Abstract The ftr gene encoding formylmethanofuran:tetrahydromethanopterin formyltransferase (Ftr) fromMethanosarcina barkeri was cloned, sequenced, andfunctionally expressed in Escherichia coli. The overpro-duced enzyme was purified eightfold to apparent homo-geneity, and its catalytic properties were determined. Theprimary structure and the hydropathic character of theformyltransferase from Methanosarcina barkeri werecompared with those of the enzymes from Methanobac-terium thermoautotrophicum, Methanothermus fervidus,and Methanopyrus kandleri. The amino acid sequence ofthe enzyme from Methanosarcina barkeri was 64%, 61%,and 59% identical to that of the enzyme from Me-thanobacterium thermoautotrophicum, Methanothermusfervidus, and Methanopyrus kandleri, respectively. A neg-ative correlation between the hydrophobicity of the en-zymes and both the growth temperature optimum and theintracellular salt concentration of the four organisms wasobserved. The hydrophobicity of amino acid compositionwas +21.6 for the enzyme from Methanosarcina barkeri(growth temperature optimum 37°C, intracellular saltconcentration ≈ 0.3 M), +9.9 for the enzyme fromMethanobacterium thermoautotrophicum (65°C, ≈ 0.7M), –20.8 for the enzyme from Methanothermus fervidus(83°C, ≈ 1.0 M) and –31.4 for the enzyme from Me-thanopyrus kandleri (98°C, > 1.1 M). Generally, a posi-tive correlation between hydrophobicity and ther-mophilicity of enzymes and a negative correlation be-tween hydrophobicity and halophilicity of enzymes areobserved. The findings therefore indicate that the hydro-pathic character of the formyltransferases compared ismainly determined by the intracellular salt concentration

rather than by temperature. Sequence similarities betweenthe formyltransferases from methanogens and an openreading frame from Methylobacterium extorquens AM1are discussed.

Key words Formyltransferase · Formyltetrahydrofolate synthase · Hyperthermophilic enzymes · Methanogenic Archaea · Methanosarcina barkeri

Abbreviations cDPG Cyclic 2,3-diphosphoglycerate · Formyltransferase or Ftr Formylmethanofuran:tetrahydromethanopterin formyltransferase · IPTGIsopropyl-β-D-thiogalactopyranoside · PCR Polymerase chain reaction

Introduction

Formylmethanofuran:tetrahydromethanopterin formyltrans-ferase (Ftr) is found in methanogenic and sulfate-reducingArchaea. In these strictly anaerobic microorganisms the en-zyme is involved in CO2 reduction to CH4, in autotrophicCO2 fixation, in C1-unit formation from CO2, and in C1-unitoxidation to CO2 (Wolfe 1991; Thauer et al. 1993; Vorholtet al. 1995). The enzyme catalyzes a reversible reaction(∆G°’ = –4.4 kJ/mol) in which two novel coenzymes are in-volved: methanofuran (Leigh et al. 1984, 1985) and tetrahy-dromethanopterin (Escalante-Semerena et al. 1984a,b). Thesoluble enzyme is composed of only one type of subunit ofmolecular mass of about 32 kDa, is stable even under oxicconditions, and exhibits a ternary-complex catalytic mecha-nism (Donelly and Wolfe 1986; Breitung and Thauer 1990;Breitung et al. 1992). The DNA sequences of the formyl-transferase gene from the thermophilic Methanobacteriumthermoautotrophicum (growth temperature optimum 65°C), from the hyperthermophilic Methanothermus fervidus(83°C), and from the hyperthermophilic Methanopyruskandleri (98°C) have been determined (DiMarco et al.1990; Lehmacher 1994; Shima et al. 1995).

Recently, the ftr gene encoding the formyltransferasefrom Methanopyrus kandleri was functionally expressed

Jasper Kunow · Seigo Shima · Julia A. Vorholt ·Rudolf K. Thauer

Primary structure and properties of the formyltransferase from the mesophilic Methanosarcina barkeri:comparison with the enzymes from thermophilic and hyperthermophilic methanogens

Arch Microbiol (1996) 165 :97–105 © Springer-Verlag 1996

Received: 7 September 1995 / Accepted: 7 November 1995

ORIGINAL PAPER

Jasper Kunow (Y) · Seigo Shima · Julia A. Vorholt ·Rudolf K. ThauerMax-Planck-Institut für terrestrische Mikrobiologie und Laboratorium für Mikrobiologie, Fachbereich Biologie, Philipps-Universität, Karl-von-Frisch-Strasse, D-35043 Marburg, GermanyTel. +49-6421-283480; Fax +49-6421-285833

in Escherichia coli in order to obtain large amounts of thehyperthermophilic enzyme for studies of structure-func-tion relationships (Shima et al. 1995). Crystals of the re-combinant enzyme suitable for structure determinationusing X-ray diffraction methods have been obtained (S.Shima, U. Ermler, and H. Michel, unpublished results).

In this communication we describe the cloning and se-quencing of the ftr gene from the mesophilic Metha-nosarcina barkeri (growth temperature optimum 37°C)and its functional expression in Escherichia coli. The aimof this work was to compare the primary sequence andproperties of this enzyme with those of the formyltrans-ferase from the thermophilic and hyperthermophilicmethanogens.

Materials and methods

Organisms, culture conditions, plasmids, and phages

Methanosarcina barkeri strain Fusaro (DSM 804) is the strain de-posited in the Deutsche Sammlung von Mikroorganismen (Braun-schweig, Germany). The Archaeon was grown on methanol, har-vested, and stored as described (Fischer and Thauer 1989; Kar-rasch et al. 1989). Escherichia coli DH5-α and Escherichia coliBL21 (DE3) are described by Sambrook et al. (1989).

