Surface-induced formation and redox-dependent staining of outer membrane extensions in Shewanella oneidensis MR-1

The metal-reducing bacterium Shewanella oneidensis MR-1 produces extensions of its outer membrane (OM) and periplasm that contain cytochromes responsible for extracellular electron transfer (EET) to external redox-active surfaces, including minerals and electrodes. While the role of multi-heme cytochromes in transporting electrons across the cell wall is well established, their distribution along S. oneidensis OM extensions is also thought to allow lateral electron transport along these filaments. These proposed bacterial nanowires, which can be several times the cell length, would thereby extend EET to more distant electron acceptors. However, it is still unclear why these extensions form, and to what extent they contribute to respiration in living cells. Here, we investigate physical contributors to their formation using in vivo fluorescence microscopy. While previous studies focused on the display of S. oneidensis outer membrane extensions (OMEs) as a response to oxygen limitation, we find that cell-to-surface contact is sufficient to trigger the production of OMEs, including some that reach >100 µm in length, irrespective of medium composition, agitation, or aeration. To visualize the extent of heme redox centers along OMEs, and help distinguish these structures from other extracellular filaments, we also performed histochemical redox-dependent staining with transmission electron microscopy on wild type and cytochrome-deficient strains. We demonstrate that redox-active components are limited to OMEs and not present on other extracellular appendages, such as pili and flagella. We also observed that the loss of 8 functional periplasmic and outer membrane cytochromes significantly decreased both the frequency and intensity of redox-dependent staining found widespread on OMEs. These results will improve our understanding of the environmental conditions that influence the formation of S. oneidensis OMEs, as well as the distribution and functionality of EET components along extracellular appendages.

The metal-reducing bacterium Shewanella oneidensis MR-1 produces extensions of its outer 12 membrane (OM) and periplasm that contain cytochromes responsible for extracellular electron transfer 13 (EET) to external redox-active surfaces, including minerals and electrodes. While the role of multi-14 heme cytochromes in transporting electrons across the cell wall is well established, their distribution 15 along S. oneidensis OM extensions is also thought to allow lateral electron transport along these 16 filaments. These proposed bacterial nanowires, which can be several times the cell length, would 17 thereby extend EET to more distant electron acceptors. However, it is still unclear why these extensions 18 form, and to what extent they contribute to respiration in living cells. Here, we investigate physical 19 contributors to their formation using in vivo fluorescence microscopy. While previous studies focused 20 on the display of S. oneidensis outer membrane extensions (OMEs) as a response to oxygen limitation, 21 we find that cell-to-surface contact is sufficient to trigger the production of OMEs, including some that 22 reach >100 µm in length, irrespective of medium composition, agitation, or aeration. To visualize the 23 extent of heme redox centers along OMEs, and help distinguish these structures from other 24 extracellular filaments, we also performed histochemical redox-dependent staining with transmission 25 electron microscopy on wild type and cytochrome-deficient strains. We demonstrate that redox-active 26 components are limited to OMEs and not present on other extracellular appendages, such as pili and 27 flagella. We also observed that the loss of 8 functional periplasmic and outer membrane cytochromes 28 significantly decreased both the frequency and intensity of redox-dependent staining found widespread 29 on OMEs. These results will improve our understanding of the environmental conditions that influence 30 the formation of S. oneidensis OMEs, as well as the distribution and functionality of EET components 31 along extracellular appendages. 32

Introduction 33
Shewanella oneidensis MR-1 is a Gram-negative, facultative anaerobic heterotrophic bacterium with 34 versatile respiratory capabilities: in its quest for energy, it can utilize an array of soluble and insoluble 35 electron acceptors, from oxygen to extracellular solid surfaces such as minerals and electrodes. This 36 ability to couple intracellular reactions to the respiration of external surfaces, known as extracellular 37 electron transfer (EET), allows microbial catalytic activity to be harnessed on the electrodes of 38 bioelectrochemical technologies ranging from microbial fuel cells to microbial electrosynthesis 39 (Nealson, 2017;Schröder and Harnisch, 2017 understanding of the role of S. oneidensis OMEs will therefore require challenging in vivo 60 measurements of their specific impact on extracellular respiration and observations of the membrane 61 protein dynamics that allow inter-cytochrome electron exchange and redox conduction (Zacharoff and  62 El-Naggar, 2017). 63

