SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. Author information
  8. References
  9. Supporting Information

Growth of Methylacidiphilum fumariolicum SolV, an extremely acidophilic methanotrophic microbe isolated from an Italian volcanic mudpot, is shown to be strictly dependent on the presence of lanthanides, a group of rare earth elements (REEs) such as lanthanum (Ln), cerium (Ce), praseodymium (Pr) and neodymium (Nd). After fractionation of the bacterial cells and crystallization of the methanol dehydrogenase (MDH), it was shown that lanthanides were essential as cofactor in a homodimeric MDH comparable with one of the MDHs of Methylobacterium extorquens AM1. We hypothesize that the lanthanides provide superior catalytic properties to pyrroloquinoline quinone (PQQ)-dependent MDH, which is a key enzyme for both methanotrophs and methylotrophs. Thus far, all isolated MxaF-type MDHs contain calcium as a catalytic cofactor. The gene encoding the MDH of strain SolV was identified to be a xoxF-ortholog, phylogenetically closely related to mxaF. Analysis of the protein structure and alignment of amino acids showed potential REE-binding motifs in XoxF enzymes of many methylotrophs, suggesting that these may also be lanthanide-dependent MDHs. Our findings will have major environmental implications as metagenome studies showed (lanthanide-containing) XoxF-type MDH is much more prominent in nature than MxaF-type enzymes.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. Author information
  8. References
  9. Supporting Information

Members of the bacterial phylum Verrucomicrobia can be found in high numbers in a wide range of habitats including soils, insect guts, aquatic systems, marine sediments and hot springs (Wagner and Horn, 2006). Their physiology is largely unknown, and until recently, the (few) cultivated strains were described as anaerobic or aerobic heterotrophs. This changed with the discovery of methanotrophic Verrucomicrobia in geothermal regions (Dunfield et al., 2007; Pol et al., 2007; Islam et al., 2008) for which the novel genus ‘Methylacidiphilum’ was proposed (Op den Camp et al., 2009). They all are likely autotrophs (Khadem et al., 2011; Sharp et al., 2012) that use methane as an energy source and grow optimal between pH 2 and 5 at temperatures of 50–60°C. Besides representatives of the NC10 phylum, which have an intra-aerobic pathway (Ettwig et al., 2010), the verrucomicrobial methanotrophs are the only known groups of aerobic methanotrophs outside of the Proteobacteria phylum. The representatives of the genus Methylacidiphilum are by far the most acidophilic (< pH 1) bacteria capable of methane oxidation.

When cultivating the verrucomicrobial methanotrophs, unknown factors seemed to limit growth under laboratory conditions in defined media, and this may have prevented their more frequent manifestation during enrichment and isolation studies of methanotrophs and methylotrophs from freshwater, marine and soil ecosystems (Chistoserdova et al., 2009).

Rare earth elements (REEs) are essential for many high-tech devices like solar cells, mobile phones and computers. Although an exact role for them in biology is not known, some lanthanides, a group of REE such as lanthanum (Ln), cerium (Ce), praseodymium (Pr) and neodymium (Nd), were shown to affect various metabolic processes. Lanthanides are used as fertilizers in agriculture, and their positive effects have been attributed to interactions with ribulose-1,5-bisphosphate carboxylase/oxygenase (Liu et al., 2011) and with the photosynthetic system, but the underlying mechanisms are poorly understood (Tyler, 2004). REEs are abundantly present in the earth's crust, but because of their low solubility, their concentration in most ecosystems does not exceed the nanomolar range. It has been suggested that the low biological availability of lanthanides has precluded the evolutionary development of lanthanide variants of metallo-enzymes, which were argued to be better alternatives to calcium (Franklin, 2001; Lim and Franklin, 2004). Recently, stimulation of xoxF gene-encoded methanol dehydrogenase (MDH) expression in Bradyrhizobium sp. MAFF211645 and Methylobacterium radiotolerans by Ce(III) and La(III) has been reported (Fitriyanto et al., 2011; Hibi et al., 2011). Although the authors discussed the possibility of gene-level regulation by Ce, an Ln(III)-dependent MDH (XoxF-type) like the recently reported XoxF1 of Methylobacterium extorquens AM1 (Nakagawa et al., 2012) seems to be a more plausible explanation. This bacterium possesses also Ca(II)-dependent MDH (MxaF type), being the main enzyme catalysing methanol oxidation during growth on methanol. Lanthanides (III) have some similarities to calcium (II), such as ionic radius, coordination and ligand preferences. Especially for Ce, it is known that it can replace calcium in biomolecules or act as an antagonist, and as such is used in pharmacology (Jakupec et al., 2005).