The lambda ZAP II genomic library of Methanosarcina bar-keri strain Fusaro (4–6 kbp EcoRI fragments) was a gift from K.Fiebig (Berlin, Germany; Massanz et al. 1993). The helper phageExAssist and Escherichia coli strains XL1-Blue-MRF´ and SOLRwere from Stratagene (Heidelberg, Germany). The expression vec-tor pET11d was from AMS-Biotechnology (Bioggio-Lugano,Switzerland).

Isolation of chromosomal DNA from Methanosarcina barkeri

Chromosomal DNA was isolated according to the method of Jar-rell et al. (1991). After 1.5 g of frozen Methanosarcina barkericells was ground in a mortar under liquid nitrogen with a pestle,the broken cells were transferred to a 50-ml conical tube and sus-pended at room temperature in 5 ml of 20 mM Tris-HCl (pH 7.5)containing 1 mM EDTA and 0.25 M sucrose. SDS (1% final con-centration) and Proteinase K (50 µg/ml) were added and the mix-ture was incubated at 60°C for 30 min. After addition of 0.5 MNaCl, the partially lysed cells were put on ice for 1 h. Subse-quently, the lysate was cleared by centrifugation at 27,000 × g for15 min at 4°C. After RNase treatment of the supernatant (10mg/ml RNase previously boiled for 10 min to inactivate DNases,30 min, 37°C), the DNA was extracted twice with an equal vol-ume of phenol-chloroform-isoamyl alcohol (25:24:1), followed bytwo extractions with chloroform. Following addition of 0.5 MNaCl, the DNA was precipitated with 2.5 volumes of cold ethanol,collected by centrifugation, and dissolved in 3 ml 20 mM Tris-HCl(pH 7.5) containing 1 mM EDTA.

Cloning of the ftr gene

A homologous probe for the ftr gene was obtained by PCR usingprimers derived from the N-terminal amino acid sequenceMEINGVEIED of Ftr [ATG GA(AG) AT(ACT) AA(CT) GG(GT)GT(GT) GAA AT(ACT) GAGG; Breitung et al. 1992] and fromthe internal amino acid sequence CPAEAGID deduced from thehomologues ftr genes from Methanothermus fervidus and Me-thanobacterium thermoautotrophicum [TCG AT(GT) CC(GT)GC(CT) TC(GT) GC(ACGT) GG(AG) C; Lehmacher 1994; Di-Marco et al. 1990]. The 25 µl reaction mixture contained: 10 ng ofgenomic DNA of Methanosarcina barkeri, 2.5 U Taq DNA poly-

merase, 100 µM each dNTP, 1.5 mM MgCl2, and 800 pM eachprimer (annealing temperature 50°C). The 170-bp PCR fragmentobtained was cloned into the pCRII vector using the Invitrogencloning kit and sequenced.

The sequence data allowed the synthesis of a homologousoligonucleotide (31-mer), that was tailed with digoxigenin-dUTPaccording to the protocol given by Boehringer Mannheim and hy-bridized with Methanosarcina barkeri genomic DNA digested tocompletion with either EcoRI, EcoRV, HindIII, AccI, or PstI in or-der to test the specificity of the probe. Southern blot hybridizationanalysis revealed only one strong signal in each digest when thehybridization was performed in 15 mM sodium citrate (pH 7.0)containing 0.15 M NaCl and 0.1% SDS (55°C for hybridizationand 60°C for washing).This finding suggests, that the homologousoligonucleotide probe specifically hybridizes with the ftr gene andthat Methanosarcina barkeri strain Fusaro contains only one ftrgene.

The homologous oligonucleotide probe was used to screen alambda ZAPII genomic library of Methanosarcina barkeri strainFusaro. One positive phage clone out of 1600 was identified andisolated. Excision and recircularization of pBluescript SK– con-taining the cloned insert (Short and Sorge 1992) generated thephagemid pJK1 (Fig. 1). After digestion of the recombinantphagemid DNA with EcoRI, a 1.6-kb fragment, which hybridizedwith the ftr probe, and a 3.2-kb fragment, which did not hybridize,were obtained. Hence, a 4.8-kb DNA fragment of Methanosarcinabarkeri genomic DNA had been cloned. Sequencing revealed thatthe 4.8-kb fragment harbored the complete ftr gene (Fig.1).

DNA sequencing

DNA sequencing was performed with denatured double-strandedplasmid DNA according to the dideoxynucleotide chain termina-tion method (Sanger et al. 1977; Chen and Seeburg 1985) usingSequenase version 2.0. The sequence of the ftr gene and of itsflanking regions was determined independently for both strands.

Heterologous expression of the ftr gene in Escherichia coli

For incorporation of the ftr gene into the expression vectorpET11d, ftr was amplified by PCR using pJK1 (Fig. 1) as template.For amplification, the following two primers were used: CAAATG GGA GAT TTC CAT GGA AAT CAA CGG AGT AGAAAT CG (sense) and GAA AAG GCT TGG ATC CTC AGAAGA GTT CGT GAA GTT TAA ACT G (antisense). The senseprimer was designed such that the amplified gene started withATG and a NcoI restriction site was located one nucleotide up-

