Leberecht

Received: 27 April 2021 | Revised: 24 February 2022 | Accepted: 7 March 2022
DOI: 10.1111/gbi.12492
ORIGINAL ARTICLE
Microbially promoted calcite precipitation in the pelagic

redoxcline: Elucidating the formation of the turbid layer
Kerstin M. Leberecht1 | Simon M. Ritter2 | Christian J. Lapp1 | Lukas Klose3 |

Julian Eschenröder | Christian Scholz | Sebastian Kühnel | Wolfgang Stinnesbeck2,4 |
2 2 2
Arnulf Kletzin5 | Margot Isenbeck-Schröter2,4 | Johannes Gescher1
1
Institute of Technical Microbiology,
Hamburg University of Technology, Abstract
Hamburg, Germany
Large bell-shaped calcite formations called “Hells Bells” were discovered underwater
2
Institute of Earth Sciences, Heidelberg
University, Heidelberg, Germany
in the stratified cenote El Zapote on the Yucatán Peninsula, Mexico. Together with
3
Department of Physics & Earth Sciences, these extraordinary speleothems, divers found a white, cloudy turbid layer into which
Jacobs University Bremen, Bremen, some Hells Bells partially extend. Here, we address the central question if the for-
Germany
4 mation of the turbid layer could be based on microbial activity, more specifically, on
Heidelberg Center for the Environment
(HCE), Heidelberg University, Heidelberg, microbially induced calcite precipitation. Metagenomic and metatranscriptomic pro-
Germany
5
filing of the microbial community in the turbid layer, which overlaps with the pelagic
Department of Biology, Microbiology;
Sulfur Biochemistry and Microbial redoxcline in the cenote, revealed chemolithoautotrophic Hydrogenophilales and
Bioenergetics, Technical University of unclassified β-Proteobacteria as the metabolic key players. Bioinformatic and hydro-
Darmstadt, Darmstadt, Germany
geochemical data suggest chemolithoautotrophic oxidation of sulfide to zero-valent
Correspondence sulfur catalyzed by denitrifying organisms due to oxygen deficiency. Incomplete
Johannes Gescher, Institute of Technical
Microbiology, Hamburg University of sulfide oxidation via nitrate reduction and chemolithoautotrophy are both proton-
Technology (TUHH), Kasernenstraße 12, consuming processes, which increase the pH in the redoxcline favoring authigenic cal-
Hamburg 21073, Germany.
Email: [email protected] cite precipitation and may contribute to Hells Bells growth. The observed mechanism
of microbially induced calcite precipitation is potentially applicable to many other
Funding information
Deutsche Forschungsgemeinschaft, stagnant sulfate-rich water bodies.
Grant/Award Number: STI128/28 and
STI128/36; CONACYT-FONCICYT- KEYWORDS
DADC, Grant/Award Number: biogeochemistry, chemolithoautotrophy, microbially induced calcite precipitation, redoxcline,
000000000278227 sulfide oxidation
1 | I NTRO D U C TI O N were discovered underwater by cave divers in the cenote El Zapote

(20°51′13.8″N 87°07′31.2″W; Figure S1) on the Yucatán Península,
Microbially induced calcium carbonate precipitation (MICCP) is an Mexico (cenote, local term for sinkhole; Figure 1). They termed these
example for the potential effects of microbial metabolic activity on extraordinary structures as Hells Bells due to their appearance in
geochemistry (Görgen et al., 2020). The results of this process are darkness and the vicinity to a sulfidic halocline. Hells Bells are char-
widely distributed in nature and can be found in diverse environ- acterized by their bell-or cone-like shape and downward extension
ments (Dupraz et al., 2009). Large bell-shaped calcite structures with a hollow interior reaching sizes of up to ~4 m. Hells Bells are
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in
any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.
© 2022 The Authors. Geobiology published by John Wiley & Sons Ltd.
498 |
wileyonlinelibrary.com/journal/gbi Geobiology. 2022;20:498–517.
|
14724669, 2022, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/gbi.12492 by Nat Prov Indonesia, Wiley Online Library on [29/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
LEBERECHT et al. 499
built up by alternating layers of micritic and sparry or dog-tooth- for example, dripstone formation in caves (Huggett, 2003) or cal-
like calcites with crystal sizes of up to several mm, which indicate cite raft formation at the water surface of subsurface cenotes and
inorganic precipitation (López-Martínez et al., 2020; Ritter et al., caves (e.g. Kovacs et al., 2017). This raises many questions regarding
2019; Stinnesbeck et al., 2018). They apparently grow underwa- the processes leading to the subaqueous formation of Hells Bells.
ter (Stinnesbeck et al., 2018) in an otherwise carbonate dissolving Lately, Hells Bells formation was suggested to initially result from
karstic aquifer environment, where the water masses are gener- CO2-degassing of ascending gas bubbles that accumulate at cave ir-
ally only saturated or even undersaturated with respect to calcite. regularities and further growth of these structures in the carbonate-
In karstic settings, calcite precipitation usually occurs due to CO2- saturated freshwater layer of cenote El Zapote (López-
Martínez
degassing in areas where water is in contact with the atmosphere, et al., 2020). However, this hypothesis is solely based on the
(a) (b)
(c) (d) (e)
S N
10 m
0
host rock
debris mount
winch-profile
organic-material
10 Hells Bells -
speleothems
turbid layer
water depth [m]
fallen tree
20 (f)
30
fresh water
40
di
halocline
ve
r-p
ro
file
50 salt water
F I G U R E 1 Cenote El Zapote hosts extraordinary bell-shaped calcite formations. (a) View on the cenote water surface from above.
(b) Hells Bells speleothems growing down the ceiling and inclined walls of the cenote. They can reach a size of ~4 m. (c) Schematic cross
section of cenote El Zapote. Hells Bells speleothems cover the cave ceiling and wall in a freshwater zone above the anoxic, sulfidic halocline
in water depths of 28–38 m. (d) Technical diver studying Hells Bells speleothems above the turbid layer. (e) A large specimen of a Hells
Bells speleothem that reaches into the white, cloudy turbid layer, the water zone in which Hells Bells grow. (f) Hells Bells speleothems
are characterized by a bell-like downward extension with a hollow interior. They show a horizontal growth pattern in form of horizontal
laminations and swellings. Photographs (b, d–f ) by Valentina Cucchiara (Liquid Jungle Media, Playa del Carmen, Mexico; 2020)
|
500 LEBERECHT et al.
observation of structures in the cave ceiling that were interpreted 2 | M ATE R I A L S A N D M E TH O DS

