Enzymes of an alternative pathway of glucose metabolism in obligate methanotrophs

Aerobic methanotrophic bacteria utilize methane as a growth substrate but are unable to grow on any sugars. In this study we have shown that two obligate methanotrophs, Methylotuvimicrobium alcaliphilum 20Z and Methylobacter luteus IMV-B-3098, possess functional glucose dehydrogenase (GDH) and gluconate kinase (GntK). The recombinant GDHs from both methanotrophs were homotetrameric and strongly specific for glucose preferring NAD+ over NADP+. GDH from Mtm. alcaliphilum was most active at pH 10 (Vmax = 95 U/mg protein) and demonstrated very high Km for glucose (91.8 ± 3.8 mM). GDH from Mb. luteus was most active at pH 8.5 (Vmax = 43 U/mg protein) and had lower Km for glucose (16 ± 0.6 mM). The cells of two Mtm. alcaliphilum double mutants with deletions either of the genes encoding GDH and glucokinase (gdh─/glk─) or of the genes encoding gluconate kinase and glucokinase (gntk─/glk─) had the lower glycogen level and the higher contents of intracellular glucose and trehalose compared to the wild type strain. The gntk─/glk─ knockout mutant additionally accumulated gluconic acid. These data, along with bioinformatics analysis, demonstrate that glycogen derived free glucose can enter the Entner–Doudoroff pathway or the pentose phosphate cycle in methanotrophs, bypassing glycolysis via the gluconate shunt.

Aerobic methanotrophs are a highly specialized group of bacteria that utilize methane as an energy and carbon source. Methanotrophs belong to the Alpha-or Gamma-classes of Proteobacteria, as well as to the phylum Verrucomicrobia. Only few representatives of alphaproteobacterial methanotrophs can utilize a limited range of polycarbon compounds and none of the methanotrophs can use sugars as growth substrates [1][2][3] . Nevertheless, bioinformatics and biochemical studies have shown that at least three pathways for cleavage of hexose phosphates are operative in gammaproteobacterial (Type I) methanotrophs: the modified pyrophosphate-dependent glycolysis, the Entner-Doudoroff pathway and the oxidative pentose phosphate cycle; the phosphoketolase glycolytic pathway is also encoded by the methanotrophs 4 . These routes start from hexose phosphates which are the first metabolites of the ribulose monophosphate (RuMP) pathway for C1 assimilation and lead to the formation of C2, C3 and C4 compounds, as well as to the redox cofactors NAD(P)H and ATP generation 5,6 . Moreover, the enzymes for glucose and fructose phosphorylation (ATP-glucokinase and ATP-fructokinase and, additionally, polyphosphate-dependent glucokinase in several species) have been characterized in Type I methanotrophs [7][8][9] .
In spite of such a high versatility of carbohydrate metabolism, Type I methanotrophs demonstrate obligate dependence on C1 compounds and the biochemical basis of their inability to grow on polycarbon compounds remains unclear. Moreover, the genomic analysis also indicates that some Type I methanotrophs encode the putative glucose-1-dehydrogenase (GDH) and gluconate kinase (GntK), whose functionality still has to be elucidated (Fig. 1).
