Porewater concentrations of dissolved oxygen and nutrients
The sampling location and appearance of the microbial mats used in this study in cross section are shown in Fig. 1. Profound changes in dissolved oxygen concentration were observed over the diel cycle because of high rates of oxygenic photosynthesis in the daytime and oxygen-requiring respiration at night (Table 1). Briefly, Layer 1 was characterized by oxygen concentration fluctuations in the range of 200–800 µM. Layers 2 and 3 ranged from 0–1200 µM and 0–200 µM, respectively. Mat Layer 4 (3–4 mm below the surface) may contain some dissolved oxygen near noon on days when there is high solar irradiance but stays anoxic for most hours of most days. Layers 5–7 (4–7 mm from the surface) remain anoxic.
Concentrations of ammonium (Table 1) reveal a pattern of increasing concentration with depth (34–124 µM) through the layers examined here. Nitrate concentrations ranged between 26–33 µM, with low variation across depths. The concentration of phosphate ranged between 3–6 µM, with the highest concentration detected in Layer 1 (0–1 mm from surface) at 5.5 µM.
Analysis of genes and transcripts in mat layers by qPCR and RT-qPCR
Gene-copy number ranges for both DNA and cDNA across all layers for all genes examined are summarized as follows: Bacteria, 104−1010 per g mat and 101−105, per ng nucleic acid; Archaea, 106−108 and 102−104; nifH, 108−1011 and 104−107; archaeal-amoA, 104−105 and 2–3; bacterial-amoA, 104−107 and 3–335; Nitrospira-nxrB, 105−107 and 27–372; nosZ, 103−105 and 2–10; nirS, 105−107 and 33–1941; Planctomycetes-16S rRNA gene and cDNA of transcripts, 104−106 and 6–66 (Fig. 2, S1).
Number of copies of DNA and cDNA genes recovered for Bacteria (A), Archaea (B), nifH (C), Archaeal-amoA (D), Bacterial-amoA (E), Nitrospira-nxrB (F), nosZ (G), nirS (H) and Planctomycetes-16S rRNA gene marker (anammox proxy) (I), per g of microbial mat, quantified by qPCR and RT-qPCR in hypersaline microbial mat profiles from different depths. P-values from Kruskal–Wallis test are overlain on each, and different letters indicate significantly different values for the given gene based on a Conover-Iman test p-value of < = 0.05.
The number of DNA and cDNA copies of 16 S rRNA genes attributable to Bacteria and to Archaea exhibited vertical changes across depths examined (Fig. 2A, B and S1). The highest copy numbers of 16 S rRNA gene markers for Bacteria per g of mat were detected in Layers 2 and 3 (1–3 mm from surface), while the highest transcript numbers per g were detected in Layers 3 and 4 (2–4 mm from surface). The highest number of 16 S rRNA gene and 16 S transcript markers for Archaea were detected in Layers 3–5 (2–5 mm from surface), being one order of magnitude higher than the other layers (Fig. 2B).
The abundance of nifH genes was highest in Layer 2 (1–2 mm from surface). The greatest number of transcribed nifH genes were found in Layers 2 and 3 in the data normalized by mass of mat, while highest number of transcripts were detected in Layer 2 when the data was normalized by ng of cDNA (Fig. 2C; Fig. S1C).
Archaeal amoA genes and transcripts were only detected in Layer 3 (2–3 mm from surface) (Fig. 2D). Layer 3 (2–3 mm) had the highest numbers of gene copies and transcripts of Archaeal-amoA, Bacterial-amoA and Nitrospira-nxrB (DNA: 7.5 × 104, 4.2 × 106 and 1.2 × 107; cDNA: 1.3 × 105, 1.4 × 107 and 1.6 × 107 copies per g mat, respectively; Fig. 2D–F). Moreover, the transcript number of Bacterial-amoA and Nitrospira-nxrB were one order of magnitude higher in Layer 3 than in the other layers.
