Inhibition of photoferrotrophy by nitric oxide in ferruginous environments

Geomicrobiology, University of Tübingen, Tübingen, Germany 2 Hydrogeology, University of Tübingen, Tübingen, Germany 3 School of Geographical Sciences, University of Bristol, Bristol, UK 4 Department of Earth System Science, Stanford University, Stanford, USA 5 School of Life Sciences, Arizona State University, Tempe, AZ, USA. 6 Department of Earth & Atmospheric Sciences, University of Alberta, Edmonton, Canada 7 School of Earth & Environmental Sciences and Centre for Exoplanet Science, University of St Andrews, St Andrews, UK


Abstract:
Anoxygenic phototrophic Fe(II)-oxidizers (photoferrotrophs) are thought to have thrived in Earth's ancient ferruginous oceans and played a primary role in the precipitation of Archean and Paleoproterozoic (3.8-1.85 Ga) banded iron formations (BIF). The end of BIF deposition by photoferrotrophs has often been interpreted as being the result a deepening of water column oxygenation below the photic zone concomitant with the proliferation of cyanobacteria. We suggest here that a potentially overlooked aspect influencing BIF precipitation by photoferrotrophs is competition with another anaerobic Fe(II)-oxidizing metabolism. It is speculated that microorganisms capable of coupling Fe(II) oxidation to the reduction of nitrate were also present early in Earth history when BIF were being deposited, but the extent to which they could compete with photoferrotrophs when favourable geochemical conditions overlapped is unknown. Utilizing microbial incubations and numerical modelling, we show that nitrate-reducing Fe(II)-oxidizers metabolically outcompete photoferrotrophs for dissolved Fe(II). Moreover, the nitrate-reducing Fe(II)-oxidizers inhibit photoferrotrophy via the production of toxic nitric oxide (NO). Four different photoferrotrophs, representing both green sulfur and purple non-sulfur bacteria, are susceptible to this toxic effect despite having genomic capabilities for NO detoxification. Indeed, despite NO detoxification mechanisms being ubiquitous in some groups of phototrophs at the genomic level (e.g. Chlorobi and Cyanobacteria) it is likely they would still be influenced by NO stress. We suggest that the production of NO during nitrate-reducing Fe(II) oxidation in ferruginous environments represents an as yet unreported control on the activity of photoferrotrophs in the ancient oceans and thus the mechanisms driving precipitation of BIF.

INTRODUCTION
Anoxygenic photoautotrophic Fe(II)-oxidizing bacteria, or "photoferrotrophs" (Equation 1), are 1 thought to have thrived in Earth's oceans prior to the rise of O2 and contributed to the deposition 2 of banded iron formations (Hartman, 1984;Widdel et al., 1993;Konhauser et al., 2002). As O2 3 began to rise, these microbes would have seen their habitats shrink, yet they are still thought to 4 have been capable of out-competing abiotic Fe(II) oxidation by O2 or respiration by 5 microaerophilic Fe(II)-oxidizers while the oxycline remained in the photic zone (Kappler et al.,6 2005), i.e., when photons could reach deeper anoxic waters. In modern anoxic environments containing both Fe(II) and nitrate (NO3 -), nitrate reduction 20 coupled to Fe(II) oxidation (Equation 2) is widespread (Bryce et al., 2018). During this process, 21 Fe(II) oxidation can be enzymatically driven (Straub et al., 1996;He et al., 2016) and/or occur abiotically (Klueglein and Kappler, 2013), catalyzed by reactive N-intermediates produced during 23 enzymatic reduction of nitrate, such as nitrite and nitric oxide (NO) (known as 24 chemodenitrification) (Klueglein and Kappler, 2013). In modern environments, such as sediments 25 (Melton et   exchange, the flushing of the headspace led to uninhibited growth in the mixed culture ( Figure 3).

84
N2O is also a potential toxin (Drummond and Matthews, 1994), therefore we directly tested 85 whether N2O could inhibit Fe(II) oxidation by R. ferrooxidans SW2, but did not observe any 86 inhibition, even at concentrations higher than those observed in Figure 1 (up to 90 µM N2O(aq)) 87 ( Figure S3). Our simulation and cultivation results combined confirm that inhibition of R. conditions ( Figure 4).

