bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1 Marine Fe-oxidizing Zetaproteobacteria: Historical, ecological, and genomic

2 perspectives

3

4 Sean M. McAllister,a Ryan M. Moore,b Amy Gartman,a* George W. Luther, III,a David

5 Emerson,c and Clara S. Chana,d#

6

7 aSchool of Marine Science and Policy, University of Delaware, Newark, Delaware, USA

8 bCenter for Bioinformatics and Computational Biology, University of Delaware, Newark,

9 Delaware, USA

10 cBigelow Laboratory for Ocean Sciences, East Boothbay, Maine, USA

11 dDepartment of Geological Sciences, University of Delaware, Newark, Delaware, USA

12

13 Running Head: The marine Fe-oxidizing Zetaproteobacteria

14

15 #Address correspondence to Clara S. Chan, [email protected].

16 *Present address: Amy Gartman, USGS Pacific Coastal and Marine Science Center,

17 Santa Cruz, California, USA

1 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

18 Abstract

19 The Zetaproteobacteria are a class of bacteria typically associated with marine

20 Fe oxidizing environments. First discovered in the hydrothermal vents at Loihi

21 Seamount, Hawaii, they have become model organisms for marine microbial Fe

22 oxidation. In addition to deep sea and shallow hydrothermal vents, Zetaproteobacteria

23 are found in coastal sediments, other marine subsurface environments, steel corrosion

24 biofilms, as well as saline terrestrial aquifers and springs. Isolates from a range of

25 environments all grow by Fe oxidation. Their success lies partly in their microaerophily,

26 which enables them to compete with abiotic Fe oxidation at the low O2 concentrations

27 common to Fe(II)-rich oxic/anoxic transition zones. Also, Zetaproteobacteria make a

28 variety of biomineral morphologies as a repository for Fe(III) waste, and as attachment

29 structures. To determine the known diversity of the Zetaproteobacteria, we have used

30 16S rRNA gene sequences to define 59 operational taxonomic units (OTUs), at 97%

31 similarity. While some Zetaproteobacteria taxa appear to be cosmopolitan, various

32 habitats enrich for different sets of Zetaproteobacteria. OTU networks show that certain

33 Zetaproteobacteria co-exist, sharing compatible niches. These niches may correspond

34 with adaptations to O2, H2, and nitrate availability, based on genomic analyses. Also, a

35 putative Fe oxidation gene has been found in diverse Zetaproteobacteria taxa,

36 suggesting that the Zetaproteobacteria evolved as specialists in Fe oxidation. In all,

37 culture, genomic, and environmental studies suggest that Zetaproteobacteria are

38 widespread, and therefore have a broad influence on marine and saline terrestrial Fe

39 cycling.

40

2 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

41 Introduction

42 Fe in marine environments is a study in contrasts. It is often a limiting nutrient in

43 the open ocean, while the basaltic ocean crust and many sediments are replete in Fe.

44 This stark difference is due to the redox chemistry of Fe, which is present as Fe(II) in

45 basalt and anoxic groundwater, but rapidly oxidizes to Fe(III) in oxic ocean water,

46 precipitating as Fe(III) minerals. This oxidation was assumed to be dominated by rapid

47 abiotic oxidation at circumneutral pH, but the discovery of the marine Fe-oxidizing

48 Zetaproteobacteria gave proof that the process can be driven by microbes. First

49 proposed as a class in 2007 (1), Zetaproteobacteria have since been widely observed in

50 deep sea and coastal environments. All isolates couple Fe oxidation to oxygen

51 respiration, producing highly reactive Fe(III) oxyhydroxides that can adsorb or

52 coprecipitate nutrients and metals. However, despite the biogeochemical importance of

53 marine Fe oxidation, we are just beginning to learn how Fe-oxidizers function and

54 influence marine ecosystems. With a recent surge of culturing and sequencing, there is

55 now a significant set of data from which we can glean broader insights into microbial

56 marine Fe oxidation.

57 The goal of this paper is to review our current knowledge of marine Fe-oxidizers

58 through the lens of this increasingly well-established class of Fe-oxidizing bacteria

59 (FeOB). We begin by describing the discovery of this novel class at an Fe-rich

60 hydrothermal system at the Loihi Seamount in Hawaii. We lay out the evidence for

61 microbially-driven Fe oxidation in this marine system, including new kinetics results from

62 experiments with the Loihi isolate and model Zetaproteobacteria Mariprofundus

63 ferrooxydans PV-1. Work at Loihi has inspired numerous studies of Zetaproteobacteria

3 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

64 isolates, biominerals, and environmental distribution. In addition to reviewing these, we

65 present a comprehensive reanalysis of Zetaproteobacteria diversity and ecology,

66 enabled by the newly developed ZetaHunter classification program (2). We then use

67 current genomic evidence to evaluate whether all members of this class have the

68 potential to oxidize Fe and discuss inferred Zetaproteobacteria niches. Finally, we

69 discuss our perspectives on open questions in Zetaproteobacteria evolution, ecology,

70 and impacts on geochemical cycling. This article was submitted to an online preprint

71 archive.

72

73 Zetaproteobacteria: a novel class of marine Fe-oxidizing bacteria

74 The discovery of Zetaproteobacteria is a story that began decades before the

75 class was proposed. The unusual morphology of biogenic Fe oxides have long been

76 used to recognize microbial Fe oxidation in terrestrial environments. The twisted stalks

77 of Gallionella and hollow sheaths containing cells of Leptothrix were described in

78 terrestrial Fe-rich environments as early as the mid-1800s (3, 4). However,

79 Winogradsky (5) was the first to confirm that Leptothrix required Fe(II) for growth, thus

80 linking microbial activity with iron deposition in terrestrial environments (6). In the 1980s,

81 similar structures were found in Fe-rich marine environments, including the Red

82 Seamount of the East Pacific Rise, the Explorer Ridge, and Loihi Seamount (Figure S1)

83 (7–11). This led to the assumption that these structures were made by Gallionella and

84 Leptothrix, though these organisms were not detected in subsequent studies of marine

85 Fe mats based on small subunit ribosomal RNA (16S) marker gene surveys (e.g. 12–

86 14). Instead, Moyer et al. (12) discovered the first sequence of the novel

4 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

87 Zetaproteobacteria class, though it wasn’t recognized at the time because there were

88 no isolates or other closely related sequences. The first isolates, Mariprofundus

89 ferrooxydans strains PV-1 and JV-1, were obtained from samples collected at Loihi

90 Seamount near Hawaii in 1996-98 (1, 15). Additional surveys from Fe-rich environments

91 provided related 16S gene sequences, which helped establish the Zetaproteobacteria

92 as a monophyletic group within the Proteobacteria (1, 13, 16–18). The association of

93 the Zetaproteobacteria and Fe-rich marine environments has been strengthened since

94 these initial observations, with continued discovery of Zetaproteobacteria within Fe-rich

95 saline environments.

96

97 Zetaproteobacteria isolates: model systems for microbial Fe oxidation

98 The difficulty of culturing FeOB has been one of the main challenges in

99 demonstrating microbial marine Fe oxidation. The first Zetaproteobacteria isolates were

100 obtained using liquid and agarose-stabilized gradient tubes and plates designed to

101 provide both Fe(II) and O2 in opposing gradients (15, 19). With this setup, Fe(II) is

102 gradually released by dissolution of solid reduced Fe minerals (e.g. Fe(0) or FeS) at the

103 bottom of the tube or plate while O2 diffuses from the headspace above. These culturing

104 techniques make it difficult to control O2 and Fe(II) concentrations. To date,

105 Zetaproteobacteria have not been culturable on solid media, so isolation requires serial

106 dilution to extinction, with transfers every ~2-3 days due to increasing autocatalytic Fe

107 oxidation over time (20). In all, these challenges likely account for why so few

108 Zetaproteobacteria have been isolated.

5 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

109 Despite these hurdles, Zetaproteobacteria representatives from two genera and

110 eight separate OTUs have been successfully isolated (Table 1). These include seven

111 isolates from Fe-rich hydrothermal Fe mats, and eight from coastal environments.

112 Mariprofundus ferrooxydans PV-1 is the type strain of the most frequently isolated

113 genus, and is an obligate neutrophilic autotrophic Fe-oxidizer. All but two other isolates

114 are similarly obligate Fe-oxidizers. These two, Ghiorsea bivora TAG-1 and SV-108, are

115 facultative Fe-oxidizers that are also capable of growth by H2 oxidation (21). Except for

116 this instance, isolates vary primarily in their physiological preferences (Table 2), which

117 are related to characteristics of their source environments.

118 Zetaproteobacteria isolates are generally microaerophiles originating from oxic-

119 anoxic transition zones, where O2 concentrations are low, i.e. micromolar to

120 submicromolar. Abiotic Fe oxidation is slow at these low O2 concentrations (22, 23),

121 which allows the Zetaproteobacteria to compete. In terrestrial environments, kinetics

122 experiments near 25˚C suggest that biotic Fe oxidation is a significant component of

123 total Fe oxidation below 50 µM and can outcompete abiotic Fe oxidation at 15 µM O2

124 (24). However, there is no kinetics data from marine FeOB. To understand the

125 conditions where marine biotic Fe oxidation is competitive, we measured Fe oxidation

126 kinetics using M. ferrooxydans PV-1 as a model (methods below). With this experiment,

127 we have shown that PV-1 outpaces abiotic oxidation below 49 µM O2, and accounts for

128 up to 99% of the Fe oxidation at 10 µM O2 (Figure 1; Table 3). In cultures of M.

129 aestuarium CP-5 and M. ferrinatatus CP-8, oxygen concentrations ranged from 0.07-2.0

130 µM O2 within the cell growth band (25). This range of O2 growth conditions is well below

131 the level at which almost all Fe oxidation was biotic for PV-1, suggesting that many

6 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

132 Zetaproteobacteria are well adapted to compete and thrive under micromolar and

133 submicromolar O2 concentrations. Such low O2 concentrations are common within the

134 oxic-anoxic transition zones where the Zetaproteobacteria are found (e.g. 26, 27).

135

136 Biomineral morphologies: form follows function

137 M. ferrooxydans PV-1 has been a model system for biomineralization by an

138 obligate Fe-oxidizer. PV-1 cells form a twisted stalk (Figure 2A-C), so similar to the one

139 formed by the terrestrial Fe-oxidizer Gallionella ferruginea that it could be mistaken for a

140 Gallionella stalk (15). The stalk consists of individual filaments made of nanoparticulate

141 Fe oxyhydroxides and acidic polysaccharides, controlling Fe mineral growth near the

142 cell surface (28). Stalk growth was measured to be 2.2 µm h-1, or nearly 5x the width of

143 a PV-1 cell per hour (28). The combination of this directed mineralization and a near-

144 neutral cell surface charge explains how the cell remains remarkably free of

145 encrustation (29). These encrustation avoidance mechanisms are important for any Fe-

146 oxidizing microbe to avoid cell death by Fe mineral growth inside and outside the cell.

147 While most Zetaproteobacteria isolates form a stalk, some make other biomineral

148 morphologies (Table 1; Figure 3A). Mariprofundus ferrinatatus CP-8 and M. aestuarium

149 CP-5 form shorter filaments that resemble the dreadlock hairstyle (Figure 3C) (25).

150 Dreads were originally observed in terrestrial FeOB Gallionellaceae Ferriphaselus spp.,

151 which makes both stalk and dreads (Figure 3B) (30). In both Ferriphaselus and the CP

152 strains, the dreads are shed from cells. This suggests that dreads and similar structures

153 are used specifically to avoid encrustation, whereas the stalk has other functions. PV-1

154 cells use the stalk as a holdfast to anchor the cell to surfaces (31). As the cell oxidizes

7 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

155 Fe and produces new stalk the cell moves forward, leaving stalk behind. Since the stalk

156 is rigid and anchored, this is a means of motility. Experiments in controlled Fe(II) and O2

157 gradients showed that PV-1 cells use their stalks to position themselves at an optimum

158 position within that gradient, often forming filaments oriented toward higher O2 (31).

159 In the environment, such oriented filaments are common. At Loihi Seamount,

160 curd-type mats (cohesive Fe mats with a bumpy surface reminiscent of cheese curds)

161 often form directly above a vent orifice (Figure 4A) (27). Micrographs of intact curd mats

162 showed centimeters-long, highly directional twisted stalks forming the mat architecture

163 (Figure 2A-C). These stalks record the synchronous movements of a community of cells

164 all growing and twisting in the same direction, as well as shifts in directionality in

165 response to changes in the environment (27). The mechanism by which these cells

166 actively control their directionality through stalk production is currently unknown.