The number of DNA and cDNA copies of nosZ and nirS genes across all layers ranged from 6.9 × 103 to 2.8 × 105; 1.9 × 105 to 5.3 × 107; copies per g mat, respectively (Fig. 2G, H). For both genes, the highest number of DNA copies was detected in Layer 3 (2.7 × 105; 4.2 × 107; copies per g mat, nosZ and nirS genes, respectively), while the highest number of transcripts were found in Layer 4 (3-4 mm): 2.8 × 105; 5.3 × 107 copies per g mat, nosZ and nirS genes, respectively.
The abundance and transcripts numbers of Planctomycetes-16S rRNA genes in the different layers varied in a range from 3.23 × 104 to 1.81 × 106 per g mat (or 6.23 to 40.5 copies per ng of nucleic acid). The highest numbers of copies were detected in Layer 3 (2–3 mm from surface), while the greatest number of transcripts were detected in Layer 4 (Fig. 2I).
cDNA/DNA ratios across depths in the mat
The cDNA/DNA ratio (a proxy for gene transcription) for all genes quantified by qPCR is shown in Fig. 3. The cDNA/DNA ratio of all genes exhibited significant changes across depths. For all genes examined, the ratio was maximal in either Layer 3 or Layer 4. The cDNA/DNA ratios for the domain-specific genes for Bacteria and Archaea had a more even distribution across depths than the nitrogen cycling genes, with the exception of Nitrospira-nxrB, which was also more evenly distributed. The genes nifH, Archaeal-amoA, and Bacterial-amoA, all had a maximal cDNA/DNA ratio in Layer 3. The amplicons Nitrospira-nxrB, nosZ, nirS, and Planctomycetes 16 S rRNA genes all had a maximal cDNA/DNA ratio in Layer 4. The overall magnitude of the cDNA/DNA ratio was also different between the genes, with nirS and nosZ both exhibiting the highest ratios (over 5 and 7, respectively) and the greatest differences between the peak layer (Layer 4) and the other layers in the mat.
Ratios of cDNA/DNA for Bacterial (A), Archaea (B), nifH (C), Archaeal-amoA (D), Bacterial-amoA (E), Nitrospira-nxrB (F), nosZ (G), nirS (H) and Planctomycetes-16S rRNA gene marker (anammox proxy) (I), per g of microbial mat, quantified by qPCR and RT-qPCR in hypersaline microbial mat profiles from different depths. Different letters indicate significantly different values based on layers for each marker gene based on Conover-Iman tests with a p-value of < = 0.05.
NMDS ordination analysis of genes
NMDS ordination analysis of the cDNA/DNA ratio of genes involved in nitrogen transformations was conducted to examine the relationship of these ratios to each other and to the dissolved oxygen and nutrient concentrations measured in these layers (Fig. 4). nifH and amoA genes were positively correlated (r ≥ 0.60) and moreover, the highest ratios of nifH and amoA genes were detected in Layer 2 and 3 (2–3 mm from surface). Nitrospira-nxrB displayed a strong positive correlation with nirS (r = 0.96), nosZ (r = 0.78), and Planctomycetes-16S rRNA genes (r = 0.96). The highest cDNA/DNA ratios of denitrifying genes (nirS and nosZ) were found in Layer 4 and strong positive correlation was detected between them (r > 0.8). Furthermore, the ratio of Planctomycetes-16S rRNA genes (Fig. 4A) was positively correlated with denitrifying organisms’ genes (r > 0.8, Table 2).
Genes associated with the following nitrogen transformations were examined: nitrogen fixation (nifH), nitrification (Bacterial-amoA, Archaeal-amoA, Nitrospira-nxrB), denitrification (nosZ, nirS) and Planctomycetes-16S rRNA gene marker (anammox proxy). The biotic data was standardized, and a sample resemblance matrix was generated using Bray-Curtis coefficient of similarity. In order to analyze the influence of abiotic variables (porewater nutrient and oxygen concentration) on the patterns of the biotic data, monotonic correlations of the abiotic variables were performed. In the plots, the distance between the samples’ points reflects their relative similarity, according to Bray-Curtis similarity matrices based on cDNA/DNA ratios of nitrogen genes examined. The vectors in panel A represent the cDNA/DNA ratios of nitrogen gene examined. In panel B, the vectors represent the environmental variables.
nifH, Bacterial-amoA and Archaeal-amoA were positively correlated with oxygen concentration (r ≥ 0.22, Table 2), while Nitrospira-nxrB was negatively correlated with oxygen (r = −0.68, Table 2). Denitrification genes (nosZ, nirS) and Planctomycetes-16S rRNA genes were all positively correlated with ammonium (r ≥ 0.5) and orthophosphate (r ≥ 0.13) and negatively correlated with oxygen (r > −0.70).
