Expression and heat-responsive regulation of a TFIIB homologue from the archaeon Haloferax volcanii


  • Dorothea K. Thompson,

    1. Department of Microbiology, The Ohio State University, 484 West 12th Avenue, Columbus, OH 43210, USA.
    Search for more papers by this author
    • *Present address: FDA, Center for Biologics Evaluation and Research, Division of Bacterial Products, HFM-428, 1401 Rockville Pike, Rockville, MD 20892, USA

  • John R. Palmer,

    1. Department of Microbiology, The Ohio State University, 484 West 12th Avenue, Columbus, OH 43210, USA.
    Search for more papers by this author
  • Charles J. Daniels

    1. Department of Microbiology, The Ohio State University, 484 West 12th Avenue, Columbus, OH 43210, USA.
    Search for more papers by this author

Charles J. Daniels. E-mail; Tel. (+1) 614 292 4599; Fax (+1) 614 292 8120.


Multiple divergent genes encoding the eukaryal-like TFIIB (TFB) transcription initiation factor have been identified in the archaeon Haloferax volcanii. Expression of one of these TFB-encoding genes, referred to here as tfb2, was induced specifically in response to heat shock at the transcription level. A time course for tfb2 induction demonstrated that mRNA levels increased as much as eightfold after 15 min at 60°C. A transcription fusion of the tfb2 promoter region with a stable RNA reporter gene confirmed the heat responsiveness of the tfb2 core promoter, and immunoblot analysis using antibodies generated against a recombinant His-tagged TFB2 showed that the protein levels of one TFB increased slightly in response to elevated temperatures. An archaeal consensus TATA element (5′-TTTATA-3′) was located 110 bp upstream of the translation start site and appeared to be used for both basal and heat shock-induced expression. The long DNA leader region (79 bp) preceding the predicted AUG translation start codon for TFB2 contained a T-rich sequence element located 22 bp downstream of the transcription start site. Using an in vivo transcription termination assay, we demonstrated that this T-rich element can function as a sequence-dependent transcription terminator, which may serve to downregulate expression of the tfb2 gene under both non-heat shock and heat shock conditions.


Archaea constitute a phylogenetic domain whose members exhibit features that are both bacterial and eukaryal in nature. Among the features that are shared between archaeal and eukaryal cells, the most striking is the similarity between their transcription systems. Archaea possess a single RNA polymerase (RNAP) with 10 or more polypeptide subunits. These proteins display remarkable sequence similarity and serological cross-reactivity with the multiple subunits of eukaryal nuclear RNAPs, in particular RNAP II (Schnabel et al., 1983; Puhler et al., 1989; Klenk et al., 1992; Zillig et al., 1993; Bult et al., 1996; reviewed by Langer et al., 1995; Bell and Jackson, 1998; Soppa, 1999a). This is in contrast to the comparatively simple α2ββ′σ subunit composition of the bacterial holoenzyme. Archaea also possess a subset of the eukaryal general transcription factors (GTFs), namely the TATA-binding protein (TBP) and TFIIB (TFB) (Ouzounis and Sander, 1992; Creti et al., 1993; Marsh et al., 1994; Rowlands et al., 1994; Qureshi et al., 1995a,b; Bultet al., 1996; Klenk et al., 1997; Smith et al., 1997; Kawarabayasi et al., 1998; J. R. Palmer and C. J. Daniels, unpublished). Functional studies using in vitro transcription systems have shown that TBP, TFB and RNAP are sufficient for directing accurate initiation of promoter-specific basal transcription in Archaea (Wettach et al., 1995; Hausner et al., 1996; Qureshi et al., 1997; Hethke et al., 1996). This finding is not surprising, as archaeal homologues of other RNAP II transcription initiation factors (for example, TFIIA, TFIIF, TFIIE and TFIIH) have not been found in archaeal genome databases (Bult et al., 1996; Klenk et al., 1997; Smith et al., 1997; Kawarabayasi et al., 1998). In addition, the large subunit of archaeal RNAP lacks the carboxy-terminal repeat domain (CTD) that undergoes a phosphorylation–dephosphorylation cycle directed by TFIIH and TFIIF, respectively, during the transition from initiation to elongation. The initiation of transcription by the archaeal RNAP does not require hydrolysis of the beta–gamma phosphoanhydride bond of ATP, another indication that a TFIIH-like homologue is not required (Bell et al., 1998).

Based on comparative sequence studies, the archaeal TBP and TFB proteins possess the highly conserved features that define the eukaryal factors (reviewed by Thomm, 1996; Reeve et al., 1997; Bell and Jackson, 1998; Soppa, 1999a). In general, archaeal TBP contains only those sequences that correspond to the two imperfect direct repeats that compose the carboxy-terminal domain of eukaryal TBPs. The archaeal TBP proteins are distinguished from their eukaryal counterparts by having a higher degree of sequence similarity between their direct repeat elements (≈40% versus ≈30% respectively). The archaeal TFB proteins possess the conserved sequence domains of eukaryal TFIIB proteins. These include the amino-terminal zinc binding (or zinc finger domain) and the carboxy-terminal imperfect repeated motif. They also have the highly conserved carboxy-terminal, helix–turn–helix region (H4′ and H5′) that has been proposed to interact with DNA sequences upstream of the TBP-binding region, the TFIIB recognition element, BRE (Lagrange et al., 1998). Functional and structural studies also support this close relationship. Archaeal TBPs have been shown to recognize and bind specifically to TATA-like sequence elements upstream of archaeal genes (Gohl et al., 1995; Qureshi et al., 1995a; Kosa et al., 1997). Both in vitro and in vivo studies have demonstrated that the central TATA-like motif in archaeal promoters plays a crucial role in accurate and efficient transcription (Reiter et al., 1988; Reiter et al., 1990; Hausner et al., 1991; Hain et al., 1992; Palmer and Daniels, 1995; Danner and Soppa, 1996). In addition, in vitro studies with the Sulfolobus shibatae transcription system (Qureshi and Jackson, 1998) have shown that mutations at the positions 3–6 nucleotides upstream of the TATA element affect promoter strength and that these sequences are protected by TFB; this is similar to the TFIIB–BRE interaction. Recently, Kosa et al. (1997) determined the X-ray crystal structure of the Pyrococcus woesei TBP and TFB proteins in complex with TATA-DNA. These data verified that the archaeal proteins possessed the characteristic structural features of the eukaryal proteins and that the ternary complex retained the TBP–DNA and TBP–TFB interactions. One unexpected difference in this complex was the observation that the archaeal TBP protein bound to the TATA-DNA in the reverse orientation, directing the amino terminus of TFB towards the upstream region and away from the transcription start site. Based on the many similarities between the archaeal and eukaryal TBP and TFB proteins, it is likely that the orientation of TBP on TATA-DNA is the same in the archaeal and eukaryal RNAP II preinitiation complexes and that the observed difference in orientation reflects a specific binding preference to the TATA-DNA used in their study. These data, together with the observation that archaeal genomes do not appear to encode eukaryal-like TBP-associated factors (TAFs), suggest that the Archaea possess a simplified version of the eukaryal RNAP II transcription system.