NO detoxification capability is widespread in phototrophs but inadequate to avoid inhibition
138 This is a pre-print submitted to Earth ArXiv and has not been peer reviewed 8 The inhibition effect we report here is not unique to R. ferrooxidans SW2. We additionally 139 tested whether inhibition of Fe(II) oxidation would occur when an alternative freshwater 140 photoferrotroph, Chlorobium ferrooxidans strain KoFox, was incubated with culture KS. In this 141 case, Fe(II) oxidation was delayed but not completely inhibited ( Figure S8). We also observed that  Figure S6). The Fe(II) oxidation mechanism in this case is of the type depicted in Figure 4c. 147 Sensitivity of these marine strains highlights that we also expect to observe a similar effect in the 148 marine realm. Interestingly, for both the freshwater and marine strains, the green sulfur bacteria 149 tested appeared to tolerate slightly higher nitrite concentrations than the purple non-sulfur bacteria 150 tested, in turn, suggesting a higher tolerance to NO. This may be the result of physiological 151 differences between the green sulfur and purple non-sulfur bacteria, or it could be because both 152 green sulfur bacteria strains do not exist in pure culture and thus may be "helped" by a partner 153 strain.   Kappler. Media recipes for other strains used can be found in Table S1.  Table S1.   Headspace dilution due to sampling was also considered. 499 The aqueous concentration changes for Fe(II), N-species and KS are given by: respectively. An addition of N2 as a result of headspace volume replacement during sampling is 509 accounted for by the addition of headspace gas at "atmospheric" (80% N2) partial pressure, 2 . 510 (Note: experiments were run under an anerobic 90:10 N2:CO2 atmosphere.) 511 All model variants were setup as well-mixed batch reactors. Partitioning of NO, N2O and N2 512 between aqueous and gas phases was considered in model variants that simulated nitrate-reducing 513 iron oxidation. The coupled system of ordinary differential equations was solved in MATLAB 514 using the built-in ordinary differential equation solver, ode15s. We fitted both the SW2-only and 515 KS-only models to measured concentration, cell density and partial pressure data using the least 516 squares MATLAB fitting tool, lsqnonlin. We fitted the logarithms of the parameters rather than the 517 parameters themselves, thereby alleviating the discrepancy between nominal values differing by 518 orders of magnitude. Our fitting scheme was based on minimizing the sum of squared differences 519 between measurements and simulated output. Additional weight was allocated to NO partial 520 pressure measurements. We justify increasing the importance of those measurements as they 521 represent a key feature in the observed toxicity response of both KS and phototrophs to NO 522 accumulation. Calibrated parameter values for photoferrotrophy and NDFO catalyzed by the KS 523 culture are presented in Table S1. To obtain a better readability of the tree, the branches in Figure 5 and Figure S9 were 556 collapsed based on a 16S % identity threshold. In Figure 5 we used a threshold of 97%, while in 557 Figure S9 we used a threshold of 90%. Additionally, in Figure 5 all non-photosynthetic strains were pruned from the tree. In all figures, the tips representing the four strains that were cultured in 559 this study were always kept as individual tips, regardless of their identity score.     showing that the mixed culture behaves similarly to the nitrate-reducing culture with regards to all 849 parameters, and Fe(II) oxidation in the mixed culture is incomplete.

Time [days]
KoFox Mix KS 26.4 g L -1 NaCl, 6.8 g L -1 MgSO4.7H20, 5.7 g L -1 MgCl2.6H2O, 1.5 g L -1 CaCl2.2H2O, 0.66 g L -1 KCl, 0.09 g L -1 KBr 0.4 g L -1 KH2PO4, 0.25 g L -1 NH4Cl Additives: 1 mL L -1 sterile filtered 7-vitamin solution (Widdel and Pfennig, 1981   Parameters with a relative error close to 1 have a low uncertainty. Supplementary Data S1. Supplementary information for mapping of nitric oxide detoxification abilities 978 in bacteria, particularly in phototrophs.    Table 4 BLAST_queries: contains accession numbers for the sequences that were used as queries in BLAST 995 searches to determine the presence/absence of genes in genomes. Additionally, for each group of genes, i.e., 996 genes for photosynthetic reaction centres (psaB, pshA/CT2020, psbA1, pufL); norV; norB (cnorB, qnorB); 997 and hmp, a Maximum-Likelihood phylogenetic tree is presented, showing the relationships between the 998 query sequences. As an inset in each tree, the alignment of the corresponding query sequences is shown, 999 highlighting gap regions and including a summary of % identity at each position in the alignment.