Bacteria use a catabolic patchwork pathway of apparently recent origin for degradation of the synthetic buffer compound TRIS

Abstract The synthetic buffer compound TRIS (2-amino-2-(hydroxymethyl)propane-1,3-diol) is used in countless applications, and no detailed information on its degradation has been published so far. Herein, we describe the discovery of a complete bacterial degradation pathway for TRIS. By serendipity, a Pseudomonas strain was isolated from sewage sludge that was able to grow with TRIS as only carbon and nitrogen source. Genome and transcriptome analyses revealed two adjacent gene clusters embedded in a mobile genetic element on a conjugative plasmid to be involved in TRIS degradation. Heterologous gene expression revealed cluster I to encode a TRIS uptake protein, a TRIS alcohol dehydrogenase, and a TRIS aldehyde dehydrogenase, catalyzing the oxidation of TRIS into 2-hydroxymethylserine. Gene cluster II encodes a methylserine hydroxymethyltransferase (mSHMT) and a d-serine dehydratase that plausibly catalyze the conversion of 2-hydroxymethylserine into pyruvate. Conjugational plasmid transfer into Pseudomonas putida KT2440 enabled this strain to grow with TRIS and with 2-hydromethylserine, demonstrating that the complete TRIS degradation pathway can be transmitted by horizontal gene transfer. Subsequent enrichments from wastewater purification systems led to the isolation of further TRIS-degrading bacteria from the Pseudomonas and Shinella genera carrying highly similar TRIS degradation gene clusters. Our data indicate that TRIS degradation evolved recently via gene recruitment and enzyme adaptation from multiple independent metabolic pathways, and database searches suggest that the TRIS degradation pathway is now globally distributed. Overall, our study illustrates how engineered environments can enhance the emergence of new microbial metabolic pathways in short evolutionary time scales.


Introduction
Since the beginning of the industrial revolution hundreds of thousands of new pesticides, herbicides, medical drugs, and other industrial chemicals were synthesized to meet the growing demand of newly emerging agricultural, pharmaceutical, and industrial applications [1].Many of these anthropogenic chemicals eventually end up in the environment, where some of them are removed quickly by existing biological degradation pathways [2], whereas others persist and accumulate in the environmental due to the absence of efficient degradation mechanisms [2][3][4].In recent decades, more and more degradation pathways for such persistent chemicals that were introduced in the early 20th century were discovered in microbes, suggesting that functional mineralization pathways for recalcitrant compounds can evolve within decades after the initial exposure of the microbial community [5][6][7][8].Two prominent examples are the anthropogenic herbicide atrazine and the artificial sweetener acesulfame, for which environmental accumulation and marginal degradation rates were reported in the years after their introduction ( [3,4] and references therein).However, after several decades of intense human use, increasing evidence for efficient biological degradation pathways for these compounds started to pile up, suggesting that these pathways evolved within a time frame of 30-50 years [3,4].In habitats exposed to such chemicals, novel pathways can evolve as a result of sufficient positive selective pressure for either their detoxification or their use as a new substrate source [2].Typically, novel degradation pathways emerge in a process called patchwork assembly by the recruitment, adaptation, and mutation of one or more genes encoding promiscuous enzymes from other metabolic pathways, either from within the host organism via gene duplication or via horizonal gene transfer (HGT) from the microbial community [1,2,9].If the mutations happen to increase the substrate affinity and catalytic activity toward the respective new substrate and if the new catalytic activity enables the host to transform the new substrate into metabolites of an already existing metabolic pathway, it will gain a selective advantage over competing microbes by being able to use this compound as a new substrate and energy resource [1,2].The genetic information encoding such novel catabolic pathways is often located on mobile genetic elements, such as plasmids or transposons, enabling transfer of the novel pathway to other microbes via HGT [5,10,11].Eventually, positive selection for the new catabolic trait can lead to a loss of mobility and manifestation of the new pathway in the genome of its host.However, the environmental concentrations of synthetic compounds may be too low to provide enough selection pressure for the evolution of novel catabolic pathways.Especially in wastewater treatment plants where the availability of other, readily degradable organic compounds is high, many low-concentrated synthetic compounds remain in the purified wastewater as so-called micropollutants [12,13].Even some synthetic and natural compounds for which efficient biodegradation pathways exist, such as steroid hormones, are sometimes not fully metabolized in wastewater treatment plants due to their low concentration and contribute to the micropollutant pool.To enhance the biodegradation of such micropollutants, bioaugmentation of water treatment facilities is increasingly being considered, which implies the establishment of microorganisms in the water purification processes that have been isolated for their ability to efficiently degrade the respective organic compounds [14,15].
The study presented here was originally set up to isolate bacteria capable of degrading steroid hormones in low concentrations from activated sludge.To increase the chances of isolating strains that can later be established for bioaugmentation in real wastewater treatment plant conditions, we used an artificial wastewater medium to mimic typical wastewater nutrient conditions.To our surprise, our enrichment procedure resulted in the isolation of a bacterial strain that did not grow with the provided steroid hormones, but with the synthetic medium buffer TRIS (2-amino-2-(hydroxymethyl)propane-1,3-diol).Herein, we pursued this finding further and present the identification and characterization of a bacterial catabolic pathway for TRIS, for which no detailed information on its degradation was available so far [16].

