Chlorine is abundant in cells and biomolecules, yet the biology of chlorine oxidation and reduction is poorly understood. Some bacteria encode the enzyme chlorite dismutase (Cld), which detoxifies chlorite (ClO2−) by converting it to chloride (Cl−) and molecular oxygen (O2). Cld is highly specific for chlorite and aside from low hydrogen peroxide activity has no known alternative substrate. Here, we reasoned that because chlorite is an intermediate oxidation state of chlorine, Cld can be used as a biomarker for oxidized chlorine species. Cld was abundant in metagenomes from various terrestrial habitats. About 5% of bacterial and archaeal genera contain a microorganism encoding Cld in its genome, and within some genera Cld is highly conserved. Cld has been subjected to extensive horizontal gene transfer. Genes found to have a genetic association with Cld include known genes for responding to reactive chlorine species and uncharacterized genes for transporters, regulatory elements, and putative oxidoreductases that present targets for future research. Cld was repeatedly co-located in genomes with genes for enzymes that can inadvertently reduce perchlorate (ClO4−) or chlorate (ClO3−), indicating that in situ (per)chlorate reduction does not only occur through specialized anaerobic respiratory metabolisms. The presence of Cld in genomes of obligate aerobes without such enzymes suggested that chlorite, like hypochlorous acid (HOCl), might be formed by oxidative processes within natural habitats. In summary, the comparative genomics of Cld has provided an atlas for a deeper understanding of chlorine oxidation and reduction reactions that are an underrecognized feature of biology.
Chlorine is converted between organic and inorganic forms in a biogeochemical cycle . Oxidation and reduction of chlorine produces different inorganic species, including the chlorine oxyanions hypochlorite (ClO−) (and its conjugate acid hypochlorous acid, HOCl), chlorite (ClO2−), chlorate (ClO3−), and perchlorate (ClO4−) [2,3,4,5,6,7]. In biology, oxidized chlorine is presently understood to be a source of energy (ClO4−, ClO3−), an intermediate in chlorination (HOCl), a chemical weapon (HOCl), and a source of oxidative stress (ClO3−, ClO2−, HOCl) [2, 6, 8, 9]. This biology reflects the high reduction potential of oxidized chlorine and the higher reactivity of its lower oxidation states: perchlorate is stable in aqueous solution, but hypochlorous acid is very reactive. Compared to other elements, less is known about the processes that produce and consume oxidized chlorine and how oxidized chlorine interacts with biology.
The production of oxidized chlorine species within biological habitats depends on oxidation state. Hypochlorous acid can be produced intracellularly and extracellularly from chemical or biochemical oxidation of chloride by enzymes like chloroperoxidase [4, 5, 9,10,11]. However, no biological oxidation of chlorine to chlorite has been observed, likely due to the high reduction potential of the redox half-reactions involved (E0’ > 1 V) . While (photo)chemical oxidation of aqueous hypochlorous acid to chlorate and perchlorate has been observed experimentally , production of perchlorate and chlorate is thought to occur predominantly in the atmosphere [7, 13, 14]. The degree of oxidation that occurs within biological habitats could be clarified by identifying the chlorine species encountered by microorganisms from different habitats.
The consumption of oxidized chlorine species, aside from the highly reactive hypochlorous acid, is thought to occur predominantly through dissimilatory (per)chlorate reduction, a specialized anaerobic respiratory pathway wherein high-affinity perchlorate reductases (Pcr) or chlorate reductases (Clr) reduce perchlorate or chlorate to provide energy in anoxic habitats [7, 15, 16]. Reduction may instead occur through co-metabolism: due to the structural and chemical similarity between oxyanions like nitrate and chlorate and perchlorate, enzymes such as nitrate reductase can reduce perchlorate or chlorate [17,18,19,20,21]. Laboratory studies have shown this inadvertent reduction of perchlorate or chlorate produces chlorite and damages cells unless chlorite is degraded . An unanswered question is if co-metabolic (per)chlorate reduction occurs at a meaningful extent at the low concentrations of perchlorate and chlorate found in natural environments. If so, many more organisms would contribute to perchlorate and chlorate reduction than presently understood.
A promising approach to answer these questions is to use a biomarker for oxidized chlorine molecules. Chlorite dismutase (Cld) is a heme-containing enzyme that catalyzes a chlorite: oxygen lyase reaction wherein a single molecule of chlorite is cleaved into chloride and molecular oxygen, which detoxifies chlorite and yields oxygen [23,24,25]. First identified as necessary enzyme in canonical dissimilatory (per)chlorate-reducing bacteria [26, 27], Cld has since been found in bacteria not known to produce chlorite as part of their metabolism . Subsequent investigations have defined the amino acids required for Cld activity  and found that aside from low hydrogen peroxidase activity, Cld has no activity towards other compounds, including nitrite, nitric oxide, hydroxylamine, and thiocyanate [28, 30]. These properties make the gene cld a useful, specific biomarker for chlorite. Because chlorite is an intermediate oxidation state of chlorine, microorganisms encoding Cld in their genomes have likely experienced not only chlorite but also more-oxidized chlorine species that can be reduced to chlorite and more-reduced chlorine oxyanion species to which chlorite is reduced.
Here, we use cld as a biomarker for chlorite in microbial genomes to expand what is known about the biology of chlorine oxyanions and redox chemistry. This comparative genomics approach adopts only two assumptions: that microorganisms encoding Cld experienced chlorite, and that genetic proximity to cld means a gene is more likely to be functionally related to cld [31, 32]. Beyond expanding the biology of chlorine oxidation and reduction, these results provide an extensive catalogue of genes potentially involved in chlorine biology for future research.
Results and Discussion
Distribution of Cld
Cld proteins belong to the protein family Pfam 06778 . Non-Cld proteins in Pfam 06778, from which Cld evolved , are mostly iron-coproporphyrin oxidative decarboxylases (HemQ) required for heme biosynthesis in monoderm bacteria . The use of chlorite dismutase (Cld) as a biomarker requires an accurate definition of proteins with Cld activity, as non-Cld proteins are often incorrectly annotated as Cld or Cld-like proteins in public databases. Here, Cld was defined as proteins in Pfam 06778 that contain the key residues required for Cld activity [23, 29]. Cld proteins formed a monophyletic clade (Fig. 1A), consisting of two major lineages, confirming previous analyses with smaller datasets [22, 34, 36, 37]. Cld proteins were primarily found in diderm phyla (Fig. 1A) and were sparsely distributed across the tree of life (Fig. 1B).
