Overlapping activator sequences determined for two oppositely oriented promoters in halophilic Archaea.

Transcription of the genomic region involved in gas vesicle formation in Halobacterium salinarum (p-vac) and Haloferax mediterranei (mc-vac) is driven by two divergent promoters, P(A) and P(D), separated by only 35 nt. Both promoters are activated by the transcription activator GvpE which in the case of P(mcA) requires a 20-nt sequence (UAS) consisting of two conserved 8-nt sequence portions located upstream of BRE. Here, we determined the two UAS elements in the promoter region of p-vac by scanning mutageneses using constructs containing P(pD) (without P(pA)) fused to the bgaH reporter gene encoding an enzyme with beta-galactosidase activity, or the dual reporter construct pApD with P(pD) fused to bgaH and P(pA) to an altered version of gvpA. The two UAS elements found exhibited a similar extension and distance to BRE as previously determined for the UAS in P(mcA). Their distal 8-nt portions almost completely overlapped in the centre of P(pD)-P(pA), and mutations in this region negatively affected the GvpE-mediated activation of both promoters. Any alteration of the distance between BRE and UAS resulted in the loss of the GvpE activation, as did a complete substitution of the proximal 8-nt portion, underlining that a close location of UAS and BRE was very important.

INTRODUCTION

The extremely halophilic Halobacterium salinarum and the moderately halophilic Haloferax mediterranei produce gas vesicles depending on environmental factors (1). We are interested in mechanisms of gene regulation in halophilic Archaea and use the gvp gene clusters encoding the proteins required for gas vesicle formation as model system. Gas vesicle formation involves the two gene clusters gvpACNO and gvpDEFGHIJKLM, which are oppositely oriented in the vac region (2). Two different vac regions of H. salinarum PHH1 are known, the plasmid pHH1-encoded p-vac and the chromosomal c-vac region, and in addition the chromosomal mc-vac region of H. mediterranei. The 14 gvp genes found in each of these vac regions encode slightly different variants of the two gas vesicle structural proteins GvpA and GvpC, and of additional Gvp proteins required for the assembly of this structure. GvpD and GvpE regulate the expression, with GvpE serving as transcriptional activator and GvpD being involved in the repression of gas vesicle formation (2–5). GvpD and GvpE are able to interact, and the interaction results in the breakdown of GvpE (4,6).

Transcription in Archaea is driven by a multi-subunit RNA polymerase (RNAP), and the initiation of basal transcription requires in addition to RNAP the TATA-box-binding protein TBP and the transcription factor TFB. A typical archaeal promoter consists of a TATA-box (= binding site of TBP) centred 24–28 nt upstream of the transcriptional start site, and the TFB-recognition element BRE located upstream and adjacent to the TATA-box (7,8). In all three vac-regions, the two promoters P and P are located upstream of the oppositely oriented gvpA and gvpD reading frames, and in the case of p-vac and mc-vac, the two BRE sequences are separated by 35 nt only (Figure 1A) (2). In contrast, the respective two promoters of c-vac are separated by 56 nt due to a 21-nt insertion upstream of P–BRE. Except for P, all of the P and P promoters are activated by GvpE (9–11). Besides studying the mRNAs by northern analysis in the original wild-type strains, these promoters have been analysed in Haloferax volcanii transformants using promoter fusions with bgaH as reporter gene encoding an enzyme with β-galactosidase activity (11–13). From the analyses of transformants containing the P- and P constructs it appears that cGvpE is the strongest activator followed by mcGvpE and pGvpE (11).

Figure 1.

Comparison of the three intergenic regions between P and P in different vac-regions, and alignment of the 8-nt UAS portions. (A) The vac regions are designated mc (mc-vac), p (p-vac) and c (c-vac). The TATA-box and BRE elements (in italics) are shaded in grey. The sequence determined as GvpE-responsive element (UAS) in P is in bold and underlined, and the two conserved 8-nt portions are labelled 1 and 2. Related sequences in p-vac and c-vac are also underlined. (B) Divergent double-stranded P promoter sequence with the two putative UAS underlined. Arrows point to the direction of transcription. (C) Alignment of the 8-nt portions of the putative UAS elements found in the P and P promoters. The conserved sequences are shaded in grey. The distal (1) and the proximal (2) 8-nt portions of the UAS are indicated.

Two scanning mutageneses have been performed to determine the sequence elements required for the basal transcription and for the GvpE-mediated activation in vivo. The first series of 4-nt mutations encompassed a 50-nt region upstream of the transcription start site of p-gvpA, and the results of this study clearly define the TATA-box and BRE in P, but also a region around position –10 that might serve as additional contact site of TFB (9,14). Mutations close to the 5′-end of this 50-nt sequence also affected the GvpE-mediated activation in P (most strongly by altering the sequence AACCA upstream of BRE), but the mutations between the TATA-box and the start site of transcription had no effect on the GvpE-mediated activation (9).

