In Methanobacterium thermoautotrophicum, the fmdECB operon encoding the molybdenum formylmethanofuran dehydrogenase is directly preceded by an open reading frame tfx predicted to encode a DNA binding protein. The 16.1 kDa protein has an N-terminal basic domain with a helix–turn–helix motif for DNA binding and a C-terminal acidic domain possibly for transcriptional activation. We report here on the DNA binding properties of the Tfx protein heterologously overproduced in Escherichia coli. Tfx was found to bind specifically to a DNA sequence downstream of the promoter of the fmdECB operon, as shown by electrophoretic mobility shift assays and DNase I footprint analysis. Northern blot hybridizations revealed that transcription of tfx is repressed during the growth of M. thermoautotrophicum in the presence of tungstate. Based on its structure and properties, the DNA binding protein Tfx is proposed to be a transcriptional regulator composed of a basic DNA binding domain and an acidic activation domain.
Transcription in archaea is more closely related to transcription in eukarya than to that in bacteria (Baumann et al., 1995; Reeve et al., 1997). The archaeal RNA polymerase is as complex as the eukaryotic RNA polymerase (pol) I, II and III and is composed of 8–13 different subunits (Zillig et al., 1993; Langer et al., 1995), whereas the eubacterial core RNA polymerase consists of only three subunits. Most archaeal genes are preceded by a TATA box-like A-box sequence TTTA[A/T]A containing a promoter that is located 27–30 bp upstream of a weakly conserved initiator motif with the transcription start site (Reiter et al., 1990; Palmer and Daniels, 1995). Therefore, archaeal promoters are more similar to eukaryotic pol II promoters containing a TATA box located around 30 bp upstream of the transcription start site and/or a weakly conserved initiator sequence (reviewed in Novina and Roy, 1996). In contrast, most of the eubacterial promoters are composed of the consensus sequences TATA and TTGACA, located at −10 and −35 respectively. The archaeal homologue of eukaryotic TBP (TATA binding protein) and the TFIIB homologue TFB have been identified from several archaea (Ouzounis and Sander, 1992; Creti et al., 1993; Marsh et al., 1994; Rowlands et al., 1994; Qureshi et al., 1995a,b; Bultet al., 1996; Smith et al., 1997; Soppa and Link, 1997). Besides RNA polymerase, the general transcription factors TBP and the TFIIB homologue TFB are minimally required for transcription (Thomm, 1996; Qureshi et al., 1997). Given these parallels with the eukaryotic pol II system, it is tempting to speculate that transcriptional regulation in archaea also mainly depends on transcriptional regulators that bind adjacent and/or distal to the core promoter region and regulate promoter strength. However, the great variety of sequence-specific, DNA binding transcriptional regulators, as found in eukarya, has still to be demonstrated.
Several regulated systems of transcription have been described in archaea, but the details remain largely unknown at the molecular level; in particular, knowledge about DNA binding regulators of transcription is still scarce. There are a few examples of DNA binding activities in archaea that suggest that transcriptional regulation is similar to that in bacteria. In the cytoplasmic fraction of Methanococcus maripaludis, Cohen-Kupiec et al. (1997) reported DNA binding activity to a palindromic repressor binding site associated with the transcription start site regulating nif gene transcription. Another example of a repressor binding system is a repressor involved in regulating the lysogeny of the ϕH prophage in Halobacterium halobium binding to several operator sites upstream of the lysis genes, although lytic development is also regulated by antisense RNA (Stolt and Zillig, 1993; 1994; see Discussion for further examples).