98

Fig.1 Location of the ftr gene within a 4.8-kb DNA fragmentcloned into the EcoRI site of pBluescript IISK–. The cloning con-struct was designated pJK1. Square and P putative promotor; Cir-cle and T putative terminator

stream of the start codon. The antisense primer was designed suchthat a BamHI restriction site was located 5 nucleotides downstreamof the ftr gene stop codon. The 250 µl reaction mixture contained:10 ng DNA of pJK1 containing the ftr gene, 2.5 U Pfu DNA poly-merase, 100 µM each dNTP, 2 mM MgCl2 and 800 pM eachprimer. The temperature program was 1 × 3 min at 94°C, 30 cycles1 min 92°C/1 min 50°C/2.5 min 72°C, 1 × 5 min at 72°C. ThePCR product was purified via absorption/desorption to a QIAquickSpin Column (Qiagen, Hilden, Germany) according to the protocolof the manufacturer. After purification, the PCR product was di-gested with NcoI and BamHI and was ligated into the pET11d ex-pression vector previously digested with NcoI and BamHI. Theconstruct thus obtained was designated pJK2 and amplified in Es-cherichia coli DH5-α. Sequencing of the ftr gene cloned intopET11d revealed no mutation.

pJK2 with the ftr gene was used to transform Escherichia coliBL21 (DE3) (Studier et al. 1990). Genes cloned into pET11d aretranscribed by T7 RNA polymerase whose gene resides on theprophage DE3 integrated into the chromosome of BL21 under thecontrol of the lacUV5 promoter. This promoter can be induced byaddition of isopropyl-β-D-thiogalactopyranoside (IPTG). For ex-pression of the ftr gene in Escherichia coli BL21 (DE3) (pJK2) thetransformed cells were aerobically grown at 37°C in 2 l minimalmedium M9 (Sambrook et al. 1989) supplemented with ampicillin(125 µg/ml). When the OD578 nm of the culture reached 0.8, IPTGwas added to the culture to a final concentration of 0.08 mM. Af-ter 2 h, at an OD578 of 2.2, the cells were harvested by centrifuga-tion at 4,200 × g for 20 min at 4°C, yielding 8 g cells (wet mass).

Purification of formyltransferase from Escherichia coli

All steps were performed under strictly anoxic conditions. Thebuffer used was 50 mM Tricine/KOH (pH 7.5) containing 2 mMdithiothreitol.

The Escherichia coli cells (8 g wet mass), in which the ftr genewas expressed, were suspended in 16 ml buffer and disrupted byultrasonification for 10 min at 100 W (50 cycles/min). Unbrokencells and cell debris were removed by centrifugation at 30,000 × gfor 30 min at 4°C. The supernatant, which contained 22 mg pro-tein and 1,400 U formyltransferase activity per ml and which wasdesignated cell extract, was subsequently ultracentrifuged at120,000 × g for 45 min. This supernatant was then diluted with athreefold volume of buffer and heated to 65°C for 60 min. Aftercooling to room temperature precipitated material was removed bycentrifugation at 20,000 × g for 20 min. The supernatant was ad-justed to a final concentration of 0.55 M (NH4)2SO4 by adding 9ml buffer saturated with (NH4)2SO4, filtrated (pore size 0.45 µm;Schleicher und Schuell, Dassel, Germany), and applied to aPhenyl-Sepharose HiLoad HR 26/10 column equilibrated with 0.4M (NH4)2SO4 in buffer (3 ml/min). The column was washed witha linear decreasing gradient of (NH4)2SO4 in buffer (200 ml, 0.4–0M) and formyltransferase was then eluted with a linear increasinggradient of ethylene glycol in buffer (400 ml, 0–60%). Fractions of8 ml were collected. Formyltransferase activity eluted in 10 frac-tions at 12–24% ethylene glycol. These fractions, which togethercontained 27 mg protein and 13,360 U formyltransferase activity,were pooled and stored at –20°C under nitrogen gas.

RNA isolation and northern blot hybridization

A 250-ml culture of Methanosarcina barkeri strain Fusaro wasgrown to an OD578 of 1.5, cooled in an ethanol/dry-ice mixture,and then harvested by centrifugation at 4°C in a pre-cooled rotor.The supernatant was discarded and the pellet was transferred to amortar that had been cooled with liquid nitrogen. The cells (0.75 gwet mass) were ruptured by grinding in liquid nitrogen for 15 min.The RNA was isolated from the ruptured cells using the single-stepmethod for RNA isolation described by Ausubel et al. (1987) andsubjected to electrophoresis. The blot was probed with the 31-merhomologous oligonucleotide probe for the ftr gene. Hybridizations

were performed at 60°C in 15 mM sodium citrate (pH 7.0) con-taining 0.15 M NaCl and 0.1% SDS.

Determination of the specific activity of formyltransferase

Formyltransferase activity was routinely determined at 37°C un-der the standard assay conditions described by Breitung et al.(1992). The formylmethanofuran concentration was, however,only 60 µM rather than 70 µM, which is mentioned because theapparent Km for the formylmethanofuran of the enzyme is 400 µM(Breitung and Thauer 1990). Protein concentrations were deter-mined by the method of Bradford (1976) using the reagents fromBio-Rad Laboratories (Munich, Germany) and bovine serum albu-min as the standard.

Supplies and chemicals

Restriction enzymes, T4 DNA ligase, Shrimp alkaline phosphataseand the Sequenase Version 2.0 DNA Sequencing Kit were fromUnited States Biochemicals (Bad Homburg, Germany). Syntheticoligonucleotides were purchased from MWG Biotech (Ebersberg,Germany). The Digoxigenin Luminiscent Detection Kit, theDigoxigenin Oligonucleotide Tailing Kit, Taq DNA polymerase,Pfu DNA polymerase, and Proteinase K were from BoehringerMannheim (Mannheim, Germany). The TA Cloning Kit was fromInvitrogen (NV Leek, The Netherlands). Nylon membrane Hy-bond-N was from Amersham (Braunschweig, Germany). ThePhenyl-Sepherose column and the low-molecular-mass markerproteins were from Pharmacia Biotech (Freiburg, Germany).Tetrahydromethanopterin and methanofuran were purified fromMethanobacterium thermoautotrophicum (strain Marburg) as de-scribed by Breitung et al. (1992). Formylmethanofuran was syn-thesized from methanofuran and 4-nitrophenylformate (Donnellyand Wolfe 1986).

