Abstract
As methanol can be derived from either CO2 or methane, methanol economy can play an important role in combating climate change. In this scenario, rapid utilization of methanol by an industrial microorganism is the first and crucial step for efficient utilization of the C1 feedstock chemical. Here, we report the development of a methylotrophic E. coli strain with a doubling time of 3.5 hours under optimal conditions, comparable or faster than native model methylotrophs Methylorubrum extorquens AM1 (Td~4hr) and Bacillus methanolicus at 37°C (Td~5hr). To accomplish this, we develop a bacterial artificial chromosome (BAC) with dynamic copy number variation (CNV) to facilitate overcoming the formaldehyde-induced DNA-protein cross-linking (DPC) problem in the evolution process. We track the genome variations of 75 cultures along the evolution process by next-generation sequencing, and identified the features of the fast-growing strain. After stabilization, the final strain (SM8) grows to 20 g/L of cell mass within 77 hrs in a bioreactor. This study illustrates the potential of dynamic CNV as an evolution tool and synthetic methylotrophs as a platform for sustainable biotechnological applications.
Similar content being viewed by others
Introduction
Methanol is the most reduced C1 compound that exists in the liquid form in ambient conditions1,2. This property by-passes difficulties in transportation and storage of C1 compounds. As such, use of methanol as a starting point for chemical production has been proposed as a potential route to reduce greenhouse gas accumulation, since methanol can be derived from CO2 or methane. In this regard, developing industrial microorganisms that can use methanol for growth and production has been a worthwhile goal3,4,5,6,7,8,9,10,11,12,13.
Rapid growth of microorganisms in methanol minimal medium is an essential step and the first requirement in microbial conversion of methanol. However, current synthetic methylotrophic strains evolved from non-C1-utilizing microbes still grow much slower than typical industrial strains5,7,9,14, except a recent publication which appeared after submission of this work13. Laboratory evolution, a widely adopted approach in engineering microorganisms, is particularly suitable for improving growth rate since it can enrich fitter organisms4,5,6,8,14,15. Methanol-utilizing synthetic methylotrophs face a particular problem, formaldehyde-induced DNA-protein crosslinking (DPC) caused by the imbalance between formaldehyde production during the first step of methanol oxidation, and formaldehyde consumption by the downstream pathways5. These synthetic organisms may accumulate and release formaldehyde during later growth phases, ultimately leading to self-destruction and poisoning surrounding cells. This scenario impedes beneficial mutations from accumulating during the evolution in the presence of methanol. This may be one of the underlying reasons for the difficulty in engineering or evolving a fast-growing synthetic methylotroph to date.
Formaldehyde is an essential intermediate in methanol utilization, but its accumulation can induce DPC, which forms aggregates between DNA and proteins that disrupt cellular functions. Severe DPC can ultimately cause cell death16,17. Native methanol-utilizing organisms have evolved sophisticated regulatory mechanisms to avoid DPC by balancing the formaldehyde production and consumption18. One such mechanism is the glutathione-dependent formaldehyde oxidation pathway that converts formaldehyde to formic acid and serves as a surge release18. However, this pathway reduces the biomass yield of the methanol-grown strain and is not desirable for our purpose here. In fact, this protective pathway was deleted in multiple reported synthetic methylotrophs4,5,6,8,10.
Without the protective formaldehyde oxidation pathway, this intermediate is likely to accumulate, as we push for fast methanol conversion. Once formaldehyde accumulates, it causes the DPC problem that kills the producing cells. Formaldehyde can also diffuse in the medium and cause DPC in other cells. Thus, even if the cell is engineered to utilize methanol for fast growth under a particular condition, formaldehyde accumulation may occur at a later stage and cause DPC to kill itself and others. Consequently, cells with beneficial mutations are difficult to enrich or accumulate. Moreover, cells in liquid culture or colonies on solid medium cannot be re-cultured until a long lag phase has passed5. Since all DNA and proteins are susceptible to DPC, formaldehyde-resistant cells are difficult to emerge. The DPC problem is particularly severe in high methanol concentrations.
Solving the DPC problem requires gain-of-activity mutations to increase the formaldehyde consumption enzyme activities or expression levels, or loss-of-activity mutations for the formaldehyde production enzyme methanol dehydrogenase (Mdh). The latter type of mutations occurs more frequently, though in this case also reduces the methanol assimilation and growth rate. The former type of mutations occurs rarely, but is likely beneficial. However, if a traditional plasmid system is used for gain-of-activity, it does not have the ability to tune the expression level for optimal performance.
Previously, we reported a synthetic methylotroph5, SM1, that grows on methanol as the sole carbon source with a doubling time of 8 h. This strain features a chromosomal segment containing 5 copies of IS5-mediated tandem repeats, each spanning 70 kb. This result suggested that transposon-mediated copy number variation (CNV) of methanol assimilation genes may be a major breakthrough in the evolution of a synthetic methylotroph. However, increasing copy number through transposons is a random process, and cannot generate a broad range of copy numbers. While this strain has alleviated the DPC problem, its growth rate (Td = 8.5 h) is still far below that of native methylotrophs.
In this work, we develop a BAC system, called ddp-BAC, that exhibits a dynamic CNV property through the ddp operon. The presence of such a sequence enables the whole BAC to generate tandem repeat concatemers with a wide range of copy numbers. Although such a system may increase the metabolic burden because of increased copy number, it provides a tentative fix and allows the cell to find a better solution. After de-bottlenecking, the strain is able to evolve and develop various beneficial mutations. Eventually the copy number of the ddp-BAC system is reduced to avoid metabolic burden and a single colony is isolated with a doubling time (Td) of 3.5 h, demonstrating faster growth than native model methylotrophs, Methylorubrum extorquens AM1 (Td ~ 4 h)19 and Bacillus methanolicus grown at 37 °C (Td ~ 5 h)20. We characterize the metabolic and genetic features of SM7 that contribute to its rapid growth. Finally, we stabilize SM7 by restoring the wild-type mutS to achieve SM8, which maintains a doubling time of around 3.5 h.
Results
Dynamic copy number variation introduced by ddp-BAC
To address the DPC issue that hinders long-term evolution, we developed a dynamic CNV approach to tune the expression levels of formaldehyde consumption enzymes. This scheme is based on our previous discovery5 of yet another tandem repeat of 7 kb region in the strain (BB1) (see Supplementary Fig. 1), which spontaneously emerged with the synthetic methylotrophic SM1 strain that can utilize methanol as the sole carbon source for growth (see Supplementary Fig. 2 for 13C labeling data). Interestingly, this 7 kb tandem repeat is non-IS mediated, and contains the ddp operon.
We cloned this operon onto a bacterial artificial chromosome (BAC) with a gfp expression cassette. Surprisingly, we found that the presence of the ddp operon in the BAC spontaneously developed tandem repeats without selection pressure, forming a circular concatemer of the whole BAC (Fig. 1a–c). The BAC containing the ddp operon is named ddp-BAC. Furthermore, we found that the concatemer and CNV did not occur in a ΔrecA strain from the Keio collection21 (Fig. 1b, c), indicating a RecA-dependent process in CNV. This RecA-dependent process can occur dynamically within a cell population, as demonstrated by time-lapse light microscopy (Fig. 2a). Using flow cytometry, we found that the ddp-BAC::gfp exhibited a bimodal distribution (Fig. 2b) with fluorescence intensity spanning 4 orders of magnitude (102–106), which presumably reflects copy number distribution. The mechanisms underlying the dynamic CNV is currently under investigation and will be reported elsewhere.
a A schematic showing BAC with and without the ddp operon. b The copy number of gfp measured by droplet digital PCR (ddPCR). Each strain was tested with at least 3 individual colonies from a LB plate. Only ddp-BAC with the ddp operon in the recA+ host existed in high copies (BW25113 strain with ddp-BAC::gfp, n = 6; BW25113 strain with BAC::gfp and ΔrecA strain with BAC::gfp or ddp-BAC::gfp, n = 3; all the data are biological repeats; data are presented as mean values ± SD). c Nanopore sequencing of ddp-BAC. The patterns shown are raw intact nanopore reads mapped to BAC. The orange label represents the ddp operon, while the blue arrows represent the BAC backbone with gfp. Only ddp-BAC in the recA+ host showed tandem repeats with a distribution of copy numbers. Source data are provided as a Source Data file.
a Time-lapse live cell imaging of GFP fluorescence. Colonies from ddp-BAC::gfp showed varied fluorescence levels both among different colonies and within the same colony. The arrows highlight the progression of the same colony formation from the initial time point (0 min) to 200 min in each column. The red and blue arrows indicate low and high fluorescence, respectively. In contrast, colonies from strains containing BAC::gfp (magenta arrows) or the high-copy pUC19::gfp (orange arrows) demonstrated uniform fluorescence intensities. GFP represents the view of green fluorescent protein fluorescence, PH represents the view of phase contrast optical microscopy, and Merge is the composite image combining both the GFP and the PH views. b Flow cytometry analysis of BW25113 containing various versions of BAC or pUC19. Results show that only ddp-BAC::gfp exhibited a bimodal distribution. The x-axis represents the fluorescence intensity in arbitrary units (a.u.) on a logarithmic scale, and the y-axis represents the count of cells. The green dashed line indicates the lowest threshold determined by WT BW25113 strain in autoclaved water for positive signals of GFP-expressing cells. Source data are provided as a Source Data file and also available at Figshare41.
Nevertheless, this wide distribution of copy numbers within a population and the efficient dynamic tuning of copy number by RecA could potentially be useful in laboratory evolution.