In the process of characterizing the basal transcription machinery of the halophilic Archaea, we have made the unexpected discovery that Haloferax volcanii possesses multiple genes encoding TBP and TFB proteins (J. R. Palmer and C. J. Daniels, unpublished). This is in sharp contrast to other Archaea and eukaryal cells in general, in which genome sequences indicate that the occurrence of multiple TBP and TFB genes is rare (Bult et al., 1996; Klenk et al., 1997; Smith et al., 1997; Kawarabayasi et al., 1998). The presence of multiple TBP and TFB genes raises the question of whether these genes are differentially expressed. Here, we report on the characterization of the tfb2 gene and show that transcription of this gene is induced during heat shock at 60°C. The expression of this gene is accompanied by an increase in the level of a protein that co-migrates in SDS–PAGE analysis with full-length recombinant TFB2 expressed in Escherichia coli. We also show that transcription of the tfb2 gene is downregulated and that this regulation may be caused by the activity of a transcription termination element in the leader region of the gene.


Cloning of the H. volcanii tfb2 gene

Using two highly conserved sequence elements present in both archaeal TFB and eukaryal TFIIB proteins, we designed two degenerate polymerase chain reaction (PCR) primers (TFBF1 and TFBR1) for the amplification of potential TFB-encoding genes in H. volcanii. A PCR product was found to encode a TFB-like sequence, and this DNA was subsequently used to screen the H. volcanii cosmid bank (Charlebois et al., 1991) for potential TFB-encoding genes. Six separate TFB-related DNA fragments were identified by this approach. One of these fragments, localized to cosmid A5, was subcloned, sequenced and found to contain a complete TFB-encoding gene, designated as tfb2. This gene appeared to be monocistronic and contained sequences related to the archaeal TATA-like element and a T-tract transcription termination signal. This gene also had a T-tract element and a short direct repeat in the leader region, features not observed in the other tfb genes. The tfb2 gene was predicted to encode a 332-amino-acid protein with an apparent molecular mass of 37284 Da. The TFB2 protein possessed the conserved sequence elements present in archaeal and eukaryal TFB proteins. These included the sequences associated with TBP interaction and DNA binding, the CXXC-N15–17-CXXC zinc finger element in the amino-terminal region of the protein and the carboxy-terminal helix–turn–helix region proposed to bind sequences upstream of the TATA element (Fig. 1). In addition, structure predictions indicated that this protein could assume a backbone configuration similar to other reported TFB proteins. Conservation of these features, coupled with the evidence that this gene is transcribed (see below), suggested strongly that this protein is a homologue of the eukaryal TFIIB.

Figure 1.

. Nucleotide sequence of the H. volcanii tfb2 gene. Sequence of the gene was derived from cosmid A5 DNA and is available as GenBank entry AF143963. The archaeal consensus TATA-like promoter element (single underline), the mapped transcription start site (asterisk), putative transcription termination sites in the leader (double underline) and downstream regions (dashed underline) and a short repeat in the leader region (arrows) are indicated. The predicted protein sequence is indicated below the DNA sequence. The highly conserved amino-terminal zinc finger and carboxy-terminal helix–turn–helix motifs are indicated in the shaded boxes. The DNA sequence presented in bold corresponds to the sequence region complementary to the probe TFB2, which was used in Northern analysis with total cellular RNA (see Fig. 2).

Transcriptional induction of the H. volcanii tfb2 gene in response to heat shock

The occurrence of multiple TFB genes in H. volcanii suggested that their expression might be differentially regulated in response to various environmental stimuli. As we had shown previously that H. volcanii exhibits a typical heat shock response (Daniels et al., 1984; Kuo et al., 1997), we were particularly interested in determining whether any of the transcription factor genes responded to this stress. To investigate this possibility, antisense synthetic oligonucleotides were designed for each tfb gene and used to probe Northern blots generated from non-heat-shocked and heat-shocked H. volcanii cellular RNAs. As shown in Fig. 2A, tfb2 mRNA levels were induced preferentially as early as 10 min after heat shock at 60°C and retained a high level of expression (five- to eightfold induction) during 45 min of heat shock treatment at 60°C. An RNase P-specific probe was used as an internal control to evaluate RNA recoveries (Fig. 2A). The size of the tfb2 transcript was estimated to be 1.1 kb, similar to the predicted size of the open reading frame (ORF), and supported the proposal that TFB2 is encoded by a monocistronic mRNA (Fig. 1). A similar, but weaker, response was observed for tfb1 mRNA, whereas the transcript levels for the other transcription factor genes decreased or showed no change (data not shown). The response of the tfb2 gene was chosen for further analysis. To determine whether expression of the TFB2 protein also changed in response to increased temperature, polyclonal antibodies were raised against recombinant H. volcanii TFB2 carrying a carboxy-terminal polyhistidine [(His)6] tag. Western analysis was performed on crude extracts prepared from H. volcanii cells subjected to no shock (37°C) and heat shock (60°C) conditions. Immunoblot analysis revealed three strong signals ranging in size from 34 to 40 kDa (Fig. 2B). As the H. volcanii TFBs show high identity (51–78%) to each other at the amino acid level, it was expected that the anti-TFB2 antibodies would cross-react with some or all of the other TFB proteins. The sizes of the three protein bands in 2Fig. 2B are consistent with the predicted sizes (35–38 kDa) of the halobacterial TFBs based on their deduced amino acid sequences. In addition, the middle protein (Fig. 2B) exhibited an ≈twofold increase in synthesis under heat shock conditions. This protein was found to co-migrate with full-length recombinant TFB2 (37 kDa) expressed in E. coli without any terminal peptide fusions (Fig. 2B), suggesting that the heat-responsive protein in H. volcanii cell-free extracts is TFB2. We were not able to determine if there was a similar increase in the level of the TFB1 protein during heat shock, as this protein co-migrated with other TFB protein present in the 35 kDa size range.