as bubble trails. Prior to that, Ritter et al. (2019) reported detailed
hydrogeochemical profiles, chemical and optical analyses on Hells 2.1 | Site description
Bells carbonate speleothems and SEM analyses on filtrates of the
turbid layer of El Zapote. They showed that Hells Bells most likely The cenote El Zapote is localized to the southeastern part of Mexico
grow in the lowermost freshwater body as indicated by uranium iso- on the Yucatán Peninsula in the federal state of Quintana Roo
topes of Hells Bells carbonate samples and the comparison of Mg/ (20°51′13.8″N 87°07′31.2″W; Figure S1). The vertical profile of El
Ca ratios of the water body and Hells Bells carbonates. Accordingly, Zapote shows that the cenote has a bottle-like shape with a narrow
the actual growth of Hells Bells is most likely restricted to the pe- shaft that reaches from the water surface to a water depth of 28 m,
lagic redoxcline, a 1–2-m-thick intermediate zone of steep redox which then turns into a dome-like contour (diameter: 60–100 m;
gradients of electron acceptors and reduced chemical species above Figure 1). The cenote has a maximum depth of about 54 m, however,
a sulfidic halocline. The redoxcline overlaps with a distinct milky, in the center there is a debris mound that reaches a water depth of
white horizontal cloud, the turbid layer, into which some Hells Bells about 36 m. Right on the peak of the debris mound, a fallen tree
partly reach or hang completely inside (Figure 1b–e). Based on these was found that shows small Hells Bells grown on the stem surface.
results, a biologically induced authigenic calcite precipitation within A high-resolution 3D model of the cenote walls including the Hells
the turbid layer was hypothesized (Ritter et al., 2019). This would Bells speleothems is available on https://doi.org/10.11588/heidi
imply a chemolithoautotrophic community that promotes basifica- con/1264341 and illustrates the complexity of the underwater cave.
tion via CO2 fixation and the catalysis of a nitrate-dependent incom- For a more detailed description of El Zapote, see Stinnesbeck et al.
plete sulfide oxidation to zero-valent sulfur. The permanent fixation (2018).
of this calcite in Hells Bells over a wide vertical range of ~10 m water
depth might be the result of a variable elevation of the halocline
and with it the redoxcline and thus the zone of calcite precipitation 2.2 | Water sampling and on-site measurements
(Ritter, 2020; Ritter et al., 2019). Also, such a repeated elevation of
the calcite dissolving halocline could explain the observed internal Sampling at the cenote El Zapote was carried out during the period
layered structure of Hells Bells speleothems, which most likely rep- from May 31 to June 6, 2018. The in situ parameters pH (±0.1; relative
resents repeated phases of calcite growth and dissolution. The halo- accuracy ±0.01), EH (±20 mV), dissolved oxygen (±0.1 mg L−1, detec-
cline elevation might vary on both, long and short time scales due tion limit 0.1 mg L−1), electrical conductivity (±0.05% of value), tem-
to changes in sea level over glacial periods over tens of thousands perature (±0.01 °C), and turbidity (±2% of value) were determined
of years, extended droughts over hundreds of years, annual tidal with the multiparameter sonde EXO 1 (Xylem Analytics, Norway). All
variations, seasonal variations of the thickness of the freshwater parameters, including the water depth via pressure measurement,
layer over months, and short-term aquifer responses on heavy pre- were internally logged by the sonde (±0.04 m). Water depths were
cipitation events over days and weeks. Furthermore, age-dating of corrected to the ambient air pressure and to the increasing salinity
Hells Bells specimens from different water depths indicated varying below a water depth of 36 m (Ritter, 2020). Multiple winch-operated
growth rates over time, which might be related to variations of the profiles of in situ parameters were conducted from the surface to
halocline position on centennial–millennial scale (Ritter, 2020). the top of the debris mound. To complete the measurement of the
Nitrate-driven sulfide oxidation was reported before in the con- water column, technical divers carried the EXO 1 sonde with them
text of calcite precipitation in sediments (Bailey et al., 2009; Himmler during sampling. Winch-operated water samples were also retrieved
et al., 2018), in which microorganisms thrive in gradients of sulfide with a polyethylene FreeFlow bottle (HydroBios Kiel). The sampling
as electron donor and nitrate as electron acceptor. Here, precipi- depth of samples taken with the FreeFlow bottle represents the
tation is based on a combined effect of the dissimilatory reduction center of the sampling bottle and was determined by cable length
of nitrate, sulfide oxidation, and anaerobic sulfate-driven methane with a depth counter attached to the winch. Furthermore, the EXO
oxidation. In contrary, the here described ecosystem is comprised of 1 probe was attached to the FreeFlow bottle to obtain the related in
a very different microbiology that catalyzes an incomplete oxidation situ parameters and exact water depth of each water sample. Water
of sulfide coupled to denitrification. samples taken by technical divers were collected by drawing up
In this study, we elucidated how a complex interaction of chem- the water of the desired sampling depth into 140 ml sterile poly-
olithoautotrophic microorganisms catalyzing reactions of the sulfur ethylene Luer-Lock syringes with an attached three-way valve. The
and nitrogen cycle could induce local calcite precipitation within the water depth and in situ parameters of samples taken by technical
pelagic redoxcline forming a distinct turbid layer. Taxonomic and divers were determined with the EXO 1 sonde, which was attached
functional profiling combined with high-resolution in situ parame- to a side-mounted gas bottle and pointed toward the front. Water
ters revealed strong geochemical and biological evidence that micro- samples between 0 and 36 m were taken with the FreeFlow bot-
bial metabolic activity in the planktonic phase promotes authigenic tle, whereas samples taken between 36 and 52 m were collected
calcite precipitation in the redoxcline by increasing the local pH by technical divers. All water samples were treated on-site immedi-
through the depletion of sulfide, nitrate, and carbon dioxide. ately after retrieval. Water samples taken with the FreeFlow bottle
|
were transferred into 140 ml sterile polyethylene Luer-Lock syringes a rate of 0.3 L min−1 by a submersible Geo-Plus pump (COMET,
with an attached three-way valve. From a total sampling volume of Pfaffenschwende, Germany). The water flow was bypassed at the
140 ml, subsamples were taken for the determination of dissolved surface into a flow-through cell for continuous recording of in situ
organic carbon (DOC) and dissolved inorganic carbon (DIC), cations, parameters with the multiparameter EXO 1 sonde (Xylem Analytics,
anions, on-site photometric analyses, and dissolved gases. Samples Norway). Each depth level was pre-pumped for at least 10 min in
for the determination of dissolved ions and for subsequent pho- order to get a minimum twofold exchange of the water volume in the
tometric analysis of ammonium, nitrite, and sulfide were filtered system (pump, tube, and flow-through cell), and water samples were
through a cellulose acetate filter (0.22 µm) and transferred into 15 taken once the in situ values stabilized. The suspended particulate
ml Falcon polypropylene centrifuge tubes. Samples for cation deter- matter in the water was sampled by filtering a defined amount of
mination were acidified with 150 µl HNO3 (32%) to adjust a pH <2. water for each sample (20–60 ml) through syringe insert-t ype poly-
carbonate filters (0.22 µm). The loaded filters were subsequently
flushed with 5–10 ml of distilled water in order to prevent the pre-
2.3 | Water analysis cipitation of salts from the remaining sample solution. Three filters
were taken at each sampling depth, two for the quantification of bulk
Cation concentrations of water samples were determined by optical chemistry and one filter for the investigation with scanning electron
emission spectroscopy with an Agilent 720 ICP-OES. Quality control microscopy (SEM). Immediately after sampling, the loaded filters
was conducted using the reference materials SPS-SW1, SPS-SW2, were put in 50 ml centrifuge tubes (Sarstedt, Nümbrecht, Germany)
and TMDA-70.2. Recovery rates were in the range of 94.5% and for subsequent bulk quantitative element analyses. The SEM filters
102.4% for the analyzed elements. Concentrations of anions Cl−, were placed into polystyrene containers and air-
dried for about
SO2−
4
, and NO−3 were determined via ion chromatography (Dionex one day. Back in the laboratory, 2 ml of aqua regia (analytical grade
ICS-1100) with a recovery rate between 95% and 98% and a relative HNO3 and HCl) was added to the centrifuge tube with the loaded
standard deviation (RSD) of <3% derived from long-term repeated filters and heated for 60 min at 50 °C in order to dissolve the aqua
analysis of reference material SPS-W W1 NUTR. The concentrations regia dissolvable fraction of the particulate matter on the filters. The
of DIC and DOC were determined with a Total Carbon Analyzer polycarbonate filter material is not dissolved as indicated by unde-
(Shimadzu TOC-CPH) with an RSD of <2% derived from repeated tectable or very low metal and trace metal contents of blank filter
analysis of an in-house standard water. Sulfide, ammonium, and ni- and aqua regia controls. The minor contributions of the analyzed
trite were determined on-site within 8 h after sample collection by particulate element content might be attributed to the used aqua
photometric analysis using Merck Spectroquant® test kits (Sulfide regia solution and leaching of polycarbonate filters (Supplementary
Test for 0.020–1.50 mg L−1, Ammonium Test for 0.010–3.00 mg L−1, data, https://doi.org/10.11588/data/GYLDH5). Subsequently, the
−1
and Nitrite Test for 0.007–3.28 mg L ) and a Hach Lange DR200 tubes were filled with Milli-Q water up to a volume of 10 ml. This
Photometer. For quality control, charge balances of all water sam- diluted supernatant solution was then subtracted and analyzed with
ples were calculated and the relative errors were below 5% in all Agilent 720 ICP-OES for Ba, Ca, Fe, Mg, Mn, P, S, Si, and Sr contents.
analyzed samples, usually around 1% in freshwater samples and The SEM filters were coated with carbon for scanning electron (SE)
−
1–5% in saline water samples. HS and HCO3− activities and calcite imaging and energy-
dispersive X-
ray (EDX) analyses. After SEM
and gypsum saturation were calculated with PhreeqC (Parkhurst analysis, two carbon-coated filters were selected and additionally
and Appelo, 1999) version 3.5.0 using phreeqc.dat. Diffusive ion coated with gold for better resolved images of the particles. SE imag-
fluxes (J) toward the redoxcline were calculated from linear regres- ing and EDX analyses were performed with a Leo 440 at 20 kV with
sions with the first Fick's law with diffusion coefficients (taken from an X-Max 80 mm2 detector.
phreeqc.dat; Parkhurst and Appelo, 1999) of DO2, DNO3−, DHCO3−,
and DHS− of 2.1, 1.9, 1.2, and 1.8 * 10 −9 −2 −1
m s , respectively.
2.5 | Microbiome sampling
2.4 | Determination of suspended To characterize the functional microbial diversity throughout the
particulate matter diverse sub-biotopes of the cenote, water samples were collected
(during the sampling campaign from May 31 to June 6, 2018) from
The particulate matter content in water depths around the redox- the respective water zones and depths: freshwater (32 m), central
cline was determined during the field trip from 13 to 25 of February redoxcline (35.8 m), and halocline (38 m). Up to a water depth of
2020. Water samples with a high vertical resolution were obtained 36 m, where the debris mount peaks, the water column was sampled
by a sampling device similar to that described by Jorgensen et al. using a winch-operated 1 L polyethylene FreeFlow bottle (HYDRO-
(1979). This pump-based system allows for near-laminar inflow into BIOS, Kiel, Germany). Sampling of saltwater (below 36 m water
the pump and thus minimizes vertical disturbances, which increases depth) was conducted by technical divers. On site, 200 ml of the
the vertical depth resolution to ~5 cm. The water was pumped at water samples were quickly filtered through a MCE membrane filter
|
with a pore size of 0.22 µm (Merck KGaA, Darmstadt, Germany) to 2.8 | Bioinformatic data analysis
concentrate the microbiome. Subsequently, the filter was preserved
in Allprotect Tissue Reagent/RNAprotect Bacteria Reagent (Qiagen, After sequencing, the 16S rRNA gene amplicon data were pro-
Hilden, Germany) at 4 °C to stabilize DNA/RNA until long time stor- cessed using the Illumina software MiSeq Reporter (MSR) v2.5.1.3
age at −80 °C and further analysis. For 16S rRNA amplicon analysis, and the Illumina Sequence Analysis Viewer (SAV) v1.9.1 for de-
biological duplicates were taken; the metagenomic and metatran- multiplexing and trimming of adapter sequences. The analysis of the
scriptomic data are based on one sample set that was sequenced 16S rRNA amplicon datasets was conducted with the CLC Genomic
with high depth due to limited possibilities to retrieve the samples Workbench Software 12.0 including the “Microbial genomic mod-
with technical divers. ule 4.0” plugin. The amplicon data were quality-
trimmed with a
threshold of 0.05, followed by a primer trimming step and merging
of paired reads. For taxonomic profiling, the OTU clustering was per-
2.6 | DNA/RNA extraction formed against the SILVA 16S database release v132 97%. For the
duplicates of every sample, a mean value was calculated for every
To extract the DNA, the filter membranes were washed once with OTU. Very low abundant OTUs (minimum combined abundance less
1x PBS to remove the Allprotect Tissue Reagent (Qiagen, Hilden, than 50) were removed. The first step in both metagenomic and
Germany). This solution was centrifuged for 5 min at 16,000 g to pel- metatranscriptomic analysis was the quality and adapter trimming
let and recover detached cells. The DNA pellet and the DNA attached of raw reads using the software tool Trimmomatic (v0.38) (Bolger
to the filter membrane were purified using the DNeasy PowerWater et al., 2014) with the following parameters: “LEADING:3 TRAILING:3
Kit (Qiagen, Hilden, Germany) according to the manufacturer's pro- SLIDINGWINDOW:4:15 MINLEN:105.” Next, the metagenome high-
tocol. IMGM Laboratories GmbH (Martinsried, Germany) conducted quality paired-end reads were assembled using Megahit (v1.1.4–2)
the isolation of RNA of water samples by means of the RNeasy with the assembly settings of k-mer × 21–141 and steps of 20 (Li
PowerWater Kit (Qiagen, Hilden, Germany). Due to the very limited et al., 2015, 2016). As a next step, the prodigal software (v2.6.1)
yield of RNA, no rRNA depletion was performed. For RNA, the total was used in -p meta mode to determine the coding sequences (CDS)
concentration and quality were measured with a Qubit 2.0 fluorom- within the assembled metagenome and to translate these into pro-
eter (Life Technologies, Carlsbad, CA, USA) and a Bioanalyzer 2100 tein sequences (Hyatt et al., 2010). For the functional metagenome
(Agilent Technologies, Santa Clara, CA, USA). For genomic DNA, a annotation, two approaches were used in combination. The protein
NanoDrop 2000 Spectrophotometer (Thermo Scientific, Waltham, sequences gained via prodigal were blasted against the Uniprot
MA, USA) and Qubit 2.0 was used. The samples were stored at database (entry version 36, state: 28.05.2019) using DIAMOND
−80 °C until further processing. (v0.9.24) blastp in sensitive metagenome mode allowing only one hit
per query (Buchfink et al., 2015). Additionally, the eggNOG-mapper
(v2.0.0) was used for orthology prediction and functional annota-
2.7 | Sequencing tion via a protein-protein BLAST against the latest KEGG database
(Huerta-Cepas et al., 2017, 2019). If the protein annotation differed
16S rRNA gene amplicon, metagenome, and metatranscrip- between these two databases, the protein and nucleotide sequence
tome sequencing was conducted by IMGM Laboratories GmbH of the respective gene was verified by another BLAST search
(Martinsried, Germany). 16S rRNA amplicons were generated using against the NCBI database. For quantification of gene expression,
the following primer sets: 341F (CCTACGGGNGGCWGCAG)/ the quality-trimmed metatranscriptomic reads were aligned against
805R (GACTACHVGGGTATCTAATCC) for Bacteria and A519F the metagenome CDS as reference index using Kallisto v0.45.0 (Bray
(CAGCMGCCGCGGTAA)/ A906R (CCCGCCAATTCCTTTAAGTTTC) et al., 2016) in default mode. For comparative analysis, the gene ex-
for Archaea (Stahl and Amann, 1991; Klindworth et al., 2013). pression level was calculated in transcripts per million (TPM), which
Amplification and sequencing of bacterial and archaeal 16S rRNA represents a normalization to gene length and sequencing depth.
genes was conducted on an Illumina MiSeq platform (Illumina, San To determine the diversity of the microbial community in the
Diego, USA) with paired-end 250 nt reads (PE250). For sequencing of redoxcline water column based on the metagenomic and metatran-
metagenomic samples, the library preparation was conducted with scriptomic data, the quality-trimmed raw reads were mapped against
the NEBNext Ultra II FS DNA library preparation kit (New England the assembled metagenome using bowtie2 (v2.3.4.3) (Langmead
Biolabs, Frankfurt am Main, Deutschland). The libraries were se- and Salzberg, 2012) in default mode. The mapping result along
quenced on the Illumina NextSeq 500 platform (Illumina, San Diego, with the quantitative information (abundance) was joined with the
USA) with paired-end 150 nt reads (PE150) and a sequencing depth functionally annotated metagenome to include the taxonomic gene
of at minimum 100 million reads. For metatranscriptome sequenc- annotation achieved by the DIAMOND blast against the Uniprot
ing, the TruSeq total RNA Sample Prep Kit (Illumina, San Diego, USA) database. Further, the number of reads was summarized on the
was used for library preparation and sequencing was performed on taxonomic level of order and family. For these two steps, a custom
the Illumina NextSeq 500 platform with single-read 150 nt reads python-based script was written (provided Data S1 in Supporting
(SR150) and a sequencing depth of at minimum 100 million reads. Information, file name “lappleb.py”). To verify the result, another 16S
|
rRNA-based approach was followed by using phyloFlash (v3.3beta1) The pH values were circumneutral in the freshwater and de-
that reconstructs and analyzes the SSU rRNAs genes to improve creased marginally to 6.84 immediately above the increase in tur-
the determination of the phylogenetic composition (Gruber-Vodicka bidity at the water depth of 35.3 m. Within the redoxcline, the pH
et al., 2020). values showed a distinct peak of 6.91 at the center of the layer.
To amend the functional analysis with further taxonomic infor- Below, pH values continuously decreased to minimum values of 6.56
mation, a binning of the contigs was conducted using MetaBAT2 at 42 m water depth. The concentration of dissolved oxygen (DO)
with default configuration after bowtie2 mapping of filtered reads to decreased downwards toward the turbid layer and was below the
the assembly (Kang et al., 2019). The quality of recovered genomes detection limit at the top of the turbid layer (Figure 2). Similarly, the
was validated using CheckM (Table S1) (Parks et al., 2015). The bin- concentrations of nitrate (~30 µmol L−1 at 34 m) and nitrite (~1.5 µmol
ning information was added to the metagenome annotation (based L−1 at 34 m) decreased from above the turbid layer toward its cen-
on unbinned contigs; Tables S2–S5). The metatranscriptomic reads ter. In contrast, the concentrations of ammonium (~0.22 mmol L−1 at
were also mapped against the bins using Kallisto to verify the tran- 38 m), sulfide (~1.1 mmol L−1 at 38 m), and bicarbonate (~7.25 mmol
scriptional activity of taxonomic units (Figure S2). L−1 at 38 m) decreased from below the turbid layer toward its center.
Importantly, the saturation indices (SI) indicated calcite saturation
in the freshwater body with SI values of 0.02–0.04 above the turbid
3 | R E S U LT S layer and calcite undersaturation in the halocline with SI values of
down to −0.32 at ~40 m water depth. A distinct peak toward over-
3.1 | Hydrogeochemistry saturation was identified in the center of the turbid layer with SI
values of up to ~0.11 at 36 m (Figure 2).
Previous studies proposed the redoxcline, which overlaps with the
turbid layer, as the water layer in which calcite is precipitated (Ritter,
2020; Ritter et al., 2019). Here, we conducted a specific study that 3.2 | Particulate matter in the redoxcline
targeted the hydrogeochemical profile of the turbid layer at the time
of microbiome sampling. The stratified water column of the cenote In connection with the central question of whether the turbid layer
comprises a freshwater body (0–36.6 m), a halocline (36.6–46 m), is formed by microbially induced inorganic calcite precipitation and
and a saltwater body (below ~46 m) (Figure 1c). Water temperatures whether the process of nitrate-driven sulfide oxidation plays a role
within the freshwater body and redoxcline are around 25 °C show- in this process, particulate calcium and sulfur were quantified in the
ing only little variation throughout the year (Ritter, 2020). At the top redoxcline. The respective solid phases were analyzed by scanning
of the halocline, a 1.6-m-thick white, cloudy turbid layer in the water electron microscopy (SEM) and energy-dispersive X-ray spectros-
depth of ~35.0–36.6 m is indicated by the peak of turbidity of up to 7 copy (EDX). As the pH is a central controlling factor of calcite pre-
FNU (formazin nephelometric units) in the center of the turbid layer cipitation in the redoxcline, the question arises whether zero-valent
(Figure 2). The turbid layer overlaps with the redoxcline, in which the sulfur (as an intermediate product or end product of an incomplete
redox potential (EH) changes significantly from +400 to −150 mV. oxidation) is indeed present in higher concentrations in the electron
F I G U R E 2 Hydrogeochemistry of cenote El Zapote showing the pelagic redoxcline (gray), the oxic fresh water above and the anoxic
halocline below. Besides the electric conductivity (EC) in mS cm−1, the turbidity in FNU (formazin nephelometric units), the pH and the redox
potential (EH) in mV, the concentrations of O2, NH+4 , NO−3, HS−, NO−2, SO−4, and HCO−3 were determined (data from 2018). SIcalcite was calculated
with PhreeqC using phreeqc.dat. The fluxes J are given in µmol m−2 s−1
|
acceptor-limited turbid layer or rather sulfate (as product of the com- depth) particulate calcium was measured, a strong increase from the
plete oxidation). The reactions leading to these two products would top of the turbid layer to the maximum value of ~28 µmol L−1 was
have opposite effects on the pH. Whereas the incomplete oxidation detected in the center of the turbid layer. The concentration de-
of sulfide to elemental sulfur would result in an increased local pH creased from there on toward the lower end of the turbid layer. The
(Kamp et al., 2006) and, thus, promote calcite precipitation, the pro- same trend could be measured for particulate sulfur with a maximum
duction of sulfate may induce acidification and calcite dissolution. value of ~29 µmol L−1. Both concentration curves coincide very well
The in situ parameters (Figure 2) suggest that no sulfate accumu- with the measured turbidity; also the peaks overlap at a water depth
lated in the turbid layer. The particulate matter in the redoxcline was of ~34.9 m. The highest pH value, however, was detected slightly
sampled in another sampling campaign using an optimized sampling below the maxima in turbidity, particulate calcium, and sulfur. Just
system (see Materials and Methods section 2.4) that allows minimal as predicted from the SI values for calcite (Figure 2), these data point
vertical disturbances during sampling and thus an increased vertical toward the redoxcline as the location for calcite precipitation with
water profile resolution of ~5 cm. Similar to Figure 2, Figure 3 shows the focus point in the center of the turbid layer.
the same pattern for the electrical conductivity and redox potential SEM and EDX analysis of parallel filters showed that the particu-
as described. However, the center of the redoxcline was about 0.8 m late Ca is derived from numerous 4–10-µ m-sized Ca-carbonate parti-
higher (~34.9 m) than in 2018 (~35.7 m; Figure 2). With the imple- cles (Figure 4b). Additionally, minor amounts of silicate particles such
mentation of the optimized water sampling system, the pH profile as diatom remains, and sand or dust grains, were present (Figure 4h).
could be resolved much higher, ranging from about 6.88–7.04 and The Ca-carbonate particles were interpreted as well-crystallized cal-
falling back to ~6.9 with increasing water depth toward the halo- cite minerals showing various rhombic (Figure 4b,f) and aggregations
cline. Regarding turbidity, the values were twice as high as measured of small sparry minerals (Figure 4a, e and g) that often appear to
before in 2018 at the most pronounced point (Figure 2). Ca and S have grown or formed around crystallization seeds (Figure 4e and h).
were by far the most abundant elements found on the filters (100– Pure S-phases or other solid phases containing enough sulfur to ex-
−1
1,100 µg L ), while Mg, Sr, and Ba as minor constituents of carbon- plain the analyzed bulk S contents could not be clearly identified by
ate minerals were also found on every filter (1–200 µg L−1). Fe and EDX analyses. Although every calcite phase contained several per-
Mn showed only trace contents (0.002–8 µg L−1), which could also cent sulfur (Table S7), incorporation of sulfur into calcite does not
be due to incorporation into carbonate minerals or contamination explain the apparent lack of sulfur phases on the SEM-analyzed fil-
as indicated by the high blank values for Mn (Table S6). Si contents ters. Accumulations of smaller ~0.1–0.3-µ m-sized globules were fre-
around 20 µg L−1 indicate that some silicate phases were also pre- quently observed on the filters (Figure 4a–d and g). When analyzed
sent on the filters, which might be detrital. Both particulate calcium by EDX, these spots with globule accumulations generally show ele-
and sulfur were clearly detectable in the redoxcline water filtrates vated amounts of sulfur compared with areas on the filters with no
−1
(Figure 3). While in the adjacent freshwater only < 2 µmol L (~33 m particles (Table S7). The vacuum conditions of the SEM chamber can
EC [mS cm-1] pH Capar [µmol l-1] Spar [µmol l-1]