The P promoter of mc-vac was used for the second 4-nt scanning mutagenesis encompassing the sequences upstream of the P-BRE, and this study defined the 20-nt sequence TGAAACGG-n4-TGAACCAA adjacent to BRE as important for the GvpE-mediated activation (Figure 1A) (13). A significant drop in the GvpE-induced activation is observed when two 8-nt portions of this element (labelled 1 and 2 in Figure 1) are mutated in steps of 4-nt, but a residual activity of 6–20% is still observed (BgaH activity quantified by ONPG assay). In silico analysis suggests similar 20-nt elements in all three P promoters and implies that they might serve as upstream activator sequence (UAS) for the GvpE-mediated activation (Figure 1A). The AACCA sequence mentioned above is part of the proximal 8-nt portion of this element in P, and mutations up to CTGG had only a minor effect on the GvpE-induction in p-vac (9). Whether similar 20-nt UAS elements also occur in the GvpE-activated P promoters is not yet known and topic of the present report. In silico analysis indicated two similar elements in P and P (Figure 1C). If a similar 20-nt UAS is found for P, the two divergent UAS in the P promoter region must overlap considerably in the centre of the 35-nt of intervening sequence as shown for p-vac in Figure 1B.

In this report, we analysed the P promoter region by 5-nt scanning mutagenesis to determine the respective UAS element required for the GvpE-mediated activation of both promoters in H. volcanii transformants. Using pGvpE, cGvpE or mcGvpE to activate P again defined cGvpE as strongest activator protein. A 5-nt scanning mutagenesis performed with P clearly yielded an UAS element of similar size as determined in the previous P-study. The effect of mutations was also analysed in the dual reporter construct pApDUGA (containing 3 UGA stop codons in gvpA) by monitoring the activity of both promoters simultaneously. These experiments confirmed the UAS of P, but also the hypothesized GvpE–UAS of P and proved that both UAS elements almost entirely overlap with their distal 8-nt portions in the centre of the P region. In addition, variations in the UAS–BRE distance, a complete substitution of the entire UAS element, and complete substitutions of either the distal or the proximal region were constructed. The results underlined that the close location of UAS–BRE is required for the GvpE-induced promoter activity.

MATERIALS AND METHODS

Constructs used for transformation of H. volcanii WFD11

The H. volcanii medium and growth conditions, the various Eex constructs (gvpE fused to the ferredoxin promoter P in the expression vector pJAS35), and also the P–bgaH construct have been described (4,5,11). The construct pA–pD contains the leaderless p-gvpA reading frame fused to the P promoter region, and bgaH fused to the transcript start site of p-gvpD (without the p-gvpD mRNA leader region) in the Escherichia coli vector pBlusescriptSK+. The bgaH reading frame includes six codons for additional amino acids at the N-terminus to ensure the production of an active BgaH enzyme (11). The construction of pA–pD involved the amplification of the leaderless p-gvpA in two PCR reactions using p-vac as template and the primer pairs 3′pD2ΔL–bgaH plus pA-Dlo-18, or 5′p-gvpA–XbaI plus pA-Dlu-18 (Table 1). The resulting subfragments served as templates in the third PCR with primer pair 5′p-gvpA-XbaI plus 3′pD2ΔL-bgaH to yield the fusion product that was used to substitute the P promoter in P–bgaH–pBluescriptSK+ resulting in pA–pDBluescript. This construct served (i) as template for the amplification of the XbaI–BamHI fragment P using the primers 3′-PpD(2) and 5′-PpD–HindIII (Table 1), (ii) for the construction of the initial pApDpWL102 vector to investigate the activities of both oppositely oriented promoters that turned out to be not useful due to the GvpA production and (iii) as template to construct pApD to investigate the P and P promotor activities simultaneously in H. volcanii transformants.

Table 1.