Glucose 1-dehydrogenases are the cofactor-dependent enzymes catalyzing oxidation of the first hydroxyl group of D-glucose to form D-glucono-1,5-lactone. The bacterial GDH (EC1.1.1.47) that catalyzes β-D-glucose oxidation, simultaneously converting NAD(P) + to NAD(P)H, belongs to the extended superfamily of shortchain dehydrogenases/reductases (SDR) 10 . The enzymes of this family show the typical GXGXXG⁄A fingerprint motif found in most NADP + -binding proteins 11,12 . Then D-glucono-1,5-lactone undergoes spontaneous hydrolysis in water (or by gluconolactonase) to gluconic acid, and the latter can be further phosphorylated to 6-phosphogluconate by gluconokinase (GntK) (EC 2.7.1.12). NADP + -GDHs have been described in grampositive heterotrophic bacteria and archaea 13 . In archaea, however, GDH reaction product is further converted Results GDH-His 6 expression, purification and structure. The sequences encoding the proteins with ~ 39% identity to that of GDH characterized in Bacillus subtilis (BSU02830) were found in the Mtm. alcaliphilum and Mb. luteus genomes. The full sequences of the respective ORFs (CCE23386 and WP_027157932) were expressed in E. coli Rosetta (DE3). The recombinant proteins, MtmGDH and MbGDH with 6 histidines at the N-termini, were purified by one-step affinity metal-chelating chromatography. SDS-PAGE of each enzyme revealed one band corresponding to a molecular mass of ~ 30 kDa, which is in accordance with the theoretically calculated masses of 28.6 kDa and 29.1 kDa for GDH from Mtm. alcaliphilum or Mb. luteus, respectively. Gel filtration of the purified enzymes indicated a tetrameric structure of both methanotrophic GDHs (with Mr 112 kDa for MtmGDH and 100 kDa for MbGDH) (Fig. S1).  of NAD + or NADP + , but they were inactive with other sugars tested (L-glucose, fructose, mannose, galactose, sorbose and N-acetyl-glucosamine). The enzyme from Mtm. alcaliphilum was most active at pH 10.0 ( Fig. 2A) and 10-20 °C (Fig. 2C). It lost 52% or 87% of activity after 1-h exposure at 40 °C and 50 °C, respectively. 50 mM NH 4 Cl stimulated the enzyme activity 1.5-fold (Supplementary Table S1). In the presence of 0.2 M NaCl or 2 M KCl, the enzyme activity decreased by 50% and 88%, respectively, but chlorides of bivalent metals had no significant effect. The dependence of activity on the concentrations of substrates obeyed the Michaelis-Menten kinetics. The maximum reaction rate with NAD + was higher than with NADP + (95 versus 39 U/mg protein) ( Table 1). The apparent K m values measured at 30 °C and pH 10 were as follows: 92 ± 4 mM for glucose, 0.36 ± 0.02 mM for NAD + , and 0.15 ± 0.01 mM for NADP + . For testing of influence of histidines on the kinetic parameters of MtmGDH, we carried out the cutting off His 6 -tags and found that the apparent K m values for the substrates did not change compared to the His 6 -tagged enzyme (Table 1). GDH from Mb. luteus was most active at pH 8.0 ( Fig. 2A) and 20-30 °C (Fig. 2C). The enzyme lost 37% and 61% of the activity after heating for 1 h at 50 °C and 70 °C, respectively. The activity slightly increased in the presence of 50 mM KCl, NaCl or NH 4 Cl. However, in the presence of 0.25 M NaCl or 2 M KCl, the residual activity was 84 and 39%, respectively. The chlorides of Ni 2+ , Cd 2+ , Zn 2+ and Co 2+ completely inhibited Mb. luteus GDH (Supplementary Table S1). The maximum of activity with NADP + was higher than with NAD + (43 vs 31 U/mg protein) (Table 1). MbGDH had a higher affinity for glucose compared to the Mtm. alcaliphilum enzyme (K m 16 mM vs. 92 mM) but a lower affinity for NADP + (K m 1.6 mM vs. 0.01 mM) ( Table 1). The calculated catalytic constants k cat and k cat /K m showed that both enzymes had low specificity for glucose and exhibited an order of magnitude higher specificity for NAD + than for NADP + ( Table 1). The activities of both GDHs did not significantly  Expression, purification and structure of GntK-His 6 from methanotrophs. The genes encoding the gluconate kinase from Mtm. alcaliphilum (CE23379) and Mb. luteus (WP_051299846) were heterologously expressed in E. coli Rosetta (DE3). The recombinant proteins MtmGntK and MbGntK with 6 histidines at the C-terminus were purified by affinity Ni-NTA chromatography. SDS-PAGE of each enzyme showed one band corresponding to a molecular mass of ~ 20 kDa, which is consistent with the theoretically calculated masses. Gel filtration of the purified enzymes indicated a monomeric structure of both GntKs (Fig. S1).

