Molecular Mechanism of N,N-Dimethylformamide Degradation in Methylobacterium sp. Strain DM1

DMF is a hazardous pollutant that has been used in the chemical industry, pharmaceutical manufacturing, and agriculture. Biodegradation as a method for removing DMF has received increasing attention. Here, we identified an efficient DMF degrader, Methylobacterium sp. strain DM1, and characterized the complete DMF mineralization pathway and enzymatic properties of DMFase in this strain. This study provides insights into the molecular mechanisms and evolutionary advantage of DMF degradation facilitated by plasmid pLVM1 and redundant genes in strain DM1, suggesting the emergence of new ecotypes of Methylobacterium.

ABSTRACT

N,N-Dimethylformamide (DMF) is one of the most common xenobiotic chemicals, and it can be easily emitted into the environment, where it causes harm to human beings. Herein, an efficient DMF-degrading strain, DM1, was isolated and identified as Methylobacterium sp. This strain can use DMF as the sole source of carbon and nitrogen. Whole-genome sequencing of strain DM1 revealed that it has a 5.66-Mbp chromosome and a 200-kbp megaplasmid. The plasmid pLVM1 specifically harbors the genes essential for the initial steps of DMF degradation, and the chromosome carries the genes facilitating subsequent methylotrophic metabolism. Through analysis of the transcriptome sequencing data, the complete mineralization pathway and redundant gene clusters of DMF degradation were elucidated. The dimethylformamidase (DMFase) gene was heterologously expressed, and DMFase was purified and characterized. Plasmid pLVM1 is catabolically crucial for DMF utilization, as evidenced by the phenotype identification of the plasmid-free strain. This study systematically elucidates the molecular mechanisms of DMF degradation by Methylobacterium.

IMPORTANCE DMF is a hazardous pollutant that has been used in the chemical industry, pharmaceutical manufacturing, and agriculture. Biodegradation as a method for removing DMF has received increasing attention. Here, we identified an efficient DMF degrader, Methylobacterium sp. strain DM1, and characterized the complete DMF mineralization pathway and enzymatic properties of DMFase in this strain. This study provides insights into the molecular mechanisms and evolutionary advantage of DMF degradation facilitated by plasmid pLVM1 and redundant genes in strain DM1, suggesting the emergence of new ecotypes of Methylobacterium.

INTRODUCTION

N,N-Dimethylformamide (DMF) is an important and versatile solvent used in the chemical industry, pharmaceutical manufacturing, and agriculture (1–4). The global consumption of DMF was approximately 285,000 metric tons in 2001, and there has been increasing demand in recent years (5). Large quantities of DMF are introduced into the environment through sewage discharge and waste disposal, and the strong water miscibility of this compound makes it difficult to remove from wastewater (6). In marine systems, DMF can be metabolized by microbes to dimethylamine (DMA) and methylamine (MA), which are known to be precursors of marine aerosols and important sources of greenhouse gases (7). DMF may also threaten aquatic animals in water bodies (8). In addition, long-term exposure to DMF, which is a volatile organic chemical, can harm human beings (9–11). In particular, it can cause chromosomal aberrations, gastric irritation, abnormal liver function, and damage to the respiratory system and can increase the risk of cancer (12–15).

Biodegradation of DMF has been established over the past 30 years. Several DMF-degrading bacteria, such as Pseudomonas (16), Paracoccus (6, 17–20), Alcaligenes (21), Ochrobactrum (22), Mycobacterium, and Methylobacterium (23, 24) strains, have been isolated and identified. However, except for Paracoccus aminophilus strains JCM7685 and JCM7686, little genomic information is available for these degraders. Methylobacterium is famous for its ability to use C₁ compounds such as methanol and MA as sole carbon and energy sources, but most Methylobacterium strains lack the ability to utilize DMF (24–26). Only one Methylobacterium aminovorans strain was reported previously to have DMF-degrading ability, and its genomic information and mechanism of DMF degradation have not been elucidated (23, 24). Current knowledge about the DMF catabolic pathways and related genes in Methylobacterium species is lacking.

DMF degradation pathways were studied. One common pathway begins with the hydrolysis of DMF by N,N-dimethylformamidase (DMFase), which yields DMA and MA (16, 22). DMA can be subsequently converted to MA by DMA dehydrogenase or DMA monooxygenase (DmmABCD) (27, 28). MA dehydrogenase encoded by the mau gene cluster then converts MA to formaldehyde and ammonia (29). MA can also be transformed to methylene tetrahydrofolate (CH₂=THF) through the N-methylglutamate pathway (NMGP), which consists of γ-glutamylmethylamide synthetase (GmaS), N-methylglutamate synthase (MgsABC), and N-methylglutamate dehydrogenase (MgdABCD) (30). The other pathway, which begins with two steps of successive demethylation, converting DMF into methylformamide and formamide, is speculated to be possessed by some Pseudomonas species (16). Two DMFases, one from Pseudomonas sp. strain DMF 3/3 and one from Alcaligenes sp. strain KUFA-1, were purified and characterized in previous studies (21, 31, 32). The protein DmfR, which is encoded on the plasmid pAMI2 in Paracoccus aminophilus JCM7686, is the only example of a regulator of the DMFase operon (19). Although the metabolites and main degradation pathways for DMF degraders are similar, they may vary in genetic composition among different genera. Therefore, it is important to study the pathways and genes in another genus to expand the DMF degradation gene and enzyme libraries. On the basis of existing genetic information, it would be of interest to identify the diverse genotypes of DMF degradation among different strains.