Primers used to construct the P and pApD promoter mutants

Primer	Sequence in 5′–3′ directiona
3′pD2ΔL-bgaH	CCAACTGCCATGGAATCTGGTTGCGC CATCTAAGAAGCTTTACACTCTCCG
pA-Dlo-18	TAGTTAGAGATGATGGCGCAACC
pA-Dlu-18	GGTTGCGCCATCATCTCTAACTA
5′-PpD-HindIII	GCAAGCTTGTGTATGGTTTCACCAGTC GTTATGTC
5′p-gvpA-XbaI	ATAGTATTCTAGACAAGCGATTCACCTCCC
3′-PpD(2)	CGTAAGGGAGGTGAATCGCTTG
3′-PpD (3)	GTGAAACCATACACAAGCTTGCCG
3′-PpD (4)	TACACAAGCTTGCCGTAAGGGAGGTG
5′-pD-M1	TGGTTTCACCAGTCGTTATGTCTCCTGTA GATCATCGTACTTCTAAG
5′-pD-M2	TGGTTTCACCAGTCGTTATGTCTCCCTCG GTGAGTCGTACTTC
5′-PpD-M3II	TGGTTTCACCAGTCGTTATGTACAGCGTA ATGAGTCGTAC
5′-pD-M4	TGGTTTCACCAGTCGTTGGACATCCTGTA ATGAG
5′-pD-M5	TGGTTTCACCAGTATCCGTGTCTCCT GTAATG
5′-PpD-M6	TGGTTTCACAGTGAGTTATGTCTCCTG TAATG
5′-PpD-M7	TGGTTGTCTAAGTCGTTATGTCTCCTG TAATG
5′-PpD-M8	TCTACGCACCAGTCGTTATGTCTCCTG TAATG
5′-PpD-M9	GTGTGGCGTTTCACCAGTCGTTATG
3′-PpD-M9	ATCCTTATGTGATGCCCGAG
3′-PpA(WT)	TACACATCCTTATGTGATGCCC
gvpA+UGA 5′	CTTGTGAGTGATTGATGATCGTGTAC TAGAC
gvpA+UGA 3′	CGATCATCAATCACTCACAAGCCTGAA GAATCTG
5′-PpD+3	CAGTCGTTATGTCTCCTGTAATGAGTG TTCGTACTTCTAAG
5′-PpD+6	CAGTCGTTATGTCTCCTGTAATGAGGA CTGATCGTACTTCTAAG
5′-PpD-Δ3I	CAGTCGTTATGTCTGTAATGAGTCGTA CTTCTAAG
5′-PpD-Δ3II	CAGTCGTTATGTCTCCTGTAATTCGTAC TTCTAAG

aSubstitutions marked in bold, insertions underlined, restriction sites in italics.

The P fragment inserted into pWL102 served as wild type in the P mutation studies. The P–bgaH promoter mutants pD-M1 through pD-M8 each contained a 5-nt substitution upstream of P–BRE. All mutants were constructed by PCR using P in pBluescriptSK+ as template and primer 3′-PpD(4) plus the respective 5′-mutation primer harbouring the 5-nt substitutions (Table 1). The promoter mutants pD+3, pD+5, pDΔ3I and pDΔ3II with insertions or deletions upstream of BRE were constructed using P and primer 3′-PpD(3) plus the respective 5′ mutation primers (Table 1). All mutant-P fragments were transferred to pWL102 as XbaI–BamHI fragment.

Construct pApDUGA (3 UGA-stop codons introduced in p-gvpA starting at position 28) was constructed using primer pair gvpA-UGA-5′ and gvpA-UGA-3′ for the PCR (Table 1). Promoter mutants harbouring 5-nt substitutions (pApD–M1 through M8, identical to those described for pD above) and an additional mutant pApD–M9 were constructed by PCR using pApDUGA as template together with primer 3′-PpA(WT) or 3′PpD–M9 plus the respective 5′-mutation primer 5′-PpD–M1 through M9 (Table 1). The resulting XbaI–BamHI gvpA–P fragments were transferred to pWL102 and used to transform H. volcanii.

The P–bgaH mutant pA1+1 (substitution of the proximal UAS portion 2 by portion 1) and pA1+Δ2 (proximal UAS portion mutated) were produced by PCR using pA–bgaH in pBluescriptSK+ as template and primer 3′-pA1 plus the mutation primers 5′-pA1+1 or 5′-pA1+Δ2 (Table 2). Similarly, the mutants pA2+2 (substitution of UAS portion 1 by 2) and pAΔ1+2 (mutation of the UAS portion 1) were constructed using P as template and primer 5′-pA2 plus the mutation primers 5′-pA2+2 or 5′-pAΔ1+2 (Table 2). Construct pA+10 contains a 10-nt insertion between UAS and BRE and was constructed using P as template and primer pair pA+10nt-5′ and pA+10nt-3′ (Table 2).

Table 2.

Primers used to construct mutants with insertions, deletions or substitutions

Primer	Sequence in 5′–3′ directiona
mcA-XbaI	CCAAACTATCTAGATGTTTGAC
mcA-InsI-forward	CAACTGTACGAATGATTTTGTTAC
mcA-InsI-reverse	TCGTACAGTTGGTTCAGCAACC
mcA-InsII-forward	TTGCGACTGAACCAACACGAATG
mcA-InsII-reverse	TTCAGTCGCAACCGTTTCACCTC
mcA-Del-reverse	GTTCA/ACCGTTTCACCTCTCG
mcA-Del-forward	ACGGT/TGAACCAACACGAATGTGA
mcA-Sub-forward	GGGTCTCACCTTGCGTCTGATTGAC GAATGATTTTGTTAC
mcA-Sub-reverse	GTCAATCAGACGCAAGGTGAGACCCT
5′-Xba-PmcD	GCAAGTAACATCTAGATTCGTGTTGG TTCAGCAACCG
3′-Nco-pA(1-5)-PmcD	CTGCCATGGAATCTGGTTGCGCCATG GATGACGCAC
pA+10nt 5′	CCATTGTGACTGAGACACATCCTTAT GTGATGCC
pA+10nt 3′	GTGTCTCAGTCACAATGGTTTCACC AGTCGTTATG
5′-pA1+1	GGCATAACGATACACATCCTTATGT GATGC
5′-pA2+2	GATGAAACCACTGGTGAAACCATAC ACATC
5′-pA1+Δ2	GGGTCTGATTGACACATCCTTATGT GATGC
5′-pAΔ1+2	GAGTCTCACCCTGGTGAAACCATAC ACATC
3′-pA1	AGTCGTTATGTCTCCTGTAATGAGTC
3′-pA2	TCCTGTAATGAGTCGTACTTCTAAG TACG

aSubstitutions marked in bold, insertions underlined, slash marks position of deletion, restriction sites in italics.