167 Beyond stalks, Loihi Seamount also hosts sheath-rich veil-type mats, which form

168 millimeters-thick Fe mat draped over rock or older Fe mat in diffuse venting

169 environments (Figure 4B). These mats are created by organisms that form hollow Fe

170 oxide sheaths (Figure 2D-F), similar to those produced by the terrestrial

171 Betaproteobacteria Leptothrix. In the marine environment, however, these sheaths are

172 formed by Zetaproteobacteria, informally called zetathrix (32). From studies based on

173 the terrestrial Leptothrix, sheaths function similarly to stalks, with tens of cells producing

174 the sheath and leaving it behind as the cells move forward (27). In Loihi Seamount

175 intact veil mats, sheaths also leave a record of highly directional growth, despite oxygen

176 profiles of these mats showing a shallow O2 gradient with O2 present throughout the mat

177 (27). In both curd- and veil-type mats, Zetaproteobacteria work together to form a highly

8 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

178 porous mat almost completely composed of biomineral filaments formed by cells,

179 making these Fe mats different from other commonly studied mats or biofilms, which

180 feature cells embedded in an exopolysaccharide matrix. The biomineral filaments

181 forming the structure of the mat also frequently have Fe biominerals attached to them,

182 suggesting FeOB also colonize the mat interior (Figure S1). Short branching hollow

183 tubes, referred to as “y-guys”, are formed by the Zetaproteobacteria (Figure 2G,H) (33).

184 However, the organisms forming other Fe biominerals have yet to be identified,

185 including nest-like structures reminiscent of freshwater Siderocapsa-like organisms

186 (Figure 2I) (34). The range of biomineral morphologies is related to differing biomineral

187 functions, which likely correspond to different niches within the Fe mat habitat.

188

189 Habitats of the Zetaproteobacteria

190

191 Loihi Seamount hydrothermal vents: a Zetaproteobacteria observatory. Most of

192 what we know about Zetaproteobacteria is based on work at the Loihi Seamount (also

193 spelled Lō’ihi Seamount), a Hawaiian submarine volcano, from long-term studies

194 including the Iron Microbial Observatory (FeMO). Loihi Seamount is an ideal habitat for

195 Zetaproteobacteria, with hydrothermal fluids rich in CO2 and Fe(II), and low in sulfide

196 (<50 µM in vent fluids; undetectable in Fe mats) (10, 35, 36). Background seawater

197 oxygen concentrations are ~50 µM at the summit of Loihi Seamount, due to its location

198 within the oxygen minimum zone (36). At the base of Loihi Seamount, the Ula Nui site

199 has higher ambient O2 concentrations (145 µM), but lower venting temperatures (1.7˚C

200 average compared to ~42˚C average at the summit) (37). Low temperatures and low

9 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

201 ambient O2 concentrations favor biotic Fe oxidation by reducing the abiotic rate at and

202 below the mat surface (23, 38). Thus, the conditions at Loihi Seamount have favored

203 the growth of microbial Fe mats ranging from centimeters to meters thick and up to 15

204 km2 (Figure 4A-C) (27, 37). The extensive Fe mats at Loihi Seamount may reflect

205 years- to decades-long stable Fe mat production by the Zetaproteobacteria, based on

206 productivity estimates (27, 33).

207 Loihi Seamount studies have provided the cornerstones of Zetaproteobacteria

208 ecology. Since the discovery of Zetaproteobacteria in the 1990s (12, 15), five research

209 expeditions from 2006-2013 have focused on Zetaproteobacteria succession, niche and

210 species diversity, and genetic potential. Colonization experiments over 4-10 days

211 showed that Zetaproteobacteria prefer low- to mid-temperature (from 22-60˚C, avg

212 40˚C) Loihi hydrothermal vents (14). This preference was reflected in longer term

213 observations following the 1996 Loihi eruption, which showed Zetaproteobacteria

214 increasing in abundance as high temperature vents cooled to pre-eruption temperatures

215 and transitioned from sulfide-rich to Fe-rich fluids (36, 39–41). The bulk of the omic

216 information on Zetaproteobacteria originates from Loihi Seamount, with the first isolate

217 genome, single cell genomes, metagenome, and proteome all from Loihi sources (42–

218 45).

219

220 Other hydrothermally-influenced habitats. Beyond Loihi, Zetaproteobacteria are

221 hosted by many other hydrothermal systems. Extensive Fe mats form around vents at

222 seamounts and island arc systems (Figure 4D) (41, 46–48). However, Fe mats have

223 also been found at spreading ridge systems, within diffuse flow at the periphery of high-

10 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

224 temperature chimneys and vents (49–52). Most hydrothermal Fe mats consist of

225 biomineral morphologies similar to those at the Loihi Seamount (twisted stalks, sheaths,

226 y-guys, etc.) (50, 51). However, mat textures and lithification can vary as a function of

227 geochemistry (e.g. Mn, Si concentration) and rates of hydrothermal discharge (53, 54).

228 Zetaproteobacteria are also found in the marine subsurface. There, oxygenated

229 seawater can mix with anoxic Fe-rich fluids, providing a favorable environment for Fe

230 oxidation. Zetaproteobacteria have been observed up to 332 meters below the sea

231 floor, within both hydrothermal recharge and cold oxic circulation cells (Figure 4E) (55–

232 57). In many near surface sediments, shallow mixing introduces O2 into an Fe-rich

233 environment, leading to abundant Zetaproteobacteria populations (13, 18, 58). As

234 hydrothermal systems age and cool, basalts and the minerals within inactive sulfide

235 mounds can also serve as Fe(II) sources for Zetaproteobacteria (59–62). These studies

236 show that the Zetaproteobacteria are abundant members of shallow and deep marine

237 subsurface environments, one of the largest underexplored habitats in the oceans.

238

239 Coastal and terrestrial habitats. Zetaproteobacteria have only recently been

240 discovered in coastal and terrestrial environments. Colonization experiments showed

241 that Zetaproteobacteria biofilms grow on Fe(II) released from mild and carbon steel that

242 is commonly used in ships and docks, suggesting that these FeOB contribute to

243 corrosion (Figure 4H) (63–67). [For a review on the role of FeOB in biocorrosion, see

244 Emerson (D. Emerson, submitted for publication; 67).] Fe(II) can also come from natural

245 sources in coastal environments, originating from mineral weathering and transported in

246 anoxic groundwater. Fe redox cycling at the oxic-anoxic transition zone of stratified

11 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

247 estuaries can support the growth of Zetaproteobacteria, as evidenced by the isolation of

248 M. ferrinatatus CP-8 and M. aestuarium CP-5 (25, 26). In near shore sediments, Fe-rich

249 groundwater can support microbial communities with Zetaproteobacteria at the

250 sediment surface (69–71). Also in these sediments, bioturbation from plant roots and

251 animal burrows provides conduits of O2 to this Fe-rich groundwater. Biotic and abiotic

252 Fe oxidation in these environments leads to the formation of Fe oxides, which coat

253 sands, salt grass and mangrove roots, and burrows (Figure 4G) (64, 72–74). Beam et

254 al. (74) found the abundance of Zetaproteobacteria within Fe oxide-coated worm

255 burrows to be an order of magnitude higher than surrounding bulk sediment, suggesting

256 that Zetaproteobacteria growth and biotic Fe oxidation can be favored in these

257 bioturbated sediments. The Fe oxides produced in these environments can sequester

258 toxins that adsorb to the mineral surface (75). Thus, FeOB activity could affect coastal

259 water quality.

260 Zetaproteobacteria have generally been considered marine FeOB, detected at

261 salinities up to 112 ppt in hypersaline brines (16, 76). Their occurrence in coastal

262 environments provides the opportunity to delineate their minimum salinity requirements.

263 McBeth et al. (77) surveyed Fe mats along the tidal Sheepscot River, Maine, as it

264 entered the estuary, finding that Zetaproteobacteria only appeared in environments with

265 5 ppt salinity or higher. This explains why Zetaproteobacteria are not commonly found

266 nor expected to be found in most terrestrial environments. This is further supported by

267 deep sequencing of Fe-rich samples: our analyses of SRA data (Dataset S1) and Scott

268 et al. (51) also determined that Zetaproteobacteria appear to be restricted to marine

269 environments.

12 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

270 Thus, it was surprising and novel to find abundant populations of

271 Zetaproteobacteria in CO2-rich terrestrial springs. Surveys of the 16S genes from

272 carbonic springs at Tierra Amarilla Spring, New Mexico (~9 ppt salinity) revealed a

273 microbial population up to one third Zetaproteobacteria (78). Similarly, 16S and genomic

274 work in CO2-rich Crystal Geyser, Utah, (~11-14 ppt salinity; Figure 4F) found the

275 Zetaproteobacteria to be both abundant and consistently present over a year of

276 observation (79–81). These springs represent the first habitat with abundant populations

277 of both Zetaproteobacteria and Betaproteobacteria FeOB (Gallionellaceae), whose

278 abundance is likely driven by cycles of freshwater and saline subsurface groundwater

279 mixing (81). The work at Crystal Geyser has produced full-length 16S sequences and

280 the only terrestrial Zetaproteobacteria genomes (79–81). Our analysis of 16S

281 phylogenetic placement and genomic clustering (by average nucleotide identity)

282 suggests that Zetaproteobacteria populations in terrestrial subsurface environments are

283 primarily novel and deeply branching Zetaproteobacteria (see further discussion of

284 habitat selection and niche below).

285

286 Common habitat characteristics. In all, Zetaproteobacteria have been found in

287 habitats sharing the following characteristics: 1) brackish to hypersaline water, 2) a

288 supply of Fe(II), and 3) predominantly micro-oxic conditions. These conditions are

289 widespread and found in diverse habitats, likely supplying multiple niches for the

290 diversification and evolution of the Zetaproteobacteria.

291

292 Zetaproteobacteria diversity

13 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

293 Zetaproteobacteria diversity has been defined using 16S gene

294 Zetaproteobacteria operational taxonomic units (ZOTU; 97% similarity), based on

295 sequences from isolates and environmental samples (Table 1; Dataset S1). Since their

296 initial description, the Zetaproteobacteria class has remained a robust taxonomic group

297 within the Proteobacteria (82, 83). A systematic analysis of 227 Zetaproteobacteria full-

298 length 16S gene sequences yielded 59 ZOTUs (2), an increase from 28 ZOTUs in 2011

299 (84). The majority of these ZOTUs are contained within two families of the

300 Zetaproteobacteria, based on sequence similarity (Figure 5A, Figure 6). Figure 5 shows

301 key ZOTUs, which are frequently sampled and abundant in the environment and are

302 primarily distinct monophyletic taxonomic groups by 16S (see detail in Figure S2).

303 These 15 ZOTUs represent 83% of sequences found in the environment. ZOTUs 1 and

304 14 are the one exception to monophyly by 16S, yet do form distinct monophyletic

305 groups in a concatenated tree of 12 ribosomal proteins (Figure 6; phylogenetic methods

306 in Supplemental Methods). ZOTUs that include isolates represent only 20% of

307 environmental sequences (Table 1), showing that the Zetaproteobacteria are largely

308 uncultivated.

309 In order to compare Zetaproteobacteria diversity across studies, we used

310 ZetaHunter, a classification pipeline designed to rapidly and reproducibly classify

311 ZOTUs (2). We classified publicly available Zetaproteobacteria full- and partial-length

312 16S gene sequences from SILVA, IMG, and NCBI SRA (total of 1.2 million sequences

313 from 93 studies; summary of samples in Dataset S1; methods in Supplemental

314 Methods) (85–88). This work provided the basis for the habitat and diversity analysis

315 below, while also allowing us to correct previous ZOTU assignments (Table S1).

14 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

316

317 Connecting Zetaproteobacteria diversity, habitat, and niche. Zetaproteobacteria

318 diversity is driven primarily by the variety of niches they inhabit. A niche is the set of

319 conditions favorable for growth, which are further influenced, or partitioned, by inter- and

320 intra-species population dynamics in the environment (89). A challenge in microbial

321 ecology is to tease apart the niche of an organism through sampling at the appropriate

322 resolution. For most marine environments, we lack the highly resolved chemical and

323 spatial information to describe niches. However, we can look for patterns of

324 associations between different Zetaproteobacteria and their habitats to understand

325 where and the extent to which Zetaproteobacteria niches may overlap.

