Cerium Regulates Expression of Alternative Methanol Dehydrogenases in Methylosinus trichosporium OB3b

Abstract

Methanotrophs have multiple methane monooxygenases that are well known to be regulated by copper, i.e., a “copper switch.” At low copper/biomass ratios the soluble methane monooxygenase (sMMO) is expressed while expression and activity of the particulate methane monooxygenase (pMMO) increases with increasing availability of copper. In many methanotrophs there are also multiple methanol dehydrogenases (MeDHs), one based on Mxa and another based on Xox. Mxa-MeDH is known to have calcium in its active site, while Xox-MeDHs have been shown to have rare earth elements in their active site. We show here that the expression levels of Mxa-MeDH and Xox-MeDH in Methylosinus trichosporium OB3b significantly decreased and increased, respectively, when grown in the presence of cerium but the absence of copper compared to the absence of both metals. Expression of sMMO and pMMO was not affected. In the presence of copper, the effect of cerium on gene expression was less significant, i.e., expression of Mxa-MeDH in the presence of copper and cerium was slightly lower than in the presence of copper alone, but Xox-MeDH was again found to increase significantly. As expected, the addition of copper caused sMMO and pMMO expression levels to significantly decrease and increase, respectively, but the simultaneous addition of cerium had no discernible effect on MMO expression. As a result, it appears Mxa-MeDH can be uncoupled from methane oxidation by sMMO in M. trichosporium OB3b but not from pMMO.

INTRODUCTION

It is well known that microorganisms have diverse mechanisms to sense and respond to metals in their environment. These mechanisms typically include strategies to regulate gene expression in response to the presence or absence of metals such as copper, zinc, iron, manganese, arsenic, and mercury (see, for example, references 1, 2, 3 and 4). One such phenomenon is the “copper switch” in methanotrophs. That is, these microbes utilize methane as their sole growth substrate but have two different monooxygenases for the initial oxidation of methane to methanol. One, the particulate methane monooxygenase (pMMO), is found in the intracytoplasmic membranes of these microbes, and its expression and activity increases with increasing availability of copper. The second, the soluble methane monooxygenase (sMMO), is found in the cytoplasm and is only expressed when copper is unavailable (5). These two forms of MMO have very different structures, activities, and substrate ranges (5–14), and so careful consideration of the form of MMO expressed is critical for understanding methanotrophic ecology, as well for various applications of methanotrophy, including the removal of chlorinated solvents and methane, a potent greenhouse gas (5, 10, 12, 15).

Further, interest in the commercial application of methanotrophy has dramatically accelerated in recent years, in part due to increased methane supplies given advances in hydraulic fracturing of shale formations. As a result, methane prices have become quite low, with the wellhead price of natural gas dropping from $10.79 per 1,000 ft³ in July 2008 to $3.38 per 1,000 ft³ in December 2012 (16). A great deal of effort has thus been put forward to determine how to best valorize methane, e.g., through the use of methanotrophs to convert methane into products such as single-cell protein, bioplastics, biofuels, and osmoprotectants, among other compounds (5, 17, 18). Given that the two MMOs have very different activities and affinities for methane, any use of a methanotrophic platform to valorize methane is strongly affected by the form of MMO present.

Recent findings, however, indicate that the subsequent step in the general pathway of methane oxidation, i.e., the oxidation of methanol by the methanol dehydrogenase, may also play a critical role in the application of methanotrophy for environmental and commercial purposes. First, it was reported in 2006 that the pMMO likely forms a supercomplex with a pyrroquinolone quinone (PQQ)-linked methanol dehydrogenase (MeDH). Cryoelectron microscopy work done by Myronova et al. (19) indicated that the PQQ-linked MeDH likely forms a “cap” to the pMMO “body,” with the PQQ-linked MeDH residing in the periplasmic space. Subsequent studies support this conclusion and indicate that the PQQ-linked MeDH and pMMO supercomplex is anchored via the intracytoplasmic membranes (20). Other research also suggests that electron transfer from the PQQ-linked MeDH to pMMO may occur in vivo (21, 22). As such, any attempt to utilize pMMO-expressing cells for any specific application should also consider efforts to stabilize the PQQ-linked MeDH-pMMO supercomplex and also ensure effective back transfer of electrons from the MeDH to pMMO to drive pMMO activity.