There are, however, also indications that transcriptional regulation in archaea is different from that in bacteria. Recently, Krueger et al. (1998) reported that a transcriptional activator involved in the regulation of gas vesicle synthesis in halophilic archaea resembles basic leucine zipper proteins typically involved in eukaryotic gene regulation. The TFIIB homologue TFB from Sulfolobus shibatae mediates sequence-specific interaction of TFB and TBP with DNA flanking the A-box of the strong S. shibatae viral T6 promoter. This sequence directly upstream of the A-box influences promoter strength. If this sequence is placed upstream of the much weaker rRNA promoter, this promoter is as active as the T6 promoter (Qureshi and Jackson, 1998). This suggests that sequence-specific interactions of the general transcription factors TFB and TBP with promoter DNA are strong determinants of promoter strength in archaea. Similarly in Haloferax volcanii, it has been shown that regulatory sequences involved in the heat-induced transcription of the cct1 gene are located within the A-box and closely adjacent sequences upstream and downstream of the A-box (Thompson and Daniels, 1998). The gene products of the brp and bat genes in Halobacterium halobium are required for the transcription of the bacterio-opsin-encoding bop gene and might function as positive activators. The details are still not clear, but bop transcription has been shown to be sensitive to DNA supercoiling and suggests an involvement of a non-B-DNA structure (Yang et al., 1996).
During studies of the genes encoding the tungsten formylmethanofuran dehydrogenase Fwd (Hochheimer et al., 1995) and of the molybdenum formylmethanofuran dehydrogenase Fmd (Hochheimer et al., 1996) and their transcriptional regulation (Hochheimer et al., 1998) in the thermophilic archaeon Methanobacterium thermoautotrophicum, we found an open reading frame tfx located directly upstream of the fmdECB operon (Fig. 1A), which encodes a putative DNA binding protein Tfx (Fig. 1B). The 16.1 kDa protein is predicted to have an isoelectric point (pI) of 9.6 and to be composed of an N-terminal basic (pI = 11.05) DNA binding domain with a helix–turn–helix motif and a C-terminal acidic (pI = 5.53) domain. The putative DNA binding protein thus has a domain structure resembling that of a group of eukaryotic transcriptional regulators such as GAL4, p53 or VP-16 composed of a DNA binding domain (reviewed in Nelson, 1995) and an acidic domain for transcriptional activation (reviewed in Triezenberg, 1995; Ptashne and Gann, 1997). It should be noted that some bacterial transcription factors display a distinct domain structure too, e.g. many response regulators from two-component regulatory systems (reviewed in Hakenbeck and Stock, 1996), and some bacterial regulators can also use acidic domains for transcriptional activation (Ptashne and Gann, 1997).
In this paper, we report on our studies on the properties of Tfx from M. thermoautotrophicum and provide evidence that Tfx might represent a eukaryotic-type transcriptional activator in this archaeon.
Heterologous overexpression of tfx and purification of the overproduced Tfx protein
Most of the strongly basic DNA binding proteins are almost insoluble at pH 7.0. Therefore, the gene encoding the 16.1 kDa basic Tfx protein (pI = 9.6) was fused to the 3′ end of the malE gene encoding the neutral 42.7 kDa maltose binding protein for heterologous expression in E. coli. The fusion was designed with a factor Xa cleavage site, yielding the protein Tfx without an N-terminal modification after cleavage of the fusion protein. The transformed E. coli cells were found to overproduce the 58.8 kDa fusion protein, which was recovered in the soluble fraction (not shown). After purification by affinity chromatography on an amylose resin, the 58.8 kDa MBP-Tfx was obtained (Fig. 2, lane 1). Approximately 15 mg of MBP-Tfx was obtained from 2 g of transformed E. coli cells (wet mass). MBP-Tfx could be cleaved proteolytically by endoproteinase factor Xa into the 42.7 kDa maltose binding protein and the 16.1 kDa Tfx, which showed an apparent molecular mass of 18 kDa in SDS–PAGE (Fig. 2, lane 2). The insoluble Tfx was separated from the soluble maltose binding protein and soluble non-cleaved fusion protein by centrifugation (Fig. 2, lanes 3 and 4). The Tfx in the pellet was dissolved in 6 M guanidinium hydrochloride and renatured by stepwise dialysis against 50 mM glycine–HCl, pH 3.0, containing 0.1% IGEPAL C-630 and guanidinium hydrochloride in decreasing concentrations.
After dialysis, part of the Tfx was still denatured and insoluble. If the dialysis was not performed stepwise and if the detergent IGEPAL C-630 was omitted from the renaturation buffer, the yield of renatured Tfx was decreased.
The renatured Tfx was not very soluble in 50 mM glycine–HCl, pH 3.0, containing 0.1% IGEPAL C-630. At concentrations higher than 0.4 mg ml−1, the protein precipitated irreversibly.