326 Each habitat displayed distinct and abundant ZOTUs, indicating that habitats can

327 host a specific set of niches that support these ZOTUs (Table 4; Dataset S1). In

328 particular, dominant ZOTUs differ between habitat types, suggesting that each habitat

329 has a set of dominant niches that favor the growth of those particular ZOTUs. ZOTU54

330 is a striking example of a dominant ZOTU clearly successful within the terrestrial

331 subsurface fluid environment. ZOTU54 is a deep-branching ZOTU that is primarily

332 limited to this environment (78–81). The distribution of ZOTU54 suggests that it is

333 endemic or adapted to thrive within terrestrial carbonic Fe-rich springs. However, while

334 it is frequently found at high abundance in terrestrial springs, ZOTU54 is also found in

335 other habitats at very low abundance, including hydrothermal Fe mats (Dataset S1). In

336 fact, many ZOTUs span habitats (Figure 5B; Dataset S1), suggesting that similar niches

337 supporting these ZOTUs can exist in multiple habitats.

15 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

338 Next, we looked for patterns in ZOTU associations with each other, mapping

339 connections in a ZOTU network (Figure 7). This network shows which ZOTUs are found

340 in isolation and which co-occur in the same sample, with connection thickness showing

341 the strength of those connections. Further, the network layout is based on the frequency

342 of co-occurrence, so when two ZOTUs commonly co-occur, they are closer together

343 and connected by thicker lines. Most ZOTUs co-occur with others (Figure 7), and these

344 connections are not random. Some ZOTUs co-occur more frequently, forming clusters

345 of interconnected ZOTU nodes (Clusters 1-3, Figure 7). The most abundantly sampled

346 ZOTUs form a central cluster (Cluster 2), sharing a common set of niches most

347 frequently sampled in the hydrothermal Fe mat environment. Cluster 3 centers around

348 ZOTUs found together in hydrothermal Fe mat samples from the Mariana Arc, but which

349 are not common in other environments. Cluster 1 is dominated by ZOTUs from samples

350 associated with metal corrosion and mineral weathering, which suggests that these

351 habitats host niches distinct from those in hydrothermal Fe mat habitats. Overall, these

352 clusters highlight ZOTUs with niches that frequently overlap, suggesting those niches,

353 and thus the growth requirements of these ZOTUs, are compatible.

354 A combination of ZOTU environmental distribution, habitat characteristics, and

355 isolate physiology can help us understand niche. Here, we use this approach to

356 describe ZOTU9, as an example. This ZOTU is a key player in marine subsurface fluids,

357 metal corrosion, and mineral weathering habitats (see Cluster 1, Figure 7; Table 4). In

358 mineral weathering habitats, ZOTU9 is frequently the only ZOTU (visualized as a

359 colored circle around the ZOTU node; Figure 7). For example, ZOTU9 was the only

360 Zetaproteobacteria detected within a basaltic glass weathering enrichment, making up

16 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

361 39% of the bacterial community by 16S sequencing estimates (61). These habitat

362 associations suggest that growth of ZOTU9 is favored when metal corrosion or mineral

363 weathering is a source for Fe(II). This association likely relates to these sources

364 producing H2 as well as Fe(II). Metals composed of zero valent Fe produce H2 as a

365 byproduct of anaerobic corrosion (90). Fe minerals can be a source of H2 because they

366 catalyze hydrolysis and induce radiolysis of water (91, 92). The production of H2 is

367 known to benefit some members of ZOTU9, including the Fe- and H2-oxidizing isolates

368 Ghiorsea bivora TAG-1 and SV-108 (21). The genetic machinery required for H2

369 oxidation has also been found in two ZOTU9 single amplified genomes (43, 51), which

370 suggests that H2 oxidation may be common in other members of ZOTU9. From these

371 observations, we conclude that both Fe(II) and H2 play a central role in the niche of

372 ZOTU9 and the habitats where it can be found. By combining isolate physiology with

373 habitat distribution patterns, we can identify key features of a ZOTU’s niche.

374

375 Spatial and taxonomic resolution in Zetaproteobacteria ecology. Hydrothermal Fe

376 mats have opposing gradients of Fe(II) and O2 and a complex internal structure (e.g.

377 Figure S1) (27, 36). This heterogeneity leads to multiple niches at small spatial scales,

378 suggesting that high-resolution sampling could help us better understand ZOTU niches

379 in this habitat. Initial bulk techniques for sampling collected liters of mat material, and a

380 single sample could contain all major ZOTUs (84). Therefore, new collection devices

381 were engineered to sample small volumes (50-75 mL) at centimeter spatial resolution

382 (50). From these discretely sampled Fe mats, we increased the resolution of our ZOTU

383 network (Figure S3). Of the 29 ZOTUs found within Fe mat habitats, 17 showed a

17 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

384 preference for a specific Fe mat type. However, the high connectivity of abundant

385 ZOTUs remained, even when considering more highly resolved sampling. This result

386 suggests that these ZOTUs share compatible niches at the centimeter scale in Fe mat

387 habitats.

388 Taxonomic resolution can also affect our understanding of Zetaproteobacteria

389 ecology through the lumping or splitting of ecologically-distinct groups. The ZOTU

390 classification may in certain cases be too coarse, representing multiple related

391 populations that have different niches. For example, Scott et al. (93) found multiple

392 oligotypes (ecological units defined by informative sequence variability) within each

393 ZOTU. While multiple oligotypes do not necessarily suggest each has a distinct niche,

394 for ZOTU6, only one oligotype differed in abundance over a transect approaching the

395 hydrothermal vent orifice. This abundance change suggested that a subpopulation of

396 ZOTU6 prefers higher flow conditions, warmer temperatures, and/or the differing

397 geochemistry found near the vent (93). Results like this warrant a more resolved

398 Zetaproteobacteria taxonomy, which could be aided by whole genome comparisons.

399

400 Using genomics to understand metabolic potential and niche

401

402 Carbon fixation

403 All Zetaproteobacteria isolates are obligate autotrophs, using the Calvin-Benson-

404 Bassham (CBB) cycle to fix carbon. Similarly, all ZOTUs sampled to date have the

405 ribulose-1,5-bisphosphate carboxylase oxygenase gene (RuBisCO; key enzyme in the

406 CBB cycle), suggesting carbon fixation by this pathway is a shared capability across the

18 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

407 class. The isolates of Mariprofundus ferrooxydans, strains PV-1, JV-1, and M34, all

408 encode the genes for both Form I (O2-insensitive) and Form II (O2-sensitive) RuBisCO

409 (42, 94). Similarly, both forms are encoded by Mariprofundus micogutta DIS-1, which

410 was specifically isolated to be more aerotolerant (67). However, most

411 Zetaproteobacteria SAG and MAG genomes only encode Form II RuBisCO (43, 80, 95),

412 suggesting most Zetaproteobacteria are specifically adapted to lower O2 concentrations.

413

414 Energy metabolism: Are all Zetaproteobacteria Fe-oxidizers?

415 The Zetaproteobacteria are often associated with high Fe environments, and all

416 isolates of the Zetaproteobacteria are capable of Fe oxidation. These observations have

417 led to the current assumption that all Zetaproteobacteria are capable of Fe oxidation. To

418 test this assumption, we first have to understand the mechanism of Fe oxidation in the

419 marine environment.

420 Initial genome analysis of PV-1 led to the proposal of the alternative complex III

421 (ACIII) as part of an iron oxidase complex (42). Follow up studies later changed this

422 model, suggesting ACIII was involved in reverse electron transport (30, 44, 45).

423 However, Field et al. (43) and Chiu et al. (25) isolated Zetaproteobacteria isolates that

424 lacked ACIII but were still capable of Fe oxidation. Furthermore, only 2 of 23

425 Zetaproteobacteria SAGs have the ACIII gene, and these 23 SAGs represent the

426 majority of Zetaproteobacteria diversity (43). Combined, this evidence showed that ACIII

427 is not a critical component of the Fe oxidation pathway.

428 The putative Fe oxidase, Cyc2, and another cytochrome Cyc1 were first

429 identified in Zetaproteobacteria by Barco et al. (45) through a proteome analysis of PV-

19 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

430 1. They were initially identified through comparison of the proteome with the closely

431 related M ferrooxydans M34 genome. Their presence in the proteome suggested that

432 the cyc1 and cyc2 genes were missing from the PV-1 draft genome due to gaps in the

433 assembly, which was confirmed by resequencing (45). The Cyc2 protein from PV-1 is a

434 homolog of the biochemically-characterized Cyc2 Fe oxidase from Acidithiobacillus

435 ferrooxidans (22% amino acid identity) (96). Based on this, the Fe oxidation pathway

436 model for the Zetaproteobacteria was revised (Figure 8). Cyc2 homologs have been

437 found in other Zetaproteobacteria and other neutrophilic FeOB, strengthening the

438 proposed pathway (97, 98). In fact, every single ZOTU that has a genomic

439 representative, including ZOTUs without isolates, has a homolog of this putative Fe

440 oxidation gene, consistent with the notion that all Zetaproteobacteria are Fe-oxidizers

441 (43, 95).

442

443 Genomic clues to niche based on O2 and nitrogen

444 Genomic evidence suggests that adaptation to differing O2 conditions plays a

445 role in ZOTU niches. Three terminal oxidases have been found within

446 Zetaproteobacteria genomes: cbb3- and aa3-type cytochrome c oxidases and the

447 cytochrome bd-I ubiquinol oxidase. These have different affinities for oxygen, which

448 would influence the niche of each ZOTU; Km’s of 230-300 nM (cbb3, bd-I) to 4.3 µM

449 (aa3) are reported (99, 100). The cbb3-type terminal oxidase gene is found in most of

450 the Zetaproteobacteria, sometimes in multiple copies, suggesting a predominant

451 preference for very low O2 concentrations (submicromolar) (43). However, the complete

452 genomes of M. aestuarium and M. ferrinatatus contain only the higher-O2 adapted aa3-

20 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

453 type terminal oxidase gene, which helps explain their adaptation to frequently higher O2

454 concentrations of their tidally-mixed water column habitat (25). Many Zetaproteobacteria

455 genomes have multiple terminal oxidases, suggesting they are adapted to fluctuating

456 oxygen conditions (43, 95). ZOTU10 and the isolate M. micogutta DIS-1 may have a

457 higher tolerance for such conditions with increased numbers of genes for O2 radical

458 protection (43, 67).

459 The genetic potential for nitrogen transformations differentiates marine and

460 terrestrial Zetaproteobacteria. In the marine environment, most ZOTUs are capable of

461 assimilatory nitrate reduction to ammonium (nasA, nirBD) (43, 95). In terrestrial Fe-rich

462 springs such as Crystal Geyser, Zetaproteobacteria genomes lack these genes, but

463 many possess nitrogen fixation genes (e.g. nifH) (79, 80). In contrast, only three marine

464 isolates (Mariprofundus strains DIS-1, EKF-M39, and M34) and one MAG outside of

465 Mariprofundus possesses nif genes (43, 67, 95), and, as yet, it has not been

466 experimentally shown that these isolates fix N2. Supporting these genomic

467 observations, the nifH gene is rarely detected in marine Fe mats (101). From these

468 patterns, the differences between these Zetaproteobacteria likely correspond with

469 differences in nitrate and ammonium availability in these habitats; nitrate is below

470 detection at Crystal Geyser compared to concentrations up to 32 µM within Loihi Fe

471 mats (79, 102). Nitrogen transformations and O2 tolerance obviously play a role in many

472 Zetaproteobacteria niches, but there are likely other conditions driving niche diversity

473 yet to be discovered.

474

475 Outstanding questions and opportunities

21 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

476 Over the last two decades, Zetaproteobacteria have been established as a

477 diverse, taxonomically-robust class, which thrive in a wide range of Fe-rich habitats.

478 Environmental studies, isolate experiments, and genomic analyses have given insight

479 into how they use biomineralization and metabolic strategies to succeed. Building on

480 this work, we are poised to address a number of intriguing questions.

481 How did the Zetaproteobacteria come to specialize in Fe oxidation? Thus

482 far, genomic evidence suggests that all Zetaproteobacteria are Fe-oxidizers. If this is

483 true, the Zetaproteobacteria would be an interesting model system in which to explore

484 the selection and evolution of a particular metabolic specialty. The answer to this

485 question likely rests on the complex challenges of Fe oxidation at circumneutral pH.