Use of ddp-BAC to de-bottleneck evolution
The synthetic methylotrophic E. coli SM1 that we previously developed5 has a doubling time of 8.5 h. This relatively slow growth was presumably due to the residual DPC issue. The residual DPC issue was amplified when the strain was grown in high methanol concentrations (Fig. 3a). We initially evolved the SM1 strain for 81 passages, but the improvement of growth rate was marginal, and genomic changes were minor (Fig. 3b and Supplementary Tables 1 and 2). Additionally, this strain grew poorly in high methanol medium (1000 mM) and could not continue to grow in the second passage (Fig. 3c), suggesting a severe DPC problem in high methanol medium that killed the cells. Therefore, we decided to use the ddp-BAC as a tool to tune the copy number of the formaldehyde consumption enzymes.
a Formaldehyde concentration in SM1 cultures at an OD600 of around 1.0, grown in cell culture tubes with low (400 mM, M400) and high (1000 mM, M1000) methanol concentrations. (n = 5 for each group, biological repeats; data are presented as mean values ± SD). b Doubling time of the SM1 strain evolved up to passages 77 showed minor improvement. c SM1p with various plasmids harboring rpe, hps, tkt, tal, phi (RHTTP) grown and passed in 1000 mM methanol. Each strain was tested with at least 3 colonies from a LB plate. d The growth curves and the doubling time of SM7 grown in 400 mM methanol. The different colors represent different biological repeats (n = 3). Source data are provided as a Source Data file.
We overexpressed the genes for formaldehyde utilization in the ribulose monophosphate (RuMP) cycle—rpe (E. coli), hps (Methylotuvimicrobium buryatense 5GB1C), tkt (E. coli), tal (E. coli), and phi (Methylobacillus flagellatus) (collectively RHTTP) on ddp-BAC, BAC, and pUC19, and transformed into strain SM1p, which is a descendent of SM1 after 81 passages (equivalent to about 284 generations) of laboratory evolution in 400 mM of methanol minimal medium (Supplementary Table 1). The ddp-BAC can dynamically generate tandem repeats with a broad range of copy numbers, which provides a library of various RHTTP copy numbers for finding optimal expression levels to bypass the DPC problem.
The presence of ddp-BAC::RHTTP system (pLY147) enabled the culture to grow in 1000 mM of methanol immediately and can continue to grow without a lag phase in the subsequent passages (Fig. 3c). As a comparison, the strain containing BAC::RHTTP (pLY143, without the ddp operon) grew slower and exhibited a lag phase in the second passage (Fig. 3c), except one colony. The beneficial effect of ddp-BAC::RHTTP is presumably attributed to the high copy number of RHTTP on the ddp-BAC system. To test if a high copy number plasmid (pUC19) can achieve the same effect as the ddp-BAC system, we also expressed RHTTP using pUC19. Interestingly, the pUC19::RHTTP system did not allow the strain to grow in 1000 mM of methanol (Fig. 3c), presumably due to high metabolic burden. High-copy-number plasmids often inhibit growth by diverting cellular resources away from essential components22,23,24,25. In contrast, the ddp-BAC system can adjust copy number for optimal growth. As another control, the strain without any RHTTP overexpression did not grow under the same condition (Fig. 3c).
Evolution to develop a fast-growing methylotrophic E. coli strain
We used the ddp-BAC::RHTTP system to continue the evolution in 1000 mM methanol for 18 passages, and observed the improvement of growth rate (Fig. 4a). We then raised the methanol concentration to 1200 mM to increase the selection pressure. Again, the growth rate in high methanol concentration continued to improve. At passage 132, we sequenced the strain and found that the recA gene on the chromosome was interrupted by an insertion sequence (IS2), and the copy number of the RHTTP gene on the ddp-BAC decreased to 1. This result suggested that during the evolution, the initial high copy number of RHTTP (Fig. 4a) allowed the strain to avoid DPC and continued to evolve, even though the high copy of RHTTP may introduce a metabolic burden. During the subsequent evolution process, the cell managed to accumulate other mutations and did not need the high copy RHTTP. Thus, an IS2 insertion of recA was enriched and dominated the culture.
a The copy number of ddp-BAC and growth rate changes along the evolution line towards SM8. The copy number of ddp-BAC was determined from Illumina sequencing. The purple line corresponds to the copy number of ddp-BAC. The red line indicates the growth rate in 400 mM methanol, while the gray line represents the growth rate in high methanol (1000 mM or 1200 mM). The red dashed line represents the growth rate where mutS is not restored. (Each passage is defined as a subculturing step, growing from OD600 0.05–0.1 to OD600 1). kb refers to kilobases, where it depicts the unit size of IS5-flanked tandem repeats derived from the original 70 kb tandem repeats. b The overall accumulation of different kinds of mutations throughout the evolution line, including SNP, INDEL, and Silent mutations. The numbers depict the passage number of the strain. All dots represent samples that were sequenced by NGS. Source data are provided as a Source Data file.
During the evolution in 1200 mM methanol, the strain presumably has evolved more efficient ways to avoid the DPC problem. We then focused on increasing methanol consumption by introducing additional copies of mdh (an improved mdh from Cupriavidus necator N-1 through directed evolution26) into the ddpA site on the chromosome and the ddp-BAC. As the DPC issue has been apparently alleviated, we returned to the evolution process in low methanol (400 mM) medium to evolve for faster growth. At passage 259, we isolated a culture by fluorescence-activated cell sorting (FACS), which had a doubling time of less than 4 h and named it SM6. From this culture, we continued to evolve for 28 passages and isolated a strain (SM7) with an even better growth rate, having a doubling time of 3.5 h (growth rate = 0.2 h−1) (Fig. 3d and 4, Supplementary Figs. 3, 4a). This growth rate is faster than model methylotrophs Methylorubrum extorquens AM1 (Td ~ 4 h) and Bacillus methanolicus grown at 37 °C (Td ~ 5 h). We measured the dry weight of the biomass and the methanol consumed in the Falcon tube culture, and obtained the biomass yield of 0.39 g biomass/g methanol.
Tuning of copy number by recA
To further understand the genomic variations during the evolution process, we sequenced 75 cultures using Illumina Mini-seq along the main evolution line and in the branched evolution lines (Fig. 4 & 5).
At passage 81 when ddp-BAC::RHTTP was first introduced, its copy number increased and reached 30 copies within about 20 passages in high methanol concentration (1000 mM or 1200 mM), then decreased to 1, where an IS2 insertion to the recA gene was also observed (Figs. 4, 5). Presumably during this period of time, the cell accumulated sufficient beneficial genomic changes, and do not need the high copy RHTTP anymore. Thus, a recA inactivated population was enriched to reduce the copy number to 1 and avoid metabolic burden. The genomic changes include accumulation of various mutations and the shrinkage of the IS5-mediated 70 kb repeat regions that emerged in the evolution of SM1 (Fig. 5 and Supplementary Fig 4b). These may contribute directly or indirectly to the growth phenotype, and allow the cell to reduce the copy number of ddp-BAC::RHTTP.
The upper shows the evolution trajectory with branching evolution lines. The numbers depict the passage number of the strain. All dots represent samples that were sequenced by NGS. The numbers in boxes depict the unit size of IS5-flanked tandem repeats derived from the original 70 kb tandem repeats. kb refers to kilobases. The lower part marks all changes in the genome along the evolution trajectory. Only changes greater than 25% variant frequency are shown. The genomic changes shown in SM1 are relative to its parental strain BW25113. The “+” & “-” signs represent mutations appeared or disappeared along the evolution trajectory, relative to the previous strain. The IS2-inserted mutS inactivation that occurred at passage 258 coincided with the increase of mutation events. The font colors in the graph are defined as follows: red for important events, green for SNPs, purple for INDELs, and gray for changes that were not retained along the evolutionary trajectory. Source data are provided as a Source Data file.
Having apparently solved the DPC issue, we cloned mdh onto ddp-BAC to form ddp-BAC::MRHTTP and introduced a recA-containing plasmid pKY29 at 165 passages (Fig. 4a and Supplementary Fig. 4a). The introduction of the active recA gene indeed increased the copy number of MRHTTP to 12 in 30 passages. Within 5 additional passages, the cell optimized the copy number on ddp-BAC to 2 and purged the recA-containing plasmid (Fig. 4a). Thus, it is evident that the cell used RecA and ddp-BAC to tune the copy number of MRHTTP that fits the specific genetic background. Since genomic changes continued to emerge along the evolution line while the copy number of ddp-BAC decreased (Figs. 4a, 5 and Supplementary Fig. 4a), it is possible that the ddp-BAC provided a favorable background for the cell to accumulate beneficial mutations while decreasing its copy number for better fitness.
Genomic changes during the evolution process
Next, we examine the genomic landscape during the evolution process. The parental strain SM1 contained 5 copies of tandem repeats each spanning 70 kb flanked by IS5 inserted next to yggE and yghO. Interestingly, the IS5-mediated tandem repeat unit shrank from 70 kb to 46 kb at passage 121 without changing the copy number (Figs. 4a,5 and Supplementary Fig. 4) during the presence of ddp-BAC::RHTTP and an active recA gene. One of the IS5 boundaries next to yghO gene was present in the ancestral BW25113 strain and the parental SM1 strain, and remained unchanged. The other IS5 boundary next to yggE was transposed from its original position to endA. Interestingly, the endA gene flanked the gshB gene which codes for glutathione synthetase that may be useful in formaldehyde detoxification. This tandem repeat fragment also contains mdh, tkt, tal, hps, phi, which were originally introduced during the development of SM1. Thus, the reduced tandem repeat region appeared to retain beneficial genes and eliminate redundant genes that cause metabolic burden. The shrinkage of the 70 kb tandem repeats was also observed in other branches of evolution, resulting in tandem repeat units with size ranging from 32 kb to 46 kb (Fig. 5 and Supplementary Fig. 4). Interestingly, these events all occurred after introducing the ddp-BAC::RHTTP system. It appears that the introduction of the ddp-BAC::RHTTP allowed the evolution of genome sequence by avoiding formaldehyde induced DPC.
Another key genomic change during the main evolution line was the emergence of the IS2 insertion into mutS which occurred at 258 passage (Figs. 4, 5, Supplementary Table 2). MutS is responsible for DNA mismatch repair, thus its inactivation caused an elevated mutation rate27. After this event, the number of SNP and INDEL rapidly increased, and the growth rate improved (Figs. 4, 5). We then isolated the fastest growing strain from 258 passage, which is SM7. Using the rifampicin assay, we found that the mutation rate of SM7 is ~80-fold higher than that of the BW25113 wild-type strain (Supplementary Fig. 5). Further evolution of SM7 continued to accumulate mutations, but the growth rate eventually decreased. To stabilize the strain, the wild-type mutS was restored in SM7 using the Lambda-Red system21, resulting in strain SM8, which is genetically stable with a mutation rate similar to the wild-type strain (Supplementary Fig. 5). After sub-culturing for 15 passages in methanol, the growth rate of SM8 remained around 3.5 h and no more mutations were observed (Figs. 4, 5 and Supplementary Table 2).