1 2 3 6.8 6.9 7 0 10 20 30 0 10 20 30
32 32
33 33
water depth [m]
34 34
35 35
36 36
0 250 500 0 0.05 0.1 0 5 10 15 0 5 10 15
EH [mV] O2 [µmol l-1 ] Turbidity [FNU] Turbidity [FNU]
F I G U R E 3 Quantification of particulate sulfur and calcite in the redoxcline of El Zapote. The presence of particulate sulfur and calcium (in
µmol L−1) was analyzed in the redoxcline in the context of the following environmental parameters (data from 2020): the redox potential (EH)
in mV, the electric conductivity (EC) in mS cm−1, the pH, the turbidity in FNU (formazin nephelometric units), and the oxygen concentration
in µmol L−1. The red dashed lines indicate the water depth below which free sulfide was detected. The gray marking represents the turbid
layer
|
cause sublimation of S(0), and the electron beam can rapidly trans- 3.3 | Microbial diversity
form or burn sulfur material (Nims et al., 2019). Since the electron
beam covers an area of at least ~2 µm in diameter, the elevated sulfur We hypothesized that the pelagic turbid layer is a specific reaction
content in areas with globule accumulations is so far the best analyt- compartment, which is stabilized in its sharp gradients by the ac-
ical evidence for the identification of the sulfur-bearing solid phases. tivity of microorganisms. Moreover, microbial activity could set the
The two gold-coated filters revealed that the globule accumulations foundation for calcite precipitation and the formation of the turbid
were enclosed into 2–4-µ m-sized oblate structures (Figure 4a, c and layer. Hence, a 16S rRNA-based community analysis was conducted,
g). The globules were brighter than the large calcite grains nearby, to determine whether the turbid layer is a zone that is simply char-
indicating that the globules consist of an element heavier than Ca acterized by a gradual transition of one community composition into
and C, like S. the other or alternatively, whether it contains unique groups that
F I G U R E 4 SEM images of gold-coated

