# Fileset

[GRANDFINAL-Manuscript.pdf](https://mdr.nims.go.jp/filesets/8a08404e-d3f4-4085-8f51-934f60cf542e/download)

## Creator

John A. Ciemniecki, Chia-Lun Ho, Richard D. Horak, [Akihiro Okamoto](https://orcid.org/0000-0002-8102-4316), [Dianne K. Newman](https://orcid.org/0000-0003-1647-1918)

## Rights

[Creative Commons BY-NC-ND Attribution-NonCommercial-NoDerivs 4.0 International](https://creativecommons.org/licenses/by-nc-nd/4.0/)

## Other metadata

[Mechanistic study of a low-power bacterial maintenance state using high-throughput electrochemistry](https://mdr.nims.go.jp/datasets/d3b78f65-a02f-41e5-bd18-94240df988b7)

## Fulltext

GRANDFINAL-Manuscript  1 Mechanistic study of a low-power bacterial maintenance state using high-1 throughput electrochemistry 2  3 John A. Ciemniecki1, Chia-Lun Ho2,4, Richard D. Horak1, Akihiro Okamoto2,3,4,5*, Dianne K. 4 Newman1,6* 5  6 1 Division of Biology & Biological Engineering, California Institute of Technology, Pasadena CA 7 91125 8 2 Research Center for Macromolecules and Biomaterials, National Institute for Materials Science, 9 1-1 Namiki, Tsukuba, Ibaraki 305-0044, Japan 10 3 Faculty of Life and Environmental Sciences, University of Tsukuba, 1-1-1 Tennodai, Tsukuba 11 Ibaraki 305-8577, Japan  12 4 School of Chemical Sciences and Engineering, Hokkaido University, 13 Kita, 8 Nishi, Kita-ku, 13 Sapporo, Hokkaido 060-8628, Japan. 14 5 Living Systems Materialogy (LiSM) Research Group, International Research Frontiers Initiative 15 (IRFI), Tokyo Institute of Technology, Yokohama, Kanagawa 226-8501, 669 Japan. 16 6 Division of Geological & Planetary Sciences, California Institute of Technology, Pasadena CA 17 91125 18  19 *co-corresponding authors: dkn@caltech.edu, okamoto.akihiro@nims.go.jp 20  21 SUMMARY 22  23 Mechanistic studies of life’s lower metabolic limits have been limited due to a paucity of tractable 24 experimental systems. Here we show that redox-cycling of phenazine-1-carboxamide (PCN) by 25 Pseudomonas aeruginosa supports cellular maintenance in the absence of growth with a low 26 mass-specific metabolic rate of 8.7 x 10-4 W (g C)-1 at 25˚C. Leveraging a high-throughput 27 electrochemical culturing device, we find that non-growing cells cycling PCN tolerate conventional 28 antibiotics but are susceptible to those that target membrane components. Under these 29 conditions, cells conserve energy via a noncanonical, facilitated fermentation that is dependent 30 on acetate kinase and NADH dehydrogenases. Across PCN concentrations that limit cell survival, 31 the cell-specific metabolic rate is constant, indicating the cells are operating near their 32 bioenergetic limit. This quantitative platform opens the door to further mechanistic investigations 33 of maintenance, a physiological state that underpins microbial survival in nature and disease.  34   2 INTRODUCTION 1  2 The physiology of bacteria in non- or slowly-growing states is poorly understood relative 3 to its ubiquity in natural, clinical, and engineered systems1,2. A fundamental question underlying 4 the study of this state is how much energy is needed to maintain cell viability in the absence of 5 growth3. Previous laboratory attempts to quantify this value inferred maintenance metabolic rates 6 by extrapolating the energy needed to support zero growth from values measured during slow 7 growth in continuous culture4. Yet these estimates are consistently two to four orders of magnitude 8 higher than metabolic rates measured in stable natural environments where sources of energy or 9 nutrients are limiting5. How can we bridge the gap between the two to study the cellular strategies 10 underpinning non-growth maintenance metabolism? 11 While environmental measurements inspire an interest in this attenuated energetic state, 12 it is important to note that metabolic rates estimated from environmental samples, by necessity, 13 also rest on inferences rather than direct measurements. First, it is assumed that all the cells 14 counted in the sample are metabolically active. Second, the fraction of cells using the measured 15 nutrient is often estimated using educated guesses. Third, it is unclear which percentage of cells 16 are in a normal vs. viable-but-non-culturable (VBNC) state with a diminished metabolism6. 17 Overestimates of any of these parameters can inflate the relevant cell count, leading to 18 underestimates of the cell-specific metabolic rate. Accordingly, the development of laboratory 19 systems in which to study metabolic rates at or near zero growth coupled to measurements of 20 single-cell metabolic activity is needed7. While progress is being made towards defining new 21 systems through which to understand the biology of growth arrest8–10, to our knowledge, no high-22 throughput experimental systems have been established that permit direct measurements of 23 metabolic power output during this state. 24 A powerful motivation for doing so comes from the fact that non-growing cells inhabit the 25 cores of biofilms, multicellular aggregates found in many chronic infections. Most antibiotics used 26 in the clinic fail to eradicate biofilms, contributing to the deaths of millions of people annually11,12. 27 Cells within biofilms tolerate many conventional drugs11, and tolerance can beget antibiotic 28 resistance13. The mechanisms driving this tolerance are complex14–17; biofilm-associated 29 metabolic changes are one important factor. The ability of diverse biofilm-forming bacteria to 30 withstand standard antibiotics is not surprising because screens for novel antibiotics have mainly 31 been performed using growth-based assays18 and most antibiotics target the cellular processes 32 that sustain actively growing cells, such as cell wall synthesis, translation, and DNA replication. 33 While conventional antibiotics are highly effective at inhibiting or killing fast-growing bacterial 34   3 pathogens that have not acquired resistance, the ability of biofilm bacteria to tolerate their 1 presence remains an unsolved and medically important challenge11. Quite simply, we struggle to 2 eradicate cells that exist in attenuated metabolic states because we do not have a mechanistic 3 understanding of maintenance physiology due to a scarcity of tractable experimental systems. 4 We hypothesized that anaerobic extracellular electron transfer (EET) could be leveraged 5 to measure the metabolic flux powering maintenance (i.e. the minimum energy needed to support 6 cellular integrity and preserve a culturable state in the absence of growth) and investigate its 7 bioenergetic underpinnings. Pseudomonas aeruginosa, an opportunistic pathogen and model 8 biofilm-forming organism, produces phenazines (C12H8N2R, e.g. Figure 1B right), a widespread 9 class of colorful, heterocyclic, redox-active metabolites19. Phenazines promote biofilm 10 development and attenuated metabolism (sometimes referred to as “dormancy”) in biofilm regions 11 where oxygen is unavailable, rendering cells physiologically tolerant to conventional drugs20,21. 12 Previously, using low-throughput electrochemical reactors, we showed that the phenazines 13 pyocyanin, phenazine-1-carboxylic acid and 1-hydroxy phenazine act as extracellular electron 14 shuttles that facilitate an anaerobic survival metabolism22, one that supports survival but not 15 growth of the bulk population. While P. aeruginosa cannot survive anaerobically via glucose 16 fermentation23, when provided with phenazines and a re-oxidizing potential, intracellular redox-17 balance is achieved, permitting glucose catabolism and substrate-level phosphorylation (SLP) via 18 the AckA-Pta pathway24 (Figure 1B). In this way, ATP is generated via a noncanonical form of 19 energy conservation where reduction of the extracellular electron shuttle facilitates SLP and 20 organic acid excretion24, similar to energy conservation strategies found in other bacteria in the 21 presence of humic substances or flavins25–29. Given this fermentative energy conservation 22 pathway requires redox-balancing using an exogenous oxidant, we refer to it as a facilitated 23 fermentation (Figure 1A). Because it supports survival without growth, we reasoned that the rate 24 of this facilitated fermentation may reveal the metabolic flux necessary for cell maintenance under 25 these conditions. 26 In this study, we investigated phenazine EET survival metabolism using a recently 27 developed high-throughput electrochemical system that enables nearly 100 measurements of 28 current generated by planktonic cultures to be measured concurrently30. By leveraging this system 29 to survey a wide range of conditions and directly measure power output, we asked: What is the 30 metabolic rate of phenazine EET-mediated survival and how does it compare to the basal 31 metabolic rates measured for other organisms across the tree of life? At the single-cell level, are 32 cells surviving without growth in a maintenance state or is the population surviving via a balance 33 of growth and death? Which types of antibiotics can kill cells in this physiological state? Which 34   4 enzymes contribute to its catabolism? And does phenazine EET power cell survival at or above 1 the maintenance power requirement? Broadly, our results introduce a tractable experimental 2 platform that permits quantitative and mechanistic investigations into maintenance physiology. 3 More specifically, they shed light on the minimal metabolic requirements and pathways that 4 contribute to powering P. aeruginosa cells under anoxic conditions relevant to the antibiotic 5 tolerant core of biofilms. 6  7 RESULTS 8  9 To study phenazine-dependent survival metabolism at the lowest thermodynamic potential 10 possible for the metabolism, we performed our studies using phenazine-1-carboxamide (PCN), 11 the phenazine derivative with the lowest standard midpoint potential (-140 mV E˚’, 31) of those 12 produced in significant quantities by P. aeruginosa32,33. Due to its low potential, PCN reduction 13 should permit EET-based survival metabolism with the smallest possible energy conserved per 14 phenazine molecule reduced. In biofilms, PCN is the phenazine produced and retained at the 15 highest levels33, making it of additional interest to our understanding of cell survival. In all 16 experiments, a ∆phz1/2 mutant was used that cannot synthesize phenazines, and PCN was 17 exogenously provided at physiologically-relevant concentrations. Data analysis details and the 18 rationale for using electrons sec-1 units, as opposed to Watts, when reporting metabolic rates are 19 provided in the Supplemental Information. 