Isolation, characterization, and genetic manipulation of cold-tolerant, manganese-oxidizing Pseudomonas sp. strains

ABSTRACT Manganese-oxidizing bacteria (MnOB) produce Mn oxide minerals that can be used by humans for bioremediation, but the purpose for the bacterium is less clear. This study describes the isolation and characterization of cold-tolerant MnOB strains isolated from a compost pile in Morris, Minnesota, USA: Pseudomonas sp. MS-1 and DSV-1. The strains were preliminarily identified as members of species Pseudomonas psychrophila by 16S rRNA analysis and a multi-locus phylogenetic study using a database of 88 genomes from the Pseudomonas genus. However, the average nucleotide identity between these strains and the P. psychrophila sp. CF149 type strain was less than 93%. Thus, the two strains are members of a novel species that diverged from P. psychrophila. DSV-1 and MS-1 are cold tolerant; both grow at 4°C but faster at 24°C. Unlike the mesophilic MnOB P. putida GB-1, both strains are capable of robustly oxidizing Mn at low temperatures. Both DSV-1 and MS-1 genomes contain homologs of several Mn oxidation genes found in P. putida GB-1 (mnxG, mcoA, mnxS1, mnxS2, and mnxR). Random mutagenesis by transposon insertion was successfully performed in both strains and identified genes involved in Mn oxidation that were similar to those found in P. putida GB-1. Our results show that MnOB can be isolated from compost, supporting a role for Mn oxidation in plant waste degradation. The novel isolates Pseudomonas spp. DSV-1 and MS-1 both can oxidize Mn at low temperature and likely employ similar mechanisms and regulation as P. putida GB-1. IMPORTANCE Biogenic Mn oxides have high sorptive capacity and are strong oxidants. These two characteristics make these oxides and the microbes that make them attractive tools for the bioremediation of wastewater and contaminated environments. Identifying MnOB that can be used for bioremediation is an active area of research. As cold-tolerant MnOB, Pseudomonas sp. DSV-1 and MS-1 have the potential to expand the environmental conditions in which biogenic Mn oxide bioremediation can be performed. The similarity of these organisms to the well-characterized MnOB P. putida GB-1 and the ability to manipulate their genomes raise the possibility of modifying them to improve their bioremediation ability.

to degrade complex organics to digestible byproducts; the minerals themselves can serve as reservoirs of organic carbon (8,9).Work from Yu and Leadbetter has shown that some species of bacteria can derive energy directly from the thermodynamically favorable oxidation of Mn (10).
In the environment, Mn oxidation may play a role in the breakdown of plant material.The concentration and redox state of Mn in leaf litter strongly correlate with the rate of litter decomposition (11), and the ability of forest ecosystems to store carbon is negatively correlated with Mn concentration (12).A significant component of plant litter is the cell wall component lignin.Lignin is a large, three-dimensional polymer of phenylpropanoid subunits; its large size and irregular structure render it, especially, difficult to degrade enzymatically (13).However, several species of fungus and bacte ria are capable of lignin degradation (13)(14)(15).These organisms employ both laccase enzymes and a variety of heme-containing peroxidase enzymes, including Mn perox idase.A major mechanism by which lignin-degrading enzymes work is through the production of soluble Mn(III) species via an oxidation reaction (15,16).
Biogenic Mn oxides (BMO) and Mn-oxidizing bacteria (MnOB) are actively being investigated for their possible applications in bioremediation due to the highly reactive and sorptive nature of the BMO.BMO generated by E. coli cells genetically modified to express a non-blue laccase from Bacillus sp.GZB have been shown to degrade the endocrine disruptor bisphenol A (17).BMO from the naturally Mn-oxidizing strain Pseudomonas sp.QJX-1 can degrade the herbicide glyphosate, and the bacteria can use the resulting breakdown products as a carbon, phosphate, or nitrogen source (18).Oxidation of pollutants is not the only mechanism of bioremediation by BMO.They have also been shown to remove arsenic from wastewater through precipitation of metal arsenates or adsorption on ferromanganese minerals (19).Breakdown of 17α-ethinyles tradiol (EE2) by BMO was increased 15-fold by the presence of the MnOB Pseudomonas putida MnB1 (20).Thus, optimal bioremediation may require living MnOB, not just the oxides they produce, making it important to identify MnOB that can thrive under a variety of growth conditions.
One of the best studied MnOB is Pseudomonas putida GB-1.This gram-negative gamma-proteobacterium has been shown to possess three genes encoding Mn oxidase enzymes that each appear to oxidize Mn(II) to Mn(IV).Two of the oxidases belong to the multi-copper oxidase family of enzymes, encoded by the genes mnxG and mcoA (21).The third oxidase, MopA, is an animal heme peroxidase (22).Mn oxidation in this species is also dependent on a two-component regulatory pathway comprising two sensor kinases -MnxS1 and MnxS2-and a σ 54 -dependent response regulator MnxR (23).Regulation of Mn oxidation in this species appears to be linked to the motile vs biofilm lifestyle switch since the deletion of regulatory gene fleQ results in altered Mn oxidation (22,24).
If a physiological function of Mn oxidation is the breakdown of recalcitrant organic carbon (ROC) for use as a food source, MnOB would be predicted to be found in areas with high concentrations of ROC, such as a compost pile.Sampling a compost pile on the campus of the University of Minnesota, Morris successfully resulted in the isolation of two MnOB strains, Pseudomonas sp.DSV-1 and MS-1.Both strains exhibit cold-tolerant growth and manganese oxidation down to 4°C.Genome sequence shows the two strains are very similar and have genes identified as important for Mn oxidation in the wellcharacterized MnOB Pseudomonas putida GB-1.We further demonstrate that the two strains can be genetically manipulated, illustrating the possibility of using these strains for bioremediation and studies of the evolution of Mn oxidation in the pseudomonads.

