Introduction

Methanogenic archaea are a crucial component of the global carbon cycle, producing approximately 1 billion metric tons of methane annually1,2. These microorganisms thrive on a limited range of substrates, deriving energy from methanogenic processes3. While most methanogens rely solely on H2/CO2 or formate for growth, members of the Methanosarcina genus exhibit greater metabolic diversity. For instance, certain species can utilize a broader range of substrates, including H2/CO2, simple carbon compounds (C1-C2, such as methanol, methylamines, dimethylamines, trimethylamines, methanethiol, and methyl-sulfides), and acetate4. Notably, Methermicoccus shengliensis, a thermophilic methanogen isolated from a deep subsurface environment5, has been found to directly convert a variety of methoxylated aromatic compounds (MACs) into methane6. This unique ability sets it apart from other methanogens and highlights its potential role in the global carbon cycle.

Transcriptomic analysis indicated that two highly expressed genes, BP07_RS03225 (annotated as putative efflux transporter) and BP07_RS04910 (annotated as PAS domain S-box-containing protein), may be involved in substrate transport when M. shengliensis uses methoxybenzoate as substrate, by comparison with methanol. However, the biological function of these two genes has not yet been determined7.

To gain a deeper understanding of these two proteins, we conducted primary structural analysis. The protein encoded by BP07_RS03225 was found to belong to the MATE family of transporters (we named the MatE), consisting solely of the MatE domain. This domain functions as a transmembrane transporter, responsible for transporting substrates across the cell. In contrast, the protein encoded by BP07_RS04910 was classified as a methyl-accepting chemotaxis protein or transduction-like protein (Tlp). Tlp often contains several distinct regions, including transmembrane helices, Per-ARNT-Sim (PAS), PAS-associated C-terminal (PAC), histidine kinases, adenyl cyclases, methyl-accepting chemotaxis proteins (MCPs), and Phosphatases (HAMP)8,9, as well as a methyl-accepting chemotaxis-like (MA) domain. The PAS domain is capable of sensing environmental factors that cross the cell membrane and/or affect cell metabolism10, while the HAMP domain is frequently found in proteins associated with signal transduction in prokaryotic and lower eukaryotic organisms. The HAMP domain plays a critical role in transmitting conformational changes from ligand-binding domains to signaling domains8. The MA domain, present in chemotaxis sensory transducers, contributes to chemotactic behavior.

In this study, we employed bioinformatics analysis to predict that Tlp senses the substrate 2-methoxybenzoate, which is then transported into the cytoplasm by MatE. To validate this hypothesis, we conducted bacterial two-hybrid and structural model assays to elucidate the mechanism of the MatE-Tlp transport system. Furthermore, the poor growth and methane production of the matE knock-out mutant provided additional evidence for the critical role of MatE in methoxydotrophic methanogenesis. Our comprehensive analysis, combining genetic, biochemical, and structural model approaches, revealed the intricate interactions between Tlp and MatE, as well as their impact on substrate uptake and subsequent methane emission behavior. This study provides the first analysis of the transport model of MatE-Tlp in M. shengliensis, shielding light on methane emission reduction and methane production regulation in specific ecosystems.

Results and Discussion

Cell growth and methane production of M. shengliensis on different substrates

The growth curves of M. shengliensis varied significantly depending on the substrate used. Specifically, the growth rate of methanol was significantly higher than that of 2-methoxybenzoate. When methanol was used as the substrate, M. shengliensis reached the logarithmic growth phase within 4 days, whereas it took 5 days to reach this phase when 2-methoxybenzoate was used (Supplementary Fig. 1a). Furthermore, M. shengliensis produced more methane from methanol than from 2-methoxybenzoate. After 7 days of cultivation, approximately 13 mmol of methanol and 10 mmol of 2-methoxybenzoate were consumed for methane production, respectively (Supplementary Fig. 1b). These results provide strong evidence that M. shengliensis has a preference for methanol as a substrate for both growth and methane production.