486 Zetaproteobacteria must position themselves at specific environmental interfaces to

487 gain energy from Fe oxidation. Meanwhile, they must compete with or tolerate abiotic

488 reactions of Fe(II) with O2 and nitrogen compounds, which can form O2 radicals and

489 toxic nitric and nitrous oxides (103, 104). They produce intricate biomineral structures,

490 which allow them to avoid encrustation, control motility, and construct mats. Thus,

491 microbial Fe oxidation appears to be a complex physiological trait, which is much more

492 likely to be inherited vertically through descent rather than transmitted horizontally

493 (105). Since Fe oxidation is a complex trait, the Zetaproteobacteria probably acquired

494 the Fe oxidation capability before the group diverged. It is unclear where the Fe

495 oxidation trait originated, but as we determine the genetic basis for the trait and

496 adaptations to its challenges, comparative genomics of various FeOB will allow us to

497 understand these evolutionary relationships.

22 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

498 What are the drivers of Zetaproteobacteria diversification? The

499 Zetaproteobacteria have diversified into at least 59 operational taxonomic units, which

500 we can now track using ZetaHunter (2). Given the increasing number of available

501 genomes, the next logical step is to develop a systematic taxonomy based on both 16S

502 gene and phylogenomics analysis. Ultimately, diversification is driven by the range in

503 environmental niches. We will improve our understanding as we continue to study

504 environmental distribution, physiology of new isolates, and genomes, especially as we

505 focus our explorations beyond the well-studied hydrothermal vents. We may be able to

506 define niches better via discrete sampling, though there are practical lower limits to

507 sample size and spatial resolution. Although intact samples are challenging to obtain,

508 the effort is worthwhile in order to use imaging-based techniques (e.g. FISH), coupled to

509 chemistry and activity measurements (e.g. elemental mapping, SIP) to discern

510 millimeter- and micron-scale associations. We are just beginning to discover the variety

511 of adaptations across genomes. As genome analyses progress, patterns of functional

512 genes and phylogeny will elucidate the drivers of Zetaproteobacteria diversification. In

513 turn, genomic clues can help us culture novel organisms, which will be key to

514 demonstrating particular roles. The integrated results of these studies will show how

515 these organisms have evolved to occupy particular niches, and how they could work

516 together to influence the geochemistry of Fe-rich habitats.

517 How do Zetaproteobacteria affect geochemical cycling, and how can we

518 track these effects? Now that we know the basics of Zetaproteobacteria metabolisms

519 and potential geochemical effects, we can move toward detecting this influence in the

520 environment and determining the controls on those effects. The key will be developing

23 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

521 ways to track Zetaproteobacteria activity, and relating this to quantitative effects. There

522 is no clear biotic Fe isotopic signature that can be used to assess the activity of

523 microbially-mediated Fe oxidation (106). An alternative is to track activity via gene

524 expression. Traditionally, this would be done via a marker gene for Fe oxidation. The

525 cyc2 gene may work if its expression proves to be specific to Fe oxidation. However,

526 now with (meta)transcriptomic approaches, we can use multiple genes (e.g. the whole

527 Fe oxidation pathway, linked with C fixation and other pathways). With the

528 Zetaproteobacteria, this will be an iterative exercise, as we are still

529 determining/validating the genes involved in Fe oxidation and other metabolisms. This

530 will be most straightforward in Zetaproteobacteria-dominated hydrothermal Fe mat

531 environments, but work in other environments will improve our understanding of the

532 range of their effects on geochemical cycling. As Zetaproteobacteria are widespread in

533 diverse environments, continued work will most likely reveal their broad influence on Fe

534 cycling in marine and saline terrestrial environments.

535

536 Essential Methods

537 M. ferrooxydans PV-1 Fe oxidation kinetics. Prior to each kinetics experiment, fresh

538 PV-1 cells were grown up overnight (approx. 19 hours) on Fe(0) plates following

539 Emerson and Floyd (19). For each 20 mL electrochemistry cell, between 1 and 13 mL of

540 cell suspension were added to fresh ASW (avg. 4.1•104 cells mL-1). Following this, the

541 electrochemistry cell was set up with O2 maintained at the desired concentrations using

542 a gas regulator (air:N2 gas mix) with continual stirring. At the start of each 1-2 hour

543 experiment, 300 µL of 10.6 mM ferrous ammonium sulfate was added to the

24 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

544 electrochemistry cell (final concentration 154 µM Fe(II)). Fe(II) was measured

545 throughout the experiment using cyclic voltammetry following Luther et al. (107). At the

546 end of each run, abiotic oxidation was measured by killing the cells with 1 mM sodium

547 azide, adding additional ferrous ammonium sulfate, and running the experiment again.

548 At the end of the biotic and abiotic runs, cells were fixed in a 1.6% paraformaldehyde

549 solution before being Syto 13-labeled and counted using a Petroff-Hausser counting

550 chamber.

551

552 ZOTU-association networks. For the ZOTU network in Figure 7, all full and partial

553 length sequences from the SILVA database (release 128) were included (85). For the

554 ZOTU network in Figure S3, only discrete Fe mat samples (50-75 mL) were included

555 (32, 43, 47, 51, 93, 95). Network connections were assigned using ZetaHunter (2),

556 which produces edge and node files as an output. ZOTU-association network

557 visualization was done in Cytoscape version 3.4.0. Figure 7 was organized spatially

558 using edge-weighted spring embedded layout, which caused ZOTUs found together in

559 multiple samples to be clustered closer together. Figure S3 was organized spatially

560 using attribute circle layout, which placed ZOTUs with a similar dominant sample type

561 close together.

562

563 Acknowledgements

564 S.M.M. and C.S.C. drafted the manuscript. All authors contributed to editing.

565 A.G., G.W.L., and C.S.C. designed and implemented M. ferrooxydans PV-1 kinetics

566 experiments. S.M.M. and R.M.M. developed ZOTU assignments and networks.

25 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

567 We would like to thank Jarrod Scott and Erin Field for their help on structuring

568 and improving this manuscript. This work was funded by NSF OCE-1155290 (C.S.C.),

569 NASA Exobiology NNX12AG20G (D.E., G.W.L, C.S.C) and NNX15AM11G (D.E.), NSF

570 MGG OCE-1558712 (G.W.L.), Office of Naval Research N00014-17-1-2640 (C.S.C)

571 and N00014-17-1-2641 (D.E.), Agriculture and Food Research Initiative grant number

572 2012-68003-30155 from the USDA National Institute of Food and Agriculture (R.M.M.),

573 and a University of Delaware Doctoral Dissertation Fellowship (S.M.M.). Computational

574 infrastructure support by the University of Delaware CBCB Core Facility funded by

575 Delaware INBRE (grant number NIH P20 GM103446) and the Delaware Biotechnology

576 Institute. The authors declare no conflict of interest.

577

578 References

579 1. Emerson D, Rentz JA, Lilburn TG, Davis RE, Aldrich H, Chan C, Moyer CL.

580 2007. A novel lineage of proteobacteria involved in formation of marine Fe-

581 oxidizing microbial mat communities. PLoS One 2:e667.

582 2. McAllister SM, Moore RM, Chan CS. 2018. ZetaHunter, a reproducible

583 taxonomic classification tool for tracking the ecology of the Zetaproteobacteria

584 and other poorly resolved taxa. Microbiol. Resour. Announc. 7:e00932-18.

585 3. Ehrenberg CG. 1838. Die infusionsthierchen als vollkommene organismen.

586 Leopold Voss, Leipzig.

587 4. Kützing FT. 1843. Phycologia generalis oder anatomie, physiologie und

588 systemkunde der tangeBrockhaus. F. A. Brockhaus, Leipzig.

589 5. Winogradsky S. 1888. Ueber Eisenbakterien. Bot. Zeitung 46:261–270.

26 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

590 6. Harder EC. 1919. Iron-deposting bacteria and their geologic relations. US Geol.

591 Surv. Prof. Pap. 113 101.

592 7. Alt JC, Lonsdale P, Haymon R, Muehlenbachs K. 1987. Hydrothermal sulfide

593 and oxide deposits on seamounts near 21°N, East Pacific Rise. Geol. Soc. Am.

594 Bull. 98:157–168.

595 8. Alt JC. 1988. Hydrothermal oxide and nontronite deposits on seamounts in the

596 Eastern Pacific. Mar. Geol. 81:227–239.

597 9. Juniper SK, Fouquet Y. 1988. Filamentous iron-silica deposits from modern and

598 ancient hydrothermal sites. Can. Mineral. 26:859–869.

599 10. Karl DM, McMurtry GM, Malahoff A, Garcia MO. 1988. Loihi Seamount, Hawaii:

600 a mid-plate volcano with a distinctive hydrothermal system. Nature 335:532–535.

601 11. Karl DM, Brittain AM, Tilbrook BD. 1989. Hydrothermal and microbial processes

602 at Loihi Seamount, a mid-plate hot-spot volcano. Deep Sea Res. Part A,

603 Oceanogr. Res. Pap. 36:1655–1673.

604 12. Moyer CL, Dobbs FC, Karl DM. 1995. Phylogenetic diversity of the bacterial

605 community from a microbial mat at an active, hydrothermal vent system, Loihi

606 Seamount, Hawaii. Appl. Environ. Microbiol. 61:1555–1562.

607 13. Davis RE, Stakes DS, Wheat CG, Moyer CL. 2009. Bacterial variability within an

608 iron-silica-manganese-rich hydrothermal mound located off-axis at the Cleft

609 segment, Juan de Fuca Ridge. Geomicrobiol. J. 26:570–580.

610 14. Rassa AC, McAllister SM, Safran SA, Moyer CL. 2009. Zeta-Proteobacteria

611 dominate the colonization and formation of microbial mats in low-temperature

612 hydrothermal vents at Loihi Seamount, Hawaii. Geomicrobiol. J. 26:623–638.

27 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

613 15. Emerson D, Moyer CL. 2002. Neutrophilic Fe-oxidizing bacteria are abundant at

614 the Loihi Seamount hydrothermal vents and play a major role in Fe oxide

615 deposition. Appl. Environ. Microbiol. 68:3085–3093.

616 16. Eder W, Jahnke LL, Schmidt M, Huber R. 2001. Microbial diversity of the brine-

617 seawater interface of the Kebrit Deep, Red Sea, studied via 16S rRNA gene

618 sequences and cultivation methods. Appl. Environ. Microbiol. 67:3077–3085.

619 17. Dhillon A, Teske A, Dillon J, Stahl DA, Sogin ML. 2003. Molecular

620 characterization of sulfate-reducing bacteria in the Guaymas Basin. Appl. Environ.

621 Microbiol. 69:2765–2772.

622 18. Kato S, Yanagawa K, Sunamura M, Takano Y, Ishibashi J, Kakegawa T,

623 Utsumi M, Yamanaka T, Toki T, Noguchi T, Kobayashi K, Moroi A, Kimura H,

624 Kawarabayasi Y, Marumo K, Urabe T, Yamagishi A. 2009. Abundance of

625 Zetaproteobacteria within crustal fluids in back-arc hydrothermal fields of the

626 Southern Mariana Trough. Environ. Microbiol. 11:3210–3222.

627 19. Emerson D, Floyd MM. 2005. Enrichment and isolation of iron-oxidizing bacteria

628 at neutral pH. Methods Enzymol. 397:112–123.

629 20. Lueder U, Druschel G, Emerson D, Kappler A, Schmidt C. 2018. Quantitative

2+ 630 analysis of O2 and Fe profiles in gradient tubes for cultivation of microaerophilic

631 Iron(II)-oxidizing bacteria. FEMS Microbiol. Ecol. 94:fix177.

632 21. Mori JF, Scott JJ, Hager KW, Moyer CL, Küsel K, Emerson D. 2017.

633 Physiological and ecological implications of an iron- or hydrogen-oxidizing

634 member of the Zetaproteobacteria, Ghiorsea bivora, gen. nov., sp. nov. ISME J.

635 11:2624–2636.

28 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

636 22. Stumm W, Lee GF. 1961. Oxygenation of ferrous iron. Ind. Eng. Chem. 53:143–

637 146.

638 23. Millero FJ, Sotolongo S, Izaguirre M. 1987. The oxidation kinetics of Fe(II) in

639 seawater. Geochim. Cosmochim. Acta 51:793–801.

640 24. Druschel GK, Emerson D, Sutka R, Suchecki P, Luther GWI. 2008. Low-

641 oxygen and chemical kinetic constraints on the geochemical niche of neutrophilic

642 iron(II) oxidizing microorganisms. Geochim. Cosmochim. Acta 72:3358–3370.