Isolation and identification of DSV-1 and MS-1
To test the prediction that Mn oxidation allows bacteria to degrade plant matter, samples were taken from a compost pile consisting of an approximate ratio of 75% plant material and 25% food waste (Troy Otsby, personal communication).Three samples were taken from the surface of the pile and at depths of ~15 cm and ~30 cm.From an initial set of ~10 Mn-oxidizing candidate species, four isolates were purified, and isolates MS-1 and DSV-1 were chosen for further characterization.At 24°C on solid media, both develop the brown colony color indicative of Mn oxidation (Fig. 1).The presence of Mn oxides was confirmed using leucoberbelin blue [LBB; data not shown, (25,26)].This oxidation behavior is similar to the well-characterized MnOB P. putida GB-1 (GB-1, Fig. 1).To tentatively identify the isolates, their 16S rRNA gene was amplified by colony PCR and sequenced.The resulting sequences were 100% identical to each other, and the best match in GenBank was to Pseudomonas psychrophila type strain E-3 (27) (NR_028619.1, 99% coverage, 99.66% identity).

Phylogenetic tree of Pseudomonas genus
To further investigate the relationship of DSV-1 and MS-1 to other pseudomonads, multi-locus sequence alignment was performed to construct a phylogenetic tree comparing 88 species of the Pseudomonas genus and the new isolates (Table 1).The species chosen represent members of each of the major clades in the genus: the fluorescens, aeruginosa, and pertucinogena lineages (28).MS-1 and DSV-1 group with P. psychrophila and P. fragi in the P. fragi group within the fluorescens lineage (Fig. 2).This lineage not only contains other known MnOB, P. putida GB-1, P. entomophila L48, and P. fluorescens PfO-1, but also species not known to oxidize Mn, including P. syringae pv.tomato str DC3000 (21).

Complete genome sequence of MS-1 and DSV-1
Using a combination of Illumina and Nanopore sequencing approaches, complete genome sequences were generated for MS-1 (NCBI accession #: JAYMYF000000000) and DSV-1 (NCBI accession #: JAYMYG000000000).The MS-1 genome is smaller than DSV-1 (5.3 vs 5.7 Mb), with fewer predicted genes (Table 2).Genetic relatedness and species attribution can be determined using the genome-wide average nucleotide identity (gANI) and alignment fraction (AF) metrics (29)(30)(31).Using tools available at the Integrated Microbial Genomes and Microbiomes website at the Joint Genome Institute (https://img.jgi.doe.gov/)(32), it was determined that the DSV-1 and MS-1 pairwise gANI is 99.1% and 88.6% for AF (Table 3), supporting their identification as two strains within the same species.However, their best match to a P. psychrophila strain was P. psychrophila CF149 with an gANI of 92.6%-92.8%and an AF of 85.8%-87.4% (Table 5).Because a gANI of >95% is commonly used to assign strains to the same species, DSV-1 and MS-1 have diverged sufficiently from P. psychrophila to be considered members of a different species.

