gas, most of the O₂ added during titration steps can escape to the headspace. This will lead to a major overestimation of total dissolved O₂ concentration and will yield dramatically overestimated K_d values. Given that the O₂ ligand can escape the assay as gas, mere estimations based on titrant buffer concentrations are not enough. In order to accurately determine the O₂ K_d values of DosT and DosS, we employed a method that measures free dissolved O₂ at every titration step using a dip-probe O₂ optode [20]. Applied to the DosT/DosS system, our method of O₂ K_d determination is accurate and establishes a direct correlation between the O₂ sensing capability of these heme-based sensors and the observed stages of non-replicating persistence.

Highlights.

• DosT and DosS are heme proteins that sense O₂ tension in Mycobacterium tuberculosis.

• O₂ affinities of DosT and DosS are 3.3 ± 1.0 μM and 0.46 ± 0.08 μM, respectively.

• These K_d values are 6 to 8-fold stronger than those widely referred to in literature.

Data availability

Data will be made available on request.