Physiological and genomic evidence of cysteine degradation and aerobic hydrogen sulfide production in freshwater bacteria

ABSTRACT The sulfur-containing amino acid cysteine is abundant in the environment, including in freshwater lakes. Biological cysteine degradation can result in hydrogen sulfide (H2S), a toxic and ecologically relevant compound that is a central player in biogeochemical cycling in aquatic environments. Here, we investigated the ecological significance of cysteine in oxic freshwater, using isolated cultures, controlled experiments, and multiomics. We screened bacterial isolates enriched from natural lake water for their ability to produce H2S when provided cysteine. We identified 29 isolates (Bacteroidota, Proteobacteria, and Actinobacteria) that produced H2S. To understand the genomic and genetic basis for cysteine degradation and H2S production, we further characterized three isolates using whole-genome sequencing (using a combination of short-read and long-read sequencing) and tracked cysteine and H2S levels over their growth ranges: Stenotrophomonas maltophilia (Gammaproteobacteria), S. bentonitica (Gammaproteobacteria), and Chryseobacterium piscium (Bacteroidota). Cysteine decreased and H2S increased, and all three genomes had genes involved in cysteine degradation. Finally, to assess the presence of these organisms and genes in the environment, we surveyed a 5-year time series of metagenomic data from the same isolation source (Lake Mendota, Madison, WI, USA) and identified their presence throughout the time series. Overall, our study shows that diverse isolated bacterial strains can use cysteine and produce H2S under oxic conditions, and we show evidence using metagenomic data that this process may occur more broadly in natural freshwater lakes. Future considerations of sulfur cycling and biogeochemistry in oxic environments should account for H2S production from the degradation of organosulfur compounds. IMPORTANCE Hydrogen sulfide (H2S), a naturally occurring gas with both biological and abiotic origins, can be toxic to living organisms. In aquatic environments, H2S production typically originates from anoxic (lacking oxygen) environments, such as sediments, or the bottom layers of thermally stratified lakes. However, the degradation of sulfur-containing amino acids such as cysteine, which all cells and life forms rely on, can be a source of ammonia and H2S in the environment. Unlike other approaches for biological H2S production such as dissimilatory sulfate reduction, cysteine degradation can occur in the presence of oxygen. Yet, little is known about how cysteine degradation influences sulfur availability and cycling in freshwater lakes. In our study, we identified diverse bacteria from a freshwater lake that can produce H2S in the presence of O2. Our study highlights the ecological importance of oxic H2S production in natural ecosystems and necessitates a change in our outlook on sulfur biogeochemistry.