The P–bgaH mutants were constructed in a similar way by PCR. Mutant mcA–InsI (TGT-insertion adjacent to BRE) was constructed using primer pair mcA–XbaI and InsIforward for the first PCR, and mcA–NcoI2a and InsI-reverse for the second PCR (Table 2). The third PCR was done with primers mcA–XbaI and mcA–NcoI2a. The mcA–InsII mutant was produced using primers InsII-forward and InsII-reverse together with the primers mcA–XbaI and mcA–NcoI2a (Table 2). Construct mcA–Del was produced using the primers Del-reverse and Del-forward, and construct mcA–Sub using Sub-reverse and Sub-forward. Construct P was obtained using the mc-vac region in pBluescriptSK+ as template and primer pair 5′-Xba–PmcD and 3′-Nco–pA(1–5)-PmcD (Table 2). The fragment contains the P promoter region plus the region encoding the mRNA leader of mc-gvpD. All these promoter fragments were fused as XbaI–NcoI fragments to the bgaH reading frame in pBluescript and transferred as XbaI–BamHI fragment to pWL102 that was used to transform H. volcanii. All promoter sequences and fusions were controlled by DNA sequencing. The accession numbers of the p-vac subfragmentsin Genbank are X64729 (p-gvpACNO), X55648 (p-gvpDEFGHIJKLM) and X64701 of the mc-vac region.

Selection of transformants and BgaH assays

Prior to the transformation of H. volcanii each construct was passaged through the E. coli dam strain GM 1674 to avoid a halobacterial restriction barrier. Transformation was done as described (15), and transformants were selected on agar plates containing 6 µg mevinolin ml−1 (pWL102) and/or 0.2 µg novobiocin ml−1 (pJAS35). Mevinolin is a derivative of lovastatin that was obtained as a generous gift of MSD Sharp & Dohme GmbH (Haar, Germany). The amount and presence of the desired constructs in each transformant were controlled by the analysis of the isolated plasmids. The BgaH activity was measured by ONPG assay in cell lysates of the respective bgaH transformants (16). The protein concentration was determined by the Bradford assay (17) using BSA as standard.

Isolation of RNA and transcript analysis

RNA was isolated from H. volcanii transformants according to (18) from cultures at OD600 = 1.8–2.0. Northern analyses involved electrophoresis of 2 or 0.02 µg RNA on denaturing, formaldehyde-containing, 1.2% (w/v) agarose gels, followed by capillar transfer to nylon membranes (17). A strand-specific RNA probe complementary to bgaH mRNA was synthesized using the 2.2-kb bgaH fragment inserted in pBluescriptSK+ as template for T7 polymerase. Símilarly, a strand-specific p-gvpA RNA probe was produced using p-gvpA inserted in pBluescriptSK+. The RNA was labelled using the DIG RNA Labeling Kit (Roche, Germany). Northern hybridization was carried out as described by Ausubel et al. (17), but the hybridization solution contained 10% (w/v) dextran sulphate (Sigma, Germany), 1% (w/v) SDS and 0.5% (w/v) skim milk powder.

Western analysis and DNA sequence analysis

Samples of H. volcanii and H. volcanii transformants were taken in early stationary growth (after 48–54 h of growth) and processed for protein isolation as described (6). Soluble proteins (20 µg) were separated on 12% tricine SDS–polyacrylamide gels (19). Western analyses were performed as described (17). Two different antisera raised against cGvpE (10) and mcGvpE (4) were used in a dilution of 1:1000 (mcGvpE) or 1:10000 (cGvpE antiserum) in blocking buffer, and the reacting antibodies were detected using the ECL detection system (GE Healthcare, Germany). DNA sequence determination of the fusion constructs was done according to the Sanger method using the Sequi-Therm EXCEL II Long-Read DNA Sequencing Kit-LC protocol (Biozym, Germany). The fragments were separated using a Licor DNA sequencer.