An expansive search for Cld identified 2411 Cld proteins in 2297 genomes/metagenome-assembled genomes and 6469 Cld in 1575 metagenomes (Fig. 1C, Supplementary Data). Cld was identified in 14 phyla and 143 genera, including the bacterial phyla Actinobacteria, Verrucomicrobia, Firmicutes, Chloroflexi, and Spirochaetes in which Cld has not previously been reported (Supplementary Data). For the first time, Cld was identified in the Archaea and Eukarya. The low percent identity to bacterial Cld sequences and the similarity of neighboring genes to non-bacterial genes corroborated their assignment to these taxa. The eukaryote with Cld was the unicellular green alga Monoraphidium neglectum . Cld was previously reported in a different eukaryote, the poplar tree (Populus) , but this was later determined to be contamination by bacterial genomic DNA and removed (personal communication, Joint Genome Institute).
Overall, Cld was observed in approximately 1% of genomes, 5% of genera and 15% of phyla in the NCBI taxonomy among the prokaryotes sampled. Genomes from the phyla Nitrospirae, Planctomycetes, and Nitrospinae are most likely to contain Cld, followed by Proteobacteria and Cyanobacteria (Fig. 1D). At the genus level, typically only a fraction of genomes had Cld, although Cld could be highly conserved within a genus (Fig. 1E). The widespread nature of Cld was further supported by its distribution across a dataset of 6961 IMG/M metagenomes encoding 10.8 billion genes. cld were a very low proportion of genes in host-associated systems and a greater proportion in freshwater and soil systems (Fig. 1F). Comparing metagenomes with greater than 10 million genes with this metric indicated that cld was most enriched in environments such as oligotrophic rocks, sediment, and ice followed by oxic wastewater, surface freshwater, and dryland soils. Within aquatic environments, cld appears least frequently where chlorine is most concentrated: estuary, ocean, and hypersaline waters. The shared features of environments where cld is the highest proportion of coding genes is that they are predominantly oxic; many are also exposed to high amounts of sunlight [39, 40]. The broad distribution of the cld gene in genomes and metagenomes indicates chlorite is experienced by diverse microorganisms in various habitats.
Evolution of Cld
After its initial evolution from non-Cld proteins (Fig. 1A), Cld has continued to evolve. To better understand the history of Cld, a phylogeny was built from 8924 total Cld proteins, summarized as 60 clades formed by branch distance (Fig. 2). About half the diversity of Cld (29 clades) was constituted entirely of metagenomic sequences (Fig. 2A). Just one clade contains all Cld found in previously described bacteria that perform dissimilatory perchlorate or chlorate reduction (clade 4). Many of the largest clades include Cld proteins with biochemically verified chlorite: O2 lyase activity, and the key residues for activity were conserved in all clades except clade 38 (n = 21) and the under-sampled clades 10 (n = 2) and 32 (n = 1) (Fig. 2B), indicating the conservation of activity through geologic time.
The expanded phylogeny revises the previous understanding of the vertical evolution of Cld, wherein lineage 1 Cld are larger, periplasmic proteins and lineage 2 Cld are smaller, cytoplasmic proteins [23, 41]. First, tree topology showed two distinct, diverse, strongly supported (>99% bootstraps) sublineages within lineage 2 Cld, which we term lineage 2a and 2b. The only cultivated microorganism with lineage 2b Cld is Nitrospina gracilis . Lineage 2b proteins are an intermediate length of 229 amino acids, and considerable variation in protein size was observed within the shorter lineage 2a Cld (Fig. 2B): 4% of Cld had larger (>20 aa) N- and C-terminal extensions that could either be artifacts of protein prediction or fusion proteins that augment the function of Cld . Second, signal peptide prediction suggested that the more basal branching clades of group 1 Cld are not periplasmic, while two clades of lineage 2b Cld are periplasmic and a small number of Cld from various lineage 2a clades are periplasmic (Fig. 2B). The acquisition of peptide signals for export by Cld in many lineages indicates periodic selection for the degradation of extracellular chlorite.
Horizontal gene transfer of Cld across evolutionary time is evident from its taxonomic distribution. A single Cld clade can be found in taxonomic groups spanning phyla or even domains of life. (Fig. 2C). Cld has also been subject to recent transfer between genera: multiple Cld clades consisted of Cld from different genera (Fig. 2D). Alone, this metric reflects the combined signal of vertical and horizontal inheritance, but a detailed view shows that horizontal inheritance is a large component. For example, within clade 4, there are two instances where Cld from (per)chlorate-reducing proteobacteria appeared to have been acquired by nitrite-oxidizing Nitrotoga (Supplemental Fig. 2) . Representing the most recent horizontal gene transfer: genomes from different genera possessed identical Cld proteins (Fig. 2D). In one case the same Cld protein (WP_011514928.1) was found in genomes from 18 genera. Anthropogenic sources of chlorine now appear to contribute to Cld’s evolution, evidenced by the first acquisition of Cld by an ammonia-oxidizing microorganism, Nitrosomonas mobilis Ms1, isolated from a wastewater treatment plant that used chlorine-based disinfectants (personal communication, Hirotsugu Fujitani) [45, 46]. The pattern of horizontal gene transfer suggests occasional yet strong selection for the ability to degrade chlorite. That can be reconciled with the rare conservation of Cld within phylogenetic groups (Fig. 1) by invoking a selective pressure for loss of Cld, possibly related to the heme requirement for this enzyme.
Comparative genomics of Cld
Diverse microorganisms have experienced enough chlorite to select for the cld gene. Genes co-located with cld might be involved in chlorine redox biology. To identify genes correlated with cld, 61 215 proteins encoded in 8751 genomic neighborhoods of Cld (defined as the set of genes within 10 genes upstream or downstream of cld) were clustered into 11 081 protein subfamilies, whose distribution relative to each other can be described by the clustering coefficient (Methods, Fig. 3A). Subfamilies with a low clustering coefficient are found in many different types of genomic neighborhoods with Cld and, therefore, are more likely to have a function related to chlorine redox biology, rather than be co-located by chance. Only a small fraction of protein subfamilies in Cld genomic neighborhoods had a very low clustering coefficient (Fig. 3B, Table 1). Many of these proteins were already known to be related to chlorine redox biology, such as (per)chlorate reductases, oxidative stress response, signaling, and genetic mobility (e.g. tRNA as an insertion site) [47,48,49]. Additionally, since this work began, an alkylhydroperoxidase AhpD-like protein (subfamily 84), identified here as having a low clustering coefficient, was found to be the enzyme RcsA involved in hypochlorous acid degradation in Pseudomonas aeruginosa . Therefore, this method identified true genetic associations with Cld.