In this study, we show the dependency on lanthanides for growth of Methylacidiphilum fumariolicum SolV. In addition, we purified, characterized and crystallized the Ln(III) containing MDH of this bacterium, and we discuss the environmental relevance of our findings.

Results and discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. Author information
  8. References
  9. Supporting Information

During our studies of M. fumariolicum SolV, we noticed that growth was very poor unless the medium was supplemented with mudpot water (MPW) from the Solfatara, its original habitat. No trace element with a known physiological role in biology could replace MPW. The final density obtained in the exponential growth phase was proportional to the amount of MPW added. Different batches of MPW resulted in varying responses, but addition of 1% v/v MPW (collected in 2011) resulted in exponential growth until an optical density of 0.5 at 600 nm (OD600), after which growth continued in a linear mode. The growth-promoting compound in the MPW was evidently a mineral, as after evaporation of the water and ashing the residue (4 h at 550°C), it still promoted the growth of M. fumariolicum SolV. In search for the identity of this essential mineral, we analysed the trace metal composition of MPW, and of the supplemented medium before and after growth of M. fumariolicum SolV, by inductive coupled plasma mass spectrometry (ICP-MS). Thus, it could be established that REEs were present in MPW at concentrations of 2–3 μM (Table 1) much higher than in other ecosystems. Such high REE concentrations are typical for very acidic waters and are 100–1000 times higher compared with circumneutral surface or soil pore water and reflect their low solubility rather than their environmental abundance (Grawunder and Merten, 2012). The MPW and surrounding soils of the Solfatara have a very low pH (1–2) as a result of microbial production of sulphuric acid by oxidation of hydrogen sulfide present in the fumaroles. Hot sulfuric acid is actually used as extraction method in REE mining.

Table 1. REE in Solfatara mudpot water
Collection yearConcentration (μM)
20112009
La0.70.47
Ce1.40.95
Pr0.130.09
Nd0.450.29
Sm0.070.045
Eu0.010.01
Gd0.060.04
Total REE2.81.9

After exponential growth in medium containing 2% MPW, the REE had completely been removed by the M. fumariolicum SolV cells. Individually added REE salts (66 nM) could replace the MPW; SolV had consumed the supplemented La(III), Ce(III) or Pr(III) from the medium after the exponential growth had ceased. When the cells were grown with a range of Ce(III) concentrations, the stimulating activity was proportional to Ce(III) from 0 to 80 nM (Fig. 1). The OD600 obtained at the end of the exponential growth phase with 40 nM Ce(III) was 0.6, and this matched well with that obtained on MPW, considering the sum of La, Ce, Pr and Nd, the major REE present in the MPW. When Ce(III) was added in excess, the uptake from the medium was 10-fold higher than needed for optimal growth. This may indicate that M. fumariolicum SolV may actively store this element. Another explanation would be adsorption to the cell wall. During growth in MPW-supplemented medium, Ce and La seem to disappear more quickly compared with Pr and Nd (Fig. S1). When tested individually, the maximum growth rates with these four REEs were identical. Salts of REE with higher atomic numbers than Nd were less supportive of growth. For example, with gadolinium, the doubling time of SolV had increased fourfold to 30 h (Fig. 2). The observation of a concentration dependent linear growth phase (Fig. 1) can only be explained assuming a limitation in the energy generating system with all structural biosynthesis routes unaffected.

figure

Figure 1. Cerium-dependent growth of Methylacidiphilum fumariolicum SolV. At t = 0 bottles with different Ce(III) concentrations were inoculated with washed cells from a preculture until an optical density of 0.05. The preculture was grown on 100 nM of Ce(III). Data points are the average of two independent experiments.

Download figure to PowerPoint

figure

Figure 2. Maximum growth rates with La, Ce, Pr, Nd and REE with higher atomic numbers tested individually at a concentration of 250 nM. Values are the mean of duplicate experiments, differing not more than 5%.

Download figure to PowerPoint

In order to elucidate the exact role of the REE in M. fumariolicum SolV, we fractionated the proteins of SolV cells by cation exchange chromatography. Fractions containing MDH activity showed high REE concentrations. The pure MDH has an ultraviolet–visible (UV-Vis) spectrum (Fig. S2) with the features of dehydrogenases that contain cofactor PQQ (in its semiquinone oxidation state) (Anthony, 2000; 2004), but the peak typically present at around 345 nm had shifted to 354 nm. Calcium, the major metallic constituent of other MDHs, was below 10 mole % (Table 2) despite its presence (low mM range) both in the medium and MPW. Zn, Cu, Ba and Sr were also detected in the fractions. Of these, only Ba and Sr have been shown to be able to substitute Ca and produce active MDH (Goodwin and Anthony, 1996), but both metals did not stimulate growth of strain SolV. The highest signals in ICP-MS analysis were observed for the REEs La, Ce and Nd, which in total accounted for 0.6 atoms per MDH (monomer).