RESULTS

Stimulation of the P promoter activity by three different GvpE proteins

In silico analysis of the region upstream of P (and P) indicated a sequence element related to the UAS of P that could be a putative GvpE-responsive element (Figure 1C). In contrast, the P promoter of c-vac lacking the activation by GvpE did not show such a conserved sequence element (Figure 1C). An initial P construct (containing P plus the region encoding the mRNA leader of mc-gvpD fused to bgaH) did not yield detectable BgaH activities in H. volcanii transformants. However, the P construct containing the direct fusion of the bgaH reading frame to P (‘leaderless’) was useful. Transformants containing P together with pEex (p-gvpE reading frame expressed under ferredoxin promoter control) produced well detectable amounts of BgaH (35 mU/mg protein). To determine the P activity in response to the two other GvpE proteins, cGvpE and mcGvpE, P was tested in transformants carrying cEex or mcEex. The presence of each GvpE protein was confirmed by western analyses using an antiserum raised against cGvpE (10). The basal P promoter activity measured as BgaH activity in the lysate of the P transformants was with 1.2 mU/mg rather low, but much larger GvpE-induced activities were observed in the lysates of transformants harbouring in addition to P construct pEex (35 mU/mg), mcEex (410 mU/mg) or cEex (830 mU/mg). These results suggested that cGvpE was the strongest and pGvpE the weakest activator protein, similar to the results already described for the P and P promoters (9,11).

Substitution mutagenesis to identify the UAS of the P promoter

A 5-nt scanning mutagenesis was performed with P to identify the UAS element required for the GvpE-mediated activation of P. The mutations started upstream of the PBRE and extended to the BRE of the oppositely oriented P promoter (Figure 2C). The PTATA-box and further p-gvpA gene sequences were not present in this construct to exclude any activity derived from P. Eight mutants (pD-M1 through pD-M8) were obtained and investigated for their BgaH activities in H. volcanii transformants (Figure 2A and C). Transformants harbouring the wild-type (or mutant) P construct plus the native pJAS35 (i.e. without gvpE) were used to determine the respective basal P promoter activities, and P+pEex transformants were used to quantify the pGvpE-induced promoter activities. In each case, the presence of pGvpE was assessed by western analysis using the antiserum raised against cGvpE, and all of them contained pGvpE in comparable amounts (Figure 2B). Similar basal promoter activities (<3.8 mU/mg protein) were determined for the wild-type P and all P mutants, indicating that none of these mutations affected the basal promoter activity. With respect to the GvpE-mediated activation, similar BgaH activities as found for wild-type (>84 mU/mg protein) were observed with the pD-M6, M7 and M8, all of which harboured the mutation upstream of the putative P–UAS element and close to the opposite P–BRE (Figure 2A and C). Mutants with the alteration closer to P–BRE yielded reduced BgaH activities. The strongest reductions were found for pD-M5, M4 and M2, whereas only a minor reduction was observed for pD-M3 carrying the alterations in the 4-nt non-conserved sequence between the two hypothesized 8-nt UAS portions (Figure 2A and C). A minor reduction was also observed with pD-M1 (nucleotide substitutions adjacent to P–BRE). These results proved the existence of the expected 20-nt P–UAS consisting of two 8-nt portions. Alterations in the distal 8-nt portion of the P–UAS (pD-M5, M4) imposed a stronger effect on the P promoter activity than alterations in the proximal 8-nt portion (pD-M2, M1) (Figure 2A). In summary, the region affecting the GvpE-mediated activation of P corresponded very well to the putative UAS element deduced from the comparison to P.

Figure 2.

Analysis of mutants derived from P. (A) BgaH activities determined for the P–bgaH wild type and the mutants in PpEex transformants, including standard deviations (n = 4). (B) Western analysis to confirm the presence of similar amounts of pGvpE in each of these transformants. The label pE depicts GvpE. (C) Sequence of the intergenic region of p-vac with the two oppositely located BRE elements and P–TATA-box shaded in grey. The arrow points in the direction of P transcription. The UAS element determined for P is underlined. Dots in the lines underneath depict nucleotides identical to wild type, and the respective nucleotide substitutions are given. The basal and GvpE-induced activities quantified by BgaH assay are given on the right.

Substitution mutagenesis and simultaneous activation of P and P

A similar 5-nt scanning mutagenesis was performed using a dual reporter gene construct containing the native P–P promoter region. The initial pApD construct harboured this promoter region with P fused to bgaH and P to the leaderless p-gvpA reading frame. Transformants containing pApD by itself were obtained in large numbers, but only a few colonies of the pApD+pEex transformants were found. All of them produced a low level of BgaH (7 mU/mg; pApD, Figure 3A) and contained strongly reduced amounts of the plasmids (data not shown). Most likely these transformants were unable to tolerate the large amounts of the hydrophobic GvpA protein produced after the induction by pGvpE. To circumvent this problem, three UGA stop codons were introduced in all three reading frames in the 5′ region of p-gvpA in pApD to prevent the translation of the p-gvpA mRNA. The transfomants containing pApD+pEex were easily obtained and produced the expected enhanced amounts of BgaH (UGA, Figure 3A). This pApD construct was used as template for the 5-nt scanning mutagenesis encompassing the 35-nt of intervening sequences located between P–BRE and P–BRE. The nucleotides altered in constructs pApD-M1 through pApD-M8 were identical to the ones altered in the pD-M1 through pD-M8, and for completion mutant pApD-M9 with an alteration close to P–BRE was constructed (Figure 3B). All these pApD mutant constructs were investigated in H. volcanii transformants in the presence or absence of pEex. The P promoter activities were determined by ONPG assays, and northern analysis was performed to determine the amounts of the p-gvpA transcript indicative for the P activity. The presence of pGvpE was assessed by western analysis using the antiserum raised against cGvpE, and all of the pEex transformants contained pGvpE in comparable amounts (data not shown).