Results

The gene encoding the formyltransferase from Me-thanosarcina barkeri was identified on a 4.8-kb EcoRIfragment which was cloned into the phagemid pBluescriptSK– (Fig. 1) and sequenced. The sequence is available un-der accession number X91143 in the EMBL database. Af-ter recloning into the expression vector pET11d, the ftrgene was expressed in Escherichia coli BL21 (DE3),yielding an active formyltransferase that was subse-quently purified and characterized.

Nucleotide sequence of the ftr gene

The ftr gene was found to be 894 bp in length with ATGand TGA as initiation and termination codons, respec-tively. The N-terminal amino acid sequence deduced fromthe ftr gene sequence corresponded to that determined byEdman degradation (27 amino acids) for the formyltrans-ferase purified from Methanosarcina barkeri strainFusaro (Breitung et al. 1992). A comparison revealed twodifferences at positions 24 (I versus L) and 26 (A versusT). It is known that the reliability of N-terminal aminoacid sequence determinations by Edman degradation de-creases with the distance from the N-terminus.

The G+C content of the ftr gene was 46 mol%, whichis consistent with the overall 42 mol% G+C content of the

99

Methanosarcina barkeri genome (Boone et al. 1993). Thecodon usage followed the pattern seen for otherMethanosarcina barkeri genes except that GAT was themost often found codon for aspartate and AAA was themost often found codon for lysine in ftr, whereas GACand AAG, respectively, are the codons usually preferredin other Methanosarcina barkeri genes, e.g. mcr, argG,atpA, and atpB (Bokranz and Klein 1987; Morris andReeve 1988; Inatomi et al. 1989).

It is predicted that the ftr gene encodes a protein of 297amino acids with a calculated molecular mass of 31701Da and a pI of 4.9. The molecular mass is in agreementwith that estimated by SDS/PAGE for the protein purifiedfrom Methanosarcina barkeri strain Fusaro (Breitung andThauer 1990). The pI of 4.9 is somewhat lower than the pIof 4.5 determined experimentally by isoelectric focusing(Breitung and Thauer 1990).

Nucleotide sequence of the ftr gene flanking regions

A sequence GGGAGA located 5 bp upstream of the ATGstart codon of the ftr gene is partially complementary tothe 3´ terminal sequence of the 16S rRNA frommethanogens and is, therefore, considered to be the ribo-some binding site (Østergaard et al. 1987; Brown et al.1989; Reeve 1992). The sequence CTGA located 53-bpupstream of the translation initiation codon ATG corre-sponds to the ATGC consensus sequence for the transcrip-tion start site in methanogens (Hain et al. 1992; Reeve1992). The sequence CTTATA 27 bp upstream of the pu-tative transcription start site is located at an appropri-ate distance for the TATA box component of a promoterthat could direct ftr transcription initiation. The consen-sus sequence for the TATA-box in methanogens is(C/T)TTA(A/T)A (Hain et al. 1992; Reeve 1992). Down-stream of the ftr gene, the DNA sequence showed an oligoT sequence TTTTCCTTTTTT beginning 12 bp, and anoligo T sequence TTTTTTATT beginning 40 bp after thestop codon. Such sequences have been proposed as tran-scription terminators for many methanogenic genes(Brown et al. 1989; Reeve 1992; Thomm et al. 1994).Within the DNA region sequenced (Fig. 1), no other openreading frame was identified.

Evidence for a monocistronic transcript of the ftr gene

The results described above show that the ftr gene isflanked by sequences indicative of regions of transcrip-tion initiation and termination. Hence, the chromosomalarrangement suggests that the 894-bp ftr gene is tran-scribed into a monocistronic mRNA of about 1 kb. Thiswas confirmed by northern blot analysis in which a ho-mologous digoxigenin-dUTP-labelled DNA probe de-duced from the ftr gene was used for hybridization againsttotal RNA extracted from Methanosarcina barkeri strainFusaro cells. Only one strong band at the 1.1-kb positionwas observed (results not shown).

Expression of the ftr gene in Escherichia coli

Escherichia coli BL21 (DE3) was transformed with pJK2,which was obtained by cloning the ftr gene into pET11d(see Materials and methods). Cell extracts of IPTG-in-duced Escherichia coli (pJK2) cells were found to exhibitformyltransferase activity. This activity was not present incell extracts of Escherichia coli BL21 (DE3) and Es-cherichia coli BL21 (DE3)(pET11d). The activity in Es-cherichia coli (pJK2) was found only after IPTG induc-tion. Induction was accompanied by an accumulation of aprotein with an apparent molecular mass of 32 kDa as de-termined by SDS/PAGE (Fig. 2, lane C). This protein wasnot dominant when induction with IPTG was omitted(Fig. 2, lane B). The formation of active formyltransferaseindicates that the recombinant enzyme correctly folded inEscherichia coli and that the enzyme did not require thepresence of a methanogen-specific cofactor for activity.