64
Beyond the detailed mechanism of electron transport along these structures, additional questions 65 remain regarding the physical and environmental conditions that trigger their formation. The S. 66 oneidensis OMEs can extend to several times the cell length, and have been observed with a range of 67 morphologies from chains of interconnected outer membrane vesicles to membrane tubes (Pirbadian 68 et al., 2014). Since early reports suggested that they form in response to electron acceptor limitation, 69 particularly oxygen limitation (Gorby et al., 2006), subsequent studies involving these OMEs have It was previously shown that S. oneidensis membrane vesicles, which form the basis of OMEs, are 84 redox-active, and that this activity likely stems from the cytochromes present on the purified vesicles 85 (Gorby et al., 2008 which might influence OME production in S. oneidensis MR-1. We find that cell-to-surface contact is 99 sufficient to trigger the formation of S. oneidensis OMEs under a wide range of conditions. To assess 100 the extent of cytochrome-dependent redox activity in these structures, we implemented heme-101 dependent staining with transmission electron microscopy to compare OMEs in wild type and 102 cytochrome-deficient strains. In doing so, we also probed 3 types of extracellular filaments (OMEs,  103 flagella, and pili) for these EET components. We find that periplasmic and outer membrane 104 cytochromes are responsible for most of the redox activity detected using this assay, and that these 105 components are limited to OMEs and do not associate with flagella or pili. The reactor consisted of a clean glass tube (thickness 1.5 mm, interior diameter 24.7 mm, and length 140 50 mm) glued on to a clean 43 mm × 50 mm no. 1 thickness glass coverslip (Thermo Scientific) using 141 waterproof silicone glue (General Electric). The autoclaved reactor was placed on the inverted 142 microscope, and a peristaltic pump (Cole-Parmer Masterflex L/S Easy-Load II) was used to control 143 injection of filtered air at a rate of 3.6 mL/min into the reactor. The air inlet (22G 3" sterile needle) 144 was placed 1-2 mm from the coverslip bottom of the reactor so as to ensure oxygen availability and 145 good mixing near the focal plane. Time-lapse imaging was started immediately following introduction 146 of 10 mL of the cell-media mixture into the reactor and continued for 2 h with images acquired in 5 147 min increments. Oxygen levels in the reactor were measured by a dissolved oxygen probe (Milwaukee 148 Instruments MW600) at various levels (e.g. 1 mm from bottom, middle, and 1 mm from top) over time 149 after cells were added. To check whether the planktonic cells also displayed OMEs, imaging was 150 stopped after the surface-attached cells produced OMEs, and 400 µL of the planktonic mixture 151 (obtained within 1-2 mm from the top solution-air interface) was gently pipetted to a new clean 152 coverslip, and immediately imaged for another 2 h. 153

Heme-Reactive Staining and Transmission Electron Microscopy 154
All heme staining experiments were performed on cells attached to electron microscopy grids 155 recovered from the perfusion flow imaging platform after confirmation of OME production using 156 fluorescence microscopy (Subramanian et al., 2018). To accomplish this, an X-thick holey carbon-157 coated, R2/2, 200 mesh Au NH2 London finder Quantifoil EM grid was glued to the glass coverslip, 158 with the carbon film-coated side facing away from the glass, before sealing the perfusion chamber.  . 1), but 199 also in near-saturating oxygen conditions (6.5-7.5 ppm O2) provided by a glass-bottomed reactor that 200 allowed air injection during in vivo microscopy (Fig. 2). Though it can take up to several hours for a 201 majority of surface-attached cells to produce OMEs, we can observe production of OMEs as early as 202 10 min after cells contact the surface of a glass coverslip (Figs. 2, S1). To further examine the role of 203 surface contact, planktonic cells from the bulk oxygenated reactor were sampled 2 h after the reactor 204 was inoculated (approx. 1.5 h after OMEs started being produced by surface-attached cells) and 205 transferred to clean coverslips for observation. These previously planktonic cells showed no evidence 206 of OMEs at the time of sampling, but then also went on to begin to display OMEs within 35 min after 207 contacting the surface (Fig. 2). These observations were not limited to the defined minimal medium 208 used, a particular surface chemistry, or mixing conditions; post-attachment OME production was also 209 observed in rich (LB) medium or in buffer (PBS), on different surfaces (glass coverslips and carbon-210 coated electron microscopy grids), and regardless of liquid flow or agitation (Figs. S1, S2). To ensure 211 that the used cell density did not result in O2-limiting conditions selectively at the surface, we also 212 experimented with sparse coverage, down to 5-20 cells per field of view (112 µm × 112 µm) in a well-213 mixed and oxygenated reactor, and confirmed that these cells also produced OMEs (Fig. S2C). 214 Taken collectively, these observations of OME production by surface-attached cells, but not by 216 planktonic cells until subsequent attachment, and regardless of medium composition, surface type, and 217 oxygen availability, point to surface contact as the primary determinant of OME production by S. produced OMEs during 3.5 h of perfusion culture (Fig. 1). This precise quantification is to-volume ratio that these structures present (Pirbadian et al., 2014). Consistent with this proposal, we 259 occasionally captured multiple extensions from single cells (Fig. S3) as well as in vivo fluorescent 260 observations of remarkably long OMEs, likely the longest observed to date. Fig. 3 and Movie S1 261 captures a cell producing a >100 µm OME at a rate over 40 µm/h, at the same time that the cell surface 262 area appeared to shrink by an amount consistent with the newly displayed OME. 263