Diversity of phylogenetic MatE and Tlp

Previous studies have shown that MatE and Tlp are highly expressed during methoxydotrophic methanogenesis7. Our quantitative real-time PCR (qRT-PCR) analysis revealed that the relative transcription levels of both genes were significantly upregulated (Supplementary Fig. 2). To further explore the distribution of MatE and Tlp in archaea, we constructed a phylogenetic tree. Our results showed that MatE and Tlp are widely distributed across archaea (Supplementary Figs. 3, 4). Notably, while many archaea possess MatE and Tlp, only M. shengliensis is able to utilize methoxylated aromatic compounds (MACs) to produce methane, aside from the non-methanogenic archaeon Archaeoglobus fulgidus (Supplementary Tables 1, 2).

MatE interacts directly with the Tlp

M. shengliensis has been shown to utilize methoxylated aromatics, such as 2-methoxybenzoate for growth (Supplementary Fig. 1). Our primary structural model assay revealed that Tlp functions as a substrate signal sensing protein, transporting the substrate to MatE, which then moves the compound into the cytoplasm (Fig. 1a, b). In other words, Tlp senses the signal to the transporter by interacting with MatE. To confirm the protein interaction, we constructed a bacterial two-hybrid system for Tlp and MatE.

Fig. 1: Tlp interacts with MatE and promotes its sensing to the substrate.
figure 1

a The domain structures of Tlp and MatE were analyzed using the Pfam and SMART programs. b The locations of domains in Tlp and MatE were analyzed using the TMHMM and TMpred programs. c Bacterial two-hybrid assay analysis was conducted to examine the interaction between Tlp and MatE in vivo. d To investigate the interaction between MatE and Tlp’s domains, a bacterial two-hybrid assay was conducted. This analysis aimed to examine the in vivo interactions between MatE and the HAMP, LBD, and SD domains of Tlp.

Full size image

Tlp consists of a ligand-binding domain (LBD), transmembrane (TM) helices, and a cytoplasmic signaling domain (SD) that interacts with downstream regulatory proteins (Fig. 1b)11,12. This enables Tlp to directly recognize ligands and mediate cellular responses13,14. To further understand how Tlp interacts with MatE, we used the bacterial two-hybrid system to confirm the interactions between LBD, TM, SD, and MatE, respectively. The interaction results showed that Tlp and HAMP interacted with MatE, and the blue spots in the experimental group were consistent with the control group (Fig. 1c, d). Additionally, the activity of the adenylate cyclase was restored when the fragments were brought into proximity by the interaction of the two chimeric proteins, indicating a direct interaction between Tlp and MatE (Supplementary Fig. 5).

To investigate how Tlp and MatE transport 2-methoxybenzoate into M. shengliensis, we designed the structure of Tlp and MatE using AlphaFold2. MatE mainly consists of two groups of α-helix, each with five α-helix (Fig. 2a). A substrate tunnel formed between these two groups of α-helix. Tlp consists of a ligand-binding domain, transmembrane helices, and a cytoplasmic signaling domain, which were predicted with high confidence (pLDDT scores >90) (Fig. 2b). Molecular docking analysis showed that MatE interacts with the HAMP domain of Tlp (Fig. 2c), which matches with protein interaction assays (Fig. 1d). Additionally, in membrane-associated proteins, HAMP domains usually lie near the cytoplasmic side of the membrane, where they convert transmembrane and intracellular sensory inputs into output response signals15. This suggests that the HAMP domain is crucial in the substrate transport process. The molecular docking result also showed that 2-methoxybenzoate has a high affinity (described by a lower binding energy of -4.8 kcal mol–1) for MatE. 2-methoxybenzoate binds to the substrate tunnel by hydrogen bonds with residues Ser302 and Arg279 (Fig. 2d). Furthermore, another methoxylated aromatic, 3,4,5-trimethoxybenzoate, docked also well with MatE (Supplementary Fig. 6), while methanol did not interact with MatE, indicating that MatE and Tlp prefer to interact with methoxylated aromatics.