In this study, we report the isolation and characterization of Methylobacterium sp. strain DM1, an efficient DMF degrader that can utilize DMF as the sole source of carbon and nitrogen. Whole-genome sequencing and transcriptome sequencing (RNA-seq) data for this strain were analyzed (Fig. 1 and 2). The megaplasmid pLVM1 of the RepABC family was found to harbor the gene clusters that specifically facilitate the initial steps of DMF degradation. Putative genes and pathways responsible for mineralization of DMF were elucidated. We also investigated the enzymatic properties of DMFase, the initial enzyme of DMF degradation. Comparative genomic analysis with Methylobacterium species and other DMF degraders revealed the distinctive genetic diversity of DMF catabolic components and the genetic redundancy of strain DM1.

FIG 1.

Molecular mechanisms of DMF degradation in strain DM1. (A) Genomic analysis revealing the circular replicons in the genome of strain DM1. The genome contains a plasmid, named pLVM1. From the inside to the outside of the plasmid (left) and chromosome (right), the circles are as follows: circle 1, scale; circle 2, G+C skew; circle 3, G+C content; circles 4 and 7, clusters of orthologous groups; circles 5 and 6, locations of coding sequence, tRNA, and rRNA genes. (B) Clusters of putative DMF degradation genes in strain DM1. Genes are colored based on their different functions (pink, DMF hydrolysis; yellow, DMA oxidation; blue, MA utilization; green, NMGP; light gray, formaldehyde oxidation; dark gray, formate oxidation; light blue, serine cycle). The arrows indicate the sizes of the genes and the direction of transcription. The full names of the enzymes and corresponding genes are summarized in Table S1 in the supplemental material. (C) Putative DMF mineralization pathway and enzymes involved. The pathway is based on analogy with previously established pathways and bioinformatics identification of putative genes. The solid and dashed lines indicate pathways catabolizing via one step and multiple steps, respectively. The colors of the arrows correspond to the functional genes in panel B. (D) Transcriptional analysis by RT-qPCR of genes related to DMF degradation in strain DM1; 16S rRNA was used as the reference gene. The fold change in mauD (in the absence of DMF) was set to 1 as a reference, and the relative fold changes in the expression of other genes in the presence of DMF (colored bars) and in the absence of DMF (gray bars) were calculated according to this reference. The colors of the bars correspond to the functional genes in panel B. DMF, treatment group with 2,000 mg liter⁻¹ DMF induction; glycerol, control group. (E, F, and G) Identification of intermediates of DMF degradation. GC-MS mass spectra of benzene sulfonyl chloride derivatives of DMA (E), MA (F), and ammonium (G) generated in a resting cell reaction are shown.

FIG 2.

Transcriptome analysis of genes in strain DM1 in response to DMF. (A) Total numbers of genes showing upregulation or downregulation within the chromosome and the plasmid pLVM1. (B) Heat map of the hierarchical cluster analysis of gene expression in strain DM1. RNA_DMF, treatment group cultured with 2,000 mg liter⁻¹ DMF; RNA_Gly, control group cultured with 1% (vol/vol) glycerol. (C) Heat map of putative genes predicted to be involved in DMF degradation according to RNA-seq data. The formula used to obtain the data for drawing the heat map was as follows: log₂ basemean RNA_Gly − [(log₂ basemean RNA_Gly + log₂ basemean RNA_DMF)/2]. (D) KEGG pathway enrichment analysis of differentially expressed genes in response to DMF.

RESULTS AND DISCUSSION

Isolation and identification of efficient DMF-degrading strain DM1.

An aerobic, pink-pigmented bacterium was isolated from contaminated soil in the Shanghai Laogang industrial district and was named strain DM1. This strain is rod shaped, Gram negative, and able to form biofilms when grown in liquid medium. Strain DM1 is able to use DMF, DMA, MA, methanol, ethanol, formamide, toluene, glycerol, and trisodium citrate as the sole source of carbon or carbon and nitrogen (Fig. 3B and D). The 16S rRNA gene of strain DM1 shares the highest level of sequence identity (99%) with Methylobacterium sequences. Phylogenic analysis revealed that strain DM1 is most closely related to Methylobacterium suomiense and Methylobacterium aminovorans (see Fig. S1A in the supplemental material). Methylobacterium sp. strain DM1 has been deposited at the China Center for Type Culture Collection (CCTCC) under accession number CCTCC M2018663.

FIG 3.

Curing of the plasmid pLVM1 from Methylobacterium sp. strain DM1 and phenotype analysis. (A) Comparison of the growth abilities of DM1ΔpLVM1 and the wild-type strain on MSM agar plates containing 2,000 mg liter⁻¹ DMF. Quadrants 1 and 4 represent the wild-type strain, and quadrants 2 and 3 represent the plasmid-free strain DM1ΔpLVM1. (B) Effects of eliminating plasmid pLVM1 from strain DM1 on growth and degradation of 2,000 mg liter⁻¹ DMF in MSM. ▲, growth curve for DM1ΔpLVM1; ♦, DMF degradation curve for DM1ΔpLVM1; ■, growth curve for the wild-type strain; ●, DMF degradation curve for the wild-type strain. (C) PCR verification of plasmid curing. Genes dmfA1A2 and dmmD are plasmid-borne genes, and genes chr1 and chr2 are chromosomal genes. ΔP, DM1ΔpLVM1; WT, wild-type strain (positive control); M, molecular size markers. (D) Phenotype analysis of carbon or carbon and nitrogen source utilization of DM1ΔpLVM1 (−P), compared with the wild-type (WT) strain. The concentrations of substrates were as follows: 5 mg liter⁻¹ DMA, 5 mg liter⁻¹ MA, 3% (vol/vol) methanol, 3% (vol/vol) ethanol, 1% (vol/vol) toluene, 1% (vol/vol) formamide, 1% (vol/vol) glycerol, and 4% (vol/vol) trisodium citrate.