Redox-Dependent Staining of Extracellular Filaments 264
The localization of the multi-heme cytochromes responsible for EET to OMEs has been previously 265 demonstrated by immunofluorescence observations of MtrC and OmcA (Pirbadian et al., 2014), as 266 well as electron cryotomography observations of outer membrane and periplasmic electron densities 267 consistent with cytochrome dimensions (Subramanian et al., 2018). To examine the distribution and 268 activity of the heme iron redox centers along the OMEs, we applied the heme-reactive 3,3'-269 diaminobenzidine (DAB)-H2O2 staining procedure (McGlynn et al., 2015), where the iron centers 270 catalyze the oxidation of DAB, forming a localized dark precipitate that can be observed with the 271 resolution of transmission electron microscopy (TEM). As expected, the OMEs clearly stained for 272 heme, with a noticeable <50 nm band of dark precipitate lining the vesicles that compose the entire 273 structure (Fig. 4). Staining was clearly limited to the OMEs and was absent from the other extracellular 274 filaments observed, demonstrating that the cytochromes do not associate with pili and flagella (Fig. 4).

275
The absence of staining in these structures, even when observed in contact with the OMEs (Fig. 4) 60 OMEs analyzed per condition. Using image processing to compare OME staining to background 289 intensities (see Materials and Methods), we found that the majority (92%) of wild type OMEs stained 290 for heme, but none stained in the chemical control where H2O2 was omitted (Fig. 5). In contrast, a 291 fraction (39%) of OMEs in the mutant strain exhibited heme staining, 2.4-fold less than in wild type 292 (p < 0.0001, Pearson's chi-squared test) (Fig. 5). While lacking all cytochromes necessary for EET, 293 staining in the mutant OMEs is likely due to the additional periplasmic cytochromes, including the 294 flavocytochrome FccA present that functioned as the terminal fumarate reductase to support respiration 295 of fumarate in our anaerobic cultures. Consistent with this interpretation, staining intensity was 3.6-296 fold stronger in the wild type than in the mutant (p < 0.0001, Student's t-test, two-sample assuming 297 equal variances) (Fig. 5). Relative to the mutant control, the observed wild type increase in both 298 staining frequency and intensity indicates that the periplasmic and outer membrane cytochromes 299 necessary for EET contribute much of the redox capacity of the OMEs. 300 301 Given its ability to discriminate between cytochrome-containing and cytochrome-free extracellular 302 filaments, and to examine the effect of specific mutations, this heme visualization strategy may hold 303 promise for understanding the presence of redox centers in a variety of microbial systems. However, a 304 detailed understanding of the extent to which these redox centers enable long-distance electron 305 transport along OMEs requires: (i) applying electrochemical techniques, recently used to measure 306 redox conduction in biofilms (Xu et al., 2018;Yates et al., 2016), specifically to OMEs or their MV 307 constituents; and (ii) measurements of the diffusive dynamics of redox molecules along membranes, 308 to test the hypothesis that these dynamics facilitate a collision-exchange mechanism of inter-protein 309 electron transport over micrometer length scales (Subramanian et al., 2018). We are actively pursuing 310 these electrochemical and dynamics measurements. 311

4
Conclusions 312 In summary, we investigated physical contributors to the production of OMEs by Shewanella 313 oneidensis MR-1 and applied heme-reactive staining to examine the extent of the redox centers along 314 the extensions. While previous studies focused on the role of oxygen limitation in triggering the 315 formation of these structures, we demonstrated that surface contact is sufficient to trigger production 316 of OMEs under a variety of medium, agitation, and aeration conditions. In addition, we show that the 317 multi-heme cytochromes necessary for EET contribute much of the redox-dependent staining 318 widespread on OMEs, and that these EET components do not associate with other extracellular 319 filaments. In addition to describing some reproducible microscopic and histochemical techniques to 320 observe redox-functionalized membrane extensions, these observations motivate additional studies to 321 understand the extent to which Shewanella oneidensis OMEs can contribute to EET and long-distance 322 redox conduction. 323

Conflict of Interest 324
The authors declare that the research was conducted in the absence of any commercial or financial 325 relationships that could be construed as a potential conflict of interest. We are grateful to Professor Jeffrey A. Gralnick for providing the cytochrome mutant strain. 335 Transmission electron microscopy at 200 kV was performed at the University of Southern California's 336 Core Center of Excellence in Nano Imaging. We also thank Christopher Buser for imaging our electron 337 microscopy samples at 80 kV at the Huntington Medical Research Institutes. were visualized with the red membrane stain FM 4-64FX. Time (t = 0) marks time since OME 473 production (white arrow). Images depict progression of events, such as (A) cell contacting surface, (B) 474 OME first visible (white arrow), (C) OME elongates and begins to fold, (D) further OME elongation, 475 (E) OME reaches longest point visible during this experiment. (Scale bars: 10 µm.) 476