A system where archaeal gene expression could be controlled by simple manipulation of growth conditions would enable the construction of conditional lethal mutants in essential genes, and permit the controlled overproduction of proteins in their native host. As tools for the genetic manipulation of Haloferax volcanii are well developed, we set out to identify promoters with a wide dynamic range of expression in this organism. Tryptophan is the most costly amino acid for the cell to make, so we reasoned that tryptophan-regulated promoters might be good candidates. Microarray analysis of H. volcanii gene expression in the presence and absence of tryptophan identified a tryptophanase gene (tna) that showed strong induction in the presence of tryptophan. qRT-PCR revealed a very fast response and an up to 100-fold induction after tryptophan addition. This result has been confirmed using three independent reporter genes (cct1, pyrE2 and bgaH). Vectors containing this promoter will be very useful for investigating gene function in H. volcanii and potentially in other halophilic archaea. To demonstrate this, we used the promoter to follow the consequences of depletion of the essential chaperonin protein CCT1, and to determine the ability of heterologous CCT proteins to function in H. volcanii.
Interest in the archaea has grown steadily as they were recognized as the third domain of life (Woese et al., 1990). Metagenomic studies (Xu, 2006) have shown archaea to be far more abundant in both extreme and non-extreme environments than previously thought, and they are now known to constitute a significant fraction of the total biomass (DeLong and Pace, 2001). They have features of both eukaryotic and of bacterial cells, and their study, as well as being of considerable interest in its own right, is giving key insights into how life evolved and into aspects of many complex processes which can be hard to study in eukaryotic cells. Research on the archaea has lagged behind that of the other domains of life because their importance was not recognized until very recently, and tools for their detailed analysis, especially at the molecular genetic level, are still relatively poorly developed when compared with those available for bacteria and eukaryotes. Genetic systems, which have been developed for some thermophilic, methanogenic and halophilic archaea, include efficient transformation, shuttle vectors, multiple resistance markers and reporter gene systems (Allers and Mevarech, 2005; Rother and Metcalf, 2005; Soppa, 2006). DNA microarray technology is also available for transcriptome analysis in several diverse archaeal species (e.g. Zaigler et al., 2003; Schut et al., 2003; Hovey et al., 2005; Snijders et al., 2006; Williams et al., 2007). Several promoters have been used for expression and transcriptional studies in archaea (e.g. Danner and Soppa, 1996; Gregor and Pfeifer, 2005). However, to date, there are no reported examples of promoters which can be turned on or off by simple changes in media components and where the basal activity is effectively zero.
Haloferax volcanii is particularly suited for genetic analysis in archaea as tools for its genetic manipulation are well developed. Transformation is straightforward and there are several plasmid vectors available. There are a number of antibiotic and auxotrophic selective markers (Allers et al., 2004), and a halophilic β-galactosidase has been developed as a reporter gene (Holmes and Dyall-Smith, 2000). Methods for constructing gene knockouts without leaving a selective marker in the final strain are established (Bitan-Banin et al., 2003). Sequencing of the genome is largely complete, with full annotation expected in the near future (A.L. Hartman et al., unpublished). A shotgun DNA microarray has been produced and validated, which is based on a PCR product library with a onefold coverage of the genome (Zaigler et al., 2003). However, there are currently no vectors that allow controlled gene expression in this organism. Here we report on the development of replicating and integrating vectors containing the promoter of the tryptophanase gene that allow controlled gene expression in H. volcanii. We demonstrate using two different genes that this system can be used to produce a conditional lethal mutant, and show that it can be used in studies of complementation or depletion of essential gene products.
Isolation and characterization of succinate-inducible promoter
Haloferax volcanii is capable of growth on defined media using a single carbon source (Mevarech and Werczberger, 1985; Kauri et al., 1990). We hypothesized that gene expression in H. volcanii is regulated in response to specific sugars, akin to the arabinose-inducible araBAD promoter of Escherichia coli (Guzman et al., 1995). To screen for sugar-inducible promoters we constructed the plasmid pTA425 (Fig. 1A), which features a promoter-less pyrE2 reporter gene for uracil biosynthesis (Bitan-Banin et al., 2003) that is insulated from readthrough transcription by a terminator sequence (previously identified upstream of the L11e ribosomal gene of H. volcanii) (Shimmin and Dennis, 1996). A partial AciI digest of genomic DNA from H. volcaniiΔpyrE2ΔtrpA strain H53 (Allers et al., 2004) was size-selected and fragments of ∼500 bp were inserted into pTA425 between the terminator and pyrE2 gene. The library was used to transform H53 and selection was carried out on minimal agar containing 0.5% of either glucose, glycerol, succinate or lactate. As minimal media lacks uracil, only cells harbouring a pTA425 derivative with a promoter upstream of pyrE2 will grow.
We obtained ∼105 transformants per microgram DNA, and these were subjected to a secondary screen on Hv-Ca (casamino acids) agar, to select for conditions under which promoters would be tightly repressed. One promising candidate emerged that grew on Hv-Minsuc (minimal media + 0.5% succinic acid) but not on any other carbon source tested. Furthermore, the candidate failed to grow on Hv-Ca, regardless of whether succinate was added. Thus, the putative promoter is induced specifically by succinate and repressed by (cas)amino acids. Sequence analysis of the plasmid pTA469 (Fig. 1A) revealed that the insert was a 454 bp AciI fragment (bp 816 925–817 378 of the main chromosome). It comprised 297 bp of the 5′ part of ORF010523 (annotated as branched-chain amino acid ABC transporter amino acid binding protein), and 157 bp of upstream sequence that should contain the succinate-inducible promoter.