20  21 PCN EET promotes anaerobic survival at extremely low metabolic rates 22  23 We first conducted weeklong PCN anaerobic survival assays as previously reported22 to 24 compare survival phenotypes on PCN to those previously made for other phenazines and make 25 metabolic rate measurements. Cells were incubated anaerobically in large, well-mixed, custom 26 glass electrochemical reactors connected to potentiostats poised to re-oxidize the PCN reduced 27 by the cells (Figure 1C). Samples were removed periodically from the reactors to measure colony 28 forming units (CFUs). After one week of incubation at 33˚C, cultures provided with PCN and a re-29 oxidizing potential (+PCN/pot) lost less than an order of magnitude of their initial viability while 30 anaerobic cultures incubated in the same medium -PCN/pot lost viability by three orders of 31 magnitude (Figure 1D). At the end of the incubation, we measured the cumulative acetate 32 produced by the cells (as a consequence of the oxidative arm of the facilitated fermentation, 33 pyruvate Þ acetate, Figure 1B), finding it to be 122±34 µmoles (n=2), approximately balanced 34   5 with the total PCN reduced (the reductive arm), 107±4 µmoles (n=2), as expected (See 1 Supplemental Information for details). The amount of acetate produced was also an order of 2 magnitude higher in the +PCN/pot than the -PCN/pot cultures (Figure 1E), in agreement with our 3 working model (Figure 1B). To assess the percentage of cells that were metabolically active in 4 these different conditions, we stained culture samples on day 5 with the fluorescent viability dye 5 propidium iodide (PI is taken up when cells are metabolically and/or physically compromised), 6 finding that the majority (88%, n=794 cells) of -PCN/pot cells fluoresced brightly whereas very few 7 (2.4%, n=830 cells) +PCN/pot cells were fluorescent (Figure 1F). Notably, the population 8 sustained by PCN cycling appeared to be homogeneous with respect to PI staining. Taken 9 together, these results are consistent with PCN EET powering a facilitated fermentation that 10 enables anaerobic survival, similar to what we have observed for other phenazines22. 11 Recently, Hoehler et al.34 collated a large dataset of metabolic rates measured across all 12 life on Earth to estimate the average mass-specific basal metabolic rate of the biosphere. We 13 wondered how the metabolic rate of PCN survival would compare to both bacterial measurements 14 and other organismal measurements in their published dataset. We summarize our findings here, 15 with a more detailed accounting of our analysis of our samples and the Hoehler et al. dataset 16 found in the Supplemental Information. Consistently, the current produced by the +PCN/pot 17 cultures during their survival is characterized by an initial spike, followed by a decrease in current 18 to a steady state held between days one and seven (Figure 2A). Given the spike occurs before 19 the onset of death in the -PCN/pot condition, this likely reflects adjustment of the cells to a new 20 condition and thus we interpret the lower, stable current as the survival metabolism. Averaging 21 this current and the associated CFUs across days one to five, we estimated the bulk cell-specific 22 metabolic rate of the surviving cells to be 1.6 x 103 electrons sec-1 cell-1 at 25˚C (lower 95% CI 8.6 23 x 102, upper 3.1 x 103). To transform units of cell-1 to (g C)-1, we measured the average cell dry 24 weight of +PCN/pot cells. Contrary to expectations for cell size when challenged by extreme 25 energy limitation7,35 the cells were not especially small; we measured their dry weight on day five 26 to be 2.0±0.1 x 10-13 g cell-1, comparable to measurements of various species in the early 27 stationary phase of growth36. Assuming a ∆G’˚ = -65.6 kJ (mol PCN reduced)-1 (see Supplemental 28 Information), our estimate of the cell-specific metabolic rate is equal to a mass-specific power of 29 8.7 x 10-4 W (g C)-1 (lower 95% CI 4.7 x 10-4, upper 1.7 x 10-3). This value is about one order of 30 magnitude below the average basal metabolic rate across all organisms reported by Hoehler et 31 al., 1.2 x 10-2 W (g C)-1. Under the conditions of our assay, the PCN EET mass-specific metabolic 32 rate falls within the lower 2% of measurements across all organismal groups, and among those 33 groups, only the gelatinous invertebrates are noticeably lower34. An alternative analysis that 34   6 includes the contribution of the initial current spike via an integration approach (across days 0 to 1 5) resulted in a slight increase in this estimate with an average cell-specific metabolic rate of 1.7 2 x 103 electrons sec-1 cell-1 and a corresponding mass-specific power of 9.3 x 10-4 W (g C)-1. 3 We then sought to further contextualize this metabolic rate across the hundreds of 4 bacterial measurements in the dataset. To avoid the accumulation of potential errors stemming 5 from assumptions about how much energy is conserved per electron transferred, we analyzed 6 the cell-specific metabolic rates in units of electrons sec-1 cell-1. Notably, across disparate 7 catabolic pathways the ratio of ATP generated per electron transferred varies by less than an 8 order of magnitude, a fact that to our knowledge has not been recognized. For example, the PCN 9 survival metabolism using glucose has an energy conservation ratio of 0.5 ATP electron-1 while 10 aerobic catabolism of glucose is 1.3 ATP electron-1 (Supplemental Information). Given these 11 values represent disparate metabolisms yet do not deviate far from unity, we conclude that within 12 this dataset that is dominated by measurements of glucose metabolism and aerobic respiration 13 rates, units of electrons have an approximate equivalency to ATP units of energy, and therefore 14 transformed the values originally reported in the primary publications collected by Hoehler et al. 15 to units of electrons sec-1 cell-1 (see Supplemental Data Table for full dataset). 16 Within the categories of bacterial measurements, those reporting on endogenous 17 metabolism, or the oxidation of endogenous carbon sources, contained the lowest and most 18 heterogeneous values; the PCN EET cell-specific metabolic rate was in the lower 15% of them 19 (Figure 2B). The maximum cell-specific metabolic rate of P. aeruginosa growing aerobically on 20 glucose has been measured to be approximately 106 electrons sec-1 cell-1 37,38, three orders of 21 magnitude higher than PCN EET, revealing that this organism has a remarkably plastic metabolic 22 pace. We also found that the cell-specific metabolic rate of PCN EET, while supporting the 23 maintenance of the population (Figure 1D), was nonetheless 2-3 orders of magnitude lower than 24 maintenance metabolic rates inferred from continuous culture studies. This finding supports those 25 of prior studies showing the metabolic rates of microbes in oligotrophic natural environments are 26 significantly lower than the maintenance requirements predicted from continuous culture5. We 27 conclude that PCN facilitated fermentation occurs at an extremely low cell-specific metabolic rate, 28 a facet of this metabolism that was unappreciated previously. 29  30 PCN facilitated fermentation powers a non-growth maintenance state 31  32  Because PCN metabolism is positioned at the lower end of measured metabolic rates 33 documented for any organism, we wondered whether the weeklong population survival phenotype 34   7 promoted by PCN EET reflected a non-growing survival state or a dynamic balance of growth and 1 death, akin to the growth advantage in stationary phase phenomenon observed in E. coli under 2 nutrient limitation39. To distinguish between these possibilities, we performed a pulse-chase 3 experiment using the fluorescent cell wall stain 7-hydroxycoumarincarbonylamino-D-alanine 4 (HADA), a stain that has been previously used in P. aeruginosa to monitor cell growth40. During 5 bacterial growth, HADA is covalently incorporated into the cell wall, and after unincorporated 6 HADA is washed away, new growth can be visualized as a loss of fluorescence in the non-pole 7 regions of the cell41 (Figure 3A). As a positive control, we checked to see whether we could 8 observe HADA loss during anaerobic growth on nitrate42. As expected, HADA fluorescence was 9 lost in a topological manner consistent with non-pole incorporation of new, non-fluorescent cell 10 wall material (Figure S1) and was mostly diminished save for a few dimly fluorescent poles by 11 day four (Figure 3B). HADA distribution over the cell body was unchanged after four days in the 12 absence of nitrate (Figure 3B, S1). The presence of HADA in the cell wall did not affect the growth 13 rate of the culture (Figure 3C). 14  Having validated HADA for P. aeruginosa in our medium, we applied it to observe cells in 15 our electrochemical reactors. HADA-stained cells in the +PCN/pot condition retained HADA 16 throughout the entirety of their cell wall over the course of seven days (Figure 3D, S1). Notably, 17 some cells that had not finished dividing during the inoculum transfer from oxic to anoxic 18 conditions had HADA-labeled septa that remained intact even after the seven days of anaerobic 19 incubation (Figure 3D, inset). The presence of HADA did not affect the survival or viability loss of 20 the +PCN/pot and -PCN/pot cultures, respectively (Figure 3E). Lastly, the HADA-labeled and 21 unlabeled poles of at least 400 cells from each day and condition were counted to estimate the 22 number of divisions across the population (Figure 3A). These values for each day were close to 23 zero (Figure 3F). Our results indicate that cells are statically surviving without growth during 24 anaerobic PCN EET, and we therefore conclude that facilitated fermentation, under these 25 conditions, powers a non-growth maintenance state, a fact that hitherto was ambiguous. 26  While imaging cells from the -PCN/pot condition, we noticed in phase contrast that a 27 significant number of cells displayed dark foci in their cytoplasm (Figure 3D), reminiscent of 28 protein aggregates associated with ATP depletion and the viable but non-culturable (VBNC) 29 state43. To test if this state explained the loss of cell viability in the -PCN/pot condition, throughout 30 the anaerobic experimental time course, we subsequently incubated a sample of -PCN/pot cells 31 in aerobic resuscitation conditions for two days before plating (see methods). At day seven of the 32 anaerobic incubation, the CFU counts from the resuscitated cells were about two orders of 33 magnitude higher than the counts from cells that were immediately plated (Figure S2B). The 34   8 remaining order of magnitude discrepancy between the +PCN/pot and -PCN/pot conditions, 1 representing most of the population, was not recovered (Figure S2A). Consistent with this finding, 2 cell pellets of the cultures on day seven showed a clear size disparity between the +PCN/pot and 3 -PCN/pot condition, indicating significant amounts of cell lysis had occurred in this sample during 4 anaerobic incubation (Figure S2C). We also confirmed that cells in the -PCN/pot condition were 5 ATP-depleted relative to the +PCN/pot condition (Figure S2D), consistent with the presence of 6 cells in the VBNC state43. We conclude that the majority of CFU loss in the -PCN/pot condition is 7 caused by death, with a minor subpopulation entering a VBNC state. 8  9 A high throughput platform can advance our understanding of non-growth metabolism 10  11 While the preceding large reactor experiments enabled us to determine the pace of PCN 12 EET metabolism relative to life’s average basal metabolic rate and confirm it relies on a facilitated 13 fermentation that fuels a non-growth metabolism, large reactors are cumbersome and low-14 throughput, limiting our ability to explore diverse conditions and mutant strains—variables that are 15 essential to study when seeking a mechanistic understanding of a biological phenomenon. 