Figure 3.

Analysis of pApD and of mutants derived from the dual reporter construct pApD. (A) BgaH activities determined for the P–bgaH in pApD+pEex wild type and mutants, including the standard deviations (n = 4). pApD designates the original construct, and UGA the construct containing the 3 UGA stop codons in p-gvpA. The GvpE-induced activities quantified by BgaH assay are given on the right. (B) Sequence of the P promoter region (given in opposite orientation as in Figure 2) with the two BRE elements and TATA-boxes shaded in grey. The arrows point to the directions of transcription. The two UAS elements determined are underlined, and the UAS of P is shaded in yellow. The various nucleotide substitutions are indicated underneath, and dots depict identical sequences. The mutations were identical to the ones introduced into the promoter region of P (Figure 2). The mutations affecting P are shaded in grey, and mutations affecting P are shaded in yellow. The relative amounts of p-gvpA mRNA in the transformants + pEex (as determined by northern analysis, Figure 4) are given on the right. (+) = presence and (−) = absence of gvpA mRNA.

The basal P promoter activities of the pApD wild-type and mutant constructs were 5–20 mU/mg (data not shown). Only mutant pApD-M7 did not yield any basal P promoter activity, although the bgaH mRNA was observed in a comparable amount (data not shown). Since the pApD-M7+pEex transformant did not show any BgaH activity as well, we assumed that construct pApD-M7 was defect in the production of an active BgaH (Figure 3A). The other eight pApD-mutant constructs confirmed the results described above for the various P mutants: pApD-M3 (intervening 4-nt sequence altered) and pApD-M6, M8 and M9 (mutations upstream of the PUAS) yielded similar GvpE-induced P activities as found for the wild-type pApD (Figure 3A). Again, the strongest reductions in the BgaH activities were observed with mutants pApD-M5, M4 and M2 (<31 mU/mg). All these results confirmed the 20-nt UAS upstream of the P–BRE as important for the GvpE-induced P activity.

Total RNA was isolated from each transformant and northern analysis carried out to assess the activity of P (Figure 4). The stable 320-nt p-gvpA transcript is always produced in large amounts. Thus, only 0.02 µg of total RNA were applied in each case for better visualization of the differences. The analysis of the basal transcription (without GvpE-stimulation) yielded similar amounts of p-gvpA mRNA except for the pApD-M6 transformants that did not produce p-gvpA mRNA (data not shown). This mutant also lacked the GvpE-induction and was thus regarded as being defective (Figure 4). With respect to the GvpE-mediated induction of P in the other mutants, the transformants pApD-M1, M2 and M3 contained similar amounts of p-gvpA mRNA as found for wild-type pApD indicating that the region altered in these mutants was not important for the GvpE-mediated activation of P. These mutations were located upstream of the putative 20-nt P–UAS and close to P–BRE (Figures 3B and 4A). The strongest reduction in the amount of p-gvpA mRNA was seen with mutants pApD-M5 and M4, whereas less strongly reduced amounts of p-gvpA mRNA were detectable in mutants pApD-M8 and M9. Mutant pApD-M7 produced almost similar amounts as the wild-type pApD (Figure 4B). These results confirmed the P–UAS derived from the comparison with P (Figure 1C), and also demonstrated the simultaneous use of both 20-nt UAS elements in pApD (Figure 3B). The overlapping distal portions of both elements in the centre of this region were most sensitive with respect to mutations affecting the GvpE-mediated activation. The region altered in mutant M5 led to the lack of both P and P promoter activities and proved to be very important for the expression of both gvp genes.

Figure 4.

Northern analysis to determine the amount of p-gvpA mRNA in pApD+pEex transformants. (A) Nucleotide sequence of the P promoter region with the two BRE elements and TATA-boxes shaded in grey. The two UAS elements are underlined. Lines and numbers above the sequence depict the various 5-nt mutations in mutants pApD-M1 through M9. (B) Total RNA of 0.02 µg (bottom) derived from each transformant were separated on a formaldehyde containing, 1.2% agarose gels and hybridized a gvpA-specific probe. Amounts and quality of the RNA were checked by staining the 16S and 23S RNA (top; 2 µg RNA applied in this case).

