Genomes of Thaumarchaeota from deep sea sediments reveal specific adaptations of three independently evolved lineages

Marine sediments represent a vast habitat for complex microbiomes. Among these, ammonia oxidizing archaea (AOA) of the phylum Thaumarchaeota are one of the most common, yet little explored, inhabitants, which seem extraordinarily well adapted to the harsh conditions of the subsurface biosphere. We present 11 metagenome-assembled genomes of the most abundant AOA clades from sediment cores obtained from the Atlantic Mid-Ocean ridge flanks and Pacific abyssal plains. Their phylogenomic placement reveals three independently evolved clades within the order Nitrosopumilales, of which no cultured representative is known yet. In addition to the gene sets for ammonia oxidation and carbon fixation known from other AOA, all genomes encode an extended capacity for the conversion of fermentation products that can be channeled into the central carbon metabolism, as well as uptake of amino acids probably for protein maintenance or as an ammonia source. Two lineages encode an additional (V-type) ATPase and a large repertoire of DNA repair systems that may allow to overcome the challenges of high hydrostatic pressure. We suggest that the adaptive radiation of AOA into marine sediments occurred more than once in evolution and resulted in three distinct lineages with particular adaptations to this extremely energy-limiting and high-pressure environment.


44
Ammonia oxidizing archaea (AOA) comprise one of the most successful archaeal phyla having zone), and 2.5 mbsf (Mn-reduction zone) were selected from GS14-GC08. Detailed information about sampling sites, sampling procedure, 16S rRNA gene profiles and porewater analysis was published in >80 % completeness, all with contamination levels £ 5%, which we consider high quality MAGs in this study, as well as four additional medium quality MAGs (66 to 76% completeness, up to 6.3% 183 contamination level) (Table1). The MAGs genome sizes (0.61 to 1.52 Mb) and GC contents (34.66 ± 184 0.72%) are in accordance with previous reports of free-living Ca. Nitrosopumilales [56] .

186
In order to study the evolution of AOA and place our deep marine sediments-derived MAGs in a 187 phylogenetic context, we reconstructed a maximum-likelihood (ML) phylogenomic tree (Fig. 2a) using 188 79 concatenated single-copy markers from our entire dataset of 85 complete genomes, MAGs and SAGs 189 representing a broad diversity of habitats (Table S1).

191
In addition, we performed an amoA-based phylogeny as in [28] in order to assign a taxonomical rank  sediments and also found to be abundant in the crust below [14,28] . Three MAGs are affiliated to 204 amoA-NP-delta, the second most abundant AOA clade in marine sediments, and are the first marine 205 sediment representatives of this clade, which includes a single other MAG (Archaeon CSP1) assembled suggested that this clade might be restricted to deep-sea sediments, an ecological specialization only found in Ca. Nitrosopumilales. Taken together, this early-branching clade seems to be a new NP While MAGs and SAGs of AOA from bathypelagic (1000-4000m) to abyssopelagic (4000-6000m) and 225 hadopelagic (6000-11000m) environments have been reported previously [23,24,58] and shotgun 226 metagenomic analyses of deep-sea sediments have been performed [10,59,60], the MAGs reported in 227 this study represent to our knowledge the first high-quality AOA genomes from bathyal and abyssal

231
While our phylogenomic tree supports that AOA likely appeared in terrestrial habitats first (Fig. 2a), 232 the sequence and number of colonization events of marine environments seem to be more complex than 233 previously proposed [29]. Importantly, our results suggest that both the deep-water adapted AOA as 234 well as the deep-sediment adapted AOA are polyphyletic. The colonization of deep waters, i.e. pelagic 235 organisms, might have occurred independently at least twice in the evolution of the Ca.