To evaluate the promoter, we used the bgaHβ-galactosidase reporter gene construct pTA487 and the ΔpyrE2ΔtrpAΔbgaH strain H557 (Fig. 1A). When the AciI genomic DNA fragment from pTA469 was cloned in pTA489, very little blue staining was observed with Xgal (Fig. 1B). To confirm that this promoter was not particularly strong, the 157 bp region upstream of ORF010523 was analysed separately (Fig. 1A). The bgaH plasmid pTA491 demonstrated that this region is a weak promoter and the pyrE2 plasmid pTA480 indicated it is not repressible by amino acids (Fig. 1B). This suggests that control sequences responsible for repression by amino acids are located in the first 297 bp of ORF010523.
We therefore generated a fusion construct consisting of the 5′ part of ORF010523, downstream of a 43 bp strong (synthetic) promoter based on the H. volcanii consensus tRNA promoter sequence (C. Daniels, pers. comm.). The fusion construct was inserted upstream of pyrE2 to generate pTA569, or upstream of bgaH to generate pTA601 (Fig. 1A). Transformants containing pTA569 could grow on Hv-Minsuc but not Hv-Ca media, and cells containing pTA601 showed bgaH expression only on Hv-Minsuc (Fig. 1B). Control plasmids pTA543 and pTA599 contained the synthetic promoter only, and exhibited strong expression of pyrE2 or bgaH, respectively, on both Hv-Minsuc and Hv-Ca media. Thus, we were able to confirm that critical regulatory sequences reside in the 5′ part of ORF010523. However, the level of expression promoted by pTA601 was not particularly high, suggesting that the fusion construct was either suboptimally configured or that we had not determine the correct conditions for induction. A further disadvantage of this promoter is that induction by succinate requires the use of defined media, and thus far, attempts to develop a defined minimal broth for H. volcanii have met with little success. Therefore, use of a succinate-inducible promoter would be limited to solid media.
Cloning and analysis of promoters for the trp genes
Tryptophan is an expensive amino acid for cells to make, and trp genes in other organisms frequently show tight regulation for this reason. We therefore next examined expression from the promoters of the genes for tryptophan biosynthesis in H. volcanii, which in this organism are organized in two operons (Lam et al., 1990; 1992). A total of 350 bp of the trpCBA and 440 bp of the trpDFEG promoter sequences were cloned into plasmid pRV1 to give plasmids pRV1–ptrpCBA–bgaH and pRV1–ptrpDEFG–bgaH. DS70/pRV1–ptrpCBA–bgaH was found to have high β-galactosidase activity in rich and minimal media. When grown in a defined medium, activity from this promoter was reduced strongly in the presence of histidine and less so in the presence of tryptophan, but was most strongly reduced when both amino acids were present, to less than 10% of the value in the absence of amino acids (see Fig. 2). Surprisingly, we found no detectable β-galactosidase activity in the strain transformed with pRV1–ptrpDFEG–bgaH. We investigated the possibility that there might be elements required for expression upstream or downstream of the fragment that we had initially cloned. A 850 bp fragment including the first 99 bp of the coding sequence of was fused to bgaH, but no activity was found (data not shown). In conclusion, although the trpCBA promoter shows clear evidence of regulation by tryptophan and histidine, the lowest level of expression was still substantially above background and the dynamic range was limited, making it unsuitable for our purposes.
Identification of inducible promoters using a DNA microarray
The next attempts to identify a suitable promoter made use of a shotgun DNA microarray with a onefold coverage of the genome (Zaigler et al., 2003). An ideal promoter would be inactive, or have very poor activity, in the uninduced state and be highly induced by a substance or condition that would not change the expression of any other gene of the cell. The first rationale was to find a substance that would be toxic to the cell and that would very specifically induce the expression of an efflux pump. Doxorubicin was tested because it had been reported that H. volcanii possesses a transporter for this drug (Komatsubara et al., 1996). In addition, several heavy metals were tested, i.e. cobalt, copper, iron, mercury and nickel. In each case, a concentration series was applied to determine the minimal inhibitory concentration. Subsequently, non-inhibitory concentrations were added to exponentially growing cultures, and the transcriptomes in the presence and in the absence of the inhibitors were compared with the DNA microarray. However, this approach did not lead to the discovery of a suitable promoter (data not shown). Several additional approaches were unsatisfactory, e.g. adding glucose to a culture growing in complex medium or on amino acids.
In contrast, the addition of tryptophan turned out to be very successful. H. volcanii was grown in synthetic medium with glucose to a cell density of 1 × 108 cells ml−1. Then the culture was split into two independent cultures and 100 μg ml−1 tryptophan was added to one. After 4 h, the transcriptomes of both cultures were compared. Three biological replicates were performed, as described in Experimental procedures. Only two spots of the microarray showed a high tryptophan-induced differential expression. The underlying two genomic clones overlapped and both contained part of the tna gene encoding a tryptophanase. Thus, this was the only differentially regulated gene. Moreover, the signal intensity in the absence of tryptophan was very low, showing that the tna gene was hardly expressed in the absence of tryptophan. The difference in signal intensity of all other clones was not higher than about 2 and not lower than 0.5 (Table S1). Therefore, the microarray results led to the detection of a promoter that seemed to ideally fit the above-mentioned constraints.
Characterization of differential expression of the tna gene
The tryptophan-dependent expression of the tna gene was further characterized by quantitative reverse transcription real-time PCR (qRT-PCR). H. volcanii was grown to mid-exponential phase (1 × 108 cells ml−1) in three different media, i.e. complex medium and synthetic media with glucose or with casamino acids. Aliquots were removed before the addition of tryptophan (100 μg ml−1) and at various times thereafter, ranging from 5 min to 6 h. The tna transcript levels were determined by qRT-PCR. Three biological replicates were performed. The results are shown in Fig. 3. Before the addition of tryptophan, the tna gene is hardly expressed in glucose medium, and the expression is very low in casamino acids medium. The transcript level in complex medium with yeast extract and tryptone is considerably higher. In glucose-grown cells the addition of tryptophan led to a very fast induction of tna gene expression: more than 50-fold after 5 min and 100-fold after several hours. Taken together, the data show that the tna promoter is ideally suited to drive differential transcription in H. volcanii.