Strain DM1 was cultured with different initial concentrations of DMF, at different pH values and temperatures, to determine the optimal growth and degradation conditions. Optimal growth and DMF degradation were observed at 30°C, with 2,000 mg liter⁻¹ DMF, at pH 7.0 (Fig. S2). Under these conditions, 2,000 mg liter⁻¹ DMF could be completely degraded in 28 h (Fig. S1B). Inasmuch as the DMF concentration in synthetic wastewater was elevated to 2,000 mg liter⁻¹ (33), the outstanding degradation capability of strain DM1 indicates that potentially it can be applied in bioremediation of DMF-contaminated environments.

Multireplicon genome of strain DM1.

Whole-genome sequencing was performed, and analysis of the data revealed that the genome of strain DM1 has a multireplicon structure, with a single circular chromosome and a megaplasmid (Fig. 1A). The circular chromosome is 5,664,348 bp in size, with a G+C content of 68.3% and a coding percentage of 82.9%; it contains 5,312 protein-coding sequences, 5 rRNA operons, and 69 tRNA genes. The megaplasmid pLVM1 is 240,873 bp in size, with a G+C content of 67.4%; it contains 256 protein-coding genes. A putative repABC operon responsible for the replication and stable maintenance of the replicons is found within the megaplasmid. The proteins encoded by repA, repB, and repC (C0214_27440, C0214_27445, and C0214_27450, respectively) share 58%, 38%, and 43% amino acid sequence identity, respectively, with the corresponding proteins in Agrobacterium rhizogenes (34). Thus, the megaplasmid pLVM1 likely is a RepABC plasmid.

The genes encoding the initial steps of DMF degradation are predicted to colocalize on a 162.5-kbp region flanked by mobile elements in the 200-kbp plasmid (Fig. 1A and B). This region includes dmfA1A2 genes encoding enzymes for the hydrolysis of DMF, putative dmmABCD genes encoding enzymes for the oxidation of DMA, mau genes arranged in a large cluster that are presumably essential for growth with MA, and gmaS, mgsABC, and mgdABCD genes predicted to encode enzymes in the NMGP of MA utilization (Fig. 1B). Genes encoding annotated ammonium transporters and multidrug efflux pumps are dispersed within pLVM1. The remaining open reading frames (ORFs) are mobile elements and hypothetical proteins. Thus, pLVM1 presumably serves as a DMF catabolic plasmid facilitating the growth of Methylobacterium sp. strain DM1 with DMF. The chromosome harbors genes that are predicted to be involved in the subsequent methylotrophic pathways of C₁ transfer and assimilation, utilization of other C₁ compounds like methanol, nitrate respiration, and amino acid synthesis.

Plasmid-determined phenotypes of strain DM1.

Curing of plasmid pLVM1 from strain DM1 (DM1ΔpLVM1) was carried out by sequentially transferring the cultures into fresh Luria-Bertani (LB) medium for >20 generations and was verified by PCR analysis (Fig. 3C). DM1ΔpLVM1 was unable to grow on or to degrade DMF (Fig. 3A and B), indicating that plasmid pLVM1 is crucial for DMF catabolism. We further analyzed its ability to utilize intermediates in DMF degradation and other carbon sources (Fig. 3D). Results indicated that DM1ΔpLVM1 showed less ability than the wild-type strain to grow on DMA. In contrast, MA utilization was not weakened. The loss of pLVM1 inhibited growth on formamide, ethanol, and toluene. Metabolism of methanol, glycerol, and trisodium citrate showed no difference, compared with the wild-type strain. The prevention of growth on DMF and formamide is probably due to loss of the genes encoding DMFase (32). Plasmid pLVM1 does not encode genes that are predicted to be involved in the utilization of toluene and ethanol; however, it carries genes (C0214_27250 and C0214_27265) encoding enzymes with >40% identity to characterized homologs that are possibly involved in the transport of these compounds (35, 36). The results also indicated that there are extra copies of genes involved in the utilization of DMA and MA present on the chromosome. Genes encoding enzymes for metabolism of methanol, glycerol, and trisodium citrate are probably within the chromosome. From the findings described above, the presence of the critical plasmid-borne DMF catabolic pathway endows the microbe Methylobacterium sp. strain DM1 with the additional ability to utilize DMF and may also facilitate tolerance to toluene and ethanol. This may provide an evolutionary advantage of supplementing microbial metabolism by utilizing and tolerating toxic xenobiotic compounds in the environment.

Plasmid pLVM1 in strain DM1 is unique among DMF degraders, with approximately 80% of its overall length representing genes involved in the initial steps of DMF degradation. The plasmid pAMV1 in the P. aminophilus isolate JCM7685 is not specific to DMF catabolism (20). In addition to a small proportion of ORFs encoding enzymes for DMF, DMA, and MA utilization, pAMV1 encodes genes facilitating the metabolism of other carbon sources, such as methanol and l-arabinose. In another isolate, P. aminophilus JCM7686, genes encoding enzymes for initial DMF degradation are separated on different replicons (pAMI2, pAMI5, and pAMI6) (19). In strain DM1, however, those genes are colocalized on the plasmid, which might increase the efficiency of genetic control in DMF degradation.

Clarification of the DMF degradation pathway.