Groups of highly similar genomic neighborhoods were defined by unsupervised clustering of neighborhoods with at least 10 genes (Fig. 3A). Neighborhood groups were distinguished by their most abundant non-Cld protein subfamilies (Fig. 3C) and recapitulated known differences (e.g. lineage 1 vs. 2 Cld, presence or absence of perchlorate and chlorate reductases, taxonomy) (Fig. 3D). (Lineage 1 and 2 Cld were only found together in the genomes of dissimilatory chlorate-reducing bacteria). Cld is found in a diverse set of genomic contexts: among the 20 neighborhood groups produced with this method, one single group contained all known genomic islands and composite transposons for respiratory perchlorate and chlorate reduction – the only natural pathways Cld has been confirmed participating in (Clark et al 2013, Melnyk and Coates 2015).
Clustering coefficients for protein subfamilies and grouping of neighborhoods were used to consider different aspects of chlorine biology: redox reactions, oxidative stress response, cellular transport, and chlorination. Because of their biochemistry, putative oxidoreductases with unknown function were among the most interesting proteins with genetic associations to Cld. Such oxidoreductases accounted for many of the subfamilies with the lowest clustering coefficients (Table 1): cupin domain-containing protein (subfamily 3), NADPH: quinone reductase-like Zn-dependent oxidoreductase (subfamily 85), and SDR family NAD(P)-dependent oxidoreductase (subfamily 8). Neighborhood-group 2 contained the SDR family NAD(P)-dependent oxidoreductase  as well as reactive chlorine-sensing regulatory elements that also had low clustering coefficients (subfamilies 37 and 62), further implicating a role this subfamily in reactive chlorine stress response. Fitness data for a protein in Sphingomonas koreensis DSMZ 15582 with 47% amino acid identity to a related SDR family NAD(P)-dependent oxidoreductase (subfamily 827, clustering coefficient 0.15) showed a deleterious effect when this protein was disrupted only in chlorite stress conditions or when glutamic acid was the carbon source . The cupin domain protein was one of the most common subfamilies in the dataset, being found with Cld in 1487 genomes among 40 genera. In fact, encoded with transposases in neighborhood groups 12, 15, and 21, the cupin domain protein could be found in 90% of genomic neighborhoods with the most extreme form of horizontal gene transfer: encoding Cld proteins that are identical across different genera. The cupin domain protein had been suspected to have a role in reactive chlorine species response in (per)chlorate-reducing bacteria . These data point to a far more common and important role for the cupin domain protein and other oxidoreductases in chlorine redox biology.
Cld genomic neighborhoods were searched for genes known to participate in the response to reactive chlorine species: methionine sulfoxide reductases, sulfur homeostasis proteins, protein chaperones, regulatory systems, and scavenging of reactive byproducts like peroxides, aldehydes, and glyoxals [6, 53, 54]. One or more of these genes could indeed be found in Cld neighborhood groups (Supplemental Data), and Cld was routinely found with methionine sulfoxide reductase systems (Supplemental Fig. 3). Responses to reactive chlorine species have largely been studied in laboratories; this is supporting evidence that use of these genes by microorganisms to respond to reactive chlorine species is not an experimental artifact but a natural phenomenon. Additionally, it provides evidence that chlorite is produced at a sufficient flux in the environment to contribute to oxidative damage in microorganisms.
The transport of chlorine oxyanions across the cellular membrane appeared to be a defining feature of two types of Cld genomic neighborhoods (Fig. 3C). No specific transporters for chlorine oxyanions are known. Neighborhood group 11 contained ABC transporter subunits, some of which are annotated as ATP-driven nitrate transporters. Such transporters could be involved in the transport of chlorate, a structural analogue of nitrate, an activity previously identified for nitrate transporters by genetic selection for chlorate resistance ((for example, see: )). Neighborhood group 17 was distinguished by a formate-nitrite transporter (FNT) family protein, an MsrP protein involved periplasmic reactive chlorine stress response (see below), and cytoplasmic Cld. As with nitrate and chlorate, formate (HCO2−) and nitrite (NO2−) are structural analogues of chlorite (ClO2−), and the potential for FNT family proteins to transport chlorite as well has been shown by the deleterious nature of FocA formate transporters and NirC nitrite transporters in chlorite stress conditions . Curiously, the FNT-Cld-MsrP gene cluster belonged to metagenomic Mycobacteria found in seasonally low-oxygen lakes [56, 57]. The combination of a chlorite-permeable transporter and cytoplasmic Cld might act to import extracellular chlorite to be converted to oxygen inside the cell. Microorganisms benefitting from the production of oxygen by Cld is a trait thus far observed only in (per)chlorate-reducing bacteria or engineered strains [7, 58].
Chlorination and dechlorination are only known to be related to hypochlorous acid, not higher oxidation states of chlorine like chlorite. Relatively low clustering coefficients with Cld for two protein subfamilies suggested otherwise: non-heme chloroperoxidase (subfamily 122), which chlorinates organic molecules by producing hypochlorous acid, and a putative subfamily of haloacid dehalogenases (subfamily 172), which removes chlorine from organic molecules. Chlorite produces and may be produced by hypochlorous acid, which generates stable chlorinated products like chlorotyrosine ; perhaps microorganisms use dehalogenases to reverse that chlorination.
Chlorite from chlorine reduction
One source of chlorite experienced by microorganisms could be the co-metabolic reduction of perchlorate and chlorate. Comparative genomics can be used to evaluate this hypothesis: if an enzyme reduces perchlorate or chlorate to chlorite, Cld can provide a benefit by degrading chlorite, and the selective pressure to co-express the Cld with would lead to its genetic co-location with the reductase in some genomes (Fig. 4A). To test this hypothesis, Cld genomic neighborhoods were searched for enzymes in the DMSO reductase family of molybdopterin enzymes (Pfam 00384), which contains both the specialized perchlorate and chlorate reductases (Pcr, Clr) and the enzymes that inadvertently reduce perchlorate and chlorate .
A total of 105 proteins in the DMSO reductase family were found in Cld genomic neighborhoods. Many genomic neighborhoods contained cytoplasmic Cld and enzymes with documented in vitro (per)chlorate reductase activity (Fig. 4B–E): assimilatory nitrate reductases (NasA), a cytoplasmic dissimilatory nitrate reductase (NarG) , and a tetrathionate reductase (TtrA) . Cld was also found on soil metagenome contigs with uncharacterized enzymes most similar to periplasmic nitrite oxidoreductases (pNxr) of nitrite-oxidizing bacteria (Nitrospira and Nitrotoga), anammox bacteria, and other microorganisms . As it is found with either periplasmic or cytoplasmic Cld, this uncharacterized reductase is not likely to function in dissimilatory (per)chlorate reduction pathways, which are only known to occur in the periplasm . The co-occurrence of Cld with formate dehydrogenase (FdhN) and an uncharacterized Fdh-like protein (YdeP) was unexpected but could be related to the structural similarity of chlorite and formate. The genetic association of Cld with diverse enzymes with co-metabolic (per)chlorate reductase activity indicates that co-metabolism contributes to perchlorate and chlorate reduction.