Catalytic properties of GntKs from methanotrophs. The recombinant MtmGntK demonstrated
activity 26 U/mg protein. The enzyme was active in the pH range from pH 7.5 to pH 9.5 with the optimum at pH 9.0 ( Fig. 2B) and in the temperature range from 10 to 30 °C (Fig. 2C). It was quite stable, withstanding 2-h heating at 30-50 °C without a decrease in activity. Nonetheless, its activity was fully lost after 5-min exposure at 70 °C. No allosteric regulators were found for MtmGntk; however, various metals significantly reduced the activity (Supplementary Tables S1, S3). Thus, CuCl 2 , CdCl 2 , NiCl 2 , ZnCl 2 and BaCl 2 at a concentration of 1 mM almost fully inhibited the activity, while CaCl 2 and CoCl 2 decreased the activity by 60 and 70%, respectively. 50 mM NH 4 Cl and KCl reduced the activity by 30%. 50 mM NaCl had no influence on the enzyme activity whereas in the presence of 0.5 M NaCl the activity of MtmGntK decreased by 80%. The dependence of the MtmGntK activity on gluconate concentration obeyed the Michaelis-Menten kinetics: the apparent K m value was 0.60 ± 0.03 mM ( Table 2). In response to the changes in ATP concentration, the enzyme behavior did not obey the Michaelis-Menten kinetics: the S 0.5 value, 1.3 ± 0.2 mM; the Hill coefficient, n = 1.6. The maximal activity of MbGntK under optimal conditions (pH 9.0, 40 °C) was 56 U/mg protein (Fig. 2). Unlike MtmGntK, the enzyme from Mb. luteus fully lost the activity after 5-min exposure at 50 °C, while the activity did not change after 2-h exposure at 30 °C. No allosteric regulator of the enzyme was found. CuCl 2 almost completely suppressed the MbGntK activity, while CdCl 2 and ZnCl 2 reduced its activity by 80 and 70%, respectively (Supplementary Table S1). The activity slightly increased in the presence of monovalent cations Na + and  www.nature.com/scientificreports/ NH 4 + (50 mM); however, in the presence of 0.5 or 1 M NaCl, the activity decreased by 70 or 10%, respectively. The apparent K m value for gluconate was 0.23 ± 0.01 mM. Similarly to the enzyme from Mtm. alcaliphilum, the dependence of MbGntK activity on ATP concentration did not obey the Michaelis-Menten kinetics; S 0.5 was 0.042 ± 0.002 mM, the Hill coefficient n = 2.3.

The activities of GDH and Glk in cell-free extracts of methanotrophs. The activity of GDH in cell-
free extracts of Mtm. alcaliphilum was found to be ~ 20 mU/mg protein in the presence of 200 mM glucose in the reaction mixture, but the reaction quickly stopped if 20 mM glucose concentration was used to measure the activity. It corresponded with the extraordinary high apparent K m value (92 mM) found for the purified enzyme. This implies the in vivo expression of glucose dehydrogenase in Mtm. alcaliphilum. The specific activity of GDH in the cell-free extracts of Mb. luteus in the presence of 200 mM glucose was ~ 28 mU/mg protein; at 20 mM glucose, it was ~ 14 mU/mg protein.
The specific activity of glucokinase (Glk) in cell-free extracts of Mtm. alcaliphilum was ~ 30 mU/mg protein. In response to the increase of glucose concentration from 5 to 200 mM in the reaction mixture, the enzyme activity increased to 60 ± 5 mU/mg protein. This increase can be explained by appearance of the low affinity GDH.
