A Shuttle-Vector System Allows Heterologous Gene Expression in the Thermophilic Methanogen Methanothermobacter thermautotrophicus ΔH

ABSTRACT Thermophilic Methanothermobacter spp. are used as model microbes to study the physiology and biochemistry of the conversion of molecular hydrogen and carbon dioxide into methane (i.e., hydrogenotrophic methanogenesis). Yet, a genetic system for these model microbes was missing despite intensive work for four decades. Here, we report the successful implementation of genetic tools for Methanothermobacter thermautotrophicus ΔH. We developed shuttle vectors that replicated in Escherichia coli and M. thermautotrophicus ΔH. For M. thermautotrophicus ΔH, a thermostable neomycin resistance cassette served as the selectable marker for positive selection with neomycin, and the cryptic plasmid pME2001 from Methanothermobacter marburgensis served as the replicon. The shuttle-vector DNA was transferred from E. coli into M. thermautotrophicus ΔH via interdomain conjugation. After the successful validation of DNA transfer and positive selection in M. thermautotrophicus ΔH, we demonstrated heterologous gene expression of a thermostable β-galactosidase-encoding gene (bgaB) from Geobacillus stearothermophilus under the expression control of four distinct synthetic and native promoters. In quantitative in-vitro enzyme activity assay, we found significantly different β-galactosidase activity with these distinct promoters. With a formate dehydrogenase operon-encoding shuttle vector, we allowed growth of M. thermautotrophicus ΔH on formate as the sole growth substrate, while this was not possible for the empty-vector control.


count of plated microbial cells [Materials and Methods])
. We achieved dense growth with spot plating, but individual colonies were barely distinguishable with this plating technique, while we achieved plating efficiencies between 1 and 5% with spread plating, and 50% and higher with pour plating (see Text S1A and Fig. S1 in the supplemental material).
M. thermautotrophicus DH is sensitive to antibiotics commonly used in methanogen genetic systems. To find a suitable selection pressure for the positive selection of genetically modified cells, we analyzed several antibiotics such as simvastatin and neomycin (Text S1B). For both antibiotics, thermostable selectable markers are available, which have been successfully used in thermophilic nonmethanogenic microbes, such as Pyrococcus furiosus and Thermococcus kodakarensis (simvastatin) (21,22), as well as Thermoanaerobacter spp. (neomycin) (23), but recently also in the thermophilic methanogens Methanocaldococcus jannaschii (simvastatin) and Methanoculleus thermophilus (neomycin) (24,25). Both simvastatin and neomycin efficiently inhibited growth of M. thermautotrophicus DH cells in liquid culture at concentrations of 21.5 mg/ml and 250 mg/ml, respectively (Text S1B and Fig. S2 and S3). While these antibiotics inhibited growth of M. thermautotrophicus DH, we observed the appearance of spontaneously resistant M. thermautotrophicus DH cells for both antibiotics (Text S1B and Fig. S2 and S3). The incubation period for the appearance of spontaneously resistant M. thermautotrophicus DH cells differed for each antibiotic compared to the incubation period for nonselective growth of wild-type cells (16 to 24 h). We observed inhibition of growth for at least 48 h with simvastatin and 60 h with neomycin at the concentrations indicated above (Text S1B). Because we found that neomycin inhibits growth of M. thermautotrophicus DH for a longer incubation period than that for simvastatin, we decided to focus on neomycin as the selection pressure to develop the first selectable marker for M. thermautotrophicus DH.
A modular plasmid design enables the fast generation of shuttle vectors to test genetic elements for functionality in M. thermautotrophicus DH. Before we focused on the neomycin-selectable marker, we had constructed a subset of plasmids that would allow us to test various approaches for the transfer of DNA into M. thermautotrophicus DH and for positive selection (Materials and Methods). To ease exchangeability of genetic elements in shuttle vectors, and to allow fast adaptation to new findings, we decided on a modular plasmid design. Inspired by the pSEVA system for Gram-negative bacteria (26), and the pMTL80000 system for Clostridia (27), we established the Methanothermobacter vector system (pMVS) design.
The pMVS design consists of five modules, which are separated by rare 8-bp recognition sequences for the restriction enzymes PmeI, AsiSI, FseI, AscI, and PacI ( Fig. 1). To follow the pMVS design, these rare restriction enzyme-recognition sequences need to stay unique to grant exchangeability of the modules by restriction/ligation cloning. The five modules are (restriction enzyme boundaries are given in parentheses) (i) the replicon for E. coli (AsiSI, PmeI), (ii) the selectable marker for E. coli (PmeI, FseI), (iii) the replicon for M. thermautotrophicus DH (AsiSI, PacI), (iv) the selectable marker for M. thermautotrophicus DH (AscI, FseI), and (v) an application module that can be used to include any genetic cargo such as a reporter gene or another gene of interest (PacI, AscI) ( Fig. 1 and Text S1C).
Based on the first shuttle vector (pMVS-V1) that led to successful DNA transfer and selection protocols, as described below, we defined pMVS-V1 as our archetype shuttle vector (Fig. 1A), with a combination of (i) the ColE1-derived replicon for E. coli in combination with the conjugational transfer function (tra region) from RK2 (27), (ii) the chloramphenicol-selectable marker (Cam r ) for E. coli (27), (iii) the entire cryptic plasmid pME2001 from M. marburgensis as the replicon for M. thermautotrophicus DH (14), and (iv) the thermostable neomycin-selectable marker (Neo r ) (28) for M. thermautotrophicus DH under the control of the P synth promoter sequence (22), but without an application module and a PacI recognition sequence (Fig. 1A). After we had demonstrated the functionality of pMVS-V1, our first complete shuttle vector, pMVS1111A-P synth -bgaB, was then constructed based on this archetype shuttle vector. pMVS1111A-P synth -bgaB contains the b-galactosidase-encoding gene bgaB (see below) and an additional PacI site, which completes the application module (Fig. 1B, Materials and Methods, and Text S1C).