Most commonly, the reductases detected with Cld in metagenomes were not co-metabolic reductases but Pcr and the newly characterized group 3 Clr  (Fig. 4D, E). Most reductases were closely related to reductases from organisms selected for the ability to respire perchlorate or chlorate. The search also identified a metagenome-assembled genome, GWC2_42_11, assembled from a groundwater sediment metagenome , encoding phylogenetically divergent copies of both Pcr and group 3 Clr (Fig. 4C). Two genes for Cld from this microorganism are found in Cld clade 6, which share a more recent ancestor with nitrite-oxidizing Nitrospira (clade 5, clades 7–9) than perchlorate-reducing bacteria (clade 4) (Fig. 2). As a member of the class Deltaproteobacteria (phylum GWC2–55–46 in GTDB taxonomy), GWC2_42_11 is the most evolutionary distinct (per)chlorate-reducing bacterium identified to date, and its equally divergent reductases and Cld might help in understanding the earliest forms of perchlorate and chlorate respiration.
The ancient form of dissimilatory (per)chlorate reduction may have resembled co-metabolic (per)chlorate reduction by a microorganism with Cld. Cld has been shown to be inessential for removing any chlorite produced if habitats with sufficient amounts of reduced iron or inorganic sulfur species [64,65,66] (or with large populations of other microorganisms that can degrade chlorate or chlorite [16, 58]), allowing a scenario where chlorite dismutase could evolve after specialized (per)chlorate reductases. Yet the above results show that chlorite stress from co-metabolic (per)chlorate reduction is a common enough phenomenon that Cld has repeatedly evolved to be co-located with co-metabolic reductase in genomes. This is a contemporary example of how respiratory metabolisms for oxidized chlorine could have first arose from the association between chlorite dismutase and a co-metabolic reductase followed by specialization for respiratory perchlorate or chlorate reduction . In that scenario, chlorite dismutase would have evolved before specialized (per)chlorate reductases.
Chlorite from chlorine oxidation
The oxidation of chlorine to chlorite is another possible source, other than co-metabolic reduction of (per)chlorate or chemical reduction of chlorate , of chlorite in oxic habitats. If it occurs, Cld should be present in microorganisms unable to co-metabolically reduce (per)chlorate. Using profile-HMMs representing the broad parts of the DMSO reductase family phylogeny that have perchlorate or chlorate reductase activity (PCRA) (Fig. 4A), we identified genomes with Cld that do not have enzymes that reduce perchlorate or chlorate (Fig. 5A). Despite the commonality of such enzymes as assimilatory nitrate reductases, this search identified 27 putative “non-(per)chlorate reducers” among isolate genomes (Supplementary Table 1). These strains represent 6 of the 19 phyla with Cld and 15 of 151 genera (Fig. 5B). All were reported aerobes (Fig. 5B). They were isolated from diverse habitats, often characterized by high sunlight (lakes and ponds, desert rocks and sediments, growing with diatoms, cyanobacteria, or mosses) or by high amounts of reactive chlorine species (human body, wastewater treatment plant, swimming pool, showerhead biofilm) (Fig. 5B). Therefore, the known mechanisms for the enzymatic reduction of chlorate and perchlorate appeared insufficient to explain the prevalence of chlorite and chlorite-degrading microorganisms.
One plausible mechanism of oxidative chlorite formation is oxidative chemistry (Fig. 5C). Habitats with high sunlight are very oxidizing due to the combined effects of oxygenic phototrophy and UV photochemistry [69, 70]. Several non-(per)chlorate reducers were isolated from high sunlight habitats. UV tolerance genes were present in Cld genomic neighborhoods from several bacteria from high sunlight habitats. A putative deoxyribodipyrimidine photo-lyase, which is a light-activated protein that repairs UV-damaged DNA, is encoded in 21 Cld genomic neighborhoods and several different neighborhood groups (groups 2, 20, and 7), such as one in a betaprotebacterium in culture with Leptolyngbya glacialis TM1FOS73 (GCA_003242045.1) . In a sunlight photobioreactor metagenome, cld is found in four of 34 MAGs (UBA7691, UBA7681, UBA7678, UBA7677), including a Planctomycetaceae bacterium that has periplasmic group 2a Cld encoded near carotenoid biosynthesis genes for limiting UV photodamage . The production of chlorite from oxidative chemistry might also explain the presence of Cld in the nitrite-oxidizing bacteria (Fig. 5C). Like the products of photochemistry such as hydrogen peroxide (H2O2, E0’ + 1.32 V) , the reactive nitrogen species peroxynitrite (ONOO−, E0’ + 1.3 V) has a high enough reduction potential to oxidize hypochlorous acid to chlorite (E0’ + 1.26 V) [74, 75].
Another plausible mechanism of chlorine oxidation is the biochemical oxidation of hypochlorous acid to chlorite (Fig. 5C). An enzyme that oxidizes hypochlorous acid to chlorite would be a major fitness benefit to microorganisms with Cld, ultimately yielding harmless chloride and oxygen. Oxidation of hypochlorous acid to chlorite only requires transfer of 2 electrons and produces a less reactive product. An analogous system would be nitric oxide dioxygenase, which uses oxygen to oxidize nitric oxide to less-toxic nitrate . Instead of spending cellular reducing equivalents to reduce hypochlorous acid or repair oxidative damage, the enzymatic oxidation of hypochlorous acid might produce reducing equivalents. Furthermore, the removal of hypochlorous acid would limit the inhibition of Cld by hypochlorous acid . Thus, the enzymatic oxidation of hypochlorous acid to chlorite would pose major selective benefits.
Experimental support for this capability would be the enrichment of microorganisms with Cld in habitats with high hypochlorous acid. One such real-word setting appeared to be a drinking water distribution system in which 47 of 89 strains isolated from a showerhead biofilm encoded Cld, and 10 were non-(per)chlorate reducers (Fig. 5D) [78, 79]. This demonstrates a strong selection for Cld within the microbial community by the chlorine residuals present in the water. The water distribution system was expected to contain 0.8 mg/liter free residual chlorine form of hypochlorous acid and hypochlorite residuals; however, it is unclear if chlorine dioxide (ClO2) was used in water treatment and produced chlorite residuals (personal communication, Jorge Santo-Domingo). Except for this uncertainty, this system would meet the criteria of a habitat that selects for the ability to degrade chlorite due to only high hypochlorous acid exposure. Enzymatic oxidation of hypochlorous acid to chlorite remains an unproven hypothesis.