The analysis of activities of the putative promoter regions of the gntk and amy genes. We were unsuccessful to find activity of gluconate kinase in the cell-free extracts from both methanotrophs. Therefore, using cat as a reporter gene (encoding chloramphenicol acetyltransferase), we analyzed the activities of the putative promoter region of the gntk gene in Mtm. alcaliphilum. In cell-free extracts of Mtm. alcaliphilum pMHA:Pgntk-cat grown under methane or in the presence of methanol, the activities of chloramphenicol acetyltransferase were 2.6 ± 0.8 and 5.1 ± 0.3 mU/mg, respectively.
In Mtm. alcaliphilum, 10 bp upstream the gntk gene the amy gene encoding alpha amylase is located (Fig. S2). The activity of the amy promoter was also analyzed. When Mtm. alcaliphilum pMHA:Pamy-cat was grown on methanol or under methane, the activities of chloramphenicol acetyltransferase were found (9.7 ± 0.5 or 6.0 ± 0.3 mU/mg respectively). Consequently, one may expect that the gntk and amy genes are functional.

Phenotypic characterization of insertional mutants.
To prove the in vivo glucose metabolism through the GDH/GntK reactions in Mtm. alcaliphilum, the strain with insertion in the gdh gene was obtained but changes in growth characteristics in the mutant strain were not identified. This can be due to the operation of glucokinase highly specific to glucose (K m 0.32 mM) 8 . In the double mutated strain gntk ─ /glk ─ , the GDH activity was 22 ± 5 mU/mg protein but Glk activity was not detected. The gdh ─ /glk ─ strain lacked both GDH and Glk activities.
In the presence of methane or methanol, the double mutation strains grew 1.5 times slower compared to the wild type (WT) strain 20Z (Fig. 3S). The same effect was observed in a single glk ─ mutant 8 . In contrast to the glk ─ strain, where the content of intracellular glucose did not actually differ from the WT strain 8 , the intracellular pools of glucose in the gdh ─ /glk ─ strain grown on methane or methanol increased 50-and 32-fold, respectively (Table 3). At the same time, the glucose content in cells of the gntk ─ /glk ─ strain increased only 3-4 times regardless of the growth substrate. Only a slight increase in glucose in the gntk ─ /glk ─ mutant indicated the functioning of GDH, which converts excess glucose into gluconate. Another evidence of the functionality of GDH and GntK enzymes is accumulation of gluconic acid in the gntk ─ /glk ─ strain (Table 3). In cells of these mutants grown in either methane or methanol, glycogen accumulation dramatically decreased (Table 3). It should be noted that glycogen concentration in the methane grown gdh ─ /glk ─ strain dropped 690 times compared to the WT strain, whereas a minor glycogen decrease was observed in the methanol-grown gntk ─ /glk ─ strain (~ 6 times). The gntk ─ / glk ─ strain growing on methanol accumulated the highest level of intracellular trehalose, but least of all when growing under methane (Table 3). In cells of the gntk ─ /glk ─ and gdh ─ /glk ─ strains growing on methanol a mixture Table 3. Accumulation of carbohydrates (mg per g of DCW) in the wild type (WT) cells of Mtm. alcaliphilum and mutant strains deficient in glucose dehydrogenase/glucokinase (gdh -/glk -) and gluconate kinase/ glucokinase (gntk -/glk -). ND, not detected. The samples for analyses were drawn at late exponential phase of the culture growth.   www.nature.com/scientificreports/ of malto-derivatives of carbohydrates were detected, but it was very difficult to identify these compounds (the individual components of the mixture) by the H 1 -NMR method.