DNA transfer into M. thermautotrophicus DH is possible via interdomain conjugation with E. coli S17-1. Following our broad approach, we investigated several published protocols (and modified versions) to transfer DNA into M. thermautotrophicus DH by using our various plasmids and shuttle vectors (Materials and Methods). These DNA transfer protocols included (i) natural competence protocols (16,24), (ii) chemically/physically induced transformation protocols, such as with elevated calcium and magnesium concentrations in the mineral medium or with low-temperature incubation of precultures to induce stress conditions, (iii) an electroporation protocol (29), and (iv) an interdomain conjugation protocol with E. coli (30). Most attempts did not result in cells with the expected selectable phenotype and, if so, could not be linked to the respective anticipated genotype and appeared to be spontaneously resistant cells. The protocol that finally led to a successful DNA transfer into M. thermautotrophicus DH was an interdomain conjugation protocol with E. coli S17-1 ( Fig. 2 and Materials and Methods), which was a modified version of the protocol for conjugational DNA transfer into Methanococcus maripaludis (30).
To achieve a successful conjugational DNA transfer of the archetype shuttle vector pMVS-V1 from E. coli S17-1 (donor) to M. thermautotrophicus DH (recipient), we increased the recipient cell concentration to ;1.6 Â 10 9 cells from a culture in the early stationary growth phase, which is a 5-fold-higher cell concentration than that in the work of Dodsworth et al. (30). Furthermore, we used a spot-mating procedure to allow close physical contact between donor and recipient cells during an incubation period at 37°C on solidified medium plates, which supported metabolic activity of E. coli S17-1 ( Fig. 2A), in contrast to direct spread plating as in the work of Dodsworth et al. (30). consists of four modules, which are intersected by the 8-bp restriction enzyme recognition sites AscI, FseI, PmeI, and AsiSI. The four modules are the replicon for E. coli (ColE1, tra), the selectable marker for E. coli (Cam r ), the replicon for M. thermautotrophicus DH (pME2001), and the selectable marker for M. thermautotrophicus DH (Neo r ). (B) pMVS1111A:P synth -bgaB consists of five modules, which are intersected by the 8-bp restriction enzyme recognition sites AscI, FseI, PmeI, AsiSI, and PacI. The five modules are as in pMVS-V1 with the following differences: the replicon for M. thermautotrophicus DH is flanked by AsiSI and PacI, and the shuttle vector contains the bgaB gene in the application module, which is flanked by PacI and AscI. P synth , synthetic promoter sequence, which is based on P hmtB ; T mcr , terminator sequence from the mcr operon of M. voltae. The nomenclature for the pMVS design is realized by adding a four-digit code after pMVS for the definition of the first four modules, which defines the plasmid backbone with basic functions for replication and selection in E. coli (modules 1 and 2) and M. thermautotrophicus DH (modules 3 and 4). Additional large capital letters can be amended to each digit to define differences, such as varying promoter sequences, in a given module. For the fifth (application) module, a descriptive name is added after the four-digit code, which allows staying as flexible as possible with the genetic cargo used (without a limitation to nine digits), while staying within the pMVS design boundaries.
Fink et al.

®
After resuspending the spot-mated cells (donor 1 recipient) from the solidified medium plate, we recovered M. thermautotrophicus DH in liquid mineral medium without any complex medium additions (no organic carbon source) and without antibiotic additions at 60°C under a molecular hydrogen/carbon dioxide atmosphere ( Fig. 2B and Materials and Methods). These incubation conditions decreased the viability of E. coli S17-1 to a very minimum, and therefore, no counterselection with antibiotics against donor cells was required. After a short incubation period (3 to 4 h) to recover M. thermautotrophicus DH under nonselective conditions, the cells were transferred to liquid medium, which contained neomycin for a selective-enrichment step. Importantly, the required incubation period for the cells to grow in this step was key for a successful identification of transconjugants (i.e., recipient cells that received the shuttle vector). Neomycin at a concentration of 250 mg/ml inhibits growth of M. thermautotrophicus DH for at least 60 h in liquid medium (Text S1B). Therefore, when the cells did not grow within less than 60 h (typically growth appeared after 24 to 48 h in a successful experiment) in the selective-enrichment step, the conjugation experiment was regarded as unsuccessful, because the number of spontaneously resistant cells is considerably higher than potential transconjugants, which renders screening essentially impossible.
The shuttle-vector DNA confers the observed antibiotic-resistant phenotype and is maintained in M. thermautotrophicus DH with high segregational stability. After we found selective growth of putative transconjugant cells, we confirmed the successful DNA transfer into M. thermautotrophicus DH via two site-specific PCRs, which amplified a fragment of the pMVS-V1 shuttle vector and a fragment of the M. ) and the shuttle vector-carrying E. coli were harvested by centrifugation, mixed, and spotted on solidified medium plates that support growth of both microbes. During the spot-mating step at 37°C, the DNA transfer process via conjugation takes place (small scheme). (B) The process to isolate and identify individual colonies of genetically modified M. thermautotrophicus DH in the standard protocol. After the spot-mating step, M. thermautotrophicus DH cells were recovered in nonselective liquid mineral medium at 60°C, and afterward, transconjugants were enriched in neomycin (Neo)-containing selective liquid mineral medium at 60°C. Individual colonies were obtained from plating the enrichment culture. Those colonies were analyzed by PCR and Sanger sequencing. (C) PCR analysis of four respective transconjugants (1 to 4) with primer combinations specific for the shuttle vector pMVS-V1 replicon (1-kb fragment) and for genomic DNA (gDNA) of M. thermautotrophicus DH (2.8-kb fragment). N, water negative control; WT, control with wild-type M. thermautotrophicus DH; M, GeneRuler 1-kb DNA ladder (Thermo Scientific, Waltham, MA, USA). (D) Experimental conditions for the confirmation of conjugation as the mechanism for DNA transfer were (from left to right) DNase I treatment, heat inactivation of E. coli S17-1, conjugation with nonconjugative E. coli NEB stable, and addition of free plasmid DNA directly to M. thermautotrophicus DH cell culture.