Figure 2.

. Induction of H. volcanii tfb2 RNA and protein in response to heat shock. A. Northern analysis of cellular RNAs isolated from H. volcanii cells under preshock and heat shock conditions. H. volcanii cells were grown to mid-log phase at 37°C (0 min at 60°C) and then incubated for the indicated times at 60°C. tfb2 mRNA levels were detected using the end-labelled TFB2 oligonucleotide. The oligonucleotide ArchaeaP was used to detect chromosomally encoded RNase P RNA, used here as an internal control. The relative level of induction, compared with 0 min at 60°C, is presented below each lane. B. Detection of TFB proteins by Western analysis. H. volcanii cells were grown at 37°C (0 min at 60°C) and then incubated at 60°C for the indicated times. Crude cell extracts were prepared for each time point and the proteins separated by SDS–PAGE. Rabbit anti-TFB2 IgG was used as the primary antibody. Lane M is a crude cell extract prepared from E. coli cells containing the pET3aTFB2 construct, which encodes the recombinant H. volcanii TFB2 protein without terminal extensions.

Analysis of tfb2 promoter activity using an in vivo reporter assay

A survey of sequences upstream of the tfb2 gene indicated the occurrence of an archaeal consensus TATA sequence, 5′-TTTATA-3′, ≈100 bp upstream of the presumed AUG translation start codon (Fig. 1). As previous experiments indicated that heat shock induction of the H. volcanii cct1 gene required only the core promoter region (Kuo et al., 1997), we chose to examine the transcription properties of a DNA fragment that included the TFB2 start codon and 140 bp of upstream sequence. This region carried the TATA element, a short pyrimidine-rich sequence (5′-TTATTCTTT-3′; referred to as the T-tract) and a short direct repeat (5′-GTCTTC-3′) (Fig. 1). The H. volcanii, plasmid-based, promoter yeast tRNAProM reporter expression system (Palmer and Daniels, 1994) was used to monitor the in vivo promoter activity of this DNA fragment. We have shown previously that this system can accurately monitor in vivo transcription and that the initiation site of inserted promoters is not altered by coupling of the reporter gene (Kuo et al., 1997; Thompson and Daniels, 1998). Two promoter–reporter fusions were constructed to confirm the temperature-dependent regulation of tfb2 gene expression. Both contained a 5′ terminus corresponding to the position 140 bp upstream of the start codon (−60 relative to the mapped transcription start site; see below). However, the 3′ termini of the promoter regions varied. The truncated promoter–reporter construct (Fig. 3A, insert II) contained sequences to position +12 relative to the mapped transcription start site. This construct carried the TATA element but lacked the downstream T-tract and direct repeat sequences. The full-length construct (Fig. 3A, insert I) contained the entire 140 bp tfb2 DNA leader region, including the predicted ATG start codon for the gene. Primer extension analysis of RNAs isolated from cells carrying the full-length construct demonstrated that transcription of the yeast tRNAProM reporter gene initiated from the same tandem adenine residues under both non-heat shock and heat shock conditions (Fig. 3B). The minor extension products present in 2Fig. 2C correspond to stops at individual A residues in the leader region and are not likely to represent alternative transcription start sites or RNA processing sites. In addition, the same AA dinucleotide start site was also used by the native gene (data not shown). The mapped initiation site was located 26 bp downstream from the centre of the archaeal consensus TATA motif (5′-TTTATA-3′, positions −29 to −24), suggesting that the single TATA-like sequence is used for both basal and heat shock-induced transcription.

Figure 3.

. Primer extension and Northern blot analysis of tfb2 promoter–reporter fusions. A. Plasmid-based promoter–reporter fusion constructions used in primer extension and Northern analysis. Two expression plasmids were used, pWLTFBI-ProM (insert I) and pWLTFBII-ProM (insert II). Both plasmids had inserts that began at position −60 of the tfb2 promoter region. Inserts I and II had 3′ termini corresponding to positions + 82 and +12 respectively. B. Primer extension. Lane 1, 15 μg of total RNA isolated from pWLTFBI-ProM-bearing H. volcanii cells grown at 37°C; lane 2, 10 μg of total RNA isolated from pWLTFBII-ProM-bearing H. volcanii cells heat shocked at 60°C for 45 min. The PROEXI oligonucleotide was used for transcript analysis and DNA sequencing. Primer extension products (lanes 1 and 2) were electrophoresed alongside dideoxy-terminated sequencing products (lanes T, C, G and A) on a 6% denaturing polyacrylamide gel. The mapped initiation site is marked (arrow). C. Northern analysis of the promoter fusions. Transcription activity of the promoters was measured under non-heat shock (0 min at 60°C) and heat shock (45 min at 60°C) conditions. Hybridization signals corresponding to the reporter tRNAProM transcript and an internal control, H. volcanii chromosome-encoded tRNALeu transcript, are indicated. The relative level of induction is expressed as the fold increase in RNA when compared with the level of tRNAProM present in cells carrying the plasmid pWLTFBI-ProM (insert I) under non-heat shock conditions.

