Proteomic analysis of the thermophilic methylotroph Bacillus methanolicus MGA3

Proteomics 2014, 14, 725–737 725DOI 10.1002/pmic.201300515

RESEARCH ARTICLE

Proteomic analysis of the thermophilic methylotroph

Bacillus methanolicus MGA3

Jonas E. N. Muller1∗, Boris Litsanov1∗, Miriam Bortfeld-Miller1, Christian Trachsel2,Jonas Grossmann2, Trygve Brautaset3 and Julia A. Vorholt1

1 Institute of Microbiology, ETH Zurich, Zurich, Switzerland2 Functional Genomics Center Zurich, University of Zurich/ETH Zurich, Zurich, Switzerland3 Department of Biotechnology, SINTEF Materials and Chemistry, Trondheim, Norway

Bacillus methanolicus MGA3 is a facultative methylotroph of industrial relevance that is ableto grow on methanol as its sole source of carbon and energy. The Gram-positive bacteriumpossesses a soluble NAD+-dependent methanol dehydrogenase and assimilates formaldehydevia the ribulose monophosphate (RuMP) cycle. We used label-free quantitative proteomicsto generate reference proteome data for this bacterium and compared the proteome of B.methanolicus MGA3 on two different carbon sources (methanol and mannitol) as well astwo different growth temperatures (50�C and 37�C). From a total of approximately 1200different detected proteins, approximately 1000 of these were used for quantification. Whilethe levels of 213 proteins were significantly different at the two growth temperatures tested,the levels of 109 proteins changed significantly when cells were grown on different carbonsources. The carbon source strongly affected the synthesis of enzymes related to carbonmetabolism, and in particular, both dissimilatory and assimilatory RuMP cycle enzyme levelswere elevated during growth on methanol compared to mannitol. Our data also indicatethat B. methanolicus has a functional tricarboxylic acid cycle, the proteins of which aredifferentially regulated on mannitol and methanol. Other proteins presumed to be involved ingrowth on methanol were constitutively expressed under the different growth conditions. AllMS data have been deposited in the ProteomeXchange with the identifiers PXD000637and PXD000638 (http://proteomecentral.proteomexchange.org/dataset/PXD000637,http://proteomecentral.proteomexchange.org/dataset/PXD000638).

Keywords:

Bacillus methanolicus / Biotechnology / Methylotroph / Microbiology

Received: November 20, 2013Revised: December 21, 2013

Accepted: December 27, 2013

� Additional supporting information may be found in the online version of this article atthe publisher’s web-site

1 Introduction

Methylotrophic bacteria utilize reduced one-carbon (C1)sources such as methanol as their sole carbon and energysource [1]. They have in common an initial oxidation stepconverting methanol to formaldehyde, which is subsequentlyused for the production of energy and biomass via various

Correspondence: Professor J.A. Vorholt, Institute of Microbiology,ETH Zurich, Vladimir-Prelog-Weg 4, 8093 Zurich, SwitzerlandE-mail: [email protected]: 0041-44-633-1307

Abbreviations: MSV, minimal salt vitamin; TCA, tricarboxylic acid

pathways [2,3]. In recent years, methylotrophic bacteria havecome into focus as potential producers for biotechnologicalprocesses [4]. Compared to sugar molasses, methanol is apure, raw material, its price is comparable, and it can be gener-ated as a renewable carbon source from methane [5]. Bacillusmethanolicus is a Gram-positive, thermophilic, spore-formingbacterium that is able to grow in the presence of methanolat temperatures of up to 60�C, with optimum growth be-tween 50 and 53�C [6]. Several B. methanolicus strains areknown to naturally produce and secrete high amounts ofL-glutamate [7], and ethyl methanesulfonate-mutagenized

∗These authors contributed equally to this work.

C© 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.proteomics-journal.com

http://proteomecentral.proteomexchange.org/dataset/PXD000637

http://proteomecentral.proteomexchange.org/dataset/PXD000638

726 J. E. N. Muller et al. Proteomics 2014, 14, 725–737

cells overproducing L-lysine have been obtained [6]. Inaddition, transformation methods and expression plasmidsfor metabolic engineering have been described for this or-ganism [8–10]. All these attributes make B. methanolicus apromising candidate for white biotechnology [5].

B. methanolicus MGA3 (ATCC 53907) is a restricted methy-lotroph, i.e. is able to use a limited number of alternativecarbon sources in addition to methanol, including manni-tol and glucose [11, 12]. During methylotrophic growth, anNAD+-dependent methanol dehydrogenase (Mdh) oxidizesmethanol to formaldehyde and the ribulose monophosphate(RuMP) cycle is employed for formaldehyde assimilation[13–15]. The RuMP cycle is initiated by formaldehyde con-densation with ribulose 5-phosphate, which results in hexu-lose 6-phosphate, and is catalyzed by 3-hexulose-6-phosphatesynthase (Hps). The resulting C6 compound is isomer-ized to fructose 6-phosphate by 6-phospho-3-hexuloisomerase(Phi), which is subsequently phosphorylated to fructose-1,6-bisphosphate by phosphofructokinase (Pfk). Fructose-1,6-bisphosphate is cleaved into glyceraldehyde 3-phosphateand dihydroxyacetone phosphate by fructose-bisphosphatealdolase (Fba). Glyceraldehyde 3-phosphate and fructose 6-phosphate are metabolized in a series of reactions knownas the “rearrangement phase,” involving Pfk, transketo-lase (Tkt), Fba, ribose-phosphate 3-epimerase (Rpe), and ri-bose 5-phosphate isomerase (Rpi) to regenerate ribulose 5-phosphate as an acceptor for formaldehyde [16, 17]. Afterthree subsequent rounds of formaldehyde fixation and re-generation of ribulose 5-phosphate, a triose phosphate unit(glyceraldehyde 3-phosphate or dihydroxyacetone phosphate)remains available for biomass formation and can be metab-olized to pyruvate via lower glycolysis. Two variants of the“rearrangement” reactions of the RuMP cycle are known:a transaldolase and a sedoheptulose-1,7-bisphosphate vari-ant [17]. In the first, erythrose 4-phosphate and fructose6-phosphate are converted to sedoheptulose 7-phosphateand glyceraldehyde 3-phosphate in a transaldolase (Tal) cat-alyzed reaction, while in the second variant sedoheptulose 7-phosphate is produced from sedoheptulose-1,7-bisphosphateby sedoheptulose-1,7-bisphosphatase (GlpX); sedoheptulose-1,7-bisphosphate is a product of the Fba-catalyzed reactionof erythrose 4-phosphate and dihydroxyacetone phosphate.B. methanolicus possesses genes for enzymes of both vari-ants, and although recent data indicates that the GlpX variantis most likely used during growth on methanol [18, 19], it isstill unknown which enzymes are necessary for C5 regener-ation (see below).

For generation of energy in the form of reduction equiva-lents, methylotrophic bacteria oxidize formaldehyde. Differ-ent cofactors are known to be used as C1 carrier molecules;bound to these, C1 units can be oxidized via formate to carbondioxide in linear pathways [3]. Direct evidence for the catalyticcapacity of such a pathway in B. methanolicus was providedby the production of 13C-labeled formate from 13C-labeledmethanol and 13C-labeled CO2 from 13C-labeled formate [20].Genome sequencing revealed that genes encoding for a

tetrahydrofolate-based linear pathway are present [19]. How-ever, experimental evidence for catabolic tetrahydrofolate-dependent oxidation is still missing. In addition to a linearformaldehyde oxidation pathway, some methylotrophic bac-teria are known to possess a cyclic formaldehyde dissimila-tion variant of the RuMP pathway [1, 21], which is partiallysimilar to the general oxidative part of the pentose phos-phate pathway. The first two steps are identical to the assimi-latory RuMP cycle (see above). Then, fructose 6-phosphateis further converted into glucose 6-phosphate by glucose-6-phosphate isomerase (Pgi), and glucose 6-phosphate isoxidized by glucose-6-phosphate dehydrogenase (Zwf) to6-phospho-glucono-1,5 lactone. The latter metabolite is hy-drolyzed by 6-phosphogluconolactonase (Pgl) to generate 6-phosphogluconate, which is decarboxylated to ribulose 5-phosphate by 6-phosphogluconate dehydrogenase (Gnd). De-pending on the organism, the dehydrogenases used in thispathway can be NAD+- or NADP+-dependent [1]. For allthese enzymes, with the exception of Pgl, high activities arereported to be present in crude cell extracts of methanol-grown B. methanolicus cultures, and both dehydrogenasesuse NADP+ as a cofactor [22]. The involvement of this cyclicdissimilation pathway during methylotrophic growth thusseems likely.