Three intermediates that accumulated at different times during the degradation process were identified as benzenesulfonyl chloride derivatives of DMA, MA, and ammonium, based on comparison of gas chromatography-mass spectrometry (GC-MS) retention times and mass spectra with those of standard compounds (Fig. 1E to G; also see Fig. S3A). Strain DM1 could also use DMA or MA as the sole source of carbon and nitrogen in growing cultures (Fig. 3D). DMA was transformed to MA, and MA could be further degraded in a resting cell reaction (Fig. S3B and C). Methylformamide and formamide could not be detected. These results indicated that strain DM1 possesses the most common pathway of DMF degradation, with the first step being the hydrolysis of DMF to DMA (Fig. 1C).

Enzymes related to this pathway in strain DM1 were identified based on amino acid sequence identity with their characterized homologs (Fig. 1C; also see Table S1). DmfA1 and DmfA2 greatly resembled their homologs in Alcaligenes sp. strain KUFA-1 and Paracoccus aminophilus strain JCM7685 (sharing >90% amino acid identity), while having lower levels of identity with those in Paracoccus aminophilus strain JCM7686 (84% and 73%, respectively) (Fig. 4A and B) (19, 20, 31). A putative LuxR family transcriptional regulator was found upstream of dmfA1A2, displaying 57% amino acid identity with its homolog in Paracoccus aminophilus strain JCM7686 (19). The mauNMGLJCADEBF genes are suggested to encode MA utilization enzymes, which show high levels of identy to their homologs in Methylobacterium extorquens strain AM1 (>75% amino acid sequence identity) (26). The predicted NMGP enzymes share at least 67% amino acid sequence identity with those in Methylobacterium extorquens strain DM4 (30). Two operons of genes encoding DmmABCD are found next to the mau genes and the NMGP genes, and both have at least 40% amino acid sequence identity to those of Methylocella silvestris strain BL2 (37).

FIG 4.

Phylogenetic relationships of enzymes predicted to be involved in DMF degradation. Maximum likelihood phylogenetic trees of the amino acid sequences of DmfA1 (A) and DmfA2 (B) and related proteins, with 1,000 bootstrap replications, are shown. The marked colors correspond to the functional genes and enzymes in Fig. 1. Accession numbers are from the NCBI database.

Genes predicted to be involved in the subsequent metabolism of formaldehyde and formate in strain DM1 are mostly found on the chromosome (Fig. 1B), as is the case in Methylobacterium extorquens strain AM1 (26, 38). Three enzymes (methylene-tetrahydrofolate [H₄F] dehydrogenase, methenyl-H₄F cyclohydrolase, and formyl-H₄F ligase) are predicted to convert CH₂=THF to formate via the H₄F-linked pathway, which starts with the spontaneous reaction of formaldehyde with H₄F; these enzymes are encoded by the genes mtdA, fch, and ftfl, respectively. The tetrahydromethanopterin (H₄MPT)-linked pathway includes enzymes involved in formaldehyde oxidation, i.e., formaldehyde-activating enzyme, methylene-H₄MPT dehydrogenase, methenyl-H₄MPT cyclohydrolase, and formylmethanofuran dehydrogenase, which are encoded by fae/fae2, mtdB, mch, and fhcABCD, respectively. Formate oxidation is catalyzed by formate dehydrogenases that are thought to be encoded by four separate gene clusters in strain DM1, i.e., fdhAB, fdh2ABCD, fdh3ABC, and fdh4AB. Ten genes encoding predicted homologous enzymes for C₁ assimilation are all found on the chromosome of strain DM1. By combining the predicted genes on the DM1 chromosome with those on the megaplasmid, we are able to predict the complete mineralization pathway of DMF in strain DM1 (Fig. 1C).

Transcriptional analysis of genes specific to DMF degradation.

RNA-seq was used to identify the genes of strain DM1 that were differentially expressed in response to DMF. According to the results, 1,538 genes were identified as differentially expressed, with log₂ fold changes of >1 or <−1 (in reads per kilobases per million reads), with P values of <0.05 (Fig. 2A). Of these, 867 genes were upregulated and 671 genes were downregulated, including 1,476 genes on the chromosome and 62 genes on the plasmid. For pLVM1, 50 of these genes showed upregulation, suggesting the importance of this plasmid in DMF utilization. Genes with the same or similar expression patterns were grouped together, to identify the unknown functions of genes by cluster analysis (Fig. 2B). Based on KEGG pathway enrichment analysis, we found that differentially expressed genes were mainly involved in methane metabolism and two-component system pathways (Fig. 2D). Real-time quantitative PCR (RT-qPCR) of 12 genes showing differential expression validated the results of RNA-seq (Fig. 1D).

As shown in Fig. 2C and Table S2, the plasmid-borne genes encoding the initial steps of degradation were highly upregulated in response to DMF. Specifically, the transcriptional activities of genes in the dmmABCD cluster were much higher than those of genes in the dmmABC2 cluster, indicating the major responsibility of dmmABCD in DMA oxidation. Similarly, for MA degradation, the mau genes were more highly expressed than genes encoding enzymes in the NMGP. Among the three copies of putative genes encoding enzymes in the NMGP, the plasmid-borne gmaS, mgsABC, and mgdABCD showed upregulation patterns similar to those of gmaS2, mgsABC2, and mgdABCD2, respectively. In contrast, genes in the mgdABCD3 cluster were downregulated, indicating the low functional importance of this gene copy. Although the folKBPEC gene cluster predicted to encode enzymes for H₄F synthesis was downregulated, most of the other putative genes in C₁ transfer and assimilation showed upregulation. For formate metabolism, it was interesting to discover that two clusters, fdhAB and fdh3ABC, encoding formate dehydrogenase were slightly downregulated, while the other two, fdh2ABCD and fdh4AB, were upregulated. Taken together, the transcriptome results indicated that most of the genes predicted to be involved in the proposed pathway of DMF degradation were induced, while some of the genes that we anticipated to be necessary for DMF degradation showed no differential expression or downregulation. We speculate that this may be due to gene preferences or their different response times.