A holistic model for chlorine redox biology
The different biological processes that involve Cld suggest that the biology of chlorine reduction and oxidation should be considered as a single, bidirectional pathway (Fig. 6). In this model, based on the above data and previous studies, microorganisms in many habitats can experience any chlorine oxyanion and some anthropogenic oxidized chlorine species (Cl2, ClO2, NH2Cl, NHCl2, and NCl3) [80, 81]. Transporters allow oxidized chlorine species to enter cells, where the molecules or their byproducts may be sensed and lead to changes in gene regulation. Oxidized chlorine species have a propensity to be reduced in cells to the next lower oxidation state. Chlorite is produced in microorganisms with enzymes that can reduce perchlorate and chlorate through metabolism or co-metabolism. Cld can be considered a shunt in the reductive pathway that, when present, prevents the formation of hypochlorous acid and produces beneficial oxygen. Any hypochlorous acid formed reacts rapidly with biomolecules, producing a combination of chloride, chlorinated carbon and nitrogen, and oxidized byproducts [2, 3, 5, 82]. Cells degrade hypochlorous acid and respond to damage caused by it. A hypothetical possibility is that microorganisms detoxify hypochlorous acid by oxidizing hypochlorous acid to chlorite and using the chlorite dismutase shunt to degrade chlorite. In oxidizing settings, chlorite can be further oxidized (photo)chemically to chlorate or perchlorate, which are also deposited into habitats from atmosphere. The relatively stable end products of this bidirectional cycle are perchlorate and chloride. These compounds are only as inert, however, as the surrounding chemistry and biology allow.
Cld, a biomarker for chlorite once thought unique to anaerobic perchlorate- and chlorate-reducing bacteria, is found in various microorganisms from both oxic and anoxic microbial habitats. This distribution suggests microorganisms experience significant enough amounts of chlorite in the environment to acquire Cld. The sources of chlorite are the dissimilatory reduction of (per)chlorate but also the co-metabolic reduction of (per)chlorate and, genomics suggests, the oxidation of chlorine’s lower oxidation states. That Cld participates in these pathways and in general response to reactive chlorine species justifies a model wherein oxidized chlorine species are part of a continuous, bidirectional biological pathway. Cld is subject to intermittent selection for its gain and loss, highlighting how much remains to be learned about the concentrations and fluxes of oxidized chlorine species in different environments. The expansive inventory of genes associated with Cld-encoding loci identified here provides targets for subsequent research in the biology of oxidized chlorine from regulation, transport, and repair to direct enzymatic action on chlorine-containing molecules.
Identification of chlorite dismutase (Cld)
A maximum likelihood phylogenetic tree of the protein family containing Cld was constructed using FastTree v. 2.1.9 with default settings from the Pfam 06778 alignment of representative proteomes, at the 15% comembership threshold to limit the number of redundant proteins [83,84,85]. The presence of key residues for Cld activity were identified by comparing the positions in the alignment corresponding to the distal heme arginine (R127) and proximal heme lysine (K92), histidine (H114), and glutamic acid (E167) in Nitrobacter winogradskyii Nb-255 . Cld and non-Cld proteins were annotated on a precomputed bacterial tree of life . Cld was identified through BLASTP search of genomes deposited in NCBI and genomes and metagenomes deposited in JGI [87,88,89] and confirmed with profile-hidden Markov models built from the two major lineages of Cld from Pfam 06778 . All data were acquired prior to 2020. See supplementary information for additional methods.
For the phylogeny of Cld, proteins were aligned using MUSCLE v3.8.1551 with default settings  and built into a maximum likelihood phylogenetic tree using FastTree with default settings . The Python package ETE v. 3 was used to plot trees and to form clades of proteins at trees nodes in which the average distance to a protein was less than a selected value . N-terminal and C-terminal extensions were defined as amino acids in the alignment beyond the positions within which the average Cld protein had amino acids. Signal peptides were assigned using SignalP v. 5 with default settings, accepting a positive result from any type of organism . See supplementary information for additional methods.
Genes within +/− 10 positions of cld on the same contig were defined as part of the Cld genomic neighborhood. Proteins from Cld genomic neighborhoods were clustered into subfamilies using MMSEQs v.7–4e23d set to a coverage of 0.5 and an E-value of 0.001 . The Python package networkx was used to compute the clustering coefficient for each node. This is a simple statistic for gene linkage to cld obtained by representing each subfamily as a node and each connection between subfamilies found in the same genomic neighborhood as edges in a network. The clustering coefficient for a node equals the number of edges between a node’s neighbors divided by the total number of edges possible between a node’s neighbors (for more information, see supplementary information).
Genomic neighborhoods with 10+ genes were grouped by similar gene content using unsupervised machine learning methods in the Python package SciKit-learn. The features were the presence  or absence (0) of each subfamily in the neighborhood. An initial dimensional reduction was performed with Principle Components Analysis, and the resulting 50 dimensions were subjected to t-Distributed Stochastic Neighbor Embedding (t-SNE) with a perplexity of 50 and 5000 iterations. Neighborhoods were then grouped by proximity (maximum distance of 4 units in the two t-SNE dimensions) with the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm.
Supplementary data are available on FigShare and include: Supplementary Data 1, information on genes and genomes used in this work including accessions, taxonomy, subfamily assignments, etc. (https://doi.org/10.6084/m9.figshare.16978561); Supplementary Data 2, information on subfamilies and their clustering coefficients (https://doi.org/10.6084/m9.figshare.16980601); protein sequences found in Cld genomic neighborhoods (https://doi.org/10.6084/m9.figshare.16980613); a phylogenetic tree and alignments for Cld (https://doi.org/10.6084/m9.figshare.16982077); and profile-HMMs to identify key proteins for perchlorate, chlorate, and chlorite biology (https://doi.org/10.6084/m9.figshare.19836151).
Atashgahi S, Liebensteiner MG, Janssen DB, Smidt H, Stams AJM, Sipkema D. Microbial synthesis and transformation of inorganic and organic chlorine compounds. Front Microbiol. 2018;9:3079.
Bengtson P, Bastviken D, de Boer W, Oberg G. Possible role of reactive chlorine in microbial antagonism and organic matter chlorination in terrestrial environments. Environ Microbiol. 2009;11:1330–9.
Winterbourn CC, Kettle AJ. Redox reactions and microbial killing in the neutrophil phagosome. Antioxid Redox Signal. 2013;18:642–60.
Agarwal V, Miles ZD, Winter JM, Eustáquio AS, El Gamal AA, Moore BS. Enzymatic halogenation and dehalogenation reactions: Pervasive and mechanistically diverse. Chem Rev. 2017:acs.chemrev.6b00571.
Comba P, Kerscher M, Krause T, Schöler HF. Iron-catalysed oxidation and halogenation of organic matter in nature. Environ Chem. 2015;12:381–95.
Gray MJ, Wholey W-Y, Jakob U. Bacterial responses to reactive chlorine species. Annu Rev Microbiol. 2013;67:141–60.