Methylotrophy in B. methanolicus MGA3 is known to bedependent on a 19 kbp plasmid designated pBM19, whichencodes mdh and five RuMP cycle genes: glpX, fba, tkt,pfk, and rpe. Transcription of all these genes is upregu-lated during growth on methanol, and the loss of the plas-mid results in the loss of the ability to grow on methanol[8,19,23]. Recent genome sequencing showed that homologsfor all five RuMP cycle genes are also found on the chro-mosome, and transcriptome data revealed that only theplasmid-encoded versions are upregulated on methanol [19].In vitro analysis of some of the RuMP cycle enzymes demon-strated that they are differentially regulated and also dif-fer in their biochemical properties. The plasmid-encodedFba has a higher affinity for glycerinaldehyde-3-phosphateand dihydroxyacetone phosphate than the chromosomal-encoded enzyme, thus preferentially catalyzing the reac-tion in the gluconeogenetic direction [24]. The chromoso-mally encoded GlpX enzyme appears to be responsible forthe phosphorylation of fructose 6-phosphate to fructose-1,6-bisphosphate, while the plasmid-encoded GlpX enzymedemonstrates higher catalytic activities in the dephosphory-lation of sedoheptulose-1,7-bisphosphate to sedoheptulose 7-phosphate, indicating a role in the GlpX variant of the RuMPcycle [18].

Inspection of the B. methanolicus MGA3 genome sequencealso resulted in the identification of two additional chromo-somally encoded mdh genes coding for two enzymes, Mdh2and Mdh3, which are 96% identical to each other and share61% and 62% identity with the plasmid-encoded Mdh, re-spectively [25]. The individual roles of these three enzymesduring growth on methanol are unclear, as transcriptomeanalysis showed no upregulation on methanol, while in vitro


Proteomics 2014, 14, 725–737 727

studies with purified enzymes revealed that Mdh2 and Mdh3were at least as efficient in oxidizing methanol as Mdh [25].Activities of all three Mdh enzymes in vitro are stronglypromoted by the activator enzyme Act, which belongs tothe class of nucleoside diphosphate linked to a “moiety X”(NUDIX) type hydrolases. The stimulating effect of Act re-sults from hydrolytic cleavage of Mdh-bound NAD+ intoadenosine monophosphate and nicotinamide ribonucleotide(NMN), resulting in an activated form of Mdh [13, 15, 25, 26].The biological impact of Act in B. methanolicus methylotro-phy remains unclear. RuMP cycle-utilizing methylotrophshave often been described as lacking a complete tricarboxylicacid (TCA) cycle [27], but B. methanolicus possesses genesthat putatively encode enzymes comprising a complete TCAcycle [19].

In this study, we used label-free quantitative proteomics toanalyze the adaptation of B. methanolicus to different growthregimes in more detail. We focused on the differences de-tected between the use of methanol (methylotrophic growth)versus mannitol (nonmethylotrophic growth), and we pro-vide further evidence that B. methanolicus only uses one of itsthree Mdh enzymes and the GlpX-variant of the RuMP cycleduring growth on methanol. We also investigated how theproteome changes upon temperature shifts, since the culti-vation temperature has impact on the industrial applicationof this organism with respect to cooling costs.

2 Materials and methods

2.1 Strain and cultivation conditions

B. methanolicus MGA3 strain was used throughout thestudy and was routinely cultivated in minimal saltvitamin (MSV) medium modified after [6]: K2HPO4,0.49 g/L; NaH2PO4, 0.14 g/L; (NH4)2SO4, 0.21 g/L.Trace elements were introduced into the minimal saltmedium as a stock solution with final concentrations of:FeSO4·7H2O, 2 mg/L; CuSO4·5H2O 40 �g/L; CaCl2·2H2O,5.3 mg/L; CoCl2·6H2O, 40 �g/L; MnCl2·H2O, 200 �g/L;ZnSO4·7H2O, 200 �g/L; Na2MoO4·2H2O, 48 �g/L; H3BO3,31 �g/L. To complete the MSV medium, vitamin stocksolution was added prior to use, resulting in thefollowing final concentrations: thiamine hydrochloride,100 �g/L; riboflavin, 100 �g/L; pyridoxine hydrochloride,100 �g/L; pantothenate, 100 �g/L; nicotinamide, 100 �g/L;p-aminobenzoic acid, 20 �g/L; folic acid, 10 �g/L; cobal-amine, 10 �g/L; lipoic acid, 10 �g/L. All chemicals wereobtained from Sigma-Aldrich, Buchs, Switzerland. Startercultures for the inoculation of bioreactors were inoculated di-rectly from frozen glycerol stocks into 500-mL baffled shakeflasks containing 100 mL MSV medium supplemented witheither 200 mM methanol or 55 mM mannitol as a carbonsource. The starter cultures were grown at temperatures of37�C or 50�C. Prior to inoculation, the media was incubatedfor 1 h at the temperature used for growth.

2.2 Growth conditions for proteome sampling

B. methanolicus MGA3 cultures for proteome sampling werecultivated in 0.6 L bioreactors (Multifors Multi-FermenterSystem, Infors AG, Bottmingen, Switzerland) at 50�C and37�C. Each bioreactor contained 300 mL MSV medium with200 mM methanol or 55 mM mannitol as the sole car-bon source. Bioreactor aeration was conducted with an airflow of 0.3 L/min. Oxygen saturation was measured contin-uously with a polarimetric oxygen electrode (Mettler Toledo,Giessen, Germany) and remained constantly above 80%. ThepH was determined online using a standard pH electrode(Mettler Toledo) and adjusted to pH 7.2 with 1 M ammonia.The B. methanolicus bioreactor cultures were inoculated at anOD of approximately 0.15 from corresponding flask startercultures grown with the same carbon source and at the sametemperature. Culture samples for proteome analysis wereharvested from exponentially growing cells at an OD of 1.5.Three independent fermentations with corresponding pro-teome sampling were performed for each condition tested.B. methanolicus exhibited similar growth rates on mannitoland methanol in MSV medium at 50�C, resulting in growthrates of 0.44 h−1 and 0.40 h−1, respectively, whereas at 37�Ccells grown in methanol-supplemented MSV medium exhib-ited a growth rate of 0.14 h−1.

2.3 Preparation of proteins for MS

Bacterial cell suspensions (50 mL) from the bioreactors werepelleted at 8000 × g for 5 min in 50 mL tubes. After resuspen-sion of the cells in 2 mL TE-buffer (10 mM Tris, 1 mM EDTA,pH 7.5) containing Proteinase Inhibitor Cocktail (cOmplete,EDTA-free, Roche Applied Science), cells were lysed threetimes in a French pressure Mini-cell under 21 700 psiinternal cell pressure. Cell debris was removed by centrifu-gation at 4�C for 30 min at 20 800 × g (Eppendorf cen-trifuge). The supernatants were transferred into fresh 2 mLtubes. The proteins were mixed with Laemmli sample buffer(Bio-Rad) and denatured for 10 min at 95�C. Proteins (30 �gper lane) were loaded onto a NuPAGE Bis-Tris polyacrylamidegel (4–12% linear gradient, 8 cm × 8 cm, 1.0 mm, ten wells,Life Technologies). Electrophoresis was performed for 2 h15 min at 100 V in NuPAGE buffer (Life Technologies). Sub-sequently, the gel was incubated for 30 min in fixing solution(20% methanol, 1% phosphoric acid) and stained overnightin Roti-Blue Staining (Roth). The gel was transferred to aclean box and incubated for 30 min in washing solution (25%methanol, 75% water). For each sample, the correspondinggel lane was cut into seven pieces. Gel pieces were washedthree times in 50% ACN and dried for 10 min under vacuum(SpeedVac, Thermo Fisher Scientific). Proteins were reducedfor 45 min at 56�C in 10 mM DTT (in 25 mM ammoniumhydrogen carbonate, pH 8.0) and carbamidomethylated for60 min at room temperature in the dark with 50 mM iodoac-etamide (in 25 mM ammonium hydrogen carbonate, pH 8.0).