The in vivo transcription activity of the two promoter–reporter constructs was measured by Northern blot hybridization using the tRNAProM-specific oligonucleotide PROEXI to detect transcript levels of the reporter gene. 3Figure 3C shows that the full-length version (insert I) exhibited a sixfold induction in transcription after incubation at 60°C for 45 min. Interestingly, the basal transcription activity of the shortened promoter–reporter fusion (insert II) was 20-fold higher than that observed for the full-length promoter fragment (Fig. 3C). The truncated promoter (insert II) was also induced by heat shock. However, the level of induction was only twofold compared with sixfold for the full-length promoter–reporter fusion. These data verified that the tfb2 promoter was induced by heat shock, and they also revealed that the promoter is downregulated by an element located between positions +13 and + 82 in the DNA leader region.

Evidence for a transcription termination element in the tfb2 leader region

Transcription fusions of the tfb2 promoter region with the yeast tRNAProM reporter gene suggested the occurrence of a negative regulatory element downstream of the initiation site. The presence of a T-tract sequence element, which resembled an archaeal termination element (Thomm et al., 1994; Kuo, 1997), suggested that a premature termination event could be responsible for the downregulation observed with the full-length promoter construct (see Fig. 3C). To evaluate this possibility, a modified pWL plasmid-based expression module, referred to as sptProM, was used to measure the in vivo termination activity of the putative transcription termination element in the tfb2 leader region. This in vivo termination assay has been described previously (Kuo, 1997). The sptProM expression cassette contains the constitutive H. volcanii tRNALys promoter, the terminatorless yeast tRNAProM reporter gene, a short sequence region containing restriction enzyme sites and the native RNAP III transcription terminator from the yeast tRNAPro gene (Fig. 4A). The termination efficiency of a sequence element inserted into the BamHI site downstream of the yeast tRNAProM gene can be evaluated by comparing the level of termination occurring at the inserted sequence with termination at the yeast tRNAPro RNAP III terminator (Fig. 4A).

Figure 4.

. Functional analysis of the putative T-track termination element in the tfb2 5′ DNA leader region. A. Schematic representation of the sptProM termination module (Kuo, 1997) carried on the H. volcaniiE. coli pWL201 shuttle plasmid. In this assay, active terminators produce short transcripts, whereas inactive terminators produce readthrough transcripts terminating at the tRNAProM terminator. Restriction sites: H, HindIII; B, BamHI; and E, EcoRI. B. Terminator sequence elements. The TATA-less TFB2 leader region (positions −6 to + 82) was cloned in both forward (T1) and reverse (T2) orientations into the BamHI site of the sptProM expression module. Two mutants of the forward element were also examined, a TT to GG double mutant at positions 25 and 26 (T3) and a deletion mutant, ΔTCTT (T4). Asterisks indicate the potential T-track terminator element, and the dashed arrows designate the repeat sequence. C. Northern blot analysis of the wild-type and mutant termination constructs. Total RNA was isolated from construct-bearing H. volcanii cells grown under normal or heat shock (45 min at 60°C) conditions, and transcripts were detected using the tRNAProM-specific probe, PROEXI. Insert sequences are indicated above the lanes. Hybridization signals corresponding to transcripts terminated at the tRNAProM terminator (readthrough transcripts) and at the putative tfb2 terminator are shown. The relative termination efficiency of the putative tfb2 terminator (percentage of termination directed by the element versus readthrough) is given below the corresponding lane.

The TATA-less tfb2 gene leader region (positions −6 to + 82, with respect to the transcription start site) was ligated into the BamHI site of the sptProM expression module (Fig. 4A). The DNA fragment was cloned in both the forward and the reverse orientations (Fig. 4B; inserts T1 and T2 respectively) to determine whether the inhibitory activity was dependent on its orientation. Termination efficiency was also monitored under non-heat shock (0 min at 60°C) and heat shock (45 min at 60°C) conditions to determine whether premature termination was temperature dependent. As shown in 4Fig. 4C, ≈80% of the yeast tRNAProM transcripts terminated within the forward-oriented (T1) tfb2 leader region, while 20% constituted readthrough transcripts that terminated within the T-tract of the yeast tRNAPro RNAP III terminator. Based on the size of the transcript, the major termination event was estimated to be within the 5′-TTATTCTTT-3′ sequence of the inserted tfb2 DNA element (see Fig. 4A). In contrast, when the same tfb2 leader region was placed in the reverse orientation so that the putative terminator was now on the template strand, 100% of the yeast tRNAProM transcripts terminated at the tRNAPro RNAP III terminator, indicating that the termination event occurred only when the T-tract was present on the coding strand. Heat shock had no significant effect on termination efficiency of the tfb2 element. To examine further the sequence requirements for the termination event directed by the tfb2 leader, we prepared two mutant forms of this DNA. Previously, Kuo (1997) had demonstrated that T → G mutations within the T-tracts of H. volcanii terminators severely reduced termination efficiency. PCR-based mutagenesis was therefore used to create a dinucleotide change in the putative termination element, altering the sequence from 5′-TTATTCTTT-3′ to 5′-TTATTCGGT-3′ (Fig. 4B; insert T3). Additionally, a portion (5′-TCTT-3′) of the first repeat of the direct repeat motif immediately downstream of the putative terminator was deleted using PCR-directed site mutagenesis (Fig. 4B; insert T4). As expected, the TT → GG mutation nearly abolished the termination activity of the 5′-TTATTCTTT-3′ sequence element in the tfb2 leader region (Fig. 4C). However, deletion of the direct repeat had no detectable effect on the general termination activity (Fig. 4C).


In this study, we report on the sequence and expression properties of one of the six tfb genes detected in H. volcanii. By all comparative sequence criteria, the H. volcanii TFB protein appears to be a functional homologue of the eukaryal TFIIB protein. This protein is 40–50% identical to other characterized archaeal TFB proteins and possesses the zinc finger and helix–turn–helix motifs, which are hallmarks of this protein. The gene possess an archaeal TATA-like promoter element centred 110 bp upstream of the first codon and appears to encode a monocistronic mRNA (Fig. 1). Prompted by the occurrence of multiple tfb genes in H. volcanii and previous results indicating that heat shock-induced transcription occurs in this organism, we examined the heat shock responsiveness of the tfb2 gene. Using a gene-specific oligonucleotide probe, we observed that tfb2 mRNA levels increased substantially as early as 10 min after heat shock at 60°C and remained high (five- to eightfold induction) during 45 min of heat shock treatment (Fig. 2A). Western analysis indicated that there was a concomitant increase (twofold) in the amount of an anti-TFB2 cross-reactive protein, which was similar in size to the predicted TFB2 protein (Fig. 2B). These data established that the expression of the tfb2 gene and its protein product are subject to regulation by heat stress.