(left) and carbon-coated (right) filter
samples of the redoxcline of cenote El
Zapote. (a) 2–9-µ m-sized calcitic particles;
black arrows point toward bright 0.1–0.3-
µm-sized globules, which agglomerate
in 2–4-µ m-sized oblate structures most
likely representing intracellular elemental
sulfur. (b) 3–5-µ m-sized calcite crystals
surrounded by many 1–2-µ m-sized
smaller calcite crystals; white arrows
point toward globule agglomerations
similar to (a). (c) 1–4-µ m large partially
coherent calcitic particles of different
shapes; black arrows point toward globule
agglomerations within oblate structures
similar to (a). (d) Differently sized calcite
particles and globule agglomerations
(marked by white arrows). (e) Approx.
30-µ m-long calcitic particle with finely
structured grain boundaries. (f) 8-µ m-
sized rhombic calcite crystal with idio-to
hypidiomorphic surfaces. (g) Approx.
7-µ m-long calcitic particle with mostly
rounded 1–2-µ m calcitic grain edges and
globule agglomerations similar to (a) (black
arrows). (h) Approx. 16-µ m-sized siliceous
shell structure with calcitic particle
growing on its lower edge
|
strongly increase in concentration compared with the adjacent water to the Hydrogenophilaceae (Figure 6). Surprisingly, the family
layers. Based on the 16S rRNA gene amplicons (Tables S8 and S9), Nitrosopumilaceae (belonging to the Thaumarchaeota) was under-
the relative abundance of different microbial phyla (Tables S10 and represented in the metagenome (<2%) but appeared to be the sec-
S11) was determined in the freshwater layer (sampled at 32 m), the ond most abundant in the metatranscriptome (13%) following the
redoxcline/turbid layer (sampled at 35.8 m), and the halocline (sam- Hydrogenophilales/Hydrogenophilaceae (together 35%). The abun-
pled at 38 m). The microbiome detected in the redoxcline differed dance of other families varied only marginally when both data sets
markedly from the surrounding water layers by the dominance of were compared. According to the 16S rRNA gene sequence from the
Proteobacteria (63%), which were significantly less abundant in the amplicon data set, the Hydrogenophilales were not further classifi-
adjacent freshwater layer (30%) and the halocline (17%) (Figure 5). able than to the level of “uncultured Hydrogenophilaceae.” The se-
The abundance of the phylum Actinobacteria was particularly high quence alignment to the SILVA database (v132) revealed an identity
in the freshwater with 49% and decreased from the redoxcline with of 93.7% to “uncultured Hydrogenophilales” and 92.8% Annwoodia
11% to 3% in the halocline. In contrast, Chloroflexi (17%) and es- aquaesulis (of the order Nitrosomonadales) according to the NCBI
pecially Epsilonbacteraeota (32%) were dominant in the halocline, 16S ribosomal RNA database. The abundance and high activity of
but their abundance decreased toward the turbid layer, while they members related to the Hydrogenophilales were further verified by
were absent in the freshwater. Bacteroidetes were almost equally a metagenomic binning. Bin 86 comprised the highest translational
present in both redoxcline (5%) and halocline (4%) and were absent activity (35.4%) and contained the key genes for the major metabolic
in the freshwater. The archaeal diversity was highest in the halocline. processes described in the section below.
Bathyarchaeota (27%), Euryarchaeota (26%), and Thaumarchaeota
(47%) were detected, whereas the freshwater exclusively contained
Thaumarchaeota (100%). In the redoxcline, Thaumarchaeota domi- 3.4 | Metatranscriptome-based metabolic profiling
nated the archaeal composition with 99% relative abundance, while
Bathyarchaeota comprised only 1%. The overlapping redoxcline and Metatranscriptomic and metagenomic analysis was used to address
turbid layer thus represent a microbial transition zone between the the question whether microbial activity could be a central catalyst
freshwater and the halocline that is dominated by Proteobacteria. for subaqueous calcite precipitation by increasing the local pH in the
In agreement with the 16S amplicon sequencing data, the turbid layer. The metagenome assembly of the redoxcline revealed
dominant taxon identified by metagenomic and metatranscrip- 1,068,412 contigs with a total of 849,570,422 bp. A summary of the
tomic analyses was the order Hydrogenophilales belonging to quantitative sequencing results (metatranscriptome and metagen-
the Proteobacteria. Some of the reads were further assigned ome) is given in the Supporting Information, Table S12. The average
Bacteria Archaea
100 100
80 80
Relative abundance [%]
Relative abundance [%]
60 60
40 40
20 20
0 0
e
e
r
er
e
lin
lin
lin
lin
at
at
xc
xc
oc
oc
hw
hw
o
o
al
al
es
es
ed
ed
H
H
Fr
Fr
R
F I G U R E 5 Relative abundance of OTUs of Bacteria (left) and Archaea (right) throughout the water column of the cenote. The microbial
diversity (based on 16S rRNA gene amplicon sequencing) in the planktonic phase was analyzed in the different water layers (freshwater,
redoxcline, and halocline). Depending on the respective sample, 100% = 71–102,000 reads for Bacteria and 24–117,000 reads for Archaea in
OTUs (Table S8–S11)
|
Metatranscriptome Metagenome
F I G U R E 6 Microbial diversity in the water column of the pelagic redoxcline of cenote El Zapote. Relative abundance based on the
taxonomic annotation of metatranscriptomic (left) and metagenomic raw reads (right)
contig size is 795 bp (minimum 200 bp and maximum 312,582 bp). nirK/S (335 TPM), the nitric oxide reductase norB (195 TPM), and
The N50 value was 996 bp. Binning of metagenomic contigs gener- nitrous oxide reductase nosZ (507 TPM). In fact, nosZ was the most
ated 132 bins. Only 3% of all contigs were binned, which represent abundant gene transcript of the denitrification pathway. On the
24% of the assembled nucleotides. Only 27.8% of the metatranscrip- contrary, dissimilatory nitrate reduction to ammonium (DNRA)
tome could be mapped to the binned metagenome but 60% to the and anaerobic ammonium oxidation (anammox) seemed to play a
unbinned metagenome. The metatranscriptome analysis based on minor role. Transcripts encoding the catalytic subunits of two dif-
the binned metagenome revealed many deficits in the elucidation ferent nitrite reductases, nrfA and nirB, involved in the DNRA were
of the central metabolic pathways. Therefore, the data presented only weakly abundant (8 and 11 TPM). Hardly any transcripts were
are based on the unbinned approach, which provided significantly found for anammox-specific genes such as the hydrazine synthase
more holistic information, as a much larger proportion of the me- (hzs) and the hydrazine dehydrogenase (hdh). Apparently, denitrifi-
tatranscriptome was mapped and integrated. For further taxonomic cation and oxidation of ammonium are the main reaction cascades
assignment of these contigs, the binning information was subse- of the nitrogen cycle in the turbid layer.
quently added. Nevertheless, important information like several key Two genes of the sulfur cycle exceeded all other genes in their
genes in the nitrogen cycle was discarded in the binning process. transcript abundance by far (Figures 7 and 9). They encode the cen-
Therefore, this analysis focuses on the annotation of all (including tral component of the periplasmic sulfur oxidation (Sox) system, the
unbinned) contigs. heterodimeric SoxYZ carrier protein. A total of 2,919 and 3,684 TPM
Figure 7 (based on Tables S2–S5) compares the expression levels could be assigned to soxZ and soxY, respectively. In this complex,
(in transcripts per million (TPM)) of genes involved in the nitrogen SoxY covalently binds reduced sulfur species at its C-terminal cys-
and sulfur cycles, photosynthesis, and autotrophy. In case of the ni- teine persulfide residue (Figure 9) (Sauvé et al., 2007). The complete
trogen and sulfur cycles, which seemed to dominate in the turbid Sox system also comprises the proteins SoxB (146 TPM; Figures 7
layer, the abundances of the respective gene transcripts are addi- and 9), SoxXA (372 TPM for soxX, and 767 TPM for soxA), and SoxCD
tionally shown in the context of their metabolic pathways (Figure 8, (soxC, 2 TPM; soxD 0 TPM) (Rother et al., 2001). SoxXA catalyzes the
scheme based on Alvarez et al. (2014) and Rodionov et al. (2005); initial step of the Sox cycle, in which sulfide, thiosulfate, elemental
Figure 9, scheme based on Grabarczyk and Berks (2017)). sulfur, and other inorganic sulfur compounds are added to the SoxY-
Within the nitrogen cycle, the gene transcripts of the ammonium cysteine persulfide. In the next step, SoxB hydrolyzes the terminal
monooxygenase (subunit A, amoA) reached the highest abundance sulfonate group and releases sulfate. SoxCD in turn oxidizes the ter-
(548 TPM) among all marker genes of the nitrogen cycle (Figures 7 minal sulfane sulfur at SoxY to another sulfonate group, which is also
and 8). Transcripts for the hydroxylamine oxidoreductase (hao), the hydrolyzed by SoxB with the release of sulfate and the SoxYZ per-
second enzyme relevant for the complete oxidation of ammonium sulfide (Grabarczyk & Berks, 2017; Rother et al., 2001). Compared
to nitrite, were only low abundant with 4 TPM. A similarly low abun- with soxXA, the respective transcription level for soxB was distinctly
dance (14 TPM) was measured for the nitrite oxidoreductase (nrxA) lower. In line with this result, we could not detect increasing sul-
that catalyzes the oxidation of nitrite to nitrate, the last step in nitri- fate concentrations in the redoxcline (Figure 2). Transcripts of genes
fication. The microbial community also expressed all marker genes encoding the SoxCD protein were almost not detectable. Instead,
involved in the process of denitrification, such as the nitrate reduc- organisms that lack the soxCD genes transfer the sulfane sulfur to
tases narG (48 TPM) and napA (246 TPM), the nitrite reductases sulfur globules to regenerate the carrier protein SoxYZ (Frigaard &
|
F I G U R E 7 Metatranscriptomic profiling of the planktonic microbiome in the redoxcline of cenote El Zapote. Abundance of gene
transcripts (in transcripts per million [TPM]) involved in the N cycle (green), S cycle (orange), photosynthesis (violet), and autotrophy (blue)
Dahl, 2008; Sauvé et al., 2007). Of note, sulfur particles have been periplasmic flavocytochrome c FccAB transfers electrons at the
previously detected in the redoxcline and turbid layer of El Zapote cytochrome c level into electron transport chains (Oh-
oka &
(Ritter et al., 2019). Blankenship, 2004), whereas the membrane-bound sulfide:quinone
Sulfide/H2S cannot be oxidized only by the Sox complex but oxidoreductase (SQR) catalyzes the same reaction, only coupling
also by two other proteins producing polysulfide and/or S0. The sulfide oxidation to the quinone pool (Chen et al., 1994; Griesbeck
|
F I G U R E 8 Gene expression levels DENITRIFICATION

of genes involved in pathways of the
narG 48 nir 335 norB 195 nosZ 507
nitrogen cycle. Metatranscriptomic
NarGHI NirK/S NorBC NosZ
profiling of the microbiome in the
NO3- NO2- NO N 2O N2
planktonic phase of the pelagic redoxcline
NapAB
of El Zapote. Abundance of gene
nxrA 14 napA 246
transcripts (in transcripts per million
NITRIFICATION
[TPM]) of the catalytic subunits of the NxrAB
NirBD NrfAH
respective enzyme is shown in dark blue.
NO2- nirB 11 nrfA 8 Hdh
Scheme based on Alvarez et al. (2014) and hdh 0
Rodionov et al. (2005) hao 4 DNRA ANAMMOX
Hao
NH2 OH NH 4+ N 2H 4
AmoABC Hzs
amoA 548 hzs 0
F I G U R E 9 Abundance levels of soxZ 2919

soxY 3684
gene transcripts involved in pathways
of the sulfur cycle. Metatranscriptomic SoxZY -S-S -
profiling of the microbial community SO42- S 2O 32-
in the planktonic phase of the pelagic 372
soxB 146 SoxB SoxXA soxX
soxA 767
redoxcline of El Zapote. Abundance
-
of gene transcripts (in transcripts per S0 HS
million [TPM]) is shown in dark blue.
Reactions with dashed arrows occur in the S0 ]S 0 ] SoxZY -S-S-SO3- SoxXA SoxZY -S-S-S-SO 3-
cytoplasm, the ones with solid arrows in soxX 372
soxA 767
the periplasm. Sox cycle according to the
revised version by Grabarczyk and Berks SoxCD SoxB soxB 146
(2017) soxC 2
soxD 0 SO42-
SoxZY -S-S-S -
fccA 738
fccB 549
FccAB
HSO3- SO42- n HS - HSn-
SO42- SorAB Sqr
hdrA 34 sorA/B 0 sqr 31
hdrB 139 SoeABC
hdrC 99
soeA/B/C 0
HdrABC
HS - SO32- APS SO42-
DsrAB AprAB Sat
dsrA 560 aprA 565 sat 252
dsrB 298 aprB 265
et al., 2002; Marcia et al., 2009; Oh-