Table 2. Metals in purified MDH of Methylacidiphilum fumariolicum SolV
 Mole % of metals in MDH
MPW-MDHaPr-MDHb
Analysis 1Analysis 2No EDTAEDTA
  1. a

    MDH purified from M. fumariolicum SolV after growth in medium supplemented with MPW; two samples were independently destructed with hot nitric acid and analysed by ICP-MS.

  2. b

    MDH purified from M. fumariolicum SolV after growth in medium supplemented with Pr(III). One sample with and one without EDTA treatment were destructed with hot nitric acid and analysed.

  3. c

    For Ca, the analysis was done by ICP-AES, which had lower background values than ICP-MS.

  4. ICP-AES, inductively coupled plasma atomic emission spectroscopy; n.a., not analysed.<, below indicated value of the background; –, below detection level (0.1%).

Cac< 10< 5< 8< 8
Mn54n.a.1
Cu1517385
Zn1292n.a.
Sr1.81< 1< 1
Ba8922
La1921
Ce2627
Prn.a.1.960–7050–70
Ndn.a.4.5
Smn.a.0.8
Eun.a.0.1
Gdn.a.

When M. fumariolicum SolV was cultivated in a medium without MPW but with only PrIII, a similarly active MDH with the same UV-Vis spectrum was purified from the resulting biomass. Metal analysis showed 0.5–0.7 Pr atoms per MDH monomer (Table 2). Again, Cu was found in relatively high amounts but seemed to be non-specifically bound because, in contrast to Pr, the copper content could be eightfold reduced upon washing the MDH with ethylenediaminetetraacetic acid (EDTA).

Sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE) of the purified MDH revealed a single band at 60 kDa (Fig. S2), and the gene encoding this protein was identified by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) analysis of the band to be the xoxF gene (Mfum_190005; 63 586 Da after SignalP (http://www.cbs.dtu.dk/services/SignalP/) predicted splicing) of the xoxFGJ operon (Op den Camp et al., 2009). This operon is also present in the genome of Methylacidiphilum infernorum V4 (Hou et al., 2008). The xoxF gene is the orthologue of mxaF that encodes the large subunit of typical MDHs. These MDH's are heterotetramers that contain a large and a small subunit (α2β2). The small subunit is encoded by mxaI, which apparently is lacking from the xox operon of verrucomicrobial methanotrophs. In addition, the xox-operon of these methanotrophs lacks any accessory genes like mxaACKLD shown to be involved in maturation/calcium insertion (Richardson and Anthony, 1992; Goodwin et al., 1996; Toyama et al., 1998). XoxFs are found in all available methylotroph genomes, but a clear role in methanol and/or formaldehyde oxidation has not been assigned as yet (Schmidt et al., 2010; Chistoserdova, 2011).

The oligomeric state of purified MDH was assessed by analytical ultracentrifugation analysis which yielded an S-value of 7.1 which fitted best to a (dimeric) protein of 127 kD (Fig. S3). In the dye-linked assay, all MDHs reported so far showed hardly any activity without ammonium and required high pH (8–10) values (Anthony, 2000; 2004). In contrast, the MDH of M. fumariolicum SolV showed highest activity at neutral pH values (Fig. S4) and was completely independent of ammonium. Purified MDH displayed only very little endogenous substrate activity, and residual substrate was converted within a few minutes to produce virtual zero background rates. This made it possible to accurately measure the MDH kinetic properties for various substrates. The purified MDH had a temperature optimum around 60°C with a maximum specific activity of about 4 μmol·min−1 mg protein−1 which compares favourable with other MDHs. As reported for some other MDHs, primary alcohols and formaldehyde were oxidized at similar rates as methanol (Table 3). However, the affinity for methanol and the primary alcohols tested was much higher than those reported for other MDHs (affinity constants one order of magnitude lower). For methanol, an affinity constant as low as 0.8 μM was measured (Fig. S5), which is the lowest value reported ever for an MDH. When limiting amounts of methanol were added in the assay, almost twice the molar amounts (> 180%) of dye (dichlorophenolindophenol, DCPIP) were reduced, which indicates oxidation continues until to the level of formic acid. Other primary alcohols were oxidized to the respective aldehydes, reducing almost equimolar (> 90%) amounts of reduced DCPIP. Secondary alcohols and acetaldehyde were not oxidized. Our results, together with those reported by Nakagawa and colleagues (2012), clearly show that the XoxF-type enzymes are real MDH.