Shuttle-Vector System for M. thermautotrophicus ® thermautotrophicus DH genomic DNA, respectively, with liquid cultures derived from individual colonies (Fig. 2C). In addition, we extracted plasmid DNA from several independent M. thermautotrophicus DH transconjugant cultures, transformed E. coli with this plasmid DNA extract, reextracted the plasmid DNA again from E. coli, and finally, performed restriction-enzyme digestions and Sanger sequencing to confirm the plasmid DNA integrity and sequence, without finding any differences from the original pMVS-V1 shuttle vector (Text S1D and Fig. S4). With different shuttle vectors in independent experiments, we achieved reliable DNA transfer into M. thermautotrophicus DH with our standard protocol (Materials and Methods), which includes a selectiveenrichment step (Fig. 2B). However, this protocol was not suitable to calculate the conjugation frequency, due to the selective-enrichment step. To determine the conjugation frequencies, we performed experiments without the selective-enrichment step but with a prolonged nonselective-recovery step, which resulted in 5 6 4 colonies (n = 6) and from which we calculated conjugation frequencies of approximately 4 Â 10 29 to 6 Â 10 26 transconjugants per initial recipients (see equation S1 in Text S1F). With this assessment, we showed that experimental variations considerably influenced the conjugation frequency (Text S1F and Table S1). Nevertheless, when following our standard protocol, a reliable transfer of plasmid DNA was achieved. Once the plasmid DNA was transferred into M. thermautotrophicus DH, it was maintained with high segregational stability over many cell divisions in an experiment under nonselective growth conditions, and we did not observe loss of pMVS-V1 (Text S1G and Fig. S5).
Free plasmid DNA is not resulting in DNA transfer into M. thermautotrophicus DH. By having demonstrated DNA transfer into M. thermautotrophicus DH, we further analyzed whether this transfer was indeed depending on conjugational DNA transfer from E. coli or whether it was rather by uptake of free DNA under the utilized cultivation conditions during the conjugation protocol (Materials and Methods). E. coli S17-1 donor cells contain a large amount of pMVS-V1, because of the high-copy-number ColE1 replicon, which might be released into the liquid medium from lysing cells. Therefore, we conducted control experiments with free pMVS-V1 plasmid DNA, heatinactivated E. coli S17-1 cells, and nonconjugative E. coli NEB stable cells that carry pMVS-V1 (Fig. 2D). None of these experiments resulted in DNA transfer into M. thermautotrophicus DH (Fig. 2D, Text S1H, and Fig. S6). In contrast, DNase I treatment of the E. coli S17-1 donor cells did not negatively influence the success of a conjugational DNA transfer into M. thermautotrophicus DH (Fig. 2D, Text S1H, and Fig. S6). Thus, we concluded that DNA transfer occurs due to conjugational mobilization activity from E. coli S17-1 to M. thermautotrophicus DH.
A thermostable b-galactosidase (BgaB) from Geobacillus stearothermophilus is a functional reporter to investigate promoter sequences in M. thermautotrophicus DH. With a DNA transfer protocol and a functional shuttle vector at hand, we proceeded with adding a genetic cargo (i.e., gene of interest) to the application module of the archetype pMVS-V1 shuttle vector. To enable the analysis of the effects from different promoter sequences on gene expression in M. thermautotrophicus DH, we decided to implement a reporter gene as our first gene of interest. We chose the bgaB gene from G. stearothermophilus, which encodes a thermostable b-galactosidase (31). We placed a codon-optimized version of the bgaB gene under the control of the nonnative P synth promoter (Materials and Methods). We transferred the resulting shuttle vector pMVS1111A:P synth -bgaB ( Fig. 1B) into M. thermautotrophicus DH via conjugation. In a qualitative preliminary experiment with cell lysate from pMVS1111A:P synth -bgaB-carrying M. thermautotrophicus DH cells (and pMVS-V1-carrying cells as an empty-vector negative control), we found that, indeed, the b-galactosidase BgaB is produced in M. thermautotrophicus DH and results in a color reaction in an enzyme assay with 3,4cyclohexenoesculetin-b-D-galactopyranoside (S-Gal) only in the presence of the bgaB gene ( Fig. 3B and Materials and Methods).
This result sparked us to establish a quantitative b-galactosidase enzyme activity assay with o-nitrophenyl-b-D-galactopyranoside (ONPG) as the chromogenic substrate for the b-galactosidase (Text S1I), which allowed us to investigate promoter sequences for their relative in-vivo effect on gene expression in M. thermautotrophicus DH during a growth experiment (Fig. S1I). Overall, we selected four distinct promoter sequences [P synth , P synth(BRE) , P hmtB , and P mrt(M.t.) ] based on our previous results and the peer-reviewed literature and compared the effects of these promoters on gene expression with the established enzyme assay ( Fig. 3 and Text S1I) (22,32). With our optimized quantitative b-galactosidase enzyme activity assay, we found that the P synth promoter, without a transcription factor B recognition element (BRE) sequence, resulted in a significantly higher b-galactosidase enzyme activity (510 6 50 Miller units) than did the empty-vector control (  Sequence repeats in P mrt(M.t.) are underlined. The transcription start site is indicated by "11," highlighted in bold, and underlined. TATA box sequences of P synth , P synth(BRE) , and P hmtB are surrounded by a box. BRE sequences are highlighted in italics and ribosome-binding sites in red. Dashes are used as spacers, while dots indicate additional base pairs, which are left out here for visualization. Differences between P synth , P synth(BRE) , and P hmtB between the TATA box sequence and transcription start site are highlighted in bold. (B) Qualitative analysis of BgaB activity with S-Gal as chromogenic substance in an in vitro assay with cell lysate of empty-vector-carrying M. thermautotrophicus DH (pMVS-V1) or pMVS1111A:P synth -bgaB-carrying M. thermautotrophicus DH (P synth -bgaB) cells. (C) Quantitative analysis of BgaB activity with ONPG as chromogenic substance in an in vitro assay with cell lysate of M. thermautotrophicus DH strains that carry plasmids with the bgaB gene under the control of the four distinct promoters [pMVS-V1, empty-vector control; P mrt(M.t.) -bgaB, pMVS1111A:P mrt(M.t.) -bgaB; P synth -bgaB, pMVS1111A:P synth -bgaB; P synth(BRE) -bgaB, pMVS1111A:P synth(BRE) -bgaB; P hmtB -bgaB pMVS1111A:P hmtB -bgaB]. Average (n = 3) with error bars indicating standard deviation. Significance was tested with Student's t test (two-tailed): *, significant difference (P , 0.05); **, highly significant difference (P , 0.01); n.s., no significant difference (P . 0.05).