16 Recently, a 96-well electrochemical system was developed30. This system employs printed carbon 17 electrodes on the bottom of each well, allowing for independent, multiplexed potentiostatic 18 measurements (Figure 4A). To determine whether we could conduct PCN anaerobic survival 19 assays in similar ways between the large and high throughput reactors, we compared survival 20 phenotypes quantitatively. Under the same conditions (i.e. medium and PCN concentration) in the 21 high throughput platform, cells survived several orders of magnitude better in the presence of 22 PCN and an oxidizing potential, as observed for the large reactors, yet we were additionally able 23 to test many other concentrations of PCN because of the increased throughput (Figure 4B). While 24 the primary trends were conserved between the reactors, cells died at an even faster rate in the 25 absence of PCN in the high throughput system, and all cultures, regardless of the PCN 26 concentration, lost viability after day 5. Accordingly, we terminated experiments in the high-27 throughput reactor platform at day 5. While the amount of survival measured in this system at 75 28 µM PCN was about half of that measured in the large reactor system by the end of the 29 experiments, it was still less than an order of magnitude total loss in survival (Figure 4B, table). 30 Encouraged by these findings, we leveraged the high throughput platform to gauge the 31 antibiotic tolerance of cells in the non-growth state supported by anaerobic PCN-cycling.  After a 32 day of incubation in the 96-well device, we added five different drugs representing distinct 33 antibiotics that target different cellular processes: ceftazidime, a b-lactam that targets cell wall 34   9 synthesis; ciprofloxacin, a fluoroquinolone that disrupts DNA gyrase; tobramycin, an 1 aminoglycoside that interferes with protein synthesis; N,N-dicyclohexylcarbodiimide (DCCD), an 2 FOF1 ATP(synth)ase inhibitor; and colistin, a polymyxin that disrupts the cell membrane. Cell 3 viability was mildly impaired by ceftazidime and ciprofloxacin, moderately by tobramycin, and 4 greatly impacted by DCCD and colistin (Figure 4C). These results resemble tolerance profiles that 5 have been observed for cells in the non-growing anoxic cores of P. aeruginosa biofilms that are 6 metabolically sustained by phenazine EET11,20,44, pointing to the importance of maintaining 7 membrane integrity45,46 and bioenergetic activity to preserve viability in the absence of growth.  8  9 Cells are energy-limited and surviving near their maintenance requirement 10  11 Having validated the utility of our high throughput PCN-cycling assay to study a clinically-12 important cellular state, we proceeded to measure the survival of hundreds of cultures across 13 various mutant strains over a range of PCN concentrations with multiple tiers of replication to 14 begin to mechanistically dissect it (Figure 5A). First, we sought to define the facilitated 15 fermentation metabolic pathway more precisely by comparing mutant strains lacking enzymes of 16 bioenergetic importance that we hypothesized might contribute to PCN-based survival24,47. 17 Second, we sought to determine whether the amount of PCN cycling was limiting survival. We 18 achieved these goals by leveraging knowledge that the amount of current generated via EET is 19 proportional to the concentration of electron shuttle provided48. Hypothesizing that the PCN 20 concentration would therefore limit the total culture’s metabolic rate, we made measurements 21 across PCN concentrations spanning three orders of magnitude (Figure 5A). 22 After five days of anaerobic incubation in our high-throughput platform, we found that 23 survival of the ∆phz1/2 strain was dose-dependent up to at least 75 µM, with increasing 24 concentrations having only marginal effects on survival (Figure 5B). Phenazine reduction can 25 occur promiscuously via multiple metabolic flavoproteins49,50, but we have shown recently that 26 PCN is predominantly reduced in stationary phase cells by two NADH dehydrogenases in the cell 27 membrane, Nuo and Nqr47 (Figure 1B). While the current produced by a ∆phz1/2 ∆nuoF ∆nqrF 28 mutant was initially lower, after a day it approximately equaled that of the ∆phz1/2 cells, 29 suggesting other PCN reductases compensate after this time (Figure 5C). However, full survival 30 across PCN concentrations nonetheless depended upon functional NADH dehydrogenases 31 (Figure 5B) with the loss of viability in their absence occurring steadily over the entire course of 32 the incubation (Figure S3A), consistent with our working model (Figure 1B). Complementation of 33 either NADH dehydrogenase restored survival at most concentrations of PCN, indicating they 34   10 serve largely redundant functions (Figure S3B). As a negative control, and in agreement with our 1 previous study24, knocking out the AckA-Pta pathway (the energy-conserving step of the 2 metabolism) had a larger consequence upon survival. Moreover, our platform allowed us to 3 measure survival of this mutant across many PCN concentrations efficiently, something that would 4 have been logistically impossible using large reactors. Interestingly, increasing PCN 5 concentrations resulted in increased survival of the ∆phz1/2 ∆ackA-pta mutant, albeit at 6 significantly lower levels (Figure 5B). We hypothesized that this alternative mode of phenazine-7 dependent survival might also rely on the NADH dehydrogenases, and therefore tested a ∆phz1/2 8 ∆nuoF ∆nqrF ∆ackA-pta mutant. Though this strain’s survival was lower than the ∆phz1/2 ∆ackA-9 pta mutant, it still improved up to 75 µM PCN, yet did not improve in survival at higher 10 concentrations of PCN, indicating that the NADH dehydrogenase complexes play a significant 11 role at these concentrations (Figure 5B). Overall, our data are consistent with substrate-level 12 phosphorylation being the primary energy conserving mechanism under these conditions (Figure 13 1B) but also indicate the existence of a previously undetected, alternative PCN-dependent energy 14 conservation pathway that supports much lower levels of survival. 15 We also measured the long-term current generated by ∆phz1/2 as a function of PCN 16 concentration and found the two were largely correlated, as expected48 (Figure 5D). As with the 17 large reactors, we observed an early spike in current generation dependent on the phenazine 18 concentration provided that then subsided to a lower, more stable current, though the decay 19 happened at a slower rate than in the large reactors (Figure 5C, Figure 2A). The one exception 20 was at 750 µM PCN, which produced highly variable initial currents despite stable currents later 21 (Figure 5C). By integrating the current generated between days one and five (i.e. after the initial 22 current spike), we calculated the average PCN oxidation rate across PCN concentrations and 23 found it unexpectedly saturated above 75 µM (Figure 5D). At 375 µM PCN, the current measured 24 on days 1-5 was below what the electrode was capable of oxidizing, exemplified by the initial high 25 current spike (Figure 5C). What might cause this later saturation?  We hypothesized that either 26 PCN oxidation at the electrode was limiting due to biofouling, or the rate of PCN reduction was 27 limiting due to a change in the cells that occurs after a day of anaerobic incubation. 28 To distinguish between these scenarios, we used scanning electron microscopy to 29 visualize cells on the electrodes, finding that the electrodes were mostly clear with a sparse 30 monolayer of cells present on the working electrode surface, with the densest cell monolayer 31 found in the 375 µM PCN condition (Figure S4A). While the absence of biofilms on the electrode 32 suggested that biofouling was unlikely to explain the observed reduction rate plateau, we sought 33 a more direct test of whether the cells or the electrode was limiting. Accordingly, on day 5 of the 34   11 survival assay, we pooled ∆phz1/2 cells from multiple technical replicate wells in the 375 µM PCN 1 condition, pelleted them, resuspended them in a small volume and added them to other wells 2 containing ∆phz1/2 cells and 375 µM PCN such that the total cell concentration was approximately 3 doubled. We observed a 1.5x increase in current generation following the increase in cells that 4 was sustained for at least 10 hours (Figure S4B), demonstrating that the electrode was not 5 limiting. Accordingly, the observed PCN oxidation rate at the electrode (Figure 5D) can be 6 interpreted as equivalent to the cellular PCN reduction rate. Our results suggest that after about 7 one day in anaerobic conditions, the cellular reduction rate above a certain PCN concentration 8 becomes attenuated. This finding is intriguing, as bacteria are believed to always maximize their 9 metabolic rate to maximize fitness during growth51,52. What explains this attenuation is a worthy 10 subject for future studies, especially given that these PCN concentrations are in the range found 11 in P. aeruginosa biofilms33. 12 Finally, we sought to determine whether the power output of the PCN facilitated 13 fermentation was at or above the minimum necessary for maintenance metabolism under PCN-14 cycling conditions. We reasoned that we would observe a relatively constant cell-specific 15 metabolic rate over the PCN concentration range if this were true, as that would indicate that 16 surviving cells require the same metabolic rate. Indeed, this is what we observed (Figure 5E). The 17 cell-specific metabolic rate measured at 75 µM PCN was 3.0 x 103 electrons sec-1 cell-1 at 25˚C 18 (lower 95% CI 1.6 x 103, upper 5.4 x 103), slightly higher than the rate measured in the large 19 electrochemical reactors. Given the PCN concentration limits survival between 0.75 and 75 µM 20 (Figure 5B), this consistent metabolic rate suggests that PCN limitation imposes an energy 21 limitation, reducing the populations’ viability in a PCN-dependent manner. While a fraction of cells 22 may be in the VBNC state that are not accounted for in this analysis, particularly at lower PCN 23 concentrations, such cells would be expected to make a minor contribution to the overall current 24 measured6. We therefore conclude that the culturable cells are operating near their minimum 25 maintenance power requirement for these conditions. 26  27 DISCUSSION 28  29 That bacteria can survive for remarkably long periods of time when limited for nutrients or 30 energy is well known, yet the question of how cells maintain viability in the absence of growth has 31 been challenging to answer due to technical limitations. Answering this question is important for 32 both fundamental and practical reasons: growth arrest not only typifies bacterial existence in most 33 natural habitats, but also in disease contexts. Indeed, our inability to successfully treat chronic 34   12 infections wherein cells are metabolically active yet doubling very slowly or not at all, stems in 1 large part from our ignorance of the molecular strategies that sustain life during growth arrest. 2 Our work here demonstrates that it is possible to gain mechanistic insight into the maintenance 3 physiology of P. aeruginosa by investigating how it cycles the phenazine PCN under anoxic 4 conditions—a non-growth state operating at an extremely low power output that lends itself to 5 quantitative studies in high throughput. 6  When surviving anaerobically via PCN cycling, P. aeruginosa cellular integrity and 7 metabolic activity is sustained at the cell-specific metabolic rate of 1.6 x 103 electrons sec-1 cell-1 8 at 25 ̊ C, equivalent to a mass-specific metabolic rate of 1.6 x 1016 electrons sec-1 (g C)-1 or roughly 9 103 ATP sec-1 cell-1. The term basal power requirement has been introduced to define the minimum 10 bioenergetic power per unit biomass required to sustain metabolic activity5. In our system, cell 11 survival is limited by a required minimum cell-specific metabolic rate (Figure 5E) suggesting this 12 rate may represent the basal power requirement for P. aeruginosa under these conditions. 