Cloning and analysis of the tna promoter
Comparison of the sequence of the tna gene with its homologue in H. marismortui identified the likely ATG start codon. We amplified and cloned 320 bp of upstream sequence of this gene (which included 100 bp encoding the C-terminus of the adjacent gene, aspartate kinase) into pRV1 to give the plasmid pRV1–ptna–bgaH. No β-galactosidase activity could be detected above background in strains containing this plasmid in minimal medium unless tryptophan was added. Maximal expression was found at 4 mM tryptophan (see Fig. 4). Shorter tna promoter sequences of 253 bp and 220 bp were also screened for activity and were found to give rise to equal expression of β-galactosidase in the presence of tryptophan to the above full-length sequence, with no detectable activity in its absence (data not shown). As the tna promoter showed much lower minimal activity than the trpCBA promoter, it was decided to utilize the 320 bp ptna sequence for subsequent experiments to determine whether this promoter allowed sufficiently tight regulation of downstream genes for the construction of conditional mutants.
Controlling pyrE2 expression with the tna promoter
To determine whether the tna promoter was sufficiently tightly repressed to construct a conditional mutant, we placed the pyrE2 gene under the control of ptna, either on a multicopy plasmid, or on the chromosome. When the ptna–pyrE2 fusion was transformed into H26 on the plasmid pRV–ptna–pyrE2, cells were able to grow in the absence of tryptophan. This is consistent with sufficient PyrE2 product being produced in the absence of tryptophan to enable cells to synthesize sufficient uracil for growth, showing that expression from the tna promoter on the multicopy plasmid is not fully repressed, perhaps because of titration of a repressor protein (the copy number of pRV1-derived plasmids is approximately 6 per genome equivalent). We also placed the ptna–pyrE2 fusion onto the chromosome, using a suicide plasmid with a mevinolin resistance marker and a non-functional cct1 gene to provide a region of significant homology (pOK-MC–ptna–pyrE2). The plasmid was successfully integrated into the chromosome at the cct1 locus and the location of the insertion was confirmed by PCR. The resulting strain, H26–ptna–pyrE2, was unable to grow on minimal media that was not supplemented with tryptophan (Fig. 5), confirming that the tna promoter was tightly off in the absence of tryptophan.
Controlling expression of chaperonin CCT1 using the tna promoter
The chaperonin complex of H. volcanii is encoded by three genes (cct1, cct2 and cct3), and one of either cct1 or cct2 must be present for viability (Kapatai et al., 2006). This makes them good candidates for further testing the utility of ptna for constructing conditional lethal mutants. We constructed a multicopy plasmid, pRV–ptna–cct1, by ligating the coding region of cct1 into pRV1–ptna–bgaH, replacing much of the bgaH gene with the cct1 gene. We also constructed a suicide plasmid, p131–ptna–cct1, by ligating the ptna–cct1 fusion into the suicide vector pTA131. Each plasmid was separately transformed into H26 Δcct1, Δcct3 (Kapatai et al., 2006). Transformants were grown in minimal media in the presence or absence of 5 mM tryptophan, and CCT expression was detected by Western blotting. Some residual expression of CCT1 was apparent in the absence of tryptophan in the strain carrying pRV–ptna–cct1 (the multicopy plasmid; see Fig. 6), consistent with the result described above with pRV–ptna–pyrE2. In the presence of tryptophan, expression of CCT1 in this strain was higher than wild-type. This overexpression of CCT1 caused CCT2 levels to fall, consistent with our previous finding that these two genes are coregulated (Kapatai et al., 2006). However, in cells where the ptna–cct1 construct was integrated on the chromosome, no CCT1 expression could be detected in the absence of tryptophan, while 5 mM tryptophan induced CCT1 expression to a level comparable to that of CCT2 under the control of its own promoter. This confirmed promoter probe data that the induced tna promoter has approximately the same strength as the constitutive levels of the cct1 and cct2 promoters (data not shown).
We then constructed two strains lacking all three endogenous cct genes, with cct1 under the control of the tryptophanase promoter, either on the chromosome or on a multicopy plasmid. First, H26 Δcct2, Δcct3 was transformed with p131–ptna–cct1, and a strain where this plasmid had integrated via a single cross-over at the cct1 locus on the chromosome was obtained. The location of the insertion and the structure of the integrated DNA was confirmed by PCR to be as expected (not shown). This strain was then transformed with pOK-NC99, a suicide plasmid carrying a truncated cct1 gene disrupted with a mevinolin resistance cassette, which we had previously used to knock out the cct1 gene by homologous recombination (Kapatai et al., 2006). This construct contains the upstream and downstream regions from the cct1 gene but little of the cct1 gene itself. Cells were selected on rich media plus 5 mM tryptophan and mevinolin. Resistant cells arising from this transformation will either have a copy of the disrupted cct1 gene alone, or else have both the disrupted version and the wild-type gene, depending on whether a single or a double-crossover event has occurred between the plasmid and the endogenous cct1 gene chromosome (recombination with the ptna–cct1 gene will be rarer, as this does not have the upstream flanking region of cct1). Individual colonies were picked and screened by PCR, and two out of 16 screened were found to have lost the wild-type cct1 gene (not shown). This strategy is shown diagrammatically in Fig. S1. To create a strain lacking all three chromosomal cct genes with ptna–cct1 on a multicopy plasmid, H26–Δcct1, Δcct3/pRV–ptna–cct1 was transformed with the suicide plasmid p131–Δcct2, which carries a constitutively expressed copy of the pyrE2 gene. Transformants were selected on minimal plates. Individual single colonies were then grown in liquid culture containing uracil, which removes the selective pressure for the presence of the pyrE2 gene and hence allows the growth of cells where recombination has removed pyrE2 from the chromosome. After several days' growth without selection, cultures were plated onto plates containing 5-fluoro-orotic acid/uracil and tryptophan to select for colonies that had lost pyrE2 from the chromosome (Bitan-Banin et al., 2003). Individual colonies were picked and screened by PCR, and in this case two out of 15 screened were found to have lost the wild-type cct2 gene. The loss of cct2 expression was confirmed by Western blotting (data not shown).