Materials and methods
Additional materials and methods can be found in the Supplementary material.[17], and 10 mM TRIS.For solid media plates, 1.5% agar was added prior to autoclaving.Enrichment cultures (5 ml) were inoculated with fresh samples (1 ml) from local wastewater treatment plants (activated sludge or activated carbon), a water purification plant (activated carbon), or different freshwater habitats (Table S1).When enrichment cultures showed growth, 100 μl of the culture supernatant were transferred to 5 ml fresh minimal medium with TRIS, and this was repeated two to three times before cultures were spread onto solid minimal medium plates containing TRIS as sole carbon, energy, and nitrogen source to isolate individual TRIS-degrading colonies.Growth experiments in liquid cultures were carried out at 30 • C and 200 rpm in 1 ml or 5 ml medium in 24-well plates or in reaction tubes, respectively.Experimental cultures were inoculated from washed, over-night starter cultures grown with TRIS.Growth was followed by measuring the optical density at 600 nm (OD 600 ).For transcriptome experiments, TRIS was replaced with 20 mM pyruvate and 20 mM NH 4 Cl were added as nitrogen source.For substrate tests, TRIS was replaced with 10 mM d-or l-serine, 10 mM choline, 10 mM betaine, 10 mM 2-amino-2-methylpropanol, 10 mM 2-amino-2-methylproandiol, 10 mM sarcosine, or 10 mM serinol, and 20 mM NH 4 Cl were added as nitrogen source.Pseudomonas putida KT2440 and its derivatives were grown in the same minimal medium with succinate as carbon source at 30 • C. Escherichia coli strains were grown in halfconcentrated lysogeny broth (LB) at 37 • C. If required, kanamycin (20-50 μg/ml), tetracycline (10 μg/ml), or gentamycin (20-40 μg/ml) antibiotics were added to the media after autoclaving.To test whether 2-hydroxymethylserine (2-HMS) can be used as a growth substrate for strain Teo1, cell suspensions of an E. coli pBBR1MCS-2::tupA_taoB_ taoA (Table S2) strain that had produced 2-HMS from TRIS were centrifuged and filter-sterilized (0.2 μm), and the culture supernatant was inoculated with strain Teo1 cells.

TRIS quantification
TRIS, its degradation intermediates, and ammonium were identified and quantified by high-performance liquid chromatography (HPLC) coupled to a mass spectrometer and an ultraviolet/visible (UV/VIS) detector [18] after derivatization of their amino group with diethyl ethoxymethylenemalonate (DEEMM, [19]).For this, 100 μl culture supernatants were mixed with 130 μl borate buffer (1 M, pH 9.0), 75 μl methanol, and 3 μl DEEMM reagent, followed by 30 min sonication and 60 min incubation at 70 • C before injection into the HPLC.A gradient (f low rate 0.3 ml/min) of ammonium acetate buffer (10 mM, pH 3.0, with 0.1% (v/v) formic acid, eluent A) and acetonitrile (eluent B) were used for product elution starting with 10% eluent B for 2 min, increasing to 90% eluent B within 22 min, remaining at 90% eluent B for 2 min, and returning to 10% eluent B within 1 min, followed by an equilibration of 5 min.Products were identified based on their mass spectra after electrospray ionization in positive ion mode.The capillary voltage was 2500 V, the end plate offset 500 V and the capillary temperature 300 • C. Nebulizer pressure was 22.5 psi, dry gas f low 12 l/min, and dry temperature 200 • C. Quantification was performed based on the product peak areas at 280 nm UV absorbance, and concentrations were calculated based on concentration curves of authentic standards treated in the same way.The concentration of 2-HMS was calculated using the same standard curve as for TRIS.

Transcriptome analysis
Pseudomonas hunanensis Teo1 was grown in minimal medium with TRIS as only carbon and nitrogen source or with pyruvate and ammonium and cultures were harvested during mid-log phase.Harvested cells were resuspended in 800 μl RLT buffer (RNeasy Mini Kit, Qiagen) with β-mercaptoethanol (10 μl ml −1 ) and cell lysis was performed using a laboratory ball mill.Subsequently, 400 μl RLT buffer (RNeasy Mini Kit Qiagen) with βmercaptoethanol (10 μl ml −1 ) and 1200 μl 96% (vol./vol.)ethanol were added.For RNA isolation, the RNeasy Mini Kit (Qiagen) was used as recommended by the manufacturer, but instead of RW1 buffer, RWT buffer (Qiagen) was used in order to isolate RNAs smaller than 200 nucleotides also.For sequencing, the strandspecific cDNA libraries were constructed with a NEB Next Ultra II Directional RNA library preparation kit for Illumina and the NEB Next Multiplex Oligos for Illumina (New England BioLabs, Frankfurt am Main, Germany).Sequencing was performed by using the HiSeq2500 instrument (Illumina Inc., San Diego, CA, USA) using the HiSeq Rapid SR Cluster Kit v2 for cluster generation and the HiSeq Raid SBS Kit (50 cycles) for sequencing in the single-end mode and running 1 × 50 cycles.Raw reads have been deposited in the Sequence Read Archive (SRR25447477-SRR25447482).More detailed information can be found in the supplementary material.