Table 3. Kinetic parameters of purified MDH of Methylacidiphilum fumariolicum SolV
SubstrateAffinity constant (μM)aMaximum rate %b
  1. a

    Affinity constant were obtained by best curve fitting of Michaelis–Menten kinetics as shown in Fig. S5. Values represent the average of a duplicate determination of which the individual values did not differ more then 20%. For methanol, the value is the average of five independent determinations of which the standard deviation is given in brackets.

  2. b

    Rate as compared with rate obtained for methanol as a substrate.

Methanol0.8 (± 0.3)100
Ethanol3100
1-Propanol7100
1-Butanol6100
1-Hexanol2100
2-Propanol0
Formaldehyde7100
Acetaldehyde> 200< 50
Formate0

All of these highly unusual properties prompted us to determine the crystal structure of this MDH. The purified MDH was crystallized, and its structure was solved from a 1.6 Å resolution dataset (see Table S1) by molecular replacement using a model constructed from the Methylobacterium extorquens MDH structure (Ghosh et al., 1995). The native enzyme appeared to be an asymmetric homodimer containing the PQQ cofactor (Fig. 3A). Calculations predict the sedimentation coefficient of the MDH dimer to be 7.0, which is very close to the experimental value (see above). Very high electron density consistent with a metal ion was present in each monomer of the MDH dimer. During refinement, it became apparent from both the density and the distances to the coordinating atoms (Table 4, Fig. 3B) that calcium, zinc, manganese and copper, found in MDH in low concentrations, could not account for the density in the metal binding sites. However, when Ce or Ln ions were placed in these positions instead, this resulted in good convergence when an occupancy of 0.6 was assumed for the lanthanide ions. No distance restraints were used at any time between the Ce or Ln ions and its coordinating atoms. Also, the planarity restraints on the PQQ O4 atoms had to be released to conform to the density, suggesting the PQQs to be in the radical semiquinone state (Anthony, 1996). Residual density in the active sites, coordinating the Ce/La-ions, was modelled as a (poly)-ethylene glycol fragment. A second dataset was collected at a wavelength of 2.01074 Å to search for the anomalous signal from any lanthanide ions present. An anomalous difference Fourier map calculated from these data indeed showed 35σ peaks at the positions of the ions (Fig. 3C), lending further support to the notion that these are lanthanide ions. In principle, barium ions could explain the electron density, the anomalous signal and the coordination distances; however, barium is present at 6- to 30-fold lower amounts (Table 2). We therefore conclude that the metal ions observed in the MDH crystal structure are lanthanide ions.

figure

Figure 3. X-ray crystal structure of Methylacidiphilum fumariolicum SolV methanol dehydrogenase.

A. Dimer structure with the PPQ shown in blue and the cerium ion in purple.

B. Close-up of the active site of MDH. The PQQ is shown in blue, the cerium ion in purple, the solvent molecule in orange and the coordination of the cerium ion is shown by red dashes.

C. Anomalous difference density (green mesh) calculated from data collected with 2.01074 Å wavelength radiation, contoured at 30 σ. A strong peak (35 σ) supports the presence of a cerium ion in the active site of MDH.

Download figure to PowerPoint

Table 4. Distances to metal ion (Å)
AtomDistance to CeDistance to Ca in MDH M. extorquensb
Ce AaCe Ba
  1. a

    A and B refer to the two MDH monomers in the asymmetric unit.

  2. b

    Williams et al. (2005).

Glu172 OE12.72.62.4
Glu172 OE22.92.92.8
Asn256 OD12.72.73.1
Asp301 OD12.52.6
Asp301 OD22.82.7
Asp299 OD12.92.9 
PQQ O52.62.52.3
PQQ N62.82.82.3
PQQ O72.72.72.4
(poly)-ethylene glycol fragment O2.83.0 

We speculate that the superior substrate affinity of lanthanide-containing MDH (Ln-MDH) is caused by the presence of lanthanides instead of Ca, as this was the most conspicuous change compared with Ca-MDH. All of the amino acids that coordinate PQQ and the metal ion were conserved, but the lanthanide is additionally coordinated by an asparagine (Asp301), which replaces an alanine that is found in all MxaFs. Lanthanide ions are slightly bigger than calcium, and the extra space required seems to be created by some specific amino acid changes; proline (Pro259), conserved in MxaF, is replaced by a threonine causing small changes in backbone conformations, which in turn force the upstream, coordinating Asn256 further away from the metal. A similar role is assumed for Gly171, which replaces an alanine.