Replicates of both strains were grown in media supplemented with tryptophan and a series of 10-fold dilutions were spotted onto plates with and without 5 mM tryptophan (Fig. 7). Both grew equally well in the presence of tryptophan. In the absence of tryptophan, H26 Δcct1, Δcct2, Δcct3 ptna–cct1 showed very little growth. H26 Δcct1, Δcct2, Δcct3/pRV–ptna–cct1 showed some growth up to 10−4 dilution, but was significantly less viable than the H26 Δcct2, Δcct3 control strain. These results are consistent with the results of the Western blot shown in Fig. 6, and show that while the low level of CCT1 produced in the absence of tryptophan when pRV1–ptna–cct1 is present (presumably because of the high copy number of the plasmid relative to the genome) is sufficient to allow limited growth, essentially no growth can occur when ptna–cct1 is on the chromosome, because levels of CCT1 are negligible. We believe that the limited growth in Fig. 7 (rows D to G on the plate lacking tryptophan) seen at high cellular concentrations of H26 Δcct1, Δcct2, Δcct3 ptna–cct1 is due to a few cell doublings that occur while CCT1 levels are falling because of depletion, as is seen in liquid medium (see below).
Application of the tna promoter
To demonstrate the applicability of this system, and to advance our understanding of chaperonin function in the archaea, we undertook two further studies.
First, we investigated the abilities of homologous cct genes from a related halophile, H. marismortui, to complement for loss of all of the endogenous cct genes. H. marimortui has four cct genes: cctA2, cctB, cctA3 and cctA1. Comparison of these seven proteins using the phylogenetics package phylip (Felsenstein, 1989) showed that cctA2 is most closely related to cct1, cctB to cct2 and cctA3 to cct3 (see Fig. S2). The proteins CCTA2 and CCTA3 could be readily expressed in several H. volcanii strains using the H. volcanii cct1 promoter in plasmid pRV1 (see Fig. 8). Surprisingly, CCTA3, although one of the smallest of the CCT proteins, ran the slowest in SDS-PAGE gels, presumably because of its very high level of charge. Plasmids for the constitutive expression of H. marimortui CCTB and CCTA1 could not be transformed in to any strain, suggesting that expression of these proteins in H. volcanii may be lethal; expression of these proteins also could not be confirmed in H. marimortui (Fig. 8, lane 2). Strain H26 Δcct1, Δcct2, Δcct3 ptna–cct1 was successfully transformed with pRV–pcct1–cctA2 and pRV–pcct1–cctA3. Transformants expressing CCTA2 grew readily in the absence of tryptophan demonstrating that this chaperonin could complement for the loss of the endogenous cct genes. Conversely, although CCTA3 was expressed at a high level (Fig. 8), it was unable to complement, and tryptophan was required for growth.
Second, we used the strain H26 Δcct1, Δcct2, Δcct3 ptna–cct1 to investigate the effects of depleting CCT protein from growing cells. We confirmed that CCT1 levels were depleted by transferring cells grown with tryptophan to a media without tryptophan (Fig. 9). After 48 h cells were pelleted and resuspended in fresh media to an optical density of ∼0.1, to prevent them from reaching saturation phase before the CCT levels in the cell had been completely depleted. We compared the growth profiles for this strain in the presence and absence of tryptophan. The results (Fig. 10) shows that growth is severely restricted in the absence of tryptophan, as would be expected when an essential gene product becomes depleted from cells. After approximately 60 h of growth, the optical density of the cells became impossible to measure accurately as the cultures became very amorphous with substantial clumping which suggests that cell damage or lysis was occurring on a large scale, confirming that CCT1 is essential for growth in liquid medium. We used SDS-PAGE to compare the soluble and insoluble protein profiles of the strain grown with and without tryptophan during the course of CCT1 depletion, but did not observe any reproducible differences between them (data not shown). This latter observation implies that CCT, assuming it is working as a chaperone, may fold a relatively small subset of the proteins in the cell.
Conditional mutants are powerful tools in genetic analysis, as they permit hypotheses to be generated about the role of specific genes. The advent of whole genome sequencing means that the rate at which putative genes are identified significantly outpaces the rate at which functions can be assigned to them, and a means to rapidly generate conditional mutants is needed to close this conceptual gap. Conditional mutants of specific genes can most efficiently be generated by cloning the genes under control of a promoter which can be turned off, and then deleting the original copy of the gene. A promoter which can be turned off without the need for major changes to the growth conditions of the cell is clearly required for this approach to be successful. In this paper we have identified such a promoter which can be used in this way for the halophilic archaeon H. volcanii, an organism for which several other genetic tools are already well developed.
In our initial attempt to isolate a regulatable promoter, we used a genetic screen for sugar-inducible promoter sequences that would lead to transcription of a pyrE2 reporter gene. We were able to identify a promoter that is induced by growth on succinic acid as a carbon source, and is repressed by amino acids. However, further experiments using a bgaHβ-galactosidase reporter showed that this promoter was rather weak. Attempts to increase expression by fusing the regulatory sequences we had identified, which are necessary for induction by succinate and repression by amino acids, to a strong synthetic promoter met with only limited success. We therefore focused our efforts on identifying promoters regulated by exogenous tryptophan.