Conjugation experiments
Conjugation experiments were carried out using strains Teo1 and Teo8 as donors and gentamicin-resistant P. putida KT2440::eyfp-gm as recipient.Approximately 1 × 10 9 donor cells and 3 × 10 9 recipient cells were mixed and spotted onto sterile cellulose acetate filter disks (∼3 cm 2 , 0.45 μm pores) on LB agar plates.Conjugation plates were incubated at 30 • C for 24 h before cells were washed off the filter disks with 50 mM phosphate buffer (pH 7.2).Decimal dilutions were plated onto minimal medium agar plates with TRIS as only carbon and nitrogen source containing 40 μg/ml gentamicin.Transconjugants were screened using a ChemiDocTM Imaging System (Bio-Rad) using blue epi illumination and an emission filter at 530 nm.The conjugation rate was calculated by dividing the number of positive transconjugants by the number of recipient cells in each assay.The successful transfer of the p1_Teo1 and p2_Teo8 plasmids was confirmed by colony PCR amplifying the tupA, taoA, taoB, glyA, and dsdA genes (Table S3).To test whether P. putida KT2440::eyfp-gm and its transconjugant carrying the p1_Teo1 plasmid can use 2-HMS as a growth substrate, 2-HMS-containing E. coli supernatant was prepared as described above.

Isolation and characterization of the TRIS-degrading strain P. hunanensis Teo1
The first TRIS-degrading organism was isolated from an enrichment culture that was originally designed to isolate testosteronedegrading bacteria from activated sludge (Supplementary material).We repeatedly observed growth in the enrichment cultures in TRIS-buffered artificial wastewater medium without removal of the provided carbon source testosterone (not shown).Plating of the enrichment cultures onto solid medium containing testosterone resulted in the isolation of a bacterial strain that was not able to grow with testosterone but showed growth in liquid artificial wastewater medium without the addition of any extra carbon or nitrogen source.Subsequent growth experiments in phosphate buffered minimal medium with different concentrations of TRIS confirmed a linear relationship between the provided TRIS concentration and the amount of biomass produced by the isolate (Fig. 1A), suggesting that TRIS was the growth substrate in these cultures.Growth experiments coupled with HPLC-MS analysis confirmed that TRIS was completely removed from the culture medium, suggesting its utilization as a carbon and nitrogen source for growth (Fig. 1B-D).The isolate had an exponential growth rate of 0.36 ± 0.09 h −1 and a doubling time of 0.44 ± 0.09 h (n = 6).The only detectable metabolite accumulating in the medium was ammonium, with around 40% of the nitrogen introduced with the TRIS substrate being released as ammonium into the medium (Fig. 1C and D).The strain did not grow with the amino alcohols serinol, 2-amino-2-methyl-1-propanol, and 2-amino-2methyl-1,3-propanediol (Fig. S1A), whereas it did grow with the quaternary ammonium compounds choline and betaine, and the secondary amine sarcosine, and with the amino acids d-and lserine, but not with glycine.Based on its whole-genome sequence (Supplementary material), the isolate was taxonomically classified as a representative of the species P. hunanensis and was named strain Teo1 (Table S1).The genome of strain Teo1 consists of a circular chromosome (6.13Mbp, GenBank ID CP131127.1 ) and two large, circular plasmids, p1_Teo1 (156.3 kbp, GenBank ID CP131128.1)and p2_Teo1 (65.0 kbp, GenBank ID CP131129.1 ).

Identification of TRIS degradation genes
TRIS growth experiments and time-resolved analysis of TRIS degradation in cell suspensions of strain Teo1 showed that TRIS degradation was induced in cells pregrown with TRIS, whereas it was not induced in pyruvate pregrown cells (Fig. S2), suggesting that TRIS catabolism is regulated at the transcriptional level in strain Teo1.Based on this observation, we used transcriptomics of TRIS-grown and pyruvate-grown Teo1 cells to identify potential TRIS degradation genes.A total of 398 genes were differentially regulated in TRIS-compared to pyruvate-grown cells (fold change (FC) > 4, P adj < 0.05, Fig. 1E, Dataset SD1), of which 65 were upregulated in TRIS-grown cells.Eleven of the 12 most upregulated genes in TRIS-grown cells were found in two adjacent gene clusters (Fig. 1E-G) localized on the p1_Teo1 plasmid.The proteins encoded in cluster I were annotated as a solute-binding protein, an amino acid/polyamine transporter, a choline dehydrogenase (CDH), a NAD/NADP-dependent aldehyde dehydrogenase, a (1)-pyrroline-2-carboxylate/ (1)-piperideine-2-carboxylate reductase, and an alpha-ketoglutaric semialdehyde dehydrogenase (Table S2).The proteins encoded in cluster II were annotated as a serine hydroxymethyltransferase, a d-serine dehydratase, two proteins presumably involved in the transformation of methylene tetrahydrofolate into tetrahydrofolate and formate, and a d-serine transporter.The clusters are separated by a gene that encodes a helix-turn-helix-type transcriptional regulator.Based on the similar upregulation of their genes, both gene clusters appear to be expressed as operons (Fig. 1G).Due to their strong upregulation and their colocalization in putative operons, we proposed that these genes or a subset of them are involved in TRIS degradation in strain Teo1.

