Nitrospina bacteria in a rocky intertidal habitat (Quintay Bay, central Chile)

Abstract Nitrospina bacteria are among the most important nitrite oxidizers in coastal and open‐ocean environments, but the relevance of the genus contrasts with the scarceness of information on their ecophysiology and habitat range. Thus far, Nitrospina bacteria have been the only nitrite oxidizers detected at high abundance in Chilean coastal waters. These levels are often higher than at other latitudes. In this study, the abundance of 16S‐rRNA gene transcripts of Nitrospina (hereafter just transcripts) was measured by reverse transcription quantitative PCR in a rocky intertidal gradient and compared with the nearshore counterpart off central Chile (~33°S). Rocky pond transcripts were also compared with the taxonomic composition of the macrobiota and bacterioplankton (by 16S‐rRNA gene‐based T‐RFLP) in the intertidal gradient. Transcripts increased from warmer, saltier, and low‐nitrite ponds in the upper intertidal zone (19.5 ± 1.6°C, 39.0 ± 1.0 psu, 0.98 ± 0.17 μmol/L) toward cooler, less salty, and high‐nitrite ponds (17.8 ± 2.6°C, 37.7 ± 0.82 psu, 1.23 ± 0.21 μmol/L) from middle and low zones. These varied from ~1,000 up to 62,800 transcripts. This increasing trend in the number of transcripts toward the lower zone was positively associated with the Shannon's diversity index for the macrobiota (r = .81, p < .01). Moreover, an important increase in the average number of transcripts was observed in ponds with a greater number of fish in the upper (7,846 transcripts during 2013) and lower zones (62,800 transcripts during 2015). Altogether, intertidal and nearshore transcripts were significantly correlated with nitrite concentrations (r = .804, p ˂ .01); rocky pond transcripts outnumbered nearshore ones by almost two orders of magnitude. In summary, rocky ponds favored both the presence and activity of Nitrospina bacteria that are tolerant to environmental stress. This in turn was positively influenced by the presence of ammonia‐ or urea‐producing macrobiota.


| INTRODUC TI ON
Canonical nitrification is a two-step aerobic reaction that has a key role in marine nitrogen cycling (Ward, 2008). The first reaction is the oxidation of ammonia and/or ammonium into nitrite by ammoniaoxidizing bacteria and archaea. The second reaction is nitrite-fueled and catalyzed by nitrite-oxidizing bacteria (NOB) to produce nitrate.
Recently, one-step complete nitrification (comammox Nitrospira) has been confirmed in laboratory cultures (Daims et al., 2015;van Kessel et al., 2015). This was also suggested for some yet-uncultured Nitrospira-like bacteria through metagenomic approaches (Pinto et al., 2016;Wang et al., 2017). However, there is no research offering sound evidence on the presence of comammox genes in marine samples, and most nitrate production in marine environments is thought to be NOB-driven.
In contrast to ammonia oxidizers, Nitrospina-NOB has received much less attention; hence, there is a general lack of information on their ecophysiological features and habitat amplitude (Daims et al., 2016) outside of coastal and open-ocean environments. In this context, rocky intertidal systems offer an ideal setting to study the physiological versatility of yet-uncultured Nitrospina, because these environments are characterized by strong physical-chemical gradients (temperature, salinity, desiccation) and complex biotic interactions among the organisms living there (e.g., competence, predation, commensalism) (Bertness, Leonard, Levine, Schmidt, & Ingraham, 1999;Bruno, Stachowicz, & Bertness, 2003;Hay et al., 2004). Herein, the 16S-rRNA gene-based transcriptional activity of Nitrospina was studied in a rocky intertidal habitat and compared with its nearshore counterpart and environmental changes in the upper, middle, and low intertidal zones as well as with intertidal biotic conditions.

| Rocky intertidal gradient and sampling
A rocky intertidal habitat in "El Litre" Beach of the Quintay Bay in central Chile (33°11′22″ S -71°41′39″ W) was visited twice at midday in autumn of 2013 (April) and summer of 2015 (January). Samples were collected from three rocky ponds (~1-m 2 area and ≤50 cm depth) within three zones across an intertidal gradient: low (ponds L1 to L3), middle (ponds M1 to M3), and upper zones (ponds U1 to U3) ( Figure 1). The intertidal zones were chosen based on their exposure to air between high and low tides (Stephenson & Stephenson, 1949) so that the low zone was exposed to air at lower tides. The upper zone was submerged only at higher tides, and the middle zone was characterized by an intermediate period of exposure to air and interruptions in the distribution of Ulva sp. and other algae ( Table 1) Duplicate seawater samples (100 ml) for nitrite determinations were collected and filtered using a GFF filtration unit in the field and stored in sterile 125 ml Nalgene bottles inside cooler containers with frozen gel packs during transport to the laboratory (~2 hr in travel distance). They were then stored at −20°C for 3 days until analysis. Seawater samples were thawed and analyzed spectrophotometrically using standard colorimetric analysis (detection limit ~0.01 μmol/L) following the procedure described by Strickland and Parsons (1972). Environmental data were available only for the 2013 sampling.  This was estimated from the number of sub-quadrants within a 0.25-m 2 quadrant that were occupied by an individual alga and expressed as percentage of total sub-quadrants.