13 Remarkably, this rate lies 3 to 4 orders of magnitude below the energy demand during fast aerobic 14 growth (Figure 2B), highlighting the extraordinary range of power output that a single bacterial 15 species can use to support its metabolism and helping to explain its success as an agent of both 16 acute and chronic infection. Recently, multiple environmental measurements of microbial cell-17 specific metabolic rates under energy-limitation have also been made or modeled53, adding to a 18 short list of basal power requirement estimates. Measurements of oxygen consumption below the 19 seafloor of the North Pacific Gyre were found to reach an asymptote at depth around 25 electrons 20 sec-1 cell-1, with the temperature unreported but presumably cold54. The cell-specific metabolic 21 rate of anaerobic sulfate-reducing bacteria found below the seafloor at multiple sites were 22 measured to reach an asymptote at depth to about 2.8 x 102 electrons sec-1 cell-1 at temperatures 23 around 5˚C55. Bacteria residing beneath permafrost in the cryopeg brines of Utqiaġvik, Alaska 24 have been modeled to sustain metabolic rates between 18 and 1.7 x 103 electrons sec-1 cell-1 at 25 subzero temperatures56. When these measurements are adjusted to 25˚C assuming a typical Q10 26 normalization57, they fall surprisingly close in order of magnitude to our measurements in the lab 27 (Figure 2B). Given the uncertainties underpinning assumptions needed to assign metabolic rates 28 to environmental datasets, it is conceivable that our lab measurement and true environmental 29 basal power requirements converge upon the fundamental limit of what is needed to power 30 bacterial survival, or otherwise, converge to the limit of what is measurable with current methods. 31 More quantitative studies of diverse maintenance states will be needed to determine whether 32 such a conserved minimum exists. 33   13 It is generally believed that organisms in the environment metabolizing at low rates are 1 growing, albeit slowly; indeed, environmental doubling times on the order of days to 1000s of 2 years have been estimated by multiple studies58. Accordingly, anticipating that PCN cycling would 3 power slow growth, we were somewhat surprised that cell growth was not detectable using our 4 HADA assay (Figure 3D). It may be that we simply did not wait enough time to observe slow 5 growth in the lab and/or that estimates of slow growth in the environment are inexact. To parse 6 this discrepancy, we note that it has been estimated that the energetic cost of aerobic division for 7 a bacterial cell is approximately 4 x 109 ATP per division59. Based on the dominant ATP-generating 8 pathway used under these conditions (i.e. oxidation of glucose to acetate coupled to substrate-9 level phosphorylation, Figure 1B), we estimate that P. aeruginosa’s ATP output is about 8.0 x 102 10 ATP sec-1 cell-1. We can then estimate that if division were possible under PCN facilitated 11 fermentation, cells would be doubling at a rate of once every two months, and perhaps even faster 12 because biosynthetic costs are thought to be less under anoxic conditions60 and the fraction of 13 energy lost to maintenance processes is presumed to be less during slow growth61. Such a growth 14 rate should have been detectable with our assay, corresponding to a steady increase to at least 15 ~0.125 divisions in a week, yet instead we observed division values consistently centered around 16 0 (Fig 3F). Thus, it seems more likely this ATP output is the maintenance energy requirement. 17 Together, this reasoning prompts us to question whether microbes are capable of doubling on 18 exceedingly-long time scales, or if they instead undergo long-term survival, opportunistically 19 waiting for punctuated bursts of local nutrients to fuel relatively fast growth as has been suggested 20 recently62. In this way, the classic “feast-famine cycle” may apply to more than just enteric 21 bacteria63, which would carry implications for biogeochemical modeling64. Future studies of the 22 growth associated with extremely slow metabolic rates are needed to determine which model is 23 more accurate. 24 Leaving such speculation aside, a consequential feature of PCN cycling fueling a non-25 growth state is that it opens the door to the identification of cellular strategies that sustain 26 maintenance more broadly. In our system, EET provides a convenient way to quantify metabolic 27 rate and stabilize a non-growth state, yet some of the physiological strategies underpinning 28 maintenance during PCN-cycling are likely to overlap with those required during non-EET 29 survival65. Given that non-growing cells tolerate conventional antibiotics but are sensitive to drugs 30 that target membrane and bioenergetic components (Figure 4C), we elucidated the key catabolic 31 enzymes supporting survival during PCN cycling as a proof-of-concept that our high throughput 32 platform can be used to gain mechanistic insight into what underpins a low-power lifestyle. By 33 focusing on PCN’s role as an electron acceptor, we expected to identify machinery that specifically 34   14 interacts with PCN, yet we anticipate that future studies of the maintenance state supported by 1 PCN will identify more general physiological strategies that sustain a low-power energy economy. 2 While we previously established that the AckA-Pta enzymes are necessary for phenazine 3 facilitated fermentation24, our work here revealed that either of the NADH dehydrogenases Nuo 4 and Nqr are additionally necessary to support full survival of cells via PCN facilitated fermentation 5 (Figure 5B). This may only apply to the phenazine PCN, as we have shown that it is the only 6 derivative made by P. aeruginosa that is predominantly reduced at the inner membrane47. Indeed, 7 phenazines are known to promiscuously react with many metabolic flavoproteins in the 8 cytosol49,50. This promiscuity likely explains our finding that increasing concentrations of PCN 9 correlated with increasing survival in the ∆phz1/2 ∆nuoF ∆nqrF mutant tested (Figure 5B). We 10 speculate that the benefit the NADH dehydrogenases provide is a direct coupling to the NADH 11 pool and, by extension, pyruvate oxidation. While other flavoproteins can dissipate reducing 12 equivalents onto the PCN pool in the absence of the NADH dehydrogenases, the reductants 13 involved may not be as closely coupled to the NADH pool and be wasted energetically. The PCN-14 dependent compensation of the ∆phz1/2 ∆ackA-pta mutant (Figure 5B) is more difficult to explain 15 and suggests the participation of unidentified alternative energy-conserving pathways. These 16 pathways may still be NADH-coupled, which could explain the loss of enhanced survival in the 17 ∆phz1/2 ∆ackA-pta mutant at high PCN concentrations in the ∆phz1/2 ∆nuoF ∆nqrF ∆ackA-pta 18 mutant (Figure 5B). Future metabolomic experiments will shed light on the nature of these PCN-19 dependent compensations.  20 In their normal function during respiratory growth, the Nuo and Nqr complexes 21 chemiosmotically pump protons across the cytosolic membrane with H+/2e- of 4 and 2, 22 respectively66,67. This is driven by a large ∆E˚’ of 420 mV between the NADH and ubiquinone 23 pools. For PCN reduction by NADH, ∆E˚’ is less than half that value at 180 mV, corresponding to 24 a ∆G˚’ of -34.7 kJ/mol. Previous work has estimated the minimum biological energy quantum 25 (enough to drive the translocation of one proton across the cytoplasmic membrane) to be -20 26 kJ/mol68. These values may change some amount depending on the reaction context, namely 27 concentrations of reactants and products and the PMF level, but provide a useful quantitative 28 baseline. From these values, we infer that PCN reduction by an NADH dehydrogenase 29 theoretically provides only enough free energy to translocate one proton, less than the normal 30 proton coupling of Nuo and Nqr. Additionally, phenazines have been shown to be reduced via 31 flavin active sites of other metabolic redox enzymes49, and if this also applies to the NADH 32 dehydrogenases (i.e. reduction at the NADH binding site), it seems unlikely that PCN reduction 33 would trigger the enzymatic conformational changes necessary to pump protons. However, if 34   15 phenazines were reduced at the ubiquinone binding site (permitting the proper conformational 1 changes for proton pumping) and the cell were under low PMF conditions, proton pumping might 2 be possible, as metabolic reactions operating near their thermodynamic equilibrium are known to 3 occur69. Further studies using purified Nuo and Nqr with phenazine substrate are needed to 4 determine if this mechanism of energy conservation is plausible. 5 Outside of its fundamental interest to cell biology, understanding the minimal energy 6 requirements and physiology of cells in a non- or slow-growth state has important practical 7 implications. Despite the fact that estimated microbial doubling times in nature typically are on the 8 order of days to weeks to years70–73, the vast majority of microbial physiology studies in pure 9 culture have been conducted using model species grown with doubling times under an hour. The 10 consequences of this nature-to-lab metabolic rate mismatch has resulted in our inability to control 11 non- and slow-growing microbes in nature and disease. In addition to the deleterious impact this 12 knowledge gap has for biofilm control in chronic infections as we have discussed, if synthetic 13 biologists are to succeed in catalyzing desirable changes in the field, such as applying pro-biotics 14 to the rhizosphere to stimulate the growth of crops, they will require a deeper understanding of 15 how cells survive periods of growth arrest due to desiccation or nutrient limitation74,75. Moreover, 16 identifying the slowest viable metabolic rates and what permits them would help geobiologists 17 and climate scientists better understand the fate of organic matter soils and sediments, where the 18 crucial roles of microbes are widely recognized but challenging to model64,76,77. In all these cases, 19 an understanding of life “in the slow lane”, is needed62. Our PCN-cycling system opens the door 20 to quantitative and mechanistic studies of a maintenance state that comes much closer to 21 approximating real-world non-growth metabolisms than the typical growth conditions used in the 22 lab. 23 Finally, while bioenergetic comparisons often rely on thermodynamic arguments to 24 compare the energy conservation efficiency of different metabolisms, our results highlight the 25 importance of kinetics in explaining the remarkable plasticity of microbial metabolism.  To wit: per 26 glucose molecule, fermentation generates roughly 10 times less ATP than aerobic respiration, yet 27 the cell-specific metabolic rate during facilitated fermentation we measure is 1,000 times slower 28 than P. aeruginosa’s fast aerobic growth. In terms of power, this disparity is dominated not by 29 differences in energy conservation efficiency but by differences in rates. Indeed, it seems likely 30 that slow kinetics may dominate the bioenergetic reality of low-power maintenance metabolism, 31 as previously suggested7. If, how, and when metabolic rate is controlled by the cell at the 32 molecular level or limited by the environmental conditions imposed are important questions to be 33   16 addressed in future research and key to an understanding of the dominant pace of microbial life 1 on the planet. 