Fig. 2: Ligand-MatE interactions in the middle membrane region of the transportation cavity.
figure 2

a An overview of the MatE model structure (cyan). b An overview of the Tlp model structure (purple). c The best interaction modes of MatE (cyan) and Tlp (purple) were created using AutoDock VinA. d Molecular docking methods were used to identify 2-methoxybenzoate (red) sites on the surface and in the center of the protein cavity. Hydrogen bonds are shown as black dashed lines.

Full size image

Among the thirty MatE homologous proteins in archaea, we found that residue Arg279 was highly conserved (Supplementary Fig. 7), suggesting that the MatE-Tlp system might function for all methylated aromatics. In summary, the LBD domain binds an extracellular substrate and transfers it to the HAMP domain, and then the HAMP domain interacts with MatE to transport the substrate into the cell.

The matE gene affects growth and methane production when M. shengliensis grown on 2-methoxybenzoate

The lack of selection markers has limited the development of genetic techniques in methanogens. To address this, we tested the effectiveness of puromycin, nourseothricin, and neomycin on M. shengliensis by determining their minimum inhibitory concentration (MIC) in liquid media at 45 °C. Our results showed that growth of M. shengliensis was completely inhibited by 2.5 μg/mL of puromycin, 100 μg/mL of nourseothricin, and 1 mg/mL of neomycin, respectively (Fig. 3a and Supplementary Fig. 8). We also found that higher concentrations of puromycin were required to inhibit growth at 65 °C, and the inhibitory effect persisted for 10 days (Fig. 3b).

Fig. 3: The matE gene influences growth and methane production in M. shengliensis when grown under 2-methoxybenzoate.
figure 3

a Growth curves of M. shengliensis under varying concentrations of puromycin (0–5 μg·mL−1) in methanol under 45 °C. b Growth curves of M. shengliensis under varying concentrations of puromycin (0–10 μg·mL−1) in methanol under 65 °C. c Schematic representation of the marker mutagenesis technique for the M. shengliensis strain. d A schematic of the generation of the M. shengliensis ΔmatE mutant. A linear DNA fragment containing pac, flanked by 700 bp regions homologous to the upstream and downstream regions of matE, was transformed into M. shengliensis. Incorporation of the pac-containing DNA at the matE locus gives rise to mutants that are resistant to puromycin. Primers used to verify ΔmatE mutants are represented by the F-R in the schematic. e Growth curves for the wild-type M. shengliensis and the ΔmatE mutant when grown with methanol and 2-methoxybenzoate. f Accumulation of methane in the headspace of bottles with wild-type M. shengliensis and the ΔmatE mutant when grown with methanol and 2-methoxybenzoate over an incubation period of 10 days. All data represent the mean of n = 3 biologically independent samples (a, b, e, f).

Full size image

To explore the utility of the pac gene (encoding puromycin N-acetyltransferase) as a selectable marker for creating chromosomal deletions in M. shengliensis, we designed a plasmid to replace the matE gene with the pac marker (Fig. 3c and Supplementary Fig. 9). We expected that incorporation of the pac marker into the chromosome would render the cells resistant to puromycin (Fig. 3d). To verify the integration of the pac marker, we designed primers outside the homologous recombinant plasmid region and on the pac cassette for PCR and sequencing verification (Supplementary Fig. 10). Our results showed that a ΔmatE mutant was successfully obtained in the transformants (Supplementary Fig. 11).

To assess the genotype of ΔmatE mutant, we observed the growth and methane production of the ΔmatE mutant under different substrate conditions. Our results showed that the deletion of the matE gene greatly inhibited the growth of M. shengliensis when 2-methoxybenzoate was used (Fig. 3e). Additionally, methane production in the ΔmatE mutant decreased by about 95% compared to the wild-type, suggesting that the deletion of matE negatively impacts methanogenesis (Fig. 3f). Our results also showed that the deletion of matE affected M. shengliensis growth and methane accumulation when 2-methoxybenzoate was used as the substrate, but not when methanol was used as the substrate.