Although there is an ongoing debate on the exact mechanism of methanol oxidation by MDH, the catalytic centre is at carbon atoms C4 and C5 of the PQQ (Zhang et al., 2007; Li et al., 2011). Ca(II) is coordinated by protein carboxylate groups and three atoms of PQQ, one being the oxygen of the carbonyl at C5. It is commonly argued that Ca(II) acts as a Lewis acid and polarizes the carbonyl C5 atom of PQQ. Lanthanides (III) are much stronger Lewis acids than Ca(II) and thus are likely to render the C5 a stronger electrophile, making it easier to remove electrons from the substrate. This may explain the high rates of Ln-MDH observed at pH 7, reflecting the natural condition in the cell, without the need of ammonium activation at pH 9–11, and may be of great advantage for research on the mechanism of methanol oxidation. The high affinity and specific activity for formaldehyde of strain SolV Ln-MDH make a direct conversion of methanol to formate by Ln-MDH a likely route, without the need of additional methanopterin- or folate-dependent enzymes, which were shown to be absent or incomplete in the genome of strain SolV (Khadem et al., 2012). In addition, we could not measure NAD(P)H or glutathione-depenedent formaldehyde conversion (results not shown).

As XoxF's are found much more frequently than MxaF-type MDHs and are present in all methylotrophs (Chistoserdova, 2011), it is tempting to speculate that besides the MDHs from strain SolV (this study) and Methylobacterium extorquens AM1 (Nakagawa et al., 2012), more XoxF's are in fact lanthanide-containing high affinity MDH variants rather than low activity MDHs. This is further supported by the fact that the xox-operons lack accessory proteins required for maturation/calcium insertion (see above) and the conservation of the lanthanide-coordinating Asp301 in all XoxF branches. The occurrence of MxaF-type MDH in a phylogenetic tree as a subgroup in the XoxF-type enzymes (Chistoserdova, 2011) suggests that the latter type may be the primordial MDH.

A high affinity is crucial in environments where substrate concentrations are very low. In this respect, it is interesting to note that very high expression of the MDH-like XoxF proteins were encountered in the proteome of coastal oceanic microbial plankton (Sowell et al., 2011). Recently, methanol and other one carbon compound were demonstrated to be important substrates for bacterioplankton (OM43 clade) in coastal ecosystems (Giovannoni et al., 2008). The genome of the Methylophilaceae strain HTCC2181, a representative of this OM43 clade, was suggested to define a minimal gene set for methylotrophy and only possessed a XoxF type MDH (Giovannoni et al., 2008; Chistoserdova, 2011). Metagenome studies of Lake Washington sediment incubated with 13C-methane suggested that the abundant Methylophilaceae phylotypes possessed multiple copies of xoxF but lacked mxaFI genes (Kalyuzhnaya et al., 2008; Beck et al., 2013). These observations all point to an important role for the XoxF-type MDH in methanol oxidation under natural oligotrophic conditions.

The same very high expression of the MDH-like XoxF proteins was also reported for the phyllosphere of plants colonized by Methylobacterium spp. and Methylobacterium extorquens AM1 (Delmotte et al., 2009). Plants can bioconcentrate REE very efficiently (Weltje et al., 2002) and may provide the necessary amounts of lanthanides. In contrast, the low expression and activity of XoxF proteins in Methylobacterium under laboratory conditions are most likely caused by the lack of lanthanides in the medium used (Nakagawa et al., 2012). One might expect that the presence of Ln-MDH is not limited to methanotrophs or methylotrophs in acid geothermic locations and phyllospheres but will be more widely distributed in the environments where lanthanide concentrations are low but replenishable by sand, a practically infinite source of REE (Tyler, 2004 and references therein).

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. Author information
  8. References
  9. Supporting Information

Strain, medium and culture conditions

Methylacidiphilum fumariolicum strain SolV was originally isolated from the central mudpot of the Solfatara crater (Pozzuoli, Italy) (Pol et al., 2007). The strain was grown on a mineral medium containing (in mM): NaH2PO4, 1.0; MgCl2, 0.2; CaCl2, 0.2; Na2SO4, 1; K2SO4, 2; and (NH4)2SO4, 2; trace elements (0.1% by volume) adjusted to pH 2.7 with sulfuric acid. MPW from the Solfatara crater (Pozzuoli, Italy) was centrifuged and filtered before use, and 1–4% was added to the medium. Incubation took place under a gas phase of air with 10% CH4 and 5% CO2 at 55°C and shaking at 400 r.p.m. Large-scale (10 l) cultivation was done as described by Khadem and colleagues (2012).

