Feammox Acidimicrobiaceae bacterium A6, a lithoautotrophic electrode-colonizing bacterium

An Acidimicrobiaceae bacterium A6 (A6), from the Acitnobacteria phylum was recently identified as a microorganism that can carry out anaerobic ammonium oxidation coupled to iron reduction, a process also known as Feammox. Being an iron-reducing bacterium, A6 was studied as a potential electrode-reducing bacterium that may transfer electrons extracellularly onto electrodes while gaining energy from ammonium oxidation. Actinobacteria species have been overlooked as electrogenic bacteria, and the importance of lithoautotrophic iron-reducers as electrode-reducing bacteria at anodes has not been addressed. By installing electrodes in soil of a forested riparian wetland where A6 thrives, as well as in A6 bioaugmented constructed wetland (CW) mesocosms, characteristics and performances of this organism as an electrode-reducing bacterium candidate were investigated. In this study, we show that Acidimicrobiaceae bacterium A6 is a lithoautotrophic bacterium, capable of colonizing electrodes in the field as well as in CW mesocosoms, and that it appears to be an electrode-reducing bacterium since there was a boost in current production shortly after the CWs were seeded with Acidimicrobiaceae bacterium A6. IMPORTANCE Most studies on electrogenic microorganisms have focused on the most abundant heterotrophs, while other microorganisms also commonly present in electrode microbial communities such as Actinobacteria have been overlooked. The novel Acidimicrobiaceae bacterium A6 (Actinobacteria) is an iron-reducing bacterium that can colonize the surface of anodes and is linked to electrical current production, making it an electrode-reducing candidate. Furthermore, A6 can carry out anaerobic ammonium oxidation coupled to iron reduction, therefore, findings from this study open up the possibility of using electrodes instead of iron as electron acceptors as a mean to promote A6 to treat ammonium containing wastewater more efficiently. Altogether, this study expands our knowledge on electrogenic bacteria and opens up the possibility to develop Feammox based technologies coupled to bioelectric systems for the treatment NH4+ and other contaminants in anoxic systems.

(CW) mesocosms, characteristics and performances of this organism as an electrode-reducing 23 bacterium candidate were investigated. In this study, we show that Acidimicrobiaceae bacterium 24 A6 is a lithoautotrophic bacterium, capable of colonizing electrodes in the field as well as in CW 25 mesocosoms, and that it appears to be an electrode-reducing bacterium since there was a boost in 26 current production shortly after the CWs were seeded with Acidimicrobiaceae bacterium A6. 27

IMPORTANCE 28
Most studies on electrogenic microorganisms have focused on the most abundant 29 heterotrophs, while other microorganisms also commonly present in electrode microbial 30 communities such as Actinobacteria have been overlooked. The novel Acidimicrobiaceae 31 bacterium A6 (Actinobacteria) is an iron-reducing bacterium that can colonize the surface of 32 anodes and is linked to electrical current production, making it an electrode-reducing candidate. 33 Furthermore, A6 can carry out anaerobic ammonium oxidation coupled to iron reduction, 34 therefore, findings from this study open up the possibility of using electrodes instead of iron as 35 electron acceptors as a mean to promote A6 to treat ammonium containing wastewater more 36 efficiently. Altogether, this study expands our knowledge on electrogenic bacteria and opens up 37 the possibility to develop Feammox based technologies coupled to bioelectric systems for the 38 treatment NH 4 + and other contaminants in anoxic systems. 39