To define the mechanism of this regulatory response better, we examined the transcription activity of the isolated tfb2 promoter element using an in vivo transcription, plasmid-based, reporter assay system (Palmer and Daniels, 1994). Primer extension analysis of the reporter fusion indicated that tfb2 transcripts initiated 82 bp upstream of the predicted TFB2 start codon and that the same start site was used under non-heat shock and heat shock conditions (Fig. 3B). This suggested that a single promoter was likely to be used for expression under both conditions. Analysis of the sequence upstream of the transcription start site indicated the presence of an archaeal consensus promoter (−35 AAAAACTTTATA −24); this included a TATA element and a short purine tract centred at positions −31 and −35 from the transcription start site respectively (Fig. 1). Inclusion of the TATA-proximal purine sequence in the consensus promoter is based on two observations: these nucleotides are needed for in vivo expression of some halophile genes (Palmer and Daniels, 1995; Thompson and Daniels, 1998); and recent in vitro data have shown that both archaeal and eukaryal TFB proteins interact with this region (Lagrange et al., 1998; Qureshi and Jackson, 1998). Use of a consensus promoter by a heat shock-regulated gene in the Archaea is not unexpected. By analogy with the eukaryal RNAP II system, the TATA sequence and upstream purine element would direct the binding of TBP and TFB to the promoter; binding of transcription factors at other DNA sites would bring about induction by increasing preinitiation complex formation or increasing the rate of transcription initiation. Both in vivo and in vitro data support the notion that archaeal and eukaryal TBPs and TFBs have similar functions; however, there is no evidence that Archaea use eukaryal-like heat shock transcription factors to regulate gene expression under heat shock conditions. Genes encoding the highly conserved eukaryal heat shock factors (for a review, see Morimoto, 1998) are absent in the sequenced archaeal genomes, and their corresponding DNA binding sequence, inverted repeats of the sequence 5′-nGAAn-3′, are not present in archaeal heat shock genes. The H. volcanii cct genes, which are also induced by heat shock, require only sequences within the core promoter region for heat shock induction (Thompson and Daniels, 1998). This does not preclude the participation of heat shock transcription factors; however, their binding sites, whether DNA or protein, must lie within this core promoter region. It is noteworthy that the H. volcanii cct (Kuo et al., 1997; Thompson and Daniels, 1998) and tfb2 genes have a consensus archaeal TATA element, 5′-TTTATA-3′. The presence of this sequence appears to be rare in halophile promoters (Palmer and Daniels, 1995; Soppa, 1999b) and may indicate that the cct genes and tfb2 genes have a different, perhaps more active, promoter than the average halophile gene. The precise mechanism of heat-induced transcription in the Archaea is not yet understood.

In addition to the functions of the core promoter sequences, promoter–reporter fusion studies have shown that sequences within the tfb2 leader region could also play a role in regulating the expression of this gene. The 82 bp leader region contained a short T-tract sequence that resembled an archaeal transcription termination signal (Thomm et al., 1994; Kuo, 1997) and a short direct repeat (Fig. 1). A comparison of the relative transcription activities of promoter–reporter constructs with and without these elements indicated that a promoter lacking the T-tract and repeat elements had an unexpectedly high level of basal transcription and a slightly lower level of induction by heat shock compared with a construct carrying all sequences (Fig. 3C). This suggested that there was a downregulation in the basal level of transcription and a higher relative level of heat shock induction when the T-tract and direct repeat were present. Evaluation of the T-tract as a potential transcription terminator indicated that this sequence was capable of directing termination in vivo ; however, the termination event was not affected by heat shock (Fig. 4C). Although the T-tract may function as a terminator, heat shock did not cause the termination signal to be ignored. Thus, it appears from these data that the increase in transcription observed with the tfb2 promoter after heat shock is not solely the result of antitermination. The independence of the potential termination event and heat shock responsiveness of the promoter is also supported by the observation that a promoter construct lacking the T-tract still retains heat shock inducibility (Fig. 3C; insert II). The short repeat element located downstream of the T-tract does not appear to play a direct role in either process, as its loss did not decrease termination by the T-tract in the termination assay (Fig. 4C; insert T4), and its absence did not abolish heat shock inducibility of the tfb2 promoter (Fig. 3C; insert II).

Based on these results, it appears that TFB2 protein production is controlled by at least two independent events: a positive regulatory event involving induction of the tfb2 promoter; and a negative event in which the level of transcription into the coding region is downregulated by the presence of a T-tract element. The precise level of induction that results from an increase in transcription initiation is not yet fully defined and may vary from two- to eightfold. We cannot distinguish whether the twofold induction observed with the tfb2 core promoter–reporter fusion construct (Fig. 3C; insert II) is a true maximum, as we cannot exclude the possibility that a positive factor may be limiting during expression from the multicopy (12 copies/cell) pWL-based plasmid used in the promoter–reporter fusion studies. The observed twofold increase in the level of TFB2 protein during a 45 min heat shock period is consistent with the observed level of tfb2 mRNA induction. As H. volcanii has a doubling time of approximately 5 h in complex medium, a sixfold increase in tfb2 mRNA would be expected to lead to a twofold increase in the protein over a 45 min time period. The role of the T-tract element is unresolved. Although the T-tract resembles a typical archaeal transcription termination signal and acts as an efficient terminator in an in vivo termination assay, it is possible that the T-tract element does not act as a termination site. This sequence could function as a pause site stalling RNAP and poising the tfb2 gene for a rapid transcription response to stimuli. A similar pausing event in the leader region has been observed for the Drosophila hsp70 gene (Rasmussen and Lis, 1995) and appears to be a common phenomenon in these cells. In the case of the hsp70 gene, the downregulation event occurs under both non-heat shock and heat shock conditions, similar to that observed for the H. volcanii tfb2 gene