643 25. Chiu BK, Kato S, McAllister SM, Field EK, Chan CS. 2017. Novel pelagic iron-

644 oxidizing Zetaproteobacteria from the Chesapeake Bay oxic-anoxic transition

645 zone. Front. Microbiol. 8:1280.

646 26. Field EK, Kato S, Findlay AJ, MacDonald DJ, Chiu BK, Luther GW, Chan CS.

647 2016. Planktonic marine iron oxidizers drive iron mineralization under low-oxygen

648 conditions. Geobiology 14:499–508.

649 27. Chan CS, McAllister SM, Leavitt AH, Glazer BT, Krepski ST, Emerson D.

650 2016. The architecture of iron microbial mats reflects the adaptation of

651 chemolithotrophic iron oxidation in freshwater and marine environments. Front.

652 Microbiol. 7:796.

653 28. Chan CS, Fakra SC, Emerson D, Fleming EJ, Edwards KJ. 2011. Lithotrophic

654 iron-oxidizing bacteria produce organic stalks to control mineral growth:

655 implications for biosignature formation. ISME J. 5:717–727.

656 29. Saini G, Chan CS. 2012. Near-neutral surface charge and hydrophilicity prevent

657 mineral encrustation of Fe-oxidizing micro-organisms. Geobiology 11:191–200.

658 30. Kato S, Ohkuma M, Powell DH, Krepski ST, Oshima K, Hattori M, Shapiro N,

29 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

659 Woyke T, Chan CS. 2015. Comparative genomic insights into ecophysiology of

660 neutrophilic, microaerophilic iron oxidizing bacteria. Front. Microbiol. 6:1265.

661 31. Krepski ST, Emerson D, Hredzak-Showalter PL, Luther GWI, Chan CS. 2013.

662 Morphology of biogenic iron oxides records microbial physiology and

663 environmental conditions: toward interpreting iron microfossils. Geobiology

664 11:457–471.

665 32. Fleming EJ, Davis RE, McAllister SM, Chan CS, Moyer CL, Tebo BM,

666 Emerson D. 2013. Hidden in plain sight: discovery of sheath-forming, iron-

667 oxidizing Zetaproteobacteria at Loihi Seamount, Hawaii, USA. FEMS Microbiol.

668 Ecol. 85:116–127.

669 33. Emerson D, Scott JJ, Leavitt A, Fleming E, Moyer C. 2017. In situ estimates of

670 iron-oxidation and accretion rates for iron-oxidizing bacterial mats at Lō’ihi

671 Seamount. Deep. Res. Part I Oceanogr. Res. Pap. 126:31–39.

672 34. Emerson D, Fleming EJ, McBeth JM. 2010. Iron-oxidizing bacteria: an

673 environmental and genomic perspective. Annu. Rev. Microbiol. 64:561–583.

674 35. Sedwick P, McMurtry GM, Macdougall J. 1992. Chemistry of hydrothermal

675 solutions from Pele’s vents, Loihi Seamount, Hawaii. Geochim. Cosmochim. Acta

676 56:3643–3667.

677 36. Glazer BT, Rouxel OJ. 2009. Redox speciation and distribution within diverse

678 iron-dominated microbial habitats at Loihi Seamount. Geomicrobiol. J. 26:606–

679 622.

680 37. Edwards KJ, Glazer BT, Rouxel OJ, Bach W, Emerson D, Davis RE, Toner

681 BM, Chan CS, Tebo BM, Staudigel H, Moyer CL. 2011. Ultra-diffuse

30 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

682 hydrothermal venting supports Fe-oxidizing bacteria and massive umber

683 deposition at 5000 m off Hawaii. ISME J. 5:1748–1758.

684 38. Emerson D, Scott JJ, Benes J, Bowden WB. 2015. Microbial iron oxidation in

685 the Arctic tundra and its implications for biogeochemical cycling. Appl. Environ.

686 Microbiol. 81:8066–8075.

687 39. Davis RE, Moyer CL, McAllister SM, Rassa AC, Tebo BM. 2005. Spatial and

688 temporal variability of microbial communities from pre- and post-eruption microbial

689 mats collected from Loihi Seamount, Hawaii. AGU Fall Meet. Abstr. PS.01.015.

690 40. Moyer CL, Davis RE, Curtis AC, Rassa AC. 2007. Spatial and temporal

691 variability in microbial communities from pre- and post-eruption microbial mats

692 collected from Loihi Seamount, Hawaii: an update. AGU Fall Meet. Abstr. B23G–

693 02.

694 41. Emerson D, Moyer CL. 2010. Microbiology of seamounts: common patterns

695 observed in community structure. Oceanography 23:148–163.

696 42. Singer E, Emerson D, Webb EA, Barco RA, Kuenen JG, Nelson WC, Chan

697 CS, Comolli LR, Ferriera S, Johnson J, Heidelberg JF, Edwards KJ. 2011.

698 Mariprofundus ferrooxydans, PV-1 the first genome of a marine Fe(II) oxidizing

699 Zetaproteobacterium. PLoS One 6:e25386.

700 43. Field EK, Sczyrba A, Lyman AE, Harris CC, Woyke T, Stepanauskas R,

701 Emerson D. 2015. Genomic insights into the uncultivated marine

702 Zetaproteobacteria at Loihi Seamount. ISME J. 9:857–870.

703 44. Singer E, Heidelberg JF, Dhillon A, Edwards KJ. 2013. Metagenomic insights

704 into the dominant Fe(II) oxidizing Zetaproteobacteria from an iron mat at Lō’ihi,

31 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

705 Hawai’i. Front. Microbiol. 4:52.

706 45. Barco RA, Emerson D, Sylvan JB, Orcutt BN, Jacobson Meyers ME, Ramírez

707 GA, Zhong JD, Edwards KJ. 2015. New insight into microbial iron oxidation as

708 revealed by the proteomic profile of an obligate iron-oxidizing

709 chemolithoautotroph. Appl. Environ. Microbiol. 81:5927–5937.

710 46. Bortoluzzi G, Romeo T, La Cono V, La Spada G, Smedile F, Esposito V,

711 Sabatino G, Di Bella M, Canese S, Scotti G, Bo M, Giuliano L, Jones D,

712 Golyshin PN, Yakimov MM, Andaloro F. 2017. Ferrous iron- and ammonium-

713 rich diffuse vents support habitat-specific communities in a shallow hydrothermal

714 field off the Basiluzzo Islet (Aeolian Volcanic Archipelago). Geobiology 15:664–

715 677.

716 47. Hager KW, Fullerton H, Butterfield DA, Moyer CL. 2017. Community structure

717 of lithotrophically-driven hydrothermal microbial mats from the Mariana Arc and

718 Back-Arc. Front. Microbiol. 8:1578.

719 48. Kato S, Kobayashi C, Kakegawa T, Yamagishi A. 2009. Microbial communities

720 in iron-silica-rich microbial mats at deep-sea hydrothermal fields of the Southern

721 Mariana Trough. Environ. Microbiol. 11:2094–111.

722 49. Dekov VM, Petersen S, Garbe-Schönberg C-D, Kamenov GD, Perner M,

723 Kuzmann E, Schmidt M. 2010. Fe–Si-oxyhydroxide deposits at a slow-spreading

724 centre with thickened oceanic crust: The Lilliput hydrothermal field (9°33′S, Mid-

725 Atlantic Ridge). Chem. Geol. 278:186–200.

726 50. Breier JA, Gomez-Ibanez D, Reddington E, Huber JA, Emerson D. 2012. A

727 precision multi-sampler for deep-sea hydrothermal microbial mat studies. Deep

32 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

728 Sea Res. Part I Oceanogr. Res. Pap. 70:83–90.

729 51. Scott JJ, Breier JA, Luther, III GW, Emerson D. 2015. Microbial iron mats at the

730 Mid-Atlantic Ridge and evidence that Zetaproteobacteria may be restricted to iron-

731 oxidizing marine systems. PLoS One 10:e0119284.

732 52. Vander Roost J, Thorseth IH, Dahle H. 2017. Microbial analysis of

733 Zetaproteobacteria and co-colonizers of iron mats in the Troll Wall Vent Field,

734 Arctic Mid-Ocean Ridge. PLoS One 12:e0185008.

735 53. Li J, Zhou H, Peng X, Wu Z, Chen S, Fang J. 2012. Microbial diversity and

736 biomineralization in low-temperature hydrothermal iron-silica-rich precipitates of

737 the Lau Basin hydrothermal field. FEMS Microbiol. Ecol. 81:205–216.

738 54. Johannessen KC, Vander Roost J, Dahle H, Dundas SH, Pedersen RB,

739 Thorseth IH. 2017. Environmental controls on biomineralization and Fe-mound

740 formation in a low-temperature hydrothermal system at the Jan Mayen Vent

741 Fields. Geochim. Cosmochim. Acta 202:101–123.

742 55. Yanagawa K, Nunoura T, McAllister SM, Hirai M, Breuker A, Brandt L, House

743 CH, Moyer CL, Birrien J-L, Aoike K, Sunamura M, Urabe T, Mottl MJ, Takai K.

744 2013. The first microbiological contamination assessment by deep-sea drilling and

745 coring by the D/V Chikyu at the Iheya North hydrothermal field in the Mid-Okinawa

746 Trough (IODP Expedition 331). Front. Microbiol. 4:327.

747 56. Meyer JL, Jaekel U, Tully BJ, Glazer BT, Wheat CG, Lin H-T, Hsieh C-C,

748 Cowen JP, Hulme SM, Girguis PR, Huber JA. 2016. A distinct and active

749 bacterial community in cold oxygenated fluids circulating beneath the western

750 flank of the Mid-Atlantic ridge. Sci. Rep. 6:22541.

33 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

751 57. Tully BJ, Wheat CG, Glazer BT, Huber JA. 2017. A dynamic microbial

752 community with high functional redundancy inhabits the cold, oxic subseafloor

753 aquifer. ISME J.

754 58. Handley KM, Boothman C, Mills RA, Pancost RD, Lloyd JR. 2010. Functional

755 diversity of bacteria in a ferruginous hydrothermal sediment. ISME J. 4:1193–

756 1205.

757 59. Sylvan JB, Toner BM, Edwards KJ. 2012. Life and death of deep-sea vents:

758 Bacterial diversity and ecosystem succession on inactive hydrothermal sulfides.

759 MBio 3:e00279-11.

760 60. Kato S, Ikehata K, Shibuya T, Urabe T, Ohkuma M, Yamagishi A. 2015.

761 Potential for biogeochemical cycling of sulfur, iron and carbon within massive

762 sulfide deposits below the seafloor. Environ. Microbiol. 17:1817–1835.

763 61. Henri PA, Rommevaux-Jestin C, Lesongeur F, Mumford A, Emerson D,

764 Godfroy A, Ménez B. 2016. Structural iron (II) of basaltic glass as an energy

765 source for Zetaproteobacteria in an abyssal plain environment, off the Mid Atlantic

766 Ridge. Front. Microbiol. 6:1518.

767 62. Barco RA, Hoffman CL, Ramírez GA, Toner BM, Edwards KJ, Sylvan JB.

768 2017. In-situ incubation of iron-sulfur mineral reveals a diverse

769 chemolithoautotrophic community and a new biogeochemical role for

770 Thiomicrospira. Environ. Microbiol. 19:1322–1337.

771 63. Dang H, Chen R, Wang L, Shao S, Dai L, Ye Y, Guo L, Huang G, Klotz MG.

772 2011. Molecular characterization of putative biocorroding microbiota with a novel

773 niche detection of Epsilon- and Zetaproteobacteria in Pacific Ocean coastal

34 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

774 seawaters. Environ. Microbiol. 13:3059–3074.

775 64. McBeth JM, Little BJ, Ray RI, Farrar KM, Emerson D. 2011. Neutrophilic iron-

776 oxidizing “Zetaproteobacteria” and mild steel corrosion in nearshore marine

777 environments. Appl. Environ. Microbiol. 77:1405–1412.

778 65. Lee JS, McBeth JM, Ray RI, Little BJ, Emerson D. 2013. Iron cycling at

779 corroding carbon steel surfaces. Biofouling 29:1243–1252.

780 66. McBeth JM, Emerson D. 2016. In situ microbial community succession on mild

781 steel in estuarine and marine environments: Exploring the role of iron-oxidizing

782 bacteria. Front. Microbiol. 7:767.