Identification of putative Mn oxidation genes
Using the genome sequences, it was possible to identify orthologs of Mn oxidation genes from the well-characterized MnOB P. putida GB-1.This organism encodes three separate Mn oxidase enzymes in its genome, mnxG, mcoA, and mopA (21,22).MS-1 and DSV-1 both carry homologs to mnxG and mcoA but not mopA (Table 4).Three genes predicted to encode parts of a two-component regulatory pathway-mnxS1, mnxS2, and mnxR-are also essential for Mn oxidation in P. putida GB-1 (23).Each of these genes are present in both DSV-1 and MS-1 (Table 4).
The spacing between the genes and their orientation on the P. putida GB-1 chromo some suggests that mnxG and mcoA represent the first gene of two operons, respec tively, while mnxS1, mnxS2, and mnxR form a third putative operon (23) (Fig. 3).All six of the putative mnxG operon genes are found in both DSV-1 and MS-1, and they are found in the same orientation and organization on the chromosome (Table 4; Fig. 3).The mcoA putative operon contains five genes in P. putida GB-1 (Fig. 3); however, only mcoA and the gene immediately downstream, a predicted SCO1/SenC copper chaperone, are conserved in DSV-1 and MS-1 (Table 4; Fig. 3).The genome organization of mnxS1/S2/R is also somewhat conserved between GB-1, MS-1, and DSV-1, except that the mcoA SCO1/  SenC gene pair of the putative mcoA operon is located in the space between mnxS1 and mnxS2 in both MS-1 and DSV-1 (Fig. 3).

Pseudomonas sp. MS-1 and DSV-1 growth at low temperature
Because of the close association of DSV-1 and MS-1 to P. psychrophila (Fig. 2), a psychro philic species that can grow at temperatures as low as −1°C (27), we compared the ability of DSV-1 and MS-1 to grow at low temperature to that of the model MnOB P. putida GB-1.At 24°C, GB-1 grew somewhat faster than either strain (Table 5), although it reached a lower final optical density (Fig. 4).Growth slowed for all three strains at 14°C (Fig. 4), with GB-1 still doubling at a slightly faster rate than MS-1 or DSV-1 (Table 5).However, at 4°C, GB-1 grew very slowly, with roughly a 24-h doubling time (Fig. 4).Both MS-1 and DSV-1 grew detectably at this low temperature, with doubling times of 7.5 and 8.1 h, respectively (Table 5).Therefore, MS-1 and DSV-1 are capable of growth at low temperature but grow more slowly than at more moderate temperatures.During growth at 4°C, the optical density of the MS-1 culture dropped dramatically once the culture reached stationary phase.This could suggest a defect in survival at low temperature for this strain.However, the Mn-oxidizing MS-1 4°C cultures began to form aggregates once they reached stationary phase (data not shown), so much of this decrease in optical density may be due to this aggregation.

Pseudomonas MS-1 and DSV-1 oxidize Mn at low temperature
The growth curve experiments were performed in the presence of reduced Mn, so it was possible to observe that all three strains accumulated Mn oxides during the course of the experiment (data not shown).To verify this observation, each strain was incubated on solid Lept media at 24°C, 14°C, and 4°C (Fig. 5).After 5 days, at 24°C, all three strains grew and oxidized Mn, as seen by the brown colony color.At 14°C, again all three strains oxidized Mn, with GB-1 producing a lighter brown color than MS-1 or DSV-1.At 4°C, both  MS-1 and DSV-1 produced brown colonies, but GB-1 produced barely detectable growth.After 10 months, GB-1 had still failed to form detectable Mn oxides, while DSV-1 and MS-1 continued to grow and accumulate Mn oxides.

