Identifying and targeting the Mg-Fe-PmrAB regulatory circuit reverses phosphate starvation-driven polymyxin resistance in Enterobacteriaceae

Reviewer #1 (Public review):

This manuscript by Zhang et al addresses how Pi scarcity/depletion drives PMB resistance in Enterobacteriaceae, because it proposes a mechanistically distinct pathway from the better-known PhoBR-linked phospholipid-remodeling responses in other Gram-negatives. The authors also suggest an intervention strategy based on Mg repletion or Fe chelation. The results are substantial and include genetic analyses, mass spectrometry, reporter assays, phospho-signaling readouts, metal quantification, and comparative analyses across enterobacterial species.

The paper reads well with the emphasis on the Mg loss followed by Fe mobilization during Pi depletion that induces PmrAB TCS activation for lipid A modification through transcriptional activation of ugd and arn genes. However, PmrAB is a well-known TCS responsible for PMB resistance through lipid A modification in the extensive studies by the Groisman lab. PmrA is a well-known transcriptional regulator to activate the transcription of the ugd gene in Salmonella and Yersinia by Mg depletion and Fe mobilization. Therefore, the current paper should focus more on the upstream signaling to connect the dots between Pi depletion and Mg loss. This is important because Ugd gene expression is not affected by PmrAB in Pi depletion. It should also be considered that Mg loss is temporally associated with Fe mobilization, but the manuscript does not quantitatively show that Mg dissociation/redistribution is sufficient to trigger Fe mobilization in the absence of Pi depletion, considering that Mg is a macronutrient, whereas Fe is a micronutrient.

Second, the relationship between arn and ugd regulation needs a clearer mechanistic resolution to orchestrate the synthesis of the L-Ara4N during Pi depletion, because the manuscript shows that arn activation is PmrAB-dependent, whereas ugd is only partially PhoBR-dependent and not dependent on PmrAB. Yet the current model and narrative treat the system as a unified "ugd-arn" output. This should be carefully addressed, given that Pi depletion and Mg depletion might trigger different signaling modules.

Third, the manuscript argues that this is a "conserved" circuit in Enterobacteriaceae. The evidence for conservation is presently strongest in E. coli MG1655 and includes supportive observations in E. coli O157, one UTI strain by lipid A MS, several UTI isolates by killing assay, and S. Typhimurium for key phenotypes. No direct mechanistic validation is shown in other important genera belonging to Enterobacteriaceae, which include Klebsiella, Enterobacter, Citrobacter, Yersinia, Serratia, or other clinically important Enterobacteriaceae.

Fourth, the reversal and translational claims are a bit stronger than the current evidence supports. The title and Abstract state that identifying and targeting the circuit reverses Pi depletion-driven PMB, and the manuscript suggests a pharmacological intervention framework based on Mg supplementation or Fe chelation. The actual intervention evidence is limited to in vitro killing assays under acute Pi-depleted minimal-medium conditions in E. coli and S. Typhimurium, without in vivo testing, in that the experiments are performed under an acute 3-hour starvation in MOPS medium, not in host-mimicking or infection-relevant environments. The reversal needs to be shown not only at the level of survival curves, but also by the quantitative MIC/MBC measurements.

More importantly, the authors demonstrated that the signaling module upon Pi limitation in Enterobacteria differs from that in other Gram-negative bacteria such as Pseudomonads. However, they did not discuss why this difference would impact the life of Enterobacteria. The authors should consider the glycolytic pathways (i.e., EMP pathway for enterobacteria vs ED pathway for pseudomonads), in that the ED pathway requires less Pi, whereas the EMP pathway requires more Pi. It should be noted that Pi supply is highly limited in the natural environment for the free-living bacteria, rather than in the host environment for the commensals.

Reviewer #2 (Public review):

Summary:

Using E. coli K-12 as a model system, the authors investigated how phosphate (Pi) depletion induces polymyxin resistance in Enterobacteriaceae, which notably lack the canonical phospholipid remodeling pathways commonly associated with phosphate starvation responses. They demonstrated that low-phosphate conditions promote L-Ara4N modification of lipid A, thereby enhancing polymyxin resistance. Proteomic analyses revealed significant upregulation of the arn operon and ugd under phosphate-limited conditions, and promoter activity assays further confirmed that both promoters are strongly induced during Pi depletion. Through gene deletion experiments, the authors showed that arn expression is regulated by the PmrAB two-component system, whereas ugd is controlled by PhoBR under low-phosphate conditions. Using ICP-MS analysis, they further found that phosphate limitation increases cell-associated Fe levels, and that reducing Fe availability abolishes PmrAB-dependent activation of the arn operon. Finally, the study demonstrated that Mg supplementation and Fe chelation can suppress polymyxin resistance, highlighting the critical role of metal homeostasis in phosphate depletion-induced antimicrobial resistance.

Strengths:

Overall, I found this study to be well conducted, with convincing results that strongly support the proposed model. Through comprehensive genetic analyses and detailed characterization of metal ion homeostasis and membrane lipid modifications, the authors uncovered a novel regulatory connection among Mg²⁺, Fe³⁺, and the PmrAB pathway, a key driver of polymyxin resistance. These findings are highly interesting and have important implications for understanding the evolution of the Fe-sensing PmrAB system, as well as the broader role of nutrient availability in shaping antibiotic resistance.

Weaknesses:

I did not identify any particular weaknesses.

Reviewer #3 (Public review):

Summary:

This manuscript examines how phosphate limitation primes E. coli and Salmonella for defense against polymyxin antibiotics. Other environmental signals, such as altered levels of extracellular Mg or Fe, were previously shown to induce polymyxin resistance in Enterobacteriaceae, and phosphate limitation was known to augment polymyxin resistance in other organisms such as A. baumannii and P. aeruginosa; however, whether phosphate limitation boosted polymyxin resistance in Enterobacteriaceae was not known. This study shows that this indeed occurs, and the mechanism is distinct from that in A. baumannii and P. aeruginosa. The model proposed is: (1) low phosphate causes bacteria to jettison Mg to balance cellular P/Mg ratio, (2) extracellular Fe3+ associates with the cell envelope to replace Mg as LPS-bridging cation, and (3) envelope Fe3+ activates PmrAB, which mediates a transcriptional response leading to L-Ara4N modification of lipid A and protection from polymyxin B. Flooding with Mg or chelating the surface Fe3+ blocks the protective response to low phosphate in E. coli and Salmonella but not in P. aeruginosa despite Fe still mobilizing in the latter. The differential response between Enterobacteriaceae and P. aeruginosa is connected to the presence/absence of Fe-sensing motifs in the PmrB periplasmic domain.

Strengths:

The strengths of the study are the wide array of approaches used and the thorough characterization of a novel stress-response mechanism involving metal mobilization. Combined with the analysis of multiple bacterial families, the results clarify how different strategies have evolved to defend against polymyxins during phosphate starvation.

Weaknesses:

Controls are needed in some of the genetic experiments, namely complementation, to verify linkage of defective survival phenotypes to the genes mutated and to rule out protein stability defects for the PmrB variants tested. In addition, the generalizability of the metal mobilization feature of the model would be strengthened by examining media with differing metal composition. Claims about antibiotic resistance would be strengthened by data examining bacterial growth in the presence of an antibiotic.