Isolation of further TRIS-degrading bacteria and genomic analysis of TRIS degradation
Subsequent targeted enrichment of TRIS-degrading bacteria resulted in the isolation of seven Pseudomonas strains (Teo2, Teo3, Teo4, Teo6, Teo8, Teo10, and Teo11) and one Shinella zoogloeoides strain (Teo12) (Table S1).All isolates were able to grow with TRIS as only carbon and nitrogen source and completely removed TRIS from the culture medium (Fig. S3).The Pseudomonas strains Teo2, Teo4, Teo6, Teo8, and Teo11 and the Shinella isolate strain Teo12 were isolated from different municipal wastewater treatment plant samples, Pseudomonas sichuanensis Teo3 from an activated charcoal basin of a water purification plant, and Pseudomonas sp.strain Teo10 from a freshwater lake sample.Enrichment cultures from two additional wastewater treatment plants and from five freshwater habitats did not develop growth (Table S1).
We sequenced the genomes of five Pseudomonas isolates (Supplementary material) and found homologous TRIS degradation gene clusters to the ones identified in strain Teo1 in all of these strains with protein similarities >98.5% (Fig. 2, Fig. S4).In Pseudomonas strains Teo2, Teo4, Teo6, and Teo8, the clusters are also localized on plasmids (p1_Teo2, p1_Teo4, p1_Teo6, and p2_Teo8), and the plasmids have high similarities to p1_Teo1 (Fig. 2A, minimum identity 99.99%, minimum coverage 87%).In Pseudomonas strain Teo3, the clusters are encoded in the chromosome (Fig. 2B).Whereas all gene cluster II sequences are highly similar (minimum similarity 99.6 5, minimum coverage 86%), only the amino acid/polyamine transporter, the CDH, and the NAD/NADP-dependent aldehyde dehydrogenase genes are conserved in all cluster I sequences (Fig. 2 and Fig. S4), suggesting that the solute-binding protein, the (1)-pyrroline-2-carboxylate/ (1)piperideine-2-carboxylate reductase-like protein, and the alphaketoglutaric semialdehyde dehydrogenase-like protein are not required for TRIS degradation.In all these strains, the TRIS degradation gene clusters are f lanked by sequences similar to known bacterial insertion elements ( Fig. 2B), suggesting that the TRIS degradation gene clusters are located within a mobile genetic element [11].A search in the NCBI nucleotide database identified one Priestia genome (accession CP065422.1)and three Pseudomonas genomes (CP069081.1,CP097105.1,CP114115.1)with homologous gene clusters (Fig. 2A, Fig. S4, protein similarities >98.5%).In Priestia, and two of the Pseudomonas strains, the clusters are also plasmid-borne, whereas they are encoded chromosomally in Pseudomonas sp.strain GXZC (Fig. 2B).
We also sequenced the genome of the Shinella isolate (Supplementary material), which consists of one circular chromosomes (3.54 Mbp, Genbank ID CP131130.1 ) and three circular plasmids, p1_Teo12 (1.89 Mbp, Genbank ID CP131131.1 ), p2_Teo12 (221.54 kbp, Genbank ID CP131132.1 ), and p3_Teo12 (140.01 kbp, Genbank ID CP131133.1 ), which have no apparent similarity to the TRIS degradation-encoding Pseudomonas plasmids.However, a DNA region on p2_Teo12 encodes homologs of all TRIS degradation genes in two adjacent gene clusters with protein similarities between 64.9% and 77.9%, except for the genes encoding the solute binding protein, the amino acid/polyamine transporter, the (1)-pyrroline-2-carboxylate/ (1)-piperideine-2-carboxylate reductase-like protein, the alpha-ketoglutaric semialdehyde dehydrogenase-like protein, and the d-serine transporter ( Fig. 2B, Fig. S4).In these clusters, the CDH and the NAD/NADP-dependent aldehyde dehydrogenase homologs are encoded in the same orientation as in cluster I of the Pseudomonas TRIS degradation gene clusters, whereas the genes are differently oriented in cluster II.
A homologous regulator gene to the one found in the Pseudomonas gene clusters is located upstream of this cluster.A search in the NCBI nucleotide database identified a closely related Rhizobium oryzihabitans strain M15 (CP048637.1)carrying a homologous DNA segment on a plasmid (p5, Fig. 2B).The predicted TRIS degradation gene clusters in S. zoogloeoides Teo12 and strain M15 are separated by a truncated transposase gene and f lanked by additional transposase genes and insertion elements.