The sequence and distribution of GDH in methanotrophs. As can be judged from protein sequences, GDHs from Mtm. alcaliphilum and Mb. luteus belong to the superfamily of short-chain dehydrogenases/reductases (SDR). These enzymes show the typical GXGXXG fingerprint motif found in NADP + -binding proteins 11,12 (Fig. 4S). There are some differences in the composition of amino acids forming the active center in GDH from Bacillus megaterium and those from Mtm. alcaliphilum and Mb. luteus (marked * in Fig. 4S). The amino acids at the C-terminus of B. megaterium GDH (Q257, A258, R260 and G261) directly interact with glucose and thus determine the kinetic properties of the enzyme. The C-termini at GDHs from Mtm. alcaliphilum and Mb. luteus are significantly different from the C-terminus of the B. megaterium enzyme, which may be reflected in the kinetic properties of methanotrophic enzymes (Fig. 4S). The only difference in amino acids forming hydrogen bonds with the hydroxyl groups of glucose between the Mtm. alcaliphilum and Mb. luteus GDHs is the replacement of aspartate (D219) by glutamate (E219). The GDHs from other methanotrophs have more differences in amino acids presumably forming the active center and interacting with glucose (Fig. 4S). Analysis of the genome databases available showed that only a number of methanotrophs have GDH homologues (Fig. 3, Supplementary Table S4). GDH homologs are present in some other Type I methanotrophs: Methylosarcina fibrata (60% identity of translated amino acids to MtmGDH) and some representatives of the genera Methylocaldum (62% identity) and Methylomonas (29% identity) (Fig. 3). However, the genomes of Methylomonas and Methylosarcina have no gene for gluconate kinase. At the same time, gluconate kinases from Mtm. alcaliphilum, Methylocaldum szegediense and Methylocella silvestris have only 32-37% identity with each other.
A number of other methanotrophs have putative GDHs, which have very low similarity (14-27% identity) to Mtm. alcaliphilum and Bacillus megaterium enzymes, but belong to the SDR family (Fig. 3).

Discussion
In this study we have characterized for the first time glucose dehydrogenase and gluconate kinase, the starting enzymes for glucose metabolism, in aerobic methanotrophs, bacteria which unable to use sugars as growth substrates. The GDHs from Mtm. alcaliphilum and Mb. luteus are strongly specific for D-glucose and use NAD + as the preferred cofactor. To date, microbial NAD(P) + -GDHs have been characterized in gram-positive bacteria, cyanobacteria, archaea, fungi and yeasts, as well as in plants, whereas GDHs investigated in gram-negative bacteria were PQQ-dependent 23 . GDHs from Bacilli can oxidize a broad range of substrates 24 . The NAD(P) + -GDH from B. thuringiensis oxidizes glucose (K m 14 mM) and 2-deoxy-D-glucose (K m 12.2 mM) 17 , and the enzyme from Lysinibacillus sphaericus also oxidizes disaccharides such as maltose, lactose, and sucrose 21 . GDHs from the archaea Thermoproteus tenaxis and Sulfolobus solfataricus have K m 0.3-0.4 mM for glucose, K m values for the enzymes from bacteria utilizing the sugar as growth substrate are in a wide range (0.5-42 mM) 23,25 and K m of the enzyme from Mb. luteus is within the same range. Curiously, GDH from Mtm. alcaliphilum remains the absolute record holder for the highest K m (92 mM). Similarly to the GDHs from Bacillus sp. ZJ and L. sphaericus 18,21 , methanotrophic GDHs are tetramers.
Previously, a single glk ─ mutant exhibiting a drastic decrease in glycogen level (50 times compared to the WT strain) has been characterized while its intracellular glucose level remained almost unaltered 8 . In this work, the gdh ─ /glk ─ mutant lacking genes for both enzymes responsible for activation of free glucose demonstrated considerable increase in intracellular glucose content (Table 3). This indicated that, despite the high K m , GDH in Mtm. alcaliphilum is involved in maintaining the intracellular pool of glucose. Only a slight increase of glucose content in the gntk ─ /glk ─ strain additionally confirmed participation of GDH in utilization of the sugar. The occurrence of gluconate in the gntk ─ /glk ─ strain where the gene for gluconate kinase is turned off and GDH working, proves the in vivo participation of both enzymes in free glucose removal.