2  3 Limitations of the Study 4  5 In this study, we have leveraged an EET metabolism to develop a quantitative, high-6 throughput system for studying a low-power maintenance state. While the mechanistic studies we 7 described focused on identifying catabolic machinery necessary for survival via PCN EET (e.g. 8 NADH dehydrogenases), parts of these pathways are needed to sustain survival in other contexts 9 (e.g. the AckA pathway is also necessary for survival when oxygen is limiting in the absence of 10 EET65). Future studies will reveal which other pathways (e.g. anabolic, regulatory) sustain the 11 needs of cells in the PCN-maintenance state and whether they are more broadly conserved 12 across bacteria existing in a non-growth state, be they other opportunistic human pathogens or 13 bacteria in arid soils or deep-sea sediments. While it is likely that some cellular maintenance 14 strategies will be specific for particular organisms in specific contexts, we expect others may be 15 more generalizable. 16 The system used to assess this metabolism, both in large reactors and the 96-well 17 platform, is an artificial planktonic reconstitution of a metabolism that is more naturally relevant to 18 biofilms33,78. Thus, it is important to acknowledge there may be differences in the metabolic 19 dependencies of cells in this planktonic state versus the more natural biofilm state. However, 20 many key features that characterize biofilm cells are recapitulated in this system, including a lack 21 of growth in the anoxic core and lowered metabolic activity79, features that contribute to antibiotic 22 tolerance11.  23 Finally, we note that the anaerobic survival assessed in this study is not a state of 24 perpetual survival: cells begin to die after 7-9 days in the large reactors and after 5 days in the 25 96-well plate system. Therefore, the cells have been assessed in the middle of a quasi-steady-26 state that may or may not be representative of cells under a maintenance state in nature or 27 disease, though they are probably much more representative than cells cultivated in chemostats 28 that achieved stable but slow, constant growth. Cells were assessed in only one medium during 29 this study, so variations in metabolic rate and active metabolic pathways as a function of 30 physiological conditions were unexplored. Bacteria do not have a defined basal metabolic rate, 31 so our finding that phenazine EET is an order of magnitude slower than the basal metabolic rate 32 of the biosphere is an imperfect comparison—one that we view as informative for understanding 33 the relative positioning of phenazine facilitated fermentation among all metabolic rates. While the 34   17 rate is slow, as we note in the results, it falls in the ~15th percentile of endogenous metabolic rates 1 previously measured in bacteria. Therefore, it is not the slowest metabolic rate measured, but to 2 our knowledge it is currently the only one proven to be coupled to a non-growth maintenance 3 state at the single-cell level.  4  5 ACKNOWLEDGEMENTS. This work was supported by NIH Grant (2R01AI127850-06A1) to 6 DKN. This work was supported by JSPS KAKENHI (22H02265 and 22KK0242), and JST GteX 7 (JPMJGX23B4) to AO. We are grateful to Georgia Squyres and Sean Wilson for their help with 8 the HADA experiment, and Avi Flamholz and other members of the Newman lab for constructive 9 feedback throughout the study. 10  11 AUTHOR CONTRIBUTIONS 12  13 Conceptualization, J.A.C., A.O., and D.K.N.; Methodology, J.A.C., C-L.H., R.D.H., A.O., and 14 D.K.N.; Investigation and Validation, J.A.C., C-L.H., and R.D.H.; Formal Analysis and 15 Visualization, J.A.C.; Resources, A.O. and D.K.N. Writing – Original Draft, J.A.C. and D.K.N.; 16 Writing – Review & Editing, J.A.C., C-L.H., R.D.H., A.O., and D.K.N.; Supervision and Funding 17 Acquisition, A.O. and D.K.N. 18  19 DECLARATION OF INTERESTS 20  21 The authors declare no competing interests. 22   23   18  1 Figure 1. PCN redox-cycling promotes anaerobic energy conservation and survival via a facilitated fermentation. (A) Simplified schema of heterotrophic metabolisms and their predominant mode of energy conservation, exemplifying how facilitated fermentation is a noncanonical form of energy conservation. SLP – substrate level phosphorylation, OxPhos – oxidative phosphorylation. (B) Working model of the active metabolic pathways catabolizing glucose during PCN facilitated fermentation. The NADH dehydrogenases Nuo and Nqr are putatively involved in the metabolism because of their previously established function as key PCN reductases47. (C) Schematic of large glass electrochemical reactors used in this study to continuously re-oxidize PCN anaerobically reduced by the cell culture. A stir bar in the main chamber kept the culture well-mixed. CE – counter electrode, WE – working electrode, RE – reference electrode. (D) Anaerobic survival of ∆phz1/2 strain in glucose minimal medium with or without 75 µM PCN and an oxidizing potential. 100% survival represents approximately 8 x 108 CFU mL-1. Data are representative of 4 biological replicates. (E) Quantification of acetate production after 7 days of culture incubation. Data are averaged across 2 biological replicates, error bars represent the standard deviation. Welch’s t-test was used to compare samples, *p<0.05. (F) Propidium iodide staining of anaerobic cultures on day 5 of survival imaged on a fluorescence microscope. Images are representative of 2 biological replicates. Scale bar is 10 µm.    19  1 100 101 102 103 104 105 106 107 1080.00.20.40.60.81.0Cell-Specific Metabolic Rate(e- sec-1 cell-1)ECDFPCN FacilitatedFermentationEndogenousMaintenance(continuous culture)Fast Growth-2000 1-1 2 3 4 5 6 720406080DayI (µA)Mean e- sec-1days 1-5GeomeanCFU mL-1days 1-5x  100 mL=Cell-Specific Metabolic Rate(e- sec-1 cell-1)A (           )(                 )0 1 2 3 4 5 6 7105106107108109DayCFU mL-1∆phz1/2, + PCN/pot∆phz1/2, - PCN/potBFigure 2. The metabolic rate of PCN facilitated fermentation is extremely low. (A) Cell-specific metabolic rate during PCN facilitated fermentation was quantified from large electrochemical reactor experiments. Cells are added on day 0. The mean current generated by the culture across days 1 to 5 was averaged and divided by the geometric mean of total CFUs across the same period. Data are representative of 4 biological replicates. (B) Empirical cumulative distribution functions of lab-measured bacterial metabolic rates collated previously by Hoehler et al.27, with units transformed to electrons sec-1 cell-1 at 25˚C. Fast growth includes bacteria growing at or near their maximum growth rate. Maintenance includes inferred maintenance energy metabolic rates (i.e. during zero growth) extrapolated from slow-growing continuous culture studies. Endogenous includes metabolic rates measured in the absence of any exogenous electron donor and using endogenous stores of donors instead. Each data point is an independent experimental measurement. The dataset encompasses 90 species and 198 measurements.   20   1 Grow aerobicallyin LB + HADAWash awayunincorporated HADAResuspend in GMM inbalch tubes or anoxicelectrochemical reactor,incubateRemove samples periodicallyto imageO2O2EchemGMM- PCN/potEchemGMM+ PCN/potDay 0Day 7BalchGMM- KNO3BalchGMM+ KNO3Day 0Day 4...HADA Fluorescence0.010.1110100% Survival0 1 2 3 4 5 6 7Day+ HADA, + PCN/pot+ HADA, - PCN/pot- HADA, - PCN/pot- HADA, + PCN/pot0 1 2 3 40.11DayOD 500nm - HADA, - KNO3+ HADA, - KNO3- HADA, + KNO3+ HADA, + KNO3AB CDFEtotal polesfluorescent poles# Divisions = log2(                    )0 2 4 6-0.50.00.5Day# Divisions- PCN/pot+ PCN/potFigure 3. Single cells do not grow during anaerobic survival via PCN facilitated fermentation. (A) Experimental design and expected topological dilution of HADA cell wall stain during cell division. (B) HADA fluorescence imaging of cells during anaerobic growth on glucose and nitrate in sparged balch tubes. GMM – glucose minimal media. Scale bar is 5 µm. (C) Anaerobic growth of cultures imaged in (B) compared to cultures that were not stained with HADA. (D) HADA fluorescence imaging of cells during anaerobic survival on glucose, PCN, and an oxidizing potential in large electrochemical reactors. Scale bar is 5 µm. Red arrows indicate phase-dark granules that accumulate in cells that are not provided with PCN and an oxidizing potential. Green arrows indicate HADA-labeled septa that remained intact over the week-long incubation - inset is a zoomed example, each edge is 3.75 µm. (E) Anaerobic survival of cultures imaged in (D) compared to cultures that were not stained with HADA. (F) Estimated number of divisions of the +PCN/pot culture imaged in (D) via manual counting of fluorescent vs total cell poles and use of the equation in (A). Each datapoint is n≥400 cells. All data in figure are representative of two biological replicates.   21  1 2 Figure 4. High-throughput electrochemical reactor system allows for multiplexed phenazine-dependent anaerobic survival assays. (A) 96-potentiostat plate reader used for experiments. Each well of the plate is outfitted with an independent working electrode (WE), counter electrode (CE), and reference electrode (RE) that are connected to the pins of the potentiostats via contacts on the bottom of the plate. (B) Anaerobic survival of ∆phz1/2 strain in glucose minimal medium at various concentrations of PCN, with an oxidizing potential. Table – quantitative comparison of survival at experimental end (5 days in high-throughput reactor, 7 days in the large reactor), where n represents the number of biological replicates. (C) Anaerobic survival of ∆phz1/2 strain in glucose minimal medium with 75 µM PCN and an oxidizing potential after various antibiotics were added on Day 1. Ceftazidime – 100 µg/mL, Ciprofloxacin – 10 µg/mL, Tobramycin – 100 µg/mL, DCCD – 2 mM, Colistin – 10 µg/mL. In both (B) and (C), 100% survival represents approximately 8 x 108 CFU mL-1. Each datapoint represents the geometric mean of 8 biological replicates and error bars represent the 95% confidence interval.    22  1  2   3 Figure 5. High-throughput PCN redox-cycling experiments reveal a dependence on NADH dehydrogenases and survival near the minimum maintenance requirement. (A) Example experimental design of high-throughput PCN redox-cycling survival experiments. (B) Survival of ∆phz1/2 and metabolic mutant strains after 5 days of anaerobic incubation in a glucose minimal medium across various concentrations of PCN. Each datapoint is the geometric mean of at least 5 biological replicates. Shaded regions represent 95% confidence intervals. 100% survival represents approximately 8 x 108 CFU mL-1. (C) Average current produced by ∆phz1/2 and ∆phz1/2 ∆nuoF ∆nqrF strains over the course of 5 days, with zoom-ins on the initial 5 hours of current production. Curves are the average of 5 biological replicates. (D) Average PCN oxidation rate at the working electrode surface of ∆phz1/2 cultures over days 1 to 5 of the anaerobic incubation. Each data point is the average of 5 biological replicates and error bars represent the 95% confidence interval. (E) Cell-specific metabolic rate of ∆phz1/2 cells during days 1 to 5 of the anaerobic incubation across various concentrations of PCN. Each data point is the geometric mean of 5 biological replicates and error bars represent the 95% confidence interval.   