Characterization of TRIS degradation reactions
Based on the predicted protein functions encoded in the upregulated gene clusters, we hypothesized that TRIS is initially oxidized by the proteins encoded in cluster I. To test this, we cloned different combinations of the genes encoding the putrescine importer-like protein, the CDH-like protein, and the aldehyde dehydrogenase from Pseudomonas extremaustralis strain Teo8 into the expression plasmid pBBR1MCS-2 and transformed the plasmids into E. coli BL21 (DE3) (Supplementary material).Then we tested high-density cell suspensions of the resulting strains for TRIS transformation activity.The E. coli strain carrying all three genes transformed all available TRIS into a new compound within 14 h (Fig. 3A).This compound was identified as 2-HMS(Fig.3B) based on its molecular weight, which was 14 Da heavier than that of TRIS, indicating the oxidation of one hydroxymethyl group into a carboxylic group.We tested whether 2-HMS can be further utilized by strain Teo1, using filter-sterilized E. coli culture supernatants containing 2-HMS as growth medium for Teo1, confirming that strain Teo1 can grow with and fully degrade 2-HMS (Fig. S1B).Small amounts of 2-HMS were also formed when only the two dehydrogenase genes were expressed in E. coli, whereas strains carrying combinations of the transporter gene and the individual dehydrogenase genes Figure 2. TRIS degradation gene clusters in the genomes of the sequenced TRIS-degrading isolates from this study and of homologous gene clusters found in the NCBI nr nucleotide database.(A) Most TRIS degradation gene clusters are located on large circular plasmids with high sequence similarities to the pseudomonas plasmids pND6-2 and pS04-90, which lack the TRIS degradation clusters.The outer black line indicates the location of the putative mobile genetic element including the TRIS degradation genes.Like pND6-2 and pS04-90, the TRIS degradation-encoding plasmids encode orthologs to mobile elements groups (inner circle) including a type IVB secretion system and a type IV pilus for plasmid transfer as well as a type II toxin-antitoxin system.(B) All TRIS degradation gene clusters, including plasmid-and chromosome-encoded clusters, are f lanked upstream and downstream by transposon-or insertion element-like structures.The plasmid p2_Teo12 from Shinella sp.strain Teo12 and the homologous plasmid p5 in a rhizobium strain have no similarity to the other TRIS degradation-encoding plasmids but also carry two gene clusters which encode homologs to the other TRIS degradation proteins.

Figure 3. TRIS transformation activity of E. coli strains carrying different combinations of the predicted TRIS degradation genes encoded in clusters I
and II on the expression plasmid pBBR1MCS-2.(A) Representative HPLC-UV/VIS chromatograms after 72 h of incubation and TRIS quantification in cell suspension supernatants of E. coli strains expressing different combinations of the tupA, taoB, and taoA genes in TRIS-containing medium.TRIS (peak I, red symbols) was completely transformed into a compound (peak II, orange symbols) in the E. coli strain that carries all three genes and in minor amounts in the E. coli strain that carries the taoA and taoB genes.The cloned genes were under the activity of the lac promotor (gray arrow).Average and standard deviation of three independent biological replicates are shown.(B) Mass spectra of TRIS (peak I) and its transformation product (peak II).The new product was identified as 2-HMS based on the molecular mass of its DEEMM (gray molecule moiety) derivative of 305 Da, which is 14 Da heavier than the molecular mass of 291 Da of the DEEMM derivative of TRIS.(C) Representative HPLC-UV/VIS chromatograms after 48 h of incubation and 2-HMS quantification in cell suspension supernatants of coli strains expressing cluster II with and without the dsdX d-serine transporter gene.2-HMS was completely degraded in the E. coli strain that carries the full cluster II and ammonium (gray bar) was released.The cloned genes were under the activity of the lac promotor (gray arrow).Average and standard deviation of three independent biological replicates are shown.
showed no significant TRIS transformation activity ( Fig. 3A).Thus, we propose that these genes encode a TRIS uptake protein (TupA), and two TRIS-oxidizing proteins, namely a TRIS alcohol dehydrogenase (TaoA) and a TRIS aldehyde dehydrogenase (TaoB), which together produce 2-HMS.Based on the predicted protein functions encoded in cluster II, we further hypothesized that these proteins are responsible for further 2-HMS degradation.
To test this, we cloned all genes of cluster II into pBBR1MCS-2 and transformed the plasmid into E. coli.The resulting E. coli strain degraded all available 2-HMS within 24 h, accompanied by an increase in free ammonium and no transformation products were detectable (Fig. 3C).When we cloned all genes of cluster II except for the dsdX gene, the resulting strain showed no 2-HMS transformation activity (Fig. 3C).This supports our hypothesis that the genes in cluster II are responsible for the degradation of 2-HMS and shows that externally supplied 2-HMS needs to be translocated into the cells by the predicted d-serine transporter.
A phylogenetic analysis of the TaoA proteins with closely related proteins of the glucose-methanol-choline (GMC) oxidoreductase family (InterPro IPR012132) revealed that the TaoA proteins form a unique cluster that diverges from a clade containing CDH protein sequences (Fig. S5A).In contrast, other CDH proteins of the TRIS-degrading isolates encoded outside of the TRIS degradation gene clusters group within the CDH clade, except for two CDH proteins from S. zoogloeoides Teo12, which are located at the base of the TaoA cluster.A phylogenetic analysis of the TaoB proteins with closely related proteins containing an aldehyde dehydrogenase domain (InterPro IPR015590) showed that the TaoB proteins also group together (Fig. S5B).TaoB proteins are most closely related to the aldehyde dehydrogenase PuuC (Accession P23882), which is involved in putrescine catabolism in E. coli and to other PuuC-like proteins encoded outside of the TRIS degradation gene clusters of the TRIS-degrading isolates.