Genetic manipulation of DSV-1 and MS-1
The ability to genetically manipulate DSV-1 and MS-1 would make it possible to use these strains to investigate low-temperature Mn oxidation and generate strains optimized for bioremediation at low temperature.As a first step, we screened both strains for antibiotic sensitivity.DSV-1 and MS-1 are both resistant to ampicillin and penicillin but sensitive to gentamicin and kanamycin (data not shown).The selective medium Pseudomonas Isolation Agar (PIA, CRITERION, Hardy Diagnostics) is often used to isolate the Pseudomonas strains from environmental samples and during triparen tal mating (23).The basis of this selection is the presence of the broad-spectrum antimicrobial drug triclosan, which inhibits fatty acid synthesis.Pseudomonas spp.are naturally resistant to triclosan due to the presence of the FabV alternative fatty acid synthesis enzyme (33).However, neither DSV-1 nor MS-1 possesses a fabV homolog in their genomes, and neither strain can grow on PIA (data not shown).
Conjugation with E. coli is routinely used to introduce foreign DNA into P. putida GB-1 (22,23).To demonstrate that conjugation can be used to move plasmids into DSV-1 and MS-1, we performed triparental conjugation to move the plasmids pBR322MCS-5 and pUCP22 (Table 7) into both strains.These plasmids both carry aacC1, the gentami cin-resistant marker gene; the successful transfer of the plasmid into DSV-1 and MS-1 was detected by the ability of the conjugants to grow on media containing gentamicin (data not shown).P. putida GB-1 can be made chemically competent and made to take up plasmids by heat shock transformation (23); a similar approach was successful with MS-1 but has not yet been tried with DSV-1.To demonstrate the ability to generate mutations in the DSV-1 and MS-1 genomes, we conjugated into each strain the plasmid pRL27, which encodes the Tn5 transposon carrying a Kan R -resistant gene marker (Table 7).This plasmid has an oriR6K origin of replication and therefore requires the presence of the pir gene on the chromosome in order to be maintained as a plasmid (34).Since DSV-1 and MS-1 lack the pir gene, the only way to obtain Kan R colonies after conjugation is if the Tn5 transposon carrying Kan R has transposed into the chromosome.The ability to isolate Kan R colonies after conjugation into DSV-1 and MS-1 (data not shown) therefore demonstrates that these strains can be manipulated by insertion of Tn5 into the chromosome.
After successfully isolating kanamycin-resistant colonies, the colonies were screened for their Mn oxidation phenotype, and 13 mutant isolates were identified with altered Mn oxidation activity (11 in the DSV-1 strain and 2 in MS-1).Mapping the site of insertion revealed that several different genes had been targeted by transposition of Tn5 (Table 6).The oxidation phenotypes ranged from slight increase (KG271) to slight decrease (KG274) to no oxidation (KG272, 277, and 278; Fig. 6A).In the non-oxidizing strain KG278 (Fig. 6A), the transposon was inserted into the gene rpoN (Table 8), which encodes the alternative sigma factor σ 54 .To verify that the oxidation defect of KG278 is due to the rpoN::Tn5 mutation, complementation was performed using a plasmid carrying the P. putida GB-1 rpoN gene (pKG228, Table 7).Complementation was successful; pKG228 restored Mn oxidation to rpoN::Tn5 (Fig. 6B).

DISCUSSION
In this work, novel MnOB were isolated from compost, supporting a possible role for Mn oxidation in the breakdown of complex organic molecules.Both MS-1 and DSV-1 were shown to grow and oxidize Mn at low temperature.A complete genome sequence and phylogenetic characterization showed these strains to be closely related to Pseudomonas psychrophila but genetically distinct.Therefore, they have been named Pseudomonas sp.DSV-1 and MS-1.Both strains are amenable to genetic manipulation and carry, in their genomes, genes homologous to those previously identified as important for Mn oxidation in P. putida GB-1.

Low-temperature growth
DSV-1 and MS-1 both grow well at 4°C but grow faster at 24°C (Fig. 4; Table 7).Psychro philic organisms are commonly defined as those that grow best at temperatures below 20°C and thus are confined to environments that are continuously cold.Conversely, psychrotrophic or psychrotolerant organisms grow best at 20°C or above but grow well at temperatures below 20°C (39).Given this definition, DSV-1 and MS-1 are best described as psychrotolerant.Cold-tolerant species have previously been identified in the genus Pseudomonas from environments as diverse as Antarctic sea ice and food spoilage (40,41).For example, Pseudomonas psychrophila HA-4 was isolated by its ability to degrade the antibiotic sulfamethoxazole at low temperature (42), and Pseudomonas fragi strains were isolated from the leaves of cold-adapted plants (43).MS-1 and DSV-1 are closely related to P. fragi and P. psychrophila (Fig. 2).Thus, the cold tolerance of Pseudomonas strains isolated from a compost pile located outside in Minnesota in winter was not unexpected.

Low-temperature Mn oxidation
While MnOB have generally been characterized as mesophiles (35,44), Mn oxidation at low temperature has been observed before.Brevibacillus brevi MO1 has been shown to oxidize Mn at 4°C but not to the same extent as it does at 37°C (45).Arthrobacter sp.NI-2 normally oxidizes Mn at 30°C; a mutation in this strain allows it to oxidize at 10°C (46).The dormant spores of Bacillus sp.SG-1 are capable of producing Mn oxides over a very wide range of temperatures, from 0°C to 80°C (47).Pseudomonas sp.MOB-449 grows well and exhibits its maximum Mn oxidation capacity at 18°C (48).At this low temperature, Mn stimulates biofilm growth and expression of the c-type cytochrome biosynthesis enzyme CcmE, leading to the proposal that the Mn oxidation supplements the cell's energy needs (49).Thus, while P. psychrophila DSV-1 and MS-1 are not the only MnOB capable of low-temperature Mn oxidation identified so far, they are the first characterized that actively grow and robustly oxidize at temperatures as low as 4°C.