Shuttle-Vector System for M. thermautotrophicus ® Miller units) (Fig. 3C). Therefore, this promoter has to be considered inactive under the tested conditions (Fig. 3C). We did not include the commonly used P mcrB(M.v.) promoter for methanogen genetic systems in this comparison, because we already had found that this promoter is not functional in driving the neomycin-selectable marker (Text S1I).
The metabolism of M. thermautotrophicus DH can be amended to enable growth on formate as an alternative substrate. Formate as the sole carbon and energy substrate can be utilized by several methanogens, such as Methanococcus spp. (33), Methanobacterium spp., and also Methanothermobacter spp. (34). For example, the strain M. thermautotrophicus Z-245 can grow with only formate, instead of molecular hydrogen and carbon dioxide, while M. thermautotrophicus DH is not able to grow with only formate ( Fig. 4 and Fig. S8) (34,35). It was hypothesized by Nölling and Reeve (35) that the genetic reason for this is the missing formate dehydrogenase (fdh) operon in the genome of M. thermautotrophicus DH compared to the same genomic region in M. thermautotrophicus Z-245. We argued that we can test this hypothesis in vivo, by providing the fdh operon as a genetic cargo in the application module of our shuttle vector. Thus, we constructed the shuttle vector pMVS1111A:P hmtB -fdh Z-245 that contains the entire fdh operon from M. thermautotrophicus Z-245, including the genes (in this order) fdhC, fdhA, and fdhB and additionally an open reading frame with unknown function (orf3), as indicated in the work of Nölling and Reeve (35), under the control of the constitutive P hmtB promoter (Fig. 3). In control experiments with M. thermautotrophicus Z-245, we confirmed growth of this microbe with either formate or molecular hydrogen and carbon dioxide as substrates (Fig. S8). Growth on formate was possible with M. thermautotrophicus DH cells that carry pMVS1111A:P hmtB -fdh Z-245 but not with cells that carry the empty-vector control pMVS-V1, and thus, no growth was observed from the small amount of yeast extract that we had added to provide any potentially missing micronutrients ( Fig. 4 and Materials and Methods). The formate dehydrogenase-producing strain had a prolonged lag phase with formate but reached a final optical density at 600 nm (OD 600 ) comparable to growth on molecular hydrogen and carbon dioxide. Hence, the hypothesis of Nölling and Reeve (35) was proven to be correct.

DISCUSSION
Here, we reported a robust method for genetic manipulation of M. thermautotrophicus DH (Fig. 2), including replicating shuttle vectors, which we based on the cryptic  (12). To achieve this, in a first step, we investigated different plating protocols to obtain clonal populations. Importantly, while we achieved plating efficiencies of up to 100%, the plating conditions considerably influenced the outcome over a range of 3 orders of magnitude (see Text S1J in the supplemental material). Thus, for specific purposes the plating technique has to be carefully considered to avoid misleading results, such as false interpretations of DNA transfer events. At a too-high plating efficiency, spontaneously resistant cells might overgrow genetically modified M. thermautotrophicus DH.
In general, it was noticeable that spontaneously resistant M. thermautotrophicus DH cells appeared readily with the antibiotics simvastatin and neomycin. Once those spontaneously resistant cells appeared, they were not inhibited or delayed in growth in subsequent transfers in liquid medium as well as on solidified medium plates. This has been already reported with neomycin for M. thermautotrophicus DH (19), M. maripaludis, and Methanococcus vannielii (36). However, in comparison to M. maripaludis for which spontaneously neomycin-resistant colonies appeared to be smaller than genetically modified colonies (which were found to carry a neomycin-selectable marker) (37), we cannot see this difference in colony size for M. thermautotrophicus DH. Therefore, we implemented a selective-enrichment step in liquid medium, which provided enough time for growth of the genetically modified M. thermautotrophicus DH, while at the same time it excluded the onset of growth of the spontaneously resistant cells by substrate limitation in the gas phase, because those cells appeared only after a longer incubation period (Fig. S2). After the selective-enrichment step, the genetically modified M. thermautotrophicus DH cells outnumbered the spontaneously resistant cells sufficiently to obtain and select genetically modified individual colonies on selective solidified medium plates.
Yet, we were not successful with genetically manipulating M. thermautotrophicus DH when we fused various potential selectable markers with the P mcrB(M.v.) promoter, which is commonly utilized in other methanogen genetic systems. However, after we switched to the P synth promoter to drive the neomycin-selectable marker, positive selection of shuttle vector-carrying cells with neomycin was possible in combination with the selective-enrichment step. Taken together, we believe that the most important parameters for the successful implementation of our genetic tools for M. thermautotrophicus DH, compared to previous attempts during the last four decades, were (i) the construction of the shuttle vector with restriction/ligation-independent cloning to fuse the pME2001 replicon with the other modules precisely at the IF5 sequence, to not interrupt any open reading frame or potential sequence of the origin of replication, (ii) the adaptations to the published conjugation protocol for Methanococcus spp. by Dodsworth et al. (30), specifically the applied temperature, medium, and headspace gas conditions during the spot-mating procedure, in combination with the selectiveenrichment step with limited gas supply that facilitated the selection for genetically modified cells over spontaneously resistant cells, and (iii) the utilization of a constitutive native promoter sequence, which was demonstrated to be functional in vitro and in other thermophilic archaea, because the classical P mcrB(M.v.) promoter turned out to be inactive in M. thermautotrophicus DH.
With the reporter gene bgaB, which encodes a thermostable b-galactosidase, we further confirmed the activity of this P synth promoter, as well as the activity of the P synth(BRE) and P hmtB promoters, which are all based on the promoter region upstream of the hmtB gene of M. thermautotrophicus DH. The BRE sequence upstream of the P synth sequence, which we had implemented in P synth(BRE) , and differences in the transcription-initiation region in P hmtB compared to both P synth and P synth(BRE) influenced the promoter strength significantly (Fig. 3). This was not surprising, because the critical role of the BRE sequence in archaea is well known, and similar observations were made with archaeal promoters in Saccharolobus solfataricus and Haloferax volcanii, where promoter strength was influenced by the presence or absence of a BRE sequence (38,39). Furthermore, modifications in the transcription-initiation region in the promoter sequence were shown to influence the strength of gene expression in Sulfolobus acidocaldarius and S. solfataricus (38,40). Thus, we now have established constitutive promoters of different strength. The thermostable b-galactosidase as a reporter provides an adequate basis to further investigate promoter sequences of M. thermautotrophicus DH and to establish inducible-promoter systems.