I n most natural environments, hydrogen sulfide gas (H 2 S) production is usually attributed to defined groups of bacteria and archaea (1,2) and occurs primarily in anoxic environments. During the process of dissimilatory sulfate reduction, sulfate acts as a terminal electron acceptor and is converted to hydrogen sulfide. However, other pathways for H 2 S production exist, namely assimilatory sulfate reduction, in which H 2 S contributes to cell growth and increased biomass, and the desulfurylation (desulfurization) of sulfur-containing amino acids, such as cysteine, which can lead to production of pyruvate, ammonia, and H 2 S (3). It is believed that assimilatory sulfate reduction contributes to growth but does not release H 2 S from the cell, while dissimilatory sulfate reduction and cysteine degradation can both contribute to growth and release of ecologically relevant nitrogen and sulfur compounds into the ecosystem.
Microbes are responsible for several steps of the assimilatory and dissimilatory pathways, which enable sulfur intermediates (sulfate, thiosulfate, and hydrogen sulfide) to flow in the environment. Sulfur cycling in freshwater ecosystems can be ecologically significant, especially in places where strong redox gradients exist (4). For example, in high arctic lakes, intermediate sulfur compounds (other than sulfate and hydrogen sulfide) are suggested to serve as biogeochemical hubs, because more organisms had genes to transform the intermediates, than for sulfate reduction and H 2 S oxidation (5). Cysteine, a sulfur-containing amino acid, is proposed to be an overlooked source of both carbon (6) and sulfur. Additionally, seston (particles in water comprised both living and non-living organisms) contain organosulfur-containing lipids, which settle into the sediments, and contributes to the sulfur pool, even in highly oligotrophic lakes such as Lake Superior (7). We also note that some of these transformations can occur abiotically as well. For example, measuring sulfur isotopes can reveal if a compound has been formed through biological or abiotic processes, and abiotic transformations of sulfur compounds have been observed in geothermal spring environments (8). Therefore, the sulfur cycle can be complex to parse out because it is driven by both biotic and abiotic factors and has multiple intermediates, for which the details of multiple specific pathways remain to be studied.
In seasonally stratified lakes consisting of oxygenated warm water (epilimnion) floating atop colder anoxic waters (hypolimnion), H 2 S is often abundant in the hypolimnion (9, 10) due to oxygen demand driving terminal electron acceptor depletion. However, an overlooked player in the pool of available H 2 S is the use of organosulfur compounds, such as cysteine, by microbes. Cysteine is required to produce many proteins and is also important for protein structure. It is one of the two amino acids (methionine being the other) that contains a sulfur group; however, the sulfhydryl group on cysteine is more reactive and can lead to H 2 S formation. Methionine can also lead to H 2 S formation, for example, in the human body (11,12).
Like all amino acids, cysteine also contains an amine group that will form ammonia once the molecule is degraded. As such, cysteine degradation (desulfurylation) by microbes leads to H 2 S production. H 2 S is ecologically relevant because it can be toxic to plants and animals. During periods of anoxia, H 2 S can accumulate to levels beyond the threshold for living organisms and can cause massive fish kills (13). Unlike other H 2 S sources, cysteine desulfurylation can occur under oxic conditions (such as in Escherichia coli) (14), thereby expanding the environmental scope of this sulfur pool. Indeed, cysteine can be desulfurylated under oxic conditions in the laboratory, but the natural prevalence of this process in freshwater lakes and other oxic environments remains unknown. We expect that H 2 S production in oxic environments (during the mixed water column periods of the year and throughout the stratified period in the mixed epilimnion) could result from cysteine breakdown by microbes.
In this study, we investigated the prevalence of organosulfur degradation (desulfurylation) in a freshwater lake, using both laboratory and genomic evidence, to advance our understanding of oxic sulfur cycling in aquatic ecosystems (Fig. 1). First, we grew bacterial isolates enriched from Lake Mendota's oxic epilimnion to quantify H 2 S and ammonia production, which informs the potential for organosulfur degradation in an oxygenated aquatic environment. We found 18 isolates producing H 2 S under oxic conditions. We selected three H 2 S-producing isolates for detailed characterization using full-genome sequencing and chemical analyses to track cysteine concentrations and H 2 S accumulation during their growth: Stenotrophomonas maltophilia (Gammaproteobacteria), S. bentonitica (Gammaproteobacteria), and Chryseobacterium piscium (Bacteroidota).

FIG 1
Overview of physiological and genomic methods to characterize hydrogen sulfide production from isolates from freshwater lakes. Isolates were enriched from oxygenated lake water, resulting in about over 60 isolates. Isolates were screened for H 2 S and ammonia production using qualitative H 2 S and ammonia production assays. Isolates that produced H 2 S were selected for 16S rRNA gene sequencing. Then, based on the taxonomic results from 16S rRNA gene sequencing, three distinct isolates were selected for whole-genome sequencing using a combination of short-and long-read sequencing. Genome characterization of functional potential and taxonomic classification was conducted on the isolates. Screening of genes involved in cysteine degradation (Table S4 at

Research Article mSystems
In all three isolates, cysteine decreased and H 2 S increased over their exponential growth curve under oxic conditions. Finally, we contextualized our laboratory results using a time series of metagenomic data from the same isolation source (Lake Mendota, Madison, WI, USA) to study the temporal importance of organosulfur degradation. We found that genes for cysteine desulfurylation were present and abundant throughout the time series, suggesting that the ability to degrade cysteine is well represented in Lake Mendota.