Purification of the active formyltransferase from Escherichia coli

Cell extracts of Escherichia coli (pJK2) catalyzed thetransfer of the formyl group of formylmethanofuran totetrahydromethanopterin at specific rates of 64 U/mg.Upon ultracentrifugation, no activity was detected in theparticulate fraction. Purification was achieved by heating

100

Fig.2 Expression of the ftr gene from Methanosarcina barkeri inEscherichia coli BL21 (DE3) carrying pJK2 as analyzed bySDS/PAGE. Cells were grown aerobically and induced with IPTG;cell extracts were prepared as described in Materials and methods.Cell extract protein was separated on a 16% polyacrylamide gel,which was subsequently stained with Coomassie brilliant blueR250 (Laemmli 1970). Lanes A, F Molecular mass standards(Pharmacia, Freiburg, Germany); lane B, 8 µg of cell extract pro-tein of Escherichia coli (pJK2) grown in the absence of IPTG; laneC 20 µg of cell extract protein of Escherichia coli (pJK2) inducedby IPTG; lane D 4 µg protein of the 20,000 × g supernatant of ex-tract of IPTG induced cells after heat treatment; lane E 3 µg of re-combinant formyltransferase purified by chromatography onphenyl sepharose (see Table 1)

the Escherichia coli cell extract for 60 min at 65°C and,after removal of precipitated protein, by performing hy-drophobic interaction chromatography on Phenyl-Sepha-rose. The formyltransferase activity was found to copurifywith the 32-kDa protein (Fig.2). Via this simple proce-dure, an eightfold purification to apparent homogeneitywith a yield of 60% was achieved (Table 1). The formyl-transferase thus represented approximately 12% of the to-tal cellular Escherichia coli protein, assuming that the en-zyme was not partially inactivated during purification.The enzyme preparation obtained had a specific activityof 495 U/mg and contained 27 mg pure formyltransferase.The specific activity of the enzyme purified from Me-thanosarcina barkeri cells has been reported to be 610U/mg (Breitung and Thauer 1990). Also the temperatureactivity optimum, the thermostability, and the effect of saltson enzyme activity were found to be essentially identical.

Discussion

The properties of the formyltransferase from Metha-nosarcina barkeri are summarized in Table 2 and com-pared with those of the enzyme from Methanobacteriumthermoautotrophicum, Methanothermus fervidus, and

Methanopyrus kandleri. The table also contains informa-tion on the G+C content, the growth temperature opti-mum, and the intracellular concentrations of salts andcyclic 2,3-diphosphoglycerate (cDPG) of the organismsfrom which the enzymes were isolated.

A comparison of the data compiled in Table 2 revealsthat the formyltransferase from Methanosarcina barkerihas almost the same molecular mass as the enzymes fromMethanobacterium thermoautotrophicum, Methanother-mus fervidus, and Methanopyrus kandleri. The fourformyltransferases differ, however, in their isoelectricpoint, hydrophobicity of amino acid composition, temper-ature activity optimum, thermostability, and dependenceof the enzyme activity on the presence of salts. The tem-perature activity optimum and the thermostability in-crease with increasing growth temperature optimum ofthe organisms from which the enzyme is derived. In-versely, the hydrophobicity of amino acid composition de-creases with increasing growth temperature optimum. Thelatter finding is of interest since generally the hydropho-bicity of amino acid composition increases with increas-ing thermophilicity of a soluble enzyme (Argos et al.1979; Menéndez-Arias and Argos 1989; Zwickl et al.1990). The finding may be explained by the fact that theintracellular salt concentration of the four organisms in-

101

Table 1 Purification of formyl-transferase from Methano-sarcina barkeri overproduced inEscherichia coli. Cell extractwas prepared from 8 g cells(wet mass). Enzyme activityand protein were determinedunder standard assay conditionsas described in Materials andmethods

Purification step Protein Activity Specific activity Yield Purification(mg) (U) (U/mg) (%) (-fold)

Cell extract 352 22,400 64 100 1

120,000 × g supernatant 304 21,440 71 96 1.1

Heat treatment 77 19,679 256 88 4

Phenyl sepharose 27 13,360 495 60 8

Parameter Ms. barkeri Mb. thermoauto- Mt. fervidus Mp. kandleritrophicum

FormyltransferaseMolecular mass (Da) 31,701 31,401 31,835 31,664Isoelectric point 4.9 4.5 5.2 4.2Amino acid sequence identity (%) 100 64 61 59Hydrophobicity + 21.6 + 9.9 – 20.8 – 31.4Temperature activity optimum (°C) 65 65 70 ≈ 90Stimulation of activity by 1.5 M K2HPO4 (-fold) 4 80 3 > 1,000Thermostability in 1.5 M K2HPO4 up to (°C) 70 80 n. d. > 90

OrganismG+C content (mol%) 42 48 33 60Growth temperature optimum (°C) 37 65 83 98Intracellular salt concentration (M) ≈ 0.3 ≈ 0.7 ≈ 1.0 > 1.1Intracellular concentration of cDPG (M) 0.001 0.065 0.3 1.1

Table 2 Properties of formyltransferases from Methanosarcinabarkeri, Methanobacterium thermoautotrophicum, Methanother-mus fervidus, and Methanopyrus kandleri (this work; Donelly andWolfe 1986; Breitung and Thauer 1990; DiMarco et al. 1990; Bre-itung et al. 1992; Lehmacher 1994; Shima et al. 1995). The tablealso contains information on the G+C content, the growth temper-ature optimum, and the intracellular concentrations of salts andcyclic 2,3-diphosphoglycerate (cDPG) of the organisms, from

which the enzymes were isolated (Hensel and König 1988; Booneet al. 1993; Zinder 1993). The molecular mass, the isoelectricpoint, and the hydrophobicity were deduced from the amino acidsequence. Hydrophobicity of amino acid composition equals thesum of amino acid hydropathies (Kyte and Doolittle 1982). Ther-mostability means no inactivation after 30-min incubation at thetemperature indicated

creases with increasing growth temperature optimum andthat adaptation of enzymes to salt is generally associatedwith a decrease in hydrophobicity of amino acid composi-tion (Zaccai and Eisenberg 1990; Jaenicke and Závodsky