Omics analysis and genotypic characterization of SM7
To confirm that these evolved strains had significantly alleviated DPC issue, we compared the extracellular formaldehyde concentration of SM1 and SM7 at different growth stages (Fig. 6a). In contrast to SM1, the concentration of formaldehyde in SM7 cultures consistently remained low, approximately around 20 µM. This suggests a well-balanced flux between formaldehyde production and consumption, regardless of the growth stage. The reduced DPC complexes in different growth stages of SM7 were then confirmed by transmission electron microscopy (TEM) (Fig. 6b), which was able to image the DPC complex after extracting the DNA5.
a Time courses of SM1 and SM7 grown in 400 mM MeOH. Samples were collected at time points I to VI for SM1, and a to f for SM7 respectively, for extracellular formaldehyde measurement using LC-MS/MS. b Transmission electron microscopy (TEM) images of DPC products extracted from both SM1 and SM7 at different growth stages by negative staining. The experiment was repeated twice independently with similar results. c Enzymatic assays of Mdh and enzymes in the RuMP cycle in SM1 and SM7. Mdh activity was measured by the colorimetric Nash assay. The remaining enzymes were measured using coupled enzyme assays with a NADPH readout at 340 nm (n = 3, biological repeats; data are presented as mean values ± SD). Source data are provided as a Source Data file.
We also assayed the activities of enzymes related to formaldehyde production and consumption, including Mdh, hexulose-6-phosphate synthase (Hps), 6-phospho-3-hexuloisomerase (Phi), transketolase (Tkt), transaldolase (Tal) and ribose-5-phosphate isomerase (Rpi) (Fig. 6c). The results indicated that in SM7, Hps and Phi exhibited a significant increase as the strains entered the stationary phase. This increase in formaldehyde consuming enzyme expression in the stationary phase avoided the DPC problem. Apparently, the evolution process altered the expression profiles of these enzymes by tuning regulatory mechanisms. Such a regulatory change would not be possible without the initial introduction of ddp-BAC::RHTTP to alleviate the DPC problem. In contrast, the SM1 strain showed a significant decrease in Hps, Tal, Tkt, while Mdh increased during the stationary phase, causing the imbalance of formaldehyde production and consumption.
We also characterized the proteome of SM7. Strikingly, the ribosome abundance increased 50% compared to SM1 (Fig. 7a, b). The increase in ribosome abundance may be a key reason underlying the fast growth of SM7. It has been reported that ribosome amount is positively associated with the cell’s growth rate due to an increase in the overall translation rate in cells28,29. In addition, the amino acid pathways were more abundant in SM7 than in SM1 (Fig. 7b and Supplementary Fig. 6), which may also contribute to the fast growth of SM7. Comparing the proteomes of SM7 and the wild-type parental BW25113 grown in 1% glucose, we found that the RuMP cycle enzymes were increased as expected. However, the AceEF, PflB, GltA decreased while PoxB, Acs, and GlcB increased (Fig. 7c). These changes suggest that SM7 possibly derives 2-carbon compounds from pyruvate oxidase (coded by poxB) and acetate utilization (coded by acs, glcB), although further flux measurements are required to confirm.
a Abundance ratio of the 44 E. coli ribosomal subunit proteins in SM7, SM1 (grown in 400 mM methanol), and wild-type parental strain BW25113 (grown in 1% glucose). The boxplot is generated with their whiskers defining the maximum and minimum value, while the box encloses the first (25%) and third (75%) quartiles, with an additional line at the second quartile marking the median. Error bars represent SD. b Functional enrichment analysis identified biological processes by comparing SM7 to SM1 grown with 400 mM methanol until OD600 reached 1. The gene ratio on the x-axis represents the proportion of genes from a given gene set that are found to be differentially expressed or enriched in the analyzed dataset. A higher gene ratio indicates that a larger proportion of the gene set is involved in the indicated biological pathway. The adjusted P-value with Benjamini-Hochberg Procedure are shown with the color code (p. adjust). The data is generated by using GSEA with a permutation-based two-tailed test with 1000 permutations, with a FDR calculated P-value of 0.05 cutoff. c Metabolic protein abundance ratios. The left and right boxes represent SM7/WT and SM1/WT log2 ratios, respectively. The circles represent the log2 ratios of SM7/SM1 non-native proteins (Mdh, Hps, and Phi). Hps was cloned from either Methylomicrobium buryatense 5 GB1S (5G) or Bacillus methanolicus (BM). Abbreviations are listed in Supplementary Table 4. All data were done with n = 3, biological repeats. Source data are provided as a Source Data file.
The final SM7 strain contains a ddp-BAC::MRHTTP, with 2 copies. To probe the effect of this system on SM7 growth, we cured ddp-BAC::MRHTTP from SM7 (renamed SM7n) and found that the strain grew slower (Td = 7 h) (Fig. 8a). This result validated the contribution of ddp-BAC::MRHTTP in SM7.
a Complementation of mutated genes with BAC harboring a wild-type copy of the indicated genes in SM7n showed that gltA, ihfA, and rhlB mutations were important to the high growth rate of SM7. Growth rates of various strains with and without BAC complementation were measured in 400 mM methanol (n = 4 for SM7n strains with ihfA, ompR, cra, and gadW expression, n = 3 for SM7n strains with gltA, rhlB, gtrS, quuD, ylcG, gatY and rusA expression; All are biological repeats; data are presented as mean values ± SD; significance was tested by 2-side Welsh-test; *P < 0.05, **P < 0.01). b High-cell-density cultivation of SM8 (SM7 with WT mutS restored) in 500 mM methanol performed in a 2.0 L bioreactor or a 50 mL Falcon tube. The strain was cultivated in a fed-batch mode with continuous supply of methanol and NH4OH. The M9 and MOPS media have been modified with an adjustment of an initial NH4Cl concentration of 25 mM. SM8 reached 20 g DCW L−1 within 77 h in the bioreactor with M9-methanol medium. The orange line represents SM8 grown in the bioreactor with M9-methanol medium, the blue line represents growth in the bioreactor with MOPS-methanol medium, and the red line represents growth in the Falcon tube with MOPS-methanol medium. c Growth of SM8 in a logarithmic scale. The doubling time of SM8 grown in a 50 mL falcon tube with modified-MOPS minimal medium containing 400 mM methanol reached 3.6 h. d The profile of fermentation products of SM8 in the bioreactor with the modified M9 methanol medium. (n = 3, sampling repeats; data are presented as mesan values ± SD) (e) Growth of SM7 in Falcon tubes containing MOPS minimal medium with 400 mM methanol (red line), 1% glucose (dark yellow line), 0.1% glucose (blue line), 400 mM methanol plus 0.1% glucose (purple line). All data were done with n = 3, biological repeats. Source data are provided as a Source Data file.
There are several mutations on SM7 (Fig. 5 and Supplementary Table 2). We tested the effect of a few, including gene (gatY, gtrS, quuD, rusA, ylcG) and promoter interruption by insertion sequences (gltA), codon change in protein coding regions (cra, gadW, ihfA, ompR) and frameshift (rhlB). To probe the effect of these mutations on growth, we complemented these disrupted genes with a copy of the intact gene on BAC in SM7n and tested their effects on growth rate. Results (Fig. 8a) showed that the complementation by BAC-borne gltA, rhlB, and ihfA genes significantly reduced the growth rate of SM7n. This result indicates that the gltA insertion by IS2 in the promoter region and rhlB frameshift were beneficial for growth. GltA is the first enzyme in the TCA cycle, which produces NADH. As methanol is a more reduced substrate than glucose, a strong TCA cycle would increase the burden of respiration. In the SM7 strain the gltA promoter was interrupted by IS2, but it did not completely abolish the gltA expression. Proteome data suggested that GltA protein was still produced, albeit at about 10% level compared to the wild-type strain (Fig. 7c). Thus, it is possible that by properly knocking down gltA, the cell would not generate too much NADH in the electron-rich methanol medium. The same promoter mutation actually existed in the SM1 strain, however, re-introducing a wild-type gltA copy did not affect SM1 growth significantly5.
The rhlB gene encodes an ATP-dependent RNA helicase. In SM7, this gene contains a frameshift mutation and abolishes its function. It has been reported that overexpression of rhlB can buffer deleterious mutations to increase fitness in E. coli mutator strains30. The inactivation of rhlB probably increased the effect of mutS mutation and accelerated evolution. However, its exact role in SM7 requires further investigation.
We also tested whether the mutations in global regulatory genes—cra (E190Q), ompR (R17C), ihfA (V78E), and gadW (frameshift)—played a critical role in the enhanced growth rate of SM7 (Fig. 5 and Supplementary Table 2). The complementation of cra (E190Q), ompR (R17C), and gadW (frameshift) by their respective wild-type copy using BAC did not show any difference in growth, suggesting that these mutations did not contribute to the fast-growth phenotype. However, the complementation of ihfA (V78E) by its wild-type version using BAC significantly decreased the growth rate of SM7n. Thus, this mutation appears to be beneficial for the growth phenotype. The ihfA gene codes for a global transcriptional regulator affecting the expression of hundreds of genes, including those involved in amino acid and fatty acid metabolism, making its role difficult to elucidate at this point31.