Enrichment cultures of isolates from a temperate freshwater lake
Lake Mendota (43°06′24″ N 89°25′29″ W) is a temperate eutrophic lake in South Central Wisconsin, in Madison, WI, USA. Lake Mendota is part of the North Temperate Lakes Long-Term Ecological Research Network (https://lter.limnology.wisc.edu/about/lakes). Lake Mendota encounters annual stratification and annual seasonal anoxia in the hypolimnion. Lake water was collected on 14 September 2018 from an integrated water sample (0-12 m) from Lake Mendota at the "Deep Hole" station (43°05′54″, 89°24′'28″), where the maximum depth is 23.5 m. The water samples were collected during stratification from the oxygenated epilimnion. The lake water was collected in preacid washed 2-L sampling bottles using a flexible PVC tube and brought back on shore within hours for immediate processing. Serial dilution was performed, and lake bacteria were grown on PCB (plate count media broth) agar media at room temperature (~21°C), in the lab under the light. The PCB media was made of: 1 L water, 5 g/L of yeast extract, 10 g/L of tryptone, and 2 g/L of dextrose/D-glucose. If grown on solid media, 10 g of agar per 1 L media was added. Enrichment resulted in about 60 isolates.

Screening for cysteine degradation into H 2 S and ammonia
Isolates were able to grow on PCB and Reasoner's 2A agar (R2A media). R2A media is a culture medium for bacteria that typically grow in freshwater. It is less "nutrient-rich" than PCB media and, therefore, slightly closer to natural lake water than PCB. For the screening of the isolates for H 2 S production, we grew them on R2A media. Each isolate had two treatments: grown in R2A media without cysteine for the control and grown in R2A media with amended cysteine as the treatment. R2A media consisted of (1 L of water) 0.5 g of casein, 0.5 g of dextrose, 0.5 g of starch, 0.5 g of yeast extract, 0.3 g of K 2 HPO 4 , 0.3 g of sodium pyruvate, 0.25 g of peptone, 0.25 g of beef extract, 0.024 g of MgSO 4 , and autoclave. To make the same media for plates, we added 15 g of agar before autoclaving. For controls, isolates were grown in the media without cysteine amendments. For "treatments, " 2 mM cysteine was added.
To assess the amount of cysteine degradation into H 2 S and ammonia, we screened each of the 60 isolates for H 2 S and/or ammonia accumulation. To test H 2 S production, we grew the strains individually in liquid media, in culture tubes with loose fitting caps, thus letting oxygen in. They were also shaken at room temperature. We used lead acetate test strips (Fisher Scientific, Waltham, Massachusetts, USA) to qualitatively assess H 2 S accumulation in the headspace. A darkening of the strip shows that H 2 S was produced. To test ammonia concentrations after 24 h, we measured samples at time 0 and 24 h using Ammonia Salicylate Reagent Powder Pillows and Ammonia Cyanurate Reagent Powder Pillows (Hatch Reagents) and used spectrophotometry at the 655-nm wavelength.
The three isolates were tested by thioglycolate broth test and were determined to be obligate aerobes.

Identification of H 2 S-producing bacteria using 16S rRNA gene sequencing
Colony PCR and DNA extractions were conducted using the EtNa Crude DNA Extraction and ExoSAP-IT PCR Product Cleanup protocols on the isolates that tested positive for Research Article mSystems producing H 2 S (10). Full-length 16S rRNA gene products were generated for sequencing using universal 16S rRNA primers (27f, 1492r) (15). DNA concentration yields were measured using the QuBit dsDNA HS assay kit (QuBit). DNA was sequenced at the University of Wisconsin-Madison Biotechnology Center (Madison, WI, USA). The program 4Peaks (16) was used to clean the base pairs by quality checking followed by homology search using BLASTn against the NCBI Genbank database (accessed December 2019) (17) to identify the sequences.