Conjugational transfer of TRIS degradation plasmids
The TRIS degradation-encoding Pseudomonas plasmids have high similarities to the previously described large plasmids pND6-2 from P. putida ND6 and pS04-90 from Pseudomonas aeruginosa S04-90 [20,21].Both these plasmids lack the putative mobile genetic element that encodes TRIS degradation (Fig. 2A and Fig. S4).However, pS04-90 carries a carbapenem resistance-encoding class 1 integron at almost the same site in the pND6-2 backbone as the TRIS degradation-encoding elements in the Pseudomonas plasmids.Both pND6-2 and pS04-90 are conjugative plasmids, which can be transferred from their hosts into other Pseudomonas strains.Like pND6-2 and pS04-90, p1_Teo1 and its homologs carry genes encoding type IV pili and a type IVB secretion system (Fig. 2A and Fig. S4), suggesting that these plasmids are also conjugative plasmids.To test this, we set up conjugation experiments with strain Teo1 and strain Teo8 as plasmid donors and P. putida KT2440 as a non-TRIS-degrading recipient.To be able to differentiate between donor and recipient cells, we labeled strain KT2440 with a chromosomally integrated gentamycin resistance associated with a yellow f luorescent protein (YFP) gene.After the conjugation of Teo1 and Teo8 with strain KT2440::eyfp-gm, growth of f luorescing colonies was observed on gentamycin-containing minimal medium plates with TRIS as only carbon and nitrogen source (Fig. 4A-B), suggesting that conjugational transfer of the p1_Teo1 and p2_Teo8 plasmids also transferred the ability to use TRIS as growth substrate to KT2440::eyfp-gm.The resulting transconjugants were also able to grow with TRIS in liquid culture (Fig. 4C), confirming that both gene clusters together encode the complete degradation pathway sufficient to use TRIS as sole carbon and nitrogen source.Additionally, the transconjugant KT2440::eyfp-gm p1_Teo1 was also able to grow with the TRIS degradation intermediate 2-HMS, which was not the case for the recipient strain KT2440::eyfp-gm (Fig. 4D), suggesting that conjugational transfer of the p1_Teo1 plasmid also transferred the ability to use 2-HMS as growth substrate.