Characterization of DMFase, the initial enzyme in the DMF degradation pathway.

The plasmid-borne DMFase subunits encoded by dmfA1 and dmfA2 were heterologously expressed in Escherichia coli. The enzyme activity was verified in vivo in a resting cell reaction (Fig. S4). The transformation of DMF to DMA was detected when the subunits were coexpressed, but E. coli expressing either DmfA1 or DmfA2 could not catalyze this reaction. The recombinant protein His₆-DMFase was expressed in E. coli BL21(DE3), purified, and characterized. The enzyme activity of His₆-DMFase was assayed in vitro using an alkyl-amine-specific color reaction (Fig. 5E). Native PAGE analysis demonstrated that the overall molecular size of His₆-DMFase was approximately 200 kDa (Fig. S5). Based on SDS-PAGE analysis, the molecular masses of the α-subunit and the β-subunit are approximately 20 kDa and 80 kDa, respectively (Fig. 5A); thus, His₆-DMFase is predicted to be an α₂β₂ enzyme. The maximal specific activity of the enzyme (35.2 U mg⁻¹) was observed at pH 6.5 and 50°C in 50 mM phosphate buffer (Fig. 5B and C). Cu²⁺ and Fe²⁺ slightly enhanced the activity of His₆-DMFase, while Co²⁺, Cr²⁺, and Ca²⁺ slightly inhibited the enzyme activity (Fig. 5D). The enzyme was stable after being incubated overnight at temperatures below 50°C (Fig. 5F). Circular dichroism spectral analysis showed a minor change in the secondary structure of His₆-DMFase at 50°C, and the protein quickly denatured when temperatures were ≥55°C (Fig. 5G). Based on these results, the critical degeneration temperature of His₆-DMFase is about 50°C. The characteristic peaks at 192 nm, 222 nm, and 208 nm in the circular dichroism spectra suggest that His₆-DMFase may have an all-α-helix structure in liquid (Fig. 5G). The K_m and V_max values for this enzyme at 37°C were 27.6 ± 2.50 mM and 51.1 ± 1.16 U mg⁻¹, respectively (Fig. 5H). The V_max of His₆-DMFase was between those of the DMFases purified from Alcaligenes sp. strain KUFA-1 and Pseudomonas sp. strain DMF 3/3, and the K_m was lower than those of both enzymes (Table 1) (21, 32). Since the amino acid sequences of DMFases are highly conserved, the differences in kinetic constants may be due to the different purification methods used.

FIG 5.

Purification and characterization of the His₆-DMFase protein. (A) SDS-PAGE analysis of His₆-DMFase. Lane 1, sonicated E. coli BL21(DE3)-pETDuet-1; lane 2, sonicated E. coli BL21(DE3)-pETDuet-1-dmfA1A2; lane 3, E. coli BL21(DE3)-pETDuet-dmfA1A2 pellet; lane 4, E. coli BL21(DE3)-pETDuet-dmfA1A2 supernatant; lane 5, purified His₆-DMFase; lane M, molecular size markers. (B) Effect of pH on His₆-DMFase activity. (C) Effect of temperature on His₆-DMFase activity. (D) Effect of metal ions (1.0 mM) on His₆-DMFase activity. (E) In vitro assays of enzyme activity using an alkyl-amine-specific color reaction, with DMF used as the substrate (a) or no DMF added (b). (F) Specific enzyme activities for His₆-DMFase incubated overnight at different temperatures. (G) Circular dichroism spectra of DMFase at increasing temperatures from 25°C to 75°C. (H) Kinetic analysis of His₆-DMFase (fitted to a Michaelis-Menten curve).

TABLE 1.

Comparison of kinetic constants of DMFases

Strain with DMFase	Km (mM)	Vmax (U mg−1)	Source
Methylobacterium sp. strain DM1	27.6	51.1	This study
Alcaligenes sp. strain KUFA-1	1.28	83.3	21
Pseudomonas sp. strain DMF 3/3	11.8	2.23	32

Comparative genomic analysis of methylotrophic pathways.

To obtain a comprehensive understanding of methylotrophic pathways in strain DM1, we compared its genome with those of two other Methylobacterium strains (AM1 and DM4) and two well-studied DMF degraders (P. aminophilus JCM7685 and JCM7686) (19, 20, 26, 30, 39). Distinct from most Methylobacterium strains, strain DM1 possesses additional methylotrophic genes encoding DMFase, termed dmfA1 and dmfA2, within the plasmid pLVM1. The G+C content of dmfA1A2 (59.3%) is relatively lower than that of the whole DM1 genome (68.3%), as well as that of pLVM1 (67.4%). Insertion elements (C0214_26845 and C0214_26895) predicted to be derived from ISPme2 in the IS5 family of Paracoccus are located upstream and downstream of dmfA1A2 within a predicted genomic island. These observations suggest that dmfA1A2 might have been obtained by lateral gene transfer.