After quenching the reactions with 50 mM DTT gel pieceswere washed twice in 50% ACN and dried for 15 min un-der vacuum. The proteins of each gel piece were digested byadding 50 ng trypsin (Promega) for 16 h at 37�C in 25 mMammonium hydrogen carbonate, pH 8.0. To stop the digesttrifluoroacetic acid (5% in 50% ACN) was added until pH wasbelow 2. Peptides were transferred to fresh tubes and solventswere evaporated. Finally, peptides were resuspended in 20 �L3% ACN and 0.1% formic acid, and desalted with C18 ZipTip(Millipore Corporation).

2.4 MS analysis

MS-based shotgun analysis was performed on a LTQ Orbi-trap XL instrument (ThermoScientific, Bremen, Germany)coupled to an Eksigent LC system (Eksigen, Redwood City,USA). Solvent compositions of the two channels were 0.1%formic acid in H2O for solvent A and 0.1% formic acid in ACNfor solvent B (both solvents from Brunschwig Chemie AG,Basel, Switzerland). Peptides were resuspendend in 20 �L3% ACN and 0.1% formic acid, and 4 �L were loaded di-rectly onto an in-house (FGCZ) packed reverse phase col-umn (75 �m, 15 cm, ProntoSIL C-18 AQ, 3 �m, 200 A,Bischoff Chromatography, Leonberg, Germany) with a flowrate of 500 nL/min. Peptides were separated with a flow rateof 300 nL/min and a gradient from 5 to 9% solvent B in 1min followed by 9–40% solvent B over 55 min. Solvent B wasincreased to 95% in 2 min and the column was cleaned for15 min followed by a reconditioning step at 5% solvent Bfor 8 min prior to the next injection. Peptides were ionizedby electro spray using a spray voltage of 1.8 kV and the iontransfer tube was operated at 200�C. MS scans were acquiredin the Orbitrap from 300 to 1800 m/z with a resolution of30 000 (at 400 m/z) and an automated gain control tartgetvalue of 5e6 or a maximum injection time of 500 m s. Fromeach MS scan, the five most intense ions were selected forcollision-induced decay fragmentation. The minimum pre-cursor signal intensity threshold was set to 500 and chargestate screening was rejecting singly charged ions and ionswith an undetermined charge state. Precursors were isolatedwith a 2 Dalton window and accumulated to a target value of1e5 or a maximum injection time of 100 ms. MS/MS spec-tra were recorded in the linear trap after activation of theprecursor with normalized collision energy of 35% (activa-tion Q = 0.25, activation time 30 ms). Each ion was put for60 s onto the dynamic exclusion list (maximum of 500 entries)after one fragmentation event with a window of 20 ppm.

2.5 Experimental sequence

To optimize MS measurements and to prevent biased read-outs, the samples were analyzed in a mixed order followingthe loading pattern on the gel. Three biological replicateswere always loaded on the gel in the following order: man-nitol 50�C, methanol 50�C, and methanol 37�C. The MS-MS

analysis of the samples started with all slices of the lowestmolecular weight (least complex) followed by the next nineslices of higher molecular weight. Additionally, for quantifica-tion with Progenesis LCMS a pooled sample (composed of analiquot from each of the nine previously analyzed samples)was included at the end of each batch of samples with thesame molecular weight that served as the aligning reference.

2.6 Protein identification and quantification using

Progenesis LCMS

Raw data of gel slices with equal molecular weight wereloaded together into the software package Progenesis LCMSVersion. 4.1 (www.nonlinear.com), a software tool developedfor label-free quantification of LC-MS data. Data were an-alyzed according to Progenesis LCMS analysis in Weisseret al. [28]. In brief, for data loading, the option “High MassAccuracy Instrument” was selected. LC-MS data were normal-ized and aligned according to manufacturer’s specifications.In the aligning step manual seeding of “three to five” vectorsalong the retention time gradient was performed followed byautomatic alignment.

For the identification of peptide features in the corre-sponding MS files, Mascot generic files (.mgf file format)are generated with Progenesis LCMS (using up to five tan-dem mass spectra for each feature with the top 200 fragmention peaks, charge deconvolution, and deisotopting activated).Mgf files were searched by Mascot 2.3 search engine againstan amino acid database containing 20 399 entries including3418 MGA3 annotated proteins (downloaded from Uniproton August 2013), 6651 yeast proteins, 261 known MS con-taminants as forward entries and from all forward entriesexcept the contaminants, a reverse decoy sequence. Parame-ters for precursor tolerance and fragment ion tolerance wereset to ± 10 ppm and ± 0.6 Da, respectively. The search pa-rameters were set to one missed cleavage for trypsin, fixedmodifications (carbamidomethylation at cysteine), and vari-able modifications (oxidation at methionine).

Mascot results were loaded into Scaffold v4.05, using theoptions for protein clustering and high mass accuracy Pep-tideProphet scoring. A spectrum report from Scaffold wasreimported into Progenesis LCMS.

Quantification was performed on nonconflicting featuresonly and grouping of proteins was applied. For quantification,proteins were only accepted when two peptides with a peptideprophet score above 80% were identified.

Seven Progenesis experiments were first performed indi-vidually for each molecular weight fraction and afterwardscombined within the software using the “combine ana-lyzed fractions” approach. Identified peptide features weresummed into protein volumes. Statistical testing (T-test)was done on normalized, arcsinh transformed protein vol-umes, as suggested by Progenesis LCMS. Criteria for signif-icant changes on protein levels were p-values < 0.05, foldchange > 2.


http://www.nonlinear.com

Proteomics 2014, 14, 725–737 729

2.7 Preparation of crude extracts from

B. methanolicus MGA3 and enzyme assays

Crude extracts of B. methanolicus cells were prepared as de-scribed for the proteome samples, with the difference thatthe cell pellets were resuspended in 50 mM Tris/HCl, pH 8,containing Proteinase Inhibitor Cocktail (cOmplete, EDTA-free, Roche Applied Science) before cell disruption. After celldisruption, the cell debris was removed by centrifugation at50 000 × g for 1 h at 4�C. The resulting supernatant was usedas crude lysate in enzyme assays. Methanol dehydrogenaseactivity was determined as reported before [25]. The activ-ity of Hps was determined in a discontinuous assay basedon the formation of 3-hexulose-6-phosphate from formalde-hyde and ribulose 5-phosphate [29]. The Hps assay containedthe following final concentrations: 50 mM potassium phos-phate buffer, pH 7.4, 5 mM MgCl2, 5 mM formaldehyde,5 mM ribose-5-phopshate, and 2 U of phosphoriboisomerase(Sigma-Aldrich). The reaction mixture was incubated at 37�Cfor 15 min and was started by the addition of 10 �L crudelysate. For each sample, a negative control reaction was per-formed containing the same volume of ddH2O instead ofribose 5-phopshate. At time point zero, a sample (100 �L)was withdrawn and added to 100 �L 0.1 M HCl solution thatwas kept cold to stop any enzymatic activity. Subsequently,a new sample was taken every 60 s. After collecting all sam-ples, protein precipitates were pelleted by centrifugation at16 000 × g and 4�C for 10 min, and 100 �L of the supernatantwas mixed with 400 �L of ddH2O and used to determinethe formaldehyde concentration. As an assay read-out, thedecreased formaldehyde concentration was determined us-ing Nash reagent (2 M ammonium acetate, 0.05 M aceticacid, 0.02 M acetylacetone) [30]. Briefly, the Nash reagentwas mixed in a 1:1 ratio with formaldehyde-containing sam-ples and incubated for 5 min at 58�C to form yellow di-acetyldihydroutidin (DDL), whose concentration was photo-metrically detected at 412 nm. The amount of DDL formedwas quantified based on external standards created by react-ing Nash reagent with samples containing known formalde-hyde concentrations. The absorption of blank samples wassubtracted from the absorption of the samples containingribose 5-phosphate. Decreases in formaldehyde concentra-tion were plotted against time, and specific enzyme activitiesfor samples were determined in U/mg total protein, where1 U was defined as 1 �mol formaldehyde consumed per 1minute. The activity of Phi was determined in a directionthat opposed the physiological reaction by measuring the for-mation of formaldehyde from fructose 6-phosphate with acoupled discontinuous enzyme assay with functional Hps(demonstrated with the corresponding assay). The Phi assaycontained a final concentration of 50 mM potassium phos-phate buffer at pH 7.4, 5 mM MgCl2, and 50 mM fructose6-phosphate (Sigma-Aldrich). The reaction was incubated at37�C for 10 min and was started by the addition of 10 �L crudelysate. For each sample, a blank reaction containing the samevolume of ddH2O instead of fructose 6-phosphate was also

performed in parallel. Sampling and determining assay read-outs were performed as described above for the Hps assay.The absorption of the blank samples was subtracted from theabsorption of the samples containing fructose 6-phosphate.Increases in formaldehyde concentration were plotted againsttime, and the specific enzyme activities for the samples weredetermined as U/mg total protein, where 1 U was defined as1 �mol formaldehyde produced per 1 minute.