In cells of the glk ─ , gdh ─ /glk ─ and gntk ─ /glk ─ strains, trehalose level was drastically increased as compared to the WT cells. The genomic analysis allows us to propose that only glycogen can be a source of trehalose. Thus, the genome of Mtm. alcaliphilum harbors the treY and treZ genes encoding putative maltooligosyltrehalose synthase (CCE23905) and trehalohydrolase (CCE23904) which convert the terminal unit of the linear maltodextrin (glycogen or starch) into trehalose via maltooligosyltrehalose (Fig. 3). It also has the gene for putative maltose alpha-D-glucosyltransferase (trehalose synthase, TreS, CCE22655) which catalyzes the reversible conversion of maltose into trehalose as described for the first time in mycobacteria 26 . At the same time, Mtm. alcaliphilum lacks the otsA-otsB genes encoding trehalose 6-phosphate synthase and trehalose 6-phosphate phosphatase responsible for trehalose synthesis at many (micro)organisms. The methanotroph has gene encoding 4-alphaglucanotransferase (MalQ, CCE25166) which recognizes maltose and longer maltodextrins and preferentially removes glucose from the reducing ends of maltose and from small maltodextrins and transfers the remaining enzyme-bound dextrinyl residue to the non-reducing ends of other maltodextrins, thus forming longer maltodextrin chains 27 . Two Glg enzymes are essential for endogenous maltotriose formation in Mtm. alcaliphilum: glycogen phosphorylase GlgP (CCE22814; CCE22130; CCE25654) which shortens the branch chains of glycogen to the lengths of maltotetraosyl and maltotriosyl units, and GlgX (CCE25166), which releases the branch chains. Some GlgPs can act as maltodextrin phosphorylases MalP 28 (Fig. 1). Therefore, glycogen serves as the main source of free glucose for GDH and Glk reactions. Since glucose and trehalose were found in wild type cells (Table 3), the process of glycogen degradation in Mtm. alcaliphilum most likely goes along with glycogen synthesis. The concerted action of glycogen synthesizing and degrading enzymes has been recently revealed in Corynebacterium glutamicum and E. coli 29 . The simultaneously occurring glycolysis and the oxidative pentose phosphate pathway in addition to the Entner-Doudoroff pathway shunt including glucose dehydrogenase and gluconate kinase has been proven for plants and cyanobacteria 15 www.nature.com/scientificreports/ In Mtm. alcaliphilum 20Z, glucose derived from glycogen degradation is utilized mainly by glucokinase due to its higher specificity to the sugar than GDH. The glucose 6-phosphate formed in Glk reaction is further converted by phosphoglucose isomerase into fructose 6-phosphate, which enters glycolysis after phosphorylation by PPidependent phosphofructokinase and/or the Entner-Doudoroff pathway via glucose 6-phosphate dehydrogenase. In the case of glycogen decomposition and/or inhibition of glucokinase activity by ADP 8 (K i 2.34 mM), GDH can be involved in the glucose utilization. It should be noted that 6-phosphogluconate is a substrate for 6-phosphogluconate dehydrogenase reaction which produces ribulose 5-phosphate, an acceptor of formaldehyde in the key reaction of carbon assimilation catalyzed by hexulosephosphate synthase (Fig. 1). Therefore, metabolism of glucose via the gluconate shunt can be an anaplerotic mechanism that replenishes the RuMP cycle by tapping cellular carbohydrate reservoirs under fluctuations in content of carbon substrate in the environment.
The ability of these bacteria to obtain the key central metabolites from glycogen could allow them to survive the lack of the growth substrate in the environment at the expense of their own internal resources. Previous studies have shown that thermotolerant methanotroph Methylococcus strain NCIB 11083 can metabolize glycogen under aerobic starvation conditions, and this process could generate sufficient energy for protein synthesis in the absence of methane 31 . Recently we have shown that inactivation of the genes encoding the key enzymes of sucrose and glycogen synthesis reduced growth rate of Mtm. alcaliphilum mutant on methane or methanol and increased the protein content in cells of the methanotroph 32 .