The presence of multiple TBP and TFB proteins in H. volcanii and evidence for differential expression of at least one of these proteins indicates that alternative transcription factor pairing may play a role in altering global gene expression in these organisms. Although the occurrence of multiple sigma factor proteins in bacterial cells and their influence on gene expression is well known, the occurrence of multiple RNAP II-associated transcription factor genes in eukaryal cells is rare. Two separate, but nearly identical, TBP-encoding genes have been identified in Arabidopsis thaliana (Gash et al., 1990) and maize (Haass and Feix, 1992). In addition to these examples, there have been recent reports that TBP-related factor proteins (TRFs) are present in eukaryal cells (Crowley et al., 1993; Ohbayashi et al., 1999; Rabenstein et al., 1999). In the case of the Drosophila TRF1 protein, there is evidence that this protein is involved in neural-specific gene expression (Hansen et al., 1997). Eukaryal cells also use TBP in the three related RNAP complexes (RNAP I, II and II), and there is a TFIIB-related factor protein (BRF) involved in RNAP III transcription (Colbert and Hahn, 1992). This indicates that eukaryal cells have used gene duplication of core RNAP subunits and, to a limited extent, duplication of the GTFs to direct different patterns of gene expression. In the case of Archaea, the available genome sequences indicate that there are generally single TBP- and TFB-encoding genes. The exceptions are Pyrococcus horikoshii and the halophiles. Analysis of the P. horikoshii genome sequence indicates that this organism has two tfb-related genes, one encoding a complete TFB protein and a second encoding a partial TFB protein that lacks the amino-terminal zinc finger region (Kawarabayasi et al., 1998). There is no information available on the expression of the P. horikoshii genes. Similarly, sequence analysis of the Halobacteria sp. NRC1 plasmid indicates that this DNA element has three TBP-encoding genes (Ng et al., 1998), and H. volcanii has multiple TBP- and TFB-encoding genes (J. R. Palmer and C. J. Daniels, unpublished). It is interesting to consider the possibility that the Archaea have used a strategy similar to the use of multiple sigma factors by bacterial cells to expand the regulatory capabilities of their eukaryal-like RNAP II machinery.

Experimental procedures

Bacterial strains, culture conditions and plasmids

E. coli ED8767 harbouring cosmid clones of an H. volcanii DS2 genomic library (Charlebois et al., 1991) were provided by R. L. Charlebois and grown in Luria–Bertani (LB) medium (Sambrook et al., 1989) supplemented with kanamycin (50 μg ml−1). E. coli strains DH5α and JM110 were cultured routinely in LB or circlegrow medium (Bio101) with ampicillin (100 μg ml−1) to select for pUC19- or pWL-based plasmids (Lam and Doolittle, 1989). E. coli BL21 (DE3) was obtained from K. Sandman and cultivated in LB medium containing ampicillin (100 μg ml−1) to select for the pET3 expression vector (Novagen; a gift from K. Sandman). H. volcanii strain WFD11 (Charlebois et al., 1987) was grown aerobically in complex low-salt medium (Nieuwlandt and Daniels, 1990) either at 42°C (for solid medium) or at 37°C (for liquid medium) and supplemented with 20 μM mevinolin (a gift from Merck) when cells carried pWL-based expression plasmids. H. volcanii transformations were performed essentially as described previously (Palmer and Daniels, 1995).

Cosmid clone A5 from the H. volcanii genomic library contains the complete tfb2 gene and its adjacent 5′ flanking regulatory region. Construction of the yeast tRNAProM reporter module in the H. volcaniiE. coli shuttle vector pWL201 has been reported previously (Palmer and Daniels, 1994). To examine the potential termination activity of a sequence element in the tfb2 DNA leader region, the pWL-based sptProM expression module, provided by Y.-P. Kuo, was used as the cloning vector. Construction of this expression module for in vivo termination analysis has been described previously (Kuo, 1997). The E. coli expression plasmids pTrcHis2A (Invitrogen) and pET3a (Novagen) were used to overexpress the TFB2 protein.

Cloning the H. volcanii tfb2 gene

A segment of tfb encoding DNA was PCR amplified from H. volcanii genomic DNA using a pair of degenerate oligonucleotides complementary to highly conserved regions of TFB-encoding genes: TFBF1, 5′-GGCCCSGARTGGCGSGCSTTC; TFBR1, 5′-GCGATGTASAGSGCSGCSGC (S = G + C; R = A + G). Amplification reactions contained 5 μg of H. volcanii genomic DNA, dNTPs at 500 μM, primer at 2 μM and 5 units of Taq DNA polymerase (Boehringer Mannheim). The reaction was carried out for 30 cycles; one cycle was composed of the steps: denaturation at 94°C for 1 min, template/primer annealing at 50°C for 1 min and primer extension at 72°C for 2 min. The products were separated by agarose electrophoresis, and the expected 0.7 kb PCR product was ‘TA’ cloned into the pCR2.1 vector (Invitrogen) to give the plasmid pCR0.7/TFB. The insert was sequenced using Sequenase 2.0 (US Biochemical) to confirm amplification of a DNA region possessing significant homology to TFB-encoding genes. The construct pCR0.7/TFB was then random primer labelled with [α-32P]-dCTP using the High Prime DNA labelling system (Boehringer Mannheim) and used to probe a set of nylon membrane (Nytran Plus; Schleicher & Schuell) dot blots containing the set of overlapping cosmid clones representing 96% of the H. volcanii genome (Charlebois et al., 1991). Hybridizations were carried out in a solution containing 0.25 M Na2HPO4 plus 7% SDS at 60°C for 24 h; membranes were then washed three times in 2 × SSC (0.3 M NaCl, 0.03 M trisodium citrate) plus 0.5% SDS at room temperature for 30 min each wash. The dot blots were analysed by electronic autoradiography using the Packard InstantImager 2024. Six potential TFB-encoding loci, distributed across the H. volcanii genome, were identified. Southern analysis was used to identify restriction fragments carrying each of the tfb genes (termed tfb1 to tfb6 ), which were subsequently cloned into the multiple cloning region of pUC19.1 (pUC19 possessing an MluI restriction site engineered between the BamHI and XbaI sites) and sequenced. The gene located on cosmid A5 was designated tfb2 and selected for further study.