Interestingly, not only are there two known forms of MMO in methanotrophs, most methanotrophs also have an alternative methanol dehydrogenase. As mentioned above, it is well known that methanol is oxidized to formaldehyde via a periplasmic PQQ-linked MeDH. This enzyme is a heterotetrameric protein (α₂β₂) with two 66-kDa (α) subunits (MxaF) and two 8.5-kDa (β) subunits (MxaI). In this MeDH, calcium is in the active site and is coordinated with the PQQ group (23). It was initially believed that this enzyme was critical for methylotrophic growth on methanol since no methanol dehydrogenase activity was observed in mutants defective in the production of this protein (24). Subsequently, however, it was found that there is a homolog to the large subunit, termed XoxF, with 50% sequence identity to MxaF (25, 26). This also encodes a PQQ-dependent methanol dehydrogenase that is associated with the periplasm (26, 27) but appears to be composed only of a single subunit with a predicted mass of 65 kDa (28) or associated with the small subunit of MxaI (26), depending on the microbe. Further, it is often observed that multiple homologs of XoxF are found in the genome of a variety of methylotrophs and methanotrophs (25, 26, 29, 30).

Xox-MeDH, however, appears to have a rare earth element in its active site. Studies in Methylobacterium radiotolerans and Methylobacterium extorquens AM1 showed that cerium and lanthanum both increased methanol oxidation by Xox-MeDH (31, 32). Such increase was not due to increased expression of xoxF but was more likely due to posttranslational activation (32). Further, simple yet elegant studies showed that growth and overall MeDH activity of a M. extorquens AM1 mutant in which mxaF was disrupted was severely limited in the absence of lanthanum but growth recovered in its presence, regardless of whether calcium was simultaneously present or not (32). Such results indicate that lanthanum was required for the activity of Xox-MeDH. Subsequent studies supported these findings, i.e., it was found that growth of the Methylacidiphilum fumariolicum SolV was enhanced in the presence of multiple rare earth elements (e.g., cerium, lanthanum, praseodymium, and neodymium [33]). Purification of the active MeDH of M. fumariolicum SolV grown in the presence of praseodymium showed approximately 0.5 to 0.7 atoms of praseodymium per monomer, indicating that this may be part of the active site. Subsequent crystallization of this MeDH revealed it to be encoded by XoxF and that rare earth metals were in the crystal structure (33).

Given these findings, we speculated that under selective growth conditions, differential expression of not only genes encoding polypeptides of sMMO and pMMO would vary but also genes encoding the two alternative forms of MeDH. Here, we report on the effect of various amounts of copper and cerium on gene expression and the growth of Methylosinus trichosporium OB3b.

MATERIALS AND METHODS

Bacterial growth conditions.

M. trichosporium OB3b was grown on nitrate mineral salt (NMS) medium (34) using >18 MΩ·cm H₂O at 30°C in 250-ml Erlenmeyer flasks under constant shaking at 200 rpm. Various amounts of copper (as CuCl₂) and/or cerium (as CeCl₃) prepared in >18 MΩ·cm H₂O were added, and cultures harvested in the late exponential phase for analysis. Copper and cerium stock solutions were filter sterilized using 0.2-μm-pore-size polyethersulfone membranes. All chemicals used were of American Chemical Society grade or better. All conditions were performed using biological triplicates.

Protein quantification.

The procedure outlined by Semrau et al. (35) was used to quantify protein concentrations. Briefly, 5-ml portions of cultures of M. trichosporium OB3b were concentrated to 1 ml and digested in 2 M NaOH (0.4 ml of 5 M NaOH per 1.0 ml of culture) at 98°C for 15 min. A Bradford assay (Bio-Rad Laboratories, Hercules, CA) was then used to determine the protein concentration according to the manufacturer's instructions. A plot of protein concentrations versus different optical densities at 600 nm (OD₆₀₀) of cultures of M. trichosporium OB3b yielded a linear regression with an OD₆₀₀ of 1.0 equal to 850 μg of protein per ml (R² = 0.995). This correlation was used to calculate protein concentration for all cultures.