INTRODUCTION
the electrode (cathode) in the more oxidized soil. The well-controlled and monitored CW 152 mesocosms were able to provide insights that were not clearly resolved by the field studies. 153 Phylogenetic analyses and microbial community structure. 154 The phylogenetic diversity at the phylum level found in the biofilm and soil samples taken 155 from the field as well as the CW mesocosms studies (Figure 3) show that Proteobacteria and 156 Acidobacteria represent on average more than 70% of the diversity found in all samples, 157 followed in abundance by Chloroflexi and Bacteroidetes in the field study, and Firmicutes and 158 Bacteroidetes in the CW mesocosm study. These highly abundant groups make up more than 80% 159 of the population found in all samples. All these phyla are commonly found in soil and in 160 bioelectrochemical systems due to their ERB ability (12, 13, 31, 32), therefore, they are common 161 subjects of study. Actinobacteria, the phylum to which Acidimicrobiaceae bacterium A6 belongs 162 to, represents as little as 2, 4, 5 and 3.5 % of the relative abundance found at the three different 163 sites and CW mesocosms respectively. Nonetheless, Actinobacteria ranks in the top 5 most 164 abundant phyla found in each field site and the CW mesocosms, and makes it to the third 165 position for some electrode biofilm samples from the field study. 166

Field experiment phylogenetic analyses 167
The Actinobacteria phylum contains Acidimicrobiaceae bacterium A6, described as an 168 unidentified_Acidimicrobiales at the genus level in the samples from the field sites, because its 169 16s rDNA sequence was not available in the public data bases at the time of the field study. The 170 OTU annotated as unidentified_Acidomicrobiales had ≥ 97 % sequence identity with A6, thus 171 confirming the presence of this Feammox bacterium in our samples. A total of 316 genera were 172 annotated in the phylogenetic analysis, however, between 51% (site 1-2) to as much as 69% (site category. Among the top 100 most abundant genera, the unidentified Acidomicrobiales ranked 175 56 th (Figure 4). The genera with the highest relative abundance at site 1-2, characterized by its 176 waterlogged condition, were Sideroxydans (Proteobacteria), an Fe(II) oxidizer (33), and Geothrix 177 (Acidobacteria), a known ERB (34). At sites 3-4 and 5-6, the most relative abundant genera were 178 the Bryobacter (Acidobacteria) an aerobic heterotroph, candidatus_Solibacter (Acidobacteria), 179 Acidibacter (Proteobacteria) an FeRB, the autotroph Acidothermus (Actinobacteria), and 180 ( Figure S3) in the CW mesocosms, their population was still enriched on the anodes compared to 196 the surrounding sediments ( Figure 2). In addition to Geobacter, 10 other genera that include 197 known electrogenic bacteria species were also detected in all CW mesocosm samples among the 198 top 100 genera, including Geothrix, Desulfobulbus, Desulfovibrio, Pseudomonas, Clostridium 199 (2), Desulfotomaculum (35), Enterobacter (36), Bacillus, Rhizomicrobium (37) and 200 Anaeromyxobacter (31). Most of the mentioned electrogenic bacteria are more abundant in the 201 electrode's biofilms than in the nearby soil samples (paired two sample t-test, p < 0.05; Table S3). 202 In fact, the electrogenic bacteria form a substantial portion of the microbial community from the 203 CW mesocosm samples, making up 3.0 -8.5 % of the total sequences for the CW soil samples 204 and 4.5 -14.4 % for CW electrode biofilm samples. Those electrogenic bacteria that are 205 enriched on electrodes compared to the nearby soil are all much more abundant than A6, 206 resulting in lower relative abundance of A6 on the electrode biofilms compared to the nearby soil, 207 even though the numbers of A6 are higher on the electrodes than the soil. 208

Current pulse after the injection of A6 enrichment culture into the CWs 209
Though the high and low Fe level CW mesocosms had similar current profiles before the 210 injection of the A6 enrichment culture, the current profiles right after the injection showed a 211 noticeable difference ( Figure 5). The electrical current between electrode pairs in the low Fe CW 212 mesocosms increased after five days following the A6 enrichment culture injection, and then 213 descended to the previous level after 50 days. However, currents between electrode pairs in the 214 high Fe CW mesocosms remained within a similar range as prior to the A6 enrichment culture 215 transferred through the electrode pairs in the low Fe mesocosm than in high Fe mesocosm as the 218 same amount of bacterium A6 was introduced into both mesocosms. 219 It should be noted that the samples for DNA extraction were taken almost four months after 220 the injection of A6 and three months after the current pulse disappeared in the low Fe mesocosm. 221 Therefore, phylogenetic results and A6 numbers discussed above may not properly capture the 222 microbial community at the time of the pulse in the current. 223