783 67. Mumford AC, Adaktylou IJ, Emerson D. 2016. Peeking under the iron curtain:

784 development of a microcosm for imaging the colonization of steel surfaces by

785 Mariprofundus sp. strain DIS-1, an oxygen-tolerant Fe-oxidizing bacterium. Appl.

786 Environ. Microbiol. 82:6799–6807.

787 68. Emerson D. In revision. The role of iron-oxidizing bacteria in biocorrosion – a

788 review. Biofouling.

789 69. Rubin-Blum M, Antler G, Tsadok R, Shemesh E, Austin, Jr. JA, Coleman DF,

790 Goodman-Tchernov BN, Ben-Avraham Z, Tchernov D. 2014. First evidence for

791 the presence of iron oxidizing Zetaproteobacteria at the Levantine continental

792 margins. PLoS One 9:e91456.

793 70. Laufer K, Nordhoff M, Schmidt C, Behrens S, Jørgensen BB, Kappler A.

794 2016. Co-existence of microaerophilic, nitrate-reducing, and phototrophic Fe(II)-

795 oxidizers and Fe(III)-reducers in coastal marine sediment. Appl. Environ.

796 Microbiol. 82:1433–1447.

35 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

797 71. Otte JM, Harter J, Laufer K, Blackwell N, Straub D, Kappler A, Kleindienst S.

798 2018. The distribution of active iron-cycling bacteria in marine and freshwater

799 sediments is decoupled from geochemical gradients. Environ. Microbiol. 20:2483–

800 2499.

801 72. Taketani RG, Franco NO, Rosado AS, van Elsas JD. 2010. Microbial

802 community response to a simulated hydrocarbon spill in mangrove sediments. J.

803 Microbiol. 48:7–15.

804 73. McAllister SM, Barnett JM, Heiss JW, Findlay AJ, MacDonald DJ, Dow CL,

805 Luther GWI, Michael HA, Chan CS. 2015. Dynamic hydrologic and

806 biogeochemical processes drive microbially enhanced iron and sulfur cycling

807 within the intertidal mixing zone of a beach aquifer. Limnol. Oceanogr. 60:329–

808 345.

809 74. Beam JP, Scott JJ, McAllister SM, Chan CS, McManus J, Meysman FJR,

810 Emerson D. 2018. Biological rejuvenation of iron oxides in bioturbated marine

811 sediments. ISME J. 12:1389–1394.

812 75. Charette MA, Sholkovitz ER, Hansel CM. 2005. Trace element cycling in a

813 subterranean estuary: Part 1. Geochemistry of the permeable sediments.

814 Geochim. Cosmochim. Acta 69:2095–2109.

815 76. Guan Y, Hikmawan T, Ngugi D, Stingl U. 2015. Diversity of methanogens and

816 sulfate-reducing bacteria in the interfaces of five deep-sea anoxic brines of the

817 Red Sea 166:688–699.

818 77. McBeth JM, Fleming EJ, Emerson D. 2013. The transition from freshwater to

819 marine iron-oxidizing bacterial lineages along a salinity gradient on the Sheepscot

36 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

820 River, Maine, USA. Environ. Microbiol. Rep. 5:453–463.

821 78. Colman DR, Garcia JR, Crossey LJ, Karlstrom K, Jackson-Weaver O,

822 Takacs-Vesbach C. 2014. An analysis of geothermal and carbonic springs in the

823 western United States sustained by deep fluid inputs. Geobiology 12:83–98.

824 79. Emerson JB, Thomas BC, Alvarez W, Banfield JF. 2016. Metagenomic

825 analysis of a high carbon dioxide subsurface microbial community populated by

826 chemolithoautotrophs and bacteria and archaea from candidate phyla. Environ.

827 Microbiol. 18:1686–1703.

828 80. Probst AJ, Castelle CJ, Singh A, Brown CT, Anantharaman K, Sharon I, Hug

829 LA, Burstein D, Emerson JB, Thomas BC, Banfield JF. 2017. Genomic

830 resolution of a cold subsurface aquifer community provides metabolic insights for

831 novel microbes adapted to high CO2 concentrations. Environ. Microbiol. 19:459–

832 474.

833 81. Probst AJ, Ladd B, Jarett JK, Geller-McGrath DE, Sieber CMK, Emerson JB,

834 Anantharaman K, Thomas BC, Malmstrom RR, Stieglmeier M, Klingl A,

835 Woyke T, Ryan MC, Banfield JF. 2018. Differential depth distribution of microbial

836 function and putative symbionts through sediment-hosted aquifers in the deep

837 terrestrial subsurface. Nat. Microbiol. 3:328–336.

838 82. Hug LA, Baker BJ, Anantharaman K, Brown CT, Probst AJ, Castelle CJ,

839 Butterfield CN, Hernsdorf AW, Amano Y, Ise K, Suzuki Y, Dudek N, Relman

840 DA, Finstad KM, Amundson R, Thomas BC, Banfield JF. 2016. A new view of

841 the tree of life. Nat. Microbiol. 1:16048.

842 83. Parks DH, Chuvochina M, Waite DW, Rinke C, Skarshewski A, Chaumeil P-A,

37 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

843 Hugenholtz P. 2018. A proposal for a standardized bacterial taxonomy based on

844 genome phylogeny. Nat. Biotechnol. doi:10.1038/nbt.4229.

845 84. McAllister SM, Davis RE, McBeth JM, Tebo BM, Emerson D, Moyer CL. 2011.

846 Biodiversity and emerging biogeography of the neutrophilic iron-oxidizing

847 Zetaproteobacteria. Appl. Environ. Microbiol. 77:5445–5457.

848 85. Glöckner FO, Yilmaz P, Quast C, Gerken J, Beccati A, Ciuprina A, Bruns G,

849 Yarza P, Peplies J, Westram R, Ludwig W. 2017. 25 years of serving the

850 community with ribosomal RNA gene reference databases and tools. J.

851 Biotechnol. 261:169–176.

852 86. Chen I-MA, Markowitz VM, Chu K, Palaniappan K, Szeto E, Pillay M, Ratner

853 A, Huang J, Andersen E, Huntemann M, Varghese N, Hadjithomas M,

854 Tennessen K, Nielsen T, Ivanova NN, Kyrpides NC. 2017. IMG/M: Integrated

855 genome and metagenome comparative data analysis system. Nucleic Acids Res.

856 45:D507–D516.

857 87. Leinonen R, Sugawara H, Shumway M. 2011. The sequence read archive.

858 Nucleic Acids Res. 39:D19–D21.

859 88. Lagkouvardos I, Joseph D, Kapfhammer M, Giritli S, Horn M, Haller D, Clavel

860 T. 2016. IMNGS: A comprehensive open resource of processed 16S rRNA

861 microbial profiles for ecology and diversity studies. Sci. Rep. 6:33721.

862 89. Holt RD. 2009. Bringing the Hutchinsonian niche into the 21st century: Ecological

863 and evolutionary perspectives. Proc. Natl. Acad. Sci. 106:19659–19665.

864 90. Matheson LJ, Tratnyek PG. 1994. Reductive dehalogenation of chlorinated

865 methanes by iron metal. Environ. Sci. Technol. 28:2045–2053.

38 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

866 91. Bach W, Edwards KJ. 2003. Iron and sulfide oxidation within basalt ocean crust:

867 Implications for chemolithoautotrophic microbial biomass production. Geochim.

868 Cosmochim. Acta 67:3871–3887.

869 92. Dzaugis ME, Spivack AJ, Dunlea AG, Murray RW, D’Hondt S. 2016. Radiolytic

870 hydrogen production in the subseafloor basaltic aquifer. Front. Microbiol. 7:76.

871 93. Scott JJ, Glazer BT, Emerson D. 2017. Bringing microbial diversity into focus:

872 high-resolution analysis of iron mats from the Lō’ihi Seamount. Environ. Microbiol.

873 19:301–316.

874 94. Fullerton H, Hager KW, Moyer CL. 2015. Draft genome sequence of

875 Mariprofundus ferrooxydans strain JV-1, isolated from Loihi Seamount, Hawaii.

876 Genome Announc. 3:e01118-15.

877 95. Fullerton H, Hager KW, McAllister SM, Moyer CL. 2017. Hidden diversity

878 revealed by genome-resolved metagenomics of iron-oxidizing microbial mats from

879 Lō’ihi Seamount, Hawai’i. ISME J. 11:1900–1914.

880 96. Castelle C, Guiral M, Malarte G, Ledgham F, Leroy G, Brugna M, Giudici-

881 Orticoni M-T. 2008. A new iron-oxidizing/O2-reducing supercomplex spanning

882 both inner and outer membranes, isolated from the extreme acidophile

883 Acidithiobacillus ferrooxidans. J. Biol. Chem. 283:25803–25811.

884 97. He S, Barco RA, Emerson D, Roden EE. 2017. Comparative genomic analysis

885 of neutrophilic iron(II) oxidizer genomes for candidate genes in extracellular

886 electron transfer. Front. Microbiol. 8:1584.

887 98. Chan CS, McAllister SM, Garber A, Hallahan BJ, Rozovsky S. 2018. Fe

888 oxidation by a fused cytochrome-porin common to diverse Fe-oxidizing bacteria.

39 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

889 bioRxiv doi:10.1101/228056.

890 99. Bekker M, de Vries S, Ter Beek A, Hellingwerf KJ, Teixeira de Mattos MJ.

891 2009. Respiration of Escherichia coli can be fully uncoupled via the

892 nonelectrogenic terminal cytochrome bd-II oxidase. J. Bacteriol. 191:5510–5517.

893 100. Arai H, Kawakami T, Osamura T, Hirai T, Sakai Y, Ishii M. 2014. Enzymatic

894 characterization and in vivo function of five terminal oxidases in Pseudomonas

895 aeruginosa. J. Bacteriol. 196:4206–4215.

896 101. Jesser KJ, Fullerton H, Hager KW, Moyer CL. 2015. Quantitative PCR analysis

897 of functional genes in iron-rich microbial mats at an active hydrothermal vent

898 system (Lō’ihi Seamount, Hawai’i). Appl. Environ. Microbiol. 81:2976–2984.

899 102. Sylvan JB, Wankel SD, LaRowe DE, Charoenpong CN, Huber JA, Moyer CL,

900 Edwards KJ. 2017. Evidence for microbial mediation of subseafloor nitrogen

901 redox processes at Loihi Seamount, Hawaii. Geochim. Cosmochim. Acta

902 198:131–150.

903 103. Winterbourn CC. 1995. Toxicity of iron and hydrogen peroxide: the Fenton

904 reaction. Toxicol. Lett. 82/83:969–974.

905 104. Jones LC, Peters B, Lezama Pacheco JS, Casciotti KL, Fendorf S. 2015.

906 Stable isotopes and iron oxide mineral products as markers of

907 chemodenitrification. Environ. Sci. Technol. 49:3444–3452.

908 105. Martiny AC, Treseder K, Pusch G. 2013. Phylogenetic conservatism of

909 functional traits in microorganisms. ISME J. 7:830–838.

910 106. Anbar AD. 2004. Iron stable isotopes: Beyond biosignatures. Earth Planet. Sci.

911 Lett. 217:223–236.

40 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

912 107. Luther GW, Glazer BT, Ma S, Trouwborst RE, Moore TS, Metzger E, Kraiya C,

913 Waite TJ, Druschel G, Sundby B, Taillefert M, Nuzzio DB, Shank TM, Lewis

914 BL, Brendel PJ. 2008. Use of voltammetric solid-state (micro)electrodes for

915 studying biogeochemical processes: Laboratory measurements to real time

916 measurements with an in situ electrochemical analyzer (ISEA). Mar. Chem.

917 108:221–235.

918 108. Takai K, Mottl MJ, Nielsen SHH, IODP Expedition 331 Scientists. 2012. IODP

919 expedition 331: Strong and expansive subseafloor hydrothermal activities in the

920 Okinawa trough. Sci. Drill. 13:19–27.

921 109. Moore RM, Harrison AO, McAllister SM, Wommack KE. 2018. Iroki: automatic

922 customization and visualization of phylogenetic trees. bioRxiv

923 doi:10.1101/106138.

924 110. Makita H, Tanaka E, Mitsunobu S, Miyazaki M, Nunoura T, Uematsu K,

925 Takaki Y, Nishi S, Shimamura S, Takai K. 2017. Mariprofundus micogutta sp.

926 nov., a novel iron-oxidizing zetaproteobacterium isolated from a deep-sea

927 hydrothermal field at the Bayonnaise knoll of the Izu-Ogasawara arc, and a

928 description of Mariprofundales ord. nov. and Zetaproteobacteria classis. Arch.