Binding of Tfx to the fmdECB promoter region
The gene tfx is located directly upstream of the fmdECB operon encoding the molybdenum formylmethanofuran dehydrogenase (Fig. 1A). The expression of the fmd operon is transcriptionally regulated by molybdate. We therefore tested whether Tfx binds to the fmdECB promoter region. A 630 bp DNA fragment was analysed, which comprised the A-box containing fmd promoter and 270 bp DNA upstream and 330 bp DNA downstream of the fmd promoter. The 5′ end of the fragment comprised the 3′ end of the tfx gene and the 3′ end of the fragment the 5′ end of the fmdE gene (Fig. 1A). The binding of Tfx to the 630 bp fragment was determined by electrophoretic mobility shift analysis (Fig. 3).
In Fig. 3, lane1, the migration distance during electrophoresis of the 630 bp fmd promoter region in the absence of Tfx is shown. Upon addition of Tfx, the migration of the DNA was retarded. At the lowest Tfx concentration, only one retardation band was observed. At higher concentrations, additional bands appeared, the migration distance of the bands decreasing with increasing Tfx concentrations. The result is interpreted as indicating that Tfx binds to the fmd promoter region and that, at high Tfx concentrations, the stoichiometry of Tfx binding increased. Whether the increase in stoichiometry is caused by multiple binding sites or oligomerization of Tfx cannot be deduced from the experiment.
Molybdate and tungstate up to 1 mM concentration had no effect on Tfx binding to the 630 bp promoter region.
The binding of Tfx to other promoter regions was also investigated. The regions tested were those of the fwd operon encoding the tungsten formylmethanofuran dehydrogenase, the nif operon encoding the nitrogenase proteins and tfx encoding the protein under investigation (data not shown). Tfx did not bind to these promoter regions. Apparently, of the four promoter regions tested, Tfx binds specifically only to the fmd promoter region.
The Tfx binding site within the fmd promoter region was narrowed down by analysing three fragments from the region. Fragment A comprises the 257 bp downstream of the fmd promoter, fragment C the 248 bp upstream of the promoter and fragment B the 248 bp upstream of the fmd promoter plus 129 bp containing the promoter (see Fig. 1A). Electrophoretic mobility shift analysis revealed that Tfx only binds to fragment A at the given concentrations. The stoichiometry of binding increased with increasing Tfx concentrations (Fig. 4, lanes 1–4).
The Tfx binding site
To narrow down the Tfx binding site within fragment A of the fmd promoter region, DNase I footprint experiments were performed with Tfx (Fig. 5). Within the 257 bp fragment, a sequence of 42 bp, 134 bp downstream from the start codon of fmdE, 167 bp downstream of the fmd transcription start site and 192 bp downstream of the promoter was efficiently protected from DNase I cleavage by Tfx. The DNA sequence of the protected Tfx binding region is shown in Fig. 6.
The 42 bp region within fragment A identified by DNase I footprint mapping was analysed for Tfx binding in electrophoretic mobility shift experiments. In accordance with the footprint analysis, Tfx was found to bind very efficiently to the 42-mer (Fig. 7A).
The DNase I footprint analysis (Fig. 5) indicates the sites within the 42-mer most efficiently protected from DNase I cleavage. These sites were mutated (Fig. 6) and the mutated 42-mer tested again for Tfx binding by electrophoretic mobility shift analysis. Tfx was found to bind to the mutated 42-mer at least at 20-fold higher concentrations (Fig. 7B).
Monocistronic transcription of tfx and transcriptional regulation
Northern blot analysis revealed that the tfx gene was transcribed monocistronically. With a hybridization probe specific for the tfx gene, a 0.5 kb transcript was identified that conforms in length to that expected for the tfx gene and the transcription start site (Fig. 8A). The latter was determined by primer extension analysis (Fig. 8B) and was found to be located 82 bp upstream of the translational start codon, which is preceded by an appropriate ribosome binding site (Reeve, 1992), and 25 bp downstream of an AT-rich sequence AAATCCAAAAATT only imperfectly conforming to the consensus sequence of the A-box of methanogenic archaea (Reeve, 1992; Palmer and Daniels, 1995). Directly downstream of the tfx gene, no typical terminator sequence of transcription is found, except an oligo(dT) stretch within the promoter of the fmdECB operon, which is preceded directly by the stop codon of the tfx gene.