A6 colonization and electron transfer to electrodes 225
The number of A6 quantified on the biofilm formed on the electrodes confirmed the 226 hypothesis that this bacterium is able to colonize electrodes to use them as an alternative electron 227 acceptor to Fe(III), thus enhancing its number compared to its surroundings ( Figure 1). The CW 228 permitted us to stablish A6's preference over the electrodes in the more reduced soil (anodes) 229 than over the rest of the environment (Figure 2). The consistent trend of A6 enrichment on 230 electrodes was found on the anodes compared to surrounding sediments but not always on the 231 electrodes that have been operated for a time period as cathodes (electrode 1 and 2 in the CW). 232 This indicates that A6 is able to colonize and be enriched on the surface of anodes. 233 It is interesting to note that injection of A6 into the low Fe CW resulted in a pulse in current 234 while this was not observed in the high Fe CW. This indicates that when little bioavailabe Fe was 235 present, A6 showed a more immediate affinity to colonize and transfer electrons to the electrodes. 236 The decrease in the current after about 50 days of the injection indicates that A6 numbers or 237 activity on the electrodes in the low Fe CW decreased over time after the injection. much lower in the CW samples, being outranked by other FeRB, with most of which A6 showed 241 negative correlations between their relative abundances in the field ( Figure 6). This is not the 242 case for the relative abundance with other non-metal reducing bacteria, also from field soil and samples. This negative correlation should be taken into account when implementing A6 in bioelectrochemical system such as MFC, particularly those that feed on organic carbon as the 262 electrode donors since these are systems where bacteria such as Geobacter spp. thrive and could 263 affect A6's population negatively. 264 The results from this study show that Acidimicrobiaceae bacterium A6's, which is an iron 265 reducer, is capable of colonizing electrodes in the field as well as in constructed wetland 266 mesocosoms, resulting for the conditions studied, in higher cell counts on the electrodes than on 267 the soil. Thus, Acidimicrobiaceae bacterium A6 is a novel anaerobic litoautothroph from the 268 Actinobacteria phyla capable of using electrodes as its terminal electron acceptor. Altogether, and rinsed three times in distilled water (38). Each electrode set was connected by a titanium (Ti) 280 wire cleaned with sandpaper (ultra-corrosion-resistant Ti wire, 0.08cm in diameter, McMaster-graphite plate to ensure a tight connection between the wire and the graphite plates to allow for 283 low contact resistance <0.5 Ω. The Ti wire was long enough to allow for 10 or 30 cm separation 284 between the graphite plates ( Figure 8). 285 Two pairs of electrodes were placed at three different sites. Each pair consisted of a shallow 286 electrode placed no deeper than 5cm into the soil, connected to another electrode with either 10 287 cm or 30 cm separation, i.e. a total of 6 sets (Table 2) in a temperate forested riparian wetland 288 located at Assunpink Wildlife Management Area in New Jersey, USA. This is the location were 289 the Feammox reaction was first discovered (8), and later the Feammox bacteria A6, was 290 identified in samples from this site (7) and isolated (10). Detailed physicochemical characteristic 291 of the soil have been described in previous studies (8, 39). Electrodes sets 1 and 2 were place in a 292 fully flooded location, sets 3 to 6 were placed in a wet but unsaturated location. The electrode 293 sets were left in the field for 52 days between June 13 and August 03, 2016. 294

Field electrodes recovery and sampling. 295
After 52 days, the electrode pairs were recovered by digging them out of the soil and placing 296 each electrode individually in a sealed bag. All the electrodes were surrounded by soil. 297 Furthermore, a soil sample was taken from a depth of ~20 cm from each site and placed in a 298 sealed bag. All samples were transported to the laboratory within 2 hours and immediately stored 299 at 4C until processed for analysis. 300 The samples obtained from the electrodes and soil are enumerated in Table 2. To analyze 301 the biomass attached to the electrodes, first, the loosely bound soil was removed by gently 302 shaking the electrode. Second, duplicate samples were taken from the soil layer (< 2 mm thick)