929 Microbiol. 199:335–346.

930 111. Laufer K, Nordhoff M, Halama M, Martinez RE, Obst M, Nowak M, Stryhanyuk

931 H, Richnow HH, Kappler A. 2017. Microaerophilic Fe(II)-oxidizing

932 Zetaproteobacteria isolated from low-Fe marine coastal sediments: Physiology

933 and characterization of their twisted stalks. Appl. Environ. Microbiol. 83:e03118-

934 16.

41 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

935 112. Makita H, Kikuchi S, Mitsunobu S, Takaki Y, Yamanaka T, Toki T, Noguchi T,

936 Nakamura K, Abe M, Hirai M, Yamamoto M, Uematsu K, Miyazaki J, Nunoura

937 T, Takahashi Y, Takai K. 2016. Comparative analysis of microbial communities

938 in iron-dominated flocculent mats in deep-sea hydrothermal environments. Appl.

939 Environ. Microbiol. 82:5741–5755.

940 113. Hassenrück C, Fink A, Lichtschlag A, Tegetmeyer HE, de Beer D, Ramette A.

941 2016. Quantification of the effects of ocean acidification on sediment microbial

942 communities in the environment: The importance of ecosystem approaches.

943 FEMS Microbiol. Ecol. 92:fiw02.

944 114. Gonnella G, Böhnke S, Indenbirken D, Garbe-Schönberg D, Seifert R,

945 Mertens C, Kurtz S, Perner M. 2016. Endemic hydrothermal vent species

946 identified in the open ocean seed bank. Nat. Microbiol. 1:16086.

947

948 Table legends

949 Table 1. Isolates of the Zetaproteobacteria and their assigned ZOTUs, with

950 representation in the environment, biomineral and metabolic properties, and references.

951

952 Table 2. Growth preferences of Zetaproteobacteria isolates, including optimal growth

953 salinity, temperature, pH, and oxygen concentrations.

954

955 Table 3. Biotic and abiotic Fe oxidation rates of Mariprofundus ferrooxydans PV-1 under

956 a range of O2 concentrations.

957

42 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

958 Table 4. Summary of the habitats where the Zetaproteobacteria are found in high

959 abundance using data from Dataset S1.

960

961 Figure captions

962 Figure 1. Relative rates of biotic versus abiotic Fe(II) oxidation at varying oxygen

963 concentrations, using the model Zetaproteobacteria isolate, M. ferrooxydans PV-1.

964 Range of 6.5 – 6.7 pH for experimental conditions. Further details in Essential Methods.

965

966 Figure 2. Morphologies of FeOOH biominerals known or suspected to be formed by the

967 Fe-oxidizing Zetaproteobacteria. (A-C) Twisted stalks from individual cell to intact Loihi

968 Fe microbial mats. (D-F) Sheaths from individual tubes to intact Loihi mats. (G-I) Fe

969 biominerals attached to stalks and sheaths in Loihi mats. (A) A single M. ferrooxydans

970 PV-1 cell (arrow) and its fibrillar stalk. (B) Intact curd-type mats (Figure 4A) are

971 composed of parallel stalks. (C) Intact mat showing directional stalks that are

972 interrupted (arrow), before biomineral production resumes. See Chan et al. (27) for

973 details. (D) Hollow sheaths formed by the Zetaproteobacteria. (E) Intact veil-type mats

974 (Figure 4B) are composed of sheaths. (F) Zoomed out intact mat with multiple sets of

975 sheaths oriented in different directions (direction corresponds to color). (G-H) Short, Y-

976 shaped tubular biominerals formed by Zetaproteobacteria. (H) Arrows show attached

977 cells. (I) Siderocapsa-like nest-type biominerals; it is not known if they are formed by the

978 Zetaproteobacteria. Scale bars: 0.5 µm (A,D), 2 µm (B,E,G-I), 100 µm (C,F). Images

979 reproduced with permission: (A) from (28), (F, H) from (27). (B-I) from the samples

980 described in (27). (D, I) imaged on JEOL-7200 field emission SEM.

43 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

981

982 Figure 3. Colorized SEM images of stalk and dread biominerals of select marine and

983 freshwater FeOB. (A) Twisted stalk (blue) formed by a M. ferrooxydans PV-1 cell

984 (yellow). Most Zetaproteobacteria isolates form stalks. (C) In contrast, M. aestuarium

985 CP-5 and M. ferrinatatus CP-8 produce only short dreads (red), which are easily shed

986 from the cell (inferred cell position indivated by dashed yellow line). (B) Stalk (blue) and

987 dreads (red) of the freshwater Betaproteobacteria FeOB Ferriphaselus sp. R-1

988 resembled the structures formed by marine FeOB. Scale bars = 1 µm. Images

989 reproduced with permission: (A) from (27), (B) from (30), (C) from (25).

990

991 Figure 4. Photographs of Zetaproteobacteria habitats. (A-D) Marine hydrothermal vent

992 mats, where Zetaproteobacteria have been found in highest abundance. (A) Curd-type

993 and (B) veil-type Fe mats, from Loihi Seamount. (C) Mn-crusted Fe mat from the Ula

994 Nui site, Loihi. Fe mat visible under broken surface (bottom right). (D) Fe mats on the

995 Golden Horn Chimney, at the Urashima vent site, Mariana Trough. (E) Transition from

996 reduced to Fe oxide-stained marine sediments (dashed line) in 26 mbsf core from the

997 hydrothermal circulation cell of Iheya North vent field, Okinawa Trough. See (108) for

998 details. (F) Terrestrial saline CO2-rich spring at Crystal Geyser, UT. (G) Fe oxide-coated

999 worm burrows from the beach at Cape Shores, DE. (H) Mild steel corrosion biofilm

1000 formed by isolate M. sp. GSB-2. (H) Reproduced with permission from Joyce M.

1001 McBeth. (F) Original photography by Chris T. Brown; reproduced with permission.

1002

44 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1003 Figure 5. Maximum likelihood phylogenetic tree showing Zetaproteobacteria 16S gene

1004 diversity, (A) colored by ZOTU and (B) colored by habitat type where the sequences

1005 were sampled. A total of 59 ZOTUs have been classified, though only the most

1006 frequently sampled are shown in the figure above. ZOTUs 1 and 14 are poorly resolved

1007 by the 16S gene. Published isolates of the Zetaproteobacteria are starred and labeled.

1008 Phylogenetic trees were colored automatically using Iroki (109).

1009

1010 Figure 6. Maximum likelihood phylogenetic tree showing a more robust phylogenetic

1011 placement of ZOTUs based on the concatenated alignments of 12 ribosomal proteins.

1012 Data from isolate, SAG, and MAG genomes (see Supplemental Methods). In this tree,

1013 ZOTUs 1 and 14 are monophyletic entities. Taxa near ZOTU9 and near ZOTU54 are

1014 only represented by genomes; these clades have only been found in the terrestrial CO2-

1015 rich spring waters of Crystal Geyser, UT.

1016

1017 Figure 7. Zetaproteobacteria OTU network showing the association of known ZOTUs

1018 within the specified habitats. Lines connect ZOTUs that are found in the same sample,

1019 with thickness representing the frequency of that association in multiple samples.

1020 Colored circles surrounding ZOTU nodes show samples where only a single ZOTU was

1021 found. ZOTU nodes are sized according to their environmental abundance. Placement

1022 of ZOTUs in the network was determined automatically, based on the frequency of co-

1023 occurrence (Cytoscape’s edge-weighted, spring embedded layout). Dotted lines denote

1024 ZOTU clusters common to the following habitats: Cluster 1, metal corrosion and mineral

45 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1025 weathering; Cluster 2, ubiquitous Fe mats; Cluster 3, Mariana Trough Fe mats. Isolate

1026 sequences are not shown. Dataset based on SILVA release 128.

1027

1028 Figure 8. Model for Fe oxidation in the Zetaproteobacteria modified from (45). An

1029 electron from Fe(II) is passed from Cyc2 to a periplasmic electron carrier (Cyc1 and/or

1030 other c-type cytochrome) before being passed to the terminal oxidase (cbb3- or aa3-type

1031 cytochrome c oxidases), generating a proton motive force. For reverse electron

1032 transport, the electron from the periplasmic carrier is passed to the bc1 complex or

1033 alternative complex III before being passed to the quinone pool where it is used to

1034 regenerate NADH.

1035

1036 References in tables only:

1037 (110) – Makita et al., 2017

1038 (111) – Laufer et al., 2017

1039 (112) – Makita et al., 2016

1040 (113) – Hassenrück et al., 2016

1041 (114) – Gonnella et al., 2016

46 bioRxiv preprint not certifiedbypeerreview)istheauthor/funder,whohasgrantedbioRxivalicensetodisplaypreprintinperpetuity.Itmadeavailable doi: https://doi.org/10.1101/416842

Essential Methods Furtherdetailsin experimental conditions. PV-1. Rangeof6.5–6.7pHfor Zetaproteobacteria model concentrations, usingthe at varyingoxygen abiotic Fe(II)oxidation Figure 1.Relativeratesofbioticversus % Biotic Fe(II) oxidation under a ; this versionpostedSeptember16,2018. CC-BY 4.0Internationallicense O 2 concentration(μM) . isolate, M.ferrooxydans R 2 =0.995 . The copyrightholderforthispreprint(whichwas bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

A B C

D E F

G H I

Figure 2. Morphologies of FeOOH biominerals known or suspected to be formed by the Fe-oxidizing Zetaproteobacteria. (A-C) Twisted stalks from individual cell to intact Loihi Fe microbial mats. (D-F) Sheaths from individual tubes to intact Loihi mats. (G-I) Fe biominerals attached to stalks and sheaths in Loihi mats. (A) A single M. ferrooxydans PV-1 cell (arrow) and its fibrillar stalk. (B) Intact curd-type mats (Figure 4A) are composed of parallel stalks. (C) Intact mat showing directional stalks that are interrupted (arrow), before biomineral production resumes. See Chan et al. (27) for details. (D) Hollow sheaths formed by the Zetaproteobacteria. (E) Intact veil-type mats (Figure 4B) are composed of sheaths. (F) Zoomed out intact mat with multiple sets of sheaths oriented in different directions (direction corresponds to color). (G-H) Short, Y-shaped tubular biominerals formed by Zetaproteobacteria. (H) Arrows show attached cells. (I) Siderocapsa-like nest-type biominerals; it is not known if they are formed by the Zetaproteobacteria. Scale bars: 0.5 µm (A,D), 2 µm (B,E,G-I), 100 µm (C,F). Images reproduced with permission: (A) from (28), (F, H) from (27). (B-I) from the samples described in (27). (D, I) imaged on JEOL-7200 field emission SEM. bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

A

B

C

Figure 3. Colorized SEM images of stalk and dread biominerals of select marine and freshwater FeOB. (A) Twisted stalk (blue) formed by a M. ferrooxydans PV-1 cell (yellow). Most Zetaproteobacteria isolates form stalks. (C) In contrast, M. aestuarium CP-5 and M. ferrinatatus CP-8 produce only short dreads (red), which are easily shed from the cell (inferred cell position indivated by dashed yellow line). (B) Stalk (blue) and dreads (red) of the freshwater Betaproteobacteria FeOB Ferriphaselus sp. R-1 resembled the structures formed by marine FeOB. Scale bars = 1 µm. Images reproduced with permission: (A) from (27), (B) from (30), (C) from (25). bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. A D E

10 cm

B 10 cm C

2 cm

F G

1 cm

H 2 mm

35 cm 50 cm 4 cm

Figure 4. Photographs of Zetaproteobacteria habitats. (A-D) Marine hydrothermal vent mats, where Zetaproteobacteria have been found in highest abundance. (A) Curd-type and (B) veil-type Fe mats, from Loihi Seamount. (C) Mn-crusted Fe mat from the Ula Nui site, Loihi. Fe mat visible under broken surface (bottom right). (D) Fe mats on the Golden Horn Chimney, at the Urashima vent site, Mariana Trough. (E) Transition from reduced to Fe oxide-stained marine sediments (dashed line) in 26 mbsf core from the hydrothermal circulation cell of Iheya North vent field, Okinawa Trough. See (108) for details. (F) Terrestrial saline CO2-rich spring at Crystal Geyser, UT. (G) Fe oxide-coated worm burrows from the beach at Cape Shores, DE. (H) Mild steel corrosion biofilm formed by isolate M. sp. GSB-2. (H) Reproduced with permission from Joyce M. McBeth. (F) Original photography by Chris T. Brown; reproduced with permission. bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