Conservation of Mn oxidation mechanism
Many of the genes identified as playing a role in Mn oxidation in P. putida GB-1 are also present in DSV-1 and MS-1.Each has orthologs to the Mn oxidase genes mnxG and mcoA but lack clear orthologs to mopA (Table 4).This suggests that Mn oxidation in these strains depends on the multi-copper oxidases MnxG and McoA but not the heme peroxidase MopA.DSV-1 and MS-1 also carry orthologs to the Mnx two-component regulatory pathway comprising MnxS1, MnxS2, and MnxR.MnxR in P. putida GB-1 is required for Mn oxidation and is predicted to be a σ 54 -dependent transcription factor, based on its domain composition (23).The MnxR orthologs in MS-1 and DSV-1 are also predicted to contain σ 54 interaction domains.This suggests that the expression of Mn oxidation genes is driven by RNA polymerase containing σ 54 in all three strains.Supporting this conclusion, a Tn5 insertion in the predicted rpoN gene of MS-1 resulted in a strain completely defective for Mn oxidation when assayed on solid media (Fig. 6) and in liquid culture (data not shown).This oxidation defect could be complemented with the GB-1 rpoN gene, reinforcing the conclusion that Mn oxidation in this strain is σ 54 dependent.
Previous work has shown that Mn oxidation in P. putida GB-1 can be disrupted by Tn5 insertions in genes encoding components of the TCA cycle, including the succi nate dehydrogenase complex (sdhABC), lipoate acetyltransferase (aceA), and isocitrate dehydrogenase (icd) (50).Insertion of Tn5 into the fumarate hydratase class I gene of DSV-1 resulted in moderately decreased Mn oxidation (KG274, Fig. 6); fumarate hydratase catalyzes the conversion of fumarate to malate in the TCA cycle.KG266-269 all have Tn5 inserted in a predicted thiol-disulfide isomerase (Table 8).In Bradyrhizobium japonicum, a similar protein called TlpA is involved in cytochrome c oxidase maturation (51).In DSV-1, the gene is in a putative operon between dsbD and dsbG genes, raising the possibility of polar effects on these neighboring genes.In Shewanella oneidensis, DsbD facilitates the transfer of electrons to the protein CcmG during the cytochrome c maturation (CCM) process (52).In P. putida MnB1, the CCM genes ccmA, E, and F have previously been identified as playing a role in Mn oxidation (50), and CcmE has been implicated in low-temperature Mn oxidation in Pseudomonas sp.MOB-449 (49).Thus, the function and regulation of Mn oxidation in the new isolates are likely similar to that in other Mn-oxidizing pseudomonads.

Low-temperature bioremediation
There are many potential applications for MnOB and biogenic Mn oxides in bioremedia tion.Cold-tolerant bacteria and their enzymes are also valuable tools for bioremediation and other industrial applications (42,57,58).Therefore, Pseudomonas ssp.MS-1 and DSV-1 expand the conditions under which MnOB can be used for bioremediation due to their ability to form Mn oxides at low temperature.Our preliminary results suggest the two strains differ in the effect of temperature on their ability to accumulate oxidized Mn.As judged by the intensity of brown oxides formed, MS-1 robustly formed Mn oxides at all three temperatures tested, while DSV-1 best formed oxides at the intermediate temperature of 14°C (Fig. 5).MS-1 also tolerates growth at temperatures above 24°C better than DSV-1 (data not shown), which suggests this strain will be the better target for bioremediation applications.
At cold temperatures, bacteria experience stress due to decreased membrane fluidity, decreased enzyme activity, altered redox state, and increased stability of RNA and DNA structures, which interfere with replication and gene expression (59)(60)(61).The MS-1 and DSV-1 genomes are very similar to one another (Tables 2 and 3); comparing these genomes may make it possible to determine the genetic basis for their differences in oxidation and temperature sensitivity phenotypes.Preliminary characterization of cold shock genes in GB-1, DSV-1, and MS-1 (Table 2 and data not shown) failed to reveal a genetic basis for the cold tolerance of DSV-1 and MS-1 since all three genomes possess six putative cold shock protein genes (cspA).Both strains can be made to take up foreign DNA by conjugation and transformation; they can express foreign genes from plasmids and can have their genomes mutated with a transposon.The apparent conservation of Mn oxidation and its regulation between the new isolates and the well-characterized MnOB P. putida GB-1 will guide future efforts to generate cold-tolerant strains optimized for Mn oxidation under various conditions.