Constructed wetland mesocosms and electrodes set up 308
Controlled conditions were required to gain further insights into the electrogenesis of A6. 309 Therefore, electrode pairs were installed in two constructed wetland (CW) mesocosms that were 310 operated in a growth chamber (Environmental Growth Chambers, www.egc.com). The 311 constructed wetlands were designed as continuous up-flow mesocosms with water surface above 312 the sediment. Standard-wall PVC pipes, pipe fittings (pipe size 6, inner diameter 6 inch, 313 McMaster-Carr code 48925K25, 4880K852 and 4880K131) were used to assemble the 314 mesocosms. The dimension of the mesocosms and the design of sampling ports are shown in 315 Figure 9. The inflow was injected from the bottom and an opening with 1-inch diameter was 316 drilled for the effluent to maintain constant water levels. Along the longitudinal axis were five 317 sampling ports, spaced 6 cm apart from each other, and lysimeters were used for pore water 318 sampling from the CW mesocosms. Nylon meshes (70 µm opening, opening area 33%) were 319 placed at 15 cm and 40 cm depth to separate the root zone/non-root zone respectively. At the 320 bottom of the mesocosms, glass beads were used as bed material to disperse the inflow evenly. 321 The CW substrate was a mixture of ASTM standard 20-30 sand, peat moss, and wetland 322 sediments from Assunpink Wildlife Management Area in New Jersey, USA (the same location 323 where the field electrodes were placed) with 1: 1: 0.5 ratio (by weight). 324 To investigate if A6 could colonize the electrodes and transfer electrons onto anodes, five 325 graphite electrodes were installed in each mesocosm at the same depths as the five sampling ports. The electrodes were made of rectangular graphite rods [10.16 (L) x 1.28 (W) x 1.28 cm 327 (H)]. The preparation of the electrodes and wires was the same as described above. Each of the 328 four electrodes that were placed into the more reduced zones (anode) was then connected via a 329 cleaned titanium wire to the electrode that was in the most oxidized zone (cathode) of the CW as 330 shown in Figure 9. peat moss both had some Fe, even the CW that was not augmented with ferrihydrite had Fe. As 339 ferrihydrite is amorphous and highly bioavailable, the Fe(III) source for A6 would mostly come 340 from the ferrihydrite added in the CW substrate. Examination of several Fe(III) sources as the 341 electron acceptor for A6 showed that this ferrihydrite yielded the highest Feammox activity 342 among the Fe(III) sources studied (10). 343 To inoculate the CW substrate with A6, 250 mL of an A6 enrichment culture (10 9 -10 10 344 CFU/g sludge, 70% A6 in biomass) was added to the CW substrate for each column. The 345 substrate was then thoroughly mixed prior to loading into the mesocosm columns. The mesocosms and pumps were placed in a growth chamber that is simulating the summer 363 climate of New Jersey (Table S2). Since the A6 enrichment culture was blended in open-air with 364 the CW substrate and the CW mesocosm conditions were rather oxidized at the beginning of 365 their operation, it was uncertain that a viable A6 population did get established in the mesocosms. 366 Therefore, after four months of operation, when the mesocosm became more reduced, another 367 A6 enrichment culture (~ 10 7 A6 count/mL culture, 250 mL per CW mesocosm) was injected 368 from the bottom into each CW on day 124 to ascertain the colonization of A6 in the mesocosms. 369 from sampling ports and collecting effluents from the top of the CWs. At the end of the 374 experiment, the CW mesocosms were dismantled and soil samples were taken for analysis of 375 microbial communities. Soil samples were collected at depths bracketing the sampling ports (6 -376 9 cm, 12 -15 cm, 18 -21 cm, 24 -27 cm, 30 -33 cm) as well as at the top layer of soil (0 -3 377 cm). For each depth, a 1 -1.5 g wet soil sample was collected and frozen at -20 °C before 378 proceeding with the DNA extraction. Electrodes installed in the CW mesocosms were removed 379 carefully and biofilm samples from the electrodes were obtained using the same procedure as 380 described above (see Supplemental Materials, Table S1. B for details). Since soil and electrode 381 sampling requires sacrificing the mesocosms, no soil samples were collected during the 382 operation of the CW mesocosms. Samples obtained from the CW mesocosm sediments and 383 electrodes are enumerated in Table 3. 384