Phenotypic characterization of SM7 and SM8
As the strain SM7 acquired the ability to grow rapidly in methanol minimal medium compared to SM1 (Figs. 3d and 6a), we then characterized the performance of the stabilized strain SM8. By manually supplying methanol and ammonia to maintain concentrations above 300 mM and 20 mM, respectively, SM8 can grow to an OD600 of 7 with a doubling time of 3.6 h in a 50 mL Falcon tube with an air-permeable membrane (Fig. 8b, c). In a bioreactor using two kinds of methanol minimal medium: MOPS-based and M9-based, SM8 can reach maximal dry cell weight of 19.9 and 12.2 g DCW L−1, respectively (Fig. 8b, c). The growth rate of SM8 in bioreactors reaches a doubling time of 6 h in modified-MOPS medium and 9 h in modified-M9 medium, respectively. Further optimization of cultivation conditions in the bioreactor to achieve faster growth rates remains to be explored. During fermentation, SM8 produced acetate up to 3 g L−1, which was consumed in the later phase (Fig. 8d). The consumption of acetate was consistent with our proteomic data, which showed that acetate consuming enzymes Acs and GlcB were increased (Fig. 7c). Formate, ethanol, lactate, and succinate were also detected at levels less than 1 g L−1 (Fig. 8d).
Interestingly, SM7 grows faster in methanol than in 1% glucose minimal medium, a concentration that is sufficient to support fast growth of wild-type E. coli in the MOPS medium (Fig. 8e). However, SM7 did not exhibit diauxic growth in a medium containing both 0.1% glucose and 400 mM methanol (Fig. 8e and Supplementary Fig. 7). The lack of diauxic growth control between glucose and methanol was expected, as such a mechanism must be evolved or designed specifically. In the presence of glucose in methanol medium, SM7 consumes glucose and methanol simultaneously for growth demonstrating a better growth rate with a doubling time of 2.9 h (Fig. 8e). This result suggest that the strain can be used in a mixed substrate fermentation for optimal performance.
Discussion
Methanol growth using the RuMP pathway requires only three enzymes (Mdh, Hps, and Phi) that are not present in the wild-type E. coli. It may deceivingly appear quite simple to convert E. coli to a synthetic methylotroph by introducing these three enzymes. However, the challenge was more significant than expected. Many authors reported methanol-dependent growth of E. coli32,33,34, which can assimilate methanol but still requires other carbon sources. The next step was the development of methanol auxotrophy that requires methanol to grow, even though other carbon sources were also needed4,6. A complete synthetic methylotroph took much longer to develop5,7,8.
We previously identified that the methanol metabolic intermediate, formaldehyde, could accumulate during the growth or stationary phases5. Formaldehyde is a potent chemical that causes DPC and kills the cell. Thus, cells with severe DPC are not evolvable. We solved this problem and evolved a first-generation synthetic methylotroph (SM1) that can grow with methanol as the sole carbon source5. However, its growth rate is still far below that of native methylotrophs.
We hypothesize that the residual DPC in SM1 is responsible for the slow growth. To solve the DPC problem requires balancing formaldehyde production from methanol and formaldehyde consumption through the RuMP cycle. The production of formaldehyde that causes DPC renders the cells difficult to evolve, if not impossible, because most of the cells were killed in the stationary phase. This phenomenon was demonstrated in Fig. 3a, b, which showed that SM1 still produced measurable amounts of formaldehyde and did not improve the growth rate much even after about 80 passages. It appeared that the formaldehyde-induced DPC still remains as a bottleneck in evolution.
In this work, we used the ddp-BAC system that exhibits a dynamic CNV property to accomplish the gain-of-activity function, which is difficult to emerge through evolution. The presence of ddp sequence causes the whole BAC plasmid to generate tandem repeat concatemers with a wide range of copy numbers through a recA-dependent mechanism (Figs. 1, 2). Such a system provides an expedient fix to bypass a bottleneck in evolution while the cell seeks for a better solution. In parallel to our effect in understanding the mechanistic details, we employed this system to solve the DPC problem in synthetic methylotrophic E. coli. The dynamic CNV is reminiscent of the copy number engineering technique35 but with a different mechanistic basis.
Since the descendent of SM1 after 81 passages (SM1p) did not improve growth much and still produced formaldehyde in high methanol concentrations (Fig. 3a, b), we over-expressed the formaldehyde consuming enzymes RHTTP using ddp-BAC, BAC, or pUC19. The ddp-BAC produces tandem repeats concatemers with a wide range of copy numbers, which provides a library of various RHTTP copy numbers for competition during evolution. Without ddp, the BAC system exists in a single copy in the cell. In contrast, plasmid pUC19 consistently exists in high copy (~500). We found that only ddp-BAC systems expressing RHTTP can cause the SM1p to grow in high methanol concentrations (1000 mM) consistently (Fig. 3c). This result suggests that the SM1p was able to find the acceptable copy numbers of RHTTP in the ddp-BAC library for growth in such a condition, and removed the bottleneck in evolution. The metabolic burden introduced by the concatemers of ddp-BAC was tentatively tolerated.
During this period, the cells was able to grow and evolve. Beneficial mutations that improved growth or reduced formaldehyde production using more effective ways occurred. For example, the 70 kb IS5-mediated tandem repeat region that emerged during the evolution of the parental strain SM1 shrank to 46 kb, which greatly reduced the protein burden and possibly allowed more balanced expression of the formaldehyde consuming enzymes. Other mutations that occurred during this time may directly or indirectly contribute to the balanced consumption of formaldehyde. After 50 passages of evolution, the ddp-BAC copy number was reduced by inactivating recA to reduce the metabolic burden. Presumably the cells had found a better way to balance the formaldehyde metabolism.
After other mutations accumulates, the copy number returned to a low level, demonstrating the dynamic tuning of copy numbers to fit the particular genetic background. This phenomenon was demonstrated again at passages 165 by reinstalling recA using a multicopy plasmid. At this genetic background, the cells evolved to find the optimal copy number of MRHTTP on ddp-BAC to be 2, and purged the recA plasmid at passage 200 (Fig. 4). When the 2-copy ddp-BAC was removed from SM7, the strain grew slower, indicating the contribution of the ddp-BAC system.
The formaldehyde-consuming enzymes Hps and Phi on SM7 exhibited a higher expression level in the late growth (or stationary) phase compared to the early growth phase (Fig. 6c). This appears to be the most direct reason for much reduced formaldehyde accumulation and diminished DPC problem. Such a change was through fine tuning of regulatory systems that became possible after introduction of RHTTP on a system that allows dynamic copy number tuning. Thus, the dynamic copy number tuning can be used as a tentative mechanism for expression tuning before a proper regulatory mechanism can be evolved or designed.
The reduction in the length of the IS5 mediated tandem repeats (Fig. 5 and Supplementary Fig. 4b) may be another possible reason for enhanced growth. During the evolution of the parental SM1 strain5, a 70 kb IS5-mediated tandem repeats appeared. These repeats span a region that include several RuMP pathway enzymes and gshB. It was thought to be beneficial for the growth of SM1. However, because of the large size, the presence of such tandem repeats increased metabolic burden. After we introduced ddp-BAC::RHTTP to de-bottleneck evolution, the length of these IS5-mediated tandem repeats reduced to 46 kb (Figs. 4, 5), allowing the cells to reduce metabolic burden and redistribute the proteome profile to ribosomes and amino acid pathways that may contribute to the fast growth (Fig. 7b).
Additionally, we observed that rhlB (frameshift), ihfA(V78E) and gltA promoter insertion by IS2 had a significant beneficial effect on SM7. Previous studies have reported that rhlB and ihfA genes modulate the proteome of bacterial cells. Specifically, rhlB deletion modulates ribosome biogenesis36, and ihfA deletion elevates amino acid biosynthesis31. These genotypic changes align with our observations on proteomic changes. Interestingly, the rhlB mutation also appeared during the evolution of recently published synthetic methylotrophic E. coli with a high growth rate13. The disruption of mutS is another important mutation that facilitates the evolution of SM6 towards SM7 by conferring hypermutated phenotype. A similar hypermutation phenotype also played a significant role in the evolution of the synthetic strain from refs. 8,13 due to a dnaQ mutation.
From the proteomic analysis, we found that AceEF, PflB, GltA decreased while PoxB, Acs and GlcB increased. GltA expression decrease was caused by an IS insertion in its promoter region that reduced its expression. The PoxB route is actually less energy efficient than the AceEF route, as it produces acetate as a carbon drain, which requires more ATP for reuptake. As methanol is richer in electron (and energy) than glucose, E. coli which typically grows in sugar-abundant environment has to find a way to deal with surplus electrons and energy. Thus, the PoxB route and the reduced TCA cycle was evolved to dump excess energy to acetate. Fermentation in bioreactor also demonstrated that the cells produced acetate and consumed it later. The glyoxylate shunt enzymes were essential for acetate consumption, and their expression levels were increased for acetate reuptake.
The final stabilized SM8 strain grows with a doubling time of around 3.5 h under optimal conditions, and exhibits no lag phase in re-growth. This growth rate is faster than that of native model methylotrophs, including Methylorubrum extorquens AM1 (Td ~ 4 h) and Bacillus methanolicus at 37 °C (Td ~ 5 h), as well as the synthetic methylotroph (Td ~ 4.3 h) reported by Reiter et al. 13. The maximal biomass that SM8 achieved is 19.9 g/L, which is comparable to the previously reported strain13. The SM8 strain contains about 50 mutations compared to its parental strain SM1 which already had 21 mutations. The strain reported by Reiter et al. 8,13 contains more than 1000 mutations, which may be due to a dnaQ mutation that caused hypermutation. We observed that hypermutation led to a decrease in the growth rate of SM7 after successive passages. By restoring the wild-type mutS gene in SM7 to create SM8, SM8 maintained a doubling time of around 3.5 h.
Overall, our dynamic copy number tuning approach demonstrates significant promise in the evolutionary engineering of C1-utilizing microorganisms. The resulting methylotrophic E. coli strains, SM7 and SM8, potentially provide a foundation for exploiting methanol-based bioconversion.
Methods
Plasmid construction
KODone polymerase (TOYOBO) was utilized for all PCR reactions. The NEBuilder 2x HiFi DNA Assembly Master Mix (NEB) was employed for Gibson assembly to construct plasmids. Subsequently, the plasmids were introduced into DH5α cells (NEB) cultivated in lysogeny broth (LB medium) or Terrific Broth (TB medium) supplemented with appropriate selective antibiotics. The plasmid sequences were verified by Sanger sequencing or nanopore sequencing. All the plasmids used in this study are listed in Supplementary Table 3, and their sequences are provided in Supplementary Data 1. The primers used for cloning and sequencing in this study are provided in Supplementary Data 2.