oka & Blankenship, 2004; adenylyl-sulfate reductase (AprAB) to adenosine-p hosphosulfate
Reinartz et al., 1998). The high abundance of fccA/B transcripts (fc- (APS) (Frigaard & Dahl, 2008; Stockdreher et al., 2014; Tanabe
cA/B 738/549 TPM) indicates that this is potentially a further mech- et al., 2019). The final enzyme sulfate adenylyltransferase (Sat)
anism that sustains the observed sharp sulfide gradient. Transcripts completes the formation of sulfate coupled to ATP synthesis by
for sqr were much lower in abundance (31 TPM; Figure 9). Thus, substrate-level phosphorylation (Frigaard & Dahl, 2008). These
FccAB and the Sox complex seem to represent the major sulfide three enzymes are involved in both sulfide oxidation and sul-
sinks in the turbid layer community, with less contribution from SQR. fate reduction (Figure 9). To clarify which direction of these two
Although we could not detect increasing sulfate concentrations metabolic processes prevails, dsrAB sequences of the main key
in the redoxcline and revealed the presence of particulate sulfur, players (see below; Figure 10) were compared against the dsrAB-
it became also evident that cytoplasmic processes will at least to database established by Müller et al. (2015) using the blast al-
some extent lead to a further oxidation of sulfur species toward gorithm. Via this classification system, we determined that the
sulfate. Core of this oxidative reaction cascade are the enzymes dominant dsr-expressing orders Hydrogenophilales, Thiotrichales,
DsrAB, APS, and Sat. The dissimilatory sulfite reductase (rDsrAB) and Nitrosomonadales (Figure 10) most probably express the re-
transfers electrons to acceptors such as NAD+ producing sulfite verse, oxidizing form of DsrAB (Tables S13–S15). Hence, microbes
(Löffler et al., 2020). Sulfite is subsequently converted by the in the planktonic phase of the redoxcline/turbid layer are likely not
|
F I G U R E 1 0 Taxonomic assignment of transcripts of identified key genes. Metatranscriptomic profiling of the microbiome in the
planktonic phase of the pelagic redoxcline of El Zapote. To identify the microbial key players, the determined most relevant genes for the
metabolism in the redoxcline were analyzed taxonomically. Abundances in transcripts per million [TPM]
conducting dissimilatory sulfate reduction. The genes for the het- expressed (<1 TPM). Likewise, the expression of four marker genes
erodisulfide reductase Hdr, which is also proposed to be involved for anoxygenic photosynthesis was—
if at all—
negligible (Imhoff,
in sulfite production (Boughanemi et al., 2016; Koch & Dahl, 2018), 2016) (Figure 7). Thus, light as energy source seems to be limited
were moderately expressed (hdrA/B/C 34/139/99 TPM). Overall, and photosynthesis is not a major process in the turbid layer.
cytoplasmic processes leading to sulfate production seem to Autotrophic microbes can increase the pH in their microenvi-
be active but—considering the fccAB and sox gene expression— ronment, which is relevant for calcite precipitation. This seems to
probably on a lower level compared with periplasmic sulfide ox- be particularly due to the uptake of HCO3− and the consequent
idation. It is important to note that all of these sulfur compound intracellular conversion to CO2 by carbonic anhydrases (Görgen
oxidation processes work in the absence of O2 (most also in its et al., 2020). Bacteria and Archaea can fix carbon dioxide via vari-
presence) and can thus be coupled to nitrate/nitrite reduction. ous pathways, and the individual occurrence correlates with certain
Accordingly, transcripts for oxygen-d ependent enzymes such as niche characteristics (Figure 7). Transcripts of genes encoding the
sulfur oxygenase reductase were not found (sor 0 TPM; not to be phosphoribulokinase (prk 400 TPM) and especially the ribulose-
confused with sorAB). bisphosphate carboxylase (rubisCO; rbcL 935 TPM, rbcS 27 TPM) of
We also addressed the question if the process of photosynthesis the Calvin cycle exceeded the abundance of all marker gene tran-
still plays a role in the center of the turbid layer in the water depth scripts of other autotrophic pathways by far (reverse TCA cycle,
of 35.8 m. The genes of apocytochrome f of the cytochrome b6f 3-hydroxypropionate (3-HP) bicycle, Wood-Ljungdahl pathway, and
complex (petA), the photosystem II P680 reaction center D1 protein 3-hydroxypropionate/4-hydroxybutyrate (3-HP/4-HB) cycle). Next
(psbA) (Mulo et al., 2009), photosystem I P700 chlorophyll a apopro- to the soxYZ gene transcripts, the rbcL mRNA was the second most
tein A1 (psaA)(Fromme, 1996), and the geranylgeranyl diphosphate/ abundant. However, a group of microorganisms also seemed to fix
geranylgeranyl bacteriochlorophyllide reductase (bchP), which is rel- carbon dioxide via the reverse TCA cycle as the key transcripts en-
evant for the synthesis of bacteriochlorophyll (Addlesee & Hunter, coding the 2-oxoglutarate ferredoxin oxidoreductase (korA 74 TPM,
1999; Shi et al., 2005), were apparently not or only marginally korB 65 TPM) and ATP citrate lyase (aclA 7 TPM; aclB 31 TPM) were
|
also detectable. Although transcripts that could be assigned to redoxclines typically occur in water bodies partially isolated from
members of the Thaumarchaeota were highly abundant in the me- the open ocean, for example, marine bays (e.g., Cariaco, Venezuela
tatranscriptome, we could detect only low transcript levels of key (Taylor et al., 2001)), fjords (e.g., Mariager Fjord, Denmark (Zopfi
genes for the 3-HP/4-HB cycle. et al., 2001); Saanich Inlet, Canada (Walsh et al., 2009)), and (semi-)
enclosed, stratified inland seas (Baltic Sea (Grote et al., 2007), Black
Sea (Glaubitz et al., 2010; Jørgensen et al., 1991)). Continental lo-
3.5 | Microbial key players cations include sulfate-
rich freshwater environments (e.g., Lake
Banyoles, Spain (Noguerola et al., 2015); Lake Cadagno, Switzerland
Determination of the potential microbial key players in the redox- (Camacho et al., 2001)) and even Antarctic meromictic lakes (Lake
cline involved in the formation of the turbid layer was based on a tax- Vanda, Antarctica (Schutte et al., 2020)). Oxygen deficiency or com-
onomic assignment of transcripts encoding key enzymes (Figure 10; plete anoxia in deeper water layers either favors the production of
based on Tables S16). The main focus was on the genes of the domi- hydrogen sulfide during microbial decomposition of organic matter
nant pathways such as sulfur oxidation, denitrification, oxidation of by sulfate-reducing bacteria or intensifies that process so that an-
ammonium, and carbon dioxide fixation via the Calvin cycle. oxic zones might expand. In the context of increasing environmen-
Notably, soxY transcripts of the SoxYZ sulfur carrier protein for sulfur tal eutrophication and climate change, which is accompanied by
oxidation were dominantly expressed by the order Hydrogenophilales rising temperatures and enhanced stratification, the occurrence of
with 2472 TPM, followed by unassigned β-Proteobacteria with oxygen-depleted water zones will expand (Diaz & Rosenberg, 2008;
1164 TPM. Also, soxA and soxB were dominantly expressed by Stramma et al., 2008). This creates the prerequisites for an increased
Hydrogenophilales. Periplasmic FccA was mainly expressed by unas- emergence of toxic hydrogen sulfide (Diaz & Rosenberg, 2008). In
signed β-Proteobacteria and the order Hydrogenophilales. Moreover, this regard, it is of particular interest to understand the diverse pro-
most dsrA reads could also be assigned to the Hydrogenophilales, fol- cesses in redoxclines as biogeochemical transition zones and bar-
lowed by the orders Thiotrichales and Nitrosomonadales and were riers for sulfide diffusion toward upper water layers (Walsh et al.,
classified as belonging to the oxidative type of DsrAB (Tables S13– 2009). With this study on the formation of the turbid layer at cenote
S15). Both Nitrosomonadales and Hydrogenophilales also expressed El Zapote, we add a new aspect to the complex biogeochemistry of
aprA and sat. The nitrate reductase catalytic subunit gene napA was al- redoxclines by showing that chemolithoautotrophic organisms pro-
most equally expressed by Hydrogenophilales and Nitrosomonadales. mote authigenic calcite precipitation in an aphotic, pelagic redox-
Besides napA, the order Hydrogenophilales also expressed all other cline via nitrate-dependent oxidation of sulfide. Certainly, we are
denitrification marker genes, such as nir, norB, and nosZ. Surprisingly, aware that the expression of genes does not necessarily correlate
not only the amoA gene for ammonium monooxygenase subunit A with a prevailing enzymatic activity. However, as will be discussed
was mostly expressed by Thaumarchaeota, and more specifically by in the next paragraphs, the combination of transcription-based and
the family Nitrosopumilaceae, but also the nir genes encoding the ni- hydrogeochemical data provided coherent proof for the assumption
trite reductase. The highest proportions of large subunit transcripts of microbially catalyzed depletion of sulfide, nitrate, carbon dioxide,
of the RubisCO (rbcL) were assigned to the orders Hydrogenophilales and oxygen in the redoxcline.
and Nitrosomonadales and to a minor part to Rhodocyclales and The microbial diversity within the cenote differs significantly
Burkholderiales. between the oxic, nitrate-containing freshwater, the hypoxic/anoxic
Concerning the key genes of the denitrification and sulfur oxida- redoxcline and the anoxic, sulfidic halocline, indicating a vertical
tion pathway within the binned metagenome, the highest expression microbial stratification. The turbid layer, which overlaps with the
values were assigned to bin 86, which in turn most likely comprises redoxcline, is strongly dominated by Proteobacteria. As in the fresh-
a population of the order Hydrogenophilales (Tables S2–S5). The water layer, Thaumarchaeota are the dominant archaeal phylum in
Hydrogenophilales dominate the sulfur cycle as indicated by the the redoxcline (Figure 5). The metagenome and metatranscriptome-
high soxY and dsrA expression levels. They also expressed all genes based diversity analysis points out that the members of the order
relevant for the complete reduction in nitrate to dinitrogen and seem Hydrogenophilales are the main key players in the pelagic redoxcline
to fix carbon dioxide via the Calvin cycle. The same applies to the and thus in the turbid layer (Figure 6). The 16S rRNA analysis of the
Nitrosomonadales; however, it is unclear whether the high levels of Hydrogenophilales extracted from the amplicon data revealed a se-
soxY transcripts from unassigned β-Proteobacteria can also be as- quence similarity of 93.7% to “uncultured Hydrogenophilaceae” and
signed to them. In the nitrogen cycle, the Thaumarchaeota seem to of 92.8% to Annwoodia aquaesulis (of the order Nitrosomonadales
be key players as they predominantly express amoA and nir genes. within the β-Proteobacteria) according to the SILVA and NCBI 16S
ribosomal RNA database, respectively. The fact that the identity val-
ues are below the 94% identity cutoff indicates that these sulfur ox-
4 | DISCUSSION idizers from the cenote El Zapote may form a new genus within the
Hydrogenophilaceae family (Schloss & Handelsman, 2005). Future
The formation of a pelagic redoxcline requires a seasonal or per- analyses will provide more detailed information on their phyloge-
manent stratification of the water body. Consequently, pelagic netic assignment and genetic features.
|
The Hydrogenophilales and unassigned β-Proteobacteria were If the oxidant availability is high, sulfur is oxidized further to sulfate,
identified in the metatranscriptome as the dominant sulfur oxidiz- which leads to a calcite dissolution due to a pH decrease.
ers. The high transcription of soxYZ and soxXA, together with the Two reaction equations were established to provide an over-
lack of soxCD reads, indicates that the oxidation of reduced sulfur view of the stoichiometry and energetic aspects (calculated based
species proceeds via elemental sulfur as an intermediate product on Thauer et al., 1977) of the incomplete (Equation 1) vs. complete
to regenerate the carrier protein SoxYZ again (Frigaard & Dahl, nitrate-driven chemolithotrophic oxidation of HS− (Equation 3). Due
2008; Sauvé et al., 2007). Both unassigned β-Proteobacteria and to the pH of ~6.9, not only HS− but also H2S should be present in the
Hydrogenophilales also seem to oxidize sulfide via periplasmic redoxcline; therefore, the same is shown for the oxidation of hydro-
FccAB. Especially the Hydrogenophilales, but also β-Proteobacteria gen sulfide (Equations 2 and 4).
expressed the genes encoding the three cytoplasmic enzymes that
are responsible for the complete oxidation of sulfide to sulfate 1. Incomplete oxidation of HS−/H2S to S0
(dsrAB, aprAB, sat; Figure 10). In the context of calcite precipitation,
this raises the question whether the main product of their meta- 5HS− + 2NO−3 + 7H+ → 5S0 + N2 + 6H2 O (1)
bolic conversion of sulfide is zero-valent sulfur or rather sulfate. In
contrast to the incomplete oxidation of sulfide to zero-valent sul- ΔG0’ = −982 kJ/mol
fur that would increase the local pH (Kamp et al., 2006) and pro- Per mole nitrate: ΔG 0’ = −491 kJ/mol
mote calcite oversaturation, the production of sulfate may result in
acidification and rather counteract calcite precipitation. Based on 5H2 S + 2NO−3 + 2H+ → 5S0 + N2 + 6H2 O (2)
the hydrogeochemical profile (Figure 2), it becomes evident that
sulfate is not the predominant product of sulfide oxidation in the 2. Complete oxidation of HS−/H2S to SO42−
redoxcline as no accumulation or increase in sulfate concentration
was detected. A cryptic sulfur cycle that would obscure sulfate pro- 5HS− + 8NO−3 + 3H+ → 5SO2−
4 + 4N2 + 4H2 O
(3)
duction via the immediate reduction by sulfate-reducing organisms
seems unlikely so far as sat genes were predominantly expressed ΔG0’ = −3,722 kJ/mol
by Hydrogenophilales and Nitrosomonadales instead of sulfate- Per mole nitrate: ΔG 0’ = −465 kJ/mol
reducing bacteria (Figure 10). Above all, the presence of particulate
sulfur, which was quantitatively analyzed in this study in redoxcline 5H2 S + 8NO−3 → 5SO2−
4 + 4N2 + 4H2 O + 2H
+
(4)
water filtrates, emphasizes that the incomplete oxidation of sulfide
to elemental sulfur prevails. It could be clearly demonstrated that Comparing the redox equations 1 and 3, it is evident that the
there is an enhanced occurrence of particulate S in the center of the incomplete oxidation of HS− (Equation 1) leads to a much higher pro-
redoxcline, which most likely results from elemental sulfur formation ton consumption, which would contribute to the observed pH shift.
(Figure 3). The emergence of particulate sulfur might be even higher Likewise, protons are consumed when H2S serves as electron donor
as particles smaller than the filter pore size of 0.22 µm are not cov- (Equation 2). However, this does not apply to the complete oxidation
ered by the data. Regarding potential electron acceptors for sulfide of hydrogen sulfide to sulfate (Equation 4) during which protons are
oxidation, the metatranscriptomic data provide strong evidence that produced instead. This supports the assumption that the incomplete
the Hydrogenophilales couple the oxidation of sulfide to a complete oxidation to S0 is of major significance. Moreover, the incomplete ox-
reduction of nitrate to dinitrogen. This is corroborated by the con- idation of HS− (Equation 1) requires a quarter of the nitrate demand
verging concentration profiles of sulfide and nitrate, which indicate of the complete oxidation to sulfate (Equation 3)—a relevant factor
their simultaneous consumption (Figure 2). If the final main product in the electron acceptor-limited redox regime. Furthermore, the aer-
was elemental sulfur, this would imply that the redoxcline and tur- obic ammonium-oxidizing activity of Thaumarchaeota could lead to
bid layer are oxidant-limited concerning nitrate as electron acceptor. a constant nitrite supply that can also be used by microorganisms. At
The hydrogeochemical data confirm this assumption, as the upward the same time, Thaumarchaeota will deplete the exceedingly limited
flux of HS− with about 7.9 * 10−4 µmol m−2 s−1 is more than one order amount of downwards diffusing oxygen at the top of the redoxcline
of magnitude higher than the downward fluxes of dissolved oxygen and create anoxic conditions in its center. This also directs the me-
with about 9.1 * 10−5 and nitrate with about 3.0 * 10−5 µmol m−2 s−1 tabolism of the Hydrogenophilales to nitrate instead of oxygen re-
(Figure 2). Furthermore, it is very likely that the accumulations of duction (Boden et al., 2017).
small globules within oblate structures found on the filter material Incomplete sulfide oxidation can be thermodynamically fa-
from the redoxcline (Figure 4a, c and g) represent intracellular ele- vorable under electron acceptor-limited conditions. Certainly, the
mental sulfur. Of note, it was established previously that the link be- complete oxidation of sulfide to sulfate (Equation 3) coupled to de-
tween sulfur oxidation, oxidant availability, and calcite precipitation nitrification yields a higher overall Gibbs free energy under standard
can even exist in a single bacterium. Yang et al. (2019) could establish state conditions compared with the partial oxidation to S0 (Equation
for Achromatium that the organism oxidizes sulfide under oxidant 1). However, when the free energy is normalized to the number
limitation first to sulfur which is combined with calcite precipitation. of molecules of nitrate as the limiting substrate in this metabolic
|
pathway, then it turns out that the energy yield is slightly higher for adapted to substrate limitations (Könneke et al., 2005; Martens-
0
the incomplete (ΔG ’ = −491 kJ/mol; Equation 1) compared with Habbena et al., 2009; Walker et al., 2010) and have already been
the complete oxidation of sulfide (ΔG 0’ = −465 kJ/mol; Equation 3). identified in other pelagic redoxclines and oxygen minimum zones
Nevertheless, in the environmental context the availability of the (Berg et al., 2015; Labrenz et al., 2010; Muck et al., 2019). Like the
different substrates and the kinetics of their conversion will be de- Hydrogenophilales, Thaumarchaeota may also contribute to the
cisive for the dominance of one metabolic routine over the other. local pH shift, and finally calcite precipitation, by their autotrophic
A parallel important process in the redoxcline is mainly con- carbon dioxide fixation (Figure 11). They harbor the most efficient
ducted by the Hydrogenophilales and unassigned β-Proteobacteria aerobic autotrophic CO2 fixation pathway (3-HP/4-HB) (Könneke
via the Calvin cycle (Figures 7 and 10). Microbial autotrophic deple- et al., 2014). Hence, the lower expression of genes involved in this
tion of carbon dioxide also favors precipitation of calcite due to a pathway compared with the Calvin Cycle should not be directly cor-
shift in the carbonate balance (Castanier et al., 1999, 2000; Castro- related with the fraction by which these organisms contribute to the
Alonso et al., 2019). In this layer, autotrophic bacteria and archaea overall process of carbon dioxide fixation. The 3-HP/4-HB pathway
strongly depend on the upwards diffusion of carbon dioxide, which is highly efficient, and lower gene expression might still lead to a
is released by bacterial decomposition of organic matter in the sedi- significant proportion of Thaumarchaeota-based primary produc-
ment of the debris mound in the center of the cenote (Figure 1). The tion. Moreover, the Thaumarchaeota will most likely also support
flux of bicarbonate (Figure 2), together with the local accumulation the observed denitrification-dependent sulfide oxidation not only
13
of C bicarbonate (Ritter et al., 2019), points to a dominant micro- be oxygen depletion but also nitrite production.
bial carbon dioxide assimilation in the redoxcline. Combining the hydrogeochemical data with the model of mi-
Thaumarchaeota (family Nitrosopumilaceae) were the second crobial contribution to inorganic calcite precipitation, it can be con-
most dominant group in the metatranscriptome and stand out by cluded that the pH (as a result of microbial metabolic activity) is the
their high expression of amoA that is involved in the oxidation of central controlling factor. Potentially, the pH and SI values are higher
ammonium. At the upper edge of the redoxcline, it is noticeable that on the microscale close to the cells than the ones measured in the re-
oxygen and ammonium, both substrates of Thaumarchaeota, disap- doxcline, but this aspect still needs to be addressed. However, in the
pear at the same water level (Figure 2). Thaumarchaeota are well water column, it is noticeable that the pH peak is actually situated
Thaumarchaeota
O2
NH4+ Aerobic chemolitotrophic