As transcription of the fmd operon is positively regulated by molybdate (Hochheimer et al., 1998), we tested whether the transcription of tfx underlies the same regulation. The transcriptional regulation of tfx, however, turned out to be different. Northern blot analysis of total RNA extracted from Methanobacterium thermoautotrophicum cells grown in the presence of molybdate and/or tungstate (Fig. 8A) revealed that the tfx transcript was predominant only in molybdate-grown cells (Fig. 8A, lane 1) and hardly detectable in both tungstate-grown cells (Fig. 8A, lane 2) and molybdate plus tungstate-grown cells (Fig. 8A, lane 3). As molybdate did not induce tfx transcription during growth in the presence of tungstate (Fig. 8A, lane 3), it has to be concluded that tfx transcription is not affected by molybdate but is repressed by tungstate. Unfortunately, it is impossible to grow M. thermoautotrophicum in the absence of both molybdate and tungstate in order to show this directly.
In 8Fig. 8A, the transcription of the hmd gene in M. thermoautotrophicum is shown as a control. Hmd is a metal-free hydrogenase (Thauer et al., 1996), the synthesis of which is not expected to be regulated by molybdate or tungstate. The concentration differences in the hmd transcript (Fig. 6A, lanes 4–6) are not considered to be significant.
Tfx was shown to be a DNA binding protein that bound specifically to a 42 bp DNA sequence 192 bp downstream of the promoter of the fmdECB operon within the fmdE gene. Tfx did not bind to the A-box of the fmd promoter or closely adjacent sequence elements. From these findings, it is concluded that Tfx does not belong to the group of general transcription factors TBP and TFB, which both bind to the A-box of the promoter and closely adjacent sequence elements (Rowlands et al., 1994; Danner and Soppa, 1996; Hausner et al., 1996; Kosa et al., 1997; Qureshi and Jackson, 1998). Sequence comparisons with TBP and TFB indicate further that Tfx is presumably not a general transcription factor. Tfx is structurally related to neither TBP nor TFB. Together with RNA polymerase, both TBP and TFB are minimally required for transcription initiation in archaea and are phylogenetically closely related to their eukaryotic counterparts TBP and TFIIB respectively (Thomm, 1996; Qureshi et al., 1997; Soppa and Link, 1997).
Besides TBP and TFB, other DNA binding proteins have also been described in archaea. Among these are the archaeal histones, a homologue of the bacterial leucine-responsive global transcription regulator Lrp in Sulfolobus solfataricus (Charlier et al., 1997), a repressor protein involved in regulating lysogeny of the ϕH prophage in Halobacterium halobium by a phage λ-like system (Stolt and Zillig, 1993; 1994) and DNA binding activities in cell extracts of Methanococcus maripaludis (Cohen-Kupiec et al., 1997) and Methanosarcina barkeri (Chien et al., 1998). The archaeal DNA–histone complexes and nucleosome assembly in vivo and in vitro have been studied extensively (reviewed in Reeve et al., 1997). The DNA binding of histones is not sequence specific. The properties of the Lrp homologue can only be deduced from the DNA sequence as the protein has not yet been isolated. This is also true for the halobacterial repressor protein and for the cytoplasmic proteins binding to the nif promoter region in Methanosarcina barkeri and Methanococcus maripaludis.