When testing REE dependency (salts > 99% pure), it was observed that standard serum bottles resulted in a highly variable growth, although exponential growth phase did never exceed OD600 0.1 when no MPW or REE was added. Sand is one of the major raw materials of glass and may contain considerable amounts of REE, and Ce may be used as an additive during glass manufacturing. It was concluded that REEs in glass are extractable, at least partly, by the acidic media used. Better results were obtained with 250 ml polycarbonate Erlenmeyer flasks (Nalgene, Rochester, NY, USA), but it appeared that REE of previous usage remained attached to the inside of the these bottles as even after extensive machine washing and hot diluted acid rinsing variable background growth were observed. Best results were obtained with single use disposable polyethylene water bottles. In the screw caps, a hole was drilled in order to accommodate a red rubber septum needed for sampling. Contact of the acidic medium with needles used for sampling was minimized as the metal seems to release REE as well. For these experiments, concentrations of trace elements were (in μM): NiCl2, 1; CoCl2, 1; Na2MoO4, 1; ZnSO4, 1; FeSO4, 5; and CuSO4, 10.

Purification of the MDH

Bacterial cells were harvested at late exponential phase by centrifugation, and the pellet was suspended in twice its volume of 10 mM phosphate buffer and adjusted to pH 7.2. DNA-ase I was added (30 mg l−1), and cells were broken by two passages through a French press at 20 000 psi. Cell debris and membranes were removed by centrifugation at 30 000 × g for 1 h (Sorvall, GSA rotor at 4°C). The resulting clear supernatant, referred to as the cell-free extract, was used for purification of MDH and contained typically 30 mg ml−1 of protein. Up to 600 mg of cell-free extract protein was applied to an ice cooled SP-Sepharose-FF (GE Healthcare, Diegem, Belgium) cation exchange column (1.6 × 20 cm) (Liu et al., 2006) that was equilibrated in 10 mM sodium phosphate buffer, containing 1 mM methanol. The enzyme bound well to the column at pH values tested from pH 6 up to pH 7.5, despite a predicted pI value of 6.36. The column was washed at 2 ml min−1 with the equilibration buffer. Most of the proteins were not retained, and the absorbance (280 nm) went back to baseline after three column volumes have passed. Bound proteins were subsequently eluted with seven column volumes of a linear gradient of 0–500 mM NaCl in the same buffer. Best separation results were obtained at pH 7.2 at which MDH left the column at 190 mM NaCl as a symmetrical peak and with an activity yield of 85%. SDS-PAGE and MALDI-TOF MS analysis proved this peak to be virtually pure MDH. When methanol was not included during this procedure, very little active MDH eluted from the column and a prominent turbid peak was eluted a bit later.

Specific activity of MDH increased 25-fold during purification. This factor is in accordance with the relatively large amount of MDH present in crude extract. The apparent MDH appeared as a dominant band in native PAGE of cell-free extract (Fig. S3), which after extraction and SDS-PAGE could be identified by MALDI-TOF MS analysis of a tryptic digest. Analysis of the native purified protein on MALDI-TOF MS revealed an average mass peak of 63 502 + 176 Da (SD, n = 7) corresponding with the calculated molecular mass (63 586 Da) of the protein encoded by gene Mfum_190005 after splicing of the signal peptide.