The trpCBA promoter showed initial promise as a regulatable promoter. Surprisingly repression by histidine was greater than that by tryptophan at the same concentration. It was most effectively repressed by a combination of the two amino acids. In the archaeon Methanobacterium thermoautotrophicum several genes for tryptophan biosynthesis have been demonstrated to be repressed by both tryptophan and amino acids of other families (Gast et al., 1994), suggesting the possibility of specific repression by tryptophan, underpinned by a global regulation of amino acid biosynthesis. We were unable to find conditions where this promoter was not still residually active, and it had a rather limited dynamic range, so it was not studied further. The reason for the lack of activity of the trpDEFG promoter is not known, and this was also not studied further.
The tna gene was identified by comparison of cells grown in the absence versus the presence of tryptophan using a shotgun DNA microarray. The tna gene was the only gene that was differentially induced and thus the specificity of tryptophan as an inducer is extremely high. For comparison, nearly 300 genes were found to be differentially expressed in exponentially growing H. volcanii cultures with glucose versus casamino acids as energy and carbon sources (Zaigler et al., 2003). The absence of a more general effect on gene expression makes the tna promoter an important tool for studying the in vivo effects following the introduction of a single RNA or single protein at a specific time into H. volcanii. The induction is very fast, with more than 50-fold increase in tna transcript level is reached within 5 min after tryptophan addition. Most importantly, the tna gene is hardly expressed in glucose medium, which indicated that the promoter would be ideally suited to be applied as a molecular genetic tool to drive the regulated expression of other genes.
The activity of the region upstream of the tna gene, which was presumed to contain the tna promoter, was studied using three different reporter genes. The pyrE2 gene is useful as it provides a simple genetic screen for promoter activity. The bgaH gene enabled quantitative measurements of promoter activity to be made, and compared with qRT-PCR data. (The reporter enzyme β-galactosidase has previously been shown to quantitatively reflect promoter activity in H. volcanii, although it was found that the enzyme is considerably more stable than its mRNA; increasing β-galactosidase activity was in accordance with an increase of mRNA during exponential phase, but it stayed at a high level at stationary phase (Gregor and Pfeifer, 2001). Therefore, all assays were done with cultures before they reached stationary phase). Finally, the use of the cct1 gene was a useful model to show how a conditional mutant of a gene known to be essential (in the absence of the cct2 homologue) could be generated using the tna promoter.
Studies using all three of these reporter genes confirmed that the tna promoter is strongly induced by tryptophan, and that expression in the absence of tryptophan is extremely low, particularly when the promoter is on the chromosome. The weak expression seen when the tna promoter is on a plasmid may be due to titration out of a hypothetical repressor protein, or may be due to readthrough from an upstream promoter, which could occur if the termination sequence of the aspartate kinase gene is not fully terminating transcription. Both these possibilities are currently under investigation. Many archaea, and halophiles in particular, show marked polyploidy: Halobacterium may have up to 25 copies of the chromosome, and H. volcanii is known to have in the region of 18 copies during active growth (Breuert et al., 2006). The plasmid copy number for pRV1 and its derivatives is about 6–8 per genome (M. Dyall-Smith, pers. comm.) and we reproducibly found expression from a promoter on pRV1 to be higher than when on a chromosome, though not dramatically so.
Conveniently, ptna shows a similar level of expression in the presence of tryptophan to that of the uninduced cct1 and cct2 promoters, as judged by β-galactosidase reporter assay. This allowed us to make a triple chromosomal knockout of the three cct genes, leaving only a ptna controlled copy. The fact that this strain required tryptophan to grow confirmed our earlier finding that at least one copy of a cct gene is essential for growth (Kapatai et al., 2006). The conditional lethal mutant strain H26 Δcct1, Δcct2, Δcct3 ptna–cct1 gave us the ability to deplete CCT and follow the effects of this on growth. This is an approach which has also been used to analyse the function of the type I chaperonin gene groEL in E. coli using the tightly repressible pBAD promoter (Ivic et al., 1997). As expected, growth slowed considerably as CCT1 was depleted from cells in which cct2 and cct3 had been deleted, but no large-scale aggregation of proteins was seen, as might have been predicted if the CCT chaperonin is responsible for the folding of many of the intracellular proteins of H. volcanii. We are currently using 2-D gels to see whether significant differences in the profiles of the soluble and insoluble proteins can be seen as CCT levels fall in growing cells.
The ability to construct conditional mutants also enables complementation studies with homologues to the gene of interest. We demonstrated this by looking at the ability of chaperonins from H. marismortui to function as the sole chaperonin in H. volcanii. Intriguingly, only one of the four CCT proteins from H. marismortui (the one with the highest homology to the CCT1 protein) could replace CCT1. Two of the four H. marismortui cct genes could not be expressed in H. volcanii, and we also were unable to obtain clear evidence that they are expressed in their host organism under normal growth conditions. This result fits well with our previous demonstration that the three cct genes of H. volcanii are not completely functionally equivalent in complementation tests (Kapatai et al., 2006) and that cct1 is the best gene in complementation assays. Uncovering the reason why different H. volcanii and H. marismortui cct genes vary in function (as detected by complementation assays) is likely to give us insights into the structure/function relationship in archaeal CCT proteins that may, in turn, help us to understand the more complicated eukaryotic CCT complex. Comparison of sequence alignments of the seven proteins shows that the conservation is lower in the apical domain part of the protein, as defined in the structure of the thermosome (Ditzel et al., 1998), which is also the region implicated in substrate binding, which raises the possibility that different archaeal CCT proteins may have a degree of specificity for different groups of substrates.