Metagenome analysis of nitrogen cycling
A total number of 922 324 genes were identified; 1305 of these genes were annotated with KOs that are part of KEGG’s Nitrogen Metabolism pathway (Table S2, S3). A dendrogram based on Bray-Curtis similarities of normalized coverages of all recovered nitrogen metabolism genes is shown in Fig. 5A. Overall, the similarity between the layers was >75%. According to SIMPROF analysis, there was a significant difference in the N-related gene coverages (based on an alpha value of 0.05) between Layers 1-Layer 2, Layer 3, and Layer 4 (p = 0.001) and Layer 2-Layer 3, and Layer 4 (p = 0.001), but not between Layers 3 and Layer 4 (p = 1), where the similarity was >90%.
A Cluster analysis illustrating the similarity of normalized coverages of all recovered nitrogen metabolism genes across the uppers 4 layers examined [(Layer 1 (0–1 mm from surface), Layer 2 (1–2 mm from surface), Layer 3 (2–3 mm from surface), Layer 4 (3–4 mm from surface)]. Red lines show non-significant differences, according to SIMPROF analysis (p > 0.05). B The bar plots show the genes of the metabolic pathways in the nitrogen cycle identified in the mat, according metagenome analysis, with relative coverage of each nitrogen cycling gene across depths examined (Fraction of Depth Integrated Coverage, FDIC). 355 unique genes were recovered from KEGG’s Nitrogen Metabolism pathway: 60 annotated as involved in nitrogen fixation, 15 in assimilatory nitrate reduction, 38 in dissimilatory nitrate reduction to ammonia (DNRA), 52 in hydroxylamine dehydrogenase EC 1.7.2.6, 121 in hydroxylamine reductase, 69 in denitrification pathway. C Values of Nitrogen-focused Coverage per Million (N-CPM). The following enzymes perform nitrogen transformation in the mat: nitrogenase molybdenum-iron protein alpha chain (nifD), nitrogenase iron protein NifH, nitrogenase molybdenum-iron protein beta chain (nifK), hydroxylamine dehydrogenase EC 1.7.2.6 (hao), hydroxylamine reductase (hcp), nitrate reductase/nitrite oxidoreductase, alpha subunit (narG, narZ, nxrA), nitrate reductase/nitrite oxidoreductase, beta subunit (narH, narY, nxrB), nitrate reductase (cytochrome) (napA), nitrate reductase (cytochrome), electron transfer subunit (napB), nitrite reductase (NO-forming) / hydroxylamine reductase (nirS), nitrogenase molybdenum-iron protein beta chain (nirK), nitric oxide reductase subunit B (norB), nitric oxide reductase subunit C (norC), nitrous-oxide reductase (nosZ), nitrate reductase gamma subunit (narI, narV), cytochrome c nitrite reductase small subunit (nrfH), nitrite reductase (cytochrome c-552) (nrfA), ferredoxin-nitrite reductase (nirA), ferredoxin-nitrate reductase (narB), MFS transporter, NNP family, nitrate/nitrite transporter (NRT, nark, nrtP, nasA). D Nitrogen cycling genes recovered in this study and the transformation that they catalyze.