23  1 MATERIALS AND METHODS 2  3 KEY RESOURCES TABLE 4  5 REAGENT or RESOURCE SOURCE IDENTIFIER Bacterial and Virus Strains   Pseudomonas aeruginosa UCBPP-PA14 ∆phz1/2 80Dietrich et al., 2006 N/A P. aeruginosa UCBPP-PA14 ∆phz1/2 ∆nuoF ∆nqrF 47Ciemniecki and Newman, 2023 N/A P. aeruginosa UCBPP-PA14 ∆phz1/2 ∆ackA-pta 24Glasser et al., 2014 N/A P. aeruginosa UCBPP-PA14 ∆phz1/2 ∆nuoF ∆nqrF ∆ackA-pta This study N/A P. aeruginosa UCBPP-PA14 ∆phz1/2 ∆nuoF ∆nqrF, attTn7::rhaPBAD-nuoF This study N/A P. aeruginosa UCBPP-PA14 ∆phz1/2 ∆nuoF ∆nqrF, attTn7::rhaPBAD-nqrF This study N/A E. coli S17 (chemically competent for transformation) ATCC ATCC-BAA-2428 E. coli SM10/pTNS1 (chromosomal insertion helper strain) 81Choi and Schweizer, 2006 N/A Chemicals, peptides, and recombinant proteins   Difco LB Broth BD Cat#244620 Difco LB Agar BD Cat#244520 MOPS Sigma-Aldrich Cat#RDD003 Sodium Chloride Fisher Cat#BP358-212 Ammonium Chloride Fisher Cat#A649-500   24 Potassium Phosphate Monobasic Fisher Cat#P382-500 Magnesium Sulphate Heptahydrate Fisher Cat#M63-500 D-Glucose Fisher Cat#D16-500 Iron(II) Sulphate Heptahydrate Fluka Cat#44970 Phenazine-1-carboxamide ChemScene Cat#CS-W022931 Sodium Acetate Trihydrate Fisher Cat#S209-500 Potassium Nitrate Fisher Cat#P263-500 HADA Tocris Cat#6647 Propidium Iodide Invitrogen Cat#P1304MP DCCD Sigma-Aldrich Cat#D80002 Colistin Sulfate Salt Sigma-Aldrich Cat#C4461 Ciprofloxacin Abcam Cat#ab141917 Ceftazidime Hydrate Sigma-Aldrich Cat#A6987 Tobramycin Sigma-Aldrich Cat#T4014 Oligonucleotides (5’-3’)   ackA-pta upstream forward: ACGACGGCCAGTGCCAAGCTTTCAGGCTGCAGAAGGACTG This study N/A ackA-pta upstream reverse: GCCCACTGGGCGGCGTTCCTTCACTGCTCCTTGGTCTGCT This study N/A ackA-pta downstream forward: AGCAGACCAAGGAGCAGTGAAGGAACCCCGCCCAGTGGGC This study N/A ackA-pta downstream reverse:  CATGATTACGAATTCGAGCTTACTGATCGCGGCCTGGAAGAAAAAGC This study N/A Recombinant DNA     25 pMQ30 82Shanks et al., 2006 N/A pJM220 83Jeske and Altenbuchner, 2010 N/A pMQ30 ∆ackA-pta This study N/A pJM220 rhaPBAD-nuoF 47Ciemniecki and Newman, 2023 N/A pJM220 rhaPBAD-nqrF 47Ciemniecki and Newman, 2023 N/A Software and Algorithms   GraphPad Prism 10 v10.1.1  https://www.graphpad.com/ ImageJ v2.0.0-rc-69/1.52p  https://imagej.net OriginPro 2021  https://www.originlab.com  1 RESOURCE AVAILABILITY 2  3 Lead Contact 4 Further information and requests for resources and reagents should be directed to and will be 5 fulfilled by the Lead Contact, Dianne Newman (dkn@caltech.edu). 6  7 Materials Availability 8 All bacterial strains are available upon request. 9  10 Data and Code Availability 11 Raw data are available upon request. 12   26  1 EXPERIMENTAL MODEL AND SUBJECT DETAILS 2  3 Bacterial growth conditions 4 All Pseudomonas aeruginosa UCBPP-PA14 strains were plated on LB agar from -80˚C glycerol 5 stocks and grown overnight at 37˚C. Plates were stored at 4˚C for up to a week and were used to 6 inoculate liquid cultures. Initial liquid cultures were grown in 7 mL of LB medium in glass culture 7 tubes (VWR #47729-583) in an orbital shaker (New Brunswick, Innova 44) at 37˚C shaking at 250 8 rpm on a slant for 20 hours (final OD500nm ~5).  9  10 Bacterial survival conditions in large electrochemical reactors 11 Initial cultures were used to inoculate a large 250 mL LB culture to an initial OD500nm of 0.06. This 12 large culture was grown under the same conditions as the initial culture for 5.5 hours to an OD500nm 13 of 3.0±0.2. These cells were pelleted (10 min, 6800 xg) and washed twice in a minimal medium 14 (100 mM MOPS pH 7.2 with NaOH, 43 mM NaCl, 93 mM NH4Cl, 3.7 mM KH2PO4, 1 mM MgSO4), 15 then resuspended to an OD of 75 and transferred into an MBraun nitrogen-only atmosphere glove 16 box (Unilab model) containing the prepared glass reaction vessels (custom made) and held at 17 33˚C. 1 mL of the concentrated culture was then diluted into the main chamber of the large 18 reaction vessels containing 99 mL of N2-sparged minimal medium with 20 mM D-glucose and 3.6 19 µM FeSO4 added (hereafter referred to as glucose minimal medium), along with 75 µM PCN in 20 the relevant samples. This culture was connected to a potentiostat (Gamry Reference 600 model) 21 via a graphite rod working electrode (Alfa Aesar #14738), platinum mesh counter electrode 22 (custom-made using Alfa Aesar platinum gauze #10283), and Ag/AgCl3M NaCl reference electrode 23 (Basi #MW-2030). The counter electrode was held in a small side-chamber separated from the 24 main by a glass frit containing 9 mL minimal medium and 75 µM PCN. The working electrode was 25 poised at a constant 0 mV vs. Ag/AgCl3M NaCl and the cultures were incubated with stirring using a 26 stir rod in the vessel. 100 µL samples were removed periodically via a port over the main chamber 27 of the vessel for CFU counting. All plastics used were left in the chamber to degas for at least 28 three days before use. 29  30 Bacterial survival conditions in 96-potentiostat electrochemical plate 31 Initial cultures were pelleted and washed twice in minimal medium, resuspended to an OD of 75, 32 and transferred into an anerobic Coy Chamber held at 33˚C and 2-3% H2 containing the prepared 33 96-well electrochemical plate (custom-made as previously described30) with a screen-printed 34   27 carbon working electrode and counter electrode, and Ag/AgCl reference electrode. The culture 1 was then diluted in N2-sparged minimal medium to an OD of 15, and 10 µL of the concentrated 2 culture was then diluted into each well of the plate containing 190 µL of N2-sparged glucose 3 minimal medium. The plate was sealed using both a slit-seal cover (BioChromato #R80.120.00) 4 and an aluminum seal cover (DiversifiedBiotech #ALUM-1000) to prevent evaporation of the 5 cultures between sampling. The plate was incubated at 33˚C without shaking in a custom 96-6 potentiostat system originally developed at the National Institute for Material Science, Japan. 7 Every well’s working electrode was held at a constant 0 mV vs. Ag/AgCl. Cyclic voltammograms 8 measured with PCN as a standard showed that a potential +50 mV more positive was applied in 9 the high-throughput reactor system than in the large reactor. On days 1, 3, and 5 of the incubation, 10 the aluminum seal was removed and the well contents were mixed by pipetting a 150 µL volume. 11 10 µL was then removed for CFU dilution counting (20 µL on Day 5 to capture the lowest CFU 12 concentrations). The plate was resealed with a fresh aluminum seal and then placed back in the 13 96-potentiostat system until the next CFU sampling. All plastics used were left in the chamber to 14 degas for at least three days before use. 15  16 METHOD DETAILS 17 Mutant strain construction and complementation  18 All plasmids used in this work are listed in the Key Resources Table. Primers were synthesized 19 by Integrated DNA Technologies and are also listed in the Key Resources Table. For all molecular 20 cloning, plasmids were constructed using Gibson assembly reactions (NEB). Plasmids were 21 chemically transformed into E. coli strain S17 and then conjugated into P. aeruginosa PA14. 22 Mutants were constructed using standard homologous recombination using 1 kb regions flanking 23 the gene(s) of interest in the pMQ30 plasmid. Genetic complement strains were constructed by 24 reintroducing the deleted gene in trans at the attTn7 site downstream of glmS in the P. aeruginosa 25 genome using the pJM220 plasmid. For each complementation, the gene was designed to have 26 its expression driven by a rhamnose-inducible promoter, and accordingly, complementation 27 survival experiments were conducted using glucose minimal medium supplemented with L-28 rhamnose (0.005% w/v, 305 μM).  29  30 CFU counting 31 Phosphate buffered saline (PBS) was used for diluting cell cultures prior to plating for CFUs. 32 Dilutions spanning 6-7 orders of magnitude were plated on LB agar using 5 µL spots and 33 incubated at 37˚C overnight. Colonies were counted with the aid of a dissection microscope 34   28 (Nikon SMZ800). The furthest dilution spot containing at least 10 colonies was counted. Plates 1 were kept for at least another 48 hours at room temp to check for delayed colony appearance; 2 counts were updated when this rarely occurred. To accurately assess the degree of anaerobic 3 killing caused by antibiotics, in that experiment (Figure 4C) samples for CFU counting were diluted 4 anoxically in N2-sparged PBS and plated on anoxic LB agar plates supplemented with 20 mM 5 KNO3. The plates were incubated at 37˚C anoxically for two days, then counted. 6  7 High-performance liquid chromatography 8 Samples were collected by centrifuging 1 mL of culture and passing the supernatant through a 9 0.22 µm filter. Samples were stored at -80˚C until analysis. Acetate was quantified using a Waters 10 e2695 Separations Module equipped with a 2998 PDA Detector and run through an Aminex HPX-11 87H column (Bio-Rad) held at room temperature at a flow rate of 0.5 mL min-1 for 25 mins. The 12 mobile phase used was 5 mM H2SO4. Retention time for acetate was validated with single species 13 standards ranging from 0 to 20 mM, to calibrate peak area against a known concentration. 14  15 Cell dry weight measurements 16 On day 5 of the anaerobic survival in the large electrochemical reaction vessels, 2 mL samples of 17 +PCN/pot culture were collected into pre-weighed tubes. The cultures were pelleted and 18 resuspended twice in deionized water, then pelleted again. The supernatant was removed, and 19 the pellet was dried overnight in a 55˚C drying oven. The dried pellets were then weighed on a 20 microbalance, and the weight was divided by 2 * CFU / mL counts for that day to estimate the 21 grams cell dry weight per cell. 22  23 Propidium iodide staining and imaging 24 Samples of culture from the large electrochemical reaction vessels were collected on day 5 of 25 anaerobic incubation. The cells were pelleted and washed twice in anoxic PBS, then stained with 26 70 µM propidium iodide (Invitrogen) for 20 minutes in anoxic conditions. The cell suspension was 27 then brought out of the MBraun chamber and spotted onto 1% (w/v) agar pads and imaged on a 28 Nikon Eclipse Ti-2 inverted microscope using an ORCA-Flash4.0 V3 camera. Excitation light (555 29 nm) was supplied through a SpectraX LED light engine for a 200-ms exposure, and emission light 30 was passed through a standard TRITC filter cube (Semrock) with the excitation filter removed. 31 Identical LUTs were applied across all fluorescence images for comparison. The number of cells 32 stained was counted manually with these LUTs applied using ImageJ’s ROI manager; cell husks 33 (those with no discernable phase-dark aspect) were excluded. 34   29  1 HADA single-cell growth measurements 2 250 mL LB cultures of cells were grown with 60 µM HADA label and washed as described in the 3 Experimental Model. After the excess HADA was washed away the cells were inoculated into the 4 anaerobic electrochemical reaction vessels as described above. For the nitrate control, smaller 5 LB cell cultures containing HADA were used (7 mL) and washed identically before being 6 inoculated into N2-sparged balch tubes containing 10 mL of glucose minimal medium with 0.05% 7 casamino acids and with or without 40 mM KNO3 added. The balch tubes were incubated in the 8 dark and at room temperature to ensure a slow growth over approximately the same time scale 9 as the electrochemical survival experiment. On each day of either incubation (balch or echem), 10 0.5 mL of cells were sampled, pelleted, and resuspended in oxic PBS. On later days, the -PCN/pot 11 samples were resuspended in smaller volumes of PBS to normalize cell density. 