The distribution and organization of DMF catabolism-related genes differed among these bacteria. For example, dmfA1A2 was not found in any of the sequenced Methylobacterium genomes that have been uploaded to NCBI, except for strain DM1. These genes are plasmid borne in strains DM1 and JCM7686, while they are encoded within the chromosome in strain JCM7685 (Fig. 6). The pattern of abundant gene copies for DMF degradation in strain DM1 is special among these isolates, with one set of the mau gene cluster, two sets of functional redundant NMGP gene clusters for MA degradation, and two sets of putative fae genes for formaldehyde metabolism. However, strains AM1 and JCM7686 have only one gene copy of the NMGP gene cluster. Strains JCM7685 and DM4 lack the mau genes and are dependent on the NMGP (Fig. 6). Additionally, strain DM1 possesses an extra copy of the putative dmm genes and the fae gene within the plasmid. The redundant genes might increase DMF utilization efficiency and provide an evolutionary advantage in adaption to environments with DMF (Fig. 6).

FIG 6.

Comparison of methylotrophic gene organization in Methylobacterium sp. strain DM1, Methylobacterium extorquens strain AM1, Methylobacterium extorquens strain DM4, Paracoccus aminophilus strain JCM7685, and Paracoccus aminophilus strain JCM7686. Colored blocks represent genes encoding different functional enzymes as in Fig. 1. Homologs of genes in different strains are linked by straight lines or indicated by the color and length of the blocks.

Methylobacterium sp. strain DM1 metabolizes formaldehyde to formate via distinct routes, compared with Paracoccus aminophilus strains JCM7685 and JCM7686 (Fig. 6). C₁ transfer in strain DM1 occurs through the H₄F-linked pathway and the H₄MPT-linked pathway, which is similar to C₁ transfer in other Methylobacterium strains. In Paracoccus, the reactions converting formaldehyde and CH₂=THF to formate are separate. Formaldehyde is catalyzed by S-(hydroxymethyl)glutathione synthase (Gfa), S-(hydroxymethyl)glutathione dehydrogenase (FlhA), and S-formylglutathione hydrolase (FghA), and the oxidation of CH₂=THF is performed by a bifunctional enzyme encoded by folD and a formyltetrahydrofolate deformylase encoded by purU; these enzymes have not been found in Methylobacterium (20). Although the genes involved in the serine cycle in these five strains are almost the same, they are dispersed on the chromosome of Methylobacterium sp. strain DM1, whereas they are concentrated in a compact gene cluster in Paracoccus.

In summary, this study describes the molecular mechanism of DMF degradation in Methylobacterium sp. strain DM1, as determined through analysis of genomic information, transcriptome analysis, and investigation of the enzymatic properties of DMFase. As shown in the genetic composition and molecular pathways of DMF catabolism (Fig. 1), the megaplasmid pLVM1 is a crucial DMF-catabolic plasmid that carries genes specifically involved in the DMF, DMA, and MA utilization pathways, whereas the downstream intermediate formaldehyde is proposed to be assimilated via the chromosome-borne H₄MPT and H₄F pathways. The plasmid-borne DMF degradation pathway in strain DM1 suggests that the ability to utilize DMF may be acquired by lateral gene transfer in Methylobacterium. Transcriptome analysis leads us to gain insights into redundant genes in response to DMF and its metabolites. The multicopy genes of DMA and MA degradation are unique among DMF degraders and Methylobacterium, suggesting the emergence of a new genotype. Thus, this study develops our understanding of DMF catabolism in microbes.

MATERIALS AND METHODS

Strains and culture media.

Methylobacterium sp. strain DM1 was identified through 16S rRNA sequence comparisons. The 16S rRNA gene was amplified using the universal primers 27F and 1492R, and sequence alignment and phylogenetic analysis were performed using BLAST and MEGA 7.0, respectively. Escherichia coli DH5α was used for plasmid construction, and Escherichia coli BL21(DE3) was used for protein expression. Information about the strains is shown in Table 2. Strain isolation was carried out by enrichment culture in mineral salts medium (MSM) (40) containing 0.1% yeast extract and 2,000 mg liter⁻¹ DMF as the sole source of carbon and nitrogen, followed by isolation on MSM plates without yeast extract. Yeast extract was used only in the initial enrichments. MSM used for culturing strains with DMF, DMA, and MA was devoid of NH₄Cl. DMF was purchased from Shanghai Lingfeng Chemical Reagent Co. LB medium was used for the cultivation of E. coli and was supplemented with 100 μg ml⁻¹ ampicillin when necessary. Solid agar plates were prepared by adding 1.5% (wt/vol) agar to the liquid medium.

TABLE 2.

Strains and plasmids used in this study

Strain or plasmid	Descriptiona	Source
Methylobacterium sp.
DM1	DMF-degrading strain; Gram negative	CCTCC (accession no. M2018663)
DM1ΔpLVM1	Strain DM1 depleted of plasmid pLVM1	This study
Escherichia coli
DH5α	F recA1 endA1 thi-1 hsdR17 supE44 relA1 deoR(lacZYA-argF) U169 80dlacZM15	TransGen
BL21(DE3)	F− ompT hsdS(rB− mB−) gal dcm(DE3)	TransGen
Plasmids
pETDuet-1	Ampr; expression vector
pETDuet-1-dmfA1	Ampr; pETDuet-1 containing dmfA1	This study
pETDuet-1-dmfA2	Ampr; pETDuet-1 containing dmfA2	This study
pETDuet-1-dmfA1A2	Ampr; pETDuet-1 containing dmfA1 and dmfA2	This study

^aAmp^r, ampicillin resistance.

Strain DM1 cultivation and DMF degradation conditions.