Media and culture conditions
Strains and plasmids used in this study are listed in Table 7. Pseudomonas strains were grown in LB or Lept liquid and solid media made according to the procedure of reference (25).Strains were grown at 24°C, 14°C, or 4°C.Escherichia coli strains were grown in LB medium at 37°C.The following concentrations of antibiotics were used: ampicillin (100 µg/mL), gentamicin (50 µg/mL), and kanamycin (30 µg/mL).For oxidation assays, MnCl 2 was added to Lept medium at a final concentration of 100 µM.Phosphate-buf fered saline was made according to standard protocols (62).

Sample collection
Samples for cultivation were collected from a compost pile on the University of Minnesota, Morris campus that is composed of a 3:1 ratio of plant material to food waste (Ostby, Personal Communication).Samples were taken in February 2019 using sterile, plastic 50 mL tubes.The tubes were opened and immediately used to scoop material from the compost surface, ~15 cm, or ~30 cm below the surface.After collection, the tubes were sealed, immediately transported back to the lab, and stored at 4°C.Further more, 1 g of sample was incubated in PBS pH 7.3 for 5 min at RT, with shaking.The PBS/ compost mixture was allowed to settle 10-15 min, and then 100 µL of the supernatant was spread onto Lept plates.After incubation at RT for 7 days, thousands of colonies were visible, with a subset of brown, putative Mn-oxidizing colonies.Mn oxidation was confirmed using a leucoberbelin blue spot test (25).LBB-positive colonies were selected and subcultured onto fresh Lept lates.After several rounds of re-streaking, DSV-1 and MS-1 were shown to be pure via microscopic observation.

Identification of isolates by 16S rRNA sequencing
To obtain 16S amplicons from our bacterial sample, colony PCR was run using iProof High-Fidelity DNA Polymerase using the following concentrations of reagents: 200 µM dNTP mix, 1 µM forward primer, 1 µM reverse primer, 0.5 U of iProof High-Fidelity DNA Polymerase per 50 µL reaction, 10 µL of 5× iProof HF Buffer per 50 µL reaction, and 1 µL of overnight culture in NB broth as the DNA template source.Primers 8F and 519R (Table 8) were used to generate an ~500 bp amplicon.Reaction conditions were initial denaturation at 98°C for 3 min followed by 25 cycles of 98°C for 30 s, 55°C for 1 min, 72°C for 1 min, followed by 72°C for 5 min.
PCR amplicons were purified using DNA Clean & Concentrator-5 according to manufacturer's instructions (Zymo Research, Irvine CA).The concentration of DNA samples was determined using a Qubit 3.0 Fluorometer from manufacturer Invitro gen (Carlsbad, CA).Furthermore, 50 ng of 500 bp length sequences and 200 ng of 1,500 bp sequences were added to new tubes for sequencing by the University of Minnesota Genomics Center (http://genomics.umn.edu/).Also included in the samples were 6.4 pmol of the appropriate primers (Table 2).Short amplicons were sequenced using 8F and 519R, while the long amplicons were sequenced with 8F, 519R, 1492R, 533F, and CDR (Table 2).
Amplicons of the 16S SSU gene were conjoined using GeneStudio (https://source forge.net/projects/genestudio/) to produce a consensus sequence of 1,467 bp.This consensus sequence was then used to query the 16S ribosomal RNA (bacteria and archaea) database using BLASTN (63,64).

DNA extraction and genome sequencing
Cultures were grown on solid R2A medium, and a single colony was transferred to 10 mL of tryptic soy broth and grown for 48 h with shaking.Five milliliters of each culture were then centrifuged for 10 min at 2,000 × g in a swinging bucket rotor, and the supernatant was removed.The cell pellets were then resuspended in 0.5 mL of sterile PBS pH 7.4 (Gibco-Thermo Fisher, Waltham MA), and DNA was extracted using the QIAamp UCP Pathogen Kit (QIAGEN, Germantown MD) following the standard protocol with the final elution in molecular biology grade water.Purified DNA was quantified using a Qubit 4 fluorometer using dsDNA HS Assay (Invitrogen-Thermo Fisher, Waltham MA).Illumina sequencing was performed using the Nextera DNA Flex Library prep following the standard protocol and sequenced with a 600-cycle MiSeq v3 Reagent Kit (Illumina, San Diego, CA).Long-read sequencing was performed using a 1D2 R9.2 Sequencing Kit on an Oxford Nanopore Minion sequencer (Oxford Nanopore, New York, NY).Read coverage was approximately 120× for Illumina sequencing and 30× for Nanopore sequencing.Illumina reads were trimmed for quality, and adapters were removed using Trimmomatic V0.39 (65).Illumina and Nanopore reads were then used to assemble the genomes using the Unicycler assembly pipeline V0.4.8 (66) with Spades V3.13.0 (67).