Metal measurements.

Copper and cerium associated with biomass and remaining in the supernatant after growth were determined as described earlier (36). Briefly, cultures of M. trichosporium OB3b were centrifuged at 5,000 × g for 10 min at 4°C. The supernatant was then transferred to sterile plastic tubes and stored at −80°C. Cell pellets were resuspended in 1 ml of fresh NMS medium without any added metals and then also stored at −80°C. For subsequent metal analyses, inductively coupled plasma mass spectrometry (Agilent Technologies, Santa Clara, CA) was used. All samples were first thawed at room temperature with supernatant samples and then diluted in NMS with 5% (vol/vol) HNO₃ to achieve a final concentration of 2% (vol/vol) HNO₃. Cell suspensions were first acidified in 1 ml of 70% (vol/vol) HNO₃ and then digested for 2 h at 95°C. Digested cell suspensions were then diluted with sterile NMS medium to achieve a final HNO₃ concentration of 2% (vol/vol). Triplicate biological samples were analyzed with each sample measured five times.

Isolation of methanobactin.

M. trichosporium OB3b was cultured for methanobactin in 0.2 or 1.0 μM CuSO₄-amended NMS medium in sequential batch reactors and methanobactin purified from the spent medium as previously described (37).

Nucleic acid extraction and cDNA preparation.

Total RNA extraction from M. trichosporium OB3b was performed as described earlier (36). Briefly, 2.5 ml of stop solution (5% buffer equilibrated phenol [pH 7.3] in ethanol) was first added to cultures (22.5 ml) to stop synthesis of new mRNA. Cell pellets were then collected by centrifugation at 5,000 × g for 10 min at 4°C. The cells were resuspended in 0.75 ml of extraction buffer before lysis using 20% sodium dodecyl sulfate (SDS), 20% Sarkosyl, and bead beating. Subsequent steps of RNA extraction were then performed as described previously (35, 38). Total RNA was then subjected to RNase-free DNase treatment until free of DNA contamination as via PCR amplification of the 16S rRNA gene. The purified RNA was quantified using a NanoDrop ND1000 spectrophotometer (NanoDrop Technologies, Inc., Wilmington, DE). RNA samples were stored at 80°C and used for cDNA synthesis within 2 days of extraction. DNA-free total RNA (500 ng) was treated with Superscript III reverse transcriptase for reverse transcription of mRNA to cDNA (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions.

RT-qPCR.

Gene specific primers (Table 1) were used for the reverse transcription-quantitative PCR (RT-qPCR) analyses of pmoA, mmoX, mxaF, mxaI, xoxF1, xoxF2, and 16S rRNA in M. trichosporium OB3b grown in different concentrations of copper and cerium. Gel electrophoresis and sequencing analyses were performed to verify the specificities of these primers. Amplifications were performed in 96-well reaction PCR plates using Mx3000P QPCR systems (Stratagene, La Jolla, CA) as previously described (39). Each qPCR was carried out in a 20-μl total volume that contained: cDNA (0.8 μl), Brilliant III SYBR green qPCR Mastermix, 1×; Agilent Technologies, Santa Clara, CA), ROX dye (15 nM), forward and reverse primers (0.5 μM each), and nuclease-free sterile water (Ambion/Life Technologies, Grand Island, NY). The thermal cycler program for qPCR consisted of 40 cycles of denaturation (95°C for 30 s), annealing (58°C for 20 s), and extension (68°C for 30 s) after an initial denaturation at 95°C for 10 min. After the completion of amplification cycles, qPCR products were subjected to melting curve analysis with temperature ranging from 55°C to 95°C to confirm their specificity. The threshold amplification cycle (C_T) values were then imported from MxPro (Stratagene, La Jolla, CA) into Microsoft Excel to quantify the relative expression of different genes. The relative gene expression levels were calculated by using a comparative C_T method (2^−ΔΔCT) (40) with 16S rRNA as the housekeeping gene. Calibration curves for examined qPCR products are shown in Fig. S1 in the supplemental material.