Strain construction
All plasmids transformed into the host cells were carried out by electroporation. The electroporation method was performed as follows: 1 mL of bacterial culture was washed twice with 1 mL of ice-cold 10% glycerol. The bacterial pellet was then resuspended in 100 μL of 10% glycerol and mixed with 50−100 ng of plasmid DNA for electroporation at 1.8 kV. Finally, the cells after electroporation were recovered for 30 min and plated on LB agar plates containing the appropriate antibiotics.
By adding additional copies of mdh into the synthetic methylotroph, we used Tn7-like transposons-mediated CRISPR system from ref. 37. The mdh was targeted to be inserted into ddpA gene where the guide RNA was designed using the tool (INTEGRATE guide RNA tool). In brief, the host strain is electroporated with the plasmid harboring the Tn-7 CRISPR system, cultivated at 30 °C for 1–2 days. After colony formation, colony PCR is conducted for confirmation of cassette insertion.
To repair the mutS mutation in SM7, we used a Lambda-Red recombineering system to introduce wild-type mutS along with a gentamycin resistance gene. A native MutS expression cassette with a gentamycin resistance marker was amplified by PCR, and 100 ng of the purified PCR products were transformed into SM7 following anhydrotetracycline-induced activation of the Lambda-Red system. Strains with successful mutS integration were picked from LB plates with gentamycin.
All the strains used in this study are listed in Supplementary Table 1 and Supplementary Table 2. The primers and gRNA used in this study are provided in Supplementary Data 2.
Media for methylotrophic growth and glucose growth of SM strains
MOPS-based methanol minimal medium was composed of 400–1200 mM methanol, 50 mg/mL chloramphenicol, 1 mM IPTG, and MOPS EZ buffer (Teknova; 40 mM MOPS, 3.02 nM CoCl2, 0.962 nM CuSO4, 50 mM NaCl, 9.5 mM NH4Cl, 0.525 mM MgCl2, 4 mM tricine, 1.32 mM K2HPO4, 0.276 mM K2SO4, 0.01 mM FeSO4, 0.5 mM CaCl2, 40 nM H3BO3, 8.08 nM MnCl2, 0.974 nM ZnSO4, and 0.292 nM (NH4)2MoO4.). Chloramphenicol was dissolved in pure methanol. A trace vitamin mix is prepared where the following final concentration is reached in the medium: 8.19 μM biotin, 10.49 μM calcium pantothenate, 4.53 μM folic acid, 40.94 μM nicotinamide, 13.29 μM riboflavin, 14.82 μM thiamine hydrochloride, and 0.07 μM vitamin B12.
M9-based methanol minimal medium is composed of M9 salts (3 g/L KH2PO4, 0.5 g/L NaCl, 6.8 g/L Na2HPO4 and 1.0 g/L NH4Cl), 400–500 mM methanol, 1 mM MgSO4, 0.1 mM CaCl2, 50 mg/mL chloramphenicol, 1 mM IPTG, trace elements and a trace vitamin mix. The final concentrations of trace elements were as follows: 13.4 mM Na2EDTA, 13.1 mM FeCl3-6H2O, 0.62 mM ZnCl2, 76 μM CuCl2-2H2O, 62 μM CoCl2-2H2O, 162 μM H3BO3 and 8.1 μM MnCl2-4H2O. The trace vitamin mix is prepared as follows: 8.19 μM biotin, 10.49 μM calcium pantothenate, 4.53 μM folic acid, 40.94 μM nicotinamide, 13.29 μM riboflavin, 14.82 μM thiamine hydrochloride, and 0.07 μM vitamin B12.
For high-cell-density cultivation experiments, the NH4Cl concentration for both media is adjusted to 25 mM. In the main text and figure legends, these media are named modified M9/ MOPS medium for better clarity.
The MOPS-based glucose minimal medium was composed of MOPS EZ buffer, 50 mg/mL chloramphenicol, trace vitamin mix, 1 mM IPTG, and either 1% or 0.1% glucose.
High-cell-density cultivation of SM8
All pre-culture strains were grown to an OD600 = 3 in 200 mL MOPS-based methanol medium in shake flasks at 250 rpm and 37 °C. The pre-cultured strain was then inoculated in a 2 L bioreactor (Sartorius) containing 0.8 L of modified-M9 500 mM methanol medium and modified-MOPS 500 mM methanol medium to an initial OD600 of 0.6 and 0.2, respectively. The condition for fed-batch cultivation was 1200 rpm stir-rate, 3 vvm air flow, and pH = 7. pO2 was set to 100% saturation by elevating gas flow of pure O2. The pH was maintained at 7.0 throughout growth by using 3 M H2SO4 and 2.94 M NH3 for adjustment. During the fed-batch cultivation, the methanol concentration was maintained above 300 mM by continuously feeding methanol. And the ammonium concentration was maintained above 20 mM by constantly feeding ammonia. Trace elements mix (13.4 mM Na2EDTA, 13.1 mM FeCl3-6H2O, 0.62 mM ZnCl2, 76 μM CuCl2-2H2O, 62 μM CoCl2-2H2O, 162 μM H3BO3, and 8.1 μM MnCl2-4H2O) were added once when OD600 of SM8 exceeded 7 in both modified MOPS and modified M9 medium. Additionally, 10 mM of K2HPO4 was added once when the OD600 exceeded 7 in the modified MOPS medium. To monitor the concentrations of methanol and ammonium, culture samples from bioreactors were collected throughout the growth. A 0.5 mL aliquot of collected samples were centrifuged for 5 min at 15,000 × g and supernatants were collected for methanol and ammonium detection. Methanol was detected by GC-FID, and the ammonium was measured using a YSI2950 Biochemistry Analyzer. OD600 was measured using by GENESYS™ 30 Visible Spectrophotometer with appropriate dilution of the samples.
The fed-batch cultivation of SM8 was performed in a 50 mL Falcon tube as well. SM8 was inoculated in modified-MOPS 400 mM methanol medium at an initial OD600 of around 0.1. The temperature for incubation was 37 °C, and spinning rate for a tube was 2700 rpm. Pure methanol and 2 M ammonium were supplied manually to maintain their concentrations above 300 mM and 20 mM, respectively.
Digital droplet PCR (ddPCR) to analyze copy number
Genomic DNA from E. coli was extracted either manually using the Puregene Kit from Qiagen or automatically purified using the magnetic bead system from Taco™ mini (GeneReach company, TW). The concentration of genomic DNA (gDNA) was determined using the BioTek Synergy H1 Plate Reader (Agilent Technologies) by measuring absorbance at 260 nm, 280 nm, and 230 nm. The concentration and purity of gDNA were calculated using the ratios of A260/A280 and A260/A230. 500 ng of gDNA was fragmented by incubating with the restriction enzymes PstI/HindIII at 37 °C for 15 min in a total volume of 50 μL. The concentration of DNA fragments was diluted to 25 pg/μL, and quantification was performed using the Qubit™ 1X dsDNA High Sensitivity assay kit (Thermo Fisher Scientific).
A ddPCR solution was prepared containing ddPCR master mix, probes, and primers for the target gene labeled with FAM, reference gene 1 labeled with HEX, and reference gene 2 labeled with Cy5. ddPCR was performed using the Naica® system for Crystal Digital PCR™ (Stilla Technologies). 10–25 pg of DNA was mixed with the ddPCR solution and injected into either a Sapphire chip for 12 samples or a Ruby chip for 48 samples. Partitioning and PCR were carried out using the Geode device, followed by chip reading using the Prism 6 device. Data analysis was conducted using the Crystal Miner data analysis software. Copy number was calculated by dividing the absolute concentration of the target gene by the absolute concentration of the reference gene.
Next-generation sequencing and Nanopore sequencing
DNA samples for sequencing are extracted either using the Nanobind CBB kit (PacBio, USA), the Puregene kit (Qiagen, Germany), or automatically extracted using the QIACUBE DNeasy (Qiagen, Germany). For NGS, 200 ng of DNA is used for library preparation with the Illumina DNA Prep Kit (Illumina, USA). The quality of DNA libraries is assessed using the 5200 Fragment Analyzer and sequenced on the Illumina MiniSeq or NextSeq with a read length of 150 bp paired-end. For normal reads (N50 about 7 kb) nanopore sequencing, 400 ng of DNA is used for library preparation with the Rapid Barcoding Kit 96 (Nanopore, UK). For long reads (N50 about 15 kb) nanopore sequencing, 1 µg of DNA is used for library preparation with the Native Barcoding Kit 24. Sequencing is performed using a MinION Flow Cell (R9.4.1 or R10.4.1) and MinION MK1B, with real-time basecalling carried out using MinKNOW software on a DELL workstation computer equipped with a high-performance GPU RTX 3090. All SNP calling and variant detection, annotations were performed by Geneious Prime software (2022.0.1).
Microscopy and live cell imaging
The real-time imaging of E. coli was conducted using an IX83 Inverted Microscope (IX83, Olympus, Tokyo, Japan) equipped with an Olympus UPlanXApo 60x/1.42 Oil Ph3, Infinity/0.17 Microscope Objective, and an ORCA-Flash 4.0 V3 Digital CMOS camera (Hamamatsu Photonics). Following the cultivation of E. coli to OD600 to 0.1, a 10 μL aliquot was applied onto a µ-Dish 35 mm, high Glass Bottom (Cat. No: 81158, ibidi) loaded with a 1% (w/v) agarose base, and subsequently positioned within the microscope setup for imaging at 37 °C for 4 h. Cell image processing was conducted using Fiji software 1.54 g.