In summary, our results suggest that MatE is important in methoxydotrophic methanogenesis in M. shengliensis, facilitating the entry of substrates into the cells from the extracellular environment. Unfortunately, we were unable to successfully establish a complementation mutant of M. shengliensis, highlighting the need for the development of genetic tools. Therefore, a better understanding of substrate transporters and methane metabolism in M. shengliensis is essential, and the development of more sophisticated genetic tools is urgently required to fill knowledge gaps and facilitate further research.

Conclusions

In this study, we constructed a phylogenetic tree of MatE and Tlp homologs, revealing that these two proteins are widely distributed among archaea. Through bacterial two-hybrid assays and molecular docking models, we propose that Tlp recognizes 2-methoxybenzoate and interacts with MatE to facilitate substrate transport, whereas methanol does not interact with the Tlp-MatE complex. Furthermore, targeted knockout of the matE gene significantly impacted methane emission in M. shengliensis when 2-methoxybenzoate was used as the substrate. Our findings demonstrate that the MATE family transporter system can influence methanogenesis in archaea by regulating substrate transport. This knowledge is crucial for developing more sustainable bioprocesses.

Methods

Microbial strains and culture conditions

The complete list of strains, primers and plasmids used in this study was shown in Supplementary Tables 3, 4, respectively. Escherichia coli DH5α and BTH101 were cultured on LB agar plates or LB broth supplemented with ampicillin (100 μg·mL-1) or kanamycin (50 μg·mL-1) at 37 °C. M. shengliensis ZC-1 (DSM18856) was stored in our laboratory (Chengdu, China) and cultivated in modified DSM medium 10845. Sludge fluid was replaced by trace element solution (100 × trace element solution), as previously reported7. The medium was sparged with N2 : CO2 in an 80:20 ratio before autoclaving. As substrate either 30 mM methanol or 30 mM 2-methoxybenzoate were used. The cultures were incubated at 65 °C. The growth was determined by measuring the optical density at 600 nm (OD600) and the methane production at different times. The identity of the microorganism was checked by 16s rRNA gene sequencing of DNA from methanol and 2-methoxybenzoate-grown cells with primers shown in Supplementary Table 3.

Determination of the minimal inhibitory concentration (MIC)

Liquid minimal inhibitory concentrations (MIC)s were determined as follows: saturated liquid cultures (OD600 = 0.35 for M. shengliensis) were used to inoculate liquid media containing different concentrations of puromycin (0–5 μg·mL-1), neomycin (0–5 mg·mL-1), and nourseothricin (0–200 μg·mL-1) at a 1:100 final dilution of cells. Growth on liquid containing the different concentrations of antibiotics was monitored over time and compared with growth on media lacking antibiotics for MIC determination.

Total RNA extraction and quantitative real-time PCR (qRT-PCR)

M. shengliensis was cultivated in 120 mL serum vials containing 40 ml of the medium with 30 mM methanol, or 2-methoxybenzoate as the substrates. Cells in the mid-exponential growth phase were collected from all treatments for total RNA extraction using the RNAprep pure Cell/Bacteria Kit (TIANGEN, China). RNA purity and concentrations were assessed through gel electrophoresis and measured using a NanoDrop2000 spectrophotometer (Thermo Scientific, USA). Ribosomal RNA was removed from total RNA using the RiboMinus TM kit (Invitrogen, USA)16.

For the qRT-PCR analyses, total RNA was extracted following the method described above and treated with DNase I to eliminate DNA contamination. The RNA was then reverse-transcribed into cDNA using the iScript™ cDNA Synthesis Kit (Bio-Rad, USA). Transcript levels were quantified using the SsoAdvanced™ Universal SYBR Green Supermix and a CFX96 ConnectTM Real-Time PCR Detection System (Bio-Rad, USA). The PCR cycling conditions were as follows: initial denaturation at 95 °C for 30 s, followed by 40 cycles of 94 °C for 15 s and 60 °C for 30 s. After the main program, melt curve analysis was performed from 65 to 95 °C, with an increment of 0.5 °C and 0.5 s plate reading at each step. If necessary, sample DNA was diluted approximately 50-fold. Positive controls (using chromosomal DNA for qPCR) and negative controls (using RNA without reverse transcriptase) were included in the experiment and subjected to identical amplification conditions. The 16S rRNA gene served as an internal standard to normalize the results. Relative transcript levels of target genes were calculated using the Quantitation-Comparative CT(2−△△CT) method16,17.