| Nearshore stations and sampling
TA B L E 1 Relative abundance (% mean ± SD) of organisms that structured the intertidal macrobiota in the autumn of 2013 instrument equipped with an oxygen sensor (RBR Ltd., Ottawa, Canada).

| Microbial and macrobiota samplings
Triplicate seawater samples (100 ml) were collected from coastal and intertidal sites and then concentrated by filtration on GVWP filters (pore size 0.22 μm, MF-Millipore) using disposable syringes (60 ml) and sterile swinnex filter holders (48 mm

| Extraction of total RNA
The RNA was extracted from microbe-loaded filters using the

| Two-step RT-qPCR assay for Nitrospina bacteria
DNase-treated RNA samples were used to synthetize cDNAs with the ImProm-II Reverse Transcription System (Promega, Madison, WI, USA) and reverse primer NitSSU_286R (5′-CCYCT CAGGCCGGCTA-3′; Levipan et al., 2014). The resulting cDNAs were used as templates for qPCR experiments targeting Nitrospina 16S-rRNA genes using the primer combination NitSSU_130F (5′-GGGTGAGTAWCACGTGAATAA-3′; Levipan et al., 2014) and NitSSU_286R. This primer set is specific for Nitrospina-like bacteria and was modified from the primer pair designed by Mincer et al. (2007). Nitrospina 16S-rRNA gene amplicons were quantified from triplicate samples following procedures previously described (Levipan et al., 2014) regarding the chemistry of qPCR reactions, amplification program, and device used. For quantification, the clone ST180811-50 m11 (access number KF452035) was used to prepare a 6-point standard curve that started from 1 × 10 7 gene copies using 10-fold dilutions (efficiency and r 2 values of 83.7% and .99, respectively). The copy number of the respective clone was calculated by dividing its DNA concentration (in ng/μl) by its mass (in ng) calculated using the following formula: mass = [(n) × (M/N A )] × 10 9 where n is the clone size (vector plus insert, in bp); M, is the average molecular weight of a base pair (660 g/mol); N A , is the Avogadro's number (60,221 × 10 23 bp/mol); and 10 9 is the factor to convert grams to nanograms. Nitrospina amplicons with the expected size were verified by analyzing melting curves and visualized through standard agarose gel electrophoresis.

| Terminal restriction fragment length polymorphism (T-RFLP) analysis of active bacterial communities
The The T-RFLP fragment peak (50-500 bp) data were extracted into excel format with the Peak Scanner Software. We only describe results generated with the MspI enzyme because the two restriction enzymes showed the same pattern.

| Data analysis
The 16S Significant differences of physical-chemical data between intertidal zones were computed by using a one-way-ANOVA followed by the Tukey's HSD post-hoc test (*p < .05).
Fluorescent T-RFLP peaks were assigned to specific OTUs and expressed as a percentage of an individual OTU to the total fluorescence in each sample (i.e., as bacterial abundance percentage).
The cluster analysis of square-root-transformed OTU data was performed using the Jaccard's proximity coefficient and UPGMA clustering algorithm, with a bootstrap value set of 1,000 times. The Jaccard's coefficient was equal to zero when there were no species or OTUs to share among the rocky ponds. The value was 1 when the ponds shared identical compositions in species or OTUs. The thermocline in the deepest coastal station was between 10 and 20 m depth (St2.; Figure 3b). The salinity in coastal stations varied between 34.2 and 34.7 psu (Figure 3a, b, and c). Dissolved oxygen concentrations ranged from 0.09 to 0.33 mg/L in the water column of the three coastal stations (Figure 3a, b, and c), while nitrite concentrations did not exceed the 0.76 μmol/L threshold and showed a slight increase with depth (Figure 3a, b, and c).

| Macrobiota composition and bacterial community structure in the intertidal habitat
The macrobiota composition in the intertidal gradient was domi-  OTUs (i.e., with relative abundances <1%) reached higher counts in the low zone ( Figure 4). In addition, low-frequency OTUs included ribotypes (e.g., OTU 77 and OTU 95) that changed their contribution to bacterial communities along the intertidal gradient. These were rare in the upper zone but abundant in low and middle zones.
Community structure changes based on the clustering analysis indicated a higher similarity between bacterial communities from low and middle zones compared with the upper zone ( Figure S1).