In addition to demonstrating heterologous production of functional b-galactosidase enzymes in M. thermautotrophicus DH, we demonstrated the production of an active formate dehydrogenase enzyme complex from the fdh operon from M. thermautotrophicus Z-245 (35), which amended the metabolism of M. thermautotrophicus DH for the ability to utilize formate as an alternative growth substrate (Fig. 4). These results are a cornerstone for heterologous and homologous (over)expression of genes in M. thermautotrophicus DH and enable researchers to further expand the genetic toolbox with methodology for chromosomal gene deletions and integrations such as allelic exchange, markerless mutagenesis (41), and CRISPR/Cas technology (42). We will now be able to target modifications in the metabolism of M. thermautotrophicus DH not only on the substrate but also on the product side. The possibility to change the product spectrum of hydrogenotrophic methanogens has been demonstrated already for M. maripaludis with a genetic modification that resulted in geraniol production (43). Our genetic tools for heterologous gene expression enable us to broaden the product spectrum of M. thermautotrophicus DH and to utilize this industrially relevant and robust microbe for power-to-chemicals (i.e., power-to-x) applications.

MATERIALS AND METHODS
Microbial strains, media, and cultivation conditions. M. thermautotrophicus DH (DSM 1053), M. thermautotrophicus Z-245 (DSM 3720), and M. marburgensis (DSM 2133) were obtained from the DSMZ (Braunschweig, Germany) and were cultivated in mineral medium according to basic principles for methanogen cultivation as stated in the work of Balch et al. (44), with modifications to the medium composition when required and with adjustments to state-of-the-art anaerobic handling equipment. The mineral medium contained (per liter) sodium chloride, 0.45 g; sodium hydrogen carbonate, 6.00 g; dipotassium hydrogen phosphate, 0.17 g; potassium dihydrogen phosphate, 0.23; ammonium chloride, 0.19 g; magnesium chloride hexahydrate, 0.08 g; calcium chloride dihydrate, 0.06 g; ammonium nickel sulfate, 1 ml (0.2 wt%); iron(II)chloride pentahydrate, 1 ml (0.2 wt%); resazurin indicator solution, 4 ml (0.025 wt%); and trace element solution, 1 ml (10-fold as stated in the work of Balch et al. [44]). All chemicals were per analysis grade. No vitamins were added. For the formate growth experiments, the mineral medium was supplemented with 200 mM sodium formate, 10 mM sodium molybdate, 1 mM sodium selenite, and 0.125 wt% yeast extract for all tested strains, including the empty-vector-carrying M. thermautotrophicus DH strain, to provide a source of potentially limiting micronutrients. The medium was gassed using N 2 /CO 2 (80/20 vol%) to eliminate dissolved oxygen. The pH value was adjusted to 7.2 using hydrochloric acid. As reducing agent and sulfur source, 0.5 g/liter cysteine hydrochloride, and for solid mineral medium additionally 0.3 g/liter sodium sulfide monohydrate, was added. Afterward, the mineral medium was dispensed into serum bottles inside an anaerobic chamber with a 100% N 2 atmosphere (UniLab Pro Eco; MBraun, Garching, Germany). The headspace of the serum bottles was exchanged to 200 kPa H 2 /CO 2 (80/20 vol%) and autoclaved (100 kPa, 20 min, 121°C). For the formate growth experiments, the headspace of the serum bottles was exchanged to 152 kPa H 2 /CO 2 (80/20 vol%) as positive control, which provides the same electron equivalents as 200 mM formate, or 152 kPa N 2 /CO 2 (80/20 vol %), when formate was the substrate. All Methanothermobacter strains in liquid medium were incubated at 60°C with shaking at 150 rpm (Lab Companion ISS-7100R; Jeio Tech, Republic of Korea). For cultivation of genetically modified M. thermautotrophicus DH strains, 250 mg/ml neomycin sodium salt was added. For solidified mineral medium, 1.5 wt% Bacto agar (BD Life Science, Berkshire, United Kingdom) was added as a supplement prior to autoclaving. Afterward, solidified medium plates were poured and dried for 2 h inside the anaerobic chamber. M. thermautotrophicus DH was applied to solidified medium plates by spot plating, spread plating, or pour plating. For spot plating, 50 ml of M. thermautotrophicus DH culture was spotted on a solidified medium plate. Incubation was started after the drop was completely absorbed. For spread plating, 50 ml of diluted or undiluted liquid culture was applied to a solidified medium plate and spread out with a Drigalski spatula until the liquid was completely absorbed. For pour plating, 5 ml mineral medium containing 0.8 wt% Bacto agar (soft agar) was mixed with a liquid M. thermautotrophicus DH culture and poured on top of a solidified medium plate, which contained 1.5 wt% Bacto agar. For better gas-solid mass transfer and to avoid sealing of the plates by water, paper clips were added to the petri dish prior to incubation in a custom-made stainless-steel jar (Raff 1 Grund, Freiburg am Neckar, Germany) inspired by the work of Balch et al. (44). The gas phase of the stainlesssteel jar was exchanged to 200 kPa of H 2 /CO 2 /H 2 S (79.9/20/0.1 vol%). The pressurized anaerobic jar was incubated without shaking at 60°C (Memmert, Schwabach, Germany).
For general cloning and gene manipulation, E. coli NEB stable (New England Biolabs, Frankfurt/Main, Germany) was used. E. coli S17-1 for conjugational DNA transfer was kindly provided by Wolfgang Fink et al.

®
Wohlleben of the Department for Biotechnology at the University of Tübingen, Germany (45). E. coli BL21(DE3) with pME2508 (Archaea Center of the University of Regensburg, Germany) was used to produce recombinant pseudomurein endoisopeptidase (PeiP) enzyme. E. coli was cultivated in LB medium, which contained (per liter) sodium chloride, 10 g; tryptone, 10 g; and yeast extract, 5 g, and which was supplemented with appropriate amounts of chloramphenicol (30 mg/ml), ampicillin (100 mg/ml), or kanamycin (50 mg/ml). For cultivation of E. coli S17-1, trimethoprim (10 mg/ml) was added to stabilize the genome-integrated tra module, which is responsible for mobilization of plasmid DNA (45). Solidified LB medium plates contained 1.5 wt% of Kobe I agar (Carl Roth, Karlsruhe, Germany) and were incubated at 37°C. Liquid E. coli cultures were incubated at 37°C with shaking at 150 rpm.