MDH activity assay

Initially, the assay was performed at high pH according to Anthony and Zatman (1967), but at lower pH values, the buffer was changed into 100 mM Tris-HCl or 20 mM sodium phosphate. Although at higher pH values the strain SolV MDH activity was stimulated by ammonium (40 mM), utmost twofold at pH 8, activity slowed down immediately from the start. At pH 9, MDH activity lasted just for a few seconds. Both phenazine methosulphate (PMS) and phenazine ethosulphate (PES) were used as primary electron acceptor. Both exhibited half saturation around 2–3 mM. Inhibition was observed at high concentration (> 5 mM). As PES gives much less background reduction of DCPIP, it was used in most cases, routinely at 1 mM. The assay was performed in a stirred 3 ml cuvette that contained 1.2 ml preheated reagents and was placed in a Cary spectrophotometer in a thermostatted holder. After adding the enzyme, DCPIP reduction was followed at 600 nm. When a stable rate was obtained, substrate was added and DCPIP reduction followed for maximally 1 min. The obtained rate was normalized to the maximum rate that was obtained with excess methanol added afterwards. In this way, we corrected for changes in enzyme activity during the day and for small variations like temperature between assays. For DCPIP, an extinction coefficient of 21.5 cm−1 mM−1 was taken at pH 8. The extinction (600 nm) change with pH was determined in order to correct for it. The estimated temperature optimum was around 60°C. Maximum specific activity was 4 μmol mg−1 min−1 when 2 mM PES was used, whereas at PES saturation, 8 μmol mg−1 min−1 was found. Higher temperatures were not tested because of the increasing background reactions. As reported for many other MDHs, reduction of DCPIP was observed without addition of a substrate (electron donor) like methanol, and this makes the determination of kinetic parameters cumbersome (Anthony, 2000). The source of this so-called endogenous substrate is still unclear. Buffers have been mentioned, and in our case, organic buffers like HEPES, TES, MES and PIPES apparently contained high amounts and prevented the measurement of methanol-dependent DCPIP reduction. With TRIS or phosphate buffer in the assay, endogenous reduction lasted for a few (2–4) minutes only. In contrast to other reports, cyanide did not inhibit this oxidation of endogenous substrate nor did it stimulate activity, but it stabilized MDH activity in the absence of methanol. The kinetics for substrates were studied at pH 7, and temperatures were between 40 and 45°C in order to limit chemical background reduction and reoxidation rates of DCPIP, especially with PMS. The choice for these parameters considerably increased the accuracy of slow rates at very low substrate concentrations. Noteworthy to mention that affinity constantly reported here were measured at 1 mM PES; however, occasionally, 2 mM was used, which seemed to result in about two times higher affinity constants. Initially, we measured the affinity constant for methanol with PMS in the assay and measured affinity constant values of around 0.5 μM; this is close to the detection limit of a spectrophotometer, and because of the higher background rates, accuracy was low. Again higher PMS concentration resulted in a higher affinity constant. Such an effect has been reported for PES with MDH of Methylophilus methylotrophus (Hothi et al., 2003), and the authors stated that subtraction of background ‘endogenous’ activity is inappropriate. However, the latter is done in many MDH kinetic studies and has significant influence on the outcome of kinetic parameters and makes direct comparison of reported values difficult.

Crystal structure

For crystallization, the protein was concentrated to an A2801 cm of 6.3 (corresponding to ∼ 2.5 mg ml−1) by ultrafiltration in 25 mM HEPES/NaOH, pH 7.5, 25 mM KCl, 1 mM methanol. Initial crystals were obtained in 100 + 100 nl sitting drops using a reservoir solution containing 20%w/v polyethylene glycol (PEG) 8000, 0.2 M NaCl, 0.1 M N-cyclohexyl-3-aminopropanesulfonic acid (CAPS) (NaOH) pH 10.5. This condition was optimized in 1 + 1 μl hanging drops using 600 μl of 22%w/v PEG 8000, 0.2 M NaCl, without buffer, in Linbro plates. Crystals were cryoprotected by briefly soaking them in reservoir solution with 30% (v/v) ethylene glycol and then flash-cooled in liquid nitrogen. All crystallographic data were measured at the X10SA beamline (at a wavelength of 0.9718 Å and 2.01074 Å) of the Swiss Light Source at the Paul-Scherrer Institut in Villigen, Switzerland, and data were processed with XDS (Kabsch, 1988a,b; 1993; 2010a,b). The structure was solved from the 1.6 Å resolution dataset (see Table 1) by molecular replacement with PHASER (McCoy et al., 2005) using a model constructed by MolRep (Vagin and Teplyakov, 1997) from the Methylobacterium extorquens MDH structure Ghosh et al., 1995). Refinement was performed using Refmac (Murshudov et al., 1997) and PHENIX (Adams et al., 2010) with iterative rebuilding done in COOT (Emsley and Cowtan, 2004).

Riding hydrogen atoms and anisotropic B-factors for the Ce ions were used to obtain the final model. Despite the use of anisotropic B-factor refinement for the Ce ions, some residual difference density remains around the Ce ions, likely because of either the partial occupancy or the fact that the metal binding sites likely contain a mixture of lanthanides, or both.

The final, refined model shows excellent geometry, although 0.3% of residues have unfavourable Ramachandran angles as determined by MolProbity (Chen et al., 2010). However, inspection of the electron density for these residues showed no discrepancies between density and model. Structural figures were prepared using PyMol (http://www.pymol.org).