The alpha diversity parameters of bacterial communities indicated
less diversity in the upper and low zones compared with the middle intertidal zone (  OTUs were defined as OTUs with abundances of < and ≥ 1%, respectively. Samples from ponds U1 and M1 were lost during the procedure or biotic parameters, but they did show a positive association with changes in alpha diversity of the macrobiota (Table 2). in co-occurrence with both a differential structure of the bacterial community ( Figure S1) and conspicuous presence of G. laevifrons (Table 1). A second sampling was performed during the summer of 2015, and this suggested that the transcriptional activity of

| Nitrospina 16S-rRNA gene transcripts in rocky intertidal ponds and nearshore stations
Nitrospina increases in ponds with fish compared to ponds where fish were at low abundance or absent independent of the zone along the intertidal gradient (Mann-Whitney test; p ˂ .001, Figure 4b).
Two mechanisms could explain a potentially fish-stimulated response in the number of Nitrospina transcripts. The ammonia excretion from fish can favor the first reaction of nitrification (Eikebrokk & Piedrahita, 1997 Nitrospina transcripts and nitrite concentrations in intertidal and nearshore seawaters ( Figure 6). This suggests the crucial role of Nitrospina in the nitrate pool dynamics depending on catalytic substrate offers.
In fact, there is a recent report on nitrite as the major environmental forcing in modulating the ecological niche of putative Nitrospina-like NOB in the ETNP (Sun, Ji, Jayakumar, & Ward, 2017).
Intertidal areas with strong physical-chemical gradients likely favor versatile microbial groups such as Nitrospina. The wet-dry alternation during tidal cycles might enhance the Nitrospina's activity and generate a rapid protein turnover triggered by stress conditions in the intertidal habitat versus their nearshore counterparts.
Significant correlations were found between Nitrospina transcripts in intertidal-coastal seawaters and potential stressors as salinity (Pearson's r = .74, p < .01) and dissolved oxygen (Pearson's r = .69, p < .01). Interestingly, the primer set used here detects two singlecell-amplified genomes (SAGs) that were studied by Ngugi et al. (2016) in a brine-seawater interface (SCGC AAA799-A02 and SCGC AAA799-C22). These were designated as "Candidatus Nitromaritima" (or clade 1) and show a genomic potential to couple nitrite oxidation at salinities as high as 11.2% (~112 psu). In fact, closely related sequences to Ca. Nitromaritima were numerically dominant in OMZ metagenomes in the ETSP off northern Chile as well as in metatranscriptomes in the subtropical ESP off Concepción (Léniz et al., 2017).
The above-mentioned findings and our results suggest that a large fraction of coastal Nitrospina bacteria can thrive at salinities close or slightly above 4% without drawbacks. This threshold is 40 g NaCl/L and is the optimum salinity reported for the most halophilic nitrifying strain (Koops, Böttcher, Möller, Pommerening-Röser, & Stehr, 1990). Nitrospina bacteria may tolerate wide fluctuations in oxygen concentrations that could vary from anoxia (Garcia-Robledo et al., 2017) to well-oxygenated conditions as reported here. Moreover, Nitrospina bacteria are resilient to solar radiation (Levipan et al., 2016), which is an important environmental driver in the upper intertidal zone. However, this was not measured here.
Therefore, more ecological and biochemical insights into yetuncultured members of the genus Nitrospina are needed, including techniques for generating axenic cultures to unveil their physiological and biogeochemical potential. We propose that rocky intertidal habitats are ideal ecosystems to explore Nitrospina ecophysiology because these areas are directly subjected to multiple climate change parameters. There are also practical logistical reasons to collect data here, for example, access, manipulability, and low-cost for sampling. (Laboratorio de la Escuela Veterinaria, UNAB). We are also grateful to CIMARQ staff for their support during coastal samplings. We especially thank Dr. Erika Poblete for her help during the cruise.
Drone mapping of the study site was kindly provided by Dr. Carlos Romero (Departamento de Geografía, UPLA). We thank two anonymous reviewers for comments that helped improve the manuscript.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.