As shown in 1Fig. 1B, Tfx is composed of two domains, a basic domain with a helix–turn–helix motif and an acidic domain. The basic domain is considered to be the domain binding specifically to the 42 bp DNA sequence 192 bp downstream of the fmd promoter. Interestingly, the 42 bp DNA sequence lacks obvious inverted repeats frequently present in DNA binding sites for proteins with a helix–turn–helix motif. Whether Tfx binds as a monomer, dimer or tetramer to its DNA binding site cannot be unambiguously concluded from the experiments. However, the electrophoretic mobility shift assays (Figs 3 and 4) indicate that, at the Tfx concentrations used in the DNase I footprint experiments (Fig. 5), more than 1 mol of Tfx binds to the target DNA. The determination of the oligomeric state of Tfx by ultracentrifugation was severely hampered by the low solubility of Tfx (< 0.1 mg ml−1) at pH 7.5, the pH of the electrophoretic mobility shift assays. The function of the acidic domain can only be deduced at present from a comparison with a group of eukaryotic DNA binding proteins containing an acidic domain as well as a DNA binding domain. In these DNA binding proteins, which are all transcriptional activators, the acidic domain has a function in transcriptional activation, and some acidic activation domains have been shown to interact with general transcription factors such as TFIID (containing TBP) and TFIIB (reviewed in Triezenberg, 1995; Ptashne and Gann, 1997). TBP from Pyrococcus woesei has been shown to interact specifically with eukaryotic activation domains derived from the adenovirus E1A protein and from p53, which contains an acidic activation domain (Rowlands et al., 1994). These findings suggest that archaeal and eukaryotic TBPs are functionally analogous and further suggest that eukaryotic-type transcriptional activators are present in archaea, possibly interacting with TBP and/or TFB.
The finding that the Tfx binding site is located downstream of the fmd promoter within the fmdE gene suggests that Tfx might function in the transcriptional regulation of the fmd operon. This is substantiated by the result that the transcription of the fmd operon is induced by molybdate (Hochheimer et al., 1998) and that of the tfx gene is repressed by tungstate (Fig. 8A), which is required for the growth of M. thermoautotrophicum in the absence of molybdate. The finding that the transcription of tfx is repressed by tungstate rather than induced by molybdate (Fig. 8A) indicates that the transcriptional regulation of the fmd operon cannot solely involve Tfx, if Tfx is involved. Interestingly, in the fwdHFGDACB operon encoding the tungsten-containing isoenzyme of formylmethanofuran dehydrogenase, the fwdH gene encodes a basic iron–sulphur protein with a helix–turn–helix motif. The putative DNA binding protein does not co-purify with the enzyme and could, therefore, be a further component involved in the transcriptional control of the formylmethanofuran dehydrogenase isoenzymes in M. thermoautotrophicum.
Unfortunately, a genetic system for Methanobacterium thermoautotrophicum is not yet available allowing in vivo testing of the influence of Tfx on transcriptional activation at the fmd promoter region using a suitable reporter gene system. An in vitro test is hampered by the requirement of purified RNA polymerase, TBP and TFB from M. thermoautotrophicum. However, in view of the relatedness of the archaeal and the eukaryotic transcription machinery, an in vivo test of the Tfx function in a eukaryotic system should be possible. In the yeast Saccharomyces cerevisiae, heterologous transactivation assays for the function of eukaryotic transcriptional activators are commonly performed. Whether this system can also be used to test the function of archaeal transcriptional activators is under study.
A search in the complete genome sequence of M. thermoautotrophicum (Smith et al., 1997) revealed that the open reading frame tfx is not the only example of an open reading frame encoding a putative transcriptional regulator in this archaeon. In the genome of the archaeon Methanococcus jannaschii (Bult et al., 1996), fewer open reading frames than in M. thermoautotrophicum were identified that might encode putative transcriptional regulators. However, at least one open reading frame is present encoding a protein with 60% sequence identity to Tfx from M. thermoautotrophicum.
Materials, bacterial strains and plasmids
Methanobacterium thermoautotrophicum strain Marburg (DSM 2133) was from the Deutsche Sammlung für Mikroorganismen (Braunschweig, Germany). The archaeon was grown at 65°C on 80% H2, 20% CO2, 0.1% H2S in a completely mineral salt medium (Schönheit et al., 1980). E. coli strain DH5α (Hanahan, 1983) was used for plasmid amplification and overproduction of the MBP-Tfx fusion protein. E. coli strain DH5α was cultured on Luria–Bertani medium at 37°C (Sambrook et al., 1989). Ampicillin (Sigma) was used at concentrations of 100 μg ml−1. Taq DNA polymerase, RNasin (RNase inhibitor), avian myeloblastosis virus reverse transcriptase, digoxigenin polymerase chain reaction (PCR) labelling mix, digoxigenin luminescent detection kit for nucleic acids, DNase I, endoproteinase factor Xa and poly-(dI–dC) were purchased from Boehringer Mannheim. Pfu DNA polymerase and pBluescript KS+ were from Stratagene, and sonicated salmon sperm DNA from Gibco BRL. All DNA-modifying enzymes and the DNA sequencing kit were from Amersham. Synthetic oligonucleotides were obtained from MWG-Biotech, and the protein fusion and purification system was from New England Biolabs.