1990; Jaenicke 1987, 1991). Apparently, the hydropathiccharacter of the formyltransferase is mainly determinedby the adaptation of these enzymes to their intracellularsalt concentration rather than to temperature. It has to beconsidered, however, that the hydropathic character of asoluble protein is not only determined by temperature andsalt (Böhm and Jaenicke 1994a,b), but also by other fac-tors such as the genomic G+C content (Oshima 1988) andnutrient availability (Mazel and Marlière 1989). Thus, therelatively high lysine content of the formyltransferasefrom Methanothermus fervidus may have been influencedby the fact that the DNA of this organism has a low G+C

102

Fig.3 Alignment of the amino acid sequence of the formyltrans-ferases from Methanosarcina barkeri, from Methanobacteriumthermoautotrophicum, from Methanothermus fervidus, and fromMethanopyrus kandleri. The sequences were aligned according tothe Clustal method using the Lasergene program (DNAStar, Lon-don, England). Gaps are marked by dashes. Identical amino acidresidues occurring in all four proteins are boxed

content of only 33% and that lysine is encoded by AAA orAAC, codons low in G and C (Shima et al. 1995). An ex-ample of the effect of nutrient availability is the findingthat the phycobiliproteins induced in cyanobacteria have areduced cysteine and methionine content when deprivedof sulfur (Mazel and Marlière 1989). A better understand-ing of the differences in the properties of the formyltrans-ferases will have to await the elucidation of the tertiarystructure of at least two of the four enzymes and thedetermination of temperature-induced conformationalchanges in these proteins. For these studies we will usethe formyltransferases from Methanosarcina barkeri andfrom Methanopyrus kandleri since the differences in theproperties of these two enzymes are the most pronounced(Table 2). Heterologous expression in Escherichia colihas made these two enzymes available in large amounts.

All four formyltransferases have 46% of their aminoacid sequence in common (Fig. 3). This is noteworthysince the four organisms from which the formyltrans-ferases are derived not only grow at different tempera-tures, but also are phylogenetically very distantly related,the phylogenetic distance being largest between Me-thanosarcina barkeri and Methanopyrus kandleri andsmallest between Methanothermus fervidus and Me-thanobacterium thermoautotrophicum (Boone et al.1993). The high degree of similarity therefore indicatesthat a major part of the sequence of the formyltransferasesis determined by the functions of the enzyme, one ofwhich is to interact with the two large coenzymesmethanofuran and tetrahydromethanopterin.

The amino acid sequences of the four formyltrans-ferases show no significant sequence similarity to anyprotein in the databases. Unpublished results show, how-ever, that Methylobacterium extorquens AM1 contains anopen reading frame that encodes a 59-kDa protein, the N-terminal half of which exhibits an overall sequence iden-tity of about 43% with the formyltransferases frommethanogens (M. E. Lidstrom, A. Springer, L. Chistoser-dova). A mutant constructed in this gene by allelic re-placement was unable to grow on C1 compounds andlacked formyltetrahydrofolate synthase activity suggest-ing that this open reading frame could encode a formylte-trahydrofolate synthase. [For a review on the moleculargenetics of methylotrophic bacteria, see Lidstrom (1992)].In favor of this interpretation is the fact that tetrahy-dromethanopterin and tetrahydrofolate, which are coen-zymes of the two enzymes, are not only structurally, butalso biosynthetically closely related (White and Zhou1993). Arguing against this interpretation is the fact thatsimilarities between the amino acid sequence deducedfrom the DNA sequence of the open reading frame andthe sequences of formyltetrahydrofolate synthases fromBacteria and Eucarya were not identified. Similaritieswould have been expected since all formyltetrahydrofo-late synthases analyzed in this respect show a high degreeof sequence similarity (Staben and Rabinowitz 1986;Hum et al. 1988; Shannon and Rabinowitz 1988; White-head and Rabinowitz 1988; Lovell et al. 1990; Nour andRabinowitz 1992; Rankin et al. 1993). Nevertheless, the

observed sequence similarity between the formyltrans-ferases from methanogens and the protein encoded by theopen reading frame in Methylobacterium extorquens AM1is of interest since methanogens and Methylobacteriumextorquens AM1 both rely on an energy metabolism basedon C1-unit interconversions (Rokem and Goldberg 1991;Weiss and Thauer 1993). This is all the more interestingwhen one considers that the C-terminal half of the proteinfrom Methylobacterium extorquens AM1, which does notshow sequence similarities to the formyltransferases, ex-hibits similarities to the γ-subunit FwdC of formyl-methanofuran dehydrogenase from methanogens (Hoch-heimer et al. 1995). This enzyme catalyzes the first step inmethanogenesis from H2 and CO2. Note that the formyl-transferase catalyzes the second step (Weiss and Thauer1993).

Acknowledgements This study was supported by a grant fromthe Deutsche Forschungsgemeinschaft and by the Fonds derChemischen Industrie.

References

Argos P, Rossmann MG, Grau UM, Zuber H, Frank G, TratschinJD (1979) Thermal stability and protein structure. Biochem-istry 18:5698–5703

Ausubel FM, Brent R, Kingston RE, Moore DD, Seidmann JG,Smith JA, Struhl K (1987) Current protocols in molecular bi-ology. Wiley, New York

Böhm G, Jaenicke R (1994a) Relevance of sequence statistics forthe properties of extremophilic proteins. Int J Pept Protein Res43:97–106

Böhm G, Jaenicke R (1994b) A structure-based model for thehalophilic adaption of dihydrofolate reductase from Halobac-terium volcanii. Protein Eng 7:213–220

Bokranz M, Klein A (1987) Nucleotide sequence of the methylcoenzyme M reductase gene cluster from Methanosarcinabarkeri. Nucleic Acids Res 15:4350–4351