Molecular cloning and plasmid construction. All primers, gBlock DNA fragments (IDT, Coralville, IA, USA), and plasmids used in this study are summarized in Tables 1 to 3. PCR was performed with Q5 Hot Start high-fidelity polymerase (NEB, Ipswich, MA, USA) according to the manufacturer's guidelines and with the required primer combinations (Table 1). Primer concentrations were reduced 10-fold, and elongation time was prolonged by 1 min. Resulting PCR products were DpnI digested when required and purified using a PCR purification kit (Qiagen, Hilden, Germany). For initial fusion of the first shuttlevector construct pSV1_1, as described below, we used the Gibson Assembly Ultra kit (Synthetic Genomics, La Jolla, CA, USA). All follow-up constructs were assembled with Gibson Assembly master mix (New England Biolabs [NEB], Ipswich, MA, USA) or restriction/ligation cloning with the aid of the implemented modular restriction enzyme-recognition sites (Fig. 1). E. coli cells were transformed with DNA via chemical transformation by following a standard heat shock protocol (46). All plasmids were confirmed by Sanger sequencing (MPI Genomics Center, Tübingen, Germany). pCF200, which contains the puromycin acetyltransferase (pac) gene (Pur r ) from Streptomyces alboniger as a codon-optimized version for M. thermautotrophicus DH under the control of the P mcrB(M.v.) promoter and the T mcr terminator from Methanococcus voltae (47), was completely synthesized (BioCat, Heidelberg, Germany). The pac gene in pCF200 was exchanged to the 3-hydroxy-3-methylglutaryl-coenzyme A reductase (HmgA)-encoding gene (Sim r ) from Thermococcus kodakarensis by using pYS3 (21) and pCF200 as the templates for Gibson Assembly, resulting in pCF203. pCF203, pME2001 (extracted from wild-type M. marburgensis), and pBBR1-MCS2 (Addgene catalog no. 85168) were used as the templates for Gibson Assembly with the Gibson Assembly Ultra kit (Synthetic Genomics, La Jolla, CA, USA) and resulted in the putative shuttle vector pSV1_1. pSV1_1 was the basis for further shuttle vectors. For the introduction of a high-copy-number replicon for E. coli, a tra region for plasmid mobilization, and an additional AsiSI restriction enzyme-recognition sequence, the pBBR1-MCS2 backbone was exchanged to the E. coli vector backbone from pMTL83151 (27), including Cam r , ColE1, and the tra minigene for mobilization, via Gibson Assembly resulting in pSV1_2. To implement the thermostable neomycin phosphotransferase gene (Neo r ) (28), the Pur r from pCF200 was exchanged to Neo r from pMU131 (23) by Gibson Assembly. Afterward, based on pCF404, the Neo r under the control of the P mcrB(M.v.) promoter and the T mcr terminator was used to construct a putative integration plasmid for the exchange of an annotated pyrF gene in M. thermautotrophicus DH, using AscI and FseI as restriction enzymes and T4 ligase for ligation, resulting in pCF407. In pCF407 the P mcrB(M.v.) promoter was exchanged to P synth by inverse PCR resulting in pSB1. The fragments P mcrB _Neo r _T mcr from pCF407 and P synth _Neo r _T mcr from pSB1 were used to substitute for the Sim r in pSV1_2 by restriction-ligation cloning using AscI and FseI, resulting in pSV1_3 and pMVS-V1, respectively. To generate pMVS1111A:P synth -bgaB, the PCR-amplified gBlock with the thermostable b-galactosidase (bgaB) gene, which was codon optimized for M. thermautotrophicus DH and which was placed under the control of the P synth promoter, was fused to AscI-digested pMVS-V1 with Gibson Assembly. The AscI restriction enzyme-recognition sequence was recovered at the intersection with the selectable-marker module, and a PacI sequence was introduced at the intersection with the M. thermautotrophicus DH replicon module. Further promoters [P synth(BRE), P hmtB , P mrt(M.t.) ] were amplified via overlap-extension PCR of the b-galactosidase gBlock and promoter gBlock and inserted by restriction/ligation cloning using restriction enzymes PacI and AscI. pCF201 was constructed by amplifying the fdh Z-245 operon from M. thermautotrophicus Z-245 genomic DNA and introducing the fragment into pUC19 by Gibson Assembly. Gibson Assembly was used to exchange the Neo r -coding region in pCF204 with the fdh Z-245 operon from pCF201, resulting in plasmid pLM201. The promoter P mcrB(M.v.) in pLM201 was exchanged to P hmtB by inverse PCR of the complete plasmid, except of the P mcrB(M.v.) sequence, with primers containing overlapping parts of P hmtB in the overhangs, and direct transformation of E. coli with the linear PCR product. The resulting plasmid pLM202 was used to amplify the P hmtB -fdh Z-245 cassette by PCR to include PacI and AscI restriction enzyme-recognition sequences. Restriction/ligation cloning with the restriction enzymes PacI and AscI was used to exchange the b-galactosidase gene in pMVS1111A:P synth -bgaB for the P hmtB -fdh Z-245 cassette to give pMVS1111A:P hmtB -fdh Z-245 .
Plasmid DNA extraction from Methanothermobacter spp. For plasmid DNA extraction from Methanothermobacter spp., 10 ml of liquid cell culture was centrifuged at 3,700 rpm for 15 min at room temperature (Centrifuge 5920 R, rotor S-4x1000; Eppendorf, Hamburg, Germany). The supernatant was discarded, and the cell pellet was resuspended in 150 ml of sucrose (30 wt%)-containing buffer P1 (from QIAprep Spin miniprep kit; Qiagen, Hilden, Germany). For lysis of Methanothermobacter cells, alkaline lysis was combined with enzymatic lysis by adding a final concentration of 100 ng/ml of pseudomurein endoisopeptidase (PeiP) to the sample prior to incubation for 1 h at 60°C. The pseudomurein-degrading enzyme PeiP, which lyses pseudomurein-containing Methanobacteriales cell walls, was produced as a heterologous 6ÂHis-tagged protein from pME2508-carrying E. coli BL21(DE3) as described previously (48). The recombinant protein was purified via a Protino Ni-TED column according to the manufacturer's guidelines (Macherey1Nagel, Düren, Germany). After the PeiP treatment, the QIAprep Spin miniprep kit  Shuttle-Vector System for M. thermautotrophicus ® (Qiagen, Hilden, Germany) manufacturer's guidelines were followed with final elution in 40 ml nucleasefree water.