236
Nitrosopumilales, once at the origin of amoA-NP-alpha clade and the other during the diversification 237 of amoA-NP-gamma (Fig. 2a)

244
Regarding the origin of deep sediments-dwelling AOA, the amoA-NP-theta lineage branches within 245 mostly marine NP clades (Fig. 2a), suggesting that this lineage might have evolved from a pelagic 246 marine ancestor. However, there is no evolutionary link between any pelagic AOA and the NP-delta 247 clade, as the latter have been mostly retrieved in estuarine and deep marine sediments [57]. The most 248 parsimonious evolutionary scenario would be that this group underwent a direct transition from 249 estuarine sediments to marine sediments during its diversification.

251
Similarly, the newly proposed clade NP-iota, the earliest branching Nitrosopumilales which has so far 252 exclusively been detected in marine sediments [28] does not seem to be closely related to pelagic Although, it is possible, that pelagic AOA closely related to amoA-NP-iota may be detected in further a wide variety of ecological environments (see Materials and Methods and Table S1). From these, 262 12,137 have representatives from at least two different genomes. In our analysis, the AOA core

278
To identify possible specific adaptations of AOA to deep marine sediments, we searched for families 279 present in at least two of the three marine sediments clades represented by our 11 MAGs (i.e. amoA- Table S1). 72 families were identified (Fig. 3), of which only 25 % (18 families) could be functionally 282 annotated (Table S3) and were classified into the following categories: information processing systems 283 (7), metabolism (5) and cellular processes (6). Some of these 18 families had functional equivalents in 284 most if not all AOA (e.g. RadA homologs). We additionally found 41 families shared predominantly The full annotations for all genes and pathways discussed in the following section can be found in Table   300 S4.   Table S3). Together with a putative nitrilase (Nit1, conserved in AOA) and a 313 putative omega-amidase (Nit2, present in amoA-NP-theta, -delta, -eta, -gamma), these genes indicate

319
All three sediment clades encode full gene sets for electron transfer to O2 via NADH dehydrogenase 320 (complex I), type bc1 complex III, and a heme-copper terminal oxidase (complex IV) (Fig. 5, Table S3).

321
No alternative complexes using a different electron acceptor were identified.

332
Complete or near-complete amino acid biosynthesis pathways as well as vitamins (including vitamin 333 B12) are present in all three sediment clades, as in other AOA (Table S3)

355
This indicates the capacity to use formate for supplementing CO2 needs while concomitantly supplying 356 reducing equivalents (as in methylotrophs [74,75]).

358
A putative 3-aminobutyryl-CoA aminotransferase (Kat, EC 2.6.1.111) and a 3-aminobutyryl-CoA 359 ammonia lyase (Kal, EC 4.3.1.14) were identified in the NP-theta and NP-delta bins, and are also found 360 in some NP-alpha, NP-gamma and NT lineages (Fig. 4, 5). These enzymes participate in lysine could be scavenged from fermenting microorganisms in the sediment community. Both enzymes can remove ammonia from 3-aminobutyryl-CoA either by transferring it to α-ketoglutarate resulting in the and could be processed accordingly, generating reducing potential in the subsequent steps.

369
The presence of a putative lactate racemase family protein (LarA), specific to the NP-theta, delta and 370 iota clades (Fig. 4, 5) suggests that lactate is another fermentation product that could be utilized by these 371 lineages. This is one of the very few protein families with a putative function prediction shared

488
All AOA adapted to marine sediments and investigated in this study are able to perform ammonia 489 oxidation in combination with CO2 fixation like all other described AOA. In addition, all three lineages 490 seem to be capable of utilising exogenous organic fermentation products that they convert into 491 intermediates of their central carbon metabolism, a feature they share with pelagic AOA from the deep 492 ocean and a few other AOA. This, together with the capability of taking up aminoacids, putatively for 493 recycling into proteins or utilization of amine groups, would support growth in this extremely 494 oligotrophic environment and contribute to organic nitrogen and carbon turnover in the sediments. In 495 the absence of any components indicating increased capacity of amino acid degradation in these AOA 496 we argue that recycling of amino acids rather than catabolism as otherwise suggested in [10,23,113]