Discussion
In this study, we identify and characterize a degradation pathway for the synthetic buffer compound TRIS in several Pseudomonas strains and a Shinella strain, which were all isolated from wastewater purification systems with TRIS as only carbon source.Our results suggest that TRIS degradation proceeds via a pathway that originated recently via a combination of genes from different, formerly independent metabolic pathways and their assembly into two coregulated gene clusters within a mobile genetic element.Although not all individual functions for the proteins encoded in clusters I and II were biochemically confirmed, conjugational transfer of the plasmid containing both clusters into P. putida KT2440 cotransferred the ability to degrade TRIS as well as its degradation intermediate 2-HMS, confirming that they together encode the complete degradation pathway sufficient to use TRIS as sole carbon and nitrogen source.Furthermore, heterologous expression of both gene clusters in E. coli confirmed their individual involvement in TRIS degradation.
The bonding of the amino group of TRIS to a tertiary carbon atom prevents its metabolism via transamination or oxidative deamination, which are the most common reactions in the degradation of amines.However, bacterial degradation of α-amino acids with a tertiary α-carbon atom is known for a long time and has been demonstrated for 2-HMS via a tetrahydrofolate-dependent mSHMT with d-serine and formaldehyde as products [22].All of our TRIS-degrading isolates encode a protein in cluster II with highest similarities to confirmed damino acid-producing mSHMTs [23] as well as a protein for the degradation of d-serine to pyruvate, suggesting that these proteins together can catalyze the degradation of 2-HMS, the product of TRIS oxidation by the TaoAB proteins.Thus, we propose a pathway for the catabolism of TRIS, in which TRIS is translocated into the cytoplasm by the TRIS uptake protein TupA and subsequently oxidized to 2-HMS by the TRIS alcohol dehydrogenase TaoA and the TRIS aldehyde dehydrogenase TaoB via a suspected aldehyde intermediate (Fig. 5).2-HMS is further transformed into d-serine by the mSHMT protein, which is then transformed into pyruvate by the d-serine dehydratase.The latter can be channeled into the tricarboxylic acid cycle via oxidative decarboxylation to acetyl-CoA.The formaldehyde-accepting tetrahydrofolate (THF) cofactor of the mSHMT protein can then be recycled by the bifunctional methylenetetrahydrofolate dehydrogenase/methenyltetrahydrofolate cyclohydrolase FolD and the formyltetrahydrofolate deformylase PurU encoded in cluster II [24,25].
The genetic context of the tupA, taoA, and taoB genes is not found outside of cluster I, suggesting that these genes were assembled from multiple origins.The TaoA proteins have highest similarities to CDH, which catalyze the oxidation of choline into betaine aldehyde [26] and belong to the GMC oxidoreductase family.Within a phylogenetic GMC protein tree, the TaoA proteins form a distinct clade that diverges from the CDH sequences with strong bootstrap support (98%, n = 1000), suggesting that the TRIS alcohol dehydrogenase functionality might have evolved from enzymes with CDH activity.Strikingly, two CDH proteins encoded outside of gene cluster I in S. zoogloeoides Teo12 localize basal to the TaoA protein clade, suggesting that they could be ancestral to TaoA.It is tempting to speculate that the TaoA proteins have originated in Shinella or a similar organism through the duplication and mutation of CDH-encoding genes.TupA and TaoB have highest similarities to proteins involved in the degradation of polyamines like putrescine [27,28], suggesting that these proteins might have been recruited from bacterial polyamine degradation or similar pathways [29].The fact that Teo1 was not able to use other α-substituted amino alcohols such as 2-amino-2-methylpropanol and 2-amino-2-methylproandiol as growth substrates suggests that either the substrate uptake by TupA or the oxidation reactions catalyzed by TaoAB evolved to be highly specific for TRIS.
Expression of the TRIS uptake protein encoded by tupA was strictly required for efficient TRIS transformation by TaoA and TaoB in E. coli cells, suggesting that TRIS translocation is required for TRIS uptake in Gram-negative cells.With a pKa of 8.07 TRIS is mainly positively charged at neutral and acidic pH, preventing its diffusion across the cytoplasmic membrane.The absence of a tupA gene homolog in S. zoogloeoides Teo12 suggests that this gene was recruited later to the pathway.It remains unknown how TRIS is translocated by Teo12; however, alternative putrescine uptake transporters [27] are also encoded in the genome of Teo12, which might play a role in TRIS uptake in this strain.It is not known why TRIS transformation activity was not detectable when taoA or taoB were expressed individually with tupA in E. coli.One possible reason could be that TaoA and TaoB are only active as a protein complex that allows substrate channeling between the two proteins, which might explain why the predicted aldehyde intermediate was never detected in these assays.Degradation of TRIS (dashed lines) and growth (solid lines) of P. putida KT2440::Eyfp-gm p1_Teo1 and p2_Teo8 transconjugants in liquid medium.The recipient strain P. putida KT2440::Eyfp-gm without the conjugated plasmids was not able to grow with or degrade TRIS.(D) Degradation of 2-HMS (dashed lines, open symbols) and growth (solid lines) of the P. putida KT2440::Eyfp-gm p1_Teo1 transconjugant in liquid medium.The recipient strain P. putida KT2440::Eyfp-gm without the conjugated plasmid was not able to grow with 2-HMS.
The genetic context of the genes encoded in cluster II is also present in other Pseudomonas strains that do not encode cluster I, suggesting that this gene cluster may also play a role in the degradation of other compounds.Whereas SHMT and mSHMT proteins are often accompanied by THF cofactor recycling proteins [ 23,30], their genetic association with d-amino acid degradation genes such as the d-serine dehydratase gene dsdA and the d-serine transporter gene dsdX in cluster II in a putative operon was not reported so far.In pathogenic bacteria, d-serine dehydratases and transporters play important roles in the detoxification and catabolism of d-amino acids [31,32].Our analysis suggests that the transporter protein DsdX is strictly required for the uptake of 2-HMS in Gram-negative cells; however, it is not yet known whether DsdX is actively involved in the native TRIS degradation pathway.Based on our findings, we hypothesize that the second part of the TRIS degradation pathway might have been assembled by a combination of genes involved in the transformation of α-substituted amino acids and in the degradation of d-amino acids.The presence of mostly incomplete insertion elements upand downstream of the TRIS degradation gene clusters suggests that the whole DNA segment was originally able to be mobilized as a composite transposon, allowing its transposition between plasmids and the chromosome.The finding of an additional transposase gene between the two gene clusters in Shinella and the closely related Rhizobium strain corroborates our hypothesis that the pathway was assembled from two independent metabolic routes encoded in the respective gene clusters I and II.
TRIS is used in countless applications in research, industry, medicine, and household settings.Its most prominent applications include its use as a buffer in molecular biology and biochemistry labs, as an excipient in drug formulations, and as an active agent for the treatment of metabolic acidosis and an enhancer for antibiotics and antiseptics for the treatment of Gram-negative infections.Based on its first large scale production in the middle of the twentieth century [33,34], we hypothesize that the TRIS degradation pathway evolved within the last 30 to 50 years, which is corroborated by the fact that utilization of TRIS as a growth substrate for bacteria has not been reported until the 1990s [16], despite its intense use in research laboratories.The organization of the TRIS degradation genes in two compact gene clusters and the fact that expression of these presumable operons was regulated in response to TRIS in strain Teo1 are signs for a highly evolved pathway [6], suggesting its turn-around in the environment for a while.The regulator gene between the two clusters is presumably involved in this regulation.Its protein product is similar to the transcription factor GltC, which regulates glutamate synthase expression in Bacillus subtilis [35].
Specific microbial habitats such as sinks and wastewater treatment facilities of large research laboratories, hospitals, or from the TRIS-manufacturing industry could be locations where the TRIS degradation pathway might have originated.First, such habitats encounter significant TRIS concentrations on a regular basis.For example, the antibiotic fosfomycin, which is produced and administered as a TRIS salt, was found at concentrations of 0.13 mM in the inf luent and 0.015 mM in the eff luent of a wastewater treatment plant of a fosfomycin production plant [36], suggesting the presence of similar concentrations of its counterion TRIS.Second, sinks and water purification systems are hotspots for HGT [11,37,38], which additionally favors the evolution of novel catabolic pathways.
The identification of identical TRIS degradation gene clusters in bacteria isolated in Germany, the Republic of Korea, and in China corroborates our findings, that TRIS degradationencoding plasmids can be readily transferred among bacteria by conjugation-driven HGT.Our isolation studies suggest that the TRIS degradation pathway is frequently present in the microbiome of municipal wastewater treatment plants but less frequent in freshwater habitats, indicating that environments with high carbon and nitrogen loads better support the maintenance and dissemination of the TRIS degradation pathway than more oligotrophic environments.Although the average TRIS concentration in the sampled wastewater treatment plants is presumably very low, its metabolism as an additional nitrogen and carbon source alongside other recalcitrant molecules might provide sufficient positive selection pressure for bacteria with broad substrate ranges such as Pseudomonas and Shinella [39][40][41][42] to maintain the pathway.
Overall, our results provide a further example of how HGT and mobile genetic elements can contribute to the evolution of novel catabolic pathways and illustrate how engineered environments such as wastewater treatment plants can enhance the emergence of new microbial metabolic pathways.