TABLE 1.

Primers used in this study

Targeted gene	Primer	Sequence (5′-3′)	Source or reference
pmoA	qpmoA_FO	TTCTGGGGCTGGACCTAYTTC	45
	qpmoA_RO	CCGACAGCAGCAGGATGATG	45
mmoX	qmmoX_FO	TCAACACCGATCTSAACAACG	45
	qmmoX_RO	TCCAGATTCCRCCCCAATCC	45
16S rRNA	q16S rRNA_FO	GCAGAACCTTACCAGCTTTTGAC	45
	q16S rRNA_RO	CCCTTGCGGGAAGGAAGTC	45
mxaF	qmxaF_FO	CTACATGACCGCCTATGACG	This study
	qmxaF_RO	ATTGGCCTTGTTGAAGTCGT	This study
mxaI	qmxaI_FO	TACGATCCCAAGCATGACCC	This study
	qmxaI_RO	CGTAGATCCATTTGCCGCTC	This study
xoxF1	qxoxF1_FO	TCAAGGACAAGGTGTTCGTC	This study
	qxoxF1_RO	CGAGCCGTCCTTGATGTTAT	This study
xoxF2	qxoxF2_FO	GCGCGAAGGATTGGGAATAT	This study
	qxoxF2_RO	GCCTCGTAATTCATGCACAG	This study

Electrophoresis and N-terminal sequence.

SDS-PAGE was performed on precast NuPAGE 4 to 12% bis-Tris gradient gels from Invitrogen (Life Technologies) with MES SDS running buffer. Gels were stained for total protein with Coomassie brilliant blue R or blotted for N-terminal sequencing. Proteins were blotted onto polyvinylidene difluoride (PVDF) Plus transfer membranes (Micron Separations, Inc., Westboro, MA) using an Xcell II Blot Module (Invitrogen) according to the manufacturer's specifications.

Amino acid sequence.

Amino acid sequence analyses were performed via Edman degradation with a Perkin-Elmer Biosystems model 494 Procise protein/peptide sequencer with an on-line Perkin-Elmer Applied Biosystems model 140C PTA amino acid analyzer. Sequence analysis was performed on samples electroblotted to PVDF membranes as described above.

Spectroscopy.

UV-visible absorption spectra were determined on a Cary 50 (Varian, Inc., Palo Alto, CA) spectrophotometer. Cerium titration experiments were performed using 50 μM aqueous solutions of methanobactin prepared in >18 MΩ·cm H₂O.

ITC.

Isothermal titration calorimetry (ITC) was performed at 25°C using a GE Microcal ITC200 microcalorimeter (GE Healthcare, Piscataway, NJ). The titration solution was 1 mM CeCl₃ and was prepared using >18 MΩ·cm H₂O. The injections were added at 180-s intervals and, based on the injection volume, the duration of the injection was predetermined by the software. An injection volume of 1.5 μl into a cell containing 100 μM methanobactin in >18 MΩ·cm H₂O with a stir rate of 800 rpm was used for all injections. The instrument was cleaned between experiments, and the sample cell washed according to the manufacturer's recommendation. The sample cell was then conditioned with 100 μM methanobactin to remove residual metal. The data were analyzed using nonlinear least-squares curve fitting in Origin 7.0 software (GE Healthcare) after subtraction of the heat of dilution of CeCl₃ into >18 MΩ·cm H₂O.