3 Results and discussion

3.1 Analysis of differential proteomes

Here, we established for the first time a proteomic analy-sis of B. methanolicus MGA3, and we used this approachto compare cells grown on methanol and mannitol at 50�Cas well as cells grown on methanol at 37�C and 50�C.B. methanolicus growth rates on mannitol and methanol inMSV medium at 50�C were similar, which allowed us toconduct proteomic comparisons of the metabolism of thesecarbon sources without any observed growth rate-induceddifferences (see also [8]). These two growth conditions repre-sent a basis to study regulatory, biochemical, and biotechno-logical characteristics of this organism. To demonstrate re-producibility and for statistical reasons, we performed threeindependent biological replicates for each growth condition.When the growth temperature of B. methanolicus MGA3 isshifted from 50 to 37�C, sporulation is initiated, while vegeta-tive growth is maintained when the cells were cultivated at aconstant temperature of 37�C [6]. These two growth tempera-tures therefore represented a second proteomic comparativeanalysis of interest in this study. A total of 1244 differentproteins were detected in the three independent biologicalreplicates for each of the three different growth conditions.Pairwise comparisons resulted in the detection of 1189 differ-ent proteins in cell cultures growing on methanol and manni-tol at 50�C, while 1140 different proteins were detected in cellcultures growing on methanol at 37�C and 50�C (SupportingInformation Tables 1 and 2). In both cases, the detected pro-teins constituted approximately one-third of the total numberof predicted protein coding sequences in this organism [19].After filtering the protein lists (a minimum of two peptidesper protein), 1048 proteins were quantified with Progenesis(see Section 2). When comparing methanol and mannitol,109 proteins exhibited significantly altered levels (fold change> 2, T-test < 0.05) (Fig. 1A). The same procedure was appliedto compare growth at 37�C and 50�C on methanol, and ofthe 1000 proteins that passed through the filters, 213 exhib-ited significant regulation (fold change > 2, T-test < 0.05)(Fig. 1A).

Proteins that were significantly affected were cate-gorized into functional groups according to the Ky-oto Encyclopedia of Genes and Genomes classification(http://www.genome.jp/kegg/). Depending on the growthcomparison, the relative contributions of the Kyoto


http://www.genome.jp/kegg/


Figure 1. Overview of detected proteins and their distribution according to cellular functions. The number of detected, quantified, andsignificantly affected proteins is shown in (A). Kyoto Encyclopedia of Genes and Genomes (KEGG) functional classification of significantlyaffected proteins is shown in (B). Distribution of significantly affected metabolic proteins is shown in (C). Central metabolism includes theRuMP cycle, glycolysis, and TCA cycle proteins.

Encyclopedia of Genes and Genomes major categories weredifferent, as summarized in Fig. 1B. A total of 62% of all thesignificantly affected proteins identified on different growthsubstrates had a function in cell metabolism, while 49% of thesignificantly affected proteins identified at different growthtemperatures belonged to the same group. The relative distri-bution of affected proteins classified as cellular processes andsignaling and genetic information processing were similar in thegrowth substrate and the growth temperature comparisons.In contrast, the proportion of functionally unknown proteinsaffected was 2.5-fold higher in the growth temperature com-parison versus the growth substrate comparison.

Proteins functionally categorized within metabolism(Fig. 1B) accounted for the largest fraction in both of thepairwise comparisons, and these proteins were therefore fur-ther classified (Fig. 1C). In particular, we noted a high pro-portion of affected proteins involved in central metabolism(33%), amino acid metabolism (18%), and cofactor/vitaminmetabolism (10%) in the growth substrate comparison. Forthe growth temperature comparison, proteins involved inamino acid metabolism (29%), cofactor/vitamin metabolism(20%), and lipid metabolism (11%) constituted more than halfof the significantly affected proteins. In both cases, proteinsnot assigned to one of the fractions mentioned constituted21% and 25%, respectively.

3.2 Influence of different growth substrates on

enzymes involved in central carbon metabolism

As previously mentioned, the proteome of B. methano-licus differed when the organism was grown either onmethanol or mannitol, particularly with respect to proteinsinvolved in central carbon metabolism (Fig. 1C). In agree-ment with previous transcriptome data [19], proteins com-prising the complete RuMP cycle, with the exception ofribose 5-phosphate isomerase (RpiB), were significantly af-fected (upregulated) in cells growing on methanol (Fig. 2).In the case of pBM19 plasmid-encoded and chromosomallyencoded paralogs, only the plasmid-encoded versions wereupregulated (Table 1). B. methanolicus has only one chro-mosomally encoded RpiB, and this protein was present atsimilar levels for both growth substrates. It therefore ap-pears feasible to assume that the catalytic activity of thisenzyme is high enough to support growth on methanol with-out further upregulation. As previously reported, a strongupregulation of fructose-1,6-bisphosphatase/sedoheptulose-1,7-bisphosphatase (GlpX) was observed, while the transal-dolase I3EBM5 was not among the significantly affected pro-teins. Together, these findings support the assumption thatB. methanolicus uses the sedoheptulose-1,7-bisphosphatasevariant of the RuMP cycle rather than the transaldolase


Proteomics 2014, 14, 725–737 731

Table 1. Regulation of proteins involved in the central carbon metabolism of Bacillus methanolicus during growth on methanol versusmannitol

Common name Abbreviation Uniprotaccessionnumber

Locustag

Log2foldchange

T-testp-value

Methanol oxidation (MO)

Methanol dehydrogenase Mdh I3DTM5 MGA3_17392 0.5 0.28Methanol dehydrogenase 3 (2) Mdh3 (2) I3E2P9 MGA3_10725 −1.5 <0.05Formaldehyde oxidation I (FOI)

Glucose-6-phosphate isomerase Pgi I3DUN2 MGA3_16421 −0.1 0.37Glucose-6-phosphate 1-dehydrogenase 2 Zwf2 I3DZR1 MGA3_15311 3 <0.05Glucose-6-phosphate 1-dehydrogenase 1 Zwf1 I3EA81 MGA3_09280 −0.4 0.176-Phosphogluconolactonase Pgl I3E313 MGA3_11305 −1.2 <0.056-Phosphogluconate dehydrogenase Gnd I3EA83 MGA3_09290 −0.1 0.78Formaldehyde oxidation V (FOV)

Bifunctional protein FolD FolD I3EAB7 MGA3_09460 −0.3 0.26Formate-tetrahydrofolate ligase Fhs I3E9N7 MGA3_08300 −0.4 0.31Putative formate dehydrogenase Fdh I3E8Y3 MGA3_07000 −0.6 0.24Formate dehydrogenase alpha chain FdhA I3E8Q8 MGA3_06625 −0.9 0.07Ribulose monophosphate (RuMP) cycle

3-Hexulose-6-phosphate synthase Hps I3DZR0 MGA3_15306 1.4 <0.056-Phospho-3-hexuloisomerase Phi I3DZQ9 MGA3_15301 1.3 <0.05Phosphofructokinase Pfk I3DTN8 MGA3_17457 3.4 <0.05Phosphofructokinase 2 Pfk2 I3ECJ8 MGA3_03000 0.3 <0.05Fructose-bisphosphate aldolase Fba I3DTM2 MGA3_17377 3.8 <0.05Fructose-bisphosphate aldolase 2 Fba2 I3EBM6 MGA3_01355 −0.04 0.77Transketolase Tkt I3DTN9 MGA3_17462 3.7 <0.05Transketolase 2 Tkt2 I3DZN5 MGA3_15171 −0.1 0.68Fructose-1,6-bisphosphatase GlpX I3DTM3 MGA3_17382 3.6 <0.05Ribose-phosphate 3-epimerase Rpe I3DTN4 MGA3_17437 2.8 <0.05Ribose 5-phosphate isomerase RpiB I3EBL1 MGA3_01280 0 0.97Glycolysis (EMP)