DNA manipulations and DNA sequence determination
Routine DNA manipulations were performed as described by Sambrook et al. (1989). Genomic DNA from Methanobacterium thermoautotrophicum was isolated as described by Kiener et al. (1987). Plasmid DNA was purified according to the method described by the Qiagen Plasmid Handbook. Nucleotide sequences were determined by the dideoxy-chain termination method of Sanger et al. (1977).
RNA isolation and Northern blot hybridization
For RNA isolation, M. thermoautotrophicum strain Marburg was grown on a molybdate-depleted medium supplemented with either 1 μM Na2WO4, 1 μM Na2MoO4 or 1 μM Na2WO4 plus 1 μM Na2MoO4. A 500 ml culture, grown to a ΔA578 of 2, was cooled in an ethanol–solid CO2 mixture and harvested by centrifugation at 4000 × g in a rotor cooled to 4°C. After discarding the supernatant, the pellet was transferred to a mortar precooled with liquid nitrogen. The cells were ruptured by grounding in liquid nitrogen for 15 min. RNA was isolated from the ruptured cells using the single-step method for RNA isolation according to Ausubel et al. (1994).
The procedure for generating Northern blots was that described by Ausubel et al. (1994). RNA was denatured by glyoxal–dimethyl sulphoxide treatment. Hybridization with digoxigenin-labelled DNA probes was performed at 50°C and pH 7.0 in 0.75 M NaCl, 0.075 M sodium citrate, 0.1% SDS. The subsequent stringent washing procedure was performed at 68°C in 0.015 M NaCl, 0.0015 M sodium citrate, 0.1% SDS after an initial washing step at room temperature in 0.3 M NaCl, 0.03 M sodium citrate, 0.1% SDS. The digoxigenin-labelled DNA probes were generated by PCR using the digoxigenin labelling mixture (Boehringer Mannheim) and pairs of primers derived from the tfx gene (primer 1: 5′-CTGAGTAAAAAAACTTTCCTAACCGAAAGA-3′ and primer 2: 5′-AAATTTTTTTATGAATTCTAGGAATTTAGATCG-3′) and the hmd gene (primer 1: 5′-GGGCCTGTGAAGTCGCTG-3′ and primer 2: 5′-GCAACCGCTTCCTTGTCG-3′).
Mapping of the 5′ end of the tfx mRNA
The 5′ end of the mRNA encoding the tfx gene was mapped using the primer extension method (Boorstein and Craig, 1989) with two oligonucleotide primers complementary to the 5′-end of the tfx gene (primer 1: 5′-ATTATCCCTCCTAGCTCG-3′, primer 2: 5′-CCCTCATTTCAAGAACGG-3′). RNA from molybdate-grown cells (20 μg) was incubated with 0.5 pmol [γ-32P]-dATP-labelled primer. After annealing by slowly cooling from 70°C to 42°C, the primer–RNA hybrids were extended with 10 U avian myeloblastosis virus reverse transcriptase in the presence of 20 U RNase inhibitor and 0.15 mM dNTPs at 42°C for 30 min. The extension products were purified by phenol extraction, subjected to denaturing polyacrylamide gel electrophoresis and visualized by autoradiography.