RESULTS

In the presence of various amounts of copper and cerium, little difference of the growth of M. trichosporium OB3b was observed (see Fig. S2 in the supplemental material). As found earlier (36, 39), copper associated with biomass significantly increased (approximately three orders of magnitude; P < 6 × 10⁻³) with the addition of copper (Fig. 1A). Interestingly, as cerium was added, the amount of cerium associated with biomass also increased significantly (by 3 to 4 orders of magnitude; P < 1.1 × 10⁻⁴; Fig. 1B), and in fact, most of the added cerium (>98%) was cell associated. Cerium binding by methanobactin was determined via spectral changes and isothermal calorimetry. The spectral changes in methanobactin were minor following cerium addition and suggest that association is only to the enethiol groups and not to the oxazolone rings of methanobactin (see Fig. S3A in the supplemental material). The binding of cerium by methanobactin, K_Ce = 4.35 × 10³ M⁻¹ (see Fig. S3B and S3C in the supplemental material) was orders of magnitude lower than that found earlier for copper, K_Cu = 10¹⁸ to 10⁵⁸ M⁻¹ (41–43) and copper displaced cerium associated with methanobactin (results not shown).

FIG 1.

Metals associated with the biomass of M. trichosporium OB3b grown in the presence of various amounts of copper and cerium. (A) Copper; (B) cerium. Errors bars represent the standard deviations of triplicate samples. Columns in each plot labeled by different letters are significantly different (P < 0.05).

Quantitative PCR (qPCR) of cDNA was then performed to determine whether various amounts of copper and cerium affected expression of the various forms of both methane monooxygenase and methanol dehydrogenase. As found previously (35, 36, 39), the addition of copper reduced mmoX expression by ∼4 orders of magnitude (Fig. 2A), while pmoA expression increased ∼54-fold (Fig. 2B), with both changes significant (P = 0.03 and 7.6 × 10⁻³, respectively). The addition of cerium, however, did not significantly affect either mmoX or pmoA expression in either the presence or absence of copper.

FIG 2.

RT-qPCR of mmoX (A), pmoA (B), mxaF (C), mxaI (D), xoxF1 (E), and xoxF2 (F) genes in M. trichosporium OB3b grown in the presence of various amounts of copper and cerium. Errors bars represent the standard deviations of triplicate samples. Columns in each plot labeled by different letters are significantly different (P < 0.05).

The expression of mxaF and mxaI, however, did respond to the addition of copper or cerium (Fig. 2C and D). When 25 μM cerium was added in the absence of copper, both mxaF and mxaI expression decreased >50-fold compared to no added cerium and copper (P = 6.3 × 10⁻³ and 1.9 × 10⁻³, respectively). In the presence of 10 μM copper, the simultaneous addition of 25 μM cerium caused mxaF and mxaI expression to be reduced by >2.5-fold each compared to when no cerium was added in the presence of copper (P = 7.6 × 10⁻³ and 8.4 × 10⁻³, respectively). Further, mxaF expression increased ∼2-fold in the presence of 10 μM copper compared to no added copper (P = 0.038), whereas mxaI expression also increased ∼2.4-fold (P = 0.05). The expression of mxaF and mxaI in the presence of both copper and cerium, however, was similar to that found in the absence of both metals (P = 0.4 and 0.5, respectively).

The expression of both xoxF1 and xoxF2 (Fig. 2E and F) increased more than an order of magnitude when cerium was added compared to when both metals were absent (P = 3.3 × 10⁻⁴ and 4.7 × 10⁻³, respectively). In the presence of 10 μM copper, the expression of both xoxF1 and xoxF2 was not significantly different from that observed in the absence of both metals (P = 0.7 and 0.6, respectively). With the simultaneous addition of 25 μM cerium, however, xoxF1 and xoxF2 expression increased by approximately 9- and 3.5-fold, respectively, and again such increases were significant (P = 7.0 × 10⁻³ and 0.02, respectively).

Given the response of mxaF, xoxF1, and xoxF2 expression to the presence of cerium in the absence of copper, it appears that under some conditions, i.e., sMMO-expressing conditions, that Xox-methanol dehydrogenase could replace the Mxa-methanol dehydrogenase. This was examined more closely through SDS-PAGE protein gels. Given the similar sizes of XoxF1/F2 and MxaF (65 and 66 kDa, respectively [23, 28, 44]), the presence of the small subunit of Mxa-methanol dehydrogenase, MxaI (8.5 kDa [23]), was tracked under various growth concentrations of copper and cerium. As can be seen in Fig. 3, in the absence of copper and cerium, a strong band at ∼8.5 kDa was observed in the cell extract of M. trichosporium OB3b, and this band was absent when 25 μM cerium was added. This band was visible, however, in the presence of copper regardless of the presence or absence of cerium. The N-terminal sequence of this band was determined, and all identified amino acids (10 of the first 11, with one unidentified residue) aligned with the predicted amino acid sequence of MxaI from M. trichosporium OB3b (see Fig. S4 in the supplemental material).