1 µL of the 12 suspension was spotted onto 1% (w/v) agar pads and imaged on a Nikon microscope as 13 described in Propidium iodide staining and imaging, using a 395 nm excitation, 200-ms exposure 14 time, and a standard DAPI filter cube (Semrock) with the excitation filter removed. Due to global 15 loss of fluorescence brightness at later timepoints, fluorescence LUTs were adjusted 16 independently for each day to maximize visualization of cell body vs. pole staining. LUTs were set 17 the same across images from the same day. The number of fluorescent poles vs. total poles was 18 counted manually with these LUTs applied using ImageJ’s ROI manager. 19  20 VBNC resuscitation 21 On days 3, 5 and 7 of the large reactor anaerobic incubation, samples of -PCN/pot culture from 22 the vessels were pelleted and resuspended twice in oxic PBS of equal volume. A sample of this 23 cell suspension was immediately plated for CFUs while another sample was diluted 1:10 in PBS 24 in culture tubes and incubated at 33˚C aerobically without shaking. A sample of this aerobic 25 resuscitation suspension was plated for CFUs after 1, 2, and 3 days of incubation. 26  27 ATP quantification assay 28 ATP was quantified using a Promega BacTiter-Glo kit according to the manufacturer’s protocol. 29 On day 7 of the large reactor anaerobic incubation, samples of culture from the vessels were 30 pelleted and resuspended twice in anoxic PBS. Cells from the +PCN/pot condition were 31 resuspended at a final concentration of 1x, while the -PCN/pot condition were resuspended at a 32 10x concentration to acquire sufficient luminescence from the assay. ATP quantification was 33 divided by the CFUs for that day; resuscitated CFUs were used for the -PCN/pot condition. 34   30  1 Scanning electron microscopy 2 At the end of day 2 of the 96-potentiostat electrochemical plate survival experiment, after current 3 had already subsided following the initial spike in the 375 µM PCN condition, samples were gently 4 mixed by pipetting a 150 µL volume, then removed from the wells. 2.5% glutaraldehyde in PBS 5 was added to each well and incubated for 10 min at 25˚C. The samples were carefully rinsed 6 three times with PBS, then dehydrated by an ethanol wash gradient with 5 min incubations. The 7 dehydrated samples were then washed with t-butanol twice for 5 minutes. The electrochemical 8 board on the bottom of the plate was removed, and freeze-dried in t-butanol under vacuum for 2 9 days. The board was then platinum coated and observed by a Keyence VE-9800 scanning 10 electron microscope at 10 or 20 kV. 11  12 QUANTIFICATION AND STATISTICAL ANALYSIS 13 To calculate an average PCN reduction rate in the 96-potentiostat system, current vs time traces 14 were trimmed by 3 hrs immediately following the times when the plate was removed and returned 15 to the potentiostat: these times following reconnection resulted in spurious, large spikes in current 16 that settled back down to levels prior to removal of the plate. The current between days one and 17 five was integrated, giving values with units nA*hr. This was then converted to units of electrons 18 using standard conversion factors, divided by 90 hrs (4 d minus 6 hrs surrounding times of plate 19 removal), divided by the geometric mean of the day 1-5 CFU (0.2 mL)-1 counts, and reported as 20 units of electrons sec-1 cell-1. 21 For all experiments, Prism 10 was used for analysis except for integration of current, which was 22 done in OriginPro 2021. Technical replicates were averaged to generate single biological 23 replicates before statistical analysis; therefore, all error ranges reported represent variation from 24 biological replicates. Since CFUs were counted in log space, geometric means were used as a 25 summary statistic. Any reported values that incorporated CFU counts (such as cell-specific 26 metabolic rates) are also reported as a geometric mean. Otherwise, replicates were averaged 27 using a linear mean. 28  29 SUPPLEMENTAL INFORMATION 30 Document S1. Figures S1-S4; Analysis of bacterial metabolic rate data from Hoehler et al. 2023, 31 PNAS; ∆G˚’ of PCN Facilitated Fermentation 32 Table S1. Excel file containing Hoehler et al. PNAS 2023 bacterial metabolic rates dataset with 33 units transformed to electrons sec-1 cell-1. 34   31   1   32 REFERENCES 1   2   3 1.    Bergkessel, M., Basta, D.W., and Newman, D.K. (2016). The physiology of growth 4 arrest: uniting molecular and environmental microbiology. Nat. Rev. Microbiol. 14, 549–5 562. 10.1038/nrmicro.2016.107. 6 2.    Morita, R.Y. (1988). Bioavailability of energy and its relationship to growth and starvation 7 in nature. Can J Microbiol 34, 436–441. 8 3.    LaRowe, D.E., and Amend, J.P. (2015). Power limits for microbial life. Front. Microbiol. 6, 9 718. 10.3389/fmicb.2015.00718. 10 4.    Tijhuis, L., Van Loosdrecht, M.C., and Heijnen, J.J. (1993). A thermodynamically based 11 correlation for maintenance gibbs energy requirements in aerobic and anaerobic 12 chemotrophic growth. Biotechnol. Bioeng. 42, 509–519. 10.1002/bit.260420415. 13 5.    Hoehler, T.M., and Jørgensen, B.B. (2013). Microbial life under extreme energy 14 limitation. Nat. Rev. Microbiol. 11, 83–94. 10.1038/nrmicro2939. 15 6.    Jin, X., Zhang, X., Ding, X., Tian, T., Tseng, C.-K., Luo, X., Chen, X., Lo, C.-J., Leake, 16 M.C., and Bai, F. (2023). Sensitive bacterial Vm sensors revealed the excitability of 17 bacterial Vm and its role in antibiotic tolerance. Proc Natl Acad Sci USA 120, 18 e2208348120. 10.1073/pnas.2208348120. 19 7.    Lever, M.A., Rogers, K.L., Lloyd, K.G., Overmann, J., Schink, B., Thauer, R.K., Hoehler, 20 T.M., and Jørgensen, B.B. (2015). Life under extreme energy limitation: a synthesis of 21 laboratory- and field-based investigations. FEMS Microbiol. Rev. 39, 688–728. 22 10.1093/femsre/fuv020. 23 8.    Yin, L., Ma, H., Fones, E.M., Morris, D.R., and Harwood, C.S. (2023). ATP is a major 24 determinant of phototrophic bacterial longevity in growth arrest. MBio 14, e0360922. 25 10.1128/mbio.03609-22. 26 9.    Robador, A., Amend, J.P., and Finkel, S.E. (2019). Nanocalorimetry Reveals the Growth 27 Dynamics of Escherichia coli Cells Undergoing Adaptive Evolution during Long-Term 28 Stationary Phase. Appl. Environ. Microbiol. 85. 10.1128/AEM.00968-19. 29 10.   Riedel, T.E., Berelson, W.M., Nealson, K.H., and Finkel, S.E. (2013). Oxygen 30 consumption rates of bacteria under nutrient-limited conditions. Appl. Environ. Microbiol. 31 79, 4921–4931. 10.1128/AEM.00756-13. 32 11.   Stewart, P.S. (2015). Antimicrobial tolerance in biofilms. In Microbial Biofilms, pp. 269–33 285. 34 12.   Thi, M.T.T., Wibowo, D., and Rehm, B.H.A. (2020). Pseudomonas aeruginosa Biofilms. 35 Int. J. Mol. Sci. 21. 10.3390/ijms21228671. 36 13.   Levin-Reisman, I., Ronin, I., Gefen, O., Braniss, I., Shoresh, N., and Balaban, N.Q. 37 (2017). Antibiotic tolerance facilitates the evolution of resistance. Science 355, 826–830. 38 10.1126/science.aaj2191. 39 14.   Stewart, P.S., and Costerton, J.W. (2001). Antibiotic resistance of bacteria in biofilms. 40 Lancet 358, 135–138. 10.1016/s0140-6736(01)05321-1. 41   33 15.   Hall, C.W., and Mah, T.-F. (2017). Molecular mechanisms of biofilm-based antibiotic 1 resistance and tolerance in pathogenic bacteria. FEMS Microbiol. Rev. 41, 276–301. 2 10.1093/femsre/fux010. 3 16.   Chiang, W.-C., Nilsson, M., Jensen, P.Ø., Høiby, N., Nielsen, T.E., Givskov, M., and 4 Tolker-Nielsen, T. (2013). Extracellular DNA shields against aminoglycosides in 5 Pseudomonas aeruginosa biofilms. Antimicrob. Agents Chemother. 57, 2352–2361. 6 10.1128/AAC.00001-13. 7 17.   Borriello, G., Werner, E., Roe, F., Kim, A.M., Ehrlich, G.D., and Stewart, P.S. (2004). 8 Oxygen limitation contributes to antibiotic tolerance of Pseudomonas aeruginosa in 9 biofilms. Antimicrob. Agents Chemother. 48, 2659–2664. 10.1128/AAC.48.7.2659-10 2664.2004. 11 18.   Mohr, K.I. (2016). History of antibiotics research. In Current Topics in Microbiology and 12 Immunology, pp. 237–272. 13 19.   Glasser, N.R., Saunders, S.H., and Newman, D.K. (2017). The colorful world of 14 extracellular electron shuttles. Annu. Rev. Microbiol. 71, 731–751. 10.1146/annurev-15 micro-090816-093913. 16 20.   Schiessl, K.T., Hu, F., Jo, J., Nazia, S.Z., Wang, B., Price-Whelan, A., Min, W., and 17 Dietrich, L.E.P. (2019). Phenazine production promotes antibiotic tolerance and metabolic 18 heterogeneity in Pseudomonas aeruginosa biofilms. Nat. Commun. 10, 762. 19 10.1038/s41467-019-08733-w. 20 21.   VanDrisse, C.M., Lipsh-Sokolik, R., Khersonsky, O., Fleishman, S.J., and Newman, D.K. 21 (2021). Computationally designed pyocyanin demethylase acts synergistically with 22 tobramycin to kill recalcitrant Pseudomonas aeruginosa biofilms. Proc Natl Acad Sci USA 23 118. 10.1073/pnas.2022012118. 24 22.   Wang, Y., Kern, S.E., and Newman, D.K. (2010). Endogenous phenazine antibiotics 25 promote anaerobic survival of Pseudomonas aeruginosa via extracellular electron 26 transfer. J. Bacteriol. 192, 365–369. 10.1128/JB.01188-09. 27 23.   Eschbach, M., Schreiber, K., Trunk, K., Buer, J., Jahn, D., and Schobert, M. (2004). 28 Long-term anaerobic survival of the opportunistic pathogen Pseudomonas aeruginosa via 29 pyruvate fermentation. J. Bacteriol. 186, 4596–4604. 10.1128/JB.186.14.4596-30 4604.2004. 31 24.   Glasser, N.R., Kern, S.E., and Newman, D.K. (2014). Phenazine redox cycling enhances 32 anaerobic survival in Pseudomonas aeruginosa by facilitating generation of ATP and a 33 proton-motive force. Mol. Microbiol. 92, 399–412. 10.1111/mmi.12566. 34 25.   Hunt, K.A., Flynn, J.M., Naranjo, B., Shikhare, I.D., and Gralnick, J.A. (2010). Substrate-35 level phosphorylation is the primary source of energy conservation during anaerobic 36 respiration of Shewanella oneidensis strain MR-1. J. Bacteriol. 192, 3345–3351. 37 10.1128/JB.00090-10. 38 26.   Benz, M., Schink, B., and Brune, A. (1998). Humic acid reduction by Propionibacterium 39 freudenreichii and other fermenting bacteria. Appl. Environ. Microbiol. 64, 4507–4512. 40 10.1128/AEM.64.11.4507-4512.1998. 41   34 27.   Flynn, J.M., Ross, D.E., Hunt, K.A., Bond, D.R., and Gralnick, J.A. (2010). Enabling 1 unbalanced fermentations by using engineered electrode-interfaced bacteria. MBio 1. 2 10.1128/mBio.00190-10. 3 28.   Tejedor-Sanz, S., Stevens, E.T., Li, S., Finnegan, P., Nelson, J., Knoesen, A., Light, S.H., 4 Ajo-Franklin, C.M., and Marco, M.L. (2022). Extracellular electron transfer increases 5 fermentation in lactic acid bacteria via a hybrid metabolism. eLife 11. 6 10.7554/eLife.70684. 7 29.   Okamoto, A., Tokunou, Y., Kalathil, S., and Hashimoto, K. (2017). Proton Transport in the 8 Outer-Membrane Flavocytochrome Complex Limits the Rate of Extracellular Electron 9 Transport. Angew. Chem. Int. Ed 56, 9082–9086. 10.1002/anie.201704241. 10 30.   Miran, W., Huang, W., Long, X., Imamura, G., and Okamoto, A. (2022). Multivariate 11 landscapes constructed by Bayesian estimation over five hundred microbial 12 electrochemical time profiles. Patterns (N Y) 3, 100610. 10.1016/j.patter.2022.100610. 13 31.   Wang, Y., and Newman, D.K. (2008). Redox reactions of phenazine antibiotics with ferric 14 (hydr)oxides and molecular oxygen. Environ. Sci. Technol. 42, 2380–2386. 15 10.1021/es702290a. 16 32.   Jo, J., Price-Whelan, A., Cornell, W.C., and Dietrich, L.E.P. (2020). Interdependency of 17 Respiratory Metabolism and Phenazine-Associated Physiology in Pseudomonas 18 aeruginosa PA14. J. Bacteriol. 202. 10.1128/JB.00700-19. 19 33.   Saunders, S.H., Tse, E.C.M., Yates, M.D., Otero, F.J., Trammell, S.A., Stemp, E.D.A., 20 Barton, J.K., Tender, L.M., and Newman, D.K. (2020). Extracellular DNA promotes 21 efficient extracellular electron transfer by pyocyanin in Pseudomonas aeruginosa biofilms. 