Interdomain conjugational DNA transfer. DNA transfer via interdomain conjugation was achieved between E. coli S17-1 and M. thermautotrophicus DH. E. coli S17-1 was transformed with the respective shuttle vector. Overnight cultures of the respective E. coli S17-1 donor strains were inoculated. At the same time, 20 ml of liquid mineral medium was inoculated with wild-type M. thermautotrophicus DH (recipient). The overnight culture of E. coli S17-1, which contained the shuttle vector, was diluted into 10 ml of fresh LB medium in a sterile 50-ml baffled flask for better aeration to give an OD 600 of 0.3 to 0.5.  TACTATTTTGCCCAATCTAATAATAAATTAATAATAAATTAGTAATTAACTTAGATTTTAGCAGGTAAGTGGGGT  CGTGCGCTCCTTTCGGTCGGGCGCTGCGGGTCGTGGGGCGGGCGCTAGACCTTGCCGGCTTCGTCGTGTTCCC  TAAGGACAGCGACGTCGACGCCCTGAATCCTGAGTTCACCCCCCCTGAAGCATTTGCCATCTATCATATTCTGG  TAGATCTTATCTTCCGGAAGGGAGAGTGTGACCTCATAGTCGTTGTGGTTAATTATAATAAGGTACTTCCATTCA  TCGGTCTCCCTCTGCTGAACTTCGACATTCTCAGCAACCTCCAGTATAGGATTTATGTGGTGTTTAGCAAACACC  TGTTCGAGAAGCCTGCCAAGGTAGTTGCTGTCAGGGTATGTTCCTACGTATATGCCCTCCCCCTTTCCGTAGCAG  TTCCTGGTAACAGCAGGAAGGCCGGCATACCAATCACCTTTGAATGTGGCGAGAGGCTCAGCACCTTCCAGCC  TTATTATATCGGCCCATGTGGTACAGTCATACTCGCCGTCGTTTGAGTAGATCTTATTCACCTTTGTCTCGGGATA  AGGAACGAATTCCTCCACGAAGATGCCGAGAATGTCCCTCAGCGGTCCTGGATATCCCCCGAGGTGCACTCTA  TCGTTCTCATCGACTATCACACTGAAAAAGCTTACAATCAGGGTTCCGCCGTTTGCGACAAACTGCCTAAGGTT  CTCATCTTCTCCCTCTTTCACCATATACAGCATCGGTGCAATAACAACCTTATATTTTGTGAGATCGTCGGACGG  TCTTACAAAGTCGACTGCTATGTTTCTCTTGTAAAGCTCTCTATAATATGCCTCTACTATGGGAATATATCTGAGC  TTGTTGTGCGGTTTGGAACTGAGCTCAACTGCCCACCAGTTTTCCCAGTCAAAGATAATTGCCACCTCTGCCTTT  ATTCTACTCCCCACGAGGCAGTCAAGTTTTTTCAGCTCCTGGCCAAGCTGGGTAACTTCCCTGTATATTCTATTG  TTTTCGTTAAGAAAGTGGGGCACCATTGCTCCGTGAAACTTCTCAGCTCCTGCTCTGGACTGCCTCCACTGAAA  GAACATTATCCCATCGGCACCCCTGGCGATTGTTGCGTAACTCCAGAGTCTCATAACCCCCGGCGGCTTTGGCA  CATTGATATCTCTCCAATTAACGTGACTGGTGACCTGCTCCATAAGAATGAACGGCTGCCCCTTCCTAAGTGAC  CTCATGAGGTCATTCATCATTGCGTGCTGTATAGGGAGTCCCTCCCTGGGATCTGGGTAGCTATCCCAGGTAAC  GATATCTACGTGCTGAGCCCACTGAAAGTAGTTGAGTGGCTTGAATGATCCCATGAAATTTGTGGAGACCGGG  ATATCGGGGGTTACTTCCCTGAGGATCTCCTTTTCTGTAAGGAAGAGTTTGAGGATTGAATCATTCATGAATCTG  TAGTAATCAAGCTCCTGGCTGGGGTTAATAAATGTTGGTGCCTTCCTAGGGGGATTAATCTCATCCCAGTGGTT  ATATCTCTGGCCCCAGAAGTTTGTACCCCATCTTTCATTAAGTTCATCAATGGTCTTATACCTTTCTTTAAGCCAT  TTTCTGAAAGCAACTGCGCAATTCTCACAGAAACACTTACTTACATGGCAAGCGTATTCGTTATTTACGTGCCAC  ATTTTGAGGGCTGGATGATTTTTGTATCTCTCAGCTATAGCCCTTACCAGCCTCTTTATATGTGTTATAAGCTGAG  GGTGATTTGGGCAATAATGCTGTCTACTCCCGAAACTCAGTATCACACCGGACTCGTCAATAGGGAGTGAATCA  GGGTATTTCTTCACGAACCAGGCGGGTGTGGTTGCGGTGGCGGTCCCCAGATTTATGTATACCCCATGATCGTA  GAGGATGTCTATCACTTTGTCGAGCCATTCAAAATCAAATACACCGTCTGATGGCTCGATTTTGGACCAGCTAA  AGATTCCGAGTGAAACAAGATTAACACCGGCCTTCTGCATAAGTTTTGCGTCCTCGTACCATATCTCCTCGGGCC  ACTGTTCTGGGTTGTAATCCCCCCCATAGCAGATGGAACTGAGAACGTTCATATGCATCACCTGCTAGCCGGTG  GTTACCGTGATATTATCTATTACTATATCCCTATATAAAGGCATTCCTCCAGGCTTAATTAAC   This study Fink et al. When this culture reached an OD 600 of 2.0 to 2.5, the incubation was stopped and the culture was harvested aerobically at 3,700 rpm for 10 min at room temperature (Centrifuge 5920 R, rotor S-4x1000; Eppendorf, Hamburg, Germany). The supernatant was discarded, and the E. coli S17-1 pellet was transferred into the anaerobic chamber. Wild-type M. thermautotrophicus DH was grown to early stationary growth phase (OD 600 of 0.25 to 0.35). 8 ml of M. thermautotrophicus DH culture were centrifuged stepwise at 12,500 rpm for 4 min at room temperature (MySPIN 12 minicentrifuge; Thermo Scientific, Waltham, MA, USA) inside the anaerobic chamber. The final pellet was resuspended in 250 ml of the original nonconcentrated M. thermautotrophicus DH culture and gently mixed with the E. coli S17-1 pellet. 100 ml of cell suspension were anaerobically spotted on solid LB-MS medium, which was a mixture that consisted of 50 vol% of mineral medium and 50 vol% of LB medium without the 10 g/liter sodium chloride. The spot was dried, while the lid of the petri dish was kept slightly open for 1 h at 37°C in the incubator (Coy Laboratory Products, Grass Lake, MI, USA) within the anaerobic chamber. When the spot was completely absorbed, the plates were provided with paper clips and transferred to a stainless-steel jar. To determine the conjugation frequency, the following modifications to the standard protocol were made: (i) the cell count of M. thermautotrophicus DH in liquid culture was determined by counting in a Petroff counting chamber (the initial recipient cell number in 100 ml of the stepwise-concentrated culture was calculated based on this cell count), (ii) the 5-ml nonselective recovery culture from the washed spot after spot mating was incubated for ;16 to 20 h instead of 3 to 4 h, and (iii) 100 ml of the nonselective recovery culture was directly spread plated on selective solidified medium plates, without a liquid selective-enrichment step.
