NOTE-656: A. baumannii PmrB oxidative-stress response memory

Contribution

The paper proposes that a two-component signal transduction system does three things not previously ascribed to one together: senses oxidative stress through a metal cofactor, and specifically a nickel cofactor held by a histidine box in the sensor’s periplasmic domain; converts single-electron oxidation of that nickel into a conformational change that transduces the signal; and encodes a short-lived “response memory” whereby a sublethal priming dose keeps the pathway active after the stimulus is gone, so the cell survives a later lethal dose of peroxide or antimicrobial peptide. It ties all three to virulence in a mouse lung model and to hypervirulence in a clinically isolated carbapenem-resistant strain. If it holds up, the nickel-on-histidines oxidative sensor would be a new sensing chemistry for this protein family, and the “response memory” would be a concrete molecular case of signal priming in a bacterial pathogen.

Key insight

A sensor kinase can be turned into a redox detector by hanging a redox-active metal (Ni2+/Ni3+) on a cluster of histidines in its sensing domain: oxidation of the metal by reactive oxygen species reshapes the domain and flips the pathway on, and the pathway then stays on long enough (tens of minutes) to act as a memory of a stress the cell has already seen. The single thing to retain is the proposed mechanism “sublethal ROS oxidises a nickel cofactor on a histidine box, which switches the sensor and primes defence for a short window” — and that this chain is assembled from binding assays, reporters and simulation rather than observed end to end.

Assumptions

• A. baumannii lacks the conventional defences. The whole motivation rests on the stated absence of RpoS, Nif and Suf homologs (Appendix dataset S1, not inspected here), with the Isc system as the sole Fe-S assembly machinery.
• Reporter and binding assays stand in for transcription. Gene regulation is argued from luciferase fusions, putative PmrA-box motifs, and EMSA, not from direct measurement of the native transcripts across conditions.
• Simulation stands in for conformational measurement. The redox switch and the metal pocket come from AlphaFold3 (with Fe2+ substituted for Ni2+, since AlphaFold3 lacks a Ni2+ ligand) and from MD with bonded metal models, not from an experimental structure of Ni2+/Ni3+-PmrB.
• Finite, defined conditions. Metal-specificity and sensing experiments use M9 minimal medium stripped of metal cations; the physiological Ni2+ pool available to PmrB in the host is not measured.