Bacterial two-hybrid (bacterial two-hybrid) and β-galactosidase assay in 96-well arrays

The bacterial two-hybrid assays as described previously18. To study protein-protein interactions, the tlp gene was cloned into the pKNT25 vector, resulting in a fusion of the C-terminal catalytic domain of CyaA with the T25 fragment. The matE gene was cloned into the pUT18C vector, resulting in a fusion with the T18 catalytic domain fragment of CyaA at either the C-terminus or N-terminus. Cloning was performed using E. coli BTH101 competent cells. Co-transformed cells were plated on LB agar containing final concentrations of 100 μg/ml ampicillin, 50 μg/ml kanamycin, 0.5 mM IPTG, and 50 μg/ml X-gal16. The collected cells were cultivated overnight at 30 °C. The pKNT25-zip and pUT18C-zip plasmids were used as positive controls, and the pKNT25/pUT18C (without insert) and pUT18C-zip/pKNT25-zip (fused with a leucine zipper protein) plasmids were used as negative controls, respectively. In the experimental group, the two fragments were fused into two interacting proteins. The bacterial two-hybrid assay primers and plasmids are listed in Supplementary Tables 3, 4, respectively.

For β-galactosidase assay was adapted to the 96-well array by modification of the Battesti & Bouveret’s method18. Briefly, an overnight culture was 100-fold diluted with fresh LB liquid and grown at 37 °C. When the culture reaches an OD600 of 0.4–0.6, add IPTG to a final concentration of 0.5 mM. After inducing for 1–3 h, 1 ml of cultures were used for β-galactosidase activity assay using the reported method following Battesti et al. The difference is that enzymatic reaction is carried out at 37 °C for 0–500 min with a measurement of OD420 every 30 min in the microplate reader. The relative β-galactosidase activity in each sample was then calculated by simple excel file manipulation. The following calculation method is ((OD420 at time t2-OD420 at time t1)/t2-t1 (min))/OD600. The t2 and t1 time points are selected to be located in the linear part of the kinetic.

M. shengliensis knock-out strain construction

The construction of M. shengliensis knock-out mutant strains was carried out in the following manner. Initially, the flanking regions of the selected open reading frames (ORFs) were amplified using PCR. A puromycin resistance cassette was then inserted in the middle of these regions. The entire element containing the cassette was then cloned into a pBluescript II SK plasmid using the Gibson assembly protocol19. The resulting plasmids were used to integrate the mutated alleles into the chromosomes of M. shengliensis through liposome-mediated transformation and allelic recombination. Notably, deletions of individual genes removed most of the coding sequence, resulting in an in-frame fusion peptide consisting of the first and last ten amino acids, thus creating non-polar mutations. Specifically, the cassette lacks both a promoter and a terminator16. It contains translational stop codons in all reading frames preceding the 5′-end of the ribosome binding site (RBS) of the pac encoding sequence, and another RBS at the 3′-end to prevent translational coupling with the downstream matE gene. The primers and plasmids used to generate the mutants are listed in Supplementary Tables 3, 4, respectively.

Transformation methods

Escherichia coli strains were transformed by heat shock at 42 °C. For M. shengliensis, liposome-mediated transformation was used. Briefly, cells from log-phase cultures (OD600 = 0.35) were collected by centrifugation and resuspended in 0.85 M sucrose to a density of about 109 cells per milliliter. Plasmid DNA-liposome complexes were formed by mixing 15 μL of DOTAP (Boehringer Mannheim) in 100 μL of 20 mM Hepes (pH 7.6) with 1.5 μg of plasmid DNA in 50 μL of 20 mM Hepes (pH 7.4). This was followed by a 30-min incubation at 65 °C. A 1.0-mL portion of the resuspended cells was added to the plasmid DNA-liposome suspension and incubated for 6 h at room temperature. With these cells and DNA concentrations, the maximum transformation frequency of M. shengliensis was achieved with 15 μL of DOTAP reagent. Following all methods, cells were transferred to 10 mL of broth medium after transformation, incubated at 45 °C for 16 h, and then plated on a medium with 2.5 μg/mL puromycin. All M. shengliensis transformants were single colony purified on solid media containing puromycin before undergoing PCR verification and phenotypic analysis.