DNA manipulation and sequencing

DNA fragments cloned into the in vivo promoter and termination assay modules, the pTrcHis2 vector and the pET3 vector were generated by PCR amplification using the H. volcanii cosmid clone A5 as the DNA template. Amplification reactions were carried out using 2 U of Vent DNA polymerase (New England Biolabs), 100–200 ng of template DNA, 1 μM each primer, 200 μM dNTP and 10 μl of 10 × Vent buffer (New England Biolabs) in a total volume of 100 μl. PCR reactions were programmed for 30 sequential cycles of denaturation (94°C, 1 min), primer annealing (50°C, 1 min) and extension (72°C, 2 min) followed by a final extension reaction (72°C, 5 min).

For the in vivo analysis of tfb2 promoter activity, the forward PCR primer pTFB2.F (5′-TGCGATTCACCCACACAG-3′) was used to generate a 5′ terminus corresponding to position −60 relative to the mapped transcription start site. The following two reverse PCR primers were used to create promoter fragments with a 3′ terminus corresponding to positions +12 (truncated version, construct II) and + 82 (full-length version, construct I) respectively: pTFB2.R1 (5′-TGGACTGGCCTTACCAAC-3′) and pTFB2.R2 (5′-CATGGTGGGGTGGAAGGC-3′). For cloning purposes, PCR oligonucleotide primers were designed to append non-complementary HindIII and XbaI endonuclease restriction sites to the 5′ and 3′ termini, respectively, of the amplified DNAs. Promoter regions were cloned into the HindIII and XbaI sites of the pWL201-based expression module according to standard ligation protocols (Sambrook et al., 1989) to give the plasmids pWLTFBI-ProM and pWLTFBII-ProM.

To investigate the potential termination activity of a sequence element located within the tfb2 leader DNA, forward primers Bam.F (5′-GTTGGTAAGGCCAGTCCA-3′) and TFB2-term.F (5′-AAGCCGGATTATTCTTTC-3′) and the reverse primer Bam.R (5′-CATGGTGGGGTGGAAGGC-3′) were used in PCR reactions to amplify the tfb2 leader region containing the putative termination element. Synthetic BamHI sites were introduced to facilitate cloning of the DNA fragments into the equivalent site in the sptProM termination assay module (Kuo, 1997). The mutagenic primers TFB2.mut1 (5′-AAGCCGGATTATTCGGTCGTCTTCTGAG-3′) and TFB2.mut2 (5′-AAGCCGGATTATTCTTTCGCTGAGGTCTTCC-3′) were used in PCR amplification, respectively, to create a TT → GG mutation in the tfb2 T-tract element and a deletion mutation in the direct repeat motif.

For constructing the His-tagged TFB2, the forward primer TF2His(M)-F (5′-GATGAGCGACACGATAACCACC-3′) with a synthetic BamHI restriction site and the reverse primer TF2His-R (5′-CGCGAGCAGGGTGCCGTCTTG-3′) with a synthetic HindIII restriction site were used in a PCR reaction to amplify the complete tfb2 gene. A polyhistidine tag [(His)6]was fused to the carboxy-terminus of TFB2 by ligating the PCR-amplified tfb2 gene into the BamHI and HindIII sites of the pTrcHis2A vector (Invitrogen). Overexpression of TFB2 without any peptide fusions was achieved by amplifying the tfb2 gene using the PCR forward primer pET-F (5′-ATGAGCGACACGATAACCACC-3′) and the reverse primer pET-R (5′-TTACGCGAGCAGGGTGCCGTC-3′) and then cloning the PCR-generated tfb2 gene into the NdeI and BamHI sites of the pET3a vector (Novagen).

Expression constructs were identified by colony blot hybridizations and then sequenced by the dideoxy chain termination method (Sanger et al., 1977) using either Sequenase 2.0 (US Biochemicals) with the 7-deaza-dGTP sequencing kit (US Biochemicals) or the Bst DNA polymerase kit (Bio-Rad).

Induction of heat shock response and RNA isolation

Both non-transformed H. volcanii cells and cells containing pWL-based expression plasmids were harvested for heat shock when the cell density reached mid-log phase (typically, OD550 = 0.5–0.7). Cultures were transferred from 37°C to a 60°C water bath to induce heat shock, and aliquots of 1.5 ml were removed before stress challenge (0 min; control) and at various times after incubation at 60°C. Total RNA was isolated as described previously (Thompson and Daniels, 1998).

Northern blot analysis and transcript quantification

The in vivo activity of each promoter–reporter fusion and termination construct was measured by Northern blot hybridization using the yeast tRNAProM exon 1-specific oligonucleotide PROEXI (5′-CCCAAAGCGAGAATCATACCAC-3′) as described previously (Thompson and Daniels, 1998). In these Northern analyses, a leucine tRNA (UUA)-specific probe (5′-GGGGACGAGATTCGAACTCGCGAACCCCTACG-3′) was used as an internal control. To demonstrate the transcriptional induction of the native tfb2 gene, total RNA isolated from wild-type H. volcanii WFD11 cells was pelleted by centrifugation (7500 × g at room temperature for 5 min) and dried briefly (2 min) under vacuum. RNA pellets were resuspended directly in 20 μl of urea loading buffer [7 M urea, 10% (w/v) glycerol, 0.05% xylene cyanol, 0.05% bromophenol blue]. Cellular RNAs were separated by electrophoresis through a 1.5% formaldehyde agarose gel, as described previously (Selden, 1987), and then transferred to Zeta-Probe-GT nylon membrane (Bio-Rad) using the Turboblotter (Schleicher & Schuell). The downward transfer of RNA was carried out in RNase-free 10 × SSC for 15–20 h. The gene-specific oligonucleotide TFB2 (5′-CACTCGGGGCAGACCTGTGT-3′) was used to detect TFB2 transcript levels. The oligonucleotide ArchaeaP [5′-GG(A/T)GG(C/G)(T/G)GGACTTTCCTC-3′], which is specific for H. volcanii chromosomally encoded RNase P RNA, was used as an internal control. Each oligonucleotide probe (5–10 pmol) was end-labelled at its 5′ terminus using [γ-32P]-ATP (160 μCi μl−1, 7000 Ci mmol−1) and 5 U of T4 polynucleotide kinase (Gibco BRL). Hybridization and washing procedures were performed essentially as reported previously (Thompson and Daniels, 1998).