22 Cell 182, 919-932.e19. 10.1016/j.cell.2020.07.006. 23 34.   Hoehler, T.M., Mankel, D.J., Girguis, P.R., McCollom, T.M., Kiang, N.Y., and Jørgensen, 24 B.B. (2023). The metabolic rate of the biosphere and its components. Proc Natl Acad Sci 25 USA 120, e2303764120. 10.1073/pnas.2303764120. 26 35.   Amy, P.S., and Morita, R.Y. (1983). Starvation-survival patterns of sixteen freshly isolated 27 open-ocean bacteria. Appl. Environ. Microbiol. 45, 1109–1115. 10.1128/aem.45.3.1109-28 1115.1983. 29 36.   Bratbak, G., and Dundas, I. (1984). Bacterial dry matter content and biomass 30 estimations. Appl. Environ. Microbiol. 48, 755–757. 10.1128/aem.48.4.755-757.1984. 31 37.   Chen, S.N. (2001). Growth Kinetics of Pseudomonas aeruginosa. Montana State 32 University Masters Thesis. https://scholarworks.montana.edu/xmlui/handle/1/8162. 33 38.   Geckil, H., Stark, B.C., and Webster, D.A. (2001). Cell growth and oxygen uptake of 34 Escherichia coli and Pseudomonas aeruginosa are differently effected by the genetically 35 engineered Vitreoscilla hemoglobin gene. J. Biotechnol. 85, 57–66. 10.1016/s0168-36 1656(00)00384-9. 37 39.   Finkel, S.E., and Kolter, R. (1999). Evolution of microbial diversity during prolonged 38 starvation. Proc Natl Acad Sci USA 96, 4023–4027. 10.1073/pnas.96.7.4023. 39 40.   Yakhnina, A.A., McManus, H.R., and Bernhardt, T.G. (2015). The cell wall amidase AmiB 40 is essential for Pseudomonas aeruginosa cell division, drug resistance and viability. Mol. 41 Microbiol. 97, 957–973. 10.1111/mmi.13077. 42   35 41.   Navarro, P.P., Vettiger, A., Ananda, V.Y., Llopis, P.M., Allolio, C., Bernhardt, T.G., and 1 Chao, L.H. (2022). Cell wall synthesis and remodelling dynamics determine division site 2 architecture and cell shape in Escherichia coli. Nat. Microbiol. 7, 1621–1634. 3 10.1038/s41564-022-01210-z. 4 42.   Arai, H. (2011). Regulation and Function of Versatile Aerobic and Anaerobic Respiratory 5 Metabolism in Pseudomonas aeruginosa. Front. Microbiol. 2, 103. 6 10.3389/fmicb.2011.00103. 7 43.   Pu, Y., Li, Y., Jin, X., Tian, T., Ma, Q., Zhao, Z., Lin, S.Y., Chen, Z., Li, B., Yao, G., et al. 8 (2019). ATP-Dependent Dynamic Protein Aggregation Regulates Bacterial Dormancy 9 Depth Critical for Antibiotic Tolerance. Mol. Cell 73, 143-156.e4. 10 10.1016/j.molcel.2018.10.022. 11 44.   Jiménez Otero, F., Newman, D.K., and Tender, L.M. (2023). Pyocyanin-dependent 12 electrochemical inhibition of Pseudomonas aeruginosa biofilms is synergistic with 13 antibiotic treatment. MBio 14, e0070223. 10.1128/mbio.00702-23. 14 45.   Schink, S., Ammar, C., Chang, Y.-F., Zimmer, R., and Basan, M. (2022). Analysis of 15 proteome adaptation reveals a key role of the bacterial envelope in starvation survival. 16 Mol. Syst. Biol. 18, e11160. 10.15252/msb.202211160. 17 46.   Schink, S., Polk, M., Athaide, E., Mukherjee, A., Ammar, C., Liu, X., Oh, S., Chang, Y.-F., 18 and Basan, M. (2024). Survival dynamics of starving bacteria are determined by ion 19 homeostasis that maintains plasmolysis. Nat. Phys. 20, 1332–1338. 10.1038/s41567-20 024-02511-2. 21 47.   Ciemniecki, J.A., and Newman, D.K. (2023). NADH dehydrogenases are the 22 predominant phenazine reductases in the electron transport chain of Pseudomonas 23 aeruginosa. Mol. Microbiol. 119, 560–573. 10.1111/mmi.15049. 24 48.   Torres, C.I., Marcus, A.K., Lee, H.-S., Parameswaran, P., Krajmalnik-Brown, R., and 25 Rittmann, B.E. (2010). A kinetic perspective on extracellular electron transfer by anode-26 respiring bacteria. FEMS Microbiol. Rev. 34, 3–17. 10.1111/j.1574-6976.2009.00191.x. 27 49.   Imlay, J.A. (2013). The molecular mechanisms and physiological consequences of 28 oxidative stress: lessons from a model bacterium. Nat. Rev. Microbiol. 11, 443–454. 29 10.1038/nrmicro3032. 30 50.   Glasser, N.R., Wang, B.X., Hoy, J.A., and Newman, D.K. (2017). The Pyruvate and α-31 Ketoglutarate Dehydrogenase Complexes of Pseudomonas aeruginosa Catalyze 32 Pyocyanin and Phenazine-1-carboxylic Acid Reduction via the Subunit Dihydrolipoamide 33 Dehydrogenase. J. Biol. Chem. 292, 5593–5607. 10.1074/jbc.M116.772848. 34 51.   Belliveau, N.M., Chure, G., Hueschen, C.L., Garcia, H.G., Kondev, J., Fisher, D.S., 35 Theriot, J.A., and Phillips, R. (2021). Fundamental limits on the rate of bacterial growth 36 and their influence on proteomic composition. Cell Syst. 12, 924-944.e2. 37 10.1016/j.cels.2021.06.002. 38 52.   Basan, M., Honda, T., Christodoulou, D., Hörl, M., Chang, Y.-F., Leoncini, E., Mukherjee, 39 A., Okano, H., Taylor, B.R., Silverman, J.M., et al. (2020). A universal trade-off between 40 growth and lag in fluctuating environments. Nature 584, 470–474. 10.1038/s41586-020-41 2505-4. 42   36 53.   Hoehler, T.M., Amend, J.P., Jørgensen, B.B., Orphan, V.J., and Lever, M.A. (2024). 1 Editorial: Studies on life at the energetic edge – from laboratory experiments to field-2 based investigations, volume II. Front. Microbiol. 14. 10.3389/fmicb.2023.1351761. 3 54.   Røy, H., Kallmeyer, J., Adhikari, R.R., Pockalny, R., Jørgensen, B.B., and D’Hondt, S. 4 (2012). Aerobic microbial respiration in 86-million-year-old deep-sea red clay. Science 5 336, 922–925. 10.1126/science.1219424. 6 55.   Jaussi, M., Jørgensen, B.B., Kjeldsen, K.U., Lomstein, B.A., Pearce, C., Seidenkantz, 7 M.-S., and Røy, H. (2023). Cell-specific rates of sulfate reduction and fermentation in the 8 sub-seafloor biosphere. Front. Microbiol. 14, 1198664. 10.3389/fmicb.2023.1198664. 9 56.   Kanaan, G., Hoehler, T.M., Iwahana, G., and Deming, J.W. (2023). Modeled energetics 10 of bacterial communities in ancient subzero brines. Front. Microbiol. 14, 1206641. 11 10.3389/fmicb.2023.1206641. 12 57.   Makarieva, A.M., Gorshkov, V.G., Li, B.-L., Chown, S.L., Reich, P.B., and Gavrilov, V.M. 13 (2008). Mean mass-specific metabolic rates are strikingly similar across life’s major 14 domains: Evidence for life’s metabolic optimum. Proc Natl Acad Sci USA 105, 16994–15 16999. 10.1073/pnas.0802148105. 16 58.   Price, P.B., and Sowers, T. (2004). Temperature dependence of metabolic rates for 17 microbial growth, maintenance, and survival. Proc Natl Acad Sci USA 101, 4631–4636. 18 10.1073/pnas.0400522101. 19 59.   Stouthamer, A.H. (1973). A theoretical study on the amount of ATP required for synthesis 20 of microbial cell material. Antonie Van Leeuwenhoek 39, 545–565. 21 60.   Mccollom, T.M., and Amend, J.P. (2005). A thermodynamic assessment of energy 22 requirements for biomass synthesis by chemolithoautotrophic micro-organisms in oxic 23 and anoxic environments. Geobiology 3, 135–144. 10.1111/j.1472-4669.2005.00045.x. 24 61.   van Bodegom, P. (2007). Microbial maintenance: a critical review on its quantification. 25 Microb. Ecol. 53, 513–523. 10.1007/s00248-006-9049-5. 26 62.   Walker, R.M., Sanabria, V.C., and Youk, H. (2023). Microbial life in slow and stopped 27 lanes. Trends Microbiol. 10.1016/j.tim.2023.11.014. 28 63.   Koch, A.L. (1971). The Adaptive Responses of Escherichia coli to a Feast and Famine 29 Existence. Adv Microb Physiol 6, 147–217. 30 64.   Rothman, D.H. (2024). Slow closure of Earth’s carbon cycle. Proc Natl Acad Sci USA 31 121, e2310998121. 10.1073/pnas.2310998121. 32 65.   Basta, D.W., Bergkessel, M., and Newman, D.K. (2017). Identification of fitness 33 determinants during energy-limited growth arrest in Pseudomonas aeruginosa. MBio 8, 34 e01170-17. 10.1128/mBio.01170-17. 35 66.   Kaila, V.R.I., and Wikström, M. (2021). Architecture of bacterial respiratory chains. Nat. 36 Rev. Microbiol. 19, 319–330. 10.1038/s41579-020-00486-4. 37 67.   Raba, D.A., Rosas-Lemus, M., Menzer, W.M., Li, C., Fang, X., Liang, P., Tuz, K., Minh, 38 D.D.L., and Juárez, O. (2018). Characterization of the Pseudomonas aeruginosa NQR 39 complex, a bacterial proton pump with roles in autopoisoning resistance. J. Biol. Chem. 40 293, 15664–15677. 10.1074/jbc.RA118.003194. 41   37 68.   Schink, B. (1997). Energetics of syntrophic cooperation in methanogenic degradation. 1 Microbiol. Mol. Biol. Rev. 61, 262–280. 10.1128/mmbr.61.2.262-280.1997. 2 69.   Jackson, B.E., and McInerney, M.J. (2002). Anaerobic microbial metabolism can proceed 3 close to thermodynamic limits. Nature 415, 454–456. 10.1038/415454a. 4 70.   Caro, T.A., McFarlin, J., Jech, S., Fierer, N., and Kopf, S. (2023). Hydrogen stable 5 isotope probing of lipids demonstrates slow rates of microbial growth in soil. Proc Natl 6 Acad Sci USA 120, e2211625120. 10.1073/pnas.2211625120. 7 71.   Coskun, Ö.K., Özen, V., Wankel, S.D., and Orsi, W.D. (2019). Quantifying population-8 specific growth in benthic bacterial communities under low oxygen using H218O. ISME J. 9 13, 1546–1559. 10.1038/s41396-019-0373-4. 10 72.   Blazewicz, S.J., Hungate, B.A., Koch, B.J., Nuccio, E.E., Morrissey, E., Brodie, E.L., 11 Schwartz, E., Pett-Ridge, J., and Firestone, M.K. (2020). Taxon-specific microbial growth 12 and mortality patterns reveal distinct temporal population responses to rewetting in a 13 California grassland soil. ISME J. 14, 1520–1532. 10.1038/s41396-020-0617-3. 14 73.   Kirchman, D.L. (2016). Growth rates of microbes in the oceans. Ann. Rev. Mar. Sci. 8, 15 285–309. 10.1146/annurev-marine-122414-033938. 16 74.   Jaishankar, J., and Srivastava, P. (2017). Molecular basis of stationary phase survival 17 and applications. Front. Microbiol. 8, 2000. 10.3389/fmicb.2017.02000. 18 75.   Jones, E.M., Marken, J.P., and Silver, P.A. (2024). Synthetic microbiology in 19 sustainability applications. Nat. Rev. Microbiol. 10.1038/s41579-023-01007-9. 20 76.   Shi, Z., Crowell, S., Luo, Y., and Moore, B. (2018). Model structures amplify uncertainty 21 in predicted soil carbon responses to climate change. Nat. Commun. 9, 2171. 22 10.1038/s41467-018-04526-9. 23 77.   Liang, C., Schimel, J.P., and Jastrow, J.D. (2017). The importance of anabolism in 24 microbial control over soil carbon storage. Nat. Microbiol. 2, 17105. 25 10.1038/nmicrobiol.2017.105. 26 78.   Dietrich, L.E.P., Okegbe, C., Price-Whelan, A., Sakhtah, H., Hunter, R.C., and Newman, 27 D.K. (2013). Bacterial community morphogenesis is intimately linked to the intracellular 28 redox state. J. Bacteriol. 195, 1371–1380. 10.1128/JB.02273-12. 29 79.   Babin, B.M., Atangcho, L., van Eldijk, M.B., Sweredoski, M.J., Moradian, A., Hess, S., 30 Tolker-Nielsen, T., Newman, D.K., and Tirrell, D.A. (2017). Selective Proteomic Analysis 31 of Antibiotic-Tolerant Cellular Subpopulations in Pseudomonas aeruginosa Biofilms. MBio 32 8. 10.1128/mBio.01593-17. 33 80.   Dietrich, L.E., Price-Whelan, A., Petersen, A., Whiteley, M., and Newman, D.K. (2006). 34 The phenazine pyocyanin is a terminal signalling factor in the quorum sensing network of 35 Pseudomonas aeruginosa. Mol. Microbiol. 61, 1308–1321. 10.1111/j.1365-36 2958.2006.05306.x. 37 81.   Choi, K.-H., and Schweizer, H.P. (2006). mini-Tn7 insertion in bacteria with single attTn7 38 sites: example Pseudomonas aeruginosa. Nat. Protoc. 1, 153–161. 39 10.1038/nprot.2006.24. 40   38 82.   Shanks, R.M.Q., Caiazza, N.C., Hinsa, S.M., Toutain, C.M., and O’Toole, G.A. (2006). 1 Saccharomyces cerevisiae-based molecular tool kit for manipulation of genes from gram-2 negative bacteria. Appl. Environ. Microbiol. 72, 5027–5036. 10.1128/AEM.00682-06. 3 83.   Jeske, M., and Altenbuchner, J. (2010). The Escherichia coli rhamnose promoter rhaPBAD 4 is in Pseudomonas putida KT2440 independent of Crp–cAMP activation. Appl. Microbiol. 5 Biotechnol. 85, 1923–1933. 10.1007/s00253-009-2245-8.  6   7   8   9   10