Distance variants of the UAS and effect of a complete substitution

To determine whether the size of the 4-nt region between the two 8-nt portions, and also the close location of UAS and BRE were important for the GvpE-mediated activation, distance variants were constructed using P-, P- and P constructs (Figure 5). The UAS of P was altered by two different 3-nt deletions, one introduced in the 4-nt intervening sequence (mutant pDΔ3I) and the other one adjacent to BRE (pDΔ3II). Two additional mutants carried a 3- or 6-nt insertion next to BRE (pD+3, pD+6). Mutant pDΔ3I (deletion in the intervening sequence) showed a normal basal P promoter activity, but almost no induction by GvpE, suggesting that the size of the space between the two 8-nt portions of the UAS cannot be altered without loosing the GvpE-mediated activation (Figure 5A). The mutants pDΔ3II, pD+3 and pD+6 with alterations adjacent to BRE yielded very low basal P activities in transformants, implying that the nucleotides adjacent to BRE are still important for the initiation of basal transcription. None of these promoters was inducible by GvpE, demonstrating that the distance between UAS and BRE cannot be altered (Figure 5A). Even a 10-nt (= 1 helix turn) insertion between the UAS and BRE performed with P resulted in the lack of the GvpE-mediated activation (Figure 5C).

Figure 5.

Distance variants constructed for the P (A), P (B) or P (C) promoter region. The sequence of each promoter region is given on top, with TATA-box and BRE shaded in grey. The UAS elements are underlined. Dots in the lines underneath depict nucleotides identical to wild type. Deletions are marked by xxx, and substitutions or insertions are indicated. The basal and GvpE-induced promoter activities quantified by BgaH assay are given on the right.

Similar analyses were performed with the P construct containing the P promoter of H. mediterranei (Figure 5B). Two 3-nt insertions enlarged the distance between UAS and BRE (mcA–InsI), or the space between the two 8-nt portions (mcA–InsII). Both mutants showed a normal basal promoter activity in P transformants, but all of them lacked the induction mediated by mcGvpE (Figure 5B). A 3-nt deletion in the spacer between the two 8-nt UAS portions (mcA–Del) led to the loss of the GvpE-induced activity (Figure 5B), similar to the results obtained with P (Figure 5A). Altogether, these results demonstrated that the 4-nt distance between UAS-portions 1 and 2 and a close location of the 20-nt UAS to BRE are crucial for the GvpE-mediated activation. Mutant mcA-Sub contained a completely substituted PUAS element (Figure 5B). No effect was observed on the basal P activity, but the GvpE-mediated activation was completely abolished, demonstrating that the UAS-element is indeed essential for the GvpE-mediated activation.

Alteration of the proximal or distal 8-nt portion of the UAS

Four additional mutants were constructed with P to determine the effect of alterations in the two 8-nt portions 1 or 2 on the GvpE-induced expression (Figure 5C). Mutants pA1+Δ2 and pAΔ1+2 contained one of the 8-nt portions completely substituted, and pA1+1 and pA2+2 contained two identical 8-nt portions of the UAS (Figure 5C). The basal P promoter activities (determined in the presence of the ‘empty’ vector pJAS35 in the respective transformants) were in the range of the P transformant or at the detection limit (0.5 mU/mg), indicating that the basal transcription was not affected (Figure 5C). With respect to the GvpE-induction, pAΔ1+2 transformants yielded a BgaH activity similar to wild type, suggesting that the proximal 8-nt portion located close to BRE was sufficient for the activation in this single promoter construct. In contrast, the P promoter in pA1+Δ2 was not activated by GvpE underlining that UAS and BRE must be in close contact. The two constructs containing the identical 8-nt portions of the UAS (pA1+1 and pA2+2) yielded slightly reduced (pA2+2, 72% of the BgaH activity of wild type) or stronger reduced GvpE-induced activities (pA1+1, 40%) (Figure 5C). The latter result could be due to the fact that the UAS portion 1 is active in both orientations, and a location of this element next to BRE might disturb the direction of the GvpE-mediated activation of P. These results underlined that both UAS portions were not equal in function, and that one of these UAS portions was required in close contact to BRE to yield a GvpE-stimulated promoter activity.

DISCUSSION

In this report, we investigated the sequences required for the GvpE-mediated activation of the two oppositely oriented promoters, P and P, driving the expression of the two gvp gene clusters involved in the gas vesicle formation. A scanning mutagenesis previously performed with the related P of the mc-vac region determined a 20-nt sequence adjacent to P–BRE sequence as important for the GvpE-mediated activation (13). This 20-nt sequence appears to be conserved between the three different P promoters (Figure 1). A putative P-UAS should be located at a similar position upstream and adjacent to BRE, and in silico analysis yielded related, but slightly divergent sequences for the two GvpE-activated P promoters in mc-vac and p-vac.