The nitrogen fixation pathway was identified with nifD, nifH, and nifK genes (Fig. 5B, C, Table S4). Of the 60 genes detected in this metabolic pathway 17 genes were annotated as nifD, 22 genes as nifH, and 21 genes as nifK. The normalized coverage of these genes showed a decreasing trend with depth. Layer 1 was characterized by the highest values of Nitrogen-focused coverage per million (N-CPM, see Supplementary Text 1) of nifD, nifH, and nifK genes: 56264.7, 54934.2 and 60059.2, respectively. On average, the three genes involved in nitrogen fixation, nifD, nifH, and nifK, decreased with depth, (2.7-fold from Layer 1 to Layer 4, with a nearly 2-fold difference solely between Layer 1 and Layer 2).
Genes involved in nitrate assimilation, annotated as nirA and narB which code for ferredoxin nitrate reductase, were 3 times as abundant in Layer 1 than Layer 2, but decreased less markedly from Layer 2 to Layers 3 and 4.
Genes for dissimilatory nitrite reduction (nrfA, and nrfH) were 4 and 16 times more abundant in Layer 4 than Layer 1. Similarly, the nitrate/nitrite regulator protein genes narl and narV displayed a nearly inverse pattern, with Layer 1 having the least proportion of genes, a large increase from Layer 1 to Layer 2, and additional increases from Layer 2 to Layers 3 and Layer 4 (Fig. 5B, C, Table S4).
Genes associated with nitrification were very poorly represented in the metagenome. No genes associated with ammonia oxidation (amoA) were detected. Genes associated with nitrite oxidation (nrxA, nrxB) that were detected are so closely related to denitrifier genes (narG, narZ, narH, narY) as to be annotated with the same KEGG KO models (K00370 representing narG, narZ, nxrA; and K00371 representing narH, narY, nxrB).
The following genes involved in denitrification were detected: napA, napB, narG, narZ, narH, narY, narI, narV, nirK, nirS, norB, norC, and nosZ (Fig. 5B, C). The nitrate reduction metabolic pathway was represented by 4 genes encoding the nitrate reductase-nitrite oxidoreductase-alpha subunit (narG, narZ, nxrA genes), 6 genes encoding the nitrate reductase-nitrite oxidoreductase-beta subunit (narH, narY, nxrB genes), 31 genes encoding the nitrate reductase gamma subunit (narI, narV), 5 genes encoding the nitrate reductase -cytochrome electron transfer subunit (napB) and 7 genes encoding the nitrate reductase -cytochrome (napA) (Table S4). The N-CPM of nitrate reductase increased with depth, but with a similar proportion of those genes in Layers 3 and 4. With respect to nitrite reductase (nirk and nirS genes, 2 and 1 genes, respectively), no nirK genes were detected in Layer 1, where the highest N-CPM of nirS was recovered (Fig. 5B). In contrast, Layer 3 had no detected nirS and the highest N-CPM of nirK. Regarding nitric oxide reductase (norB and norC genes, 6 and 1 genes, respectively), the highest normalized coverage of norB was detected in Layer 3, while highest for norC was in Layer 1. Finally, nosZ (6 genes) was detected in all the layers, steadily decreasing in normalized coverage from the top layer to the deepest (Fig. 5B, C; Table S4).
DNRA metabolism was represented by nrfA (26 genes) and nrfH (12 genes), and by narI, narV (31). Layer 1 was characterized by the lowest normalized coverage of narI, narV, nrfA, and nrfH genes (6880.2, 3724.6, and 284.6 N-CPM, respectively), while Layer 3 was characterized by the greatest coverage of narI, narV, nrfA, and nrfH genes (32760.5, 14417.9 and 4504.1, respectively; Fig. 5B, C; Table S4).
Genes for hydroxylamine dehydrogenase EC 1.7.2.6 and hydroxylamine reductase (hao and hcp, respectively) were the most abundant nitrogen metabolism genes in the mat: hao having a cumulative N-CPM of ~150000 and hcp having a cumulative N-CPM of nearly 350,000 across the 4 depths (Fig. 5C). Both genes increased in abundance with depth; hcp increased two-fold between Layer 1 and Layer 2, and more gradually in Layer 3 and Layer 4. Hao exhibited a three-fold increase in relative abundance from Layer 1 to Layer 2 and remained relatively constant through Layer 3 and Layer 4 (Fig. 5B, C; Table S4).
Source: Ecology - nature.com