AlphaFold2 prediction and docking analysis

The amino acid sequences of MatE and Tlp were determined based on the M. shengliensis M18856 genome. The three-dimensional structures for the putative MatE and Tlp sequences were built using AlphaFold220,21. The molecular models in UCSF Chimera (version 1.12) were used to perform energy minimization on the structures obtained from the previous step22. Next, 2-methoxybenzoate was docked to MatE using version 1.1.2 of AutoDock Vina23. The receptor and ligand options in AutoDock Vina were set to default. The grid box used in the docking procedure was defined to encompass the HAMP domain which is functionally relevant, and the corresponding residues were observed in the binding site of crystal structures for homologous proteins. The number of binding modes, the exhaustiveness of the search, and the maximum energy difference (expressed in kcal mol–1) were set to 8, 5, and 3, respectively. All of the structures were visualized and exported as images using PyMOL (http://www.pymol.org).

Genome analysis

The whole genome of M. shengliensis was downloaded from the NCBI database, and an operon analysis was conducted24. All the candidates were then reexamined for their domain organization using the SMART database (http://smart.embl-heidelberg.de/smart/set_mode.cgi?GENOMIC=1)25. The locations of domains in Tlp and MatE were analyzed using the TMHMM and TMpred programs26,27.

Phylogenetic analysis

The MatE and Tlp sequence from M. shengliensis and homologous MatE/Tlp sequences from publicly available archaea species were aligned using MEGA 11 (Supplementary Tables 1, 2)28,29. A maximum-likelihood phylogenetic tree based on the alignments was constructed by FastTree v2.0 under the JC model, and bootstrap analysis was applied with 1000 replications30. The initial tree for the heuristic search was automatically generated.

Growth curves and chemical analyses

M. shengliensis, grown in the presence of either 2-methoxybenzoate or methanol, were maintained at 65  °C at all times. After adapting to the respective substrate, the culture is further passaged in a medium containing antibiotic, typically for three generations, with an inoculation amount of 2%. The cultures were then incubated in quadruplicates at a steady temperature of 65 °C. Growth was measured every 12 or 24 h by measuring the optical density at 600 nm (OD600) with a Spectronic 21 spectrophotometer (Milton Roy, USA).

Gas samples (200 µL) were collected from the headspace of vials or serum bottles at various time points using a pressure-lock syringe (Vici, Schenkon, Switzerland). Methane (CH4) concentrations were analyzed with an Agilent GC7820A gas chromatography system (Agilent, USA) equipped with a Porapak Q column and a thermal conductivity detector. The column, oven, and detector temperatures were set to 65, 120, and 130 °C, respectively. Argon served as the carrier gas at a flow rate of 27 mL min-1. The gas pressure in the headspace of vials or serum bottles was measured at room temperature using a barometer (Ashcroft, Stratford, CT). The total amount of gas products was calculated based on Avogadro’s law, following calibration with a standard gas mixture (composition: N2 : CH4 : C2H6 : C3H8 : 2-C4H10 : 2-C5H12 = 61.41% : 14.24% : 8.21% : 10.02% : 3.80% : 2.32%). The concentrations of methane were determined using an external calibration curve16.

Statistics and reproducibility

Every experiment was conducted with a minimum of three independent biological replicates. Data were presented as the mean ± standard deviation (SD). Data points in the figures represent biological replicates. Each graph displays individual data points. Statistical analyses (mean and SD) were performed using GraphPad Prism 10 software. Detailed information on all reagents and resources can be found in the “Material and methods” section.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.