The UAS element required for the GvpE-induction of P was determined by scanning mutagenesis with P using eight different 5-nt mutations, and all these variants were investigated in H. volcanii transformants in the presence (+pEex) or absence of pGvpE. None of these mutations abolished the basal transcription initiated at P, and the analysis of the transformants containing pEex in addition clearly defined the region required for the GvpE-mediated activation. As expected, this region was found close to the P–BRE and consisted of two 8-nt sequence portions spaced by four non-conserved nucleotides. Any mutations further upstream did not affect the GvpE-induction of P. Additional mutants constructed to determine the effect of distance variations showed that the spaces between UAS–BRE and also between the two 8-nt UAS portions were important. Similar results were also observed with distance variants of the P–UAS, where a 3-nt deletion, or several insertions between UAS and BRE resulted in the loss of the GvpE-mediated promoter activation. Thus, a very close location of UAS and BRE is required for the GvpE-mediated activation. In addition, mutants harbouring a complete substitution of the P–UAS lacked the activation by GvpE, indicating that the UAS sequence was essential for the GvpE-induced promoter activation. The consensus sequence YGAAAYGA could be derived from all 8-nt portions of the UAS elements found in the various gvp promoters activated by GvpE. This sequence is specific for the gvp gene clusters and does not occur throughout the genome sequence of H. salinarum NRC-1 (20). The close location of UAS and BRE suggests that GvpE could contact proteins of the basal transcription apparatus. It is possible that GvpE enhances the recruitment of TBP and thus activates the transcription initiation, similar to the transcriptional activator Ptr2 of Methanocaldococcus jannaschii (21,22).

Due to the 35-nt intervening sequence between the BRE sequences of P and P, the distal portions of both 20-nt UAS elements almost completely overlap in the centre of the region. The dual reporter construct pApD was used to analyse the activity of both promoters simultaneously. The P–UAS element determined by the scanning mutagenesis in pApD was identical to the one determined for P, and the northern analysis done with a p-gvpA-specific probe confirmed the P–UAS element for P that was predicted from the comparison with the P promoter region. These analyses also uncovered interesting features concerning the use and also the importance of the two 8-nt UAS portions. The mutations in both portions affected the GvpE-induced activation of the two promoters in slightly different ways. The activation of P by GvpE was more severely affected when the mutation occurred in the 3′-region of the central 8-nt portion, whereas in the case of P the alterations in the 5′-region of this UAS portion were more effective (Figure 6). The strongest effect on the GvpE-enhanced promoter activity was observed with mutants carrying 4- or 5-nt alterations in this distal 8-nt portion of the UAS element, that thus appeared to be the most important one in the dual promoter construct (Figure 6). A 4-nt sequence could be assigned in the central region where alterations most strongly affected the P-, P- or P activities; this region is located next to centre and closer to P (Figure 6A and B, marked in bold). The sequence constitutes the 3′- (P) or the 5′-portion (P) of the respective overlapping distal 8-nt UAS portions. This region might be required for the initial binding of GvpE, followed by the occupancy of the respective proximal 8-nt portions of the UAS. In contrast, alterations in the proximal 8-nt portion were less severe on the GvpE-induced promoter activity (Figure 6B). The results obtained with the dual promoter construct were very similar to the results obtained with the two single promoter constructs P (11) and P (this study), and demonstrated that both UAS elements including the overlapping portions are indeed functional in pApD. Despite the fact that mutation of UAS portion 1 imposed the strongest effect on the promoter activation by GvpE, this distal portion of the UAS could be deleted without the loss of the GvpE-mediated activation in the single promoter construct pAΔ1+2. The reason for this discrepancy is still unclear. In contrast, construct pA1+Δ2 (lacking the proximal UAS portion) did not respond to GvpE, underlining the importance of a close location of at least one UAS portion and BRE for the activation by GvpE.

Figure 6.

Comparison of the various scanning mutageneses performed with the mc-vac (A) and p-vac (B) promoter regions. The columns above the sequences indicate the specific BgaH activities determined for the various mutants (13, and this study). The respective mutations are marked as lines below each column and include the numbers of the various pApD mutants in (B). The 8-nt portions of each 20-nt UAS element are underlined, and the TATA-box and BRE are shaded in grey. Arrows point in the direction of transcription. The 4-nt next to the centre of the region that were the most important ones for the GvpE-mediated activation of both promoters are indicated in bold.

The promoter mutants carrying an UAS consisting of two identical portions yielded a slight reduction of the GvpE-induced activity (72%) in the case of pA2+2, whereas the pA-1+1 promoter version resulted in only 40% of the GvpE-induced activity determined for wild type. The latter result implied that the possible use of UAS-portion 1 in two orientations somehow disturbs the recruitment of the basal transcription apparatus by GvpE, if this sequence is placed adjacent to BRE.

In summary, our results confirmed the almost complete overlap of the two distal portions of both UAS regulatory elements in the P promoter region. This arrangement appears to be similar to overlapping operator sequences in the genome of bacteriophage lambda. To our knowledge, this is the first case of such an arrangement of regulatory elements in Archaea. Another promoter region, containing two oppositely oriented promoters that are co-ordinately regulated, has been characterized for the genes encoding the selenium-free [NiFe]-hydrogenases in Methanococcus voltae, but the distance between both TATA-boxes is with 290-nt significantly larger (23). The compact arrangement of the regulatory elements in P might determine the amount and the time point for the activation of the both divergent gvp promoters during the gas vesicle formation, and it might be interesting to investigate the activity of both P and P promoters with the natural UAS-overlap compared to a construct harbouring two separated UAS elements with respect to the time-resolved expression during growth.