Hybridization signals were quantified using the Packard InstantImager 2024. Values for the tRNAProM signals were normalized against the internal control as described previously (Thompson and Daniels, 1998). Values for hybridization signals corresponding to native tfb2 mRNA (Fig. 2) were normalized by calculating the ratio of total counts for TFB2 transcripts over the total counts for the internal control, RNase P RNA. For promoter fusion studies (Fig. 3), induction by tfb2 promoters is expressed as fold increases above the level of reporter produced with the native promoter under non-heat shock conditions. The tRNALeu RNA was used to monitor RNA recovery.

Primer extension analysis

Total RNA (10–15 μg) isolated from pWLTFBII-ProM-carrying H. volcanii WFD11 cells under non-heat shock (37°C) and heat shock (60°C for 45 min) conditions served as the template for cDNA synthesis. The yeast tRNAProM expression construct used in 5′ transcript analysis contained 139 bp of the upstream tfb2 regulatory region and the predicted translation start codon (AUG) for the gene. The tRNAProM-specific oligonucleotide PROEXI was used as the primer for both cDNA synthesis by Superscript II RNase H reverse transcriptase (Gibco BRL) and DNA sequencing. The reverse transcription reaction was carried out essentially as described previously (Thompson and Daniels, 1998). Dideoxy-terminated DNA sequencing reaction products and primer extension reaction products were separated by electrophoresis on a 6% denaturing (8.3 M urea) polyacrylamide gel and subjected to autoradiography.

Overexpression and purification of recombinant TFB2

E. coli DH5α cells containing the pTrcHis2TFB2 construct were grown in a 10 l fermentation batch culture at 37°C and induced with 1 mM IPTG at an OD600 of 0.6. The culture was allowed to grow for 2 h after the addition of IPTG. Cells pelleted from ≈5 l of the IPTG-induced culture were resuspended in 40 ml of Invitrogen denaturing binding buffer (8 M urea, 20 mM NaPO4, 500 mM NaCl, pH 7.8) and disrupted by three passages through a French pressure cell (15 000–20 000 psi). Unbroken cells and cellular debris were removed by centrifugation (9000 r.p.m. at 4°C for 30 min). Overexpression of recombinant TFB2 was detected in E. coli crude extracts by Western blot analysis using the anti-myc mouse monoclonal IgG antibody (Invitrogen), which was specific for the myc epitope in the His-tagged TFB2.

Purification of recombinant TFB2 was accomplished by chromatography using a 5 ml HiTrap nickel affinity column (Pharmacia Biotech) with the BioLogic chromatography system (Bio-Rad). His-tagged TFB2 was eluted with a decreasing linear pH gradient (pH 6–4) in a buffer containing 8 M urea, 20 mM NaPO4 and 500 mM NaCl at a flow rate of 1.0 ml min−1. Fractions (2 ml) were collected, and recombinant TFB2 was eluted at pH 4. Peak fractions contained ≈7–10 mg ml−1 recombinant TFB2. The purity of recombinant TFB2 was monitored by SDS–PAGE and Western blot analysis.

Preparation of anti-TFB2 antibodies and Western blot analysis

Antiserum against purified recombinant TFB2 was generated by immunizing New Zealand white rabbits. The column fraction used for polyclonal antibody production contained > 90% recombinant TFB2. The urea was removed from the fraction by eluting the recombinant protein from a HiTrap desalting column (Pharmacia Biotech) in native elution buffer (20 mM NaPO4, 500 mM NaCl, pH 4.0). The rabbits were injected with an emulsified mixture (0.5 ml) of His-tagged protein (1 mg) in native elution buffer and an equal volume of Freund's complete adjuvant (Sigma). A second injection (booster) of purified recombinant TFB2 (1 mg) in Freund's complete adjuvant was given to the rabbits 6 weeks after the initial immunization. Rabbits were terminally bled at week 10 to recover the antisera. An IgG fraction was prepared from the rabbit antisera using a 5 ml HiTrap Protein A column (Pharmacia Biotech) that was equilibrated with 20 mM NaPO4 (pH 7.0). IgG fractions were eluted with 0.1 M citric acid (pH 3.0) and then immediately neutralized with 1 M Tris-HCl (pH 7.0). Anti-TFB2 antibodies were stable at −70°C for an indefinite amount of time.

To detect the native TFB2 protein, H. volcanii cells were subjected to non-heat shock and heat shock conditions as described. Before electrophoretic separation, 30 μl of 2 × SDS reducing buffer (125 mM Tris-HCl, pH 6.8, 20% glycerol, 4% SDS, 10% 2-mercaptoethanol, 0.04% bromophenol blue) was added to each cell pellet (from 1.5 ml of culture), and the protein samples were heat denatured for 4–5 min in a boiling water bath. Protein samples (equal volumes per lane) were separated by electrophoresis on a 12% SDS–polyacrylamide gel according to the procedure described by Laemmli (1970). After electrophoresis, proteins were transferred to PVDF-Plus membrane (Micron Separations) in 1 × Tris–glycine buffer (pH 8.3) using the Genie electrophoretic blotter (Idea Scientific). Western analysis was performed using the Immun-Blot assay kit (Bio-Rad) according to the manufacturer's instructions. The primary antibody buffer contained a 1:50 000 dilution of an IgG fraction purified from rabbit anti-TFB2 serum. The secondary antibody was goat anti-rabbit IgG alkaline phosphatase conjugate (Sigma; 1 : 5000 dilution). Western signals were quantified using the Scion image program version 1.59 (Wayne Rasband, NIH).


  1. *Present address: FDA, Center for Biologics Evaluation and Research, Division of Bacterial Products, HFM-428, 1401 Rockville Pike, Rockville, MD 20892, USA


This work was supported by a grant from the Department of Energy, DE-FG02-91ER20041, to C.J.D. C.J.D is an associate of the Canadian Institute for Advanced Research.