Youngblut MD, Wang O, Barnum TP, Coates JD. (Per)chlorate in biology on Earth and beyond. Annu Rev Microbiol. 2016;70:435–59.
Coates JD, Achenbach LA. Microbial perchlorate reduction: rocket-fueled metabolism. Nat Rev Microbiol. 2004;2:569–80.
Winterbourn CC. Reconciling the chemistry and biology of reactive oxygen species. Nat Chem Biol. 2008;4:278–86.
Leri AC, Mayer LM, Thornton KR, Northrup PA, Dunigan MR, Ness KJ, et al. A marine sink for chlorine in natural organic matter. Nat Geosci. 2015;8:620–4.
Ortiz-Bermudez P, Hirth KC, Srebotnik E, Hammel KE. Chlorination of lignin by ubiquitous fungi has a likely role in global organochlorine production. Proc Natl Acad Sci USA. 2007;104:3895–3900.
Rao B, Estrada N, McGee S, Mangold J, Gu B, Jackson WA. Perchlorate production by photodecomposition of aqueous chlorine solutions. Environ Sci Technol. 2012;46:11635–43.
Rajagopalan S, Anderson T, Cox S, Harvey G, Cheng Q, Jackson WA. Perchlorate in wet deposition across North America. Environ Sci Technol. 2009;43:616–22.
Dasgupta PK, Martinelango PK, Jackson WA, Anderson TA, Tian K, Tock RW, et al. The origin of naturally occurring perchlorate: The role of atmospheric processes. Environ Sci Technol. 2005;39:1569–75.
Youngblut MD, Tsai CL, Clark IC, Carlson HK, Maglaqui AP, Gau-Pan PS, et al. Perchlorate reductase is distinguished by active site aromatic gate residues. J Biol Chem. 2016;291:9190–202.
Barnum TP, Cheng Y, Hill KA, Lucas LN, Carlson HK, Coates JD. Identification of a parasitic symbiosis between respiratory metabolisms in the biogeochemical chlorine cycle. ISME J. 2020;14:1194–206.
Leimkühler S, Iobbi-Nivol C. Bacterial molybdoenzymes: old enzymes for new purposes. FEMS Microbiol Rev. 2015;40:1–18.
Weiner JH, MacIsaac DP, Bishop RE, Bilous PT. Purification and properties of Escherichia coli dimethyl sulfoxide reductase, an iron-sulfur molybdoenzyme with broad substrate specificity. J Bacteriol. 1988;170:1505–10.
Hinsley AP, Berks BC. Specificity of respiratory pathways involved in the reduction of sulfur compounds by Salmonella enterica. Microbiology 2002;148:3631–8.
Riggs DL, Tang JS, Barrett EL. Thiosulfate reductase as a chlorate reductase in Salmonella typhimurium. FEMS Microbiol Lett. 1987;44:427–30.
McEwan AG, Wetzstein HG, Meyer O, Jackson JB, Ferguson SJ. The periplasmic nitrate reductase of Rhodobacter capsulatus; purification, characterisation and distinction from a single reductase for trimethylamine-N-oxide, dimethylsulphoxide and chlorate. Arch Microbiol. 1987;147:340–5.
Celis AI, Geeraerts Z, Ngmenterebo D, Machovina MM, Kurker RC, Rajakumar K, et al. A dimeric chlorite dismutase exhibits O2-generating activity and acts as a chlorite antioxidant in Klebsiella pneumoniae MGH 78578. Biochemistry 2015;54:434–46.
Schaffner I, Hofbauer S, Krutzler M, Pirker KF, Furtmüller PG, Obinger C. Mechanism of chlorite degradation to chloride and dioxygen by the enzyme chlorite dismutase. Arch Biochem Biophys. 2015;574:18–26.
Van Ginkel CG, Rikken GB, Kroon AGM, Kengen SWM. Purification and characterization of chlorite dismutase: A novel oxygen-generating enzyme. Arch Microbiol. 1996;166:321–6.
Dubois JL. O–O bond formation by a heme protein: The Unexpected efficiency of chlorite dismutase. Molecular Water Oxidation Catalysis: A Key Topic for New Sustainable Energy Conversion Schemes 2014.
Rikken GB, Kroon AGM, Van Ginkel CG. Transformation of (per)chlorate into chloride by a newly isolated bacterium: reduction and dismutation. Appl Microbiol Biotechnol. 1996;45:420–6.
Coates JD, Michaelidou U, Bruce RA, O’Connor SM, Crespi JN, Achenbach LA. Ubiquity and diversity of dissimilatory (per)chlorate-reducing bacteria. Appl Environ Microbiol. 1999;65:5234–41.
Maixner F, Wagner M, Lücker S, Pelletier E, Schmitz-Esser S, Hace K, et al. Environmental genomics reveals a functional chlorite dismutase in the nitrite-oxidizing bacterium ‘Candidatus Nitrospira defluvii’. Environ Microbiol. 2008;10:3043–56.
Hofbauer S, Howes BD, Flego N, Pirker KF, Schaffner I, Mlynek G, et al. From chlorite dismutase towards HemQ-the role of the proximal H-bonding network in haeme binding. Biosci Rep. 2016;36:e00312.
de Geus DC, Thomassen EAJ, Hagedoorn P-L, Pannu NS, van Duijn E, Abrahams JP. Crystal structure of chlorite dismutase, a detoxifying enzyme producing molecular oxygen. J Mol Biol. 2009;387:192–206.
Overbeek R, Fonstein M, D’Souza M, Gordon DP, Maltsev N. The use of gene clusters to infer functional coupling. Proc Natl Acad Sci USA. 1999;96:2896–901.
Pellegrini M, Marcotte E, Thompson MJ, Eisenberg D, Yeates TO. Assigning protein functions by comparative genome analysis: Protein phylogenetic profiles. Proc Natl Acad Sci USA. 1999;96:4285–8.
Celis AI, DuBois JL. Substrate, product, and cofactor: The extraordinarily flexible relationship between the CDE superfamily and heme. Arch Biochem Biophys. 2015;574:3–17.
Zámocký M, Hofbauer S, Schaffner I, Gasselhuber B, Nicolussi A, Soudi M, et al. Independent evolution of four heme peroxidase superfamilies. Arch Biochem Biophys. 2015;574:108–19.
Ha Dailey, Gerdes S. HemQ: An iron-coproporphyrin oxidative decarboxylase for protoheme synthesis in Firmicutes and Actinobacteria. Arch Biochem Biophys. 2015;574:27–35.
Schaffner I, Hofbauer S, Krutzler M, Pirker KF, Bellei M, Stadlmayr G, et al. Dimeric chlorite dismutase from the nitrogen-fixing cyanobacterium Cyanothece sp. PCC7425. Mol Microbiol. 2015;96:1053–68.