Detailed characterization of three H 2 S-producing isolates
We selected three isolates that could aerobically produce H 2 S for further detailed characterization. We selected these isolates because some of the 18 isolates that produced H 2 S when grown with cysteine had identical 16S rRNA sequences; therefore, we chose isolates that had distinct 16S rRNA sequences for full-genome sequencing. Additionally, using 16S rRNA gene sequencing of the isolates, only one was assigned to Stenotrophomonas sp., and we believed that whole-genome sequencing would enable us to get a higher taxonomic confirmation and more complete information. We performed DNA extraction using the PowerSoil Powerlyzer kit (Qiagen) without protocol modifications and sent the genomic DNA for whole-genome sequencing at the Microbial Genome Sequencing Center (MIGS; Pittsburg, PA, USA) with combined short-read Illumina and long-read nanopore sequencing. The data were processed by MIGS to assemble the short reads (Illumina Next Seq 2000) and long reads (Oxford Nanopore Technologies [ONT]) into full genomes. Quality control and adapter trimming were performed with Bcl2fastq (Illumina) and Porechop (https://github.com/rrwick/ Porechop) for Illumina and ONT MinION sequencing, respectively. Hybrid assembly with Illumina and ONT reads was performed with Unicycler (18). Genome annotation of the three isolates was done with Prokka v.1.14.5 (19), using the --rfam setting.
Genome completeness and contamination were estimated using CheckM v.1.1.3 (20) lineage_wf. Taxonomic classification was conducted using GTDB-tk v.0.3.2 (21) with the database release r95. The full-genome taxonomic classification agreed with the prior 16S rRNA gene sequencing results, but we were further able to identify Isolate 43 as S. bentonitica. We ran METABOLIC-G v.4.0 (22) to identify genes associated with cysteine degradation and other metabolic pathways.
Growth measurements of the three isolates were measured using OD 600 with a spectrophotometer, with measurements every 1 h. The isolates were grown in R2A broth media, shaken in an incubator at 27°C. Aliquots were collected over the growth range for cysteine measurements across the growth curves, as described below. A H 2 S microsensor (Unisense) was used to measure H 2 S over time.

Methods to measure cysteine
Cysteine concentrations were measured as cystine, as described by Hotvedt et al. (23) (https://osf.io/9k8a6/). One of the reasons for measuring cystine instead of cysteine is that in oxic environments, cysteine is oxidized rapidly into cystine (24,25). Additionally, unless liquid chromatography-mass spectrometry is used, cysteine can be difficult to measure directly. Samples were diluted 5:4: Pump. Precolumn derivatization was performed using an Agilent 1260 ALS (G1329B) with an injector program. Detection was performed using an Agilent 1260 Infinity II MWD (G7165A) at 338 nm with 10 nm bandwidth. Reference was 390 nm with 20 nm bandwidth. Recovery was tested during method development. Recoveries of cystine ranged from 87.2% to 101.5%, with an average of 92.1%.

Methods to measure H 2 S using a microsensor
Aliquots of at least 1 mL were taken from cultures at desired times after inoculation. We used the Unisense H 2 S microsensor probe (https://unisense.com/products/h2smicrosensor/) following exactly the manufacturer's method for making standards and calibrating. H 2 S concentrations were measured by suspending the H 2 S probe in the aliquot and leaving it in place until the measurement stabilized over 2 min. Because the measurement fluctuates over the course of these minutes, we excluded data gathered while the probe was stabilizing in the sample and averaged the value for each time point.

Isolates capable of H 2 S production in oxic conditions
To answer the question of whether bacteria could produce H 2 S in the presence of oxygen, we grew freshwater isolates in pure culture originally recovered from the water column of temperate eutrophic Lake Mendota. We grew the isolates in moderately rich media (R2A, see Methods) under control and treatment conditions (cysteine addition) and tracked H 2 S production after 24 h ( Fig. S1; Table S1 at https://doi.org/10.6084/ m9.figshare.21711488). Using qualitative H 2 S measurements, we found that 18 isolates produced H 2 S and ammonia when grown in the presence of amended cysteine. We performed 16S rRNA gene sequencing on the 29 isolates that produced H 2 S, regardless of whether they produced ammonia or not when amended with cysteine. Isolates that produced both H 2 S and ammonia were identified as S. rhizophila (Betaproteobacteria), S. maltophilia (Betaproteobacteria), Citrobacter gillenii (Gammaproteobacteria), and Chryseobacterium sp. (Bacteroidota), whereas those producing H 2 S but not ammonia