The strain was grown to exponential phase in 50 ml sterilized MSM with 2,000 mg liter⁻¹ DMF at 30°C, with shaking at 200 rpm, and this preculture was used as a seed broth. The inoculum size was 5% (vol/vol) in the following experiments. When optimizing the DMF concentration, the seed broth was inoculated into 50 ml fresh MSM containing different initial concentrations of DMF (400, 2,000, 4,000, and 6,800 mg liter⁻¹) and was cultivated at 30°C. In the pH optimization experiment, MSM was adjusted to different pH values (4.0, 6.0, 7.0, 8.0, and 10.0) with HCl and NaOH and supplemented with 2,000 mg liter⁻¹ DMF, and then strain DM1 was inoculated into the medium and cultivated at 30°C. At the optimal initial DMF concentration and pH (2,000 mg liter⁻¹ DMF and pH 7.0), bacterial growth and DMF degradation were monitored at different temperatures (25°C, 30°C, 37°C, and 42°C). Cell density was determined by measuring the optical density at 600 nm (OD₆₀₀) with an UV spectrophotometer. The growth curves were fitted using OriginPro 8.0 software and calculated with the built-in formula for a logistic growth model. DMF concentrations were measured by high-performance liquid chromatography (HPLC) (Agilent Technologies 1200 series) with an Agilent Eclipse XDB-C18 column (5 μm, 4.6 by 150 mm). We used the analytical method with modifications in the flow rate (0.5 ml min⁻¹) and column temperature (30°C) (22).

Determination of intermediate metabolites.

Strain DM1 was cultivated under optimal conditions in MSM with 2,000 mg liter⁻¹ DMF until the late exponential phase. Cells were collected by centrifugation at 4°C, washed twice with 10 mM phosphate-buffered saline, and resuspended. The starvation treatment was conducted by cultivating cells without substrate at 30°C for 3 h (these cells were called starved cells). DMF, DMA (purchased from Shanghai Aladdin Bio-Chem Technology Co.), and MA (purchased from J&K China Chemical) were used as substrates in resting cell reactions. A GC-MS system (Agilent 6850/5975C) was used for detection of intermediates, using the method described by Sacher et al. (41).

Genome sequencing and genetic information analysis.

Genomic DNA was extracted using the Wizard genomic DNA purification kit (product no. A1125; Promega). Whole-genome sequencing was performed on the Illumina MiSeq and PacBio sequencing platforms. Functional gene predictions and annotations were performed using GeneMarkS (http://exon.gatech.edu/GeneMark), the Rapid Annotation Subsystem Technology (RAST) database (42), BLAST (https://www.ncbi.nlm.nih.gov/BLAST), and UniProt (www.uniprot.org). Insertion elements were predicted using ISfinder (43).

Plasmid curing and phenotype determination.

Strain DM1 was sequentially transferred into fresh LB medium at 30°C. The dilution separation method was carried out on LB plates to obtain single colonies. We picked up 98 single colonies at one time and transferred them to MSM agar plates containing 2,000 mg liter⁻¹ DMF and LB plates at the same time. Colonies that could grow on the LB plates but could not grow on the plates with DMF were obtained as potential DM1ΔpLVM1 strains. PCR analyses of chromosomal and plasmid-borne genes were used to verify the plasmid-free DM1 strain (DM1ΔpLVM1). Primers for PCR verification are listed in Table 3. Both the DM1ΔpLVM1 and wild-type strains were cultured in MSM containing the following compounds as the sole carbon or carbon and nitrogen source: 2,000 mg liter⁻¹ DMF, 5 mg liter⁻¹ DMA, 5 mg liter⁻¹ MA, 3% (vol/vol) methanol, 3% (vol/vol) ethanol, 1% (vol/vol) formamide, 1% (vol/vol) toluene, 1% (vol/vol) glycerol, or 4% (vol/vol) trisodium citrate. The strains were cultured in 96-well plates with a total volume of 200 μl in each well. OD₆₀₀ values were measured with an automated microbiology growth curve analysis system.

TABLE 3.

Primers used in this study

Primer	Primer sequence (5′ to 3′)
dmfA1-F	CCGGAATTCGATGACTGAAGCCAGCGAATCCTGCG
dmfA1-R	CCCAAGCTTCTATGCAAGCTCTGCGCGCACATCT
dmfA2-F	CGCCATATGATGAAAGACATTGCCATTCGCGGAT
dmfA2-R	CGGGGTACCTCAGACCCGCGGCGCCGGCTCATCCTTG
dmmD-F	GCATGAATTCATGACCACTGACACGTTCCCG
dmmD-R	AGCTGGATCCTCAGGCCCGCTCATCCAGCG
chr1-F	CGCGGATCCGTGACGTCGGCCTCCTCCGAAC
chr1-R	CCCAAGCTTTCAACGTTCGACGACGAGGTCGCT
chr2-F	CCGGAATTCGATGGAACCGATCCCGAAGAAGG
chr2-R	CCCAAGCTTTCAGGTGATCGGCCAGTTATCGT
dmfA1-RTq-F	GGGAATGGGACGCCTATC
dmfA1-RTq-R	CGGTCTTGAATCGCTCATC
dmfA2-RTq-F	GAGGACACCGAGGATTAC
dmfA2-RTq-R	TTGGCATAGGCGAGATAG
mauJ-RTq-F	CGTGCTGGTCGGCTTCTTC
mauJ-RTq-R	CACGAAGAAGCCGACCAG
mauA-RTq-F	TACAACCCGACCGACAAG
mauA-RTq-R	CCGAAGCACCAGATGATG
mauD-RTq-F	TTCCAGGTCGGCAAGATC
mauD-RTq-R	TGTTGGTGAGACCCTTGG
mauB-RTq-F	TCGTCGGCACCTATCCTTGG
mauB-RTq-R	CTGTGGCGAGAACTGGTAGAAC
mauF-RTq-F	ATTTGCCCAGGAATACTC
mauF-RTq-R	AAGCCAATGAAGAGAACC
dmfR-RTq-F	CCTACTTGGACAATTCATAC
dmfR-RTq-R	GATCGGCTCTTCAGATAG
dmmD-RTq-F	CGGTTACGAGGTGTTCTG
dmmD-RTq-R	GACGATGTCCAGGCAATC
dmmA-RTq-F	GACATCGGCATGAACTTC
dmmA-RTq-R	AGGTTGAGCAGGAACTTC
dmmC2-RTq-F	CCTGGCTGCCGATTATCC
dmmC2-RTq-R	CGAAGACGAACGCTTGATC
dmmA2-RTq-F	GGCAAGCAGGAGGTCTTC
dmmA2-RTq-R	GACACCCCATCGTTGAAGG