Boone DR, Whitman WB, Rouvière P (1993) Diversity and taxon-omy of methanogens. In: Ferry JG (ed) Methanogenesis. Chap-man & Hall, New York, pp 209–252

Bradford MM (1976) A rapid and sensitive method for the quanti-tation of microgram quantities of protein utilizing the principleof protein-dye-binding. Anal Biochem 72:248–254

Breitung J, Thauer RK (1990) Formylmethanofuran:tetrahydro-methanopterin formyltransferase from Methanosarcina barkeri.FEBS Lett 275:226–230

Breitung J, Börner G, Scholz S, Linder D, Stetter KO, Thauer RK(1992) Salt dependence, kinetic properties, and catalytic mechanism of N-formylmethanofuran:tetrahydromethanopterinformyltransferase from the extreme thermophile Methanopyruskandleri. Eur J Biochem 210:971–981

Brown JW, Daniels CJ, Reeve JN (1989) Gene structure, organi-zation, and expression in archaebacteria. Crit Rev Microbiol16:287–338

Chen EY, Seeburg PH (1985) Supercoil sequencing: a fast andsimple method for sequencing plasmid DNA. DNA 4:165–170

DiMarco AA, Sment KA, Konisky J, Wolfe RS (1990) The formyl-methanofuran:tetrahydromethanopterin formyltransferase fromMethanobacterium thermoautotrophicum ∆H. Nucleotide se-quence and functional expression of the cloned gene. J BiolChem 265:472–476

Donnelly MI, Wolfe RS (1986) The role of formylmethano-furan:tetrahydromethanopterin formyltransferase in methano-genesis from carbon dioxide. J Biol Chem 261:16653–16659

103

Escalante-Semerena JC, Rinehart KL Jr, Wolfe RS (1984a)Tetrahydromethanopterin, a carbon carrier in methanogenesis.J Biol Chem 259:9447–9455

Escalante-Semerena JC, Leigh JA, Rinehart KL Jr, Wolfe RS(1984b) Formaldehyde activation factor, tetrahydromethan-opterin, a coenzyme of methanogenesis. Proc Natl Acad SciUSA 81:1976–1980

Fischer R, Thauer RK (1989) Methyltetrahydromethanopterin asan intermediate in methanogenesis from acetate in Methano-sarcina barkeri. Arch Microbiol 151:459–465

Hain J, Reiter W-D, Hüdepohl U, Zillig W (1992) Elements of anarchaeal promoter defined by mutational analysis. NucleicAcids Res 20:5423–5428

Hensel R, König H (1988) Thermoadaptation of methanogenicbacteria by intracellular ion concentration. FEMS MicrobiolLett 49:75–79

Hochheimer A, Schmitz RA, Thauer RK, Hedderich R (1995) Thetungsten formylmethanofuran dehydrogenase from Methano-bacterium thermoautotrophicum contains sequence motifscharacteristic for molybdopterin dinucleotide containing en-zymes. Eur J Biochem (in press)

Hum DW, Bell AW, Rozeni R, MacKenzie RE (1988) Primarystructure of a human trifunctional enzyme. Isolation of a cDNAencoding methylenetetrahydromethanopterin dehydrogenase-methenyltetrahydrofolate cyclohydrolase-formyltetrahydrofo-late synthetase. J Biol Chem 263:15946–15950

Inatomi K-I, Eya M, Maeda M, Futai M (1989) Amino acid se-quence of the α and β subunits of Methanosarcina barkeri AT-Pase deduced from the cloned genes. J Biol Chem 264:10954–10959

Jaenicke R (1987) Folding and association of proteins. Prog Bio-phys Mol Biol 49:117–237

Jaenicke R (1991) Protein stability and molecular adaptation to ex-treme conditions. Eur J Biochem 202:715–728

Jaenicke R, Závodszky P (1990) Proteins under extreme physicalconditions. FEBS Lett 268:344–349

Jarrell KF, Faguy D, Hebert AM, Kalmokoff ML (1991) A generalmethod of isolating high-molecular-weight DNA frommethanogenic archaea (archaebacteria). Can J Microbiol 38:65–68

Karrasch M, Bott M, Thauer RK (1989) Carbonic anhydrase activ-ity in acetate-grown Methanosarcina barkeri. Arch Microbiol151:137–142

Kyte J, Doolittle RF (1982) A simple method for displaying thehydropathic character of a protein. J Mol Biol 157:105–132

Laemmli UK (1970) Cleavage of structural proteins during the as-sembly of the head of bacteriophage T4. Nature 227:680–685

Lehmacher A (1994) Cloning, sequencing, and transcript analy-sis of the gene encoding formylmethanofuran:tetrahydro-methanopterin formyltransferase from the hyperthermophilicMethanothermus fervidus. Mol Gen Genet 242:73–80

Leigh JA, Rinehart KL Jr, Wolfe RS (1984) Structure ofmethanofuran, the carbon dioxide reduction factor of Me-thanobacterium thermoautotrophicum. J Am Chem Soc 106:3636–3640

Leigh JA, Rinehart KL Jr, Wolfe RS (1985) Methanofuran (carbondioxide reduction factor), a formyl carrier in methane produc-tion from carbon dioxide in Methanobacterium. Biochemistry24:995–999

Lidstrom ME (1992) The genetics and molecular biology ofmethanol-utilizing bacteria. In: Murell CJ, Dalton H (eds)Biotechnology handbooks, vol 5. Methane and methanol utiliz-ers. Plenum Press, New York, pp 183–206

Lovell CR, Przybyla A, Ljungdahl LG (1990) Primary structureof the thermostable formyltetrahydrofolate synthetase fromClostridium thermoaceticum. Biochemistry 29:5687–5694