Generation of Pseudomonas phylogenetic tree
Assembled genomes for 88 Pseudomonas species were downloaded from NCBI, and 4 housekeeping genes (16S rRNA, rpoB, rpoD, and gyrB) were extracted from each assembled genome [as in reference (28)].These four genes were concatenated and aligned with Cellvibrio japonicus as an outgroup using default parameters in MAFFT [version 7; (68)].This alignment was used to build a phylogenetic tree with RAxML [v.8.2.11; (69)].The rapid bootstrapping and search for best scoring ML tree approach was used with 1,000 replicates, the input was partitioned by each gene, and a GTR Gamma nucleotide model was implemented.All of the above took place within the Geneious Prime (v.2019.1.1)interface.

Growth curves, growth rate, and doubling times
Growth rates and doubling times were calculated using spectrophotometry at a wavelength of 600 nm.Cultures of GB-1, DSV-1, and MS-1 were grown overnight in 5 mL of Lept media with continuous agitation at 240 rpm at a temperature of 24°C.Subcultures of each strain were prepared in triplicate by diluting the overnight cultures 100-fold into 50 mL of Lept media.These subcultures were then grown at 24°C, 14°C, and 4°C with continuous agitation at 240 rpm.Furthermore, 1 mL samples were taken periodically to determine optical density using a spectrophotometer.

Transposon mutagenesis
The plasmid carrying the transposon Tn5, pRL27, was moved into DSV-1 and MS-1 by triparental conjugation (23) and transconjugants selected by plating on LB containing 30 µg/mL kanamycin and 100 µg/mL ampicillin.Colonies were replica plated onto solid Lept media to screen for variations in the manganese oxidation phenotype.Selected MS-1 and DSV-1 mutants were streaked for single colonies on Lept media and compared to wild type to confirm the variation in their manganese-oxidizing capabilities.

Mapping site of transposon insertion
Some Tn5 insertion sites were mapped according to the protocol of reference (24) with the following exceptions.Genomic DNA was isolated using the Wizard Genomic DNA Purification Kit (Promega, Madison, WI).Five micrograms of purified gDNA were digested with BamHI in a 50 µL reaction overnight at 37°C.The digested DNA was ethanol precipitated, and 100 ng was ligated using T4 DNA ligase (New England BioLabs, Ipswich, MA) in a 20 µL reaction overnight at room temperature.The ligation reactions were then transformed into E. coli GT115 commercially made competent cells (Invivogen CHEMICOMP GT115, Fisher Scientific).LB agar with Km was used to select E. coli cells transformed with the plasmid containing Tn5 and the BamHI fragment of the genome.Plasmids were purified from Km R colonies using the QIAprep Spin Miniprep Kit (Qia gen,Valencia, CA).Purified plasmids were sent to be sequenced at Functional Biosciences (https://functionalbio.com/) using primers tpnRL17-1 and tpnRL13-2 (Table 2).The sequenced genes were identified using a BLAST search against the relevant genome database on the Integrated Microbial Genomes website (http://img.jgi.doe.gov/)(70).
The remaining Tn5 insertion sites were mapped using an arbitrary PCR approach (55).Genomic DNA was prepared as above.Three reactions were performed for each mutant, each using tpnRL17-1 as the forward primer, and ARB1, ARB2, or ARB3 as the reverse primer (Table 2).Furthermore, 1 µL of genomic DNA was used as template, 1× Promega GoTaqG2 Hot Start Green Master Mix, and 0.8 µM final concentration primers in total volume of 25 µL.PCR conditions were as follows: 1 cycle of 95°C 5 min, 6 cycles of 94°C 30 s, 30°C 30 s, 72°C 2 min followed by 30 cycles of 94°C 30 s, 45°C 30 s, 72°C 2 min followed by 72°C 5 min and then stored at 4°C.One microliter of this reaction was used as template in a second PCR with ARB4 and tnp5IR-2R as the primers, 0.8 µM final concentration, and 1× Promega GoTaq G2 Hot Start Green Master Mix in a total volume of 30 µL.PCR conditions were as follows: 1 cycle of 95°C 5 min, 30 cycles of 94°C 30 s, 55°C 30 s, 72°C 2 min followed by 72°C 5 min and then stored at 4°C.Reactions were separated on a 1% low melt agarose gel; the prominent band from each reaction was excised using a razor blade and stored at 4°C.One microliter of liquid from the excised band was used as template for a third round of PCR with the same primers and conditions as the second PCR.The DNA from both the excised gel band and the third PCR was cleaned using the GeneJET Gel Extraction and DNA Cleanup Micro Kit (ThermoScientific) and was sent to Functional Biosciences (https://functionalbio.com/) to be sequenced using primer Tnp5IR-2R (Table 2).