FIG 3.

SDS-PAGE of cell extracts of M. trichosporium OB3b grown with various amounts of copper and cerium. Lane S, molecular mass standards (kDa); lane 1, M. trichosporium OB3b grown with 0 μM copper plus 0 μM cerium; lane 2, M. trichosporium OB3b grown with 0 μM copper plus 25 μM cerium; lane 3, M. trichosporium OB3b grown with 10 μM copper plus 0 μM cerium; lane 4, M. trichosporium OB3b grown with 10 μM copper plus 25 μM cerium.

DISCUSSION

We show here that multiple metals affect gene expression in M. trichosporium OB3b. That is, in addition to the canonical “copper switch” that is well known to control expression of the two forms of MMO (5), cerium also appears to regulate expression of multiple methanol dehydrogenases found in this methanotroph. Specifically, the expression of mxaF and mxaI decreased significantly in the presence of cerium but with no added copper compared to the absence of both metals, whereas xoxF1 and xoxF2 increased. These findings suggest that Xox-methanol dehydrogenase could replace MxaF-methanol dehydrogenase when M. trichosporium OB3b was expressing sMMO. Cerium, however, had little effect on mxaF or mxaI expression under pMMO-expressing conditions, i.e., when 10 μM copper was present. Such findings support earlier conclusions that the pMMO forms a supercomplex with the Mxa-MeDH (19, 20) and that this complex is critical for the oxidation of methane in methanotrophs under pMMO-expressing conditions, i.e., in the presence of copper. Our findings suggest, however, that Mxa-MeDH is not essential in sMMO-expressing conditions; rather, Xox-MeDH is sufficient for the further oxidation of methanol. Further, SDS-PAGE data and subsequent N-terminal sequencing show that in the absence of copper but the presence of cerium, the MxaI polypeptide was not evident, suggesting that, as found for methylotrophs, Xox-MeDH is active in M. trichosporium OB3b and requires only XoxF and not MxaI (28).

The finding that multiple metals affect gene expression in M. trichosporium is intriguing. It suggests that this microbe has multiple mechanisms to sense and collect both copper and cerium and that these mechanisms may play a role in controlling the relative expression of Mxa- and Xox-methanol dehydrogenases. It has been shown that copper uptake in M. trichosporium OB3b is regulated by the chalkophore, methanobactin, and a regulatory model has been proposed whereby methanobactin serves to enhance the magnitude of the “copper switch” but is not the basis of the switch (35). By analogy, there appears to be a separate mechanism by which cerium is sensed by M. trichosporium OB3b, but it should be stressed that this mechanism is still unknown. It does not appear that methanobactin is the mechanism for cerium uptake since it was only loosely bound to methanobactin and copper displaced cerium from methanobactin, and yet most of the added cerium was found to be cell associated. This is intriguing, for although cerium is considered a rare earth element, as noted elsewhere (33), such metals are not actually “rare” since they are a significant fraction of the earth's crust. Rather, they are considered “rare” given that most species of these elements are sparingly soluble. From the data presented here, it is tempting to speculate that systems for the uptake of rare earth elements exist and that the availability of and competition for these metals may have a significant effect on overall methanotrophic community composition and activity.

It is recommended that this work be extended to see what, if any, other rare earth elements also affect expression of Mxa- versus Xox-MeDH in methanotrophs and to also determine whether any growth parameters, e.g., yield and carbon conversion efficiency, are enhanced in the presence of rare earth elements, particularly in sMMO-expressing conditions. It may be that rare earth elements affect methanotrophic community composition as well as specific gene expression and that simple strategies whereby the presence of rare earth elements is controlled can enhance the utility of methanotrophs for a variety of environmental and industrial applications.