Key results

• PmrA/PmrB is required for oxidative-stress survival. pmrA/pmrB/double nulls fail to grow under H2O2 with no defect absent stress; plasmid copies rescue, empty vector does not; same pattern under the NO generator spermine NONOate (Appendix). oxyR inactivation does not impair H2O2 survival, unlike pmrA. (Fig. 1B—C, Appendix Fig. S3.)
• The effect runs through the Fenton reaction. Ferrous chelator DP and ferric/ferrous chelator DFO restore the pmrA mutant under H2O2; thiourea (HO· scavenger) restores growth; streptonigrin (kills by cytoplasmic free Fe2+) plus H2O2 kills more, and DFO removes the WT/pmrA difference. (Fig. 1D—G.)
• PmrA directly activates named defence genes. Putative PmrA boxes, ~6-fold H2O2-induced (colistin-not-induced) luciferase from the hscB-hscA-fdx (Isc) promoter, and EMSA with competition, repeated for ftnA (ferritin), ahpF1 (peroxidase) and katE (catalase); katG is a negative control not bound by PmrA. Inactivating ahpF1/hscA/ftnA/katE reduces H2O2 survival. OxyR and PmrA are argued to be independent (OxyR represses ahpF1 basally; PmrA acts only under stress). (Fig. 2.)
• Periplasmic domain senses oxidative stress; TM domain senses colistin. A periplasmic-deletion variant (PmrB1, Δ70—130) loses H2O2 survival but keeps colistin resistance; a TM1-deletion variant (PmrB2, Δ23—29) keeps H2O2 but loses colistin; both still localise to the inner membrane. (Fig. 3B—D.)
• Four conserved histidines are required for sensing, not for colistin. Ala substitution of any or all of His86/89/90/92 lowers H2O2 survival but not colistin resistance; a neighbouring Ser88Ala control still rescues; hscA/ahpF1 induction is lost in the quadruple mutant (Appendix). The histidines are Acinetobacter-specific; Salmonella/Pseudomonas PmrB (no histidine box) cannot restore H2O2 resistance but can restore colistin resistance. (Fig. 3E—F, Appendix Figs. S6—S8.)
• Nickel is the cofactor. Ni2+ (not Co2+/Mn2+/Zn2+) supplementation raises hscB-hscA-fdx and ahpF1 induction and raises H2O2 resistance in WT but not the pmrB null; ICP-MS and a fluorescent dye detect Ni2+ on WT purified periplasmic domain (aa 30—140) but not on the histidine-box mutant or BSA. AlphaFold3 models His89/90/92 as direct metal binders, His86 as a stabiliser. (Fig. 4.)
• Oxidation is modelled as a conformational switch. MD (apo/Ni2+/Ni3+, 100 ns): Ni3+ raises periplasmic-domain RMSD to ~5.5 Å (~40% above apo/Ni2+ ~4 Å) while keeping low binding-site RMSD; PCA has Ni3+ PC1 capturing 49.0% of variance (vs 38.3% Ni2+, 30.4% apo) with a directed initial→final path; differential contact maps show loop displacement at ~75—85 and expansion between ~70—75 and ~90—130 with compaction at ~30—50 and ~55—78, the histidine box at the near-zero boundary. Read as a redox-driven switch that preserves the coordination shell. (Fig. 5.)
• Response memory. Sublethal H2O2 (1—10 µM, 1 h) raises oxidative-defence and AMP-resistance genes (pmrC, naxD) and improves later survival to 30 mM H2O2 and 3 µg/mL colistin, PmrA-dependent, OxyR-independent, abolished in the histidine-box mutant. Priming persists 30—90 min after signal removal; CFU counts are unchanged over that window, offered as evidence the decay is not dilution by division. (Figs. 6—7, Appendix Figs. S11—S14.)
• Virulence and clinical relevance. Intratracheal mouse infection: pmrB mutant → less lung inflammation, fewer neutrophils (Ly6gCre;R26LSL-Tomato reporter), lower lung CFU; WT PmrB restores, the histidine-box mutant does not. Priming raises pmrC and LL-37 survival PmrA-dependently; survival in J774A.1 macrophages needs pmrA/pmrB and the histidine box. The histidine box is present in hypervirulent CRAB A0062 and absent in low-virulence A0075; A0075 is killed by 30 mM H2O2 while A0062 resists up to 120 mM; pmrB disruption in A0062 removes priming-dependent survival. (Figs. 8—9.)

Claims

id	claim	strength	support
C1	PmrA/PmrB, not OxyR/RpoS, is the principal regulator of oxidative-stress defence in A. baumannii, acting by lowering Fenton-reaction damage	moderate	knockout/complementation + chelator/scavenger epistasis (Figs. 1—2); genetic/pharmacological, not a direct ROS measurement
C2	PmrA directly activates hscB-hscA-fdx, ftnA, ahpF1, katE under oxidative stress	moderate	reporters + EMSA with competition (Fig. 2); in-vitro binding and fusions, not native-transcript profiling
C3	Four periplasmic histidines (His86/89/90/92) are required for oxidative but not colistin sensing	moderate-to-strong	domain and point mutants, cross-species swaps (Fig. 3, Appendix)
C4	PmrB uses a Ni2+ cofactor on the histidine box to sense oxidative stress	moderate	Ni2+-specific induction/survival, ICP-MS/dye on purified domain, AlphaFold3 model (Fig. 4); cofactor identity rests on correlative detection + modelling, no holo structure
C5	Single-electron oxidation Ni2+→Ni3+ drives a directional conformational switch of PmrB	weak-to-moderate	MD + PCA + contact maps (Fig. 5); entirely in silico
C6	Sublethal ROS leaves a 30—90 min PmrA/histidine-box-dependent “response memory” that primes cross-protection, not explained by growth dilution	moderate	priming reporters/survival + CFU control (Figs. 6—7)
C7	Oxidative sensing via the histidine box is required for virulence in mice and for hypervirulence in clinical CRAB	moderate	mouse burden/inflammation + two clinical isolates + one insertional mutant (Figs. 8—9); small clinical sample

Method

Bacterial genetics in A. baumannii ATCC 17978 (λ-Red/FLP chromosomal deletions, plasmid complementation in trans, 3xFLAG tagging); luciferase reporter fusions in defined M9 medium with/without metal cations; EMSA with purified PmrA and competitor DNA; purification of the PmrB periplasmic domain for ICP-MS and fluorescent Ni2+ detection; subcellular fractionation via NADH for localisation; structure prediction with AlphaFold3 (Fe2+ as Ni2+ proxy); MD of apo/Ni2+/Ni3+ PmrB with bonded metal models (RMSD, RMSF, SASA, radius of gyration, PCA, differential contact maps); priming/memory assays (sublethal H2O2 then lethal H2O2 or colistin, with CFU controls); intratracheal mouse infection with neutrophil-reporter imaging; macrophage (J774A.1) survival; and assays on clinical isolates A0062 (hypervirulent CRAB) and A0075 (low-virulence).