ammonium oxidation NO2- NO3-
O2
REDOXCLINE
HS-
S0
Calcite pH Anaerobic chemolitoautotrophic NO3-
increase N2
precipitation nitrate-driven sulfide oxidation
H+
Corg
CO2
Hydrogenophilales HS-
NH4+
CO2
F I G U R E 1 1 Interactions of microbial key players identified in the pelagic redoxcline of cenote El Zapote. Thaumarchaeota thrive in the
gradients of oxygen and ammonium at the upper edge of the redoxcline, where minimal concentrations of oxygen seem to be available
for the oxidation of ammonium to nitrite. By oxygen depletion, they provide anoxic conditions in the center of the redoxcline, where
chemolithoautotrophic Hydrogenophilales conduct anaerobic nitrate-driven sulfide oxidation. Thereby, Hydrogenophilales depend on the
upwards diffusion of sulfide and carbon dioxide, as well as the downwards diffusion of the electron acceptors nitrate and nitrite, the product
of Thaumarchaeota. Due to oxidant limitations, the oxidation of hydrogen sulfide mainly terminates at the stage of zero-valent sulfur. The
simultaneous consumption of sulfide, nitrate, carbon dioxide, and protons results in a pH shift that creates conditions for authigenic calcite
precipitation and the formation of the turbid layer in the redoxcline of cenote El Zapote
|
slightly lower than the measured maximum values of turbidity and have to address the question how MICCP in the planktonic phase
the occurrence of particulate sulfur and calcium. The periodic move- of the turbid layer of the cenote El Zapote is related to the growth
ment of the halocline of the cenote El Zapote as described by Ritter of the spectacular Hells Bells in more detail. Since sulfur-oxidizing
(2020) might be a plausible explanation. Precipitation (rain) builds bacteria are described to play a key role in many other pelagic
up the freshwater body of the Yucatán karst aquifer and with it that redoxclines, microbially promoted calcite precipitation may also
of the cenote, which leads to a compensatory movement of the un- occur in a similar way in these stratified aquatic systems but has
derlying water layers downwards and to the sides leading to low- yet to be discovered. In association with environmental changes
ering of the halocline and redoxcline. On the contrary, prolonged such as rising temperatures and eutrophication, and thus with
periods without precipitation or droughts lead to a thinner fresh- the increasing occurrence of sulfide-r ich oxygen-d epleted water
water layer and thus to an elevation of the halocline. In addition, zones, it is even more important to better understand the complex
the daily tides and general sea-level variations influence the vertical biogeochemical processes occurring in pelagic redoxclines. Future
positioning of the halocline (Ritter et al., 2019). As the halocline rises, research will show whether MICCP is similarly relevant in other
particles in the redoxcline should also be carried upwards because redoxclines.
of the density gradient toward the halocline. Thereby, their specific
gravity, which describes the density ratio of two substances, plays
a role. Small particles will react and be pushed upwards faster than 5 | CO N C LU S I O N S
larger ones, which will most likely be carried along with a time lag.
In contrast, heavier particles are expected to sink downwards faster With this study on the microbial activity in context of calcite pre-
again during the lowering of the halocline. Small particles, however, cipitation and thus the formation of the turbid layer in the cenote El
will move back more slowly (depending on their specific gravity). Zapote, Mexico, we add a new aspect to the complex biogeochem-
Those small particles, which have not sunk back to the original ini- istry of redoxclines. Based on our metagenomic and metatranscrip-
tial water level by the time of the next halocline elevation, will be tomic profiling of the microbiome, it was shown that microorganisms
pushed up further during the next elevation. As long as the calcite could metabolically promote authigenic calcite precipitation in
particles stay in the growth zone around the positive pH peak in the an aphotic, pelagic redoxcline via chemolithoautotrophic nitrate-
redoxcline, particles will mostly likely keep growing until they reach dependent oxidation of sulfide. The detected presence and accumu-
a certain mass at which they are no longer pushed out of the growth lation of particulate zero-valent sulfur in the redoxcline underlined
layer. Then, these heavy particles sink into the deeper water zones, the model that the oxidation is rather incomplete due to electron
where they may dissolve again due to the lower pH in the halocline. acceptor limitation resulting in zero-valent sulfur as the predominant
In short, the periodic vertical dynamics of the halocline, together metabolic product. The consequent positive pH shift is responsible
with the diversity of particle sizes, may explain the accumulation of for calcite oversaturation, precipitation, and finally the growth of
small particles and the maximum in turbidity just above the pH peak. calcite particles within the redoxcline. The simultaneous generation
Although there are no indications to speculate on controlled biomin- of sulfur and calcite particles transforms the water zone overlapping
eralization of calcite by the microorganisms so far, it is interesting to with the redoxcline into a white, cloudy turbid layer and may further
note that calcite biomineralization was previously also discussed as be related to the formation of Hells Bells underwater speleothems.
a mechanism to influence microbial motility by gravitaxis as calcite Furthermore, microbially induced calcite precipitation as revealed
inclusions could considerably increase the cells buoyancy (Monteil here in cenote El Zapote could potentially play a role in other pelagic
et al., 2021). redoxclines of stagnant, sulfate-rich waters.
In this study, nitrate-driven sulfide oxidation (ND-SO) (Bailey
et al., 2009; Himmler et al., 2018), combined with autotrophy, AC K N OW L E D G M E N T S
was identified as the key biogeochemical process that influences The authors very much appreciated the great support by the own-
the equilibrium between carbonate dissolution and precipitation. ers of Cenote Zapote Ecopark, Rosario Fátima González Alcocer and
The turbidity in the redoxcline does not only result from the pre- Santos Zuñiga Roque, and their team members during the fieldwork.
cipitation of calcite but also from the formation of (most likely We wish to express our thanks to the technical divers Christine
intracellular) sulfur globules. To our knowledge, neither chemo- Loew and Dirk Penzel for their excellent work during the challeng-
lithoautotrophically induced calcite precipitation in aphotic, pe- ing sampling campaigns at cenote El Zapote. Thank you very much,
lagic redoxclines nor the specific role of Hydrogenophilales in this Jerónimo Avilés, for initiating and pushing the investigation of Hells
context, has been described so far. The identified microbial key Bells. Thanks to Vicente Fito and his team for sharing the initial dis-
processes together with the described oscillatory dynamics of the covery if the Hells Bells at cenote El Zapote in 2009. The authors
halocline and thus of the calcite precipitation zone could form the also acknowledge the impressive photographic material on Hells
basis for a speleothem growth regulated by the planktonic phase, Bells provided by Valentina Cucchiara (liquidjunglemedia.com). We
as proposed by Ritter et al. (2019). Indeed, some features of Hells would like to thank Stefan and Silvia Rheinberger for their excel-
Bells such as the large size and shape of the dog-tooth calcite lent support with water analyses and the examination of hydrogeo-
crystals suggest such a mechanism. Therefore, further studies will chemical data. This research has been supported by the Deutsche
|
Forschungsgemeinschaft (grants no. STI128/28 and STI128/36) and Buchfink, B., Xie, C., & Huson, D. H. (2015). Fast and sensitive protein
the CONACYT-FONCICYT-DADC (grant no. 000000000278227). alignment using DIAMOND. Nature Methods, 12(1), 59–60. https://
doi.org/10.1038/nmeth.3176
Open Access funding enabled and organized by Projekt DEAL.
Camacho, A., Erez, J., Chicote, A., Florín, M., Squires, M. M., Lehmann,
C., & Backofen, R. (2001). Microbial microstratification, inorganic
C O N FL I C T O F I N T E R E S T carbon photoassimilation and dark carbon fixation at the chemo-
The authors declare no competing interests. cline of the meromictic Lake Cadagno (Switzerland) and its rele-
vance to the food web. Aquatic Sciences, 63(1), 91–106. https://doi.
org/10.1007/PL00001346
DATA AVA I L A B I L I T Y S TAT E M E N T Castanier, S., Le Métayer- Levrel, G., & Perthuisot, J.- P. (1999). Ca-
The complete hydrogeochemical data sets (from 2018), which are carbonates precipitation and limestone genesis -the microbioge-
the basis of the water profiles (Figure 2), are available on https:// ologist point of view. Sedimentary Geology, 126(1–4), 9–23. https://
doi.org/10.1016/S0037- 0738(99)00028-7
doi.org/10.11588/data/TMYLWS. The raw data from the sampling
Castanier, S., Le Metayer-Levrel, G., & Perthuisot, J.-P. (2000). Bacterial
campaign (in 2020) addressing the quantification of particulate sul- roles in the precipitation of carbonate minerals. In R. E. Riding, & S.
fur and calcite together with the respective in situ parameters during M. Awramik (Eds.), Microbial sediments (pp. 32–39). Springer.
sampling are available via https://doi.org/10.11588/data/GYLDH5. Castro-Alonso, M. J., Montañez-Hernandez, L. E., Sanchez-Muñoz, M.
A., Macias Franco, M. R., Narayanasamy, R., & Balagurusamy, N.
All 16S rRNA amplicon and metagenomic/metatranscriptomic raw
(2019). Microbially induced calcium carbonate precipitation (MICP)
reads included in this study are publicly available through NCBI and its potential in bioconcrete: microbiological and molecular
BioProject PRJNA635010. concepts. Frontiers in Materials, 6, 126. https://doi.org/10.3389/
fmats.2019.00126
Chen, Z.-W., Koh, M., Van Driessche, G., Van Beeumen, J. J., Bartsch, R.
ORCID
G., Meyer, T. E., Cusanovich, M. A., & Mathews, F. S. (1994). The
Kerstin M. Leberecht https://orcid.org/0000-0002-0733-7509
structure of flavocytochrome c sulfide dehydrogenase from a pur-
Simon M. Ritter https://orcid.org/0000-0002-1152-8754 ple phototrophic bacterium. Science, 266(5184), 430–432. https://
Lukas Klose https://orcid.org/0000-0001-7616-4878 doi.org/10.1126/science.7939681
Johannes Gescher https://orcid.org/0000-0002-1625-8810 Diaz, R. J., & Rosenberg, R. (2008). August). Spreading dead zones and
consequences for marine ecosystems. Science, 321, 926–929.
https://doi.org/10.1126/science.1156401
REFERENCES Dupraz, C., Reid, R. P., Braissant, O., Decho, A. W., Norman, R. S., &
Addlesee, H. A., & Hunter, C. N. (1999). Physical mapping and functional as- Visscher, P. T. (2009). Processes of carbonate precipitation in mod-
signment of the geranylgeranyl-bacteriochlorophyll reductase gene, ern microbial mats. Earth-Science Reviews, 96(3), 141–162. https://
bchP, of Rhodobacter sphaeroides. Journal of Bacteriology, 181(23), doi.org/10.1016/j.earscirev.2008.10.005
7248–7255. https://doi.org/10.1128/JB.181.23.7248-7255.1999 Frigaard, N.-U., & Dahl, C. (2008). Sulfur metabolism in phototrophic sul-
Alvarez, L., Bricio, C., Blesa, A., Hidalgo, A., & Berenguer, J. (2014). fur bacteria. In R. Poole (Ed.), Advances in microbial physiology (vol.
Transferable denitrification capability of Thermus thermophilus. 54, pp. 103–200). Academic Press. https://doi.org/10.1016/S0065
Applied and Environmental Microbiology, 80(1), 19–28. https://doi. -2911(08)00002-7
org/10.1128/AEM.02594-13 Fromme, P. (1996). Structure and function of photosystem I. Current
Bailey, J. V., Orphan, V. J., Joye, S. B., & Corsetti, F. A. (2009). Opinion in Structural Biology, 6(4), 473–484. https://doi.
Chemotrophic microbial mats and their potential for preserva- org/10.1016/S0959- 4 40X(96)80112-6
tion in the rock record. Astrobiology, 9(9), 843–859. https://doi. Glaubitz, S., Labrenz, M., Jost, G., & Jürgens, K. (2010). Diversity of ac-
org/10.1089/ast.2008.0314 tive chemolithoautotrophic prokaryotes in the sulfidic zone of a
Berg, C., Vandieken, V., Thamdrup, B., & Jürgens, K. (2015). Significance Black Sea pelagic redoxcline as determined by rRNA-based stable
of archaeal nitrification in hypoxic waters of the Baltic Sea. isotope probing. FEMS Microbiology Ecology, 74(1), 32–41. https://
The ISME Journal, 9(6), 1319–1332. https://doi.org/10.1038/ doi.org/10.1111/j.1574-6941.2010.00944.x
ismej.2014.218 Görgen, S., Benzerara, K., Skouri-Panet, F., Gugger, M., Chauvat, F., &
Boden, R., Hutt, L. P., & Rae, A. W. (2017). Reclassification of Thiobacillus Cassier-Chauvat, C. (2020). The diversity of molecular mechanisms
aquaesulis (Wood & Kelly, 1995) as Annwoodia aquaesulis gen. of carbonate biomineralization by bacteria. Discover Materials, 1(1),
nov., comb. nov., transfer of Thiobacillus (Beijerinck, 1904) from 2. https://doi.org/10.1007/s43939-020-0 0001-9
the Hydrogenophilales to the Nitrosomonadales, proposal of Grabarczyk, D. B., & Berks, B. C. (2017). Intermediates in the Sox sul-
Hydrogenophilalia class. nov. International Journal of Systematic fur oxidation pathway are bound to a sulfane conjugate of the
and Evolutionary Microbiology, 67(5), 1191–1205. https://doi. carrier protein SoxYZ. PLoS One, 12(3), e0173395. https://doi.
org/10.1099/ijsem.0.001927 org/10.1371/journal.pone.0173395
Bolger, A. M., Lohse, M., & Usadel B. (2014). Trimmomatic: a flexible trim- Griesbeck, C., Schütz, M., Schödl, T., Bathe, S., Nausch, L., Mederer,
mer for Illumina sequence data. Bioinformatics, 30(15), 2114–2120. N., Vielreicher, M., & Hauska, G. (2002). Mechanism of sulfide-
https://doi.org/10.1093/bioinformatics/btu170 quinone reductase investigated using site- directed mutagenesis
Boughanemi, S., Lyonnet, J., Infossi, P., Bauzan, M., Kosta, A., Lignon, and sulfur analysis. Biochemistry, 41(39), 11552–11565. https://doi.
S., Giudici-Orticoni, M. T., & Guiral, M. (2016). Microbial oxidative org/10.1021/bi026032b
sulfur metabolism: Biochemical evidence of the membrane-bound Grote, J., Labrenz, M., Pfeiffer, B., Jost, G., & Jürgens, K. (2007).
heterodisulfide reductase-like complex of the bacterium Aquifex Quantitative distributions of Epsilonproteobacteria and a
aeolicus. FEMS Microbiology Letters, 363(15), fnw156. https://doi. Sulfurimonas subgroup in pelagic redoxclines of the central Baltic
org/10.1093/femsle/fnw156 Sea. Applied and Environmental Microbiology, 73(22), 7155–7161.
Bray, N. L., Pimentel, H., Melsted, P., & Pachter, L. (2016). Near-optimal https://doi.org/10.1128/AEM.00466- 07
probabilistic RNA-seq quantification. Nature Biotechnology, 34(5), Gruber-Vodicka, H. R., Seah, B. K. B., & Pruesse, E. (2020). phyloFlash:
525–527. https://doi.org/10.1038/nbt.3519 Rapid small-subunit rRNA profiling and targeted assembly from
|
metagenomes. mSystems, 5(5). https://doi.org/10.1128/msyst aquifer conditions in Hoyo Negro and Ich Balam (Sac Actun Cave
ems.00920-20 System), Quintana Roo, Mexico. Quaternary Science Reviews, 175,