Molecular methods for analysis of genetically modified M. thermautotrophicus DH. PCR analysis was performed from liquid cultures and directly from individual colonies. 100 ml of liquid culture or one individual colony, which was resuspended in 40 ml of deionized water, was boiled for 12 min at 100°C (ThermoMixer C; Eppendorf, Hamburg, Germany). After cooling the sample on ice, 1 ml was added to a 10-ml PCR mix. PCR was performed using Phire plant PCR master mix (Thermo Scientific, Waltham, MA, USA). The denaturation and annealing times were increased from 5 s to 20 s and to 10 s, respectively. Thirty cycles were performed for all analyses. We observed false-positive PCR signals for shuttle-vector DNA due to plasmid DNA carryover from E. coli for two transfers after the nonselective liquid recovery step. After the third transfer, plasmid DNA from E. coli was not detectable anymore in any of our experiments. For robust PCR amplifications of individual colonies from M. thermautotrophicus DH, it was crucial to keep the agar contamination of the PCR sample as low as possible. Therefore, even though the plating efficiency is higher with pour plating, genetically modified M. thermautotrophicus DH strains were spread plated instead of pour plated. This led to a lower total colony count but to more reliable results.
Additional to PCR analysis, plasmid DNA from genetically modified M. thermautotrophicus DH strains was extracted as described above. The purified plasmid DNA was used for retransformation of E. coli NEB stable. Analysis of E. coli NEB stable colonies was performed via test restriction digestions and Sanger sequencing for further confirmation of stable replication of shuttle vectors in M. thermautotrophicus DH.
b-Galactosidase enzyme activity assays. For a qualitative b-galactosidase enzyme activity assay with the lactose analogue S-Gal, 2 ml of overnight cell cultures that carry pMVS-V1 or pMVS1111A:P Synth -bgaB was harvested by centrifugation for 4 min at 13,000 rpm at room temperature (Centrifuge 5424, rotor FA-45-24-11; Eppendorf, Hamburg, Germany). The supernatant was discarded, and the samples were stored at 220°C until further use. All samples were resuspended in 100 ml buffer P1 (from Qiagen QIAprep Spin miniprep kit) containing sucrose (30 wt%) and lysed by adding 100 ng/ml PeiP, followed by incubation for 30 min at 60°C. Fifty microliters of the cell lysate was incubated with 250 mg/ml S-Gal and 250 mg/ml ammonium ferric citrate in 1 ml LB medium, which provided any potentially required trace compounds. The samples were incubated for 1 h at 60°C. After ;30 min, a color change was visible.
For a quantitative b-galactosidase enzyme activity assay with the lactose analogue ONPG, 4 ml of cell culture was harvested anaerobically by stepwise centrifugation (Centrifuge 5424, rotor FA-45-24-11; Eppendorf, Hamburg, Germany). Afterward, the same lysis procedure for samples was applied as for the S-Gal assay. The resulting cell lysate was used for a quantitative in vitro b-galactosidase enzyme activity assay with ONPG as chromogenic substance according to the method of Jensen et al. (31). In brief, 12.5 ml of cell lysate (equal to 0.5 ml of original cell culture) was mixed with 600 ml of ONPG (1 mg/ml)-containing substrate solution. The mixture was incubated for 2 h at 60°C. Afterward, 200 ml was added to 200 ml of 1 M sodium bicarbonate stop solution in a 96-well plate. The absorbance at 420 nm was measured in a microplate reader (Multiskan Go; Thermo Scientific, Waltham, MA, USA). For the preliminary experiment (see Fig. S7 21 ]. High-performance liquid chromatography for formate analysis. Formate concentrations were analyzed via a high-pressure liquid chromatography (HPLC) system as described in the work of Klask et al. (49). For HPLC sample preparation, all culture samples were centrifuged for 5 min at 13,000 rpm (Centrifuge 5424; Eppendorf, Germany) in 2-ml reaction tubes. 450 ml of the supernatant were transferred into clean reaction tubes and stored at 220°C until use. Frozen samples were thawed at room temperature. The samples were vortexed and centrifuged again, and 400 ml of the supernatant was transferred into short-thread HPLC/gas chromatography (GC) vials (glass vial ND9; VWR, Germany) and sealed with short screw caps, which contained rubber septa (6 mm for ND9; VWR, Germany). Formate standards (1 to 500 mM) were prepared freshly for the analysis. All HPLC samples were randomized.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. TEXT S1, DOCX file, 0.1 MB.