Massanz C, Michel R, Friedrich B, Fiebig K (1993) Cloning of thestructural genes of the coenzyme F420-reactive hydrogenasefrom the methanogenic archaeon Methanosarcina barkeriFusaro. BioEngineering 1:42

Mazel D, Marlière P (1989) Adaptive eradication of methionineand cysteine from cyanobacterial light-harvesting proteins. Na-ture 341:245–248

Menéndez-Arias L, Argos P (1989) Engineering protein thermalstability. Sequence statistics point to residue substitutions in α-helices. J Mol Biol 206:397–406

Morris CJ, Reeve JN (1988) Conservation of structure in the hu-man gene encoding argininosuccinate synthetase and the argGgenes of the archaebacteria Methanosarcina barkeri MS andMethanococcus vannielii. J Bacteriol 170:3125–3130

Nour JM, Rabinowitz JC (1992) Isolation and sequencing of thecDNA coding for spinach 10-formyltetrahydrofolate syn-thetase. Camparisons with the yeast, mammalian, and bacterialproteins. J Biol Chem 267:16292–16296

Oshima T (1988) Current results on the studies of proteins fromthermophiles. Cell Technol 7: 513–518

Østergaard L, Larsen N, Lefers H, Kjems J, Garrett R (1987) A ri-bosomal RNA operon and its flanking region from the Archae-bacterium Methanobacterium thermoautotrophicum, Marburgstrain: transcription signals, RNA structure and evolutionaryimplications. Syst Appl Microbiol 9:199–209

Rankin CA, Haslam GC, Himes RH (1993) Sequence and expres-sion of the gene for N10-formyltetrahydrofolate synthetasefrom Clostridium cylindrosporum. Protein Sci 2:197–205

Reeve JN (1992) Molecular biology of methanogens. Annu RevMicrobiol 46:165–191

Rokem JS, Goldberg I (1991) Oxidation pathways in methy-lotrophs. In: Goldberg I, Rokem JS (eds) Biology of methy-lotrophs. Butterworth-Heinemann, Boston, pp 111–126

Sambrook J, Fritsch EF, Maniatis T (1989) Molecular cloning, alaboratory manual, 2nd edn Cold Spring Harbor LaboratoryPress, Cold Spring Harbor, NY

Sanger F, Nicklen S, Coulson AR (1977) DNA sequencing withchain terminating inhibitors. Proc Natl Acad Sci USA74:5463–5467

Shannon KW, Rabinowitz JC (1988) Isolation and characterizationof the Saccharomyces cerevisiae MIS1 gene encoding mito-chondrial C1-tetrahydrofolate synthase. J Biol Chem 263:7717–7725

Shima S, Weiss DS, Thauer RK (1995) Formylmethanofuran:tetrahydromethanopterin formyltransferase (Ftr) from the hy-perthermophilic Methanopyrus kandleri: cloning, sequencingand functional expression of the ftr gene and one-step purifica-tion of the enzyme overproduced in Escherichia coli. Eur JBiochem 230:906–913

Short JM, Sorge JA (1992) In vivo excision properties of bacterio-phage λ ZAP expression vectors. Methods Enzymol 216:495–508

Staben C, Rabinowitz JC (1986) Nucleotide sequence of the Sac-charomyces cerevisiae ADE3 gene encoding C1-tetrahydrofo-late synthase. J Biol Chem 261:4629–4637

Studier FW, Rosenberg AH, Dunn JJ, Dubendorf JW (1990) Useof T7 RNA polymerase to direct expression of cloned genes.Methods Enzymol 185:60–89

Thauer RK, Hedderich R, Fischer R (1993) Biochemistry. Reac-tions and enzymes involved in methanogenesis from CO2 andH2. In: Ferry JG (ed) Methanogenesis. Chapman & Hall, NewYork, pp 209–252

Thomm M, Hausner W, Hethke C (1994) Transcription factors andtermination of transcription in Methanococcus. Syst Appl Mi-crobiol 16:648–655

Vorholt J, Kunow J, Stetter KO, Thauer RK (1995) Enzymes andcoenzymes of the carbon monoxide dehydrogenase pathwayfor autotrophic CO2 fixation in Archaeoglobus lithotrophicusand the lack of carbon monoxide dehydrogenase in the het-erotrophic A. profundus. Arch Microbiol 163:112–118

Weiss DS, Thauer RK (1993) Methanogenesis and the unity ofbiochemistry. Cell 72:819–822

White RH, Zhou D (1993) Biosynthesis of the coenzymes inmethanogens. In: Ferry JG (ed) Methanogenesis. Chapman &Hall, New York, pp 209–252

104

Whitehead TR, Rabinowitz JC (1988) Nucleotide sequence of theClostridium acidiurici (“Clostridium acidi-urici”) gene for 10-formyltetrahydrofolate synthase shows extensive amino acidhomology with the trifunctional enzyme C1-tetrahydrofolatesynthase from Saccharomyces cerevisiae. J Bacteriol 170:3255–3261

Wolfe RS (1991) My kind of biology. Annu Rev Microbiol 45:1–35

Zaccai G, Eisenberg H (1990) Halophilic proteins and the influ-ence of solvent on protein stabilization. Trends Biochem Sci15:333–337

Zinder SH (1993) Physiological ecology of methanogens. In: FerryJG (ed) Methanogenesis. Chapman & Hall, New York, pp209–252

Zwickl P, Fabry S, Bogedain C, Haas A, Hensel R (1990) Glycer-aldehyde-3-phospate dehydrogenase from the hyperther-mophilic archaebacterium Pyrococcus woesei: characterizationof the enzyme, cloning and sequencing of the gene, and ex-pression in Escherichia coli. J Bacteriol 172:4329–4338

105