Triosephosphate isomerase TpiA I3DUA5 MGA3_15786 −0.6 <0.05Glyceraldehyde-3-phosphate dehydrogenase Gap I3DUA3 MGA3_15776 −0.8 <0.05Phosphoglycerate kinase Pgk I3DUA4 MGA3_15781 −0.3 0.132,3-Bisphosphoglycerate-independent phosphoglycerate

mutaseGpmI I3DUA6 MGA3_15791 −0.6 0.17

Enolase Eno I3DUA7 MGA3_15796 −0.3 0.06Pyruvate kinase Pyk I3ECJ9 MGA3_03005 0.3 0.27Pyruvate dehydrogenase E1 component alpha subunit PdhA I3DYU2 MGA3_13696 −1.3 <0.05Tricarboxylic acid (TCA) cycle

Citrate synthase CitY I3ECK3 MGA3_03025 −0.4 <0.05Aconitate hydratase AcnA I3DU49 MGA3_16693 −0.4 0.44Isocitrate dehydrogenase Icd I3ECK4 MGA3_03030 −1.1 <0.052-Oxoglutarate dehydrogenase E1 component OdhA I3E8K3 MGA3_06350 −2 <0.05Succinyl-CoA ligase subunit beta SucC I3DZ97 MGA3_14471 −1.2 <0.05Succinyl-CoA ligase subunit alpha SucD I3DZ98 MGA3_14476 −1.3 <0.05Succinate dehydrogenase flavoprotein subunit SdhA I3ECP9 MGA3_03265 −1.8 <0.05Fumarate hydratase FumA I3E2Q7 MGA3_10765 −1 <0.05Malate dehydrogenase CitH I3ECK5 MGA3_03035 −0.6 0.14Mannitol metabolism (MAN)

Phosphotransferase system (PTS) mannitol-specific IIAcomponent

MtlF I3E3X1 MGA3_12920 −3.8 <0.05

Phosphotransferase system (PTS) mannitol-specific IIBCcomponent

MtlA I3E3X3 MGA3_12930 −4.1 <0.05

Mannitol-1-phosphate 5-dehydrogenase MtlD I3E3X0 MGA3_12915 −4.1 <0.05

Proteins are listed according to their respective pathway (according to the Kyoto Encyclopedia of Genes and Genomes (KEGG) database).The fold change is given with respect to growth on methanol.



Figure 2. An integrated view of central carbon metabolism in Bacillus methanolicus MGA3. The reactions are balanced with regard to theRuMP cycle to demonstrate how three formaldehyde molecules result in the formation of one molecule of pyruvate. Green arrows indicatesignificant upregulation of proteins on methanol. Red arrows correspond to enzymes downregulated on methanol and thus upregulatedon mannitol. If more than one protein is predicted to catalyze a reaction, the regulated protein is also indicated by color-coding. All proteinsshown were detected. Fold changes and p-values for the presented proteins are shown in Table 1. Pathways: MO: methanol oxidation;FOV: formaldehyde oxidation V; FOI: formaldehyde oxidation I; RuMP: ribulose monophosphate cycle; MAN: mannitol metabolism; EMP:glycolysis; TCA: tricarboxylic acid cycle. Proteins: Mdh: methanol dehydrogenase; FolD: bifunctional methylene tetrahydrofolate dehydro-genase FolD; Fhs: formate-tetrahydrofolate ligase; Fdh: putative formate dehydrogenase; FdhA: formate dehydrogenase alpha chain; Hps:3-hexulose-6-phosphate synthase; Phi: phosphohexuloisomerase; Pfk: phosphofructokinase; Fba: fructose-bisphosphate aldolase; GlpX:fructose-1,6-bisphosphatase; Tkt: transketolase; Rpe: ribose-phosphate 3-epimerase; RpiB: ribose 5-phosphate isomerase; Pgi: glucose-6-phosphate isomerase; Zwf: glucose-6-phosphate 1-dehydrogenase; Pgl: 6-phosphogluconolactonase; Gnd: 6-phosphogluconate dehydro-genase; MtlF: phosphotransferase system (PTS) mannitol-specific IIA component; MtlA: phosphotransferase system (PTS) mannitol-specificIIBC component; MtlD: mannitol-1-phosphate 5-dehydrogenase; Tpi: triosephosphate isomerase; Gap: glyceraldehyde-3-phosphate dehy-drogenase; Pgk: phosphoglycerate kinase; GpmI: 2,3-bisphosphate-independent phosphoglycerate mutase; Eno: enolase; Pyk: pyruvatekinase; PdhA: pyruvate dehydrogenase E1 component alpha subunit; CitY: citrate synthase; AcnA: aconitate hydratase; Icd: isocitratedehydrogenase; OdhA: 2-oxoglutarate dehydrogenase E1 component; SucC: succinyl-CoA ligase subunit beta; SucD: succinyl-CoA ligasesubunit alpha; SdhA: succinate dehydrogenase flavoprotein subunit; CitH: malate dehydrogenase. Names and abbreviations are givenaccording to the Uniprot database (http://www.uniprot.org).

variant [19, 24]. Among the three methanol dehydrogenaseparalogs, Mdh was present at roughly similar amounts uponthe two growth conditions. Due to their high sequence sim-ilarity, 96% of Mdh2 and Mdh3 shared most of the detectedpeptides. Notably, MS analysis detected one unique peptidefor Mdh3 while all the other detected peptides may origi-

nate from either one of the two enzymes. In consequence,the presence of Mdh3 could be demonstrated with the ex-perimental approach used in this paper while the presenceof Mdh2 could not be unambiguously shown. The Hps andPhi proteins are encoded by a bicistronic operon, and theywere found to be upregulated by approximately 2.5-fold on


http://www.uniprot.org

Proteomics 2014, 14, 725–737 733

methanol (Table 1). In addition, we confirmed the constitu-tive synthesis of the key enzymes Mdh, Hps, and Phi by mea-suring their activities in crude cell lysates of B. methanolicusMGA3 grown on methanol or mannitol. Methanol dehydro-genase activity was found to be 210 mU/mg on methanol,while it was 120 mU/mg on mannitol. Hps activity was15 U/mg and 17 U/mg, and Phi activity was 1.2 U/mg and1.0 U/mg on methanol and mannitol, respectively. In agree-ment with previous transcriptional data, the methanol de-hydrogenase activator protein Act was not among the regu-lated proteins [8, 19]. As the activation mechanism of Act isbased on its activity as a NUDIX-hydrolase [13, 15, 26], wealso investigated whether other potential NUDIX-hydrolaseswere among the significantly upregulated proteins. Al-though two other potential NUDIX-hydrolases, annotated asNUDIX hydrolase (I3E862) and MutT/NUDIX family protein(I3DUK9), were detected, neither of them was significantlyregulated.

The enzymes required for the proposed linear dissimi-latory tetrahydrofolate pathway, FolD, Fhs, and FdhA in B.methanolicus [19], were all detected but were not upregulatedon methanol (Table 1), indicating that the pathway is eitherimportant on both substrates (tetrahydrofolate can act as aC1 carrier for biosynthetic reactions including purine biosyn-thesis) or alternatively it has no specific role in formaldehydeoxidation during growth on methanol. The latter would besurprising, as the linear dissimilation of formaldehyde is re-ported to be important for energy production and formalde-hyde detoxification during growth on methanol [2]. An al-ternative pathway for the production of energy in form ofreduction equivalents from formaldehyde is the dissimilatorycyclic RuMP pathway via 6-phosphogluconate [2]. Notably,we found here that glucose-6-phosphate 1-dehydrogenase 2(Zwf2) was upregulated by approximately eightfold, while theremaining enzymes of this pathway were not or in the caseof Pgl were even slightly downregulated on methanol. Zwf2might thus represent a key enzyme for the regulation of theflux through this pathway, and one might speculate that thecyclic rather than the linear dissimilation pathway is usedfor energy production in B. methanolicus during growth onmethanol.

All mannitol metabolism-related enzymes, includingthe components IIA and IIBC of the mannitol-specificphosphotransferase system and mannitol-1-phosphate 5-dehydrogenase, were present at much higher levels on man-nitol (Table 1). In addition, the transcriptional regulator MtlR(I3E3×2) was found to be more highly expressed by 19-foldon mannitol. All these enzymes are known to be essentialfor growth on mannitol in other organisms [31], and theirregulation was thus expected.