The tna promoter may also prove useful for studies on other halophiles such as Halobacterium and Haloarcula species, and it will be of interest to compare the activity of the promoter from the H. volcanii gene with those of the endogenous tna genes of these organisms. However, the use of this promoter in archaea other than the halophiles is unlikely, as the tna gene is not found in most archaea. Of the 29 available genomes on Genbank only the three halophiles above and the crenarcheote Aeropyrum pernix have the tryptophanase gene. The tna promoter should also prove very useful for overexpressing proteins that cause cell damage such as lysis or inclusion bodies, by growing cells to high density in media without tryptophan and then using induction with tryptophan late in exponential phase and to allow higher production of the protein.
Although we describe this promoter as being ‘repressible’, we do not have any information as to the actual mechanism that enables tryptophan to alter expression from the promoter, and it is conceivable that the promoter is off unless induced, rather than on but tightly repressed. In this study we used a 320 bp region upstream from the ATG of the tna gene for most of the experiments, as this clearly contained all the regulatory sequences necessary. Inspection of this sequence has not revealed any obvious motifs where a regulator protein might bind, and experiments are currently underway in the laboratory to identify the minimal sequence required for the regulation of the promoter, and to use this to try to identify any potential regulatory proteins that may bind to it.
The controlled expression system for H. volcanii described in this report provides a new means for the study of the molecular genetics and proteins in this archaeon, and further adds to the tools already available. It opens the possibility of constructing a conditional mutant in any gene in the organism. H. volcanii is now one of the most exploited archaea in laboratories as it needs no special equipment for incubation, grows rapidly in both solid and liquid media, is genetically tractable and can be studied with ever increasing range of techniques. H. volcanii is thus becoming a useful model organism to study many systems that would be difficult to investigate in other archaea or eukaryotes.
Strains and culture conditions
All H. volcanii strains are listed in Table 1. H. volcanii and Haloarcula marismortui were routinely grown aerobically in modified growth medium containing 20% or 23% total salt with 2% w/v agar for solid media. Where a fully defined medium was required, a minimal medium based on that of Dyall-Smith (2006) was used, but substituting alanine (25 mM) for ammonium chloride as the nitrogen source and using 0.4% succinate as the carbon source. For H26 and its derivatives the minimal media was supplemented with 50 μg ml−1 uracil. A total of 50 μg ml−1 5-fluoro-orotic acid was used for selection of ΔpyrE2 cells, 0.5 μg ml−1 novobiocin or 4 μg ml−1 mevinolin was added for antibiotic selection of transformants. H. volcanii strains H53 ΔpyrE2ΔtrpA and H557 ΔpyrE2ΔtrpAΔbgaH (derived from H53 by using pTA617 to delete the bgaH gene) were routinely grown at 45°C in rich medium (Hv-YPC), while casamino acids (Hv-Ca) and minimal succinic acid medium (Hv-minsuc) were used in selective plates. Media for these strains were prepared as described before (Allers et al., 2004), except that Hv-Minsuc used 0.5% succinic acid instead of mixed carbon source.
Escherichia coli strains were grown in Luria–Bertani media supplemented with ampicillin (100 μg ml−1) or kanamycin (50 μg ml−1), as appropriate. DH5α (F–, ø80dLacZΔM15, Δ(lacZYA-argF)U169 deoR recA1 endA1 hsdR17(rK–, mK+) phoA supE44 λ– thi-1 gyrA96 relA1) and XL1-blue MRF′ (mcrA183 mrrCB-hsdSMR-mrr173 endA1 supE44 thi-1 recA1 gyrA96 relA1 lac[F′proAB lacIqZM15 Tn10]) were used for routine plasmid construction and maintenance. Unmethylated DNA was produced by passage through either JM110 (rps, thr, leu, thi-1, lacY, galK, galT, ara, tonA, tsx, dam, dcm, supE44, Δ(lac-proAB)) or GM121 (F–dam-3 dcm-6 ara-14 fhuA31 galK2 galT22 hdsR3 lacY1 leu-6 thi-1 thr-1 tsx-78) before transformation into H. volcanii strains.
All plasmids are described in Table 2. DNA sequences were amplified by PCR using Phusion polymerase (Finnzyme, Finland) in either HF or GC buffer as supplied by the manufacturer; primer sequences are shown in Table 3. The sequences of all fragments which had been produced by PCR amplification were confirmed by DNA sequencing.
pBluescript with HindIII–AgeI fragment containing bgaHa allele, inserted at HindIII and XmaI sites. The bgaHa allele differs minimally from bgaH and will be described in a subsequent publication (S. Delmas et al. unpublished)
Integrative vector based on pBluescript II, with pyrE2 marker
pTA351 with KpnI fragment containing L11e terminator, upstream of KpnI–EcoRI fragment containing promoter-less pyrE2 gene from pGB70, inserted at KpnI and KpnI/EcoRI sites respectively
pTA425 with the 454 bp AciI genomic fragment comprising 5′ part of ORF010523 and 157 bp of upstream sequence, inserted at ClaI site between L11e terminator and pyrE2 gene
pTA425 with 157 bp region upstream of ORF010523, inserted at ClaI site between L11e terminator and pyrE2 gene
pTA351 with KpnI fragment containing L11e terminator, upstream of ClaI/SpeI fragment containing promoter-less bgaHa gene from pTA128, inserted at KpnI and ClaI/SpeI sites respectively. Two BspHI sites in pTA351 were filled-in and the direction of AmpR reversed
pTA487 with the 454 bp AciI promoter fragment from pTA469, inserted at BspHI site between L11e terminator and bgaHa gene
pTA487 with 157 bp region upstream of ORF010523, inserted at BspHI site between L11e terminator and bgaHa gene
pTA131 with HindIII-BsrGI bgaHa fragment from pTA128 subcloned in at Asp718/NotI sites (all sites filled-in)
pTA425 with 43 bp synthetic promoter based consensus tRNA promoter, inserted at ClaI site between L11e terminator and pyrE2 gene
pTA543 with PCR fragment comprising 5′ part of ORF010523 and pyrE2 from pTA469, inserted at ClaI/EcoRI sites to replace pyrE2
pTA543 with ClaI/SpeI bgaHa fragment from pTA487, inserted at ClaI/SpeI sites to replace pyrE2
pTA569 with PCR fragment comprising 5′ part of ORF010523 and bgaHa from pTA489, inserted at ClaI/SpeI sites to replace 5′ part of ORF010523 and pyrE2
pTA506 with deletion of EcoRV–BamHI bgaHa coding sequence, to generate ΔbgaHa construct (BamHI site filled-in)
Table 3. Primers used in this study.