The reasons for the inhibition of growth of the strains lacking the genes for metabolism of free glucose are still unclear. Apparently, the accumulation of glucose in the cytoplasm and dramatic drop in glycogen level in the mutants indicate a permanent process of decomposition of glycogen, and inhibition of the polymer synthesis (Fig. S5). The molecular mechanisms underlying the regulation of the synthesis and degradation of the storage polymer in bacteria remain to be elucidated. A direct non-enzymatic effect of glucose on bacterial proteins/ enzymes and the triggering of stress responses are not excluded 33 .
The genomic analyses showed that all Type I methanotrophs possess genes encoding glycogen synthase (Supplementary Table 4S). They also harbor up to two genes for glucokinases with a single exception Methylocaldum marinum S8 lacking Glk encoding gene. None of them encode PTS system for glucose transport. The GDH homologues present in Type I methanotrophs (usually along with GntK ones) with exceptions of Methylococcus capsulatus (Bath and Texas), Methylomonas methanica MC09, Methylomonas sp. MK1 and Methylotuvimicrobium kenyense AMO1 lacking both gluconate shunt enzymes. Several species, for example Methylosarcina fibrata AML-C10 and Methylomicrobium album BG8 encode gluconate dehydrogenase (GntD) instead of the GntK thuds implying another variant of non-phosphorylated gluconate shunt (Supplementary Table 4S).
Among Type II methanotrophs, synthesizing hexosephosphates mainly through gluconeogenesis, almost all representatives have GlgA encoding genes (Supplementary Table 4S). The exceptions are all Methylosinus strains and Methylocystis sp. LW5 lacking glgA genes. About half of the genomes of this group methanotrophs sequenced today also code GDH and few species also have Glk encoding genes. All Type II methanotrophs possess genes for PTS system for sugar transport. Thus, genomic analysis demonstrates high versatility of carbohydrate metabolism in methanotrophs, the bacteria unable to grow on any sugars. The redundancy of the pathways of sugar metabolism in obligate methanotrophs determines their ability to survive under the conditions of CH 4 deficiency, but the question about the causes of their inability to use glucose as a carbon and energy source for growth is still open.

Materials and Methods
Bacteria and growth conditions. Mb. luteus (IMV-B-3098) and Mtm. alcaliphilum 20Z (VKM B-2133 T = NCIMB 14124 T ) were grown at 30 °C on a nitrate salt medium. For the growth of Mtm. alcaliphilum 20Z, the medium additionally contained 0.1 M NaHCO 3 and 0.3 M NaCl 34 . Methane (50% in the gas phase) or methanol (0.5%, v/v) was used as a carbon source. Escherichia coli Rosetta (DE3) obtained from Stratagene (La Jolla) was grown at 37 °C on a Luria-Bertani (LB) agar or in broth. Kanamycin (50 μg/mL) and/or chloramphenicol (25 μg/mL) was added to the growth medium as necessary.

Cloning, preparation and purification of the recombinant enzymes.
To amplify the gene encoding ubiquitin with 6 histidines at the N-terminus from the pHUE plasmid (Supplementary Table S5), the Ub-F/ Ub-R primers (Supplementary Table S6) and vector pET28 for cloning were used. The gdh gene was amplified by PCR from the DNAs of Mtm. alcaliphilum or Mb. luteus using the primers presented in the Supplementary Table S4. The purified PCR product corresponding to the Mtmgdh gene was ligated into the pET28 or pET28Ub-shHis vectors at the NdeI/HindIII or SacII/HindIII restriction sites, respectively. The PCR product carrying the Mbgdh gene was ligated into pET28Ub-shHis at the SacII/HindIII restriction sites. The resulting vectors pET28:gdhMtm, pET28Ub-shHis:gdhMtm, or pET28Ub-shHis:gdhMb were transferred into E. coli Rosetta (DE3), enzyme synthesis was induced by 0.5 mM isopropyl-β-D-thiogalactopyranoside, and cells were incubated for 15 h at 18 °C. Recombinant enzymes with His 6 at the N-termini were purified as described 35 . To remove His-tags, the enzymes were incubated for 18 h with Usp2-His 6 protease (50 μg of protease per 1 mg of protein) at + 4 °C. The efficiency and completeness of hydrolysis were checked by SDS-PAAG electrophoresis. The final purification of GDH preparations from the cut-off His-tag and Usp2-His6 were not carried out, since GDH activity was lost during dialysis.