Heterologous overproduction of Tfx in E. coli and purification
The tfx gene was amplified with Pfu DNA polymerase. The following two primers were used: 5′-CTGAGTAAAAAAACTTTCCTAACCGAAAGA-3′ starting with the tfx start codon and 5′-AAATTTTTTTATGAATTCTAGGAATTTAGATCG-3′ introducing an EcoRI restriction site. The 437 bp PCR product was cleaved with EcoRI and ligated into the plasmid pMAL-c2 (New England Biolabs), linearized with XmnI and EcoRI, yielding a plasmid with the tfx gene fused to the 3′ end of the malE gene from E. coli encoding the maltose binding protein (MBP, 42.7 kDa). The gene fusion was designed so that the resulting fusion protein contained a factor Xa cleavage site and, after cleavage of the fusion protein, Tfx without any change in the amino acid sequence was obtained. The sequence of mbp–tfx was verified by DNA sequencing. For overexpression, E. coli strain DH5α was transformed and grown in 1 l of Luria–Bertani medium containing 0.2% glucose and ampicillin (100 μg ml−1) to a ΔA578 of 0.5, induced with IPTG (0.3 mM final concentration) and grown further for 2 h at 37°C. The cells were harvested by centrifugation and resuspended in 250 ml of 20 mM Tris-HCl, pH 7.4, 200 mM NaCl, 1 mM EDTA and disrupted by sonification. After removal of cell debris by centrifugation, the MBP–Tfx fusion protein (58.8 kDa) was purified from the cell extract by affinity chromatography on an amylose resin according to a protocol from the manufacturer (New England Biolabs). After elution, the protein was concentrated using an Amicon chamber with an exclusion limit of 30 kDa. In order to cleave the MBP (42.7 kDa) from Tfx (16.1 kDa), 1.0 mg of MBP–Tfx was incubated at pH 8.0 with 50 μg factor of Xa for several days at room temperature. The progress cleavage was followed by 16% SDS–PAGE. The released Tfx was nearly insoluble at pH 8, precipitated and could be separated from the soluble MBP–Tfx, MBP and factor Xa by centrifugation. The pellet was solubilized in 6 M guanidinium hydrochloride, and the solution was then dialysed against 50 mM glycine–HCl, pH 3.0, containing 0.1% IGEPAL C-630 and guanidinium hydrochloride at concentrations of 3 M, 1 M, 0.5 M, 0.25 M, 0.1 M and 0 M, which were reduced stepwise. Precipitated Tfx was removed by centrifugation. Soluble Tfx in the supernatant was stored at −20°C after the addition of glycerol and phenylmethylsulphonyl fluoride (PMSF) to final concentrations of 10% and 0.6 mM respectively.
Preparation of DNA for DNA binding assays
DNA fragments for electrophoretic mobility shift analysis were amplified from genomic DNA from M. thermoautotrophicum by PCR using specific oligonucleotide primers. fmd promoter region (630 bp): 5′-AAATCCCGGAACACTGTCGACTTTGTTAAATTCCTAAAATCACCTGTA-3′ and 5′-GACGGTCTTTTCAGGGTCGACTATTATTCTAAGAGCCTGGACTGT-3′. tfx promoter region (599 bp): 5′-TAGGGCAGACACAACGTCGACAGAAAGCCTTGGAAAGTGGTTCTC-3′ and 5′-TAGGAATTTAACAAAGTCGACAGTGTTCCGGGATTTCTCTATGTT-3′. fwd promoter region (534 bp): 5′-GCCCCTGCAAAT-GTCGACACAGGCGGAGAC-3′ and 5′-CTCTGCTGGCATGTCGACACTATCGTCCTT-3′. nif promoter region (512 bp): 5′-ATGAAGAAGGTTGTCGACTCAGCCTCCTCT-3′ and 5′-TCCATCATGGTTGTCGACATTTTACCGCCC-3′. For the fragments A, B and C of the 630 bp fmd promoter region, the following pairs of oligonucleotide primers were used. Fragment A (257 bp): 5′-AATTTCAGTGAATTCGCAAGCCCGGGC-3′ and 5′-GACGGTCTTTTCAGGGTCGACTATTATTCTAAGAGCCTGGACTGT-3′. Fragment B (377 bp): 5′-AAATCCCGGAACACTGTCGACTTTGTTAAATTCCTAAAATCACCTGTA-3′ and 5′-GCCCGGGCTTGCGAATTCACTGAAATT-3′. Fragment C (248 bp): 5′-AAATCCCGGAACACTGTCGACTTTGTTAAATTCCTAAAATCACCTGTA-3′ and 5′-AAATTCTAGGAATTCAGATCGACGATA-3′. The introduced restriction sites for Sal I are underlined, and those for EcoRI are underlined and italic. The purified PCR products were cleaved with Sal I or EcoRI and radiolabelled with [α-32P]-dATP by Klenow fill-in. Oligonucleotides for electrophoretic mobility shift analysis were annealed by slow cooling from 95°C to room temperature and labelled with [γ-32P]-dATP by T4 polynucleotide kinase. Oligonucleotides for the 42 bp Tfx binding site: 5′-TGTTACATGTGAAACGACAAACTGTTTACCCGACGCATTCCA-3′ and 5′-TGGAATGCGTCGGGTAAACAGTTTGTCGTTTCACATGTAACA-3′. Oligonucleotides for the mutated 42 bp mutated Tfx binding site: 5′-TGTTACATCAGACCCTCCAAACGGTTTACCATACGCCGGCCA-3′ and 5′-TGGCCGGCGTATGGTAAACCGTTTGGAGGGTCTGATGTAACA-3′.