The enzymes involved in glycolysis and gluconeogenesiswere all detected and found to be present at similar levelson mannitol and methanol (Table 1). Notably, however, Fbashould function both in the RuMP cycle and in glycolysis. Thereaction it catalyzes represents a branching point, at whichit is determined whether a C3 intermediate is entering gly-

colysis or will be used for regeneration of a C5 precursor forformaldehyde fixation via the RuMP cycle. Interestingly, thepBM19-encoded version of Fba, and not the chromosomalvariant, was upregulated on methanol. Compared with thechromosomal version of the enzyme, it had a lower affinitytoward fructose-1,6-bisphosphate as well as lower glycolyticactivity, but it exhibited higher efficiency for the gluconeo-genetic reaction [18]. Thus, it was assumed to preferentiallycatalyze the formation of fructose-1,6-bisphosphate [18]. Onemight argue that only this Fba is upregulated on methanolto counteract the loss of C3 compounds and maintain fluxthrough the RuMP cycle.

On both growth substrates, all enzymes of the TCA cyclewere detected, indicating that B. methanolicus does not onlypossess all the genes necessary for a complete TCA cycle butalso expresses these genes during growth on methanol. In-terestingly, four of the TCA cycle enzymes, Icd, Suc, SdhA,and OdhA, were upregulated (two- to fourfold) on manni-tol. These enzymes represent four of five energy-producingreactions of the TCA cycle with only malate dehydrogenasemissing. In contrast to methanol growth, energy from man-nitol is produced via the TCA cycle with these enzymes; thismight permit higher flux into this pathway compared to whencells consume methanol. Whether the TCA cycle is operatingas a full cycle during growth on methanol or is only used forcataplerosis remains to be elucidated.

3.3 Influence of growth substrates on the synthesis

of proteins required for L-glutamate

biosynthesis

It has been reported that cultures of B. methanolicus secreteup to 60 g/L of L-glutamate during batch-fed methanol fer-mentation [7], while no analogous data have been publishedfor cells growing on mannitol. For future metabolic engineer-ing approaches, it might be interesting to determine whethergrowth conditions have any influence on the synthesis of pro-teins required for glutamate biosynthesis. L-glutamate is syn-thesized from the precursor metabolite 2-oxoglutarate, whichis an intermediate of the TCA cycle, and the high productionof L-glutamate by B. methanolicus indicates that a substantialfraction of pyruvate enters the first reactions of the TCA cycle.Levels of the relevant enzymes were not affected by the growthsubstrate, as mentioned above (Fig. 2). 2-Oxoglutarate canbe transaminated to L-glutamate or alternatively oxidized by2-oxoglutarate dehydrogenase (Odh) and further metabolizedin the TCA cycle. Odh activity during growth on methanol hasbeen reported to be low in B. methanolicus, and it has beenspeculated that this is the main reason for the high produc-tion of L-glutamate [7]. Here, we found that Odh levels wereapproximately four times higher when B. methanolicus wasgrown on mannitol compared to methanol. This might causedecreased L-glutamate production compared to cell growthon methanol. Glutamate is produced from 2-oxoglutarate viareactions catalyzed by glutamate dehydrogenase or glutamate



Table 2. Top 25 most upregulated proteins of B. methanolicus during methylotrophic growth at 37�C compared to 50�C

Common name Uniprotaccessionnumber

Locustag

Log2 foldchange

Stage VI sporulation protein D I3ECT2 MGA3_03430 InfinityAnti-sigma F factor antagonist I3EA49 MGA3_09120 15.5Uncharacterized protein I3E915 MGA3_07170 9ABC transporter related protein I3EBR7 MGA3_01560 7.4Catalase I3E7W8 MGA3_05110 6Cytochrome c551 I3EAX9 MGA3_00060 5.6Uncharacterized protein I3E8C9 MGA3_05955 5.6Sulfate ABC transporter I3E882 MGA3_05710 4.950S ribosomal protein L17 I3E403 MGA3_13090 4.7Serine protein kinase I3E783 MGA3_03870 4.4Putative nicotinate phosphoribosyltransferase I3E8V9 MGA3_06880 4.3Nonspecific serine/threonine protein kinase I3E3R7 MGA3_12650 4.1Dihydroneopterin aldolase I3EC62 MGA3_02285 4Glycerophosphoryl diester phosphodiesterase family protein I3EAA7 MGA3_09410 3.9Acetaldehyde dehydrogenase (Acetylating) I3E349 MGA3_11515 3.95-Methyltetrahydropteroyltriglutamate-homocysteine methyltransferase I3E8N9 MGA3_06530 3.9Lon protease I3ECS3 MGA3_03385 3.8DNA primase, Toprim domain protein I3DUG0 MGA3_16061 4Aldehyde dehydrogenase I3E8H9 MGA3_06215 3.7Uncharacterized protein I3DZR9 MGA3_15351 3.7Luciferase-like monooxygenase I3E8B1 MGA3_05865 3.6Uracil permease I3DZ41 MGA3_14191 3.6Carbon-monoxide dehydrogenase large subunit I3E8H4 MGA3_06190 3.6Stage IV sporulation protein A I3E9Y8 MGA3_08815 3.5Acyl-CoA dehydrogenase I3E8B2 MGA3_05870 3.5

Proteins are listed based on their fold change levels. Fold change designated as “Infinity” indicates that the protein was only detectedduring growth at 37�C (T-test p-value < 0.05).

synthase. In addition, it can be produced by deamination ofglutamine by the reverse reaction of glutamine synthetase.All three enzymes have been detected in B. methanolicus,but glutamate dehydrogenase (I3EAII) and glutamine syn-thetase (I3DZK5) were found here to be present at similarlevels for both growth substrates. Glutamate synthase con-sists of a small (I3E3J7) and a large (I3E3J8) subunit, andthe large subunit was slightly upregulated by 2.3-fold duringgrowth on mannitol. It is unknown which of these pathwaysis employed by B. methanolicus for L-glutamate production.However, recent findings have demonstrated that overexpres-sion of glutamate synthase and glutamate dehydrogenase inB. methanolicus MGA3 has no or a marginally positive effecton glutamate production [32], while overproduction of the na-tive Odh protein led to a fivefold reduction in glutamate. In-terestingly, Odh overexpression in an L-lysine-overproducingB. methanolicus host resulted in a twofold increase in L-lysineproduction yield, suggesting that Odh activity plays an impor-tant role at the branching point between glutamate synthesisand the TCA cycle, eventually regenerating the L-lysine pre-cursor molecule oxaloacetate [32]. Combining these findingswith the proteome data generated for the presence of Odhin this study suggests that the potential for amino acid pro-duction could differ on both carbon sources, with methanol

being the more suitable substrate for the production of glu-tamate and mannitol being the preferred substrate for theaccumulation of lysine.

3.4 Influence of growth temperature on the

proteome of B. methanolicus

B. methanolicus is a spore-forming bacterium and a rapidshift in growth temperature from 50 to 37�C induces sporu-lation in MGA3 cell cultures [6]. The sporulation processin other organisms such as B. subtilis had been studied ingreat detail, and once sporulation is initiated, it is irreversibleand resource demanding [33]. When growing B. methanolicuson methanol at 37�C, the sporulation-related proteins stageVI sporulation protein D and anti-sigma F factor antagonistSpoIIAA were among the proteins most highly upregulatedin comparison to methylotrophic growth at 50�C. In addition,another sporulation protein, stage IV sporulation protein A,was among the top 25 most regulated proteins at 37�C (Ta-ble 2). The detected anti-sigma F factor antagonist is knownto activate sigma factor F in the forespore by titrating SpoI-IAB (I3EA48) away from sigma F, thereby initiating spore


Proteomics 2014, 14, 725–737 735

formation [34]. In B. subtilis, stage VI sporulation protein Dis required for the assembly of the spore coat and might be acomponent of the innermost layer of the spore coat [35]. StageIV sporulation protein A is an ATPase and plays a role at anearly stage of the morphogenesis of spore coat outer layers[36, 37].

Among the 25 most upregulated proteins at growth tem-perature 37�C were other enzymes that may be indirectlylinked to spore formation as a consequence of metabolicrearrangements. Acyl-CoA dehydrogenase, one of the ini-tial proteins needed for ß-oxidation, was detected as wellas catalase. In addition, other ß-oxidation related enzymeswere significantly regulated, such as 3-hydroxybutyryl-CoAdehydrogenase (I3EBN6; 3.3-fold), short-chain acyl-CoA de-hydrogenase (I3EBN4; threefold), putative acyl-CoA dehydro-genase, and short-chain-specific putative acyl-CoA dehydro-genase (I3E8K5; 2.1-fold). These findings suggest it is likelythat due to the high demand of energy during sporulation,high levels of ß-oxidation are required to supply additional en-ergy. Because H2O2 is known to be an unwanted side productof ß-oxidation, the overexpression of catalase might counter-act oxidative stress.