Flow cytometry analysis
Flow cytometry analysis was performed using a CytoFLEX S flow cytometer (Beckman), equipped with a 488 nm blue laser emitting at 51 mW, which facilitated the detection of forward scatter (FSC), side scatter (SSC), and fluorescence signals. All the samples were taken from bacterial cultures grown in LB overnight, and then were appropriately diluted in autoclaved 1x phosphate-buffered saline (PBS) and adjusted to a concentration of approximately 1.5 × 106 cells/mL. To minimize background noise, SSC thresholds for all collected samples were set at a value of 1000. Data acquisition was conducted at a rate exceeding 1000 events per second (approximately 30 µL/s), with an abort rate of less than 10%. Fluorescence distribution histograms were generated using FCS Express 7 software (De Novo Software).
Growth assay and laboratory evolution of synthetic methylotrophic strain
Methylotrophic strains were evolved in 3 mL tubes at 37 °C and 250 rpm using a New Brunswick Scientific Innova 44 shaker. These strains were then transferred to fresh medium at an initial OD600 ranging from 0.05 to 0.2 once they reached the stationary phase. OD600 values were measured using a GENESYS™ 30 Visible Spectrophotometer. We also cultivated the strain in a volume of 20 mL in a 50 mL Falcon tube for growth rate measurement. The strain was incubated in an RTS-8 Plus system (Biosan) with a spinning rate of 2700 rpm for the Falcon tubes, which automatically measured OD600. We also took samples and diluted them appropriately for accurate OD600 measurement using a GENESYS™ 30 Visible Spectrophotometer.
The detection of formaldehyde in culture medium
50 µL of bacterial culture was filtered through an Amicon Ultra-0.5 centrifugal filter (50 kDa, Merck), and then incubated with 50 µL of 0.1% 2,4-Dinitrophenylhydrazine (DNPH) and 2 µL of 85.4% H3PO4 at room temperature for 10 min. After incubation, 2 μL of the sample was injected into the HPLC system for separation by a Triart C18 ExRS column (1.9 μm, 2.1 × 100 mm, YMC) with a constant flow rate of 0.3 mL/min at 50 °C. Eluent A consisted of 50% acetonitrile in water (50 v/v) with 0.4 mM ammonium fluoride, while eluent B contained 100% methanol with 0.4 mM ammonium fluoride. The HPLC program was set as follows: 100% A (0.0–3.5 min), 0 to 98% B (3.5–5.5 min), 98% B (5.5–10.0 min), 98 to 0% B (10.0–10.5 min). A heated electrospray ionization (ESI) probe was equipped for ionization and operated in negative ion mode. The data were processed and analyzed using LabSolutions software (Shimadzu).
Transmission electron microscopy analysis to characterize DPC
The protocol was modified from our previous work5. Briefly, 2 mL cultures of E. coli strains SM1 and SM7 M400 were harvested by centrifugation at 5000 g for 5 min. Then the pellet was resuspended in 100 μL of 10 mM Tris-HCl (pH 7.5) containing 2.5 mg/mL lysozyme and incubated at 37 °C for 30 min to lyse the cells. To extract DNA-protein complex, 500 μL DNAzol reagent was added to the lysate, mixed for 5 min, and centrifuged at 12,000 × g for 10 min. The supernatant was transferred to a new tube, and 300 μL of ice-cold 100% ethanol was used to precipitate the DNA-protein complexes. After freezing at −30 °C for 20 min, the precipitate was washed twice with cold 80% ethanol. The pellet was dissolved in 190 μL of 8 mM NaOH and 10 μL of 1 M Tris-HCl (pH 7.5) and incubated at 37 °C for 30 min. To denature proteins and dissociate them from DNA, urea and SDS were added (final concentrations: 8 M urea and 2% w/v SDS) and gently shaken at 37 °C for 30 min. Proteins were salted out by adding an equal volume of 5 M NaCl, followed by shaking at 37 °C for 30 min. After centrifugation at 12,000 g for 20 min, the supernatant was filtered through an Amicon Ultra-4 filter with a 3 kDa cutoff and washed three times with 10 mM Tris-HCl (pH 7.4). The volume was concentrated to 450 μL, then 50 μL of 3 M potassium acetate and 1 mL of ice-cold 100% ethanol were added. Samples were incubated at −30 °C for 1 h. The resulting pellet was washed twice with cold 80% ethanol and dissolved in 100 μL of 10 mM Tris-HCl (pH 7.5). DNA concentration was measured using a Qubit assay. Purified DPC complexes, ~500 ng of DNA, were mounted onto activated 300-mesh copper grids coated with carbon-stabilized formvar (Ted Pella). This mounting process was operated at room temperature for 1 min. Following the removal of excess liquid with filter paper blotting, the samples were stained with a 2.5% uranyl acetate solution for 1 min and air-dried naturally at room temperature. TEM images were captured using a Talos L120C. The magnification ratio was set to range from 2700x to 15,000x for optimal visualization.
Enzyme assay
The protocol for enzyme assay was modified from previous studies3,4. Specifically, enzymes from the RuMP pathway (including Hps, Phi, Tkt, Tal, Rpe, Rpi) were purified from a lac-induced His-tagged overexpression system in E. coli. The cells were inoculated in 5 ml LB and then grown over night at 37 °C. On the following day, the overnight culture was inoculated into a 100 mL LB at 1% and then grown at 37 °C. 0.1 mM IPTG was then added when the culture reached OD600 0.4–0.6 at log phase. The culture was then grown at 20 °C 250 rpm overnight, and then collected by centrifugation. The cells were then resuspended with 10 mM phosphate buffer solution, pH 7.4, and broken by a Sonicator (Q600) on ice. After 1 h of 21,130 g centrifugation, the protein solution was then filtered by a 0.22 μM filter, and then purified by a Biorad NGC using a IMAC column (Roche). The crude extracts are prepared as followed. 2 mL of normalized OD600 1 culture were harvested from SM7 and SM1 when grown at OD600 1 and OD600 3 respectively, where dilution is done by pure PBS. The cultures are then centrifuged at 4602 × g, where supernatant is discarded, and the pellets are resuspended in 10 mM Phosphate Buffer, pH 7.4 at 700 µL. The mixture was then sonicated, and the supernatants (crude extracts) were collected by 1 h of centrifugation at 21130 g.
The enzyme assay readouts are all done by a BioTek Synergy H1 Plate Reader (Agilent Technologies). All samples are normalized to a concentration of 1 mg/mL by using Bradford assay. The Pgi and Zwf enzymes were obtained commercially from Roche. For Hps/ Phi Assay, the following enzymes were utilized: Hps, Phi, Rpi, Pgi, Zwf with a NADPH readout with 100 mM Phosphate buffer (pH7.4), 10 mM formaldehyde, 4 mM ribose 5-phosphate (R5P), 1 mM thiamine pyrophosphate, 2 mM MgCl2. For Tkt/Tal/Rpi Assay, the following enzymes were utilized: Tkt, Tal, Rpe, Rpi, Pgi, Zwf with a NADPH readout with 100 mM Phosphate buffer (pH7.4), 4 mM R5P, 1 mM thiamine pyrophosphate, 2 mM MgCl2. Mdh is tested by a Nash reaction assay, similar to our previous report26. In brief, the Nash reagent was prepared by dissolving methanol, ammonium acetate, and acetylacetone in H2O to final concentrations of 400 mM, 200 mM, 20 mM respectively.
The cell cultures were harvested and centrifuged with 5000 × g, 3 min with the supernatant removed, and replaced by the Nash solution. The Nash reaction is then dynamically measured by absorbance of 440 nm with shaking of 750 rpm at 37 °C, where the activity comparison is based on the Abs400nm, normalized by the OD600 at time zero (t0).
Proteomics analysis
Proteins were prepared similar to enzyme assay, where SM7, SM1, and WT are grown in methanol 400 mM, methanol 400 mM, and 10 mM glucose respectively until OD600 1. Each sample was collected with three biological repeats. The strain pellet collected by centrifugation was sonicated and then normalized to 1 mg/mL. Proteins were suspended in 8 M urea/50 mM HEPES (pH 8), disulfide bonds were reduced, free thiols were alkylated with iodoacetamide. Afterward, samples were digested overnight at 37 °C using LysC protease and trypsin at an enzyme-to-substrate ratio of 1:50 (w/w), and peptides were desalted by using C18 Spin Tips. Eluates were dried by vacuum centrifugation. Desalted peptides were reconstituted in 100 mM HEPES, and TMT18-plex reagents (Thermo Fisher) was added from stocks dissolved in 100 % anhydrous ACN. The peptide-TMT mixture was incubated for 1 h at room temperature with frequently mix, and the labeling reaction was stopped by addition of 5% hydroxylamine to a final concentration of 0.4% and incubation for 15 min at room temperature. Labeled peptide mixtures were pooled into 18-plexed sample, which was fractionated into 4 fractions by using high pH reversed-phase peptide fractionation kit (Pierce).
Each fraction was injected for nano-liquid chromatography-nano electrospray ionization-tandem mass spectrometry (nanoLC-nanoESI-MS/MS) analysis. NanoLC-nanoESI-MS/MS analysis was performed on a Thermo UltiMate 3000 RSLCnano system connected to a Thermo Scientific Orbitrap Fusion Tribrid Mass Spectrometer (Thermo Fisher Scientific, Bremen, Germany). Peptide mixtures were loaded onto a 75 μm ID, 25 cm length PepMap C18 column (Thermo Fisher Scientific) packed with 2 μm particles with a pore with of 100 Å and were separated using a segmented gradient in 60 min from 5% to 25% (for 55 min), 25% to 35% (for 5 min) solvent B (acetonitrile with 0.1% formic acid) at a flow rate of 300 nL/min. Solvent A was 0.1% formic acid in water. For SPS-MS3 analysis, the survey scans of peptide precursors from 400 to 1600 m/z were performed at 120 K resolution with a 2 × 105 ion count target. Peptide precursors with charge state 2–6 were sampled for MS2. Tandem MS was performed by isolation window at 0.7 Da with the quadrupole, CID fragmentation with normalized collision energy of 30, and MS2 spectra were acquired in the ion trap using an AGC target value of 1 × 104 and the max injection time was 35 ms. Afterwards, MS3 spectra were acquired using synchronous precursor selection (SPS) of 10 isolation notches. MS3 precursors were fragmented by HCD with 55% collision energy and analyzed using the Orbitrap with 50 K resolution power at 100–500 m/z scan range, a two m/z isolation window, 1 × 105 AGC, and a maximum injection time of 200 ms with one microscan.