Mlynek G, Kostan J, Fu S, Maixner F, Gysel K, Furtmu PG, et al. Unexpected diversity of chlorite dismutases: a catalytically efficient dimeric enzyme from Nitrobacter winogradskyi. J Bacteriol. 2011;193:2408–17.
Bogen C, Al-Dilaimi A, Albersmeier A, Wichmann J, Grundmann M, Rupp O, et al. Reconstruction of the lipid metabolism for the microalga Monoraphidium neglectum from its genome sequence reveals characteristics suitable for biofuel production. BMC Genom. 2013;14:926.
Cary SC, McDonald IR, Barrett JE, Cowan DA. On the rocks: the microbiology of Antarctic Dry Valley soils. Nat Rev Microbiol. 2010;8:129–38.
Maccario L, Carpenter SD, Deming JW, Vogel TM, Larose C. Sources and selection of snow-specific microbial communities in a Greenlandic sea ice snow cover. Sci Rep. 2019;9:2290.
Dubois JL, Ojha S. Production of dioxygen in the dark: Dismutases of oxyanions. Sustaining Life on Planet Earth: Metalloenzymes Mastering Dioxygen and Other Chewy Gases 2015. p. 45–87.
Lucker S, Nowka B, Rattei T, Spieck E, Daims H. The genome of Nitrospina gracilis illuminates the metabolism and evolution of the major marine nitrite oxidizer. Front Microbiol. 2013;4:27.
Enright AJ, Iliopoulos I, Kyrpides NC, Ouzounis CA. Protein interaction maps for complete genomes based on gene fusion events. Nature 1999;402:86–90.
Kitzinger K, Koch H, Lücker S, Sedlacek CJ, Herbold C, Schwarz J, et al. Characterization of the first “Candidatus Nitrotoga” isolate reveals metabolic versatility and separate evolution of widespread nitrite-oxidizing bacteria. mBio 2018;9:e01186–18.
Thandar SM, Ushiki N, Fujitani H, Sekiguchi Y, Tsuneda S. Ecophysiology and comparative genomics of Nitrosomonas mobilis Ms1 isolated from autotrophic nitrifying granules of wastewater treatment bioreactor. Front Microbiol. 2016;7:1869.
Fujitani H, Kumagai A, Ushiki N, Momiuchi K, Tsuneda S. Selective isolation of ammonia-oxidizing bacteria from autotrophic nitrifying granules by applying cell-sorting and sub-culturing of microcolonies. Front Microbiol. 2015;6:1159.
Melnyk Ra, Engelbrektson A, Clark IC, Carlson HK, Byrne-Bailey K, Coates JD. Identification of a perchlorate reduction genomic island with novel regulatory and metabolic genes. Appl Environ Microbiol. 2011;77:7401–4.
Clark IC, Melnyk RA, Engelbrektson A, Coates JD. Structure and evolution of chlorate reduction composite transposons. mBio 2013;4:e00379–13.
Melnyk RA, Coates JD. The perchlorate reduction genomic island: Mechanisms and pathways of evolution by horizontal gene transfer. BMC Genom. 2015;16:862.
Nontaleerak B, Duang-nkern J, Wongsaroj L, Trinachartvanit W, Romsang A, Mongkolsuk S. Roles of RcsA, an AhpD family protein, in reactive chlorine stress resistance and virulence in Pseudomonas aeruginosa. Appl Environ Microbiol. 2020;86:e01480–20.
Kallberg Y, Oppermann U, Persson B. Classification of the short-chain dehydrogenase/reductase superfamily using hidden Markov models. FEBS J. 2010;277:2375–86.
Price MN, Wetmore KM, Waters RJ, Callaghan M, Ray J, Liu H, et al. Mutant phenotypes for thousands of bacterial genes of unknown function. Nature 2018;557:503–9.
Melnyk RA, Youngblut MD, Clark IC, Carlson HK, Wetmore KM, Price MN, et al. Novel mechanism for scavenging of hypochlorite involving a periplasmic methionine-rich peptide and methionine sulfoxide reductase. mBio 2015;6:e00233–15.
Sultana S, Foti A, Dahl J-U. Bacterial defense systems against the neutrophilic oxidant hypochlorous acid. Infect Immun. 2020;88:e00964–19.
Stewart V. Nitrate respiration in relation to facultative metabolism in enterobacteria. Microbiol Rev. 1988;52:190–232.
Tran P, Ramachandran A, Khawasik O, Beisner BE, Rautio M, Huot Y, et al. Microbial life under ice: Metagenome diversity and in situ activity of Verrucomicrobia in seasonally ice-covered Lakes. Environ Microbiol. 2018;20:2568–84.
Tsuji JM, Tran N, Schiff SL, Venkiteswaran JJ, Molot LA, Tank M, et al. Anoxygenic photosynthesis and iron–sulfur metabolic potential of Chlorobia populations from seasonally anoxic Boreal Shield lakes. ISME J 2020;14:2732–47.
Clark IC, Youngblut M, Jacobsen G, Wetmore KM, Deutschbauer A, Lucas L, et al. Genetic dissection of chlorate respiration in Pseudomonas stutzeri PDA reveals syntrophic (per)chlorate reduction. Environ Microbiol. 2016;18:3342–54.
Chapman ALP, Hampton MB, Senthilmohan R, Winterbourn CC, Kettle AJ. Chlorination of bacterial and neutrophil proteins during phagocytosis and killing of Staphylococcus aureus. J Biol Chem. 2002;277:9757–62.
Liu Q, Song W-Z, Zhou Y-G, Dong X-Z, Xin Y-H. Phenotypic divergence of thermotolerance: Molecular basis and cold adaptive evolution related to intrinsic DNA flexibility of glacier-inhabiting Cryobacterium strains. Environ Microbiol. 2020;22:1409–20.
Singh D, Kumari A, Ramaswamy S, Ramanathan G. Expression, purification and substrate specificities of 3-nitrotoluene dioxygenase from Diaphorobacter sp. strain DS2 Biochem Biophys. 2014;445:36–DS42.
Barnum TP, Coates JD. An uncharacterized clade in the DMSO reductase family of molybdenum oxidoreductases is a new type of chlorate reductase. Environ Microbiol Rep. 2020;12:534–9.
Anantharaman K, Brown CT, Hug LA, Sharon I, Castelle CJ, Probst AJ, et al. Thousands of microbial genomes shed light on interconnected biogeochemical processes in an aquifer system. Nat Commun. 2016;7:13219.
Liebensteiner MG, Pinkse MWH, Nijsse B, Verhaert PDEM, Tsesmetzis N, Stams AJM, et al. Perchlorate and chlorate reduction by the Crenarchaeon Aeropyrum pernix and two thermophilic Firmicutes. Environ Microbiol Rep. 2015;7:936–45.