Himmler, T., Smrzka, D., Zwicker, J., Kasten, S., Shapiro, R. S., Bohrmann, 97–111. https://doi.org/10.1016/j.quascirev.2017.09.006
G., & Peckmann, J. (2018). Stromatolites below the photic zone in the Labrenz, M., Sintes, E., Toetzke, F., Zumsteg, A., Herndl, G. J., Seidler,
northern Arabian Sea formed by calcifying chemotrophic microbial M., & Jürgens, K. (2010). Relevance of a crenarchaeotal subcluster
mats. Geology, 46(4), 339–342. https://doi.org/10.1130/G39890.1 related to Candidatus Nitrosopumilus maritimus to ammonia oxida-
Huerta-Cepas, J., Forslund, K., Coelho, L. P., Szklarczyk, D., Jensen, L. J., tion in the suboxic zone of the central Baltic Sea. The ISME Journal,
von Mering, C., & Bork, P. (2017). Fast Genome-Wide Functional 4(12), 1496–1508. https://doi.org/10.1038/ismej.2010.78
Annotation through Orthology Assignment by eggNOG-Mapper. Langmead, B., & Salzberg, S. L. (2012). Fast gapped-read alignment with
Molecular Biology and Evolution, 34(8), 2115–2122. https://doi. Bowtie 2. Nature Methods, 9(4), 357–359. https://doi.org/10.1038/
org/10.1093/molbev/msx148 nmeth.1923
Huerta- C epas, J., Szklarczyk, D., Heller, D., Hernández- Plaza, A., Li, D., Liu, C.-M., Luo R., Sadakane K., & Lam T.-W. (2015). MEGAHIT: an
Forslund, S. K., Cook, H., Mende, D. R., Letunic, I., Rattei, T., ultra-fast single-node solution for large and complex metagenom-
Jensen, L. J., von Mering, C., & Bork, P. (2019). eggNOG 5.0: a ics assembly via succinct de Bruijn graph. Bioinformatics, 31(10),
hierarchical, functionally and phylogenetically annotated orthol- 1674–1676. https://doi.org/10.1093/bioinformatics/btv033
ogy resource based on 5090 organisms and 2502 viruses. Nucleic Li, D., Luo, R., Liu, C.-M., Leung, C.-M., Ting, H.-F., Sadakane, K.,
Acids Research, 47(D1), D309–D314. https://doi.org/10.1093/ Yamashita, H., & Lam, T.-W. (2016). MEGAHIT v1.0: A fast and
nar/gky1085 scalable metagenome assembler driven by advanced methodol-
Huggett, R. (2003). Karst landscapes. In Fundamentals of geomorphology ogies and community practices. Methods, 102, 3–11. https://doi.
(Routledge fundamentals of physical geography) (pp. 386). Routledge. org/10.1016/j.ymeth.2016.02.020
Hyatt, D., Chen, G.-L ., LoCascio, P. F., Land, M. L., Larimer, F. W., & Löffler, M., Feldhues, J., Venceslau, S. S., Kammler, L., Grein, F.,
Hauser, L. J. (2010). Prodigal: prokaryotic gene recognition and Pereira, I. A. C., & Dahl, C. (2020). DsrL mediates electron
translation initiation site identification. BMC Bioinformatics, 11(1). transfer between NADH and rDsrAB in Allochromatium vino-
https://doi.org/10.1186/1471-2105-11-119 sum. Environmental Microbiology, 22(2), 783–795. https://doi.
Imhoff, J. F. (2016). New dimensions in microbial ecology-functional org/10.1111/1462-2920.14899
genes in studies to unravel the biodiversity and role of functional López-Martínez, R., Gázquez, F., Calaforra, J., Audra, P., Bigot, J.,
microbial groups in the environment. Microorganisms, 4(2), 19. Pi Puig, T., Alcántara-Hernández, R., Navarro, Á., Crochet, P.,
https://doi.org/10.3390/microorganisms4020019 Corona Martínez, L., & Daza Brunet, R. (2020). Bubble trail and
Jørgensen, B. B., Fossing, H., Wirsen, C. O., & Jannasch, H. W. (1991). folia in cenote Zapote, Mexico: Petrographic evidence for abi-
Sulfide oxidation in the anoxic Black Sea chemocline. Deep Sea otic precipitation driven by CO2 degassing below the water table.
Research Part A. Oceanographic Research Papers, 38, S1083–S1103. International Journal of Speleology, 49(3), 173–186. https://doi.
https://doi.org/10.1016/S0198- 0149(10)80025-1 org/10.5038/1827-8 06X.49.3.2344
Jorgensen, B. B., Kuenen, J. G., & Cohen, Y. (1979). Microbial trans- Marcia, M., Ermler, U., Peng, G., & Michel, H. (2009). The structure of
formations of sulfur compounds in a stratified lake (Solar Lake, Aquifex aeolicus sulfide: Quinone oxidoreductase, a basis to un-
Sinai). Limnology and Oceanography, 24(5), 799–822. https://doi. derstand sulfide detoxification and respiration. Proceedings of the
org/10.4319/lo.1979.24.5.0799 National Academy of Sciences of the United States of America, 106(24),
Kamp, A., Stief, P., & Schulz-Vogt, H. N. (2006). Anaerobic sulfide oxi- 9625–9630. https://doi.org/10.1073/pnas.0904165106
dation with nitrate by a freshwater Beggiatoa enrichment culture. Martens-Habbena, W., Berube, P. M., Urakawa, H., De La Torre, J. R.,
Applied and Environmental Microbiology, 72(7), 4755–4760. https:// & Stahl, D. A. (2009). Ammonia oxidation kinetics determine niche
doi.org/10.1128/AEM.00163- 06 separation of nitrifying Archaea and Bacteria. Nature, 461(7266),
Kang, D. D., Li, F., Kirton, E., Thomas, A., Egan, R., An, H., & Wang, Z. 976–979. https://doi.org/10.1038/nature 08465
(2019). MetaBAT 2: An adaptive binning algorithm for robust and Monteil, C. L., Benzerara, K., Menguy, N., Bidaud, C. C., Michot-
efficient genome reconstruction from metagenome assemblies. Achdjian, E., Bolzoni, R., Mathon, F. P., Coutaud, M., Alonso, B.,
PeerJ, 7, e7359. https://doi.org/10.7717/peerj.7359 Garau, C., Jézéquel, D., Viollier, E., Ginet, N., Floriani, M., Swaraj,
Klindworth, A., Pruesse, E., Schweer, T., Peplies, J., Quast, C., Horn, M., S., Sachse, M., Busigny, V., Duprat, E., Guyot, F., & Lefevre, C. T.
& Glöckner, F. O. (2013). Evaluation of general 16S ribosomal RNA (2021). Intracellular amorphous Ca- c arbonate and magnetite
gene PCR primers for classical and next-generation sequencing- biomineralization by a magnetotactic bacterium affiliated to the
based diversity studies. Nucleic Acids Research, 41(1), 1–11. https:// Alphaproteobacteria. The ISME Journal, 15(1), 1–18. https://doi.
doi.org/10.1093/nar/gks808 org/10.1038/s41396-020-0 0747-3
Koch, T., & Dahl, C. (2018). A novel bacterial sulfur oxidation pathway Muck, S., De Corte, D., Clifford, E. L., Bayer, B., Herndl, G. J., & Sintes,
provides a new link between the cycles of organic and inorganic E. (2019). Niche differentiation of aerobic and anaerobic ammonia
sulfur compounds. The ISME Journal, 12(10), 2479–2491. https:// oxidizers in a high latitude deep oxygen minimum zone. Frontiers in
doi.org/10.1038/s41396-018-0209-7 Microbiology, 10, 2141. https://doi.org/10.3389/fmicb.2019.02141
Könneke, M., Bernhard, A. E., De La Torre, J. R., Walker, C. B., Waterbury, Müller, A. L., Kjeldsen, K. U., Rattei, T., Pester, M., & Loy, A. (2015).
J. B., & Stahl, D. A. (2005). Isolation of an autotrophic ammonia- Phylogenetic and environmental diversity of DsrAB-type dissim-
oxidizing marine archaeon. Nature, 437(7058), 543–546. https:// ilatory (bi)sulfite reductases. The ISME Journal, 9(5), 1152–1165.
doi.org/10.1038/nature 03911 https://doi.org/10.1038/ismej.2014.208
Könneke, M., Schubert, D. M., Brown, P. C., Hügler, M., Standfest, S., Mulo, P., Sicora, C., & Aro, E.-M. (2009). Cyanobacterial psbA gene fam-
Schwander, T., Schada von Borzyskowski, L., Erb, T. J., Stahl, D. ily: Optimization of oxygenic photosynthesis. Cellular and Molecular
A. & Berg, I. A. (2014). Ammonia-oxidizing archaea use the most Life Sciences, 66(23), 3697–3710. https://doi.org/10.1007/s0001
energy-efficient aerobic pathway for CO2 fixation. Proceedings of 8-0 09-0103-6
the National Academy of Sciences of the United States of America, Nims, C., Cron, B., Wetherington, M., Macalady, J., & Cosmidis, J.
111(22), 8239–8244. https://doi.org/10.1073/pnas.1402028111 (2019). Low frequency Raman Spectroscopy for micron-scale and
Kovacs, S. E., Reinhardt, E. G., Chatters, J. C., Rissolo, D., Schwarcz, H. P., in vivo characterization of elemental sulfur in microbial samples.
Collins, S. V., Kim, S.-T., Nava Blank, A., & Luna Erreguerena, P. (2017). Scientific Reports, 9(1), 7971. https://doi.org/10.1038/s41598-019-
Calcite raft geochemistry as a hydrological proxy for Holocene 44353-6
|
Noguerola, I., Picazo, A., Llirós, M., Camacho, A., & Borrego, C. M. Hells Bells –Unique speleothems from the Yucatán Peninsula,
(2015). Diversity of freshwater Epsilonproteobacteria and dark in- Mexico, generated under highly specific subaquatic conditions.
organic carbon fixation in the sulphidic redoxcline of a meromictic Palaeogeography, Palaeoclimatology, Palaeoecology, 489, 209–229.
karstic lake. FEMS Microbiology Ecology, 91(7), fiv086. https://doi. https://doi.org/10.1016/j.palaeo.2017.10.012
org/10.1093/femsec/fiv086 Stockdreher, Y., Sturm, M., Josten, M., Sahl, H.-G., Dobler, N., Zigann, R., &
Oh- oka, H., & Blankenship, R. E. (2004). Green bacteria: Secondary Dahl, C. (2014). New proteins involved in sulfur trafficking in the cy-
electron donor (Cytochromes). In W. J. Lennarz, & M. Lane (Eds.), toplasm of Allochromatium vinosum. The Journal of Biological Chemistry,
Encyclopedia of biological chemistry (pp. 321–324). Elsevier. 289(18), 12390–12403. https://doi.org/10.1074/jbc.M113.536425
Parks, D. H., Imelfort, M., Skennerton, C. T., Hugenholtz, P., & Tyson, G. W. Stramma, L., Johnson, G. C., Sprintall, J., & Mohrholz, V. (2008). Expanding
(2015). CheckM: Assessing the quality of microbial genomes recov- oxygen-minimum zones in the tropical oceans. Science, 320(5876),
ered from isolates, single cells, and metagenomes. Genome Research, 655–658. https://doi.org/10.1126/science.1153847
25(7), 1043–1055. https://doi.org/10.1101/gr.186072.114 Tanabe, T. S., Leimkühler, S., & Dahl, C. (2019). Chapter Seven -The func-
Parkhurst, D. L., & Appelo, C. A. J. (1999). User's guide to PHREEQC tional diversity of the prokaryotic sulfur carrier protein TusA. In R.
(Version 2): A computer program for speciation, batch-reaction, Poole (Ed.), Advances in microbial physiology (vol. 75, pp. 233–277).
one-Dimensional transport, and inverse geochemical calculations. Academic Press.
Water-Resources Investigations Report, 99(4259). Taylor, G. T., Iabichella, M., Ho, T.-Y., Scranton, M. I., Thunell, R. C.,
Reinartz, M., Tschäpe, J., Brüser, T., Trüper, H. G., & Dahl, C. (1998). Muller-Karger, F., & Varela, R. (2001). Chemoautotrophy in the
Sulfide oxidation in the phototrophic sulfur bacterium Chromatium redox transition zone of the Cariaco Basin: A significant midwater
vinosum. Archives of Microbiology, 170(1), 59–68. https://doi. source of organic carbon production. Limnology and Oceanography,
org/10.1007/s00203 0050615 46(1), 148–163. https://doi.org/10.4319/lo.2001.46.1.0148
Ritter, S. M. (2020). Unravelling the formation of Hells Bells: underwater Thauer, R. K., Jungermann, K., & Decker, K. (1977). Energy conservation
speleothems from the Yucatán Peninsula in Mexico. (Doctoral dis- in chemotrophic anaerobic bacteria. Bacteriological Reviews, 41(1),
sertation, Heidelberg University). https://doi.org/10.11588/ 100–180. https://doi.org/10.1128/br.41.1.100-180.1977
heidok.00027813 Walker, C. B., De La Torre, J. R., Klotz, M. G., Urakawa, H., Pinel, N., Arp,
Ritter, S. M., Isenbeck-Schröter, M., Scholz, C., Keppler, F., Gescher, J., D. J., Brochier-Armanet, C., Chain, P. S., Chan, P. P., Gollabgir, A.,
Klose, L., Schorndorf, N., Avilés Olguín, J., González-González, Hemp, J., Hügler, M., Karr, E. A., Könneke, M., Shin, M., Lawton, T.
A., Stinnesbeck, W. (2019). Subaqueous speleothems (Hells Bells) J., Lowe, T., Martens-Habbena, W., Sayavedra-Soto, L. A., Lang, D.,
formed by the interplay of pelagic redoxcline biogeochemistry and Sievert, S. M., Rosenzweig, A. C., Manning, G., & Stahl, D. A. (2010).
specific hydraulic conditions in the El Zapote sinkhole, Yucatán Nitrosopumilus maritimus genome reveals unique mechanisms for
Peninsula, Mexico. Biogeosciences, 16(11), 2285–2305. https://doi. nitrification and autotrophy in globally distributed marine crenar-
org/10.5194/bg-16-2285-2019 chaea. Proceedings of the National Academy of Sciences of the United
Rodionov, D. A., Dubchak, I. L., Arkin, A. P., Alm, E. J., & Gelfand, M. States of America, 107(19), 8818–8823. https://doi.org/10.1073/
S. (2005). Dissimilatory metabolism of nitrogen oxides in bacte- pnas.0913533107
ria: Comparative reconstruction of transcriptional networks. PLoS Walsh, D. A., Zaikova, E., Howes, C. G., Song, Y. C., Wright, J. J., Tringe, S.
Computational Biology, 1(5), e55. https://doi.org/10.1371/journ G., Tortell, P. D., & Hallam, S. J. (2009). Metagenome of a versatile
al.pcbi.0010055 chemolithoautotroph from expanding oceanic dead zones. Science,
Rother, D., Henrich, H. J., Quentmeier, A., Bardischewsky, F., & 326(5952), 578–582. https://doi.org/10.1126/science.1175309
Friedrich, C. G. (2001). Novel genes of the sox gene cluster, mu- Yang, T., Teske, A., Ambrose, W., Salman-C arvalho, V., Bagnell, R., &
tagenesis of the flavoprotein SoxF, and evidence for a general Nielsen, L. P. (2019). Intracellular calcite and sulfur dynamics of
sulfur-oxidizing system in Paracoccus pantotrophus GB17. Journal Achromatium cells observed in a lab-based enrichment and aerobic
of Bacteriology, 183(15), 4499–4508. https://doi.org/10.1128/ incubation experiment. Antonie van Leeuwenhoek, 112(2), 263–274.
JB.183.15.4499-4508.2001 https://doi.org/10.1007/s10482-018-1153-2
Sauvé, V., Bruno, S., Berks, B. C., & Hemmings, A. M. (2007). The SoxYZ Zopfi, J., Ferdelman, T. G., Jørgensen, B. B., Teske, A., & Thamdrup, B.
complex carries sulfur cycle intermediates on a peptide swinging (2001). Influence of water column dynamics on sulfide oxidation
arm. Journal of Biological Chemistry, 282(32), 23194–23204. https:// and other major biogeochemical processes in the chemocline of
doi.org/10.1074/jbc.M701602200 Mariager Fjord (Denmark). Marine Chemistry, 74(1), 29–51. https://
Schloss, P. D., & Handelsman, J. (2005). Introducing DOTUR, a computer doi.org/10.1016/S0304- 4203(00)00091-8
program for defining operational taxonomic units and estimating
species richness. Applied and Environmental Microbiology, 71(3),
1501–1506. https://doi.org/10.1128/AEM.71.3.1501-1506.2005 S U P P O R T I N G I N FO R M AT I O N
Schutte, C. A., Samarkin, V. A., Peters, B., Madigan, M. T., Bowles, M., Additional supporting information may be found in the online
Morgan-Kiss, R., Casciotti, K., & Joye, S. (2020). Vertical stratifica-
version of the article at the publisher’s website.
tion and stability of biogeochemical processes in the deep saline
waters of Lake Vanda, Antarctica. Limnology and Oceanography,
65(3), 569–581. https://doi.org/10.1002/lno.11327
Shi, T., Bibby, T. S., Jiang, L., Irwin, A. J., & Falkowski, P. G. (2005). Protein How to cite this article: Leberecht, K. M., Ritter, S. M., Lapp, C.
interactions limit the rate of evolution of photosynthetic genes in J., Klose, L., Eschenröder, J., Scholz, C., Kühnel, S., Stinnesbeck,
cyanobacteria. Molecular Biology and Evolution, 22(11), 2179–2189. W., Kletzin, A., Isenbeck-Schröter, M., & Gescher, J. (2022).
https://doi.org/10.1093/molbev/msi216
Microbially promoted calcite precipitation in the pelagic
Stahl, D. A., & Amann, R. (1991). Development and application of nu-
cleic acid probes. Nucleic acid techniques in bacterial systematics, redoxcline: Elucidating the formation of the turbid layer.
205–248. Geobiology, 20, 498–517. https://doi.org/10.1111/gbi.12492
Stinnesbeck, W., Frey, E., Zell, P., Avilés, J., Hering, F., Frank, N., Arps, J.,
Geenen, A., Gescher, J., Isenbeck-Schröter, M., Ritter, S., Stinnesbeck,
S., Núñez, E. A., Dahne, V. F., González, A. G., & Deininger, M. (2018).