Raw data files were processed using Proteome Discoverer v3.0 (Thermo Fisher Scientific) and the tandem MS data were then searched using SEQUEST HT with the INFERYS Rescoring algorithm against the protein sequence of Mdh, Hps (Bm), Hps (5 G), Phi (Mf), Tal (Kp), Tkt (Metbu), GapC (BL21), Cat, SopA, SopB, RepE and an E. coli-K12 fasta (4518 sequence entries, dated: May 2021) from UniProt database with common contaminant proteins. The search parameters included trypsin as the protease with maximum of two missed cleavages allowed; oxidation of methionine (+15.9949 Da) was set as a dynamic modification while static modifications included carbamidomethyl (+57.0215 Da) at cysteine and TMT as a static modification of lysine residues and peptide N-termini (+304.20175 Da). Precursor mass tolerance was set at 10 ppm and fragment mass tolerance was set at 0.6 Da. Peptide confidence was estimated with the Percolator node. Peptides were filtered at q < 0.01 based on a decoy database search. Reporter ions for TMT labeled peptides were quantified using the Reporter Ions Quantifier Node included a TMT18plex quantification method in Proteome Discoverer with a peak integration tolerance of 20 ppm and an integration method based on the most confident centroid peak at the MS3 level. Peptides with an average reporter signal-to-noise ratio >10 were used for protein quantification. Correction for the isotopic impurity of reporter quantification values was applied.
The abundance data were then processed by omitting all NA value entries and performing imputation by Bayesian PCA missing value estimator from pcaMethods. Data were then log2 transformed, and omitted any data with NA values, equal median centered, and then compared by using DeqMS. The results are then analyzed by Clusterprofiler for enrichment analysis by using GSEA and KEGG databases.
The measurement of fermentation products and methanol consumption
Ethanol, formate, methanol, and acetate were analyzed using an Agilent 7890B Gas chromatography with a 5977 mass spectrophotometer GC-MS with a SH-I-5MS column (Shimadzu). Helium is used as the gas carrier with constant pressure at 7.0633 psi. The thermal cycle is as follows: initial temperature: 40 °C, 2 min, ramp rate 10 °C/min to 60 °C, 40 °C/min to 220 °C, 2 min final hold. The data were processed and analyzed by Enhanced ChemStation F.01.03.2357 (Agilent).
Lactate and succinate were measured by triple quadrupole Shimadzu LC–MS 8045 equipped with YMC-Triart C18 ExRS. The injection volume of the samples was 2 μL. The oven temperate for the column was set to 50 °C. The separation was performed using a buffer system composed of 100 mM tributylamine and 150 mM acetic acid with 0.4 mM NH4F (solvent A) and 100% methanol with 0.4 mM NH4F (solvent B). The Initial buffer composition for gradient separation was 95% solvent A and 5% solvent B. Solvent B was then increased linearly to 25% from 0 to 4 min. Next, solvent B was increased to 98% linearly from 4 to 10 min, and was maintained at 98% for 10.2 min. At 20.2 min, solvent B was decreased to its initial 5% within 0.5 min and maintain constant 4.8 min. MS/MS was operated in negative mode with the following acquisition parameters: interface temperature of 250 °C, desolvation line temperature of 200 °C, heat block temperature of 400 °C, nitrogen as the desolvation gas, a nebulizing gas flow rate of 3.0 L/min, and a drying gas flow rate of 5 L/min38. The data were processed and analyzed by LabSolutions Version 5.93 software (Shimadzu).
Methanol consumption was measured using an Agilent 7890B Gas Chromatograph equipped with a Flame Ionization Detector (FID) and a DB-FFAP column (Agilent). The inlet temperature was set to 250 °C. The oven temperature was initially set to 40 °C for 3 min, followed by a ramp of 20 °C/min to 150 °C for 6.25 min, and then a ramp of 45 °C/min to a final temperature of 240 °C for 3 min. A 0.3 μL sample was injected with a split ratio of 20:1 and a split flow of 40 mL/min in constant pressure mode at 9.52 psi. The FID detector was set to 300 °C with an air flow of 450 mL/min, H2 fuel flow of 40 mL/min, and makeup nitrogen flow of 45 mL/min. The data were processed and analyzed using OpenLAB CDS ChemStation C.01.07 SR2 (Agilent).
Rifampicin-based mutation assay
The protocol is modified from ref. 39. Specifically, 1 mL of bacterial cultures were harvested and sprayed on rifampicin plates at the concentration of 25 µg/mL. Meanwhile, the same volume of the inoculum was seeded on LB plates to normalize cell number. The mutation rate was determined by calculating the colony-forming units (CFU) per mL on rifampicin plates, normalized with CFU/mL on LB plates.
Analysis of 13C labeling pattern of amino acids and intracellular primary metabolites from methanol-growing SM1
1 mL of SM1 strain grown in MOPS-based 400 mM 13C-methanol medium without trace vitamins was harvested by centrifugation at 5000 g for 10 min. The samples were collected with three biological repeats. After the cell pellets were suspended in 0.5 mL of 6 N HCl, the solutions were heated for 24 h at 100 °C, and then evaporated by a vacuum concentrator (Eppendorf Vacufuge plus). Dry hydrolysates were dissolved in 0.1 mL of pure water for analysis. Direct analysis of amino acids was conducted by a UHPLC-MS system with Agilent 1290 Infinity II ultra-high performance liquid chromatography (UHPLC) system coupled online to Agilent 6545 quadrupole time-of-flight (Q-TOF) mass spectrometer (Agilent Technologies). Samples were retained by using ACQUITY UPLC BEH amide column (1.7 μm, 2.1 × 100 mm, Waters), with a constant flow rate of 0.3 mL/min and the injection volume of 2 μL. The mobile phases were composed of H2O (eluent A) and 90% acetonitrile (eluent B) respectively, where both eluents contained 15 mM ammonium acetate and 0.3% ammonium hydroxide. The Dual Agilent Jet Stream ESI was selected as the ion source, and the instrument was operated in positive and negative full-scan mode, collecting m/z 60–1500. After-run analysis was processed using Agilent Qualitative Analysis 10.0 and Agilent Profinder 10.0 software (Agilent Technologies), which included chromatogram acquisition, detection of mass spectral peaks, and waveform processing. The method for derivatization of amino acids was modified from ref. 40. In brief, the amino acids were derivatized by o-phthalaldehyde (OPA, Sigma), 3-mercaptopropionic acid (MPA, Sigma) and 9-fluorenylmethyl chloroformate (FMOC, Sigma), and then analyzed by LC-MS/MS. Specifically, 30 μL of MPA and 15 μL of OPA were mixed with 5 μL appropriately diluted sample and placed at room temperature for 1 min. After primary amine derivatization, the sample were further reacted with FMOC for 2 min. The LC-MS/MS system is based on Shimadzu LCMS-8045 triple quadrupole mass spectrometer equipped with high performance liquid chromatography system (HPLC, Nexera X2 LC-30AD, Shimadzu). 2 μL of sample was injected into HPLC system, and the derivatized amino acids were separated by Triart C18 ExRS column (1.9 μm, 2.1 × 100 mm, YMC) with a constant flow rate at 0.3 mL/min, 50 °C. The eluent A consisted of 2.38 mM tributylamine, 1.28 mM acetate and 0.4 mM ammonium fluoride, while the eluent B contained 100% methanol with 0.4 mM ammonium fluoride. The HPLC program is as follows: 5–25% B (0.0–4.0 min), 25–98% B (4.0–10.0 min), 98% B (10.0–20.0 min), 98–5% B (20.0–20.2 min) and 5% B (20.2–25.0 min). Heated ESI probe was equipped for ionization and was operated in negative ion mode. The data were processed and analyzed by LabSolutions Version 5.93 software (Shimadzu).
To extract intracellular primary metabolites from SM1, 2 mL of cells at OD600 1 were centrifuged at 5000 ×g for 10 min. The cell pellets were suspended in 0.6 mL of pure water and mechanically disrupted using a Q600 Sonicator. The crude extract was centrifuged at 20,000 × g for 30 min, followed by collecting the supernatant and filtering it through an Amicon 0.5 mL MW 10 kDa filter (Millipore). The eluent was then further concentrated to 0.1 mL by a vacuum concentrator (Eppendorf Vacufuge plus) for analysis. The samples were collected with three biological repeats. Intracellular primary metabolites were detected and analyzed by Shimadzu LCMS-8045 triple quadrupole mass spectrometer. 2 μL of sample was separated by a 15 cm Triart C18 ExRS column (1.9 μm, 2.1 × 150 mm, YMC) in HPLC system with a constant flow rate at 0.3 mL/min 50 °C. The composition of mobile phase, HPLC program, ionization mode, and data processing were the same as the method for analyzing derivatized amino acids.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
The NGS data generated in this study have been deposited in the NCBI database under BioProject PRJNA1103758. The raw proteomics data have been deposited in the PRIDE database under accession ID PXD051688. The raw flow cytometry data have been deposited in Figshare [https://doi.org/10.6084/m9.figshare.27133875.v2]. Source data are provided with this paper.
References
Schrader, J. et al. Methanol-based industrial biotechnology: current status and future perspectives of methylotrophic bacteria. Trends Biotechnol. 27, 107–115 (2009).
Conrado, R. J. & Gonzalez, R. Envisioning the bioconversion of methane to liquid fuels. Science 343, 621–623 (2014).
Bogorad, I. W. et al. Building carbon–carbon bonds using a biocatalytic methanol condensation cycle. Proc. Natl Acad. Sci. USA 111, 15928–15933 (2014).
Chen, C.-T. et al. Synthetic methanol auxotrophy of Escherichia coli for methanol-dependent growth and production. Metab. Eng. 49, 257–266 (2018).