For DNase I footprint analysis, the 257 bp fragment A of the 630 bp fmd promoter region, cleaved with Sal I and EcoRI, was ligated in pBluescript KS+ and linearized with Sal I and EcoRI. Fragment A was radiolabelled at one end by Klenow fill-in of the EcoRI restriction site with [α-32P]-dATP. The radiolabelled fragment A was released from the plasmid by cleavage with Sal I and purified by agarose gel electrophoresis with subsequent excision of the radiolabelled fragment A.
Electrophoretic mobility shift assays
Radiolabelled DNA fragments (3000 c.p.m.; 0.3–1.0 ng) were incubated with purified Tfx in 10 mM Tris-HCl, pH 7.5, 100 mM KCl, 5% glycerol, 1 mM EDTA and 0.1% IGEPAL C-630 containing 2 μg of poly-(dI–dC) in a total volume of 20 μl. The binding reactions were started by the addition of Tfx. The samples were incubated for 15 min at 60°C and analysed completely on non-denaturing 5% polyacrylamide gels (acrylamide–bisacrylamide weight ratio of 37.5:1) containing 0.5 × Tris borate–EDTA (TBE) buffer. The gels (8 cm × 7 cm) were pre-electrophoresed at 100 V for 30 min and, subsequently, the protein–DNA complexes were separated at 30 mA at room temperature for 30–75 min. The gels were covered with Saran wrap (Dow Chemical) and autoradiographed.
DNase I footprint analysis
For the binding reaction, 10 ng of the radiolabelled DNA fragment (20 000 c.p.m.) was incubated at 60°C for 15 min with purified Tfx in 20 μl of 10 mM Tris-HCl, pH 7.5, 100 mM KCl, 5% glycerol, 1 mM EDTA and 0.1% IGEPAL C-630 containing 2 μg of poly-(dI–dC) (buffer A). Then, 30 μl of a solution of 8.35 mM MgCl2 and 1.67 mM CaCl2 in buffer A was added and, after adjustment to room temperature, 1 μl of DNase I solution (20 U of DNase I in 1 ml of 10 mM Tris-HCl, pH 7.5, 50 mM KCl, 5 mM MgCl2, 1 mM CaCl2 and 5% glycerol). After incubation at room temperature for exactly 1 min, DNase I cleavage was stopped by the addition of 50 μl of 0.2 M NaCl containing 25 mM EDTA, 1% SDS and 1 μg of sonicated salmon sperm DNA. After phenol extraction, the DNA was precipitated with ethanol, resuspended in formamide loading buffer (Sambrook et al., 1989) and analysed on a 6% sequencing gel. The DNase I footprint was visualized by autoradiography.
Sequence accession numbers
The sequence data referred to in this manuscript are deposited with the EMBL Sequence Data Bank and are available under accession numbers AJ009686 (tfx ), X97820 (fmdECB ) and X87970 (fwdEFGDACB ).
*Present address: Howard Hughes Medical Institute, Department of Molecular and Cell Biology, University of California, Berkeley, CA 94720-3204, USA
This work was supported by the Max-Planck-Gesellschaft, by the Deutsche Forschungsgemeinschaft and by the Fonds der Chemischen Industrie.