The vector pET30 (Novagen) was used to amplify the gntk genes from methanotrophic DNAs by the respective primers (Supplementary Table S6). The resulting vector pET30:gntkMtm or pET30:gntkMb was transferred into E. coli Rosetta (DE3). The purification of the recombinant GntKs was carried out as described above. Purified recombinant enzymes were stored in 40% glycerol at −20 °C.  Table S6) and cloned in the pCM184 vector at the BglII/Acc65I and SacII/SacI sites. The resulting vector pCM184:gdhKm was introduced into Mtm. alcaliphilum cells by electroporation 36 . The gdh ─ clones were selected by PCR using primers dGDH2_f1 and dGDHr. The gntk-gnl DNA locus was amplified by PCR using primers dGnKinF and dGnKinR (Supplementary  Table S6) containing the recognition sites for restriction endonucleases EcoRI and MluI and inserted into pCM184 with removal of the kanamycin resistance gene. The middle of this locus was removed at the sites of endonuclease PvuI and replaced by the kanamycin resistance gene. The kanamycin resistance gene was amplified by PCR from pCM184. The resulting plasmid pCM184:gntk-Km was introduced into Mtm. alcaliphilum cells by electroporation.
Glycogen was extracted from 20 mg of freeze-dried cells as described 31 . The glycogen content was estimated using "Glycogen Assay kit" according to the manufacturer's recommendations.
Analysis of the promoter activity of the gntk and amy genes. Using primer pairs PamyF/PamyR or PglnF/PglnR (Supplementary Table S6) the PCR products of 563 and 551 bp containing the sequences of putative amy or gntk promoters, respectively, were obtained. The PCR product with the cat gene was generated with primers SDglnK + CmF/CmR using plasmid p34S-Cm as the source of the cat gene. Primers PamyF/PglnF and CmR are tailed with XbaI and SphI restriction ends, respectively. PamyR/PglnR is tailed with a 20 bp sequence homologous to the first 20 bp of SDglnK + CmF containing the SD sequence of the gntK gene. PCR products containing the amy or the gntk promoter and the cat gene were used as a template in the third PCR reaction to produce the final fusion PCR products Pamy-cat or Pgntk-cat with XbaI and SphI restriction ends. The plasmid pMHA200 (Supplementary Table S5) was digested with these endonucleases to remove a 752-bp segment, which was replaced by the Pamy-cat or Pgntk-cat fragments of DNA. The correct assembly was verified by sequencing, and the new plasmids were designated as pMHA200:Pamy-cat and pMHA200:Pgntk-cat. The resulting plasmids were introduced into Mtm. alcaliphilum cells by electroporation.
The activity of chloramphenicol acetyltransferase (i.e. catalyzis the reaction of the acyl group transfer from acetyl-CoA to chloramphenicol) was assessed by recording the 5-thio-2-benzoic acid accumulation as a result of interaction between 5,5′-dithiobis-(2-nitrobenzoic acid) (DTNB) and CoASH sulfhydryl groups at 30 °C. The reaction mixture (1 mL) contained 50 mM Tris-HCl buffer (pH 7.5), 0.1 mM DTNB, 1 mM acetyl CoA, and 10 − 50 μL cell-free extract of the Mtm. alcaliphilum. The reaction was triggered by 2 mM chloramphenicol. The optical density was measured at 412 nm. The Mtm. alcaliphilum pMHA:cat and WT strains lacked the activity of chloramphenicol acetyltransferase.