Fig 1.
PhoPQ two component system senses low Mg2+ concentrations through direct interactions with PhoQ.
(A) When activated by low external Mg2+ PhoQ undergoes autophosphorylation, transfers phosphoryl group to PhoP which activates transcription of downstream genes. PhoP positively regulates transcription of phoPQ operon, as well as mgrB. MgrB binds PhoQ and suppresses kinase activity. (B) Normalized reporter output from PmgrB saturates as phoPQ operon transcription is increased. (C) Steady state normalized reporter output (PmgrB) plateaus as Mg2+ decreases, but increases further at growth limiting conditions (hypothetical normalized reporter output at growth limiting Mg2+, red square). Plot recreated from [21].
Fig 2.
PhoP-P is robust to overexpression of PhoP, PhoQ.
A—Modified PhoPQ TCS interaction network shows PhoQ binding MgrB, repressing PhoQ autophosphorylation (red arrow marks suppressed rate compared to unbound PhoQ autophosphorylation). This forms the interaction module. The interaction module takes 3 inputs (i) Stimulus (autophosphorylation rate), (ii) Total PhoP (PhoQ is assumed proportional, and 1/40 times PhoP based on actual measurements) and (iii) Total MgrB. Total protein is represented by subscript T in all figures and text. The system is numerically solved for a steady state concentration of PhoP-P as a function of varying PhoP, PhoQ total. The interaction module is coupled with a transcription module representing negative feedback with PhoP-P as input and total MgrB concentration as output. B—The system is solved numerically for steady state concentration of PhoP-P as a function of varying PhoP (and PhoQ) total (blue). PhoP-P is robust to PhoP/PhoQ concentrations, increasing further when PhoQ concentration is large enough to overcome MgrB negative feedback. Over most of the range of PhoQ concentrations, MgrB ≈ MgrB-total indicating large stoichiometric excess MgrB (orange line). Robustness breaks when MgrB is no longer in large excess of PhoQ. Dashed line indicates PhoP concentration estimated from measurements of Ref. [25].
Fig 3.
A two-state model of PhoPQ TCS can explain biphasic response.
A -Schematic of the two-state model. PhoQ exists in phosphatase (PhoQ) or kinase (PhoQ*) form, PhoQ* assumed to bind Mg2+ and switch to PhoQ. Concentration of Mg2+ in medium assumed constant, and absorbed into a pseudo-first order kinetic rate, k−1 (blue arrow). MgrB reversibly binds PhoQ/PhoQ*. B—Simulated output (normalized YFP:CFP; Methods) from the ODE model representing schematic in A with two rate constants suppressed in MgrB bound PhoQ. The pre-factor converts Mg2+ concentration (mM) to rate constant k−1(s−1). The affected rates are denoted by red arrows: switching rate from phosphatase to kinase (i.e. k1, PhoQ-Mg2+ dissociation), and autophosphorylation. Detailed balance condition is satisfied by assuming PhoQ-MgrB dissociation is suppressed by the same factor as k1. Simulated steady state output shows biphasic response to increasing signal.
Fig 4.
Understanding the biphasic dose-response possible in the two-state model of PhoPQ.
A—Reactions in PhoPQ-MgrB network. Dotted squares enclose 4 sub forms of PhoQ (Qph, Qkin, QBph, QBkin). 4 reactions outside the dotted squares have rates comparable to dilution due to growth. Each sub form dilutes with a rate kpd, while synthesis is only in the Q* form. B—Most significant fluxes at high Mg2+. Steady state [PhoP-P] is approximated by matching phosphorylation and dephosphorylation fluxes. Phosphorylation flux is proportional to 1/k−1, while dephosphorylation flux is approximately constant. C- Most significant fluxes at intermediate Mg2+. Phosphorylation flux is proportional to [B]Total and independent of k−1 while dephosphorylation flux is still approximately constant D- [PhoP-P] as a function of signal showing plateau at intermediate Mg2+. Blue line indicates simulated [PhoP-P] from the model in the previous section. Red dashed line shows approximate [PhoP-P] in the high Mg2+ range, and black dashed line shows the approximate plateau value of [PhoP-P] in the intermediate Mg2+ range. E- Fractions of total Q in the 4 catalytic forms.
Fig 5.
Combination of positive and negative feedback increases range of sensitivity to signal.
A—steady state response of simulated promoter output (YFP, normalized to high Mg2+ YFP) for models with no positive feedback (black), no negative feedback (green) and both feedback loops (blue). B- Absolute value of sensitivity (log derivative of PhoP-P with respect to k−1, ) for model with no negative feedback (green) has the shortest range of signal sensitivity, while the wild-type model displays two phases of high sensitivity to signal.
Fig 6.
Removing upregulation of mgrB can result in oscillations in the two component system.
A—Model schematic of an in silico mutant expressing mgrB constitutively instead of being expressed from the PhoP-P dependent promoter PmgrB B—Simulations of PhoP-P following a switch from high (50 mM) to intermediate (1 mM) Mg2+ show oscillations if mgrB is expressed at constant but low levels. Oscillations are absent in wild-type models, as well as models expressing mgrB at a constant high rate.