RNA-seq.

Strains grown with 2,000 mg liter⁻¹ DMF in MSM were taken as the treatment group, while strains grown with 1% (vol/vol) glycerol in MSM were taken as the control group. Strains were collected at early/mid-exponential phase, and supernatants were removed. Samples were stored at −80°C after liquid nitrogen freezing. RNA-seq was performed by Shanghai Personal Biotechnology Co.

RT-qPCR analysis.

Total RNA was extracted with the RNAprep pure cell/bacteria kit (TianGen, China). After treatment with DNase I (Thermo Scientific), reverse transcription was carried out using the Promega GoScript reverse transcription system (product no. A5001; Promega). The cDNA product was used for the RT-qPCR analysis. RT-qPCR was performed using the CFX96 real-time PCR detection system (Bio-Rad, Hercules, CA) with Real MasterMix (SYBR green) (TianGen). Primers for RT-qPCR are listed in Table 3. The RT-qPCR program was as follows: initial denaturation at 95°C for 3 min, followed by 40 cycles of denaturation at 95°C for 20 s, annealing at 59°C (16S rRNA gene, mauA, mauJ, dmmA, dmmA2, dmmD, and dmmC2), 61.4°C (mauF, dmfA1, and dmfR), or 63.3°C (dmfA2, mauB, and mauD) for 20 s, and elongation at 68°C for 15 s. The melting curve analysis was as follows: temperature maintained at 95°C for 30 s, changed to 55°C for 65 s, and finally held at 95°C. The fold change in gene expression was calculated using the 2^−ΔΔCt method, using the 16S rRNA gene as the reference gene.

Cloning and expression of genes.

The dmfA1 and dmfA2 genes were amplified from the genomic DNA of Methylobacterium sp. strain DM1 with primers dmfA1-F and dmfA1-R and primers dmfA2-F and dmfA2-R, respectively (listed in Table 3). The amplified genes were ligated into pETDuet-1 to generate the recombinant plasmids pETDuet-1-dmfA1, pETDuet-1-dmfA2, and pETDuet-1-dmfA1A2. The recombinant plasmid (listed in Table 2) was transformed in E. coli BL21(DE3) for protein expression. A final concentration of 0.1 mM isopropyl-β-d-thiogalactopyranoside was added to the medium, and the cells were further cultivated at 16°C for 12 h. Cells were disrupted using an ultrasonic cell disruptor. DMFase with an N-terminal His-tag (His₆-DMFase) was purified by Ni-chelating affinity chromatography. The residual imidazole in the eluted fractions was removed by ultrafiltration with 50 mM phosphate buffer (pH 8.0). Protein concentrations were estimated using the Bradford method, with bovine serum albumin as the standard. SDS-PAGE was performed with electrophoresis buffer (1.0 g liter⁻¹ SDS, 18.8 g liter⁻¹ glycine, and 3.02 g liter⁻¹ Tris). Native PAGE was performed with the same electrophoresis buffer without SDS, and the pH of the buffer was adjusted to 8.0.

Enzyme characterization.

His₆-DMFase activity was assayed using a modified alkyl-amine-specific color reaction (21). Buffers with different pH values, such as 50 mM citric acid/sodium citrate (for pH 4.0 to 6.0), 50 mM phosphate buffer (for pH 6.0 to 8.0), 50 mM Tris-HCl (for pH 8.0 to 9.0), and 50 mM sodium hydrogen carbonate/sodium carbonate (for pH 9.0 to 11.0), were used to evaluate the effects of pH on enzyme activity. Experiments to test the effects of temperature (16°C to 60°C) on this enzyme were carried out in 50 mM phosphate buffer (pH 6.5). For thermal stability assessment, DMFase was preincubated overnight in 50 mM phosphate buffer (pH 8.0) at different temperatures (30°C to 55°C), and then the enzymatic reaction was conducted at 50°C. The kinetic parameters were measured in phosphate buffer (pH 7.0) at a reaction temperature of 37°C. One unit of enzyme activity was defined as the amount of enzyme catalyzing the hydrolysis of 1 μmol DMF per minute under the assay conditions.

Accession number(s).

Methylobacterium sp. strain DM1 has been deposited in the China Center for Type Culture Collection (CCTCC) under accession number M2018663. The whole-genome sequence of Methylobacterium sp. strain DM1 was deposited into the NCBI database under the accession numbers CP029173.1 (chromosome) and CP029174.1 (megaplasmid pLVM1). The full names and loci of genes and enzymes in the DMF degradation pathways of strain DM1 are listed in Table S1 in the supplemental material.