Leberecht

Uploaded by

Copyright:

Available Formats

Leberecht

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Leberecht

Uploaded by

Copyright:

Available Formats

Received: 27 April 2021 | Revised: 24 February 2022 | Accepted: 7 March 2022

Microbially promoted calcite precipitation in the pelagic

Kerstin M. Leberecht1 | Simon M. Ritter2 | Christian J. Lapp1 | Lukas Klose3 |

Arnulf Kletzin5 | Margot Isenbeck-Schröter2,4 | Johannes Gescher1

1 | I NTRO D U C TI O N were discovered underwater by cave divers in the cenote El Zapote

(c) (d) (e)

observation of structures in the cave ceiling that were interpreted 2 | M ATE R I A L S A N D M E TH O DS

EC [mS cm-1] pH Capar [µmol l-1] Spar [µmol l-1]

F I G U R E 4 SEM images of gold-coated

Relative abundance [%]

F I G U R E 8 Gene expression levels DENITRIFICATION

F I G U R E 9 Abundance levels of soxZ 2919

et al., 2002; Marcia et al., 2009; Oh-

NH4+ Aerobic chemolitotrophic

You might also like

Leberecht

Uploaded by

Copyright:

Available Formats

Leberecht

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Leberecht

Uploaded by

Copyright:

Available Formats

Received: 27 April 2021 | Revised: 24 February 2022 | Accepted: 7 March 2022

Microbially promoted calcite precipitation in the pelagic

Kerstin M. Leberecht1 | Simon M. Ritter2 | Christian J. Lapp1 | Lukas Klose3 |

Arnulf Kletzin5 | Margot Isenbeck-­Schröter2,4 | Johannes Gescher1

1 | I NTRO D U C TI O N were discovered underwater by cave divers in the cenote El Zapote

(c) (d) (e)

observation of structures in the cave ceiling that were interpreted 2 | M ATE R I A L S A N D M E TH O DS

EC [mS cm-1] pH Capar [µmol l-1] Spar [µmol l-1]

F I G U R E 4 SEM images of gold-­coated

Relative abundance [%]

F I G U R E 8 Gene expression levels DENITRIFICATION

F I G U R E 9 Abundance levels of soxZ 2919

et al., 2002; Marcia et al., 2009; Oh-­

NH4+ Aerobic chemolitotrophic

You might also like

Arnulf Kletzin5 | Margot Isenbeck-Schröter2,4 | Johannes Gescher1

F I G U R E 4 SEM images of gold-coated

et al., 2002; Marcia et al., 2009; Oh-