Construction of the rpoN plasmid
The rpoN gene was PCR amplified from Pseudomonas putida GB-1 using primers rpoN_1 F and rpoN_2 R (Table 2), with a high-fidelity DNA polymerase (Phusion HotStart highfidelity DNA polymerase).The resulting PCR product was cloned into pJET1.2/blunt(CloneJet PCR Cloning Kit; Fermentas, Glen Burnie, MD).The genes were subsequently subcloned into the broad host-range plasmid pUCP22 (Table 1) using the EcoRI and XbaI restriction enzyme recognition sites engineered into the amplification primers, and the presence of the insert in the resulting plasmid was confirmed by restriction digest.The genes inserted into pUCP22 are expressed constitutively from the plasmid-borne promoter P lac .

Transformation of Pseudomonas sp. MS-1 and derivatives
Pseudomonas sp.MS-1 and the MS-1 rpoN::Tn5 mutant were made competent as follows.Bacteria were grown overnight in LB and subsequently diluted 25-fold into fresh LB and grown at room temperature for 4 h.Furthermore, 2 mL of cells was pelleted by centrifugation at 12,000 × g for 1 min and then washed with 1 mL of ice cold 0.1 M CaCl 2 .Cells were then pelleted and resuspended in 1 mL ice cold 0.1M CaCl 2 and incubated on ice for 30 min.Finally, cells were pelleted and resuspended in 100 µL ice cold CaCl 2 .Transformation was performed by adding 2 µL plasmid to the cells and incubating on ice for 30 min.Next cells were exposed to heat shock for 90 s at 37°C and then returned to ice for 2 min.Then, 400 µL SOC medium was added to each transformation, which were then incubated at room temperature with shaking for 1 h.Finally, the entire transformation was plated onto LB Gm plates and incubated at room temperature.

FIG 2
FIG 2 Phylogenetic tree of genus Pseudomonas.A total of 88 species of the genus Pseudomonas are represented by this tree, which uses a concatenated sequence of the 16S rRNA gene, rpoB, rpoD, and gyrB to construct proposed evolutionary relationships.Known MnOB are highlighted in red.

FIG 3
FIG 3 Conservation of putative Mn oxidation operons.Arrows represent predicted genes.Numbers below the arrows represent the number of base pairs (bp) between predicted genes; numbers above the arrows are the length of the predicted protein product in amino acids (aa).Written within the arrows are the gene names and/or IMG gene ID# for the gene.Genes in the putative mnxG operon are red, genes in the mcoA operon are green, and those in the mnx two-component regulatory pathway are blue.GB-1, Pseudomonas putida GB-1; MS-1, Pseudomonas sp.MS-1; DSV-1, Pseudomonas sp.DSV-1.

FIG 4
FIG 4 Growth of P. putida GB-1, and P. psychrophila strains DSV-1 and MS-1 at various temperatures.Datapoints represent the average of three replicates; error bars are the standard deviation.(A) 24°C, (B) 14°C, and (C) 4°C.After 80 h of growth at 4°C, cellular aggregation in the MS-1 culture made it difficult to measure OD 600 .

TABLE 1
Accession numbers of strains used in phylogenetic tree

TABLE 1
Accession numbers of strains used in phylogenetic tree(Continued)

TABLE 3
Average nucleotide identity and alignment fraction

TABLE 4
Putative Mn oxidation gene orthologs in DSV-1 and MS-1

TABLE 5
Growth rates and doubling times

TABLE 6
Tn5 mutations in DSV-1 and MS-1 a Cytochrome c maturation.

TABLE 7
Plasmids and strains a Amp R , ampicillin resistance; Kan R , kanamycin resistance; Gm R , gentamicin resistance.

TABLE 8
Primer list