Figure 1 .
Figure 1.Characteristics of TRIS metabolism in P. hunanensis Teo1 and identification of potential TRIS degradation genes.(A) Maximal biomass production of strain Teo1 in relation to TRIS concentration.Error bars indicate standard deviation (n = 3).(B) Growth of strain Teo1 with 10 mM TRIS as only carbon and nitrogen source (n = 6).(C) Removal of TRIS (straight lines) and formation of ammonium (dashed lines) in Teo1 growth cultures.(D) Representative HPLC chromatograms of a strain Teo1 culture growing with 10 mM TRIS.TRIS is depleted from the culture while ammonium is released.(E) Volcano plot of transcriptome data showing significantly upregulated genes (adjusted p value < 0.05) under TRIS-grown (positive FC) and pyruvate-grown conditions (negative FCs) in Teo1.Genes predicted to be involved in TRIS degradation are labeled.(F) Transcript abundances and log 2 (FCs) of the predicted TRIS degradation genes in Teo1 in TRIS-and pyruvate-grown cells (three biological replicates of each growth substrate, horizontal bar indicates median, corrected P values of all selected genes were <0.05).(G) mRNA reads (upper panel: positive strand, lower panel: negative strand) of a representative TRIS-grown culture mapped onto the DNA section containing the predicted TRIS degradation gene clusters in strain Teo1.

Figure 4 .
Figure 4. Conjugative transfer of the TRIS degradation pathway-encoding plasmids p1_Teo1 and p2_Teo8.(A) Growth of YFP-f luorescing colonies on TRIS-containing agar plates after conjugation of P. putida KT2440::Eyfp-gm with Teo1 (left panel) and Teo8 (right panel) as donor strains.(B) Conjugation rate of KT2440::Eyfp-gm with p1_Teo1, p2_Teo8, and of a control without donor cells.Black lines indicate median rates of three independent conjugation experiments, boxes indicate the interquartile range, and dashed lines the maximum and minimum range.Transconjugants of up to seven replicates from three independent experiments were counted.Datapoints are colored based on the donor strain.Outliers are not filled.(C)Degradation of TRIS (dashed lines) and growth (solid lines) of P. putida KT2440::Eyfp-gm p1_Teo1 and p2_Teo8 transconjugants in liquid medium.The recipient strain P. putida KT2440::Eyfp-gm without the conjugated plasmids was not able to grow with or degrade TRIS.(D) Degradation of 2-HMS (dashed lines, open symbols) and growth (solid lines) of the P. putida KT2440::Eyfp-gm p1_Teo1 transconjugant in liquid medium.The recipient strain P. putida KT2440::Eyfp-gm without the conjugated plasmid was not able to grow with 2-HMS.

Figure 5 .
Figure 5. Proposed TRIS degradation pathway.The active involvement of the transporter protein DsdX in the degradation of TRIS is not yet known.mTHF, methenyltetrahydrofolate; fTHF, formyltetrahydrofolate.