Chen, F. Y.-H., Jung, H.-W., Tsuei, C.-Y. & Liao, J. C. Converting Escherichia coli to a synthetic methylotroph growing solely on methanol. Cell 182, 933–946.e914 (2020).
Meyer, F. et al. Methanol-essential growth of Escherichia coli. Nat. Commun. 9, 1508 (2018).
Kim, S. et al. Growth of E. coli on formate and methanol via the reductive glycine pathway. Nat. Chem. Biol. 16, 538–545 (2020).
Keller, P. et al. Generation of an Escherichia coli strain growing on methanol via the ribulose monophosphate cycle. Nat. Commun. 13, 5243 (2022).
Zhan, C. et al. Reprogramming methanol utilization pathways to convert Saccharomyces cerevisiae to a synthetic methylotroph. Nat. Catal. 6, 435–450 (2023).
Gonzalez, J. E., Bennett, R. K., Papoutsakis, E. T. & Antoniewicz, M. R. Methanol assimilation in Escherichia coli is improved by co-utilization of threonine and deletion of leucine-responsive regulatory protein. Metab. Eng. 45, 67–74 (2018).
Woolston, B. M., King, J. R., Reiter, M., Van Hove, B. & Stephanopoulos, G. Improving formaldehyde consumption drives methanol assimilation in engineered E. coli. Nat. Commun. 9, 2387 (2018).
Chou, A., Lee, S. H., Zhu, F., Clomburg, J. M. & Gonzalez, R. An orthogonal metabolic framework for one-carbon utilization. Nat. Metab. 3, 1385–1399 (2021).
Reiter, M. A. et al. A synthetic methylotrophic Escherichia coli as a chassis for bioproduction from methanol. Nat. Catal. 7, 560–573 (2024).
Gleizer, S. et al. Conversion of Escherichia coli to generate all biomass carbon from CO2. Cell 179, 1255–1263.e1212 (2019).
Lin, P. P. et al. Construction and evolution of an Escherichia coli strain relying on nonoxidative glycolysis for sugar catabolism. Proc. Natl Acad. Sci. USA 115, 3538–3546 (2018).
Klein, V. J., Irla, M., Gil López, M., Brautaset, T. & Fernandes Brito, L. Unravelling formaldehyde metabolism in bacteria: road towards synthetic methylotrophy. Microorganisms 10, 220 (2022).
Bazurto, J. V. et al. EfgA is a conserved formaldehyde sensor that leads to bacterial growth arrest in response to elevated formaldehyde. PLoS Biol. 19, e3001208 (2021).
Barber, R. D. & Donohue, T. J. Pathways for transcriptional activation of a glutathione-dependent formaldehyde dehydrogenase gene. J. Mol. Biol. 280, 775–784 (1998).
Nayak, D. D. & Marx, C. J. Genetic and phenotypic comparison of facultative methylotrophy between Methylobacterium extorquens strains PA1 and AM1. PLoS ONE 9, e107887 (2014).
Müller, J. E. et al. Proteomic analysis of the thermophilic methylotroph B acillus methanolicus MGA 3. Proteomics 14, 725–737 (2014).
Baba, T. et al. Construction of Escherichia coli K‐12 in‐frame, single‐gene knockout mutants: the Keio collection. Mol. Syst. Biol. 2, 2006.0008 (2006).
Rajacharya, G. H., Sharma, A. & Yazdani, S. S. Proteomics and metabolic burden analysis to understand the impact of recombinant protein production in E. coli. Sci. Rep. 14, 12271 (2024).
Dekel, E. & Alon, U. Optimality and evolutionary tuning of the expression level of a protein. Nature 436, 588–592 (2005).
Bhattacharya, S. K. & Dubey, A. K. Metabolic burden as reflected by maintenance coefficient of recombinant Escherichia coli overexpressing target gene. Biotechnol. Lett. 17, 1155–1160 (1995).
Bentley, W. E., Mirjalili, N., Andersen, D. C., Davis, R. H. & Kompala, D. S. Plasmid‐encoded protein: the principal factor in the “metabolic burden” associated with recombinant bacteria. Biotechnol. Bioeng. 35, 668–681 (1990).
Wu, T. Y. et al. Characterization and evolution of an activator-independent methanol dehydrogenase from Cupriavidus necator N-1. Appl Microbiol. Biotechnol. 100, 4969–4983 (2016).
Cox, E. C., Degnen, G. E. & Scheppe, M. L. Mutator gene studies in Escherichia coli: the mutS gene. Genetics 72, 551–567 (1972).
Scott, M., Gunderson, C. W., Mateescu, E. M., Zhang, Z. & Hwa, T. Interdependence of cell growth and gene expression: origins and consequences. Science 330, 1099–1102 (2010).
Dai, X. et al. Reduction of translating ribosomes enables Escherichia coli to maintain elongation rates during slow growth. Nat. Microbiol. 2, 1–9 (2016).
Rudan, M., Schneider, D., Warnecke, T. & Krisko, A. RNA chaperones buffer deleterious mutations in E. coli. Elife 4, e04745 (2015).
Kyle Bennett, R. et al. Regulatory interventions improve the biosynthesis of limiting amino acids from methanol carbon to improve synthetic methylotrophy in Escherichia coli. Biotechnol. Bioeng. 118, 43–57 (2021).
Bennett, R. K., Gonzalez, J. E., Whitaker, W. B., Antoniewicz, M. R. & Papoutsakis, E. T. Expression of heterologous non-oxidative pentose phosphate pathway from Bacillus methanolicus and phosphoglucose isomerase deletion improves methanol assimilation and metabolite production by a synthetic Escherichia coli methylotroph. Metab. Eng. 45, 75–85 (2018).
Müller, J. E. et al. Engineering Escherichia coli for methanol conversion. Metab. Eng. 28, 190–201 (2015).
Whitaker, W. B. et al. Engineering the biological conversion of methanol to specialty chemicals in Escherichia coli. Metab. Eng. 39, 49–59 (2017).
Tyo, K. E., Ajikumar, P. K. & Stephanopoulos, G. Stabilized gene duplication enables long-term selection-free heterologous pathway expression. Nat. Biotechnol. 27, 760–765 (2009).
Carpousis, A. J., Campo, N., Hadjeras, L. & Hamouche, L. Compartmentalization of RNA degradosomes in bacteria controls accessibility to substrates and ensures concerted degradation of mRNA to nucleotides. Annu. Rev. Microbiol. 76, 533–552 (2022).
Vo, P. L. H. et al. CRISPR RNA-guided integrases for high-efficiency, multiplexed bacterial genome engineering. Nat. Biotechnol. 39, 480–489 (2021).
Luo, S. et al. A cell-free self-replenishing CO2-fixing system. Nat. Catal. 5, 154–162 (2022).
Rodríguez-Verdugo, A., Gaut, B. S. & Tenaillon, O. Evolution of Escherichia coli rifampicin resistance in an antibiotic-free environment during thermal stress. BMC Evol. Biol. 13, 1–11 (2013).
Uutela, P., Ketola, R. A., Piepponen, P. & Kostiainen, R. Comparison of different amino acid derivatives and analysis of rat brain microdialysates by liquid chromatography tandem mass spectrometry. Anal. Chim. Acta 633, 223–231 (2009).
Nieh, L.-Y. Flow cytometry analysis of BW25113 containing various versions of BAC or pUC19. https://doi.org/10.6084/m9.figshare.27133875.v2 (2024).
Acknowledgements
We thank the following core facilities at Academia Sinica: the NGS High Throughput Genomics Core Facility at the Biodiversity Research Center, the Imaging Core Facility at the Institute of Molecular Biology Center, Bio-Imaging Core Facility at the Institute of Biological Chemistry, the Comprehensive Flow Cytometry Laboratory at the Institute of Plant and Microbial Biology, and the Academia Sinica Common Mass Spectrometry Facilities for Proteomics and Protein Modification Analysis at the Institute of Biological Chemistry. The NGS Core Facility and mass spectrometry Facility are supported by the Academia Sinica Core Facility and Innovative Instrument Project, Taiwan (AS-CFII-108-114 and AS-CFII-108-107).
Author information
Authors and Affiliations
Contributions
L.Y.N., F.Y.H.C., and J.C.L. conceived the project. L.Y.N. developed the ddp-BAC system. F.Y.H.C., L.Y.N., H.W.J., K.Y.S. conducted the evolution process. K.Y.S. analyzed the ddp-BAC nanopore sequencing results. C.Y.T. conducted and analyzed time-lapse live cell imaging. C.T.L. and Y.Q.L. conducted flow cytometry analysis. F.Y.H.C. analyzed the NGS and proteomics data. L.Y.N. analyzed the formaldehyde concentration of the SM strains. F.Y.H.C. and H.W.J. worked on ddPCR. F.Y.H.C. and H.W.J. conducted the TEM DPC experiment. L.Y.N. and F.Y.H.C. worked on characterizing the mutations of the SM strain. F.Y.H.C. worked on GC-MS analysis metabolites. H.W.J. worked on the library prep for all DNA sequencing samples. All authors worked on designing constructs, cloning, or participated in writing and preparation of the manuscript and figures. J.C.L. supervised the entire project.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Communications thanks Nico Claassens, Kazuyuki Shimizu and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Source data
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Nieh, LY., Chen, F.YH., Jung, HW. et al. Evolutionary engineering of methylotrophic E. coli enables fast growth on methanol. Nat Commun 15, 8840 (2024). https://doi.org/10.1038/s41467-024-53206-4
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41467-024-53206-4
This article is cited by
-
A synthetic methylotroph achieves accelerated cell growth by alleviating transcription-replication conflicts
Nature Communications (2025)
-
A blueprint for designing the next-generation of synthetic C1 microbes
Nature Communications (2025)
-
Metabolic engineering strategies for constructing methylotrophic cell factories
Systems Microbiology and Biomanufacturing (2025)
-
Coupled biological and chemical catalysis for jet biofuel production: a review
Environmental Chemistry Letters (2025)
-
Synthetic methylotrophy: Engineering microbes for sustainable biomanufacturing from C1 compounds
Systems Microbiology and Biomanufacturing (2025)