Analytical ultracentrifugation

MDH was concentrated to A2801 cm = 6.3 in 25 mM HEPES/NaOH, pH 7.5, 25 mM KCl, 1 mM methanol and investigated in a Beckman Proteomelab XL-I (Beckmann Coulter GmbH, Krefeld, Germany) analytical ultracentrifuge at 30 000 r.p.m. and 20°C in a two-sector cell with a 1.2 cm optical path length. Absorption scan data were collected at 280 nm and evaluated using SEDFIT (Schuck, 2009). Assuming a monomer with a molecular weight (MW) of 63 586 Da as calculated from the sequence, maximum possible sedimentation coefficients Smax ≈ 0.00361.MW2/3 of 12.0, 9.1 and 5.8 can be estimated for an MDH trimer, dimer and monomer respectively (Erickson, 2009). As the main maximum in the c(S) distribution of MDH was at 7.1 Svedberg, these values correspond to Smax/S values of 1.7, 1.3 and 0.8, respectively, indicating that MDH most probably occurs as a dimer in solution (Erickson, 2009). Indeed, calculations with HYDROPRO, García de la Torre and colleagues (2000) predict the sedimentation coefficient of the MDH dimer found in the crystal structure to be 7.0, which is very close to the experimental value.

Metal analysis by ICP-MS and ICP-optical emission spectrometry (OES)

Metals were analysed by ICP-MS on a Series I ICP MS (Thermo Scientific, Breda, the Netherlands). Ca measurements were done by ICP_OES on a iCap 6300 radial (Thermo Scientific). Height point calibration was performed with a dilution series of (multi-) element standards (1000 ppm in 1% nitric acid; Merck, Darmstadt, Germany). For metals present in MDH, 50–300 μl of purified MDH (9.7 mg protein ml−1 for MPW-MDH or 7.0 mg protein ml−1 for Pr-MDH) was washed with water on a Vivaspin 500 filter (Sartorius, Goettingen, Germany) and destructed with nitric acid (10–30%) at 90°C for 30–60 min and diluted to 5 ml with water.

Protein determination

Protein in cell-free extract was determined by the bicinchonic acid (BCA) Protein Assay kit (Pierce, Rockford, IL, USA), with serum albumin as a standard. The protein concentration of purified MDH was determined by its extinction at 280 nm. The molar extinction coefficient 158 cm−1 mM−1 was calculated on basis of its deduced amino acid sequence with the ‘Protein Identification and Analysis Tools’ on the ExPASy Server (http://www.expasy.ch/tools). MDH protein concentration measured by the Biorad assay was 50% higher or lower (with gamma globulin and BSA as standards respectively)

MALDI-TOF MS analysis

The molecular mass and identification of MDH in SDS-PAGE gels were determined MALDI-TOF MS on gel spots as described previously (Khadem et al., 2011).

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. Author information
  8. References
  9. Supporting Information

The work was funded by a Mosaic grant (017.005.113) of the Netherlands Organization for Scientific Research to A.F. Khadem, by an European Research Council grant (232937) to M.S.M. Jetten and by the Max-Planck Society. T.R.M. Barends wishes to thank Ilme Schlichting for continuous support and collection of diffraction data.

Author information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. Author information
  8. References
  9. Supporting Information

Atomic coordinates and structure factor amplitudes have been deposited in the PDB with accession code 4MAE.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. Author information
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. Author information
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
emi12249-sup-0001-si.pdf159K

Fig. S1. Uptake preference within the series of REE (La, Ce, Pr, Nd) during growth in MPW-supplemented (4%) medium.

Fig. S2. SDS-PAGE (12%) of the purified methanol dehydrogenase of strain SolV (A) and the UV-Vis spectrum of the pure enzyme (B).

Fig. S3. Analytical ultracentrifugation analysis of purified MDH.

Fig. S4. Effect of pH on MDH activity. Activities refer to the μM DCPIP reduced per min. Each symbol represents an individual measurement. Squares and triangles indicate measurements with and without 40 mM ammonium chloride respectively.

Fig. S5. A. Kinetics of purified MDH. Symbols indicate individual rate measurements. Lines represent the curves obtained by fitting Michaelis–Menten kinetics. Obtained affinity constants (Km) are given in Table 3.

B. Kinetics of purified MDH. Symbols indicate individual rate measurements. Lines represent the curves obtained by fitting Michaelis–Menten kinetics. Obtained affinity constants (Km) are given in Table 3.

Table S1. Crystallographic data and model quality.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.