The increased synthesis of glycerophosphoryl diester phos-phodiesterase, a protein needed for phospholipid degrada-tion, at 37�C and the simultaneous downregulation of 1-acyl-sn-glycerol-3-phosphate acyltransferase (I3EA01), a pro-tein involved in phospholipid biosynthesis, may indicate thathigher demand for fatty acids is fulfilled by the degradationof cell membrane components at this lower temperature.

In general, sporulation requires an entirely new set of pro-teins. It therefore appears feasible that cells optimize theirmetabolism to enable rapid and efficient protein biosynthe-sis. Some ribosomal proteins were found to be upregulated,specifically ribosomal protein L17 (I3E403), 50S ribosomalprotein L14 (I3E422; 6.8-fold), and 30S ribosomal proteinS18 (I3EBY3; 6.2-fold). Another protein that is associatedwith the ribosome, ribosome-binding factor A (I3DZF5; 5.3-fold), was also significantly upregulated at 37�C, althoughit was not among the top 25 most regulated proteins. Ithas been reported that the deletion of rbfA in Escherichiacoli triggers cold shock responses, resulting in lower pro-tein synthesis, and overproduction of this protein servesas protection against cold shock [38–40]. B. methanolicustherefore might use the same mechanism to maintain pro-tein synthesis at low temperatures. Another hint that pro-tein synthesis is of high priority at 37�C is the upregula-tion of 5-methyltetrahydropteroyltriglutamate-homocysteinemethyltransferase (I3E8N9), an enzyme needed in the biosyn-thesis of methionine, which is the first amino acid of everynewly synthesized protein.

It is known that motility is reduced during sporulation inB. subtilis [41]. The same appears to be true for B. methano-licus, as two flagellar proteins were found to be among the25 most downregulated genes at 37�C. Additionally, a flagel-lar assembly factor (FliW; I3EAZ9; 4.4-fold) was significantlydownregulated.

4 Concluding remarks

Our data provide detailed information concerning the relativeproduction levels of proteins under different growth condi-tions. Growth on methanol requires a specific set of proteinswith distinct biochemical features. Some of the genes corre-sponding to the most-regulated proteins needed for methy-lotrophic growth exist in B. methanolicus in duplicate. Ourdata confirm previous reports that only those proteins thatare encoded on the pBM19 plasmid are upregulated dur-ing growth on methanol, underlining the importance of thisplasmid for growth on methanol. As the biochemical prop-erties of the two versions of Fba and GlpX as determined invitro have already been described [18, 24], it appears reason-able to assume that the different enzymes perform differentmetabolic tasks. Whether the same is true for the two ver-sions of Tkt and Zwf remains to be determined; however,because the production of the respective isoenzymes is de-pendent on the carbon source, a similar preference for differ-ent pathways seems likely. The strongly induced expressionof GlpX on methanol is in line with published transcrip-tome data [19], and these findings corroborate the assump-tion that B. methanolicus MGA3 uses the sedoheptulose-1,7-bisphosphatase variant rather than the transaldolase variantof the RuMP cycle. Although the plasmid-encoded mdh genehas been reported to be essential for growth on methanol [23]and slightly upregulated at the transcriptome level [19], wedid not detect a significant quantitative difference at the pro-teome level. The constitutive production of Mdh might thusindicate that this organism frequently encounters methanolin its natural environment, and therefore downregulation inits absence is not a preferred trait. For Mdh2, no regulationwas detected, and for Mdh3 only a slight threefold downreg-ulation on methanol was detected, making it unlikely thatthese enzymes are directly involved in methanol oxidation.Due to their reported broad substrate specificity [25] and thefact that some methylotrophic bacilli can grow on other al-cohol substrates [11], this may suggest that Mdh2 and Mdh3are involved in the oxidation of alternative carbon substrates;however, further experiments are needed to determine theirmetabolic function. The missing regulation of any proteinsassumed to be part of the linear formaldehyde oxidation path-way and the fact that of all the enzymes of the putative cyclicformaldehyde dissimilation pathway, only Zwf (i.e. Zwf2) isupregulated on methanol, may suggest a role for this latterpathway in formaldehyde oxidation. It also remains to beinvestigated whether the linear tetrahydrofolate pathway in-deed plays a catabolic role as has been assumed before [20], orwhether it serves as a formaldehyde overflow valve for detoxifi-cation comparable to the role of the tetrahydromethanopterinpathway in the obligate methylotroph Methylobacillus flagella-tus KT [21].

The data presented here also revealed the potential foramino acid production to be affected by growth on differentcarbon sources. The upregulation of Odh on mannitol



may indicate suitable L-lysine production conditions andsimultaneously suggests methanol as the preferred sub-strate for increased L-glutamate accumulation, but furtherstudies are needed to prove this assumption. In any case,B. methanolicus is a good candidate for L-glutamate produc-tion from methanol, and its natural high levels of L-glutamateproduction might even be increased by further metabolicengineering.

Sporulation is a very resource-demanding process [33]. Assporulation is initiated at 37�C in B. methanolicus [6], majormetabolic changes appear to support this process. In additionto proteins directly involved in spore formation, ß-oxidationappears to be increased in response to the higher energy de-mand. Additionally, protein synthesis is optimized to enablethe fast production of sporulation-related proteins.

To our knowledge, our data represent the first proteomestudy of a Gram-positive Bacillus species during methy-lotrophic growth. Taken together, the presented referencedata sets examining the adaptation of B. methanolicus MGA3to different growth conditions can be used to develop futurestrategies for targeted metabolic engineering approaches.

The MS proteomics data have been deposited tothe ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the PRIDE partner repository [42] withthe dataset identifiers DOI PXD000637 (carbon dependency) andPXD000638 (temperature dependency). This work was supportedby the European Science Foundation (ESF) 09-EuroSYNBIO-FP-023 SynMet project, funded in part by SNF 31SY30–131039,and EU-FP7 PROMYSE “Products from Methanol by SyntheticCell Factories.”

5 References

[1] Anthony, C., The Biochemistry of Methylotrophs, AcademicPress, London/New York 1982.

[2] Chistoserdova, L., Modularity of methylotrophy, revisited.Environ. Microbiol. 2011, 13, 2603–2622.

[3] Vorholt, J. A., Cofactor-dependent pathways of formalde-hyde oxidation in methylotrophic bacteria. Arch. Microbiol.2002, 178, 239–249.

[4] Schrader, J., Schilling, M., Holtmann, D., Sell, D. et al.,Methanol-based industrial biotechnology: current statusand future perspectives of methylotrophic bacteria. TrendsBiotechnol. 2009, 27, 107–115.

[5] Brautaset, T., Jakobsen, O. M., Josefsen, K. D., Flickinger, M.C., Ellingsen, T. E., Bacillus methanolicus: a candidate forindustrial production of amino acids from methanol at 50degrees C. Appl. Microbiol. Biotechnol. 2007, 74, 22–34.

[6] Schendel, F. J., Bremmon, C. E., Flickinger, M. C., Guettler, M.,Hanson, R. S., L-lysine production at 50 degrees C by mu-tants of a newly isolated and characterized methylotrophicBacillus sp. Appl. Environ. Microbiol. 1990, 56, 963–970.

[7] Brautaset, T., Williams, M. D., Dillingham, R. D., Kaufmann,C. et al., Role of the Bacillus methanolicus citrate synthase

II gene, citY, in regulating the secretion of glutamate in L-lysine-secreting mutants. Appl. Environ. Microbiol. 2003, 69,3986–3995.

[8] Jakobsen, O. M., Benichou, A., Flickinger, M. C., Valla, S.et al., Upregulated transcription of plasmid and chromoso-mal ribulose monophosphate pathway genes is critical formethanol assimilation rate and methanol tolerance in themethylotrophic bacterium Bacillus methanolicus. J. Bacte-riol. 2006, 188, 3063–3072.

[9] Cue, D., Lam, H., Dillingham, R. L., Hanson, R. S., Flickinger,M. C., Genetic manipulation of Bacillus methanolicus, agram-positive, thermotolerant methylotroph. Appl. Environ.Microbiol. 1997, 63, 1406–1420.

