Characterization of a thaumarchaeal symbiont that drives incomplete nitrification in the tropical sponge Ianthella basta

Abstract

Marine sponges represent one of the few eukaryotic groups that frequently harbour symbiotic members of the Thaumarchaeota, which are important chemoautotrophic ammonia-oxidizers in many environments. However, in most studies, direct demonstration of ammonia-oxidation by these archaea within sponges is lacking, and little is known about sponge-specific adaptations of ammonia-oxidizing archaea (AOA). Here, we characterized the thaumarchaeal symbiont of the marine sponge Ianthella basta using metaproteogenomics, fluorescence in situ hybridization, qPCR and isotope-based functional assays. 'Candidatus Nitrosospongia ianthellae' is only distantly related to cultured AOA. It is an abundant symbiont that is solely responsible for nitrite formation from ammonia in I. basta that surprisingly does not harbour nitrite-oxidizing microbes. Furthermore, this AOA is equipped with an expanded set of extracellular subtilisin-like proteases, a metalloprotease unique among archaea, as well as a putative branched-chain amino acid ABC transporter. This repertoire is strongly indicative of a mixotrophic lifestyle and is (with slight variations) also found in other sponge-associated, but not in free-living AOA. We predict that this feature as well as an expanded and unique set of secreted serpins (protease inhibitors), a unique array of eukaryotic-like proteins, and a DNA-phosporothioation system, represent important adaptations of AOA to life within these ancient filter-feeding animals.

Figure 1

Fluorescence in situ hybridization of a 5 μm cryosection of I. basta using double‐labelled (Stoecker et al., 2010) probe Arch915 in red and the double‐labelled probe EUB338‐I‐III set in blue. Green and white/lila structures represent autofluorescence (see white arrows). As I. basta harbours ‘Ca. Nitrosospongia ianthellae’ as the only archaeon, all red signals represent this AOA. Blue signals represent bacterial symbionts.

Figure 2

Phylogeny of Ca. Nitrosospongia ianthellae. A. Bayesian 16S rRNA gene tree. B. Bayesian phylogenomic tree based on 34 concatenated universal, single‐copy marker genes identified with CheckM (Parks et al., 2015). Bayesian posterior support values >0.5 are indicated for each branch. Outgroups for both trees consisted of all three genome‐sequenced members of the Nitrososphaera cluster, both members of the Nitrosocosmicus clade, and Ca. Nitrosocaldus icelandicus. In all trees, sequences obtained from sponges are depicted in bold.

Figure 3

Heat map showing the distribution and gene copy number per genome of selected genes and gene classes among genome‐sequenced AOA. The colour scale indicates copies per genome and MAG respectively. Sponge‐derived MAGs start on the left and are depicted in bold, followed by members of Ca. Nitrosopumilaceae, Ca. Nitrosotenuaceae, Ca. Nitrosotaleales, the Nitrososphaerales, and Ca. Nitrosocaldales. Genome sizes are listed next to each species name. An extended version of this Figure is available as Supporting Information Fig. S6.

Figure 4

A. Reconstructed metabolic pathways of genes and their expression (red) detected in Ca. N. ianthellae proposed to be involved in extracellular (and intracellular) protein degradation as well as amino acid transport and assimilation. Predicted proteins (and their respective subunits when relevant) are colour coded to denote the degree of homology among all sequenced Thaumarchaeota: green – Ca. N. ianthellae unique gene families; light blue – shared exclusively among thaumarchaeal sponge symbionts; dark blue – ubiquitously found in Thaumarchaeota. Extracellular proteins derived from the marine environment as well as the sponge mesohyl may be degraded extracellularly or in the thaumarchaeal pseudo‐periplasmic space. These resultant oligo/dipeptides and amino acids which can also be derived from the environment can then be transported by a suite of ABC transporters and major facilitator superfamily (MFS; Newstead, 2015 and references therein) permeases into the cytoplasm to be further degraded by intracellular peptidases or assimilated. Amino acids such as arginine and aspartate can be further degraded to form NH₃/NH₄ ⁺ for assimilation or export to the pseudo‐periplasmic space for ammonia oxidation. Arginine can be decarboxylated by arginine decarboxylase (PdaD) to agmatine which can then be degraded to urea by agmatinase (SpeB). Both proteins are ubiquitously distributed among the Thaumarchaeota. All but one of the branched‐chain amino acid transporter subunits (LivFGHMK operon) are exclusively found among thaumarchaeal sponge symbionts, whilst the periplasmic solute binding subunit (LivK – found to be expressed and denoted by an asterisk) can be found not only among sponge symbionts but also in N. maritimus and members of the genus Nitrosocosmicus (see also Fig. 3). B. Genetic context map depicting examples of colocalized S08A endopeptidases and serine protease inhibitors (serpins). Colour coding as in panel (A), except purple denotes the presence of a tRNA, sites at which gene insertions are common. The asterisk next to serpins colour coded as sponge‐specific (light‐blue) denotes that Ca. N. chungbukensis is the sole non‐sponge symbiont encoding a serpin and belonging to this orthologous group (see Supporting Information Fig. S4B). AspC, aspartate aminotransferase; GDH, glutamate dehydrogenase.

Figure 5

Nitrification activity of the I. basta holobiont during 7‐day incubation experiments at ambient conditions and with added ammonium (25 or 100 μM) and with/without the AOA inhibitor PTIO. A schematic overview of the corresponding experimental setup is given in Supporting Information Fig. S7. A. Depicts net nitrification (as calculated from the addition of net NO₂ ⁻ and NO₃ ⁻ flux) and gross nitrification rates (estimated using the ¹⁵NO₃ ⁻ isotope pool dilution method). B. Net fluxes of all DIN species for the ambient, +25 μM NH₄ ⁺ and +100 μM NH₄ ⁺ treatments. Rates of the different days in the ambient and +25 μM NH4⁺ incubations were averaged as they were not statistically different. C. The relationship between gross nitrification rates of the I. basta holobiont and the carbon fixation rate of sponge clones sampled on day 7. The carbon fixation rates derived from the I. basta holobiont nitrification experiments displayed a positive and significant correlation with gross nitrification rates (R = 0.665, p < 0.005). Error bars in panel C reflect the standard error of the sample mean, where n = 3 for the carbon fixation rates and for all weighted gross nitrification rates in all treatment conditions. PTIO additions started 2 days after the experiment was commenced; hence, fixation of ¹³C‐labelled bicarbonate during the first 2 days was not influenced by inhibition.