Primer names (pairs as used)
Restriction sites used for cloning are underlined.
For cloning of promoters, primers were designed to introduce an NdeI site (CATATG) at-3′ end, corresponding to the translation start site of the cognate gene, and a BglII site (AGATCT) at the 5′ end. The promoter regions of the trpCBA and trpDFEG operons of H. volcanii (Lam et al., 1990; 1992) were cloned as a 360 bp fragment and a 420 bp fragment respectively. The tryptophanase (tna) gene was located in the H. volcanii genome (A.L. Hartman et al. unpublished), the start codon identified by comparison with the H. marismortui gene sequence, and 320 bp of the upstream sequence was amplified. Shorter tna promoter sequences of 253 bp and 220 bp were similarly cloned. The promoter region for cct1 was identified from the available sequences and the 720 bp upstream of the ATG start codon was cloned. Promoter sequences were inserted into plasmid pRV1 (a kind gift of Dr Mike Dyall-Smith, University of Melbourne) cut with BglII and NdeI, immediately upstream of the coding sequence of the β-galactodidase (bgaH) gene derived from Haloferax alicantei (Holmes and Dyall-Smith, 2000). These plasmids were referred to as pRV1–pxxx–bgaH, where pxxx is the name of the individual promoter.
The pyrE2 coding region from plasmid pTA131 was cloned into the NdeI site of pRV1–ptna–bgaH, with the 3′ end incorporating a AseI site so that no restriction site would be present at the 3′ end following ligation to give plasmid pRV–ptna–pyrE2. The activity of pyrE2 was confirmed by growing transformed H26 strains on minimal media supplemented with tryptophan. The ptna–pyrE2 fusion was also ligated into plasmid pOK-MC, a suicide plasmid which was derived from pOK-mev (Kapatai et al., 2006) by ligating it with a HindIII fragment carrying a promoter-less copy of the cct1 gene to provide a substantial region of homology for recombination with the chromosome.
The coding region of cct1 was amplified with an NdeI site at the start codon and a EcoRI site several bases downstream of the putative RNA polymerase termination sequence. The amplified region was inserted into plasmid pRV1–ptna–bgaH, removing the 5′ third of the bgaH gene and putting the chaperonin gene under the control of the tna promoter. This plasmid was referred to as pRV–ptna–cct1. For insertion into the chromosome, the ptna–cct1 fusion was ligated into the suicide vector pTA131 (Allers et al., 2004) carrying the pyrE2 gene (for selection on minimal media). Similarly, the coding regions of cctA1, cctA2, cctA3 and cctB from H. marismortui were found in the annotated genome (available at http://www.ncbi.nlm.nih.gov) and the start codon identified by comparison with the equivalent H. volcanii gene. Coding regions were amplified and inserted between the NdeI and EcoR1 sites of plasmid pRV1–pcct1–bgaH, to allow constitutive expression of the heterologous CCT proteins in several H. volcanii strains.
Molecular genetic methods
The transformation protocol for H. volcanii was as described previously (Kapatai et al., 2006). Isolation of plasmid DNA from H. volcanii was carried out as described previously (Norais et al., 2007). E. coli transformations, restriction digests, DNA ligation and plasmid purification were by standard methods (Sambrook and Russell, 2001). Sequencing was by the BigDye Version 3 reaction and using a ABI 3700 Capillary Sequencer, according to manufacturer's instructions (Applied Biosystems, CA, USA).
Genomic DNA libraries
Thirteen micrograms of H. volcanii H53 DNA was digested with AciI for 30 min at the recommended temperature, using ∼0.2 units μg−1 DNA in New England Biolabs buffer 1. Fragments of ∼500 bp were excised from agarose gels and ligated with pTA425, which had previously been cut with ClaI and the DNA ends dephosphorylated. The plasmid library was used to transform GM121 and DNA was prepared directly from colonies to avoid differential amplification. DNA was used to transform H. volcanii H53 ΔpyrE2ΔtrpA and transformants were selected on Hv-Min plates containing 0.5% of either glucose, glycerol, succinic acid or lactic acid.
DNA microarray analysis
Haloferax volcanii WR340 (Bitan-Banin et al., 2003; a kind gift of Moshe Mevarech Tel Aviv University, Israel) was grown in synthetic medium with 5 mM glucose as described (Wanner and Soppa, 2002). At a cell density of 1 × 108 cells ml−1, the culture was split into a control culture and a culture that was challenged with tryptophan (100 μg ml−1). After 4 h of further aerobic growth, the cell densities had reached 1.7 × 108 cells ml−1. Aliquots of 109 cells were taken from both cultures and RNA was isolated using the RNeasy Mini Kit (Quiagen, Hilden, Germany). cDNA synthesis and labelling, competitive hybridization to the shotgun H. volcanii microarray, and scanning were performed as described previously (Zaigler et al., 2003). All spots were inspected individually and in some cases the diameters generated by the scanning software (GENEPIXPRO 3.0) were corrected and spots containing artifacts were removed. The data were normalized under the assumption that the majority of genes are not regulated by tryptophan, and thus the averages of all Cy3 and Cy5 signals were set to equity. Three independent biological replicates were performed and resulted in 1659, 1640 and 1171 spots with good signals. The results were analysed with Microsoft Excel. As the aim of this study is to identify a highly inducible promoter, all signals with a strength of less than 1000 fluorescence units were removed, resulting in the data sets with 673, 623 and 536 spots. Spots that were detected only in a single experiment were removed, leaving 575 spots. The results are summarized in Table S1 (clone identifier, signal intensities of three experiments, average values and signal ratios). Two shotgun clones, 437F10 and 457A1, showed high induction in the presence of tryptophan. Their sequences were analysed at the European Bioinformatics Institute (http://www.ebi.ac.uk) and it became clear that they overlap and both contain parts of a gene encoding a tryptophanase.