Detailed microbiological, chemical, and genomic characterization in selected isolates
Next, we selected three isolates (#43, #13, and #66) representing distinct species based on 16S rRNA sequence taxonomy (97% identity), which produced H 2 S for further characterization. These detailed characterizations include OD 600 -based growth rates and paired quantitative measurements of cysteine and H 2 S concentrations in the spent medium. Cysteine addition resulted in concommital H 2 S production over time ( Fig. 2A; Table S2 at https://doi.org/10.6084/m9.figshare.21711509). We also tested which forms of cysteine (L-cysteine or D-cysteine) the isolates used. The Gammaproteobacteria isolates used L-and D-cysteine at similar rates, but C. piscium used D-cysteine at a greater rate than L-cysteine ( Fig. 2B to D; Table S3 at https://doi.org/10.6084/m9.figshare.21711497). Next, we performed whole-genome sequencing using combined short-read and long-read sequencing on these three isolates. Isolate "13-LM-B-02-08" (referred to as #13) had a genome size of 4.18 Mbp, GC content of 66.8%, and was classified to the phylum Proteobacteria, class Gammaproteobacteria, order Xanthomonadales, family Xanthomonadaceae, genus Stenotrophomonas, and species S. maltophilia. Isolate #13 was fully circular and assembled in one scaffold. Isolate 43-LM-B-01-03 (referred to as #43) had a genome size of 4.3 Mbp, GC content of 66.5%, and was classified to the phylum Proteobacteria, class Gammaproteobacteria, and species S. bentonitica. Isolate #43 was assembled in two scaffolds. Finally, isolate LM_BA_5.2 (referred to as #66) had the smallest genome at 1.37 Mbp, GC content of 33.7%, and was classified to the phylum Bacteroidota, class Bacteroidia, order Flavobacteriales, family Weeksellaceae, genus Chryseobacterium, and species C. piscium. Isolate #66 was assembled in seven contigs. All three genomes were estimated to have ~100% genome completeness according to GTDB-tk.
We performed whole-genome sequencing because functional information such as gene content cannot be predicted reliably from 16S rRNA gene sequencing alone. The whole genome of isolate #43 was assembled into a single circular genome and taxonomically assigned to S. maltophilia. Unlike the 16S rRNA sequence that assigned it to S. rhizophila, the full genome was actually closer to S. maltophilia. The whole genome of isolate #13 could be assembled into two long contigs and was taxonomically assigned to S. bentonitica. The Chryseobacter genome was assembled into one circular genome and assigned to C. piscium. All three genomes were estimated to be 100% complete based on CheckM. Overall, the 16S rRNA gene amplicon sequencing performed prior agreed with full-genome sequencing assignment in some cases, and in others, the whole-genome sequencing assignment allowed finer taxonomic resolution (such as in the case of isolate #13).
Overall, whole-genome sequencing provided more information about the isolates' metabolic potential. The genomic content was then used to inform how or why H 2 S might be produced in oxic environments, as shown in the laboratory experiments. We used gene annotations of the three isolates to infer the presence of genes involved in cysteine metabolism, namely those involved in cysteine degradation to ammonia, pyruvate, and H 2 S: metC, malY, tnaA, cysM, cysK (which involve the use of L-cysteine as the substrate), and dcyD (which involves the use of D-cysteine as the substrate) (Table S4 at https://doi.org/10.6084/m9.figshare.21711491). However, we note that these genes may have other enzymatic activities, such as cysteine biosynthesis instead of degradation ( Fig. 3; Table S5 at https://doi.org/10.6084/m9.figshare.21711494).
Leveraging the full genomic content of the three isolates (Table S6 at https://doi.org/ 10.6084/m9.figshare.21711500), we proposed a joint cellular map based on identified metabolic functions and pathways in the genomes (Fig. 4A). All three isolates contained pathways associated with central carbon metabolism: including the TCA cycle, glycolysis, gluconeogenesis, the pentose phosphate pathway, and the glyoxylate cycle. They could all generate fatty acids using fatty acid biosynthesis and catabolize fatty acids through the beta-oxidation pathway. As expected, they had genes for cysteine metabolism, including metC, malY, tnaA, cysM, cysK, and dcyD, cysteine biosynthesis pathways from homocysteine and serine, as well as pathways for degradation of other amino acids, including methionine.  (B, D). D-cysteine also decreased over time in all samples except the negative control; however, the net amount decreased was less compared with L-cysteine. Cysteine concentrations were measured as cystine as described in the Methods section. Because Gammaproteobacteria isolates (#43 and #13) were assessed in a different experimental run than the Bacteroidota isolate (#66) but using the exact same instruments and methods, plots B, C, and D are separated by HPLC runs. Due to large sample volume, it was not possible to test all isolates and conditions in one HPLC run.