[10] Nilasari, D., Dover, N., Rech, S., Komives, C., Expression ofrecombinant green fluorescent protein in Bacillus methano-licus. Biotechnol. Prog. 2012, 28, 662–668.

[11] Arfman, N., Dijkhuizen, L., Kirchhof, G., Ludwig, W. et al.,Bacillus methanolicus sp. nov., a new species of thermotol-erant, methanol-utilizing, endospore-forming bacteria. Int. J.Syst. Bacteriol. 1992, 42, 439–445.

[12] Arfman, N., de Vries, K. J., Moezelaar, H. R., Attwood,M. M. et al., Environmental regulation of alcohol metabolismin thermotolerant methylotrophic Bacillus strains. Arch. Mi-crobiol. 1992, 157, 272–278.

[13] Arfman, N., Hektor, H. J., Bystrykh, L. V., Govorukhina, N. I.et al., Properties of an NAD(H)-containing methanol dehy-drogenase and its activator protein from Bacillus methano-licus. Eur. J. Biochem. 1997, 244, 426–433.

[14] de Vries, G. E., Arfman, N., Terpstra, P., Dijkhuizen, L.,Cloning, expression, and sequence analysis of the Bacillusmethanolicus C1 methanol dehydrogenase gene. J. Bacte-riol. 1992, 174, 5346–5353.

[15] Hektor, H. J., Kloosterman, H., Dijkhuizen, L., Identifica-tion of a magnesium-dependent NAD(P)(H)-binding do-main in the nicotinoprotein methanol dehydrogenasefrom Bacillus methanolicus. J. Biol. Chem. 2002, 277,46966–46973.

[16] Ferenci, T., Strom, T., Quayle, J. R., Purification and prop-erties of 3-hexulose phosphate synthase and phospho-3-hexuloisomerase from Methylococcus capsulatus. Biochem.J. 1974, 144, 477–486.

[17] Kato, N., Yurimoto, H., Thauer, R. K., The physiological roleof the ribulose monophosphate pathway in bacteria and ar-chaea. Biosci. Biotechnol. Biochem. 2006, 70, 10–21.

[18] Stolzenberger, J., Lindner, S. N., Persicke, M., Brautaset,T., Wendisch, V. F., Characterization of fructose 1,6-bisphosphatase and sedoheptulose 1,7-bisphosphatasefrom the facultative ribulose monophosphate cycle methy-lotroph Bacillus methanolicus. J. Bacteriol. 2013, 195,5112–5122.

[19] Heggeset, T. M., Krog, A., Balzer, S., Wentzel, A. et al.,Genome sequence of thermotolerant Bacillus methanolicus:features and regulation related to methylotrophy and pro-duction of L-lysine and L-glutamate from methanol. Appl.Environ. Microbiol. 2012, 78, 5170–5181.

[20] Pluschkell, S. B., Flickinger, M. C., Dissimilation of[(13)C]methanol by continuous cultures of Bacillus methano-licus MGA3 at 50 degrees C studied by (13)C NMR and


http://proteomecentral.proteomexchange.org

http://proteomecentral.proteomexchange.org

Proteomics 2014, 14, 725–737 737

isotope-ratio mass spectrometry. Microbiology 2002, 148,3223–3233.

[21] Chistoserdova, L., Gomelsky, L., Vorholt, J. A., Gomelsky,M. et al., Analysis of two formaldehyde oxidation pathwaysin Methylobacillus flagellatus KT, a ribulose monophosphatecycle methylotroph. Microbiology 2000, 146, 233–238.

[22] Arfman, N., Watling, E. M., Clement, W., van Oosterwijk,R. J. et al., Methanol metabolism in thermotolerant methy-lotrophic Bacillus strains involving a novel catabolic NAD-dependent methanol dehydrogenase as a key enzyme. Arch.Microbiol. 1989, 152, 280–288.

[23] Brautaset, T., Jakobsen, M. O., Flickinger, M. C., Valla, S.,Ellingsen, T. E., Plasmid-dependent methylotrophy in ther-motolerant Bacillus methanolicus. J. Bacteriol. 2004, 186,1229–1238.

[24] Stolzenberger, J., Lindner, S. N., Wendisch, V. F., Themethylotrophic Bacillus methanolicus MGA3 possesses twodistinct fructose 1,6-bisphosphate aldolases. Microbiology2013, 159, 1770–1781.

[25] Krog, A., Heggeset, T. M., Muller, J. E., Kupper, C. E. et al.,Methylotrophic Bacillus methanolicus encodes two chromo-somal and one plasmid born NAD(+) dependent methanoldehydrogenase paralogs with different catalytic and bio-chemical properties. PLoS ONE 2013, 8, e59188.

[26] Kloosterman, H., Vrijbloed, J. W., Dijkhuizen, L., Molecular,biochemical, and functional characterization of a Nudix hy-drolase protein that stimulates the activity of a nicotinopro-tein alcohol dehydrogenase. J. Biol. Chem. 2002, 277, 34785–34792.

[27] Chistoserdova, L., Kalyuzhnaya, M. G., Lidstrom, M. E., Theexpanding world of methylotrophic metabolism. Annu. Rev.Microbiol. 2009, 63, 477–499.

[28] Weisser, H., Nahnsen, S., Grossmann, J., Nilse, L. et al., Anautomated pipeline for high-throughput label-free quantita-tive proteomics. J. Proteome Res. 2013, 12, 1628–1644.

[29] Quayle, J. R., 3-Hexulose-6-phosphate synthase from Methy-lomonas (Methylococcus) capsulatus. Methods Enzymol.1982, 90, 314–319.

[30] Nash, T., The colorimetric estimation of formaldehyde bymeans of the Hantzsch reaction. Biochem. J. 1953, 55,416–421.

[31] Solomon, E., Lin, E. C., Mutations affecting the dissimilationof mannitol by Escherichia coli K-12. J. Bacteriol. 1972, 111,566–574.

[32] Krog, A., Heggeset, T. M., Ellingsen, T. E., Brautaset, T.,Functional characterization of key enzymes involved in L-glutamate synthesis and degradation in the thermotolerantand methylotrophic bacterium Bacillus methanolicus. Appl.Environ. Microbiol. 2013, 79, 5321–5328.

[33] Phillips, Z. E., Strauch, M. A., Bacillus subtilis sporulationand stationary phase gene expression. Cell. Mol. Life Sci.2002, 59, 392–402.

[34] Duncan, L., Alper, S., Losick, R., SpoIIAA governs the re-lease of the cell-type specific transcription factor sigma Ffrom its anti-sigma factor SpoIIAB. J. Mol. Biol. 1996, 260,147–164.

[35] Beall, B., Driks, A., Losick, R., Moran, C. P., Jr., Cloningand characterization of a gene required for assembly ofthe Bacillus subtilis spore coat. J. Bacteriol. 1993, 175,1705–1716.

[36] Stevens, C. M., Daniel, R., Illing, N., Errington, J., Character-ization of a sporulation gene, spoIVA, involved in spore coatmorphogenesis in Bacillus subtilis. J. Bacteriol. 1992, 174,586–594.

[37] Roels, S., Driks, A., Losick, R., Characterization of spoIVA, asporulation gene involved in coat morphogenesis in Bacillussubtilis. J. Bacteriol. 1992, 174, 575–585.

[38] Dammel, C. S., Noller, H. F., Suppression of a cold-sensitive mutation in 16S rRNA by overexpression of a novelribosome-binding factor, RbfA. Genes Dev. 1995, 9, 626–637.

[39] Jones, P. G., Inouye, M., RbfA, a 30S ribosomal binding fac-tor, is a cold-shock protein whose absence triggers the cold-shock response. Mol. Microbiol. 1996, 21, 1207–1218.

[40] Xia, B., Ke, H., Shinde, U., Inouye, M., The role of RbfA in16S rRNA processing and cell growth at low temperature inEscherichia coli. J. Mol. Biol. 2003, 332, 575–584.

[41] Nishihara, T., Freese, E., Motility of Bacillus subtilis duringgrowth and sporulation. J. Bacteriol. 1975, 123, 366–371.

[42] Vizcaıno, J. A., Cote, R. G., Csordas, A., Dianes, J. A. et al., ThePRoteomics IDEntifications (PRIDE) database and associatedtools: status in 2013. Nucleic Acids Res. 2013, 41, D1063–D1069.


Proteomic analysis of the thermophilic methylotroph Bacillus methanolicus MGA3

Documents

Transcript of Proteomic analysis of the thermophilic methylotroph Bacillus methanolicus MGA3