Quantitative real-time RT-PCR
Haloferax volcanii WR340 was grown to a cell density of about 1 × 108 cells ml−1 in complex medium or in synthetic medium with glucose (0.25% w/v) or with casamino acid (1% w/v) essentially as described (Wanner and Soppa, 2002), Molybdate was omitted, the trace element solution was replaced by the SL-6 solution (Dyall-Smith, 2006), and 8 mM FeSO4 and 0.01% (w/v) yeast extract were added. All cultivations were performed aerobically at 42°C and 250 r.p.m. l-tryptophan was added to a concentration of 100 μg ml−1 at the beginning of the time course. Two millilitres of samples was removed immediately before as well as at various time points after addition. The cells were harvested by centrifugation (1 min, 15 000 g) and RNA was isolated as described by Gauthier et al. (1997). Residual DNA was removed using RQ1 RNase-free DNase (Promega, Mannheim, Germany) following the manufacturer's instructions, a subsequent purification was carried out with the RNeasy Mini kit (Qiagen, Hilden, Germany) according to the RNA cleanup protocol or with the NucleoSpin RNA Cleanup kit (Macherey-Nagel, Dueren, Germany). Reverse transcription of 0.5 μg of RNA was carried out with 100 U M-MLV reverse transcriptase RNase H minus (Promega, Mannheim, Germany) and 0.15 μg of random hexamer primers (Sigma, Steinheim, Germany) in 10 μl 1× reaction buffer (Promega) in the presence of 0.2 mM dATP and dTTP as well as 0.3 mM dCTP and dGTP. The reaction mix without enzyme was denatured for 10 min at 65°C, cooled on ice for 2 min and after reverse transcriptase addition, the reaction was performed for 1 h at 42°C. Then, additional 50 U reverse transcriptase were added and the incubation was continued for 1 h. Finally, the reaction was heat inactivated for 5 min at 80°C.
The quantitative real-time PCR was performed in a RotorGene 3000 (Corbett Research, Melbourne, Australia) in a volume of 25 μl with DyNAmo SYBR Green qPCR Mastermix (Finnzymes Oy, Espoo, Finland), 0.4 μM each forward and reverse primer (biomers.net, Ulm, Germany) and cDNA in an appropriate dilution (usually 1.0 μl) as template. Controls without template and with RNA instead of template were included. The primers Trpase_F2_RT and Trpase_R2 were used for tna amplification. RibL10-H. v.-B-3′ and RibL10-H. v.-B-5′ were used for the ribosomal protein L10 gene rpl10 chosen as internal standard not influenced by tryptophan (sequences see Table 3). The PCR consisted of 10 min initial denaturation at 94°C, and at least 50 cycles of 30 s 94°C, 45 s 68°C, 40 s 72°C. A subsequent melting point analysis from 60°C to 99°C in 0.5°C steps was performed. Data analysis was conducted with the RotorGene 6.0 software (Corbett Research) calculating CT values from the intersection of a threshold line with the early exponential interval of the fluorescence curve. Relative tryptophanase induction levels were calculated according to the 2−ΔΔCT method (Livak and Schmittgen, 2001) first normalizing the tna CT to the corresponding rpl10 CT and then setting the amount of tna RNA on glucose minimal medium without tryptophan to 1. The measurement was performed with RNA from three independent cultivations.
β-Galactosidase assays were based on the method of Holmes et al. (1997). Assay buffer contained 2.5 M NaCl, 10 μM MnCl2, 0.1% w/v β-mercaptoethanol, 50 mM TrisHCl pH 7.2. One hundred microlitres of culture (grown for 48 h after dilution in the presence of different tryptophan concentrations) was added to 850 μl of the above media and cells were lysed by sonication. The reaction was started by the addition of 50 μl of 8 mg ml−1 2-nitrophenol-β-D-galactopyranoside (2-NPG) and the change in absorbance measured at 405 nm continuously at room temperature. Where shown, units are expressed as μmol 2-NPG hydrolysed per minute divided by the OD650 of the culture, using a molar extinction coefficient 3300 M−1 cm−1 for 2-nitrophenol. To measure β-galactosidase activity qualitatively, plates were sprayed with a fine mist of Xgal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) solution (BlueTech, Mirador) and transformants scored the next day.
Detection of CCT proteins
Cells expressing CCT proteins were harvested and lysed, centrifuged to remove cell debris and the protein precipitated with 10% w/v trichloroacetic acid. Sample preparation, SDS-PAGE, Western blotting and detection were as previously described (Kapatai et al., 2006).
We would like to thank Professor Mike Dyall-Smith and Brendan Russ, University of Melbourne, for plasmid pRV1 and the L11e terminator sequence, and helpful information. We thank Chuck Daniels for the synthetic promoter sequence, and Andrew Fenton for initial work on tryptophan-repressible promoters. We are grateful to the Leverhulme Trust, the Biotechnology and Biological Sciences Research Council, the Medical Research Council and the Royal Society for funding. The project was also supported by grants of the German Science Foundation to J.S. (DFG So264/9 and/10).