Research Article mSystems
Despite these similarities, the three isolates also have distinguishing characteristics among them (Table S6 at https://doi.org/10.6084/m9.figshare.21711500; Fig. 4B). For example, while all isolates encoded genes for sulfur oxidation (sulfur dioxygenase), genes for thiosulfate oxidation were present in the two Stenotrophomonas isolates but not in Chryseobacterium. The Chryseobacterium isolate encoded for a urease, suggesting the use of organic nitrogen in the form of urea, but this was absent in the two Stenotrophomonas isolates. Finally, genes for sugar utilization were identified in the two Stenotrophomonas isolates but not in Chryseobacterium.

Presence of cysteine-degrading organisms and genes in a 5-year metagenomic environmental time series
To put these laboratory results and lab-grown organisms into a natural environment context, we leveraged a previously published metagenomic time series collected from the oxygenated upper mixed layer of Lake Mendota spanning 2008-2012 (97 time points) (26) and reassembled and rebinned the data (see Methods) to search for the presence of cysteine degradation genes in the microbial communities.
First, we searched the time series to see if organisms in our study were also present in the time series. To do this, we linked the 16S rRNA gene sequences of the isolated organisms to the assembled metagenomes (i.e., contigs) from the time series. We found that while the 16S rRNA sequences were also present in the time series (Table S7 at Table S6. Research Article mSystems metagenomic data only. As such, the full-genome sequencing we performed was particularly helpful in understanding the full genomic structure of the H 2 S-producing organisms. Second, we searched for the six genes associated with cysteine degradation and H 2 S production (Table S5 at https://doi.org/10.6084/m9.figshare.21711494 ) (in binned and unbinned contigs) using KEGG HMMs with HMMsearch. In total, we searched over 22 million amino acid sequences and identified 1,882 hits to the five marker genes found in the isolate metagenomes; dcyD was not found (Fig. 5; Table S10 at https:// doi.org/10.6084/m9.figshare.21711527). cysK and malY were the genes with the most corresponding matches at any time point, followed by metC and cysM. Only two scaffolds contained tnaA. Overall, after normalizing for metagenome read size per sample, there was no obvious temporal trend of the genes, although genes were found throughout the 5-year time series.
Among these cysteine-degrading gene sequences, several were identified in binned MAGs ( Fig. 5B; Table S11 at https://doi.org/10.6084/m9.figshare.21711545), which allowed for the assignment of taxonomy. Overall, we identified 139 genes to be distributed in genomes of organisms from Actinobacteria, Bacteroidota, Chloroflexota, Cyanobacteria, Planctomycetes, Proteobacteria, and Verrucomicrobia, representing common freshwater lineages (36). The tnaA gene was only present in Bacteroidota, but other genes were more broadly distributed among taxonomic groups. Taxonomy of the metagenome-assembled genomes in which those genes were found.

DISCUSSION
The fate of H 2 S and ammonia, two products of cysteine degradation H 2 S blocks the binding of oxygen during aerobic respiration, making it toxic to the cell (37). Therefore, we hypothesize that most microbes aerobically degrading cysteine will excrete H 2 S from the cell, as was seen with the 29 isolates from this study producing H 2 S when grown in media containing a cysteine source. Additionally, the three isolates that underwent further characterization were quantitatively shown to deplete cysteine levels when producing H 2 S (Fig. 2). While it is possible that organisms may use H 2 S internally as a sulfur source, we did not identify any genes encoding sulfide quinone oxidoreductases, flavocytochrome c sulfide dehydrogenases, or other genes for the oxidation or transformation of H 2 S. H 2 S can also be used as a protective compound against antibiotics in aerobic environments by some bacteria; however, due to the stress H 2 S accumulation has on the cell, this defense mechanism is only used in extreme situations (37). The H 2 S-producing isolates fell into two groups when grown in media with cysteine: ammonia producing and ammonia consuming. Because ammonia is a common metabolic precursor, the production or consumption of ammonia alone does not indicate whether the cell is undergoing the cysteine degradation pathway. This is likely why we see variation in whether the 29 isolates produced or consumed ammonia despite all producing H 2 S (Fig. S1).