Mariprofundaceae A (Family 1) ZOTU6

ZOTU9 91 Family 2

100 ZOTU15 100 ZOTU1/14 ZOTU11/23/36 56 36 71 Mariprofundus ferrooxydans 40 97 ZOTU54 65 59 100 ZOTU4 100 100 ZOTU13 ZOTU2 99

ZOTU18/37 ZOTU10 0.05 Outgroups ZOTU6 B ZOTU9 91

100 ZOTU15 100 ZOTU1/14 56 71 ZOTU11/23/36 36 Habitat: 40 97 ZOTU54 Fe mats 65 59 Sediments ZOTU4 100 Marine subsurface 100 Weathering/corrosion 99 ZOTU13 Coastal ZOTU18/37 Terrestrial subsurface ZOTU2 ZOTU10 0.05 Outgroups

Figure 5. Maximum likelihood phylogenetic tree showing Zetaproteobacteria 16S gene diversity, (A) colored by ZOTU and (B) colored by habitat type where the sequences were sampled. A total of 59 ZOTUs have been classified, though only the most frequently sampled are shown in the figure above. ZOTUs 1 and 14 are poorly resolved by the 16S gene. Published isolates of the Zetaproteobacteria are starred and labeled. Phylogenetic trees were colored automatically using Iroki (109). bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

Mariprofundaceae ZOTU18 + (Family 1) ZOTU3/36/37 ZOTU14 ZOTU11

ZOTU1 100 39 Family 2 100

97 100 85 53 44 100 62 32 48 46 ZOTU9 100

100 ZOTU4 100 near ZOTU9 ZOTU13

(no 16S) 100

ZOTU6 ZOTU2 ZOTU10 0.1 Outgroups

Figure 6. Maximum likelihood phylogenetic tree showing a more robust phylogenetic placement of ZOTUs based on the concatenated alignments of 12 ribosomal proteins. Data from isolate, SAG, and MAG genomes (see Supplemental Methods). In this tree, ZOTUs 1 and 14 are monophyletic entities. Taxa near ZOTU9 and near ZOTU54 are only represented by genomes; these clades have only been found in the terrestrial CO2-rich spring waters of Crystal Geyser, UT. bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

Habitat: Hydrothermal Fe mats Hydrothermal sediments Marine subsurface fluids Terrestrial subsurface fluids Metal corrosion Mineral weathering Other

Cluster 3

Figure 7. Zetaproteobacteria OTU network showing the association of known ZOTUs within the specified habitats. Lines connect ZOTUs that are found in the same sample, with thickness representing the frequency of that association in multiple samples. Colored circles surrounding ZOTU nodes show samples where only a single ZOTU was found. ZOTU nodes are sized according to their environmental abundance. Placement of ZOTUs in the network was determined automatically, based on the frequency of co-occurrence (Cytoscape’s edge-weighted, spring embedded layout). Dotted lines denote ZOTU clusters common to the following habitats: Cluster 1, metal corrosion and mineral weathering; Cluster 2, ubiquitous Fe mats; Cluster 3, Mariana Trough Fe mats. Isolate sequences are not shown. Dataset based on SILVA release 128. bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

H2O 2+ 3+ Fe Fe FeOOH + H - Outer Membrane e Cyc2 Cyc1 ? + Periplasm H + + H H Cyt ? NADH bc or cbb or aa ATP Inner Membrane Q 1 - - 3 3 Dehydrogenase ACIII e e cyt oxidase synthase

+ + + 1/2 O2 + 2H H2O H NAD NADH ADP ATP

Figure 8. Model for Fe oxidation in the Zetaproteobacteria modified from (45). An electron from Fe(II) is passed from Cyc2 to a periplasmic electron carrier (Cyc1 and/or other c-type cytochrome) before being passed to the terminal oxidase (cbb3- or aa3-type cytochrome c oxidases), generating a proton motive force. For reverse electron transport, the electron from the periplasmic carrier is passed to the bc1 complex or alternative complex III before being passed to the quinone pool where it is used to regenerate NADH. bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

Table 1. Isolates of the Zetaproteobacteria and their assigned ZOTUs, with representation in the environment, biomineral and metabolic properties, and references. ZOTU Primary Fe(II) H Genus (if named) Species Strain ZOTU Envir. Isolation source biomineral 2 Refs. oxidation abund. morphology oxidation Mariprofundus ferrooxydans PV-1 ZOTU11 Loihi hydrothermal vents stalk X (15) Mariprofundus ferrooxydans JV-1 ZOTU11 Loihi hydrothermal vents stalk X (15) 1.58% Mariprofundus ferrooxydans M34 ZOTU11 Loihi hydrothermal vents stalk X (84) Mariprofundus ferrooxydans SC-2 ZOTU11 Big Fisherman's Cove pyrrhotite colonization stalk X (62) Mariprofundus ferrinatatus CP-8 ZOTU37 0.09% Chesapeake Bay stratified water column dreads only X (25) Mariprofundus aestuarium CP-5 ZOTU18 Chesapeake Bay stratified water column dreads only X (25) Mariprofundus micogutta ET2 ZOTU18 1.49% Bayonnaise hydrothermal vents thin filaments X (110) Mariprofundus micogutta DIS-1 ZOTU18 West Boothbay Harbor mild steel incubation stalk X (67) Mariprofundus sp. GSB-2 ZOTU23 0.26% Great Salt Bay salt marsh Fe mat stalk X (64) Mariprofundus sp. EKF-M39 ZOTU36 0.35% Loihi hydrothermal vents stalk X (43) Ghiorsea bivora TAG-1 ZOTU9 MAR hydrothermal vents none XX (21) 8.42% Ghiorsea bivora SV-108 ZOTU9 Mariana hydrothermal vents none XX (21) Zetaproteobacteria sp. CSS-1 ZOTU14 4.04% coastal sediment bloodworm microcosm stalk X (74) Zetaproteobacteria sp. S1OctC ZOTU3 Norsminde Fjord estuary sediments stalk X (111) 3.51% Zetaproteobacteria sp. S2.5 ZOTU3 Kalø Vig beach sediments stalk X (111) bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

Table 2. Growth preferences of Zetaproteobacteria isolates, including optimal growth salinity, temperature, pH, and oxygen concentrations.

Growth salinity (ppt) Growth temperature (˚C) Growth pH Growth [O2] Doubling Headspace Range

Genus (if named) Species Strain ZOTU time (h) Isolation Range Optimum Isolation Range Optimum Isolation Range Optimum (% O2) (µM) Refs. Mariprofundus ferrooxydans PV-1 ZOTU11 12 35 (ASW) 3.5-35a 28-31.5a 12-24 10-30 30 6.4-6.5 5.5-7.2 6.2-6.5 1 – (1, 15) Mariprofundus ferrooxydans JV-1 ZOTU11 12 35 (ASW) –– 12-24 10-30 30 6.4-6.5 5.5-7.2 6.2-6.5 1 – (1, 15) Mariprofundus ferrooxydans M34 ZOTU11 – 35 (ASW) – – RT – – – – – 5-15 – (84) Mariprofundus ferrooxydans SC-2 ZOTU11 – 35 (ASW) – – RT – – 6.2-6.5 –– 3.4 – (62) Mariprofundus ferrinatatus CP-8 ZOTU37 27 18 (SEM)b 7-31.5 14-17.5 RT 15-35 25-30 6.2 5.5-8.3 6.9-7.2 1 0.07-1.7 (25) Mariprofundus aestuarium CP-5 ZOTU18 19.5 18 (SEM)b 7-31.5 14-17.5 RT 10-30 20-25 6.2 5.5-8.3 6.9-7.2 1 0.31-2.0 (25) Mariprofundus micogutta ET2 ZOTU18 24 35 (ASW) 10-40 27.5 25 15-30 25 6.2-6.5 5.8-7.0 6.4 1-3 – (110) Mariprofundus micogutta DIS-1 ZOTU18 –c 35 (ASW) – – RT – – 6.1-6.4 max 8 – 5-15 max ~220 (67) Mariprofundus sp. GSB-2 ZOTU23 13d 35 (ASW) 1.75-35 – RT – 25 6.1-6.4 max 7.25 – 5-15 – (64) Mariprofundus sp. EKF-M39 ZOTU36 – 35 (ASW) ––––– 6.1-6.4 –– 0e – (43) Ghiorsea bivora TAG-1 ZOTU9 21.8f 35 (ASW) ––– 5-30 20 6.5 5.5-7.5 6.5-7.0 0.4-15 – (21) Ghiorsea bivora SV-108 ZOTU9 20.0f 35 (ASW) ––– 5-30 20 6.5 6.0-7.5 6.5-7.0 0.4-15 – (21) Zetaproteobacteria sp. CSS-1 ZOTU14 – 35 (ASW) – – RT – – – – – 5-10 opt. 60g (74) Zetaproteobacteria sp. S1OctC ZOTU3 – 23 (ASW) 6.9-23 –––– 7.1 –– 6-10 – (111) Zetaproteobacteria sp. S2.5 ZOTU3 – 23 (ASW) 6.9-23 –––– 7.1 –– 6-10 – (111)

RT = room temperature ASW = Artificial Seawater Medium a) Data from (25) b) SEM = Simulated estuary medium (50:50 MWMM:ASW); MWMM = Modified Wolfe's Mineral Medium c) Cited as comparible with other FeOB d) 13 h doubling time in standard gradient tubes; 7-8 h doubling time on metal coupons e) Trace O2 was likely introduced into the culture with aerobic vitamin/mineral solutions and mat innoculum f) Doubling time on Fe(II) shown; Doubling time on H2 was 14.1 h and 16.3 h for TAG-1 and SV-108, respectively g) Reported as approximate optimum bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

Table 3. Biotic and abiotic Fe oxidation rates of Mariprofundus ferrooxydans

PV-1 under a range of O2 concentrations. Fe(II) oxidation rate O2 conc. biotic biotic abiotic % Biotic Fe(II) µM µM Fe(II) hr-1 µM Fe(II) cell-1 hr-1 µM Fe(II) hr-1 oxidation pH 10.4 21.05 5.06E-04 0.32 98.5 6.63 10.4 24.69 5.53E-04 1.66 93.7 6.70 54.8 33.32 1.00E-03 39.43 45.8 6.67 79.7 5.57 1.25E-04 47.31 10.5 6.50 bioRxiv preprint doi: https://doi.org/10.1101/416842; this version posted September 16, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

Table 4. Summary of the habitats where the Zetaproteobacteria are found in high abundance using data from Dataset S1. No. of samples Habitat + Features ZOTU compositiona Total 16Sb NGSb Key Refs. hydrothermal Fe mats ZOTU2 ZOTU1 • flocculent Fe mats (bulk & discrete) ZOTU9 • Zetaproteobacteria are mat architects (47, 48, 51, ZOTU14 113 31 72 • 1.7˚C – 77˚C temperature range ZOTU4 84, 93, 112) • <10–96% of bacterial community ZOTU8 ZOTU6 Other hydrothermally-influenced sediments ZOTU59 • shallow sediment cores from ZOTU4 deep sea seamounts/ridges & ZOTU1 23 6 11 (13, 113) near shore hydrothermal CO2 seeps ZOTU9 • 2-32% of bacterial community ZOTU2 Other marine subsurface fluids ZOTU9 • depth of 5-332 mbsf ZOTU58 • sampled from borehole fluids 9 3 2 primarily at the Mariana ZOTU57 (18, 114) Backarc and Mid-Atlantic Ridge ZOTU15 • 11-50% of bacterial community Other terrestrial subsurface fluids ZOTU54 • CO2-rich saline terrestrial springs in Utah and New Mexico ZOTU9 3 3 0 • 10-31% of bacterial community Other (78-80)

metal corrosion incubations • carbon steel and mild steel coupons ZOTU9 incubated in near-shore ZOTU18 9 4 0 environments ZOTU2 (62-64) • 6-26% of bacterial community Other mineral weathering incubations • incubation of pyrrhotite, basalt, and ZOTU9 ZVI in water column, borehole fluid, ZOTU6 7 2 1 (61, 62) and sediment environments ZOTU22 • 31-43% of bacterial community Other a ZOTUs 8, 22, and 57-59 are colored according to their closest colored representative in Figure 5. ZOTU8 is a close relative to ZOTU2. ZOTUs 57/58 are close relatives of ZOTU9. ZOTU22 is closest to ZOTU10. ZOTU59 is closest to ZOTU6. b No. of samples with more than one sequence