Liebensteiner MG, Pinkse MWH, Schaap PJ, Stams AJM, Lomans BP. Archaeal (per)chlorate reduction at high temperature: an interplay of biotic and abiotic reactions. Science 2013;340:85–87.
Sorlini S, Collivignarelli C. Chlorite removal with ferrous ions. Desalination 2005;176:267–71.
Barnum TP, Figueroa IA, Carlström CI, Lucas LN, Engelbrektson AL, Coates JD. Genome-resolved metagenomics identifies genetic mobility, metabolic interactions, and unexpected diversity in perchlorate-reducing communities. ISME J. 2018;12:1568–81.
Brundrett M, Yan W, Velazquez MC, Rao B, Jackson WA. Abiotic reduction of chlorate by Fe(II) minerals: Implications for occurrence and transformation of oxy-chlorine species on Earth and Mars. ACS Earth Space Chem. 2019;3:700–10.
Gao Q, Garcia-Pichel F. Microbial ultraviolet sunscreens. Nat Rev Microbiol. 2011;9:791–802.
Latifi A, Ruiz M, Zhang CC. Oxidative stress in cyanobacteria. FEMS Microbiol Rev. 2009;33:258–78.
Cornet L, Bertrand AR, Hanikenne M, Javaux EJ, Wilmotte A, Baurain D. Metagenomic assembly of new (sub)polar Cyanobacteria and their associated microbiome from non-axenic cultures. Micro Genom. 2018;4:e000212.
Krohn-Molt I, Wemheuer B, Alawi M, Poehlein A, Gullert S, Schmeisser C, et al. Metagenome survey of a multispecies and alga-associated biofilm revealed key elements of bacterial-algal interactions in photobioreactors. Appl Environ Microbiol. 2013;79:6196–206.
Winterbourn CC. The biological chemistry of hydrogen peroxide. In: Cadenas E, Packer L, editors. Methods Enzymology. 528: Academic Press; 2013. p. 3–25.
Radi R. Oxygen radicals, nitric oxide, and peroxynitrite: Redox pathways in molecular medicine. Proc Natl Acad Sci USA. 2018;115:5839–48.
Moller MN, Rios N, Trujillo M, Radi R, Denicola A, Alvarez B. Detection and quantification of nitric oxide-derived oxidants in biological systems. J Biol Chem. 2019;294:14776–802.
Forrester MT, Foster MW. Protection from nitrosative stress: a central role for microbial flavohemoglobin. Free Radic Biol Med. 2012;52:1620–33.
Hofbauer S, Gruber C, Pirker KF, Sündermann A, Schaffner I, Jakopitsch C, et al. Transiently produced hypochlorite is responsible for the irreversible inhibition of chlorite dismutase. Biochemistry 2014;53:3145–57.
Soto-Giron MJ, Rodriguez RL, Luo C, Elk M, Ryu H, Hoelle J, et al. Biofilms on hospital shower hoses: Characterization and implications for nosocomial infections. Appl Environ Microbiol. 2016;82:2872–83.
Le Dantec C, Duguet JP, Montiel A, Dumoutier N, Dubrou S, Vincent V. Chlorine disinfection of atypical mycobacteria isolated from a water distribution system. Appl Environ Microbiol. 2002;68:1025–32.
Deborde M, von Gunten U. Reactions of chlorine with inorganic and organic compounds during water treatment - Kinetics and mechanisms: A critical review. Water Res. 2008;42:13–51.
Asami M, Kosaka K, Kunikane S. Bromate, chlorate, chlorite and perchlorate in sodium hypochlorite solution used in water supply. J Water Supply: Res Technol-AQUA. 2009;58:107–15.
Hwang C, Ling F, Andersen GL, LeChevallier MW, Liu WT. Microbial community dynamics of an urban drinking water distribution system subjected to phases of chloramination and chlorination treatments. Appl Environ Microbiol. 2012;78:7856–65.
El-Gebali S, Mistry J, Bateman A, Eddy SR, Luciani A, Potter SC, et al. The Pfam protein families database in 2019. Nucleic Acids Res. 2018;47:D427–D432.
Chen C, Natale DA, Finn RD, Huang H, Zhang J, Wu CH, et al. Representative proteomes: a stable, scalable and unbiased proteome set for sequence analysis and functional annotation. PLoS One. 2011;6:e18910.
Price MN, Dehal PS, Arkin AP. FastTree 2–approximately maximum-likelihood trees for large alignments. PLoS One. 2010;5:e9490.
Mendler K, Chen H, Parks DH, Lobb B, Hug LA, Doxey AC. AnnoTree: visualization and exploration of a functionally annotated microbial tree of life. Nucleic Acids Res. 2019;47:4442–8.
Camacho C, Coulouris G, Avagyan V, Ma N, Papadopoulos J, Bealer K, et al. BLAST plus: architecture and applications. BMC Bioinform. 2009;10:421.
Chen I -MA, Markowitz VM, Chu K, Palaniappan K, Szeto E, Pillay M, et al. IMG/M: integrated genome and metagenome comparative data analysis system. Nucleic Acids Res. 2017;45:507–16.
Coordinators NR, Om- GE, Viewer M, Read S. Database resources of the National Center for Biotechnology Information. Nucleic Acids Res. 2014;43:13–21.
Finn RD, Clements J, Arndt W, Miller BL, Wheeler TJ, Schreiber F, et al. HMMER web server: 2015 update. Nucleic Acids Res. 2015;43:30–38.
Edgar RC. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 2004;32:1792–7.
Huerta-Cepas J, Serra F, Bork P. ETE 3: Reconstruction, analysis, and visualization of phylogenomic data. Mol Biol Evol. 2016;33:1635–8.
Almagro Armenteros JJ, Tsirigos KD, Sønderby CK, Petersen TN, Winther O, Brunak S, et al. SignalP 5.0 improves signal peptide predictions using deep neural networks. Nat Biotechnol. 2019;37:420–3.
Hauser M, Steinegger M, Söding J. MMseqs software suite for fast and deep clustering and searching of large protein sequence sets. Bioinformatics 2016;32:1323–30.
This work would not be possible without the sequencing of microorganisms and microbial communities by countless scientists, for whom we are in continual gratitude. We also thank Shengqiang Shu and Joe Carlson of the Joint Genome Institute, Hirotsugu Fujitani, Jorge Santo-Domingo, and Kostas Konstantinidis for their personal communication regarding their work. Financial support was provided through a grant from the Energy Biosciences Institute EBI-BP program to JDC and through the NSF Graduate Research Fellowship Program to TPB.
The authors declare no competing interests.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Barnum, T.P., Coates, J.D. Chlorine redox chemistry is widespread in microbiology. ISME J (2022). https://doi.org/10.1038/s41396-022-01317-5