Genomic structure of the H 2 S-producing isolates
Overall, the three isolates selected for whole-genome sequencing revealed genes for cysteine degradation into H 2 S. Based on laboratory studies, they were able to produce H 2 S in the presence of oxygen. The isolates were obligate aerobes, presenting interesting questions about these organisms' life history.
S. maltophilia, S. bentonitica, and C. piscium have been shown to be present in natural environments. S. maltophilia is a cosmopolitan bacterium in nature and found in a range of natural environments, particularly in association with plants (38). S. bentonitica was originally characterized in bentonite formations, was predicted to have high tolerance to heavy metals (39), and has been observed in arctic seawater (40). C. piscium was isolated from a fish in the arctic ocean (41), but its ecological significance in the oceans remains unknown. This previously described C. piscium strain LMG 23089 was not reported to produce H 2 S, yet our genetic and physiological analyses suggest that it has the enzymatic machinery to degrade cysteine.
One possible explanation for this discrepancy is that LMG 23089 was previously grown on sulfur reduction, indole production, and motility (SIM) medium to test H 2 S. The SIM medium is used for physiological study of SIM. The SIM medium offers visual representation of H 2 S production via the reduction of thiosulfate; it reacts with iron salts in the media, and media changes to darker black color. This method of assessing H 2 S is of lower resolution than the modern H 2 S probes that measure µM concentrations. As a side test on isolate #66, H 2 S was not produced when thiosulfate was provided, but H 2 S was produced when cysteine was provided.
One particular finding of this study was that none of the six genes searched for cysteine degradation into H 2 S and ammonia was common to all three isolates, despite all three isolates showing the same cysteine decrease, ammonia increase, and H 2 S increase over time. This could be explained by alternative, perhaps less straightforward pathways for H 2 S production. One pathway is led by a gene named cystathionine gamma-lyase ("CTH" or "CSE"). In some bacteria and mammals, this enzyme is involved in H 2 S production (42). An HMM search for this enzyme showed that it was present in isolates #13, #43, and #66. While it was not initially included in the initial methods and study, this could hint to another commonality among oxic H 2 S-producing organisms.

Challenges associated with measuring oxic H 2 S production from organosulfur in the environment
Extrapolating these laboratory results to widespread distribution of organosulfur degradation in the natural environment necessitates several steps, namely because of the major knowledge gaps that exist concerning the sulfur cycle in freshwater lakes and because bridging the gap between cultivation-based, omics-based (43), and field-based experiments is needed. Foremost, the identity, distribution, and availability of organosulfur compounds broadly across lakes globally are currently mostly unknown. Cysteine is notoriously difficult to measure, and many previous studies characterizing the amino acid composition of the water column only measure the sulfur-containing organosulfur compound taurine (44,45). One of the difficulties in studying the fate of cysteine in oxic environments is that it can be abiotically and spontaneously oxidized into cystine (24), which E. coli has been shown to uptake (25). In a study of E. coli K-12 that lacked a cysteine transporter, cysteine could enter the cell through transporters dedicated to other amino acids, when no amino acid alternatives other than cysteine were present in the medium (14).
Organic sulfur in the form of cysteine is an important organosulfur amino acid and is important in protein folding and function (46). As such, there is a difference in the fates of cysteine when it exists bound in cell walls versus when cysteine is free in the water column and available for degradation by bacteria. While cysteine has been shown to contribute to the carbon pool and carbon flow in lakes (6), more quantitative field measurements are necessary to support whether cysteine also serves as a sulfur pool. Yet, other forms of organosulfur have important significance in aquatic environments. In marine environments, for example, dimethylsulfoxonium propionate is a critical component of the marine organosulfur cycle (47).
Additionally, current differences between computational gene similarity searches versus in vivo enzymatic functions are challenging to assess for the genes responsible for the cysteine degradation into pyruvate, ammonia, and H 2 S. One reason is that the enzymatic activity of the gene has mostly been described in model organisms such as E. coli, and it has been shown that gene expression can be induced by genetic factors or environmental factors such as metals (48). At least six genes have been proposed to have this enzymatic activity, yet each gene may serve different functions in situ, and it is difficult to assert directionality of enzymatic function based on metagenomic or genomic analyses only. To this end, the isolated bacterial strains from this study, which are non-model organisms and originate from the natural freshwater lake environment, may be used for further detailed biochemical, physiological, and microbiological studies. Further characterization of these bacterial isolates using gene knockout, gene induction, or heterologous gene expression studies may inform the functional activity of these genes in nature.