Concepts

• response memory — a signalling state that stays “on” after the initiating signal is removed, priming a stronger response to a later signal; here a 30—90 min window following sublethal ROS.
• histidine box — four conserved periplasmic histidines (His86/89/90/92) of Acinetobacter PmrB, proposed to coordinate the nickel cofactor.
• standard (sublethal) priming dose — 1—10 µM H2O2, said to match bloodstream/airway ROS levels at initial infection sites.

Connections

The paper positions PmrA/PmrB against the canonical OxyR/RpoS oxidative-stress regulators and against Salmonella/Pseudomonas PmrB (which confer colistin resistance but, lacking the histidine box, no oxidative sensing). It likens the proposed Ni-on-histidines redox switch to the Fe2+-on-histidines sensing of Bacillus PerR, and the “response memory” to IFN-γ priming of mammalian immune cells. Machine-readable lineage belongs on the LIT, not here.

Bearing on the record

• THEORY-136 (Active) is the entry this reading most directly touches. THEORY-136 holds that a lasting or abrupt change is weak evidence for a system with alternative stable states, and that only hysteresis, initial-state dependence, or a lasting shift after a temporary disturbance with slow return excluded comes close to showing one. The paper’s “response memory” is precisely a lasting-shift-after-temporary-disturbance claim, and it supplies one control THEORY-136 asks for — CFU unchanged over the memory window, excluding dilution-by-division — while not demonstrating bistability or true hysteresis (no initial-state-dependence test, no dose loop). So it is a biological instance that meets THEORY-136’s bar only partway: evidence of a persistent response, not yet of an alternative stable state.
• THEORY-058 (Proposed) — mental disorders maintained by self-reinforcing loops that let them outlast their triggers — is a looser analogue: a response that outlasts its cause through pathway-internal dynamics rather than a sustained driver. The connection is conceptual, cross-domain, and does not change that THEORY.
• THEORY-130 (coral bleaching as breakdown of a host-symbiont exchange under oxidative stress) shares the vocabulary of oxidative stress in a host-associated organism but concerns a different system and mechanism; no real bearing.
• No THEORY is filed from this reading. The claim one would state — that a bacterial TCS encodes an oxidative-stress memory via a nickel-histidine switch — rests on a single unrefereed preprint whose mechanism is assembled from reporters, correlative metal detection and simulation, and is exactly the kind of persistence claim THEORY-136 counsels holding at arm’s length.
• No instruction for machine-learning practice; nothing for the anthology. AlphaFold3, MD and PCA appear only as analysis tools.

Limitations

• Mechanism not observed end to end. No holo structure of Ni2+- or Ni3+-PmrB; the cofactor identity is correlative (ICP-MS, dye) plus an AlphaFold3 model that substitutes Fe2+ for Ni2+; the redox switch is wholly in silico (100 ns MD, single system each).
• Regulation inferred from reporters and in-vitro binding, not native transcript measurement across the regulon.
• Short and narrow memory evidence. 30—90 min window; no initial-state-dependence or reversibility/hysteresis experiment; the dilution control addresses only one alternative explanation.
• Thin clinical base. Two clinical isolates and one insertional mutant underwrite the hypervirulence-determinant claim; histidine-box presence/absence is correlated with virulence across a tiny sample.
• Unrefereed. Not peer reviewed; figures read as rendered in the PDF; appendix figures and datasets S1—S3 underpinning several statements (absence of RpoS/Nif/Suf, metal specificity, localisation) were not opened.

Open questions

• A holo structure (or spectroscopy) of Ni-bound PmrB and a direct redox measurement of the Ni2+→Ni3+ transition in the protein — would convert C4—C5 from correlative/computational to shown.
• A hysteresis or initial-state-dependence experiment on the response memory — would decide whether it is bistability (an alternative stable state, in THEORY-136’s sense) or slow relaxation.
• Native-transcript profiling of the PmrA regulon under graded ROS — would test C2 beyond reporters and EMSA.
• Whether the histidine box is a general virulence determinant across clinical Acinetobacter — needs more than two isolates.