DNA extraction, Acidimicobiaceae bacterium A6 quantification and phylogenic analysis 385
Total genomic DNA was extracted from each biofilm sample obtained in duplicate from a 386 half face or full face of each electrode deployed in the field, except for the deep electrode of set 6, 387 which was only partially recovered and only one sample could be obtained from all faces. Since 388 many electrodes in the CWs had a much lower biofilm mass than the filed electrodes, only one 389 sample was recovered for DNA extraction from the CW electrodes. DNA was also extracted 390 from each soil sample as described above (Table S1)  Paired-end reads were assembled by using FLASH V.1.2.7 (43). Raw reads were processed 429 according to QIIME V1.7.0 quality controlled process (44) and chimeric sequences were filtered 430 out using UCHIME algorithm (45). For all samples (field and CW) a minimum of 25,000 431 sequences were obtained. These resulting sequences were clustered into operational taxonomic 432 units (OTUs) using Uparse V7.0.1001 (46). Sequences with ≥97% similarity were assigned to 433 the same OTUs. A total of 3206 OTUs were produced across all field samples and 2870 OTUs 434 were produced across all CW mesocosm samples, with a range between 1422-1794 OTUs per 435 field sample and 674 -1481 OTUs per mesocosm sample. A representative sequence for each 436 samples were standardized using the least sequence number obtained from all samples so that the 443 same number of sequences were used for calculating the relative abundance of OTUs. 444

Soil surface area analysis 445
Nitrogen sorption was used to determine the surface area of the soil samples taken from the 446 field (Table 2) and the CW (Table 3). Prior to the analysis, samples were oven-dried at 56ºC until 447 the mass stabilized. Subsequently, the samples were degassed at 60ºC and 0.1 mmHg using a 448 Smart Micrometrics VacPrep (Norcross, GA, USA). The nitrogen sorption measurements were 449 conducted using a Micromeritics 3FLEX (Norcross, GA, USA), using the BET method 450 (Brunauer-Emmett-Teller) to calculate the surface area of the soil. The measurements obtained 451 were used to normalize the bacterial count data by surface area. 452

Analytical Methods 453
The sediment's iron concentrations were analyzed using the ferrozine method (50). Briefly, 454 0.5 mL sediment sample was added to 9.5 mL 0.5 M HCl and shaken for 24 hours at room 455 temperature to extract Fe(II). In total, 60 μL 6.25 M NH 2 OH·HCl was added to 3 mL of 456 extraction solution and shaken for 24 hours at room temperature to reduce Fe(III) to Fe(II). For 457 the chromogenic reaction, 60 μL of extraction solution was added to 3 mL 1 g/L ferrozine 458 solution (pH 7.0) and reacted for 30 minutes. The concentrations of Fe(II) and total Fe were 459 measured by reading the absorbance at the 562-nm wavelength using a Spectronic® Genesys TM 2 The voltages between electrode pairs in the CW mesocosms were measured using a multimeter 465 and currents were calculated accordingly. 466

Statistical Analysis. 467
The Welch t-test statistical analysis was used to determine if there were statistical 468 differences in bacterial counts per surface area between electrode and soil samples in the field 469 and CW mesocosm experiments. Paired two sample t-tests were conducted to determine if there 470 were statistical differences in relative abundance for electrode/soil pairs in the CW mesocosms.