Implications of oxic H 2 S production by microbes in freshwaters
This study demonstrates the potential for H 2 S production by microbes in lake ecosystems to occur in the presence of oxygen, using genomic and physiological evidence from pure culture bacterial isolates and screening of long-term metagenomic time series. By combining lake-to-laboratory experiments, we show that multiple bacterial strains spanning Gammaproteobacteria, Betaproteobacteria, Actinobacteria, and Bacteroidota are all able to produce H 2 S under oxic conditions and at temperatures that would be ecologically relevant for surface lake water during the summer. Surface water temperatures in Lake Mendota can reach up to 27°C, and the top few meters of water surface are saturated in oxygen. Worldwide, maximum lake surface temperature can range between 23°C and 31°C (49).
Unlike dissimilatory sulfate reduction, bacteria use cysteine to generate ammonia, pyruvate, and H 2 S and are not dependent on sulfate as an initial reaction substrate. Increased sulfate concentrations are shown to lead to higher sulfate reduction rates in shallow eutrophic freshwater, the sulfur originating from algal decay, for example (50).
While Lake Mendota is a low-iron and high-sulfate lake (51), not all lakes have elevated sulfate levels, and therefore, H 2 S production might previously not have been thought of as relevant to study. However, sulfur-containing amino acids can have many origins. In lakes, concentrations of amino acids (free dissolved and combined) often reflect the input and outputs of the lake (52,53). For example, amino acids contributed a detectable amount to the nitrogen cycle, and bacterial utilization of amino acids contributes to nitrogen pool and cycling in natural ecosystems (53), although cysteine amino acids were not measured in that study.
Freshwater lakes that are dimictic can become stratified in temperature and oxygen during the summer, and oxygen concentrations vary throughout the year. In the fall and spring, oxygen is abundant, and cysteine degradation into H 2 S could be a relevant process for the sulfur pool, and H 2 S fluxes to the atmosphere could be significant since wind is prevalent. Under ice during the winter, where oxygen is plentiful, H 2 S could be produced but could be consumed or oxidized. On the other hand, gases would be trapped under ice. During summer, the anoxic hypolimnion and sediments are known H 2 S sources due to dissimilatory sulfate reduction, but density gradients would prevent H 2 S from reaching the atmosphere. However, the oxygenated mixed epilimnion could be an H 2 S source through organosulfur degradation. If we consider the importance of oxic H 2 S production, which could occur year-round, the H 2 S pool and the scope of sulfur transformations may be greater than anticipated, if we focus solely on the anoxic hypolimnion (Fig. 6). Future work aiming to understand the broader distribution of sulfur-containing amino acids and other organosulfur compounds in freshwaters, their

DATA AVAILABILITY STATEMENT
The 16S rRNA sequences for the 29 H2S-producing isolates, and the whole-genome sequences (nucleotides and amino acids) for isolates #13, #43, and #66 are available on OSF at the following DOI: https://doi.org/10.17605/OSF.IO/G25EQ. The isolates genomes are also deposited on NCBI in the Bioprojects: PRJNA776273, PRJNA776272, and PRJNA839079. The 97 metagenomes were previously published in reference 45 and are available through JGI's IMG/M and Genome Portal.

ADDITIONAL FILES
The following material is available online. Fig. S1 (msystems.00201-23-s0001.pdf). Qualitative accumulation of hydrogen sulfide among microbial isolates enriched from a freshwater lake water column. Filled circles represent isolates grown with cysteine, and open circles represent isolates grown without cysteine. The vertical line represents values corrected for the control (natural ammonia production/consumption in the negative control). All points to the right of the vertical lines indicate an accumulation of ammonia, and all points to the left of the vertical lines refer to those that consumed ammonia after 24 h. The isolates #43, #13, and #66 (bolded) were selected for further analysis.